-------
MOBILIZATION OF BODY RESERVES
Although data on changes in body weight and length and in body
compartment sizes (liver, gut) have been recorded for eels undergoing induced
gonadal maturation, not much is known about changes in biochemical composition
and the enzymes involved in mobilization of resources. The migrating silver
eels certainly differed in biochemical composition from the yellow eel,
especially in their higher fat content, and these differences have been
attributed to natural gonadal development (Lewander et al. 1976). Starved
eels lost between 0.1 and 0.2% body weight per day (Inui and Oshima 1966; Chan
et al. 1969, 1980) with about 30% of total energy derived from protein
catabolism and 70% from oxidation of fat (Fisher 1977, Chan and Woo 1978).
The bulk of the body reserves in support of metabolism came from mobilization
of protein and fat from white muscles (Inui and Oshima 1966, Aster and Moon
1981). Starved eels have been shown to survive at least 3 years, by which
time about 70% of the initial weight had been lost (Boetius and Boetius 1967).
With increased mobiliaation of tissue protein and lipid, plasma
aminoacid and free fatty acid levels increased as expected. Loss of protein
from white muscle during starvation has been shown to originate mainly from
the insoluble fiber component (Moon 1983a,b). The soluble protein fraction in
liver and muscle that contained the metabolic enzymes was unaltered.
Gluconeogenic enzymes (glucose-6-phosphatase, fructose-1.6-bis-phosphatase and
PEP carboxykinase) and aminoacid metabolizing enzymes (glutamate-pyruvate
transaminase, glutamate-oxaloacetate transaminase and leucine transaminase) in
liver tend to be conserved while muscle FDPase and GPT increased (Moon 1983a).
Based on the NADPH-generating capacity, the liver appeared to be the
primary site of lipogenesis in the eel (Aster and Moon 1981). In short term
starvation (1-3 weeks), there was marked reduction in hepatic fatty acid
synthetase and acetyl-CoA carboxylase, but the NADPH-generating enzymes were
maintained (Abraham et al. 1982). With prolonged starvation (26 weeks),
however,hepatic glucose-6-phosphate dehydrogenase was reduced while most other
enzymes were conserved (Aster and Moon 1981). The rate of conversion of 14C-
acetate to fatty acid progressively decreaed with starvation up to 39 weeks
but significant levels of activity still remained up to at least 95 weeks of
starvation (Abraham et al. 1982). Muscle lipid content actually increased
steadily during prolonged starvation (Moon 1983a).
Blood glucose and carbohydrate supply was maintained (Dave et al. 1975,
Renaud and Moon 1980), but enzymes of the glycolytic pathway and the
tricarboxylic- acid cycle in the muscles decreased slightly (Bostrom and
Johansson 1972.). Thus in a broad sense, starvation in eels evoked changes
that tended to preserve the metabolic enzyme machinery, whereas in other fish
species, drastic decreases in liver and muscle enzyme activities would result.
The present data showed that the unique enzymatic adaptation to starvation in
the eel is put to important use during gonadal development, which takes place
without dietary nutrient supply.
The nitrogen requirements of the developing testis differ substantially
from that of the ovary, the former being rich in nucleic acids and the latter
in yolk proteins. Differences in plasma protein levels in male and female
121
-------
eels during gonadal development reflected the hepatic synthesis and release of
vitellogenin in the female and its absence in the male (Kara et al. 1980).
Vitellogenin is a phophoprotein that specifically binds Ca. In male eels, the
plasma-free aminoacid levels increased drastically but did not rise in the
female eel. This probably represented aminoacid in transit, much of which was
removed by the liver for vitellogenin synthesis in the developing female eel.
Stimulation of glutamate dehydrogenase and AMP aminohydroiase in white muscle
during gonadal development was in line with the elevated breakdown of muscle
tissue. Suppression of glutamate dehydrogenase in the liver was probably due
to the need to conserve nitrogen. In female eels treated with oestradiol-17£,
mobilization of protein without concomittant gonadal development was
accompanied by marked elevation of ammonia excretion and ammonia loss/oxygen
consumption ratio in eels maintained in freshwater (Chan and Cheung 1982).
For purine catabolism, the enzymes of the uricolytic pathway was elevated only
in male eels suggesting marked increase in purine turnover. This probably was
related to the marked elevation of nucleic acid synthesis in the developing
testis.
Substantial increases in the lipogenic enzymes (NADPH-generating
systems) were recorded in both sexes and this was probably due to the large
demand on lipids for gonadal development. Increased hepatic lipogenesis
occurred early in gonadal development in both sexes, but the large demand of
lip id in the developing oocytes during exogenous vitellogenesis later depleted
this store in the female eel. The ultimate source of the lipid was certainly
the white muscle compartment, which showed a marked decline in fat-store in
female eels. Unlike other teleost fish, the eel has little abdominal fat and
red muscle fat content was unaffected during gonadal development.
The present study showed that the key enzymes in intermediary metabolism
were unimpaired and in some cases elevated during induced gonadal development.
In contrast, the marked decline in haematocrit values placed the animal in
jeopardy and probably constituted the most important factor limiting survival.
REFERENCES
Abraham, S., H.J.M. Hansen and F.N. Hansen. 1984. The effect of prolonged
fasting on total lipid synthesis and enzyme activities in the liver of
the European eel (Anguilla anguLlla). Comp. Biochem. Physiol. 79B-285-
289.
Aida, K., R.S. Nishioka, and H.A. Bern. 1980. Degranulation of the Stannius
corpuscles of coho salmon (Oncorhynchus kisutch) in response to ionic
changes in vitro. Gen. comp. Endocr. 42:305-313.
Aster, P.L. and T.W. Moon. 1981. Influence of fasting and diet on
lipogenic enzymes in the American eel, Anguilla rostrata LeSuer J
Nutr. 111:346-354.
Bergmeyer, H.U.
Weinheim.
1974. Methods of Enzymatic Analysis. Verlag Chemie,
122
-------
Boetius, J., I. Boetius, A.M. Henuningsen, A.F. Bruun, and E. Moller-
Christensen. 1962. Studies of ovarial growth induced by hormone
injections in the European and American eel (Anguilla anguilla L and
Anguilla rostrata LeSueur). Medd. Danm. Fisheri-og Havunders. N.S.
3:183-198.
Boetius, I. and J. Boetius. 1967. Studies in the European eel, Anguilla
anguilla (L>). Experimental induction of the male sexual cycle, and
its relation to temperature and other factors. Medd. Denm. Fiskeri-og
Havunders. N.S. 4:339-405.
Bostrom, S.L. and R. G. Johansson. 1972. Enzyme activity patterns in
white and red muscle of the eel (Anguilla anguilla') at different
developmental stages. Comp. Biochem. Physiol. 426:533-542.
Butler, D.G. 1969. Corpuscles of Stannius and renal physiology in the
eel (Anguilla rostrata). J. Fish. Res. Bd Can. 26:639-654.
Carpenter, S.J. and H.L. Heyl. 1974. Fine structure of the corpuscles of
Stannius of Atlantic salmon during the freshwater spawning migration.
Gen. comp. Endocr. 23:212-223.
Chan, D.K.O., I. Chester Jones, I.W. Henderson, and J.C. Rankin. 1967.
Studies on the experimental alteration of water and electrolyte
composition in the eel (Anguilla anguills L.). J. Endocr. 37:297-317.
Chan, D.K.O. 1968. Endocrine regulation of calcium and inorganic phosphate
balance in the freshwater-adapted teleost, Anguilla anguilla and A.
japonica. Int. Congr. Ser. Excerta Med. Found. 184:709-716.
Chan, D.K.O. 1972. Hormonal regulation of calcium balance in teleost fish.
Gen. comp. Endocr. suppl. 3:411-420.
Chan, D.K.O., J.C. Rankin, and I. Chester-Jones. 1969. Influences of the
adrenal cortex and the corpuscles of Stannius on osmoregulation in the
European eel (Anguilla anguilla L.) adapted to freshwater. Gen. comp.
Endocr. suppl. 2:342-354.
Chan, D.K.O. and N.Y.S. Woo. 1978. Effect of hypophysectomy on the chemical
composition and intermediary metabolism of the Japanese eel, Anguilla
japonica. Gen. comp. Endocr. 35:169-178.
Chan, D.K.O. and A. Cheung. 1982. Hormonal and metabolic changes in the
male eel during artificial induction of sexual maturation. Gen. comp.
Endocr. 46:386-387.
Copp, D.H., S.W.Y. Ma, M.C. Qui, M. Sim, and K. Szeto. 1985. Role of the
pituitary and the corpuscles of Stannius in calcium regulation in
teleosts. In: Current Trends in Endocrinology. B. Lofts and W.N.
Holmes (eds.). Hone Kong University Press, pp. 819-821.
123
-------
Dave, G., M. Johansson-Sjobeck, A. Larsson, K. Lewander, and U. Lidman.
1975. Metabolic and haematological effects of starvation in the
European eel, Anguilla anguLlla L. I. Carbohydrate, lipid, protein and
inorganic ion metabolism. Comp. Biochem. Physiol. 52A:423-430.
Dufour, S., D.L. Belle, and Y.A. Fontaine. 1984. Stimulation de la
liberation d'hormone gonadotrope et du developpement des gonades sous
1'effet conjugue du pimozide et d'un agoniste de la LHRH chez 1'anguille
femelle argentee pretraitee a 1'oestradiol. C.r. Acad. Sci. Paris
299:321-234.
Dufour, S., E. Lopez, F. Le Menn, N. Le Belle, S. Baloche, and Y.A. Fontaine.
1987. Stimulation of gonadotropin release and of ovarian development
by the administration of a gonadoliberin agonist and of dopamine
antagonist, in female silver eel pretreated with estradiol. Gen. comp.
Endocr.
Etienne, N. 1959. Influence de la maturation sexuelle provoquee sur
1'activite thyroidienne dell' anguille eruopeenne male, Anguilla
anguillla. C.r. Soc. Biol. 153:41-44.
Fenwick, J.C. 1974a. The corpuscles of Stannius and calcium regulation in
the North American eel (Anguilla rostrata LeSueur). Gen. comp. Endocr.
23:127-135.
Fenwick, J.C. 1974b. The comparative in vitro release of calcium from
dried and defatted eel and rat bone fragments, and its possible
significance to calcium homeostasis in teleosts. Can. J. Zool. 52'755-
764.
Fenwick, J.C. and Y.P. So. 1974. A perfusion study of the effect "of
Stanniectomy on the net influx of calcium-45 across the isolated eel
gill. Can. J. Zool. 188:125-131.
Fisher, Z. 1977. Nitrogen changes in starved eel (Anguilla anguLlla) P01.
Arch. Hydrobiol. 24:289-298.
Fontaine, M. 1964. Corpuscules de Stannius et regulation ionique
(Ca,K et Na) du millieu interieur de L'anguille (Anguilla anguLlla L.).
C.r. Acad. Sci. Paris. 2590:875-878.
Fontaine, M. 1967. Intervention des corpuscules de Stannius dans 1'equilibre
phosphocalcique du milieu interieur d'un poisson teleosteen, 1'anguille.
C.r. Acad. Sci. 2640:736-737.
Fontaine, M., N. Delerue, E. Martelly, J. Marchelidon, and C. Milet. 1972.
Role des corpuscles de Stannius dans les echanges de calcium d'un
poisson teleosteen 1'anguille (Anguilla anguLlla L.) avec le milieu
ambiant. C.r. Acad. Sci. Paris. 2750:1523-1528.
Fujita, H. and Y. Honma. 1967. On the fine structure of corpuscles of
Stannius of the eel, Anguilla japonica. Z. Zellforsch. mikrosc. Anat
77:175-187.
124
-------
Hanke, W., K. Bergerhoff, and O.K.0. Chan. 1969. Histological observations
on pituitary ACTH cells, adrenal cortex, and the corpuscles of Stannius
of the European eel. Gen. comp. Endocr. 9:64-75.
Kara, A., Y. Yamauchi, and H. Hirai. 1980. Studies on female-specific serum
protein (vitellogenin) and egg yolk protein in Japanese eel (Anguilla
japonica). Comp. Biochem. Physiol. 656:315-320.
Heyl, H.L. 1970. Changes in the corpuscles of Stannius during the spawning
journey of Atlantic salmon (Salmo salar). Gen. comp. Endocr. 14:43-52.
Hirschman, A. and A.E. Sobel. 1965. Composition of the mineral deposited
during in vitro calcification in relation to the fluid phase. Arch.
Biochem. Biophys. 110:237-243.
Inui, Y. and Y. Oshima. 1966. Effect of starvation on metabolism and
chemical composition of eels. Bull. Jap. Soc. Scient. Fish. 32-491-
501.
Kaneko, T., S. Harvey, L.¥. Lkine, and P.KXT. Pang. 1988. Localization
of calcium regulatory hormones in fish. Abs. 1st Int. Symp. .Fish
Endocr. Edmonton, p. 19.
Krishnamurthy, V.G. and H.A. Bern. 1969. Correlative histological study
of the corpuscles of Stannius and the juxtaglomerular cells of teleost
fishes. Gen. comp. Endocr. 13:313-335.
Lewander, K.G. Dave, M.L. Johansson, A. Larsson, and U. Lidman. 1974.
Metabolic and haematoloical studies on the yellow and silver phases of
the European eel, Anguilla anguilla L. I. Carbohydrate, lipid, protein
and inorganic ion metabolism. Comp. Biochem. Physiol. 476:571-582.
Lopoez, E. 1970A. Etude histophysiologique des corpuscules de Stannius
de Salmo salar L. au cours des diverses etapes de son cycle vital.' Gen.
comp. Endocr. 12:339-349.
Lopez, E. 1970. L'os cellulaire d'un poisson teleosteen Anguilla anguilla L.
II Action de 1'ablation des corpuscules de Stannius. Z. Zellforsch.
190:566-572.
Lopez, E., H.S. Lee, and C. Baud. 1970. Etude histophysique de 1'os d'un
teleosteen Anguilla anguilla L. au cours d'une hypercalcemie provoquee
par la maturation, experimentale. C.r. Acad. Sci. Paris. 270:2015-2017.
Lwowski, E.S. 1978. The corpuscles of Stannius and calcified tissues in the
eel, Anguilla rostrata (L.). Comp. Biochem. Physiol. 59A:183-187.
Moon, T.W. 1983a. Metabolic reserves and enzyme activities with food
deprivation in immature American eels, Anguilla rostrata (LeSueur).
Can. J. Zool. 61:802-811.
125
-------
Moon, T.W. 1983b. Changes in tissue ion contents and ultrastrueture of
food-deprived immature American eels, Anguilla rostrata (LeSueur)
J. Zool. 61:812-821.
Can.
Nadkarni, V.B. and A. Gorbman. 1966. Structure of the corpuscles of Stannius
in normal and radiothyroideetomized chinook fingerlings and spawning
Pacific salmon. Acta Zool. Stockh. 47:61-66.
Ochiai, A., S. Umeda, and M. Ogawa. 1974. On the acceleration of maturity
of the catadromous female eel by hormone injection, and changes in the
liver and blood character. Bull. Jap. Soc. Scient. Fish. 40:43-50.
Olivereau, M. 1961. Corpuscules de Stannius et reproduction chez 1'anguille
male. C.r. Acad. Sci. Paris. 253:541-543.
Olivereau, M., P. Dubourg, P. Chambolle, and J. Olivereau. 1986. Effect of
estradiol and mammalian LHRH on the ultrastrueture of the pars distalis
of the eel. Cell Tiss. Res. 246:425-437.
Pang, K.T., R.K. Pang, and W.H. Sawyer. 1974. Environmental calcium and the
sensitivity of killifish (Fundulus heteroclitus) in bioassays for the
hypocalcemic response to Stannius corpuscles from killifish and cod
(Gadus morhua). Endocrinology. 94:548-555.
Renaud, J.M. and T.W. Moon. 1980. Starvation and the metabolism of
hepatocytes isolated from the American eel, Anguilla rostrata. LeSueur.
J. comp. Physiol. 135:127-137.
Subhedar, H.A., A. Cheung, and D.K.O. Chang. 1981. Histophysiology of the
corpuscles of Stannius of the eel Anguilla japonica with special
reference to changes during sexual maturation. Abst. 9th Int. Symp.
Endocr., p 101.
Wang, Y., C. Xhao, Z. Shi, Y. Tan, K. Zhang, Y. Li, Y. Yang, and Y. Hong.
1980. Studies on the artificial inducement of reproduction in common
eel. J. Fish. China. 4:147-156 (in Chinese).
Yamamoto, K. and Y. Nagahama. 1973. Cytological changes in the
adenohypophisys of freshwater cultivated male Japanese eels, Anguilla
japonica, induced to maturation by transfer to sea water and synahorin
injection. Bull. Jap. Soc. Scient. Fish. 39:585-594.
Yamamoto, K. and K. Yamauchi. 1974a. Sexual maturation of Japanese eel and
reproduction of eel larvae in the aquarium. Nature Lond. 251:220-222.
Yamamoto, K. , T. Morioka, 0. Hiroi, and M. Omori. 1974b). Artificial
maturation of female Japanese eels by the injection of salmonid
pituitary. Bull. Jap. Soc. Scient. Fish. 40:1-7.
126
-------
Yamauchi, K. andK. Yamamoto. 1982. Experiments o,n artificial.maturation and
fertilization of the Japanese eel (Anguilla. japonLca). In: Proc. Int.
Symp. Reprod. Physio. Fish, Wageningen, Netherlands, G.J.J. Richter &
H.J.Th. Gooe (eds.). pp. 185-187. ,
Zhang, M.L. 1988. Studies on Hormonal Inducing Maturatin in the Eel
(Anguilla. japonLca Temminck and Schlegel). M.Sc. Thesis, Zhongshan
University, Guangzhou, China.
127
-------
UTILIZATION OF INORGANIC NITROGEN BY
TILAPIA NILOTICA
by
Wang Yi Qiang1 and Shu Xue Bao1
INTRODUCTION
The development of animal culture throughout the world requires large
sources of protein. Whether the protein source can be supplied sufficiently
will affect this development directly. Thus to exploit protein sources is an
outstanding problem.
The use of NPN by organisms to synthesize protein was examined in
ruminants in the late 19th Century. Urea and other nitrogen compounds can be
utilized by ruminants. These compounds may be decomposed into ammonia by
microorganisms in the rumen of the animal; ammonia, then, becomes an effective
protein source through synthesis by microorganisms. So it is considered that
urea can be used to replace a part of protein in the diet.
Urea-contained nitrogen is about 2.8 times that of protein-contained
nitrogen, and the price of urea is low. So several hundred thousand tons of
urea are used each year in animal feed in the USA. Molimoto (1971) reported
that urea might be used to replace about 20 to 30 percent of protein,
resulting in greater economic efficiency.
Urea was used by microorganisms to synthesize protein, which was
supplied to the animal. The urea was not utilized by the animal directly.
Animals have different dietary natures requiring different levels of
dietary protein. Some animals that require little protein grow rapidly.
Obviously, a part of protein is synthesized in the animal body. It may be
that the non-protein substance in the animal body, under the condition of
existence of nitrogen, transforms into protein. Wang Yi-Qiang et al. (1987),
using an immunocytochemical method, confirmed that the grass carp
(Ctenopharvngodon idellus) and the black carp (Mylopharvngodon piceus) have
different levels of insulin in their islets of Langerhans. Evidently, their
levels of metabolism of carbohydrates are different. Grass carp use more
carbohydrates to transform into protein and, therefore, they should be
provided with more nitrogen to increase growth. So, whether the animals can
or cannot use NPN directly to synthesize into protein is a problem for
discussion.
Shanghai Fisheries University, Shanghai, PRC
128
-------
Fishes have different dietary natures and require different protein
levels in diet also. Satia (1974) proved that rainbow trout require 40 to
46%. Ogino and Saito (1970) and Takeuchi et al. (1979) discovered common carp
require 31 to 38%. Lin Ding et al. (1980) and Yang Guo-Hua et al. (1985)
reported grass carp need 20 to 25%. Garling and Wilson (1976) proved catfish
need 22 to 40%. Jauncey (1982) confirmed that tilapia require 35%. There are
contrary opinions among scientists concerning whether fish can use NPN
directly. Nilni and Keveren used a diet containing urea to feed carp and
catfish. They proved that fish cannot utilize inorganic N. Fladofska and
Valet used urea to replace the protein source to feed the carp and mullet.
They proved that urea can replace a part of protein in the diet. Chen et al.
(1986) made a special diet containing urea to feed carp. They got a high
output.
All the different scientific opinions were obtained from studies of
growth by adding NPN to the diet. The investigators examined whether the
fishes can utilize NPN directly to replace the protein in diet. None of them
investigated the physiological activity of the animal.
This paper studied the amount of NPN directly through the digestive
tract or peritoneal cavity of the fish. The utilization ratio of absorbed NPN
in fish was measured by means of stable nuclide 15N-labeled NPN. The results
will provide scientific basis for fisheries production.
MATERIALS AND METHODS
Tilapia nilotica (spawned juvenile) were taken from the fish farm of our
university and were transferred to a circulating, filtrating aquarium for
provisional rearing. The size of aquarium is 200 x 80 x 80 cm3. The length
of fishes was from 5 to 13 cm; the weight was from 3 to 33 grams.
The pellet diet was made containing the high protein level and normal
protein level by the materials according to the formula as Table 1.
TABLE 1. THE DIET FORMULA (percent)
Rape
Fish Rice Plant seed Vitamin'
meal Bran chaff Flour oil Mineral cake compound Protein
High protein
level diet
Normal protein
level diet
53
15
8
30
5
20
4 •
9
3
3
4
4
23
19
0.01
0.01
49.87
27.62
15NHAC1 (made by Shanghai Chemical Industrial Institute), abundance
95.44%, was added to the diet.
The fishes were fed a pellet diet containing 30% protein until they ate
it at once when the food was put in the water. After acclimation to the diet,
the fish were divided into three groups: a high protein level diet group, a
129
-------
normal protein level diet group, and a starved group. Before the experiment,
the diet of both the high protein level diet group and normal protein level
diet group, which did not receive 15NH4C1, were used for 15 days. The starved
group also was fed with normal protein level diet for 8 days, then starved for
7 days. Then the test began.
The fish were divided into a feeding group and an injection group for
introducing 15NHAC1 into the organisms. In the feeding group, the fish were
fed a high protein level diet and a normal protein level diet, both of which
contained 15NHAC1. The amount was 0.5 g/fish/day. From the first time of
feeding, three fish were taken for sampling from each group.
In the injection group, a 30% 15NH4C1 injection solution was introduced
at a rate of 0.3 mg NH4C1 per gram of body weight through intraperitoneal
cavity injection. After injection, three fish were taken from each group for
analysis. During the whole experimental time, the fish were fed either a high
protein level or a normal protein level diet not containing 15NH4C1.
The temperature of the water was 24 to 25°C; the dissolved oxygen
content was 3.5 to 5.5 mg/L.
Fish removed for sampling were dissected at once and all muscle weighed.
The amount of N in a certain weight of muscle was determined by Kjeldahl's
method. Another portion of the muscle was treated twice with 5%
trichloroacetic acid to precipitate the protein and then N was determined
using Kjeldahl's method. The abundance of 15N in each sample was determined
by mass spectrometry.
Formulas used in the calculation were: AN1/N1=___RG^Ri
N%=AN/N - SNX
SN2
where: AN/N = the amount of NH4C1-N absorbed (ANj/N^ or transformed
into protein (AN2/N2) within each gram protein-N of fish
muscle
Rg — abundance of 15N in muscle sample
Rf - abundance of 15NH4C1 in diet or injected solution
Ri = the natural abundance of 15N in fish muscle
13$! = the total amount of N in fish muscle
SN2 = the amount of NPN (mg) induced by eating or injection
RESULTS AND DISCUSSION
Tilapia can absorb NH4C1-N through the digestive tract, as shown in
Figure 1.
In the feeding groups, all the fishes in the high protein level diet
group, the normal protein level diet group, and the starved group, showed an
increase of N in muscle according to the continuous experimental time. The
AN/N! rose from 0.052, 0.054, and 0.036 mg N/g N at 5 hours to 0.568, 0.380,
and 0.518 mg N/g N at 48 hours, respectively.
130
-------
0.6-
z
•£ 0.5-
z
O>
"o
*- 0.3-
X
5 0.2-
Z
< 0.1-
0-
^
mi ll
5 10
-
^
24 36 48
Time, hours
Figure 1. The absorption of NI^Cl-N
through digestive tract at differ-
ent times.
Tilapia can absorb the NH4C1-N through the peritoneal cavity, as shown
in Figure 2.
Five hours after injection, the ANj/Nj. was 0.132, 0.224, and 0.155 mg
N/g N for the high protein level group, the normal protein level group, and
the starved group, respectively. This means that, in tilapia, NPN was
absorbed through peritoneal cavity very quickly. From 5 hours to 10 hours,
the ANi/Nj. decreased, meaning that, at this time, a certain amount of NPN was
excreted. From 10 hours to 48 hours, the AN.,/^ continued to decrease slowly,
which means that NPN may take part in synthesizing into protein.
z •
z
m 0.3 -
0.2-
.p
\
}
5 10
J
I
24
Time, hours
m
36
•Ito
48
Figure 2. The absorption of
through peritoneal cavity at dif
ferent times.
131
-------
After the NH4-N was absorbed, does the N take part in synthesizing into
protein? We used trichloroacetic acid to treat the muscle of fish in feeding
group in order to precipitate the protein in muscle. We then determined the
abundance of N and observed its variation and calculated the amount of NPN
transferring into the muscle. The result shows that the abundance of 15N in
muscle increased according to digestion and absorption (Figure 3).
0.6-
Z
o>
0.5-
0.4-
•o
*- 0.3-1
X
1«H
< o.,J
10 24 36
Time, hours
48
Figure 3. The amount of NPN trans-
ferring into the protein of fish
muscle and the total amount of
absorbed NPN.
In comparing the amount of NPN transferring into the protein of fish
muscle (AN2/N2) with the total amount of absorbed NPN (AN^N^ , we see that the
ratio (R) has risen. At 10 hours, the R values of the three groups are 54.1%
56.1%, and 55.5%, respectively. At 48 hours, the R values rise to 75.5%,
74.7%, and 60.3%, respectively. It means that, in tilapia, not only theW-N
can be absorbed through digestive tract but also the absorbed NPN can be
synthesized continuously into the protein of muscle.
^Using the feeding amount within 48 hours, the total amount of muscle
protein in each fish, its percentage in the total fish protein, and other
data, we established the absorptivity of NHAC1 in 48 hours by tilapia body and
muscle and its transferability of protein from the absorbed NHA-N (Table 2)
A few papers Delate metabolism of fish studied with stable nuclide 15
'N.
m — ---- - ---- - — > —-n^-.b v w TT j. *,ii ta W-U.I_/ J-^r J.II_4.*^ .L J. U-C J.1 *
Yitou studied the juvenile of common carp, crucian carp, loach, etc. absorbing
the NH3-N and N02-N from water. He confirmed that the absorbed NH3-N takes
part in metabolism entering various tissue and organs by biosynthesis.
In this paper, we proved that the amount of absorbed NPN in muscle at 48
hours is ten times that at 5 hours; that the absorptivity of NPN in muscle
after injection at 5 hours is 4 to 5 times the absorptivity after feeding at 5
hours; and that, by treating with 5 mg NH4C1, after treatment at 5 hours the
absorptivity of NPN through the peritoneal cavity is 3 times that through
digestive tract. From the R value at 48 hours, we estimate that most of the
absorbed NPN took part in the synthesis of muscle protein.
132
-------
TABLE 2. THE ABSORPTIVITY OF NHA-N BY FISH BODY AND MUSCLE PROTEIN
AND THE TRANSFERABILITY OF PROTEIN FROM NH4-N IN MUSCLE IN
48 HOURS
Absorptivity of
Fish body Fish muscle
Transferability
of NH4-N into
fish muscle, %
High protein level diet
Normal protein level diet
Starved group
3.27
2.54
1.77
1.97
1.53
1.06
75.5
74.7
60.3
Now it is sure that, in tilapia, the NHA-N can be absorbed directly and
be synthesized into protein. It can be quantified, but the amount of
inorganic N replacing protein is very little. In the ruminant, adding 2% urea
can replace 20 to 30% protein in the diet. And the dietary coefficient can be
decreased to 1.3 by adding 2% urea into the diet of rainbow trout.
Urea is the end excreta of metabolism in mammals. After they are
decomposed they can be utilized to synthesize protein by microorganisms. This
is a kind of indirect utilization but is not direct utilization as this
experiment.
CONCLUSION
In tilapia, the direct absorption of NH4-N can be both through its
digestive tract and the peritoneal cavity. At 5 hours, the absorptivity
through the peritoneal cavity is higher than that through the digestive tract.
The absorbed NHA-N took part the synthesis of body protein. In tilapia, at 48
hours, the transferability is 60.3 to 75.5% in the muscle.
In tilapia, the absorptivity of 2% NH4C1 in diet is 1.77 to 3.27%.
REFERENCES
Chen, Mui Bin et al. 1986. Studies of the application of urea in the pellet
diet of carp. Freshwater Fisheries. 3:6-9.
Garling, D. L. Jr., and R. P. Wilson. 1976. Optimum dietary protein to
energy ratio for channel catfish fingerlings (Ictalurus punctatus). J.
Nutr. 106:1368-1375.
Jauncey, K. 1982. The effects of varying dietary protein level on the
growth, food conversion, protein utilization, and body composition of
juvenile tilapias (Sarotherodon mossambicus). Aquaculture. 27:43-54.
Lin Ding et al. 1980. Optimum requirement of protein for juvenile of grass
carp (Ctenopharyngodon idellus) in growth phase. Acta Hydrobiologica
Sinica. 7(2):207-212.
Ogino, C. and K. Saito. 1970. Protein nutrition in fish. The utilization of
dietary protein by young carp. Bull. Jpn. Soc. Sci. Fish. 36:250-254.
133
-------
Satia, B. P. 1974. Quantitative protein requirements of rainbow trout
Progress in Fish Culture. 36:80-85.
Takeuchi, T., T. Watanabe, and C. Ogino. 1979. Optimum ratio of dietary
energy to protein for carp. Bull. Jpn. Soc. Sci. Fish. 45:983-987
Wang, Yi-Qiang et al. 1987. Localization of insulin in grass carp (Ct-
enopharyn^odon idellus) and black carp (Mvlopharvngodon piceusO with
immunocytochemical method (Unpublished).
Yang, Guo-Hua et al. 1980. Fish nutrition and nutrient targets for several
Chinese carp. Scientific report of the Shanghai Fisheries Research
Institute (No. 1) 3-14.
134
-------
OXYGEN. CARBON DIOXIDE AND AMMONIA TRANSFER
ACROSS TELEOST FISH GILLS
by
David J. Randall1
Fish transfer oxygen, carbon dioxide and ammonia across their gills
between water and blood. About equal amounts of oxygen and carbon dioxide are
transferred, but in opposite directions, whereas ammonia excretion is only 10
to 30% of oxygen uptake. This, of course, is a reflection of metabolic
utilization and production of these compounds. The carbon dioxide excretion:
oxygen uptake exchange ratio is usually between 0.7 and 1.0, whereas the
ammonia:carbon dioxide exchange ratio is between 0.1 and 0.3 (Randall and
Wright 1989). Variations in the exchange ratio outside these levels result
from changes in body stores of carbon dioxide and/or ammonia.
Body stores of these two compounds represent about two or three times
the excretion rate per hour. Oxygen stores, ignoring that in the swimbladder,
are only sufficient to maintain tissue requirements for about 5 minutes. If
the swimbladder contains only oxygen then, assuming the fish can utilize all
the oxygen present, swimbladder oxygen could maintain tissue oxygen require-
ments for a few hours (see Randall and Daxboeck, 1984, for review). It is
possible that swimbladder oxygen stores supplement oxygen requirements during
exposure of fish to hypoxic environments (Smith et al. 1983).
The gills consist of a series of gill arches, supporting rows of
filaments with lamellae, forming a sieve-like structure in the path of the
water flow. The blood circulation is complex having both lamellar and
filamental components. The filament circulation is part of a secondary, low
hematocrit circulation that serves superficial structures in the fish.
Surface structures in fish obtain oxygen from and excrete carbon dioxide and
ammonia directly into the surrounding water (Randall and Daxboeck 1984).
Oxygen transfer across the gills into the blood is to supply internal
organs only. The secondary circulation to the skin and gill filaments is not
to exchange gases but rather to deliver organic nutrients, remove waste-
products, and maintain osmotic balance in the tissues. In resting fish, skin
and gill oxygen consumption represents about 20 to 30% of total oxygen uptake
^University of British Columbia, Vancouver BC, Canada.
135
-------
by the fish; that is, only 70% of oxygen consumed by the fish crosses the
gills into the blood, the rest is supplied to superficial structures directly
from the water. The fraction of total oxygen uptake crossing the gills will
increase with exercise because the increase in oxygen uptake goes almost
entirely to the working muscles that receive their oxygen via the blood.
Blood flows through the lamellae countercurrent to the water flow.
Blood flows around pillar cells, which act as posts, embedded with collagen
that extends around the blood space, holding the two walls of the lamellae
parallel to each other (Figure la). The blood sheet is about 9 to 12 urn thick
and is very dependent on blood pressure. The lamallae show no increase in
length or height as blood pressure rises but the width does increase with
blood pressure (Farrell et al. 1980). Gas transfer occurs across the lamellar
wall, which consists of a pillar cell wall, a basement membrane, and two
layers of epithelial cells.
There are intracellular spaces between epithelial cells but the outer
layer of cells are bound together by tight junctions, ensuring a low osmotic
and ionic permeability of the respiratory epithelium. Mucus covers the apical
surface of the epithelial layer and water flows between the lamellae in a
sheet about 25 urn thick. Blood transit time through the lamellae is about a
second whereas that for water is about 100 milliseconds.
Oxygen uptake across the gills is both perfusion and diffusion limited
(Randall 1982). Increases in oxygen uptake are achieved by increases in both
blood and water flow and the diffusing capacity of the gills. Maximum oxygen
delivery to the tissues is determined by cardiac output and arterial oxygen
content (Jones 1971). Many teleost fish have blood with a marked Root shift,
that is, a decrease in pH reduces the oxygen capacity of the blood even at
high Po2. The Root shift plays an important role in unloading oxygen into the
swimbladder.
Acidotic conditions, therefore, could lead to a reduction in oxygen
delivery to the tissues by reducing blood oxygen content. This is
ameliorated, however, by the release of catecholamines into the blood.
Catecholamines react with beta-receptors on red blood cells and stimulate
sodium/proton exchange. This causes a rise in intracellular pH and an
increase in intracellular sodium and chloride levels, which results in
swelling of the erythrocytes as water is sucked into the cells (Nikinmaa et
al. 1987). The elevation in erythrocytic pH, due to stimulation of
sodium/proton exchange by catecholamines, offsets the effect of a reduction in
plasma pH on erythrocytic pH and contributes to the maintenance of arterial
blood oxygen content (Primmett et al. 1986, Randall et al. 1987).
Carbon dioxide is excreted as C02, but a small fraction (about 10% of
total^carbon dioxide excretion in resting fish) is converted to bicarbonate in
the gill epithelium and exchanged for chloride across the apical membrane
(Figure 2). There appears to be no functional carbonic anhydrase activity in
the plasma or associated with the respiratory endothelium (Perry et al. 1982).
In trout, and probably other fish, there is no carbon dioxide excretion in the
absence of red blood cells. Under conditions of controlled blood flow and
carbon dioxide content, excretion is proportional to hematocrit (Perry et al.
136
-------
H
•H rH
tn H
•H
-P tn
0
-------
1982) . Carbonic anhydrase activity in fish red blood cells is high and is
required to catalyse the bicarbonate dehydration reaction in blood in the
gills.
The erythrocyte membrane also contains high concentrations of the
protein associated with chloride/bicarbonate exchange (Table 1) That is red
blood cells are required to ensure adequate rates of bicarbonate dehydration
?Uri£S S°°i transit of the S111' and to maintain low levels of carbon dioxide
in the blood. The gill epithelium also contains high levels of carbonic
anhydrase, this enzyme appears to function in maintaining adequate fluxes of
protons and bicarbonate, through the catalysis of the carbon dioxide hydration
reaction for anionic and cationic exchange processes on the apical surface of
the gill (Figure 2).
c.a = carbonic anhydrase
Na
H
CO
ERYTHROCYTE
WATER
GILL
EPITHELIUM
Figure 2. Diagramme to illustrate the pattern of carbon
dioxide excretion across the gills of teleost fish. The
dotted line represents possible but unlikely pathways
of carbon dioxide excretion in fish.
138
-------
TABLE 1. CHARACTERIZATION OF - CHLORIDE/BICARBONATE
TRANSPORT SYSTEM (BAND III MOLECULES) IN TROUT RED
BLOOD CELLS COMPARED WITH THAT FOR HUMAN
RED BLOOD CELLS (from Romana and Passow 1984)
Band III molecules , cell"1
Cell surface , cm2
Band III molecules, cm"2
Half time for Cl" ion exchange, s:
0°C
10°C
15°C
38°C
Trout
8 x 106
2.67 x 10"6
30 x 1011
3.42
1.29
0.81
Human
1 x 106
1.42 x 10"6
7 x 1011
17 . 2
2.32
0.89
0.05
Carbon dioxide and ammonia are the major metabolic endproducts excreted
across the gills of fish. In general, the fish excretes about ten times as
much carbon dioxide as ammonia. These compounds exist in the body in both an
ionized and unionized form. The ratio of carbon dioxide (C02) to bicarbonate
(HC03~) and ammonia (NH3) to ammonium ion (NHA+) varies with pH, but the pKs of
these reactions are very different. The pK of the NH3/NHA+ reaction is about
9.5, whereas the apparent pK of the C02/HC03" reaction is around 6.1
(Boutiller et al. 1984). The C02/HC03" and NH3/NH4-+ ratios are equal at the pH
where the two lines cross in Figure 3A, at the midpoint between the pK of the
NH3/NHA+ reaction and the apparent pK of the C02/HC03~ reaction. The pH at
this point is similar to the pH of fish blood over a range of temperatures.
The midpoint between these two pKs varies with temperature in much the same
way as blood pH changes in fish. This is of functional significance, for the
animal must maintain the transfer of both ammonia and carbon dioxide out of
the body. [The term carbon dioxide refers to total carbon dioxide (C02 +
H2C03 + HCQ3~ + C03=) , and ammonia to total ammonia (NH3 + NH^) ] .
The gill epithelium is not very permeable to either HC03" (Perry et al.
1982) or NHA+ (see review by Randall and Wright 1987), but is very permeable
to C02 and NH3. Thus, C02 and NH3 will be the predominant forms excreted, as
long as adequate blood-to-water C02 and NH3 gradients exist. This will occur
for both metabolic endproducts if blood pH is maintained at the midpoint
between the pK of the two reactions and environmental levels of NH3 and C02
remain low. The fish must maintain C02 and NH3 excretion because increased
C02 levels result in an acidosis and because high levels of ammonia in the
body are toxic, resulting in convulsions in all vertebrates. A blood
acidosis, however, will favour CO2 excretion whereas a blood alkalosis will
augment NH3 excretion (Hillaby and Randall 1979). Both blood C02/HC03" and
NH3/NHA+ ratios are very low because blood pH is nearly two orders of
magnitude away from the point where pH equals pK for either reaction. Thus in
both instances, the unionized form represents only 1 to 5% of the total and
there are large stores of HC03" and NHA+ in the body.
139
-------
A.
15°C
1.4
CO2/HCO3
10.0
B.
pH
8.0 -
7.8 -
7.6 •
7.U
10
15
20 25
T(°C)
30
35
Figure 3. A. The effect of varying pH on the C02/HC03~ and NH3/NH4+
ratio in trout plasma at 15°C. B. The pH at which the ratios are
equal have been calculated at different temperatures and added to
a graph of variations in plasma pH with temperature for several
fish (from Randall and Wright 1988).
140
-------
Excretion of CO2 and NH3 will be influenced by the composition of water
near the gill surface, which may have a different chemical composition from
that of the bulk medium. The surface of the gills is covered by a mucous
coating and there is also a boundary layer of water next to this mucous^ layer.
Molecular C02 and NH3, along with some HC03" (Perry et al. 1982) and NHA
(Wright and Wood 1985), are excreted into the mucous and boundary water layer
next to the gill surface (Figure 4). Bicarbonate ions represent only a small
portion of the carbon dioxide excreted, probably less than 10%, and the amount
of ammonia excreted as NH3 varies between 45 and 100% (see review by Randall
and Wright 1988). As long as the pH is above 6.1, then most of the C02
excreted will form HC03" and acidify the water. Wright et al. (1986) showed
that this reaction was catalysed by carbonic anhydrase in the mucous layer.
The mucous layer contains a large number of damaged gill epithelial
cells, which are known to contain high levels of carbonic anhydrase (Lacey
1983)! C02 entering this boundary layer will acidify the mucous and boundary
Blood
Boundary
water layer
CO2
11
HCO;
H
Free flowing
water
co2
1L
HCO3
Figure 4. Schematic cross-section through the gill
epithelium, mucus and water boundary layer, contain-
ing carbonic anhydrase (•). The thickness of the
arrows denotes the approximate magnitude of the
particular process illustrated (after Randall and
Wright 1987).
141
-------
layer of the gill (Wright et al. 1986). The extent of acidification will be
determined by the rate of carbon dioxide excretion in relation to water flow
and the buffering capacity and pH of the water. Excreted NH3 will combine
with a proton and increase water pH, but because more C02 than NH3 is excreted
across the gills, the overall effect usually will be an acidification of the
boundary layer. NH3 entering this acidified layer next to the gill surface
will be converted to NH4+ and will diffuse out of the boundary water layer
into the bulk medium.
The conversion of NH3 to NHA+ and its subsequent diffusion from the
boundary layer will reduce NH3 levels in the water and, therefore, NH3 levels
in the blood (Wright et al. 1989). Thus, there is an interaction in the
boundary layer between carbon dioxide and ammonia excretion. Starving trout
have elevated blood NH3 levels (Hillaby and Randall 1979) and this can be
correlated with a decreased C02 excretion.
When water pH is below the apparent pK of the C02/HC03~ reaction, C02
excretion will^ have only a minor effect on water pH. At a water pH of 4.00,
almost no HC03 will form and, therefore, almost no further acidification of
gill water will occur. In fact, the opposite will occur, for NH3 excretion
will now cause an increase in boundary water pH (Figure 5B). At a pH of 5.1,
at 10°C, only 10% of the C02 entering the water will be converted to
bicarbonate, but essentially all NH3 will be converted to NH4+. It is at
about this pH that there is no change in the pH of water as it passes over the
gills (Figure 5B) ; that is, H+ production with HC03~ formation is equal to H+
consumption by NH4+ formation. This observation is consistent with a ratio of
10:1 for C02 to NH3 excretion. The real situation is more complicated than
indicated above, which ignores the contribution of Na+:NH4+ and C1~:HC03~
exchange to total carbon dioxide and ammonia excretion, as well as the
excretion of any other compounds that may affect the pH of gill water.
Acid conditions reduce ammonia excretion because of an inhibition of
Na :NHA exchange (Wright and Wood 1985) resulting in increased blood NH3
levels. Acid conditions, in themselves, will favour NH3 transfer because any
NH3 entering the water will be immediately converted to NH^, maintaining NH3
gradients across the gill epithelium. This effect, however, is not sufficient
to offset the effect of inhibition of Na+:NH4+ exchange, and blood NH3 levels
rise (Wright and Wood 1985). Under less acidic conditions (water pH 6.64)
when there is no inhibition of Na+:NH4+ exchange, blood NH3 levels are reduced
compared with fish exposed to water of pH 7.85 (Wright and Wood 1985).
The reverse is true under alkaline conditions. C02 excretion results in
a large acidification of the boundary layer, as nearly all the C02 is con-
verted to bicarbonate, or even carbonate, in the water (Figure 5B) This
acidification, however, may not be sufficient to maintain NH3 excretion, and
alkaline conditions can lead to increased NH3 levels in the body (Wright and
Wood 1985, Randall and Wright 1989). Elevated ammonia levels stimulate urea
production via uricolysis in some fish, but not trout (Olsen and Fromm 1971).
Oreochromis alcalicus grahami lives in Lake Magadi, an alkaline lake, but
unlike other teleost fish, is completely ureotelic, producing urea via the
ornithine-urea cycle. Thus, this fish avoids problems of ammonia buildup when
exposed to these very alkaline conditions by converting ammonia to the less
toxic urea (Randall et al. 1989).
142
-------
pH electrodes
Open-ended
block plastic box
Valve-
. Latex mask
mounted on divider
V0percular cannula
"Stirring bar
Slopped-flow
apparatus
;upport
for box
B
10-
9-
e-
o.
! 7-
Vancouver dechlorinated water
Rainbow trout
Exhalent
water
Exhalent water pH-
• pH of flowing water
A Stop flow pH
10
Inspired water pH
Figure 5. A—The apparatus for measuring inhalent and exha-
lent water pH in trout. The exhalent water transit time
from the gills to the pH electrode was less than 2 seconds.
The pH of the flowing water in the exhalent chamber was re-
corded and then flow was stopped and the equilibrium pH
(stop flow pH) was measured. B---The relationship of exhalent
to inhalent water pH in trout breathing dechlorinated Vancou-
ver tapwater (from Randall and Wright 1989) .
143
-------
REFERENCES
Boutilier,^R.G., T.A. Heming, and G.K. Iwama. 1984. Appendix:
Physiochemical parameters for use in fish respiratory physiology In
Fish Physiology, Vol. 10A. W.S. Hoar and D.J. Randall (eds.). Academic
Press Inc., New York. pp. 403-430.
Farrell, A.P., S.S. Sobin, D.J. Randall, andS. Crosby. 1980. Intralamellar
blood flow patterns in fish gills. Am. J. Physiol. 239.-R428-R436
Hillaby, B.A. and D.J. Randall. 1979. Acute ammonia toxicity and ammonia
excretion in rainbow trout (Salmo gairdneri^. J. Fish. Res Bd Canada
36:621-629.
Jones, D.R. 1971. The effects of hypoxia and anaemia on the swimming
performance of rainbow trout (Salmo gairdneril. J. Exp. Biol. 55:541-
Lacy, E.R. 1983. Histochemical and biological studies of carbonic anhydrase
activity in the opercular epithelium of the euryhaline teleost, Fundulus
heteroclitus. Am. J. Anat. 166:19-39.
Nikinmaa, M., J.F. Steffensen, B.L. Tufts, and D.J. Randall. 1987. Control
of red cell volume and pH in trout: Effects of Isoproterenol, transport
inhibitors, and extracellular pH in bicarbonate/carbon dioxide-buffered
media. J. Exp. Zool. 242:273-281.
Olson, K.R. and P.O. Fromm. 1971. Excretion of urea by two teleosts exposed
to different concentrations of ambient ammonia. Comp. Biochem Phvsiol
40A:999-1007. '
Perry, S.F., P.S. Davie, C. Daxboeck, and D.J. Randall. 1982. A comparison
of C02 excretion in a spontaneously ventilating blood-perfused trout
preparation and saline-perfused gill preparations: contribution of the
branchial epithelium and red blood cell. J. Exp. Biol. 101-47-60
Primmett, D.R.N., D.J. Randall, M.M. Mazeaud, and R.G. Boutilier. 'l986 ' The
role of catecholamines in erythrocyte pH regulation and oxygen transport
in rainbow trout (Salmo gairdneril . J. Exp. Biol. 123-139-148
Randall, D.J. 1982. The control of respiration and circulation in fish
during hypoxia and exercise. J. Exp. Biol. 100:275-288
Randall, D.J. and C. Daxboeck. 1984. Oxygen and carbon dioxide transfer
across fish gills. In Fish Physiology, Vol. 10A. W.S. Hoar and D.J.
Randall (eds.). Academic Press Inc., New York. pp. 263-314
Randall.^D.J., D. Mense, and R.G. Boutilier. 1987. The effects of burst
swimming on aerobic swimming in chinook salmon (Oncorhynchus
tshawvtscha"). Mar. Behav. Physiol. 13:77-88.
Randall, D.J., C.M. Wood, S.F. Perry, H. Bergman, G!M.O. Maloiy, and P.A.
Wright.^ 1989. Urea excretion as a strategy for survival in a fish
living in a very alkaline environment. Nature. 337:165-166
Randall, D.J. and P.A. Wright. 1987. Ammonia distribution and excretion
in fish. Fish Physiol. Biochem. 3:107-120.
Randall, D.J. and P.A. Wright. 1989. The interaction between carbon
dioxide and ammonia excretion and water.pH in fish. Can. J. Zool In
press.
Romano, L. and H. Passow. 1984. Characterization of anion transport system
in trout red blood cell. Am. J. Physiol. 246:C330-C338
Smith,_D.G., W. Duiker, andl.R.C. Cooke. 1983. Sustained branchial apnea
in the Australian short-finned eel, Anguilla australis J Exp Zool
226:37-43.
144
-------
Wright, P.A., T.A. Heming, and D.J. Randall. 1986. Downstream pH changes
in water flowing over the gills of rainbow trout. J. Exp. Biol.
126:499-512.
Wright, P.A. and D.J. Randall. 1987. The interaction between ammonia and
carbon dioxide stores and excretion rates in fish. Annls. Soc. r. Zool.
Belg. 117 (supplement 1):321-329.
Wright, P.A., D.J. Randall, and S.F. Perry. 1989. Fish gill water boundary
layer: A site of linkage between carbon dioxide and ammonia excretion.
J. Comp. Physiol. In press.
Wright, P.A. and C.M. Wood. 1985. An analysis of branchial ammonia
excretion in the freshwater rainbow trout: Effects of environmental pH
change and sodium uptake blockage. J. Exp. Biol. 114:329-353.
145
-------
CONTROL OF VENTILATION IN FISH
by
E. W. Taylor1 and D. J. Randall2
INTRODUCTION
Respiratory gas exchange in fish takes place by diffusion over the
surfaces of the gills. The gill arches are ventilated by muscular pumps
around the mouth and pharynx at a rate determined by an oscillator in the CNS,
which varies in its activity with oxygen supply and demand. The generation of
the central respiratory rhythm and its modulation by peripheral
mechanoreceptors and chemoreceptors are not completely understood. This
review concentrates on some recent evidence of a role for circulating
catecholamines in the control of the ventilatory response to hypoxemia.
RESPIRATORY RHYTHM GENERATION IN THE CNS
In fish, water is propelled continuously over the gills in one direction
by the ventilatory muscles that operate around the jaws and skeletal elements
in the gill arches, lining the pharynx (Ballintijn and Hughes 1965, Hughes and
Ballintijn 1965). These muscles are innervated by cranial nerves with their
neuron cell bodies located in the brainstem, close to the site of the central
respiratory pattern generator (CPG).
Rhythmic ventilatory movements continue in fish following brain
transection to isolate the medulla oblongata, although' changes in pattern
indicate that there are influences from higher centres (Shelton 1959).
Central recording and marking techniques have identified a longitudinal strip
of neurons with spontaneous respiration-related bursting activity, extending
dorso-laterally throughout the whole extent of the medulla (Shelton 1970,
Waldron 1972). These neurons comprise the trigeminal Vth, facial Vllth,'
glossopharyngeal IXth and vagal Xth motor nuclei, which drive the respiratory
muscles, together with the descending trigeminal nucleus and the reticular
formation (Ballintijn 1982). The respiratory rhythm is thought to originate
in a diffuse CPG in the reticular formation, which remains functional
following anaesthesia (Ballintijn 1988).
England
of Biological Sciences, University of Birmingham, Birmingham,
^Department of Zoology, University of British Columbia, Vancouver, BC
Canada
146
-------
All the motor nuclei are interconnected and each receives an afferent
projection from the descending trigeminal nucleus and has efferent and
afferent projections to and from the reticular formation (Figure 1). The
intermediate facial nucleus, which receives vagal afferents from the gill
arches that innervate a range of tonically and phasically active
mechanoreceptors, projects to the motor nuclei (Ballintijn et ail. 1983).
Finally, areas in the mid brain such as the meseneephalic tegmentum have
efferent and afferent connections with the reticular formation (Ballintijn
1982, 1988). These meseneephalic neurons are responsible for initiating
bursts of ventilation during the periodic or bout respiration shown by some
inactive or hyperoxic fish (Ballintijn 1988).
Studies of retrograde intraxonal transport of HRP along nerves
innervating the respiratory muscles revealed that the neurons in the various
motor nuclei are distributed in a sequential series in the brainstem of fish
(Withington-Wray et ail. 1986, Levings and Taylor 1987, ¥ithington-Wray et ail.
1988). Recordings of efferent activity from the central cut ends of the
nerves innervating the respiratory muscles of the dogfish Scyliorhinus
canLcula (Barrett and Taylor 1985) and ray RaLa clavata (E. W. Taylor and J.
J. Levings, unpublished) have revealed that the branches of the Vth, Vllth,
IXth and Xth cranial nerves fire sequentially in the order of the sequential
rostro-caudal distribution of their moto-nuclei in the brainstem.
IFOREBRAIN
SENSORY
Respiratory
muscle
proprioceptors
Trigeminal Vth
Nucleus
Gill Arch
Receptors
IX & Xth
•?
Facial IVth
Nucleus
| MESENCEPHALON|
CPG
IN
RF
MOTOR
Motor V
Mandibular
Motor V11
Post-spiracular
Motor
Glosso 1X
Motor
Vagal X1
X2
X3
X4
Hypobranchial
Occipi tal
Spinal
1&2
->
— »•
h*
Jaw
Muscles
Opercular
or
Spiracular
Muscles
Gill
Arch
Muscles
Accessory
Respiratory
& Feeding
Muscles
Figure 1. Summary diagram of the possible functional connections
involved in the central nervous control of ventilation in fish.
Abbreviations are: CPG, central pattern generator; RF, reticular
formation.(Taken from Taylor 1989)
147
-------
Experiments in which ventilatory movements were prevented with curare
demonstrated that the CPG in fish continues to generate rhythmic bursts of
activity that are slower than the unparalysed respiratory rhythm in teleosts
(Ballintijn 1972) but faster in elasmobranchs (Barrett and Taylor 1985).
These changes in rate following paralysis indicate that activity in the CPG
normally is modulated by peripheral receptors. This is consistent with the
fact that the CPG receives afferent information from a range of
mechanoreceptors, proprioceptors and skin stretch receptors around the gill
arches and jaws (Ballintijn 1982) with their sensory (afferent) innervation in
cranial nerves V, VII, IX and X. Part of the population of motoneurons
innervating the respiratory muscles is silent in the paralysed animal in both
fish and mammals and may be stimulated to fire by artificially induced
proprioceptive information (Ballintijn 1982). The recruitment of these silent
motoneurons may serve to increase the motor output and consequent amplitude of
contraction of respiratory muscles. In addition, fish may recruit feeding
muscles innervated by the hypobranchial nerve trunk, which contains fibres
from the occipital and anterior spinal motor nuclei (Figure 1). These insert
on the skeletal elements around the mouth and pharynx and can contribute to
active, forced ventilation when respiratory demand is high.
Thus, rhythmic contractions of the respiratory muscles are determined by
a central respiratory pattern generator that is modulated by feedback on the
force of ventilatory contractions from mechanoreceptors in the respiratory
apparatus and by inputs from elsewhere in the CNS.
CHEMORECEPTOR RESPONSES
Fish pump water over their gills at rates controlled with respect to
oxygen supply or demand. In teleost fish, ventilation increases when oxygen
supply is reduced by environmental hypoxia, or when transport in the blood is
reduced either directly by anemia or indirectly by hypercapnia, when the
resultant acidosis causes a Root effect, reducing oxygen carrying capacity
(for references see reviews by Randall 1982, Shelton et al. 1986). Similarly,
an increase in oxygen demand during vigorous swimming or following the stress'
of experimental manipulation results in an increase in ventilation rate.
Obversely, an increase in oxygen supply resulting from environmental hyperoxia
causes a decrease in ventilation rate, which may result in hypoventilatory
hypercapnia (Figure 2).
Most of these responses can be interpreted as arising from the
stimulation of peripheral or central chemoreceptors sensitive to changes in
oxygen supply (i.e., oxygen content and blood flow, Randall 1982). Evidence
for the location and characteristics of the oxygen receptors is largely
circumstantial and the mechanisms by which reflex ventilatory changes are
initiated are not clear (Taylor 1985). Both the ventilatory (Eclancher and
Dejours 1975) and cardiac (e.g., Butler and Taylor 1971) responses to sudden
exposure to hypoxia or exercise are rapid in onset indicating that nervous
pathways are involved. The motor arm of the response is of course nervous as
the ventilatory muscles are innervated by the efferent cranial nerves (Hughes
and Ballintijn 1978) . The afferent arm has not been clearly identified
(Taylor 1985) and neuropharmacology of the central connections implicated in
these reflex responses, and the possible roles for circulating hormones are
not known. Circulating catecholamines may be involved in the control of
ventilation and this review now concentrates on this point.
148
-------
5OO -i
4OO H
30O -
c
20O -
100
9.4
17.2
Hypoxia
15.0
Hypercapnia
Anemia
14.6
Normoxia
33.3
Hyperoxic
hypercapnia
i i i i i
45 6789
Arterial blood oxygen content (Vol %)
Figure 2. The relationship between rate of water
flow over the gills and the oxygen content of
arterialised dorsal aortic blood from the
rainbow trout Salmo gairdneri (mean values
± SE). The number by each point is the cor-
responding arterial oxygen partial pressure
in kPa. (Taken from Randall 1982)
BLOOD CATECHOLAMINE LEVELS
Teleost fish have low (<0.5 nmol I"1), but measurable, resting levels of
circulating catecholamines (Boutilier et al. 1988). The levels are higher in
the elasmobranch fish ScylLorhinus canicula. (above 20 nmol I"1) , possibly
because these fish lack a sympathetic connection to the heart and branchial
(gill) apparatus, which may enhance the role of circulating catecholamines in
cardiovascular and ventilatory control (Butler et al. 1978). Changes in
physiological state, particularly those induced by stressful stimuli such as
physical disturbance (Nakano and Tomlinson 1967) cause an increase in
circulating catecholamine levels. Other stimuli include hypoxia (Boutilier et
al. 1988), anemia (Iwama et al. 1987), hypercapnia (Perry et al. 1988), acid
infusion (Boutilier et al. 1986), and violent exercise (Primmett et al. 1986).
Circulating catecholamines.play a number of important roles in the
control of metabolism and respiratory gas exchange and transport, including
149
-------
stimulation of anaerobic glycolysis during exercise; increased contractility
of the heart, vasodilation or vasoconstriction of peripheral blood vessels
?TT /?!!? v nt chanSes in blood Pressure and vascular resistance to blood flow
(Wood 1975); increased permeability of the gill respiratory epithelium (Isaia
et al. 1978); release of red blood cells from the spleen, increased red cell
volume and intracellular PH (Nikinmaa 1982). In addition, it is now clear
that many stimuli causing ventilatory increases also are accompanied by
changes in circulating catecholamines (Figure 3). The simplest interpretation
of_this observation is that the ventilatory changes are in some way induced or
reinforced by changes in circulating catecholamine levels.
CATECHOLAMINES AND VENTILATION
Measurable increases in circulating catecholamine levels in fish are
associated with environmental hypoxia in both teleosts (Boutilier et al 1988)
and elasmobranchs (Butler et al. 1978). The hypoxemia resulting from anemia
(Iwama et al. 1987) and hypercapnia (Perry et al. 1988) also is associated
with increased catecholamine levels. Thus many stimuli that result in
increases in ventilation (Figure 2) also cause increased levels of circulating
catecholamines (Figure 3). • &
Evidence for a direct link between ventilatory responses and changes in
catecholamine levels have been obtained by acid infusion (K. Holmgren
unpublished observations) and hypoxia (S. Aota, unpublished observations) in
trout Both cause an increase in circulating catecholamines and in gill -
ventilation. °
The £-adrenergic blocking agent, propranolol, inhibits the ventilatory
infus?™ £5 n°^ ^ incre?Se in blood catecholamines. During hyperoxia, acid
infusion did not cause an increase in circulating catecholamines nor was there
any increase in .gill ventilation. The simplest explanation of these data is
isaLrt?VrfTa^ ^ yentilation durin§ hypoxia or following an acid injection
is mediatedjry the release of catecholamines into the circulation. The site
of action of_these catecholamines could be either peripheral or central, but
the pathway is blocked by propranolol and so involves ,9-adrenergic receptors.
SITE OF ACTION OF CATECHOLAMINES
Peyraud-Waitzenegger (1979) showed that intravenous injection of
adrenaline in the eel caused an increase in ventilation. The injection of
catecholamines into dogfish (Scyliorhinus canicula and Squalus acanthius) •
caused changes in central respiratory drive measured as the efferent, motor
activity in a branchial branch (to a gill arch) of the vagus nerve of
paralysed fish (i.e., in the absence of peripheral mechanoreceptor inputs)
The fish were force ventilated with hyperoxic seawater to reduce stress and
decrease endogenous levels of circulating catecholamines (E. ¥. Taylor and D
J. Randall-unpublished observations). Intravenous injection of adrenaline or
noradrenaline after a short delay (20 sec), caused a transient but marked
increase in efferent activity in S. canicula, which recovered to normal
activity, then after a slightly longer delay showed a further transient
increase.
150
-------
z
o
l-
<
a:
H
z
LU
U
z
to p
Ul U
< LU
_J t/)
o <
T LU
LU (J
H
60
20
16
± 12
Z
o
P 8
a:
o
a.
s *
a.
A NA
HYPOXIA
A NA
ANEMIA
A NA
HYPERCAPNIA
A NA
ACID
INFUSION
Figure 3. The proportional increase in the plasma levels of
circulating catecholamines in the trout Salmo gairdneri
elicited by exposure to hypoxia or hypercapnia by experi-
mentally induced anemia and by acid infusion. Abbrevia-
tions are: A, adrenaline; NA5 noradrenaline.
Our interpretation of this rather complex response is that the bolus of
blood containing high levels of injected catecholamines passed for two
circuits of the circulatory system over a receptive area that triggered an
increase in activity in the medulla. Subsequently, the catecholamines would
be extracted from the blood and metabolized by the gills and other tissues
(Nekvasil and Olson 1986a). The approximate location of the receptive area
can be deduced from our limited knowledge of circulation times in the dogfish
to lie efferent to the heart, possibly in the cerebral circulation. These
effects of intravenous injection of high concentrations of catecholamines were
most marked for noradrenaline. This may cross the blood-brain barrier in fish
more easily than adrenaline (Nekvasil and Olson 1986 b).
When catecholamines were injected into curarised, hyperoxic S. canLcula
in which activity in a hypobranchial nerve, which innervates feeding muscles,
151
-------
was recorded simultaneously with that in a branchial branch of the vagus,
direct evidence of nervous recruitment was observed (J. J. Levings and E. W.
Taylor, unpublished observations). In the hyperoxic fish prior to injection
of adrenaline, the branchial branch showed regular bursting activity, but the
hypobranchial nerve showed low levels of activity and only fired .
intermittently, with a respiration-related activity, injection of adrenaline
caused high levels of respiration-related, bursting activity in the
hypobranchial nerve, which resembled or exceeded that observed in the
disturbed normoxic and fictively hyperpnoeic fish. This result seems to
identify a role for circulating catecholamines in the onset of forced
ventilation.
We currently are investigating the possibility that catecholamines exert
their effects upon the central pattern generator or the respiratory
motoneurons by direct injection of catecholamines into the CNS. Injection of
small volumes- (8-20 /*!) of a 10"4 molar solution of adrenaline into the fourth
ventricle of S. acanthius caused a marked change in the pattern of central
respiratory drive (Figure 4). The response was complex but stereotyped. An
initial increase in the rate of bursting of respiratory motor units was
followed by a slowing of the rhythm accompanied by a huge increase in the
activity within each burst (Figure 4b), apparently due to the recruitment of
units having larger recorded spikes, implying larger fibre diameters, as all
vagal fibres are myelinated (Short et al. 1977). This implies that the
increase in ventilation brought about by injection of catecholamines may
result from an increase in stroke volume rather than rate of contractions of
the respiratory apparatus. The response to injected catecholamines was
blocked by simultaneous injection of propranolol, a j8-adrenergic antagonist
(Figure 4c).
These data imply that areas in the CNS, accessible from the fourth
ventricle, respond to an increase in catecholamine concentration with an
increase in central respiratory drive. The vagal respiratory motoneurons lie
in the dorsal vagal motor nucleus, which is situated bilaterally in the
medulla in a medial position close to the wall of the fourth ventricle
(Withington-Wray et al. 1986). These neurons are rhythmically active, under
the influence of the central pattern generator, and supply the efferent
innervation to intrinsic respiratory muscles in the gill arches. Their
location close to the wall of the fourth ventricle may have caused them to be
directly affected by the injection of catecholamines into this location.
Identification of the site of action of catecholamines on the respiratory
rhythm generator and associated respiratory motor neurons in the medulla must
await microelectrode studies of single, identified neurons.
CONCLUSIONS
Our current knowledge of the control of ventilation in fish is
incomplete at all levels. The respiratory rhythm.originates in a medullary
central pattern generator, which has yet to be clearly identified and
characterised. Its activity is modulated by inputs from elsewhere in the CNS
and from peripheral mechanoreceptors. The central location of respiratory
motor neurons, innervating the various respiratory muscles, has been described
in detail for some fish, particularly elasmobranchs.
152
-------
adr-
15 sec
b control
post adr-
1 sec
ADR
ADR PROP
ADR PROP
15 sec
Figure 4. The effect of injection of adrenaline into the fourth ventricle
of the dogfish Squalus acanthius upon efferent activity in the third
branchial (respiratory) branch of the vagus nerve, a—injection of 20 ul
of 10~^ molar adrenaline (adr.), after a 20-sec latent period, induced a
transitory increase in the frequency of the bursting activity recorded
from the nerve. After'-;about 75 sec, this was replaced by a slowing in
bursting rate accompanied by an increase in the amplitude of the recorded
bursts, b—detail of recording from trace in (a) to compare the rate
and amplitude of the bursts recorded before and after injection of adre-
naline, c—injection of adrenaline (ADR) into the fourth ventricle of
another experimental animal induced a similar response to (a) that was
subsequently blocked by simultaneous injection of adrenaline and propran-
olol (ADR PROP). 3
-------
We are still unclear, however, about the link between the CPG and the
sequential firing of the motor neurons, which result in coordinated
contractions of the respiratory muscles, and about the mechanisms that result
in recruitment of feeding muscles into forced ventilation. Ventilation is
matched to oxygen requirements by stimulation of chemoreceptors, which seem to
respond to oxygen content or supply. The precise location and characteristics
of these chemoreceptors are still not known.
Chemoreceptor stimulation evokes a number of reflex changes in the
respiratory and cardiovascular systems of fish that are rapid in onset and
seem adaptive (e.g., increased ventilation and a bradycardia in response to
hypoxia). Conditions that result in hypoxemia and the consequent ventilatory
changes also cause an elevation in circulating catecholamine levels. We have
explored the possibility of a causal relationship between these levels and the
ventilatory response. Strong evidence arises from experiments on hypoxia and
acid infusion, which trigger a ventilatory increase and a rise in circulating
catecholamines. Both ventilatory responses are blocked by an injection of
propranolol despite an increase in catecholamine levels.
The ventilatory response to hypoxia, at least, occurs very rapidly,
perhaps before any marked increase in circulating catecholamines and almost
certainly before any blood catecholamines could reach the respiratory neurons.
This argues for an immediate neuronal reflex based on chemoreceptors in the
gill region responding to hypoxia. Clearly, circulating catecholamines also
affect ventilation through some action in the medulla and could act in concert
with a direct neuronal chemoreceptive drive during hypoxia. .The studies on
acid infusion during hyperoxia, where there is an acidosis but no increase in
ventilation or blood catecholamines, would argue against any hydrogen ion
receptor, either peripheral or central, being involved in the reflex
ventilatory response to acidotic conditions in fish.
The release of catecholamines into the circulation, therefore, seems to
be an absolute requirement for the ventilatory response to acidosis in fish.
Present evidence supports a role for ^8-adrenergic receptors on respiratory
neurons, stimulated by changes in the levels of circulating catecholamines, in
the control of ventilatory responses to changes in blood oxygen levels in
fish.
ACKNOWLEDGMENTS
This work was supported by a grant from NATO which enabled EWT to visit
the University of British Columbia and funded the work at Bamfield Marine
Station. We wish to acknowledge this support and the help provided by the
staff at Bamfield. Our attendance at the Symposium was funded by the U.S.
Environmental Protection Agency and we are grateful to both the EPA and
Zhongshan University, Guangzhou, PRC, for organizing this successful
international meeting.
154
-------
REFERENCES
Ballintijn, C. M. 1982. Neural control of respiration in fishes and mammals.
In: Exogenous and Endogenous Influences on Metabolic and Neural Control.
A. D. F. Addink and N. Spronk (eds.). Pergamon Press, Oxford.
Ballintijn, C. M. 1988. Evolution of central nervous control of ventilation
in vertebrates. In: The Neurobiology of the Cardiorespiratory System.
E. W. Taylor (ed.). Manchester University Press, Manchester.
Ballintijn, C. M. and G. M. Hughes. 1965. The muscular basis of the
respiratory pumps in the trout. J, Exp. Biol. 43:349-362.
Ballintijn, C. M., B. L. Roberts, and P. G. M. Luiten. 1983. Respiratory
responses to stimulation of branchial vagus nerve ganglia of a teleost
fish. Respir. Physiol. 51:241-257.
Barrett, D. J. and E. W. Taylor. 1985. Spontaneous efferent activity in
branches of the vagus nerve controlling heart rate and ventilation in
the dogfish. J. Exp. Biol. 117:433-448.
Boutilier, R. G., G. K. Iwama, and D. J. Randall. 1986. The promotion of
catecholamine release in rainbow trout, Salmo gairdneri, by acute
acidosis: Interactions between red cell pH and hemoglobin oxygen-
carrying capacity. J. Exp. Biol. 123:145-157.
Boutilier, R. G., G. Dobson, U. Goeger, and D. J. Randall. 1988. Acute
response to graded levels of hypoxia in rainbow trout (Salmo gairdneri):
metabolic and respiratory adaptations. Respir. Physiol. 71:69-82.
Butler, P. J. and E. W. Taylor. 1971. Response to the dogfish (Scyliorhinus
canicula L.) to slowly induced and rapidly induced hypoxia. Comp.
Biochem. Physiol. 39A:307-323.
Butler, P. J., E. W. Taylor, M. F. Capra, and W. Davison. 1978. The effect
of hypoxia on the levels of circulating catecholamines in the dogfish,
Scyliorhinus canicula. J. Comp. Physiol. 127:325-330.
Eclancher, B and P. Dejours. 1975. Control de la respiration chex les
poissons teleosteens: existence de chemorecepteurs physiologiquement
analogues aux chemorecepteurs des vertebras superieur. C. R. Acad. Sci.
Paris Ser. D 280:451-453.
Hughes, G. M. and C. M. Ballintijn. 1965. The muscular basis of the
respiratory pumps in the dogfish. J. Exp. Biol. 43:363-383.
Isaia, J., J. P. Girard, and P. Payan. 1978. Kinetic study of gill
epithelial permeability to water diffusion in the freshwater trout,
Salmo gairdneri: Effect of adrenaline. J. Membrane Biol. 41:337-347.
Iwama, G. K., R. G. Boutilier, T. A. Heming, P. A. Wright, D. J. Randall, and
M. Mazeaud. 1988. The interaction between gill ventilation, blood
catecholamines and gas exchange in rainbow trout.
Levings, J. J. and E. W. Taylor. 1987. Vagal, preganglionic innervation of
the gut in the lesser spotted dogfish Scyliorhinus canicula. J.
Physiol. 394:99P.
Nakano, T. and N. Tomlinson. 1967. Catecholamine and carbohydrate
concentrations in rainbow trout (Salmo gairdneri) in relation to
physical disturbance. J. Fish. Res. Bd. Canada, 24: 1701-1715.
Nekvasil, N. P. and L. R. Olson. 1986a. Extraction and metabolism of
circulating catecholamines by the trout gill. Am. J. Physiol.
250:R526-R531.
Nekvasil, N. P. and K. R. Olson. 1986b. Plasma clearance, metabolism and
tissue accumulation of 3H-labelled catecholamines in trout. Am. J.
Physiol. 250:R519-R525.
155
-------
Nikirunaa, M. 1982. Effects of adrenaline on red cell volume and
concentration gradient of protons across the red cell membrane in the
rainbow trout, Salmo gairdneri. Mol. Physiol. 2:287-297.
Perry,^S. L. , P. Kincaid, P. Fletcher, and D. J. Randall. 1988. Factors
influencing the release of catecholamines (in preparation).
Peyraud-Waitzenegger, M. 1979. Simultaneous modifications of ventilation and
arterial Po2 by catecholamines in the eel, Anguilla anguilla L. :
participation of A and B effects. J. Comp. Physiol. 129:343-354.
Primmett, D. R. N. D. J. Randall, M. Mazeaud, and R. G. Boutilier. 1986.
The role of catecholamines in erythrocyte pH regulation and oxygen
transport in rainbow trout (Salmo gairdneri) during exercise. J EXD
Biol. 122:139-148.
Randall, D. J. 1982. The control of respiration and circulation in fish
during exercise and hypoxia. J. Exp. Biol. 100:275-288.
Shelton, G. 1959. The respiratory centre in the tench (Tinea tinea L/),. I.
The effects of brain transection on respiration. J. Exp Biol 36-191-
202.
Shelton, G. 1970. The regulation of breathing. In: Fish Physiology, Vol.
4. W. S. Hoar and D. J. Randall (eds.), Academic Press, New York.
Shelton, G., D. R. Jones, and W. K. Milsom. 1986. Control of breathing in
ecothermic vertebrates. In: Handbook of Physiology-The Respiratory
System, S. R. Geizer, A. P. Fishman, N. S. Cherniack and J. G.
Widdicombe (eds.) Section 3, Vol, II, pp. 857-909. American
Physiological Society, Bethesda, Maryland.
Taylor, E. W. 1985. Control and coordination of gill ventilation and
perfusion. Symp. Soc. Exp. Biol. 39:123-161.
Taylor, E. W. 1988. Cardiovascular respiratory interactions in fish and
crustaceans. In: Neurobiology of the Cardiorespiratory System, E. W.
Taylor (ed). Manchester University Press, Manchester.
Taylor, E. W. 1989. Nervous control of ventilation and heart rate in
elasmobranch fish, a model for the study of the central neural
mechanisms mediating cardiorespiratory interactions in mammals. In:
Nonmammalian Animal Models for Biomedical Research, A. D. Woodhead (ed )
• CRC Press, Florida.
'Waldron, I. 1972. Spatial organisation of respiratory neurones in the
medulla of tench and goldfish. J. Exp. Biol. 57:449-459.
Withington-Wray, D. J., B. L. Roberts, and E. W. Taylor. 1986. The
topographical organisation of the vagal motor column in the elasmobranch
fish, Scyliorhinus canicula L. J. Comp. Neurol. 248:95-104.
Withington-Wray, D. J., E. W. Taylor, and J. D. Metcalfe. 1988. The location
and distribution of vagal preganglionic neurones in the hindbrain of
lower vertebrates. In: The Neurobiology of the Cardiorespiratory
System, E. W. Taylor (ed.). Manchester University Press, Manchester.
156
-------
ACCUMULATION OF CARBON DIOXIDE IN FISH FARMS
WITH RECIRCULATING WATER
by
J. F. Steffensen1 and J. P. Lomholt2
INTRODUCTION
In northern Europe, intensive culturing of eels, Anguilla anguilla. in
heated fresh water at 25°C is economically feasible in recirculated systems
provided with mechanical and biological filtration. Because of the
consumption of energy for heating, the eels are grown in tanks with
recirculating water, located in well-insulated buildings. The culture is
based on weaning of wild glass-eels to dry feed.
The density of fish is high (75 to 150 kg/in2) in eel farms with
recirculating water. Because the eels are fed continuously and the water is
warm, the oxygen consumption of the fish is high. Some fish are reported to
grow slower at hypoxic conditions (pOo = 70 mmHg) (Stewart et al. 1967).
Hence, it is common practice to enrich the water supplying the fish tanks with
pure oxygen (to 200 to 400% 02 saturation) to ensure that the eels are never
exposed to hypoxic water.
In a typical eel farm with recirculating water (Figure 1), normoxic
conditions in the fish tanks is fulfilled by injecting pure oxygen, but carbon
dioxide produced by the fish is not removed effectively. Because high carbon
dioxide concentration is reported to inhibit oxygen consumption, to reduce the
oxygen affinity and the oxygen carrying capacity of blood (Root 1931), or even
to induce a narcotic acidicosis, it may be profitable to focus on reducing
carbon dioxide in recirculating eel farms.
THE EEL FARM
A diagram of a typical Danish eel farm is shown in Figure 2. From the
fish tanks, the water flows through an over-flow to a sludge separator and
pump sump. The water then is pumped through an aerator to submerged and
trickle filters for nitrification and denitrification. Finally the water is
pumped through an oxygenator, supplied with pure oxygen, and back to the fish
tanks. In the figure, the measured pH and partial pressures of oxygen (pO£)
are indicated, the latter in mmHG. Water is maintained at 6.8 to 7.2 pH by
adding bicarbonate or another base. Danish ground water has a high
bicarbonate concentration and a relatively high C02 concentration.
It appears that the oxygen tension is normoxic or hyperoxic in the
entire system. It also appears that carbon dioxide tension is considerably
•"-The Marine Pollution Laboratory, The National Environmental Protection
Agency, Charlottenlund, Denmark
n
•'Department of Zoophysiology, University of Aarhus, Aarhus C, Denmark
157
-------
Production
Standing stock
Number of fish tanks
Area of fish tanks
Volume of fish tanks
Total vol. of system
Reclrculatlon flow
Density of fish
Food - dry pellets
Water oxygen saturation
Oxygen supply
HCO3- supply
Production time
Temperature
Area of filter
Current market value
100 ton/year
60 ton •
60
860 m2 (4.6 m dla.)
426 m3
660 m3
660 m3/hour
60 - 100 kg/m2
660 kg/day (1.1 % b.w. pr. day)
300 % at Inlet - 100 % at out!
300 kg/day
76 kg/day
24 months (from .36 to 260 g)
26 DegC
26000 mZ
B US$/kg
Figure I. Typical eel farm with recirculating water.
elevated, hypercapnic, in the entire system, even though some CC>2 is "blown
off", especially at the trickle filter. A cC02 of 32 mmHg. For comparison,
CC>2 in the normal eel habitat rarely exceeds 5 mmHg. Values of 20 to 30 mmHg
are far above what eels normally are exposed to.
THE EEL AND THE BLOOD
Because the excretion of carbon dioxide over the gills is a passive
diffusion process, pC02 in the blood is somewhat higher than in the ambient
water. Blood pCO? in eels in natural conditions is 3 to 6 mmHg. In the fish
farm with pC02 = 32 mmHg in the fish tank, the blood pC02 must be even higher.
A pC02 above 30 mmHg for fish is far higher than normal. An increase in
C02 will cause a decrease in pH, which in fish will be compensated for by an
increase in the blood bicarbonate concentration. The coherence between pH,
pC02, bicarbonate concentration, and total C02 concentration is shown in the
Davenport diagram in Figure 3.
Assuming a pC02 of 5 mmHg and pH = 7.7 as normal values for eel,
bicarbonate concentration equals 10 meq/1, according to Figure 3. In the eel
farm situation, however, with a pC02 of 32 mmHg, the blood will have a
slightly lower pH. At pH = 7.5 and 7.4, bicarbonate concentration will be 37
and 29 meq/1, respectively. These bicarbonate concentrations are far higher
than what hitherto has been considered physiologically normal for fish.
Heisler (1984 1986) suggested that the maximal blood HC03- concentration
in hypercapnic fish, which can be used for compensating a hypercapnic
acidosis, is 30 mmol~l. Recently, however, Dimberg (1988) reported
bicarbonate concentrations of 55 to 66 mmol"-'- in rainbow trout exposed to
hypercapnic water with a pC02 of 26 to 34 mmHg.
CONCLUSION
Carbon dioxide has long been known to reduce the affinity of blood for
oxygen (Root 1931), to reduce the oxygen carrying capacity of the blood, and
to inhibit the metabolic rate of fish (Basu 1959, Saunders 1962, Beamish
1964). Basu reported that the active oxygen consumption of several freshwater
fishes decreased with increasing C02- At higher concentrations, C02 depresses
ventilation and can even induce a narcotic acidosis (Dejours 1988).
Only limited information is available on the effect of C02 and oxygen on
swimming capacity. Dahlberg et al. (1968) reported that the prolonged
158
-------
swimming speed of largemouth bass did not decrease in response to
concentrations up to 48 mg 1"1. The performance of coho salmon, on the
contrary, decreased with increasing CC>2 concentrations between 2 and 61 mg
I"1. The effect of COg was especially pronounced at high concentration of
oxygen (lOmg 1 ). This decrease of performance may be a result of the
decrease in blood oxygen affinity and carrying capacity, or an increased cost
of acid-base regulation. :
In spite of the unnaturally high CC>2 tensions in the fish-farms, the
eels seems to be in good health and growing well. It is likely, however, that
these extreme CC>2 tensions have some consequences for the well-being of the
eels. Even though the fish are growing well, they may grow even better at
more normal COo conditions, if, for example, less energy has to be used for
acid-base regulation. In one fish farm, it was observed that the eels
appetite increased after installing a trickle filter.
The results_of future research (growth studies, acid-base balance, etc.)
will determine whether more effort should be aimed at decreasing the CC>2
levels in intensive fish farms with recirculating water and oxygenators.
If it is desired to get more natural C02-conditions, it is not
sufficient to adjust the water pH by supplying bicarbonate, as practiced in
many fish farms. Only by increasing the contact between the water and the
AIR
OXYGENATOR
BIOLOGICAL
SUBMERGED FILTER
SLUDGE
& WATER
PUMP
Figure 2. Diagram of eel farm with recirculating water.
159
-------
aC02 = 0.04 mM/l : pKapp: 6.12-5.97
I 1 1 1 1
6.7
6.9
PH
Figure 3. Davenport diagram.
atmosphere, for example with trickle filters, can the diffusion of CC>2 from
the water to the atmosphere be facilitated, hence decreasing the COo-tension
in the water.
ACKNOWLEDGMENT
Financial support from the Danish Natural Science Research Council and
The University of Aarhus is gratefully acknowledged.
REFERENCES
Basu, S. P., 1959. Active respiration of fish in relation to ambient
concentrations of oxygen and carbon dioxide. J. Fish. Res. Bd. Can.
-LO * JL. / J "* ^ .L^ .
Beamish, F W. H. 1964. Respiration of fishes with special emphasis on
standard oxygen consumption. Can. J. Zool. 42:847-856
Stewart, N. E. , D. L. Doudoroff, and P. Shumway. 1967. Influence of oxygen
concentration on the growth of juvenile largemouth bass. J. Fish Res
Board. Can. 24:475-494.
Dahlberg, M. L. , D, L. Shumway, and P. Doudoroff. 1968. Influence of
dissolved oxygen and carbon dioxide on swimming performance of
largemouth bass and coho salmon. J. Fish. Res. Bd. Can 25-49-70
Dejours, P. 1988. Respiration in water and air. In: Adaptations -
Regulations-Evolution. Elsevier, Amsterdam, New York, Oxford p 179
160
-------
Dimberg, K. 1988. High blood CC>2 levels in rainbow trout exposed to
hypercapnia in bicarbonate-rich fresh water--a methodological
verification. J. Exp. Biol. 134:463-466.
Heisler, N. 1984. Acid-base regulation in fishes. In: Fish Physiology, Vol
XA. W. S. Hoar and D. J. Randall (eds.). Academic Press, New York,
London, p. 315-401.
Saunders, R. L. 1962. The irrigation of gills in fishes. II: efficiency of
oxygen uptake in relation to respiratory flow, activity and
concentration of oxygen and carbon dioxide. Can. J. Zool. 40:817-862.
Root, R. W. 1931. The respiratory function of the blood of marine fishes.
Biol. Bull. 61:441-462.
161
-------
IMPACT OF ENVIRONMENTAL ACIDIFICATION
ON GILL FUNCTION IN FISH
by
Chris M. Wood1 and D. Gordon McDonald1
INTRODUCTION
Although environmental acidification resulting from industrial and
domestic S02 and NOX emissions has been documented most strongly in north-
eastern North America and northern Europe, it is now recognized as a global
problem. Zhao and Sun (1986) have reported that much of China to the south of
the Yangtze River is subject to "acid rain" due to a combination of intensive
combustion of high-sulphur coal and low atmospheric buffer capacity (i.e., low
ammonia and alkaline particulates). Soils are generally acidic, and some
major cities report annual rainfall pHs as low (-4.0) as those in the most
seriously affected areas of eastern Canada and Scandinavia.
Damaging effects on freshwater fisheries may occur anywhere in the world
where precipitation pH is persistently below 5.6 (the pH of distilled water in
equilibrium with atmospheric C02) and the buffer capacity of the watershed is
low. Such "softwater" regions are generally characterized by low calcium
carbonate alkalinity in soils, bedrock, sediments, and the water column
itself. The softwater pH range of environmental significance to fish is 4.0
to 6.0; pHs below 4.0 rarely; if ever, occur. Most fish populations are not
adversely affected at pHs above 6.0.
Within this range, the deleterious effects of acidity are complex, and
include trophic degradation, reproductive failure, early life stage mortality,
and lethal and sublethal toxicity to juvenile and adult fish (Howells 1984).
Over the past decade, our research program has concentrated on this latter
area, investigating the physiological mechanisms of acid toxicity. The
present paper provides an overview of this research, with emphasis on recent
findings in the area of combined acid and aluminum toxicity, sublethal
effects; and acclimation.
Department of Biology, McMaster University, Hamilton ON, Canada.
162
-------
PHYSIOLOGICAL RESPONSES TO PURE ACID
STRESS AND THE INFLUENCE OF CALCIUM
Figure 1 summarizes the major effects of acute exposure to pH = 4.0 -
4.5 on salmonids such as the rainbow trout (Salmo gairdneri). For detailed
reviews of this earlier work on the lethal effects of acid alone, the reader
is referred to Wood and McDonald (1982), McDonald (1983), and Wood (1988). At
this pH range, disturbances in respiratory gas exchange.do not occur, although
they may contribute to toxicity when metals are present (see below) or at more
acidic pHs where branchial epithelial swelling, delamination, and mucification
may occur. Nevertheless, there are multiple toxic actions, including ionic
dilution and acidification of the blood and extracellular fluids (ECF),
demineralization of bone, electrolyte losses from the intracellular fluids
(ICF), especially white muscle, and disturbances of renal function.
By far the most important toxic effects, and those ultimately
responsible for the death of the fish, however, are exerted on the external
surface of the gills. This is a compound action, comprising an inhibition of
active Na+ and Cl~ uptake through the mitochondrial-rich "chloride cells," and
perhaps more importantly, a stimulation of passive diffusive losses of Na ,
Cl", and other electrolytes. The latter is thought to result from an
increased permeability of the paracellular diffusion channels in the branchial
epithelium. The net result is a large loss of electrolytes from the blood
plasma, and secondarily from the tissues. Osmotic pressure falls faster in
the ECF than in the ICF, and in compensation, water shifts from plasma into
cells, especially muscle and erythrocytes. Hematocrit, plasma protein
concentration, and blood viscosity all increase markedly as the red blood
cells swell and the plasma volume contracts. These effects are compounded by
adrenergically mediated vasoconstriction of systemic resistance vessels,
cardioacceleration, and mobilization of additional red cells from the spleen.
In the end, greatly elevated arterial blood pressures, hematocrits as high as
70%, and plasma volumes as low as half normal values result in a circulatory
failure ultimately attributable to the branchial ion loss.
In laboratory tests, elevations in water calcium protect against these
ion losses, and therefore against acid toxicity, an observation that is
confirmed by a wealth of field data (e.g., Wright and Snekvik 1978, Howells et
al. 1983). The mechanism of this protection is thought to be the well known
effect of Ca2+ in limiting membrane permeability, specifically in "plugging"
the branchial paracellular diffusion channels that are permeabilized by H+.
Those natural softwaters most sensitive to acidification due to their low
alkalinity, of course, are also low in calcium (generally in the range 25 -
400 uequiv/L =0.5 - 8 mg/L), which exacerbates the problem for the fish.
(Waters of higher calcium content are rarely acidified, except in cases of
mine acid drainage.) The ameliorative effects of "liming," therefore, are due
to the increases in water calcium, as well as in pH and alkalinity.
A point of considerable confusion is the often-stated view that
increased water calcium also should protect the fish against H+ entry and
internal acidosis. This is clearly untrue, as illustrated by the data in
Figure 2. Indeed at a given acidic pH, the higher the water calcium level,
the greater is the depression in arterial pH and metabolic acidosis, whereas
163
-------
m
iS-o
X * X +
+ + + CM
+ + + O 'to !_
X X O 2 O
164
-------
the smaller Is the plasma Na+ and Cl" deficit. This pattern has been +
confirmed by the direct measurement of ion and "acidic equivalent" (net H )
fluxes across the gills.
The explanation for this apparent dichotomy lies in simple physical
principles--the laws of electrical neutrality and the "strong ion difference"
relationships (Stewart 1983). Stated very simply, the flux of strong cations
(mainly Na+ plus K+) minus the flux of strong anions (mainly Cl") across a
biological membrane such as the gill will automatically constrain an equal
"flux" of acidic equivalents in the opposite direction. Water is an infinite
source or sink of acidic equivalents; net H"1" flux is not an independent
variable, but rather one dependent upon differential strong cation and anion
flux. Any resultant acid-base disturbance in the fish, therefore, is an
unavoidable consequence of ionoregulatory disturbance.
Higher water calcium levels protect against strong electrolyte losses at
low pH, but are more effective in limiting Cl" efflux than Na+ or K+ efflux,
probably because the diffusion channels become cation selective. At lower
calcium levels representative of natural softwaters, overall electrolyte
permeability increases, but anion permeability catches up to and may
eventually exceed cation permeability. As a result, there is negligible net
H+ uptake, or even a slight net H1" efflux across the gills. Internal acid-
base status is unchanged or may even become slightly alkalotic (cf. Figure 2).
In summary, under pure acid stress, at environmentally realistic levels
of pH (4.0 - 6.0) and calcium (25 - 400 uequiv/L), toxicity is due to
ionoregulatory disturbance, not to acidosis, and not to respiratory failure.
THE IMPORTANCE OF ALUMINUM
Unfortunately, pure acid stress itself may not be an environmentally
realistic situation. An increasing body of evidence argues for the
involvement of aluminum in acid toxicity in the field (e.g. Schofield and
Trojnar 1980, Harvey and McArdle 1986, Johnson et al. 1987). Aluminum is the
third most abundant element in the earth's crust, occurring ubiquitously in
rocks, soil, and sediments. Its solubility increases exponentially as pH
falls below about 5.6 (Figure 3); consequently, it is readily mobilized into
acidified softwaters (Driscoll 1980, Lazerte 1984). This mobilization has
been repeatedly suggested as the "missing link" explaining why mortality in
the field is often much greater than that in the laboratory at the same pH and
calcium levels, although this point remains controversial (e.g. Muniz and
Leivestad 1980, Schofield and Trojnar 1980, Howells et al. 1983, Howells 1984,
Neville 1985, Schindler 1988).
Much of the uncertainty surrounding the role of aluminum devolves from
its very complex aqueous chemistry. A highly simplified speciation model is
shown in Figure'3, which deliberately omits the fluoride, sulphate, carbonate,
and organic anion binding reactions, all of which are thought to lessen
aluminum toxicity, and the various polymerization reactions that may occur
under super-saturated conditions. The equilibria plotted are based on the
dissolution of microcrystalline gibbsite, which is probably more representa-
165
-------
+0.2
0
s -°-2
E
-0.4
S 4'°
+8.0
_i ~10
CT
UJ
E -20
-30
0
_j ~10
CT
UJ
E -20
-30
ApH
AH
m
x"" API
asma
,. _,
APIasma [crj
if'
Water
(mEq/L)
Figure 2. The relationship between water calcium concentra-
tion and the extent of various acid-base and ionic distur-
bances in the arterial blood of rainbow trout exposed to
pH 4.3 for 3 days. (A) Changes in arterial pH. (B) Changes
in blood metabolic acid load. (C) Changes in plasma sodium
concentration. (D) Changes in plasma chloride concentration.
Means ±1 SEM (n = 4 - 24) . (from Wood 1988)
166
-------
tive of field conditions (at least in eastern Canada, Lazerte 1984). A total
solubility curve based on amorphous aluminum hydroxide also is presented. The
latter is representative of freshly prepared solutions used in the laboratory.
Figure 3 illustrates that the speciation, as well as the solubility, of
inorganic monomeric aluminum is critically dependent upon pH. As pH falls,
the non-toxic anion A1(OH)A" is sequentially replaced by the Al(OH)2+,
Al(OH)2"1", and finally A13+ cations. All these cationic forms are thought to be
toxic to fish, although A13+ could also be protective through its ability to
mimic the action of Ca2+ (Baker and Schof ield 1982) . Much of the previous
work on the physiological responses of fish to aluminum has been done in
recirculating systems, often without precise control of pH or water chemistry,
or realistic water calcium levels. The resulting divergence in conclusions
may well reflect changing solubility, speciation, organic and inorganic
complexation, and calcium/aluminum interactions.
106-K
Aluminum
10
TOTAL Al (amorphous AlDH)3)
x
TOTAL Al (gibbsite)
4-0 4-2 4-4 4-6 4-8 5-0 5-2 5-4 5-6 5-8 6-0
PH
Figure 3. Speciation chemistry and solubility of aluminum (omitting complexation
reactions) as a function of pH in low ionic strength waters. The Al/pH combin-
ations tested in the present study are noted.
167
-------
PHYSIOLOGICAL RESPONSES TO ACID/ALUMINUM/CALCIUM COMBINATIONS
The goal of our own recent work has been to study the physiological
responses of juvenile and adult salmonids to various environmentally
representative combinations of pH (4.4, 4.8, 5.2), calcium (25, 400 uequiv/L),
and aluminum (0, 111, 333, 1000 ug/L) under defined artificial softwater
conditions (reverse osmosis water, 50 uequiv/L Na+, 0 DOC and fluoride, pH set
with H2S04) . The fish are previously acclimated to the artificial softwater
at pH - 6.5, Al = 0 ug/L for at least 2 weeks.
Our approach has been to "set" the water chemistry, using freshly
prepared solutions, and then to move the water past the fish as quickly as
possible to minimize pH shifts, speciation changes, and organic complexation
reactions (1/2 replacement time ~20 min. in the experimental chambers). Note
that the Al/pH combinations tested (plotted on Figure 3) are well below the
solubility limits for amorphous aluminum hydroxide (and except in one case for
gibbsite also), so supersaturated conditions are avoided, at least in the bulk
water. The brook trout (Salvelinus fontinalis) has been studied as it is the
salmonid endemic to the acid-threatened regions of North America, and the
rainbow trout as it is a widely distributed reference species. For more
detailed information on the brook trout studies, the reader is referred to
Booth et al. (1988), Wood et al. (1988a,b,c), McDonald and Milligan (1988),
and Walker et al. (1988), and on the rainbow trout studies to Goss and Wood
(1988) and Playle et al. (1989).
Upon exposure to acidic pH alone in the absence of Al, brook trout
exhibit transient and sublethal net losses of Na+ and Cl" across the gills in
proportion to the severity of the pH (Figure 4); zero balance is re-
established in 12 to 72 hours. The addition of Al to the exposures greatly
exacerbates these losses and kills many of the fish within 24 to 72 hours
(Figure 4). Terminal blood samples reflect these losses, with severely
depressed ECF ions and dramatic hemoconcentration in lethal exposures, as
previously seen in rainbow trout dying from pure acid stress. The cumulative
ion losses prior to death in acid/Al exposures are not as large, however,
suggesting that an additional toxic mechanism(s) may be involved (see below).
In some cases, the brook trout do not die in the presence of Al, but rather
return to zero or even positive balance and survive indefinitely.
Measurements of unidirectional fluxes across the gills with 22Na
demonstrate that initial losses reflect a combination of inhibited active
uptake and stimulated diffusive efflux; the latter is quantitatively much more
important. This constitutes an initial "shock" phase (0 - 12 hours) of heavy
losses that may eventually kill the fish by inducing intolerable fluid volume
and circulatory disturbances. The extent of cumulative ion loss during this
period is a reliable predictor of eventual survival or death, even though that
death may not occur for several more days. In trout that survive, there
follows a "recovery" phase in which the efflux component is returned to
control levels or below by 24 to 72 hours. Active influx remains depressed,
however, for up to 10 days. The initial large efflux is thought to result
from opening of the paracellular diffusion channels in the gills, and the
later adaptive response from an effective closing of these channels, as
discussed subsequently.
168
-------
By 48 hours, the majority of the electrolyte loss that will occur has
already done so. Cumulative 48-hour Na+ loss and mortality are strongly
correlated (Figure 4). These responses increase with Al concentration at any
given acidic pH, and increase with pH at Al = 333 ug/L. Thus Al toxicity
appears to be greatest at pH = 5.2, where total solubility is lowest and
A1(OH)2+ and A1(OH)2+ predominate (Figure 3).
The higher water calcium level (400 uequiv/L, still in the softwater
range) generally reduces the extent of electrolyte loss at any particular
pH/Al combination and removes the clear relationship between cumulative ion
18H NA+ LOSS- (1000 UEQ/KG)
Ca2'l"=25ueq/l
I
PH
Figure 4. Cumulative net sodium losses of brook trout over the
first 48 hours of exposure to various combinations of pH and
aluminum in flowing softwater (Ca = 25 uequiv/L). Means ±1
SEM (n = 6 in each treatment). Percentage mortalities at the
end of 10 days of exposure are indicated for each treatment.
(modified from McDonald et al. 1988, Booth et al. 1988)
169
-------
loss and mortality seen at lower calcium (25 uequiv/L; Figure 4). Neverthe-
less, a substantial number of fish still die, even though their ion losses and
plasma Na+ and Cl~ depressions are quite small. Under these conditions,
respiratory failure, the second toxic mechanism alluded to previously, becomes
the primary cause of death. This interference with 02 uptake and C02 excre-
tion across the gills is signalled by a pronounced hyperventilation accom-
panied by decreases in arterial blood 02 tension (PaO2; Figure 5C) and 02
content below typical venous values, and reciprocal increases in C02 tension
and blood lactate. Although this phenomenon is most marked at the higher
calcium level (Figure 5G) and higher pHs, it occurs prior to death in all
treatments where Al is present (Figure 5B,E).
Although the responses of the rainbow trout are qualitatively similar to
those of the brook trout, there are two important differences. Firstly, under
any particular acid/Al condition, rainbows are far less tolerant, showing
greater ion loss (Figure 6B), hemoconcentration (Figure 6C), and respiratory
disturbance (Figure 6D), and earlier and larger percentage mortalities. These
physiological results are in accord with field and toxicological data indicat-
ing that Salmo gairderi is the most sensitive, and Salvelinus fontinalis the
most resistant, of a range of salmonids (Grande et al. 1978). Secondly, the
contribution of respiratory disturbance to toxicity is greater in the rainbow
trout under all conditions where Al is present, not just at the higher water
calcium level. Indeed, unlike the brook trout, there is no evidence that
respiratory disturbance is exacerbated by higher water calcium, although it is
worsened by higher pH. These differences emphasize that generalization
between species, even within the same Family, must proceed with caution.
BRANCHIAL MECHANISMS OF ACTION
FOR ALUMINUM, ACID, AND CALCIUM
The preceding observations can be summarized by saying that acidity and
aluminum both cause ionoregulatory toxicity, that calcium protects against
ionoregulatory toxicity, that aluminum alone also causes respiratory toxicity,
which may or may not be exacerbated by higher calcium, and that the extent of
aluminum's ionoregulatory and respiratory effects appears to be greater at
higher acidic pHs. Clearly the situation is complex, and any explanation(s)
must take into account the chemistry of aluminum, the response(s) of the gill
epithelium, and the nature of the interaction between the two. The final
point is the critical one, about which we know very little.at present. The
ideas that follow are, therefore, highly speculative.
An important starting point is the observation that accumulation of
aluminum on the gills (Figure 6A) appears to be directly associated with
toxicity, and correlates well with physiological indices of disturbance
(Figure 6B,C,D). Gill aluminum levels increase in a time- and concentration-
dependent fashion, are greater at higher pH for any particular water Al level,
are greatest in dying fish, and are clearly related to differences in species
sensitivity (Figure 6A). We view this as a superficial uptake onto or into
the gill cells, for neither we nor Neville (1985) can find any evidence of
aluminum entry into the fish in exposures lasting up to 10 days.
170
-------
120
80
40
0"
80
40
0
80
40
0
80
40
0
80
40
pH=4.8, Ca=25 uequiv/L, Al=0 ug/L
pH=4.8, Ca=25 uequiv/L. Al=333 ug/L
pH=4.8,
Ca=400 uequiv/L, Al=333 ug/L *
PH = 4.4. Ca = 25 uequiv/L, Al=0 ug/L
pH = 4.4, Ca=25 uequiv/L. Al=333 ug/L
C 0
10
20 30
Time(h)
40
Figure 5. Changes in the Q£ tension of arterial blood (PaQ2) in
chronically cannulated brook trout during 10 days exposure to five
different pH/Ca/Al conditions in flowing softwater. Means ± SEM
(n = 9 - 28 in each treatment). Data are shown from the control
period until 48 hours for all fish that survived beyond this time in
each exposure (i.e., the most resistant fish in each group). Addi-
tionally, terminal data (T, cross-hatched bars) representing the last
measurements prior to death in fish dying at any time during the
entire 10 days exposure are shown in bar graphs at right and compar-
ed with initial measurements (I, open bars) during the control
period for these same fish. In panel A, none of the fish died, so
data at 10 days exposure have been substituted for terminal data.
Asterisks indicate significant differences (p=<0.05) from appropriate
control or initial values, (from Wood et al. 1988a)
171
-------
Why should aluminum accumulate in the first place? The branchial
surface is rich in organic anions' on mucus and cell surfaces, and these anions
may avidly bind the cationic Al species that occur at low pH (Figure 3).
Analysis of mucus fragments extruded by the fish or gently scraped from the
gills of dead fish show very high aluminum contents. Furthermore, at bulk
water pHs below about 5.5, the branchial micro-environment is much more
alkaline than the inspired water (Figure 7A) probably due mainly to the efflux
of NH3 across the gills. As inflowing water encounters this more basic
milieu, the formation of aluminum hydroxides will be favored (Figure 3), and
perhaps more importantly, supersaturating conditions may occur, resulting in
direct precipitation of polymers on the gills. The exponential relationship
between pH and Al solubility is critical here (Figure 3).
Figure 7B illustrates that an aluminum concentration (111 ug/L) that
would appear to be well below saturation at a bulk water pH of 4.8, is
effectively right on the solubility limit, because the pH at the gills is
raised to about 5.2 (Figure 7A). In turn, at a bulk water pH of 5.2, where
this concentration appears to be at the solubility limit, it is in fact
several-fold saturated (Figure 7B), because the pH is raised to about 5.5 at
the gills (Figure 7A). This may explain why aluminum is more toxic at higher
acidic pHs, although it is also possible that the less charged species are
intrinsically more toxic; the two explanations are certainly not mutually
exclusive.
Figure 8 presents a simple model of the branchial epithelium incor-
porating some of the ideas presented in this paper and recent histological
work (Karlsson-Norrgren et al. 1986, Tietge et al. 1988, M.E. Mueller, pers.
comm.). H+ by itself (Figure 8B) inhibits the active Na+ and Cl" uptake
processes through the chloride cells and, more importantly, increases their
diffusive loss by opening up the tight junctions of the paracellular channels,
perhaps by displacing bound Ca2+; respiratory gas exchange is not affected.
If aluminum also is present (Figure 8B), there is a precipitation of Al com-
plexes and/or binding of Al to organic anions on the gill surface. In turn,
this acts as an irritant that stimulates the production of additional mucus
and induces an inflammatory response comprising edema, white cell invasion,
lamellar clubbing and even fusion. A general thickening and distortion of the
branchial epithelium results, which together with the mucus/Al layer, in-
creases the transcellular diffusion distance from water to blood, thereby
reducing 02 and C02 exchange. At the same time, this distortion further opens
the paracellular channels, increasing diffusive electrolyte loss, while the
surface coating or Al cations themselves further inhibit active Na+ and Cl"
uptake. Recovery, if it occurs, presumably reflects a reduction in the
inflammatory response, a closing of the paracellular channels, and a reduction
in the transcellular diffusion distance for 02 and C02. Higher water calcium
levels (Figure 8D) also help close the paracellular channels, thereby reducing
ion efflux. In the brook trout, where calcium intensifies the gas exchange
problem, we speculate that it acts in some way to also thicken the mucus/Al
coat.
ACCLIMATION TO ACID/ALUMINUM STRESS
It was noted earlier that some individual trout may recover from an
acid/Al challenge in the continued presence of the stressors, which raises the
172
-------
Gill Aluminum
Plasma Na+
Hematocrit
70 -,
Plasma Lactate
10
20 30
Time (hours)
Figure 6. A comparison of the responses in (A) gill alumi-
num accumulation, (B) plasma sodium concentration, (C)
hematocrit, and (D) blood lactate between•juvenile brook
trout (BT) and juvenile rainbow trout (RBT) exposed for
48 hours to pH = 4.8, .Al = 111 ug/L, at Ca = 25 uequiv/L.
Means ±1 SEM (n = 10 - 20 at each point). (D.G. McDonald
and C.M. Wood, unpublished results)
173
-------
+0.6
+0.4
+0.2
X 0.0
<
°6>
-0.2
-0.4
-0.6
4.0
4.5
V5.5
6.0
5.0
6.5
—t—
-PHin-
V--,
10000
1000
D)
100
10
\\
4.0
Al solubility
in bulk water
at gill A
t i i i I i .
I I I I.
4.5
5.0
5.5
6.0
6.5
PH
in
Figure 7. (A) The relationship between inspired bulk water PH and
gill water PH in the rainbow trout in softwater. Each point re-
presents an individual measurement. (B) The effect of this shift
in water PH at the gills on the solubility of aluminum. See text
detallS- (R-°- Playle and C'M- Wood,. unpublished
174
-------
possibility that acclimation may occur. This is further suggested by field
survey data (Schofield and Trojnar 1980, Kelso et al. 1986) from chronically
acidified lakes, indicating that salmonids may survive in the field at
aluminum levels known to be lethal in laboratory trials. Of course an ^
alternate explanation is that of genetic differences and natural selection.
Acclimation is defined as increased resistance to a more severe exposure
acquired as a result of a previous sublethal exposure. Therefore, we have
evaluated the physiological responses to long-term sublethal acid/Al exposures
and subsequent more severe challenges. For more detailed information on these
acclimation studies, the reader is referred to Wood et al. (1988b,c), McDonald
and Milligan (1988), Audet et al. (1988), and Audet and Wood (1988).
B
Control
H +
Water
-------
There is no evidence that salmonids have any ability to acclimate to
pure acid. Rainbow trout exposed to sublethal low pH (4.8) in flowing
softwater show a gradual loss of electrolytes and elevation of plasma glucose
and cortisol levels, both of which are sensitive stress indicators. Although
a plateau is reached after 4 to 7 weeks in most parameters, there is no
recovery back to control levels. When these fish are challenged with a more
severe exposure (pH =4.0) they exhibit further ion losses and other
physiological disturbances that are greater than those seen in previously
unexposed trout. Thus sensitization, rather than acclimation, occurs.
The situation is very different for combined acid plus aluminum
exposure. Brook trout held for 10 weeks in flowing softwater at pH = 5.2 plus
75 or 150 ug/L Al actually appear less stressed than those held in pH - 5.2
alone, based on plasma electrolytes, glucose, and cortisol, and rates of
branchial sodium uptake. The mechanism is unknown but may be associated with
a proliferation of branchial chloride and mucous cells seen in fish chroni-
cally exposed to low levels of aluminum (Tietge et al. 1988). When these fish
are subjected to a more severe combined acid plus aluminum challenge (3 days
at pH - 4.8, Al •= 333 ug/L, which is lethal to the majority of previously
unexposed trout), they exhibit 100% survival, and much smaller physiological
disturbance. This acclimation appears to protect against both toxic mecha-
nisms (ionoregulatory and respiratory distrubance), and is reflected in stress
indicators (glucose, cortisol), respiratory parameters (ventilation, blood
gases, lactate) and ionoregulatory parameters.
For example, Figure 9 illustrates that the cumulative Na+ loss over the
first 48 hours of challenge (pH - 4.8, Al - 333 ug/L) is approximately halved
by pre-exposure to low levels of aluminum at pH = 5.2, and the depressions in
plasma Na and Cl" correspondingly reduced or abolished. Note that the Na+
loss is only reduced to the level that is caused by challenge with acid alone
(pH - 4.8) and that pre-exposure to acid alone (pH =5.2) provides no protec-
tion against acid plus aluminum challenge. In agreement with the rainbow
trout experiments, this finding indicates that the acclimation is not to acid
but rather to aluminum. '
What is/are the mechanism(s) of acclimation to aluminum? With other
metals, acclimation has been traditionally associated with internal processes
such as the induction of metal-binding proteins (e.g. metallothioneins) or the
activation of excretion mechanisms. Aluminum appears to be entirely a
surface-active toxicant, however, so these explanations appear unlikely.
As a first step in attacking this problem, we recently have followed a
variety of physiological and histological parameters over time in juvenile
brook trout exposed to sublethal aluminum. The goal was to see which para-
meters, if any, change as acclimation develops. After an initial sensitiza-
tion, a significant and sustained increase in resistance (i.e. relative LT50)
to more severe aluminum challenge occurs between days 10 and 13 of exposure to
low levels of Al (75 or 150 ug/L) at PH = 5.2 (Figure 10A). Again exposure to
acid alone (pH =5.2) is relatively ineffective in inducing acclimation. Most
internal parameters show little or no correlation with this time course, but
there is a striking relationship between gill aluminum burden and the develop-
ment of acclimation (Figure 10B). Despite continued low level exposure to
176
-------
-6000
O)
3
cr
0
3
-4000
-2000
-40
I -20
03
E
0
Cumulative 48 H Na+ Loss
pH = 6.5
AI = O
pH=6.5
Al = 0
pH=5.2
Al = 0
pH = 5.2
Al = 150
pH=5.2
Al = 75
CU APIasmafNa"*]
^ A Plasma
pH = 4.8
i 1
Al = 0
pH = 4.8
Al = 333
Challenge
Figure 9. The effects of 10 weeks of sublethal exposure to condi-
tions shown (in flowing softwater, Ca = 25 uequiv/L) on the
responses of brook trout to subsequent challenge with a more
severe low pH alone (4.8) or low pH plus Al (333 ug/L). (A)
Cumulative net sodium losses over 48 hours of challenge.' (B)
Changes in plasma sodium and chloride concentrations over 48
hours of challenge. Means ±1 SEM. (n = 6 in each group), (modified
from Wood et al. 1988b)
177
-------
the metal, the gill is in some way able to "clean" itself of bound or preci-
pitated aluminum compounds.
A pronounced hypertrophy of branchial mucous cells occurs, so this may
be associated with increased rates of mucus synthesis and sloughing as the
cleansing mechanism. As the gill aluminum burden is reduced, so may be the
inflammatory response, the paracellular permeability, and the diffusion
distance for respiratory gases. Whatever the explanation, the nature of the
gill surface appears to change so as to become less reactive to higher levels
of waterborne aluminum during challenge.
CONCLUSION
The unifying theme of the research summarized here has been that the
toxic actions of environmental acidity and associated aluminum stress, as well
as the protective effects of calcium and the adaptive responses of recovery
and acclimation, can all be interpreted at the level of gill epithelium.
Nevertheless, our understanding of the chemical, physical, and morphological
characteristics of the branchial surface remains rudimentary. A much finer
characterization of this complex and delicate epithelium, together with its
aqueous micro-environment, is required.
ACKNOWLEDGMENTS
The work reported here was supported by grants from the NSERC (Canada)
Strategic Program in Environmental Toxicology, and a contract ("Lake
Acidification and Fisheries", RP-2346-01; Dr. J. Mattice, project manager)
from the Electric Power Research Institute, Environmental Assessment Dept.,
Palo Alto, CA (U.S.A.), through a subcontract from the University of Wyoming.
We thank Dr. H.L. Bergman and the staff of the Red Buttes Fish Physiology and
Toxicology Laboratory, University of Wyoming, for extensive collaboration, and
R.C. Playle, R. Rhem, M.E. Mueller, and D.R. Mount for permission to cite
previously unpublished data.
REFERENCES
Audet, C., R.S. Munger, and C. M. Wood. 1988. Long-term sublethal acid
exposure in'rainbow trout (Salmo gairdneri) in soft water: Effects on
ion exchanges and blood chemistry. Can. J. Fish. Aquat Sci 45-1387-
1398.
Audet, C. and C. M. Wood. 1988. Do rainbow trout acclimate to low pH?
Can. J. Fish. Aquat. Sci. 45:1399-1405.
Baker, J.P. and C.L. Schofield. 1982. Aluminum toxicity to fish in acidic
waters. Water, Air, and Soil Pollut. 18:289-310.
Booth, C.E., D.G. McDonald, B.P. Simons, and C.M. Wood. 1988. The effects
of aluminum and low pH on net ion fluxes and ion balance in the brook
trout, Salvelinus fontinalis. Can. J. Fish. Aquat. Sci. 45:1563-1574.
178
-------
o
p
0)
tr
2-0 -i
1.6-
1.2-
0.8-
0.4-
0.0
pH = 5.2. Al = 75 ug/L
LT50
10 15
Time (Days)
20
25
0)
•
400 -i
300-
200-
100-
Gili Al
5.2, A! = 150 ug/L
pH = 5.2, Al = 75 ug/L
Control
-I---J/0- -J -5
• f m. *— * • * •
c
0
5
10
15
i ,
20
' 1
25
Time (Days)
Figure 10. (A) The effects of sublethal exposure to pH = 6.5
alone (control), pH = 5.2 alone, and pH = 5.2 plus Al = 75 ug/L
or 150 ug/L ( in flowing softwater, Ca = 25 uequiv/L) on the
mean lethal times (LT50, ±95% CL) of juvenile brook trout when
challenged with Al = 1000 ug/L. The exposures started on Day 1.
The LT50 data have been expressed as ratios to the control.
Asterisks indicate means significantly different (p<0.05) from
control. (B) Changes in gill aluminum content during these expo-
sures. Means ±1 SEM (n = 10 at each point). (D.G. McDonald,
C.H. Wood, R.G. Rhem, M.E. Mueller, D.R. Mount, and H.L. Berg-
man, unpublished results)
179
-------
Driscoll, C. 1980. Chemical Characterization of Some Dilute Acidified
Lakes and Streams in the Adirondack Region of New York State. Ph.D.
Thesis, Cornell University, Ithaca, NY. 309 pp.
Goss, G.G. and C.M. Wood. 1988. The effects of acid and acid/aluminum
exposure on circulating plasma cortisol levels and other blood
parameters in the rainbow trout (Salmo gairdneri") J. Fish. Biol 32*63-
76.
Grande, M., I.P. Muniz, and S. Anderson. 1978. The relative tolerance of
some salmonids to acid waters. Verh. Internat. Verein. Limnol. Biol
20:2076-2084.
Harvey, H.H. and J.M. McArdle. 1986. Physiological responses of rainbow
trout Salmo gairdneri exposed to Plastic Lake inlet and outlet stream
waters. Water, Air, and Soil Pollution 30:687-694.
Howells, G.D. 1984. Fishery decline: mechanisms and predictions. Phil.
Trans. R. Soc. Lond. 6305:529-547.
Howells, G.D., D.J.A. Brown, and K. Sadler. 1983. Effects of acidity,
calcium and aluminium on fish survival and productivity - a review. J.
Sci. Food. Agric. 34:559-570.
Johnson, D.W., H.A. Simonin, J.R. Colquhoun, and F.M. Flack. 1987. In
situ toxicity tests of fishes in acid waters. Biogeochemistry 3:181-
208.
Karlsson-Norrgren, L., I. Bjorklund, 0. Ljungberg, and P. Runn. 1986.
Acid water and aluminum exposure: experimentally induced gill lesions
in brown trout, Salmo trutta L. J. Fish. Diseases 9:11-25.
Kelso, J.R.M. , C.K. Minns, J.E. Grey, and M.L. Jones. 1986. Acidification
of surface waters in eastern Canada and its relationship to aquatic
biota. Can. Spec. Publ. Fish. Aquat. Sci. 87. 42 p.
LaZerte, B.D. 1984. Forms of aqueous aluminum in acidified catchments of
central Ontario: A methodologial analysis. Can. J. Fish. Aquat. Sci.
41:766-776.
McDonald, D.G. 1983. The effects of H+ upon the gills of freshwater fish.
Can. J. Zool. 61:691-703.
McDonald, D.G., and C.L. Milligan. 1988. Sodium transport in the brook
trout, Salvelinus fontinalis: The effects of prolonged low pH exposure
in the presence and absence of aluminum. Can. J. Fish. Aquat Sci
45:1606-1613.
McDonald, D.G., J.P. Reader, and T.K.R. Dalziel. 1988. The combined
effects of pH and trace metals on fish ionoregulation. In: Acid
Toxicity and Aquatic Animals, Society for Experimental Biology Seminar
Series. R. Morris, D.J.A. Brown, E.W. Taylor, and J.A. Brown (eds.).
Cambridge University Press.
180
-------
Muniz I P and H. Leivestad. 1980. Acidification-effects on freshwater
'fish. In: Ecological Impact of Acid Precipitation. D. Drablos and A.
Tollan (eds.). pp. 84-92, SNSF-project.
Neville C M 1985. Physiological response of juvenile rainbow trout,
Salmo gairdneri. to acid and aluminum - prediction of field responses
from laboratory data. Can. J. Fish. Aquat. Sci. 42:2004-2019.
Playle R C G G Goss, and C.M. Wood. 1989. Physiological disturbances
in rainbow trout (Salmo gairdneri) during acid and aluminum exposures in
soft water of two calcium concentrations. Can. J. Zool. In Press.
Schindler, D.¥. 1988. Effects of acid rain on freshwater ecosystems.
Science 239:149-157.
Schofield, C.L., and R.J. Trojnar. 1980. Aluminum toxicity to brook
trout (Salvelinus fontinalis) in acidified waters. In: Polluted Rain.
T.Y. Toribara, M.W. Miller, and P.E. Morrow (eds.). Plenum Press, New
York, NY. pp. 341-366.
Stewart, P.A. 1983. Modern quantitative acid-base chemistry. Can. J.
Physiol. Pharmacol. 61:1444-1461.
Tietge J E R D. Johnson, and H.L. Bergman.. 1988. Morphemetrie changes
in gill secondary lamellae of brook trout (Salvelinus fontinalis)_after
long-term exposure to acid and aluminum. Can. J. Fish. Aquat. Sci.
45:1643-1648.
Walker R.L., C.M. Wood, and H.L. Bergman. 1988. Effects of low pH and
aluminum on ventilation in the brook trout (Salvelinus fontinalis).
Can. J. Fish. Aquat. Sci. 45:1614-1622.
Wood, C.M. 1988. The physiological problems of fish in acid waters. In:
Acid Toxicity and Aquatic Animals, Society for Experimental Biology
Seminar Series. R. Morris, D.J.A. Brown, E.W. Taylor, and J.A. Brown
(eds.). Cambridge University Press.
Wood C.M., and D.G. McDonald. 1982. Physiological mechanisms of acid
toxicity to fish. In: Acid Rain/Fisheries. R.E. Johnson (ed.).
American Fisheries Society, Bethesda, MD. pp. 197-226.
Wood CM and D.G. McDonald. 1987. The physiology of acid/aluminum
' stress in trout. In: Ecophysiology of Acid Stress in Aquatic Organisms.
H. Witters and 0'. Vanderborght (eds.). Annls. Soc. r. Zool. Belg. 117
(suppl. 1):399-410.
Wood C.M., R.C. Playle, B.P. Simons, G.G. Goss, and D.G. McDonald
1988a Blood gases, acid-base status, ions, and hematology in adult
brook trout (Salvelinus fontinalis) under acid/aluminum exposure. Can.
J. Fish. Aquat. Sci. 45:1575-1586.
181
-------
Wood, C.M., D.G. McDonald, C.E. Booth, B.P. Simons, C.G. Ingersoll, and H.L.
Bergman. 1988b). Physiological evidence of acclimation to
acid/aluminum stress in adult brook trout (Salvelinus fontinalis). 1.
Blood composition and net sodium fluxes. Can. J. Fish. Aquat. Sci.
45:1587-1596.
Wood, C.M. , B.P. Simons, D.R. Mount, and H.L. Bergman. 1988c. Physiological
evidence of acclimation to acid/aluminum stress in adult brook trout
(Salvelinus fontinalis). 2. Blood parameters by cannulation. Can. J.
Fish. Aquat. Sci. 45:1597-1605.
Wright, R.F., and E. Snekvik. 1978. Acid precipitation: Chemistry and
fish populations in 700 lakes in southernmost Norway. Veh. Internat.
Verein. Limnol. Biol. 20:765-775.
Zhao, D. and B. Sun.
15:2-5.
1986. Air pollution and acid rain in China. Ambio
182
-------
AMMONIA TOXIGITY TO FISHES
Robert V.: Thurston1
INTRODUCTION
Ammonia is a naturally occurring product of metabolic degradation and, as
a consequence, is commonly found in most aquatic environments. It also is
frequently added to aquatic systems as a waste product from industrial,
municipal, and agricultural processes. Ammonia is extremely toxic to fishes
and other aquatic animals, and this toxicity is affected by several factors
including water pH, dissolved oxygen content, temperature, concentration
fluctuations, degree of salinity, presence of other chemicals, and prior
acclimation. Short-term exposure of fishes to high concentrations of ammonia
causes increased gill ventilation, hyperexcitability, loss of equilibrium,
convulsions, and then death (Smart 1978, Thurston et al. 1981c) . These
effects are most likely the result of a direct effect of ammonia on the
central nervous system. Chronic exposure of fishes to lesser concentrations
of ammonia include damage to tissues, decreases in reproductive capacity (egg
production, egg viability, spawning delay) , decreases in growth, and increases
in susceptibility to disease (Thurston et al. 1984, 1986). Chronic exposure
also may cause progressive deterioration of other physiological functions, any
one of which may be the ultimate cause of death (Randall and Wright 1987) .
AQUEOUS AMMONIA EQUILIBRIUM
Ammonia assumes two chemical forms in aqueous solution, shown by the
following equilibrium equation.
NH3 + nH20 = NH3'nH20 = NH4+ + OH" + (n-l)H20.
These forms are the un- ionized ammonia species (NH3) , hydrogen-bonded to at
least three (n > 3) water molecules (Butler 1964) , and the ionized species
(NH4+) . The toxicity of ammonia is generally attributed to NH3, which is an
extremely soluble gas. Total ammonia is the sum of NH3 and NHA*, and it is
total ammonia that is most commonly measured in aqueous solutions; the
concentration of NH3 is then calculated. Tables of NH3 concentration, as a
function of total ammonia concentration, over a range of pH values from 5 to
12 and range of temperatures from 0 to 40 °C are available (Emerson et al.
1975, Thurston et al. 1979).
•'•Fisheries Bioassay Laboratory, Montana State University, Bozeman MT USA
183
-------
The relative concentrations of ionized and un-ionized ammonia in a given
solution are a function of the pH, temperature, and ionic strength of that
solution. ,As pH increases, the concentration of NH3 increases, while that of
NHA decreases. For example, a pH increase from 7.0 to 8.0 in the temperature
range of 0 to 30 °C results in a nearly 10-fold increase in the concentration
of NH3. Temperature increase also favors the NH3 species, but to a lesser
extent; a temperature increase of 5 degrees between 0 and 30 °C at pH 7.0
results in an NH3 concentration increase of 40 to 50%. An increase in the
ionic strength of a solution, at low concentrations, favors the NH4+ species.
In natural waters with low to moderate amounts of dissolved solids (200-1000
mg/L) this effect will slightly reduce the concentration of NH3, and the
magnitude of this effect will vary with the composition of the water (Thurston
et al. 1979).
RELATIONSHIP OF AMMONIA TO NITRITE AND NITRATE
Ammonia is interrelated with nitrite (NO;,0 and nitrate (N03") through the
process of nitrification, the biological oxidation of ammonia to nitrate.
Under aerobic conditions ammonia is readily oxidized to nitrite by Nitro-
somonas bacteria, and nitrite is then oxidized to nitrate by Nitrobacter
bacteria. In well-oxygenated aquatic systems, the conversion of ammonia to
nitrite is the rate-limiting step in the total process, and the conversion of
nitrite to nitrate occurs fairly rapidly. Nitrite concentrations in most
natural systems, therefore, are usually low, which is fortunate for aquatic
life because nitrite, also, is extremely toxic to fishes. Nitrate, although
generally present in all but the most oligotrophic surface waters, is rela-
tively non-toxic to fishes. Where it is present in high concentrations, any
problem is usually one of contributing to eutrophication rather than to
toxicity.
ACUTE TOXICITY OF AMMONIA
There is some variation in the susceptibility of fishes to the ammonia
toxicity. Representative 96-hour acute toxicity values (96-hour median lethal
concentrations, or LC50 values) are listed in Table 1. From available data,
salmonids^appear to be among the most sensitive to acute ammonia exposure, and
centrarchids, catfish, and some minnows the most tolerant. In addition to
differences in susceptibility among fish species, there also can be differ-
ences in susceptibility at different life stages of a given species (Thurston
and Russo 1983). Moreover, several factors can affect the toxicity of
ammonia, including environmental pH, dissolved oxygen, temperature, salinity
and ionic composition, previous acclimation and intermittency of exposure, and
presence of other toxicants.
EFFECT OF pH
Some years ago researchers observed that the toxicity of ammonia
solutions was greater at high pH values (Chipman 1934, Wuhrmann et al. 1947,
Wuhrmann and Woker 1948), and because the percentage of NH3 in total ammonia
increases as solution pH increases, these authors concluded that NH3 was the
toxic form of ammonia in the water. They also concluded that NH4+ was non-
toxic or appreciably less toxic. More recently, however, several researchers
have reported that NH3 is more toxic at low pH, separate from the effect of pH
184
-------
Table 1. Representative acute toxicity values for un-ionized ammonia.
(Values reported are ranges.)
Species
Pink salmon
(Oncorhynchus gorbuscha)
Mountain whitefish
(Prosopium williamsoni)
Brown trout
(Salmo trutta)
Rainbow trout
(Salmo gairdneri)
Largemouth bass
(Micropterus salmoides)
Smallmouth bass
(Micropterus dolomieui)
Common carp
(Cyprinus carpio)
Red shiner
(Notropis lutrensis)
Fathead minnow
(Pimephales promelas)
Channel catfish
(Ictalurus punctatus)
Bluegill
(Lepomis macrochirus)
9 6 -hour LC50
mg/L NH3)
0.08-0.1
0.14-0.47
0.50-0.70
0.16-1.1
0.9-1.4
0.69-1.8
2.2
2.8-3.2
0.75-3.4
0.50-3.8
0.55-3.0
Reference
Rice & Bailey (1980)
Thurston & Meyn (1984)
Thurston & Meyn (1984)
Broderious & Smith (1979)
Calamari et al. (1981)
Thurston & Russo (1983)
Roseboom & Richey (1977)
Broderius et al. (1985)
Hasan & Macintosh (1986)
Hazel et al. (1979)
Thurston et al. (1983)
Colt & Tchobanoglous (1976)
Roseboom & Richey (1977)
Arthur et al. (1987)
Emery & Welch (1969)
Roseboom & Richey (1977)
185
-------
on the NHs/NH/i"1" equilibrium in the water. This greater toxic effect has been
demonstrated with prawn larvae and Daphnia. as well as with fishes (rainbow
trout Salmo gairdneri. fathead minnow Pimephales promelas. coho salmon
Onchorhvnchus kisutch. chinook salmon 0. tshawytscha. channel catfish Ictal-
urus punctatus. and green sunfish Lepomis cyanellus) (Tabata 1962, Sousa et
al. 1974, Robinson-Wilson and Seim 1975, Armstrong et al. 1978, Hillaby and
Randall 1979, Tomasso et al. 1980, Thurston et al. 1981c, McCormick et al.
1984, Broderius et al. 1985, Sheehan and Lewis 1986). Results of one such
study (Thurston et al. 1981c) showed that there was a decrease in the 96-hour
LG50 value for NH3 for rainbow trout (Figure 1) and fathead minnows as pH
decreased, indicating that fishes were more susceptible to NH3 at low pH. In
longer exposures, Broderius et al. (1985) also reported a marked effect of pH
on the toxicity of NH3 in a study on smallmouth bass (Micropterus dolomieui)
from late embryo through 32 days of exposure. The 32-day "no-observed-effect"
1000 R
100
10
x
o>
o 10
10
cc
o
I
(D
01 1.0
O.I
TOTAL AMMONIA
UN-IONIZED AMMONIA
6.5 7.0
7.5 8.0
PH
8.5 9.0 9.5
Figure 1. Acute toxicityof ammonia (96-hour LC50
values) to rainbow trout at various environmental
pH values. Error bars are 95% confidence inter-
vals. (Reprinted with permission from Thurston
et al. 1981c; copyright 1981, American Chemical
Society)
186
-------
concentrations radically decreased from 0.61 to 0.044 mg/L NH3 as pH was low-
ered from 8.68 to 6.60. In a preliminary study, we have observed that the
toxicity ,of NH3 also increased at pH values above 9 (Figure 2) (Russo et al.
1988).
EFFECT OF DISSOLVED OXYGEN
Several researchers, working with a variety of fish species, have
reported that the toxicity of ammonia increases when dissolved oxygen con-
centrations decrease (Wuhrmann 1952, Wuhrmann and Woker 1953, Downing and
Merkens 1955, Merkens and Dpwning 1957, Alabaster et al. 1979, Thurston et al.
1981b). Discharges of ammonia frequently are associated with reduced oxygen
concentrations in the receiving water, so the effect of dissolved oxygen on
ammonia toxicity can be important.
EFFECT OF TEMPERATURE
Many researchers have reported that ammonia is less toxic to fishes at
temperatures near the higher end of their normal environmental range than near
the lower end, but not all studies support this. Colt and Tchobanoglous
(1976) reported a decrease in acute toxicity of ammonia to channel catfish
with an increase in temperature over the range 22 to 30 °C, and Roseboom and
Richey (1977) reported that bluegill (Lepomis macrochirus), largemouth bass
(Micropterus salmoides), and channel catfish were more susceptible to ammonia
toxicity at 22 °C than at 28 to 30 °C. Arthur et al. (1987) conducted 96-hour
ammonia toxicity tests on white sucker (Catostomus commersoni), walleye
(Stizostedion vitreum vitreum), rainbow trout, fathead minnow, and channel
catfish at temperatures ranging from 3.4 to 26 °C. Except for channel
catfish, none of the tests showed a progressive increase in LC50 with increas-
ing water temperature, and no clear relationship was found between ammonia
toxicity and temperature; the 96-hour LC50 values for channel catfish were
0.50, 0.98, and 1.29 mg/L NH3 at 3.5, 14.6, and 19.6 °C.
FLUCTUATING CONCENTRATIONS AND ACCLIMATION
Fish are frequently subjected to fluctuating concentrations of ammonia as
a consequence of diurnal changes. Also, episodic "slugs" of ammonia, as a
result of accidental spills or intentional discharges into rivers and lakes,
are common. Thurston et al. (1981a) studied the acutely toxic effects of
fluctuating concentrations of ammonia on rainbow trout and cutthroat trout
(Salmo clarki) and reported that fish were able to withstand short-term
excursions slightly above acutely toxic concentrations without any apparent
long-term adverse effects, provided the high ammonia concentrations were
followed by periods of low ammonia concentration. However, they also reported
that fish were better able to withstand steady concentrations of ammonia, over
96 hours, than they were fluctuating concentrations having mean values
comparable to those of the steady concentration for the same length of time.
Increased resistance to acute toxicity from ammonia as a consequence of prior
exposure to low ammonia concentrations has been reported by Vamos (1963),
Malacea (1968), Schulze-Wiehenbrauck (1976), Redner and Stickney (1979),
Thurston et al. (1981a), and Alabaster et al. (1983). Soderberg and co-
187
-------
JE
_c:
~t^
CD
O
O
CD
e
1—
3uu
200
100
80
60
50
40
30
20
in
" o i
0
o
D 8
° s : * s o
- M "
W A
A
A
A
° 0~5%NaCI
a ~100mg/LCa2+
o -lOmg/LNaHCOs
A ~95mg/LNaHC03
A
a
nf\ 1 1 1 1 1 1 1 1 1
6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.510.0.10.5
PH
Figure 2. Effect of pH and different water chemistry
variables on the acute toxicity of un-ioniniaed
ammonia to coho salmon alevins. (Reprinted with
permission from Russo et al. 1988; American Fisheries
Society)
workers (1983), however, have reported that tissue damage in rainbow trout
reared for 4 months in earthen ponds was correlated with ammonia concentration
extremes, as opposed to ammonia averages.
MIXTURES OF AMMONIA AND OTHER CHEMICALS
Information on the toxic effects of ammonia as a result of exposure in
combination with other chemicals is limited, but an increase in ionic strength
appears to reduce ammonia toxicity, separate from the effect of ionic activity
on the NH3/NH4+ equilibrium. Increased calcium concentrations in the water
have been reported to decrease ammonia toxicity to channel catfish (Tomasso et
al. 1980), and an increase in water salinity has been reported to reduce the
toxicity of ammonia to Atlantic salmon (Salmo salar) (Alabaster et al. 1979).
188
-------
A decrease in the toxicity of ammonia to juvenile coho salmon with the
addition of sodium chloride to the test solution also has been observed in our
laboratories (Figure 2).
Wuhrmann and Woker (1948) and Broderius and Smith (1979) studied combina-
tions of ammonia and hydrogen cyanide and reported synergistic effects of the
two toxicants. The toxicity of ammonia and phenol combined was reported to
approximate the toxicity of either ammonia or phenol when tested separately
(Herbert 1962) Research has also been conducted on the effect of ammonia in
combination with copper (Herbert and Vandyke 1964), zinc (Ministry of Technol-
ogy, U.K. 1962), and combinations of these (Brown et al. 1969); the toxicity
of ammonia in combination with these other toxicants appears to be simply
additive.
CHRONIC TOXICITY OF AMMONIA
The longest ammonia exposure study reported in the literature is a life-
cycle test on rainbow trout, conducted over a 5-year period (Thurston et al.
1984). Fish were tested at five exposure concentrations over the range 0.01
to 0.07 mg/L NH3. Parental fish were exposed for 11 months, the first filial
generation (Fx) for 4 years, and the second filial generation (F2) for 5
months. The parental fish spawned of their own volition, and Fx fish were
manually spawned at 4 years of age. Blood ammonia concentrations in Fx fish
were positively correlated with ammonia concentrations in the test waters, and
histopathological lesions were common in parental and F1 fish at 0.02 mg/L NH3
and higher. No other major effects were observed. Viable eggs were produced
by both generations at all ammonia concentrations tested, and there was no
significant correlation between ammonia concentration and numbers of egg lots
spawned, total numbers of eggs produced, numbers of viable eggs, growth of
progeny, or mortality of parents or progeny in any of the generations, tested.
Effects of ammonia on fish reproduction and survival were observed,
however, in two life-cycle tests on fathead minnows (Thurston et al. 1986).
Fish were tested at five exposure concentrations for 1 year over the range of
0.07 to 0.96 mg/L NH3. Growth and survival of, and egg production by,
parental fish were all affected at 0.96 mg/L NH3; no effects were observed on
growth or survival of parental fish at 0.44 mg/L, or on egg production or
viability at 0.37 mg/L. Growth and survival of F-^ larvae were not affected at
0.36 mg/L NH3 (the highest concentration at which they were tested), and egg
hatching success was not affected at 0.19 mg/L, but was at 0.37 mg/L. Brain
lesions were common in parental fish at all stages of their development at
exposure concentrations of 0.21 mg/L NH3 and higher, but not at 0.11 mg/L. In
summary, the chronic effects threshold concentration for these tests on
fathead minnows, based on survival, growth and reproductive success, was
estimated to be 0.27 mg/L NH3; based on histological damage, it was estimated
to be only half that, 0.15 mg/L NH3.
Other studies at sublethal concentrations of ammonia, lasting from 1 week
to 3 months, have been reported for both salmonid and non-salmonid species
(Department of the Environment, U.K. 1972, Robinette 1976, Schulze-Wiehen-
brauck 1976, Burkhalter and Kaya 1977, Thurston et al. 1978, Rice and Bailey
1980). At exposure concentrations as low as 0.002 to 0.15 mg/L NH3, fish
139
-------
showed reduced food uptake and assimilation, accompanied by growth inhibition.
In the range of 0.04 to 0.4 mg/L NH3, effects reported for a variety of fish
species, under a wide range of water test conditions, included leucopenia and
diminished numbers of red blood cells, inflammation and degeneration of gills
and kidneys, and lowered resistance to disease (Reichenbach-Klinke 1967, Flis
1968, Smart 1976, Thurston et al. 1978, Peters et al. 1984, Soderberg 1985,
Dabrowska and Wlasow 1986, Lang et al. 1987).
Carline et al. (1987) tested the effects of domestic wastewater on the
survival, growth, swimming performance, and gill tissue of brown trout (Salmo
trutta) for 12 months at concentrations from 0.005 to 0.066 mg/L NH3. No
significant effects were reported for survival, .growth, or swimming perform-
ance, but degree of damage to the gills was directly related to effluent
concentrations. Mitchell and Cech (1983) have implicated chlorine as con-
founding ammonia-caused gill damage. Burrows (1964) tested rainbow trout for
6 weeks at concentrations as low as 0.003 mg/L NH3 (recalculated from the
author's original data) and reported extensive gill damage, but Daoust and
Ferguson (1984) tested that same species at concentrations up to 0.4 mg/L NH3
for 90 days and did not observe "characteristic" gill damage. Meade and
Herman (1986) reported that lake trout (Salvelinus namaycush), reared for 8
weeks in a water re-use system, showed gill and kidney damage of the kind
reported to be caused by ammonia, and yet the ammonia concentrations to which
these fish were exposed averaged 0.001 mg/L NH3 or less within each 2-week
period, and the highest recorded value was 0.003 mg/L. Their conclusion was
that "parameters other than ammonia concentrations are, significant factors in
branchial irritation" of fish under culture.
In an extensive review article, Meade (1985) states as part of his
summary: "The accumulating evidence indicates that gill hyperplasia, reported
as characteristic of ammonia poisoning, is probably not caused by un-ionized
ammonia. ... All end products of metabolism probably have not been identi-
fied, and certainly their interactions in water of various qualities and in
various fish culture systems.have not been determined."
LITERATURE CITED
Alabaster, J.S., D.G. Shurben, and G. Knowles. 1979. The effect of dissolved
oxygen and salinity on the toxicity of ammonia to smolts of salmon, Salmo
salar L. Journal of Fish Biology 15:705-712.
Alabaster, J.S., D.G. Shurben, andM.J. Mallett. 1983. The acute lethal
toxicity of mixtures of cyanide and ammonia to smolts of salmon, Salmo
salar L- at low concentrations of dissolved oxygen. Journal of Fish
Biology 22:215-222.
Armstrong, D.A.,^D. Chippendale, A.W. Knight, andJ.E. Colt. 1978. Interac-
tion of ionized and un-ionized ammonia on short-term survival and growth
of prawn larvae, Macrobrachium rosenbergii. Biological Bulletin of the
Woods Hole Institute 154:15-31.
190
-------
Arthur, J.W., C.W. West, K.N. Allen, and S.F. Hedtke, 1987. Seasonal
toxicity of ammonia to five fish and nine invertebrate species. Bulletin
of Environmental Contamination and Toxicology 38:324-331.
Broderius, S., R. Drummond, J. Fiandt, and C. Russom. 1985. Toxicity of '.
ammonia to early life stages of the smallmouth bass at four pH values.
Environmental Toxicology and Chemistry 4:87-96.
Broderius, S.J., and L.L. Smith, Jr. 1979. Lethal and sublethal effects of
binary mixtures of cyanide and hexavalent chromium, zinc, or ammonia to
the fathead minnow (Pimephales promelas') and rainbow trout (Salmo
gairdneri). Journal of the Fisheries Research Board of Canada
36:164-172.
Brown, V.M., D.H.M. Jordan, and B.A. Tiller. 1969. The acute toxicity to
rainbow trout of fluctuating concentrations and mixtures of ammonia,
phenol and zinc. Journal of Fish Biology 1:1-9.
Burkhalter, D.E., and C.M. Kaya. 1977. Effects of prolonged exposure to
ammonia on fertilized eggs and sac fry of rainbow trout (Salmo
gairdneri). Transactions of the American Fisheries Society 106:470-475.
Burrows, R.C. 1964. Effects of accumulated excretory products on hatchery-
reared salmonids. Bureau of Sport Fisheries and Wildlife, U.S. Fish and
Wildlife Service, Research Report 66, 12p.
Butler, J.N. 1964. Ionic equilibrium.
Inc., Reading, Massachusetts, USA.
Addison-Wesley Publishing Company,
Calamari, D., R. Marchetti, and G. Vailati. 1981. Effects of long-term
exposure to ammonia on the developmental stages of rainbow trout (Salmo
gairdneri Richardson). Rapports et Proces-Verbaux des Reunions, Conseil
International pour 1'Exploration de la Mer 178:81-86.
Carline, R.F., A.J. Benson, and H. Rothenbacher.
treated domestic wastewater on brown trout.
1987. Long-term effects of
Water Research 21:1409-1415.
Chipman, W.A., Jr. 1934. The role of pH in determining the toxicity of
ammonium compounds. Ph.D. thesis, University of Missouri, Columbia,
Missouri, USA.
Colt, J., and G. Tchobanoglous. 1976. Evaluation of the short-term toxicity
of nitrogenous compounds to channel catfish, Ictalurus punctatus.
Aquaculture 8:209-224.
Dabrowska, H., and T. Wlasow. 1986. Sublethal effect of ammonia on certain
biochemical and haematological indicators in common carp (Cyprinus carpio
L.). Comparative Biochemistry and Physiology 83C:179-184.
Daoust, P.-Y., and H.W. Ferguson. 1984. The pathology of chronic ammonia
toxicity in rainbow trout, Salmo gairdneri Richardson. Journal of Fish
Diseases 7:199-205.
191
-------
Department of the Environment, U.K.
Water Pollution Research, 1971.
p. 37.
1972. Effects of pollution on fish. In:
H.M. Stationery Office, London, England.
Downing, K.M., and J.C. Merkens. 1955. The influence of dissolved-oxygen
concentration on the toxicity of un-ionized ammonia to rainbow trout
(Salmo gairdneri Richardson). Annals of Applied Biology 43:243-246.
Emerson, K., R.C. Russo, R.E. Lund, and R.V. Thurston. 1975.
equilibrium calculations: effect of pH and temperature.
Fisheries Research Board of Canada 32:2379-2383.
Aqueous ammonia
Journal of the
Emery, R.M. , and E.B. Welch. 1969. The toxicity of alkaline solutions of
ammonia to juvenile bluegill sunfish (Lepomis macrochirus Raf.). Water
Quality Branch, Division of Health and Safety, Tennessee Valley Author-
ity, Chattanooga, Tennessee, USA.
Flis, J. 1968. Anatomicohistopathological changes induced in carp CCyprinus
carpio L.) by ammonia water. Part II. Effects of subtoxic concentra-
tions. Acta Hydrobiologica 10:225-238.
Hasan, M.R., andD.J. Macintosh. 1986.
carp fry. Aquaculture 54:97-107.
Acute toxicity of ammonia to common
Hazel, R.H., C.E. Burkhead, and D.G. Huggins. 1979. The development of water
quality criteria for ammonia and total residual chlorine for the protec-
tion of aquatic life in two Johnson County, Kansas streams. Office of
Water Research and Technology, U.S. Department of the Interior, Washing-
ton, District of Columbia, USA.
Herbert, D.W.M. 1962. The toxicity to rainbow trout of spent still liquors
from the distillation of coal. Annals of Applied Biology 50:755-777.
Herbert, D.W.M., and J.M. Vandyke. 1964. The toxicity to fish of mixtures of
poisons. II. Copper-ammonia and zinc-phenol mixtures. Annals of Applied
Biology 53:415-421.
Hillaby, B.A. , and D.J. Randall. 1979. Acute ammonia toxicity and ammonia
excretion in rainbow trout (Salmo gairdneri). Journal of the Fisheries
Research Board of Canada 36:621-629.
Lang, T., G. Peters, R. Hoffmann, and E. Meyer. 1987. Experimental inves-
tigations on the toxicity of ammonia: effects on ventilation frequency,
growth, epidermal mucous cells, and gill structure of rainbow trout Salmo
gairdneri. Diseases of Aquatic Organisms 3:159-165.
Malacea, I. 1968. Untersuchungen uber die Gewohnung der Fische an hohe
Konzentrationen toxischer Substanzen. (Studies on the acclimation of
fish to high concentrations of toxic substances.). Archiv fur Hydro-
biologie 65:74-95, (English translation).
192
-------
McCormick, J.H., S.J. Broderius, and J.T. Fiandt. 1984. Toxicity of ammonia
to early life stages of the green sunfish Lepomis cyanellus. Environmen-
tal Pollution (Series A) 36:147-163.
Meade, J.W. 1985. Allowable ammonia for fish culture.
Culturist 47:135-145.
The Progressive Fish-
Meade, J.W., and R.L. Herman. 1986. Histological changes in cultured lake
trout, Salvelinus namaycush. subjected to cumulative loading in a water
reuse system. Canadian Journal of Fisheries and Aquatic Sciences
43:228-231.
Merkens, J.C., and K.M. Downing. 1957. The effect of tension of dissolved
oxygen on the toxicity of un-ionized ammonia to several species of fish.
Annals of Applied Biology 45:521-527.
Ministry of Technology, U.K. 1962. Effects of pollution on fish. Toxicity
of mixtures of zinc sulphate and ammonium chloride. In: Water Pollution
Research, 1961. H.M. Stationery Office, London, England, pp. 90-93.
Mitchell, S.J., and J.J. Cech', Jr. 1983. Ammonia-caused gill damage in
channel catfish (Ictalurus punctatus): confounding effects of residual
chlorine. Canadian Journal of Fisheries and Aquatic Sciences 40:242-247.
Peters, G., R. Hoffmann, and H. Klinger. 1984. Environment-induced gill
disease of cultured rainbow trout (Salmo gairdneri) . Aquaculture 38:105-
126.
Randall, D.J., and P.A. Wright. 1987. Ammonia distribution and excretion in
fish. Fish Physiology and Biochemistry 3:107-120.
Redner, B.D., and R.R. Stickney. 1979. Acclimation to ammonia by Tilapia
aurea. Transactions of the American Fisheries Society 108:383-388.
Reichenbach-Klinke, H.-H. 1967. Untersuchungen uber die Einwirkung des
Ammoniakgehalts auf den Fischorganismus. (Investigations on the in-
fluence of the ammonia content on the fish organism.). Archiv fur
Fischereiwissenschaft 17:122-132, (English translation).
Rice, S.D., and J.E. Bailey. 1980. Survival, size, and emergence of pink
salmon, Oncorhynchus gorbuscha. alevins after short- and long-term
exposures to ammonia. Fisheries Bullretin 78:641-648.
Robinette, H.R. 1976. Effect of selected sublethal levels of ammonia on the
growth of channel catfish (Ictalurus punctatus). The Progressive
Fish-Culturist 38:26-29.
Robinson-Wilson, E.F., and W.K. Seim. 1975. The lethal and sublethal effects
of a zirconium process effluent on juvenile salmonids. Water Resources
Bulletin 11:975-986.
193
-------
Roseboom, D.P., and D.L. Richey. 1977. Acute toxicity of residual chlorine
and ammonia to some native Illinois fishes. Report of Investigation 85.
Illinois State Water Survey, Urbana, Illinois, USA.
Russo, R.C., D.J. Randall, and R.V. Thurston. 1988. Ammonia toxicity and
metabolism in fishes. In: Protection of River Basins, Lakes, and
Estuaries. Edited by R.C. Ryans, American Fisheries Society, Bethesda,
Maryland, USA. pp. 159-173.
Schulze-Wieheribrauck, H. 1976. Effects of sublethal ammonia concentrations
on metabolism in juvenile rainbow trout (Salmo gairdneri Richardson).
Berichte der Deutschen Wissenschaftlichen Kommission fur Meeresforschung
24:234-250.
Sheehan, R.J., and W.M. Lewis. 1986. Influence of pH and ammonia salts on
ammonia toxicity and water balance in young channel catfish. Transac-
tions of the American Fisheries Society 115:891-899.
Smart, G. 1976. The effect of ammonia exposure on gill structure of the
rainbow trout (Salmo gairdneri). Journal of Fish Biology 8:471-475.
Smart, G.R. 1978. Investigations of the toxic mechanisms of ammonia to fish
-- gas exchange in rainbow trout (Salmo gairdneri) exposed to acutely
lethal concentrations. Journal of Fish Biology 12:93-104.
Soderberg, R.W. 1985. Histopathology of rainbow trout, Salmo gairdneri
Richardson, exposed to diurnally fluctuating un-ionized ammonia levels in
static-water ponds. Journal of Fish Diseases 8:57-64.
Soderberg, R.W., J.B. Flynn, and H.R. Schmittou. 1983. Effects of ammonia
on growth and survival of rainbow trout in intensive static-water
culture. Transactions of the American Fisheries Society 112:448-451.
Sousa, R.J., T.L. Meade, and R.E. Wolke. 1974. Reduction of ammonia toxicity
by salinity and pH manipulation. Proceedings of the Fifth Annual
Workshop, World Mariculture Society 5:343-354.
Tabata, K. 1962. Suisan dobutsu ni oyobosu amonia no dokusei to pH, tansan
to no kankei. (Toxicity of ammonia to aquatic animals with reference to
the effect of pH and carbon dioxide.). Tokai-ku Suisan Kenkyusho
Kenkyu Hokoku 34:67-74, (English translation).
Thurston, R.V., C. Chakoumakos, and R.C. Russo. 1981a. Effect of fluctuating
exposures on the acute toxicity of ammonia to rainbow trout (Salmo
gairdneri) and cutthroat trout (S_. clarki) . Water Research 15:911-917.
Thurston, R.V., and E.L. Meyn. 1984. Acute toxicity of ammonia to five fish
species from the northwest United States. Technical Report No. 84-4,
Fisheries Bioassay Laboratory, Montana State University, Bozeman
Montana, USA. 13p.
194
-------
Thurston, R.V. , G.R. Phillips, R.C. Russo, and S.M. Hinkins. 1981b. In-
creased toxicity of ammonia to rainbow trout (Salmo gairdneri) resulting
from reduced concentrations of dissolved oxygen. Canadian Journal of
Fisheries and Aquatic Sciences 38:983-988.
Thurston, R.V., and R.C. Russo. 1983. Acute toxicity of ammonia to rainbow
trout. Transactions of the American Fisheries Society 112:696-704.
Thurston, R.V., R.C. Russo, and K. Emerson. 1979. Aqueous ammonia equilib-
rium - tabulation of percent un-ionized ammonia. Ecological Research
Series, EPA-600/3-79-091, U.S. Environmental Protection Agency, Duluth,
Minnesota, USA.
Thurston, R.V., R.C. Russo, R.J. Luedtke, C.E. Smith, E.L. Meyn, C. Chakou-
makos, K.C. Wang, and C.J.D. Brown. 1984. Chronic toxicity of ammonia
to rainbow trout. Transactions of the American Fisheries Society 113:56-
73.
Thurston, R.V., R.C. Russo, E.L. Meyn, and R.K. Zajdel. 1986. Chronic
toxicity of ammonia to fathead minnows. Transactions of the American
Fisheries Society 115:196-207. .
Thurston, R.V., R.C. Russo, and G.R. Phillips. 1983. Acute toxicity .of
ammonia to fathead minnows. Transactions of the American Fisheries
Society 112:705-711.
Thurston, R.V., R.C. Russo, and C.E. Smith. 1978. Acute toxicity of ammonia
and nitrite to cutthroat trout fry. Transactions of the American
Fisheries Society 107:361-368.
Thurston, R.V., R.C. Russo, and G.A. Vinogradov. 1981c. Ammonia toxicity to
fishes: effect of pH on the toxicity of the un-ionized ammonia species.
Environmental Science and Technology 15:837-840.
Tomasso, J.R., C.A. Goudie, B.A. Simco, and K.B. Davis. 1980. Effects of
environmental pH and calcium on ammonia toxicity in channel catfish.
Transactions of the American Fisheries Society 109:229-234.
Vamos, R.
97.
1963. Ammonia poisoning in carp. Acta Biologica (Szeged) 9:291-2-
Wuhrmann, K. 1952. Sur quelques principes de la toxicologie du poisson.
(Concerning some principles of the toxicology of fish.) Bulletin du
Centre beige d'Etude et de Documentation des Eaux 15:49-60, (English
translation).
Wuhrmann, K., and H. Woker. 1948. Beitrage zur Toxikologie der Fische. II.
Experimentalle Untersuchungen uber die Ammoniak- und Blausaurevergiftung.
(Contributions to the toxicology of fishes. II. Experimental investiga-
tions on ammonia and hydrocyanic acid poisoning). Schweizerische
Zeitschrift fur Hydrologie 11:210-244, (English translation).
195
-------
Wuhrmann, K., and H. Woker. 1953. Beitrage zur Toxikologie der Fische. VIII.
Uber die Giftwirkungen von Ammoniak- und Zyanidlosungen mit verschiedener
Sauerstoffspannung und Temperatur auf Fische. (Contributions to the
toxicology of fishes. VIII. On the toxic effects of ammonia and
cyanide solutions on fish at different oxygen tensions and tempera-
tures). Schweizerische Zeitschrift fur Hydrologie 15:235-260, (English
translation).
Wuhrmann, K., F. Zehender, and H. Woker. 1947. Uber die fischereibiologische
Bedeutung des Ammonium- und Ammoniakgehaltes fliessender Gewasser. (Bio-
logical significance for fisheries of ammonium ion and ammonia content of
flowing bodies of water). Vierteljahrsschrift der Naturforschenden
Gesellschaft in Zurich 92:198-204, (English translation).
196
-------
ENVIRONMENTAL TOLERANCE OF SOME MARINE
FISH: APPLICATIONS IN MARICULTURE MANAGEMENT
by
R.S.S. Wu1
INTRODUCTION
Most studies of environmental effects on marine fish have confined
themselves to evaluating physiological responses of a few species. Few
studies compare the environmental tolerances of different species. Moreover,
differences in experimental protocols make it difficult to compare tolerances
and responses between studies (Davis 1975).
Although it has long been recognized that the success or failure of
mariculture operations may hinge on the choice of species (Avault 1986), in
the absence of scientific data, the suitability of a species for a particular
hydrographic condition in a culture site can only be determined by experience
or by trial and error. Hydrographic conditions, however, vary and fluctuate
greatly both spatially and temporally. As a result, knowledge of environ-
mental tolerances and requirements of marine fish provides a valuable
scientific basis for selecting appropriate species to suit different culture
conditions and is of obvious importance from the viewpoint of mariculture
management and development.
The tolerance of a species to a particular environmental factor may be
assessed by mortality and the behavioral and physiological responses of the
fish. In this context, the time for-50% of the experimental population to
show mortality (LT50 value) , and the time for 50% of the experimental
population to exhibit abnormal behavioral changes (BC50 value) can serve as
useful tolerance indicators with which the relative tolerance of different
species in relation to a particular environmental factor can be assessed and
compared. This paper presents experimental data on oxygen, salinity, and
temperature tolerances of some common mariculture fish in the Southeast Asia
regions, and the results are related to fish kill statistics collected for 10
years in Hong Kong. The application value of these tolerance data in
mariculture management also are discussed.
MATERIALS AND METHODS
All fish used for the experiment had been cultured in sea cages
(salinity, 30 to 32%; dissolved oxygen, 5 to 8 mg 02 L"1; water temperature,
•"•City Polytechnic of Hong Kong, Kowloon Tong, Hong Kong.
197
-------
25 to 28°C) for about a year. Experimental fish weighing in the range of 150
to 250 g were selected and acclimated in the laboratory (salinity, 30%;
dissolved oxygen, 8 to 8.5 mg 02 L"1; temperature, 25 ± 1°C) for at least 48
hours prior to the experiment and were not fed throughout the entire
acclimation and experimental period.
OXYGEN TOLERANCE
A continuous flow system that provided a constant, controlled level of
ambient oxygen was set up for the hypoxic tolerance experiment (for detailed
design of the system, see Wu and Woo 1984). Twenty individuals of each of the
nine species listed in Table 1 were acclimated in the experimental system for
24 hours prior to experiments, during which period normoxic water (25 ± 1°C)
was provided. The levels of oxygen were adjusted to 4, 2.5, 1.0 and 0.5 mg 02
L"1 within 15 minutes and the fish were exposed to the desired dissolved
oxygen level for 7 hours. The time at which any fish began to show any stress
symptom (i.e. loss of balance, abnormal swimming pattern, jerking) was noted
and the time of death of any fish was recorded.
TABLE 1. TIME (minutes) FOR 50% OF VARIOUS EXPERIMENTAL FISH TO SHOW
MORTALITY (LT50) AND ABNORMAL BEHAVIOR (BC50)a-b
Species
Ghrysophrys major
Rhabdosarga sarba
Siganus oramin
Lutlanus ruselli
Epinephelus akaara
Mylio macrocephalus
Epinephelus awoara
Lates calearifer
Epinephelus tauvina
0.5 mg 0
BC50
TQC
TQC
TQC
TQC
70
30
150
373
(10%)
'„ r1
LT50
TQC
TQC
7
14
92
100
200
393
(10%)
1.0 mg 0
BC50
20
130
10
78
110
NBC
190
NBC
NBC
'2 I"1
60
225
177
(30%)
160
NM
270
NM
NM
°Actual mortalities are shown in parenthesis for mortality-behavior changes
occurring in less than 50% of the population throughout the experimental
period.
bTQC, too quick to count; NBC, no behavioral changes; NM, no mortality.
198
-------
To test whether there is any correlation between hypoxic tolerance and
metabolic rate of different species, the standard oxygen consumption rate of
each species at 25°C (ambient oxygen level) was measured. Five of each
species were placed individually in sealed Plexiglas chambers (20 x 12 x 12
cm) and were placed in experimental tanks of the continuous flow systems.
Fish were allowed to adapt to the experimental condition for 24 hours before
measurement. Water was allowed to flow into the chamber through the inlets
and siphoned out from the outlet of the chambers. The flow rate was adjusted
so that there was a measurable difference of about 0.5 mg 02 IT1 between the
inflow and outflow of the sealed chambers. An empty chamber identical with
that containing the animal was set up as a control, and the oxygen uptake of
the fish was calculated using the following equation.
q - (X0
- (X0-X2)(dV/dt)2
where q is the oxygen uptake, X0 is the oxygen concentration at the inflow, Xx
is the oxygen concentration at the outflow of the experimental chamber, X2 is
the oxygen concentration at the outflow of the chamber, (dV/dt^ is the flow
rate of the experimental chamber, and (dV/dt)2 is the flow rate of the control
chamber. The oxygen consumption rate then is calculated and expressed in
terms of mg 02 g body wet wt.'1 min."1
SALINITY TOLERANCE
After acclimation, twenty individuals of each of the 13 species listed
in Table 3 were transferred abruptly to 3 , 5 and 10% sea water (temperature =
25 ± 1°C) for 14 days in the laboratory. In addition, twenty individuals of
S . oramin E . akaara . and m. macrocephalus were transferred to freshwater (0%) .
The time at which any fish began to show any stress symptom was noted and the
time of death of any fish was recorded.
TEMPERATURE TOLERANCE
A continuous flow system in which a constant water temperature of 25 ±
0.2°C was set up for the temperature tolerance experiment. Twenty individuals
of each of the eight species listed in Table 4 were acclimated in the system
for 24 hours prior to experiments. The temperature of the system was
gradually brought to 12 ± 0.2°C within 1 hour, and the fish then were
subjected to the experimental temperature for 7 hours. The time at which any
fish began to show any stress symptom was noted and the time of death of any
fish was recorded.
FISH KILL STATISTICS
Since 1976, all major fish kills in Hong Kong have been reported to the
Agriculture and Fisheries Department. In most of these cases, investigations
were carried out to ascertain the cause of the kill and to assess the total
loss as well as the species affected. Over 40% of the major kills recorded
was attributable to oxygen depletions resulting from red tides and algal
199
-------
blooms; the total'loss between 1976 and 1988 was estimated at 124 tonnes,
valued at approximately US $1 million.
RESULTS
OXYGEN TOLERANCE
No observable stress symptom can be found for all species at 2.5 mg 02
L" and above. Most of the species turned into a noticeable lighter color
when oxygen dropped below 2.5 mg 02 L"1, and this phenomenon vraS particularly
marked for the Sparidae. Time required for 50% of each species to show stress
symptoms and mortality are shown in Table 1. Large variations were found in
hypoxic tolerance amongst the nine species tested. For example, 50% of £_..
major exhibited abnormal symptoms within 20 minutes at 1 mg 02 L"1, whereas L^
calcarifer and E. tauvina did not show any abnormal symptoms within the 7-hour
exposure. Similarly, all C. major and R. sarga died before the oxygen level
dropped to 0.5 mg 02 L"1, whereas only 10% of E. tauvina showed abnormal
symptoms and mortality under the same condition throughout the 7-hour
experimental period.
Oxygen consumption rates for the nine species tested are shown in Table
2. The metabolic rate of intolerant species (e.g. C. major and R. sarba) was
more than two times higher than those of tolerant species (e.g. E. tauvina and
TABLE 2. OXYGEN CONSUMPTION RATE OF VARIOUS SPECIES AT 25°C AND NORMOXIA
(Species arranged in ascending order according to their LT50 values at 0.5 mg
°2 i"1)•
Species
Oxygen consumption rate, (/^g 02 g"1 min.
Chrysophrvs major
Rabdosarga sarba
Siganus oramin
Lutjanus ruselli
Epinephelus akaara
Mvlio macrocephalus
Epinephelus awoara
Lates ealcarifer
Epinephelus tauvina
4.65 ± 0.81
4.88 ± 0.33
9.73 ± 1.42
5.93 ± 0.21
3.48 ± 0.12
2.84 ± 0.39
2.20 ± 0.40
1.98 ± 0.14
1.99 ± 0.53
200
-------
L. calcarifer). A Spearman rank correlation test showed a significant
correlation between the rank orders of the LT50 and BC50 values at both 1.0 and
0.5 mg 02 L"1 (r=0.986 and 0.790, respectively). A significant negative
correlation was found between oxygen consumption rate and LT50 and BC^, values
of the species tested at 0.5 mg 02 L"1 (r= -0.732 and -0.673, respectively).
SALINITY TOLERANCE
LT50 values for each species at different salinities are shown in Table
3. Except for Caranx eauula. no mortalities were found for any species at the
end of 14 days of exposure to 30 and 10%. No mortality was found in 6 species
exposed to 5%. Three species survived without any abnormal behavior in 3%,
and 10% of M. macrocephalus survived without any abnormal behavior at the end
of the 14 days of exposure to freshwater.
TEMPERATURE TOLERANCE
Except for S. oramin. which exhibited a 20% mortality, no mortality was
found for the other 7 species in a 7-hour exposure to 12°C. Time for 50% of
experimental animals to exhibit abnormal behavior is shown in Table 4. EL.
oramin and Lates calcarifer showed a very quick behavioral response to the
lowering of .water temperature, followed in order by E. tauvina. R. sarba and
C. major. No behavioral changes were found for E. akaara. E. aworara. and M.
macrocephalus during a 7-hour exposure to 12°C.
FISH KILL STATISTICS
Species involved in fish kills due to oxygen depletion in Hong Kong from
1976 to 1986 are shown in Table 5. C. major and R. sarba were involved in 17
and 11 incidents out of 19 major fish kills, respectively. L. calcarifer and
E. tauvina were never reported in any fish kills caused by oxygen depletion
although they also are commonly cultured. Table 6 shows the stock and
percentage killed of each species in a fish kill in which a dissolved oxygen
level of 0.8 mg 02 L"1 was recorded. In this incident, only C. major. R.
sarba. L. ruselli. and E. akaara were killed. M. macrocephalus and E. tauvina
were not affected although they also were cultured at the same site.
Statistics of fish kills caused by cold spells (with water temperature
<15°C for more than 7 days) in Hong Kong are shown in Table 7. L. calcarifer.
E. tauvina were found to be involved in all the five recorded incidents. E.
akaara. E. awoara. and M. macrocephalus were never reported although these
species also were cultured.
DISCUSSION
Most warm water fish survive under a wider range of environmental
conditions than do cold water species (Parker and Davis 1981). In this study,
the species tested showed large variations in their tolerance to hypoxia,
temperature, and salinity, and several species showed remarkable tolerance to
201
-------
TABLE 3. LT50 VALUES (hours) OF VARIOUS MARINE FISH KEPT UNDER DIFFERENT
HYPO-OSMOTIC SALINITIES FOR 14 DAYS (n=20).a-b
Species 0%
Slganus oramin 1.2
Caranx eouula
Parapristjpoma trilineatum
Chrvsophrvs major
Lethrinus nebulosus
Epinephelus akaara 2.5
Plotosus anguillaris
Pomadasvs hasta
Epinephelus awoara
Rhabdosarga sarba
Therapon iarbua
Lutianua russelli
Mylio macrocephalus 60.0
Salinity
3% 5%
2.3
3.1
4.4
4.8
9.1
31.3
40.0
70.0
76.0
96.0
NM
NM
NM
NM
42.5
38.0
11.1
86.0
NM
55.0
(30%)
NM
(10%)
NM
NM
NM
10%
NM
126.0
NM
NM
NM
NM
NM •
NM
NM
NM
NM
NM
NM
"Actual mortalities shown in parentheses for mortalities of less than 50%
within the experimental period (Wu and Woo 1982).
bNM, no mortality.
one or more parameters. For example, the black sea bream, Mylio
macrocephalus, was very tolerant to all three environmental parameters tested.
Results of hypoxic tolerance experiments show that Lates calcarifer.
Epinephelus tauvina and Mylio macrocephalus are tolerant of, whereas
Chrvsophrvs major and Rhabdosarga sarba are particularly susceptible to,
hypoxic conditions. Hypoxic tolerance is negatively related to the metabolic
rate^of the species. It appears that species that have a lower metabolic rate
survive better in a hypoxic environment because less oxygen would be required
to sustain essential biological functions. The hypoxic tolerance of fish
species derived from laboratory results agrees closely with these established
from analyses of fish kill statistics. C. major and R. sarba always suffer
202
-------
TABLE 4. TIME (minutes) FOR 50% OF EXPERIMENTAL ANIMALS TO EXHIBIT ABNORMAL
BEHAVIOR (BC50) AT C (n=20)
Species
BC
50
Mortality, %
Siganus oramin
Lates calcarifer
Epinephelus tauvina
Rhabdosarga sarba
Chrysorphyrs major
Epinephelus akaara
Ep inephelus awoara
Mvlio macrocephalus
TQC
1
20
50
89
NBC
NBC
NBC
20
aNBC, no behavioral changes; TQC, too quick to count
heavy loss in fish kills caused by oxygen depletions; other species normally
are affected to a much lesser extent.
Conversely, L. calcarifer and E. tauvina. which have never been reported
in any fish kills due to oxygen depletion, were found to be highly tolerant to
hypoxia in the laboratory experiments. Detailed biochemical studies on
Ephinephelus akaara and Mylio macrocephalus exposed to hypoxic conditions by
Woo and Wu (1984) revealed no increases in serum and tissue lactate and only
slight changes in other tissue metabolites and electrolytes at 4.0 to 2.5 mg
02 L"1 for these two species, indicating that they can obtain enough oxygen to
prevent anaerobiosis in this regime. Marked elevations of serum lactate and
serum Na+, K+ and Ca2+ were found and osmoregulation failure occurred, however,
when oxygen values fell below 1 mg 02 L"1.
The overall results suggest, therefore, that hypoxic tolerant species
such as Lates calcarifer and M. macrocephalus may be cultured in eutrophic
waters in which oxygen depletion is more likely to occur and that culturing
sensitive species such as C. maj or and R. sarba should be avoided in the same
environment. Nevertheless, the quick, abnormal behavioral response to hypoxia
exhibited by C. maj or and R. sarba may serve as an effective indicator with
which culturists could detect the onset of oxygen depletion at an early stage
and aerate their cages to prevent loss of their stock. Such a "biological
indicator" system has been proven to be successful in Hong Kong.
203
-------
TABLE 5. NUMBER OF INCIDENTS IN WHICH A PARTICULAR SPECIES WAS INVOLVED IN 19
MAJOR FISH KILLS CAUSED BY OXYGEN DEPLETION IN HONG KONG, 1976 to 1987
Species
No. of incidents involving
species in 19 fish kills
Chrysophrvs malor
Rhabosarga sarba
Epinephelus akaara
Seriola purpurascens
Lutianus ruselli
Mylio macrocephalus
Mvlio latus
Lutianus sanguinensis
17
11
6
5
4
2
2
2
Results of the salinity tolerance experiments show that 10 out of the 13
species tested are euryhaline and survive without any abnormal behavior and
tissue hydration for more than 2 weeks in salinities above 10% (Wu and Woo
1983), suggesting that such salinities are normally not an important limiting
factor for culturing most of these species. Further experiments carried out
on E. akaara and M. macrocephalus showed only transient disturbance of various
electrolytes and metabolites in the case of the latter at salinities above
12%, suggesting that physiological disturbance is unlikely to occur in
salinity above this regime (Woo and Wu 1982). Dendrinos and Thorpe (1985)
also found that the bass (Dicentrarchus labrax") survives in salinities between
5 and 33% for over 12 months, and food and protein conversion efficiencies
were maximal at 25 and 30%. It may be hypothesized further that, because less
energy would be required for osmoregulation in an iso-osmotic environment,
energy saved might be channelled to tissue production. The possibility of
increasing mariculture productivity by culturing eurhaline species in an
environment of reduced salinity (e.g. estuary) should be explored.
Results from temperature tolerance experiments indicated that S. oramin.
L. calcarifer, and E. tauvina are relatively sensitive to cold temperature,
whereas E. akaara. e. awoara and M. macrocephalus are more tolerant. Again,
the laboratory findings were clearly supported by fish kill statistics in Hong
Kong. High mortality of the former three species were found during cold
spells when water temperatures fell below 15°C for a prolonged period.
Conversely, Mvlio macrocephalus. Epinephelus akaara. and Epinephelus awoara.
which showed no abnormal behavior at 12°C in the laboratory, have never been
204
-------
reported upon in fish kills caused by cold spells. The results, therefore,
indicate over-wintering problems for L. calcarifer and E. tauvina in Hong
Kong. Shortening of the grown-out period by, for example, importing larger
fingerlings might minimize the risk of fish kills for these two species in a
severe winter.
Menasveta (1981) found that the range of upper lethal temperature for 24
species of marine fish in the Gulf of Thailand is fairly narrow (34°C to
37.5°C). The upper lethal temperatures of Chrysophrys major and Mylio
macrocephalus were found to be 32 and 36°C, respectively (Woo and Fung 1980,
Woo per. com). The low tolerance of C. major to high water temperature
suggests that this species is less suitable for culturing in shallow waters
(i.e. < 3m, the shallowest thermocline in the coastal waters of Hong Kong)
where water may easily be heated by solar radiation in the summer to a
temperature beyond the lethal limit of this species.
TABLE 6. INVOLVEMENT OF VARIOUS SPECIES IN FISH KILLS AT PO TOI 0 HONG KONG
IN WHICH A DISSOLVED OXYGEN LEVEL OF 0.8 mg 02 I"1 WAS RECORDED
Species
Stock (t) Loss (t) Loss (%)
Chrysophrys major
Rhabdosarga sarba
Lut janus russelli
Epinephelus akaara
Sub -total
Epinephleus awoara
Epinephleus tauvina
Mylio latus
Mylio macrocephalus
Lut janus argentimaculatus
Others
Sub -total
8.1
20.3
2.8
0.8
32.0
3.0
2.4
1.3
9.0
1.2
2.1
22.5
0.1
0.8
0.1
0.1
1.1
0
0
0
0
0
0
0
1.2
3.9
3.6
12.5
3.4
0
0
0
0
0
0
0
205
-------
TABLE 7. NUMBER OF INCIDENTS IN WHICH A PARTICULAR SPECIES WAS INVOLVED IN
FIVE MAJOR FISH KILLS CAUSED BY COLD SPELLS (water temperature less than 15°C
for more than 7 days) IN HONG KONG, 1982 to 1987
Species
No. of incidents involving species
in five fish kills
Lates calcarifer
Epinephelus tauvina
Lutjanus sanguinensis
Rhabdosarga sarba
5
5
4
1
REFERENCES
Avault, J.W., Jr. 1986. Which species to culture - A check list.
Aquaculture Magazine. 12:41-44.
Davis, J. 1975. Minimal dissolved oxygen requirements of aquatic life
with emphasis on Canadian species: A Review. Journal of Fisheries
Research Board Canada. 32:2295-2332.
Dendrinos, P. and J.P. Thorpe. 1985. Effects of reduced salinity on growth
and body composition in the European bass (Dicentrarchus labrax (L.'n
Aquaculture. 49:333-358.
Hughes, G.M. 1973. Respiratory responses to hypoxia in fish. American
Zoologist. 13:475-489.
Menasveta, P. 1981. Lethal temperature of marine fish of the Gulf of
Thailand. Journal of Fish Biology. 18:603-607.
Woo, N.Y.S. and C.Y.A. Fung. 1981. Studies on the biology of the red sea
bream Chrsophrys major. IV Metabolic effects of starvation at low
temperature. Comparative Biochemistry and Physiology. 244:1-5.
Woo, N.Y.S. and R.S.S. Wu. 1982. Metabolic and osmoregulatory changes in
response to reduced salinities in the red grouper, Epinephelus akaara
(Temminck and Schlegel), and the black sea bream Mylio macrocephalus
(Basilewsky). Journal of Experimental Marine Biology and Ecoloev
65:139-161. &y
Woo, N.Y.S. and R.S.S. Wu. 1984. Changes in biochemical composition in the
red grouper, Epinephelus akaara (Temminck and Schlegel), and the black
sea bream, Mylio macrocephalus. during hypoxic exposure. Comparative
Biochemistry and Physiology. 77A:475-482.
206
-------
Wu, R.S.S. and N.Y..S. Woo. 1982. Tolerance of hypo-osmotic salinities in
thirteen species of adult marine fish: Implication for estuarine fish
culture. Aquaculture. 32:175-181.
Wu, R.S.S. and N.Y.S. Woo. 1984. Respiratory responses and tolerance to
hypoxia in two marine teleosts, Epinephelus akaara (Temminck and
Schlegel) and Mylio macrocephalus (Basilewsky). Hydrobiologia.
119:209-217.
207
-------
STUDIES ON PHYSIOLOGY OF PHOTOTAXIS OF FTRH
AND MARINE ANIMALS IN CHINA
by
Daren He1
INTRODUCTION
LIghtfishing has a long history in China. It can be traced back to
ancient times when fishermen used pine torches to lure cuttle fish and squid
in the sea and used lamps to lure crabs and torches as an aid to catching
cormorants. As early as 1930s., the fishermen of Dongshan, Fujian, Nanao, and
Guangdong used gaslamps to lure and catch round scads and sardines. With the
depletion of demersal fish resources, many countries in the world have began
to exploit the rich resources of pelagic fish. At present, lightfishing is
an important means of catching pelagic fish. With the development of science
and technology and the improvements in lighting technology, lightfishing has
become an advanced means of production that plays an important role in marine
fishery in many countries. China's coastal waters abound in varieties and
quantities of pelagic fish, of which chub mackerel, round scads, horse
mackerel, round herring, sardines, etc. have already become the major obiects
for purse seine fishing with light.
In the early 1960s, we carried out successful experiments in underwater
lamp fishing with the fishermen of Xiamen, Dongshan, and other areas From
the late 1960s to the early 1970s, the fishermen of Xiamen, Dongshan, etc
made tremendous strides in seine fishing using lights installed on motorized
junks. In the early 1970s, Shanghai Fishery Company and other units began to
develop purse seine fishing with lights, furthering the exploitation of the
marine pelagic fish resource of our country and raising the proportion of the
pelaghic fish yield in the total. To raise the light fishing industry
further, we must solve the problems that arise in production. Several
questions are involved. What is the optimum color light source for round scad
and chub mackerel? What is their optimum illumination? Why is the gathering
effect of light under moonlight not so ideal? Why is the phototaxis so bad
during the breeding season?
With these questions in mind, we, together with our colleagues in the
Shanghai Institute of Physiology and the East China Sea Fisheries Research
Department of Oceanography, Xiamen University, Xiamen, PRC
208
-------
With these questions in mind, we, together with our colleagues in the
Shanghai Institute of Physiology and the East China Sea Fisheries Research
Institute, made comprehensive studies of the phototactic character of the main
marine pelagic fish such as round scads and chub mackerel using ethological,
electro-physiological and visual-pigment-biochemical methods in the early
1970s. We also tested new light sources such as thallium-indium lamp on the
sea and achieved excellent results. Our achievements won the China National
Science Conference Prize in 1978. Later we continued with our studies on the
theory and application of the mechanism of phototaxis of fish and marine
animals, vision physiology, and ethology. This paper gives a summary of the
study in this field in China.
He Daren, Yu Wentsao, and others have studied the phototactic behaviour
of round scad CDecapterus maruadsi), chub mackerel (Pneumatophorus japonicus),
sardine CSardinella perforata), silverside (Atherina bleekeri), needlefish
CAblennes anastomella") , cardinal fish (Apogon lineatus) , Mugil carinatus and
squid CLoliog duvaucelii") using the photogradient method; Zheng Meili, He
Daren, and others have studied the behavioural physiology of cuttlefish
fSepiella maindroni); Luo Huiming and others have studied that of young eel
CAnguilla japonica) and swimming crab (Portunus trituberculatus). Their
studies included threshold of reaction to light, optimum illumination,
reaction to colored light, and the influence of background light on
phototactic behaviour. The exposition of the above-mentioned characteristics
of animals can provide a scientific basis for the selection of the optimum
artificial light source in production, the determination of effective
attraction range, and the calculation of gathering rate.
There were three major findings. First, young fish and adults of round
scad and chub mackerel showed positive phototaxis under the horizontal
photogradient of between 1CT1 and 103|x. Second, the phototaxis of young fish
was stronger than that of adults. Third, the optimum illumination of young
• chub mackerel was 1CT2 to 14 Ix in water temperature of 24.5 to 27°C. In both
single and group experiments, the results of response of chub mackerel are
basically the same, and the response of chub mackerel are basically the same,
and the response for the group experiment is even more stable.
We also did comparison experiments on needlefish, cardinal fish and Mugil
carinatus. the results are: (1) needlefish were always distributed^in the
relatively stronger illumination region of each series of photogradient; (2)
under the same conditions, cardinal fish were constantly distributed in the
relatively weaker illumination region of each series of photogradient; (3) the
phototaxis of the young fish and adults of round scad and chub mackerel and
the young of Mugil carinatus remained between the former two. Chub mackerel
and. round scad both belong to the type of fish that show positive reaction to
light but do not tend to the strongest light. When a comparison was made
between these two, round scad showed stronger phototaxis than chub mackerel.
Therefore, both fish can be the objects of light fishing. The optimum ^
illumination regions of squid and cuttlefish were 10"1 to ICPlx and 10 to 10
Ix, respectively. Both belong to the type that tends to weak light. The
optimum illumination of swimming crab was 10"3 to 10"2lx. With growth ^
development, the optimum illumination of young eels changed from 10" to 10 Ix
209
-------
to 1CT2 to lO^lx. This shows that the character of phototaxis differs in
different stages of development.
Some fish and marine animals can show selective response and color vision
to colored light. Study in this field is of great importance to light source
color selection in production. Under dark adaptation, the phototactic rate of
young and adult round scad is highest to blue-green light while that to red
light is the lowest. After the transition from dark to light adaptation, the
phototactic rate to yellow-green light is the highest. Chub mackerel show the
highest phototactic rate to violet and red lights under both dark and light
adaptations. The peak value of its spectral sensitivity remains to blue and
green lights. The author points out that the optimum light color is different
from the sensitivity light color. The squid shows the greatest phototactic
rate to blue light.
Moonlight, functioning as background light, poses a practical problem in
production. Our tests showed that background light invariably reduced the
phototaxis of the above-mentioned animals; the influence of background light
could be restrained by increasing the light intensity of illumination; the
background light had a stronger influence on the animals of weaker phototaxis
than on those of stronger phototaxis. Within the range of 101 to 103lx,
however, squids showed a higher phototactic rate with background light 'than
without it. Such an unusual phenomenon was associated with the sexual
maturity and moonlight night reproduction of squids. Therefore, contrary to
the common phototactic fish, the yield of squids can be raised if we fish with
a hook and a line at moonlight night during .their breeding season with
properly increased light source intensity. Water temperature can influence
the phototaxis of young round scads and increase their phototaxis within a
relatively lower optimum temperature range.
He Daren and others (1983) also have studied the feeding intensity of
juvenile mullets (Mugil sp_..) under different illumination and probed into the
role of vision during a fish's feeding. The results show that the
illumination intensity has a very strong influence on the feeding intensity of
juvenile mullet to daphnia. Its feeding intensity reaches the maximum under
the illumination of 102lx; the feeding rate reaches the highest under 102lx
and reaches the maximum within 20 minutes.
The peak of feeding activity of young mullet appears only under certain
illumination conditions and is closely related to the feeding activity under
natural conditions and circadian rhythm. Our tests have confirmed the
important role of vision in their feeding activity and provided a basis for
determining the environmental brightness and bait quantity at the time of
throwing in the bait.
Since 1984, He Daren and others have studied the optomotor reaction of
fish. They studied the optomotor reaction characters of the young fish of
gray mullet (Mugil cephalus") , perch (Lateolabrax laponicus'), porgy (Sparus
latus), carp (Cyprinus carpio'), grass carp (Ctenopharvngodon idellusl , etc.
Liu Lidong and others (1986) also have made an overall study on the optomotor
reaction of Nile tilapia and its influencing factors. Their study included
the influences of environmental illumination, the screen speed of rotation,
210
-------
water temperature, structure of visual field, body length, etc., on the
optomotor reaction of tilapia. They also described the characters of reaction
of individual, population, and monoculars. The young fish of this kind has
typical optomotor reaction under the illumination of 1CT5 to 104lx. Within a
certain range, the optomotor reaction of fish increases with rising
environmental illumination and water temperature and declines with increasing
screen rotation speed and body length. The fish reacts most effectively to
black-and-white vertical stripes, less effectively to oblique stripes, and
none to horizontal stripes. The reaction of optomotor reaction intensifies
with the increasing width and number of vertical stripes within a certain
range. There is no obvious difference in the reaction of individuals and
populations. The reaction of monocular fish is evidently weaker than that of
normal-binocular ones, and the former has an obvious orientation. He Daren
and Zhou Shijie have studied the sensibility of young prawn (Penaeus
orientalis) to moving objects and its relationship with environmental
illumination. They also have studied the optomotor reaction of young porgy
under conditions of colored light.
ELECTROPHYSIOLOGICAL STUDIES
For the past dozen years and more, much has been done in the field of
electrophysiology of fish and in marine animal vision. As the phototactic
character of fish is related directly to its vision function, the
determination of the electro-activity of different levels of peripheral and
central neurons is one of the ideal indexes in the study of physiology of
phototaxis. The main work in China is to study the ERG character with rough
electrodes. After an overall study on the ERG of round scad and chub
mackerel, Young Xiongli and others have found that their ERG had the character
of the retina of mixed type and had obvious off-response and that its b-wave
was highly sensitive to the oxygen deficiency; Background light could
noticeably raise the threshold of retina sensitivity. Zheng Weiyun and others
have found through their study on red grouper (Epinephelus akaara) that its
retina has a two photoreceptor system of rod and cone, but they have not found
any typical character of mixed retina in the ERG. They believe that the cone
of such fish has degenerated and is adapted to weaklight vision without the
ability of color discrimination.
Yang Xiongli and others (1978) have determined the spectral sensitivity
of the ERG b-waves of round scads and chub mackerel and found that the peaks
of spectral sensitivity curves of the two fishes under dark vision are 490 and
480nm, respectively. The peaks of spectral sensitivity curves under light
vision shift to 520 and 525 nm, respectively. They also have studied the
induction potential of the tectum of chub mackerel. With the latent period of
induced potential as the index, the spectral sensitivity curves thus obtained
basically correspond with those under ERG dark vision. Therefore, it is
believed that the spectral sensitivity remains unchanged in the visual center.
Zheng Weiyun, Chai Minjuan, Chen Zhong, Liu Lidong, and others have studied
the ERG character of cuttle fish, swimming crab and prawn (Penaeus penillatus)
and showed that the ERGs of these invertebrates are cornea negative wave one
positive off-response. This is identical with the structure and arrangement
of their retina. Through analyzing the ERG spectral sensitivity of mud crab
211
-------
(Scylla serrsta) under different adaptation, Yang Xiongli and others have
suggested that its compound eye has only a single receptive system and
revealed the circadian rhythms of its ERG sensibility. They also have drawn a
schematic diagram for control of circadian rhythms in the ERG of the crab.
Chai Minjuan and others have made some comparative studies on the ERG
circadian rhythms of several kinds of crustaceans (mud crab, tiger head crab
Orithyia mammillaris, swimming crab and prawn). In recent years, scientists
of our country also have begun to study micro-electrodes. Yang Xiongli and
others have recorded the horizontal cell potential of retina on the body of
fresh water fish.
STUDIES ON BIOCHEMISTRY OF VISUAL PIGMENT AND HISTOLOGY OF RETINA
As visual pigment forms the material basis of photoreception of retina,
great importance has been attached to its study. Not much work has been done
in the field of biochemistry of visual pigment; however, Chen Ming and others
have determined the rod visual pigment of round scad, chub mackerel, black
carP (Mylopharyngodon piceus) grass carp, silver carp (Hypophthalmichthvs
malitrix), bighead (Aristichthvs hobilis), tilapia, and others. Black carp,
grass carp, silver carp and bighead have pigment of simple retinene2; the peak
values of their spectral absorption are 530, 528, 525 and 527 nm, respective-
ly. Chub mackerel show the pigment of retinenel and the peak value of its
absorption is 500nm. Round scad contain two kinds of rhodopsin and their peak
values of absorption are 488 and 510nm, respectively. Tilapia contains the
mixed pigment of retinenel and retinene2 and their peak values of absorption
are 500 and 522 nm, respectively.
Zheng Meili and others (1980) have studied the histological structure of
the retina of cuttlefish. Xu Yonggan and others (1986) have made histology-
physiological studies on the retina of gray mullet, porgy, red grouper, round
scad and sardine (Sardinella auritusl. They also have studied submicroscopic
structure and probed into the relationship between the retina morphology and
the function of the five fishes and the ecological environment. Xu Yonggan,
He Daren, and Zhang Houquan have made further studies on the retinomotor
reaction of the young fishes of porgy and gray mullet.
The paper has so far given a brief account of the research work that has
been done in China for the past, dozen years or more in the fields of the
phototactic physiology of fish and other marine animals. There is still much
room for further study. For example, the varieties of fish that have been
studied are still just a few. Round scads and chub mackerel have been studied
through various methods, but most of this fishes have been studied only in
some particular aspects. Few studies have been made on the biochemistry of
visual pigment; a systematic study on the factors that influence the photo-
taxis of fish is still lacking. It also is necessary to probe deeply into the
mechanism of phototaxis. As regards test methods, in the past, more work was
done in the laboratory than actual observation on the sea. Experimental
techniques must be improved--for example, automatic record-analysis equipment
should take the place of naked-eye observation in the study of ethology and
micro-electrodes and ultramicroelectrodes should be used in electrophysioloey
All these must be improved in future work.
212
-------
REFERENCES
Chai, M.J. and W.Y. Aheng. 1982. Spectral sensitivity of ERG of cuttlefish.
Acta Oceanologica Sinica. 4(6):784-787.
Chai, M.J. and W.Y. Aheng. 1983. Preliminary study on colour vision of
cuttlefish (Sepiella maindroni de Rechebrune). Marine Sciences. 5:36-
39.
Cheng, C. and L.D. Liu.
penicillatus.
1982. A preliminary study on color vision of Penaeus
1979.
He, D.R., J.H. Xiao, H.M. Luo, M.L. Zheng, J.X. Xu, andR.S. Huang.
Study on the phototactic behaviour of squid (Loligo duvaucelii
d'Orbigny). Journal of Xiamen University (Natural science) 3:99-104.
He, D.R., H.M. Luo, and M.L. Zheng. 1980. A study on the phototactic
behaviour of sardine (Sardinella perforata Cantor) and silverside
(Atherine bleekeri Gunther). Response to diffused white light and the
process of adaptation to iso-illuminative spectral colour. Universitstis
Amoiensis Acta Scientiarum naturalium. 19(2):81-88.
He, D.R., H.M. Luo, and M.L. Zheng. 1983. The feeding intensity and its
dynamics of juvenile mullet under different illumination. Transactions
of The Chinese Ichthyological Society 3:21-27.
He, D.R., S.J. Zhou, L.D. Liu, H.C. Cai, and W.Y. Zhen. 1985. Study on the
optomotor reaction of some young fishes. Acta Hydrobiologica Sinica.
9(4):365-373.
He, D.R., L.D. Liu, and W.Y. Zheng.
reaction of Tilapia niloticus.
267.
1987. Factor influencing optomotor
Acta Hydrobiologica Sinica. 11(3):255-
Liu, L.D., D.R. He, and W.Y. Zheng. 1986. Electrophysiological studies on
visual characteristics of Mugil cephalus and M. carinatus. Journal of
Xiameti University (Natural Science) 26(2):227-232.
Liu, L.D., D.R. He, and W.Y. Zheng. 1986. The characteristics of optomotor
reaction of young Nile tilapia (Tilapia nilotica). Journal of Xiamen
University (Natural Science). 25(5):580-587.
Luo, H.M., D.R. He, and M.L. Zheng. 1981. Discrimination of sardine and
silverside to spectral colours. Acta Scientiarum Naturalium Univer-
sitatis Amoiensis. 20(2):237-242.
Shanghai Institute of Physiology, Department of Oceanography, Amoy University,
East China Sea Fisheries Research Institute 1975. The preliminary report
on the studies on physiological ecology of phototaxis of round scad and
chub mackerel. Journal of Xiamen University (Natural Science). 1:75-80.
213
-------
Xu, Y.G., D.R. He, M.L. Zheng, and M.J. Chai. 1986. Histological study on
the retina of gray mullet (Mugil cephalus). Transaction of the Chinese
Ichthyological Society. 5:17-26.
Xu, Y.G. and D.R. He. 1988. Preliminary study on the retinomotor responses
of juvenile Sparus latus and Mugil cephalus. Oceanologia et Limnologia
Sinica. 19(2):109-114.
Yang, H.L., C.Y. Li, W.L. Shy, W.Y. Zheng, Y.L. Huang, and C.M. Pan. 1977.
I. General properties of the electroretinograms in the round scad
Decapterus maruadsi and the Japanese mackerel Pneumatophorus laponicus.
Acta Biochimica et Biophysica Sinica. 9(l):25-36.
Yang, H.L., C.Y. Li, W.Y. Zheng, and C.M. Pan. 1977. II. Spectral sen-
sitivities of b-wave of the round scad Decapterus maruadi and the
Japanese mackerel Pneumatophorus 1aponicus. Acta Biochimica et Bio-
physica Sinica. 9(l):37-42.
Yang, H.L., C.Y. Li, C.M. Pan, and Y.L. Huang. 1977. III. General properties
and spectral sensitivities of the visually evoked tectum potential in the
Japanese mackerel Pneumatophorus 1aponicus. Acta Biochimica et Bio-
physica. 9(1):43-51.
Yang, H.L., T.T. Shou, C.Y. Li, W.Y. Zheng, and M.J. Chai. 1978. Changes in
retinal sensitivity of crucian carp during light adaptation. Acta
Biochimica et Biophysica Sinica. 19(1):17-26.
Yang, X.L., T.X. Fan, W.Y. Zheng, M.J. Chai, and L. Zhang. 1984. Circadian
rhythm in the electroretinogram of the crab (Scylla serrata). Acta
Physiologica Sinica. 36(2):202-206.
Yang, X.L., W.Y. Zheng, M.J. Chai, and T.X. Fan. 1985. Single receptor
system in the compound eye of crabs (Scylla serrata). Acta Phyisoloeica
Sinica. 37(1):61-69.
Yang, X.LO., W.Y. Zheng, M.J. Chai, and L. Zhang. 1986. Circadian rhythms of
the ERG of the crab Scylla serrata. Acta zoologica Sinica. 32(2):122-
Yang, X.L., M.J. Chai, and W.Y. Zheng. 1986. Effects on circadian rhythms in
the ERG of the crab by sectioning the optic nerve. Acta Zoologica
Sinica. 32(2):130-136.
Yu, W.T., D.R. He, and Y.S. Chen. 1978. The influence of background light on
the phototactic reaction of round scad (Decapterus maruadsi (T.S.)) and
common japanese mackerel (Pneumatophorus laponicus (Houttuyn)). Journal
of Xiamen University (Natural Science). 4:14-18.
Yu, W.T., D.R. He, and Y.S. Chen. 1979. Phototactic reaction to the iso-
energetic spectral color of round scad (Decapterus maruadsi (T.S.)) and
common Japanese mackerel (Pneumatophorus maruadsi (T.S.)) and common
Japanese mackerel (Pneumatophorus j aponicus (Hottuyn)). Journal of
Xiamen University (Natural Science). 2:126-130.
214
-------
Yu, W.T., D.R. He, and Y.S. Chen. 1982. Studies on the phototactic behavior
of round scad and chub mackerel. Acta Oceanologica Sinica. 1(2):275-
28i;
Zheng, M.L., J.H. Xiao, W.Y. Zheng, M.J. Chai, and D.R. He. 1980. Studies on
the phototactic behaviour of cuttlefish (Sepiella maindroni de Roche-
brune). Journal of Xiamen University (Natural Science). 19(3):91-99.
Zheng, W.Y., and M.J. Chai. 1981. Electrophysiological study on the adapta-
tion of ERG of cuttlefish. Marine Sciences. 4:37-40.
Zheng', W.-Y. and M.J. Chai. 1981. I. General properties of cuttlefish
electroretinogram. Acta Oceanologica Sinica. 3(3):472-476.
Zheng, W.Y., and M.J. Chai. 1982. The retinal adaptive properties of
Protunus trituberculatus. Journal of Xiamen University (Natural
Science). 21(3):363-370.
Zheng, W.Y. 1985. Study on the visual characteristics of red grouper.
Journal of Xiamen University (Natural Science). 24(4):493-500.
Zheng, W.Y. and L. Zhang. 1985. On the receptor systems of compound eyes of
the shrimp (Penaeus penicillatus) and their adaptive properties. Journal
of Xiamen University (natural Science). 24(2):256-261.
Zhou, S.T., and D.R. He. 1987. The sensibility df young prawn to moving
object and its relation with environmental illumination. Marine Scien-
ces. 3:57-59.
215
-------
IDENTIFIED SEX PHEROMONES IN FISH:
ENDOCRINE AND BEHAVIORAL EFFECTS
by
Norman E. Stacey1 and Peter W. Sorensen2
Teleost fish use pheromones for a variety of functions--identifying
individuals, kin groups, and species; attracting conspecifics (i.e., school-
ing); triggering alarm or fright reactions that signal the injury of a
conspecific; and synchronizing reproductive activities both between and within
the sexes (see reviews by Colombo et al. 1982; Lambert et al. 1986; Liley
1982; Liley and Stacey 1983; Stacey et al. 1986, 1987). In nearly all cases,
the evidence for such pheromonal activity comes from studies in which in-
dividuals are exposed to an unpurified conspecific odor (holding water or
tissue preparation). Although such an experimental approach may convincingly
demonstrate pheromonal activity, until a pheromone is identified chemically
and available for testing, it is impossible to conduct carefully controlled
experiments to investigate the mechanism of pheromone action.
In the past few years, we have been fortunate in identifying several sex
pheromones released by female goldfish that trigger physiological and be-
havioral responses in the male (Sorensen and Stacey, in press). The fact that
these pheromones are commercially available sex hormones has greatly facili-
tated our studies. More importantly, our findings raise the possibility that
fish commonly use sex hormones and their metabolites as sex pheromones; if
true, both fundamental and applied aspects of fish sex pheromone research
could expand rapidly in the near future. Indeed, there now seems no doubt
that sex pheromones soon could be applied effectively to the problems of
controlled reproduction of cultured fishes.
FEMALE SEX PHEROMONES IN GOLDFISH
Ovarian growth and vitellogenesis in goldfish occur during fall and
winter (Stacey 1987). In the spring, increasing water temperature and aquatic
vegetation (spawning substrate) trigger ovulation in females with post-
vitellogenic ovaries. Females typically spawn only a few times during the
late spring spawning season. As is typical of many cyprinids, goldfish are
Department of Zoology, University of Alberta, Edmonton AB, Canada
Department of Fisheries and Wildlife, University of Minnesota, St. Paul MN
USA
216
-------
group- or gang-spawners, many males following or chasing an ovulated female so
as to be close when she releases her oocytes in floating vegetation. It is
highly likely that male goldfish have evolved adaptive responses to the odors
of periovulatory females as a result of selective pressures exerted by two key
aspects of goldfish reproduction--the existence of interacting mixed-sex
groups prior to spawning and the intense male competition for females during
spawning. Because many cultured cyprinids display reproductive activities
comparable to those of goldfish, they also may have evolved similar sex
pheromone systems.
The sex pheromones we have identified in female goldfish are released
during the brief periovulatory period that commences with the start of the
preovulatory gonadotropin (GtH) surge and ends with the completion of oviposi-
tion approximately 15 hours later. This ovulatory process is closely
synchronized with the light:dark cycle, the GtH surge commencing in mid-
photophase, ovulation (follicular rupture) occurring in the latter half of
scotophase, and oviposition extending over several hours during early morning
(Stacey et al. 1979). Between the onset of the GtH surge and ovulation,
females release a preovulatory pheromone with "primer" effects on the male
reproductive endocrine system (Stacey et al. in press). Later, between
ovulation and the completion of oviposition, females release a postovulatory
pheromone with "releaser" effects on male sexual behavior (Sorensen et al.
1988). A schematic model of the sequential function of these two pheromones
is presented in Figure 1.
PREOVULATORY PRIMER PHEROMONE
Evidence of a preovulatory primer pheromone in goldfish was first
provided by Kobayashi et al. (1986a) who showed that, in males placed in
contact with preovulatory females, blood GtH rises in synchrony with the
female's GtH surge. Because the odor of preovulatory females was sufficient
to trigger a GtH increase in males and because cutting the olfactory tracts
blocked the response, Kobayashi et al. (1986b) suggested that the GtH increase
in males was triggered by a female sex pheromone. Our studies have confirmed
these findings and demonstrated that the preovulatory pheromone is 17a, 20/3-
dihydroxy-4-pregnen-3-one (17,20/3-P), the ovarian steroid that induces' oocyte
final maturation (migration and breakdown of the nucleus; Nagahama et al.
1983). We have shown that preovulatory females release 17,20/3-P to the water
where it then functions as a pheromone to rapidly increase circulating GtH
levels of males (Stacey and Sorensen 1986, Dulka et al. 1987, Stacey et al. in
press). This increase in male GtH then stimulates an increase in sperm
production. Males exposed to the preovulatory female pheromone thus benefit
from increased stores of releasable sperm (and therefore increased fertility)
by the time the female ovulates and begins to spawn.
Blood levels of 17,20/3-P begin to increase several hours after the start
of the preovulatory GtH surge, reach peak levels 4 hours prior to ovulation,
and return to basal levels near the time of ovulation (Kobayashi et al. 1987,
Stacey et al. in press); peak levels are at least 30 times those preceding and
following the periovulatory surge. Because the periovulatory profile of
17,20/3-P released to the water is essentially the same as that in the blood
217
-------
FEMALES
(Maturation)
PGFs
(Ovulation)
1200
2000
preovulatory
17,206-P
pheromone
Female Sex
Behavior
postovulatory
PGF
pheromone
MALES
Spawning
Synchrony
(Milt)
T
Male Sex
Behavior
1200
2000
0400
Time of Day
1200
Figure 1. Schematic model depicting actions of two hormonal sex
pheromones produced by periovulatory female goldfish. Preovula-
fcory primer pheromone—a preovulatory surge in blood gonado-
tropin (GtH), triggered by environmental cues, stimulates ovarian
synthesis of 17a20g-dihydroxy-4-pregnen-3-one (17,203-P), which
induces oocyte final maturation (resumption and completion of
meiosis). 17,203-P also is released to the water qhere it exerts
a primer pheromone ennect on the male endocrine system, a rapid
increase of blood GtH that stimulates testicular 17,203-P syn-
thesis, leading to increased releasable sperm within 2 to 4
hours. Postovulatory releaser pheromone—mature follicles
rupture (ovulate) 10 to 12 hours after the GtH surge has begun.
Ovulated oocytes in the reproductive tract then initiate and
maintain the synthesis of F prostaglandins (PGFs), which enter
the circulation and act within the brain to trigger female
spawning behavior. PGFs and PGF metabolites also are released
to the water where they function as a pheromone-triggering
male sexual arousal.
218
-------
(Stacey et al. in press), release of the steroid must be very rapid, an
important feature if 17,20/3-P concentrations in the water are to serve as an
accurate indicator of impending ovulation. Nothing is yet known regarding the
mechanism(s) of 17,20/3-P release, although studies of other teleost species
(Colombo et al. 1982) suggest that release in urine is likely.
Extracellular electrical recording from the goldfish olfactory epithelium
(electro-olfactogram, EOG) has demonstrated that water-borne 17,20/3-P is an
extremely potent olfactory stimulant (1 pM threshold, Sorensen et al. 1987 and
unpublished). EOG experiments testing more than 30 steroids have identified
three types of steroid that are goldfish olfactory stimulants (Sorensen et al.
1987 and unpublished)--bile salts such as taurocholic acid; androgens such as
androstenedione and testosterone; and "progestogens" related to 17,20/3-P.
Cross-adaptation experiments, in which the olfactory epithelium is first
adapted to one odorant and then stimulated with another, indicate that the
olfactory receptors responding to 17,20/3-P are different from those responding
to bile salts and androgens. Both the bile salts and androgens are con-
siderably less potent than 17,20/3-P, and the magnitude of the response induced
by androgens is relatively small.
Of the progestogens tested, the three most effective are 17,20/3-P,
17a,20/3,21-trihydroxy-4-pregnen-3-one (17,20/3,21-P) , and 17a-hydroxyproges-
terone (17-P), in decreasing order of potency. Because any other change to
the 17,20/3-P structure, or to its pregnane (A-ring-reduced) derivative, almost
completely destroys activity (Sorensen, unpublished results), it seems likely
that ovarian 17,20/3-P does not require further metabolism in order to function
as a pheromone. Because ovulatory goldfish release equivalent preovulatory
surges of free 17,20/3-P and conjugated 17,20/3-P (likely 17,20/3-P glucuronide) ,
however, the possibility remains that conjugated 17,20y3-P also may have
pheromonal activity, (Stacey et al. in press). Unfortunately, 17,20/3-P
glucuronide is not available for testing. The goldfish olfactory system does
not appear to detect glucuronides of estradiol, testosterone, or etiochol-
anolone (a 5/3-reduced androgen) (Sorensen et al. 1987), despite the finding of
Colombo et al. (1982) that the latter steroid attracts male goldfish.
When 17,20/3-P is added to aquarium water to create supra-threshold con-
centrations , the blood GtH levels of males increase rapidly (within 15 min)
and milt (sperm and seminal fluid) volumes increase within 2 to 4 hours (Dulka
et al. 1987 and unpublished). Blood 17,20/3-P levels of pheromone-exposed
males also increase over this time period (Dulka et al. 1987), supporting the
proposal that testicular 17,20/3-P mediates GtH-stimulated sperm production
(Ueda et al. 1985). Of a variety of steroids tested by addition to aquarium
water, 17,20/3-P is the most potent in inducing the GtH and milt increases,
although related progestogens also are'effective at higher concentrations.
Androgens are completely ineffective in stimulating either GtH or milt volume
increase, indicating their olfactory activity (as shown by EOG) is related to
a different function (Stacey and Sorensen 1986 and unpublished). Both the GtH
and milt responses to water-borne 17,20/3-P are abolished either by complete
section of the paired olfactory tracts or by selective section of the medial
tracts (Stacey and Sorensen 1986, Dulka unpublished). These findings are
consistent with a number of studies showing that olfaction is the dominant
sensory mode for pheromone detection in fish and. that pheromonal information
219
-------
is conducted in the medial olfactory system (Stacey et al. 1986, Kyle et al
1987).
Although simple addition of 17,20/3-P to aquarium water consistently
evokes rapid and dramatic increases in GtH level and milt volume (Stacey and
Sorensen 1986, Dulka et al. 1987), it is likely that this "artificial"
stimulation is not strictly comparable to that produced by natural interaction
with a periovulatory female. For example, the GtH increase induced by chronic
17,20/J-P exposure in aquaria reaches peak levels within 30 minutes and does
not change for at least several hours (Dulka et al. 1987), whereas GtH levels
of males exposed to the odor of females throughout the periovulatory period
continue to increase until ovulation (Stacey et al. in press). Although there
are many possible explanations for this difference between the effects of
artificial and natural 17,20/3-P exposure, we suspect that during artificial
(chronic) exposure, adaptation of olfactory receptors makes the pheromone
ineffective after a short period of time, whereas during natural exposure, the
male is repeatedly stimulated by transient wisps of pheromone, each of which
could independently lead to GtH increase. Indeed, we have calculated (Soren-
sen and Stacey in press) that the amount of 17,20/3-P released by a female
prior to ovulation is sufficient to produce only a small and very short-lived
detectable odor plume, in which case males would not be chronically exposed to
stimulatory 17,20/?-P concentrations.
Also of concern in comparing the effects of artificial and natural
17,20/3-P exposure is whether behavioral interaction with the female, or
exposure to steroids other than 17,200-P, normally contribute to the male's
endocrine response. It seems unlikely that behavioral interaction with the
preovulatory female is an important factor, because direct contact with
periovulatory females induces no greater increase in GtH levels and milt
volumes than does exposure to female odor (Stacey et al. in press). The
potential involvement of steroids other than 17,20/3-P is problematical,
however, because periovulatory females also are known to release large
quantities of 17-P (Van Der Kraak et al. 1989), which, although less potent
than 17,20/3-P, is capable of stimulating both GtH and milt increase (Stacey
and Sorensen 1986 and unpublished).
Recently we have found that the male's response to water-borne 17,20/3-P
is affected by simultaneous exposure to water-borne androgens (unpublished
results). The milt increase induced by 17,20/3-P exposure is inhibited in a
dose-dependent manner by water-borne testosterone and androstenedione.
Androgens probably exert this effect by inhibiting GtH increase, but this has
not yet been examined. The biological function for a potential inhibitory
androgen pheromone is not clear because we do not understand the patterns of
androgens released by male and female goldfish. At present, the simplest
explanation would be that water-borne androgens inhibit inappropriate respon-
ses to the minor amounts of 17,20/3-P that likely are released by males and
nonovulatory females. Thus, only the preovulatory females release sufficient
17,20^-P to overcome the inhibitory effect of released androgens.
220
-------
POSTOVUIATORY RELEASER PHEROMONE
As shown in females of many teleost species (Liley 1982, Liley and Stacey
1983, Stacey et al. 1986), ovulated female goldfish release a pheromone that
rapidly stimulates male courtship behaviors (Partridge et al. 1976, Sorensen
et al. 1986). Our recent work (Sorensen, et al. 1988) indicates that this
postovulatory releaser pheromone is comprised of F2a- series prostaglandins
(PCFs).
A variety of experiments (see Liley and Stacey 1983, Stacey 1987, Stacey
and Goetz 1982) indicate that prostaglandin F2Q! first functions as a hormone
that synchronizes spawning behavior with ovulation and, then, is released to
the water where it acts as a pheromone. The presence of ovulated eggs in the
reproductive tract stimulates synthesis of PGF2o!, which then enters the
bloodstream and is carried to the brain to trigger female sexual behavior.
When oviposition is completed, PGF2o! synthesis decreases, PGF2o, is quickly
removed from the circulation, and female sexual behavior ceases.
Sorensen et al. (1986) provided the first evidence that PGFs could
function as female sex pheromones by showing that male goldfish exhibit
similar behavioral responses to the odors of ovulated and PGF2Q;- injected
females. However, because males exhibited no response when the PGF2os was
simply added to the water of the testing aquaria, they suggested that PGF2Q,
either was metabolized to an active form prior to release or stimulated
production of an unrelated product with releaser pheromone effects . Further
studies using EOG responses to evaluate the olfactory activity of PCs
(Sorensen, et al. 1988) support the former possibility.
When a variety of prostaglandins (PCs) were tested as olfactory
stimulants (Sorensen, et al. 1988), the most potent were found to be PGF2a
(100 pM threshold) and 15 -keto -prostaglandin F2o, (15K-PGF2o!; 1 pM threshold), a
PGF2Q! metabolite recently identified in goldfish (Goetz et al. 1987). Cross-
adaptation studies indicate that these two PGFs stimulate separate classes of
olfactory receptors, which are different from those that respond to 17,20/3-P,
bile salts, and amino acids (Sorensen, et al. 1988). PCs other than PGF2a! and
15K-PGF2o! also are effective in stimulating EOG responses, but only at much
higher concentrations (1-20
Both ovulated and PGF2Q,- injected female goldfish release immuno re active
PGF to the water (Sorensen, et al. 1988). The rate of PGF release increases
significantly at ovulation and continues to increase for at least 6 hours but
rapidly declines to basal rates if ovulated eggs are removed by spawning or
stripping. Because removal of eggs also abolishes female spawning behavior
(Stacey 1987), it appears that a single mechanism, activated by the presence
of ovulated eggs, produces the PGF2a, which first acts as a hormone to
stimulate female sexual behavior and later acts as a pheromone to stimulate
male sexual behavior.
Although there is as yet no direct evidence that ovulated female goldfish
release 15K-PGF2a, there is strong indirect evidence that this is the case.
For example, the potency of the odor of PGF2a- inj ected females (as measured by
evoked EOG responses) is greater than would" be expected if the PGF2fl! had been
221
-------
released without being metabolized, and similar to what would be predicted if-
the PGF2a were converted to 15K-PGF2o, prior to release (Sorensen et al. 1988).
More recently, we (Sorensen et al. 1988, Sorensen et al. in press) have
shown^that both water-borne PGF2a and 15K-PGF2o! stimulate the courtship
behaviors exhibited by males exposed to the odors of ovulated or PGF2Q,-
injected females (Sorensen et al. 1986). When PGFs were added to aquarium
water, grouped males immediately decreased feeding activity and increased
swimming activity and social contact with conspecifics. These studies also
have confirmed and extended the findings of Kyle et al. (1985) that GtH levels
of male goldfish increase during sexual interaction with PGF2a-injected
females. In particular, experiments comparing the effects of PGF exposure on
grouped and isolated males provide strong evidence that the pheromone does not
increase GtH directly, but acts indirectly by stimulating sexual activity
(Sorensen, et al. in press). Thus, the actions of water-borne PGFs are
distinctly different from those of water-borne 17,20y9-P, which induces
equivalent reproductive responses in grouped and isolated males (Stacey and
Sorensen 1986).
An unresolved and potentially important aspect of PGF pheromone function
in goldfish is whether PGF2a and 15K-PGF2a, perform the same or different
functions. Although the presence of separate olfactory receptors for these
PGFs suggests the potential for distinct effects, only qualitative differences
in responses have so far been observed (Sorensen, et al. in press). We
believe that by responding to two PGFs males are able to increase the distance
over which they can locate ovulated females. For example, were males to
employ only a single class of PGF olfactory receptor, saturation would occur
as the male approached the female, with the result that no further increase in
concentration would be detectable. By employing two classes of PGF receptors,
with different sensitivities but similar central effects, males may extend the
range of PGF concentrations over which a gradient can be detected.
FURTHER ASPECTS OF SEX PHEROMONE FUNCTION IN GOLDFISH
We are only beginning to understand the nature of sex pheromone function
in goldfish; new pheromones, and new functions for identified pheromones,
certainly await discovery. For example, the studies of Yamazaki and Watanabe
(1979) suggest that sex-specific goldfish pheromones may be induced by
estrogen and androgen treatment. Also, there are suggestions that the
pheromonal function of 17,20/3-P is more complex than originally expected. We
recently have found that low concentrations of water borne 17,20/3-P increase
the occurrence of ovulation in goldfish (Sorensen and Stacey 1987 and un-
published) , an effect consistent with the fact that 17,20/3-P induces equiva-
lent EOG responses in male and female goldfish (Sorensen et al. 1987). If
this is a pheromonal effect, then the concept of 17,20ft-P as a female sex
pheromone will need to be revised. Indeed, because exposure to water-borne
17,20/3-P increases 17,20/3-P blood levels (and presumably 17,20/3-P release) in
male goldfish (Dulka et al. 1987), it is entirely possible that 17,20/3-P is a
bisexual pheromone, produced by, and acting on, both male and female. Such a
pheromonal mechanism could synchronize the spawning of local populations.
222
-------
HORMONAL PHEROMONES IN .OTHER TELEOSTS
Our findings that goldfish sex pheromones are hormones provides strong
empirical support for theoretical arguments (Doving 1976, Colombo et al. 1982)
that fish commonly use released hormones for intraspecific reproductive
synchrony. Although the generality of hormonal pheromones in fish cannot be
assessed until more species are investigated, in all cases where the chemical
identity of a fish sex pheromone has been proposed, the active substance is a
sex hormone or metabolite (Lambert et al. 1986, Stacey et al. 1987, Sorensen
and Stacey in press).
Although pheromonal effects of progestational steroids (i.e., 17,20/3-P-
like compounds) appear to have been studied only in the goldfish, Van Den Hurk
et al. (1987) propose that glucuronated androgens released by male zebrafish
(Brachydanio rerio) acts as a primer pheromone to trigger ovulation in the
female. Another glucuronated androgen, etiocholanolone glucuronide, has been
suggested to act as a releaser, attracting the ovulated female Gobius joso to
the male's territory (Colombo et al. 1982). In fathead minnows, Pimephales
promelas (Cole and Smith 1987), and milkfish, Chanos chanos (C.D. Kelley, C.S.
Tamaru, and C.-S. Lee, Oceanic Institute, personal communication), male
courtship responses are triggered by the odor of PGF2o!-injected females and
water-borne PGF2a, respectively. Together with the finding that PGFs induce
EOG responses in arctic charr, Salvelinus alpinus (T. Sveinsson and T.J. Hara,
Freshwater Institute, personal communication), these recent studies suggest
that hormonal pheromones are widespread among teleosts.
PRACTICAL APPLICATIONS OF FISH SEX PHEROMONES
There are several reasons for considering the application of sex
pheromones to control certain reproductive processes that presently are
managed by hormone or drug injection. First, because pheromone treatment does
not require the handling of fish, both labor costs and fish stress are
reduced. Second, hormonal pheromone therapies will affect only fully mature
individuals because they simply mimic natural triggers of endogenous reproduc-
tive events (ovulation, spermiation). It should be possible, therefore, to
eliminate a difficulty often associated with injection therapies: determining
whether individual fish are in a reproductive state appropriate for treatment.
Third, present information suggests the level of technical expertise required
to carry out pheromone treatment will be less than that required for treatment
with exogenous hormones. Fourth, fish marketability is not likely to be
affected by a treatment that applies a natural product exogenously, at
concentrations lower than those present in the fish.
We believe that, in each teleost species, reproductive strategy and
reproductive endocrinology work in concert to determine what sex pheromone
functions will evolve (intersexual primer, intersexual releaser, territorial
advertisement, etc.), and what hormones will be selected to play these roles
(Sorensen and Stacey in press). Because the reproductive biology of many
cultured cyprinids is broadly similar to that of the goldfish, we also believe
that all these species use similar sex pheromone systems, and that the
pheromonal effects we have demonstrated in goldfish, therefore, can be readily
223
-------
applied to the reproductive management of 'cultured cyprinids. There seems no
reason why 17,20y3-P could not be used to stimulate spermiation of cyprinid
broodstook. On the other hand, because cyprinid ovulation often is
synchronized with daily photoperiod, further work on the ovulation inducing
effect of water-borne 17,20/3-P is required to establish the most effective
time of exposure, and minimum duration of exposure required.
It also might be possible to advance gonadal development by elevating GtH
levels with longterm progestogen treatment (e.g., daily pheromone pulses).
Prostaglandin pheromones also may be useful in carp culture, one obvious
application being the use of PGF-treated females as pheromone sources to
segregate sexually mature males from mixed-sex schools or to attract and
collect wild males for broodstock. Whether the inhibitory androgens have any
practical application is not yet clear, although they may be useful agents for
delaying spermiation and ovulation to extend the period of egg production.
The reproductive effects of hormonal sex pheromones in goldfish are
dramatic and can be reliably induced with minimal expense and experience. If
the same effects can be demonstrated in farmed cyprinids, a novel approach to
the problem of controlled finfish reproduction may be at hand.
ACKNOWLEDGMENT
This study was supported by grants from the Natural Sciences and En-
gineering Research Council of Canada (N.E.s.), the Alberta Heritage Foundation
for Medical Research (P.W.S.), and the Minnesota Agricultural Experimental
Station.
REFERENCES
Cole, K.S., and R.J.F. Smith. 1987. Release of chemicals by prostaglandin
treated female fathead minnows, Pimephales promelas. that stimulate male
courtship. Hormones and Behavior 21:440-456.
Colombo, L. , P.C. Belvedere, A. Marconato, and F. Bentivegna. 1982.
Pheromones in teleost fish. In: Proceedings of the Second International
Symposium on the Reproductive Physiology of Fish. C.J.J. Richter and
H.J.T. Goos (eds.). Pudoc, Wageningen, The Netherlands, pp. 84-94.
Doving, K.B. 1976. Evolutionary trends in olfaction.
Activity Relationships in Chemoreception. G. Benz
London, pp. 149-159.
In: Structure-
(ed.). IRL Press,
Dulka, J.G., N.E. Stacey, P.W. Sorensen, and G.J. Van Der Kraak. 1987. A
steroid sex pheromone synchronizes male-female spawning readiness in
goldfish. Nature 325:251-253.
224
-------
Goetz, F.W., M. Ranjan, A.K. Berndtson, and P. Duman. 1987. The mechanism
and hormonal regulation of ovulation: the role of prostaglandins in
teleost ovulation. In: Proceedings of the Third International Symposium
, on the Reproductive Physiology of Fish. D.R. Idler, L.W. Grim, and J.M.
Walsh (eds.). St. John's, Newfoundland, pp. 235-238.
Kobayashi, M., K. Aida, and I. Hanyu. 1986a. Gonadotropin surge during
spawning in male goldfish. General and Comparative Endocrinology 62:70-
79.
Kobayashi, M., K. Aida, and I. Hanyu. 1986b. Pheromone from ovulatory female
goldfish induces gonadotropin surge in males. General and Comparative
Endocrinology 63:451-455.
Kobayashi, M., K. Aida, and I. Hanyu. 1987. Hormone changes during ovulation
and the effects of steroid hormones on plasma gonadotropin levels and
ovulation in goldfish. General and Comparative Endocrinology 67:24-32.
Kyle, A.L., N.E. Stacey, R.E. Peter, and R. Billard. 1985. Elevations in
gonadotropin concentrations and milt volumes as a result of spawning be-
havior in the goldfish. General and Comparative Endocrinology 57:10-22.
Kyle, A.L., P.W. Sorensen, N.E. Stacey, andJ.G. Dulka. 1987. Medial
olfactory tract pathways controlling sexual reflexes and behavior in
teleosts. Annals of the New York Academy of Sciences 519:97-107.
Lambert, J.G.D., R. Van Den Hurk, W.J.E.J. Schoonen, J.W. Resink, and P.G.W.J.
Van Oordt. 1986. Gonadal steroidogenesis and the possible role of
steroid glucuronides as sex pheromones in two species of teleosts. Fish
Physiology and .Biochemistry 2:101-107.
Liley, N.R. 1982. Chemical communication in fish.
Fisheries and Aquatic Sciences 39:22-35.
Canadian Journal of
Liley, N.R., and N.E. Stacey. 1983. Hormones, pheromones and reproductive
behavior in fish. In: Fish physiology IXB: Behavior and Fertility
Control. W.S. Hoar, D.J. Randall, and E.M. Donaldson (eds.). Academic
Press, New York. pp. 1-63.
Nagahama, Y., K. Hirose, G. Young, S. Adachi, K. Suzuki, and B. Tamaoki.
1983. Relative in vitro effectiveness of 17a,20/3-dihydroxy-4-pregnen-3-
one and other pregnene derivatives on germinal vesicle breakdown in
oocytes of the ayu (Plecoglossus altivelis), amago salmon (Oncorhynchus
rhodurus), rainbow trout (Salmo gairdneri), and goldfish (Carassius
auratus). General and Comparative Endocrinology 51:15-23.
Partridge, B.L., N.R. Liley, and N.E. Stacey.
in the sexual behaviour of the goldfish.
1976. The role of pheromones
Animal Behaviour 24:291-299.
225
-------
Sorensen, P.W. , N.E. Stacey, and P. Naidu. 1986. Release of spawning
pheromone(s) by naturally ovulated and prostaglandin-injected, nonovu-
lated female goldfish. In: Chemical Signals in Vertebrates 4. D.
Duvall, D. Muller-Schwarze, and R.M. Silverstein (eds.). Plenum Press,
New York. pp. 149-154.
Sorensen, P.W., T.J. Kara, and N.E. Stacey. 1987. Extreme olfactory sen-
sitivity of mature and gonadally-regressed goldfish to a potent steroidal
pheromone, 17a,20/3-dihydroxy-4-pregnen-3-one. Journal of Comparative
Physiology A 16:305-313.
Sorensen, P.W. , and N.E. Stacey. 1987. 17o:, 20/3-dihydroxy-4-pregnen-3-one
functions as a bisexual priming pheromone in goldfish. American Zool-
ogist 27:412.
Sorensen, P.W. , T.J. Kara, N.E. Stacey, and F.W. Goetz. (1988). F pros-
taglandins function as potent olfactory stimulants which comprise the
postovulatory female sex pheromone in goldfish. Biology of Reproduction.
39:1039-1050.
Sorensen, P.W., and N.E. Stacey. (in press). Identified hormonal pheromones
in the goldfish: the basis for a model of sex pheromone function in
teleost fish. In: Chemical Signals in Vertebrates 5. D. McDonald, D.
Muller-Schwarze, and R.M. Silverstein (eds.). Oxford University Press,
London.
Sorensen, P.W. , N.E. Stacey, andK.J. Chamberlain. (in press). Differing
behavioral and endocrinological effects of two female sex pheromones on
male goldfish. Hormones and Behavior.
Stacey, N.E. 1987. Roles of hormones and pheromones in fish reproductive
behavior. In: Psychobiology of Reproductive Behavior: An Evolutionary
Perspective. D. Crews (ed.). Prentice Hall, New York. pp. 28-69.
Stacey, N.E., A.F. Cook, and R.E. Peter. 1979. Ovulatory surge of
gonadotropin in the goldfish, Carassius auratus. General and Comparative
Endocrinology 37:246-249.
Stacey, N.E., and F.W. Goetz. 1982. Role of prostaglandins in fish reproduc-
tion. Canadian Journal of Fisheries and Aquatic Sciences 39:92-98.
Stacey, N.E., A.L. Kyle, and N.R. Liley. 1986. Fish reproductive pheromones.
In: Chemical Signals in Vertebrates 4. D. Duvall, D. Muller-Schwarze,
and R.M. Silverstein (eds.). Plenum Press, New York. pp. 119-133.
Stacey, N.E., and P.W. Sorensen. 1986. 17a,20£-dihydroxy-4-pregnen-3-one: a
potent steroidal primer pheromone which increases milt volume in the
goldfish. Canadian Journal of Zoology 64:2412-2417.
226
-------
Stacey, N.E., P.W. Sorensen, J.G. Dulka, G.J. Van Der Kraak, and T.J. Kara.
1987. Teleost sex pheromones: recent studies on identity and function.
In: Proceedings of the Third International Symposium on the Reproductive
Physiology of Fish. D.R. Idler, L.W. Grim, and J.M. Walsh (eds.).
Memorial University Press, St. John's, Newfoundland, pp. 150-154.
Stacey, N.E., P.W. Sorensen, G.J. Van Der Kraak, and J.G. Dulka. (in press).
Direct evidence that 17a,20j8-dihydroxy-4-pregnen-3-one functions as a
goldfish primer pheromone: preovulatory release is closely associated
with male endocrine responses. General and Comparative Endocrinology.
Ueda, H., A. Kambegawa, and Y. Nagahama. 1985. Involvement of gonadotropin
and steroid hormones in spermiation of the amago salmon, Oncorhynchus
rhodurus. and goldfish, Carassius auratus. General and Comparative
Endocrinology 59:24-30.
Van Den Hurk, R., W.G.E.J. Schoonen, G.A. Van Zoelan, and J.G.D. Lambert.
1987. Biosynthesis of steroid glucuronides in the testis of the
zebrafish. Brachydanio rerio. and their pheromonal function as ovulation
inducers. General and Endocrinology 68:179-188.
Van Der Kraak, G.J., P.W. Sorensen, N.E. Stacey, and J.G. Dulka. 1989.
Periovulatory female goldfish release three potential pheromones:
17cc, 20/3-dihydroxyprogesterone, 17a, 20/3-dihydroxyprogesterone glucuronide,
and 17o;-hydroxyprogesterone. General and Comparative Endocrinology.
73:452-457.
Yamazaki, F., and K. Watanabe. 1979. The role of steroid hormones in sex
recognition during spawning behavior of the goldfish, Carassius auratus
L. Proceedings of the Indian National Science Academy, Part B 45:505-
511.
227
-------
EFFECT OF FEMALE-SPECIFIC PROTEIN COMPOUND
TREATMENT ON HATCHING RATE OF FISH LARVAL AND
FISH GONAD MATURATION (PROTEINS ENGINEER I")
by
Wang Hao and Liu Rongzhen
The effects of partially purified fish gonad protein on increasing oocyte
maturation and hatching rate of fish larval were studied in adult freshwater
and seawater fishes (Hypophthalmichthys molitrix. Cyprinus carpios L,
Ctenopharyngodon idellus. Mugil soiuy. Aristichthys nobilis. Parahramis
perkinensis. Carassius auratus. Monopterus albus).
In our work, female-specific protein compounds from fish gonads were
isolated by using biochemical methods. All experimental fishes received
intramuscular or intraperitoneal female-specific protein injections of 2 to 4
ml extractions for each fish in our experiment. After injection, at 3 to 7
days the hatching rate of fish larval were 89% and 90% in female Aristichthys
nobilis. and the hatching rate of Aristichthvs nobilis larval of control were
65% and 66%, respectively. The results of observation of all experimental
female fish have increased the hatching rate by about 20%. Gonad weights of
experimental fishes were higher than those of controls. Controls received
injections of saline solution only. Our extracted female-specific protein
from fish gonad was very effective. When it was used, it induced oocyte
maturation and increased hatching rate of fish larval, as well as related
phospholipids.
Department of Biology, University of Nanjing, Nanjing, PRC
228
-------
TABLE 1. INJECTING (INTRAPERITONEAL) FEMALE-SPECIFIC
PROTEIN SOLUTIONS AND THE EFFECTS ON OOCYTE
GROWTH IN FEMALE Carassius auratus
No. of fish
2
3
15
16
17
18
19
20
21
Control
6
4
7
9
10
TABLE 2.
Body wt, g Gonad .
wt, g
167 ' 15.4
172 20
170 19
190 15.8
148 14
141 9
198 16.5
200 16 . 5
209 17.5
124 13
129 4.7
147 9
196 9.5
189 6.60
INJECTED FEMALE-SPECIFIC
Mean oocyte
diameter , /un
750
780
775
790
780
805
821.4
821.4
857.7
400
450
560
688.6
610.5
PROTEIN TREATMENT
ON FEMALE Mugll soiuy AND THE EFFECTS ON
HATCHING RATE OF FISH LARVAE
Experiment
A
B
C
D
Control
1
2
3
Dose, ml Fertilization Hatching
rate, % rate of fish
larval , %
2 62.9
2 72.2
2 71.6
2 76
0 70.6
0 58.5
0 40
24
23.4
X
22.5
7.4
6.07
X
229
-------
ANTIFREEZE PROTEIN GENES: PHYSIOLOGICAL REGULATION
AND POTENTIAL VALUE TO THE GENETIC ENGINEERING
OF FREEZE RESISTANT FISH
by
Garth L. Fletcher1, Margaret A. Shears1, Madonna J. King1,
Ming H. Kao1, Peter L. Davies2 and Choy L. Hew3
INTRODUCTION
The freezing point of most aqueous solutions depends upon the
concentration of dissolved solute particles (colligative properties). The
higher the concentration of solute particles the lower the freezing point.
Thus distilled water freezes at 0°C, whereas the body fluids of most fishes
freeze at -0.5 to -0.8°C depending on their salt content. This means that
fishes inhabiting fresh water are in no danger of freezing, unless, of course
the water turns solid. Seawater has a salt concentration that is
approximately three times that of fish body fluids. Therefore, its freezing
temperature ranges from -1.7 to -2°C, depending on salinity.
Since the seawater temperatures in the polar oceans and along the
Northeast Atlantic coast of Canada approach freezing for at least part of the
year, why do not fishes inhabiting these waters freeze and die? Scholander
and his colleagues first attempted to answer the question by studying a number
of marine fish inhabiting the coast of Northern Labrador (Scholander et al.
1957, Gordon et al. 1962). They made two important observations. Some fish
do survive supercooled; that is, at temperatures below their freezing point.
If they come into contact with ice, however, they immediately freeze and die.
Therefore,-it is the combination of supercooling and ice contact that is
lethal.
The second observation made by Scholander's group was that the body
fluids of some fish had the same freezing point as seawater. They termed the
responsible solute "antifreeze." The early work of DeVries and Feeney
(Devries et al. 1970) established that the antifreezes were polypeptides.
Since that time, four distinct classes of antifreeze polypeptides differing in
Sciences Centre, Memorial University of Newfoundland, St. Johns
NF, Canada
Department of Biochemistry, Queens University, Kingston ON, Canada
Department of Biochemistry, Research Institute Hospital for Sick
Children, Toronto ON, Canada
230
-------
carbohydrate content, amino acid composition, protein sequence, and secondary
structure have been isolated from the blood sera of a diverse group of marine
teleosts. These proteins are synthesized in the liver for export to the
blood, which distributes them throughout most of the extracellular space where
they effectively protect the fish from freezing at temperatures as low as the
freezing temperature of seawater (-1.8°C) (Hew and Fletcher 1985, Scott et al.
1986, Davies et al. 1988).
WINTER FLOUNDER ANTIFREEZE PEPTIDES
We have been studying the antifreeze polypeptides (AFP) of the winter
flounder (Pseudepleuronectes americanus) (Figure 1) for the past 15 years (Hew
and Fletcher 1985, Davies et al. 1988). This marine flatfish inhabits shallow
inshore coastal waters of much of "the Northeast Atlantic coast of North
America and many populations are exposed to freezing temperatures (<-0.7°C)
and ice every winter (Fletcher et al. 1985).
The AFP in winter flounder consist of a family of at least seven
independently acting polypeptides ranging in size from 3300 to 4500 daltons.
The two major AFP (3300 daltons) found in the blood plasma have been sequenced
completely and one of them has been crystallized to reveal that they are
amphiphilic a helices. The majority of the hydrophilic amino acid side chains
project along the length of one side of the helix and the opposite side is
predominantly hydrophobic (Figure 2) (Davies et al. 1982; Pickett et al. 1984;
Fourney et al. 1984a; Hew et al. 1986 a,b; Yang et al. 1988). It is believed
that the hydrophilic residues bind (hydrogen bond) to embryonic ice crystals,
thus preventing their growth by blocking the further addition of water
molecules (DeVries and Lin 1977). Yang et al. (1988) recently refined this
model and elaborated on the mechanism by which the antifreeze peptides complex
with ice.
Figure 1. Gyotaku print of a 35-cm winter flounder
(Pseudopleuronectes americanus)(artist Ron Fourney)
231
-------
Th'r2 Asp5 fhr.
13 Asn|6 Thr24 Asn27 Thr35
WINTER FLOUNDER
AFP
Figure 2. Structure of winter flounder AFP. [The alpha-helical
secondary structure of the AFP is indicated by the coil. Residues (•)
with hydrophilic side chains are indicated and those that are postulated
to interact with the ice lattice project downwards. The broken line
between lysine and glutamate indicates an intramolecular salt bridge. The
molecule is about 50 nra long.]
In Newfoundland populations of winter flounder, mature AFP appear in the
blood plasma in November when the water temperature approximates 4 to 6°C,
reach peak values by January, and disappear during May when the temperature
begins to rise above 0°C (Figure 3) (Fletcher 1977). These seasonal .changes
in plasma AFP levels are accompanied by concomitant changes in the winter
flounder's resistance to freezing. Moreover, the winter increase in freezing
resistance over that observed during the summer is identical to the winter
increase in plasma antifreeze activity (Figure 4) (Fletcher et al. 1986).
The environmental and physiological factors regulating the annual AFP
cycle in winter flounder have been reviewed in detail by Davies et al. (1988)
and Fletcher et al. (1988a). All of the available data suggest that the
annual cycle is endogenous, with photoperiod acting via the pituitary gland to
regulate the precise time of onset of AFP production in the Fall (Fletcher
1977; Fourney et al. 1984 b,c). Water temperature does not appear to play a
major role in controlling the time of AFP production, but it does affect the
rate of AFP disappearance from the blood. At the water temperatures that
prevail in Newfoundland in the Spring (March and April) (0 to -1°C) plasma AFP
levels remain at winter values (February) despite the fact that AFP synthesis
has stopped. Only when the water temperatures increase above 0°C will AFP
levels decline (Fletcher 1977, Hew et al. 1986b). This means that the fish
are protected from freezing even when there is an unusually prolonged winter.
ANTIFREEZE PEPTIDES AND ATLANTIC SALMON CULTURE
Approximately 7 years ago, Dr. Arnie Sutterlin brought to our attention
the economic desirability of culturing Atlantic salmon in sea cages along the
Atlantic coastline of Canada. No evidence for the presence of AFP has been
found in any salmonid; therefore, the problems associated with this endeavour
were obvious to all of us: the icy marine environment was lethal to salmon
(Fletcher et al. 1988b). Because the only fish capable of surviving in such
an environment were those possessing antifreeze polypeptides, we decided to
232
-------
-1.8
O
o
I -1.6
LU
H -1.4
CD
N |,
LU '•'
LU
o:
LL.
-1.0
2
CO
5 -0.8
_j
CL
-0.6
_ , i , i , , i i i i i i _
•-,. J — Jv WATER
xX \J TEMP-/ _
" ^ / \ •
\ I \ •'
L X / 1 / "
^J I / • •
r- \'
1 \ A
/ X -' \
/ •- -' \
/—FREEZING \
I TEMP . \
/ T \
/ IK-
,.-/ A v
/ r ;
/ \ -
/ \ :
14
o
o
10 -'•
QL
LU
6 H
CC
LU
2 t
0 *
-2
la" "=
£
^
01
14 ^
LU
N
LU
10 g
LL
h-
6 <
2 1
<
ASONDJFMAMJJA
MONTH
Figure 3. Annual cycles of plasma freezing temperatures and
plasma antifreeze peptide (AFP) levels in Newfoundland winter
flounder. [All values plotted as means ± one standard error (N=5 to 10).]
attempt to make Atlantic salmon more freeze resistant by giving them a set of
antifreeze genes.
DO ANTIFREEZE PEPTIDES CONFER FREEZING RESISTANCE TO FISH?
Because we were embarking on a project to improve the freezing
resistance of Atlantic salmon by winter flounder gene transfer, we needed
direct evidence that the AFP was not species specific and that it would
increase the freezing resistance of salmon.
The lethal freezing temperature of both seawater-acclimated Atlantic
salmon and rainbow trout was approximately -0.75°C (Figure 4) (Fletcher et al.
1988b). When AFP purified from the blood plasma of winter flounder were
injected into the rainbow trout, their freezing resistance increased in direct
proportion to blood AFP levels (Figure 4) (Fletcher et al. 1986). These
results indicate that Atlantic Salmon can be engineered to tolerate subzero
water temperatures and ice if winter flounder AFP genes can be inserted into
the salmon genome and expressed in physiologically significant quantities.
WINTER FLOUNDER ANTIFREEZE GENES
During the course of our studies on the antifreeze peptides and their
regulation in winter flounder, the genes coding for these peptides were
233
-------
-1.6
o
£- -1.4
Q.
UJ
-1.2
CS
N -1.0
UJ
UJ
CO
-0.8
WINTER
FLOUNDER
TROUT
PLUS ANTIFREEZE
/+ ATLANTIC SALMON
*- RAINBOW TROUT
-0.8 -1.0 -1.2 -1.4 -1.6
PLASMA FREEZING POINT (°C)
0 2.5 5.0 12.0
PLASMA ANTIFREEZE (mg/ml)
Figure 4. Relationship between lethal freezing temperatures
and plasma freezing temperatures. [Lethal freezing temperatures
were determined on seawater-adapted fish by gradually lowering the temper-
ature of the aquaria with the use of crushed seawater ice. "Winter" floun-
der are winter flounder acclimatized to winter conditions. "Summer" floun-
der are winter flounder acclimated to summer conditions. Experiments on
salmon and rainbow trout were carried out during late winter. Crosses
represent means ± one standard error. Solid dots represent individual
samples from rainbow trout injected intraperitoneally with winter flounder
AFP. The diagonal line indicates the water temperature at which the fish
should die as predicted from the plasma freezing points. Data summerized
from Fletcher et al. 1986, 1988.]
isolated (Davies et al. 1984). The functional AFP gene is 1 kilobase (kb)
long and contains a 600 base pair (bp) intron between two exons. Genomic
Southern blots show that there are approximately 30 to 40 genes per haploid
genome. The majority of these AFP genes (>20) are present in 7 to 8 kb
elements of DNA, which codes for the major AFP found in flounder blood plasma
(Scott et al. 1985).
In preparation for our gene transfer experiments, one of the genomic
tandem (Bam HI) repeats (7.8kb) (subclone 2A-7) was cloned into the Bam HI
site of plasmid pUC-9 using E. coli as the host to produce large quantities of
the gene (Figure 5). The plasmid containing the antifreeze protein gene was
harvested from the E. coli. purified, and linearized for injection by
digestion with Eco Rl (Fletcher et al. 1988c). Because the gene used for
injection was an entire tandem repeat of the winter flounder antifreeze gene,
it should contain all of the DNA sequences responsible for controlling its
expression (Scott et al. 1985).
TECHNIQUES FOR GENE TRANSFER
From the outset of our studies, we have concentrated on microinjection
techniques to transfer the antifreeze genes to salmon eggs. The most
234
-------
effective place to inject any foreign gene is directly into the nucleus. For
a number of reasons, however, this is not easy to do with salmonid eggs.
Although salmon eggs are very large (5mm diameter) they are also opaque,
making it impossible to see the nuclear area. In addition to this, once the
eggs are fertilized in water, the chorion separates from the perivitelline
membrane and hardens, making it very difficult to penetrate with fine glass
needles. This is unfortunate, for once the perivitelline space has formed,
the germinal disc is always on top of the yolky egg where it can be readily
located for microinjection.
Salmon eggs are fertilized by the entry of a sperm through the
micropyle, which is located close to the dividing egg nucleus (Figure 6)
(Ginsburg 1968, Riehl 1980). Thus it was evident that the best means of
locating the female nucleus was to find the micropyle. Although it took us a
number of spawning seasons to perfect this technique, we can now do it with
relative ease under a dissecting microscope using reflected light. Once the
micropyle is located, a 3 to 5 /am (outside diameter) glass needle can slide
without resistance into the germinal vesicle area of the egg.
One potential problem with using the micropyle as an injection route is
that it may become altered in such a manner as to interfere with the normal
fertilization process. Thus we decided to carry out the microinjections after
fertilization.
Figure 5. Schematic diagram of tandemly repeated antifreeze
gene locus in winter flounder. [B=Bam HI site, E=EcoRl. The dis-
continuity indicates the varied number of repeats in a gene cluster. The
AFP gene (1 kb long) is indicated in the enlargement showing its single
intron and two exons. One of the tandem repeats was cloned into the Bam
HI site of the plasmid Puc 9. The plasmid containing the AFP gene (clone
2A-7) was harvested from E. coli and linearized for injection using re-
striction endonuclease EcoRl.]'
235
-------
B
> ' - - fc .• v«
-•'".:• 7.' -'• ".' •.'. 'JQu,
Figure 6. Schematic of fertilized salmon egg (A) and micropyle (B).
[A—BL=blastodisc, CH=chorion, M=micropyle, PV=perivitelline space, VM=vitelline
membrane, and Y=yolk. B--CA=cortical alveoli, CH=chorion, CY=cytoplasm, FS=fer-
tilizing spermatozoa, MII=metaphase of second maturation division, MC=micropilar
canal, and PBI=first polar body. Some of the ideas for these drawings are from
Ginsburg (1963, 1968).]
Once fertilized, the chorion immediately begins the water hardening
process, thus the time available to use the micropyle for injection is
relatively short. To extend this time, we followed up on Ginsburg's (1963)
technique of fertilizing the eggs in a salmon ringer solution. When this
procedure is used, the eggs are fertilized but not activated and the water
hardening process is delayed until the eggs are placed in.fresh water.
The injection apparatus consists of a 3 to 5 //m (O.D) glass needle
driven by a micromanipulator. The volume of DNA solution injected (2 to 3 nl)
was regulated using short bursts (100 to 1000 ms) of N2 at 200 kPa. The
timing of the bursts was controlled using a Grass Stimulator. The'eggs were
injected with approximately 1 X 106 copies of the linearized antifreeze gene
plus plasmid (pUC-9) through the micropyle within 2 hours of fertilization
and incubated in freshwater at 8°C. '
ANTIFREEZE GENE TRANSFER
Details of the antifreeze gene transfer experiments have been published
(Fletcher et al. 1988c). Of the 1800 Atlantic salmon eggs injected with the
antifreeze genes, approximately 80% survived to hatching. This survival rate
was the same as that of uninjected eggs.
Eight months after injection 30 fingerlings (1 to 2 g each) were
collected for DNA analysis. Genomic DNA was isolated from individual
fingerlings as described by Scott et al. (1985). These DNAs were screened for
the presence of the winter flounder antifreeze gene using genomic Southern
blotting procedures (Southern 1975). In this technique, the high molecular
weight DNA was cut into smaller fragments using restriction enzymes and the
resulting solution was electrophoresed in 0.7% agarose gels. The DNA
fragments, now separated according to size, then were transferred to
236
-------
nitrocellulose and probed for the presence of antifreeze genes using a short
32P labelled segment (2.7 kb) of DNA containing the complete antifreeze gene
(Figure 7). If winter flounder antifreeze gene sequences are present in the
salmon DNA, the 32P labelled probe will bind to them and they can be localized
using autoradiography.
EVIDENCE FOR ANTIFREEZE GENE TRANSFER TO ATLANTIC SALMON
A preliminary screening demonstrated that the DNA from 2 of the 30
salmon fingerlings tested contained the antifreeze genes. A more detailed
analysis of these two positive DNAs (No. 26 and No. 36) along with the DNA of
one negative control (No. 45) was carried out in order to be certain that the
complete antifreeze gene was integrated into the salmon genome (Fletcher et
al. 1988c).
When the salmon DNA was not cut with restriction enzymes, the 32P-
labelled probe only hybridized to large molecular weight DNA, indicating that
the injected genes were present within the salmon genome rather than free
within the cell. When the DNA was cut with restriction endonucleases Sst 1
and Bam HI, the probe hybridized to 2.7 and 7.8 kb DNA segments, respectively.
Segments of this size are what would be predicted if the injected DNA gene
sequence was cut with the same enzymes. When the salmon DNA was cut with
restriction enzyme Hind III, the probe hybridized to DNA fragments of 9.4 kb
and longer in one of the two positive salmon and 13.5 kb and longer in the
other. Because the gene used for injection had only one site at which this
enzyme could cut, the other cut had to be within the salmon DNA itself.
Because we cannot predict where this cut would be, the size of the fragment
would be at least 7.8 kb (the length of the injected winter flounder gene) and
have a variable upper length (Figure 7) (Fletcher et al. 1988c).
The minor bands in the Bam HI and Hind III lanes of 26 and 36 possibly
represent cleavage products of 2A-7 that have been independently incorporated
into the salmon genome. In this situation, one would expect to see fewer of
these accessory bands .in the Sst I lanes than in the other two digests, simply
because there is less chance of the breaks in 2A-7 occurring within the
central 2.7-kb Sst I fragment than within the 7.8-kb Bam HI fragment. When
the overall hybridization signal derived from the 30 to 40 genes in the Sst I
digested winter flounder standard (Figure 7) is considered, it is clear that
more than one copy of the AFP gene had been incorporated into the transgenic
salmon. The exact number of copies per cell must await further research.
The integration frequency of the injected genes in the present study (2
out of 30; 6%) appears to be similar to that observed by other investigators
using microinjection procedures. Hammer et al. (1985) found an integration
frequency of 1 to 13% for several mammals. Maclean et al. (1987) reported a
5% integration frequency of rat growth hormone-mouse metallothionein fusion
gene in the genome of rainbow trout. Similarly, Dunham et al. (1987)
demonstrated, using restriction analysis, the successful integration of a
mouse metallothionein-human growth hormone fusion gene in 2 out of 10 (20%)
channel catfish. Ozato et al. (1986) found evidence to suggest a 30% (10 of
30) integration frequency of a chick S-crystalline gene construct. McEvoy et
al. (1988) using dot blot hybridization procedures suggest that 2 out of 15
237
-------
26
H S «'
23,1 -. %,
9.4-
6.7-
4.4-
* ^-
2.3-
2,0-
ESB
T
j
s s
j | 3^^
Lo^i
probe
2.7 Kb
,S BH
J — 1
E
J
—
Figure 7. Genomic Southern blots of salmon DNAs. [26, 36, and 45 represent
individual salmon. WF=DNA from winter flounder. U=undigested DNA. B,H, and S repre-
sent DNAs digested with restriction enzymes Bam HI, Hind III, and Sst I, respectively.
The origin of the gel (0) is indicated by the arrow, and the length (kb) of DNA stan-
dards are indicated below it. On the right-hand side, the arrow indicates the point
of migration of the AFP gene containing Sst I fragment of flounder DNA and Sst I
fragment of 2A-7. The schematic diagram represents a restriction map of the integra-
ted, linearized plasma 2A-7. Salmon genomic DNA is indicated by the broken line. The
pUC-9 section of 2A-7 is indicated by the double lines. The 1-kb gene is represented
by the small rectangular box. The open areas of the box are the exons (AFP coding
, sequences) and the hatched area is the intervening sequence (intron). The cleavage
sites of restriction enzymes Bam HI (B), Sst I (S), and Hind III (H) are marked. The
2.7-kb Sst I fragment used as a radioactive probe is underlined. This figure is pro-
duced from a combination of Figure 1A and 2 of Fletcher et al. (1988c).]
238
-------
(13%) Atlantic salmon embryos (14 weeks old) contained detectable levels of
the E. coli ft galactosidase-mouse metallothionein promoter fusion gene. Zhu
et al. (1985) reported that 50% (3 out of 6) of their goldfish injected with a
human growth hormone gene construct were transgenic; Chourrout et al. (1986)
claimed an integration frequency of 75% by injecting a human growth hormone
cDNA sequence into rainbow trout eggs.
An excellent study carried out by Stuart et al. (1984) examined 547
adult zebrafish that had been injected as embryos with a linearized bacterial
plasmid (pSV-hygro) and found that only 5% of them retained the foreign DNA.
The amount of foreign DNA present averaged less than one copy per cell. When
20 of their positive (transgenic) fish were crossed to uninjected control
fish, only one consistently transmitted the foreign sequences to its
offspring. One of their Fj progeny was crossed with a control fish to yield
an F2 generation of which 50% contained the foreign DNA sequence. Their
experiments demonstrate unequivocally that injected DNA can be integrated into
the fish genome to give rise to a stable breed of transgenic fish.
ANTIFREEZE GENE EXPRESSION
We are screening the blood of potentially transgenic salmon for the
presence of antifreeze peptides using immunoblotting procedures (Burnette
1981). Samples of blood plasma, approximately 10 /il, are analyzed on a 15%
acrylamide sodium do'decyl sulfate, polyacrylamide gel electrophoresis (SDS-
PAGE) and electrophoretically transferred to nitrocellulose paper.
Immediately after the proteins are transferred, the nitrocellulose sheet is
incubated first in a solution containing bovine serum albumin and then in a
solution containing rabbit anti-flounder AFP antibodies. The antibody-AFP
complex is located on the nitrocellulose sheets by hybridization with 125I-
protein A followed by autoradiography (Hew et al. 1986).
The results obtained from screening blood plasma collected from 46
salmon fingerlings during February 1986 indicated that three of them contained
low but detectable levels of winter flounder AFP. The immunoblotting results
from one of the three AFP-positive salmon are illustrated along with several
AFP-negative fish in Figure 8.
The molecular mass of the presumptive AFP found in the salmon plasma was
approximately 6 kd, twice that of mature AFP found in winter flounder (3.3 kd)
(Figure 8). An explanation for this observation comes from our knowledge of
AFP gene expression in winter flounder. The winter flounder AFP gene used in
this study codes for an 82 amino acid pre-proAFP (Figure 9) (Pickett et al.
1984). The 23 amino acid pre-sequence is cleaved off intracellularly and the
proAFP is secreted into the blood where the pro-sequence is removed within 24
hours to yield a 37 amino acid mature AFP. The molecular size of the AFP-
immunoreactive protein found in salmon plasma is essentially the same as that
of the pro-AFP (6kd) isolated from flounder. This suggests that, although the
Atlantic salmon can properly express the inserted winter flounder gene and
secrete the proAFP into the blood, it lacks the necessary (enzyme)- systems to
process the pro-AFP to mature AFP. A similar result has been obtained in
transgenetic Drosophila melanogaster for which only proAFP was found in the
hemolymph (Rancourt et al. 1987).
239
-------
Studies by Hew et al. (1986b) have demonstrated that proAFP isolated
from winter flounder liver possess significant antifreeze activity (Figure
10). Thus although the proAFP are less effective than mature AFP at
depressing the freezing temperature, they nevertheless would be capable of
conferring increased freezing resistance to salmon.
We currently are attempting to repeat our observations of AFP gene
expression by screening all (>700) of our individually tagged, potentially
transgenic salmon. In this set of observations, DNA will be extracted from
the red blood cells of fish with detectable plasma levels of AFP and analyzed
for AFP genes using genomic Southern blotting procedures. All salmon showing
good evidence for the presence of AFP genes and their expression will be
selected for breeding experiments to determine their inheritability.
HOW CAN AFP GENE EXPRESSION IN SALMON BE IMPROVED?
It is evident from our preliminary screening that transformed salmon
express the winter flounder AFP genes and secrete the proprotein into the
blood. If these results are substantiated by further analysis, it appears
that the level of expression and secretion of the proAFP will be low and not
sufficient to confer a significant increase in the salmon's ability to resist
freezing. Because blood plasma AFP concentrations of 10 mg/ml will be
required for winter survival of sea-cage-farmed salmon in most locations along
2 3456789 10
AFP
STD
Figure 8. Immunoblots of Atlantic salmon plasma. [Blood
plasma from Atlantic salmon that developed from fertilized eggs in-
jected with winter flounder antifreeze genes (clone 2A-7) were ana-
lised on a 15% acrylamide SDS-PAGE and transferred to nitrocellulose
paper. The paper was incubated with anti-AFP antibodies followed by
1251 lableled protein A. Lane 7 AFP standard. All other lanes contain
salmon plasma. The arrow under lane 3 indicates an AFP-immunoreactive
protein of approximately 6 kd. Numbers on the right side of the gel
indicate the location of molecular weight markers.]
240
-------
MET ALA LEU SER LEU PHE THR VAL GLY GLN LEU ILE PHE LEU
23
PHE TRP THR MET ARC ILE THR GLU ALA SER PRO ASP PRO ALA
ALA LYS ALA ALA PRO ALA ALA ALA ALA ALA PRO ALA ALA ALA
44-
ALA PRO ASP THR ALA SER ASP ALA ALA ALA ALA ALA ALA LEU
THR ALA ALA ASN ALA ALA ALA ALA ALA LYS LEU THR ALA ASP
82
ASN ALA ALA ALA ALA ALA ALA ALA THR ALA ARG GLY
PRE PRO AFP = 82 AA
PRO AFP = 59 AA
MATURE AFP = 37 AA
Figure 9. Amino acid sequence of the antifreeze precursor from winter
flounder (HPLC component 6) . [The underlined region indicates the sequence
of the mature protein (Hew et al. 1986).]
O
o
(S)
(f)
LU
FLOUNDER
PRO-AFP
2 3
AFP (mM)
Figure 10. Antifreeze activity of winter flounder
pro-AFP, winter flounder mature AFP, and ocean
pout AFP. [Thermal hypteresis is a measure of. antifreeze
activity. Data summarized from Hew et al. (1986) and Kao
et al. (1986).]
241
-------
the East coast of Canada, the expression of these exogenous genes will have to
be improved.
Studies of the antifreeze genes and their expression in natural
populations provides us with a number of clues as to which way to proceed.
Our research on the antifreeze peptides in a broad variety of taxanomic groups
has led us to suggest that they evolved relatively recently in response to the
appearance of seawater ice (Scott et al. 1986). In many species, the genes
coding for these antifreezes are amplified, suggesting intense selective
pressure to produce large amounts of protein. Recently we have found evidence
that the degree of gene amplification (gene dosage) can be correlated directly
with the level of AFP produced (Fletcher et al. 1985, Scott et al. 1988a Hew
et al. 1988).
Because gene amplification appears to be a common mechanism by which
fish have increased their circulating AFP levels, it seems reasonable to look
for ways to increase AFP gene copy number in salmon.
The winter flounder DNA sequence used in the present study represents
one of the tandemly repeated genes; therefore, it may well contain elements
responsible for gene amplification. If such sequences are present, they could
lead to a similar expansion of AFP gene number in transformed salmon. Our
breeding experiments with the transformed salmon should determine whether this
will occur. These experiments, however, may take some time for Stuart et al.
(1988) found, using zebrafish, that it took until the F2 generation to produce
a stable line of transgenic fish.
It does not seem wise to rely entirely on the winter flounder AFP genes
to transfer freezing resistance to salmon. For one thing, the proAFP produced
by the transformed salmon is a less active antifreeze than the mature protein.
Furthermore, we recently have found that the winter flounder AFP genes are
regulated by growth hormone (Vaisius et al. 1988). Growth hormone must be
absent, or at low levels, in the plasma before AFP genes can be transcribed.
If this regulatory system is passed on to the salmon, there may be conditions
under which growth hormone levels may not be low enough during winter for
transferred AFP genes to be transcribed.
Recently, we have been considering the potential value of wolffish and
ocean pout AFP genes as candidates for gene transfer to salmon. Wolffish
(Atiarhichas lupus) and Ocean pout (Macrozoarces americanus") are members of
different families from the suborder Zoarcoidea. Both species produce high
concentrations (20 to 30 mg/ml) of AFP (6 kd), which differ markedly from that
of the winter flounder in that they are not a helical and do not possess an
abundance of alanine (Hew et al. 1984, 1988; Scott et al. 1988b).
There are four reasons why the AFP from wolffish and ocean pout may be
useful for gene transfer. First, both species appear to be expressing their
AFP genes constitutively for they maintain significant plasma concentrations
all year round (Fletcher et al. 1985 and unpublished data). Furthermore,
plasma AFP levels, at least for the ocean pout, are controlled to some degree
by water temperature. This is not the case with winter flounder where AFP
genes are controlled by the pituitary. Second, the fact that they are not
242
-------
biased, in the way winter flounders are, -to the use of alanine might make it
easier for the salmon to synthesize them in large quantities. Third, neither
the wolffish nor the ocean pout produce pro AFP. Therefore, no special
mechanisms would he required by the salmon to produce fully efficient mature
antifreeze. Fourth, on a molar basis, the ocean pout AFP, and most likely the
wolffish AFP, are approximately 40% more active than the mature winter
flounder AFP (Figure 10).
The AFP genes from the wolffish and the ocean pout are encoded in large
multigene families of approximately 80 and 150 members, respectively (Hew et
al. 1988, Scott et al. 1988b).. Detailed studies of AFP gene organization in
the wolffish indicate that two-thirds of the 80 to 85 genes are organized in 8
kb tandem repeats, each of which contains two genes in inverted orientation.
These repeats are clustered in groups of at least six (Scott et al. 1988b). A
restriction map of one genomic clone (Wr4A) containing two-and-one-half
repeats (five AFP genes) is illustrated in Figure 11. We are now injecting
this 16 kb stretch of wolffish DNA into the salmon eggs. By using this
relatively long sequence, we are not only attempting to directly increase the
AFP gene copy number incorporated into the salmon genome but are also
increasing the potential for gene expansion or amplification by unequal
crossing over during meiosis (Hew et al. 1986a).
In summing up, it is apparent that, as we learn more about the
regulation of antifreeze genes and their expression in natural populations, we
will devise more appropriate gene constructs for engineering freeze-resistant
fish. Although we look forward to the ideal result of finding that all
transgenic fish produce enough antifreeze protein to survive down to the
freezing point of seawater, it is more realistic, however, to anticipate that
there will be a more limited expression of the transferred genes and that the
production of freeze resistant salmon will have to rely on appropriate
selection and mating techniques.
SHSa SH SaHS
XWr4A
BH SHSa SH SaH
BH SHSa
B=BamHI H=HindIE
= SstI Sa = SalI
WOLFFISH
Figure 11. Restriction map of a wolffish genomic clone containing
five antifreeze protein genes. [Arrows indicate positions of the' genes
and the directions of transcription.
243
Redrawn from Scott et al. (1988).]
-------
ACKNOWLEDGMENTS
We would like to dedicate anything of merit in this work to Dr. Wu Shan-
Chin formally of the Institute of Oceanology Academica, Sinica, Tsingtao
China. Dr. Wu spent 6 weeks in Newfoundland where she taught us a little egg
surgery and a great deal about patience and dedication to science. This
research was supported by strategic and operating grants from the Natural
Sciences and Engineering Research Council of Canada and by the Medical
Research Council of Canada. (Ocean Sciences Centre Contribution No. 19)
REFERENCES
Burnette, W. N. 1981. Western blotting electrophoretic transfer of proteins
from sodium dodecyl sulfate-polyacrylamide gels to unmodified
nitrocellulose and radiographic detection with antibody and radio
iodinated protein A. Anal. Biochem. 112:195-203.
Chourrout, D., R. Guyomard, and L. M. Houdebine. 1986. High efficiency gene
transfer in rainbow trout (Salmo gairdneri Rich.) by microinjection into
egg cytoplasm. Aquaculture 51:143-150.
Davies, P. L., C. L. Hew, and G. L. Fletcher. 1988.
proteins: Physiology and evolutionary biology.
2617.
Fish antifreeze
Can. J. Zool. 66:2611-
Davies, P. L., C. Hough, G. K. Scott, N. Ng, B. N. White, and C. L. Hew.
1984. Antifreeze protein genes of the winter flounder. J. Biol. Chem.
259:9241-9247.
Davies, P. L., A. H. Roach, and C. L. Hew. 1982. DNA sequence coding for an
antifreeze protein precursor from winter flounder. Proc. Natl. Acad
Sci. USA 79:335-339.
DeVries, A. L., S. K. Komatsu, and R. E. Feeney. 1970. Chemical and physical
properties of freezing point depressing glycoproteins from Antarctic
fishes. J. Biol. Chem. 245:2901-2913.
DeVries, A. L., and Y. Lin. 1977. Structure of a peptide antifreeze and
mechanism of adsorption to ice. Biochim. Biophys. Acta 495:380-392.
Dunham, R. A., J. Eash, J. Askins, and T. M. Townes. 1987. Transfer of the
metallothionein-human growth hormone fusion gene into channel catfish.
Trans. Am. Fish. Soc. 116:87-91.
Fletcher, G. L. 1977. Circannual cycles of blood plasma freezing point and
Na+ and Cl" concentrations in Newfoundland winter flounder
fPseudopleuronectes americanus^: Correlation with water temperature and
photoperiod. Can. J. Zool. 55:789-795.
244
-------
Fletcher, G. L., K. Haya, M. J..King, and H. M. Reisman. 1985. Annual
antifreeze cycles in Newfoundland, New Brunswick and Long Island winter
flounder (Pseudopleuronectes americanus). Mar. Ecol. Prog. Ser. 21:
205-212.
Fletcher, G. L., D. R. Idler, A. Vaisius, and C. L. Hew. 1988a. Hormonal
regulation of antifreeze protein gene expression in winter flounder.
Fish. Physiol. Biochem. In press.
Fletcher, G. L., M. H. Kao, and J. B. Dempson. 1988b. Lethal freezing
temperatures of Arctic char and other salmonids in the presence of ice.
Aquaculture. 71:369-378.
Fletcher, G. L., M. H. Kao, and R. M. Fourney. 1986. Antifreeze peptides
confer freezing resistance to fish. Can. J. Zool. 64:1897-1901.
Fletcher, G. L., M. A. Shears, M. J. King, P. L. Davies, and C. L. Hew.
'1988c. Evidence for antifreeze protein gene transfer in Atlantic salmon
(Salmo salar). Can. J. Fish. Aquat. Sci. 45:352-357.
Fourney, R. M., G. L. Fletcher, and C. L. Hew. 1984b. The effects of long
day length on liver antifreeze mRNA in the winter flounder
(Pseudopleuronectes americanus). Can. J. Zool. 62:1456-1460.
Fourney, R. M., G. L. Fletcher, and C. L. Hew. 1984c. Accumulation of winter
flounder antifreeze messenger RNA after hypophysectomy. Gen. Comp.
Endocrinol. 54:392-401.
Fourney, R. M., C. L. Hew, S. B. Joshi. and G. L. Fletcher. 1984a.
Comparison of antifreeze peptides from Newfoundland, Nova Scotia,
New Brunswick and Long Island winter flounder. Comp. Biochem. Physiol.
78B:791-796.
Gordon, M. S., B. N. Andur, and P. F. Scholander. 1962. Freezing resistance
in some northern fishes. Biol. Bull. 122:52-62.
Ginsburg, A. S. 1963. Sperm-egg association and its relationship to the
activation of the egg in salmonid fishes. J. Embryol. Exp. Morph.
11:13-33.
Ginsburg, A. S. 1968. Fertilization in fishes and the problem of polyspermy.
Academy of Sciences of the USSR, Institute of Development Biology.
Translated from Russian by Z. Blake and edited by B. Golek. (Israel
Program for Scientific Translations, Jerusalem 1972, Keter Press
Binding: Wiener Binding Ltd., Jerusalem. 366p.)
Hammer, R. E., V. G. Pursel, C. E. Rexford, Jr., R. J. Wall, D. J. Bolt, K. M.
Ebert, and R. L. Brinster. 1985. Production of transgenic rabbits,
sheep and pigs by microinjection. Nature (Lond.) 313:680-683.
245
-------
Hew, C. L. , and G. L. Fletcher. 1985. Biochemical adaptation to the freezing
environment- structure, biosynthesis and regulation of fish antifreeze
polypeptides . In: Circulation, Respiration and Metabolism. R Gilles
(ed). Springer -Verlag, Berlin Heidelberg, pp. 553-563.
Hew, C. L. , G. K. Scott, and P. L. Davies. 1986a. Molecular biology of
antifreeze. In: Living in the Cold: Physiological and Biochemical
Adaptations. H. C. Heller et al. (eds.). Elsevier, New York. pp. 117-
Hew, C. L., D. Slaughter, S. B. Joshi, G. L. Fletcher and V. S.
Ananthanarayanan. 1984. Antifreeze polypeptides from the Newfoundland
Ocean pout, Macrozoarces americanus: Presence of multiple and
compositionally diverse components. J. Comp. Physiol. B. 155:81-88.
Hew, C. L. , N. C. Wang, S. Joshi, G. L. Fletcher, G. K. Scott, P. H. Hayes, B
Buettner, and P. L. Davies. 1988. Multiple genes provide the basis for
antifreeze protein diversity and dosage in the ocean pout, Macrozoarces
americanus . J. Biol. Chem. 263:12049-12055.
Hew, C. L., N. C. Wang, S. Yan, H. Cai, A. Sclater, and G. L. Fletcher.
1986b. Biosynthesis of antifreeze polypeptides in winter flounder:
Characterization and seasonal occurrence of precursor polypeptides
Eur. J. Biochem. 160:267-272.
Kao, M. H., G. L. Fletcher, N. C. Wang, and C. L. Hew. 1986. The
relationship between molecular weight and antifreeze polypeptide
activity in marine fish. Can. J. Zool. 64:578-582.
MacLean, N. , D. Penman, and Z. Zhu. 1987. Introduction of novel genes into
fish. Biotechnology 5:257-261.
McEvoy, T., M. Stack, B. Keane, T. Barry, J. Sreenan and F. Gannon. 1988.
The expression of a foreign gene in salmon embryo. Aquaculture. 68:27-
Ozato, K. , H. Kondoh, H. Inohara, T. Iwamatsu, Y. Wakamatsu, and T. S. Okado
1986. Production of transgenic fish: Introduction and expression of
chicken 5-crystalline gene in medaka embryos. Cell. Differ. 19-237-
244 .
Pickett, M. , G. Scott, P. Davies, N. Wang, S. Joshi, and C. L. Hew. 1984.
Sequence of an antifreeze protein precursor. Eur. J. Biochem. 143:35-38.
Rancourt, D. E. , V. K. Walker, and P. L. Davies. 1987. Flounder antifreeze
protein synthesis under heat shock control in transgenic Drosophila
melanogaster. Mol. Biol. 7:2188-2195.
Riehl, R. 1980. Micropyle of some salmonins and coregonins . Environ Biol
Fishes. 5:59-66.
246
-------
Scholander, P. F., L. Van Dam, J. W. Kanwisher, H. T. Hammel, and M. S.
Gordon. 1957. Supercooling and osmoregulation in Arctic fish. J.
Cell. Comp. Physiol. 49:5-24.
Scott, G. K., P. L. Davies, M. H. Kao, and G. L. Fletcher. 1988a.
Differential amplification of antifreeze protein genes in the
pleuronectinae. J. Mol. Evol.27:29-35.
Scott, G. K., G. L. Fletcher, and P. L. Davies. 1986. Fish antifreeze
proteins: recent gene evolution. Can. J. Fish. Aquat. Sci. 43-1028-
1034.
Scott, G. K., P. H. Hayes, G. L. Fletcher, and P. L. Davies. 1988b. Wolffish
antifreeze protein genes are primarily organized as tandem repeats that
each contains two genes in inverted orientation. Mol. Cell Biol
8:3670-3675.
Scott, G. K., C. L. Hew, and P. L. Davies. 1985. Antifreeze protein genes
are tandemly linked and clustered in the genome of the winter flounder.
Proc. Natl. Acad. Sci. USA 82:2613-2617.
Southern, E. M. 1975. Detection of specific sequences among DNA fragments
separated by gel electrophoresis. J. Mol. Biol. 98:503-517.
Stuart, G. W., J. V. McMurray, and M. Westerfield. 1988. Replication,
integration and stable germ-line transmission of foreign sequences
injected into early zebrafish embryos. Development. 103:403-412.
Vaisius, A., G. L. Fletcher, and D. R. Idler. 1988. Growth hormone represses
transcription of an antifreeze protein gene in winter flounder. Proc.
XVIth International Congress of Genetics Aug. 20-27, Toronto ON Canada.
GENOME, 30:SUPPL 1. 30.12.10. pp. 64.
Yang, D. S. C., M. Sax, A. Chakrabartty, and C. L. Hew. 1988. Crystal
structure of an antifreeze polypeptide and its mechanistic implications.
Nature (London). 333:232-237.
Zhu, Z., G. Li, L. He, and S. Chen. 1985. Novel gene transfer into
fertilized eggs of goldfish (Carassius auratus L. 1758). J. Appl.
Ichthyol. 1:31-33.
247
-------
BIOENERGETICS MODELING OF FISH GROWTH
by
Cui Yibo1 and R. J. Wootton2
INTRODUCTION
Modeling the growth of fish is of great interest to fish biologists.
Adequate growth models are needed for predicting the production of fish
populations, which is important in the management of fisheries and for an
understanding of the role of fish populations in aquatic ecosystems.
A popular approach to modeling fish growth is to regard the growth rate
as a function of body size, such as the von Bertalanffy model. This approach
does not explicitly consider the effects of causal factors. The model only
reflects the growth of the fish under the average conditions of the water body
concerned; it is descriptive rather than predictive.
Bioenergetics modeling has provided a promising alternative for
predicting fish growth in varying environments. This approach considers the
effects of causal factors--usually rate of food consumption, water temperature
and body size--on each component of the energy budget of fish and integrates
these effects to predict a growth rate. Bioenergetics models have been
applied to a number of fish species (Kitchell et al. 1974, 1977; Kitchell and
Breck 1980; Rice et al. 1983; Stewart et al. 1983; Diana 1983; Bevelhimer et
al. 1985; Cuenco et al. 1985a,b,c; Stewart and Binkowski 1986) and have been
applied to natural systems (Cochran and Rice 1982, Hurley 1986, Kitchell and
Crowder 1987, Carline 1987). There has not been much improvement in the
structure of the model, however, since it was first developed over a decade
ago. Before we can be sure of extensive applications of this approach, the
assumptions adopted in the model need to be justified, and the model has to be
validated against data from independent, rigorous experiments.
The purpose of this study was to evaluate the use -of bioenergetics
models for predicting fish growth. A bioenergetics model was developed for
the European minnow, Phoxinus phoxinus (L.). Using this model, the impacts of
some assumptions commpnly adopted in other bioenergetics models were assessed.
Data from an independent experiment was used to validate the model.
Institute of Hydrobiology, Academia Sinica, Wuhan, PRO
Department of Zoology, University College of Wales, Aberystwyth, UK
248
-------
METHODS AND RESULTS
MODEL DEVELOPMENT
According to the energy budget of fish, growth can be expressed as:
G = C-F-U-SDA-Rs-Ra
where G is growth, C is food consumption, F is fecal production, U is
excretion, SDA is specific dynamic action, Rs is standard metabolism, and Ra
is activity metabolism (Stewart et al. 1983). In the model, each of the
components on the right side of the equation is regarded as a function of a
series of predictor variables--i.e., rate of food consumption, temperature and
body weight. Sub-models also were developed for the dry matter content and
energy content of fish, so that growth in energy can be converted into wet
weight.
Most parameters in the model were derived from a series of controlled
experiments, referred as "original experiments" in this paper. Procedures for
these experiments were described elsewhere (Cui 1987; Cui and Wootton
1988a,b,c; in press a,b), and only a summary is given below. In each
experiment, 25 minnows were equally divided into five ration groups--0, 1, 2
and 4% of initial body weight per day and ad libitum. Each experiment lasted
21 days, during which each fish was kept in an individual aquarium and fed
daily the prescribed quantity of Enchytraeus worms. Food consumption, fecal
production, excretion, dry matter content and energy content were determined
directly, and metabolism was estimated indirectly as the difference between
consumption and other components.
The metabolic rate of the starved fish was assumed to be the standard
metabolism. SDA was assumed to be 15% of food energy, an assumption adopted
in most other bioenergetics models for fish. Equations relating' each of the
model components were calculated from the experimental data by multiple
regression. They are summarized in Table 1.
The relative dry matter content (% wet weight) and energy content (cal
mg"1 dry matter) of the fish at the end of the original experiment were termed
equilibrium dry matter content (DRYE) and equilibrium energy content
(ENERGYE), respectively. The dry matter content on day t (DRYt) was
calculated as:
DRYt = DRY,..! + (DRYE-DRYt.^/21
The energy content on day t (ENERGYt) was calculated in a similar way.
The total energy content on day t (Et, cal) was:
Et = Et^+G
E was converted to wet weight (W) by
W = (E/DRY/ENERGY)xlOO
A computer program, written in FORTRAN 77, was used to calculate the
growth of fish based on a time-step of 1 day.
EVALUATION OF ASSUMPTIONS
The assumptions commonly adopted in other bioenergetics models for fish
include: (1) fecal production (F) is a constant fraction of food consumption
(C), (2) excretion (U) is a constant fraction of C, (3) activity metabolism
249
-------
(Ra) is a fixed multiple of standard metabolism (Rs), (4) dry matter content
and energy content of'the fish are constant. These assumptions were not used
in the model for the minnow. To assess their implications, the minnow model
was modified to adopt these assumptions.
In the original experiments, the average fractions of food energy lost
in feces and excreta were 0.0652 and 0.0511, respectively; the average dry
matter content and energy content for all the fish used were 25.77% and 5 02
cal mg , respectively. These values were used in the modified models.
Thus, to adopt the assumption that F is a constant fraction of C,,the
sub-model for fecal production (Equation 3 in Table 1) was substituted'by
F - 0.0652C.
To adopt the assumption that U is a constant fraction of C, Equations 4
and 5 in Table 1 were substituted by
U - 0.0511C.
TABLE 1. EQUATIONS USED IN THE BIOENERGETICS MODEL FOR THE MINNOWa
No.
1
2
3
4
5
6
7
8
Equation
Cmax = o.017W°-806T1.22
F = 0.032C0.863260.1093T
U = -2. 0518+0. 0.528C+0.2338T, when C>0
U = 9xlO-5wO-703e0.549T-0.018T2, when C=0
SDA = 0.15C
Rs = 0.0162wO-8108e0.1001T
Ra - -24. 3989+0. 465C+0.017W+0.2739T2-0.00258WT
DRYE = exp(3.5117-0.05711n(FL=l)+0.03731n(FL+lUnT
r2
0.922
0.824
0.911
0.874
_
0.557
0.753
n ATI
-0.01921nWlnT)
ENERGYE = exp(5.6722-0.48161nW-l.75671nT-0.0071Ln(FL-KL)lnW'
+0.20331nWlnT+0.00491n(FL+l)InWlnT)
0.489
aAlso shown are the coefficients of determination (r2) of the equations
fitted to the original data. Cmax: maximum food consumption (cal d"1);
W: body weight (mg); T: temperature (°C); F: fecal production (cal d"1) ;
C: food consumption (cal d"1); U: excretion (cal d-1); SDA: specific
dynamic action (cal d"1); Rs: standard metabolism (cal d"1) ; Ra: activity
metabolism (cal d l); DRYE: equilibrium dry matter content (% body weight);
ENERGYE: equilibrium energy content (cal mg"1 dry matter): FL = feeding
level - 100xCmax(%).
250
-------
To adopt the assumption of constant dry matter and energy content, the
dry matter content and energy content of fish were assumed to be constant
throughout the simulation at 25.77% and 5.02 cal mg"1, respectively.
To adopt the assumption that activity metabolism is a fixed multiple of
standard metabolism, Equation 8 in Table 1 was substituted by
Ra = AxRs
where A = 0,1,2 or 3 and was constant in each simulation.
These modified versions of the model were used to predict the 21-day
growth by a fish with an initial weight of 2500 mg at 10°C. Three ration
levels were used: starvation, 3% of initial weight per day, and ad libitum.
The growth predicted from each modified model was compared with that from the
original model (Figure 1). The assumptions that fecal production and
0%
-500
2000-
1600-
1200-
800-
400
MAX
3456
Assumption
Figure 1. Effect of different assumptions on the 21-day
growth of a 2500 mg minnow at 10 °C at three rations
(% initial body weight day~l) predicted by the bioener-
getics model. Assumptions: 1. nominal model, 2. F=0.0652C,
3. U=0.0511C, 4. dry matter and energy constents are con-
stant, 5. Ra=0, 6. Ra=Rs, 7. Ra=2Rs, 8. Ra=3Rs.
251
-------
excretion are constant fractions of food energy caused only small deviations
in the predicted growth; the other assumptions caused large deviations.
VALIDATION OF THE MODEL
Data from an independent experiment, referred to as "test experiment,"
was used to validate the bioenergetics model for the minnow. The experiment
was carried out at 10°C (Cui 1987). Thirty-six minnows were equally divided
into three ration groups: starvation, 3% of initial weight per day, and ad
libitum. The experiment lasted 21 days. Growth, food consumption, fecal
production, excretion and dry matter content of each fish were determined.
The energy content was not determined but calculated from an empirical
relationship between dry matter content and energy content (Cui and Wootton
1988c). Temperature, -initial body weight, ration, and initial dry matter and
energy content (estimated from a group of control fish sacrificed at the start
of the experiment) were input into the model. The predicted growth from the
model was plotted against the observed, growth (Figure 2). The model gave
reasonable predictions for the starved fish but over-estimated growth by the
feeding fish, particularly fish fed ad libitum.
REASONS FOR MODEL FAILURE
Validation of the bioenergetics model for the minnow showed that the
model was not successful in predicting the growth of fish in the "test
experiment." As the model gave reasonable predictions for the starved fish,
failure to predict growth at higher rations may be caused by an inadequate
simulation of the growth-ration relationship. A simulated relationship
between specific growth rate (SGR), conversion efficiency (K), and ration is
shown in Figure 3. SGR was calculated as: 100x(lnWfc-lnW0)/t where Wt is the
predicted final, W0 is the initial weight of fish (mg), and t is the period of
growth (day). K is calculated as: 100x(Wt-W0)/CT where Ct is the total amount
of food consumed (mg). In the simulations, the initial weight of the fish was
assumed to be 2500 mg; the initial dry matter content and energy content were
25.77% and 5.02 cal mg"1, respectively; the growth period was 21 days; ration
ranged from starvation to ad libitum: four temperatures were used: 5, 9, 12
and 15°C.
At all temperatures, the predicted SGR increased with increased ration.
There was no approach to an asympote at high rations. The predicted
conversion efficiency also increased with increased ration without a decrease
at high rations. These predictions did not agree with results from most
empirical studies, which showed that the relationship between SGR and ration
in fish is a decelerating curve and the relationship between conversion
efficiency and ration is an inverted "U"-shaped curve (Brett and Groves 1979,
Cui and Wootton, in press b).
To find out why the model failed to predict the growth of fish in the
test experiment, observed values for some of the energy budget components in
the experimental fish were used in the model instead of the predictive sub-
models. The resulting "models" were used to predict the growth of the fish
and the predictions were compared with those from the original model. If use
of observed values for a component yields a great improvement in the
252
-------
predictions, then errors due to the sub-model for this component may be an
important source for the errors in the growth predictions.
Data for the 12 fish fed ad libitum in the test experiment were used for
this analysis. The total summed growth by the 12 fish was compared with that
predicted by each of the modified "models" as well as the original model
(Table 2). Use of observed values for fecal production and excretion resulted
in little improvement in the predictions, whereas use of observed values for
food consumption and dry matter and energy content yielded great improvements.
This indicated that errors in growth predictions may be largely caused by the
errors in the sub-models for consumption and dry matter and energy content.
-200
0>
~ -400
£ -400
O
73
-------
2 -i
tt 1
a
w
o -
-1
4 8
Ration
12
20-
10-
5
Ration
10
Figure 3. Relationship between specific growth rate (SGR, % day"1),
conversion efficiency (K, %), and ration (% body weight day1)
predicted by the bioenergetics model for the minnow at four
temperatures—• 5 °C, A9 °C, X i2 °C,H15 °C.
As food consumption is strongly related to body weight in the model, the
over-estimation of consumption may be a result of the over-estimation of
growth due to the poor estimates of other components. This was proved to be
true as the average predicted ration, when expressed as a percentage of mean
body weight, was within the range observed for the 12 fish in the test
experiment. (The average predicted ration was 6.2%, while the observed values
ranged from 4-7%.) When the observed values for all these components were
used in the model, there was still a substantial deviation in the predicted
growth, indicating that the sub-models for metabolism, which was not directly
determined in the experiment, also are important in causing the prediction
.errors. As the model gave reasonable predictions for the starved fish, the
sub-model for standard metabolism should be suitable; the errors lie in the
sub-models for SDA and activity metabolism.
254
-------
TABLE 2. ERRORS IN THE PREDICTIONS OF THE GROWTH OF THE 12 MINNOWS FED
ad libitum IN THE TEST EXPERIMENT FOR "MODELS" USING OBSERVED
VALUES FOR1 DIFFERENT COMPONENTS
Predicted growth
"Model" in which observed values
were used for the following
Observed Original component(s)a
growth model C F+U DE C+F+U+DE
Growth (mg)
Error (mg)
Relative error (%)
6465
0
0
16485
+10020
154.99
13607
+7142
110.47
16792
+10327
159.74
11232
+4767
73.74
8814
+2349
36.33
energy content.
DISCUSSION
The results of this study indicated that, in the bioenergetics modeling
of fish growth, it is relatively safe to assume that fecal production and
excretion are constant fractions of food energy; however, major errors may
arise from the assumptions that activity metabolism is a fixed multiple of
standard metabolism and that energy content of fish is constant.
Although most parameters in the bioenergetics model for the minnow were
derived from controlled experiments and the model adopted fewer assumptions
that most other models of its sort, the model failed to accurately predict the
growth of fish under laboratory conditions. The model's prediction of the
relationship between growth and ration did not agree with that observed in
empirical studies. This may be an important cause for the failure of the
model.
In most other bioenergetics models, the predicted relationship between
growth and ration was not explicitly examined. Kitchell et al. (1977) stated
that their model for Perca flavescens predicted maximum conversion
efficiencies at maximum rations. This was in agreement with the prediction
from the minnow model.
Inaccurate estimates of dry matter and energy content, SDA and activity
metabolism were the major sources of errors in the growth prediction from the
minnow model. In most other bioenergetics models for fish, simplifying
assumptions were made for these components.
This study suggested that, at the present stage, bioenergetics modeling
may have limited use for predicting fish growth. Future studies should focus
on the development of more accurate sub-models for the dry matter and energy
content of fish, SDA and activity metabolism to improve the performance of the
model.
255
-------
REFERENCES
Bevelhimer, M. S., R. A. Stein, and R. F. Carline. 1985. Assessing
significance of physiological differences among three esocids with a
bioenergetics model. Canadian Journal of Fisheries and Aquatic Sciences
42:57-69.
Brett, J. R. and T. D. D. Groves. 1979. Physiological energetics. In:
W. S. Hoar, D. J. Randall and J. R. Brett (eds.). Fish Physiology, Vol.
VIII. Academic Press, New York, NY. p. 278-352.
Carline, R. F. 1987. Simplified method based on bioenergetics modeling to
estimate food consumption by largemouth bass and northern pike.
Transactions of the American Fisheries Society 116:224-231.
Cochran, P. A. and J. A. Rice. 1982. A comparison of bioenergetics and
direct field estimates of cumulative seasonal food consumption by
largemouth bass (Micronterus salmoides). In: G. Cailliet and C.
Simenstad (eds.). Gutshop '81: Fish Food Habits Studies. Washington
Sea Grant Publications, Seattly WA. p. 88-96.
E. Grant. 1985a. Fish bioenergetics
Cuenco, M. L., R. R. Stickney, and W.
and growth in aquaculture ponds;
Ecological Modeling 27:169-190.
Cuenco, M. L., R. R. Stickney, and W.
I. individual fish model development.
E.
and growth in aquaculture ponds:
Phoxinus
Grant. 1985b. Fish bioenergetics
II. effects of interactions among
size, temperature, dissolved oxygen, unionized ammonia and food on
growth of individual fish. Ecological Modeling 27:191-206.
Cuenco, M. L., R. R. Stickney, and W. E. Grant. 1985c. Fish bioenergetics
and growth in aquaculture ponds: III. effects of interspecific
competition, stocking rate, stocking size and feeding rate on fish
productivity. Ecological Modeling 28:73-96.
Y. 1987. Bioenergetics and Growth of a Teleost, Phoxinus phoxinus
(Cyprinidae). Ph.D. Thesis, University of Wales, Aberystwyth.
Y. and R. J. Wootton. 1988a. Pattern of energy allocation in the
minnow, Phoxinus phoxinus (L.) (Pisces: Cyprinidae). Functional
Ecology 2:57-62.
Y. and R. J. Wootton. 1988b. The metabolic rate of the minnow,
phoxinus (L.) (Pisces: Cyprinidae), in relation to ration, body size
and temperature. Functional Ecology 2:157-162.
Y. and R. J. Wootton. 1988c. Effects of ration, temperature and body
size on the body composition, energy content and condition of the
minnow, Phoxinus phoxinus (L.). Journal of Fish Biology 32:749-764.
Y. and R. J. Wootton (in press) a. Bioenergetics of growth of a
cyprinid, Phoxinus phoxinus (L.): the effect of ration, temperature and
body size on food consumption, fecal production and nitrogenous
excretion. Journal of Fish Biology.
Y. and R. J. Wootton (in press) b. Bioenergetics of growth of a
cyprinid, Phoxinus phoxinus (L.): the effect of ration, temperature and
body size on growth rate and efficiency. Journal of Fish Biology.
Diana, J. S. 1983. An energy budget for northern pike (Esox lucius).
Canadian Journal of Zoology 61:1968-1975.
Hurley, D. A. 1986. Growth, diet, and food consumption of walleye
(Stizostedion vitreum vitreum): an application of bioenergetics
modeling to the Bay of Quinte, Lake Ontario, population. Canadian
Special Publications. Fisheries and Aquatic Sciences 86:224-236.
Cui,
Cui,
Cui,
Cui.
Cui,
Cui,
256
-------
Kitchell, J. F. and J. E. Breck. 1980. Bioenergetics model and foraging
hypothesis for sea lamprey (Petromyzon marinus). Canadian Journal of
Fisheries and Aquatic Sciences 37:2159-2168.
Kitchell, J. F. and L. B. Growder. 1986. Predator-prey interaction in Lake
Michigan: model predictions and recent dynamics. Environmental Biology
of Fishes 16:205-211.
Kitchell, J. F., J. F. Koonce, R. V. O'Neill, H. H. Shugart, J. J. Magnuson,
and R. S. Booth. 1974. Model of fish biomass dynamics. Transactions
of the American Fisheries Society 103:786-798.
Kitchell, J. F., D. J. Stewart, and D. Weiningen. 1977. Application of a
bioenergetics model to yellow perch (Perca flavescens') and walleye
(Stizostedion vitreum vitreum). Journal of the Fisheries Research Board
of Canada 34:1922-1935.
Rice, J. A., J. E. Breck, S. M. Bartell, andJ. F. Kitchell. 1983.
Evaluating the constraints of temperature, activity and consumption on
growth of largemouth bass. Environmental Biology of Fishes 9:263-275.
Stewart, D. J. and F. P. Binkowski. 1986. Dynamics of consumption and food
conversion by Lake Michigan alewives: an energetics-modeling synthesis.
Transactions of the American Fisheries Society 115:643-661.
Stewart, D. J., D. Weininger, D. V. Rottier, and T. A. Edsall. 1983. An
energetics model for lake trout, Salvelinus namaycush: application to
the Lake Michigan population. Canadian Journal of Fisheries and Aquatic
Sciences 40:681-698.
257
-------
ROLE OF SCIENCE AND TECHNOLOGY IN CONSERVATION OF THE STURGEON RESOURCE
by
Liu Jiankang1 and Yu Zhitang2
INTRODUCTION
The Zhonghua sturgeon (Acipenser slnensis Gray) is an anadroraous fish of
enormous size, reaching a maximum body weight of 550 kg. It inhabits the
continental shelf of the east coast of Asia. Every year in July and August,
individuals approaching maturity ascend the Chang Jiang (the Yangtse River),
where their gonads undergo further development, and spawn in the fall of the
following year. Its spawning ground is distributed mainly in the lower
reaches of the Jinsha River and the upper reaches of the Chang Jiang. The
annual catch of the sturgeon in the provinces along the Chang Jiang averaged
400 to 500 fish in the 1970s (Yu et al. 1986).
The construction of the Gezhouba Hydro-electric Project blocked the
passage for the spawning migration of this fish. The need to construct a dam
by-pass for the fish was hotly debated from 1970 to 1982. Finally, on
December 28, 1982, six leading officials sent a report to our central
government pointing out that "construction of a fish by-pass may not be
considered among the measures for the salvation of the Zhonghua sturgeon in
order to avoid serious economic loss and waste," that "more impetus should be
given to the experimentation on the induced spawning of the sturgeon below the
dam," and that "strict prohibition of commercial sturgeon fishing must be
enforced" (Academia Sinica 1984). These decisions from the central government
are well grounded.
It might be helpful here to summarize the discrepancy in opinions in
this matter. Those who advocated the construction of a fish by-pass
maintained that the natural spawning ground of the sturgeon lies in the upper
reaches of the Chang Jiang some 1000 km from Yichang. The gonads of those
sturgeon detained at the Gezhouba dam could not develop to maturity, hence it
would be impossible for them to spawn below the dam. For the same reason,
qualified brood sturgeon for induced spawning (whose gonads must attain Stage
IV development) would not be available. Therefore, the sturgeon is faced with
the crisis of extinction. In their opinion, a fish by-pass is a sort of
advanced installation that is indispensable. Even if the effectiveness of
^•Institute of Hydrobiology, Academia Sinica, Wuhan, PRC
2Institute of Reservoir Fisheries, Ministry of Water Conservancy and
Academia Sinica, Wuhan, PRC
258
-------
such an installation may not be satisfactory for sturgeon, it is nevertheless
better to construct one than none.
Those who objected to the construction of a fish by-pass held that when
the sturgeon ascend to below the dam, there was some evidence that the gonads
could develop to Stage IV, at least in a part of the spawning shoal.
Moreover, there are river sections downstream of the dam that offer bed
conditions and hydrological features similar to those in the natural spawning
ground of the Jinsha River. These sections possibly could form a new spawning
ground. Experiments on the induced spawning of sturgeon in the upper reaches
proved successful early in 1972. Induced spawning of the sturgeon below the
Gezhouba Dam should not be "beyond the capability of our generation" as
someone pessimistically predicted. What is more important is the fact that,
although fish by-passes intended for sturgeons built at hydro-power stations
abroad are generally unsuccessful (Doroshov 1977), the Zhonghua sturgeon has
even larger body-size and sparser population, and, coupled with the vast
breadth and great depth below the dam, the effectiveness of such construction
seems very doubtful. Suppose we built a fish by-pass at the cost of scores of
millions of yuan (RMB) that required millions of kWh of electric power to
operate but achieved nothing but a "monument of stupidity" (LaBounty, J. F.,
personal communication).
Besides, the construction of a fish by-pass at the Gezhouba Dam is a
matter that must be considered in the light of the proposed Three-Gorge High
Dam. To have the sturgeon pass the high dam over 100 m above the water level
is certainly much more difficult than to have it pass the low dam of Gezhouba.
And if it is decided in the end that no fish by-pass should be built at the
high dam, then what is the sense of constructing a fish by-pass at the dam of
Gezhouba.
SCIENTIFIC AND TECHNOLOGICAL EFFORTS AFTER THE
INTERCEPTION OF FLOW AT THE GEZHOUBA DAM
Interception of flow at Gezhouba was effected in January 1981.
Beginning from the fall of that year, inspection of individuals with mature
gonads was made each year at the sturgeon shoal at the Yichang River section
below the dam. The results were 1981, 13.5%; 1982, 12.5%; 1983, 25.4%; 1984,
51.1%; (Zhao et al. 1986) 1985, 44.0%; 1986, 49.6% (Deng et al. 1987). The
sturgeon aggregated below the dam normally includes individuals that will not
spawn until next fall; therefore, a percentage of nearly 50% is to be regarded
as normal (Zhao et al. 1986). In October-November 1982, members of the
Institute of Hydrobiology disclosed for the first time a new spawning ground
of the sturgeon downstream of the dam and recovered a large number of sturgeon
eggs from the intestines of benthic fishes. Thirty newly hatched, live
sturgeon fry also were collected 20 km below the Gezhouba Dam (Deng et al.
1987).
A joint investigation group dispatched by the National Economy
Committee; Ministry of Agriculture, Animal Husbandry and Fisheries; and the
Ministry of Water Conservancy and Electric Power verified this finding and
wrote in their report of investigation, "The spawning and normal hatching of
the sturgeon below the dam is a fact beyond any doubt and should be affirmed"
259
-------
(Academia Sinica). Subsequent surveys for 3 years (from 1982 to 1984) not
only confirmed the fact that the gonads of sturgeon below the dam can develop
to maturity, but also delineated the site of the new spawning ground as the
river section between the Yichang Shipbuilding Yard of the Bureau of
Changjiang Navigation and the vicinity of Aijiahe and Jiangjunmao--a section
of about 14 km in length. Besides, a brood sturgeon with running eggs was
caught near Huyatan (Deng et al. 1987).
The earliest success of induced spawning of the sturgeon at the site of
its natural spawning ground dated back to 1972. After the interception of
flow at Gezhouba, the Collaboration Group for Induced Spawning (Liu 1987)
embarked on a test for the sturgeon caught from below the dam in November
1983, adopting the method of chaining up the brood fish at the river bank.
Fertilized eggs were kept in a circular "race-course," where fry were hatched.
This implies the first success of induced spawning of sturgeon below the dam
(Liu 1987, Fu et al. 1985). During the course of the experimentation,
technical improvements were made in several ways, such as the use of the
commercially available estrogen LRH-A to replace the pituitaries of sturgeon
•and carp completely (Yi et al. 1986), the rearing of immature female sturgeon
(with ovaries of the developmental stage III) to sexually ripe parent fish in
the course of 1 year (Liu 1988), and the use of a new "Sieve net" type of
incubator. Newly hatched fry were kept temporarily in a loop-shaped pond, and
after 7 days, they were introduced into the rearing pond at the density of 500
fry per square meter. After 137 days of rearing, the fingerlings grew to a
body weight of 200 g. This result demonstrates clearly that sturgeon caught
below the dam can be used for induced spawning (Liu 1987).
The Sturgeon Release Station was organized by the Fishery Division of
the Gezhouba Bureau of Engineering in April 1982, and by 1984, the base was
already in a workable condition. From 1982 to 1984, the station engaged in
transporting the would-be spawners over the dam from below, in the hope that
they might ascend further to reach the natural spawning ground. In the three
fishing seasons, the station captured and released 82 fish to the newly formed
reservoir above the dam (Fishery Division 1984). As to the induced spawning
of sturgeon below the dam, the station obtained more than 40,000 fry from its
new spawning pond and released 6000 fingerlings 4 to 5 cm in length downstream
as the first batch. From 1984 to 1986, 467,000 7-day baby sturgeon 1.8 to 2.3
cm in length (69 -78 mg in weight) were released downstream (Liu 1987) .
Tagging of various sorts were done on fingerlings and on a limited number of
large fish (average weight 715 g), but the number was too small, and the time
interval too short, to get any returns as yet. A summary on the cultivation
techniques of fry and fingerlings has been published recently (Xiao et al.
1988).
No matter whether natural or induced spawning is concerned, the
prerequisite at present is an abundance of brood sturgeon (with ripe gonads)
in the river. Regretfully, in the first 2 years after flow interception,
batches and batches of brood sturgeon were captured and killed when the
migrating fish were impeded by the dam. In the fall and winter of 1981, more
than 800 sturgeon were caught in the Hubei Province, i.e., 5.5 times the
multiyear average of 145 before dam construction. Again, a total of nearly
400 fish were caught in 1982 and in the spring of 1983 (Yu 1986). Since the
260
-------
fall season of 1983, a number of Fishing Administration Stations have been set
up and strict protection of the brood fish has been brought into effect. The
action began to obtain good results in the second year. The number of brood
sturgeon caught since 1984 has decreased to a large extent, and, in 1984, the
season of natural spawning occurred earlier than in 1982 and 1983. The
duration of two batches of spawning was prolonged and the range of spawning
ground was extended in comparison with the previous years (Academia Sinica
1984).
Having affirmed the effects of brood fish protection, research workers
now turn to the protection for the juvenile sturgeon (Zhao 1986). Their
investigation in the estuary indicated that the number of specimens of
juvenile sturgeon they could collect (477 in 1984 and 571 in 1985) was
substantially increasing. However, juveniles descending a long way from
Gezhouba were captured in large numbers by the set-net fishery at the estuary.
The authors stress that the protection of juvenile sturgeon is an important
link in the chain of sturgeon resource management and should arouse the
attention of fishery management departments. The proposal is made to the
effect that during the peak period of juvenile "emigration" (mid-June to mid-
July) , set-net fishing should be banned for a month, so as to give the
juveniles more chance to descend to the sea for growth.
To sum up, from the refutation of the idea of building a fish by-pass
for the sturgeon to the confirmation of the possibility that the sturgeon
below the Gezhouba Dam can develop ripe gonads, and to the actual finding of
the anticipated new spawning ground below the dam and from the improvements in
the techniques of induced spawning to the protection of brood sturgeon and the
installation of release stations for the sturgeon, along with the protection
of juvenile sturgeon, all these point to the guiding role of science and
technology in the conservation of the sturgeon resource.
REFERENCES
Academia Sinica. 1984. Circumstance Report on the Work of Fish Salvation at
Gezhouba. Institute of Hydrobiology, Academia Sinica.
Deng et al. 1987.
Doroshov, S. I. 1977. Passage of sturgeon through the fishlocks in the USSR.
Fisheries Institute. 1981. Summary for the Biotechnical Research on the
Artificial Breeding of Zhonghua Sturgeon by the Method of Chaining.
Fisheries Institute of Chang Shou Lake, Chongqing.
Fishery Division of the Gezhouba Bureau of Engineering. 1984. Manuscript.
Fu, C-j. 1985. Freshwater Fisheries, No. 1, 1-5.
Liu, Y. 1987. Preliminary studies on the induced spawning and fingerling
rearing of the Zhonghua sturgeon (Manuscript).
Liu, Y. 1988. Technical study of the artificial proliferation of the Chinese
Zhonghua sturgeon--The Status Quo and Prospects. Shuiliyuye Water
Conservancy Fisheries No. 4, 20-23.
Xiao, H. et al. 1988. Preliminary studies on the techniques of cultivation
for the fry and fingerling of the Zhonghua sturgeon. Shuiliyuye (Water
Conservancy Fisheries) No. 4, 24-29.
Yi, J-f. et al. 1986. Shuiliyuye (Water Conservancy Fisheries) No. 2, 44-46.
Yu, Z-t. 1986. Transactions of the Chinese Ichthyological Society. No. 5,
1-16.
Zhao, Y. 1986. Shuiliyuye (Water Conservancy Fisheries) NO. 6, 38-41.
261
-------
PROTECTION OF WATER QUALITY IN THE UNITED STATES
by
Rosemarie C. Russo1
Since 1970 the United States Environmental Protection Agency (EPA) has
exercised primary responsibility within the Federal government to address the
environmental problems confronting our nation. The Agency administers nine
comprehensive environmental protection laws: the Clean Air Act; the Clean
Water Act; the Safe Drinking Water Act; the Comprehensive Environmental
Response, Compensation, and Liability Act (better known as "Superfund"); the
Resource Conservation and Recovery Act; the Federal Insecticide, Fungicide,
and Rodenticide Act; the Toxic Substances Control Act; the Marine Protection,
Research, and Sanctuaries Act; and the Uranium Mill Tailings Radiation Control
Act.
Some of these Acts were created since the EPA was first established;
others are amended versions of legislation first enacted as far back as 1899.
The 1899 legislation was the Rivers and Harbors Act, which was the first U.S.
Federal law in environmental protection. Subsequent legislation in water
pollution control produced the Clean Water Act, which was recently amended and
reauthorized in 1987.
Using the environmental laws as the foundation, the Agency develops
regulations and sets standards for pollutant levels in the individual environ-
mental media and then monitors compliance by, pollutant generators within the
individual states. To ensure that its activities are based on the best
available scientific and technical knowledge, the Agency conducts an extensive
research and development program.
EPA is organized into four major regulatory offices to carry out its
responsibilities to protect the environment and human health. These are
(Figure 1) Air and Radiation, Water, Pesticides and Toxic Substances, and
Solid Waste and Emergency Response. In addition, there is an Office of
Research and Development with twelve research laboratories whose function is
to conduct both fundamental and applied research in support of the regulatory
offices. There are ten regional offices whose function is to work with the
^•Environmental Research Laboratory, U.S. Environmental Protection Agency,
Athens, GA, USA
262
-------
Office of
Administration and
Resources Management
Office of
Enforcement and
Compliance Monitoring
Criminal
Enforcement
Senior
Enforcement
Counsel
Office of
Water
—
—
Water
Enforcement
and Permits
Water
Regulations
and Standards
Municipal
Pollution
Control
Drinking
Water
Marine and
Estuarlnas
Protection
Ground Water
Protection
Wetlands
Protection
Office of
Solid Waste and
Emergency Response
-
Solid Wasto
Emergency
and Remedial
Response
Waste
Programs
Enforcement
Underground
Storage Tanks
Office of
Air & Radiation
Air Quality
Planning
and Standards
Mobile
Sources
Radiation
Programs
Office of
Inspector General
—
Audit
Investigations
Management
Modeling,
Environmental
Monitoring and
Quality Assurance
Environmental
Engineering
Technology
Demonstration
Environmental
Processes and
Effects
Research
Health Research
Health and
Environmental
Assessment
Figure 1. U.S. Environmental Protection Agency Organization
263
-------
states to implement EPA's environmental regulations throughout the country.
Finally, there are several administrative offices, including Administration
and Resources Management; Enforcement and Compliance Monitoring; Policy,
Planning, and Evaluation; International Activities; General Counsel; and
Inspector General.
EPA also is empowered to enforce compliance with its regulations through
administrative, civil or criminal actions in court. By seeking and winning
large financial and criminal penalties against significant violators, EPA can
perhaps remove any incentive to non-compliance. EPA had a strong record in
enforcement in 1986 and 1987. There were 373 civil referrals in 1986, highest
in history; and 274 civil referrals in 1987, second highest in history. There
were 3,200 administrative orders issued by EPA in 1987, the highest in
history. A record $24 million in civil penalties was assessed in 1987. Sixty
percent of the total of all EPA civil penalties have been collected over the
last three years.
Since 1972, EPA has awarded $44.6 billion for sewage treatment construc-
tion grants. The 1987 Clean Water Act Amendments authorize an additional $18
billion, total, for construction grants through 1990. Of this amount, at.
least $8.4 billion and as much as $13.2 billion is authorized for capitaliza-
tion grants to establish state revolving funds. Federal money in the state
revolving funds must be matched by at least 20% state funding. The State
Revolving Funds will be used for construction of sewage treatment plants and
other water pollution control activities. They are part of the transition
from the Federal construction-grants program to a state-operated loan program.
EPA is increasingly helping states develop their own criminal enforcement
capabilities to detect and prosecute environmental crimes. In 1988 program
enforcement priorities include an emphasis on municipal compliance for water
quality. State enforcement is reaching increasingly high levels, with State
agencies prosecuting a total of 723 environmental cases in 1987--twice that of
the previous year. States also initiated 3200 administrative orders.
Clearly, water quality protection is a high environmental priority. The
keystone for water pollution control is the National Pollutant Discharge
Elimination System (NPDES). Under the NPDES, community sewage facilities must
secure permits that specify the types and amounts of pollutants that may be
discharged. Industries discharging pollutants into waterways also are subject
to control requirements. These effluent limitations are designed to reach an
ultimate goal of completely eliminating the discharge of pollutants into the
nation's waters.
The water pollution control program seems to be succeeding. Since 1970,
municipal sewage treatment has been provided for more than 80 million people.
Most industrial plants have installed water pollution control technology. As
a result, organic wastes discharged from industrial sources have been reduced
by 38 percent. When all the mandated effluent guidelines are in place,
discharges of toxic pollutants will have been reduced 96 percent from 1972
levels.
To deal with problems of contamination of surface waters from toxic
organic pollutants such as pesticides and inorganic pollutants such as lead
264
-------
and arsenic, the Agency established National Water Quality Criteria for
Protection of Aquatic Organisms and Their Uses. The process of deriving
national criteria is shown schematically in Figure 2 (U.S. EPA 1985). The
goal of these criteria is to prevent unacceptable long-term and short-term
effects on commercially and recreationally important aquatic species; other
important species; fish and benthic invertebrate assemblages in rivers and
streams; and fish, benthic invertebrate, and zooplankton assemblages in lakes,
reservoirs, estuaries, and oceans. The national criteria for each chemical
are given as two numbers: (a) criterion maximum concentration is intended to
be protective against acute adverse effects due to high, short-term concentra-
tions ; (b) criterion continuous concentration is intended to be protective
against unacceptable chronic effects of lower, long-term continuous concentra-
tions .
National criteria for a specific chemical may be applied to a particular
body of water directly, or may be modified to take into account local condi-
tions of water quality and aquatic species present ("Site-Specific Water
Quality Criteria"). Chemical characteristics of the local water may be such
as to increase or decrease the toxicity of the chemical of concern. Some
species in the local water may be very sensitive or insensitive to the
chemical of concern, or the resident species may be stressed by other factors
such as parasites, disease, predators, other pollutants, etc.
The states are responsible for establishing and enforcing standards for a
particular surface water body. The standards are set taking into account the
national criteria, any site-specific modifications of the criteria based on
data about field conditions, plus social and economic considerations and
environmental and analytical chemistry capabilities. EPA assists the states
in setting and enforcing state water quality standards.
To regulate chemicals in surface waters, we need to know: what chemical
is being discharged, the quantity being discharged, the fate of that chemical
in the aquatic environment, the short- and long-term hazard it poses to
aquatic life, and what the national criteria are, if available.
To derive National Water Quality Criteria, certain specific information
is required: the individual chemical, data on acute and chronic toxicity to
representative species of aquatic animals, toxicity to freshwater algae or
vascular plants, and bioaccumulation by aquatic organisms (if a maximum
permissible residue concentration in tissue is available). Other pertinent
data that may be available are also taken into consideration.
Data on acute toxicity to animals consist of 96-hour LC50 (concentration
causing mortality in 50% of tested animals) or EC50 (concentration causing
immobility or other appropriately defined effect) values. These LC50 or EC50
values are generated by continuous exposure in laboratory toxicity tests on
fishes and macroinvertebrates.
Data on chronic toxicity to aquatic animals consist of results of
continuous exposure laboratory tests on fish and macroinvertebrates. These
tests may use any of several endpoints to measure unacceptable effects, such
as survival, growth, reproduction, or histological or other effect. Data on
265
-------
REVIEW FOR
COMPLETENESS
OF DATA AND
APPROPRIATENESS
' OF RESULTS
NATIONAL
CRITERION
A
A~
CRITERION
UAVTIuTMIur L
1
iJ [1
£1
ss
/
J
H t
§i
/
E
;i
"1
,
J 6
ii
\
/
H >-
52
!s
^
^
§
3
q
>
k
1
s
3
r-&
A
o
fj M W
^j ^5 r~*
Sp
A
~~A
CHRONIC
TOXICITY
TO ANIMALS
A
A
A
1-1 2 <
fa Oi >
A
TOXICITY
TO
PLANTS
t
§
W I-H
SDH
O O
O
0)
4-1
O
M
Pn
a)
•s
o
1-1
cfl
0)
4-)
Cf
0)
4-1
G >~s.
O IT)
•iH oo
4-1 0\
(d iH
"a
cd W
O
a
w
M-l Q)
o w
•H -H
4J 0)
Q ttf
266
-------
bioaccumulation are used for chemicals where residues in aquatic organisms
may be harmful to the animal; or residue concentrations may adversely affect
wildlife, including fishes and birds, that consume these organisms; or resi-
due concentrations may adversely impact human health or marketability due to
flavor impairment.
For derivation of national water quality criteria (U.S. EPA 1985), the
minimum set of data required for acute toxicity to aquatic animals is results
of acceptable acute tests with at least one species of freshwater animals in
at least eight different families such that all of the following are included.
a. the family Salmonidae in the class Osteichthyes
b. a second family in the class Osteichthyes, preferably a commer-
cially or recreationally important warmwater species (e.g., blue-
gill, Lepomis macrochirus: channel catfish, Ictalurus punctatus.
etc.)
c. a third family in the phylum Chordata (may be in the class
Osteichthyes or may be an amphibian, etc.)
d. a planktonic crustacean (e.g., cladoceran, copepod, etc.)
e. a benthic crustacean (e.g., ostracod, isopod, amphipod, crayfish,
etc.)
f. an insect (e.g., mayfly, dragonfly, damselfly, stonefly, caddisfly,
mosquito, midge, etc.)
g. a family in a phylum other than Arthropoda or Chordata (e.g.
Rotifera, Annelida, Mollusca, etc.)
h. a family in any order of insect or any phylum not already
represented.
The minimum set of data for chronic toxicity to animals is information
from tests with species of aquatic animals for which acute toxicity data are
available, and the tested species must be in at least three different famili-
es. Of these three species, at least one must be a fish, at least one must be
an invertebrate, and at least one must be an acutely sensitive freshwater
species (the other two may be saltwater species).
Results are required from at least one acceptable test with a freshwater
alga or vascular plant. Furthermore, if plants are among the aquatic organ-
isms that are most sensitive to the pollutant, results of a test with a plant
in another phylum (division) also should be available. Finally, at least one
acceptable bioconcentration factor determined with an appropriate freshwater
species is required, if a maximum permissible tissue concentration is avail-
able.
These data requirements were developed based on the assumption that data
from toxicity tests with representative species provide a useful indication of
267
-------
the sensitivities of appropriate untested species. It also must be recognized
that there is a great deal of variability in toxicity testing. Species tested
are often selected on the basis of availability alone; tests are conducted in
laboratory waters that are free of pollutants and typically low in particulate
matter and organic matter. Duration of tests can vary considerably Also
tests on single species are used to represent biologically complex systems;
i.e., to reflect the response of ecosystems to pollutants. Furthermore, these
data requirements are quite expensive and time-consuming to meet. Therefore
it is worthwhile to compare the toxicity of given chemicals to a range of
aquatic species. This would help to determine whether a single species might
be used as a surrogate for others, and to determine what minimum amount of
toxicity data could adequately, reflect the susceptibilities of the aquatic
animals that we are trying to protect.
Only one experimental study, by Thurston and coworkers (1985), has been
conducted to determine, under standardized test conditions and procedures, the
toxicity of a selected set of chemicals to a selected group of aquatic
animals. The chemicals tested were selected based on their exhibiting (from
literature values) a wide range of toxicity (from 26 grams/liter to 1 micro-
gram per liter) to the fathead minnow, Pimephales promelas. and also on their
having different physiological mechanisms of toxic action. The animals tested
were selected based on their frequent use for laboratory toxicity tests their
ease of culture in the laboratory, and at the same time to provide a reason-
able diversity of species. Six fishes, two crustaceans, a chironomid, and an
amphibian were selected. Acute toxicity tests were then conducted for ten
chemicals on ten aquatic species (Table 1).
Two general conclusions were drawn from this study. First, there was no
consistent relative susceptibility, or orders of sensitivity, among the test
species for this group of chemicals. Second, if one averaged the individual
LC50 values to obtain a mean toxicity, positive deviations from the mean tox-
icity tended to be much greater than negative deviations. This suggests that
the observed variation in toxicity between any two species for any one chemi-
cal is likely attributable to one of the species being relatively insensitive
to the action of the chemical rather than one of the species being much more
sensitive than the mean sensitivity for all other species.
The study also found a relationship between the toxicity data for rain-
bow trout (Salmo gairdneri) and fathead minnows (P. promelas'), whereby
toxicity data for either organism could be relied on to classify properly ary
of the chemicals tested. In addition, equations were developed to estimate
the lethal concentrations of chemicals with each species from the toxicity
data for fathead minnows.
Finally, the largest existing compilation of aquatic toxicity data in
terms of both chemicals and species, is EPA's AQUIRE (Aquatic Information
Retrieval) toxicity data base (Russo and Pilli 1984), which is a computerized
compilation of literature toxicity data from acute and chronic tests for
freshwater and marine organisms. AQUIRE contains information from 100 000
toxicity tests for 5000 chemicals on 2300 aquatic species. Data included in
268
-------
this data base are rated for quality, based on the test procedures used.
Access to AQUIRE is available through the Scientific Outreach Program, EPA
Environmental Research Laboratory, 6201 Congdon Boulevard, Duluth, MN 55804.
TABLE 1. COMPARATIVE TOXICITY OF TEN ORGANIC CHEMICALS TO TEN COMMON
AQUATIC SPECIES (Thurston et al. 1985)
Chemical
Species
2-(2-Ethoxyethoxy)-ethanol
2-Methyl-2,4-pentanediol
2-Methyl-l-propanol
2,2,2-Trichloroethanol
2,4-Pentanedione
2-Choroethanol
Hexachloroethane
Pentachlorophenol
Permethrin
Endrin
Daphnia magna
Tanytarsus dissimilis
Orconectes immunis
Rana catesbiana
Salmo gairdneri
Lepomis macrochirus
Gambusia affinis
Ictalurus punctatus
Carassius auratus
Pimephales promelas
REFERENCES
Russo, R.C. and A. Pilli. 1984. AQUIRE: Aquatic Information Retrieval
Toxicity Data Base. EPA-600/8-84-021. U.S. Environmental Protection
Agency, Duluth, MN.
Thurston, R.V., T.A. Gllfoil, E.L. Meyn, R.K. Zajdel, T.I. Aoki, and G.D.
Veith. 1985. Comparative toxicity of ten organic chemicals to ten
common aquatic species. Water Research 19(9):1145-1155.
U.S. EPA (Environmental Protection Agency). 1985. Guidelines for deriving
numerical national water quality criteria for the protection of aquatic
organisms and their uses. National Technical Information Service
Accession Number PB85-227049. U.S. Environmental Protection Agency,
Washington, D.C.
269
-------
THE FISH EMBRYO-LARVAL PROCEDURE!
PREDICTING CHRONIC TOXICITY
AND ECOLOGICAL EFFECTS
by Jeffrey A. Black and Wesley J. Birge
INTRODUCTION
Numerous studies relating to fish toxicology generally have indicated
that piscine reproduction is one of the most sensitive target sites of aquatic
contaminants (Birge et al. 1974). Reproductive processes can be severely
impaired by toxicants at concentrations that apparently have no overt effects
on adult organisms. As a result, test procedures for establishing water
quality criteria have included determinations of chemical concentrations that
produce chronic-level effects on aquatic life.
Methodologies historically used for such purposes have involved periods
of exposure incorporating all or most life-cycle stages (Mount and Stephan
1967, Pickering 1974, McKim et al. 1976). Such tests generally require 8
months to 2 years for completion, are labor-intensive, and have impeded the
generation of chronic data. To obviate such problems, more economical early-
life-stage procedures were developed (U.S. EPA 1978). These tests span 30 to
90 days and chronic-effect endpoints appear to be in reasonable agreement with
those observed in the longer life-cycle and partial-life-cycle tests (McKim
•*•«*/// •
Because of the thousands of chemical substances requiring toxicological
evaluations, however, the need persisted for still shorter and more economical
tests for reliably estimating chronic toxicity to fish species. Two such
"short-term" methodologies have been developed for these purposes. The first
involves a 7-day test performed with larvae of the fathead minnow (Pimephales
promelas), in which organisms are fed throughout the exposure period. Testing
is initiated with 1-day-old larvae, and endpoints are quantified on the basis
of survival frequencies and growth differentials between control and
experimental animals. Specific procedures for the applications of this test
have been described elsewhere (Horning and Weber 1985, Mount et al. 1985
Norberg and Mount 1985). '
A second procedure, the short-term embryo-larval test, may be performed
with a variety of fish and amphibian species. This test is initiated shortly
after egg fertilization and continues through 4 days posthatching. The
1. Graduate Center for Toxicology and School of Biological Sciences
University of Kentucky, Lexington, KY, USA.
270
-------
exposure period is 5 to 8 days (25°C) and typically is conducted with species
such as the largemouth bass (Micropterus salmoides), bluegill sunfish (Lepomis
macrochirus), and the fathead minnow. No feeding is required. Test responses
include embryonic mortality, egg hatchability, larval mortality, and
teratogenesis.
Log problt analysis (Finney 1971) can be used to calculate median lethal
(LC50) and toxicant threshold (LCI) concentrations. These values are
generally calculated by combining frequencies of dead organisms and grossly
teratic larvae observed at the end of the test. Conventional statistical
analyses such as Dunnett's multiple comparison procedure (Horning and Weber
1985) and the chi-square Fisher's exact test (Armitage 1971) can be used to
calculate no-observed-effect concentrations (NOEC). Rainbow trout (Salmo
gairdneri) also have been used in this system, in which the total exposure
period is about 28 days (13°C).
These'latter two tests can be performed using either static-renewal or
continuous-flow procedures and have produced chronic values similar to those
determined in traditional life-cycle experiments (Birge et al. 1981, Birge and
Cassidy 1983, Birge et al. 1985, Norberg and Mount 1985).
This paper reviews the reliability and usefulness of the short-term fish
embryo-larval test for estimating chronic-level effects of contaminant stress
on aquatic organisms. Primary attention given to specific applications of the
procedure, including assessments of (1) the toxicity of single compounds and
chemical mixtures, (2) differential species sensitivity, (3) teratogencity as
a toxicological endpoint, (4) structure-toxicity relationships, (5) effluent
and receiving water toxicity, and (6) effects produced by sediment-associated
chemicals.
GENERAL ASPECTS OF THE SHORT-TERM
EMBRYO-LARVAL PROCEDURE
Procedures for performing the short-term embryo-larval toxicity test were
reviewed briefly in the above section and have been further detailed in
previous publications (Black et al. 1983, Birge et al. 1985, Horning and Weber
1985). An overview of test characteristics is given in Table 1. Selection of
the particular exposure system to be used (static-renewal, flow-through) is
dependent on several considerations. The static-renewal test, in which
solutions are changed at regular intervals (e.g., 12 or 24 hours), is
economical and can be used with toxicants that are relatively stable in
solution. Flow-through procedures are recommended for testing rapidly
degradable materials. One such system developed in our laboratory has been
used successfully in regulating exposure concentrations of volatile or
insoluble organic compounds (Birge et al. 1979a, Black et al. 1982). We
presently are developing new toxicity test procedures that will require
solution volumes of only 50 ml or less.
Biological endpoints subject to evaluation in embryo-larval tests are
summarized in Table 2. Endpoints not discussed in this paper (e.g.,
avoidance/attraction behavior, tumor formation) have been described elsewhere
(Black and Birge 1980, Hendricks et al. 1984). Growth, not typically used as
an embryo-larval test endpoint, is currently being evaluated for reliability
271
-------
Table 1. GENERAL CHARACTERISTICS OF THE SHORT-TERM EMBRYO- LARVAL TEST
1. Five or more exposure concentrations plus control
2. Two or four replicates/exposure
3. 20-50 organisms per 50-400 mL chamber
4. Dilution water: natural or reconstituted
5. Exposure through hatching plus 4 days posthatching
6. Static-renewal or flow-through system
7. No feeding requirement
8. Laboratory or field applications
9. Multiple animal species capable of being tested in one chamber
10. Multiple test endpoints
11. Regression analysis (LC50, LCI)
12. Analysis of variance, Dunnett's test, Duncan's test (NOEC, LOEC)
and reproducibility. The short-term embryo-larval procedure has a number of
applications and includes among others the evaluation of single toxicants,
effluents and other complex mixtures, and sediment-associated chemicals.
Several types of environmental media evaluated using early-life stages are
given in Table 3. The following sections .summarize various aspects of the
test, including the reliability and applicability of the data generated.
REPRODUCIBILITY AND RELIABILITY OF
EMBRYO-LARVAL PROCEDURES
The reproducibility of the short-term embryo-larval test has been
investigated using various compounds, and an example of the precision of test
endpoints is presented in Table 4 (Birge et al. 1985). Six continuous flow
tests (two replicates each) were performed with cadmium on the fathead minnow
over an 8-month period. The LC50 values varied only from 0.067 to 0.084 mg/L
and LCI values were between 0.010 and 0.014 mg/L. The 95% confidence limits
were relatively small and probit-derived LCls (threshold toxicity values) were
in good agreement with NOECs determined by independent statistical procedures.
Table 2. EMBRYO-LARVAL TEST ENDPOINTS
Acute toxicity - LC50
Chronic Toxicity - Estimated chronic values or toxicity threshold values (LCI)
Teratogenesis - defective or retarded development (embryos, larvae)
Growth - length, weight
Behavior - avoidance/attraction responses, locomotor impairment, feeding
Bioconcentration - toxicant uptake
272
-------
Table 3. ENVIRONMENTAL MEDIA SUBJECT TO EMBRYO-LARVAL TOXICOLOGICAL
EVALUATIONS
Single chemicals/substances (soluble, insoluble, volatile)
Chemical combinations - toxic interactions
Complex mixtures
Whole effluents/dilutions
Effluent fractions
Chemical leachates
Persistence of toxicity
Receiving systems
Sediment-associated chemicals
Natural bulk sediments
Chemically enriched sediments
Sediment elutriates
Similar precision was observed in static-renewal tests, using the same
compound and fish species.
The short-term procedure also has been used to perform multiple species
tests, in which embryos of different species have been maintained in the same
exposure chamber. In one such experiment, eggs of the carp (Gyprinus carpio),
the fathead minnow, and the largemouth bass were exposed to cadmium, and the
LC50 values were 138.9, 107.0, and 244.1 jig/L, respectively. The
corresponding LCls were 4.1, 11.5, and 13.9 /zg/L. To achieve a broader
toxicity base on cadmium with warmwater species, additional experiments were
performed on the channel catfish (Ictalurus punctatus) and goldfish (Garassius
auratus), and the LCI values (Mg/L) obtained with these species were 8.6 and
3.0, respectively.
The range of LCI values calculated for these 5 species was 3.0 to 13.9 pg
Cd/L. This range compared favorably to chronic values of 5.8 to 15.0 /ig/L
achieved in longer-term chronic tests reported in the EPA criterion document
for six warmwater fish species (U.S. EPA 1980). Although data for precise
species to species comparisons were not always available, these results lend
support to the premise that the toxicity threshold values (LCls) obtained in
short-term embryo.-larval tests can be used reliably to estimate chronic
toxicity.
Further evidence of this nature has been shown in tests with the rainbow
trout. Embryo-larval stages exposed in static-renewal tests to seven
different metals gave LCI values that compared favorably to maximum acceptable
toxicant concentrations (MATC) calculated from chronic tests with rainbow
trout and other salmonid species (Table 5, Birge et al. 1985).
TOXICOLOGICAL SCREENING OF CHEMICALS
On the basis of data indicating that the embryo-larval test can be used
for predicting long-term effects, this procedure may afford a practical means
273
-------
by which to screen a great number of contaminants for toxic properties,
identify those of greatest concern to aquatic ecosystems, and estimate
concentrations that produce chronic toxicity. As an example, the short-term
procedure with the rainbow trout was tested.on 33 elements found in fossil
fuels (Birge et al. 1980). In this study, the LC50 values-were less than 1
mg/L for 19 of 33 elements and LCI values were 10 //g/L or less for 12.
Results indicated the high sensitivity of trout developmental stages and the
potential application of these procedures for use in testing programs designed
to prioritize compounds for further toxicological evaluations.
The test also has been used to establish a toxicological response range
for a number of different aquatic organisms. The heterogeneity of species
response is important in understanding the degree to which aquatic biota vary
in their sensitivity to environmental contaminants. The toxicity data on
cadmium given above demonstrate the differential susceptibility among certain
warmwater fish species. Another example of this concept was reflected by
comparative tests in which inorganic mercury was administered to embryo-larval
stages of 14 amphibian species (Birge et al. 1979b). The LC50 values spanned
nearly 2 orders of magnitude, ranging from 1.3 /ig/L with the narrow-mouth toad
(Gastrophryne carolinensis") to 107.5 fig/L with the marbled salamander
(Ambvstoma opacum).
Table 4. REPRODUCIBILITY OF 8-d EMBRYO-LARVAL TESTS WITH CADMIUM USING THE
FATHEAD MINNOW IN A CONTINUOUS-FLOW SYSTEM3
Test
no.
1
2
3
4
5
6
LC50
(mg/L)
0.067
0.084
0.073
0.079
0.084
0.070
95% confidence
limits
0.056-0.078
0.073-0.097
0.062-0.085
0.068-0.091
0.072-0.098
0.060-0.081
LCI
(mg/L)
0.010
0.013
0.012
0.012
0.014
0.010
95% confidence
limits
0.006-0.014
0.008-0.017
0.007-0.017
0.008-0.017
0.009-0.020
0.006-0.015
NOEC
(mg/L)
0.012
0.011
0.013
0.011
0.010
0.014
"Reprinted with permission from Environmental Toxicoloty and Chemistry, vol.
4, W.J. Birge, J.A. Black, and A.G. Westeman, Short-term fish and amphibian
embryo-larval tests for determining the effects of toxicant stress on early
life stages and estimating chronic values for single compounds and complex
effluents, 1985, Pergamon Press.
274
-------
TERATOGENICITY
Use of the short-term embryo-larval test provides the opportunity to
quantify frequencies of teratogenesis present in hatched populations For
most toxicants, the number of teratic larvae generally has been found to
increase with exposure level, but the dose-response relationship usually is
not as precise as that observed for mortality. In most experiments, the
incidence of teratogenesis has ranged from 0 to 5% at or near the toxicity
threshold and frequencies of 10% or greater usually have not been observed
except at or above median lethal concentrations. There are exceptions to this
trend for certain metals and organic contaminants, and these occurrences have
been discussed previously (Birge and Black 1977, Birge et al 1983) Gross
anomalies most often encountered have included defects of the head and
Table 5.
METALS
LClb
Element
Cadmium
8.0
95% confidence MATC
limits
Species
Test0
5.4-10.9
1.7-3.4
3.8-11.7
Brook trout
Brown trout
clc
els
Chromium
Copper
Lead
Mercury
Silver
Zinc
21.5
3.4
10.3
0.2
0.1
216
10.3-35.2
1.6-5.9
6.9-14.6
0.1-0.3
0.1-0.2
157-275
51-105
3.0-5.0
5.0-8.0
9.4-17.4
4.1-7.6
7.2-14.6
71-146
0.29-0.93
0.09-0.17
532-1368
Rainbow trout
Brook trout
Brook trout
Brook trout
Rainbow trout
Rainbow trout
Rainbow trout
Brook trout
Rainbow trout
Brook trout
els
els
els
clc
pic
pic
els
clc
pic
pic
aLCl values determined in 28-d static-renewal tests with rainbow trout
terminated 4 d after hatching.
bMATCs taken from published data for 60- to 90-d early-life-stage (els)
partial life-cycle (pic) or complete life-cycle (clc) tests
Reprinted with permission from Environmental Toxicolnt-y and Chenngt-r-y Vol
embr™'/^'/^' J1*^' and A'G- Westeman, Short-term fish and amphibian
embryo-larval tests for determining the effects of toxicant stress on early
eff?u!^SeS,^ Timatlng chronlc values for single compounds and complex
ettluents, 1985, Pergamon Press.
275
-------
vertebral column, dwarfed bodies, partial twinning, microcephaly, absent or
reduced eyes and fins, and amphiarthrodic jaws. Approximately 80% to 90% of
the gross anomalies observed have 'been defects of the skeletal system. Most
have involved the vertebral column and included lordosis, scoliosis, kyphosis,
and rigid coiling.
Based on these observations, it was proposed that fish and amphibian
embryo-larval stages may constitute simple and effective models with which to
investigate teratogenesis and that such test systems possess high potential
for (1) determining the mechanisms of teratogenesis, (2) evaluating the impact
of environmental toxicants on aquatic biota, and (3) preliminary screening and
identification of environmental teratogens that may be of concern to human
health (Birge et al. 1983).
STRUCTURE-ACTIVITY RELATIONSHIPS
Because aquatic ecosystems will continuously be affected by a broad
spectrum of pollutants, environmental hazard assessments must rely on an
orderly and systematic process for screening toxicity of diversified
contaminants. The vast number of chemicals introduced to the environment
places constraints on the extent of toxicity testing that can be accomplished
within a practical time frame. For this reason, considerable attention has
been given to developing procedures to allow for the estimation of toxicity
based on certain structural characteristics of such compounds (Konemann 1981,
McCarty et al. 1985, Hodson et al. 1988). Given an adequate toxicological
baseline for a representative number of organic chemicals or chemical classes,
the toxicity of many untested compounds could potentially be predicted by
examination of their structures or other physical-chemical characteristics.
The usefulness of the short-term embryo-larval procedure for predicting
toxic effects of structurally related chemicals has been evaluated for certain
classes of organic compounds (Birge and Cassidy 1983, Black et al. 1983,
Millemann et al. 1984). In embryo-larval studies testing compounds within
each of three chemical classes (i.e., hydroxylated aromatic hydrocarbons,
azaarenes, polycyclic aromatic hydrocarbons), toxicity to both the largemouth
bass and rainbow trout was found to increase with increasing ring number
(Black et al. 1983). Similarly, toxicity of azaarene compounds to the leopard
frog (Rana pipens) increased with ring number and increased octanol/water
partition coefficients (Birge and Cassidy 1983).
Fish and amphibian short-term tests also were used to assess the
relationship between toxicity and degree of chlorination of certain
polychlorinated biphenyl compounds and chlorinated alkanes. Toxicity was
observed to be greater as the chlorination of these compounds increased. For
example, LC50 values with the leopard frog were 48, 4.16, and 1.64 mg/L for
methylene chloride (2 chlorine atoms), chloroform (3 chlorine atoms), and
carbon tetrachloride (4 chlorine atoms), respectively. Corresponding octanol
water partition coefficients (log P) were 1.51, 2.02, and 2.79 (Birge and
Cassidy 1983). Based on these and other data, the sensitive and economical
short-term embryo-larval procedure may be extremely advantageous for use in
broad-based testing programs designed to establish structure-toxicity
relationships for a wide variety of compounds.
276
-------
EMBRYO-LARVAL BIOMONITORING OF EFFLUENTS AND RECEIVING WATERS
The above descriptions of the short-term embryo-larval test primarily
have been concerned with assessments of single compounds. This procedure,
however, has been used successfully in a number of field investigations
involving direct toxicological monitoring of effluents and receiving waters.
In biomonitoring studies on effluents from a tannery-sewage treatment plant, a
synthetic rubber plant, a metal plating plant, and a chemical manufacturing
plant, both acute and embryo-larval tests using fathead minnows were performed
concurrently on each of these NPDES (National Pollutant Discharge Elimination
System) discharges (Birge and Black 1981).
The embryo-larval tests gave greater detection of effluent toxicity. The
LC50 and LCI (percent effluent by volume) values for the four effluents, as
determined in flow-through tests, ranged from 0.3% to 29.4% and 0.001% to
2.8%, respectively. In the acute tests, it was not possible to calculate LC50
values for three of the four effluents. The acute LC50 for the most toxic
effluent was 8%, a value about 27 times greater than the embryo-larval LC50
(0.3%) and different from the embryo-larval LCI (0.001%) by more than three
orders of magnitude.
In three of the on-site studies, tests also were conducted to determine
the utility of embryo-larval biomonitoring for identifying toxic effluent
fractions and for evaluating the effectiveness of waste treatment processes.
Results indicated that more accurate evaluations could be achieved with
embryo-larval data that reflect the "net toxicity" of complex mixtures than
could be predicted with chemical analyses of the effluents or effluent
fractions.
In another major field investigation, chemical monitoring, ecological
surveys, hydrological measurements, and fathead minnow embryo-larval tests
were conducted on an effluent from a secondary sewage treatment plant (STP)
and the affected'receiving stream (Birge et al. in press). A good correlation
existed between embryo-larval toxicity and the degree of impact determined
using faunistic survey data evaluated at the different sampling stations. In
addition, an independent flow-through toxicity test performed on the STP
effluent and 4 effluent dilutions gave an effluent LCI value of 30.6% of
volume. The LCI compared closely with the instream effluent concentration
(33%) observed for the last downstream monitoring station, which was the only
station at which ecological survey data revealed no significant impact on fish
and macroinvertebrate populations.
EVALUATION OF EFFECTS PRODUCED BY SEDIMENT-ASSOCIATED CHEMICALS
It is evident that short-term embryo-larval tests have broad uses for
evaluating the toxicity of waterborne contaminants. More recently, this
procedure has been adapted for determining chronic effects of sediment
contamination. The importance of sediment toxicity and the development of
criteria for sediment-associated chemicals are receiving considerable
attention by the U.S. EPA and other interested parties. An in-depth treatment
of this subject has been presented in a recent volume edited by Dickson et al.
(1987).
277
-------
A number of different types of sediments, both naturally contaminated and
chemically enriched (spiked), have been evaluated using fish and amphibian
embryo-larval stages. Test procedures and findings from these studies have
been presented in earlier publications (Francis et al. 1984, Birge et al.
1987) . In one such investigation, experiments were performed to determine the
effects of mercury-enriched sediment on trout eggs. Samples of sandy loam
sediments were taken from a local stream with biotic conditions characteristic
of a healthy aquatic ecosystem and were enriched with inorganic mercury at
concentrations ranging from 0.1 to 100 Mg/g. Sediments were layered to a
depth of 2 cm in 500 mL-Pyrex test chambers and covered with 400 mL of
mercury-free reconstituted water. Eyed eggs of the rainbow trout were placed
on the sediment layer and exposed from 10 days before hatching through 10 days
after hatching.
A distinct inverse correlation existed between sediment mercury
concentrations and percent survival (Table 6). Trout embryos and alevins
accumulated significant quantities of mercury during the 20-day exposure, and
tissue concentrations rose proportionally with increasing sediment levels of
mercury. It is likely that contaminated sediments, when used as spawning
substrates, could present a formidable hazard to fish reproduction.
With respect to these results and those reported by others (Chu-fa et al.
1979, Swartz et al. 1982, Malueg et al. 1984, and Nebeker et al. 1984), there
is a considerable evidence that toxicological biomonitoring is a practical and
reliable procedure by which to evaluate sediment contamination. Taking into
account results based on early-life-stage toxicity tests, it is apparent that
bulk sediment chemistry is important for evaluating the hazard of sediment-
associated chemicals and that such tests should significantly augment the
development of regulatory criteria.
SUMMARY
The embryo-.larval toxicity tests described above have important
implications in the assessment of chronic toxicity produced by environmental
contaminants (Table 7). This procedure may be performed both in the
laboratory and in the field and can be used for development of aquatic and
sediment criteria, the establishment of structure-toxicity relationships, and
effluent and receiving water biomonitoring. Test endpoints observed in field
studies have been found to correlate closely with independent ecological
parameters used to quantify effects of point-source discharges on receiving
water systems. In addition, tests conducted with an effluent dilution series
are useful in estimating the likelihood or degree of instream impact (Birge et
al. in press). Not only is the test routinely employed for assessing chronic-
level effects under the NPDES biomonitoring program, but it is also applicable
for evaluating toxicity associated with hazardous waste sites in which liquid
discharges, leachates, or surface water contamination are involved. Moreover,
this methodology may prove to be an economical and predictive means by which
to prescreen chemicals for potential effects on human health.
278
-------
H
CD
O
04
E-i
g
g
M
f§
W
S
fn
O
CO
w
O
H
OO
w
fn
M
p-I
S
3
M
<^
O
r j
^ '
w
g
I-I
Q
~
W
35
U
M
25
K-J
/••^
P
C*!
w
a
_
o
oo
E-i
Cd
Cn
fn
W
•
vo
W
LyJ
^"j
H
C
o
•H
4J
cd
1-4
4-)
S
1
O
t*
H
O.
i
r— 1
cd
4J
0
^
ra
4J i— 1
d td .
o -,-<
M ^
0) S-4
PH 3
CO
...
"w>
bO CO
^a.
< +1
a) C
01 a;
01 S
t-i
CJ
a.
M
0)
^
01
CO .
p*
^
0)
g
o
. r-l CO
' id
o
. <3
j (
n) .
tH
8
0
&
• • •
o o 'o CM r—
+1 +1 +1 -H .+1
i— i m m co o
i-l CM i-l OO •*
O O O i-l VO
CM i-H CO
o CM
IS
•o o
o H
IX
r{ * — '
Ml ^
0 -J
_c C
4J cd
i 'c
4J -r-l
d ^
C ^
^ fT-1
CD
bO ^
cd 2
JJ -G
oi •'-'
bO w
bO <"
CD *"
' 5i
*O "
SL S
m ^7*
0)
s*. M
•jj
cd "—•
CD J»
CD '^
"5 ^
4J C/J
*XJ ' " '"
a
T^J ' 4"
nl • 'O 01
rt o 'd
,™ CM . CD
4J W
'g 0 cd
•i-l
01 |o ^
CD
4J P.TJ
c! Q)
£ CD 4->
B 3 C
0) O -rH
tj f^. y)
1^ S -§
279
-------
Table 7. APPLICATIONS OF EMBRYO-LARVAL TOXICOLOGICAL EVALUATIONS
Chemical criteria development
Ranking toxicity and animal sensitivity
Animal test species (species response range)
Chemicals (toxic effects range)
Structure-toxicity relationships
Field biomonitoring (effluents, sediments, hazardous waste sites)
Predicting ecological impact (receiving streams, sediments)
Surrogate testing for prescreening health effects (teratogenesis)
REFERENCES
Armitage, P. 1971. Statistical Methods in Medical Research.
Sons, New York, NY.
John Wiley &
Birge, W.J. and J.A. Black. 1977. Sensitivity of Vertebrate Embryos to
Boron Compounds. U.S. Environmental Protection Agency, Washington, DC
EPA-560/1-76-008.
Birge, W.J. and J.A. Black. 1981. In situ Acute/Chronic Toxicological
Monitoring of Industrial Effluents for the NPDES Biomonitoring Program
using Fish and Amphibian Embryo-Larval Stages as Test Organisms. U.S.
Environmental Protection Agency, Washington, DC. OWEP-82-001.
Birge, W.J., J.A. Black, J.E. Hudson, and D.M. Bruser. 1979a. Embryo-larval
toxicity tests with organic compounds. In: Aquatic Toxicology. L.L.
Marking and R.A. Kimerle (eds.). American Society for Testing and
Materials, Philadelphia, PA. STP 667. p. 131-147.
Birge, W.J., J.A. Black, and B.A. Ramey. 1981. The reproductive toxicology
of aquatic contaminants. In: Hazard Assessment of Chemicals - Current
Developments. J. Saxena and F. Fisher (eds.). Academic Press, New York,
NY. p. 59-115.
Birge, W.J., J.A. Black, T.M. Short, A.G. and Westerman. In Press. A
comparative ecological and toxicological investigation of a secondary
wastewater treatment plant effluent and its receiving stream.
Environmental Toxicology and Chemistry.
Birge, W.J., J.A. Black, and A.G. Westerman. 1985. A short-term fish embryo-
larval test for determining the effects of toxicant stress on early life
stages and estimating chronic values for single compounds and complex
effluents. Environmental Toxicology and Chemistry 4:807-821.
280
-------
Birge, W.J., J.A. Black, A.G. Westerman, and P.C. Francis. 1987. Toxicity of
sediment-associated metals to freshwater organisms: biomonitoring
procedures. In: Fate and Effects of Sediment-Bound Chemicals in Aquatic
Systems. K.L. Dickson, A.W. Maki, and W.A. Brungs (eds.). Pergamon
Press, New York, NY. p. 199-218.
Birge, W.J., J.A. Black, A.G. Westerman, and J.E. Hudson. 1979b. The effects
of mercury on reproduction of fish and amphibians. In: Biochemistry of
Mercury in the Environment. J.O. Nriagu (ed.). Elsevier/North Holland
Biomedical Press, Amsterdam, The Netherlands, p. 629-655.
Birge, W.J., J.A. Black, A.G. Westerman, and J.E. Hudson. 1980. Aquatic
toxicity tests on inorganic elements occurring in oil shale. In: Oil
Shale Symposium: Sampling, Analysis, and Quality Assurance, March, 1979.
C. Gale (ed.). U.S. Environmental Protection Agency, Cincinnati, OH. p.
519-534.
Birge, W.J., J.A. Black, A.G. Westerman, and B.A. Ramey. 1983. Fish and
amphibian embryos - a model system for evaluating teratogenicity.
Fundamental and Applied Toxicology 3:237-242.
Birge, W.J. and R.A. Cassidy. 1983. Structure-activity relationships in
aquatic toxicology.. Fundamental and Applied Toxicology 3:359-368.
Birge, W.J., A.G. Westerman, and O.W. Roberts. 1974. Lethal and teratogenic
effects of metallic pollutants on vertebrate embryos. In: Trace
Contaminants in the Environment, Proceedings of the Second Annual NSF-
RANN Trace Contaminants Conference, 1979. T. Novakov (ed.). University
of California, Berkeley. p. .316-320.
Black, J.A. and W.J. Birge. 1980. An Avoidance Response Bioassay for Aquatic
Pollutants. Water Resources Research Institute, Lexington, KY. Office
of Water Research and Technology, U.S. Department of the Interior,
Washington, DC. Research Report 123.
Black, J.A., W.J. Birge, W.E. McDonnell, A.G. Westerman, B.A. Ramey, and D.M.
Bruser. 1982. The Aquatic Toxicity of Organic Compounds to Embryo-
Larval Stages of Fish and Amphibians. Water Resources Research
Institute, Lexington, KY. Office of Water Research and Technology, U.S.
Department of the Interior, Washington, DC. Research Report 133.
Black, J.A., W.J. Birge, A.G. Westerman, and P.C. Francis. 1983. Comparative
aquatic toxicology of aromatic hydrocarbons. Fundamental and Applied
Toxicology 3:353-358.
Chu-fa, T., J. Welch, K.Y. Chang, J. Shaeffer, and L.E. Cronin. 1979.
Bioassay of Baltimore Harbor sediment. Estuaries 2:141-153.
Dickson, K.L., A.W. Maki, and W.A. Brungs (eds.). 1987. Fate and Effects of
Sediment-Bound Chemicals in Aquatic Systems. Pergamon Press, New York,
NY.
281
-------
Finney, D.J. 1971. Probit Analysis, 3rd Edition. Cambridge Press, New York,
NY. , ;.''"." '' ..
" t ' ;
Francis, P.C., W.J. Birge, and J.A. Black. 1984. Effects of cadmium-enriched
sediment on fish and amphibian embryo-larval stages. Ecotoxicology and
Environmental Safety 8:378-^387.
Hendricks, J.D., T.R. Meyers, J.L. Casteel, J.E. Nixon, P.M. Loveland, and
G.S. Bailey. 1984. Rainbow trout embryos: advantages and limitations
for carcinogenesis research. National Cancer Institute Monograph 65:129-
137.
Hodson, P.V., D.G. Dixon, and K.L.E. Kaiser. 1988. Estimating the acute
toxicity of waterborne chemicals- in trout from measurements of median
lethal dose and the octanol-water partition coefficient. Environmental
Toxicology and Chemistry 7:443-454.
Horning, W.B. and C.I. Weber (eds.). 1985. Short-Term Methods for Estimating
the Chronic Toxicity of Effluents and Receiving Waters to Freshwater
Organisms. U.S. Environmental Protection Agency, Cincinnati, OH.
EPA/600/4-85/014.
Konemann, H. 1981. Quantitative structure-activity relationships in fish
toxicology studies. Part 1: relationships for 50 industrial pollutants.
Toxicology 19:209-221.
Malueg, K.W., G.S. Schuytema, D.F. Krawczyk, andJ.H. Gakstatter. 1984.
Laboratory sediment toxicity tests, sediment chemistry, and distribution
of benthic macroinvertebrates in sediments from the Keweenaw Waterway,
Michigan. Environmental Toxicology and Chemistry 3:233-242.
McCarty, L.S., P.V. Hodson, G.R. Craig, and K.L.E. Kaiser. 1985. The use of
quantitative structure-activity relationships to predict the acute and
chronic toxicities of organic chemicals to fish. Environmental
Toxicology and Chemistry 4:595-606.
McKim, J.M. 1977. Evaluation of tests with early life stages of fish for
predicting long-term toxicity. Journal of the Fisheries Research Board
of Canada 34:1148-1154.
McKim, J.M. , G.F. Olson, G.W. Holcombe, and E.P. Hunt. 1976. Long-term
effects of methylmercuric chloride on three generations of brook trout
(Salvelinus fontinalis): toxicity, accumulation, distribution and
elimination. Journal of the Fisheries Research Board of Canada 33:2726-
2739.
Millemann, R.W., W.J. Birge, J.A. Black, R.M. Cushman, K.L. Daniels, P.J.
Franco, J.M. Giddings, J.F. McCarthy, andA.J. Stewart. 1984.
Comparative acute toxicity to aquatic organisms of components of coal-
derived synthetic fuels. Transactions of the American Fisheries Society
113:74-85.
282
-------
Mount, D.I., A.E. Steen, and T..J. Norberg-King. 1985. Validity of Effluent
and Ambient Toxicity Testing for Predicting Biological Impact of Five
Mile Creek, Birmingham, Alabama. U.S. Environmental Protection Agency
Duluth, MN. EPA/600/8-85/015.
Mount,_D.I. and C.E. Stephan. 1967. A method for establishing acceptable
limits for fish - malathion and the butoxyethanol ester of 2,4-D.
Transactions of the American Fisheries Society 96:185-193.
Nebeker, A.V., M.A. Cairns, J.H. Gakstatter, K.W. Malueg, G.S. Schuytema, and
D.F. Krawczyk. 1984. Biological methods for determining toxicity of
contaminated freshwater sediments to invertebrates. Environmental
Toxicology and Chemistry 3:617-630.
Norberg, T.J. and D.I. Mount. 1985. A new fathead minnow (Pimephales
promelas) subchronic toxicity test. Environmental Toxicology and
Chemistry 4:711-718.
Pickering, Q.H. 1974. Chronic toxicity of nickel to the fathead minnow.
Journal of the Water Pollution Control Federation 46:760-765..
Swartz, R.C., W.A. DaBen, K.A. Sercu, and J.O. Lamberson. 1982. Sediment
toxicity and the distribution of amphipods in Commencement Bay,
Washington, USA. Marine Pollution Bulletin 13:359-364.
U.S. Environmental Protection Agency. 1978. Federal Register 43:21506-21518.
U.S. Environmental Protection Agency. 1980. Ambient Water Quality Criteria
for Cadmium. U.S. Environmental Protection Agency, Washington, DC.
EPA/440/5-80-025.
283
-------
TOXICITY OF FENVALERATE TO SIX SPECIES
OF FISH AND TWO SPECIES OF FISHFOOD ORGANISMS
by
Ding Shurong1, Zhou Fengfen, and Zhang Min
Fenvalerate is one of the pyrethroid insecticides with broad spectrum
and high activity. It is highly toxic to fish and crustaceans when it enters
rivers and aquaculture ponds. Some information is needed about its toxicity
to aquatic animals, especially fish that have economic value and the organisms
on which they feed. In the past decade, some researchers studied the toxicity
of fenvalerate to many species of aquatic animals (Mulla et al. 1978, Coats
and O'Donnell-Jeffery 1979, Mcleese et al. 1980, Anderson 1982, Clark et al.
1985, Mckee and Knowles 1986, Bradbury et al. 1987). Few common fish with
economic value in China and fishfood organisms were included in these studies.
Only a few subacute toxicity studies of the toxicity of fenvalerate to aquatic
animals have been done. The present study evaluates both the acute toxicity
of fenvalerate to six species of fish with economic value in China and two
species of fishfood organisms (Daphnia and algae) and the subacute toxicity on
the aspects of behavioral toxicity (avoidance reaction), genetic toxicity and
accumulation toxicity of fenvalerate to fish. Acute toxicity of fenvalerate
to goldfish (Carassius auratus) and silver carp (Hypophthalmichths molitrix)
was evaluated to provide some ecotoxicological data for the safety assessment
of the pesticide.
MATERIALS AND METHODS
CHEMICALS
Fenvalerate (Technical, 90%) was supplied by the Research Institute of
Hormone, Jintan, Jiangsu Province. It was dissolved in ethanol, heated
slightly and then diluted to various concentrations for testing.
^•Department of Environmental Sciences, Nanjing University, Nanjing, PRC.
284
-------
STATIC ACUTE TOXICITY TEST
Test animals
FIngerlings of silver carp (Hypophthalmichths molitrix), average length
3.10 ± 0.15 cm, average weight 0.49 ± 0.15 g.
Fry of silver carp, obtained from the Aqualculture Farm of Nanjing City,
10 to 15 days old.
Fingerlings of the common carp (Gyprinus carpio), average length 3.78 ±
0.50 cm, average weight 1.06 ± 0.09 g.
Fry of the common carp, 20 days old. The common carp were obtained from
the Aquaculture Farm of Nanjing.
Finless eels (Monopterus albus). average length 24.4 ± 0.9 cm, average
weight 13.9 g, were obtained from the market.
Goldfish CCarassius auratus) , 20 to 30 days after hatching, obtained
from the market, with average length of 1.0 to 1.5 cm and average weight of
0.3 to 0.5 g.
Grass carp (Ctenopharyngodon idellus). average length 3.96 cm, average
weight 1.1 g, obtained from the Aquaculture Farm of Nanjing.
The loach (Misgurnus anguillicaudatus). average length 11.9 cm, average
weight 13.6 g, obtained from the market.
All the fish were laboratory-acclimated for 7 to 10 days before testing.
They were not fed 24 hours preceding or during the 96-hour exposures.
Pregnant daphnia (Daphnia carinata) were cultured in 1000-ml beakers 12
hours before testing. Each beaker contained 30 animals. After culturing for
12 hours, the animals were filtered and screened with webs to obtain the
newborn daphnia with instar of 12 hours.
Algae (Scenedesmus obliquus and Chlorella pyrenoidosa1) were cultured
with Bold Basal solution (Stein 1973) and inoculated to 1000-ml flasks
containing culture media under the condition of 24 to 26°C, 2000 lux. The
cultures were transferred at each 120-hour interval. After three transfers,
the algae culture reached a stage of synchronus growth and they were ready for
testing.
Water quality for fish test: pH 6.0 to 6.5, conductivity 220 to 310 x
102 ^mho/cm, dissolved oxygen 7 mg I"1, COD^ 1.03 to 1.19 mg I"1, chloride 7.0
to 8.5 mg I"1, hardness 5.38 to 6.05 (Deutsche degree).
Water quality for daphnia test: Tap water filtered by activated
charcoal and aerated with micropumps, dissolved oxygen >7.6 mg I"1,
conductivity 1.25 To 1.60 X 102 ^mho/cm.
285
-------
Bioassay procedures: Static exposure tests using fingerling were run
using 40-L aquariums, each containing 10 randomly assigned fish. A series of
five exposure concentrations and a control were used. Mortality counts were
recorded once every day during the 96-hour exposure. LC50 values were
calculated (APHA, AWWA, WPCF, 1980). The static exposure tests using fry were
run using 1000-ml and 5000-ml beakers.
A series of five exposure concentrations and a control were used for
static acute toxicity tests for daphnia. Ten parallel groups were set up for
each concentration. Each ten animals in one 150-ml beaker containing 100 ml
of test solution were tested under the condition of 21 ± 1°C and 12 to 14-hour
light periods with the intensity of 4500 to 5000 lux. Mortality counts were
recorded once every day during the 48-hour exposures. The 48 hour-LC50s were
calculated according to the data obtained (APHA, AWWA, WPCF, 1980).
To perform the static acute toxicity test for algae, the prepared test
solution of fenvalerate was added to each 70-ml test tube containing fresh
culture media for algae to make up test solutions having different
concentrations, i.e., 1000 ppm, 100 ppm, 10 ppm, 1 ppm, 100 ppb, 10 ppb, and 1
ppb (6 tubes for each concentration). An ethanol control and a blank control
were used.
After making cell suspensions of the algae, with cell cultures attaining
the stage of synchronous growth, the cell densities were counted. The cell
densities of the cultures of chlorella pyrenoidosa and Scenedesmus obliquus
were 5.31 x 107 cells/ml and 7.15 x 106 cell ml, respectively. Cell
suspensions of 1.13 ml and 2.09 ml were inoculated to 27 tubes containing
culture medium with 3 tubes for each concentration. The initial
concentrations of chlorella pyrenoidosa and Scenedesmus obliquus were 2 x 106
cells/ml and 5 x 10s cells/ml, respectively.
Culture conditions: 25 to 29°C, 2500 lux, light/dark = 14:10, shaking
speed 110 rpm. The growth amounts during the culturing periods of 24, 48, 72,
96 and 120 hours were determined using a cell counting method.
ACCUMULATION TOXICITY TEST
Accumulation toxicity tests using loach (average length 11 to 13 cm,
average weight 10 to 12 g) were run using 401 aquariums each containing 20
randomly assigned fish, divided into treatment groups and control. The tests
were done by exposure with toxicant dose increasing progressively. The amount
of toxicant in each step was 1 to 1.5 times the amount of the preceding
period. The increase of doses continued until half of each group of fish were
dead. The accumulation coefficients, K, were calculated by the WHO formula
(WHO 1978)
K
LD50(n)
LD50(1)
286
-------
where: LD50(n) = accumulative total dose for 50% dead animals
exposed repeatedly , . . . .
LD50(1) = dose for 50% dead animals exposed once
TEST FOR INDUCTION OF MICRONUCLEI IN THE PERIPHERAL BLOOD ERYTHROCYTE OF THE
LOACH
The loaches (average length 11.9 cm, average weight 13.6 g) were
laboratory acclimated for 7 to 10 days before testing. Feeding ceased before
and during the exposure. The loaches were exposed both by contact with
fenvalerate solution and by intraperitoneal injection. Three to five dose
groups were set up according to the 96-hour LC50 of the loach. A positive
control with mitomycin and two solvent controls with ethanol and dimethyl
sulfoxide were included in the experiment. Peripheral blood was taken with a
heparinized capillary tube after severance of the caudal peduncle. Blood
smears were made and slides were fixed in methanol and stained with Feulgen.
The erythrocytes were examined with a microscope.
AVOIDANCE TEST
Test animals: silver carp (average length 3.6 Cm, average weight 0.64
g), the common carp (average length 3.1 cm, average weight 0.9 g), and grass
carp (average length 3.96 cm, average weight 1.1 g). The fish were
laboratory-acclimated for 7 to 8 days before testing. Feeding ceased before
exposure. Water for dilution was aerated and dechlorinated tap water. Four
to six treatment groups between 1/13 to 1/2 96-hour-LC50 of the carp were
chosen and a "clean" control and solvent (ethanol) control were set up.
Avoidance reaction tests were run in straight-type avoidance equipment
provided with constant water flow. Experiments were repeated four times for
each concentration. Each experiment was done with ten fish, which were put
previously into the avoidance test equipment for 15 minutes to acclimate.
Each experiment lasted for 20 minutes. The frequencies of appearance by fish
both in "clean" region and in polluted region were recorded separately each
minute. Calculated data were judged by X2 test.
TEST FOR TOXICITY OF SOAKED-OUT LIQUID THROUGH ADSORPTION BY SOIL
Test animals: fry of silver carp (10 to 15 days after hatching) were
obtained from Aquaculture Farm of Nanjing City and goldfish were obtained from
the market with average length 1.0 to 1.5 cm and average weight of 0.3 to 0.5
g. Soil from a rice field was collected from the suburbs of Nanjing City.
The air-dried soil was put into 13 beakers (5000 ml in volume) with each 400
g. Four groups were divided according to the duration of adsorption, i.e.,
24, 48, 72 and 96 hours. The concentrations for each group were 100 //g I"1,
500 ng I"1 and 1000 /ig I"1. A control group was included. The amount of 3500
ml of fenvalerate solution with different concentrations was poured into each
beaker separately. After stirring with a glass rod, it was allowed to stand
287
-------
for precipitation. The amount of 2500 ml of soaked-out liquid were drawn from
3 beakers each at the interval of 24 hours to conduct the acute toxicity test
for goldfish. Ten fish were included in each experimental group. The method
of toxicity test for fry of silver carp was just the same as that for goldfish
but with smaller containers. The experimental temperature was 23 ± 1°C.
RESULTS AND DISCUSSION
THE ACUTE TOXICITY OF FENVALERATE TO FISH, DAPHNIA AND ALGAE
Fenvalerate is highly toxic to silver carp, common carp, goldfish and
loach (Table 1) . According to the classification on the degree of grade of
toxicity proposed by an international group (Joint IMCO/FAO/UNES CO/WHO Group
of Experts 1969), it is "very toxic" to these fish, but "toxic" to finless
eel. Mcleese et al. (1980) reported that fenvalerate was highly toxic to
Atlantic salmon juvenile (Salmon salar) : the 96-hour LC50 was 1.2 ng I"1 only.
Coats and O'Donnell-Jeffery (1979) found the 24-hour LC50 for fingerling of
rainbow trout (Salmo gairdneri) tested with fenvalerate was 76.0 ppb. 'The
acute toxicity (96-hour LC50) of fenvalerate to six estuarine fish was
determined in flow- through laboratory tests by Clark et al. (1985) and 96-hour
LC50's for cyprinodon variegatus. Menidia menidia. Menidia penissulae. Menidia
beryllina. Leuresther tenuis . and Opsanus beta were 5.0 /jg I"1, 0.3 pg I"1, 1.0
fig I"1, 1.0 pg I'1, 0.3 /*g I"1 and 2.4 /jg I"1, respectively. The values of 96-
hour LC50's for five species of fish with economic value exposed to
fenvalerate were coincident with these previous studies. Therefore, there is
seemingly a potential menace to fishery and it is worthwhile to understand
much about the toxicity of fenvalerate to aquatic animals.
TABLE 1. ACUTE TOXICITY OF FENVALERATE TO FIVE
SPECIES OF FISH AND DAPHNIA
Test organism
Silver carp (fingerling)
Silver carp (fry
Common carp (fingerling)
Common carp (fry)
Goldfish (fry)
Finless eel
Loach
Daphnia
Temp . ,
°C
91 4- 1
jLL IE J.
O o 4-1
/.:> ± i
O1 +1
on 4- i
i\J 21 L
23 ± 1
OR 4- 1
OR 4- • 1
OT -4-1
ZJ_ IE L
LC50(ppb)
24 h 48 h 72 h 96 h
•} 1 S 2 35
QC;
38
70
12.10 11.55 11.40 10.2
1100 1000 81 0
169
11 /.
288
-------
As one kind of fishfood organism, Daphnia do occupy an important
situation in the food chain and food web in aquatic ecosystems. It is very
sensitive to fenvalerate. Meyer and Ellesieck (1986) found that the acute
toxicity of fenvalerate to Daphnia magna was 2.1 /ig I"1. During the 21-day
test period, survival of Daphnia magna was not significantly (p = 0.05)
affected by the 21-day exposure; however, reproduction was reduced at
fenvalerate concentration of 0.25 and 0.5 ^g I"1 (McKee and Knowles 1986).
The results obtained by Anderson (1982) indicated that fenvalerate can affect
behavior and cause death for certain stream invertebrates at constant exposure
concentration of 0.030 //g I"1. The most sensitive animals in fenvalerate
exposure were the amphipods. Within 4 to 7 days, over 50% of the gammarids at
the low concentration of 0.022 fig I"1 were dead. The results of the present
study indicated that 48-hour LC50 for Daphnia carinata exposed to fenvalerate
was 1.14 /zg I"1. We think that datum showing high sensitivity as this may be
useful in the process of comprehensive analysis for the potential hazards of
fenvalerate to aquatic ecosystems.
With the toxicity test for Chlorella pvrenoidosa and Scenedesmus
obliquus, it was found that under low concentrations, fenvalerate showed a
promotive action to these two species of algae (Tables 2 and 3). Within a
certain range of fenvalerate concentrations, this action increased with
increasing concentrations. The stimulating actions appeared at the
concentration range of 100 ppb to 10 ppm. It is analogical to the data
obtained by Chen and Zeng (1982) in their study on the toxicity of some
pyrenthroids to algae. The inhibition effects of fenvalerate appeared at the
higher concentration. The EC50 for Scenedesmus obliquus and Chlorella
pyrenoidosa ranged from 100 ppm to 1000 ppm. In general, during degradation
pesticides containing nitrogen and phosphorus can liberate plant nutrients.
The CN" group in fenvalerate molecule may play a role in stimulation.
RESULTS OF ACCUMULATION TEST
According to the assessment criteria for accumulation (WHO 1978), when
K>5, it belongs to "slightly accumulated." As the K value obtained by us in
the present study equals 6.4, which is more than 5, there is only slight
accumulation of fenvalerate in the loach. The experimental method for
accumulation toxicity adopted by us is a method denoting functional
accumulation. There is a relationship between the intensity of functional
accumulation and the strength of metabolism as well as the degree of
functional change in fish body. But information about the accumulation
toxicity of fenvalerate to fish has not yet been reported. The data would be
useful to evaluate the toxicity of fenvalerate.
RESULTS OF MICRONUCLEUS TEST
Results obtained are listed in Tables 5 and 6.
Results indicated that the basal value for the rate of micronucleus in
the loach was rather low. It basically coincides with the data obtained by
289
-------
1
P5
S
M
O
Crf
W
g
crj
W
O
•o
•H
O
£ ^
P r-l
0) r-l
^0
i r— I
6 ~
[BERS OF
17ALERATE
S »
Ig
H-3 r
_J r^H
W °
U CO
0 g
M C~j
0 gj
1
»3 o
H °
CM
1
o a
O Pi
0 P.
o a
O Pi
r-l Pi
o a
I-l Pi
Pi
r, I
Pi
8 •&
i-i P.
0 ,0
p<
1—1 1—1
o o
H B
4J C8
§£
rj H
11
o cq
O «-»
to o
13 B 4J
3 rt id
2 H
- 01
a) r-i
H Ji
CO
CM
O
i-i
o
m
m
rH
o
r-i
o
.
00
in
to
t-i
VO
CO
i-i
VO
o
VO
o
r-i
P
a)
,Q
9
B
r-l
O
$
o
CM
i-l
34
in
CM
i— i
34
CO
r-l
r-l
34
to
i-l
i— 1
34
in
CM
CM
34
00
CO
to
let
34 b S
to W)
r-l 4J
CO rj rH
1 £
(u r-t
Q) O
J-J ^
] j
tH *0
0 ^
-------
1
ry/j
[x3
}-^'
fe
Q
_j
**•*
C^
g
CO
d
"
o
•i-i
<-* <— s
°t
w > .
3 , I
p* _J
M m
UJ Qj
rrj ^
0) in
S 0
W O
U
CO
pL) ft^i
O O
M
W H
O
"i c
3 m
a)
•r*4
p-1
a
p.
a
P<
P.
a
p.
a.
a
PI
P.
&
a
p.
,0
p,
o
p,
p.
^7
O
fi
4->
pg*|
Nrf^
?
r*» >* r**
o S5 *o
in Jl| CM
l-~ O
•* x '^
CO oO **"*
CO CO
m 5^ r-l
• VO
r-l -rl ,Q 1-H
•
CM
CO
CO
in
^
0
ON
j_j
O
.0
3
B
r-l
i-l
a)
O
*
CM
P-*
34
in
CM
5*6
o
CM
r-l
in
o
CM
X
in
CM
r-l
X
in
o
r-(
X
r-l
.
in
X
0
in
m
r-1
I-l
a)
o
-------
TABLE 4. ACCUMULATION OF FENVALERATE
Group
Number
of
Fish
Dose
1/20x0 . 02ml
2/20x0 . 02ml
Number
of
Fish
0
0
Accumulation
Mortality, coefficient
% (K)
Control
(ethanol) 20
Treatment
20
4/20x0 . 02ml
8/20x0 . 02ml
16/20x0. 02ml
32/20x0. 02ml
1/20 LC50
2/20 LC50
4/20 LC50
8/20 LC50
16/20 LC50
32/20 LC50
0
0
0
0
0
2
2
3
2
1
0 No accumulation
50 6.4
Liu (1981). After exposure, micronuclei do appear, but the difference of the
rates of micronucleus between treatment group and control group was not
significant. It denotes that fenvalerate may not be a cytogenetic agent to
fish.
RESULTS OF AVOIDANCE REACTION TEST
Results were expressed as the number of observations in clean water as a
percentage of the total number of observations. Significant avoidance was
judged to occur when the number of counts in the clean arm of the trough was
statically higher than in the polluted arm (X2>3.84, p<0.05) (Table 7).
According to the data shown in Table 7, the concentration in which silver carp
and common carp could avoid the fenvalerate were the 1/10 to 1/2 values of 96-
hour LC50 for common carp, and that for grass carp were the 1/8 to 1/2 values
of 96-hour LC50 (10.3 > X2 3.84, 0.05 > P > 0.01) .
The capacity of silver carp, grass carp and common carp to avoid
fenvalerate solution related to the concentration of fenvalerate. The
concentration threshold of fenvalerate fell in the range of 0.36 mg I"1 to
0.40 mg I"1 for silver carp and common carp and that for grass carp was 0.40
292
-------
TABLE 5. INCIDENCE OF FENVALERATE TO MICRONUCLEUS OF
THE PERIPHERAL NUCLEATED ERYTHROCYTES IN THE
LOACH (By Contact Exposure)
Dose,
ppm
Number of
erythrocytes
scored
Number of
micronucleus
Rate of
micronucleus ,
%
P
0.05
0.02
0.01
0.005
0.002
Mitomycin
(1.8 mg/kg)
ethanol
(0.01 ml)
Control
50000
30000
35000
25000
25000
10000
50000
60000
1
1
1
1
1
1
1
0.02 ± 0.02 >0.05
0.03 ± 0.03 >0.05
0.028 ± 0.02 >0.05
0.04 ± 0.04 >0.05
0.04 ± 0.04 >0.05
0.6 ± 0.24 <0.01
0.02 ± 0.02 >0.05
0.017
Control: tap water
mg I"1 to 0.50 mg I"1. The avoidance capacity of silver carp and common carp
seems stronger. The avoidance capacity would be an important factor to
evaluate the actual hazard of fenvalerate to fish.
RESULTS FROM TEXT FOR TOXICITY OF THE SOAKED-OUT LIQUID THROUGH ADSORPTION BY
SOIL
The acute toxicity of soaked-out liquid from soil adsorption decreased
rapidly along with the prolongation of time intervals for soil adsorption
(Figure 1, Figure 2).
The decreasing degree was greater for those groups with higher initial
concentrations. All the soaked-out liquid adsorbed by soil for 4 days did not
show any acute toxicity. The decrease of acute toxicity may be due to both
soil adsorption and aerobic degradation of fenvalerate (Sumito Chemical Co.
1982, Ohkawa 1980). Can and Chen reported that through an isotope tracer
experiment using "C-labeled fenvalerate they found that the mud in water
system showed a strong action of adsorption and it is responsible mainly for
the water purification. Fish could not only take up but also liberate the
fenvalerate residues very quickly from and to waters.
293
-------
To sum up, although the acute toxicity of fenvalerate to aquatic animals
was very high in laboratory experiments, the actual toxicity to aquatic
ecosystems in the real world would be less than predicted.
TABLE 6. INCIDENCE OF FENVALERATE TO MICRONUCLEUS OF
THE PERIPHERAL NUCLEATED ERYTHROCYTES IN THE
LOACH (By Intraperitoneal Injection)
Dose,
ppm
0.02
0.01
0.005
Mitomycin
(1.8 mgAg)
Dimethyl
Sulfoxide
(0.01 ml)
Distilled water
Number of
erythrocytes
scored
20000
20000
25000
10000
20000
30000
Number of
micronucleus
1
1
1
6
1
1
Rate of
micronucleus ,
% P*
0.05 ± 0.05 >0.05
0.05 ± 0.05 >0.05
0.04 ± 0.04 >0.05
0.6 ± 0.24 <0.01
0.05 ± 0.05 >0.05
0.03
*P>0.05, not significant
P<0.01, significant
REFERENCES
Anderson, R.L. 1982. Toxicity of fenvalerate and permethrin to several
nontarget aquatic invertebrates. Environ. Entomol. 11:1251-1257.
APHA, AWWA and WPCF. 1980. Standard methods for the examination of water
and wastewater, 15th edition. American Public Health Association.
Bradbury, S.P., D.M. Symonik, J.R. Coats, andG.J. Atchison. 1987. Toxicity
of fenvalerate and its constituent isomers to the fathead minnow,
Pimephales Promelas and blue, Lepomis macrochirus. Bull. Environ.
Contain. Toxicol. 38:727-735.
Chen Zhitao and Zeng Zhaoqu. 1982. Toxicity of 26 kinds of pesticides to
bacteria and algae in water. China Environmental Science. 2:24-28.
294
-------
TABLE 7. CAPACITY OF THREE SPECIES OF FISH TO SEEK
WATER FREE OF FENVALERATE
Species
Silver
carp
Grass
carp
Common
carp
Concentration ,
Pg/1
2.00 (1/2 96hLC50)
1.00 (1/4 96hLC50)
0.50 (1.8 96hLC50)
0.40 (1/10 96hLC50)
0.36 (1/11 96hLC50)
0.30 (1/13 96hLC50)
ethanol 0.02 ml
(ml/1000 ml)
tap water
2.00 (1/2 96hLCso)
1.00 (1/4 96hLC50)
0.50 (1/8 96hLC50)
0.40 (1/10 96hLC50)
ethanol 0.02 ml
(ml/1000 ml)
tap water
2.00 (1/2 96hLC50)
1.00 (1/4 96hLC50)
0.50 (1/8 96hLC50)
0.40 (1/10 96hLC50)
0.36 (1/11 96hLC50)
ethanol 0.02 ml
(ml/1000 ml)
tap water
Number of fish**
In In
water fenvalerate
22 3
17 6
15 5
16 7
13 9
9 7
11 9
11 10
18 6
17 6
18 7
14 9
9 8
12 10
26 7
22 8
21 10
22 10
10 9
16 15
14 14
Percentage
in water
87.4
73.0
74.3
69.5
59.0
54.6
5 2'. 3
50.0
76.8
75.2
71.6
59.0
52.8
53.8
78.3
72.4
71.0
68.2
53.3
51.0
50.0
X2*
value
13.90
4.66
4.58
3.89
0.850
0.101
0.045
0.046
6.84
5.72
4.76
0.76
0.056
0.13
10.56
5.61
4.34
4.22
0.848
0.019
0.0005
*X2 - P (3.84 - 0.05; 6.63 - 0.01; 10.83 •= 0.001), P > 0.05 not significant;
0.05 >. P > 0.01 significant; P £ 0.01 very significant.
**Does not include fish in the holding area.
295
-------
J?
(0
*-•
o
100
80 -
6O-
40-
20-
e e 1000ugL
o— — -e 500 ugL "1
G—~—o 100ugL"1
24
48 72
Time, hours
96
Figure 1. The acute toxicity of fenvalerate to silver
carp after adsorption by soil at different intervals.
Cheng Haihong, Li Shurong, and Yin Guihua. 1986. Study on the toxicity of
fenvalerate. Pesticide. No. 4.
Clark, J.R., J.M. Patrick, Jr., D.P. Middaugh, and J.C. Moore. 1985,
Relative sensitivity of six estuarine fish to carbophenothion,
chlorphyrifos, and fenvalerate. Ecotoxicology and Environ. Safety.
10:382-390.
Coats, J.R. and N.L. O'Donnell-Jeffery. 1979. Toxicity of four synthetic
pyrethroid insecticides to rainbow trout. Bull. Environ. Contain.
Toxicol. 23:250-255.
Can Jianying and Chen Ziyuan. 1986. Dynamics of fenvalerate in rice-water-
fish system. Acta Scientiae Circumstantiae. 6:263-271.
Joint IMCO/FAO/UNESCO/WHO Group of Experts on the Scientific Aspects of
Marine Pollution. 1969. Abstract of the report of the first session,
Water Research. 3:995-1005.
Lewis III, F.G. and R.J. Livingston. 1977. Avoidance of bleached kraft
pulpmill effluent by pinfish (Lagodon rhomboides) and gulf killifish
(Fundulus grandisl. J. Fish Res. Board Can. 34:568-570.
Liu Changjie. 1981. Study on the monitoring for environmental pollution
using nucleate erythrocytes micronucleus of vertebrates. Acta of
Huazhong Normal College (Natural science edition). No. 2.
296
-------
Mayer, F.L. and M.R. Ellersieck. 1986. Manual of Acute Toxicity:
Interpretation and Data Base for 410 Chemicals and 66 Species of
Freshwater Animals. U.S. Fish and Wildlife Service, Washington DC.
Mckee, M.J. and C.O. Knowles. 1986. Effects of fenvalerate on biochemical
parameters, survival, and reproduction of Daphnia magna. Ecotoxicology
and Environ. Safety. 12:70-84.
Mcleese, D.W., C.D. Metcalfe and V. Zitko. 1980. Lethality to permethrin,
cypermethrin and fenvalerate to salmon, lobster and shrimp. Bull.
Environ. Contamin. Toxicology. 25:950-955.
Mulla, M.S., H.A. Nawab-Gojrati and H.A. Darwazeh. 1978. Toxity of
mosquito larvidical pyrethroids to four species of freshwater fish.
Environ. Entomol. 7:428-438.
Ohkawa, Hideo, Ryoichi Kikuchi and Junshi Miyamoto. 1980. Bioaccumulation
and biodegradation of the (s)-acid isomer of fenvalerate (sumicidinR) in
an aquatic model ecosystem. J. Pesticide Sci. 5:11-22.
Stein, J.R. (ed.). 1973. Handbook of Phycological Methods, Culture Methods
and Growth Measurement, Cambridge, the University Press.
Sumito Chemical Co. Ltd. 1982. Selected Papers on Pesticides--Synthetic
pyrethroides, Sumicidin.
(0
*J
o
100 -
80 -
60 -
4O -
20 -
o 1OOOugL'
— -© 500 ugL M
100ugL"1
24 48 72
Time, hours
96
Figure 2. The acute toxicity of fenvalerate to gold-
fish after adsorption by soil at different intervals,
297
-------
WHO. 1978. Principles and Methods for Evaluating the Toxicity of Chemicals,
Part 1. Geneva.
Zhu Qinghua, Liang Xinglan, and Liu Fengming. 1987. Study on the toxicity of
fenvalerate, pesticide. No. 2.
298
-------
MODELING THE EFFECTS OF TOXICANTS
ON FISH POPULATIONS
by
Thomas G. Hallam1, Ray R. Lassiter2, Jia Li1, and William McKinney1
INTRODUCTION
The relationships between toxic chemicals and their effects on
populations are intricate and generally poorly understood. One reason for
this lack of understanding is that the biology of the stressed organisms
frequently is not considered in the determination of the risk associated with
an exposure to a toxicant. The first steps in current chemical assessment
procedures are generally based upon quantitative structure-activity relations
(QSARs). QSARs are mathematical expressions that relate biological activity
(molar concentrations causing quantal effects) to descriptors of molecular
properties of a sequence of chemical compounds. These approaches are based on
properties of the chemicals and do not include biological properties of the
exposed organisms.
A theme of this article is that the present theoretical basis for
determining effects of chemicals on populations is inadequate primarily
because past developments do not account for biological detail. This
inadequacy is magnified considerably when it is noted that an improper
investigative focus at the population level is usually employed. These
deficiencies have contributed to a lack of substantial developments in
ecotoxicology. Such hindrances are especially restrictive for the
foundational work relating to the determination of effects of toxicants on a
biological system.
Another theme of this article is that a proper focal level for
ecotoxicology must account for the individual organism. Indeed, in the
classical ecological organizational scheme, the individual is special and
unique. In the hierarchy from the cell to tissue to individual organism to
population, the levels below the individual are sets of genetically identical
elements, whereas the population is structured by genetic variation (Lomnicki
1988). It is individual variation that often is missing or suppressed in
•"•Department of Mathematics, University of Tennessee, Knoxville TN, USA
Environmental Research Laboratory, U.S. Environmental Protection Agency,
Athens GA, USA
299
-------
studies of the effects of chemicals on populations. This variability'is
needed to properly investigate and develop the appropriate theoretical basis
for ecotoxicology. Variation in the genetical, physiological and physical
distributions of individuals in a population together with the biogeochemical
environment of the population determines the characteristics of the effects
resulting from chemical exposure. In this article, we will generate some of
the genetic and physiological distributions associated with a fish population
that are relevant to the chemical effects problem.
Chemical impact occurs at the level of the individual not at the
population level. Even though the target site of a chemical may be specific
tissues, the exposed affected individual is the appropriate reference point
for extrapplation to the population level. A basic concept of toxicology,
susceptibility, intrinsically implies the existence of variance-variation- that
is viewed here as structure in the population. Susceptibility to chemical
exposure, an .individual property, is not static but is a variable related to
the dynamics of the individual. Dynamic susceptibility of individuals must be
reflected in the distribution of susceptibility for the population. A basic
premise of this article is that effects of chemical stress at the population
level are the cumulation of effects of the chemical on the individuals that
compose the population.
For investigations of effects, it is essential to note distinctions
between individual properties and population properties. Individual
properties include physiological variables such as body size and composition
as well as tolerance or susceptibility to a toxicant. Population properties
include distributions of individual properties such as distribution of
tolerances or susceptibility and moments of these distributions.
For modeling purposes, techniques exploring the role of the individual
in determining population dynamics are a relatively recent phenomenon but are
now evolving at a rapid rate (Metz and Diekmann 1986). Most of the original
works in mathematical ecology unfortunately do not reflect this individual
perspective, primarily because aggregation in many early studies was done at
the population level. Developments at the population or higher organizational
levels have proved to be generally unsuccessful in ecotoxicological studies
because individual variation is lost in these representations. Recent
progress in the analysis of individual-based population models is encouraging
because there have been significant developments in the area of accessible
computing power. This allows one to track large numbers of individuals in a
reasonable computational time period. We shall utilize an approach that first
develops a physiologically structured population model and then performs the
analysis by numerical techniques.
This article presents a theoretical study of the effects of a lipophilic
narcotic on a dynamic fish population. The work is theoretical because data,
facilities and techniques are not presently available to generate the
information needed for corroboration of the basic hypotheses or outcomes. The
conclusions obtained are conjectures that must be tested; however, because the
foundations of our model formulations are solidly grounded in the biological
and toxicological literature, we believe they merit the efforts necessary to
check their consistency. The effects literature is sparse because theoretical
efforts in ecotoxicology have virtually ignored biology. Our efforts
300
-------
represent an initial attempt to include relevant biology in an assessment
procedure.
Effects are limited for the present discussion to mortality in the
population. This restriction is not necessary but it is sufficient to
illustrate the procedure that we suggest. The rudiments of the underlying
theory for mortality in a static population are given in Lassiter and Hallam
(1989a). This static theory, intuitively nicknamed "survival of the fattest,"
is developed for acute chemical exposures and a static population. It
..assesses the effects of toxic exposure by relating the n-octanol-water
partition coefficient to the partition coefficient of the lipid and aqueous
phases of the animal, by hypothesizing equilibration within the body, and by
employing quantitative structure-activity relationships as a component of the
biological response assessment. Because the chemical is lipophilic, a known
distribution of lipid in the population is necessary to apply this static
theory.' We are aware of only two distributions of lipid in an aquatic
population [Brockway (1979) and Clark (1988, personal communication)]. These
distributions, for static fish populations, indicate that there can be much
variation in lipid content of fish of the same relative size in both
laboratory and field populations.
The dynamic behavior of individuals coupled with the possibility of
chronic or multiple toxicant exposures requires a new perspective. To our
knowledge, the dynamic distribution of lipid in any aquatic population is
nonexistent at the present time. Nondestructive sampling methods currently
are being developed and will ultimately lead to progress in this area. For
our purposes, however, it is currently necessary to obtain dynamic lipid
distributions by methods other than experimental ones. The approach espoused
here--to focus at the biological-chemical interface of the individual--
necessitates development of a dynamic representation of an individual
organism. The specific individual representation employed is one developed
for the purpose of determining the effects of a chemical on an individual
fish. The individual model, based upon energetics and described in detail in
Lassiter and Hallam (1989b) , is an important part of the population model. A
brief summary of the individual model is presented below so that its role in
the dynamics of the population can be understood.
Exposure and uptake of chemical also must be modeled to determine the
effects of a toxicant on a population. We utilize an analogue of the uptake
model, FGETS, developed by Barber et al. (1988). GETS and FGETS were both
formulated for fish and need little modification for our purposes.
Population dynamics are a compilation of all individual dynamics. We
model population dynamics by employing a partial differential equation that
incorporates individual dynamics explicitly in the population representation
and describes the behavior of the population in terms of a density function
that depends on individual physiological model variables and time. The
population model,' its behavior in the absence of the chemical, and the
behavior in the presence of the toxicant will be discussed here.
The primary objectives of this paper are to indicate theoretical
developments formulated to explore the effects of a chemical on a fish
population when toxic exposure is allowed through both the environmental and
301
-------
the food chain pathways. In a recent paper (Hallam et al. 1989), we have
studied the effects of different chemicals on a Daphnia population. In this
article, we study the effects of exposure duration on a fish population.
A MODEL OF INDIVIDUAL FISH DYNAMICS
The dynamics of an individual fish are well documented in the
literature; see books in the series edited by Hoar and Randall (1969).
Shul'man (1974) explores the role of lipids and their dynamics in fish
populations. Other papers that explore energetics and model growth in fish
population include Stewart et al. (1983) and Kitchell et al. (1974, 1977). We
know of no application of energetically based models to the assessment of
effects of a chemical on a fish population other than those presented here.
Kooijman and Metz (1984), Kooijman (1986), and Hallam et al. (1989) have
developed model formulations for Daphnia that are analogous to the ones
presented here. Philosophically, our model is closely related to these
efforts. Technically, this work is similar to our previous work on Daphnia;
however, the specifics of the individual (and subsequently, the population
model) are different than the Daphnia population model.
Any modeling project must be compatible with its objectives; hence, our
individual model must allow for relevant interaction with the chemical and
must be able to account for its toxicity. Appropriate model components must
be chosen with the specific chemical and type of exposure in mind.
Most industrial chemicals and many chemicals of environmental concern
are non-ionic, are nonreactive, induce baseline narcosis and are, to some
degree, lipophilic (Veith et al. 1983). The chemicals employed in our
illustrations are assumed to have these characteristics although the procedure
does not explicitly require any of these characteristics except the
lipophilicity and the nonreactivity. A method of determining the toxicity
effects threshold such as application of a QSAR also is needed. The lipid in
an individual buffers the action of a lipophilic chemical, allowing larger
body burdens in fatter individuals than in less fat organisms to elicit
biological response. Thus, additional lipid in an individual results in an
extension of the toxicity effect thresholds for an acute exposure. Lipid
storage provides protection against toxic stress from transient exposures only
if the organism is not forced to rapidly mobilize quantities of stored lipids.
When an organism rapidly utilizes stores of lipid in a situation where high
body burdens have been obtained, internal release of the chemical can lead to
toxicity effects under conditions where there is no change in the external
environmental concentration of a chemical. These considerations indicate
that, for the class of chemicals under consideration, a lipid compartment is
important and appears necessary in any individual model that is utilized to
represent the biological-chemical interaction.
A MATHEMATICAL MODEL OF A FISH
We now summarize the model of the life history of a fish. More detail
and background information can be found in Lassiter and Hallam (1989b).
Figure 1 is the conceptual model listing associated compartments and the flow
302
-------
1L
Resource
1P
Food Lipid
\
2L ;
i
Fecal
i
3L
Energy Integrator;
Mass to Energy
i i
Food Protein
t
3P
2P
5LR IT
4M
I W
4W
4R
*•"»»
3PR _
Egg Lipid
Energy Losses
Maintenance
Work (activity)
Reproduction
Egg Protein
Figure 1. Compartment and flow'diagram for the individual fish model.
chart for an adult female fish. The model assumes that the only inputs to the
lipid and protein compartments are obtained from a decoupling of the food
lipid and structure; that is, no synthesis of fat can occur from the
carbohydrates and proteins of the resource. This is not a valid hypothesis in
most situations (Vel'tishcheva 1961, Amlacher 1961, Menzel 1960) as Shul'man
(1974) indicates. On the other hand, the diets of many fishes contain small
amounts of carbohydrates that can be converted into lipids, so in these cases
this model might be useful.
Let mL and ms denote the mass of the lipid and mass of the structure,
respectively, in an individual organism. Structure is regarded as primarily
protein and carbohydrates. Each of these components are assumed to have both
labile and nonlabile portions. The nonlabile portion of the structure is
viewed as protein and carbohydrates bound in soma and is designated in the
model by mps, the mass of the protected structure. The labile portion of the
structure component is represented by ms - mps. The nonlabile lipid is
assumed to be proportional to mps and in the model is represented by efflps;
hence, the labile lipid is mL - emps.
The density of the resource is denoted by x and we assume that x - XL +
xs where XL and xs are the lipid and structural portions of the resource
density, respectively. The resource is assumed to be utilized for growth
according to a hyperbolic uptake law (Lassiter 1986). For details on the
particulars of the representation used here refer to Lassiter and Hallam
(1989b). The losses of energy for the maintenance and the activities
compartments are assumed to operate on a continuous time scale. Hence, on
intervals where there is no reproductive loss the fish is assumed modeled by
the differential equations:
303
-------
dmL
- -
XT
TED > AE
TED <: AE
(1)
dms
dt
TED > AE
TED < AE
The quantity ad is the encounter rate coefficient (liters per day)
obtained from Gerritsen and Strickler (1977) and is given by
8.64.10Wd(v2+3v2h)/1033vh
(2)
with sd defined as the reactive distance (cm) of the fish (Breck and Gitter
1983) and given by Sd - 4.775Lf1/2, where Lf denotes the length of the fish; vp
is the velocity of the prey (cm/s) and is given by vp = folsp . Lp where Lp
denotes the length of the prey and blsp denotes the body lengths per second of
the prey; and vh is the velocity of the fish (cm/s) while hunting for prey
defined by (blsh) (Lf) with blsh denoting the body lengths per second of the
fish while hunting. The symbol Wp represents the mass of the prey (g dry wt) ;
£v is the difference in the swimming velocities of the fish and its prey: Sy -
(vc - vp)8.64.10A; k is the gut emptying rate (1/cf) and Ms is the mass capacity
of the fish's stomach (g).
A generic derivation for the feeding rate of a pursuit feeder may be
found in Lassiter (1986, Equation 66). The formulation is a hyperbolic
function F — x/(A1 + Azx) in g/d where x is the density of the prey
population. When formulated in terms of mass consumed, the above form may be
written as a hyperbolic function of numbers of prey Np:
F =
The terms in the denominator relate to characteristic times (d/g) for
pursuit, encountering prey, and digestion, respectively:
T0=Sd/(5vWp) (d/g captured) ,
(d/g captured) ,
rd=l/(ktfs/) (d/g eliminated).
304
-------
The feeding rate is thus F = (Tc + Te + T^y1. In our computations, we
assume that the number of prey is fixed. The reproductive losses are assessed
at discrete times of reproduction. Reproduction is allowed in a time window
to represent seasonality aspects of reproduction. This reproduction interval
is utilized to set, .the birthing times for the specific species and is utilized
because we do not indicate hormone triggers such as temperature-related
phenomena. The losses associated with the individual fish model include
energetic needs associated with maintenance, activity and reproduction as well
as losses of lipid and structure for egg production. Maintenance energy
requirements are assumed to be proportional to the compartment. Activities
are modeled according to Gerritsen (1984) who states a relationship for power
consumed by swimming fish as
P = 0.275Av2-5ir°-5q-1 (ergsec'y1)
where A is wetted surface area (0.4L2), v is swimming velocity (part of the
time hunting, part of the time chasing and is given as the appropriate product
of bis and L), and q is swimming efficiency (0.2 according to Gerritsen).
Thus, after converting to appropriate power units
P - 4.75.10~351s5/2L4
or when cruising and chasing are included
P - 4.75.10-3L4Mf(&lsh5/2TeF+&lscs/2TcF)
The reproductive losses are assessed at the discrete times of reproduc-
tion. It is clear that the process of allocation of bulk mass to eggs occurs
over^a continuous time frame but, because there is little information on the
specific time scales of this process and because it is small relative to
population time scales, we treat them as discrete events. The reproductive
losses include bulk allocation to eggs and the energy required to deposit this
mass in the eggs. These operations, as well as the mechanism employed to
determine the number of eggs produced, are described in detail in Lassiter and
Hallam (1989b) and Hallam et al. (1989).
Examples of the numerical solution of the individual model are given in
Figure 2. 6
UPTAKE IN AQUATIC ANIMALS
The uptake model that is employed in connection with the above individual
model is a modification of the FGETS model (Barber et al. 1988). The model,
based upon thermodynamic potential, represents the chemical exchange between a
fish^and the aqueous environment that occurs across gill membranes and the
chemical exchange that occurs across gut walls and the contaminated ingested
305
-------
9.0
Figure 2. Lipid (MI) and structure (MS) cycles in the individual
model. The dynamics illustrate the decrease of size while in the
egg sack, exponential growth as a juvenile, and the reproductive
cycles. The nonlabile structure, mps, is a nondecreasing function
of age.
food. The model in its simplest form without food uptake is of the classical
bioaccumulation form; however, in this approach the parameters of the uptake
model are represented in considerable detail and include representations of
many factors that can influence internal concentrations such as the fractions
of the organism that are lipid (PL) , aqueous (PA) and structure (Ps) ; the
partition coefficients that indicate the affinity of the chemical towards
lipid (KL) and structure (Ks) ; the conductance of the exposed membrane (Kg);
the total weight of the organism (WT) and the active (effective) exposure8 area
(Sg) of the gill. The general form of the model is thus that of a classical
bioaccumulation model plus a term for exchange across the gut:
dt
s
(3)
In Equation 3, BT represents the total toxicant in the organism; Cw and
CF represent the concentrations of toxicant in the environment and in the
306
-------
food, respectively; F and E are the mass fluxes of food and feces, respective-
ly; BCF is the bioconcentration factor (total concentration in the organ -
ism/Cw) ; kE is the partition coefficient of chemical to excrement; CE = kECA,CA
is the concentration of toxicant in the aqueous portion of the organism; and
£>A is the aqueous distribution coefficient (D^CL+PgP^Ks+P^^K^'1 . The unit
conductance kg may be calculated from the molecular weight of the chemical and
the n-octanol- water partition coefficient. Barber et al. (1988) show that kg
is approximately N^D^h^'1 where Nsh is. the Sherwood number, Dw is the toxi-
cant's diffusion coefficient, which is a function of the molecular weight of
the chemical, and hw is the characteristic dimension of interlamellar chan-
nels .
A seemingly natural approach to model the uptake of chemical from the
food chain pathway would be to proceed by employing hypotheses similar to
those imposed for the environmental pathway. The complications of this
hypothesis and other options for modeling the uptake from contaminated food
are discussed in Barber et al. (1988).
The basic assumption of this model representation, that of equilibration
of chemical between the organism's body and the gut contents, is, of course,
not necessarily true. It has been demonstrated that this requirement is a
worst case assumption during increasing body concentration when exposed to
contaminated food; that is, no additional chemical could be taken up under any
thermodynamically consistent assumption than would be taken up when food and
body equilibrate. During depuration, however, this assumption leads to
predicted minimum depuration times; that is, any other thermodynamically
consistent assumption would lead to longer depuration times. For toxicity
evaluations, this would usually not be considered the worst case scenario.
EFFECTS OF TOXICANTS ON INDIVIDUALS
The basic ideas employed to assess the effects of chemicals on an
individual and on a static population are given in Lassiter and Hallam
(1989a) . We review these ideas to set the stage for this study on the effects
of a chemical on a population. Again the particular effect focus is on
mortality of the individual but sublethal effects could be considered by the
same methods .
The assessment of mortality due to chemical action is implemented by
utilizing formulations obtained from QSARs. The processes relating to mode of
action and concentration- response relationships must be coupled to determine
the effects. The procedure then consists of combining the following into an
assessment protocol:
(a) The differential equations for development of the individ-
ual dynamics .
(b) The differential equation for the total concentration of
toxicant (or, equivalently, the total body burden) in the
organism.
(c) The determination of mortality from QSARs.
307
-------
The first two of these steps couple the dynamics of the individual with
the dynamics of the chemical in the individual. The chemical uptake is
coupled to the individual through the weight terms in each of the Equations 1
and 2. Step c, the assessment of mortality, utilizes results of Veith et al.
(1983) and Konemann (1983) (see Figure 3). These bioassays are developed for
baseline narcotic chemicals and relate a chemical property, KOW, to mortality
of the individual.
There are numerous QSARs in the literature for modes of action other than
baseline narcosis. For example, the modes of polar narcosis (Veith et al.
1985) and uncoupling of oxidative phosphorylation (Schultz et al. 1986) have
been documented. A common feature of each of these modes of action is that
lipophilicity of the chemical as measured by KQW is important. We do not
present modes of action other than narcosis. Because the analogous QSARs are
based at least in part upon KOW, however, it appears that assessment of risk
due to exposure to a chemical with these other modes also must utilize lipid
as an individual model component.
u -
Ttll * _
— p —
o
10-3-
o
_j
2-4-
_J
-5-
-6-
-7-
Q n .-I
D °^
D
R] O
0 D
Q3
DD, ft,
o t?1 ff
a
O D
©a a a1^ gg apa
a C? D nD P
° ^ DaD
D ri^ci™
a ° j3
a ° cP
D DO
O .Q
1 1 1 I "1 "
-2 ' 0 2 4 6
Log Kow
Figure 3. Relationships between log LC5Q and log KQJJ. Fitted result
is log (LC5o)=-0.8 - log (KQW)• Data from Veith (1983) and Kone-
man (1981).
308
-------
AN ACUTE EXPOSURE-STATIC POPULATION THEORY:
SURVIVAL OF THE FATTEST
In a recent article (Lassiter and Hallam 1989a), we have developed an
approach that utilizes individual variation to structure a population and to
explain the effects of an acute toxicant exposure from a lipophilic narcotic
on that population. The basic idea is that lipid provides a buffer against
toxic stress and this factor must be utilized in any consideration of effects.
According to this theory, in an assessment of mortality, an individual with a
smaller lipid fraction body content will die before another individual with a
larger lipid fraction given equal exposure (Lassiter and Hallam 1989a). The
hypotheses for this theory are related directly to the static state of the
population. This static state allows only acute toxicant exposures. The
pathway of exposure is not important for effects on static populations.
A DYNAMICS THEORY: EFFECTS OF TOXICANTS ON POPULATIONS
Our prototype dynamic population model is based on individual response so
that effects can be determined directly and the cumulative effect at the
population level ascertained. The toxicant-population model is formulated so
that a toxicant may be released at numerous times and for arbitrary exposure
lengths during the study period so that chronic as well as acute exposures may
be investigated.
First, we will sketch the approach used to model the population. Then,
we will discuss the effects of the toxicant on the population due to duration
of exposure.
An approach that allows incorporation of individual dynamics into a
dynamic population formulation is the McKendrick-von Foerster equation (Metz
and Diekmann 1986). This partial differential equation allows explicit
representation of physiological variables as they determine the dynamics of
individuals. It also keeps track of the total population through the popula-
tion density function. The assessment of effects of a chemical that is a
lipophilic narcotic mandates that the individual model should minimally
include some measures of the physiological variables age, lipid, and struc-
ture. Age is necessary to reflect the life history of an individual. Lipid,
a bioconcentration site for the toxicant, is necessary to assess the effects'
of the chemical on the individual. Structure is necessary to account for size
related measurements such as weight or length of the organism.
If p=p(t,a,mL^,ms) is the population density function--which depends upon
time and the physiological variables a, representing age; mL, representing the
mass of the lipid, and ms, representing the mass of the structure compartment-
-and gL and gs are the growth rate of the lipid and structural components of
an individual as given by the Equations 1 and 2 respectively, then an equation
that incorporates these physiological variables into a population scheme is
309
-------
The subscripts denote partial derivatives with respect to that subscript.
The birth process for the population is represented by a boundary condi-
tion, and the mortality rate is given explicitly in the differential equation
as /L*. For physiologically structured models, the particular form of the birth
process may be written in several equivalent representations. One of these
forms is
,, - III
/3 (t, a, mLo, mS(), 2
where mLo is the mass of the lipid and mso is the mass of the structure at
age-0 and 0 is the birth function which represents the number of eggs with
lipid content mLo, structure content mso born to an individual of age a with
lipid content m^ and structure content ms at time t. One can use this form
because there is a one-to-one correspondence between lipid or structure mass
and age for fixed mLo.
Several different types of mortality are represented in our numerical
model formulation. We include formulations for age dependent mortality, size
dependent mortality, and density dependent mortality. The age dependent
mortality is assessed uniformly along cohorts, whereas the density dependent
mortality is assessed uniformly across the population. The size dependent
mortality is viewed as possibly caused by predation and is determined by
weight of the individual. The population model is, in general, nonlinear but
it would be linear if the density dependent mortality term were omitted. Some
restriction on mortality generally is needed to prevent the population from
growing beyond reasonable size. The specific forms of these mortalities are
generic and have a form that can be structured to test processes. Formula-
tions of the mortalities are given in Figure 4.
This population model is a first order hyperbolic partial differential
equation that may be represented in an alternate manner by the method of
characteristics. In this method, the partial differential equation is reduced
to a set of ordinary differential equations that are valid along certain
special curves called characteristics. Specifically, in this model an equiva-
lent representation for the partial differential equation is the following
system of five ordinary differential equations.
d\
dX
310
-------
Individual Mortality:
Weight-Dependent Mortality : fj,w = jw . fiiw = jw . fjilw(W), W = weight (ing)
Vc,
continuous and linear, elsewhere
Age-Dependent Mortality:
•{
Ka, 0 < a < 50
oo
a>50
Population Mortality:
Density-Dependent Mortality : fj,D = fiD(PB), PB = total population biomass.
PB = Po
PB>PC
continuous and linear, elsewhere
Figure 4. Mortality functions used in population model.
dmL
d\
"dA "*s
dp
dX
'n>L
Our approach is to solve a transformation of this system of codes numeri-
cally. Examples of graphical representations of these numerical solutions are
given later for several situations of interest.
The behavior of the population in the absence of a toxicant is oscil-
latory on several scales. There are oscillations on the time scale of the
311
-------
(assumed periodic) reproductive time (gestation period) of the species. There
are longer term oscillations that do not seem to be related to mortality but
rather to the rate at which the organisms grow. This longer term oscillation
is not apparently due to the dynamics of the resource, which is assumed to be
at constant density. In particular, this is not a typical predator-prey
oscillation where both predator and prey have oscillatory behavior. However,
this oscillatory behavior of a consumer when the resource is at a relatively
constant level is characteristic of some aquatic populations. (Murdock and
McCauley 1985) that do oscillate in the presence of a nonoscillatory resource.
The cause of the longer term fluctuations is apparently not a result of the
assumed density-dependent regulation. We mention this because the folklore in
rudimentary nonstructured models indicates that oscillations are often caused
by inclusion of density dependence (Hallam 1986.).
Our numerical procedure essentially follows cohorts of individuals along
characteristics. This allows effects of toxicant exposures to be assessed at
the individual level, a desirable attribute as we have indicated above, but
the overall effects on the population can still be determined by an accumula-
tion of individual effects. The addition of a toxicant leads to another type
of mortality assessment. This mortality due to toxicant exposure is assessed
at the individual level according to lipid content of the individual. Hence,
in the model, there are four essentially distinct causes of death: mortality
due to the physiological process of aging, mortality due to size (as, for
example, determined by predation), mortality due to density dependence (as,
for example, determined by total population size), and mortality due to
toxicity (as, for example, determined by the lipid distribution in the popula-
tion) .
The population structure and analysis as reported here is based upon 27
different types of individuals. These individual ecotypes (morphs) are
determined by the constant level of resource at which they feed, the quality
of that resource as indicated by its lipid content, and the gut clearance rate
of the organism. Each of these three individual characteristics ranges
through three levels for the total of 27. Any number of individuals can be
employed in the model. We employ 27 because the diversity generated by the 27
ecotypes is considerable. Figures 5 and 6 indicate the graphs of several of
the ecotypes of individuals. The parameters we employ are selected from the
literature on trout although there are several requiring estimation (Lassiter
and Hallam, 1989b).
The population model records the dynamics of cohorts of individuals in
the population, assesses mortality in these cohorts, and indicates births.
Organisms are assumed to be clones of their female parent; this may not be a
proper hypothesis for fish, which are not parthenogenetic.
Our objectives are to indicate the processes that affect studies on
effects of toxicants on populations and, although the importance of many
environmental parameters on ecological systems is recognized, we have not
included important relationships such as seasonality, temperature, and pH in
process formulations. In general, these environmental variables affect only
the parameter values in our process representations and not the formulations
themselves.
312
-------
°o.o
219.0
219.0
days
21'S.O
Figure 5. The graphs of nine ecotypes of fish utilized in structuring
the population. [The differences between the ecotypes are the resource level at
which they fend and the fraction of lipid in the'food. Another variable used to
generate variation in the population is the gut clearance rate; the variation caused
by changes in this parameter are not presented here but they are used in the popula-
tion model (smallest mg).]
313
-------
Figure 6. The Climax Population. [A. The dynamic lipid density function of the
climax population on the interval 2000 to 2737.5 days. B. The dynamic structural
mass density function of the climax population on the interval 2000 to 2.737.5 days.
C. The dynamic age density function of the climax population on the interval 2000
to 2737.5 days.]
314
-------
TOXIC EFFECTS AT THE POPULATION LEVEL
As we have noted, survival of the fattest is a theory obtained by con-
sidering toxic exposures to a static population-exposures that are necessarily
acute. We have demonstrated in previous work that this theory is not valid
when the hypothesis of a static population is violated.
Survival of a population after chronic exposure is determined not only by
the lipid distribution as in the survival of the fattest but now, in a dynamic
setting, also by the growth rate of the individuals in the population. When
the model population was stressed by a chronic exposure at a concentration
where there was some mortality but also some cohort survival, the particular
ecotypes of the surviving individuals depended upon many factors including
biological aspects of the population and toxicological aspects of the chemi-
cal. The biological features were related to the intrinsic oscillations of
the population, to the assessment of mortality, and to the reproductive
characteristics. Toxicological features of importance include the length of
exposure, the initial time of exposure, as well as the strength of the chemi-
cal.
The fact that exposure may come from both the environmental and the food
chain pathways complicates determination of cause of death or, equivalently,
reason for survivorship in a dynamic population. We shall focus upon a single
aspect of the exposure, namely its duration. This approach will illustrate
our procedure to assess the effects of a toxic chemical on our model fish
population. In each instance below we shall stress the population employing
the same chemical but we investigate the outcomes of different durations of
exposure.
The dynamics of the physiological variables--lipid,
as expressed through the respective distributions of the
on the time interval from 2007.5 days to 2737.5 days are
6a, b, c. This and all subsequent population graphs are
physiological variable size determined by the ecotypes.
scaling is to allow viewing of the larger individuals in
with this scaling, the smaller individuals still dominate
of their numbers.
structure, and age--
unstressed population
presented in Figure
scaled by a maximal
The purpose of this
the population. Even
the graphics because
A 4-8 day exposure, initiated on day 2025, to a 3.2 parts-per-million
concentration of a chemical having an octanol-water partition coefficient of
105 decreases the number of cohorts of fish in the stressed model population
from 85 down to 32. The surviving cohorts primarily consist of older in-
dividuals because no individuals in the younger age classes survive the
exposure. The population continues to persist until day 2502, at which time
it goes to extinction. The reason for this extinction is that the numbers of
individuals are small and the types of individuals grow sufficiently slowly so
that a threshold biomass in the density dependent mortality function is not
exceeded. Figure 7a, b, c represent the lipid, structural, and age distribu-
tions in this stressed population.
An exposure with the same chemical at the same concentration but of
duration 3.8 days instead of 4-8 days leads to a completely different con-
315
-------
Figure 7. Extinction caused by chemical stress. [Duration of exposure is
4.8 days initiated on day 2025. See text for details about chemical properties. A.
The dynamic lipid density of the stressed model population. B. The dynamic strusture
density of the stressed model population. C. The dynamic age density of the stressed
model population.]
316
-------
elusion. This population is now persistent over the study period. The number
of remaining cohorts that survive are fish that grow relatively fast. It is
their presence that allows this population to survive and their absence in the
above stressed population that contributes to its demise. Figure 8 demon-
strates the evolution of this stressed population. A comparison with the
unstressed population of Figure 6 shows relatively small differences in the
stressed and the unstressed populations.
SUMMARY
To demonstrate a proposed protocol for assessing the effects of a toxi-
cant on a fish population, we have developed a model of an individual fish and
incorporated these individual dynamics into a population model. Accordingly,
our approach focuses upon the individual -- the interface between the biology
and the chemistry. This allows investigation of the effects of toxicants on
populations by accumulating the effects on individuals and reinforces our
premise that the basis for studies in ecotoxicology should be the individual.
The model fish population was stressed by a chemical and the effects of
the duration of exposure studied. The ecotypes of the surviving individuals
were determined by both biological succession in the population and toxico-
logical aspects of the chemical. Biological succession proceeds to the
dominant- -fas test growing- -ecotype among all ecotypes that structure the
population. Toxic stress can alter the successional dominant ecotype from the
fastest grower ecotype (Hallam et al. 1989). Here we have seen that the
duration of exposure can also drastically influence the population evolution
A situation was presented where the faster growing morphs were eliminated from
the population and the population eventually went to extinction (Figure 7) A
shortening , of the duration of exposure allowed the population to persist
(Figure 8). Even though the surviving cohorts were not the fastest growing
morphs they grew sufficiently fast to overcome the threshold of mortality
present in the density dependent mortality representation.
As we have previously demonstrated for Daphnia. in a dynamic population
model, survival of the fattest is not valid, even when exposure is only from
the environmental pathway. For fish this statement is also valid Both the
ind^HeiTSented ^ FiSUreS ? Snd 8 indlcate « -ettlng where the surviving
individuals were not the fattest; they were the slower or the faster growers
of the ecotypes. The "survival of the fastest" grower is characteristic^
the evolutionary behavior of our model. The surviving ecotypes with the
largest growth rate will eventually dominate the population
experiments necessary to study our hypotheses and our conclusions
Hor then T? /^ in some instances, not even conceptualized.
However the models studied, here are well grounded in biological and toxico-
anfre1! 1^°™^™ *?* ^^ credible. It is a clear implication of tnis
attribute f ^^ ?* assessment should not be based solely upon
fundi Tl 5 t T1C Chemica1' The bi°l°gy °f the exposed organisms is
fundamental to the determination of the effects of the toxicant on a popula-
317
-------
Figure 8. Population persistence following a chemical stress of 3.8
days duration initiated on day 2025. [The same chemical is used as in the
situation represented by Figure 7. A. The dynamic lipid density function of the
stressed model population. B. The stressed density function of the stressed model
population. C. The dynamic age density function of the stressed model population.]
318
-------
REFERENCES
Amlacher, E. 1961., Das verhalten der inneren organs und der muskulatur
dreisommeriger Karpfen aus dem teich des dresdner zwingersbei extremer
kohlenhydratfutterung. Zeitschr. f. Ficherei ,9:749-761.
Barber, M. C., L. A. Suarez, and R. R. Lassiter. 1988. Modeling bioconcen-
tration of nonpolar organic pollutants by fish. Environ. Toxicol. Chem.
7:545-558.
Barber, M. C., L. A. Suarez, and R. R. Lassiter. 1988. Modeling bioaccumula-
tion of nonpolar organic pollutants by fish. (to appear)
Breck, J. E. and M. J. Gitter. 1983. Effect of fish size on the reactive
distance of bluegill (Lepomis macrochirus) sunfish. Can. J. Fish. Aquat.
Sci. 40:162-167.
Brockway, D. L. 1972. The Uptake, Storage and Release of Dieldrin and Some
Effects of Its Release in the Fish Cichlasoma bimaculatum (Linnaeus).
University of Michigan, Ann Arbor, MI. Ph.D. Dissertation p. 104.
Gerritsen, J. 1984. Size efficiency reconsidered: A general foraging model
for free swimming aquatic animals. Am. Nat. 123:450-467.
Gerritsen, J. and J. R. Strickler. 1977. J. Fish. Res. BD. Can. 34:73-82.
Hallam, T. G. 1986. Population dynamics in a homogeneous environment. In:
Mathematical Ecology: An Introduction. T. G. Hallam and S. A. Levin
(Eds.). Biomathematics Vol 17. Springer, Berlin 61-94.
Hallam, T. G., R. R. Lassiter, J. Li, and L. A. Suarez. 1989. Determination
of effects of lipophilic toxicants on dynamics of Daphnia populations.
Environ. Toxicol. Chem.
Kitchell, J. F., J. F. Koonce, R. V. O'Neill, H. H. Shugart, Jr. J. J.
Magnuson, and R. S. Booth. 1974. Model of fish biomass dynamics.
Trans. Am. Fish. Soc. 103:786-798.
Kitchell, J. F., D. J. Stewart, and D. Weininger. 1977. Applications of a
bioenergetics model to yellow perch (perca flavescens) and walleye
(Stizostedion vitreum). J. Fish. Res. Board Can. 34:1922-1935.
Konemann, H. 1981. Quantitative structure activity relationships in fish
toxicity studies. Part I. Relationship for 50 industrial pollutants.
Toxicology. 19:209-221.
Kooijman, S. A. L. M. 1986. Population dynamics on basis of budgets. In:
The Dynamics of Physiologically Structured Populations. J. A. J. Metz
and 0. Diekmann (eds.). Lecture Notes in Biomathematics. Vol. 68,
Springer, Berlin. 266-297.
319
-------
Kooijman, S. A. L. M. and J. A. J. Metz. 1984. On the dynamics of chemically
stressed populations: The deduction of population consequences from
effects on individuals. Ecotoxicology and Environmental Safety, 8:254-
274.
Lassiter, R. R. 1986. A theoretical basis for modeling element cycling. In:
Mathematical Ecology: An Introduction. T. G. Hallam and S. A. Levin
(eds.). Vol. 17, Biomathematics, Springer, Berlin, 341-380.
Lassiter, R. R. and T. G. Hallam. 1989a. Survival of the Fattest: A Theory
for Assessing Acute Effects of Hydrophobia Chemicals on Populations.
Environ. Toxicol. Chem.
Lassiter, R. R. and T. G. Hallam. 1989b. A physiological model of a fish.
In preparation.
Lomnicki, A. 1988. Population Ecology of Individuals Monographs in Popu-
lation Biology. Vol 25. Princeton Press, Princeton, N.J.
Menzel, D. W. 1960. Utilization of food by a Bermuda reef fish Epinephlus
gullalus. J. Cons. perm. int. Explor. Mer. 25:216-222.
Metz, J. A. J. and 0. Diekmann. 1986. The Dynamics of Physiologically
Structured Populations. Lecture Notes in Biomathematics, Vol. 68.
Springer, Berlin.
Murdock, W. W. and E. McCauley. 1985. Three distinct types of dynamic
behavior shown by a single planktonic system. Nature. 316:628-630.
Schultz, T. W., G. W. Holcombe, and G. L. Phipps. 1986. Relationships of
quantitative structure-activity to comparative toxicity of selected
phenols in the Pimephales promelas and Tetrahymena pyriformis test
systems. Ecotoxicol. Environ. Saf. 12:146-153.
Shul'man, G. E. 1974. Life Cycles of Fish. Physiology and Biochemistry.
Wiley, New York.
Stewart, D. J., D. Weininger, D. v. Rottiers, and T. A. Edsall. 1983. An
energetics model for lake trout, Salvelinus namaycush: application to
the Lake Michigan population. Can. J. Fish Aquat. Sci. 40:681-698.
Veith, G. D., D. J. Call, and L. T. Brooke. 1983. Structure toxicity rela-
tionships for the fathead minnow, Pimephales promelas: Narcotic in-
dustrial chemicals. Can. J. Fish. Aquat. Sci. 40:743-748.
Veith,-G. D., D. D. Foe, and M. Knuth. 1985. Structure-activity relation-
ships for screening organic chemicals for potential ecotoxicity effects.
Drug Metab. Rev. 15:1295-1303.
Vel'tishcheva, L. F. 1961. Penetration of G" from water into the body of
fish and its distribution in the organs. Trudy VNIRO. 44.
320
-------
TOXICITY OF ISOPRQTHIOLANE TO FISHES
by
Zhai Liangan1, Zhao Xiaochun, Yao Aiqin, and Li Jifang
INTRODUCTION
It is an inevitable trend that new pesticides will be substituted for
organic chlorine ones because of the toxicity of the latter's residuals to
environments, animals, fishes and even human beings, although they have played
an important role in the control of pests and insects in China. In recent
years, we have produced and imported organophosphors, pyrethrin, and other
pesticides but reports about their toxicity to fish are quite scarce. This
paper reports a toxicity experiment using a bacteriocide--Isoprothiolane. The
goal is to provide a scientific base for pesticides application, environmental
protection, and criteria and standards development for fisheries water
quality.
MATERIALS AND METHODS
Isoprothiolane (IPT), i.e. Fuji--one, NNF--109, is a heterocyclic
pesticide that was experimentally produced by Japan Pesticides Company in
1968. Since 1983, we have imported 2500 tons (95% emulsifiable concentrate).
The name of its effective chemical composition is, Ms(isopropyl)-1.3-dithio-
lane-2-ylidene-malonate, with a structure of
(CH3)2 CH - 0 - C
\
C - C
(CH3)2 CH - 0 - C s— I
0
Molecular formula: C12 H18 04 S2, molecular weight: 290.4.
Yangtze Institute of Fisheries, Chinese Academy of Fishery Sciences,PRC.
321
-------
Purified isoprothiolane is a white crystal, and its melting point is 54°C.
Its steam pressure is 1.41 X 10~4mmHg, and it is slightly dissolved in water
and easily in benzene, acetone, alcohol and other organic solvents. Its
solubility in water at 20°C is 48 ppm. Isoprothiolane is a systemic bac-
tericide for the control of rice blast, and can prevent physiological dis-
orders caused by low temperature, over-moisture in soil, etc.
TABLE 1. TEST ORGANISMS
Species
Standards
Simple description
Source
Common carp
Average body
length, 1.72 cm;
weight, 0.0368 g
Omnivore, living in
the lower layer
From the experi-
mental fish farm
of our institute
Silver carp
Average body
length, 1.80 cm;
weight, 0.046 g
Feeder on phyto-
plankton, living
in the upper layer
From the experi-
mental fish farm
of,our institute
Fish eggs of
grass carp
From fertilized
eggs td gastrula
stage
From the experi-
mental fish farm
of our institute
Daphnia
roagna
2 to 3 days age
A kind of fish
food feeding on
phytoplankton
Cultured in
laboratory
The water used in this experiment is tap water that has been exposed to
air. Its turbidness is 3.7; pH 7.57; total hardness is 8.6 (Germany Hard-
ness); dissolved oxygen meets fisheries requirements.
ACUTE TOXICITY EXPERIMENT WITH FISH
Concentrations of isoprothiolane for experiments with common carp
(Cyprius carpio") were determined to be 8.0, 2.7, 0.9, 0.3, and 0.1 ppm by the
equidistant transformation method of logarithm; those for silver carp (Hypo-
thalmichthys molitrix") were 4.5, 2.5, 1.4, 0.8, and 0.45 ppm. Each group had
a replicate and a control. Poisoning symptoms and the mortality rate of
fishes were noted, then the 96-hour LC50 was obtained by the Karber method.
The results were multiplied by 0.1 to find the safety concentration.
322
-------
GROWTH EXPERIMENT WITH FISH
The experiment carried out by stocking 10 fish in 0.4 ppm (safety
concentration for common carp), 0.85 ppm, 0.45 ppm, 0.25 ppm (survival
concentration for silver carp) and in pesticide free water (control). Fish
were fed in fixed quantity for 28 days. The growth rate was determined by
measuring the body length and weight at the end of the experiment in
comparison with that at the beginning.
EMBRYONIC DEVELOPMENT EXPERIMENT
Concentrations of isoprothiolane for experimental groups were determined
to be 8.00, 4.50, 2.50, 1.40 and 6.88 ppm by equidistant transformation method
of logarithm. Twenty eggs at the gastrula stage were stocked in experimental
groups and a control group. Embryo lethality, hatching rate and survival rate
were noted periodically and compared with that of the control group.
EXPERIMENT WITH DAPHNIA MAGNA
Seven isoprothiolane concentrations were determined between 0.15 and 2.5
ppm, and a control and a replicate were set up. Ten Danhnia magna was stocked
in each group. Mortality was noted at 24, 48 and 96 hours, and the LC50 was
solved from those counts.
RESULTS AND DISCUSSION
In the acute toxicity experiment, the fish appear to be uncomfortable and
to swim at the water surface and to be sluggish in behaviour at the beginning
of the experiment. At high concentrations of isoprothiolane, fish would be
blackened in body color and some even died. Results in Table 2 show that
isoprothiolane has moderate toxicity to fish.
In the growth experiment (Table 3), isoprothiolane at 0.40 ppm had no
effect on the growth of common carp, but 0.80 ppm had a distinct effect on the
growth of silver carp. 0.45 ppm, 0.25 ppm and control had no effect on the
growth of silver carp, so the maximum functioning concentration of isoprothio-
lane on silver carp growth is 0.45 ppm.
In the embryonic development test (Table 4), it was shown that 2.50 ppm,
4.50 ppm and 8.00 ppm could affect development. The embryonic body in 1.40
ppm, 0.80 ppm and control group begins to wriggle slightly after 12 hours.
Hatching rates from high to low concentrations are 20%, 30%, 85%, 90%, 80% and
75%, respectively. Obviously, only 4.50 ppm and 8.00 ppm have a distinct
effect on embryonic development. Fingerlings in those concentrations with
high deformative rate all died after 96 hours. There is no significant
difference among 2.50 ppm, 1.40 ppm, 0.80 ppm and control group. The maximum
no-effect concentration for the embryo was 2.50 ppm.
In the daphnia experiment (Table 5), there was no significant difference
among 0.15 ppm, 0.27 ppm, 0.47 ppm, 0.84 ppm, 1.50 ppm and control group.
Daphnia magna should not be higher than 0.178 ppm.
323
-------
TABLE 2. ACUTE TOXICITY EXPERIMENT RESULT OF IPT TO FISH
Species
Common
carp
Silver
carp
Logarithmic
Cone en. value of
ppm Concen . (X)
8.0 0.903
2.7 0.431
0.9 -0.046
0.3 -0.523
0.1 -1.000
Total
4.5
2.5
1.4
0.8
0.45
0.25
Total
No. of Mortality
fish (r) r2 96 -hr LC50
10 10 100 4.138
10 11
10 0 0 95% confidence
10 0 0 limits
10 0 0 =4.13810.888
Sr=ll Sr2-101 SE=0.0477
3.180
95% confidence
limits
=3.180±0.360
SE=0.025
TABLE 3. GROWTH EXPERIMENT RESULT OF IPT TO COMMON CARP AND SILVER CARP
Beginning Ending Increase
Species Concen. Length, Weight, Length, Weight, Length, Weight, Expla-
ppm cm g cm g cm g nation
Common
Carp
Silver
Carp
0.40
0.00
0.80
0.45
0.25
0.00
1.72
1.72
1.80
1.80
1.80
1.80
0.0368
0.0368
0.046
0.046
0.046
0.046
2.03
2.02
2.35
2.46
2.44
2.54
0.085
0.072
0.0742
0.0940
0.0990
0.1175
0.31
0.30
0.55
0.66
0.64
0.74
0.0482
0.0352
0.0282
0.048
0.053
0.072
27 days
28 days
324
-------
TABLE 4. EMBRYONIC DEVELOPMENT RESULT OF IPT TO GRASS CARP
Concen. ,
8.00
4.50
2.50
1.40
0.80
Control
Number
of Eggs
20
20
20
20
20
20
Survival number3
6hr
20
19
19
20
20
19
9hr
20
19
19
20
20
19
12hr
18
18
19
20
20
18
24hr
18
17
19
18
17
16
28hr
4
6
17
18
17
15
34hr
4
5
17
18
17
15
72hr
4
5
15
17
17
15
96hr
All
died
5
14
17
16
15
aAt 110 hours, all had died.
SUMMARY
1. 96-hour LC50 values of isoprothiolane for common carp and silver carp
are 4.138 ppm and 3.180 ppm, respectively. 96-hr LC50 value multi-
plied by 0.1 are 0.4138 ppm and 0.1380 ppm, respectively.
2. 0.45 ppm of iosoprothiolane is the maximum no-effect concentration for
silver carp growth.
3. 4.50 ppm of isoprothiolane is the maximum no-effect concentration for
embryonic development.
4.
96-hour LC50 value of isoprothiolane for Daphnia magna is 1.780 ppm.
In conclusion, 0.318 ppm of isoprothiolane is a safe level with
respect to acute toxicity, growth embryonic development of Daphnia
rcagna survival. Therefore, it is advisable to recommend that 0.318
ppm (0.3 ppm) of isoprothiolane should be the water quality standards
for fisheries.
325
-------
TABLE 5. RESULT OF EXPERIMENT WITH Daphnia magna (96-hr)'
Concen. ,
ppm
2.50
2.00
1.50
0.84
0.47
0.27
0.15
Control
Daphnia magna
number
10
10
10
10
10
10
10
10
24 hr
10
10
10
10
9.5
10
10
10
48 hr
10
10
10
9
9
10
9
10
Result
72 hr 96 hr 96 hr LC50
0
3
9 8
9 9 1.7826168
8.5 7.5 ~ 1.80
8 9
8 9
9.5 9.5
"Water temperature 20±5°C. Results are for two replicate experiments.
326
-------
FISH ENZYMATIC INDICATORS IN AQUATIC TOXICOLOGY
by
Xu Lihong1, Zhang Yongyuan, and Wang Deming
INTRODUCTION
The response of an intact animal, such as death or a change of reproduc-
tive behavior, has been used as an index in classical toxicological tests.
Such responses, however, occur only when a toxicant concentration is relative-
ly high. The indices are comparatively simple and approximate. More and more
physiological and biochemical indices are being introduced into toxicology
now.
By measuring changes in the activities of different enzymes, changes of
metabolic or hematological parameters, or changes in the components of tissue
of animals inhabiting a certain water body or being exposed to toxicants, a
series of biochemical parameters can be obtained. Similarly, when selected
organ or tissue homogenate is exposed to toxicants, in vitro biochemical
parameters can be obtained. With these biochemical indices, toxicant action
mode could be explained, health of animal in the water could be evaluated,
early signs of potential pollution could be provided, fast screening of
toxicity of chemicals could be made, and effects of long-term exposure to low
toxicant dosage on an animal could be detected. Also, the biochemical indices
are helpful for determining the target organ of toxicant action. Some indices
can be used as specific indicators for the existence of certain pollutants.
Enzymes are one of the most important macromolecular proteins in the
body. Some enzymes have already been used as toxicological indices, and
others are potentially useful. The reaction between key enzymes of organisms
and toxicants has become a topic of mutual interest to toxicologists and
biochemists.
Adenosine triphosphatase (ATPase) plays an important role in ion trans-
portation and osmoregulation of fish. Acetylcholinesterase (AChE) inactivated
acetylcholine is released from the synamses of cholinergic nerves preventing
•"•Institute of Hydrobiology, Academia Sinica, Wuhan, PRC
327
-------
tetanic firing of the post-synaptic nerve. The effects of toxicants on ATPase
and AChE are presented in this paper. The possibility of the two enzymes
being used as toxicological indices and their mode of action are discussed.
MATERIALS AND METHODS
Grass carp (Ctenopharyngodon idellus). 11 to 14.5 cm long, were caught in
a fish pond and acclimated in dechlorinated tap water for 3 days. Paradise
fish (Macropodus opercularis) . 1-year old and 5.9 to 7.8 cm long, were
cultured in a laboratory, where temperature was kept over 15°C. For ATPase
assay (Paxton and Limminger 1983), the relevant tissue was homogenized in cold
buffer (containing 40 mM imidazole, 250 mM sucrose, and 5 mM EDTA at pH 7.1
and 25°C). The homogenate was centrifuged at 3000 rpm for 5 minutes and the
supernatant was kept for enzyme assay. Supernatant (0.2 ml) and 0.3 ml of
tested toxicant stock solution (0.3 ml of double-distilled water substituted
for the stock solution in the control and in in vivo tests) were incubated in
a reaction mixture at 25°C for 20 minutes, then 0.1 ml Na2 ATP was added to
the mixture. The reaction mixture in a final volume of 3.0 ml contained 40 mM
imidazole, 100 mM Na+, 20 mM K+, 2 mM Mg+ (without Na+ and K+ for NaK-ATPase
assay) and 2 mM NaaATP (pH 7.1, 25°C). After 15 min, the reaction was stopped
by 30% cold TCA, and the reaction mixture was centrifuged (3000 rpm) for 5
minutes. The supernatant was kept for inorganic phosphate determination.
Protein concentration of the enzyme was estimated with the Lowry method (Scope
1982). The specific activity of the enzyme was recorded in terms of micro-
molar Pimg"1 protein hr"1. For AChE assay, brain tissue was homogenated in
0.01M. phosphate buffer (pH 7.2). In the final homogenation, 2 mg of brain
tissue per milliliter of homogenate was obtained. One milliliter of this
homogenate was added to 1 ml of 0.004M bromine acetylcholine, then the mixture
was incubated for 20 minutes at 30°C. The remaining bromine acetylcholine was
measured according to the Hestrin method (Hestrin 1949). The specific
3 recorded in terms of micromolar bromine
activity
enzyme
choline mg'1 brain tissue hr"1.
jtyl-
In the Hg2+ treatment of grass carp, the fish were exposed to 0.2 ppm Hg2+
(15°C, pH 7.8). After 24 hours, NaKMg-ATPase activity of various tissues were
measured for both the experimental group and the control group.
In the Cu2* treatment of paradise fish, the fish was exposed to 0.1 ppm
Cu2* in a flow-through system (21-24°C, pH 7.2-8.0). After 4 weeks, brain
AChE activity was measured. Control (2.8 ^igCu2+/l) was measured at the same
time.
LAS (linear alkyl benzylsulfonate) and TCB (tri-chlorinated biphenyls)
were used as environmental analysis standard reagents. Others were analysis
reagents.
RESULTS
NaKMg-ATPase activities of grass carp caught at different times and
accli-ated 3 days in the laboratory were measured. Table 1 shows NaKMg-ATPase
activities of gill; Table 2 shows that of kidney. One-way variance analysis
328
-------
TABLE 1. GILL NaKMg-ATPase ACTIVITIES OF GRASS CARP CAUGHT AT DIFFERENT TIMES
Body length, cm Acclimated temperature, °C ATPase activity,
mean ± SD
11.5-12 19.3-20.5
12.5-13.5 13.0-13.5
12-12.5 13
11.5-13.5 12.2-14.0
12.5-13.5 14.5
13-13.5 17.2-19.0
11-12 19.2-20.2
13.5-14 15.5-18
18.69 ± 1.09
19.69 ± 1.00
17.41 ± 3.15
15.44 ±1.12
15.57 ± 1.07
16.25 ± 2.33
18.28 ± 1.95
19.58 ± 3.48
was carried out. When the significance level was 0.05, there is no sig-
nificant difference in activities of NaKMg-ATPase for both gill and kidney
between different groups of fish.
Plotting the probit of percentage of inhibition of toxicants on enzyme
and logarithmic concentration of toxicants, we obtained straight lines (Figure
1). I50, 50% inhibition concentration of toxicant on ATPase, was calculated
from a linear equation.
a: Hg2+ 1=4.70+0.51 InC I50 1.8 ppm (kidney)
b: LAS 1=4.14+0.63 InC I50 3.9 ppm (gill)
c: TCB 1=3.39+0.55 InC I50 18.4 ppm (gill)
d: Cd2+ 1=1.56+0.71 InC I50 127.2 ppm (kidney)
I is the probit of percentage of inhibition of toxicant. C is the concentra-
tion of toxicant.
Effects of Hg2+ on NaKMg-ATPase, NaK-ATPase, and Mg-ATPase are shown by
plotting inhibition percentage and toxicant concentration in Figure 2. As
evident from the figure, the inhibition of Hg2+ on NaK-ATPase was the strong-
est.
After 24 hours of exposure to Hg2+, gill and kidney NaKMg-ATPase in grass
carp were measured and the results are shown in Table 3. At the significance
level of 0.05, there is no significant difference between the test and the
control groups.
329
-------
TABLE 2. KIDNEY NaKMg-ATPase ACTIVITIES OF GRASS CARP CAUGHT AT DIFFERENT
TIMES
Body length, cm
Acclimated temperature, °C
ATPase activity,
mean ± SD
11.5-12
12-125
11.5-13.5
12.3-13
19.3-20.5
12-13
12.2-14
17.2-19
20.25 ± 0.94
26.09 ± 0.75
24.29 ± 4.12
24.61 ± 1.46
After 4 weeks of exposure to Cu2+, brain AChE activities of paradise fish
were measured and the results are presented in Table 4. At the significance
level of 0.05, there is no significant difference between the test and the
control groups. It seems that the non-neurotoxic chemical has no effect on
AChE.
DISCUSSION
Since the 1960s, studies of the reaction of ATPase to toxicants,both in
vivo and in vitro have been carried out in large numbers by scientists.
Different organisms were used as experimental materials, a wide variety of
chemicals were studied, and some work even delved into the study of enzyme
dynamics. In these studies, the probable mechanisms of the effect of
toxicants on ATPase have been scrutinized, and the mode of intoxication and
the target organ of the toxicants were deduced. Because quite a number of
chemicals can influence ATPase, such as organochlorine pesticide, heavy
metals, surfactants, DEHP, PCBs and petroleum refinery wastewater (Boese et
al. 1982, Davis et al. 1972, Kuhnert and Kuhnert 1976, Riadel and Christensen
1979, Srivastava et al. 1975, Verma et al. 1979a, Verma et al. 1979b, Yap et
al. 1971), some investigators proposed that ATPase activity could be used as a
nonspecific indicator in aquatic environment monitoring (Haya and Waiwood
1983). Because of -the sensitivity of ATPase to toxicants, many investigators
thought that ATPase activity could reflect the damage of a sublethal
concentration toxicant to an organism and predict potential pollution of
chemicals in the aquatic environment.
There is a major premise that fish ATPase could be an indicator. ATPase
plays a key role in .ionic and osmotic regulation in fish body, so changes in
ATPase activity might reflect the health or even the survival of the fish.
330
-------
p
o"'
p
<0
o o
- §
1
si
o
o
o o
CM i-
U)
I
•H
ft
•H
H
OJ
331
-------
TABLE 3. IN VIVO EFFECT OF Hg2+ ON GRASS CARP GILL AND
KIDNEY ATPase (n=4)
Specific activity, mean ± SD
gill
kidney
control
0.2 ppm Hg2+
17.05 ± 3.15 \
16.48 ± 4.54
22.38 ± 5.28
24.08 ± 5.83
After 3 days acclimation in the laboratory, there is no significant
change in ATPase activity of fish caught from fish-pond. This is important
for the in vivo tests. When an in vitro test is carried out, only a small
quantity of tissue is needed. So, from the view of experimental material, it
is practicable to use grass carp tissue as an enzyme source or to carry out in
vivo tests with grass carp.
All the toxicants studied in our tests have obvious effects on ATPase
under in vitro conditions. Such regularities were found on the basis of the
effects of Hg2+, LAS, TCB, CDZ+ on ATPase, that is, linear relationships were
obtained from the plotting of probit of inhibition percentage and logarithm of
toxicants concentration. This is the same as found in biological tests. This
means that an in vitro test does show some characteristics of the toxic
action.
Kg2* had greater inhibition on kidney NaKMg-ATPase than on gill NaKMg-
ATPase. In Figure 2, it is demonstrated that Hg2+ had stronger inhibition on
NaK-ATPase than on Mg-ATPase. NaK-ATPase activity in the kidney is higher
than in the gill in freshwater fish (Jampol and Epstein 1970).. So, when
NaKMg-ATPase was used as an index, the change of kidney NaKMg-ATPase is more
sensitive than that of gill NaKMg-ATPase. This indicates the need to search
for several tissues in order to select the most sensitive indicator. For the
same reason, Cd2+ had great inhibition on kidney NaKMg-ATPase than on gill
NaKMg-ATPase. Considering the properties of NaK-ATPase on the one hand and
metal on the other hand, metal should affect NaK-ATPase more readily than Mg-
ATPase .
After 24 hours of exposure to 0.2 ppm Hg2"1", grass carp gill and kidney
NaKMg-ATPase were unaffected and death occurred at 0.4 ppm Hg2+ exposure.
Sastry and Sharma (1980) reported the results of this experiment in which
freshwater fish were exposed to 1.8 ppm Hg+ for 96 hours and to 0.3 ppm Hg2+
for 15 days and 30 days. ATPase activity increased a little in the 96-hour
acute exposure but decreased slightly in the 15-day and 30-day exposures,
although death occurred in all of the test groups. Maybe the lethal toxicity
332
-------
TABLE 4. IN VIVO EFFECT OF Cu2+ ON PARADISE FISH BRAIN
AChE
Specific activity, mean ± SD
(mean ± SD)
control
100 jugCu2+/l
1.79 ± 0.36 (n=6)
1.47 + 0.20 (n=5)
of Hg2+ to fish was not the inhibition on gill and kidney ATPase but its
effect on other enzymes or tissues. For example, death probably resulted from
the suffocation of fish owing to the deposition of Hg2+ on gill epithelia, a
situation that would not be revealed in the in vitro test. So it is helpful
to integrate the in vitro and the in vivo test results in deciding the
probable mode of toxicity on organisms.
From previous reports and the present research, it is obvious that ATPase
activity is affected by varied chemicals and a consistent pattern can be
obtained. The 50% inhibition concentration of toxicant on enzyme (I50) ,
corresponding to LC50, also could be determined. So, it is feasible to use
ATPase activity as a quick nonspecific index, which is very helpful in
toxicity evaluation of chemicals. Although I50 derived from an in vitro test
cannot tell all the characteristics of chemicals, it does provide preliminary
judgment for the toxic properties of chemicals, When ATPase activity of fish
exposed to toxicants of low dosage is measured, sublethal effects can be
detected, which provides another basis for understanding the toxic properties
of chemicals.
AChE has been the earliest enzyme adopted in environmental monitoring
(Weiss 1989). In the late 1940s, the action pattern of organophosphorus
pesticide was found to provide specific inhibition on AChE. Since then, AchE
has been used as an index showing the presence of organophosphorus pesticide.
Compared with ATPase, AChE is a relatively specific indicator. It can only
reflect the action of neurotoxic materials. A specific indicator like AChE is
very important in toxicological research and environmental monitoring.
Certain biochemical indices represent particular functions. For example,
the change of ATPase shows the status of the energy metabolism system in the
body, AChE reflects the nerve impulse conduction of organism. Various
physical function are reflected from specific indices. Life is maintained by
various activities. Any process may have ill or even lethal effects on an
organism. Therefore, one or several indices would not be enough to evaluate
the whole effect of a chemical on an organism. Death and growth retardation,
of course, can explain in some way the toxicity of a chemical, but no com-
prehensive evaluation can be made from them. Similarly, the change of ATPase
333
-------
100
Concentration of Hg , ppm
Figure 2. Inhibition of Hg2+ on grass carp kidney
NaKMg-ATPase, NaK-ATPase, and Mg-ATPase.
and AChE cannot reflect the harmfulness of a chemical in a comprehensive way.
Consequently, a series of indices that could represent different activities of
an organism should be obtained for the comprehensive evaluation of toxicity of
chemicals. After establishing the indices, chemicals could be classified
according to the extent of their effects on different indices, e.g., chemicals
having greater effects on ATPase could be classified as "ATPase-toxin"
toxicants, those having greater effects on AChE, "AChE-toxic" toxicant, etc.
Then, not only toxic properties will be understood, but also a certain
knowledge of the action pattern can be gained. Comprehensive biochemical
indices also can provide parameters for water quality criteria, as well as the
basis and method for rapid screening of chemicals.
SUMMARY
Several kinds of toxicants had effects on grass carp ATPase. Plotting of
probit of inhibition percentage and logarithm of toxicant concentration
presents a linear relationship. I50 obtained from the plotting is a specific
index of toxicant. Combined with other indices, I50 can be used in a screen-
ing test of chemicals. Cu2+ had no effect on paradise fish brain AChE. AChE
is a relatively specific indicator, which can be used as an indicator monitor-?
ing pollution of neurotoxic toxicants.
334
-------
ACKNOWLEDGMENT
We are especially grateful to Dr. Jiankang Liu for his directions. The
technical assistance of Mrs. Renzheng Zhou who helped to carry out the test of
Cd2"1" is gratefully acknowledged.
REFERENCES
Boese, B.L., V.G, Johnson, D.E. Cheopman, andJ.W. Ridlington. 1982. Effects
of petroleum refinery wastewater exposure on gill ATPase and selected
blood parameters in the Pacific staghorn sculpin (Leptocattus armatus).
Comparative Biochemistry and Physiology. 71C:63-67.
Davis, P.W., J.M. Friedhoff, and G.A. Wedemeyer. 1972. Organochlorinated
insecticide, herbicide and polychlorinated biphenyls (PCB) inhibition of
NaK-ATPase in rainbow trout. Bulletin of Environmental Contamination and
Toxicology. 8:69-72. ,
Haya, K. and B.A. Waiwood. 1983. Adenylate energy change and ATPase
activity: potential biochemical indicators of sublethal effects caused
by pollutants in aquatic animals. Advances in Environmental Sciences and
Technology. 13:307-333.
Hestrin, S. 1949. The reaction of acetylcholine and other carboxylic acid
derivatives with hydroxylamine and its analytical application. Journal
of Biological Chemistry. 180:249-261.
Jampol, L.M. and F.H. Epstein. 1970. Sodium potassium-activated adenosine
triphosphatase and osmotic regulation by fishes. American Journal of
Physiology. 218:607-611.
Kuhnert, D.M. and R.B. Kuhnert. 1976. The effect of in vivo chromium
exposure on Na/K- and Mg-ATPase activity in several tissues of the
rainbow trout (Salmo gairdneri). Bulletin of Environmental Contamination
and Toxicology. 15:383-390.
Paxton, B. and B.L. Limminger. 1983. Altered activities of branchial and
renal Na/K- and Mg-ATPase in cold-acclimated goldfish (Carassius
auratus). Comparative Biochemistry and Physiology 74B:503-506.
Riadel, B. and G. Christensen. 1979. Effects of selected water toxicants and
chemical upon adenosine triphosphatase activity in vitro. Bulletin of
Environmental Contamination and Toxicology. 23:365-368.
Sastry, K.V. and K. Sharma. 1980. Effects of mercuric chloride on the
activities of brain enzymes in a freshwater teleost, Ophiocephalus
(channa) punctatus. Archives of Environmental Contamination and
Toxicology. 9:425-430.
Scope, R.K. 1982. Protein purification: principles and practice. Springer-
Verlag New York, NY. pp. 265-266.
335
-------
Srivastava, S.P., P.K. Seth, and D.K. Agarwal. 1975. Biochemical effects of
di-2-ethylhexyl phthalate. Environmental Physiology and Biochemistry.
5:178-183.
Verma, S.R. , A.K. Tyagi, N. Pal and R.C. Dalela. 1979a. In vivo effect of
the syndets Idet 5L and Swanic 6L on ATPase activity in the teleost,
Channa punctatus. Archives of Environmental Contamination and
Toxicology. 8:241-246.
Verma, S.R. , P. Mohan, A.K. Tyagi, and R.C. Dalela. 1979b. In vivo response
of ATPase in few tissues of the fish Mystus vittatus (Ham.) to the
synthetic detergent Swacofix E. (ABS). Bulletin of Environmental
Contamination and Toxicology. 22:327-331.
Weiss, C.M. 1959. Response of fish to sublethal exposure of organic phos-
phorus insecticides. Sewage and Industrial Wastes. 31:580-593.
Yap, H.H., D. Desaiah, and L.K. Cutkomp. 1971. Sensitivity of fish ATPase to
polychlorinated biphenyls. Nature. 233:61-62.
336
-------
RIVER NETWORK WATER QUALITY MODELING USING THE
ENHANCED STREAM WATER QUALITY MODEL QUAL2E
by
Qian Song1
INTRODUCTION TO QUAL2E
QUAL2E is a comprehensive and versatile stream water quality model. It
can simulate up to 13 water quality constituents in any combination desired by
the user. Constituents that can be simulated are:
1. Dissolved oxygen
2. Biochemical oxygen demand
3. Temperature
4. Algae as chlorophyll a
5. Organic nitrogen as N
6. Ammonia as N
7 . Nitrite as N
8. Nitrate as N
9. Organic phosphorus as P
10. Dissolved phosphorus as P
11. Coliforms
12. Arbitrary nonconservative constituents
13. Three conservative constituents
The model is applicable to dendritic streams that are well mixed. It
assumes that the major transport mechanisms, advection and dispersion, are
significant only along the main direction of flow (longitudinal axis of the
stream or canal). It allows for multiple waste discharges, withdrawals,
tributary flows, and incremental inflow and outflow. It also has the
capability to compute required dilution flows for augmentation to meet any
prespecified dissolved oxygen level.
The computer program permits simulation of any branching, one-dimen-
sional stream. The first step in modeling a system is to subdivide the stream
system into reaches, which are stretches of a stream that have uniform
hydraulic characteristics. Each reach is then divided into computational
elements of equal length. Thus, all reaches must consist of an integer number
of computational elements.
•"•Department of Environmental Sciences, Nanjing University, Nanjing, PRC
337
-------
There are seven different types of computational elements in the
original model: (1) headwater element, (2) standard element, (3) element just
upstream from a junction, (4) junction element, (5) last element in system,
(6) input element, and (7) withdrawal element.
By using QUAL2E, the stream can be conceptualized as a string of
completely mixed reactors (computational elements) that are linked
sequentially to one another via the mechanisms of transport and dispersion.
The basic equation solved by QUAL2E is the one-dimensional advection-
dispersion mass transport equation, which is numerically integrated over space
and time for each water quality constituent. This equation includes the
effects of advection, dispersion, dilution, constituent reactions and
interactions, and sources and sinks. For any constituent, C, this equation
can be written as:
aj
8t
ac
3x
dx _
dx
+ s
(1)
where
M
x
t
C
u
— mass
— distance (L)
- time (T)
«- concentration (ML"3)
— cross-sectional area (L )
- dispersion coefficient (L2T"X)
= mean velocity (LT"1)
- external sources or sinks (MT"1)
The constituent reactions ^nd interreactions are linked by dissolved
oxygen concentration. Figure 1 \llustrates the conceptualization of these
interactions. The arrows on the figure indicate the direction of normal
system progression in a moderately polluted environment; the directions may be
reversed in some circumstances for some constituents.
The mathematical relationships that describe the individual reactions
and interactions are presented by Brown and Barnwell (1985).
The classical implicit backward difference method is used to solve
Equation 1. The finite difference scheme is formulated by considering the
constituent concentration, C, at four points in the mnemonic scheme, as shown
in Figure 2. Three points are required at time n+1 to approximate the spatial
derivatives. The temporal derivative is approximated at distance step i.
After two steps differentiations, we obtain the equation:
in-t-l
(2)
338
-------
Atmospheric
Reaeration
/fr/w
Figure 1. Major constituent interactions in QUAL2E.
where a, = -[(AD,),,
i L \ L/i-1
ViAx,
= 1.0 + [(ADL)i
At
LAv,
The values of a±, bj., cit and z± are all known at time n, and the cirt'ii terms
are the unknowns at time step n+1. An efficient method for solving these
simultaneous linear equations is presented by Brown and Barnwell (1985).
IMPROVEMENT OF QUAL2E
One of the limits in using QUAL2E is -that the river system to be
simulated can only have one last" element. That means that the river system
339
-------
cannot have branching elements or "fork" elements, from which simulated
tributaries flow out from a main stem. And on the other hand, to improve the
water quality, we need to know how much pollutant the river system could
receive while the water quality of the river system still meets the required
level, so that we can know at least how much pollutant should be removed from
the waste discharge.
To model river system with fork elements, we first need to know the
hydraulics of both the main stem and the tributary. We assume that the
hydraulics of the river network are known. (For the fork element, the ratio
of the flow of the main stem and the flow of the tributary, FLWRT are given.)
is:
We know that, for a standard element i, the basic differential equation
(3)
in the case of a fork element, the basic equation becomes
where:
- Ci*[l-FLWRT(I)]
Two methods can be used to solve these equations . Because GJ, and f j are
several orders smaller then ^ and ~bit the first method is to assume that Ci+1
equals Cj , and divide Equation 4 into two equations :
DOWNSTREAM
UPSTREAM
element I + 1
element i
1+1
1-1
N + 1: t + At/2
N: t - At/2
I
Figure 2. Classical implicit nodal scheme.
340
-------
and:
(5)
(6)
Now it is easy to use the method presented by Brown and Barnwell (1985) to
solve the simultaneous linear equations. When back-substituting the con-
centration from the last element, Equation 4 is used to solve the fork element
concentration. .
The second method does not use the foregoing assumption because, in many
cases, Ci+1 does not equal (L.
Mathematically speaking, simultaneous linear equations are always
solvable, but their efficient solution is not easy. In this paper, an
efficient method for solving those equations is presented.
The method presented by Brown and Barnwell (1985) is applied before the
fork element(s), and when a fork element is met in the process of forward
substitution, we apply a sub-loop of forward and backward substitution which
is shown as follows
Ci-l + Wi-iCi - Gi.j.
aiCi-i+DiCi+Ci ' Ct+1+f jCj-Zi
substituting Equation 7 into Equation 8, we obtain:
ci + WtC1+1 = Gi - FiCj
where: W = G
F =
(7)
(8)"
(9)
- aiGi.1)/(bi -
and substituting Equation 9 to the next basic equation and forward substitut-
ing the obtained equation, we have:
where: F± =
k kk+1 = k
^-a^.J , Fk = a^F^
k = i+1, i+2, . . ., n-1, n
n = the order of the last element of the reach.
(10)
when k = n, k+1 = kk, and kk is the element order of the next type 3 or type 5
element, Equation 10 becomes: '
341
-------
= Gn - FnCd
•The forward substitution stopped when the next type 3 or type 5 element
is met. Then back substitution is started from the element. The back
substitute equations are shown as follows
A i>n wr (n)
Gk - Ak - BfcCj - n^i^^
where: Ak - Gk - Ak+1Wk, Bk = Fk - Bk+1Wk, Hk = -WkHk+1
^ - Gn> Bn = Fn, ^ = ¥n
k - n-1, n-2, .... i+1, i
when the back substitution stopped at the element i+1, we have:
R r H G (12)
Ci-H ~ Ai+l " "i+l^j " i+1 *&•
Substituting Equations 12 and 11 into the equations for the tributary, we
have:
(13)
where: W'k - ck/MM, G'k
MM - Bi^-a,
G'j - (Zj-
k - j+1, j+2 kk-2, kk-1
When k - kk-1, we have:
, TT» r* /-<> v
t-i + « ^-1^ - t. ujj-i S ]&-
, F'k - -akF'k_i-
-ajF'if
c <14)
where: EJ^.J, - G'^-i, L^-i = F'j^-i + Wj^-i
The back substitution of Equation 14 obtained the relationship of Cj and Ckk:
C - E- - L-Ckk (15)
where: ^ = G'k - W'kEk+1, 1^. = F'k - W'kLk+i
k - kk-2, kk-3, ..., j-1, j
Substituting Equation 15 to Equation 10, we have:
- G
(16)
where: W'B = WB-FnLd, G'n - Gn-FnEd
342
-------
and rewriting Equation 14:
Now both the main stream and the tributary down stream from the fork element
are ready to go back to the main loop of the forward and backward substitution.
By adding another two kinds of computational element --type 9 (fork) and
type 8 (element just downstream from a fork on the tributary )-- the QUAL2E
source code is modified so that it could simulate the net -like river systems.
To calculate the maximum point source loads , we first need to linearize
Equations 2 . The process of linearization is shown as follows
(1). Assume that in one reach, the concentration of the constituent at
the last element is equal to the concentration of the first element of the next
reach:
cn = c
n+i
(17)
(2). Let
and
where:
then we have:
where:
Vl - -(ba+
An-2 = (Zn-l-bn-lAn-l)/an-l
Bn-2 - -(bn-lVl+Cn-lVan-l
Ai = (Zi+i-bi+iAi+i-ci+1Ai+2)/ai+1
Bi = -(bi+iBi+i+ci+iBi+2)/ai+i
n = the number of element of the reach;
i -. 1, 2, 3, ..., n-3
ci = the concentration of the constituent at the first element
of the reach
Cn = the concentration of the constituent at the nth element
of the reach.
Rewrite Equation 18 as :
cn = Ai +
(19)
where:
At --
/%! for reach i
! for reach i
343
-------
After the linearization, we put the first basic equation of every reach
together as a set of simultaneous equations. Substituting Equation 19 into
these simultaneous equations, we have:
where:
i,i
Ci.2
the concentration of the last element of reach i-1
the concentration of the first element of reach i
the concentration of the second element of reach i
Because G£ is several orders smaller, and the b± and ai( and Citi and C1(2 are
not of very much difference, we assume that Cij2 = Cij:L. Substituting Equation
19 for Equation 20, we have:
and
(21)
or
where :
AA£ —
, i = 2,3,...,n
BB± = bi + cif i = 1,2, . . . ,n
ZZi = Z£ - a^i-i, i = l,2,...,n
Rewriting Equation 21 in matrix form,
AB * C - Z + Z
(22)
where:
AB -
BI
A/
(
(
r
*l
^2 BB2
) AA3 BB3
) 0 AA4
: 0 -
i n n_--. ----------
C = (^iiii C2>ii - - - ) Cn,
Zx - (0, -az&!, -a3A2, - - -
Z = (Zj_, Z2, - - - , Zjj)
Under steady-state conditions,
344
-------
where:
Si
P..
external sources or sinks
internal sources or sinks
Let P = (Pl, p2, ..., Pn)*, Z
+ P, S - (Slf.S2, ..., Sn)T and
v =
V.,
Substituting Z, Z
equation, we have:
lf
and V into Equation 22 and rearranging the
V*AB*C-V*Z,
(23)
When substituting the river water quality criteria C* to Equation 23,
we have the maximum point load allowed in the river network:
S* = V * AB * C* - V * Z2 (24)
The first step in computing the maximum point source load is to
redefine the reaches of the river system. The input element is the first
element of the reach.
Using Equation 19 to Equation 24, an additional subroutine was coded
for QUAL2E, which was called MAXCNT. Using the improved model, a case study
is performed.
Figure 3 shows the stream network of computational elements and
reaches. The river system contains one fork element. The input data set for
the simulation is a modification from one of the demonstration input data
sets shown at the US EPA's QUAL2E workshop at Nanjing University in 1985.
Figure 4 shows the result of BOD simulation. Figure 5 shows the river system
and the result of the BOD simulation of the above-mentioned workshop demon-
stration. Table 1 shows the BOD point source loads (P.L.) and the simulation
result (concentration) at the first element of the reach. When substituting
the concentrations in Table 1 as C*, the maximum point source loads (MPL) of
the river network should be the point source loads in Table 1. Table 2 shows
the result.
REFERENCE
Brown, L. C. and T. 0. Barnwell. 1985. Computer Program Documentation
for the Enhanced Water Quality Model QUAL2E. U.S. Environmental
Protection Agency, Athens, GA. EPA/600/3-85/065.
345
-------
Reach
Number
Junction
1
Figure 3. Stream network of computational elements
and reaches.
TABLE 1. POINT SOURCE LOADS
Reach
1 2
345 6789 10
P.L., kg/day 0 1865 0
0000
BOD, mg/L 2.19 21.58 1.97 18.78 5.26 0.69 0.71 9.79 9.31 9.14
346
-------
TABLE 2. RESULTS
Reach
12 3 45 6 7 89 10
CA, mg/L 2.2 21.6 2.0 18.8 5.3 0.7 0.7 9.8 9.3 9.1
MPL kg/day 0 1870 000 0 0 00.0
BOD (mg/L)
21.58
9.79
Figure 4. BOD simulation result.
347
-------
BOD (mg/L)
21.58
8.58
Figure 5. River system and BOD simulation results from
EPA's QUAL2E workshop, Nanjing, 1985.
348
U.S. GOVERNMENT PRINTING OFFICE: 1990— 748- 159/ 00436
-------
-------
Q) O
^»
""" !2.
5- CD
CD
C
CD
<»
m
-o
o
10
vo
o
co.::
•a m _.
Z.3-5. 5J =•»
a o ° i. 5 3
Is2 §|i
III p?l
I ^s l§
I 8 S s S
a 2 co
= - S n
^ -»
JZJ M 3" 3
2 CD <0 CO
C -t Q) JO
3| &^
03 S =
5" S 5
m Q _. o
Q] X 3 Q)
&o 3?
> me
"S 3-1-
3 s
8
o
^
o'
en o
NJ 2
o> 3
00 CD
CD
cn
CD
Q>
-------