United States
Environmental Protection
Agency
Office of
Exploratory Research
Washington DC 20460
fcPA 600/9 HI' 00,
April 1982
Research and Development
Environmental
Biology
State-of-the-Art
Seminar
-------
EPA-600/9-82-007
April 1982
ENVIRONMENTAL BIOLOGY
STATE-OF-THE-ART SEMINAR
Patricia A. Archibald, Editor
1981
Organized by Environmental Protection Agency,
Office of Research Grants and Centers
Held at College of Saint Scholastica,
Duluth, Minnesota, July 20-22, 1981
Published by Office of Exploratory Research,
Environmental Protection Agency,
Washington, D.C.
-------
Preface
By Direction of Congress, the Office of Research and Development,
Environmental Protection Agency, implemented a new process for solicitation
and evaluation of research grants in late 1979. An Office of Research
Grants and Centers (ORGC) began operation in January, 1980. Research
grants funded under the auspices of ORQC are not necessarily representative
of the kinds of projects which the Agency funds as cooperative agreements
and contracts through other programs at Headquarters and the regional
research laboratories.
In April, 1980, the first annual generalized solicitation was
issued. Initially, proposals were requested in four major areas, e.g.,
Environmental Engineering and Pollution Control Processes; Health Research;
Environmental Chemistry and Physics; and Environmental Biology. Usually,
eligibility is limited to universities and other non-profit institutions.
The program is extremely competitive with scientific merit of every
proposal being evaluated by an extramural Peer Review Panel comprised of
members nominated from the academic community, other agencies within the
Federal government, and industry. A chairperson elected from the academic
community administers the Peer Review Process. Only those proposals
recommended for consideration for funding by the Peer Review Panel are
considered further by ORGC. The final funding decision for each competitive
research grant is made within the Agency relative to scientific merit,
relevance to current Agency programs and availability of monies.
Dr. Freida Taub, of the University of Washington-Seattle, served' as
the first chairperson of the Environmental Biology Peer Review Panel.
Her skills in organization and appreciation of detail have been important
assets in the inauguration of the competitive grants problem in the area
of Environmental Biology. Procedures developed by Dr. Taub will continue
to be employed by subsequent chairpersons on both the Environmental
Biology Panel and panels in the other fields of study.
Each Principal Investigator is required to present a seminar detailing
his research. Al work presented at this seminar"for Environmental
Biology is by Principal Investigators who had their original proposals
reviewed during Dr. Taub's term of office. The discussant for each
paper, in most incidences, was one of the initial reviewers of the
proposal as first submitted.
The Agency wishes to compliment both Dr. Taub and the Principal
Investigators for the quality and depth of work presented.
-------
STATE-OF-THE-ART SEMINAR
Environmental Biology
College of St. Scholastica
Duluth, Minnesota
July 20 - 22, 1981
Page
Monday - July 20
12:00 Noon to 5:00 P.
6:30'- 7:00 P.M.
7:00 - 8:00 P.M.
8:00 - 9:00 P.M.
Tuesday - July 21
8:30 - 9:00 A.M.
M. Registration and Tour of EPA Lab, Duluth
Reception - College of St. Scholastica
Buffet Dinner - College of St. Scholastica
Speaker: Dennis Tirpak, Director
Office of Exploratory Research
U.S. Environmental Protection Agency
Washington, D.C.
Welcoming Remarks: Norbert Jaworski
Director
EPA Laboraory, Duluth
SCIENTIFIC SESSION #1: Frieda B. Taub, Chairperson
MARINE ECOSYSTEMS
9:00 - 9:45 A.M.
9:45 - 10:30 A.M.
10:30 - 10:45 A.M.
"The Sensitivity of Critical Life Stages of
Benthic Meiofauna to Drilling Muds"
John J. Lee, CUNY
Discussion Leader: Kenneth E. Biesinger
"Indicators of Current Growth Rate as Rapid
Method for Toxicity Tests with Fish Larvae"
Ira A. Adelnan, Univ. of Minnesota, St. Paul
Discussion Leader: Guenther Stotzky
Coffee Break
13
-------
State-of-the-Art Seminar Page
Environmental Biology
10:45 - 11:30 A.M. "Growth Kinetics of Aeromonas hydrophila in 32
Fresh Waters and Their Relationship to
Trophic State"
Victor J. Cabelli, Univ. of Rhode Island
Discussion Leader: Guenther Stotzky
11:30 - 12:15 P.M. "Drilling Fluid Induced Changes in Neuronal 45
Activity Monitored in an Intact Behaving
Marine Animal"
James E. Kanz, Texas A&M, Galveston
Discussion Leader: Kenneth E. Biesinger
12:15 - 1:30 P.M. Lunch
SCIENTIFIC SESSION #2: Frieda B. Taub, Chairperson
MODELING
1:30 - 2:15 P.M. "Qualitative Analysis in the Setting of East 61
Coast Marine Benthos"
Charles J. Puccia, Harvard University
Discussion Leader: Jack Waide
2:15 - 3:00 P.M. "Application of Loop Analysis Techniques in 94
Nova Scotian Ecosystems"
Patricia A. Lane, Harvard University
Discussion Leader: Jack B. Waide
3:00 - 5; 30 _P_*M.. New Approaches to Measuring the Biological 123
Impact of Environmental Toxicants
Hinsdill Group, Univ. of Wisconsin
3:00 - 3:45 P.M. "Impact of Toxicants, Disease and Climate on 126
Growth and Reproduction Using Peromyscus
maniculata"
Warren P. Porter, Univ. of Wisconsin
Discussion Leader: Nancy L. Stanton
3:45 - 4:00 P.M. Coffee Break
•4:00 - 4:45 P.M. "Immunotoxicity in Peromyscus maniculata" 150
Ronald D. Hinsdill, Univ. of Wisconsin
Discussion Leader: Maurice G. Zeeman
-------
State-of-the-Art Seminar p
Environmental Biology a9e
4:45 - 5:30 P.M. "Host-Pathogen-Toxicant Interactions: A 169
Progress Report"
Thomas M. Yuill, Univ. of Wisconsin
Discussion Leader: Richard M. Kocan
Wednesday, July 22
SCIENTIFIC SESSION #3: Frieda B. Taub,- Chairperson
TERRESTRIAL AND FRESIWATER ECOSYSTEMS
8:30 - 9:15 A.M. "Effects of Milorganite on Two Grassland 182
Ecosystems"
Gary W. Barrett, Miami University
Discussion Leader: Bernard P. Sagik
9:15 - 10:00 A.M. "The Assessment of Toxic Effects in an 205
Environmental Stream Ecosystem"
Robert H. Boling, Michigan State University
Discussion Leader: Robert J. Beyers
10:00 - 10:45 A.M. "BDlyphosphate in Natural Phytoplankton . 222
Assenblages and its Formation and Degradation
in Cultured Algal Cells"
Linda Sicko-Goad, Univ. of Michigan
Discussion Leader: Joe M. King
10:45 - 11:00 A.M. Coffee Break
11:00 - 11:45 A.M. "Oil and Related Contaminant Effects on 240
Water Fowl Immune Defenses"
Thomas M. Yuill, Univ. of Wisconsin
Discussion Leader: Holly L. Collins
11:45 - 12:00 Noon Seminar Summary and Closing
Frieda Taub
Norbert Jaworski
-------
THE SENSITIVITY OF CRITICAL LIFE STAGES OF BENTHIC
MEIOFAUNA TO DRILLING MUDS
John J. Lee, Marie E. McEnery,and Judith R. Garrison
. Department of Biology
The City College of C.U.N.Y.
Convent Avenue at 138 Street
New York, N. Y. 10031
ABSTRACT
Static bioassays using benthic meiofauna (Chromadorina germanica
[neraatode], Nitocra typica [harpacticoid copepod], Allogrotnia laticollaris
[foraminifera], and Euplotes vannus [ciliate]), as sensitivity indicators
were run testing drilling muds from Mobile (Alabama) bay. Mud sample types
were raw, desilter, and desander/desilter. Drilling mud mixtures with sand
was an unfavorable substrate for several animals, restricting their body and
feeding movements and causing them to starve to death. When the mixtures
were embedded in agar to change their physical properties the animals
moved and fed normally. The results of this preliminary study suggest that
two of the animals tested, Chromadorina germanica and Allogromia laticollaris
have the potential to be developed into bioassay tools for marine benthic
mud s.
INTRODUCTION
Microfauna (protozoa > 50 urn) and meiofauna (nematodes, copepods,
foraminifera, and similar sized [0.05-0.5 ram] small organisms) are among
the most numerous group of marine animals and play key roles in benthic
food webs (reviewed in 1). They consume and regulate benthic microfloral
assemblages and they, in turn, serve as food for larger, more familiar
animals. Because they live in intimate contact with the substrate as both.
epifauna and infauna, it has been suggested that they are among the most
sensitive indicators of environmental quality (2, 3, 4). Nutritional and
physiological ecological studies suggest that various larval (e.g. nauplei)
stages are even more sensitive to unfavorable conditions than are adults (5),
Although there are indications that biocides such as pentachlorophenol
and wetting agents such as barite adversely affect meiofaunal assemblages
-------
(6, 7) little is known of the range of sensitivies of the group to such
agents which are constituents of drilling muds. Since many species of micro-
and meiofauna are easily cultivated in ordinary laboratory glassware under
common laboratory conditions (8, 9) they have great potential as tools for
the assay of benthic sediments which may have become impacted by various
types of discharges.
MATERIALS AND METHODS
The species of meiofauna used in this study were isolated from the
sublittoral zones of salt marsh tidal paines at either Towd Point on Peconic
Bay, Southampton, Long Island, or Greater Sippewissett Marsh on Buzzards
Bay, Falmouth, Mass. (8). The four species selected, Allogromia laticollaris
(foraminifera), Nitocra typica (harpacticoid copepod), Chromadorina german-
ica (nematode), Euplotes vannus (ciliated protozoan) have worldwide distri-
butions and are easily cultivated. Laboratory procedures for raising and
manipulating the experimental animals followed those previously published
(5, 9-15). Care was taken to make sure that inocula were uniform size or
age. Ten Allogromia ~ 500 urn in diameter; 10 gravid female Nitocra; 30
Chromadorina from 15 day cultures; 100 log phase Euplotes served as inocula
in experiments. The organisms were inoculated into experimental vessels
(35 cm disposable tissue culture flasks and 60 mm petri plates or 20 x 125
ram test tubes) along with excess algal food (1 x 10° cells/ml). The effects
of the drilling muds on the algae used as food was tested separately. On
the basis of a preliminary toxicity screen the following algae were selected
from our collection as food for the animals: Navicula multica. Nitzschia
ovalis. Fragilaria shiloi. and Cvlindrotheca fusiformis. Incubation was at
25° C in an illuminated incubator (16 hr L/8 hr D cycle) (Scherer Gel 4-4).
The drilling muds from a Mobile, Ala. bay well were shipped frozen from
the EPA Environmental'' Research Laboratory, Gulf Breeze, Florida. They were
refrigerated (4° C) upon arrival and maintained at the same temperature for
the duration of the study. Collection data, depth of sample, density, vis-
cosity, pH, calcium content, water content, and diesel oil content were sent
with the samples. A copy of the metal analysis and methodology used for the
analyses (Al, Pb, Cr, Ba, Fe and Zn) were also transmitted to us.
Various muds were diluted and mixed very thoroughly with washed silica
beach sand from the National Wildlife Refuge on Peconic Bay, Southampton,
L.I., N.Y. (grain characteristics see 16). Sand was also the substrate for
the controls. In the initial experiments with the 3 raeiofaunal species, sand
was diluted with drilling mud volumetrically 1:1, 2:1 and 4:1. Peconic Bay
seawater was used as an overlay (15 ml in the 35 cm flasks, 10 ml in the
60 x 15 mm petri dishes). Later tests with the algae were done at higher
dilutions (to 1 x 10" ). The effects of drilling muds on Euplotes vannus
were studied in test tubes with the muds diluted with sand over the range of
10"2 to 10"5.
-------
In an attempt to separate the effects of surfactants in the muds from
other potentially toxic materials drilling muds were mixed (v/v) with steril-
ized agar (10 to 10 ) to create a solid suspension. As in the previous
experiment the substrate was overlain' with sea water.
The duration of the experiments varied with the organisms. A. laticol-
laris and Chromadorina germanica were observed weekly for 30 days. Experi-
ments with Nitocra typica were observed weekly and terminated after 7 weeks.
Euplotes vannus was observed after 7 and 14 days. The experiments were
terminated by adding 10 ml of Rose Bengal/formalin to the test vessels. Data
were collected separately from the liquid overlay and from the agar fraction.
With the higher concentrations of muds the agar had to be broken up and re-
suspended to enable enumeration of the Chromadorina germanica. In lower
concentrations quadrants were read and statistically compared.
All experiments were in quadruplicate. No additional food was added
during the experiment. Evaporation of media was kept to a minimum with dis-
tilled water in the bottom of the large (150 mm) Pyrex dishes which held the
smaller, 60 mm petri plates.
RESULTS
The initial dilutions of 1:1, 2:1 and 4:1 were too turbid to permit
growth of the algae tested. Bacterial growth also contributed to the
density. Concentrations above 1:100 permitted reproduction of Navicula
multica. Nitzschia ovalis. Fragilaria shiloi. and Cylindrotheca fusiformis.
Growth of the latter two species increased four-fold as the muds were
successively diluted (1:10" to 1:10"°) to the highest level tested; the
growth of N. ovalis trebled over the same range and that of N. multica was
unaffected. Growth of the algae in the 10" dilution was equal to that of
the controls.
Of all the organisms tested, Allogromia laticollaris was the most
sensitive to drilling muds. Allogromia laticollaris and Chromadorina ger-
manica did not survive in the 1:1 or 2:1 dilutions of any of the mud samples
tested; some were still viable after 14 days in some 4:1 dilutions. Speci-
mens of both of these species were behaving abnormally in the 4:1 dilutions.
However, many seemed to recover when transferred to fresh control media.
Microscopic observations suggested that A. laticollaris were starving to
death. Their color was pale and they did not extend- any pseudopods. Some
of the "near dead: had a dark brown, wrinkled appearance. Microscopic ob-
servations of Chromadorina gave the impression that the physical properties
of the media were responsible for their toxicity. The movement of the ani-
mals was abnormal, they pulsed slowly instead of rapidly moving when agitated,
Some viscous material adhered to many worms and their mouths were clogged
with sediment particles. They were probably attempting to feed on dead
algal material.
-------
TABLE 1. THE EFFECT OF DRILLING MUDS DILUTED 1:10 ON THE VIABILITY AND REPRODUCTION OF
MEIOFAUNA AND A CILIATED PROTOZOAN
fr
REPRODUCTION VIABILITY
SHAKER RAW RAW SHAKER
1-2 2-3 2-7 2-11
Chromadorina germanica 44 4 4 44
Euplotes vannus 4 4 4 44
Allogromia laticollaris 4 0 4 4
Nitocra typica 44 44 44 44
Chromadorina germanica 444 444 INI INI
Euplotes vannus 444 444 444 444
Allogromia laticollaris 4044
Nitocra typica 4444 4444 4444 4444
DESA/SI RAW DESILT
2-20 3-2 3-16
4 4
4 4
0 4
44 44
444 444
444 444
0 4
111! 1 I II
IIII I I If
44
4
4
4
444
444
4
4444
SHAKER
5-1
4
4
0
4
444
444
0
4444
0 dead
4 variable viability
44 viability (100%)
444 variable reproduction
I I I I reproductive success (several generations)
-------
TABLE 2. THE EFFECT OF VARIOUS DRILLING MUD FRACTIONS ON THE FECUNDITY
OF AN HARPACTICOID COPEPOD, NITOCRA TYPICA
CONTROL
RAW 2-7
S.D.
SHAKER 2-11
S.D.
DESA/DESI 2-20
S.D.
DESILTER 3-16
S.D.
RAW 3-2
S.D.
SHAKER 5-1
S.D.
308
20.3
350
14.6
381
12.1
396
6.9
323
2.1
356
9.1
1:1
401
48.6
398
31.4
361
5.6
388
8.1
431
4.?
298
4.6
2:1
390
40.3
360
28.6
360
8.9
388
5.3
351
2.1
261
2.1
4:1
386
36.6
358
20.6
319
10.3
369
6.3
352
1.6
270
3.4
10:1
390
21.2
378
23.6
366
6.1
359
7.3
361
4.5
235
2.9
20:1
389
3^.?
388
16.4
371
7.3
386
4.1
370
3.6
233
3.0
50:1
363
16.1
374
11.3
369
6.8
358
3.2
381
4.2
248
2.6
100:1
354
21.3
351
7.6
363
6.6
387
5.3
372
4.7
225
2.8
Incubation 7 weeks, average of 4 replicates
S . D . = «tondnrd
Heviotion
TABLE 3. THE EFFECT OF VARIOUS DRILLING MUD FRACTIONS ON THE FECUNDITY
OF A MARINE BENTHIC CILIATE, EUPLOTES VANNUS
CONTROL
RAW 2-7
S.D.
SHAKER 2-11
S.D.
DESA/DESI 2-20
S.D.
DESILTER 3-16
S.D.
RAW 3-2
S.D.
SHAKER 5-1
S.D.
Incubation 7 da1
187
14.3
78
18.6
57
19.2
75
13.9
89
16.4
74
18.4
ys, overo<
1:10
167
20.6
49
13.6
84
12.4
120
10.3
171
13.3
139
12.1
ge of 4 rep
•» /,
3 1:10
250
10.3
83
14.9
68
12.6
58
12.1
119
16.5
44
9.3
>licote< : S.C
1:10
99
8.6
55
15.3
66
13.1
54
11.6
104
14.3
42
8.6
> . = *tondord
1:10
175
9.3
49 •
12.1
48
10.2
36
12.4
38
19.6
31
II. 1
deviation
-------
TABLE 4. THE EFFECTS OF VARIOUS DRILLING MUD FRACTIONS ON THE FECUNDITY
OF A MARINE BENTHIC NEMATODE, CHROMADORINA GERMANICA
DILUTED MUD SOLIDIFIED WITH 1.5% AGAR
CONTROL 1 : 1
SHAKER 1-2
S.D.
RAW 2-7
S.D.
SHAKER 2-11
S.D.
500
5.5
480
4.8
560
19 3
DESA/DESI 2- 20525'
S.D.
DESILTER 3-16
S.D.
RAW 3-2
S.D.
SHAKER 5-1
S.D.
Incubated 30
* Deod
4.6
601
3.3
581
5.8
567
4.7
d«ys
75
4.3
162
3.4
2168
12.4
30*
40
1.7
28
3.2
39
4.8
;average of
S.D.=
2:1
60
3.1
158
4.5
2496
27.9
30*
40
5.8
29
3.8
45
5.6
4:1
70
2.8
300
3.0
2032
10.8
30*
36
50°
7.1
46
6.0
4 replicate
10:1
150
2.9
404
4.5
1759
50.4
30*
38
1&5
13.8
28
3.2
20:1
150
3.3
390
2.6
1800
61.4
30
. o.o
36
48!09
29.8
26
6.6
50:1
140
3.6
388
3.8
1740
61.6
100
1.5
40
ttl'
17.7
40
5.5
100
150
:1
3.8
400
8.
5
1896
II.
150
2.
180
&
8.
200
6.
9
5
2
4
8
cultures
"tondord deviation
TABLE 5. THE EFFECTS OF VARIOUS DRILLING MUD FRACTIONS ON THE FECUNDITY
OF A FORAMINIFERAN, ALLOGROMIA LATICOLLARIS
RAW 2-7
SHAKER 2-11
DESA/DESI 2-20
DESILTER 3-16
RAW 3-2
SHAKER 5-1
CONTROL
300
280
320
319
319
301
-3
1:10 "
40%s
10
dead
1007.8
90%s
90%s
-L
1:10
40%s
10
f'ead
100%s
80%s
80%s
-5
1:10
40%s
175
dead
1007.8
89%s
907.S
(•.
1:10
407.6
350
dead
100%s
90%s
80%s "*"
s • survival
+ = cytotomy and binary fission noted
Inoculum - 10 organisms; experiment terminated after 30 days
-------
On the other hand, Nitocra typica were viable and reproduced in most
muds (Tables 1 and 2). The animals gathered the algae in the inoculum into
balls above the surfaces of the mud and spent most of their time in the
balls or swimming between algal masses. Nauplei and copepodites also spent
most all their time in the superficial algal masses. None were observed as
infauna in the substrate muds. Euplotes vannus usually survived in dilu-
tions of mud 1:10 or greater (Table 3). Sorae reproduction was noted at that
dilution. The reproduction rate reached that of controls at dilutions above
1:10 .
When some characteristics of the drilling mud fractions were altered by
solidification in agar Chromadorina germanica survived at all the dilutions
tested and reproduced in many (Table 4). In several muds, raw 2-7, shaker
2-11, and raw 3-2, reproduction rates in dilutions of 20:1 or greater
approached those of controls. Reproduction exceeded controls in all dilu-
tions of one mud sample (2-11). Inhibition of reproduction to less than
half of controls was still noted in dilutions of 100:1 in 4 mud fractions
(raw 2-7, desand/desilt 2-20, desilter 3-16, shaker 5-1).
Most drilling mud fractions were toxic to Allogromia laticollaris even
when they were solidified with agar (Table 5). The animals reproduced only
in the shaker mud 2-11. Control levels were reached in a 1:10 dilution of
mud. The desand/desilted mud 2-20 was toxic at all concentrations tested.
There was little change in the survival rate of animals in the remaining mud
samples over a dilution range of 1:10 to 1:10 .
DISCUSSION
There was a wide range in the sensitivities of the 4 species of meio-
fauna to the drilling mud fractions which were tested. The implication from
this very limited study is that the drilling muds would greatly alter.the
meiofaunal assemblage structure but would not totally eliminate it. Those
animals able to live on the walls of the vessels (e.g. Euplotes vannus)
would most likely survive on aufwuchs and particles which settle on the sur-
faces of the mud. The growth of Euplotes in a'll the experimental and control
vessels was rather poor compared to its reproductive potential (13). This
animal is a bloom feeder. It does well on algal or bacterial food when the
densities exceed 10^ cells per ml. Below that level feeding rates are very
slow. Since the algae in the initial inoculum did not reproduce at high
concentrations of drilling mud fractions the animal quickly grazed the algae
down to below optimum feeding levels. This result is in sharp contrast to
some of our observations on New York barbor and estuary muds where bacterial
counts in some samples were quite high. The ciliates in those test vessels
quickly reached their carrying capacity in less than a week (Lee and Tietjen
work in progress). Workers in Europe (Persoone, personal communication) are
using this same ciliate for the assay of marine muds with heavy metallic
lodes.
-------
The survival of Nitocra was undoubtedly due to the large algal inoculum
which provided refuge for the animal. It is problematical whether this is
an artifact of test conditions or whether this and. similar animals could
survive under natural conditions since many harpacticoid copepods are highly
selective feeders (5, 17). The assay protocol which we used needs re-examin-
ation if this animal is to be employed in the future.
Chromadorina germanica could be used as an assay organism using either
of the 2 protocols tested. Previous studies have shown that molecules as
small as acetate are not absorbed through the cuticle of marine nematodes.
They enter only when the animal is feeding on small particles (15). If
there are soluble toxic materials in many of the muds they are not getting
into the animals. This would account for the survival noted in many tubes.
Larval and juvenile stages with thinner cuticles failed to survive. The
reproduction of the nematodes in the shaker 2-11 mud dilutions suggests
that the test conditions were favorable for both the food organisms and the
animals, otherwise the food would have been grazed down very rapidly.
Allogromia laticollaris was the most sensitive organism tested. It only
reproduced well in a 1 x 10 " dilution of shaker 2-11 muds. Survival was
noted at a 10"^ dilution. One hundred percent survival was also noted in all
the desilter 3-16 muds tested.
The progress to date on this project seems to indicate that both
Chromadorina germanica and Allogromia laticollaris could be used as bioassay
organisms for drilling muds. The results of this preliminary study suggests
that reproductive rate may be a better criterion to observe than survival
since the survivers are slowly starving to death. Much more time and effort
need to be expended to make a bioassay using these 2 animals practical, but
the potential is there.
8
-------
REFERENCES
1. Lee, J. J. 1980. A conceptual model of detrital decomposition and the
organisms associated with the process. Adv. Aquatic Microbiol.
Vol. 11:257-291.
2. Heinle, D. R. and Beaven, M. 1977. Effects of chlorine on the copepod
Acartia tonsa. Chesapeake Sci. 18:140.
3. Hoppenheit, M. and Sperling, K. 1977. On the dynamics of exploited
populations of Tisbe holothuriae (Copepoda, Harpacticoida). 'IV. The
; toxicity of cadmium: response to lethal exposure. Helgolander wiss.
Meeresunters. 29:328-336.
4. Roberts, M. H., Diaz, R., Bender, M. and Hugget, R. 1975. Acute
toxicity of chlorine to selected estuarine species. J. Fish. Res.
Bd. Canada 32:2525-2528.
5. Lee, J. J, Tietjen, J. H. and Garrison, J. R. 1976. Seasonal switching
in the nutritional requirements of Nitocra typica Boeck, an harpacti-
coid copepod from salt marsh aufwuchs communities. Trans. Amer.
Micros. Soc. 95:628-637.
6. Canteltno, F. R. and Rao, K. R. 1978. Effect of pentachlorophenol on
the meiobenthic nematodes in an experimental system. In: Penta-
chlorophenol: Chemistry, Pharmacology and Experimental Toxicity:
165-174. Environmental Science Research, Vol. 12, Plenum Press, New
York.
7. Cantelmo, F. R., Tagatz, P. and Rao, K. R. 1979. Effect of barite on
raeiofauna in a flow through experimental system. Mar. Environ. Res.
2:301-309.
8. Lee, J. J., Tietjen, J. H., Stone, R. J., Muller, W. A., Rullman, J. and
McEnery, M. 19708 The cultivation and physiological ecology of
members of salt marsh epiphytic communities. Helgolander wiss.
Meeresunters. 20:136-156.
9. Lee, J. J. and Muller, W. A. 1974. Culture of salt marsh organisms and
micrometazoa. In: Culture of Marine Invertebrate Animals. (W. L.
Smith and M. H. Chaney, eds). Plenum Press, New York. 87-107.
-------
10. Lee, J. J., Muller, W. A., Stone, R. J.. McEnery, M. E. and Zucker, W.
1969. Standing crop of foratninifera in sublittoral epiphytic communit-
ites of a Long Island salt marsh. J. Mar. Biol. 4:44-61.
11. Lee, J. J., McEnery, M. E., Kennedy, E. M. and Rubin, H. 1975. A
nutritional analysis of a sublittoral epiphytic diatom assemblage from
a Long Island salt marsh. J. Phycol. 11:14-49.
12. Lee, J. J., Tietjen, J. H., and Garrison, J. R. 1976. Seasonal switch-
ing in the nutritional requirements of Nitocra typica Boeck, an
harpacticoid copepod from salt marsh aufwuchs communities. Trans.
Amer. Micros. Soc. 95:628-637.
13. Muller, W. A. and Lee, J. J. 1977. Biological interactions and the
realized niche of Euplotes vannus from the salt marsh aufwuchs. J.
Protozool. 24:523-527.
14. Tietjen, J. H. and Lee, J. J. 1973. Life history and feeding habits of
the marine nematode Chromadora macrolaimoides Steiner. Oecologica:
12:303-314.
15. Tietjen, J. H. and Lee, J. J. 1975. Axenic culture and uptake of
dissolved organic substances by the marine nematode Rhabditis marina
Bastian. Cahiers de Biologie Marine, 16:685-694.
16. Matera, N. J. and Lee, J. J. 1972. Environmental factors affecting
the standing crop of foraminifera in sublittoral epiphytic and
psammolittoral communities of a Long Island salt marsh. Marine
Biol. 14:89-103.
17. Provasoli, L. 1977. Cultivation of animals: axenic cultivation. In:
Marine Ecology. Vol. III. (0. Kinne, ed.) John Wiley and Sons, New
York. 1295-1320.
10
-------
Comments on
The Sensitivity of Critical Life Stages
of Benthic Meiofauna to Drilling Muds
Kenneth E. Biesinger
ERL-Duluth
The grant, entitled "The Sensitivity of Critical Life Stages of
Benthic Meiofauna to Drilling Muds," was recommended for funding because
of the importance of these organisms in marine ecosystems, their sensitivity to
pollutants, the expertise of the principal investigator to culture these
organisms in the laboratory, and the possibility of developing bioassay
techniques for testing benthic marine organisms. In addition, the
proposed work was deemed tamely because of the increasing use of drilling
fluids for petroleum exploration and their possible effects on marine
organisms and because of the widespread pollution of the littoral zone
from other anthropogenic sources.
The meiofauna, which includes nematodes, copepods, foraminifera and
other small organisms, consume detritus, bacteria and algae and form an
especially important role in the marine food web. In addition, they are
sensitive indicators of environmental quality in salt marshes, tide
flats, estuaries and bays. Interactions between inorganic and organic
wastes and between sediments are poorly understood and meiofauna species
and abundance are the best indicators of environmental quality.
The principal investigator has done considerable work on the physiology,
nutritional and ecological requirements of meiofauna. He has also
successfully cultured many of these species in the laboratory and thus
is well qualified to carry out this study. The principal investigator's
expertise and facilities to test the toxicity of drilling fluids to
meiofauna were important factors in the panel's decision to recommend
funding.
The development of a bioassay procedure to test meiofauna individually
is important for determining the relative toxicity of drilling fluids;
however, the results should be used conservatively. Bioassays will give
toxicity information about the drilling fluids and needs to be tested;
but "drilling fluids" are not reagent grade and many different mixtures
are used for drilling. Perhaps more important is the composition of the
"muds" both produced from the drilling operations and those in the area
which will be affected. These will, in time, be affected by interactions
with pollutants already present. Bioassays will help elucidate the most
sensitive organisms and may be used to separate effects from toxic
substances and physical conditions such as suspended particulates and
11
-------
changes in the benthic substrates. Bfcwever, the probable area impacted
by drilling fluids, the duration of impact, and time for recovery can
best be measured under field conditions in conjunction with laboratory
experiments. For both laboratory and field studies drilling fluids and
muds should be analyzed for heavy metals (such as Pb, Cr, Hg, Ag and Ba),
appropriate organics, and physical properties which can be used as
"tracers" or "markers."
12
-------
INDICATORS OF CURRENT GROWTH RATE AS RAPID METHODS
FOR TOXICITY TESTS WITH FISH LARVAE
Ira R. Adelman and Greg P. Busacker
Department of Entomology, Fisheries and Wildlife
University of Minnesota
St. Paul, Minnesota 55108
ABSTRACT
After a series of experiments conducted for the development of
methodology and for providing an understanding of the physiological basis for
the methodologies, a few tentative conclusions have been reached. Percentage
glycine incorporation into protein by larval fathead minnow tissue (in vitro)
is reflective of protein synthesis and provides an index of growth as
evidenced by differences between fed and starved fish. Glycine uptake is
highly correlated with total available glycine and is probably the result
of active uptake and absorption. Glycine incorporation appears to follow
Michaelis-Menten kinetics. The RNA:DNA ratio also provides an index of
growth rate as evidenced by differences between fed and starved fish. Both
percentage glycine incorporation and RNArDNA ratio have been reduced at
sublethal concentrations of Cr+^I in preliminary tests and hopefully will
provide a rapid indicator of the long term effect of toxicants on growth.
INTRODUCTION
In recent years two important goals of research in aquatic toxicology
have been: 1) to find rapid techniques for the determination of the effect
of toxic materials, and 2) to find methods for determining if concentrations
of a toxicant which cause effects in laboratory assessments are the same
concentrations which cause similar effects in the natural environment of the
organism. Although exposure of a test organism to a toxicant over a full
life cycle provides the most complete information on the effect of that
toxicant and is the standard upon which the validity of other kinds of tests
are based, full life cycle tests are expensive and time consuming.
Recent studies have developed techniques for more rapid determination
of toxic effects on fish and other organisms by showing that with many
toxicants, levels of no effect determined in full life cycle tests can be
13
-------
predicted by the more rapid method, e.g. the fish cough response (1).
However, many rapid methods fail to be generally useful with a large number
of toxicants, possibly because of their failure to measure a meaningful
physiological process. McKim (2) and Macek and Sleight (3) have found that
measurement of the toxic effect during a sensitive period in the life cycle
of a fish, usually the embryonic and larval period, will provide an estimate
of toxicity very similar to that determined by a complete life cycle test
but these tests still require about 30 days. With the recently adopted
federal Toxic Substances Control Act (TSCA), it is imperative that rapid
indicators of chronic effects be developed to expedite the testing required
on an exceedingly vast array of chemicals. These rapid tests must reduce
the time and expense of the presently required embryo-larva tests, should
measure a physiological process related to the fish's well-being, and if
possible be useful in field applications as well as the laboratory testing.
Two techniques are available that with modification may enable rapid
measurement of the effects of a toxicant in both the laboratory and in
unconfined field populations. These procedures are: 1) the amino acid
incorporation growth index, and 2) the RNA-DNA ratio.
14
The C-amino acid incorporation growth index was developed by Ottaway
and Simkiss (4) as a relative measure of the rate at which a fish is growing
at the time it is sampled. The original method consists of removal of scales
from a live fish, incubation of the scales in -"-^C-labelled glycine, then
measurement of incorporation of the glycine. Close correlations have been
shown between glycine incorporation and the rate of growth of different
fish populations, different year classes, and fish in different seasons
(4,5,6). In laboratory experiments where fish were stressed by factors that
are known to influence growth rate, glycine incorporation was reduced by
handling, oxygen deprivation, and starvation (6,7,8). The method is
extremely sensitive to rapid changes in growth rate of the fish since
circadian rhythms have been detected in both laboratory (.7) and field
studies (Adelman-unpublished data).
The RNA-DNA ratio is another potentially useful indicator of growth
rate. Investigations of protein synthesis in various organisms have shown
that ribonucleic acid (RNA) increases rapidly in growing cells whereas
deoxyribonucleic acid (DNA) remains relatively constant (9,10,11). Various
studies have used the RNA-DNA ratio to effectively demonstrate rapid changes
in growth rate of fish due to changes in feeding regimes or environmental
factors (e.g. 12,13,14,15). Buckley (16) found the RNA-DNA ratio to be a
good index of growth rate in Atlantic cod (Gadus morhua) larvae.
Since these growth indicators respond rapidly to changes in growth rate
and become an immediately measurable characteristic of the fish, they are
potentially useful as rapid indicators of stress in both laboratory and
field applications. Thus, the overall objective of the present study is to
determine the "effectiveness of the amino acid incorporation growth index and
the RNA-DNA ratio as rapid methods for determination of the no effect level
of toxicants. Specific objectives to be achieved in this evaluation are to:
1) determine if the amino acid incorporation and the RNA-DNA ratio methods
will detect effects of toxicants on growth of fish in 96 hours or less, 2)
develop methodology for amino. acid incorporation by larval fish muscle tissue
14
-------
instead of scales in order to test the more sensitive life history stage
3) determine if further modification of the amino acid growth index by
exposing the muscle tissue to both the toxicant and the amino acid in-vitro
will also serve as a rapid indicator of effect on growth.
The following information is a summary of results of approximately the
first 6 months of investigation and is largely development of methodology.
AMINO ACID INCORPORATION
DEVELOPMENT AND VERIFICATION OF PROCEDURES
General Methodology
The methodology for measurement of the amino acid incorporation index
of current growth rate has previously been accomplished only on older fish
by using scales. One or more scales are removed from the fish and quickly
immersed in a physiological saline containing -^C-labelled glycine. After
incubation in the glycine the scales are rinsed, the tissue is digested, and
radioactivity that has been taken up by the scale is counted by liquid
scintillation. Final uptake data are reported as concentration of glycine
per some measurement of size such as scale weight or area in order to
normalize results among different sized scales. Glycine was chosen since
it is the amino acid of highest concentration in scale proteins.
Since part of the objective of the present study is to use incorporation
of an amino acid as an indicator of growth rate in larval fish rather than
older fish with scales, considerable development of methodology has been
required. It was assumed that the epaxial musculature including the dermal
and epidermal tissue and attached fins would be the only practical tissue to
incubate in the amino acid. Experiments have thus been directed at
determining if these tissues would take up the radiolabelled amino acid and
if uptake would be related to growth rate. Although glycine is not the most
common amino acid in muscle protein, it is one of the more common ones. To
maintain continuity with studies on uptake by scales, l^C-labelled glycine
was chosen for the larval tissue.
The larval tissue for most all of the completed experiments was
prepared by placing a fathead minnow (Pimephales promelas) larva on an iced
glass plate and cutting it transversely with a razor blade just posterior
to the gas bladder. The posterior section of the larva was then immediately
placed in the incubation medium of a physiological saline and l^C-labelled
glycine and gently agitated in a 25C water bath during the incubation period.
After incubation a variety of procedures have been attempted during the
procedural development.
Total Glycine Uptake
An initial series of experiments was conducted that ultimately produced
negative results and led to a different approach„ Three experiments were
15
-------
first conducted to: 1) determine the relationship of glycine uptake by the
larval sections to time of incubation, 2) to determine if a prolonged rinse
was necessary after incubation and 3) to determine which of three buffer
systems would result in the greatest and most linear uptake with time. The
prolonged rinse was not necessary; of the three buffers tried, two
zwiterionic buffers and one bicarbonate buffer, the latter produced the
highest and most linear uptake and was the least expensive, although pH
changes were the greatest. Uptake was approximately linear with time from
0.5 to 4 hours.
At this point an experiment was conducted to determine if a protein
synthesis inhibitor (cycloheximide) would diminish uptake of glycine by
the fathead larval sections. The cycloheximide had no effect at any of
four concentrations, thus at least most of the glycine uptake that was
being measured was not incorporated into newly synthesized protein.
Finally, an experiment was conducted to compare glycine uptake in sections
taken from fathead larvae that had been feeding with those that had been
starved for 2 days. When the size differences between the faster growing, •
fed fish and the slower growing, starved fish were normalized by surface area,
there was no difference in glycine uptake. Thus, the procedures used up to
this time would not allow the use of glycine uptake as an index of growth.
Glycine Incorporation Into Protein
Two new techniques were next considered as possibilities for development
of the procedure. Since uptake of glycine is an active process that has
been shown at least in rainbow trout intestinal cells to be a function of
glycogen content (17), an experiment was conducted to determine if the
presence of glucose along with glycine in the incubation medium would
enhance uptake and differentiate between fast growing and slow growing fish.
In addition, general methodology was altered so that glycine actually
incorporated into protein would be measured and differentiated from glycine
in the free amino acid pool. The following procedure was used:
Fathead minnow larvae (140) were divided between two tanks; one
group was fed brine shrimp and the other starved for 48 hours.
Tissue sections were incubated for 2 hours, 15 per vial, in
physiological saline which contained either glycine (0.4 yCi/ml
only or glycine and glucose (lOmM), and each of those with (100
yg/ml) or without cycloheximide. After incubation and rinsing the
larval sections were homogenized in 2 ml of distilled water. One
hundred yl of this homoge'nate was digested and radioactivity
counted to measure total uptake of glycine. Protein was precipitated
from a 500 pi sample of the homogenate by the addition of cold
(1.5 ml) 8% trichloroacetic acid (TCA) and centrifuged for 10
minutes at 10,000 rpm (4 C). The protein pellet was then washed
and centrifuged a second time with cold 6% TCA. This pellet was
digested and radioactivity counted to determine glycine incorporated
into protein. A second protein pellet was prepared from 500 ul
of the homogenate for determination of total protein by a modified
Lowry method (18).
16
-------
14
Glucose had no apparent effect on total uptake of C-glycine and in
fact may be competitive with glycine (19) (Table 1).
TABLE 1. EFFECT OF GLUCOSE AND CYCLOHEXIMIDE ON TOTAL UPTAKE AND
INCROPORATION OF 1*C GLYCINE INTO PROTEIN (pmol/yg PROTEIN) BY FED
AND STARVED FATHEAD MINNOW TISSUE SECTIONS
Feeding Cyclo-
Regime heximide*
Total
Glucose Glycine
Incor- % Inhibition
porated Incor- by Cyclo-
Glycine porated heximide (%)
Fed
Fed
Fed
Fed
Starve
Starve
Starve
Starve
Fed
Fed
Fed
Fed
Starve
Starve
Starve
Starve
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
No
Yes
Yes.
No
No
Yes
Yes
No
No
Yes
Yes
No
No
Yes
Yes
Experiment
0.192
0.235
0.206
0.263
0.360
0.559
0.541
0.342
Experiment
0.219
0.183
0.252
0.219
0.236
0.219
0.286
0.247
1
0.0088
0.0019
0.0094
0.0023
0.0148
0.0068
000031
0.0038
2
0.0114
0.0069
0.0156
0.0056
0.0070
0.0070
0.0109
0.0061
4.58
0.79
4.58
0.86
4.12
1.12
0.57
1.12
7.11
3.78
6.19
2.58
2.99
3.39
3.81
2.47
82.8
81.2
72.8
?
' 46.8
58.4
?
35.1
* 100 ug/ml in Experiment 1 and 1000 yg/ml in Experiment 2.
Cyclpheximide inhibited glycine incorporation in both fed and starved fish
by about 70-80%. Except for the starved-glucose treatment, cycloheximide
seemed to enhance total uptake of glycine. The results from the
17
-------
starved-glucose treatment were inconsistent with the other groups and are
presumed to be in error. It is apparent that total uptake masks the small
amount of ^C-glycine incorporated into protein. The starved fish generally
had a greater total uptake and incorporation of glycine after adjustment by
protein content, but the percentage incorporation seemed reduced.
A repeat of the previous experiment was conducted except that the
concentration of cycloheximide was increased 10X to 1000 yg/ml. Total
uptake across all groups was more uniform and again glucose had no effect
(Table 1). The increased cycloheximide was actually less effective at
inhibiting protein synthesis and also less effective among the starved than
the fed fish. The percentage incorporation was again less among the starved
fish indicating that percentage incorporation would provide an index of
growth rate.
Relationship of Glycine Concentration to Incorporation
The next two experiments were directed at determining the relationship
between concentration of glycine in the incubation medium and both total
uptake and incorporation. In order to minimize expenses for the large amount
of -"-^C-glycine that might be needed, it was assumed that the tissue sections
would take up and incorporate either .l^C-labelled (hot) or unlabelled (cold)
glycine without preference, i.e. in proportion to their concentration in
solution. Thus, the higher concentrations of glycine needed in these
experiments were prepared by mixing known quantitites of hot and cold glycine
and adjusting the measured radioactivi'ty taken up or incorporated by the
proportion of hot glycine in the mixture. The experimental procedures were
as follows:
Larval fathead tissue sections were incubated for 2 hours in vials
containing six different concentrations of total glycine (15 tissue
sections per vial. 2 vials per concentration). Vials were then
placed in an ice water bath and homogenization, protein precipitation,
protein concentration determination, and glycine uptake and
incorporation determination proceeded as previously described.
Glycine concentration ranged from 9 to 1800 yM concentration.
The lowest percentage of radioactive glycine was 8% in the 1800 yM
concentration. In the 9 to 44 pM glycine concentration, 100%
was radiolabelled.
Total uptake was linearly related to concentration of glycine (p < ,01),
with a regression equation of
A = 736 + 14.86 U (R2 = 99.3%)
where A is the concentration of glycine in the incubation medium and U is
the total glycine uptake. Incorporated glycine was also linearly related
to concentration (p < .01)
A = 241 + 1.64 I (R2 = 94.4%)
where I is glycine incorporated into protein. Percentage of total uptake
18
-------
incorporated into protein was erratic in relation to concentration, but
seemed to decline at higher concentration (Table 2).
The radioactive glycine used in these studies is purchased in a
hydrochloric acid (HC1) solution. After the previous experiment was
completed, it was realized that the pH of the incubation media containing
the higher glycine concentrations had been lowered by the relatively large
addition of HC1. The experiment was repeated with additional concentrations
of glycine and with the HC1 neutralized with equivalent normality ,NaOH and
the Na and K in the physiological saline increased to counteract dilution.
As in the previous experiment, total glycine uptake was again linearly
related to concentration (p < 0.01),
A = 1510 + 30.64 U (R2 = 97.7%)
Total uptake was considerably higher than in the previous experiment at
most concentrations of glycine (Table 2). The reason for this is not known
and seems unlikely to be related to the control of pH since both
relationships were linear. Glycine incorporated into protein reached a
plateau with glycine concentrations in the incubation media above 600 yM
(Fig. 1).
TABLE 2. MEAN TOTAL UPTAKE AND INCORPORATION OF GLYCINE
IN RELATION TO GLYCINE CONCENTRATION IN THE INCUBATION MEDIUM
Glycine Total glycine Incorporated glycine %
concentration (pmol/Ug) (ptnol/Pg) Incorporated
Exp. 3 Exp. 4 " Exp. 3 Exp. 4 Exp. 3 Exp. 4
9
18
36
72
144
300
600
900
1200
1500
1800
0,348
-
0.732
-
2.677
-
11.500
-
18.550
-
26.900
0.392
0.711
1.173
3.364
6.241
11.148
27.733
27.120
39.228
42.971
580120
0.062
-
0.109
-
0.561
-
. 1.775
-
2.017
-
3.138
0.053
0.083
0.173
0.402
0.476
0.739
1.110
1.377
1.354
1.344
2.003
17.8
-
14.9
-
21.0
-
15.4
-
10.9
-
' 11.7
13.6
11.6
14.7
12.0
7.6
6.6
4.0
5.1
3.5
3.1
3.4
19
-------
ro
o
CO
1.0
0.8
0.6
0.4
0.2
I I
0 0.2
0.6
1.0
1.4
1.8
Glycine (mM)
Figure 1. Incorporation of glycine into protein in relation to glycine concentration in the
incubation medium.
-------
The increased incorporation at 1800 iaM may indicate the complex kinetics
of different enzyme variants involved in the protein synthetic processes.
A Lineweaver-Burk reciprocal plot (R^ = 0.98, 9d.f.) resulted in a maximum
velocity (Vmax) of 0.665 pmol glycine per yg protein per hour. The
differences in actual glycine incorporation and percentage incorporation
between the two experiments were probably due to the neutralization of the
glycine in the latter and those results are considered more representative
of the actual relationships. (Table 2).
Relationship of Glycine Incorporation to Time
Since it had now been established that glycine incorporation rather
than uptake would be used as a growth index, it was necessary to reexamine
the relationship with time. An experiment was conducted with fed and
starved (24 hour) fathead larvae sampled from the incubation media at
15 minute intervals for 2 hours. Glyeine concentration in the incubation
medium was 290 pM with 4% radiolabelled.
Glycine incorporation increased with time, more or less linearly
through 105 minutes, then with a sudden increase during the last 15 minutes
(Table 3, Fig. 2). At this time no explanation has been found for that
increase. Percentage glycine incorporated for both fed and starved fish
generally increased with time until incorporation became asymptotic at
about 90 minutes. The reduction in percentage incorporation among starved
fish is reflective of their reduced growth rate. Tissue sections should
be incubated for at least 90 minutes to allow the rate of incorporation
of glycine to equilibrate with the rate of uptake and become constant with
time.
TABLE 3. TOTAL GLYCINE UPTAKE (pmol/Pg protein), GLYCINE INCORPORATION
INTO PROTEIN (pmoVug protein), AND PERCENTAGE INCORPORATION AT VARIOUS
TIME INTERVALS FOR FED AND STARVED FATHEAD MINNOW LARVAE
Minutes
of
Incubation
Total glycine
fed starved
Incorporated glycine
fed starved
% Incorporated
fed starved
15
30
45
60
75
90
105
120
20.571
19.619
17.904
20.571
32.380
26.875
27.428
47.238
17.905
14.667
29.333
21.714
31.238
26.875
31.048
52.000
0.524
0.848
,381
,990
,667
,286
,524
0.486
0.819
5.571
1.286
1.695
2.267
2.452
2.857
4.952
2.5
4.3
7.7
9.7
8.2
12.2
12.8
11.8
2.7
5.6
4.4
7.8
7.3
9.1
9.2
9.5
21
-------
14
& 12
.i
10
8
6
£
I 4
2
I
9
.1
30
60
Time, nwi
90
120
Figure 2. Glycine incorporation (bottom) and percentage glycine
incorporation (top) into protein in relation to time for fed
and starved larvae.
22
-------
RNA-DNA STUDIES
METHODOLOGY VERIFICATION
Chemical Analysis
Studies of RNA, DNA, and their ratio have been pursued in order to
determine if these measurements would also provide a rapid index of the
effect of a toxicant on growth rate of larval fish and to determine if this
methodology would be more sensitive than amino acid incorporation.
RNA and DNA have been analyzed in homogenized larval fathead minnows
by the Schmidt-Thannhauser method (20), modified to measure microgram
quantities of nucleic acids. RNA is measured by ultra violet (UV) absorbance
(260 run) after alkaline hydrolysis. :DNA is also measured by UV absorbance
(268 nm) after separation from protein by hot acid extraction (16). Protein
remaining after RNA and DNA removal is measured by the Lowry method (18) .
After an experimental treatment the above analytical procedures have been
performed either immediately on fresh larvae or on larvae that had been
frozen in liquid nitrogen and stored at -80C until analysis.
Relationship to Growth Rate
An initial experiment was conducted to determine differences in RNA,
DNA, and their ratio on fed and starved fathead minnow larvae. Two groups
of larvae were fed or starved for 33 hours to create a growth rate
difference,, Two duplicate samples of 10 fish from each treatment were
analyzed for RNA and DNA.
There were substantial differences in RNA, DNA, and their ratio
between the fed and starved groups (Table 4).
TABLE 4. RNA, DNA (yg/larva) AND RNAtDNA RATIO FOR GROUPS
OF FED AND STARVED FATHEAD MINNOW LARVAE
FED STARVED
Measurement Sample 1 Sample 2 Sample 1 Sample 2
RNA
DNA
RNA: DNA
11.66
6.40
1.82
12.72
6.80
1.87
1.59
2.72
0.58
2.12
2.80
0.76
23
-------
The actual quantities of RNA and DNA are based on amount per larva and
since the fed and starved fish differed in size, direct comparison cannot
be made. Later tests have measured protein present in each sample so that
size differences as reflected by protein differences can be normalized.
However, in the present test the RNA:DNA ratio is a valid comparison because
DNA is reflective of the number of nuclei present and thus theoretically
an index of size.
An experiment was next conducted to determine the effect of fish age and
size on the RNA:DNA ratio. Fish were sampled on days 1, 4, 8 and 12 for
RNA, DNA, and protein analysis. Due to increasingly large within sample
size variance with age, no statistically significant results were obtained
for RNA:DNA ratio. However, it appeared as if neither size'nor age had an
independent effect on the ratio and thus the ratio for fish up to 12 days
was not dependent on size (Table 5). DNA content was highly correlated with
protein (r = 0992, d.f = 10) and thus reflective of size.
TABLE 5. EFFECT OF SIZE AND AGE ON THE RNA:DNA RATIO DURING A 12-DAY PERIOD
Day
1
4
8
12
Mean dry
Weight of
larvae (mg)*
0.113
0.354
0.598
1.217
Mean
Protein
(U8)
42.15
72.00
156.05
499.93
RNA
(yg/mg protein)
130.7
140.7
124.6
114.5
RNA: DNA
2.00
1.89
2.08
1.95
*Determined from a different subsample of fish.
Two experiments were conducted to determine if density of larvae in
experimental chambers to be used for toxicity tests had an effect on the
RNA:DNA ratio. Results were inconclusive although there was a slight
indication that more than 30 larvae resulted in decreased ratios.
TOXICITY TEST OF HEXAVALENT CHROMIUM
' Two preliminary toxicity tests of hexavalerit chromium (Cr ) were
conducted and a 30-day larval test is underway0 In the first preliminary
test, 30 1-day old fathead minnow larvae were randomly assigned to each of
six Cr+VI concentrations and a control in a flow-through diluter modified
from Mount and Brungs (21). After 96 hours groups of fish were taken from
each treatment for the RNA,DNA analyses, and glycine incorporation
exposureso There was no clear indication of an effect of Cr+VI on any growth
index although the lowest percentage incorporation did occur at the highest
Cr concentration as did the lowest RNA:DNA ratio except for the lowest
treatment where there may have been a procedural error in subsampling (Table 6).
24
-------
+VI
TABLE 6. EFFECT OF Cr ON RNArDNA RATIOS AND
GLYCINE INCORPORATION IN FATHEAD MINNOW LARVAE
Cr Glycine uptake Glycine %
(Mg/1) RNA:DNA (pmol) Incorporation Incorporation
(pmol)
0
11
42
63
135
230
440
2.04
1.76
2.03
2.07
2.02
2.04
1.93
492
423
489
508
518
519
465
31.4
24.8
24.4
39.8
30.9
32.4
22.9
6.4
5.9
5.0
7.8
6.0
6.2
4.9
Unless fish were more sensitive in this test than the one discussed below,
the low ratio and glycine incorporation in the high treatment were probably
an artifact.
+VI
In a second experiment with hexavalent chromium, two series of Cr
concentrations, increased from the previous test, were run in duplicate.
The test chambers in one series contained 15 larvae which were sampled at
96 hours for RNA:DNA ratio. The chambers in the other series contained 35
larvae which were sampled at 120 hours for RNA:DNA ratio and glycine
incorporation.
+VI
The increased Cr concentrations resulted in a visually observable
effect on growth and some mortality occurred in the higher treatments. The
RNA-.DNA ratios were clearly affected at both 4 and 5 days at Cr+v*
concentrations between 2.4 and 4.1 mg/1 (Table 7). Differences between 4
and 5 days may be due to density. The percentage incorporation of glycine
seemed to be affected between the next higher treatment groups, i.e.
4.1 and 7.6 mg Cr+VI/l (Table 7).
With these results, a full length (30 day) embryo-larval test was
undertaken to examine sensitivity of the RNArDNA ratio at 4 days for
predicting the effect of Cr+VI on growth over 30 days. Glycine incorporation
trials will also be conducted but results of these tests are not yet
completed.
CONCLUSIONS AND FUTURE STUDIES
GLYCINE INCORPORATION
The incorporation of glycine as measured in TCA extracted larval tissue
homogenates is due to protein synthesis as evidenced by cycloheximide
25
-------
+VI
TABLE 7. EFFECT OF Cr ON RNA:DNA RATIOS AND GLYCINE
INCORPORATION IN FATHEAD MINNOW LARVAE
Cr+VI
(mg/1)
RNA:DNA
4 days 5 days
Glycine Uptake
(pmol)
Glycine
Incorporation
(pmol)
%
Incor-
poration
0 2.72 2.00 1013 82.2 8.1
1.2 2.51 2.07 934 78.7 8.5
2.4 2.31 2.12 1134 92.9 8.2
4.1 2.08 1.60 812 75.7 9.3
7.6 1.57 - 581 42.9 7.4
12.8 1.20 - 544 35.4 6.5
21.0 - - 514 33.3 6.5
inhibition of incorporation. Total uptake is dependent on size of the
exposed tissue and may be the result of both active uptake and physical
absorption; thus, total uptake is not necessarily reflective of glycine
incorporation. For example, starved fish may have equal or greater uptake
on a per size basis when compared to fed fish, but their percentage
incorporation of glycine is less. Because a measure of protein synthesis
should be a better index of growth rate than combinations of active uptake
or absorption, the analysis of glycine incorporation as indicative of
protein synthesis will provide a better index of growth rate than total
uptake. In order to normalize results for different sized fish, the
dimensionless unit of percentage incorporation should provide an index
unbaised by size. An additional experiment using groups of different sized
fish will be conducted to verify this assumption.
Glycine uptake is highly correlated with the total available glycine
in the incubation medium but incorporation is not directly proportional to
available glycine. Rather, incorporation appears to follow Michaelis-Menten
kinetics.
Results of the preliminary toxicity tests with Cr appear promising
and additional toxicity testing will be conducted over the next year at the
University of Minnesota laboratory and in conjunction with ongoing projects
at the Environmental Research Laboratory-Duluth. There will be a continued
effort devoted to determination of the biological basis for the observed
experimental results. Autoradiography of larval tissue sections will be
used to examine the location of the protein synthetic activity and
improvement in overall methodology will be sought in an attempt to understand
and reduce the cause of sample variability.
RNA:DNA RATIO
The major problem still hindering the RNA:DNA ratio experiments is
26
-------
analytical variability. Continued effort will be devoted to improvements
including measurements of analytical precision and efficiency of
extraction of the proteins and nucleic acids. Also, the experiment to
establish the size independence of percentage glycine incorporation will be
used for establishment of size independence of RNA:DNA. As with the
glycine incorporation methodology, the same toxicity tests will be used to
compare RNA:DNA ratios with toxicant effect on growth and to compare the
sensitivity of the two indices.
27
-------
REFERENCES
1. Carlson, R. W., and R. A. Drummond. 1978. Fish cough response-a
method for evaluating quality of treated complex effluents.
Wat. Res. 12:1-6.
2. McKim, J. M. 1977. Evaluation of tests with early life stages of fish
for predicting long-term toxicity. J. Fish. Res. Board Can.
34:1148-1154.
3. Macek, K. J., and B. H. Sleight, III. 1977. Utility of toxicity tests
with embryos and fry of fish in evaluating hazards associated
with the chronic toxicity of chemicals to fishes, pp. 137-146.
In F. L. Mayer and J. L. Hamelink (eds.). American Society for
Testing and Materials, ASTM STP 634.
4. Ottaway, E. M., and K. Simkiss. 1977a. "Instantaneous" growth rates
of fish scales and their use in studies of fish populations.
J. Zool. (London) 181:407-419.
5. Ottaway, E. M., and K. Simkiss. 1979. A comparison of traditional
and novel ways of estimating growth rates from scales of natural
populations of young bass (Dicentrarchus labrax). J. Mar. Biol.
Assoc. U.K. 59:49-59.
6. Ottaway, E. M. 1978. Rhythmic growth activity in fish scales. J.
Fish Biol. 12:615-623.
7. Ottaway, E. M., and K. Simkiss. 1977b. A method for assessing factors
influencing "false check" formation in fish scales. J. Fish
Biol. 11:681-687.
14
8. Adelman, I. R. 1980. Uptake of C-glycine by scales as an index of
fish growth: Effect of fish acclimation temperature. Trans.
Amer. Fish. Soc. 109: 187-194.
9. Davidson, J. H., and I. Leslie. 1950. A new approach in the
biochemistry of growth and development. Nature 165:49-53.
10. Hotchkiss, R. 1955. The biological role of the deoxypentose nucleic
acids, pp. 435-473. In E. Chargaff and J. Davidson (eds.). The
nucleic acids, chemistry and biology, Vol. 2. Academic Press, NY
11. Sutcliffe, W. H. 1965. Growth estimates from ribonucleic acid contents
in some small organisms. Limnol. Oceanogr. 10:R253-R258.
12. Bulow, F. J. 1970. RNA-DNA ratios as indicators of recent growth
rates of a fish. J. Fish. Res. Board Can. 27:2343-2349.
13. Haines, T. A. 1973. An evaluation of RNA-DNA ratio as a measure of
long-term growth in fish populations. J. Fish. Res. Board Can.
30:195-199.
28
-------
14. Raines, T. A. 1980. Seasonal patterns of muscle RNA-DNA ratio and
growth in black crappies. Environ. Biol. Fish. 5:67-70.
15. Bulow, F. J., M. E. Zeman, J. R. Winningham, and W. F. Hudson. 1981.
Seasonal variations in RNA-DNA ratios and in indicators of
feeding reproduction, energy storage, and condition in a population
of bluegill. J. Fish. Biol. 18:237-244.
16. Buckley, L. J. 19790 Relationships between RNA-DNA ratio, prey
density, and growth rate in Atlantic cod (Gadus morhua) larvae.
J. Fish. Res. Board Can. 36:1497-1502.
17. Boge, G., A. Rigal, and G. Peres. 19790 A study of energized transport
mechanisms of glycine absorption by rainbow trout. Comp. Biochem.
Physiol. 64A:537-541.
18. Hartree, E. F. 1972. Determination of protein: A modification of the
Lowry method that gives a linear photometric response. Anal.
Biochemo 48:422-427.
19. Prosser, C. L. (ed.). 1973. Comparative animal physiology. W. B.
Saunders Co., Philadelphia.
20. Munro, H. N., and A. Fleck» 1966. Recent developments in the
measurement of nucleic acids in biological materials: A
supplementary review. Analyst 91:78-88.
21. Mount, D. I., and W. A. Brungs» 1967, A simplified dosing apparatus
for fish toxicity studies. Water Res. 1:21-29.
29
-------
Comments on
Indicators of Current Growth RAte
As Rapid Methods for Tbxicity
Tests with Fish Larvae
Guenther Stotzky
New York University
One of the aims of environmental toxicology is the development of
rapid, simple, and inexpensive assays that will reflect accurately the
physiological effects of a toxicant on the biota in situ. Many such
assays exist, using test organisms ranging from microbes to animals and
in time required from hours to complete life cycles of months. Assays
using whole animals and life cycles seldom meet the above criteria,
although ultimately all assays must be correlated with complete life
cycle evaluations, and assays using indicator organisms do not always
reflect the more complex physiological responses of the higher organisms
of concern in situ.
The objectives of this project are to develop and test tWD rapid
methods to determine the "no effect level of toxicants" at the molecular
level. One method involves measuring the uptake and incorporation into
protein of a C-labeled amino acid (glycine), and the other method
involves determination of RNArDNA ratios. Both methods are sensitive
indices of cellular activity and, therefore, should quickly and efficiently
detect the toxicity of a spectrum of potential toxicants. The technique
involves the use of fat head minnow larvae tissue (the entire portion of
the larva posterior to the gas bladder), and therefore, it uses a test
organisms that is physiologically similar to a major group of higher
organism of concern in aquatic environments. An important advantage of
this assay is that it requires less than 96 hours to conduct in contrast
to the 30 days usually required for estimating toxicity during the
embryonic and larval period of fish. Another advantage is that a certain
amount of redundancy is built into the assay: i.e., amino acid incorpora-
tion and RNA synthesis, in essence, measure the same phenomenon, protein
synthesis. This redundancy, however, may provide some bonus information
on cellular mechanisms of pollutant toxicity, especially if amino acid
incorporation and RNA synthesis are not correlated. For example, the
uptake of the amino acid may be subject to catabolite repression and
other changes in metabolism (e.g., in active uptake and permeation),
whereas RNA synthesis may be unaffected, thereby indicating the cellular
level at which the toxicant is operating.
The Principal Investigator is methodically developing and verifying
the procedures, as suggested by the original review panel, and has made
reasonable progress since the grant was awarded. Additional testing is
30
-------
obviously needed to resolve a number of questions (e.g., use of RNase
and DNase to verify the RNA:DNA ratios now determined by absorbance
after alkaline and acid treatments, respectively; the influence of the
shock of beheading on uptake and incorporation of the labeled amino
acid). The major need, however, is to increase the number of larvae per
test and the number of tests so that reliable statistical analyses can
be applied to the data.
In summary, the review panel recommended support of this project,
as it concluded that the proposal was scientifically sound, that the
Principal Investigator was competent to do the work, and that the data
obtained would have high utility to EPA, especially as some cooperative
studies with EPA-Duluth and the Chesapeake Bay Institute are planned.
31
-------
GROWTH KINETICS OF AEROMONAS HYDROPHILA IN FRESH WATERS
AND THEIR RELATIONSHIP TO TROPHIC STATE
Scott R. Rippey, Savas C. Danos, and Victor J. Cabelli
Department of Microbiology, University of Rhode Island
Kingston, Rhode Island 02881
ABSTRACT
A previously reported study, showing that the in situ Aeromonas
hydrophila densities in bodies of fresh water are correlated with their
trophic states, was extended to examine the potential of the water to
support the growth of seeded A. hydrophila. No growth of the seeded
bacteria occurred in untreated (raw) water samples incubated at room tem-
perature irrespective of the trophic state of the water body or the season
when the samples were collected. Multiplication to relatively high
densities was observed, however, in filtered-autoclaved water samples.
Moreover, the maximum A. hydrophila densities attained greatly exceeded the
summer in situ densities for the same Rhode Island ponds.
The generation times of the organism, determined for the ten Rhode
Island ponds utilized in the growth studies, were strongly correlated with
the Relative Trophic Index (RTI), a measure of trophic state. Of those
parameters comprising this index, total phosphorus, chlorophyll a^ and Secchi
depth correlated strongly with the generation time for the organism. Water
conductivity was also correlated with generation time. Maximum
A. hydrophila densities observed in the filtered-autoclaved water samples
were not correlated with any individual parameter, the RTI, or generation
time.
INTRODUCTION
The authors, in an earlier publication (1), reported that the in situ
densities of an autochthonous, heterotrophic bacterium, Aeromonas hydrophila,
were inversely correlated with the Relative Trophic Index, RTI (2), of fresh
water environments. These findings suggested that this bacterial parameter
could be useful in trophic state assessments and possibly in the early
prediction of long term limnetic changes in these aquatic environments.
However, the reasons for the observed correlation between the densities of
_A. hydrophila and the trophic state of fresh waters have yet to be deter-
mined .
32
-------
A number of factors could directly or indirectly influence the
A. hydrophila densities attained in limnetic waters. Included are nutrient
availability; physical parameters such as temperature, pH, dissolved oxygen,
suspended solids, etc.; antibiosis from and competition with other biota;
and predation by protozoa, bacteria, or bacteriophage. Of these, only the
effects of temperature on the in situ densities of A^. hydrophila in natural
waters (1) and the physiological activities of the organism in culture (3)
have been well documented. Some data are available, however, on the growth
kinetics of the organism in nutrient supplemented tap waters (4). These
data indicate that the generation time is substantially influenced by added
sources of carbohydrate when nitrogen and phosphorus are not limiting.
The results presented herein consider nutrient availability as one
explanation for the observed correlation between in situ A. hydrophila
densities and trophic state. One objective of this aspect of the overall
project was to determine the conditions required for the growth of seeded
^. hydrophila in environmental water samples. The second was to examine
the growth kinetics of ^. hydrophila seeded into waters of varying quality.
This was done in order to determine which, if any, of the growth parameters
are correlated with the various physical, chemical or biological parameters
(including the RTI) used in defining the trophic state of fresh waters.
MATERIALS AND METHODS
Growth Studies
Water samples for the growth studies were taken from various Rhode
Island.ponds at approximately 0.5 m depth and at selected limnetic sites
away from the vicinity of aquatic marcrophytes. They were collected in
acid-cleaned (overnight immersion in 10% HC1 followed by six deionized-
distilled water rinses), autoclaved, polypropylene containers. Upon return
of the samples to the laboratory, those portions to be used in growth
studies with raw (untreated) water were dispensed in 200 ml quantities to
acid-cleaned, sterile, 500 ml Erlenmeyer flasks. Filtered-autoclaved
water was obtained by passing portions of the samples through pre-washed
(200 ml of deionized water), 0.22 ym GS Millipore filters (Millipore Corp.,
Bedford, MA.), at a vacuum of less than 13 cm Hg. The filtrates were
dispensed to flasks as described; and the flasks were autoclaved at 121 C
for 15 min, allowed to cool to room temperature, and vigorously shaken to
aerate the water. They were then inoculated with the A., hydrophila test
suspensions to a final density of approximately 10 organisms/ml, incubated
at room temperature, and monitored periodically using the mA procedure for
A. hydrophila (5). The interval between collection of the water sample and
its treatment generally did not exceed six hours.
In those instances in which the ability of a given body of water to
support the growth of the seeded ^. hydrophila was examined against the
"nutrient" levels in the water, 1 L water samples were collected from each
of six, randomly selected, limnetic sites. These were pooled; the pool was
filtered; the filtrate was dispensed into triplicate flasks; and the flasks
of water were autoclaved and treated as- described earlier.
33
-------
The log-in densities of ^. hydrophila were plotted against the incubation
time in hours; two points during maximum exponential growth were identified;
and these were used to calculate the generation (g) time using the formula
(t2 - tj_) In 2
S = In D2 - In DX
where t, and t2 and D, and D~ are the incubation times in hours and den-
sities per mil f-or the first and second points.
Inoculum Preparation and Maintenance
An A. hydrophila strain (N-l) isolated from a highly oligotrophic body
of water (Bay of Naples, Maine) was selected for the growth studies. The
cells from a nutrient broth culture of the strain (incubated for 18 h at
37 C) were washed three times by alternate centrifugation and resuspension
of the cells in 0.85% NaCl. A flask of filtered, autoclaved, oligotrophic
pond water was inoculated with an appropriate dilution of the _A. hydrophila
suspension and incubated as described above. Log-phase cells from this
culture were serially passed into flasks of the same medium in order to
provide a continuous supply of log-phase cells for use as the test inocula
in all growth experiments.
Physical, Chemical and Biological Parameters
Six parameters were measured in assessing the trophic states of the ten
Rhode Island ponds examined. They included hypolimnetic dissolved oxygen,
water transparency (Secchi disc), total phosphorus, dissolved reactive
phosphorus, inorganic nitrogen (NH, + N02 + NO- as N) , and chloroyphyll a_.
Water temperature and conductivity were also measured. The six parameters
were used in calculating a relative trophic index (RTI) for each body of
water using a previously described procedure (1,2).
In situ dissolved oxygen, conductivity, and temperature measurements
were made with a calibrated YSI model 54 oxygen meter (Yellow Springs
Instrument Co., Yellow Springs, OH.) and YSI model 33 S-C-T conductivity
and temperature meter, respectively. Water transparency was determined
using a Secchi disc.
Samples for chemical analyses were collected at a "deep-station" with a
messenger tripped, 1 liter, teflon-coated Van Dorn sampler (General
Oceanics, Miami, FL.)« Under conditions of stratification, two epi-, two
meta-, and two hypolimnetic samples were taken; otherwise three samples
were collected at representative depths. Samples were transferred into
500 ml acid-washed sample bottles. All samples were held at ambient tem-
perature in the dark and processed in the laboratory within four hours.
Subsurface and pooled surface water samples for chemical and chlorophyll
a. analyses were treated as follows: A 100 ml portion of the sample was
filtered through pre-washed (200 ml of deionized water) 0.45 ym HA Millipore
filters. The filtrate was delivered into an acid-cleaned 125 ml poly-
propylene bottle and immediately frozen for future dissolved nutrient
34
-------
analyses; the filter was frozen for later chlorophyll ji analysis. The
second 100 ml portion was delivered (unfiltered) into an acid-cleaned 125 ml
sample bottle and frozen for later total phosphorus analysis.
Colorimetric analyses for ammonia-nitrogen (NH- - N), nitrate-nitrite-
nitrogen (NO. - N0« - N), dissolved reactive phosphorus (PO, - P), and total
phosphorus (TP) were made using a Technicon Auto Analyzer II by the follow-
ing methods (6); NH- - N, Berthelot reaction; NO- - NO- - N, copper-cadmium
reduction followed By sulfanilamide diazotization, N-(I-napthyl)-ethylene
diamine method; PO, - P, phospho-molybdenum blue method; TP, potassium
persulfate digestion followed by PO, - P analysis. Fluorometric analysis of
chlorophyll £ utilizing 90% acetone extraction was according to Strickland
and Parsons (7).
Statistical Analyses
The degree of association between variables was tested by the non-
parametric Spearman rank correlation analysis utilizing the SAS computer
statistical package (8).
RESULTS
Growth of the indigenous or seeded A. hydrophila at room temperature did
not occur in the untreated (raw) water samples collected during the fall and
early winter from eutrophic as well as oligotrophic waters in Rhode Island
(Fig. 1). Pursuant to the identification and circumvention of those factors
in raw water which prevented the growth of the seeded A. hydrophila,
membrane-filtered (0.22 ym Millipore GS) water samples were examined for
their abilities to support the growth of the seeded bacteria. This treat-
ment was consistent with one objective of the study, the correlation of the
growth kinetics of seeded A. hydrophila in the water to the concentrations
of dissolved nutrients therein. However, membrane filtration alone was
often found to be unsatisfactory for treatment of the water because the
sterility of the filtrate could not be assured. When contamination of the
filtrate occurred, the growth of A. hydrophila was generally retarded or
precluded. Moreover, sterile filtrates, even from eutrophic bodies of
water, did not always support growth of seeded A., hydrophila. This inhib-
itory effect could be eliminated by autoclaving the water at 121 C for
15 minutes. Therefore, the water samples used in all subsequent growth
studies were filtered and then autoclaved. When the water samples were so
treated, growth occurred; and relatively high densities of the organism
resulted (Fig. 1).
35
-------
DAYS
Fig. 1. Growth characteristics of ^. hydrophila in untreated (o) and
filtered-autoclaved (•) oligotrophic and in untreated (A) and filtered-
autoclaved (A) eutrophic pond waters.
Water samples from ten Rhode Island ponds were taken during the months
of March through early June for the conduct of growth studies and the
physical, chemical, and biological analyses necessary for trophic state
assessment. The results from some typical growth experiments using
filtered-autoclaved water samples from oligotrophic, mesotrophic, and
eutrophic ponds are shown in Figure 2. It can be seen that lag periods
were minimal or absent, presumably because cells from exponentially growing
cultures in oligotrophic water were used as the inocula. Irrespective of
the trophic state of the water used, the maximum densities (D . ) achieved
in the growth experiments markedly exceeded those observed in situ during
a previous summer; however, they were considerably less than those attained
in nutrient broth (Table 1).
36
-------
DAYS
Fig. 2. Growth kinetics of A. hydrophila in filtered-autoclaved water from
eutrophic (•), mesotrophic (o), and oligotrophic (A) water bodies, (points
are means of 3 replicates). The mean generation times were 1.7 h, 3.5 h
and 10.3 h, respectively. '
37
-------
TABLE 1. COMPARISON OF IN VITRO, MAXIMUM GROWTH DENSITIES OF A. hydrophila
FOR RHODE ISLAND PONDS TO THEIR RESPECTIVE IN SITU DENSITIES
Pond
Wo r den
30 Acre
Tucker
Deep
Yawgoog
Nutrient
broth
Trophic
State3
E
E
E
M
M
_
A. hydrophila
in situ
2.0 x 103
1.5 x 103
3.6 x 102
8.5 x 101
8.5 x 101
_
Density/100 ml
in vitro0
1.4 x 107
1.3 x 107
3.2 x 107
3.6 x 107
2.1 x 107
3.8 x 1010
E - eutrophic; M - mesotrophic
b From 1978 data ( 1 )
From growth studies as described in text.
Large differences in generation (g) time were evident among the ponds
of varying water quality (Fig. 2), and the g time appeared to correlate
with trophic state. Therefore, the degrees of association of g time and
DMAX achieved in the growth studies to those parameters indicative of
trophic state (as well as the RTI) were examined using the Spearman rank
correlation test. The correlation coefficients (r) and their significance
(P) are shown in Table 2. Only the generation time was significantly
correlated to any of the individual trophic state parameters or the RTI;
the highest r value (0.91) obtained was for g time vs. the RTI calculated
for the ten Rhode Island ponds using the National Eutrophication Survey's
209 lake data set (2). Highly significant correlations were also found
for g time with total phosphorus and chlorophyll a_; significant correlations
were obtained with both Secchi depth and conductivity; no correlation was
observed with dissolved inorganic nitrogen or dissolved reactive phosphorus.
38
-------
TABLE 2. ASSOCIATION OF A. hydrophila GROWTH PARAMETERS TO PHYSICAL,
CHEMICAL, AND BIOLOGICAL TROPHIC STATE MEASUREMENTS AND THE RTI
n
Parameter r Value for Parameter
G Dmax
RTI (National base)d 0.91** -0.18 NH-j + NO" + N0~(as N)
RTI (Local base)6 0.82** -0.23 Dissolved React.
Phosphorus
Total Phosphorus -0.80** 0.34 Conductivity
Chlorophyll a_ -0.81** 0.14 Dmax
Secchi Disc 0.73* -0.16
* P <0.0005 ** P <0.0001
a
Spearman rank correlation coefficients
b c
Generation time; Maximum density achieved
d
a
r Value for
G Dmax
-0.20 -0.17
-0.34 0.35
0.77* -0.06
-0.18
Relative Trophic Index based on ranking the 10 Rhode Island ponds using
the 209 lake data base from the 1972 NES study ( 2 )
Relative Trophic Index based on ranking the 10 Rhode Island ponds among
themselves
The longest generation time observed was 12.8 hours, and the shortest
was 1.25 hours. The generation time for the nutrient broth culture, treated
identically to the test cultures, was 0.56 h. Furthermore, from the gener-
ation time-RTI relationship shown in Figure 3, it would appear that a g
time .comparable to that observed with the nutrient broth cultures would not
be achieved even in water samples from the most highly eutrophic waters.
39
-------
12
UJ
5
h-
UJ
z
UJ
UJ
100
200 300
LRTI
400
Fig. 3. Relationship of the generation (g) time for the growth of
A., hydrophila to the Relative Trophic Index derived from the 10 Rhode Island
ponds themselves (r=0.82, P<0.0005). Generation times for _A. hydrophila
are the means from three replicate growth experiments. The broken line is
the g time in nutrient broth.
DISCUSSION
An earlier report described the relationship of the in situ
A. hydrophila densities found within a given body of water to its trophic
state (3). However, this bacterial measurement cannot be used in the late
fall, winter, and early spring in temperate zones, presumably because of
temperature limitations on the growth of the organism. A practical objec-
tive of the present study was to examine the growth characteristics of the
organism in laboratory studies as an alternative or adjunct to the in situ
bacterial measurements.
The mean generation time, as determined from growth studies using
filtered, autoclaved water, appears to be a reasonable alternative to the
in situ densities, especially for oligo- to early mesotrophic bodies of
water. However, the discriminating ability of this measurement decreases
as the waters become more eutrophic; and it would appear that, even in the
40
-------
most eutrophic water, the exponential growth rate would not be comparable
to that obtained in nutrient broth. One possible explanation for this
observation is that the availability of one or more specific nutrients
(e.g. vitamins, amino acids, energy sources), other than those included in
the RTI, limited the growth rate of A. hydrophila in these waters. The
lack of correlation between the maximal density attained and any of those
parameters indicative of trophic state may be explained in a similar way.
This explanation is being pursued against the possibility that, once these
nutrients are identified and added to the test waters, the generation times
for eutrophic waters and the maximum densities achieved will be better
correlated with trophic state and approach those densities observed in
nutrient broth. Moreover, the supplemented water would then provide the
"base medium" for examining the effects of added trophic state associated
nutrients (e.g. nitrogen, phosphorus) on the growth of seeded A. hydrophila.
The lack of growth of A. hydrophila in any of the raw waters collected
in Rhode Island during the fall and winter was a consistent phenomenon. It
is possible that k. hydrophila is multiplying in these waters but that
pressures such as predation, toxicity, competition, etc. are so intense
that density increases are never observed. This explanation probably
accounts for the observation that the maximum A. hydrophila density obtained
in the growth studies for a given pond consistently exceeded the in situ
densities observed during the summer. It is also consistent with the
observation that the densities of seeded and indigenous A. hydrophila in raw
waters are markedly reduced within 8-10 days, while the high densities
achieved in filtered-autoclaved water from the same pond are maintained for
extended periods of time, at least 23 days (data not shown). This aspect
of the study also is being pursued. Some preliminary experiments are in
agreement with the hypothesis that predation, toxicity, and competition are
important in limiting the in situ A. hydrophila densities.
The generation time of A. hydrophila seeded into filtered-autoclaved
water samples does show promise as a trophic state indicator for bodies of
fresh water. This promise may be more fully realized for the generation
time and even the maximum density attained in growth studies once more
information is obtained on the nutrients which limit the growth of the
organism in natural waters. Moreover, these growth parameters could be
more indicative of trophic state than are the in situ densities of the
organism since they are independent of predation, competition, toxicity,
etc. In addition, as noted earlier, samples may be taken for assay during
any season of the year.
REFERENCES
1. Rippey, S.R. and V.J. Cabelli. 1980. Occurrence of Aeromonas
hydrophila in Limnetic Environments: Relationship of the Organism to
Trophic State. Microb. Ecol. ^:45-54.
2. United States Environmental Protection Agency, National Eutrophication
Survey, Pacific Northwest Environmental Research Laboratory, Corvallis,
Oregon, Working Papers No. 24 (1972) and No. 900 (1977).
41
-------
3. Cavari, B.Z., D.A. Allen, and R.R. Colwell. 1981. Effect of Tempera-
ture and Growth and Activity of Aeromonas spp. and Mixed Bacterial
Populations in the Anacostia River. Appl. Environ. Microbiol. 41:
1052-1054.
4. van der Kooij, D., A. Visser, and W.A.M. Hijnen. 1980. Growth of
Aeromonas hydrophila at Low Concentrations of Substrates Added to Tap
Water. Appl. Environ. Microbiol. 39:1198-1204.
5. Rippey, S.R. and V.J. Cabelli. 1979. Membrane Filter Procedure for
Enumeration of Aeromonas hydrophila in Fresh Waters. Appl. Environ.
Microbiol. _38_: 108-113.
6. Technicon. 1973. Industrial Methods. Technicon Instruments Corp.,
Tarrytown, N.Y.
7. Strickland, J.D. and T.R. Parsons. 1972. A Practical Handbook of
Seawater Analysis, 2nd ed. Fish. Res. Bd. Can. Bull. 167.
8. SAS User's Guide. 1979. Ed. by J.T. Helwig and K.A. Council. SAS
Institute, Inc., Raleigh, North Carolina.
42
-------
Comments on
Growth Kinetics of Aerononas hydrophila
in Fresh Waters and Their Relationship
to Trophic State
Guenther Stotzky
New York University
Cne of the major factors that causes deterioration of bodies of
freshwaters is an increase in the^trophic (i.e., nutritional) level.
Such changes in the trophic state can result in enhanced growth of
phyto- and zooplankton, in a decrease in the dissolved oxygen content
with a concomitant decrease in the oxygen-requiring biota, in an increase
in microorganisms including species pathogenic to animals, in production
of toxic substance, in aesthetic deterioration, and in numerous other
conditions that can render the waters unfit for consumption and recreational
uses. Although increases in trophic states occur naturally, these are
slow and long-term processes. More frequently, increases in trophic
states are anthropogenic and usually result from inputs of sewage or
industrial wastes. Because of the fragility of most freshwater systems,
it is important that even subtle changes in trophic states be detected
early and accurately.
Numerous methods, including chemical, physical, and biological
assays, have been developed for detecting changes in trophic states.
Although many of these methods, especially when used in combination,
have been successful, most are time-consuming, expensive, and require
specialized equipment. The development of a microbiological assay using
a sensitive indicator species would eliminate many of these problems, as
such assays are relatively rapid, inexpensive, reproducible, require
relatively little equipment, and can be conducted by essentially untrained
personnel.
The purpose of this research project is to develop and test such a
microbiological assay, using the bacterial species, Aercmonas hydrophila.
There are several characteristics that recommend the use of this species:
it is a strict heterotroph and, therefore, dependent for survival on
organic matter either introduced into or synthesized by autotrophs
(which require inputs of inorganic nutrients) in the aquatic systems; it
is a pathogen of human beings; and it does not form spores and, therefore,
does not overwinter well nor survive long in the absence of nutrients.
Consequently, the presence of A._ hydrophila in waters can suggest that
the nutrient loadings are of sewage origin.
Three other important factors that recommended this research
project were that preliminary studies by the principal investigator had
43
-------
shown that densities of A._ hydrophila were correlated with other indicators
of enhanced trophic states (e.g., chlorophyll a_ and total phosphorus
contents, turbidity, and the "relative trophic~~index"); the development,
by the principal investigator, of a method for quantifying this bacterium
in waters; and the competency, qualifications, and productivity of the
principal investigator.
Inasmuch as this is a basic research project, the studies will
extend beyond just demonstrating further the correlation between trophic
states and densities of A. hydrophila in freshwaters. They will attempt
to determine which specifTc nutrients in nutritionally-enriched waters
are responsible for the enhanced growth of this specific indicator
organism and whether there are significant nutritional differences
between biotypes of this species which would differentially affect their
survivial and growth. In addition, factors other than nutrition that
could influence the density of the indicator organism (e.g., anti-microbial
agents, predation) will be studied, inasmuch as such non-nutritional
factors could indirectly affect population densities and, thereby, the
correlation. The project will also determine whether the growth of
A. hydrophila can be used to measure and predict the relative effectiveness
oT procedures to restore (i.e., decrease the trophic state) deteriorated
bodies of freshwaters.
As with all research proposals, the review panel had a number of
questions about this proposal. However, these were, for the most part,
minor, and the concensus was that the principal investigator, given his
experience and capabilities, would be able to resolve these.
In summary, the panel recommended support of this project, as it
concluded that the proposal was scientifically sound, would result in
significant new basic information, and that the Environmental Protection
Agency would gain valuable information to aid it in assessing trophic
states of freshwaters and which the Agency could use in its monitoring
and regulatory functions.
44
-------
DRILLING FLUID INDUCED CHANGES IN NEURONAL
ACTIVITY MONITORED IN AN INTACT, BEHAVING MARINE ANIMAL
James E. Kanz and Michael D. Duvall
Marine Biology Department
Texas A&M University at Galveston
Building 311, Fort Crockett
Galveston, Texas 77550
INTRODUCTION
The greater need and utilization of domestic oil reserves has resulted
in an increase in off-shore drilling particularly near the edge of the conti-
nental shelf. The importance of the continental shelf, and its associated
coral reefs, to the ecological productivity of surrounding areas (1), has
heightened concern over the environmental impact that such drilling operations
pose to the surrounding fauna and flora.
The effects that oil-rig related pollution parameters have on the biota
can be ascertained by subjecting the biota to various concentrations of one
or more of the .polluting parameters and monitoring changes in selected be-
havioral, physiological and developmental systems. This approach has been
employed to determine the toxicity of such drilling-related environmental
pollutants as petro-hydrocarbons (2, 3, 4, 5, 6, 7). More recently, research
has begun to determine the toxic effects of drilling fluids on biota endemic
to off-shore communities (8, 9, 10, 11, 12, 13, 14). Drilling fluids are
heavy mixtures which can include fungicides, bacteriocides and pesticides and
are used to facilitate drilling-related functions, e.g., drill-cutting re-
moval and drill bit lubrication (15, 16).
We report here the first experiments to monitor changes in an animal's
(the marine gastropod Mollusc, Aplysia californica) nervous system activity
induced by drilling fluid. Our research utilized a recently developed extra-
cellular electrophysiological recording technique (17) that makes it possible
to monitor the neuro -physiological correlates of behavioral patterns in
intact, freely-behaving Aplysia (18, 19). Changes in neuronal activity
always precede behavioral changes and can occur without any visible .behavior
modifications (20). Therefore, detection of neuronal activity changes eli-
cited by the presence of an environmental pollutant (such as drilling fluid)
provides an important, sensitive method for determining toxicological effects
below and above the threshold of observable behavioral modifications.
45
-------
METHODS
Aplysia californica obtained from Pacific Bio-Marine Laboratories, Inc.,
were housed in 16 x 19 x 19 cm perforated plastic chambers within a 130-liter
Instant Ocean tank (with automatic aeration and filtration) at 15-17°C under
a 12:12 hr. light-dark schedule. Animals were fed dried sea weed and each
was assigned a number (CC = chronic californica). Drilling fluid (Sample
JX-FL 3-1 from July 11, 1980) was obtained from the Gulf Bfceeze Laboratories
of the U. S. Environmental Protection Agency, Gulf Breeze, Florida. A 10 cc
volume of 10 ppt whole drilling fluid (wet weight) minus tha barium sulfate
was introduced via silastic cannulation tube (2 mm O.D) into the body cavity
of experimental animals at 0.5 cc/sec. Control animals received either 5 or
10 cc of sterilized isotonic saline (sea water) by cannulation.
Surgery for implantation of the in vivo extracellular recording
electrodes and cannulation tube was performed according to Kanz et al. (18).
Following surgery, pre-operative body weight was restored by injecting
sterilized isotonic sea water and the animal was allowed to recover overnight.
The silastic cannulation tube was sutured to the internal body wall muscu-
lature. The monopolar electrodes were implanted on the Aplysia's siphon
nerve, a peripheral nerve from the abdominal ganglion which innervates the
siphon, an organ used in expelling sea water and excretory products from
the mantle cavity, (Figure LA). The double cuff electrode assembly (Figure
IB) consisted of two independent stainless steel wires each referenced to an
indifferent lead in the chamber; each pair (i.e., monopolar and indifferent
lead) was attached to the head stage of a differential amplifier. Signals
from the amplifier passed through a 60 and 120 Hz Notch Filter (to eliminate
some of the background noise), then to an oscilloscope, FM tape recorder and
pen recorder. Thus, two independent electrophysiological records, were
obtained from loci on the siphon nerve 10 mm apart (Figure 1C).
Two neurophysiological parameters were monitored: 1) the frequency
and pattern of an identified burst of neuronal activity from the siphon
nerve, Interneuron II (INT II) associated with a nearly simultaneous con-
traction of the gill, parapodia and siphon and believed to function in re-
placing "old" sea water with "new" in the mantle cavity (21, 22) and 2) the
level of background spontaneous activity in the siphon nerve (Figure 1C).
The experimental protocol used 7 animals into which was cannulated
10 cc of 10 ppt. drilling fluid and 6 control animals 5 of which received
10 cc of sterilized saline and 1 received 5 cc of sterilized saline. The
animals were monitored for 24 hours prior to drilling fluid (DF) or saline
cannulation and a minimum of 24 hours after cannulation. Interneuron II
46
-------
IN VIVO RECORDING
prox.
dlit
200 jiV
2O sec
Figure 1. IN VIVO RECORDING
A. Schematic of Aplysia showing funnel-shaped
siphon projecting dorsally between tissue
(parapodia). Laminated gill is situated
within the mantle cavity. The abdominal
ganglion is sketched along with a double-
cuff assembly in place on siphon nerve.
Cannulation tube is seen emerging posteriorly
(to left).
B. Schematic of double-cuff electrode assembly.
Left-cross-sectional view of cuff where one
of electrode wires (W) passes through sil-
astic tubing. Opening at bottom represents
trough in tubing to permit placement around
siphon nerve. Right - Longitudinal view
of double cuff assembly. N = nerve, L =
ligature to afix assembly onto nerve, S =
silastic tubing (O.D. = 1 mm). PROX. and
DIST. refer to two independent monopolar
electrode wires proximal and distal, re-
spectively, to abdominal ganglion.
C. Five-minute example of analog double-cuff
recording from siphon nerve of intact, un-
restrained Aplysia. Both electrode records
(prox. and dist.) displayed. Animal was
quiescent but note the great amount of
background spontaneous activity. Arrowhead
indicates the occurrence of a spontaneous
interneuron II (INT II) burst. An INT II
burst consists of two phases: 1) increased
activity from small and intermediate amplitude
units and inhibition of large amplitude units
(dashed horizontal line) and 2) decreased
activity from small and intermediate amplitude
units and rebound firing from large amplitude
units (solid horizontal line).
-------
burst frequencies were compared using the non-parametric Wilcoxon Matched
Pairs, Mann-Whitney U and Spearman Rank correlation tests (.23) at a con-
fidence level of < 5%, two-tailed.
RESULTS
Six of the 7 Aplysia administered 10 ppt drilling fluid (DF) died within
2 days whereas the 6 control animals survived a mean of 8 days following
saline cannulation.
However, prior to death the experimental animals (i.e., DF cannulation)
showed a significant increase (N = 14, U = 8, p = 0.038) in the number of
Interneuron II (INT II) neuronal bursts per hour following DF cannulation
(X = 11.97 ± 13.29) compared to^ the INT II burst frequency over the 24-hour
period before DF cannulation (X = 4.94 ± 11.17). No significant change
(N = 12, U = 17,__p = 0.93) in INT II frequency was seen in the control
animals before (X = 4.75/hr ± 6.79) vs. after (X = 4.96/hr ± 5.54) saline
cannulation. The pre-cannulation and post-cannulation INT II burst frequencies
(i.e., number per hour) for individual control and experimental animals is
shown in Table I and Figure 2. Experimental animals showed a consistent
and highly significant increase in INT II rate following DF cannulation.
Control animals demonstrated inconsistent or insignificant changes in INT II
rate after saline cannulation. Figure 2B also shows that the mean post-
cannulation INT II rate seen in individual animals is correlated with the
mean pre-cannulation rate (N = 7, rg = 0.92, p < 0.02). Thus the animals
with the highest, intermediate, or lowest ongoing INT II rate before intro-
duction of DF also displayed the highest, intermediate, or lowest INT II
rate, respectively, after administration of DF. An example of the hourly
INT II burst rate for a control animal, CC 24 (Figure 3A), and an experi-
mental animal, CC 19 (Figure 3B), before vs. after cannulation illustrates
the large increase elicited by the presentation of DF compared to saline.
Figure 3B shows that DF triggers an immediate increase in INT II bursts over
the first 3 hours following cannulation compared to the ongoing mean INT II
frequency of 9.0 ± 6.76 over the 24-hours before cannulation. In contrast,
Figure 3A shows that the INT II rate immediately after saline cannulation
decreased slightly compared to the ongoing mean rate of 5.32 ± 6.35 over the
24-hours prior to introducing saline. For experimental animal CC 19 (Figure
3B), the immediate DF-elicited increase in INT II frequency was followed by
a decline over the next 6 hours (until 2200 hours) after which the INT II
rate rose again and remained elevated for the remainder of the experiment.
Such a time-course of DF effect on INT II burst rate (i.e., initial increase
followed by a decrease, then a sustained increase) was seen in all experi-
mental animals. No such substantial and significant increase in INT II
rate was seen for the control animals.
In addition to the change induced by DF in the rate of INT II bursts,
DF also resulted in a qualitative change in over-all siphon nerve activity.
Figures 4 and 5 show 5-minute samples of neuronal activity from repre-
sentative control (CC 18) and experimental (.CC 19) animals respectively.
Part A of Figures 4 and 5 give pre-injection (i.e., cannulation) neuronal
48
-------
TABLE I. INTERNEURON II FREQUENCIES FOR CONTROL (SALINE) AND EXPERIMENTAL
(10 PPT. DF) ANIMALS BEFORE VS. AFTER CANNULATION
Control
Animals
Experimental
animals
X frequency (No./hr)
before after
N*
cc
cc
cc
cc
cc
cc
7
13
18
24
26
27
2.
4.
8.
5.
2.
4.
0 ±
46 ±
34 ±
32 ±
64 ±
48 ±
1.6
4.36
11.88
6.35
3.52
5.41
1.75 ±
5.13 ±
9.43 ±
2.90 ±
4.39 ±
5.56 ±
0.89
5.93
8.64
4.15
3.89
5.98
51
52
78
77
77
77-
1.
0.
2.
2.
2.
1.
10
83
30
39
53
22
<0
<0
<0
<0
<0
<0
.32
.41
.021
.017
.0.11
.222
CC
CC
CC
CC
CC
CC
CC
11
12
14
19
20
21
23
1.
3.
8.
9.
2.
3.
2.
96 ±
18 ±
68 ±
0 ±
12 ±
54 ±
7 ±
3.07
6.82
11.04
6.76
3.02
4.82
2.61
3
7
20
22
5
17
13
.39 ±
.60 ±
.28 ±
.7 ±
.55 ±
.69 ±
.93 ±
3.51
10.4
12.06
16.21
5.21
12.85
13.41
78
78
46
78
78
53
39
2
3
3
4
4
4
3
.28
.71
.57
.78
.26
.97
.97
<0.023
<0.0002
<0.0005
<0. 00006
<0. 00006
<0. 00006
<0.0001
* ii is the combined number of observations before and after cannulation
t Z is the statistical value computed from the Mann-Whitney test
fy p is the calculated significance, two tailed; p <.05 is taken as
significant.
49
-------
A.
PRE
INJECTION
post
INJECTION
ANIMALS (CC NUMBERS)
B.
Figure 2.
A. Histogram of mean number of INT II bursts
per hour before (unhatched) and after
(hatched) cannulation of 5 cc or 10 cc of
sterilized saline into six control animals
(CC 7, CC 26, CC 13, CC 27, CC 24, CC 18).
Animals were ranked along abscissa accord-
ing to their mean pre-cannulation INT II
burst rate. Post-cannulation INT II rate
showed no tendency toward a correlation
N = 6, rs = 0.34, p < 0.30) with the
respective pre-cannulation rate.
B. Histogram of mean INT Il's/hr pre- and post-
DF cannulation for seven experimental
animals (CC 11, 20, 23, 12, 21, 14, 19.)
ranked according to their pre-cannulation
mean INT II rate. Pre-DF cannulation rate
was positively correlated (N = 7, r = 0.96,
p <0.02) with the respective post-DF
cannulation rate.
en
o
ANIMALS (CC NUMBERS)
-------
CC 24 SALINE
s i i I I I
CENTRAL OATLICHT TIME IMHS.)
I 8
i i !
CENTflAL OATULCHT T]rt£ (MRS.)
Figure 3.
A. Hourly histogram for a saline-control animal (CC 24) of number of INT II
bursts over the 77-hour course of the experiment. Hatching (beginning
at 1300 hours) represents the post-saline cannulation period. The INT
II rate actually decreased for this animal following saline cannulation.
B. Hourly histogram for a DF experimental animal (CC 19) showing changes
in the number of INT II bursts during the 77 hrs. of the experiment.
Immediately after DF cannulation (i.e., within 40 sec.) the INT II rate
increased significantly over the ongoing rate before cannulation; this
increase lasted for 3 hrs. (until 1600 hrs.), declined and then returned
to a significantly higher level (2200 hrs.) for the remainder of the
experiment.
51
-------
A. Before Infection
SALINE CANNULATION
•*•
JO MRS.
B. After Injection
prox.
24 HRS.
1OOpV
Figure 4. SALINE CANNULATION
A. Before Injection - Five-minute samples of siphon nerve activity
mal cuff electrode only) from a control animal (CC 18). 0 HRS..(1300
hrs CDT), 2 HRS. and 20 HRS. represent times after experiment began.
Experiments were initiated one day after surgery for implantation of
the double-cuff electrode assembly and cannulation tube. The star in
the top record (0 HRS.) indicates an INT II burst. Characteristically,
variability was seen in the general level of neuronal activity across
the three samples.
B. After Injection - Four samples (proximal cuff record only) of siphon
nerve activity over a 48-hr, period following cannulation (vertical
arrow) of 10 cc of sterilized isotonic sea water into the body cavity of
animal CC 18. Amplitudes of units were unchanged compared to samples
in part A and the duration and pattern of INT II bursts (e.g., that _
"starred" in the 2 HRS. sample) showed no modification following admini-
stration of saline. The animal appeared healthy (based on the criteria
of parapodia closed tightly and animal attached by foot to substrate)
throughout the course of the experiment.
52
-------
DRILLING FLUID CANNULATION
A. B«for« Injactlon
prox.
0 HRS.
3 HRS.
30 HRS.
B. AM»r Injection
pro*.
4* MM.
100 w
Figure 5. DRILLING FLUID CANNULATION
A. Before Injection - Three 5-min. samples of spontaneous siphon nerve
(proximal electrode record only) activity from animal CC 19 prior to
cannulation of drilling fluid. 0 HRS. corresponds to 1300 hrs., CDT.
Variability was seen in the level of background spontaneous activity.
No INT II bursts occurred during these sampling periods. Animal
appeared healthy throughout this period.
B. After Injection - Four 5-min. post-DF cannulation samples of neuronal
activity from animal CC 19. Following DF (10 ppt. wet weight) cannu-
lation (vertical arrow in top sample) at a latency of approximately 40
sec., 7 INT II bursts were triggered within 5-mins. The duration
(approximately 15 sec.) of the second triggered INT burst (star over
first phase of burst in all samples) is indicated by a double-ended
arrow. The level of background activity was also heightened. The
second 5-min. sample (2 HRS., after DF cannulation) showed a continued
high frequency of INT II bursts and heightened background activity.
One day later (24 HRS.), background activity had decreased although the
INT II burst frequency was still high; the duration of INT II bursts had
now doubled to approximately 30 sec. ("starred" INT II burst with double-
ended arrpw). Two days (48 HRS.) following DF cannulation, unit ampli-
tudes had decreased slightly and INT II burst durations had increased
to 80-100 sec. Prolongation of both phases of the INT II burst charac-
terized its extended time-course. By 48-hrs. post-DF cannulation, the
animal's parapodia were not closed tightly; substrate attachment was
still good.
-------
samples from the 24-hours before cannulation while Part B of each figure
presents post-cannulation neuronal samples. In Figure 4A, the first sample
(0 HRS) shows a spontaneous INT II burst. The first sample after saline
cannulation (0 HRS in Figure 4B) contains a single INT II burst as do the
second and fourth samples (2 HRS and 48 HRS., respectively). Overall,
Figure 4 illustrates no qualitative changes in neuronal activity over the
pre-injection and post-injection period. Furthermore, the pattern of the INT
II bursts (e.g., overall duration and duration of the two phases of the
INT II burst) in the control animal is unchanged throughout the experiment;
the amplitude of the units also showed no change across the 77 hours of the
experiment. In contrast, DF elicited a qualitative increase in the level of
siphon nerve activity, (Figure 5A vs. 5B) which was high-lighted by the en-
hanced rate of INT II bursts (see 0 HRS., 2 HRS., and 24 HRS., in Figure
5B). In the presence of DF, the neuronal pattern of the INT II burst was
also seen to change (compare "starred" INT II burst in Figure.5B for 0 HRS.,
24 HRS., and 48 HRS.).
The modification of the neuronal pattern of INT II bursts following
cannulation of DF is illustrated in Figure 6 which displays (at a faster
time-base) representative examples of INT II bursts for 3 control and 3
experimental animals. The INT II bursts for the control animals are
basically unchanged throughout the course of the 77-hour experiments (Figure
6A); however, in the experimental animals INT II bursts changed substantially
following DF cannulation (Figure 6B). Specifically, by 24 HRS., post-DF
cannulation, the duration of both the first and second phase (cf. Figure 1C)
of the INT II burst have increased in two of the animals. The typical pre-DF
cannulation INT II bursts for each of the experimental animals was 10-15
seconds in duration. Twenty-four hours after administering DF, the duration
had increased to approximately 35 seconds and 20 seconds for animals CC 14
and CC 19, respectively; and by 48 HRS. INT II bursts for CC 19 were typically
50-60 seconds in duration. The INT II burst pattern for experimental animal
CC 21 in Figure 6B did not show the typical duration increase illustrated by
animals CC 14 and CC 19; but the burst pattern did change in that by 24 HRS.,
post-DF cannulation, the large amplitude rebound units were not seen.
Therefore, the experiments demonstrate that 10 ppt DF: 1) is lethal
(usually within 48 hours) to adult Aplysia californica, 2) effects a change
in siphon nerve activity most notably a significant increase in the rate of
INT II bursts, and 3) results in a change in the neuronal pattern of the
INT II burst that precedes the first observable behavioral changes.
DISCUSSION
The Interneuron II burst seen in the siphon nerve is composed of activity
from a population of motoneurons in the abdominal ganglion with processing
in the siphon nerve that innervate the muscles of the siphon. (24, 18).
Interneuron II has been shown to be a network of cells in the abdominal
ganglion consisting of one or more burst-generator cells which are presynaptic
on premotor cells which are in turn presumably presynaptic on siphon and gill
motoneurons (22). The behavioral response (i.e., gill, parapodial and siphon
54
-------
INTERNEURON II
A. SALINE Before In).
CC18
pros.
CC27
prox.
100 ||V
CC 28
After Inj.
14 MRS.
pros,
CC21
dlst.
After Inj.
ie MRS.
MHUS.
94 HRS.
48 MRS.
I I
Figure 6. INTERNEURON II
Comparative illustration of time-course of INT II burst at faster time-sweep.
A. SALINE - INT II burst patterns and durations for control animals CC 18,
CC 27 and CC 26 Before Inj. (injection) vs. After Inj. (injection) were
within the range of variability typically seen in INT II bursts over
time. Total burst durations were approximately 15-25 sec. consisting
of phase one (approximately 2-5 sec.) and phase two (approximately 10-
15 sec.).
B. DF - Three examples (from experimental animals CC 14, CC 19, CC 21) of
INT II bursts were compared before (Before Inj.) and After Inj.) DF
cannulation. Throughout the post-cannulation monitoring period INT II
bursts became distorted. For CC 14 (16 HRS.) and CC 19 (24 HRS. and
48 HRS.), the INT II burst pattern was prolonged and each phase showed
a considerably lower level of activity than before DF was presented.
The change in INT II bursts in CC 21 was characterized by a slightly
decreased overall duration and absence of rebound firing of large
amplitude units in phase two.
55
-------
contraction) associated with INT II activity is believed to function in
Aplysia respiration and excretion (21). Thus, by recording INT II rate and
neuronal changes in INT II burst patterns, it is possible to monitor the
status of one or more physiological systems of primary importance to an
animal.
Kanz et al. (18) and Eberly et al. (19) have reported other types of
exogenous stimuli that influence the rate of INT II activity. For example,
tactile siphon stimulation triggers short-latency (< 5 seconds) INT II bursts
(18). Light and food presentation resulted in a longer-term (10-20 minutes)
increase in INT II activity (19); and the increase in INT II activity
following light presentation was found to be correlated with the INT II rate
prior to' lights-on. The results of this study show that DF induces a
longer-lasting increase in INT II rate than tactile siphon stimulation,
lights-on or food presentation. And significantly, a positive correlation
was seen between the pre-DF and post-DF cannulation INT II burst frequencies.
This suggest that DF, like the lights-on stimulus, may be effecting the
INT II network at the level of the burst-generator rather than premotor cells.
Typically, the greatest and longest-lasting increase in INT II rate
began approximately 5 hours after DF cannulation (see Figure 3B).. A short-
term rate increase was also seen within the first hours after DF cannulation
(Figure 3B) but characteristically with a latency of 30-40 seconds (see
Figure 5B, 0 HRS.). Thus, the animals appeared to "detect" (using INT II
activity as the criterion) the presence of the DF soon after its cannulation
but did not respond systemically for several hours after cannulation.
In summary, our results indicate that drilling fluid cannulation
into the body cavity of Aplysia has a significant effect on neuronal activity
that is believed to be directly involved in the animal's respiratory/
excretory system; and this effect is manifested hours before observable
deterioration in the animals condition.
We are currently bringing on-line computer procedures (25) for quanti-
fying the firing pattern of individual units making up the analog spike
train (e.g., as that seen in Figures 1C, 4, 5 and 6). This will make it
possible to determine even subtler neuronal changes that might result from
environmental toxins.
56
-------
REFERENCES
1. Sorokin, Yu. I. 1973. Microbiological aspects of the productivity of
coral reefs. In: Biology and Geology of Coral Reefs, pp. 17-45.
' 0. A. Jones and R. Endean (eds.). Academic Press. New York.
2. Mileikovsky, S. A. 1970. The influence of pollution of pelagic larvae
of bottom invertebrates in marine nearshore and estuarine waters.
Mar. Biol. £: 350-356.
3. Lee, R. F., R. Sauerheber and A. A. Benson. 1972. Petroleum hydro-
carbons: Uptake and discharge by the marine mussel. Mytilus
edulis. Science 177:344-346.
4. Carr, R. S., and D. J. Reish. 1977. The effects of petroleum hydro-
carbons on the survival and life history of polychaetous annelids.
In: Fate and Effects of Petroleum Hydrocarbons on Marine
Ecosystems and Organisms, pp. 168-173. D. A. Wolfe (ed.).
Pergamon Press, New York.
5. Anderson, J. W., J. M. Neff, B. A. Cox and H. E. Hightower. 1974.
Characteristics of dispersions and water soluble extracts of crude
and refined oils and their toxicity to estuarine crustaceans and
fish. Mar. Biol. 2]_:75-88. .
6. Hawkes, J. W. 1977. The effect of petroleum hydrocarbon exposure on
the structure of fish tissues. In; Fate and Effects of Petroleum
Hydrocarbons in Marine Ecosystems and Organisms, pp. 115-128.
D. A. Wolfe-(ed.). Pergamon Press, New York.
7. Sabo, D. J., and J. J. Stegeman. 1977. Some metabolic effects of
petroleum hydrocarbons in marine fish. In: Pollution and
Physiology of Marine Organisms. Vol. II. A. Calabrese and J. F.
Vernberg (eds.). Academic Press, New York.
8. Ray, J. P., and E. A. Shinn. 1975. Environmental effects of drilling
mud and cuttings. In EPA-sponsored conference: Environmental
Aspects of Chemical Use in Well-drilling Operations. Houston, Texas,
9. Richards-, N. L. 1977. Effects of chemicals used in offshore well-
drilling operations. Program review proceedings of: Environmental
Effects of Energy Related Activities on Marine/Estuarine Ecosystems
EPA Report.
57
-------
10. Brannon, A. C. and K. Ranga Rao. 1979. Barium, strontium and calcium
levels in the exoskeleton, hepatopancreas and abdominal muscle of
the grass shrimp, Palaemonetes pugio : Relation to molting and
exposure to barite. Comp. Biochem. Physiol. 63; 261-274.
11. Tagatz, M. E. , J. M. Ivey, H. K. Lehman and J. L. Oglesby. 1978.
Effects of a lignosulfonate-type drilling mud on development of
experimental estuarine macrobenthic communities. Northeast Gulf
Science 2^35-42.
12. Tagatz, M. E. , and M. Tobia. 1978. Effects of barite (63804) on
development of estuarine communities. Estuarine Coastal Mar. Sci.
7_: 401-407.
13. Sprague, J. B. , and W. S. Logan. 1979. Separate and joint toxicity
to rainbow trout of substances used in drilling fluids for oil
exploration. Environ. Pollut. 269-281.
14. Crawford, R. B., and J. D. Gates. 1981. Effects of a drilling fluid
on the development of a Teleost and an Echinoderm. Bull. Environ.
Contamin. Toxicol. 26; 207-212.
15. Gray, G. R. 1970. Where the industry's mud money goes. Oil and Gas
£8:157-159.
16. Monaghan, P. H., C. D. McAuliffe and F. T. Weiss. 1976. Environmental
aspects of drilling muds and cuttings from oil and gas extraction
operations in offshore and coastal waters. Report prepared for
EPA Offshore Operations Committee.
17. Cobbs, J., and H. M. Pinsker. 1978. In vivo responses of paired
giant mechanoreceptor neurons in Aplysia abdominal ganglion.
J. Neurobiol.
18. Kanz, J. E. , L. B. Eberly, J. S. Cobbs and H. M. Pinsker. 1979.
Neuronal correlates of siphon withdrawal in freely behaving
Aplysia. J. Neurophysiol. 42; 1538-1556.
19. Eberly, L., J. Kanz, C. Taylor and H. Pinsker. 1981. Environmental
modulation of a central pattern generator in freely behaving
Aplysia. Behavioral and Neural Biol. (in press) .
20. Pinsker, H. M. 1980. Neuroethological analysis o.'f information
processing during behavior. In; Information Processing in the
Nervous System, pp. 285-312. H. M. Pinsker and W. D. Willis Jr.
(eds.). Raven Press, New York.
21. Kandel, E. R. 1976. Cellular Basis of Behavior; An Introduction to
Behavioral Neurobiology. W. H. Freeman and Co., San Francisco.
58
-------
22. Byrne, J., and J. Koester. 1978. Respiratory pumping: Neuronal
control of a centrally commanded behavior in Aplysia. Brain
Res. 143:87-105.
23. Siegel, S. 1956. Nonparameteric Statistics. McGraw-Hill Co., New
York.
24. Perlman, A. J. 1979. Central and peripheral control of the siphon
withdrawal reflex in Aplysia californica. J. Neurophysiol.
42:510-529.
59
-------
Comments on
Drilling Fluid Induced Changes in
Neuronal Activity Monitored in an
Intact, Behaving Marine Animal
Kenneth E. Biesinger
ERL-Duluth
The above grant was recommended for funding primarily because using
an electrophysiological technique to monitor neuronal response to stresses
in an intact animal may provide a sensitive technique to assess toxicity
to Aplysia and other organisms from drilling fluids and other pollutants.
The respiratory pumping frequency, kidney efflux and locomotion behavior
of Aplysia as affected by sublethal concentration of drilling fluid is
to be determined. In addition, the influence of this potential toxicant
on the activity of the siphon and pericardial nerves, the output from
the network of cells in the abdominal ganglion responsible for respiratory
pumping behavior> and the activity of abdominal ganglion cells capable
of endogeneous bursting on neuronal activity will be studied.
Aplysia inhabit areas likely to be impacted by drilling fluids and
other pollutants, have an accessible model nervous system from which
neurobiological principles can be studied and extrapolated to other
organisms and may be.useful for predicting adverse effects to marine
benthic organisms. The proposed work, although thought to have a high
risk, was also thought to have a high gain potential. The ability to
monitor neuronal output ir± situ from intact animals might prove to be a
, sensitive tool to detect sublethal effects.
Some of the criticism of the proposed work included the high
concentrations of drilling fluids selected for the initial experiments
and no intent to measure effects from individual components (i.e.,
barium sulfate, bentonite clay, chromium lignosulfonate, etc.) in the
drilling fluids. The reviewers were also concerned about how neuronal-
behavioral responses in the laboratory could be related to adverse
effects in the marine environment.
60
-------
QUALITATIVE ANALYSIS IN THE SETTING OF EAST COAST MARINE BENTHOS
Charles J. Puccia and William A. Bennett
Department of Population Sciences
School of Public Health
Harvard University
665 Huntington Avenue
Boston, Massachusetts 02115
ABSTRACT
Soft sediment benthic ecology is described and forms the basis of six
general, core qualitative models. Loop analysis (1) of these models indicates
they provide robust predictions that are insensitive to the details of the
models. The effect of predators and the use of specialized predators on three
functional groups of benthic organisms do produce differences in predictions
among the models. Seasonal variation is important to the benthos. The spe-
cial considerations necessary for benthic environmental assessment are ex-
plained by the loop model analysis.
INTRODUCTION
We are examining benthic ecosystems along the western Atlantic from St.
Margaret's Bay, Nova Scotia to Chesapeake Bay. Despite latitudinal ranges
which produce differences in salinity, temperature ranges, tidal variation,
sediment composition, etc., there is, on the whole, a similar benthic commu-
nity. Although species composition does change along this gradient, with the
region north/south of Cape Cod, Massachusetts frequently cited as the divide
point, this difference is marginally important when considering species aggre-
gated into functional groups. Consequently, we are presenting a caricature of
the benthos that will illustrate the special considerations for the western
Atlantic. At any specific location, a detailed description of the benthic com-
munity is necessary.
Our theoretical models highlight the benthic organism interactions likely
to give counter-intuitive results. It alerts us to look for the surprises in
environmental effects. The general models are not intended to surplant local
information and detailed measurements. Rather, it tells us the kinds of local
information that can aid environmental assessment, which measurements must be
made, and which can be ignored because they don't change the predicted outcome.
61
-------
Our general benthic models do include particular data, firstt the models
are restricted to the soft-bottom sediments: sand, sand-mud and mud. Second,
the models consider patches within an estuarine bay, tidal flat, or a small
part of a community along the shallow continental shelf. Thus, Buzzard's Bay
in southeastern Massachusetts is about 19.5 kilometers wide and 46 kilometers
long. Within this area are patches of "identifiable" dominant organisms. For
example, Sanders (2, 3) identifies especially dominant burrowing polycheate-
small bivalve areas as Nephtys-Nucula communities. Mills (4) finds areas in
nearby Barnstable Harbor dominated by Ampelisca, a tube-building amphipod.
Any benthic ecosystem will contain a variety of patches with an individual
patch expanding, shrinking or becoming extinct (5). • The consequence of scale,
i.e., models for patches versus entire bays, will be considered below espe-
cially in the effect on community or subcommunity stability.
BENTHIC MODELS
Soft-bottom benthic organisms can be divided into five functional groups
(6) according to the way in which organisms alter the environment. Through
sediment mediation, benthic animals may negatively effect one another. Over-
simplified, there are organisms which bind together particles, hence, stabi-
lize the soft sediment substrate versus those which disrupt or resuspend
material into the water column, hence, disturbing sediments. Organisms which
bind sediments are primarily tube-building annelids and crustaceans. Many
vascular plants or submerged aquatic vegetation with complex root systems can
be classified as sediment stabilizers. Destabilizers include all mobile bur-
rowing animals like bivalves, gastropods and many polychaete worms which move
either just under the sediment surface or forage deep in the sediments.
Model 1 (Figure 1) represents this simple view of benthic ecology. Double
negative links between the two functional groups indicate a competitive re-
lationship. Tube-builders construct vast arrays of tubes, resembling a mat.
Mats inhibit or prevent foraging of the mobile-burrowers (3, 7). Constant
sediment reworking by mobile burrowers inhibits the ability of tube-builders
to construct tubes or to get established in an area (3, 7, 8).
Densities of both tube-builders and mobile-burrowers effect their own
population growth rate, hence, are self-damped (1). The pool of baywide lar-
val production is independent of adult populations in the patch. But larval
settlement is necessary for survival of either functional group in the patch.
This input of larvae is the reason for self-damping. See (1) for mathematical
explanation. Tube-builders are also self-damped because of migration or in-
creased mortality as a consequence of increasing density.
The next four models increase in complexity over Model 1 by the addition
of the functional group filter-feeders. Filter-feeders are composed of
filter-feeding bivalves such as Mulinia laterialis, Gemma gemma or Mya
arenaria. Filter-feeders interact only with tube-builders in Model 2. Models
3 through 5 differ from Model 2 by including alternative interactions of
filter-feeders with mobile-burrowers. Assumptions about all the models are
listed in Table 1.
62
-------
TABLE 1. MODEL ASSUMPTION LIST
Model
Assumption
First Assump-
Occurs tion No.
Description of Assumption
Subsystem TB-MB is unstable when part of a larger com-
munity that has filter feeders present. Mills (4)
showed that mobile burrowers, Nassarius, did not occur
with tube builder, Ampelisca.
Subsystem FF-TB is stable. Mills (4), Watling (14)
have shown that Gemma gemma increases when Ampelisca
is present, but not with mobile burrowers.
(-) effect of DF on FF times self damping on TB is
greater than (+) effect of MB on FF through TB. In
the mud environment, this corresponds to the trophic
amensalism hypothesis (8).
(-) effect of MB on TB times self damping on FF is
greater than (+) effect of MB on TB through FF. In
mud areas direct negative effects of disturbance on
TB by MB times strong self damping on FF would be
stronger than the strong effect of MB on FF times the
relatively weak effect of FF on TB due to larval
filtration. TB larva not very succeptible to filtra-
tion in mud sediment with a low density of suspension
feeders.
Assumption 1 plus the assumptions that (-) effect of
FF on both TB and MB are equal. In sand, filter
feeders can occur in dense assemblages and probably
have an equal effect on TB and MB through filtering
of larvae.
Assumptions 1 and 2 combined: the (-) effect of TB
on MB times self damping on FF is greater than (-)
effect of TB on FF times (-) effect of FF on MB.
Extended, this assumption also means that because the
(-) effect of FF on MB and FF on TB are equal (assump-
tion 1), and the effect of FF on TB is less than the
self dampong on TB (assumption 2) and because the self
damping on TB is less than the effect of TB on MB; the
63
-------
TABLE 1 CONTINUED
Model
Assumption
First Assump-
Occurs tion No.
Description of Assumption
effect of TB on MB is greater than the effect of FF
on MB. Also, the effect of TB on MB is composed of
adult-adult and adult-larval interactions, whereas
the effect of FF on MB is composed solely of an
adult-larval interaction.
Note that the (-) effect of DF on FF is one way due
to trophic amensalism hypothesis; (-) effect of FF on
MB is adult-larval filtration.
Assumes the positive feedback subsystem of MB and
"Free mud" is a stable system.
Needs to be verified; results under this assumption
could not be clearly seen with information at hand.
Interactions need to be relatively quantified.
10 The (+) effect of "Free sand" on TB times the self
damping on FF is greater than the (=) effect of "Free
sand" on FF times the (-) effect of FF on TB. If we
assume sand helps TB and FF equally, we can see that
in a mud patch, FF will not have a strong siphoning
effect on TB, at least not as strong as rate of input
of FF to patch (i.e., self damping on FF).
11 The (-) effect of TB on MB times the (-) effect of
MB on FF (i.e., the positive effect of TB on FF
through MB) is greater than the direct (-) effect of
TB on FF times the self damping effects of MB on
itself. Link TB on MB is assumed strong relative to
self damping on MB, and the trophic amensalism effect
of MB on FF is probably stronger than the adult-
larval-mat interaction between TB and FF.
64
-------
Figure 1. Model 1, two benthic functional groups.
TABLE 2. MODEL 1 PREDICTIONS ON DIRECTION OF CHANGE
IN ABUNDANCE THROUGH CHANGE IN RATE OF GROWTH
(+) Input to
Effect on
Tube Builders Mobile Burrowers
Tube
Builders
Mobile
Burrowers
65
-------
Figure 2. Model 2, an expanded version of model 1. Filter-
feeder functional group in competition with
tube-builders.
TABLE 3. MODEL 2 PREDICTION ON DIRECTION OF CHANGE
IN ABUNDANCE THROUGH CHANGE IN GROWTH RATE.
SIGNS IN PARENTHESES INDICATE ASSUMPTIONS
"ERE MADE ABOUT RELATIVE MAGNITUDE 0? LIHKS.
ASSUMPTIONS ARE SUMMARIZED BY SUBSCRIPT IN TABLE 1.
Effect on
(+) Input to
Filter
Feeders
Tube
Builders
Mobile
Burrowers
Filter Tube Mobile
Feeders Builders Burrowers
<->!
-
+
-
+
-
+
-
<+>2
66
-------
Model 2 (Figure 2) has interference between tube-builders and mobile-
burrowers as in Model 1. In addition to the reasons for interference stated
previously, Woodin (9) adds that settling larvae are consummed by tube-
builders or are buried in tube-builders fecal material. Thus, tube-builders
negatively effect mobile-burrowers and filter-feeders. On the other hand,
settling tube-builder larvae are unable to construct tubes, often become
buried in the very flocculent upper sediment layer, and are possibly consumed
by foraging burrowers.
In dense assemblages, filter-feeders can effectively siphon the majority
of larva (even their own) attempting to settle in their area (9). Therefore,
filter-feeders negatively effect tube-builders.- Because of a high rate of
larval recruitment from outside the patch, filter-feeders are self-damped;
density-dependent consumption of their own larvae enhances this self-damping.
Sediment composition can be, roughly speaking, either mud, sand, or sand-
mud and the benthic communities of each are represented in Models 3, 4 and 5,
respectively.
In muddy areas, mobile-burrowers rework the sediments, resuspending the
upper layers of the sediment into the water column. Floculent material inter-
feres with the respiratory and feeding mechanisms of filter-feeders (8).
Thus, mobile-burrowers will negatively effect filter-feeders, and this is
depicted in Model 3 (Figure 3).
A sandy area, as represented in Model 4 (Figure 4), reverses the negative
link of Model 3; now, it is from filter-feeders to mobile-burrowers. Sand is
not easily resuspended into the water column, but it is on excellent substrate
for and is often found with dense assemblages of filter-feeders. In high
abundance, filter-feeders may increase mortality on mobile-burrowers1 settling
larvae (9).
At a mixture of sand-mud, the filter feeders and mobile-burrowers are
true competitors negatively effecting each other. Thus, Model 5 (Figure 5)
has a negative pair of links between all the functional groups.
Model 6 (Figure 6) distinguishes direct adult-larvae interactions from
sediment mediated interactions. Restricted to a mud sediment, as in Model 3,
we include two sediment variables. Free-mud and free-sand represent the
percentage of mud and sand not amalgamated into a tube or tube mat. Free-mud
benefits mobile-burrowers, facilitating foraging. The mobile-burrowers dis-
rupt tubes hence "freeing" mud; .mobile-burrowers and free-mud have paired
mutualistic links. The negative link from mobile-burrowers to tube-builders
is larval mortality, unlike Model 3 and 5 where it includes adult-adult
interactions via sediment disturbance.
Tube-builders cannot construct durable tubes in mud alone. Tube-mat
density is highest in sand-mud environments, increasing sharply in mostly mud
habitats (2). Therefore, free-mud inhibits tube-builders, while free-sand
assists them. Tube construction, of course, decreases free-mud and free-sand.
Tube mats, although no different in particle composition than the original
substrate, nevertheless alter sediment physical characteristics, changing
67
-------
Figure 3. Model 3, the same as model. 2 but includes negative
interaction of mobile-burrowers on filter feeders.
TABLE 4. MODEL 3 PREDICTIONS ON DIRECTION OF CHANGE
IN ABUNDANCE THROUGH CHANGE IN RATE OF GROWTH.
SEE TABLE 3' CAPTION FOR EXPLANATION OF
PARENTHESIS SIGNS AND SUBSCRIPTS
Effect on
(+) Input to
Filter
Feeders
Tube
Builders
Mobile
Burrowers
Filter Tube Mobile
Feeders Builders Burrowers
(->l
?(+)
?(-),
-
+
?(-)2
+
-
(+)2
68
-------
Figure 4. Model 4, with reverse direction from model 3
of negative interaction between filter-feeders
and mobile-burrowers.
TABLE 5. MODEL 4 PREDICTIONS ON DIRECTION OF CHANGE
IN ABUNDANCE THROUGH CHANGE IN RATE OF GROWTH.
SEE TABLE 3 CAPTION FOR EXPLANATION OF
PARENTHESIS SIGNS AND SUBSCRIPTS
Effect on
(+) Input to
Filter
Feeders
Tube
Builders
Mobile
Burrowers
Filter Tube Mobile
Feeders Builders Burrowers
(->l
-
+
(+>5
+
-
<+>5
(->6
(+)2
69
-------
Figure 5. Model 5 includes all pairwise competitive links
between functional groups.
TABLE 6. MODEL 5 PREDICTIONS ON DIRECTION OF CHANGE
IN ABUNDANCE THROUGH CHANGE IN RATE OF GROWTH.
SEE TABLE 3 CAPTION FOR EXPLANATION OF
PARENTHESIS SIGNS AND SUBSCRIPTS
Effect on
(+) Input to
Filter
Feeders
Tube
Builders
Mobile
Burrowers
Filter Tube Mobile
Feeders Builders Burrowers
<->!
?<+>ll
<->!
(+)5
<->7
7
(+)5
<->6
(+)2
70
-------
permeability and hydrodynamics (4, 10). Free-sand can increase filter-feeder
populations either through inducing larval settlement or curtailing mud
resuspension.
Filter-feeders and tube-builders have adult-larvel interactions (9) as
in Model 3. There is, however, a long negative path of tube-builders through
free-sand on filter feeders. Physiological/morphological constraints of sand-
mud substrate on filter-feeders become excentuated by tube-builder activity.
Thus, it may be that benthic organism activity creates patchiness in nearly
homogeneous physical regimes.
As.in Model 3, mobile-burrowers resuspend sediments that clog filter-
feeder apparatus. Model 6 includes direct negative links from mobile-
burrowers to filter-feeders for easy comparison with Model 3. The negative
link should, however, be from free-mud to the filter-feeders. Predictions
will not change if we redraw the negative link because the mobile-burrowers
and free-mud variables are symmetrical in the model. Thus, this model cannot
distinguish between which negative link or if both should be present. If it
seems important to do this, then another model must be developed which lumps
variables according to other criteria such as separating adults from larvae.
MODEL PREDICTIONS
Tube-builders bind sediments while mobile-burrowers disrupt sediments;
these two functional groups are interference competitors. Competition can
lead to potential instability through positive feedback. On the other hand,
negative feedback of the effects on the two organismic types enhances the
likelihood of co-occurrence. Broadly speaking, any soft-sediment patch will
have tube-builders and mobile-burrowers (Sanders, pers. comm.). In the
absence of filter-feeders, another competitive group, we assume the two former
groups are part of a stable community; hence the overall community has nega-
tive feedback. In the presence of the filter-feeders, we reverse this assump-
tion, as will be pointed out in the analysis of our later models. All assump-
tions made are listed in Table 1 and refer to subscripts in all subsequent
tables.
For Model 1, we show that (Table 2) a positive change in the population
growth of tube-builders will increase the tube-building fauna while decreasing
the mobile-burrowing fauna. For short, a positive (or negative) change in the
growth rate of a variable we call a "parameter input." Likewise, a positive
parameter input to mobile-burrowers increases their own abundance and de-
creases that of the tube-builders. This is the stereotypic pure competitive
interaction result.
Model 2 recognizes a third faunal benthic group, the filter-feeders.
This group siphons larvae of both mobile-burrowers and tube-builders, reducing
self-damping especially on mobile-burrowers. However, community stability
requires strong self-damping. The mobile-burrower/tube-builder subsystem is
probably unstable when imbedded in a larger community containing filter-
feeders. Indirect support for this assumption comes from Mills' (4)
71
-------
Figure 6. Model 6 which includes physical characteristics
of the sediment as the variables free-sand and
free-mud.
TABLE 7. MODEL 6 PREDICTIONS ON DIRECTION OF CHANGE
IN ABUNDANCE THROUGH CHANGE IN GROWTH RATE.
SEE TABLE 3 CAPTION FOR EXPLANATION OF PAREN-
THESIS SIGNS AND SUBSCRIPTS
Effect on
TJ
(+) Input to
:o
Filter
Feeders
Tube
Builders
Mobile
Burrowers
Free
Sand
Free
Mud
03
0) 0)
4J »T3
.H (U
•H 0)
(-)
1
?
(+)
7
-a
en g
rH 0 *
•HMO) 8
(+)9
(-)u
?
(+)9
^^O
72
-------
observation that tube-building amphipods of the genus Ampelisca do not coexist
in Barnstable Harbor with the burrowing gastropod Nassarius.
Filter-feeders siphon their own larvae, tube-builders show adult density-
dependent effects, so both probably have strong self-damping. Gemma gemma,
the filter-feeding bivalve, is abundant when tube-building Ampelisca is active
and mobile-burrowers are relatively rare (4). Consequently, we assume the
filter-feeder/tube-builder subsystem is stable.
Because we assume an unstable tube-builder/mobile-burrower subsystem, a
Model 2 prediction (Table 3) is non-intuitive. An environmental change that
increases the growth rate Of_filter-feeders causes a decrease in their own
abundance. Of course, their turn-over rate has increased. A question mark
is placed in the table and a negative sign in parentheses to emphasize that an
assumption is made about relative strengths of links. In a following section,
we describe the kinds of data that are necessary to test the assumption.
The fine mud sediment region of the benthos is described in Model 3.
Mobile-burrowers disturb the mud, clogging filter-feeder siphons (8, 9).
Model 3 predictions (Table 4) agree closely with that for Model 2 (Table 3)
except that three question marks appear. These question marks, unlike that
of Model 1, are not stability assumptions but occur because at least two path-
ways exist from one species group to another with opposite effects. For ex-
ample, a parameter input entering through mobile-burrowers can negatively
effect filter-feeders through a direct path and positively via tube-builders.
As in the first three models, we have calculated predictions for Models
4, 5, and 6 (Tables 5, 6, 7).
All six model predictions are summarized in Table 8. The models all
agree in the effect of a parameter input to tube-builders on mobile-burrower
abundance and the effect of a parameter input on mobile-burrowers on their
own abundance. There is nearly complete agreement, except for Model 5, in
predicting abundance change to tube-builders through a parameter input to
themselves. And, while Model 3 and 5 give ambiguous predictions, all other
models agree in predicting a negative change in tube-builder abundance from
a parameter input to mobile-burrowers. The fact that added model complexity
conforms with a simple model's predictions is reassuring.
The five complex models predict a positive input to filter-feeders caus-
ing a decline in their abundance. This is a consequence of the single assump-
tion that the tube-builder/mobile burrower subsystem, in the presence of
filter-feeders, is unstable. Even if this is untrue, all the models predict
the same result, only reversed from above. To get different results between
these models requires that either (a) the parameter input to filter-feeders is
so strong as to swamp network effects in some patches, but not others, or (b)
there exists an unspecified unstable subcommunity which links only with filter-
feeders and that this is true in some patches and not others, or (c) in some
patches, the tube builder/mobile burrower subcommunity is stable, and in other
patches it is unstable. This last possibility is the most likely, and under-
scores the caution that these general models cannot be applied indiscriminately
for environmental assessment, but must be verified according to local conditions,
73
-------
TABLE 8. SUMMARY OF ALL SIX CORE BENTHIC MODEL PREDICTIONS.
A SIGN IN PARENTHESIS FOLLOWED BY AN * INDICATES
THE MODEL IS FORCED TO THIS PREDICTION. A (?)*
MEANS FORCING CHANGES THE INDICATED PREDICTION TO
AMBIGUOUS PREDICTION
Filter
Feeders
Tube
Builders
Mobile
Burrowers
Model
No.
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
Filter Feeders
n/a
_
_
_
_
—
n/a
-
+ (?)*
-
+ (?)*
?
+
-(?)*
+
-(?)*
+
Tube Builders
n/a
_
_
+
(+)*
_
+
+
+
+
- (?)*
+
-
-
(-)*
-
(-)*
-
Mobile Burrowers
n/a
+
+
+
+(?)*
+
-
-
-
-
-
-
+
+
+
+
+
+
74
-------
Question marks in Model 3 and 5 can be used to make predictions about
community interactions rather than assembling them in counter-point to the
above. We call this "forcing the model" because we presume Models 1, 2, 4,
and 6 give the "correct" result, "force" the question marks to give the same
result, and examine the consequence of doing so. Hence, replacing the ques-
tion marks in the mobile-burrower/tube-builder box with a negative sign
implies the negative effect of mobile-burrowers directly on tube-builders out-
weighs the benefits to the tube-builder by having its filter-feeding compet-
itors reduced. This conclusion is reasonable since we already have indicated
competition between mobile-burrowers and tube-builders is strong and that
filter-feeders have strong density-dependent effects.
BENTHIC PREDATORS
The six core models do not include predators. The major predators on the
benthos include fish, epifaunal Crustacea and some of the benthic infauna.
Benthic predators, especially fish, are commonly assumed to be generalists
because they can and sometimes do eat most taxa. However, in terms of effect-
ing populations, the generalist predator among our functional groups is un-
likely. Winter flounder, Pseudopleuronectes americanus, have been shown to
select size-specific prey (12). In St. Margaret's Bay, Nova Scotia, during
the early spring, the amphipod Photis reinhardi and the polychaete,
Micronephthys minuta, are abundant. Winter flounder, however, consume the
amphipod but not the polychaete. During the summer, as amphipods decline and
the smaller-sized ostracod Nurmanicythese leisderma increases, the flounder
does not switch. Yet, by fall, the flounder does consume the dominant
Micronephthys. Clearly, prey size is the chief selection criteria except in
very low abundances, and this changes over seasons.
Models 1 through 5 are reconsidered with specialized predators for 'each
of the functional groups (Figures 7-11; Tables 9-13). In all cases, the
central predictions on tube-builders/mobile-burrowers does not change. In
other words, the predator does not alter the "network dynamics" for these two
functional groups. However, predators effect themselves and filter-feeders.
The specific prey functional group determines the direction of effect.
Concordance in model predictions for all six models is reassuring. Yet,
it also indicates that the general description of the benthic community
through the set of alternative network connections is inadequate for distin-
guishing between benthic patches. Important information about benthic ecol-
ogy is missing, and the nature of this information will be discussed in the
next section.
DISCUSSION
Mobile-burrowers, tube-builders, and filter-feeders are functional groups
which lump species either in competition with each other or having a negative
effect on at least one other group. Lumping goes beyond taxonomic similarity.
75
-------
(a)
(b)
Figure 7. Model 7, which has model 1 as the core and
specialist predator on (a) the tube-builder
or (b) mobile-burrower.
76
-------
TABLE 9. MODEL 7a,b PREDICTIONS ON DIRECTION OF CHANGE
IN ABUNDANCE THROUGH CHANGE IN RATE OF GROWTH.
SEE TABLE 3 FOR EXPLANATION OF PARENTHESIS
SIGNS AND SUBSCRIPTS
Filter
Feeders
Tube .
Builders
Specialist
Predators
SP 1
SP 2
Tube
Builders
Mobile
Burrowers
Specialist
Predators
SP 1 / SP 2
77
-------
(b)
V [Specialist!
(c)
Figure 8. Model 8, which has model 2 as the core and
specialist predator on (a) the tube-builders,
(b) the mobile-burrowers and (c) filter-feeders.
73
-------
TABLE 10.. MODEL 8a,b,c PREDICTIONS ON DIRECTION OF CHANGE IN
ABUNDANCE THROUGH CHANGE IN RATE OF GROWTH. SEE
TABLE 3 CAPTION FOR EXPLANATION OF PARENTHESIS SIGNS
AND SUBSCRIPTS
Filter
Feeders
Tube
Builders
Mobile
Burrowers
Specialist
Predators -
SP1/SP2/SP3
Filter
Feeders
Tube
Builders
Mobile
Eurrowers
Specialist
Predators
SP 1/SP2/SP3
79
-------
(c)
Figure 9. Model 9, which has model 3 as the core and specialist
predator on (a) the tube-builders, (b) the mobile-
burrowers and (c) filter-feeders.
80
-------
TABLE 11. MODEL 9a,b,c PREDICTIONS ON DIRECTION OF CHANGE IN
ABUNDANCE THROUGH CHANGE IN RATE OF GROWTH.. SEE
TABLE 3 CAPTION FOR EXPLANATION OF PARENTHESIS SIGNS
AND SUBSCRIPTS
Filter
Feeders
Tube
Builders
Mobile
Burrowers
Filter
Feeders
Tube
Builders
Specialist
Mobile Predators
Eurrowers SP1/SP2/SP3
Specialist
Predators
SP1/SP2/SP3
(?)*
(?)*
(?)*
(-)*
(-)*
(-)*
(+)*
(?)*
(-)*
•'2
* Forced to negative
81
-------
Figure 10. Model 10, which has model 4 as,the core and specialist
predator on (a) the tube-builder, (b) the mobile-burrower
and (c) filter-feeder.
82
-------
TABLE 12. MODEL 10a,b,c PREDICTIONS ON DIRECTION OF CHANGE IN
ABUNDANCE THROUGH CHANGE IN RATE OF GROWTH. SEE TABLE 3
CAPTION FOR EXPLANATION OF PARENTHESIS SIGNS AND SUBSCRIPTS
Filter
Feeders
Tube
Builders
Specialist
Mobile Predators
Eurrowers SP 1/SP2/SP3
Filter
Feeders
Tube
Builders
Mobile
Burrowers
Specialist
Predators
SP1/SP2/SP3
83
-------
(c)
Figure 11. Model 11, which uses model 5 as the core and
specialist predator on (a) the tube-builder,
(b) the mobile-predator and (c) the filter-feeder.
84
-------
TABLE 13. MODEL 1la,brc.-PREDICTIONS ON DIRECTION -OF CHANGE IN
ABUNDANCE THROUGH CHANGE IN RATE OF GROWTH. SEE TABLE 3
CAPTION FOR EXPLANATION OF PARENTHESIS SIGNS AND SUBSCRIPTS
Filter
Feeders
Tube
Builders
Mobile
Burrowers
Filter
Feeders
Tube
Builders
Mobile
Eurrowers
Specialist
Predators
SP 1 SP 2
Specialist
Predators
SP1/SP2/SP3
(+Uv
'5,*
(?)*
(?)*
(?)*
(?)*
(?)*
(?)*
(?)*
'(+),
'5,1]
?)*"
(-)*
(-)*
(-)*
(+)*
'5,11
(?)*
(-)*
85
-------
The burrowing polychaete Nephtys incisa is grouped with the small mobile-
burrowing bivalve Nucula proxima.
Scale of applicability can change a model and results. Local patches
will receive recruitment of settling larvae from a pool of larvae that are
produced by adults outside the patch. In a large estuary or enclosed limited-
interchange bay, recruitment is from larvae produced by adults within the
bounds of the region. In the latter case, self-damping may be removed or
diminished in magnitude. When strong self-damping is scale-dependent and
necessary for stability as in the filter-feeder/tube-builder subsystem, then
increasing scale from a patch to the bay can bring on subsystem instability.
This not only effects the subsystem but changes the effect of parameter in-
puts to other, components of the ecosystem.
Seasonal variation in predator selection, settling of larvae, and
strength of composition can effect environmental assessment. For a specific
time of year, different models may apply. A mud patch in spring has winter
flounder consuming the tube-building amphipod, Photis. Any input to winter
flounder during this time will cause an increase in mobile-burrowers and de-
crease in tube-builders. The same mud patch in fall has winter flounder
switching prey, consuming the burrowing polychaete, Micronepthys. During this
season, an input to winter flounder increases tube-builders and decreases
mobile-burrowers. Thus, an assessment of benthic patches that receive impacts
through winter flounder requires seasonally adjusted models.
•Benthic impact sources are not always known, but they may be identified
through the use of a series of benthic models. For example, data showing a
positive correlation between mobile-builders and a fish predator corresponds
to predictions of Model 3b; hence, the impact is a substance or activity that
only effects tube-builder predators. When an impact source is known, then
using the summary of predictions, we compare models with seasons.
Environmental assessment of benthic impacts requires two types of consid-
erations. One is knowing which functional group is affected by the impact.
Weathered No. 2 fuel oil increases mortality on all benthic infauna (13) and
resulting analysis adds all three rows in a table of predictions.
Another consequence of an impact to the benthos is that the community
dynamics are changed. Evidence from the Florida oil spill in Buzzard's Bay,
Massachusetts suggests organisms are inhibited from using the most economical
biochemical pathways (13). The fiddler crab, Uca pugnax can molt, show breed-
ing coloration out of season and burrow too shallow for winter survival; the
small estuarine fish, Fundulus heteroclitus, shows a lower rate of hepatic
lipogensis and, in general, oil induced physiological stress increases demand
on stored energy; petroleum contaminated sediments show the clam, Mya arenaria,
and two species of mussels have reduced carbon flux.
We can extend the above hypothesis to consider for illustrative purposes
that the tube-building amphipod, Ampelisca, through non-lethal concentrations
of oil contaminated sediments, will impair tube construction or build tubes of
reduced structural integrity. However, since it is the tubes of Ampelisca
that interfere with mobile-burrowers, then reduced tube density or construction
86
-------
strength will effectively weaken the competitive link of Ampelisca on the
mobile-burrowers. This reverses the previous assumption that the subsystem of
tube-builders/mobile burrowers is unstable and consequently reverses many of
the predictions. It is not sufficient to identify the organisms receiving the
direct impact of human activity. There also must be some knowledge of how the
impact effects the community strengths of interactions.
In summary, environmental assessment of the benthos requires data that
goes beyond distribution and abundance of organisms. Specific site location
and sediment characteristics will determine interconnections between species
functional groups. Relative quantification through experimentation is neces-
sary for the links between tube builders to filter-feeders compared with the
link from mobile-burrowers to filter-feeders. Seasonal characteristics of
predators and the selection of prey must be obtained or deduced from correla-
tion data and the models. Finally, an estimate of source of settling larvae,
whether local or global relative to the impact site, must be understood. The
general benthic models presented here are to a first approximation accurate
descriptors of the known or currently accepted hypothesis of benthic organism
interactions. Understanding these models is perhaps a first guide to detailed
data collection and constructing site specific models.
87
-------
REFERENCES
1. Levins, R. , 1974. The qualitative analysis of partially specified
systems. Ann. N.Y. Acad. Sci. 231:123-^138.
2. Sanders, H.L., 1958. Benthic studies in Buzzard's Bay. I. Animal-
sediment .relationships. Limnol. and Oceanogr. 3:245-258.
3. Sanders, H.L., 1960. Benthic studies in Buzzard's Bay. III. The structure
of soft-bottom community. Limnol. and Oceanogr. 5:138-153.
4. Mills, E.L., 1967. The biology of an ampeliscid amphipod crustacean sib-
ling species pair. J. Fish. Res. Board Can. 24(2):305-355.
5. Levin, S.A., Paine, R.J., 1974. Disturbance, patch formation, and com-
munity structure. Proc. Nat. Acad. Sci. (USA) 71:2744-2747.
6. Woodin, S.A. and Jackson, J.B.C., 1979. Interphyletic competition among
marine benthos. Amer. Zool. 19:1029-1043.
7. Woodin, S.A., 1974. Polychaete abundance patterns in a marine soft-
sediment environment: the importance .of biological interactions. Ecol.
Monogr. 44:171-187.
8. Rhoades,, D.C., Young, O.K., 1970. The influence of deposit-feeding
organisms on sediment stability and community trophic structure. J. Mar.
Res. 28:150-178.
9. Woodin, S.A., 1976. Adult-larval interactions in dense infaunal
assemblages: patterns of abundance. J. Mar. Res. 34:25-41.
10. Mills, E.L., 1969. The community concept in marine zoology with comments
on continua and instability in some marine communities: A review.
J. Fish. Res. Board Can. 26:1415-1428..
11. Gray, J.S., 1974. Animal-sediment relationships, Oceanogr. Mar. Biol.
Ann. Rev. 12:223-261.
12. Levings, C.S., 1974. Seasonal changes in feeding and particle selection
by winter flounder. (Pseudopleunonectes americanus). Trans. Am. Fish.
Soc. 103:828-832.
-------
13. Sanders, H.L., Grassle, J.F., et. al., 1980. Anatomy of an oil spill:
long term'effects from grounding of the barge Florida off West Falmouth,
Mass. J. Mar. Res. 38:265-379.
14. Watling, L., 1975. Analysis of structural variations in a shallow
estuarine deposit-feeding community. J. Exp. Mar. Biol. Ecol. 19:275-313.
89
-------
Comments on
Qualitative Analysis in the
Setting of East Coast Marine Benthos
Jade B. Waide
Ecosystem Research and Simulation Division
Environmental Laboratory
USAE Waterways Experiment Station
Vidcsburg, MS 39180
One of the most pressing problems in contemporary environmental
biology is the need to develop reliable approaches, firmly based in
ecological theory, for predicting and thereby mitigating human impacts
on natural ecosystems. In spite of this pressing need, most current
approaches in "environmental toxicology11 focus on responses of intact
environmental systems. Although toxicological research on less-than-
whole ecosystems is important and should be intensified/ its results
cannot be extrapolated to natural ecosystems in the field. The constraints
imposed on species or species groups by their coupling within networks
of matter and energy flow in ecosystems alter their dynamic behaviors in
ways which are neither intuitively obvious nor predictable a_ priori.
Thus, reliable impact assessment for natural ecosystems requires research
focused directly at the system level rather than on components or sub-
systems of ecosystems.
Among the various approaches presently employed by environmental
biologists to study ecosystem dynamics in natural and perturbed states,
the qualitative technique known as loop analysis—originally developed
by Richard Levins and explicated in the present paper by Puccia and
Bennett (as well as in the following paper by Lane)—offers many
advantages. This approach, conceptually similar to the earlier work of
Mason on signal flow graphs, but more general and motivated by ecological
concerns, derives its theoretical basis from the theory of differential
equations, as represented near equilibrium points by matrices and associated
linear graphs. The method focuses on the information that can be gained
only from knowledge of the directions and signs of interactions among
components of ecological networks. Thus, the approach avoids the many
frustrations and reductionist aspects of quantifying the plethora of
coefficients and input functions involved in large-scale ecosystem
simulation models. More importantly, loop analysis can focus attention
on critical connections or pathways in ecosystems, and thereby provide
important direction to ecosystem research programs. 'In this sense, loop
analyses may eventually form an integral step in the total development
of more quantitative simulation models. Moreover, current uses of loop
analysis are firmly grounded in ecological theory, at least at the level
90
-------
of the biotic community if not at the ecosystem level. Analyses of loop
models allow one to investigate the effects of network topological
structure on system dynamics and stability independently of the compositional
details of that structure. This is perhaps the major advantage of the
approach for both ecological theory and management application.
The present paper by Puccia and Bennett nicely summarizes the basis
of loop analysis and illustrates its use in investigating qualitative
changes in benthic marine communities found along the Atlantic Coast of
North America to altered parameter inputs. Discussions in the paper
clearly demonstrate how current understandings of interactions among
functional groupings of benthic organisms can be translated into loop
models having wide geographic applicability. The authors further demonstrate
how the models can be manipulated to generate hypotheses or predictions
concerning both correlations among state variables and changes in standing
crops as a consequence of altered parameter inputs to the model. In a
very nice twist, alternative structural models are studied in the hope
of deriving "robust" results which do not depend on any one specific
model structure. This approach should be taken more often; too much of
"systems ecology" has generated predictions which are more strongly
related to the properties of the mathematical formulation employed in a
given model than to the properties of the ecological system being modeled.
Finally, the authors correctly stress the importance of scale and seasonal
effects to their model analyses. (However, it is not clear how loop
models offer advantages over other types of ecological models in
investigating scale-dependent or seasonal effects on model results.)
In spite of the many advantages inherent in loop analysis, it
remains a relatively new approach to studying ecosystems. Thus, problems
in its application remain to be resolved. Also, some reservations exist
concerning the specific way in which loop analysis is used in this
paper. In the remainder of my response I will briefly summarize several
of these problems-reservations. It should be emphasized, however, that
research described by Puccia and Bennett is preliminary and in progress,
so that my comments should not be taken as any final criticism of the
loop model approach. Many of these points may be resolved in subsequent
work. Indeed, most of these issues are ones that require additional
research for their resolution.
First, in spite of comments in the paper to the contrary, the
models developed here are not models of ecosystems, but rather of benthic
invertebrate communities. The benthic community is treated as though it
were isolatable from the remainder of the ecosystem network structure.
Effects of other biotic and physiccchemical variables are assumed to
enter only as so-called "parameter inputs" to the benthic network structure.
In other words, the remainder of the ecosystem is treated as forming the
separate (and separable) environment of the benthic community. I would
suggest that this is a flawed conceptual approach, that the benthic
community is not isolatable in the above sense, and the meaningful
ecosystem analysis and impact prediction require that ecosystems be
treated as intact, functional, biogeochemical systems rather than as
91
-------
discussion on illustration of the use of loop analyses for studying
benthic communities. The loop analysis approach also have promise for
assessing human impacts on natural ecosystems. However, for the reasons
summarized above (and others), the ability to use loop analysis in
routine environmental impact assessment remains to be demonstrated
conclusively.
92
-------
networks of only select biotic state variables. Note that this is a
criticism of the approach taken here rather than of loop analysis per se.
A second problem concerns the need for what might be called ad hoc
hypotheses to resolve ambiguities in loop model predictions of changes
in benthic standing crops following altered parameter inputs (e.g., see
Tables 2-6 and Appendix II). The use of such hypotheses is certainly an
integral part of the research process, and forms a vital link between
the conceputal model of a given system and experimentation on that
system. However, the need for such hypotheses may make loop analysis
only a research tool, limiting its utility in routine impact assessment.
Moreover, many of the hypotheses employed by Puccia and Bennett seem a
bit tentative and require verification or further justification before
results based upon them can be accepted unequivocally.
A third problem involves the dichotomy between the direction and
magnitude of predicted change in an ecological state variable. Loop
analyses can reveal whether a given state variable will increase, decrease,
or remain unchanged following a specific change in some parameter input.
Yet it cannot reveal whether a variable will increase by 10 percent or
1000 percent. Such information, however, is vital in both theoretical
ecology and impact assessment. An increase of 2 percent in some state
variable might be an insignificant change in a given system, whereas a
change of 200 percent would probably be highly significant. Clearly, we
need to be able to discriminate between such cases.
Fourth, since loop analysis is a near-equilibrium technique, it is
questionable whether it can provide needed information to make reliable
impact assessments. I do not mean to criticize the local equilibrium
assumption which is inherent in loop analysis; it has been an exceedingly
powerful approach in many areas of science. However, human impacts on
ecosystems are qualitatively and quantitatively different from "natural"
dynamics in ecological processes at various space-time scales. Human
interventions can alter entire network structures, and push ecosystems
into entirely new and perhaps unanticipated basins of attraction.
Whether loop analysis can provide useful predictions in such cases is a
question requiring additional attention.
Finally, most (though not all; see following paper) of the work on
loop analysis to date has been theoretical. Additional work on the
validation or verification of predictions of loop models is needed
before the approach gains wide use. Such work should focus not only on
actually performing experiments to validate loop model predictions, but
also in developing appropriate methods for validating such predictions
in a rigorous sense. Validation of loop models is likely to be quite
different from validation of quantitative simulation models.
In summary, it is probably fair to conclude that loop analysis
remains a valuable research tool for investigating effects of ecosystem
structure on functional dynamics. The present paper provides a stimulating
93
-------
USING QUALITATIVE ANALYSIS TO UNDERSTAND PERTURBATIONS TO MARINE ECOSYSTEMS
IN THE FIELD AND LABORATORY
by
Patricia A. Lane
Department of Population Sciences
Harvard School of Public Health
Harvard University
Boston, Massachusetts
and
Department of Biology
Dalhousie University
Halifax, Nova Scotia
ABSTRACT
Qualitative analysis is underutilized in many areas of science. This
study involves an application of one qualitative approach (loop analysis) to
a marine ecosystem that is undergoing perturbation. In loop analysis, signed
diagraphs are used to represent a complex system and variables are connected
based on the sign of their interaction coefficients. The study system is
Bedford Basin, Nova Scotia, which is a.small semi-enclosed marine basin.
Analysis of the field community showed that there is a high level of organi-
zation that is trackable over the annual cycle. One loop diagram could be
used to explain much of the observed correlation patterns among the ecological
variables that were measured. The basic ecological structure consists of a
two layered system with nine functional groups (two nutrients, two phytoplank-
ton and five zooplankton variables). The parameter input to the network
enters low in the trophic hierarchy through the nitrogen complex.
The field loop diagram was studied by establishing two laboratory com-
munities in a large divided tank. One side of the tank was enriched with
nitrogen, phosporus and silica. The enriched side of the tank exhibited
higher levels of biomass for most groups of organisms than the unenriched
side. The loop diagram was similar for both sides of the tank and it
differed in five ways from the field diagram. There were two time periods
in which the plankton dynamics differed in both sides of the tank. This
temporal sequence was similar to those observed in the field for spring
bloom conditions. Interestingly, the major parameter input entered at the
algal level and not at nutrients in the first part of the experiment.
During Time 2, parameter inputs occurred through silica for the enriched
side and through flagellates for the unenriched side of the tank.
Loop analysis is a useful approach for analyzing ecological complexity
which is needed for meaningful environmental impact assessment.
94
-------
INTRODUCTION
At present, a satisfactory method of anticipating the impact of multiple
human interventions in ecosystems does not exist. Ecological complexity is
staggering. Coastal marine ecosystems may include several hundred variables
(biotic and abiotic) which are interconnected in thousands of feedback path-
ways. Each pairwise interaction can result in +, - or 0 change in the rate
of increase of each variable. Thus, for a ten species ecosystem there are
100 pairwise interaction coefficients (10^) and 3^00 patterns of ecosystem
interaction. Even though many of these patterns are not ecologically plau-
sible, the number of possible networks is much greater than that which could
be systematically studied and experimentally verified. May (1) conducted a
computer study of stability and complexity in randomly-generated ecological
networks; he found there were billions or googols (10^00) of biologically
reasonable systems. Lawlor (2) demonstrated that for 40 species, there would
be 10^64 networks of which 10^00 would be biologically reasonable. They would
be so sparse that random sampling would never find any of them. Lawlor (2)
further estimated that for a 20 species network there is a 95% expectation of
never encountering an ecological system in ten years if 1 million hypothetical
networks were generated per second by computer.
In addition to delineating a feasible network, there is the problem of
determining the major entry points of the perturbation (for example, an oil
spill) into the network and then to predict the impact of this stress on
individual components as well as on the total network. Even in relatively
simple systems ( and few ecosystems are simple) there can be many complex
pathways of effect between an impact (parameter input) and a particular
variable. Often direct effects, that are traditionally measured in controlled
laboratory experiments involving isolated sub-systems, are nullified or
reversed when they occur in ecological networks (3). For example, pesticides
kill insects directly in laboratory containers but rarely control these pests
in nature.
To make accurate predictions for environmental impact assessments the
multiple causal pathways that exist between the stress and its effects on
ecosystem components need to be understood for realistic ecological networks.
The relationship between correlation and causality has been an uneasy one in
ecological research. Although significant correlations may be found between
pairs of ecosystem variables, at best, ecologists can only infer causal
mechanisms and often unknowingly, nonsense correlations are supported by their
inferences. In aquatic environments many types of parameters enter the
ecological network in different places. For example, temperature and light
are natural parameter inputs. Humans affect ecosystems in many other ways:
thermal pollution, toxic chemicals, physical alterations, oil spills and
natural parameter inputs are interrelated with these human interventions.
Several current controversies center on the basic question of whether the main
parameter enters above (vertebrate predation) or below (nutrient enrichment)
in the aquatic trophic heirachy. Thus, given that aquatic ecosystems are
always partially-specified, it is not surprising that correlation patterns of
aquatic variables and more importantly, the causal mechanisms underlying these
patterns, have not been rigorously delineated.
95
-------
Much of current ecosystem-level research relies on large-scale quantifi-
cation efforts. Often masses of data are interpreted using computer simula-
tion models which are based on complicated differential equations used to
represent the system. Precise values of parameters are usually required.
Qualitative methodology departs markedly from those quantitative approaches
in that it is mainly concerned with delineating the qualitative structure of
complex networks without formulating the equations. Quantification is used
to resolve ambiguity and to gain further insight. The qualitative approach is
useful for understanding ecosystem complexity and in particular, cause-effect
relationships in perturbed ecosystems. While quantification is often desir-
able, the initial use of qualitative analysis enables the investigator to be
more selective in subsequent quantification efforts.
The main objective of the present study is to develop a qualitative
methodology for application to environmental impact assessment problems.
Qualitative networks of several coastal marine communities from the Davis
Strait in Northern Canadian waters to the Chesapeake Bay bordered by Maryland
and Virginia are being developed and studies. This paper reports on one
example of using a qualitative technique (loop analysis) to understand the
network of a plankton community both in the field and laboratory. Microcosm
and large enclosure experiments involving whole communities have proved use-
ful in aquatic ecology since they can be used to integrate field observation
with laboratory experimentation (4,5,6,7,8). Loop analysis is applied here
to determine if the laboratory network is similar to its natural counterpart
and to test loop analysis predictions in a semi-controlled situation.
METHODS
BEDFORD BASIN
Study Area
The study was conducted in Bedford Basin, Nova Scotia, which is a semi-
enclosed marine basin separated from the Atlantic Ocean by a channel, Halifax
Harbour, 10 km long, 20 m deep, and 400 m wide at the narrowest point. The
basin is 5 km long with a maximum depth of 72 m and a surface area of 17 km .
Freshwater input from the Sackville River varies between 10^ and 10' m-Vd.
The tides are of a semi-diurnal type, 0.6-2.0 m in amplitude. The Basin has
been subject to considerable stress from human intervention (9, 10) and is
prone to anoxia. There have been many studies conducted on Bedford Basin
because of the proximity of scientists at the Bedford Institute of Oceano-
graphy and Dalhousie University including basic oceanography, nutrient-
phytoplankton-detritus relationships (11-17) and zooplankton dynamics (18-24).
Field Sampling and Laboratory Processing
All samples and measurements were taken from a fixed navigational buoy
at approximately six week intervals from May, 1974 to March, 1975, on seven
24 hour cruises. Plankton, particulate organic carbon and nitrogen samples
96
-------
were collected from eight depths (0, 5, 10, 15, 20, 25, 40 and 55 m) four to
six times per 24 hour period. (Four to six temperature, salinity, sigma T,
oxygen, primary production and chlorophyll profiles- were determined for each
cruise, but these data are not discussed here).
Triplicate one-liter samples were drawn from van Dorn bottles at the
eight sampling depths for P.O.C. and P.O.N. analysis. The water was filtered
with a 191 my mesh prior to vacuum filtration onto 0.8 my silver membrane
filters and sodium sulfate was added to the sample to remove inorganic salts.
Immediately after filtration, the samples were placed over dry ice, and stored
in a freezer until analysis took place. Carbon and nitrogen values were
obtained using a Hewlett Packard model 185 CHN analyzer. These nutrients were
collected at the eight depths with van Dorn water samplers. 75 ml of water
were preserved in 25 ml of Transeau's solution and a 5 or 25 ml aliquot
(depending on phytoplankton abundance) was enumerated using an inverted micro-
scope. The counts were then converted to cell numbers/m3. Phytoplankton
abundances were combined into the following groups: small diatoms (0-40 my),
large diatoms (> 40 my), small flagellates (> 40 my), and Ceratium spp. The
categories used were A^ for the two size classes of diatoms and A£ for the
flagellates.
Zooplankton were collected by ten minute horizontal tows of two calibrated
Clarke-Bumpus plankton samplers outfitted with No. 10 (158 my) nets. The
samplers had a 12.7 cm diameter. Generally, six complete profiles were
sampled during the 24 hour period of each cruise. Samples were concentrated
and preserved in a 4% formalin solution buffered by hexamethylenetetramine.
(The smaller forms were not adequately sampled and are therefore not included
in this report). Cladocerans and copepods were enumerated by counting at
least two sub-samples, representing .03% of the total sample. A Wild M-7
dissecting microscope was used. Large forms, Calanus finmarchicus and
Sagitta elegans, were enumerated in total. The latter species was divided
into two groups: small (less than 5 mm) and large (more than 5 mm). Nine
other species were identified. Copepodites, nauplii^ Centropages spp, and Acartia
spp, were not enumerated to species. There were sixteen species or groups used
in this study with copepods, cladocerans and small S_. elegans grouped in the
Z2 and Z^ categories and adult j^. elegans and C^. finmarchicus grouped in the
S category.
TOWER TANK
The tower tank is a single cylinder, 10.5 m deep and 3.7 m wide, contain-
ing over 100 m3 of water; it is a key component of the Dalhousie University
Aquatron Facility. The tank and its design are described by Balch et al. (25).
A divider was installed so that control and experimental treatments could be
run simultaneously. In this study thermal and salinity stratifications
developed within the water column.
In the nutrient enrichment experiment, side A (Tower A) was enriched at
approximately 2X ambient concentrations of total nitrogen, phosphorous, and
silica as measured in nearby waters, while side B (Tower B) served as a
control. The communities were collected in Bedford Basin and added
(after careful mixing) in similar densities to Tower A and Tower B. It was
not possible to obtain growth of Sagitta elegans, an arrowworm, which is a
97
-------
major invertebrate predator in Bedford Basin. Six depths were sampled weekly
for six weeks. Animals were then monitored for several additional weeks.
Temperature and salinity were measured at each depth using a Beckman RS-5
temperature and salinity probe. A hose and pump arrangement were used to
sample the water column, with the hose being attached to a specially-designed
horizontal sampler which made it possible to sample several locations at one
time. Approximately one minute was allowed for the hose to clear for each
sample and then one 12 1 and 3 one 1 bottles were filled along with a 1000 ml
graduated cylinder. The sides were sampled alternately at each corresponding
depth. 50 ml of the water collected at each depth was used for a nutrient
sample which was later analyzed on the auto-analyzer for N03, P04, and SiO,.
PO^ has been classified under N^, NO 3 under N£, and SiO^ under Si.
500 ml of water was filtered through a lOmu mesh, and the residue was
then placed in 100 ml of Transeau's solution for phytoplankton enumeration.
The phytoplankton were counted using two 25 ml setting chambers, and transects
were counted on a Zeiss inverted microscope. Diatoms were grouped together
and are referred to as A^, while flagellates are grouped into A£. A 500 ml
sample was taken for chlorophyll, and 2-250 ml replicates were filtered on
45my Millipore filters and analyzed by fluorescence. The 12 1 sample was
filtered through a 40mu mesh, and this residue was preserved in 4% formalin
for zooplankton enumeration. The animals were counted on a Wild M-7
dissecting microscope. Cladocerans and copepods were enumerated by counting
at least two sub-samples, while larger forms were enumerated in total.
Polychaete larvae were put into the grouping M, and the rest of the zooplank-
ton are grouped under 1^ and Z3- 15 species of zooplankton and 10 species of
phytoplankton were identified. Microzooplankton (Z;L) were not counted.
LOOP ANALYSIS AND DATA CALCULATIONS
Loop analysis is a qualitative, network technique that uses signed
diagraphs to represent sets of interacting variables. The mathematical
formalism is described in detail elsewhere (26-28). Lane and Levins (3) have
studied hypothetical freshwater plankton communities using this methodology
and Briand and McCauley (29) conducted lake manipulation studies and tested
loop analysis predictions with their results.
Figure 1 illustrates a four variable aquatic ecosystem. The large circles
enclose the variables (N,A,H and C). The arrowheads indicate a positive
effect on the variable the arrowhead touches and similarly the circleheads
indicate a negative effect on the variable they touch. The signs (+ and -)
represent the qualitative effects of one variable on the rate of change of the
second variable. (Thus, the signs are equivalent to the partial derivatives
of the equation of each variable's rate of change, as a function of all other
variables, evaluated at equilibrium). For example, herbivores(H) consume
algae(A) causing a (positive) increase in H and a (negative) decrease in A.
The following definitions and rules summarize loop analysis as it is
used here:
98
-------
Figure 1. Loop diagram of a four variable aquatic ecosystem. Nutrient(N) and
carnivore(C) variables are self-damped. Algal(A) and herbivore(H)
variables are not self-damped. A circlehead indicates a negative
effect on the adjacent variable and an arrowhead represents a
positive effect on the adjacent variable. N-A, A-H, and H-C inter-
actions represent predator-prey relationships. For example,
carnivores(C) are increased by feeding on herbivores(H) whereas
herbivores are decreased through carnivore predation.
TABLE 1. PREDICTIONS OF DIRECTED CHANGES IN STANDING CROPS FOR VARIABLES
SHOWN IN FIGURE 1
Increase in:
T
H
Variable
N
A
H
C
TABLE 2. PREDICTED CORRELATIONS BETWEEN PAIRS OF VARIABLES SHOWN IN FIGURE 1
FOR INCREASES IN NUTRIENT(N) AND CARNIVORE(C)
Variable Pair
T
Increase in:
N
N-A
N-H
N-C
A-H
A-C
H-C
99
-------
1. A loop of length k is a simple, closed path from a variable to itself
through k steps which visits each variable on the loop only once. The value of
a loop is the product of the a^-s of its links, and the sign is the sign of that
product. A loop of length 0 is oy convention positive and has the value +1.
Feedback is defined as the effect of a variable on itself by way of intervening
variables.
2. Mathematically, the feedback at level k, (F^) , in a system of n > k
variables is defined by F^ = Z (-^"^Lfrn.k) . Feedback at level k is summed
over all sets of the products of m disjunct loops that total k elements.
Disjunct loops have no variables in common (L = loops).
3. Loops of length 0 have a value of +1 and FQ = -1. This is an algebraic
convenience.
4. A path P-y ' ' is a product of (k - 1) alpha values from Xj to X^
involving k variables, none of which are visited more than once. PJJ_ = 1.
5. The complement of a path is the set of variables not on the path.
6. Let Cv be any of s parameters of the system dX-^/dt = f£(X-j,X2>X3 ...,
Xn; C^,C2,C3 ...,Cg). Then the effect of a change in C^ on the equilibrium
level of any variable (X) in the system is
3X.j/3Ch = S Ofj/aCh) X P.^ X Fn_k[Comp
that is, if Cjj is a positive input to X^; then its effect on X.i will have the
sign of the sum of the products of each path from X^ to X^ , each multiplied by
the feedback of its complement, and all divided by the feedback of the whole.
The input along that path has no effect.
In Figure 1, the path from nutrient to carnivore is positive, whereas
the path from carnivore to nutrient is negative. There are three loops of
length two representing predator prey interactions and two self-damped loops
(length one). There are no loops of higher length (level) that is, involving
more than two variables. Table 1 gives the qualitative predictions of changes
in standing crops for four parameter inputs (an increase in N, A, H and C)
shown in the left-hand margin. The changes are read across each row. For
example, an increase in A, perhaps reflecting a better environment for algae
(more optimal temperature or light regime) , results in a decrease in N and
increases in A, H and C.
These directed changes can also be used to predict correlations between
changes in pairs of variables (Table 2). For example, if N is increased then
there will be positive correlations between the changes in standing crops of
all variable pairs. If C is increased, only N-H and A-C pairs will be
positively correlated and all other pairs will exhibit negative correlation.
Both directed changes in standing crops and the resultant correlation values
generated by loop analysis can be tested with field and laboratory data sets.
In this study, the field data are used to study correlation predictions and
the laboratory data are tested against directed change predictions.
TOO
-------
The data correlations for the variables measured in the field cruises on
Bedford Basin were generated as Pearson correlation coefficients using SPSS
and the Dalhousie CDC 6AOO computer. The loop analysis calculations were
done by hand.
RESULTS
BEDFORD BASIN
A summary of the data correlations taken from a standard correlation
matrix is given in Table 3. Most pairs of variables are strongly correlated
and consistent within functional groups of organisms. (Functional groups
are loosely-defined ecological units. They- are used as variables in the loop
diagrams and also to organize the data sets. The definitions are based on
several criteria including: toxonomic affinities, feeding preferences and
behaviors, chemical similarities, developmental stage, and size of organisms
and their foods). The correlation values were generated over all data points
for each pair of variables for the total year of sampling. Preliminary
analysis of individual correlation matrices per date showed that they were
consistent with the all-dates correlation matrix. These correlation patterns
persist over the annual cycle despite high spatial and temporal variation.
This indicates a well-organized community structure which is trackable. Only
those data correlations that refer to variables that are included in the
following (Bedford Basin) loop diagram, and can therefore be tested, are
included in this summary.
For testing the loop predictions, the correlation values were grouped as
negative at the 75 and 95% significance levels and grouped similarly for
positive signs. In general, there was a lot of positive correlation which
indicates that the major parameter input is entering low in the trophic web at
the nutrient or algal levels. (See Figure 1 which shows how positive corre-
lation is generated upwards in a food chain). There was little negative
correlation exhibited in this data set.
The loop diagram representing the Bedford Basin ecosystem is illustrated
in Figure 2. It contains 9 variables and the major parameter input enters at
N2» nitrogen complex. The network is essentially a two-layered system. The
generation times of the organisms are effective in isolating subsystems in the
diagram. The top layer of the diagram is more effectively isolated from the
bottom layer during periods of thermal stratification. The standing crop pre-
dictions for Figure 2 are given in Table 4. The question marks indicate that
there are both negative and positive pathways or complements that result in
ambiguous (?) predictions. Z-^, microzooplankton, functions as a satellite
variable; it changes little itself except when it receives a direct parameter
input. The predicted correlation matrix for the Bedford Basin loop diagram is
given in Table 5. Only the correlation signs for the predictions arising
from a parameter input at N£ are given here. Each parameter input to a loop
diagram variable usually yields different correlation predictions. Also, one
or two link changes in a loop diagram can result in marked changes in both the
standing crop and correlation predictions.
101
-------
TABLE 3. SUMMARY OF CORRELATION COEFFICIENTS FOR BEDFORD BASIN DATA SET*
Variables** + (+) 0 (-) - Sum Correlation sign
Z x Z
Z x S
Z x A-L
Z x A2
Z x N
S x S
S x A^
S x A2
S x N
A^ x AI
AI x A2
AI x N
A2 x A2
A2 x N
N x N
63
2
0
20
11
1
0
0
0
1
0
3
2
6
1
3
0
1
3
2
0
0
0
0
0
2
1
0
0
0
23
7
20
19
14
0
2
2
2
0
4
0
1
0
0
2
7
4 '
0
1
0
2
1
2
0
0
0
0
0
0
0
12
3
0
0
0
0
3
0
0
0
0
0
0
0
91
28
28
42
28
1
4
6
4
1
6
4
3
6
1
253
* + and - represent positive and negative correlation values significant at
the 95% level and signs in parentheses indicate a 75% significance level.
** Key: Z = zooplankton, S = Sagitta - Calanus complex,
A^ = diatoms, A2 = dinoflapjellates, and N = nutrients (nitrogen and
carbon).
To test a loop diagram for correlation relationships, a table of correlation
predictions versus data correlation values is developed (not shown). Of the
253 data correlation values given in Table 3, six represented self-correlation
terms that could not be tested with loop analysis (S-S, A-j^A-^ A2-A2, N-N)
Z-Z correlations could be tested because Z2 and Z$ are separate variables in
loop diagram and they a. predicted correlation value. 157 data correlation
values were of the correct sign and 21 were of the wrong sign. 67 values were
zeroes, that is, statistically insignificant when a + or - value was predicted
by loop analysis. Almost all of the zero data correlation values were in the
direction of the prediction. That a correlation is measured as zero in the
field when loop analysis yields a + or - prediction does not necessarily
indicate that the theoretical predictions are in error; it often means that the
variables were not measured adequately. Only 8% of the data correlation values
were predicted as definitely wrong by these loop analysis results.
102
-------
Figure 2. Loop diagram of Bedford Basin ecosystem. The network includes two
nutrients complexes (N^ = phosphate-silica complex and N2 = nitrogen
complex), two algal groups (AI = diatoms and A2 = flagellates), and
five zooplankton groups (Z^ = microzooplankton, Z2 and Z$ =
cladocerans and copepods, M = macrozooplankton and S = deep living
adult Sagitta elegans and Calanus spp. group).
TABLE 4. PREDICTIONS FOR DIRECTED CHANGES IN STANDING CROPS FOR VARIABLES
SHOWN IN FIGURE 2*
Increase in:
T
Nl
N2
A!
A2
Z2
Z3
S
M
Nl N2 Al A2
+ - 0 +
4- 0 +
- 0 +
0 +
+ + 0 -
+ + 0
0 +
? ? 0 ?
Z1 Z2 Z3 S M variable — *•
0 - - + ' +
0 + + - +
0 - - + -
0 + + - +
? + + - +
0 - + - +
0 + - + ?
0 ? ? - +
* See legend for Figure 2 for meaning of the symbols.
? = Ambiguous result.
103
-------
TABLE 5. PREDICTIONS OF CORRELATION SIGNS FOR FIGURE 2.
ASSUMING PARAMETER INPUT ENTERS AT N2*
Variable
T
N,
i
N2
Xx
A2
Zl
22
23
S
M
Nl N2 Al A2 Zl
x - 0 - 0
x 0 + 0
x 0 0
x 0
x
Zo Zo S
+
4- +
000
+ + -
000
X +
x -
x
M ^_ »• variable
+
0
-t-
0
+
+
-
x
* Key: x = Self correlation term could not be calculated, other symbols are
given in Table 4.
TOWER TANK
The tower tank was used as an experimental tool for the following purposes:
1) to determine if loop diagrams could be constructed with laboratory data,
2) to compare these diagrams with the ones developed for the field data,
3) to delineate how nutrient enrichment affected the laboratory community, and
4) to evaluate if laboratory testing of loop analysis is feasible and useful.
The data set in the tower tank experiment (6"dates x 6 depths =• 36 samples) for
each side is substantially different than its field counterpart.
Depth and Data Variability
Temperature, salinity, oxygen concentration and conductivity levels were
similar in the tower tank on each date. Because of this similarity, only data
from Tower A is shown (Figure 3). Temperature gradually shifted 1% °C in the
bottom water during the experiment. Heating from the overhead lights caused
the surface water to warm more quickly and two thermal'periods were defined
(November 14 - November 28 and December 12-19). Oxygen concentration exhibited
a more marked surface shift in the latter part of the experiment than did
temperature. Undoubtedly, high abundances of phytoplankton at the surface con-
tributed to these high oxygen concentrations.
Nutrients were added to Tower A as P-P04 + N - N03 + Si-Si03 in the ratio
1:5:8 on November 11, 24, and December 12 and 19. Their concentrations changed
markedly with depth throughout the experiment although depth distributions were
similar in the two sides of the tank for each date (Figure 4). There was a
substantial decrease in nutrients in the middle portion of the water column on
December 5 and 12, followed by a decrease in the surface water on December 19.
Nutrients ratios in Tower B remained more consistent than they did in Tower A.
The mean values over dates of all nutrients were higher in Tower A than in
Tower B, however, on particular dates some nutrient concentrations were higher
in Tower B.
104
-------
TOWER A
.0-1
2-
4-
£1 6^
"a.
Q
8-
10-
8-
10-
|
cvjcvjp
>> o u 6
OO 41 V 9>
22QQQ
7 8 9 10 II 12 13
Temperature (°C)
TOWER A
O-i
2-
4-
5 6 7 8. 9 10 II 12 13
Oxygen (ppm)
Figure 3. Vertical profiles of A) temperature (°C) and B) oxygen concentration
(ppm) for each sampling date in Tower A. Profiles of the same
factors in Tower B were similar.
105
-------
TOWER A --- Silicate
0-
2-
4-
6-
8-
= 10-
Nov. W
Nitrate
Nov.21
—• Phosphorus
Nov. 28
2-
4 -
6-
8-
10-
Dec.05
0246 8 10
1.0
0«c. 19
0246 8 10
Nitron mg m3
0 .5 1.0 0
Phosphorus mg m*9
036 0
Silicate mg m"s
0 2 4 6 8 10
1.0
TOWER B Silicate
o •
2-
4-
6-
8-
£ 10-
NOV.M
Nitrate
Nov.21
\
• Phosphorus
Nov. 28
a.
Q
Z-
4-
6-
8-
10-
OK. OS
OK. 12
024 6 8 10 0 2 4 6 8 10 0246810
Nitrate ma m-s
1.0
0 .5 1.0 _
Ptncphorus rnQ m
0 ' 3 ' 6 '
Silicate mg m'5
1.0
Figure 4. Vertical profiles of nutrient concentrations in Tower A and Tower B
from November 14, 1980' to December 19, 1980.
106
-------
Plankton abundances fluctuated with date in both sides of the tank
(Figure 5 and Table 6). Both diatoms and flagellates initially decreased,
then diatoms became dominant in Tower A and Tower B. Flagellates declined to
almost imperceptible levels in Tower A on November 28 and December 5. At the
end of the experiment, in both sides of the tank, diatoms decreased to approx-
imately one-half of their peak abundances and flagellates increased over mid-
experiment levels.
Zooplankton in both sides of the tank exhibited peak abundance by develop-
mental stage. By depth, copepodite maximum abundances coincided with maximum
peaks of chlorophyl concentration. Nauplii were most abundant near the surface
of the water column. Interestingly, the copepodites were not as apparent in
Tower A as they were in Tower B. This may indicate that there was heterchrony
in their development. In terms of mean abundances over all dates, Pseudocalanus
minutus was the dominant adult copepod and nauplii and copepodites, which were
not identified to species, were the next most abundant animal groups. Other
adult zooplankton species from highest to lowest mean abundance were: Oithona
similis. Temora longicornis, Acartia spp. Eurytemora affinis, Centropages spp.-
Tortanus discaudatus and Evadne normanni. Generally, the relative abundances
of the zooplankton species were similar in both sides of the tank. Macro-
zooplankton consisted of polychaete larvae which were more numerous in Tower B
than in Tower A.
In summary, there was considerable variability with depth and date in both
-sides of the tank. All variables except macrozooplankton exhibited higher mean
levels in Tower A than in Tower B (Table 6). This increase is probably a
direct-result of the nutrient' enrichment. There appeared to be two distinct
time periods in the dynamics of both communities. With the exception of marco-
zooplankton, both plankton communities exhibited similar relative abundance
patterns and this similarity did not seem to be swamped by nutrient enrichment
effects. In comparing the laboratory and field communities, the tower tank
communities possessed substantially fewer species than did the field.
The loop diagram of the tower tank communities is given in Figure 6. There
are five differences between it and the field diagram. First, S (Sagitta-
Calanus) was not included in the laboratory diagram because these animals did
not occur; the tower tank is probably too shallow for the depth distinction
found in the field to be important. The self-damping of S was transferred to
M (macrozooplankton). Second, a silica (Si) variable has been added because
silica was not functioning in a similar way to phosphate in the tower tank.
The inclusion of the Si variable does not alter any of the other results,
however, its presence allows predictions of the fate of this nutrient in the
tower tank. In the field data set, there were no silica data and thus, its.
inclusion as a separate variable in the field loop diagram, was not motivated
by the data or a need to explain particular theoretical results. Third, the
interaction between Z-^ and Z£ is a two-way one in the laboratory diagram.
Fourth, Z^ is not self-damped at least when macrozooplankton are present. The
majority of copepod individuals are consuming phytoplankton and detrital foods
(18, 21, 22) especially when populations have high relative abundances of
juvenile stages. Fifth, the interaction between N^ and A£ is also diagrammed
as a two-way flow. Biologically, this means that the flagellates are increas-
ing through their consumption of phosphate and that their rate of change in
107
-------
TOWER A
10-
1
5-
§•
0-1
Phytoplankton
Adults
14 21 28
Nov.
05 12 19
Dec.
1980
29
i—i 1 r
02 06 12 19
'Jon.
1981
TOWER B
m
10- '6'
O
,§5H
f
1
Q.
I
Phytoplankton
'V
Copepodites
14 21 28
Nov.
05 12 19
Dec.
I960
Adults
29
02 06 12 19
Jan.
1981
Figure 5. Changes in phytoplankton and zooplankton abundances in Tower A and
Tower B. Phytoplankton were not enumerated after December 19, 1980.
-108
-------
TABLE 6; ABUNDANCES AND CONCENTRATIONS OF VARIABLES IN THE EXPERIMENTAL (A)
AND CONTROL (B) PARTS OF THE TOWER TANK*
Variable
NT (A)
NI(B)
N2(A)
N2(B)
Si (A)
Si(B)
A! (A)
A]_(B)
A2(A)
A2(B)
Z(A)
Z(B)
M(A)
M(B)
11/14
1.35
1.27
9.29
8.58
6.63
6.45
768.
715.
74.5
24.9
29.2
23.6
542.
555.
11/21
1.13
1.33
8.03
8.48
5.95
6.38
183.
96.0
9.28
6.19
21.1
16.4
208.
264.
11/28
1.47
1.32
10.5
8.52
6.68
6.13
2,440.
1,410.
.1
6.19
17.8
11.7
28.0
57.0
12/5
1.14
1.19
6.19
6.48
5.68
5.51
2,180.
2,260.
.1
6.19
10.6
9.34
0.0
28.0
12/12
1.01
1.28
5.91
5.82
5.20
4.28
1,310.
1,730.
6.19
.1
21.8
14.0
0.0
7.0
12/19
1.53
0.94
5.88
5.58
4.19
4.50
1,200.
749.
12.4
24.8
50.9
45.8
0.0
14.0
X Date-*
1.27
1.22
7.63
7.24
5.72
5.54
1,350.
1,160.
17.1
11.4
25.2
20.2
130.
154.
* Key: N]_ (phosphate), N2 (nitrogen), and S^ (silica) concentrations are
calculated as mg - at N-N03/m3, rag - at P-P04/m3 and mg - at Si-Si03/m3,
and A.I (diatoms), A2 (flagellates), ZA (.zooplankton) are calculated in 10
organisms / m^, macrozooplankton are given as number /m3.
relation to this nutrient is more apparent in the laboratory than in the field.
It is unknown how changes in species composition and diversity affect the
interaction terms. Alternatively, the need to include the N^ to A2 pathway may
be a result of the change in location of the entry of the parameter input to
the community network. In both sides of the tower tank, the major parameter
inputs occurred at A-^ (diatoms) for Time 1 as compared to an N£ entry point in
the Bedford Basin diagram. The ecological nature of the parameter input is
not known, however, it undoubtedly relates to factors influencing the vertical
heterogeneity of A]_. Some realistic, potential factors are light, temperature
and turbulence. From the viewing ports of the tower tank, it was possible to
observe the stratification directly. At 2m it was not possible to see the
divider, whereas at 3m the water was clear and the view unobstructed.
The testing of the loop diagram shown in Figure 6 was accomplished by
comparing the directed changes in standing crop values determined experimentally
with those predicted by the theoretical analysis (Table 7). Whereas, it was
possible to compare predicted and actual correlation values for the field
diagram, the laboratory system fluctuated markedly and correlation matrices in
the early weeks were not similar to those calculated for the final weeks. As
Figure 5 illustrates, there was a great deal of negative correlation between
abundance curves of pairs of planktonic components. Thus, the directed change
method of testing gave a more effective comparison with a temporal sequence
data frame. The theoretical predictions for Time 1 were the same for both sides
of the tower tank and were consistent with the known and observed biology. The
109
-------
Figure 6. Loop diagram of the plankton community in the tower tank during
Time 1. The curved arrow indicates where a positive parameter
input enters the network. Predictions for this input are given
in the first column of Table 7. All symbols are the same as those
given in Figure 2, except that Si (silica) is separated from N.
(phosphate complex). The brackets around the negative loop of
length two between
in the laboratory.
and Z indicate that this loop may weaken
no
-------
TABLE 7. COMPARISON OF DIRECTED CHANGES IN STANDING CROPS OF
TOWER TANK VARIABLES (A = EXPERIMENTAL, B = CONTROL)
WITH PREDICTIONS FROM LOOP ANALYSIS OF DIAGRAM IN
FIGURE 5. TIME 1 INCLUDES DATA FROM 11/14/80 TO
11/28/80, AND TIME 2 INCLUDES DATA FROM 12/12/80
12/19/80
Increase in: Predicted Change Data Changes
f Time 1 Time 2 Time 1 Time 2
A and BAB A B A B
~^"A —Si. -HAo
N _+___+_
N2 ?* _
Si --+--- +
A! +._++__
A O «^ l_ i _i i
A2 ?*-+-(- - -+ +
Z, +__**** ** **
4 _++-..+ +
M -0+0-0 +
Key: Symbols are given in the legends to Figure 2 and Table 4.
* If the A2 - Z3 loop weakens, then these changes would not be
ambiguous; N£ would increase and A2 would decrease.
** Z^ was not measured in these experiments.
changes in N~ and A~ levels were ambiguous because there is a positive pathway
of effect through Z2 and Z-$ and a negative pathway through N]_.
In Time 2, the network of Tower A changed because the macrozooplankton
were not present (not shown). The self-damping of the macrozooplankton variable
was probably transferred to Z-j, The loop between A2 and Z-j was stronger than it
had been in Time 1. The major parameter input entered through a negative change
in Si concentration. The predictions are compared with the actual qualitative
data changes in Table 7. If Z was not self-damped then the change in A2 would
be predicted to be 0 instead of the increase that is predicted with Z<$ self-
damped and which was also found by the empirical results. Increased self-
damping may have arisen because of the proportionally higher abundances of adults
in the zooplankton community that occurred at the end of the experiment. The
presence of self-damping, however, was not experimentally determined. In the
control side (Tower B), the network did not change but the parameter input
entered at A2. The most likely explanation is related to the vertical hetero-
geneity of the phytoplankton in regard to factors such as light, turbulence, and
temperature.
Ill
-------
DISCUSSION
Two examples have been presented to demonstrate how a qualitative method-
ology (loop analysis) can be applied to marine plankton communities in the field
and laboratory. The underlying motivation of this study is to develop better
ways of understanding complex ecological systems without resorting to a "measure
everything" approach which is always costly and often ineffective. Improved
understanding and methodology are essential for environmental assessment. The
Discussion is centered around some of the advantages and problems with using
loop analysis in an ecological context. In some cases, these categories are
not discrete entities. Since the purpose of the paper is to illustrate "by
example," the discussion is also organized this way rather than presenting a
full literature review of the many biological results and conclusions of the
study.
ADVANTAGES OF LOOP ANALYSIS
Coping with Complexity
Lawlor (2) estimates that there is less than one change in a googol of
constructing an ecosystem with a random-number generator. An aquatic environ-
ment can possess a 1000 species; thus in the sense of Lawlor (2) the networks
in the water may exceed the sand grains on the beach. The complexity problem
is enormous. Lawlor (2) suggests that since the problem cannot be tackled by
computer, it is more reasonable to confine exploration and study to.those
systems that are biologically acceptable. If a large enough group of these
systems can be analysed, then basic ecological properties and structural
similarities should emerge. Loop analysis has tremendous potential to make
complexity manageable at a low cost relative to other methods. Loop diagrams
can be constructed and tested with minimum data requirements. The results of
loop analysis are more easily-transferred among ecosystems than are those of
the more traditional quantitative methodologies. (Computer simulations often
work only within a narrow set of parameter values and initial conditions found
in a single ecosystem). It will also be possible to use loop analysis to
develop a set of ecosystem descriptors for community structure and stability;
for example, the average connectivity for a system of N variables or ratios of
feedback at each level of N - K variables would be useful.
Understanding Biological Phenomena
Because the qualitative nature of the ecosystem is emphasised in loop
analysis, it is easier to focus on the critical components and pathways of
effect in understanding ecosystem structure and function in contrast to
quantitative methodologies. Many different types of variables can be included
in one diagram and there is no need to standardize everything into a single
unit such as kilocalories of energy or grams of carbon. For example, abiotic
and biotic variables were included in the Bedford Basin diagram. Variables
can be also used which cannot be measured. Human actions such as management
decisions or institutions such as regulatory agencies can be diagrammed with
plankton populations.
112
-------
In this study, the Bedford Basin loop diagram explained most of the corre-*-
lation patterns that were observed during an annual cycle. The plankton
community appears to be well-organized and receiving a major parameter input
through the ^ variable. In the laboratory, it was possible to study the net-
work further and to delineate important structural differences between the field
and laboratory communities-. The laboratory experiment provided a detailed
analysis of a small section of the annual cycle in comparison to the field data that
integrated the annual results-. In many regards, the dynamics of the tower tank
communities resembled those of field communities- undergoing spring bloom con-
ditions (30-32). At the beginning of the bloom there are low temperature and
light levels that begin to rise. Turbulence has often been involved as a
mechanism to explain nutrient increases in the surface waters. Seasonal
succession proceeds with a shift from small to large diatoms to flagellates
and from microzooplankton to larger zooplankters . Flagellates appear after
nutrient levels become depleted and thermal stratification occurs. In both the
experimental and control sides of the tower tank, major parameters enter through
the algal populations and not through nutrient variables. Thus, despite the
differences in nutrient input and the resultant differences in biomass levels,
the qualitative plankton dynamics were similar for both Tower A and Tower B.
Events other than nutrient enrichment were driving the networks. This is
interesting in contrast to many of the disputes in the literature on the roles
of particular factors in the spring bloom. (The shift from the major parameter
input entering at nutrient to one entering at the algal level has been found in
a qualitative study of eutrophication in Gull Lake (Lane, unpubl.); perhaps
this parameter shift represents a fundamental phenomena in aquatic ecosystems) .
By integrating the field and laboratory results with the theoretical frame-
work provided by loop analysis, it is possible to understand the complex
community dynamics of systems undergoing stress more effectively than by using
a single data set. Extrapolation from laboratory results to field situations
is often difficult. The laboratory diagram is similar to the field diagram
and differences in the variables and linkages were directly related to the
measured biological differences. This also confirmed the sensitivity of the
methodology.
Guide to Research
Loop analysis also provides a guide to research in that it generates test-
able hypotheses and elucidates particular variables and interaction links that
need further study. A few examples are given from the present study; the list
is not inclusive. 1) The tower tank system experienced little turbulence; it has
often been given an important role in spring succession dynamics. Is turbulence
important? Was the low level of turbulence related to the increased stratifi-
cation of phytoplankton which in turn led to a shift in the node of entry of
the main parameter input? Conversely, does the parameter input at algae occur
regardless of the levels of turbulence and/or prevailing nutrient ratios in
the field? These questions could be answered with more experiments and
theoretical analysis. 2) What are the major mechanisms underlying parameter
input to algal variables? 3) Do subtle shifts in nutrient ratios change either
the structure or the node of parameter entry in these networks? 4) How much
self ^-damping is occurring; for example, how many animal populations exhibit
cannibalism within their functional group? 5) Under what circumstances is
A-Z a strong loop and when can it be disregarded? 6) What is the role of
113
-------
biomass level in the qualitative dynamics that were observed in both sides of
the tank? When do enhanced levels of biomass lead to structural changes in
the network? 7) In this experiment, when'M dis-appeared in Tower A, there
were few changes in the predictions. The loss of other -variables, however,
could have more dramatic consequences on the network and loop analysis pre-
dictions. When will the loss of a variable be important?
Applying loop analysis to these field and laboratory systems provides a
wealth of interesting ideas for future research and study. Often analysis of
loop diagram yields important predictions about particular links in the net-
work or delineates an omitted variable. Lane and Levins (3) demonstrate how in
a 4 variable-9 link system only four links need to be measured to understand
the critical pathway of effect. Data sets are not used to correct equations or
achieve better fitting parameters but rather to find new ways to look at
systems and to find new variables of interest. In subsequent research, there
can be more selection in collecting the missing information and the "measure
everything" approach- does not need to be invoked.
Ecological Causality and Prediction
Loop analysis helps greatly in understanding cause and effect in
ecological systems. "Causality, however, is always a theoretical construct:
as Hume observed, causation cannot be proved empirically. Whatever the
observed evidence, there remains the possibility that some uncontrolled
factor was responsible for the occurrence" (33). Because loop analysis gives
direct predictions in regard to the standing crops, turnover rates (not shown,
(3) and their correlation patterns, it is useful for a variety of assessment
and management problems. Loop diagrams provide needed focal points, and causal
pathways can be drawn out and examined. Loop analysis also provides an inte-
grative tool for predicting the effects of multiple indirect pathways between
pairs of variables that are not directly related to each other.
PROBLEMS WITH LOOP ANALYSIS
Integration of Theory .and Biology
Whereas, loop analysis permits an effective integration of theory and
biology, the results reported in this study should also serve as a warning.
The subtle shifts in parameter inputs over small time scales and the qual-
itative changes in interaction signs can occur even when the species list and
relative abundance patterns are fairly consistent. A great deal of biology
is needed to elucidate these subtle effects. Loop analysis assumes that
variables with similar time dynamics are grouped together in the networks.
This assumption needs substantially more testing than it has received to date.
For example, from both freshwater and marine results, it appears that most
fish populations are acting as parameters, not variables, in regard to the
plankton populations. It may be that fish are sufficiently slower so as to
be "disconnected" from the plankton dynamics. Whereas, loop analysis can
help delineate where biological detail needs to be supplied, it is not a
magic antedote that will solve all of our environmental problems. Good bio-
logical insight is needed as much as it has ever been.
114
-------
Testing Loop Analysis Results
This study used the most simple, direct form of comparing theoretical
and empirical results possible-inspection. Statistical testing is presently
receiving considerable effort in our research group and will be reported in
more detail later. At present, it is not possible to state if the best
diagram for the observed results has been constructed. 8% of the field
results were in error by a complete sign change. It is not known 1) what the
universe of possible networks is like, 2) when a particular level of agreement
is satisfactory, and 3) how to separate errors of prediction from sampling
errors. There is also much work needed in delineating the attributes of
particular types of data sets and how they can be interrelated in regard to
testing loop predictions. For this study the two data sets were fairly con-
sistent in regard to the networks that were generated. This may not always
occur.
ACKNOWLEDGMENTS
I wish to thank William Atherton who coordinated the 1974-1975 field
data collection in Bedford Basin. Anna Atherton, Kenneth Lee, David MacDonald,
Paul Phinney and Joseph Salter participated in the cruises. Roanne Conover
enumerated the phytoplankton. Dr. Robert Conover of the Bedford Institute
of Oceanography provided invaluable advice and support during the study. I
am indebted to Terrance Collins and Gregg Hitman who provided excellent
technical assistance in making the project possible and for many ideas which
enhanced the final results. John Wright helped immensely in data analysis and
computer programming. Dr. Noval Balch, Director of the Aquatron Facility at
Dalhousie University and his technical staff maintained the tower tank and
assisted in operational details. Dr. Richard Levins has been a long-term
collaborator on the theoretical aspects of this study. Dr. Brian Marcotte
and Terrance Collins illustrated the manuscript. Dr. Brian Marcotte and
Gregg Mitman suggested several points in regard to interpreting the data.
All errors in interpretation are my own. Gregg Mitman helped in manuscript
preparation and Jean Joseph typed the manuscript. N.S.E.R.C. Grant No. A9621
supported the field study and Grant No. G0223 funded the tower tank experiment.
E.P.A. Grant No. R807713-01 facilitated the subsequent data analysis and
development of the application of the theory to marine ecosystems.
115
-------
REFERENCES
1. May, R.M. 1973. Stability and complexity in model ecosystsms. Monographs
in Population Biology, No. 6. Princeton University Press, Princeton, N.J.
2. Lawlor, L.R. 1978. A comment on randomly constructed model ecosystems.
Amer. Natur. 112:445-446.
3. Lane, P.A. and R. Levins, 1977. The dynamics of aquatic systems 2. The
effects of nutrient enrichment on model plankton communities. Limnol.
Oceanogr. 22:454-471.
4. Mullin, M.M. and P. Evans, 1974. The use of a deep tank in plankton
ecology, 2. Efficiency of a planktonic food chain. Limnol. Oceanogr.
19:902-911.
5. Oviatt, C.A., H. Walker and M.E.Q. Pilson, 1980. Analysis of microcosm
and ecosystem behavior using multivariate techniques. Helgolander Wiss.
Meeresuntersuchungun (in press).
6. Parsons, T.R., W.H. Thomas, D. Seibert, J.R. Beers, P. Gillespie and
C. Bowden, 1977. The effect of nutrient enrichment on the plankton community
in enclosed water columns. Int. Revue ges Hydrobiol. 62:565-572.
7. Parsons, J.R., K. von Broukel, P. Koeller, M. Takahashi, M.R. Reeve, and.
0. Holm-Hansen, 1977. The distribution of organic carbon in a marine
planktonic food web following nutrient enrichment. J. Exp. Mar. Biol.
Ecol. 26:235-247.
8. Stickland, J.O.H., 0. Holm-Hansen, R.W. Eppley, and R.J. Lirin, 1969.
The use of a deep tank in plankton ecology 1. Studies of the growth and
composition of phytoplankton crops at low nutrient levels. Limnol.
Oceanogr. 14:23-34.
9. Riley, G.A. 1974. Oceanography 'of the Halifax inlet in relation to
problems of sewage disposal. Report to the Department of the Env.,
Province of Nova Scotia, Inst. for Env. Studies, Dalhousie University,
Halifax, Nova Scotia.
10. Platt, T., R.J. Conover, R. Loucks, K.H. Mann, D.L. Peer, A. Prakash and
D.D. Sameo.to, 1970. Study of a eutrophicated marine basin. Fish Rep.
99:1-10.
11. Conover, S.A.M., 1974. Nitrate, ammonium, and the urea and nitrogen
sources for marine diatoms in culture and in spring blooms. Ph.D. Thesis,
Dalhousie University, Halifax, Nova Scotia, 175p.
12. Conover, S.A.M., 1975. Nitrogen utilization during spring blooms of
marine phytoplankton in Bedford, Nova Scotia, Canada, Mar. Biol. 32:247-261.
116
-------
13. Duerden, F.C., 1973. Aspects of the phytoplankton production and
phosphate exchange in Bedford Basins, Nova Scotia, Ph.D. Thesis,
Dalhousie University, Halifax, Nova Scotia, 242pp.
14. Hargrave, B.T. and S. Taguchi, 1978. Origin of deposited material
sedmented in a marine bay. J. Fish Res. Bd. Can. 35:1604-1613.
15. Krauel, D.P. , 1969. Bedford Basin data report - 1967. FBR Tech. Rep. 120,
84pp.
16. Lee, K., 1975. Diurnal fluctuations in photosynthetic pigments in Bedford
Basin. B.Sc. Thesis. Dalhousie University, Halifax, Nova Scotia, 36pp.
17. Platt, T. and R.J. Conover, 1971. Variability and its effect on the
24 hour chlorophyll budget of a small marine basin. Mar. Biol. 10:52-65.
18. Gibson, C.G., 1980. Some aspects of the feeding biology of the marine
copepod Temora longicornis. M.Sc. Thesis, Dalhousie University, Halifax,
Nova Scotia, 104pp.
19. Gifford, P., 1979. The population biology and community ecology of the
marine cladoceran of Bedford Basin, Nova Scotia, with special reference to
the feeding biology of Podon leuckarti. G.O. Sans, 515pp.
20. Pearre, S., 1973. Vertical migration and feeding on Sagitta elegans
Verrill. Ecology 54:300-314.
21. Poulet, S.A., 1973. Grazing of Pseudocalanus minutus on naturally
occurring particulate matter. Limnol. and Oceanogr. 18:564-573.
22. Poulet, S.A., 1976. Feeding of Pseudocalanus minutus on living and non-
living particles. Mar. Biol. 34:117-125.
23. Sameoto, D.D., 1973. Annual life cycle and production of the chactognath
Sagitta elegans in Bedford Basin, Nova Scotia. J. Fish Res. Bd., Conn.
30:333-344.
24. Zo, Z. , 1973. Breeding and growth of the chaetognath Sagitta elegans in
Bedford Basin. Limnol. and Oceanogr. 18:750-756.
25. Balch, N., C. Boyd, and M. Mullen, 1978. Large-scale tower tank systems.
Rapp. R.V. Reun. Cons. Int. Explor. Mer., 173:13-21.
26. Levins, R., 1973. The qualitative analysis of partially-specified systems.
Ann., N.Y. Acad. Sci. 231:123-138.
27. Levins, R., 1975. Evolution in communities near equilibrium, pp. 16-50 in
Ecology and Evolution of Communities, eds. M.L. Cody and J.M. Diamond,
Harvard University Press.
28. Lane, P.A. and R. Morison (eds.) 1981. Analyzing perturbations to marine
ecosystems using loop analysis and time-averaging by the Harvard Marine
Ecosystems Research Group. E.P.A. report, 60pp.
117
-------
29. Briand, F. and E. McCauley, 1978. Cybernetic mechanisms in lake plankton
systems: how to control undesirable algae, Nature, 273:228-230.
30. Pratt, D.M., 1965. The winter-spring diatom flowering in Narragansett
Bay. Limnol. Oceanogr. 10:173^184.
31. Raymont, J.E.G., 1963. Plankton and Productivity in the Oceans.
Pergamon Press, New York.
32. Smayda, T.J. and G. Hitchcock, 1977. Bioassay of lower Narragansett Bay
waters during the 1972-1973 winter-spring bloom using the diatom
Skeletonema costatum. Limnol. Oceanogr. 22:133-139.
33. Lave, L.B. and E.P. Seskin, 1979. Epidemiology, causality, and public
policy. Amer. Sci. 67:178-186.
118
-------
Comments on
Using Qualitative Analysis to Understand Perturbation
to Marine Ecosystems in the Field and Laboratory
Jack B. Waide
Ecosystem Research and Simulation Division
Environmental Laboratory
USAE Waterways Experiment Station
Vidcsburg, MS 39180
The work summarized in this contribution by Lane complements that
discussed in the previous paper by Puccia and Bennett. Both studies
involve the application of loop analysis methodologies to the study of
marine ecosystems, and both are components of a single, overall research
program funded by the Environmental Protection Agency. Because of the
interrelatedness of the two studies, many of my comments in response to
the previous paper apply to the present one also. My comments here will
focus primarily on additional insights into the ecological utilization
of loop analysis brought out by this paper. As before, work described
here is in progress and not complete. My comments should again be
interpreted in light of this fact.
Vfoereas the previous paper focused on the structure of marine
benthic communities, Lane is concerned here with planktonic sytems in a
small, semi-enclosed marine basin located offshore from Nova Scotia.
Also, in contrast to the largely theoretical nature of the previous
paper, Lane demonstrates how theoretical loop analyses can be combined
with both field and laboratory data into an overall research program on
plankton dynamics. Correlations derived from field data on abundances
of planktonic organisms and essential elements are used to test predictions
of state variable correlation patterns generated by loop analyses of a
network model of the plankton community. Similarly, data resulting
from laboratory manipulations of plankton community dynamics using
experimental "towers" are used to test loop model predictions of changes
in standing crops of select state variables following altered parameter
inputs. Thus, Lane provides examples of ways in which loop model
predictions can be tested, and shows how the loop model approach can
provide direction to field research programs, thereby avoiding the
pitfalls of what she correctly terms the "measure everything" philosophy
which has characterized much ecosystem research in the past (especially
that which has had simulation modeling as a major goal or component). In
this way, Lane demonstrates how loop analysis can provide an essential
link between theory and experimentation in ecology. Another contrast
between this paper and the previous one relates to the structure of the
models studied: the models used here are structurally more elaborate, and „
119
-------
include essential elements and several categories of autotrophs as state
variables in addition to strictly invertebrate compartments.
As with the previous paper, this contribution raises several questions
concerning the ecological utilization of loop analysis which should be
addressed in subsequent work. First, the question of the isolatability
of a portion of the ecosystem for study via loop analysis again requires
consideration. Although the network models used here are more "complete"
than those used in the previous paper, they still ignore key components
of ecosystems (e.g., pathways and state variables involved in the breakdown
of detritus and the regeneration of essential elements). Such variables
again enter the present analysis only indirectly as "parameter inputs."
Yet inclusion of such variables in the network structure could grossly
change model predictions. For example, inclusion of pathways for detritus
production and remineralization would add indirect, positive pathways
from algae to essential elements, and might thereby alter model predictions.
Again, it would seem as though all major ecosystem components required
for a given analysis should be included explicitly in the loop models to
allow meaningful predictions to be made. Conceptually treating portions
of the ecosystem as being separable or isolatable fron the remainder of
the network strucutre is a questionable procedure, and may lead to
erroneous predictions.
A second problem in this use of loop analysis again involves what
might be considered "extra" or ad hoc hypotheses. In her paper, Lane
stresses the need to determine tEe major entry point(s) for human inter-
ventions into ecosystems—i.e., the state variables though which the
altered parameter input(s) is(are) "entering" the system. Yet little
discussion is provided as to how this is done in a rigorous or objective
fashion. How, for example, does one determine that the major parameter
input to the "field" loop model is through variable N_, or through
variable A, to the "laboratory" loop model? On what objective basis
does one determine that the major parameter input for "Time 2" has
changed, or that it differs for the two towers? Also, does the approach
taken preclude multiple "parameter inputs" to the system of interest?
Perhaps due to the brevity of the discussion, the approach taken here
appears tentative and arbitrary. Again, this problem could be resolved
by including all ecosystem components relevant to a given analysis
explicitly as state variables rather than implicitly (and arbitrarily)
as parameter inputs.
A third problem raised by this paper concerns the model dependence
of loop analysis results. Lane correctly emphasizes that, as is the
case in all modeling studies, results of her loop analyses are dependent
on the structure of the models employed. Even small changes in model
structure could possibly lead to substantial changes in model predictions.
However, in contrast to the previous paper, this work does not include
an analysis of alternative model structures, so that one does not know
how sensitive (or robust) these results are. Additionally, different
loop diagrams for the field and laboratory planktonic assemblages are
presented, without thorough discussion of how the model structures were
120
-------
constructed from the available data in each case. Also the author
suggests that the connection between variables A» and Z., may "weaken" in
the laboratory assemblages, without explaining how this was determined.
All of these questions suggest that additional attention must be given
to objective means for constructing loop diagrams, and to investigating
the sensitivity of model predictions to network structure.
The use of correlation analyses to verify or text the prediction of
loop analyses is another area requiring additional study and clarification.
The validation of models, whether qualitative or quantitative, is a non-
trivial problem which has generally received inadequate attention by
systems ecologists. Although the fact that a large number of correlations
in the field data set were correctly predicted by the associated loop
model may be reassuring, it is not necessarily a rigorous test of model
validity. Because of the large number of direct and indirect pathways
involved in a network model of any degree of structural complexity/ an
"incorrect" model structure could "correctly" predict a number of inter-
variable correlations. Thus, the whole issue of validating loop models
requires further study. As suggested in my response to the Puccia-
Bennett paper, methods for validating such qualitative models will
probably be quite different from methods employed to validate quantitative
simulation models.
Finally, some comments concerning the issue of ecosystem complexity
as raised by Lane must be made. It should be pointed out that complexity
is a property of a scientific description of a given system; it is not
in any sense "inherent" in the system. Models of natural phenomena
which are broadly recognized as being useful or successful are such
because they focus on essential details of a given phenomenon, and ignore
other related but "nonessential" information. Ecosystem models (in the
broadest sense) continue to appear complex because ecologists persist in
describing ecosystems in terms of species-level properties and interactions,
and not as a functionally intact system with indentifiable macroscopic
properties. The number of species or interaction pathways says nothing
about ecosystem complexity. Nor is the improbability of randomally
constructing a "reasonable" biological network structure via computer
experiment a major challenge to ecosystem theory. Such approaches
ignore two essential sources of constraint on current ecological networks:
constraints imposed on contemporary biotic assemblages by historically
antecedent structures, and constraints imposed on biotic evolution by
system-level matter and energy exchanges with the surrounding physico-
chemical environment. Biotic structures of ecosystems evolve over long
time periods from prior structures under the constraints of biogeochemical
interactions between the biota as a whole and the physicochemical environ-
ment to which that biotic structure is intimately coupled. They do not
emerge de noyo via any sort of random arrangement of members of a local
species pooT. It is a worthy goal to seek "non-complex" descriptions of
ecosystems. However, current ecosystem models are complex because
ecologists have generally failed to achieve an acceptable conceptual
simplification of these systems and not because ecosystems are inherently
more complex than other systems studied by science.
121
-------
To summarize, Lane has provided a stimulating illustration of the
coupling of theoretical loop models with field and laboratory experimental
data in a study of marine plankton dynamics. Her work nicely complements
that of Puccia and Bennett, and demonstrates again that loop analysis is
an effective research tool for elucidating ecosystem properties. Ihe
approach also has real potential for impact analysis of ecosystems, but
a potential which largely remains to be demonstrated conclusively.
122
-------
NEW APPROACHES TO MEASURING THE BIOLOGICAL
IMPACT OF ENVIRONMENTAL TOXICANTS
R. D. Hinsdill, W. G. Hunter, W. P. Porter and T. M. Yuill
Environmental Toxicology Center
University of Wisconsin
Madison, Wisconsin 53706
OVERVIEW AND GENERAL REMARKS
Current methods for assessing the impact of polluting chemicals and
environmental contaminants are limited in scope and generally are not well
suited to predicting biological impact under real life situations, since
the exposure is generally occurring simultaneously with a variety of other
types of stress (nutritional, chemical, pathogenic). The work to be des-
cribed is part of a collaborative study to construct an integrated numerical
predictive model of wild small mammals outdoors from components recently
developed by the group. The model can be used to study the complex inter-
actions between physiological stress due to climate, chemical toxicity, and
pathogens. Predictions of the effects of climate, toxicity and pathogens
on growth, reproductive potential and activity periods as well as food and
water needed for survival, growth and reproduction can be generated from
the model and tested experimentally. The theoretical-experimental model
should be useful ultimately as a quick screening device and field monitor.
The model incorporates a state-of-the-art breakthrough in predicting heat
and mass transfer of animals with furry insulation in almost any outdoor
environment. It also includes a microclimate model, a predictive outdoor
respiratory water loss model, a physiological thermoregulation model and
both fractional and full factorial experimental designs that provide rapid,
effective, simultaneous screening of many variables whose effects must be
empirically determined. Most of the first six months were spent getting
needed preliminary data in the laboratories of the four participating
colleagues so as to launch our first series of integrated, multifactoral
experiments.
Although the division of labor in making the presentation may give the
impression that each group is working independently, we can assure you that
we are integrating our efforts and findings to a high degree in order to
complete our original objectives. The work will be discussed in three parts
as shown in your program and we will attempt to provide sufficient back-
ground information to make clear why certain approaches, techniques and
agents have been chosen over others in developing the model.
123
-------
Comments on
New Approaches to Measuring
the Biological Iirpact of
Environmental Toxicants
Richard M. Kocan
University of Washington
The objective of this proposal is to develop a model system which
would predict the effect of a combination of naturally occurring environ-
mental insults on the growth and reproductive success of a small mammal
under laboratory and field conditions.
The three variables in the study were to be climate, pathogens and
toxic substances. In order to carry out a study with such varied
disciplines, it was necessary to recruit the expertise of several
different investigators. Dr. Porter was to carry out the heat loss
studies and climatic stress difference, Dr. Yuill had responsibility for
pathogens and their effect on the host's general health and reproductive
capacity. Dr. Hinsdill was to assess the effects of toxic chemicals on
the same parameters. All of the data and experimental design were to be
examined by Dr. Hunter, who has a wide range of experience in statistics.
Dr. Porter's group was to establish certain baseline data on
deer-mice in metabolic chambers. These data were to include heat
production, 0_ consumption, CCL production and evaporative water loss
under various climatic conditions, disease states and toxicant exposures.
Once these data were collected, long-term experiments in the Biotron
(controlled environmental chamber) could be carried out to determine the
ultimate effect of the various combinations of insults on the growth and
reproductive success of the deer-mice.
Dr. Yuill's group was responsible for determining the effective
non-lethal levels and routes of administration (infection) for acute and
chronic virus and bacterial infections. They were then to collaborate
with Dr. Hinsdill's group to determine appropriate combinations of
pathogen and toxicant that would allow the animals to survive for the .
duration of the study period thus allowing the assessment of these
sublethal infections and intoxications on the long-term health and
reproductive success of deer-mice. The pathogens chosen for these
studies included viruses which produced both acute and chronic infections
and bacteria which produce acute infections. By taking blood samples at
various intervals throughout the experiment it would be possible to
monitor the production of antibody and interferon in the animals under
different conditions of environmental stress.
124
-------
The third phase of the study was to be carried out by Dr. Hinsdill
and consisted of determining immunosupressive non-lethal dose levels of
potential chemical toxicants. Two compounds (cyclophosphamide and
azathioprine) were selected as positive controls due to their known
immunosupressive action. The unknown chemicals chosen belong to a group
of chemicals which are relatively unknown in terms of their effects on
mammalian systems. These are the plant growth regulators Maleic hydrazide,
cycloheximidefand triiodobenzoic acid.
125
-------
IMPACT OF TOXICANTS, DISEASE AND CLIMATE
ON GROWTH AND REPRODUCTION USING
i IS MAMTr.l II
Warren P. Porter and James Jaeger
Department o-f Zoology
University of Wisconsin
Madison, WI 53706
INTRODUCTION
Many biotic and abiotic variables that affect an
animal's capacity to grow and reproduce have been
identified. How to efficiently study and model the effects
of such a large number of variables in a biological system
is a difficult problem. We have found that an energy and
mass budget framework provides an overall context in which
we can examine and model many diverse factors affecting
the capacity of animals to grow and reproduce.
One cornerstone to this approach is the development
of mechanistic models of heat and mass exchange between an
orgainism and its environment. These exchanges and the
energy and mass balances which result are of prime
importance to survival, growth, and reproduction. By
modeling these exchanges we can understand the limits they
impose upon many aspects of an organism's life.
These exchanges take place toy fairly simple physical
mechanisms and can be described by mechanisic models <1).
These models describe heat and mass exchange as a function
of the microclimate and certain relevant physical and
physiological features of the animal. They are then used
to drive other models of interest, such as food and water
requirements, activity patterns, body temperature, and
dessication rate. This approach has been used to study
temperature and water regulation in ectotherms and the
constraints those regulations impose upon an ectotherm's
use of time and space (1,2,3,4,5,6), and species
distributional limits <7).
126
-------
A second cornerstone to our approach is the use of
highly efficient full and fractional factorial designs to
<=>•.•:pi ore biological variables that can not be defined with
heat and mass balance equations. We have completed
experiments using those fractional factorial designs, to
test model predictions of climatic effects and to identify
the significant "black box" variables affecting the
ability of female deer mice, Pp»romy«sri.ic; mani ml atu«s , to
raise young to weaning (8,9,10). All significant
variables were then used in full factorial designs to
establish a response surface. These designs are ideally
suited for screening large numbers of variables
simultaneously and then iteratively adding resolution when
the data suggest that additional information is needed.
We are now integrating these designs with our heat
and mass transfer models to predict effects of toxicants
and disease on maintenance, growth, and reproduction
requirements for animals in any physical environment. To
do this we are working in three major areas: 1. Heat and
water loss for small mammals 2. Food and water
assimilation rates for small mammals 3. Growth and
reproduction studies of small mammals. Heat and mass
transfer equations have been used to predict maintenance
food and water requirements of healthy adults of several
species (1,2,3), but prediction of climatic effects of
growth and reproduction potential has only been done for
healthy invertebrate ectotherms (10,11) and fish <12).
MODELS OF GROWTH AND REPRODUCTION
Our model of animal growth and reproduction is
derived from the conservation of energy and mass for
animals. Simply stated, the energy available for growth
and reproduction is the difference between the energy
assimilated from food and that which is metabolized to
carbon dioxide, water and heat. Figure 1 (Porter and
McClure, in preparation) is a system of word equations
representing the mass and energy balances for an animal.
It shows the linkage between the physical environment,
physiological processes and their biological
consequences. We expect toxicants and diseases to affect
several variables described below.
HEAT BALANCE EQUATION
127
-------
Qsolar
ro
oo
mF,COg
KEY
O'HEAT
m* MASS
F • FOOD
('INGESTED
D» DEFECATED
A' ABSORBED
U'URINE
G* GROWTH
R' REPRODUCTION
W* WATER
S' STORED
•out
•f
Qconv
+
Qcond
•f
Q,
Figure 1. Coupled heat and mass balance equations (Porter
and McCl'ure, 1981) .
-------
The equation on the diagonal is the heat transfer
equation -for an animal. Once each term on the diagonal has
been evaluated using the physics o-f heat and mass transfer
(13,14,15). The equation can be solved -for body
temperature or metabolic rate. The variables, Dmetab and
CQevap, are the heat generation and loss terms due to
physiological processes. They are also influenced by
environmental heat inputs and losses. Examples of
environmental heat inputs are solar and infrared
radiation, and losses, Qsolar and QIR,in. An example of
heat loss is infrared emission from the animal, DIR,out.
Heat exchange also occurs by the mechanisms of convection
and conduction, Qconv and Qcond. The storage of heat, Qs,
may be either positive, negative or zero, depending on
whether body temperature is rising, falling, or at steady
state. The physiological variables Qmetab and Qevap each
have a flow of mass associated with them. The top
horizontal equation is the flow of mass due to the dry
component of food. The bottom horizontal equation is the
mass flow of water.
MASS BALANCE EQUATIONS
The mass of food ingested less that defecated is the
amount of mass absorbed (mF,I - mF,D = mF,A). Absorbed
mass that is oxidized (symbolized by the slash) produces
heat (Qmetab), metabolic water, (mF,H20), protein
decomposition products (mNH2+), and carbon dioxide,
(mF,C02). The remaining mass may be used for activity,
growth, reproduction or stored in the form of fat. The
metabolic rate requires a second mass flow other than
food, namely oxygen, which accepts hydrogen ions at the
end of the cell mitochondrial respiratory chain to
synthesize water (mF, H20).
The mass of water ingested less that lost in
defecation equals water absorbed
-------
into account. By modeling the total resources assimilated
(left hand side in Figure 1) and the obligatory heat and
mass losses
-------
ENDOTHERM
O
o
u.
Maximum Assimulation
(floppy lid)
Maximum Available Mass
Minimum Metabolism
(floppy dish)
Low
High
'environment
(a)
Max.- ^r-——_
o
en Min.
CO
Maximum Assimulation
(floppy lid)
Actual Available Mass
Minimum Metabolism
(floppy dish)
Low High
'environment
(b)
Figure 2. Two dimensional plot of maintenance and available
food mass vs temperature.
131
-------
FOOD FOR
MAINTENANCE
FOOD
REQUIRED
FOOD AVAILABLE FOR
GROWTH, REPRODUCTION,
ACTIVITY, OR INTERNAL
STORAGE
MAX. PROCESSING
CAPACITY
Figure 3. Three dimensional plot o-f potential growth or
reproduction vs climatic variables o-f air temperature
and solar and thermal radiation (Porter and McClure,
1981).
132
-------
increase's. The maintenance surface can be visualized
starting with the solid diagonal line from the "cold"
corner to the "hot" corner. The diagonal is the minimum
metabolism curve of Figure 2, where air temperature equals
radiant temperature. This is the standard environment
•found in metabolic chambers. When it is cold, metabolism
and -food required is maximal (cold corner). As
environmental temperatures increase, metabolism decreases,
until the "flat" thermoneutral zone is reached. As
temperatures continue to increase, the thermoneutral zone
is traversed, and metabolism begins to increase with heat
stress as the "hot" corner is approached. The remaining
dish surface is the food required when air and radiant
temperature are not equal. The edges of the surface that
drop off represent lethal conditions. New coupled
mechanism fur models developed by our group at Wisconsin
were instrumental in .generating surfaces like this
(18,19,20,21).
The upper surface farming the top of Figure 3
represents the maximum food processing capacity of the
animal given unlimited food. This physiological lid is
represented as being flat, but drops somewhat at high
temperatures for deer mice, Figure 4. Actual consumption
may not. reach the lid unless the animal is a lactating
femaTe (22, 23),. is exceptionally active, is severely cold
stressed (17), or laying down fat for hibernation.
The space between the two surfaces, the availability
space, represents the "uncommited" food mass which can be
used for growth, reproduction, storage or activity. The
rules governing the trade-offs between growth,
reproduction, storage and activity are still largely
unknown.
We have developed models of the maintenance surface
for fur-covered endotherms. These models have three
substantial improvements over the models used to generate
the climate space of (24). First, the convective
component for the floppy dish was adjusted for Reynolds
number range (a combined velocity, body size effect) and
the convection coefficient was adjusted for turbulence
outdoors, which may affect heat transfer by 50"/. or more
(25,26). Second, the fur heat transfer model is radically
different. The fur model for the climate space was only a
conduction model based on still air conductivity. We now
know that radiation losses through the fur can account for
SOX or more of the heat loss (27). Also radiative and
conductive heat losses through fur are coupled in complex
133
-------
o
•o
T
o»
5
TJ
O
JQ
UJ
CO
o
o
o
or
id
UJ
H-
O
i i i i i i i
30%
50%
I
i }
(2) (6)
1
(6)
Peromyscus maniculatus
Lactating Females
21 Day Average
Litter Size = 4-5
20% Relative Humidity
Steady State
i
i
12 18 24
AIR TEMPERATURE (°C)
30
36
Figure 4. Maximum ad lib energy consumption during
lactation vs air temperature. Radiant temperature
equals air temperature.
134
-------
ways that we have recently have been able to model
(19,20,21). Third, we have developed a model that relates
body temperature, limb blood flow and posture to skin
temperature. Using these improvements, we can now predict
metabolic rates of deer mice to within +/- 67. of measured
values in metabolic chambers (Figure 5a, Conley and
Porter, in preparation). Individual variation in data
points can now be identified as due to variation in body
temperature and limb blsed 'flaw., Figure 5b illustrates
postural effects on heat loss at different temperatures
(Conley and Porter, in preparation).
EXPERIMENTAL METHODS
The deer mOUSe., Ptaynmygr-i na man i <-| il at-1 id.. is a
ubiquitious small mammal living throughout North America.
It naturally reproduces in a wide variety of climatic
conditions from Mexico to Alaska and from sea level to
14,000 feet. It also reproduces readily in laboratory
conditions and thus is an ideal animal for our studies of
toxicant, disease, and climatic effects upon growth and
reproduction.
Food consumption is a discontinuous,
process whose basic rate controlling mechanisms
understood. Therefore, we are approaching this
our research in a more empirical fashion than
heat and water loss.
voluntary
are poorly
phase of
we did for
In an ongoing series of experiments we have been
measuring food and water consumption of growing mice,
non-pregnant adults, pregnant and lactating females under
a variety of experimental conditions (B,9), Figure 4.
We have been measuring the energy allocated to
reproduction in terms of litter size, birth weight, and
growth of young to weaning. By exposing a female and her
young to various temperatures and humidities, we can move
the maintenance costs up or down. By restricting food or
water, introducing toxicants or disease, or by varying
food quality (i.e., presence or absence of sprouts) we can
raise and lower the assimilation level. This gives us a
direct measure of how climate, toxicants, disease,, food,
and water interact to affect growth and reproduction
(Figure 2).
Designs we have used for studies of growth and
reproductive capacity of pregnant and lactating females
135
-------
-2
> i u -
6
7
o>
*
*s
LL) *^
h-
OL
0 4
O
m
K 3
2
E 2
U
Q.
M 1
°c
ill)
•
•
© ®
•
_ •
m
- -
•
•
•
•'
©
• ®
(•)
I l i i .
) 8 16 24 32 40
TOO
A. C
(a)
136
-------
7x10
-2
cc
o
o
CD
UJ
o
u.
= 2
o
UJ
Q.
(0
I I I I I I
THERMOREGULATORY HEAT TRANSFER
Peromyscus maniculatus
PREDICTED
• - Cylinder
A-Sphere
EXPERIMENTAL
This Study (Restrained) -
Wickler 1980
(Unrestrained)
0 5 10 15 20 25 30 35 40
CORE-AIR TEMPERATURE GRADIENT (TC-TA),°C
(b)'
Figure 5. Comparisons of observed and predicted metabolic
rates for different limb blood flow rates and postures
vs air temperature
-------
are shown in Tables 1,2,3,and 4. Justification for these
designs is given in (29,29,8). These highly efficient
fractional factorial designs are used iteratively to
establish first, the role of available food and water and
sprouts on growth and reproduction at room temperature
(Table 1). Then additional variables are added (Table
2'). Then we manipulate levels of significant variables
that show up (Table 3). Finally we get complete
resolution of significant variables (Table 4), We have
found that food and water shortage have their greatest
impact during lactation. Temperature and temperature
oscillation also affect survivorship, but not weights of
young when weaned.
The females and young for these experiments have been
randomly selected from our breeding colony of
approximately 1000 individual mice. The colony is annually
augmented by captured wild deer mice. The mice are
screened for disease and their offspring are used for
outbreeding in the colony.
The experimental design for simultaneous effects of
climate, disease and toxicants on growth and reproduction
potential in deer mice is illustrated in Table 5. The
results of this experiment were presented at the meeting
in Duluth and showed that. mice are vulnerable to the
disease and immune suppressant only when low on food or
water or both. When ad lib food and water are present,
the animals compensate by consuming more food than healthy
animals and are able to raise young to weaning as
successfully as animals with no disease or immune
suppressant. Preliminary indications are that serious
effects of this disease and toxicant show up at times of
food or water shortage and are most severe when both food
and water are in short supply. Thus screening tests done
on non-reproducing animals fed ad lib food and water may
likely give misleading predictions that the agents are
harmless.
138
-------
TABLE 1. INITIAL DESIGN FOR ASSESSING
THE EFFECTS OF FOOD AND WATER ON
DEERMOUSE GROWTH AND REPRODUCTIVE SUCCESS
Var i abl e
1
2
•TJ
4
Amount
Type of
Amount
Sprouts
of
food
dry
of
food
water
<-) Levels (•+•)
807.
Regul ar
607.
None
Ad
1
i
b
Breeder
Ad
Ad
1
1
i
i
b
b
Center
Repl i cate
907.
1/2 ?<
807.
507.
1/2
Variable
1234
Run #
1 - -
2 -f- - +
3 - + - +
4 + +
5 + +
6 + +
7 - + . +
a + + + +
T39
-------
TABLE 2
DESIGN FOR PRECOPULATION TO
WEANING EXPERIMENT IN THE BIOTRON
.mani CLI! at us 1
£« MM « ~ H — 3S 55 SS 52 —
Vari able
1 Temperature time
dependence
2 Temperature (C)
3 Relative Humidity
4 Number of young left
with mother
5 Nest box
6 Amount of food
7 Amount of wheat sprouts
8 Amount of water
<-) Levels <•+•>
Center
Repli cate
Steady
state
11<+/-5)
907.
No
807.
None
807.
Oscillating 1/2 Osc
17.5 •+•/- 2.5
507.
No roof
907.
507.
907.
23
207.
4
Yes
Ad lib
Ad lib
Ad lib
Vari able
123*
124*
134*
234*
Run
1
3
4
8
o
7
8
9
10
11
12
13
14
15
16
-t-
•+•
+
-t-
*Specific three factor
effect of the variable
interaction con-founded with the main
140
-------
TABLE 3. SECOND ITERATION OF DESIGN IN TABLE 2
i
2
"*•»
4
5
6
7
8
Run
1
2
3
4
5
6
7
0
9
10
11
12
13
14
15
16
= := =
Variable <-)
Temperature time Steady
dependence state
Temperature (C) ll<4-/-5)
Relative Humidity 907.
Number of young left 2
with mother
Nest box No
Amount of food 807.
Amount of wheat sprouts None
Amount of water* 607.
Variable
123
4-
12345
#
_ - _ _ _
+ - - - +
_.+.__ +
•+•+---
+ _ +
+ _+•__
-+•+•--
•f + + - +
„ _ _ +
+ - - + +
4. _ + +
•+••+•-•»•-
--•+•+• +
4. - + +
- + + + -
"(•+•+• + •+•
= = =:=:=: = = =; = = =:=;=:=: = =:=: = = =:=: = =: = = =: = = = = =:=: =
Center
Levels <+) Replicate
Oscillating 1/2 Osc
23<+/-5) 17.5 +/- 2.5
207. 507.
4 3
Yes No roof
Ad lib 907.
Ad lib 507.
807. 707.
124 134 234
•+••+•+•
67 a
_ _ _
+ +• -
+ - 4-
4- +
+ +
+ - 4-
4- + -
_ _ _
+ 4- +
- +
- + -
+ - ~
+ _ _
4-
- - 4-
4- 4- 4-
S3 = =i = !=ss = =: = = 5::= = = i=: = = = = = =:=5=5 = = = =:
The blocking pattern is the same as in previous runs.
#The levels of this variable are the only ones diffrartsmt -from
Table 2
141
-------
TABLE 4. THIRD ITERATION OF DESIGN IN TABLE 2
Center
Variable <-) Levels < + > Replicate
1 Temperature time Steady
dependence state
2 Temperature (C) ll(-t"/-5
3 Relative Humidity 907.
4 Number of young left 2
with mother
5 Nest box No
6 Amount of food 807.
7 Amount of wheat sprouts None
8 Amount of water 807.
Vari abl e
1
1234
Run #
1 _ _.
'•> j- _ _ _
±. ~
,^t "*" 4* "~ "™*
4 + *
KT _ ^ ,w
i i. i
O **• —• "r
7 H- 4-
8 + H- •+• -
(3 _ _ ._ 4.
10 + - - +
11 - H- - +
12 ++-•+•
13 - •+• +
14 4- - + +
15 + + •+•
16 + -l- + +
# Signs for the level of food admin
from Table 2 for full resolution of
with food
-------
TABLE 5. FIRST EXPERIMENTAL DESIGN TO ASSESS MAIN
EFFECTS AND INTERACTIONS OF FOOD, WATER, SPROUTS,
AN IMMUNE SUPPRESSANT AND A DISEASE
Var i abl e
1
2
T
4
5
Amount
Amount
Amount
Immune
Vi rus
of
of
of
food
water
sprouts
suppressant *
**
<-) Levels <•+•)
807.
(307.
None
No
No
Ad
Ad
Ad
20
ID
lib
lib
lib
mg/kq
100
Center
Repl i cats
907.
907.
507.
10 mg/k
ID 50
g
Vari able
Run #
1
1234
4 + +
5 - - 4-
7 -•+••+•
8 + + +
o ._ _ —
10 +
11 - +
12 •+• +
13 - - +
14 •+• - +
15 - + +
16 + + 4-
"*•" 'T*
j, Tl-l
•+• 4-
* Cyclophosphamid
** Venezuelan equine encephalitis (vaccine strain)
ID100 means infectious dose yielding 100'/. infected
ani mals
143
-------
Re-f erences
Porter . W. P. , J . W. Mitchell. W. A. Bee: kman and C.B.
DeWitt. 1973, Behavioral implications of mechanistic
ecology. Oecoiogia 13: 1—54.
Tracy,. C.R. 1976. A model of the dynamic exchanges of
water and energy between a terrestrial amphibian and it;
e n v i r o n m e n t „ E c a 1 „ ri o n o a . 46 (3) r, 2 9 3-3 2 0 „
J a m e s. I- . U. a n d W. P. P o r t e r . 19 / 9. behavior-
M i croc 1 i rn a t e r e 1 a t. ion s h i p s i n t h e A f r i c: a n R a i n b o w
1i z ar d, Ag a. ma a g am a. Cop e i a 1979(4); 5S3-593„
4 Ws.l dschmi dt, 3. and C.R. Tracy. IvSl. Interactions
between a lizard and its thermal environments
Implications -for sprint performance and space
utilization in the lizard Lita stansburi_ana. Ecology
(submitted).
5 M u t h ,, A. 19 7 7,. T h e r m ore g u 1 a t o r y p o s t u r e s and
orientation to the sun: A mechanistic evaluation for
the zebra—tailed lizard. Cal^lisaurus dra.coQoi.cles.
Cooeia 1973(4)s 710-720.
6 Wai dschmi dt, S. 1980,, Orientation to the sun by the
i g u a n i d lizard s U t a s t a n s b u r i a n a and S c e 1_ g D o r u s
yQdul.atus: Hourly and monthly variations. Copeia
19 S 0 (3) s 4 5 S - 4 6 2.
Por t er, W.P. and C.R. Tracy. 19Q0„ Biophysical an a1y sei
of eniergeti cs, time-space utilization, and
distributional limits of lizards. in Lizard Ecology
Symposium, Huey, Schoener and Pianka. eds. Harvard U.
P r e s s ( i n p r e s s) .
8 Porter, W.P. and R. L. Busch. 1973. Fractional
f a c t o r i a 1 a n a 1, y s is o f g r o w t h and w e a n i n g s u c c e s s i n
P e r o m v s c u s m a n'i c u 1 a t u s, Science 2 0 2: 9 0 7 - 910.
144
-------
9 Jaeger, J.W. and W.P. Porter. 19bl. Deer mouse growth
or r epr od LIC t i ve potential I: Life history
characteristics, J. Mammal, (submitted)
10 Reichert, S. and C.R. Tracy. 1975. The roles of
thermal environment and prey availability on the
dispersion of the desert spider, Agengl.gp_si.s QE§.rta
(Getsch.). Ecology 56(2): 265-2847'""'"
1.1 Kingsolver, J.6. 1979. Thermal and hydric aspects of
environmental heterogeneity in the pitcher plant
m o s q u i t o. Ec o1. Monog. 4 9(4) s 357-376.
12 KitcheH, J.F.,, D.J. Stewart and D. Weininger. 1977.
A pplications of & B i o e n e r g e t i c s M o d e 1 t o V" e 11 o w F' e r c h
^P-§C£S f_.l.§yss;cens) and Walleye (S/ti.^gstedi.gn YjLtreuni
yi.treuni} „ -j". Fish. Res. Board Can. 34(10):
1922-1935.
13 Bird. R.B., W.E. Stewart and E.N. Light-foot. 1960.
Transport F'hen omen a. Wiley and Sons. New York,, 7 8 Op.
14 K'reith, F. 1965. Principles of Heat Transfer. Int.
Textbook Co. Scranton,, F'a. 620p.
15 Gates. D.M. 1980. Biophysical Ecology. £pringer-Ver1ag,
Mew York. 61Ip.
16 Brady, 3. 1945. Bioenergetics and Growth. Reinhold
P LI b 1 .. C o r p „ 9 5 3 p .
7 Kendei gh, b.c. 1949. L-ffect of temperature and season
on energy resources of the English Sparrow. Auk 66:
i13—127.
IS Kowalski. G.J. 197B. An Analytical and Experimental
Investigation of the Heat Loss Through Animal Fur.
Ph.D. Thesis, Dept. Mechanical Engineering, University
o f Wi s c on s in, Madison.
145
-------
.9 Kowal ski ., 6.J. and J,W. Mitchell. 1979, An analytics]
a n d e x p e r i m e n t a 1 i n v estigati o n o f t h e h e a t t. r a n s f e r
m e c h a n i s m s w i t h i n f i b r o us m e d i a. T r a n s. A S i'*1 E P a p (-2 r N o,.
79-WA/HT-40: 1-7.
Kowalski, 6.J. and J.W. Mitchell. 1980, An
experimental investigation of the convective heat
transfer mechanisms within a fur layer,. Trans. ASME
8O-WA/HT-2S.
:i Gebremedhin, !<., C.O,, Cramer and W,P, Porter,, 1981,
Predictions and measurements of heat production and
food and water requirements of Hoi stein calves in
different environments,, Trans. Am. Soc. Ag. Eng«,
24 (3) ;: 715 720.
22 Randolph (McClure), p.A., J.C. Randolph, K. Mattingly,
and M.M. Foster,. 1977. Energy costs of reproduction in
the cotton rat, Si.gmgdgn hi.!iBi.dus. Ecology 58: 31-45,,
McClure, P.A. and J.C. Randolph, I960,, Relative
allocation of energy to growth and development of
homeothermy in the eastern wood rat ''Ng.ot.gma f.lgr.i.da)
and hispid cotton rat (Si_gmgdgn hi.sp.i.dus) . Ecol „
Moriog ,. 50 (2) s 199-219.' ""
24 Porter, W.P. and D.M. Gates. 1969, Thermodynamic
equilibria of animals with environment, Ecol. Monog,
'39: 227-244.
Pear man, G.I., Weaver, H.L. and C.E<. Tanner, 1972,
Boundary layer heat transport coefficients under field
conditions. Aq. Meteor. 10(1-2): 83-92.
Kowalski. G.J. and J.W. Mitchell. 1976. Heat transfer
from spheres in the naturally turbulent, outdoor
en v i r on men t. J. He a t Tr ansf er 98(4); 649-653,
146
-------
7 Skuldt. D.J., W.A. Beckman, J.W.Mitchell and W.P.
Porter. 1975. "Conduction and Radiation in Artificia]
Fur" in Perspectives o-F E,, A review o-F response
s u r -f a c e m e t h o d o 1 o g y f r o m a ta i o m e t r i c v i e w p c i n t „
Biometrics 31; S03--S51.,
Box, G.E.P. W.G. Hunter and J.S. Hunter 197E
S t a t i s t i c: s f c r E ;•; p e r i m e n t e r s,. Wiley a n d S o n s
York, 653 pp.
i Mew
147
-------
Comments on
Impacts of Toxicants, Disease, and
Climate on Growth and Reproduction
Using Peromyscus maniculatus
Nancy L. Stanton
University of Vifyoming
The objectives of this research include the following:
1) To develop a quantitative model to assess impacts of
toxicants, pathogens and climate on small mammal growth and
reproduction in the Laboratory and in the field.
2) To identify physiological parameters that can be used
a) to screen toxicants in the laboratory and field, b) to assess
mechanisms of interaction where synergistic effects are found.
These objectives are particularly relevant to EPA objectives since
the goal of the research is not only to understand an animal's response
to the stress of toxicants and disease but to predict the response.
There is a critical need to make realistic predictions concerning the
effect of a host of potential toxicants on a variety of animals. Standard
toxicity tests are not adequate. For one reason, the laboratory results
are not easily translatable to natural populations because of the complexity
of biotic and abiotic stresses that field populations are exposed to.
Thus, the capability of predicting synergistic effects will be of particular
value.
Theoretically, the model proposed is applicable to any endothermic
(and with modification, ecothermic) animal. However, the choice of the
deer mouse is judicious: 1) the deer mouse is ubiquitous and abundant,
2) it can prosper in highly disturbed habitats such as mine spoils and
croplands, and 3) thus it is exposed to a variety of chemical and physical
insults. There are probably field populations living under every type
of condition imposed by our chemical society. The deer mouse may be
viewed as the field version of the white rat. This mechanistic model
based upon heat and mass transfer equations (if validated) should be
able to predict population fitness of natural populations of deer mice
exposed to a variety of contaminants in combination with natural stresses
imposed by disease—and over a wide range of altitude and latitude.
The idea is exciting! The approach is innovative! And the goals, if
achieved, will contribute both to our understanding of species distribution
and abundance and to how these parameters are affected by both natural
and man-created stresses.
148
-------
Finally, this research exemplifies good, solid science. There is a
theoretical and quantitative model. The methodology is both statistically
sound and efficient, using full and fractional factorial designs, and
provides rigorous testing. If the model is validated, the theoretical
framework is broad enough to permit prediction to be made under a wide
range of circumstances. It is good science with a high utilitarian
potential.
149
-------
IMMUNOTOXICITY IN PEROMYSCUS MANICULATUS
R. D. Hinsdill, L. J. Olson and D. Weltman
Environmental Toxicology Center and Department of Bacteriology
University of Wisconsin
Madison, Wisconsin 53706
ABSTRACT
Both Swiss-Webster outbred white mice and Peromyscus maniculatus were
exposed to a number of plant growth regulators and known immunosuppressive
drugs at different levels by two routes. A number of immunological para-
meters were monitored, including plaque-forming cell counts in the spleen,
lymphocyte counts, hematocrit, total white blood cell counts, antibody titers
and weights of various organs as well as total body weight. The doses used
(mg/kg/day) were as follows: gibberellic acid (GA), 25, 50, 100; chlorocho-
line Cl (CCC), 1.0, 2.5, 5.0; triiodobenzoic acid (TIBA), 30, 45, 60; maleic
hydrazide (MH), 35, 70, 140; Dinoseb (DNBP), 0.4, 0.7, 1.4; and cycloheximide
(CH), 3.5, 7.0, 14.0. Cyclophosphamide (CP, 10 mg/kg/day) and azathioprine
(AZ, 10 mg/kg/day) served as known immunosuppressant controls. CH, CCC and
TIBA most strongly and/or consistently immunosuppressed the mice. CH lowered
PFC levels by 60%, reduced hemolysin titers by nearly 50% and lowered thymus
weights. CCC and TIBA had less affect but decreased up to 6 different para-
meters; however, only CH approached AZ or CP suppression levels. DNBP and
MH had mixed results. GA, conversely, caused moderate overall enhancement
in white mice. On the basis of these results certain compounds and exposure
levels were suggested as appropriate for further studies both in Dr. Yuill's
laboratory and in the integrated studies supervised by Dr. Porter.
INTRODUCTION
SELECTION OF AGENTS FOR STUDY
A large number of agents were considered for use in developing and tes-
ting the model described by Dr. Porter, e.g. heavy metals, pesticides, di-
oxins, etc. After considering all aspects, including safety for workers,
the decision was made to study a selected group of chemicals known as plant
growth regulators and to include as known positive-type controls two immuno-
suppressive drugs which are effective orally and parentally. Plant growth
regulators currently account for only a small share of the total worldwide
agrochemical market (5-10%). However, a number of experts and business
. 150
-------
organizations feel that these growth regulators are still in their "infancy",
compared to the pesticide market, and that within a few years production and
usage will outstrip that of the herbicides. For practical purposes, plant
growth regulators can be defined as organic chemicals, other than nutrients,-
that are applied directly to a plant to alter its life processes or structure
in some beneficial way, i.e. to enhance yields, improve quality or, particu-
larly important, to facilitate harvesting. Certain herbicides, when applied
in small amounts to induce a specific beneficial change can also be consid-
ered plant growth regulators. Some specific examples of usage of plant
growth regulators include flowering agent, color enhancer, dormancy breaker,
abscission stimulant, fruit-drop controller, fruit-thinner, ripener, and
increaser of pod set in soybeans. No one at present can predict accurately
to what extent these substances will ultimately contaminate food or our aqua-
tic environment.
At present there are roughly thirty different plant growth regulators
being marketed by various chemical manufacturers. Certain of the regulators
are produced by more than one chemical company and marketed under different
trade names, e.g. maleic hydrazide (MH-30, Uniroyal; Sprout Stop, Sucker
Stuff, Drexel; De-Cut, Fairmount). Approximately thirty chemical manufac-
turers in the U.S. are producing these growth regulators and include such
large firms as Dupont, Chevron, Uniroyal, Monsanto, Aldrich, Dow and Ameri-
can Cyanimid. With the continuing need to increase efficiency and decrease
costs in agriculture, it is inconceivable that these products will not be
heavily promoted. Unfortunately, a literature search reveals that only very
limited information is available on the potential toxicity of these compounds
for humans or animals. To the best of our knowledge few have ever been
examined for their immunotoxicity, nor have they been tested for interactions
by the methods proposed here. Table 1 lists six plant growth regulators
selected for initial study.
IMMUNOTOXICITY
Toxic agents may obviously affect a variety of organs and tissues, and
depending on how such agents are metabolized, affect different species to
varying degrees. Almost any agent, however, which would significantly
suppress the host defense mechanism, including the immunologically specific
cells, could have serious consequences in the animal's ability to deal with
foreign material, particularly pathogens. Although chemicals in the environ-
ment will continue to be screened for their mutageni.c, carcinogenic and
teratogenic effects, many investigators are now arguing that these same sub-
stances should be screened for immunotoxicity as well. There the consensus
ends, however, for there is no simple test or battery of tests which can be
relied upon to show the exact nature and degree of immunotoxicity. Part of
the reason is the complexity of the immune system as illustrated in Figure 1.
Not only are there a great many cell types involved, but there are a number
of stages in the immune response which could be affected (Figure 2). Even
when it is clearly established that a given compound influences a particu-
lar immunological parameter, one is rarely able to predict how this will
affect the body's ability to withstand challenge or carry on reproduction.
Our model is one of the few that considers this most important aspect of
immuno t oxi c i ty.
151
-------
TABLE 1. PLANT GROWTH REGULATORS AND CONTROLS TESTED
Uses
Crop Applications
Toxicity (Oral)
Azathioprine:
Chlorocholine Cl:
Cyclohexand.de:
Cyclophosphamide:
immunosuppressanc,
chemotherapy
ripener, flowering
agent, color enhancer,
lodging reducer
abscission agent
immunosuppressant,
chemotherapy
sugarcane, pineapple,
apples, cabbage, wheat,
barley, potatoes,
grapes, oranges
most citrus, cotton
LD
5Q
.5-l.Og/kg
(mouse)
133mg/kg
(mouse)
Dlnoseb :
Gibberellic Acid:
Maleic Hydrazide:
Triiodobenzoic
Acid:
herbicide, corn yield
enhancer
flowering initiator,
shoot stimulator,
fruit enlarger,
dormancy breaker
growth retardant,
sprout inhibitor,
lodging reducer
stem thickener,
flower induction
corn, many crops as
herbicide
grapes, cherries, peas,
lemons, wheat, rye,
lettuce, potatoes, oats,
oranges, beans
potatoes, tobacco, rye,
onions, sugarcane,
citrus
soybeans , apples
LD_n » 58mg/kg
50 (rat)
LD.- »25g/kg
50 (rat)
LD,n = 1400mg/kg
50 (rat)
LD,. = 813mg/kg
30 (mouse)
The first step in measuring immunotoxicity or in identifying agents
which may have a great biological impact via the immune system is to obtain
sufficient information to create an immunological profile similar to that
established for Cyclophosphamide and shown in Figure 3. Here we see that
not only is the production of IgM and IgG immunoglobulin suppressed, but
that IgE production has been enhanced. The latter type of antibody is
involved in "immediate type" anaphylactic (allergic) reactions and also con-
stitutes a serious threat to the animal. Measuring only immunoglobulin
levels may not be sufficient, however, as some animals with depressed levels
of lymphocytes may ultimately give normal antibody levels. For this reason,
we always include a measurement of the plaque-forming cell ability of the
spleen in order to assess whether there are a normal number of cells able
to respond to an antigenic stimulus. The assay procedure is summarized in
Figure 4.
RESULTS OBTAINED TO DATE
Because of the initial shortage of Peromyscus and because of the fairly
large amount of preliminary data needed to begin the multifactoral experi-
ments, outbred Swiss-Webster mice were used for the early experiments. Table
152
-------
ANTIGENS
\
MACROPHAGE
Bone marrow
Sensitized
B-Cell
T-Killer
Cell
V.
T-Effector
Cell
Activated T-Memory
Macrophage Cell
J
Plasma Cell
Cell-Mediated Immunity Memory
FIGURE 1, TYPES OF CELLS INVOLVED IN THE IMMUNE RESPONSE.
Humural
Immunity
153
-------
I, FORMATION OF LYMPHOID CELLS (BONE MARROW & LYMPHOID TISSUES)
II, MATURATION OF LYMPHOID CELLS (B's AND T's)
III. PHAGOCYTOSIS AND ANTIGEN PROCESSING
IV, ANTIGEN RECOGNITION BY B AND T CELLS
V, BLAST TRANSFORMATION AND PROLIFERATION
VI, ANTIBODY AND LYMPHOKINE PRODUCTION
VII, RETENTION OF MEMORY CELLS ( B's AND T's)
VIII, INDIRECT EFFECTS (HORMONES, GLUCOCORTICOIDS)
FIGURE 2, PHASES OF THE IMMUNE RESPONSE,
154
-------
1 ON TYPE III PNEUMOPOLYSACCHAfllOE ANTIBODY TtTERS
10 '00
•«/>•
IN. CPt * «41 Kf AbM/iM
2. ON TETANUS ANTITOXIN TITERS (IgGl
10 too )oo
3 ON PCA TITEHS (IgE) ANTIBODY TO TETANUS TOXOIO
Figure 3. Effects of various doses of intraperitoneally injected
cyclophosphamide on several classes of antibodies. Such
data constitute an important part of an "immune profile".
From R. S. Speirs, D. W. Roberts, R. D. Hinsdill and
E. E. Speirs, 1978, Immunotoxicity Assessment, Proc. 4th
FDA Science Symposium on Inadvertent Modification of the
Immune Response.
155
-------
1st injection of compounds
into ? Swiss Webster outbred
mice (IP in saline)
Each antibody producing lymphocyte
forms a clear area (plaque), count and
quantify to determine effect of
compound on PFC's
24 days
(2 injections/week)
Mixture transferred to modified
Cunningham Monoiayer Plate,
incubated .5-1.0 hour at 37°C
Separated spleen cells mixed
with SRBC's and complement
Inject .2 ml of 10% Sheep
Red Blood Cells as antigen, IP
Last compound
injection
4 day's
Spleen cells separated via
squeezing through nylon mesh
in culture media
Spleen to plaque forming
cell assay (PFC)
Sacrifice mouse, remove
spleen, liver and thymus,
take blood sample
Other parameters taken •
Bodywt. Hematocrit Hemolysin liter
Spleen wt. Total WBC Organs saved for histology
Thymus wt. Differential WBC Plasma saved for further
Liver wt. Plasma protein testing
FIGURE 4. SEQUENCE OF EVENTS IN THE PGR IMMUNOTOXICITY ASSAY.
156
-------
2 lists the six plant growth regulators and two known immunosuppressive
drugs (acting as positive-type controls) which were injected twice weekly
intraperitoneally (IP) for four weeks. To monitor the immunological func-
tions, sheep red blood cells (SRBC) were administered IP on day 24 and the
immunological assays were performed on day 28.
TABLE 2. AGENTS AND DOSAGES USED
V
Agent
Cyclophosphamide (CP)
Azathioprine (AZ)
Gibberellic acid (GA)
Chlorocholine chloride (CCC)
Dinoseb (DNBP)
Cycloheximide (CH)
Maleic hydrazide (MH)
Triiodobenzoic acid (TIBA)
Dosage
10
10
25, 50, 100
1.0, 2.5, 5.0
0.4, 0.7, 1.4
3.5, 7.0, 14.0
35, 70, 140
30, 45, 60
mg/kg/day
mg/kg/day
mg/kg/day
mg/kg/day
mg/kg/day
mg/kg/day
mg/kg/day
mg/kg/day
Injected IP twice weekly for four weeks
Figure 5 shows the effect on the number of PFC's per gram of spleen as
a percent of the control value. Note that azathioprine (AZ), gibberellic
acid (GA) and maleic hydrazide (MH) appear to stimulate this particular
function. This cannot be interpreted as either good or bad. Such findings
often indicate that such agents are capable of acting as immunomodulators
and that the specific effect at any given time will depend mostly on the
dose, but also on the timing of immunization relative to exposure and tes-
ting.
Figure 6 shows the effects of the same chemicals on total white blood
cell counts, Figure 7 the hematocrit values and Figure 8 the mean lymphocyte
counts per gram of spleen. Table 3 is an attempt to interpret and summarize
the immunological effects of these chemicals in the usual test setting.
Figure 9 compares feed studies with several agents in both Swiss-Webster
outbred mice and Peromyscus. On the basis of these experiments, CP at a
feed level of 20 mg/kg/day was recommended as the first toxicant for the
integrated studies in both Dr. Yuill's and Dr. Porter's laboratory. 10 mg/
kg/day will be used for the center replicate.
157
-------
240
180
2 160
•*—
o
o
H- 140
O
3 120
LU
CO
100
c
O)
180
CO
o
o>
o
u_
0_
IX
60
40
20
0
37
fl I.
i
CP AZ GA CCC DNBP CH MH TIBA
10 10 Z025 SO .00 LO Z.5 3.0 0.4 0° ..4 3.5 7.O ,4.0 3S TO !4O 3O 45 SO
Compounds and Doses Tested (in mg/kg/day)
FIGURE 5. EFFECTS OF PGR's ON PLAQUE FORMING CELLS (PFC's)XGRAM
OF SPLEEN ± SEM, EXPRESSED AS A % OF THE CONTROL (N=10),
158
-------
160 r
c
O
o
^
O
CD
120
100
UJ 80
60
IX 20
I I
CP AZ GA CCC DNBP CH MH TIBA
10 10 25 50 100 1.0 2.5 5.0 0.4 0.7 1.4 3.5 7.0 14.0 35 70 140 30 45 60
Compounds and Doses Tested (in mg/kg/day)
FIGURE 6. RESULTS OF PGR EFFECTS ON X TOTAL WHITE BLOOD CELL COUNT
(WBO+SEM, EXPRESSED AS A I OF CONTROL (N=10),
159
-------
I60r
j: 140
c
o
o
^o 120 -
0 100
'^^^
^^^
LU
+1 80
05
— *"T
>
® 60
o
o 40
o
£
V
x 20
IX
E*:l E*l Iwl t:'::l rSl rk I'Sq t$!:] c-H (S
|$-a j:?;
Kv*4 1*X'
fev'l "••'•'
9 I
^
-
:: *3? ESJ {-Si p&:j •*• [rrj
•* P'Pl n*ifi*i l"""""*3 ^T^
•* fctl *33 H*n •"•
^
;I
I;
I;
Jj
i
-
ii i i i i
CP AZ GA CCC DNBP CH MH
10 10 20 25 50 100 1.0 2.5 5.0 0.4 0.7 1.4 3.5 7.0 14.0 35 70 140
Compounds and Doses Tested (in mg/kg/day)
FIGURE 7, PGR EFFECTS ON X HEMATOCRIT LEVELS + SEM AS A % OF
CONTROL (N=10).
160
-------
IOU
0
•*"•
C
o 160
O
<*-
o
3^ 140
in
a
CO
+i
c 100
^^
(A
0)
•£ 40
u
O
J=
CL
£ 20
^
lx o
,
-
-
I
'.,
.
i
i
i
-
:
,
i
»
i
'
'
. •
'
":
:
:
f
V
i
j.
J
%
..
._
.
:
.
i
I
,
':
'
: '••
ft'
....
ri
•
i
i*
!
1
:•:•
j5:|
Jii
:$:
Sij
:?•:
:¥:
1
|
1
;§
ur MA t?A CCC DNBP CH MH TTRA
10 10 25 50 ,00 1.0 2.3 5.0 0.4 0.7 1.4 3.5 r" !4.0 35 7!? ,40 30 49 «
Compounds and Doses Tested (in mg/kq/day)
FIGURE 8, PGR EFFECTS ON X TOTAL LYMPHOCYTES/GRAM OF SPLEEN,
EXPRESSED AS A I OF CONTROL (N=10).
161
-------
TABLE 3. SUMMARY OF NET AFFECTS OF TEST COMPOUNDS ON VARIOUS PARAMETERS
COMPOUNDS TEST PARAMETERS
Cyc loph 03 ph amide
Azathioprine
Cycloheximide
Dinoaeb
Malelc Hydrazide
Triiodobenzoic
Acid
Glbberelllc Acid
Chlorcholine Cl
PFC's/gm Total lymphocytes/ PFC's/10 Hematocrit Total Plasma Hemolysin Thymus Spleen Liver
of spleen gm of spleen viable cells WBC protein titer Wt. Wt. Wt.
II 1 II t 1 1 n 1 1 1 1 A.
*K V T * T v V V ™ » V V \V \
t t 1 f •li ]f ^ - IJ/^0
1^ o || o to^V ll/ol"
1 0 v.y *V~ --00
T T 1 o|-^ >Jcoo
t t variable t t 0 0 0 0
J, ^ | 1tJ,|'-000
Body
Wt.
0
0
0
0
0
0
0
0
Legend : T T strong elevation
"f moderate elevation
0 no apparent affect
+ moderate suppression
\ i strong suppression
- data not available
(significance levels not yet determined)
162
-------
__
2 280 r
c
-------
QUESTIONS TO DR. HINSDILL (DR. ZEEMAN, DISCUSSANT LEADER)
ZEEMAN:
Why did you measure the hematocrlt? Is this measure of the red blood
cells of any immune significance?
HINSDILL:
We were simply looking at as many parameters as possible as we were
screening so that nothing, important would be missed. The relevence of a
drop in red blood cells to either cellular or humoral immunity is not clear,
except that it may indicate suppression of stem cells in the bone marrow and
thereby suggests other potential effects. The hematocrit was easily added
to the other tests and represented no hardship.
ZEEMAN:
Of the 6 test plant growth regulators, do you know to what extent they
will persist and perhaps accumulate in organisms in the environment?
HINSDILL:
No, that is something we did not address. One of our problems is know-
ing to what extent any one of these agents is going to be used. These agents
are just now coming into use and I think that within the next five or ten
years you'll see a marked increase in application. It was very hard for me,
since I'm not a soil scientist nor an organic chemist, to figure out what
kinds of factors would interact to prolong the "half-life" of a particular
plant growth regulator or decrease it. So the initial selection was made
more on the basis of chemical structure and whether we thought they were
likely to interact with DNA replication or the immune system. For purposes
of the model itself, it was not critical to select the most persistant
agents. What we wanted were some chemicals that we could plug into the model
to produce varying degrees of immunosuppression or immunomodulation and which
worked by different means. We must not lose sight of the fact that we are
trying to develop a predictive model which will give important information
about the biological impact of a particular chemical. Much more work will
be required to answer questions regarding persistance and bioaccumulation
of the plant growth regulators under a variety of environmental conditions.
KOCAN:
What effects other than immunosuppression are known to result from ex-
posure to the test compounds you propose to use? Do any of the compounds
164
-------
you are using require metabolic activation by the mice?
HINSDILL:
There have been a number of studies carried out with some of them.
Maleic hydrazide, for example, has been looked at for its ability to produce
carcinogenesis and changes in DNA synthesis. There is some confusion when
you go back and read the literature, because the results are dependent on
the type of test system used. It was difficult to reach conclusions as to
their biological impact. I presume that the usual kinds of safety testing
were done by the parent chemical companies before they marketed the product,
but you don't see much of this in the literature; again, it is very hard to
determine what they found or looked for in the two year or similar feed
study. We were primarily concerned with the immuno-suppressive ability of
these compounds rather than liver toxicity or what they would do in an Ames
mutagenicity test, but there is something in the order of 75 to 80 papers
that have been published concerning other biological effects of these same
chemicals.
FISHER:
Since certain immunosuppressants used in chemotherapy destroy platelets
and can lead to chronic internal bleeding, is this type of stress and re-
sponse being examined in your studies?
HINSDILL:
No, but none of the agents used to date have produced any overt inter-
nal bleeding detectable at the time of posting.
FISHER:
The reason I asked that question is that some of the other parameters
that youwill be measuring might affect the metabolism under certain condi-
tions, production of the enzymes required to metabolize inactive forms of
these agents might actually be shut off and thus the toxic effects would
not be observed.
HINSDILL:
Activation of potentially dangerous chemicals can be very important.
This is why we are doing feed studies primarily to make sure that the oppor-
tunity for activation in animals is present. There is a flaw and I will
admit it. Maleic hydrazide, for example, if tested by the Ames mutagenicity
test is not mutagenic, but if you first activate it with plant enzymes, it
is mutagenic in this test. Now I don't know where plant activation comes
into the picture, but it is obviously something we haven't built into the
model yet. You can certainly envision a situation where the plants would
take up the agents and activate them and then animals would eat the plants
and get the fully activated.agents.
165
-------
Comments on
Immunotoxicity in Percmyscus maniculata
Maurice G. Zeeman
Food and Drug Administration
I was asked by Dr. Taub to briefly discuss why Dr. Hinsdill's
portion of this three-part research project was worthy of interest and
support by the Environmental Protection Agency. I will attempt to do so
by first digressing for a moment into the subject of comparative immunology,
then by returning to Dr. Hinsdill's part of this entire project, and
finally by briefly discussing this research in relation to the entire
project.
Around the turn of the century, the Russian researcher,
Elie Metchnikoff, wrote a book while working at the Pasteur Institute in
Paris. In this book, Metchnikoff focused attention on the fact that
phagocytic cells were present and active inside all kinds of animals in
the environment, from water fleas to goldfish. He postulated the value
of these cells in resistance to disease. These phagocytic-type cells
now turn out to be a very basic part of an immune-system that appears to
be universal throughout multicellular animals. Something very much like
a cell-mediated immune response has been discerned in most of the invertebrate
phyla. In addition, basic research over the last 10-15 years has demonstrated
that mammalian-like cell-mediated and humoral immune systems have been
found to extend throughout all of the vertebrates.
Enough about comparative immunology, now back to Dr. Hinsdill's
work and the EPA. I have been aware of Dr. Hinsdill's interest in
Immunotoxicology for about three or four years now. His previous research
on the PCB's and dioxins in lab rodents was of interest to me because I
was conducting research on a persistent chlorinated hydrocarbon pollutant
and looking for its effects on the immune system of organisms in the
environment. .
Over the last decade, Federal support of research on Immunotoxicology
has often appeared to focus on research that had direct applications to
man and/or his economically important food species. To me, support of
Dr. Hinsdill's type of project by EPA indicates the recognition by a
Federal agency that organisms in the environment do have immune systems
that can be compromised by toxic agents. It is refreshing to realize
that there exists in the EPA an awareness that organisms in the environment
have immune systems which are probably just as essential to their health
and survivial as our immune systems are to our own health and survival.
166
-------
It is becoming more and more obvious that (under various conditions)
environmental contaminants can compromise some elements of the complex
(or simple) immune systems of animals residing in a contaminated environment.
Dr. Hinsdill's research has just started to examine the effects of
a group of six potentially widely used plant growth regulators on a
variety of humoral and cell-mediated immune parameters of the deer
mouse/ a rodent species widely distributed in nature. Attempts to make
the chemical exposures as much like those expected in the field will
eventually make this approach as natural as possible, given the constraints
of a laboratory-type setting.
Dr. Hinsdill properly emphasizes that very little information is
available on the potential toxicity of these compounds. Some of this
information is summarized in publications such as the NIOSH Registry of
Toxic Effects of Chemical Substances (1980) and in the Merck Index
(1976). But toxicity is just one of the relevant characteristics of
these chemicals that must be determined. How such chemicals act in the
environment can be very important in assessing their potential for
affecting various animals living in that environment. Are these chemicals
capable of transport throughout the environment? Are they persistent?
Do they bioaccumulate? DDT is not very acutely toxic to most mammals
and birds, yet does appear capable of causing some degree of immuno-
suppression in several vertebrate species. The persistence and bioaccumulation
potential of DDT would appear to endow this chemical (and other chemicals
with similar properties) with considerably more potential -for long-term
impact than does the acute toxicity. Knowing that the plant growth
regulator dinoseb is not tightly adsorbed to agricultural soils and can
leach into rainwater (Herbicide Handbook Committee, 1979), but that it
also appears to degrade quite rapidly in soil and on plants (Brown,
1978; Herbicide Handbook Committee, 1979) is probably just as relevant
to evaluating dinosebs potential for environmental effects as is knowing
it is highly toxic to several species of fish (Herbicide Handbook Committee,
1979; Johnson and Finley, 1980) and only moderately toxic to several
bird and mammal species (Herbicide Handbook Committee, 1979). All such
characteristics of chemicals may be of value in a first determination of
which specific chemicals have the most potential for initiating effects
(such as immunosuppression) in organisms in the environment.
In their research proposal, Dr. Hinsdill and the other investigators
briefly discussed the practical and theoretical environmental significance
of their overall project. The following three ideas on environmental
contaminants and the potential value of the project seemed worthy of
repeating.
First, this research could provide the basis for rapid screening of
the effects of toxicants in the context of climate and disease, both
common parts of natural systems. Second, this research should allow
predictions of the consequences of environmental contamination on
populations of animals at different times of the year and different
latitudes and elevations. And lastly, this project could assist in
167
-------
reclaiming environments that have been mined or otherwise contaminated,.
since the necessary climates and quantities of food and water needed for
persistence of species in a given area at all times of the year could
then be estimated.
Dr. Hinsdill's project is only one of three parts, all of which
will finally be integrated and synthesized into a model that will be
useful in examining the interrelated effects of climate, disease organisms,
and environmental chemicals on the growth and reproduction of a natural
rodent population. I support the concept of such a research project, am
pleasantly surprised at the volume of data generated in the six-month
span this project has been funded, wish the researchers continued luck
in their work, and finally applaud the EPA's vision in funding this type
of research project.
Brown, A.W.A. 1978. Ecology of Pesticides. John Wiley & Sons, Inc.
New York. P. 330.
Herbicide Handbook Committee. 1979. Dinoseb, pp. 173-177 in Herbicide
Handbook of the Weed Science Society of America, Fourth EdTtion.
Weed Science Society of America, Champaign, IL.
Johnson, W.W. and M.T. Finley. 1980. Dinoseb, pp. 32-33 ^n Handbook
of Acute Ttoxicity of Chemicals to Fish and Aquatic Invertebrates.
U.S. Dept. of Interior, Fish and Wildlife Service, Resource Publication
137. Washington, D.C.
Merck Index. 1976. An Encyclopedia of Chemicals and Drugs, Ninth
Edition. M. Windholz, S. Budvari, L.Y. Stroutsos and M.N. Fertig,
Eds. Merck and Co., Inc., Rahway, NJ.
NIOSH. 1980. Registry of Toxic Effects of Chemical Substances, 1979
Edition. R.J. Lewis and R.L. Tatken, Eds., U.S. Dept. Health and
Human Services, National Institute for Occupational Safety and Health,
Cincinnati, OH. Two volumes.
168
-------
NEW APPROACHES TO MEASURING THE BIOLOGICAL IMPACT OF
ENVIRONMENTAL TOXICANTS
III. HOST-PATHOGEN-TOXICANT INTERACTIONS: A PROGRESS REPORT
A. Glicken and T. M. Yuill
Department of Veterinary Science
University of Wisconsin
Madison, Wisconsin 53706
ABSTRACT
The interactions between host, pathogen, and toxicant were examined as
part of a study to determine the effects of physiologic stresses on growth
and reproduction of deer mice (Peromyscus maniculatus). Breeding colony
conditions are described. Two arboviruses which cause acute infections,
Northway virus and Venezuelan equine encephalitis virus (VEEV), were tested
for pathogenicity in deer mice. Northway virus caused no mortality, no
viremias, and a weak (£1:20) antibody response. VEEV caused significant
mortality (57%) at high doses and scattered mortality in all doses.
Measurable viremias (_>1:10) and antibody responses were present at all
doses. An ID^QO an<^ an ^50 were determined. The effect on the host
response to VEEV infection in mice continually fed 20 mg/kg of mouse of
cyclophosphamide was examined. Mortality rates were increased with longer
survival times, viremia onset showed no change, the duration of the viremia
was lengthened, and peak viremia titers increased.
MOUSE BREEDING COLONY
The colony is housed in a separate room in the semi-isolation facility
at Charraany Experimental Farms. The mice are kept in opaque polypropylene
boxes with stainless steel tops and wood shavings as bedding. They are fed
standard rodent lab chow and water ad lib. The light cycle is maintained
at 14 light:10 dark and temperatures are approximately 23°C. No attempt is
made to control humidity. Mice are weaned at 3 weeks and separated by
sex. They are paired for breeding at 60 to 90 days of age.
It was originally thought that maximum reproduction could be obtained
through the use of triad matings, one male and two females. However,
discussion with other investigators and personal experience has shown that
169
-------
pair mating is more efficient for producing the maximum numbers of litters
per month. Leaving the male with the female at all times assures breeding
at the post-partum estrus, thus producing a litter every 3.5 weeks.
We presently have 100 breeding pairs and expect to maintain this
number to assure a constant supply of experimental animals and replacement
breeders. Breeding success has markedly improved since switching to pair
mating and with the onset of spring (Figure 1). We are maintaining a
continual cull and replacement process to achieve highest possible
success. An attempt is made to keep inbreeding to a minimum and wild-
caught pairs are periodically introduced to the colony to increase genetic
heterogeneity.
co 50 mm
r 40 —
- 30 —
(A
, 20 —
i 10.
U.
triads
J
^
V
° III
D J F
pairs
l/^
^V
1 1
M A
-^
1
M
Month
Figure 1. Breeding success of deer mice kept as triads (December 1980 -
March 1981) and pairs (March - May 1981).
Dr. Glicken is in charge of medical surveillance of the colony. Any
animals dying from other than experimental causes are necropsied and
appropriate samples are taken for histopathology, bacteriology, and
virology. Records are maintained for all necropsies. So far, we have seen
one 3 year old female with a uterine prolapse and subsequent uremic
syndrome due to blockage of the urethra, and two old females with extensive
mammary tumors. One of these mice also had cerebral hemorrhage which had
caused central nervous system signs prior to euthanasia. Interestingly,
there appears to be no previous reports in the literature of deer mice with
mammary tumors.
Health screens have also been performed to establish what potential
pathogens are already present in our colony. In addition, all animals
170
-------
entering our breeding colony are bled and a serum sample is stored for
future reference in the event of a major disease outbreak. Serological
tests have shown the mice to be free of antibodies to 11 viruses commonly
found in laboratory mice (Table 1).
TABLE 1. VIRUSES COMMONLY FOUND IN LABORATORY MICE
Polyoma virus
Reovirus, type 3
Mouse poliovirus, strain GD VII
Ectromelia
K-virus
Sendai virus
Mouse adenovirus
Mouse hepatitis virus
Lymphocytic choriomeningitis virus
Pneumonia virus of mice
Mammary tumor virus of mice
However, 20 pairs of mice were recently received from the colony of Dr. C.
Terman, College of William and Mary, VA to help expand our colony. These
mice have low level antibody titers (1:10) to Sendai virus. Bacterio-
logical cultures of organ swabs and tracheal washings were negative for
Mycoplasma spp. and positive for Staphlococci spp. (lung, liver),
Streptococci spp. (tracheal wash) and Corynebacterium spp. (liver). These
bacteria were encountered in small numbers and their presence was not
associated with any lesions. All mice were negative for internal
parasites. Health screens still need to be done on a representative sample
of the wild-caught mice. Generally, we feel that we have a very healthy
colony.
The importance of this aspect of the work cannot be overstated. These
mice are quite different than normal laboratory white mice in terms of
growth and reproduction. Any serious viral or bacterial outbreak in one or
more of the colonies would seriously interrupt the breeding schedule and
could, if not detected, influence experimental groups in the combined
physiologic, toxic and pathogenic stress experiments.
HOST-PATHOGEN INTERACTION
Development of the model of the effects of physiologic stresses on
growth and reproduction of deer mice will require establishment of several
non-lethal bacterial and viral infections. The bacterial and viral
diseases of this species have not been well studied. We wish to use at
least two bacterial infections, four acute viral infections, and one
chronic viral infection in the model. Each pathogen establishes its own
171
-------
relationship with its host, which may be affected to varying degrees by
toxicants or environmental factors. Thus, the more pathogens used, the
more useful the model will be. We recently did a field study in northern
Alberta and have isolated a variant of Modoc virus from an apparently
healthy deer mouse. Adult deer mice infected in the laboratory with this
virus survived without apparent disease (1). We believe that Modoc virus
can be used to assess effects of chronic viral infection. Rodent-
associated (at least in part of their geographic range) viruses that will
be used for acute infection include Northway, Klamath, _C. gapperi agent,
Peromyscus virus, western equine encephalitis, and the TC-83 (vaccine)
strain of Venezuelan equine encephalitis virus. All but TC-83 were
isolated from wild North American rodents and all are sufficiently safe to
handle in the laboratory.
PATHOGEN SELECTION
The first virus selected for testing was Northway virus. Little is
known about this virus. It was the first arbovirus of the Bunyarawera group
to be found in Alaska (2) where the original isolation was from mosquitoes
(Aedes spp.) and sentinel rabbits (Oryctolagus cuniculus). McLean, et al
(1978) (3) at the University of British Columbia have been studying the
transmission of the virus by the Arctic mosquitoes, but no work has been
done on its infectivity and pathogenicity in mammals.
Stock virus was grown from seed stock obtained from Dr. Calisher,
Center for Disease Control, Fort Collins, Colorado. Two separate lots of
Alaska-origin virus were passed in suckling mouse brains. This was only
the second passage of the field isolation in a vertebrate so there is
little likelihood that the virus changed very much from its original
form. Both lots were titered in cell culture by a standard plaque assay.
Lot 1 had 2.2 x 10 plaque forming units (pfu) per ml and Lot 2 had
1.0 x 10 pfu/ml. Thus, infectivity of the second lot was about 10-fold
greater than the first.
The virus identity was confirmed with reference antiserum obtained
from the Center for Disease Control, utilizing the plaque-reduction test.
A working stock of hyperimmune mouse ascitic fluid produced for future
reference work titered 1:10 against 100 TCID^Q of virus.
Two experiments were performed to find a dose and route of virus
inoculation that would infect 100 percent of the mice while killing none.
In the first experiment, the first lot of virus was serially diluted
10-fold from 10~ to 10" . Each of six mice per dilution were inoculated
subcutaneously (SC) with 0.2 ml. They were bled daily for 7 days to
collect whole blood for virus isolation and then bled every 3-7 days for 4
more weeks to collect serum for antibody assay. Since there were no
mortalities other than those due to the trauma of bleeding, only the lowest
dilution group (10 pfu/ml) was tested for the presence of viremias and
antibody response. Testing of the whole blood for virus was attempted in
the microtiter cell culture system, but the red blood cells made the cell
cultures difficult to see. Therefore, the blood was assayed for virus by
172
-------
intracranial inoculations of suckling mice. Brain suspensions of any dead
or sick suckling mice were tested in cell culture to confirm the presence
of virus. Only 1 of the 6 deer mice had a measurable viremia—-with a titer
of 2.3 log^Q. Sera tested by microtiter serum neutralization (SN) showed
an antibody response in all 6 mice on days 8, 14, and 28 post inoculation.
The second lot of virus was injected SC as well as intraperitoneally
(IP), in an effort to achieve more rapid adsorption along with a depot for
slow release to maintain virus levels. Since Lot 2 had 10-fold more virus
than Lot 1, higher dosages were attained. Serial 10-fold dilutions were
made and injected into 7 mice per group. The same bleeding schedule was
followed as for Lot 1. Once again there were no mortalities so only the
lowest dilution group was tested. Whole blood diluted 1:10 was tested in
the microtiter cell culture system for the presence of virus without the
problem of visibility. Again, no viremias were found. Sera from mice
sampled on days 9, 14, and 20 post inoculation were titrated in the SN test
and found to have maximum titers of 1:10 for days 9 and 14 and 1:20 for day
20.
We concluded that Northway virus is not very pathogenic for deer mice
although the mice apparently become infected and are able to mount an
antibody response to it. It may be that deer mice are not natural hosts
for this virus, even though their range includes the parts of Alaska and
Yukon Territory where the virus has been found. Rather than trying to
adapt the virus by repeated passage in deer mice we considered it to be
more advantageous to work with another virus that is known to be.more
pathogenic. We sought a virus that would consistently induce a measurable
viremia in deer mice to determine if that viremia would be increased or
decreased when the mice were subjected to the stresses of toxicants and
climatic factors.
The next virus selected for testing was another arbovirus, Venezuelan
equine encephalitis virus (VEEV). Wild deer mice are infected by VEEV (4)
and deer mice infected in the laboratory produced good viremias and
antibody responses (5,6). These previous studies used a field strain of
virus which resulted in high rates of mortality as well. Therefore, we
elected to use the vaccine strain of the virus to find an infectious but
non-lethal dose. In addition, this strain is of lower pathogenic!ty to
humans and can be used in all phases of the study with minimal risks and
handling restraints.
A working stock of virus was grown in VERO cells and the usual
virological standardization tests were performed. The working stock .had an
infectivity of 8 x 10 pfu/ml and a tissue culture 50% infective dose
(TCID^) of 5.58 logiQ. Suckling mouse intracranial inoculations resulted
in a SMICLDcQ of 4.83 log^Q. Thus, the virus is about 10 times more
pathogenic for suckling mice than for cell cultures. A positive plaque
reduction test with reference VEEV antiserum verified the identity of the
virus and a supply of rabbit hyperimmune serum was produced for future
reference.
173
-------
The virus was serially diluted 10-fold to contain 8 x 10' to 8
pfu/ml. Five to 10 mice per dilution were each inoculated SC and IP with a
total dose of 0.4 ml. This dose caused mortality in all dilutions except 8
pfu/ml (Table 2a). Tissue culture raicrotiter tests showed viremias in 100
percent of the mice in dilutions 8 x 10, 8 x 104, 8 x 103, and 8 x 102
pfu/ml (8 x 10 and 8 x 10 were not tested). Eighty percent of the mice
in the 80 pfu/ml dilution had measurable viremias (_>_!: 10) as did 50 percent
of those in the 8 pfu/ml dilution. Antibody titers were determined at
selected intervals from 8 to 39 days post inoculation for mice receiving
dilutions 8 x 104 and 8 x 103 pfu/ml (Table 3). VEEV antibodies were
present beginning on day 8, with titers ranging from 1:10 to 1:80, with
peaks on days 19 to 24.
TABLE 2. NUMBERS OF DEER MICE (_P. MANICULATUS) DYING POST INOCULATION WITH
0.4 ML SC AND IP OF VEE VIRUS STRAIN TC-83
Virus
dilution
(pfu/ml)
8 x 107
8 x 106
8 x 105
8 x 104
8 x 103
8 x 102
80
8
Virus
dilution
(pfu/ml)
8 x 104
8 x 102
80
8
a. No cyclophosphamide in feed.
Days post inoculation
7 8 9 10 ... 18 20 21-62
11 1 1
2
3
1
1*
1
1
b. Cyclophosphamide in feed, 20 mg/kg
Days post inoculation
789 10 11 12 17-21 31
11 1 1
1 2
1 . lf 1
Total
Dead
4
2 .
3
1
1
1
1
0
mouse.
Total
Dead
4
0
3
3
number
Total .
7
6
7
7
7
10
10
8
Number
Total
10
10
10
10
%
Dead
57
33
43
14
14
10
10
0
Dead
40
0
30
30
* Virus isolated from brain
t Missing, presumed dead
174
-------
TABLE 3. MEAN* VIREMIA AND ANTIBODY TITERS IN DEER MICE (JP. MANICULATUS)
INOCULATED SC AND IP WITH 0.4 ML VEE VIRUS STRAIN TC-83
Days post
inoculation
1
2
3
4
5
6
7
8
10
14
19
24
32
39
N
8 x l'07
141*
89
141
43
0
0
0
§
§
§
§
§
§
§
6
Virus dilution (pfu/ml)
8 x 104
Viremia
91
16
33
3
0
0
0
Antibody
4
15
35
52
56
34
25
5
8 x 103
46
2
4
0
0
0
0
6
13
8
10
11
8
15
5
* Geometric means (x )
§ Not measured
t Reciprocal of dilution
SC and IP inoculation of 0.4 ml of the 8 x 10 pfu/ml dilution of this
working stock of virus was selected as the test dose for the mice in the
physiological stress experiments. This dose infected all of the mice with
measurable viremia titers and relatively high antibody production. It
appears that some mortality can be expected at any dose (Table 2a),
probably due to occasional "supersusceptible" mice that are unable to
contain the infection regardless of the size of the initial dose. The
8 x 10 pfu/ml dilution also infected all mice but was not selected because
the viremia and antibody levels were not as high as those in the 8 x 10
pfu/ml group.
An IDsf) (infectious dose of 50 percent of the animals) was needed for
use as the center replicate in the physiological stress experiment.
was readily shown to be the 8 pfu/ml dilution of the working stock
(Table 4). Determination of antibody titers at this level is still in
progress.
This
175
-------
TABLE 4. PERCENT DEER MICE (P. MANICULATUS) WITH MEASURABLE VIREMIAS «1:10) AFTER SC AND IP
INOCULATION
WITH 0.4 ML OF
VEE VIRUS STRAIN TC-83 AND + OR -20
MG/KG MOUSE
CYCLOPHOSPHAMIDE (CP) IN THE FEED
Days post
inoculation
1
2
3
4
5
6
7
Total %
viremic
N
8 x 10
Control
100
100
100
100
0
0
0
100
6
7
+CP
§
§
§
§
§
§
§
§
§
8 x 10
Control
100
60
80
20
0*
. 0
0
100
5
4
+CP
80
80
90
50
40*
20*
20*
100
10
8 x 103
Control +CP
100 §
33 §
50 §
17* §
17* §
17* §
17* §
100 §
6 §
8 x 10
Control
90
70
80
90
10**
§
§
100
10
2
+CP
80
90
80
90
30*
0
0
90
10
80
Control
80
70
70
60
10**
§
§
80
10
+CP
50
70
90
60
304
20
0
100
10
8
Control
20
30
40
30
10
§
§
50
10
+CP
30
70
33t
44t
22t
10
0
90
10
* Died
T N = 9
i 1.1 *
Significant difference between +CP and Control percentage (Student's t-test P<0.05)
§ Not measured
-------
PATHOGEN-TOXICANT INTERACTION
A review article by Nathanson and Cole (1971) (7) discusses the
effects of immunosuppression caused by a single injection of cyclo-
phosphamide (CP) on acute virus infections in mice. In summary, viremia
titers, mortality rate, and survival time were all increased. Work done in
a collaborative project in Dr. R. Hinsdill's laboratory (8) has shown that
continuous feeding of lower levels of CP can also result in immuno-
suppression; specifically, a decrease in number of circulating lymphocytes
and antibody-producing cells in the spleen. Since arbovirus infections are
terminated by the production of neutralizing antibodies (9) it was
necessary to determine what effect this chronic low level of CP in the feed
would have on a sublethal infection of VEEV in the deer mice. An
appropriate adjustment in virus dose could then be made, if necessary,
prior to using these compounds as stressors in the physiological
experiment.
Forty mice were fed CP at 20 mg/kg mouse for 23 days. Groups of mice
were then injected SC and IP with 0.4 ml per mouse of VEEV at 8 x 10 ,
8 x 10 , 80, or 8 pfu/ml dilutions (10 mice per dilution). Ten control
mice were fed feed mixed with saline instead of CP and inoculated with the
same dilutions of virus plus one group inoculated with diluent only. The
40 experimental mice were fed CP-feed for the remainder of the study. Day
23 was chosen for inoculation as this is the time used for antigen
stimulation in the prior CP studies done in Dr. Hinsdill's laboratory.
Viremias and mortality rates of the mice fed CP were compared with those of
the mice in the experiment with VEEV alone, described above.
Table 2b shows the mortality pattern for mice injected with VEEV while
being fed CP. A comparison with Table 2a shows that feeding CP results in
higher mortality rates and increased survival time.
The time of onset of viremia was not affected by CP feeding. The
percentage of mice with viremia onset on Day 1 was not significantly
different (Student's t-test, PjOO.05) between mice fed CP (CP treated) and
mice not fed CP (controls) for any virus dilution (Table 4). However, the
duration of viremia was increased as a result of CP feeding. Viremias
lasted for a maximum of 4 days in the control mice whereas the CP treated
mice had viremias persisting for up to 5 days for all virus dilutions
except 80 pfu/ml, where the viremia persisted for a maximum of 6 days
(Table 4).
The geometric mean (xg) peak viremia titer for each virus dilution is
shown in Figure 2. It is evident that CP significantly (P<0.05) increased
the peak viremia titer, except for the 8 x 10 pfu/ml dilution of virus.
The reason for this discrepancy is unknown.
177
-------
120 •*
8 X 10T 8 X 102 80 8
Virus dilution (pfu/ml)
Figure 2. Mean (x ) peak viremia titers of deer mice (P_. maniculatus)
inoculated SC + IP with 0.4 ml of VEE virus strain TC-83 and
fed + or -20 mg/kg mouse cyclophosphamide (CP).
In summary, chronic low levels of CP in the food has an immuno-
suppressive effect on deer mice. This causes an acute, sublethal infection
of the arbovirus VEEV to attain higher viremia levels of longer duration
and results in increased mortality rates. The effects of CP on the
antibody response to an acute infection is still being evaluated.
178
-------
REFERENCES
1. Zarnke, R. L. 1978. Occurrence of selected microbial pathogens in
Alberta wild mammals. Ph.D. Thesis, University of Wisconsin-Madison.
72 pages.
2. Calisher, C. H., H. S. Lindsey, D. B. Ritter, and D. M. Sommerman.
1974. Northway virus: a new Bunyamwera group arbovirus from
Alaska. Can. J. Microbiol. 20:219-23.
3. McLean, D. M., P. N. Grass, B. C. Judd, K. J. Stolz, and K. K. Wong.
1978. Transmission of Northway and St. Louis encephalitis viruses by
Arctic mosquitoes. Arch. Virol. 57:31-5-22.
4. Bigler, W. J., A. K. Ventura, A. L. Lewis, F. M. Wellings, and N. H.
Ehrenkranz. 1974b. Venezuelan equine encephalomyelitis in Florida:
endemic virus circulation in native rodent populations of Everglades
hammocks. Am. J. Trop Med. Hyg. 23(3):513-521.
5. Bigler, W. J., A. L. Lewis, and F. M. Wellings. 1974a. Experimental
infection of the cotton mouse (Peromyscus gossypinus) with Venezuelan
equine encephalomyelitis virus. Am. J. Trop. Med. Hyg.
23(6):1185-1188.
6. Bowen, G. S. 1976. Experimental infection of North American mammals
with epidemic Venezuelan encephalitis virus. Am. J. Trop. Med. Hyg.
25(6):391-399.
7. Nathanson, N. and G. A. Cole. 1971. Immunosuppression: a means to
access the role of the immune response in acute virus infections.
Immunol. Soc. Sytnp. Fed. Proc. 39(6): 1822-1830.
8. Olson, J. and R. Hinsdill. 1981. Immunotoxicity in Peromyscus
maniculatus. Personal coramuncations.
9. Allison, A. C. 1972. Immune responses in persistent virus
infections. J. Clin. Path. 25, Suppl. (6):121-131.
179
-------
Comments on
Host-Pathogen-Tbxicant Interactions:
A Progress Report
Richard M. Kocan
University of Washington
At the time of the presentation, this project had only been underway
for nine months. Therefore, only the preliminary studies had been
completed and much of the proposed research had not yet been initiated.
Mouse Breeding Colony
A breeding colony of deer-mice (Peronyscus maniculatus) has been
established at the University of Wisconsin's semi-isolation facility in
Madison, Wisconsin. The colony currently consists of 100 breeding pairs
and the breeding success has been increased to 30% (up from 10%) since
the number of mice per cage was reduced from three to two. This success
also coincided with the beginning of spring. All of the animals have
been or will be checked for the presence of pathogenic organisms and
regularly monitored for signs of disease or parasites.
Host-Pathogen Interaction
Several strains of low virulence virus stock were tested in mice to
obtain a non-fatal/high infectivity test virus for future studies with
toxicant. The third strain tested (VEE vaccine) produced the desired
combination of effects and was tested further.
Pathogen-Toxicant Interaction
Tb determine the combined effects of virus and toxicant (cyclo-
phosphamide CP), two groups of mice were given inoculations of virus
ranging from 8 x 10 to 8 plague forming units per ml. One group was
fed a diet of 20 mg/Kg mouse for 23 days prior to virus inoculation.
The data collected after these treatments showed a definite supression
of an antibody response to virus in the CP treated group as well as a
higher virus titer. There was, however, some discrepancy in the mortality
figures which showed all 10 animals surviving a virus dose of 800 pfu
and CP, while animals receiving only 8 pfu's suffered a 30% mortality.
Otherwise, the virus and CP group had a higher overall mortality rate
than the virus only group.
180
-------
This project seems to be off to a good start and with time should
provide some very useful data concerning the combined effects of pathogens
and toxic chemicals.
181
-------
EFFECTS OF MILORGANITE ON TWO GRASSLAND ECOSYSTEMS
Gary W. Barrett, Department of Zoology and
Institute of Environmental Sciences
Miami University, Oxford, Ohio 45056
INTRODUCTION
Numerous studies have been concerned with the effects of sludge on young
(e.g., agricultural) or mature (e.g., forest) ecosystems (1, 2, 3). Few
studies have focused upon the intermediate (e.g., old-field) stages of eco-
system development (4) as a feasible and efficient community-type for the
disposal and recycling of sewage waste products. Such old-field communities
are highly productive (5, 6, 7) and are efficient at recycling nutrients (8).
We hypothesized in the present study that the intermediate stages of commun-
ity development (i.e., the old-field or grassland meadows) may well be the
best community-type for processing sewage sludge.
Previous sludge studies have been mainly concerned with the lower levels
of biological organization, i.e., with the cellular (9), the histological
(10), and organismic (11) levels of organization. Most of these studies fo-
cused upon individual plant or animal species, or their respective component
parts, without attempting to analyze the system as a whole (12). Odum (13)
has pointed out that problems often arise when stressors (e.g., sludge) are
tested at one level (e.g., individual) and then used without further study a±
another level (e.g., ecosystem). Also, by studying the ecosystem as a whole,
man should be able better to predict the long-term response of stress Or nu-
trient effects on natural ecosystems (14, 15, 16). Barrett et_ al^. (17) out-
lined specific guidelines for testing stressors, such as sludge, at the eco-
system level. Thus, another major objective of this investigation was to
identify indicator "white mice" indices (17) regarding the effects of sludge
on total ecosystems. Both structural (e.g., plant biomass and species diver-
sity) and functional (e.g., productivity and developmental) parameters are
being measured and evaluated within two different grassland (old-field) com-
munities in order to recommend properly such indices for future studies.
At present, no studies have been conducted which simultaneously attempt
(a) to measure, evaluate, and compare the effects of sludge applications on
contrasting community-types treated in an idential manner; (b) to evaluate
sludge effects on several different levels of biological organization; (c) to
isolate indicator parameters which may be indicative of total stress response;
182
-------
and (d) to identify the energy flow (fate) pathways of heavy metals (cadmium,
zinc, lead, and copper) as they move from the sludge through the producer
trophic level into the primary consumer (Microtus) small mammal populations
functioning within these old-field communities. We hypothesizes that wild
mammalian populations tend to depict better sludge effects (e.g., biological
magnification) than by extrapolating strictly from laboratory populations.
BACKGROUND INFORMATION
The present long-term study was initiated in 1978 at the Miami Univer-
sity Ecology Research Center located on the Bachelor Wildlife Reserve near
Oxford, Ohio. The initial phase (1978-1980) of the study was supported by a
subcontract of EPA No. R 805445-01 entitled "Evaluation of the Health Risks
Associated with the Treatment and Disposal of Municipal Wastewater and Sludge"
awarded to the Department of Environmental Health, University of Cincinnati,
Dr. C. Scott Clark, Principal Investigator. I served as Principal Investi-
gator of the subcontract to Miami University. The present EPA Grant (R807370
-01) became effective 13 May 1980. Thus, three (3) years of field data have
been collected; only one (1) year of data has been collected concerning the
present grant.
EXPERIMENTAL DESIGN
The study site consists of an 0.8-ha grassland (old-field) community
presently (1981) in the third year of secondary succession and an 0.8-ha (old-
field) community in the seventh year of succession. Each community-type was
divided into eight 0.1-ha plots by enclosing each plot with 0.2 x 2.0 m
sheets of galvanized steel. Enclosures of this type have been previously
described (14, 16, 18).
All eight plots in the third-year old-field were originally plowed and
disked and a commercial fertilizer (12-12-12) applied at the rate of 56.5 kg
ha"1 (300 Ibs/acre) on 24 October 1977. Winter wheat (Triticum aesveticum
Ranger variety) was sown at a rate of 22.6 kg ha"1 on 25 October. These
plots were then left fallow the following years, i.e., up to the present time.
In October 1974, the present seven-year old-fields were plowed, disked,
and all eight (8) plots sown with a mixture of grasses consisting of 6.8 kg
fescue (Festuca elatior), 11.3 kg Kentucky bluegrass (Poa pratensis), and 4.5
kg rye grass (Lolium perenne) (19). These plots have been left fallow since
1974.
In each community-type, three replicate plots were treated with dried
sludge (Milorganite; 6-2-0, N-P-K), three replicate plots were treated with
fertilizer (6-2-0, N-P-K), and two plots were left untreated to serve as con-
trols (Fig. 1). Milorganite is a heat-dried, anaerobically digested sludge
that has been commercially available since 1927 (20). Milorganite for each
of the past three years (1978-1980) has been applied at an enrichment rate
183
-------
EXPERIMENTAL DESIGN
SECOND-YEAR OLD-FIELD GRASSLAND ECOSYSTEMS
F
F
S
F
C
C
S
S
F
C
S
S
F
F
S
C
SEVENTH-YEAR OLD-FIELD
GRASSLAND
ECOSYSTEMS
Figure 1. Experimental design for the proposed investigation,
Treatments were selected at random within each com-
munity type. Each plot represents a quarter-acre
(0.1 ha) area. Sludge and fertilizer will each be
applied at monthly intervals from May through Sep-
tember in a 6-2-0 (N-P-K) ratio.
184
-------
equivalent to 1792 kg/ha (1600 Ibs/acre) monthly. Sludge is applied five
times (May-September) per year. An equivalent amount of fertilizer is applied
to each community-type the same time sludge is applied. The same design
(i.e., the same application rates) is also being used this summer (1981). It
is only from long-term studies of this nature that chronic effects of munici-
pal sludge may be evaluated in a quantitative manner.
RESULTS
PRODUCER TROPHIC LEVEL
In 1978, no significant differences due to treatment were found in spe-
cies richness in the annual (wheat) field or in the perennial (fourth-year
old-field) field. There was a significant increase in productivity in both
community-types in the fertilizer and sludge treatments; however, the in-
crease was larger and occurred earlier in the fertilizer treatments. Esti-
mates of net primary produtivity (NPP) in the agricultural field were 1690 g
m-2 yr-l ± 34 (SE) in the fertilizer treatment, 1240 g m~2 yr"1 ± 58 (SE) in
the sludge treatment, and 772 g m~2 yr"1 ± 3 (SE) in the control. Estimates
of NPP in the fourth year old-field were 1110 g tir2 yr-1 ± 39 (SE) in the
fertilizer treatment, 684 g m~2 yr"1 ± 69 (SE) in the.sludge treatment, and
424 g m~2 yr"1 ± 37 (SE) in the control. It appeared that the nutrient en-
richment effects on plant community structure and function were greater in
the agricultural field than in the old-field.
In 1979 and 1980, NPP in July was significantly higher (P <_ 0.05) in the
one-year old-field fertilized and sludge plots as compared to control plots
(Fig. 2). Interestingly, the two-year old-field control plots exhibited sig-
nificantly higher (P <_ 0.05) values than the sludge plots, but not signifi-
cantly higher than the fertilized plots. Only one significant difference in
NPP was observed in the five-year or six-year old-fields; nutrient-enriched
plots in the six-year old-field were significantly greater than control plots
in July (Fig. 2).
Control plots exhibited significantly greater (P £ 0.05) values for spe-
cies richness and apportionment than the nutrient-enriched plots in the one-
year and two-year old-fields. No differences in species richness or appor- ;
tionment were observed in the five-year or six-year old-fields (Fig. 3).
Competition for nutrients appeared to account for these differences. Differ-
ences in community function (e.g., NPP) were found to relate directly to
changes in community structure (e.g., species diversity). Changes in plant
species composition and NPP for 1978-1980 are summarized in Tables 1 and 2.
PRIMARY CONSUMER TROPHIC LEVEL
Meadow vole (Microtus pennsylvanicus) populations have been monitored
for three years by live-trapping 16 0.1-ha enclosures. Vole population den-
sities were affected by treatment only in the wheat plots in 1978 (Fig. 4).
Sludge-treated wheat plots had significantly higher population densities than
either fertilizer or control plots. However, survivorship, longevity, and
percentage of breeding adults were not affected by sludge treatment. Sex
185
-------
CO
«
«
E >
1-Vr. Old-Fltld
i ,.
4-JI l-ll «-M T-M t-ll (-11
S-Yr. Old-FUId
Ftrtlllitr
Sludgt
Control
2-Yr. Old-FUId
4-M t-M «-!• Mi *•!• •-««
/' ^ A
. / '-^ / \
N. / / V \ / *,
10-11
« i 1 i
4-n i-tt i-n »-n •-!• »-n
t-» 1-11 «-M T-M t-n no
Figure 2. Changes in NPP (g m~2 day'1) in 1979 and 1980 for both community types.
-------
00
Figur 3.
'
Changes in total plant species (number of species/m~2) in
1979 and 1980 for both community-types.
-------
TABLE 1
SUMMARY OF PLANT SPECIES COMPOSITION AND NPP CHANGES IN THE WHEAT,
1-YEAR, AND 2-YEAR OLD-FIELDS (1978-1980)
FERTILIZER
SPECIES
Aster pilosus
Erigeron annuus
•y
Daucus carota
__*
00
oo Trifollum pratense
Triticum aestivum
Ambrosia artemisiifolia
Cirsium arvense
Polygonum pennsylvanicum
Setaria faberii
Chenopodium album
Agr
*
A
*
14
536
877
48
14
*
56
1-yr
43
A
*
*
109
253
293
99
376
296
2-yr
*
*
*
A
*
270
350
349
309
108
Agr
*
*
*
38
326
456
52
10
A
5
SLUDGE
1-yr
*
*
A
A
A
365
164
222 •
275
A
2-yr
24
A
A
A
A
12
483
349
173
21
Agr
A
A
A
110
364
151
89
A
A
6
CONTROL
1-yr
199
82
176
112
71
A
179
A
29
A
2-yr
840
293
A
A
A
A
76
A
42
A
*Specles with NPP values less than 5 g-m 2-yr~1
All values represent annual net primary productivity (g-m~2-yr-1)
-------
10
TABLE 2
SUMMARY OF PLANT SPECIES COMPOSITION AND NPP CHANGES.IN THE
4-YEAR, 5-YEAR, AND 6-YEAR OLD-FIELDS (1978-1980)
FERTILIZER
SPECIES
Festuca elatior
Poa pratensis
Solidago canadensis
Ambrosia artemisiifolia
Barbarea vulgaris
Setaria faberii
Oenothera biennis
4-yr
350
120
120
*
A
11
5-yr
354
66
448
169
8
390
14
6-yr
301
18
270
397
161
415
76
4-yr
177
100
112
A
A
13
A
SLUDGE
5-yr
247
77
569
A
12
137
6-yr
135
85
865
227
234
262
62
4-yr
125
75
21
A
A
A
A
CONTROL
5-yr 6-yr
128 246
119 172
142 159
A A
A A
A A
A A
*Species with NPP values less than 5 g-m 2-yr 1
All values represent annual net primary productivity (g-m~2*yr~1)
-------
Wheat Field 0978)
ra
g
50
40
30
20
10
50
40
30
20
10
First-Year Old-Field (1979)
Fourth-Year Old-Field (1978)
sludge
fertilized
---- control
Fifth-Year Old Field (1979)
Jun Jul Aug Sept Oct Nov Jun Jul Aug Sept Oct Nov
Figure 4. Mean meadow vole population densities per 0.1 ha.
-------
ratios favored females in nutrient-enriched plots and males in control plots
in 1978 and 1979; adult sex ratios did not differ significantly among treat-
ments in either community-type in 1980. Recruitment rates .(number of recruits/
lactating female), however, were higher in control plots than in sludge or
fertilizer enclosures during September and October in 1980. Also, fertilizer-
treated plots had significantly lower population growth rates than sludge or"
control plots within the six-year old-fields.
METAL CONCENTRATIONS IN MEADOW VOLE TISSUES
Food-chain uptake of cadmium (Cd), lead (Pb), zinc (Zn), and copper (cu)
by meadow voles was assessed in both community-types from 1978-1980; these
metals will also be assessed in 1981. Thus far, only the 1978-1979 data have
been completely analyzed.
Table 3 provides a summary of metal concentrations in Milorganite and
commercial urea-phosphate fertilizer for 1978-1979. The 1980 concentrations,
presently being analyzed, are comparable. There were no significant differ-
ences (P > 0.05) between years for cadmium, lead, or copper concentrations in
either the fertilizer or the sludge. Zinc was significantly lower in the
second year sludge samples compared to the first year. Cadmium was approxi-
mately 30 times higher in the sludge than in the fertilizer; lead 45 times
higher; zinc 55 times higher; and copper 200 times higher.
Mean Cd concentrations in livers and kidneys of all sludge-treated
groups in the second year of application (1979) were significantly higher (P
<_ 0.05) than controls; in the first year (1978), only females from sludge-
treated wheat plots had higher levels than controls. Tables of cadmium con-
centrations in vole tissue for both community-types are summarized in Tables
4 and 5. Remaining heavy metal data present in vole tissues for copper, lead,
and zinc have been summarized by Anderson and Barrett (21). Ranges of mean
Cd concentrations in sludge-treated voles were 0.1-1.1 yg/g wet weight in
liver samples and 0.4-6.6 yg/g in kidneys; ranges in controls were 0.01-0.4
yg/g in liver and 0.01-1.6 yg/g in kidneys. Cd concentrations in fertilizer-
treated voles did not differ from controls. These data have been submitted
for publication (21).
Liver and kidney Cu concentrations were elevated in some sludge and
fertilizer-treated groups. There were no significant differences in lead or
zinc concentrations, nor in organ or whole-body weights.
Despite accumulation of metals in vital organs, short-term effects on
small mammals due to sludge application appear to be benign, i.e., detected
amounts are less than those volumes reported for laboratory feeding studies
(22, 23).
METAL CONCENTRATIONS IN SOIL AND EARTHWORMS
Earthworms are a good indicator species for investigating heavy metal
accumulation (24). They also play an important role in maintaining aeration,
water permeability, and mineral turnover of the soil. Approximately 90% of
the worms collected in this study were Lumbricus rubellus. Earthworms can
191
-------
TABLE 3
METAL CONCENTRATIONS IN MILORGANITE AND COMMERCIAL UREA-PHOSPHATE FERTILIZER.
VALUES REPRESENT 95% CONFIDENCE INTERVALS BASED ON THREE SAMPLES PER
APPLICATION DATE (i.e., n = 15 FOR EACH INTERVAL)
to
Milorganite
1978
1979
Fertilizer
1978
1979
Cd
58.4-59.6
56.1-59.9
1.8-2.2
1.2-2.0
Pb
463.6-482.4
443.1-474.9
1.2-9.6
0-25.0
Wg/g
Zn
1043.9-1118.1
922.3-1013.7
19.3-24.3
12.6-20.2
Cu
312.8-327.
315.7-342.
1.6-2.2
1.2-2.0
2
3
-------
TABLE 4
CADMIUM CONCENTRATIONS1 IN VOLES FROM THE WHEAT FIELD (1978) AND FIRST-YEAR OLD-FIELD (1979)
10
CO
Control
1978 1979
Fertilized
1978 1979
Sludge
1978 1979
Liver
adult males
adult females
sub-adult males
sub-adult females
Kidneys
adult males
adult females
sub-adult males
sub-adult females
Lungs
adult males
adult females
sub-adult males
sub-adult females
Gonads
adult males
0.01
0.04
0.01
0.02
0.01
0.20
0.01
0.05
0.04
0.02
0.02
0.04
0.01
(3) e
(4) e
(4) e
(3) e
(1) f
(2) f
(2) f
(2) f
(2) ab
(2) b
(2) b
(2) ab
(2) a
0.03
0.06
0.02
0.04
0.16
0.59
0.12
0.14
0.01
0.01
0.01
0.01
0.01
(3) e
(2) de
(3) e
(2) e
(2) f
(2) ef
(2) f
(1) f
(2) b
(1) b
(1) b
(1) b
(3) a
0.03
0.09
0.03
0.04
0.09
0.31
0.06
0.04
0.02
0.01
0.01
0.01
0.01
(5) e
(5) de
(6) e
(4) e
(3) f
(3) f
(3) f
(2) f
(3) b
(3) b
(3) b
(2) b
(4) a
0.04
0.09
0.05
0.04
0.31
0.88
0.10
0.17
0.01
0.01
0.01
0.01
0.01
(3)
(4)
(5)
(3)
(2)
(3)
(3)
(3)
(2)
(2)
(2)
(2)
(3)
e
de
e
e
f
def
f
f
b
b
b
b
a
0.14
0.56
0.20
0.29
0.90
2.16
0.63
1.37
0.01
0.03
0.02
0.02
0.01
(4) cde
(5) b
(5) cde
(6) c
(3) def
(3) be
(2) ef
(3) cde
(3) b
(3) b
(2) b
(3) b
(4) a
0.28
0.96
0.47
0.83
1.72
3.59
1.71
2.34
0.02
0.02
0.01
0.07
0.01
(4) cd
(4) a
(4) b
(4) a
(3) bed
(3) a
(3) bed
(3) b
(2) b
(2) b
(2) b
(!) a
(4) a
of treatment replicates (yg/g wet weight); the number of samples is in parentheses; means within
each organ followed by the same letter are not significantly different (p > 0.05) (Duncan's multiple
range test) •
-------
TABLE 5
:1
CADMIUM CONCENTRATIONS1 IN VOLES FROM THE FOURTH-YEAR (1978) AND FIFTH-YEAR (1979) OLD-FIELDS
Control
1978 1979
Liver
adult males
adult females
sub-adult males
sub-adult females
Kidneys
adult males
adult females
sub-adult males
sub-adult females
Lungs
adult males
adult females
sub-adult males
sub-adult females
Gonads
adult males
0.02
0.26
0.11
0.14
0.15
-
0.15
0.70
0.01
-
0.01
0.01
0.01
(4) d
(1) bed
(3) cd
(4) cd
(2) e
(2) e
(2) cde
(2) b
(1) b
(2) b
(3) a
0.04
0.44
0.02
0.04
0.40
1.61
0.01
0.18
0.01
0.02
0.01
0.01
0.02
(2)
(3)
(3)
(3)
(2)
(2)
(2)
(2)
(2)
(2)
(1)
(1)
(2)
cd
b
d
cd
de
bcde
e
e
b
b
b
b
a
Fertilized
1978 1979
0.03
0.11
0.01
0.05
0.14
0.59
0.34
0.12
0.01
0.01
0.01
0.01
0.01
(7) d
(6) cd
(4) d
(4) cd
(3) e
(3) cde
(2) de
(2) e
(2) b
(3) b
(2) b
(2) b
(4) a
0.04
0.12
0.06
0.07
0.44
1.30
0.33
0.55
0.01
0.01
0.01
0.01
0.01
(4) cd
(4) cd
(5) cd
(4) cd
(3) de
(3) bcde
(3) e
(2) cde
(3) b
(3) b
(2) b
(2) b
(2) a
Sludge
1978 1979
0.17
0.45
0.07
0.23
0.81
2.27
0.39
1.10
0.03
0.02
0.01
0.14
0.01
(5) cd
(5) b
(4) cd
(4) bed
(3) cde
(3) b
(2) de
(2) bcde
(3) b
(3) b
(2) b
(2) a
(4) a
0.27
1.06
0.44
0.40
1.80
6.57
1.93
2.24
0.01
0.02
0.02
0.02
0.01
(4) be
(4) a
(4) b
(4) b
(3) bed
(3) a
(3) be
(3) b
(2) b
(3) b
(2) b
(2) b
(2) a
of treatment: replicates (pg/g wet weight); the number of samples is in parentheses; means within
each organ followed by the same letter are not significantly different (p > 0.05) (Duncan's multiple
range test) .
-------
considerably influence the trace element content of a soil by redistribution
and bioaccumulation. Bioaccumulation of heavy metals by earthworms has been
found to occur naturally in unpolluted.soils and the accumulation is relative
to the concentration found in that soil (24). Cadmium, nickel and zinc have
been found to be accumulated in concentrations greater than those present in
the soil. This accumulation is significant when there are large amounts of ~
metal being introduced into the soil. A study was begun in May, 1981, to in-
vestigate the earthworm population and its accumulation of heavy metals in
sludge- and fertilizer-treated plots.
Two control, two fertilized and two sludge-treated plots were randomly
selected from the three-year field. Five 10 cm3 samples were taken from each
plot for each sample date. Samples were hand sorted and the worms placed in
petri dishes on moist filter paper for four days to void their guts of soil.
Soil samples were oven-dried at 80° C for 24 hours and screened through a 1
mm sieve. Soil and worms were digested with nitric acid and aqua regia to
remove organic material. Atomic absorption spectrophotometry was used to
quantify concentrations of metals.
A total of four sample dates will be treated in this manner. Early re-
sults indicate a significant difference in biomass between the sludge-treated
and the control or fertilized plots. Biomass collected in the sludge plots
was almost twice that found in the fertilized or control plots (Table 6).
Their high density may be an important factor in the conditioning of the
sludge-treated soil.
Data for concentrations of cadmium, lead, and copper are presented in
Table 7 (data for zinc are not available at this time). Copper and lead show
very little accumulation in earthworms collected from the sludge-treated
plots. However, cadmium is concentrated by a factor greater than 10. The
possibility of biomagnification through the food chain to birds or small mam-
mal consumers awaits further investigation.
BEHAVIORAL ANALYSIS
Urinary marking patterns were used to establish a dominance index for
Microtus pennsylvanicus. Scrotal males, weighing 25 g or more, were placed ,
in 46 x 46 cm boxes with a wire mesh bottom. Filter paper was used to obtain
the marking patterns. Animals were left in the marking boxes for a 12-hour
period without food or water. The filter papers were analyzed with an ultra-
violet light and an acetate overlay consisting of 36 squares.
A dominance index was calculated from the equation proposed by Noyes et
al. (25), where the index value is equal to the sum of the relative density,
relative area, and relative frequency of urine marks. The index values ranged
from 0.74 to 40.09. Plotting cumulative percent vs the dominance index values,
a breakpoint in the curve of 16 was determined. Animals with an index value
greater than 16 were considered to be dominant. Those animals that had an
index value less than 16 were said to be subordinate. These social rank pre-
dictions were supported by random pairing tests in which predicted dominants
won 15 or 16 matches. The establishment of a dominance index allows for a
195
-------
TABLE 6
BIOMASS (GRAM WET WEIGHT) OF EARTHWORMS COLLECTED FROM FERTILIZED,
SLUDGE AND CONTROL PLOTS
SAMPLE I
26 May 1981
SAMPLE II
16 JUNE 1981
SAMPLE III
25 JUNE 1981
SLUDGE
22.25
20.42
10.75
FERTILIZER
10.31
9.26
6.97
CONTROL
11.28
10.59
4.10
TABLE 7
SOIL AND EARTHWORM CONCENTRATIONS (PPM; uG/G) OF CADMIUM, COPPER
AND LEAD AND CONCENTRATION FACTORS; ESTIMATED ACCORACY ±0.1
SOIL
EARTHWORMS
Cu
Sludge
Fertilizer
Control
15,
9,
8.9
2.9
1.4
0.1
Pb
CF*
Cd
Sludge
Fertilizer
Control
1.8
1.0
0.4
24.6
3.7
1.3
13.67
3.70
3.25
0.18
0.15
0.12
Sludge
Fertilizer
Control
37.0 .
25.4
23.4
2.9
2.0
1.6
0.08
0.08
0.07
*CF = Concentration Factor; amount present in earthworm divided by amount
present in soil.
196
-------
22
20
18
16
o
Q
12
10
8
s
1 6
Figure 5«
t
CONTROL
FERTILIZER SLUDGE
Comparison of the mean dominance index
values and standard errors for Microtus
males in each field treatment.
197
-------
relatively quick method of constructing a social hierarchy in these microtine
populations.
Figure 5 illustrates the mean dominance index values for Microtus popu-
lations in each of the experimental fields. The mean value in the control
field is 9.80. The fertilized field shows a mean of 9.64 while the sludge-
treated fields show a mean of 12.74. This difference between sludge and con-
trol plots is not significant but does suggest a possible trend.
198
-------
REFERENCES
1. Hinesly, T. D., and B. Sosewitz. 1969. Digested sludge disposal on
cropland. . J. Water Poll. Contr. Fed. 41:822-830.
2. Sopper, W. E., and L. T. Kardos. 1973. Recycling treated municipal
wastewater and sludge through forest and cropland. Penn. St. Univ.
Press, University Park, Pa. 492 pp.
3. Sabey, B. R., N. N. Agbim, and C. D. Markstrom. 1977. Land application
of sewage sludge; IV. Wheat growth, N content, N fertilizer value, and
N use efficiency as influenced by sewage sludge and wood waste mixtures.
J. Environ. Qual. 6:52-58.
4. Odum, E.. P. 1969. The strategy of ecosystem development. Science 164:
262-270.
5. Odum, E. P. 1960. Organic production and turnover in old-field succes-
sion. Ecology 41:34-49.
6. Kurd, L. E., M. V. Mellinger, L. L. Wolf, and S. J. McNaughton. 1971.
Stability and diversity at three trophic levels in terrestrial succes-
sional ecosystems. Science 173:1134-1136.
7. Mellinger, M. V., and S. J. McNaughton. 1975. Structure and function
of successional vascular plant communities in central New York. Ecol.
Monog. 45:161-182.
8. Golley, F., and J. Gentry. 1965. A comparison of variety and standing
crop of vegetation on a one-year and twelve-year old-field. Okios. 15:
185-199.
9. Rose, F. I., and J. C. Harshbarger. 1977. Neoplastic and possibly re-
lated skin lesions in neotenic tiger salamanders from a sewage lagoon.
Science 196:315-317.
10. Tipton, I. H., and M. J. Cook. 1963. Trace elements in human tissue:
Part II. Adult subjects from the United States. Health Physics 9:103.
11. Kinzell, J. H., R. P. Cheeke, and R. W. Chen. 1977. Tissue heavy metal
accumulation, pentobarbitol sleeping times and multigeneration reproduc-
tive performance of rats fed activated sewage sludge. Nutr. Rep. Int.
15:645-650.
12. Odum, E. P. 1977. The emergence of ecology as a new integrative disci-
pline. Science 195:1289-1293.
13. Odum, E. P. 1968. Testing pesticides at the ecosystem level. ASB Bull.
15:48.
14. Barrett, G. W. 1968. The effects of an acute insecticide stress on a
semi-enclosed grassland ecosystem. Ecology 49:1019-1035.
199
-------
15. Wakefield, N. G., and G. W. Barrett. 1979. Effects of positive and
negative nitrogen perturbations on an old-field ecosystem. Am. Midi.
Nat. 101:159-169.
16. Suttman, C. E., and G. W. Barrett. 1979. Effects of Sevin on arthro-~
pods in an agricultural and an old-field plant community. Ecology 60
(3):628-641.
17. Barrett, G. W., G. M. Van Dyne, and E. P. Odum. 1976. Stress ecology.
BioScience 26(3):192-194.
18. Bulan, C. A., and G. W. Barrett. 1971. The effects of two acute stres-
ses on the arthropod component of an experimental grassland ecosystem.
Ecology 52:597-605.
19. Stueck, K. L., and G. W. Barrett. 1978. Effects of resource partition-
ing on the population dynamics and energy utilization strategies of
feral house mice (Mus musculus) populations under experimental field
conditions. Ecology 59:539-551.
20. Anderson, M. S. 1959. Fertilizer characteristics of sewage sludge.
Sew. Ind. Waste. 31:678-682.
21. Anderson, T. J., and G. W. Barrett. 1981. Effects of dried sewage
sludge on meadow vole (Microtus pennsyIvanicus) populations in two
grassland communities. J. Appl. Ecol. (submitted for publication).
22. Schroeder, H. A., J. J. Balassa, and W. H. Vinton, Jr. 1964. Chromium,
lead, cadmium, nickel and titanium in mice: Effects on mortality, tu-
mors and tissue levels. J. Nutr. 83:239-250.
23. Lener, J., and B. Bibr. 1970. Cadmium content in some foodstuffs in
respect of its biological effects. Int. J. Vitalstoffe-Zivilisations-
krankheiten 15:139.
24. Van Hook, R. I. 1974. Cd, Pb, and Zn distributions between earthworms.
and soils: Potentials for biological accumulation. Bull. Environ. Con-
tamination and Toxicology 12:509-511.
25. Noyes, R. F., G. W. Barrett, and D. H. Taylor. 1981. Social structure
of feral house mouse (Mus musculus L.) populations: Effects .of resource
partitioning. Behav. Ecol. and Sociobiology (in press).
200
-------
Comments on
Effects of Milorganite
on TWo Grassland Systems
Bernard P. Sagik
Drexel Lfriiversity
In an unusual experiment, the Environmental Protection Agency has put
its omniscient reviewers to the test: they have been asked to examine
the preliminary results of an on-going research project which they
themselves found worthy of support one or two years ago. What Barrett
proposed at that time was:
a. to measure and evaluate the effects of chronic sludge
application on several different levels of biological
organization;
b. to compare two different grassland communities (types)
treated in an identical manner;
c. to analyze different structural (e.g., biomass and
species diversity) and functional (e.g., energy flow
and community resilience) ecosystem parameters -
measurements which may be indicative of total
systems repsonse;
d. to evaluate sludge effects (via ecosystem food chains)
on a natural small population (Microtus pennsylvanicus)
functioning within these community types (we feel that
wild mammal populations illustrate better the potential
sludge effects on man than by extrapolating strictly
from laboratory populations); and
e. to develop a feasible ecosystem approach, including
the identification of "white mice" field parameters,
for testing stressors (e.g., sludge) on intact
ecosystems.
Individual aspects of these objectives had been examined earlier but no
studies which attempted to do all this simultaneously had been undertaken.
In effect, we were seeing a holistic approach and, therefore, a most
ambitious one. As Barrett noted, Odum has observed that problems are
created when stressors are tested at the individual level and then used
without further study at the ecosystem level.
201
-------
In addition to the prior stated goals anent total systems response
and effects of sludge application on several different levels of biological
organization, the PI hoped that the natural animal population, specifically
Microtus pennsylvanicus, would serve better to illustrate the potential
sludge effects on man than merely extrapolating from laboratory feeding
experiments. In my own notes at the panel review I find that I raised
the question as to whether Milorganite was equal to all sludges; implicit
in this question was the idea that somehow Milorganite was better,
cleaner (the psychological effect on me of commerical packaging, no
doubt). A second question was whether the voles would breed at a
differential rate sufficient to detect treatment effects on reproduction
and behavior, as well.
The test plots were a 0.8 ha old-field (perennial) grassland and a
0.8 ha wheat field in its first year of old-field succession. Comparisons
were made among 0.1 ha subunits enclosed with sheet metal and treated
with Milorganite or commercial fertilizers or left as untreated controls
and seeded with winter wheat or a mixture of grasses. No differences
were observed in the first year between annual or perennial fields in
species richness. Both commercial fertilizer and Milorganite increase
productivity, the former earlier and to a greater extent. In the next
two years, net primary productivity (NPP) was greater in one-year old-
field fertilized and sludge plots as compared with control plots. This
advantage was lost in two-year old-field studies; and in five-year and
six-year old-fields only nutrient-enriched plots were greater in NPP
than control plots. In contrast, control plots showed significantly
greater species richness than nutrient-enriched plots in one-year and
two-year old-fields. The differences disappeared in five-year or
six-year old-fields.
At the primary consumer trophic level, Barrett found vole populations
higher only in sludge-treated wheat plots. Although some significant
differences in sex ratios were occasionally seen and recruitment rates
appeared higher in control than treated plots. These data appear too
preliminary to be taken seriously yet, as do lessened growth rates in
fertilizer-treated plots in six-year old-fields. What is striking at
this early stage of the study are the metal concentrations found in
meadow vole tissues. It should be noted that Cadmium (Cd) was 30 times
higher in Milorganite than in fertilizer, lead 45 times higher, zinc 55
times higher, copper 200 times higher. Analyses show Cd ranges in
Milorganite-exposed voles to be 0.1 - l.ljug/g wet wt in liver,
0.4 - 1.6jug/g in kidneys. Controls, by contrast, had 0.01 - 0.4;ug/g
in liver and 0.01 - 1.6;ug/g in kidneys. In a preliminary experiment,
Barrett compared Lumbricus rubellus harvested from control with those
fertilized or sludge-treated plots and found, not unexpectedly, a mean
concentration of Cd 10 times greater in the Milorganite-treated soils,
the implications for biomagnification through birds, and small animal
predation needs further investigation.
202
-------
I have hesitated, being normally of a pacific nature, to comment on
the effects of sludge on dominance patterns. Although Barrett's data
hint at increased dominance as shown by urinary marking patterns, I
frankly don't know how to interpret the complex interactions of Milorganite
spreading, elevated cadmium in livers and kidneys and urinary frequency, _
volume, and pattern. In the Philadelphia area, suburbanites in whose
parks we attempt to spread composted sludge seem to react in similar
fashion even before we dump the stuff.
I have done Dr. Barret somewhat of a disservice by summarizing his
paper in unsophisticated fashion. What I wanted to note, however, is
that there appear to be certain agricultural benefits to the use of some
sludges. Concomitantly, we must recognize that not all sludges are
created equal. Earthworms and vole kidneys make it apparent that heavy
metal contributions from roads and parking lots as well as industrial
processes are a significant problem. Nightsoils in Canton, while rich
in pathogenic bacteria and parasites, don't have that particular problem.
There are fields in the Braunschweig district in Germany which use
liquid sludges in batch lots only after testing for the presence of
excessive levels of heavy metals. Point sources can be controlled;
street and road pollution cannot be in our older cities, in particular
where common storm and sanitary sewer systems create a near unsuperable
problem. For these cities, a return to selective ocean dumping, composting
and subsequent application to mine spoils and parkways may yet be a more
desirable choice.
I'd like to close with my favorite quotation from William Lowrance's
book Of Acceptable Risk, as in a sense, I have highlighted some of
Barrett's findings which have negative connotations. Lowrance wrote
"In 1900 the principal insecticide, sprayed on everything
from apples and grapes to strawberries and potatoes, was
"Paris green" - lead arsenate. The first canned foods were
being preserved with sulfites, boric acid, and formaldehyde
at rather high concentrations. The leading antiseptic for home
and hospital use was corrosive carbolic acid. The red robustness
of teas, candies, and other commercial food products was imparted
by lead chromate, which today's biochemist would prefer not
even to handle, much less feed to anybody. Medicinals were
unreliable: "Adulteration threatened the healing art. Even
more alarming was the spread of proprietary or patent medicines.
Some had genuine merit, but most were nostrums, often
containing dangerous, habit-forming ingredients. Such
fraudulent remedies were far from new, but in the latter
part of the 19th century they increased in number and
resort to consumption cures, lost-manhood tonics, soothing
syrups, liver pills, and other so-called remedies became
a craze." Foods were often adulterated; the laws
controlling food composition were weak. Tinctures of opium
such as laudanum and paregoric were readily available in
203
-------
corner drugstores and were unchecked for indiscriminate
use. Children frequently became habituated.
As to the general environment, some rivers were so filthy
with raw sewage and industrial waste that, as the saying
went, "bait died on the hook." Industrial towns were black
with coal soot, as were people's, lungs. Workers labored at
their own peril."
"Since the taking of both personal and societal risks
is inherent in human activity, there can be no hope of
reducing all risks to zero. Rather, as when steering
any course, we must continually adjust our heading so
as to enjoy the greatest benefit at the lowest risk and
cost."
204
-------
THE ASSESSMENT OF TOXIC EFFECTS IN
EXPERIMENTAL STREAM ECOSYSTEMS
Robert H. Boling, Jr., William E. Cooper, R. Jean Stout
Department of Zoology
Michigan State University
East Lansing, Michigan 48824
ABSTRACT
Fate and effects of toxic stress on large scale outdoor
stream channels were studied from 1979 to 1981 at the Monticello
Ecological Research Station, Monticello, Minnesota. P-cresol, an
energy-related phenolic compound, was introduced at 8 ppm into
closed-cycle stream ecosystems from 13 hours in 1979, for 24 and
then 48 hours in 1980, and for 96 hours in 1981. This report
presents only results from work in 1980. In 1980, microbiology
and microcosm laboratories were added to the on-site facility to
relate results from flask, microcosm, and pseudo-natural stream
ecosystems to one another. The fate of p-cresol in microcosm
tanks tracked that of the outdoor streams, but results from flask
experiments did not track the fate of p-cresol in microcosm or
stream ecosystems. The effects of p-cresol on the biological
community during the 48-hour dose differed in each tier: flask,
microcosm and outdoor stream channels. During the second night
of the 48-hour dose in the stream channel, the concentration of
dissolved oxygen sank to less than 1 ppm. By daybreak, several
fish species were severely stressed as were amphipods, isopods
and some species of mayflies. Neither severe oxygen depletion
nor biotic stress occurred in microcosm or flask experiments.
Changes in microcosm design were therefore made for the 1981
field season. Chemical and engineering advances were made in the
on-line rapid detection of phenols. Sampling and behavioral
methods were developed for the testing of effects of toxic
compounds on natural biological systems.
INTRODUCTION
In the fall of 1978, Michigan State University was awarded a
contract by the U.S. Environmental Protection Agency to design
and implement a modification involving five of the eight 500
meter artificial stream channels at the Monticello Ecological
Research Station, Monticello, MN., to study the fate and effects
205
-------
of toxic stress on large scale experimental stream ecosystems. The
fundamental objective of the project has been, from the outset, to
characterize the functional relationship between laboratory studies of
toxicity and pseudo-natural ecosystem responses; i.e., to better
define the rules of transferability and predictive robustness of
flask, aquaria, and microcosm studies.
The results of that interdisciplinary design project are two
independent, isolated, recycling stream ecosystems that physically,
chemically, and biologically are typical of warm water, grassland,
third-to-fifth order streams. Containing a water volume of 1.5 to 1.8
million liters each, the 1000 x 3 meters channels are driven by
recirculation pumps that propel the stream at up to 11000 1/m
resulting in mean velocities up to 12.2 cm/sec (riffle velocities
range up to 25 cm/sec). The earthen channels have alternating riffles
and pools of 30m lengths. The control and monitoring of the channel
flows and water chemistry (pH, DO, temperature), local meteorological
conditions, and toxicant injection are executed by a dedicated on-site
computer system.
An organic chemistry laboratory was established on-site geared to
our prototype toxicant, p-cresol. Thus, we have established, at the
Monticello site, a set of large experimental stream ecosystems, and
facilities for analyzing toxicant concentrations in all ecosystem
components and microbial activities.
The work supported by the EPA Office of Research and Development
involved the basic systems refinement, operation, and analysis for a
24-hour and a 48-hour ecosystem exposure to the toxicant., p-cresol,
during the 1980 experimental season.
EXPERIMENTAL OPERATIONS
24-HOUR EXPOSURE
P-cresol was added at 8 ppm from 1330 hours on July 19th to 1300
hours on July 20th (Figure 1). The retention time for the channels
was approximately 6 hours at a flow rate of 4000 liters per minute.
By 2200 hours an equilibrium profile was established with a slight
oscillation of concentration rippling through the system.
Figure 2 presents the ambient concentrations at the tail end of
the system. Both the flow-through phenol analyzer and the gas
chromatographic analyses produced similar estimates. The estimated
half-life was approximately 64 hours for the first 24 hours during the
dosing. After dosing terminated, accidental events and a misjudgement
of the process of p-cresol degradation prevented us from monitoring
until the following morning. At 1100 hours we picked up the analysis
at 2 ppm ambient levels and measured half-lives less than one hour.
206
-------
10
8
6-1
24
a;
2
i7* •*'
£*• i
24 HOUR DOSE, 8 PPM
Head End
-••.- ,~. . ••
... — ••- .
1200 2400 1200
19 JULY 20 JULY 1980
TIME. HOURS
2400
21 JULY
Figure 1. Ambient Concentrations of p-Cresol at Head End of Channel B
During and After Dose of p-Cresol at 8 ppm. July 19-21, 1980.
24 DOSE, 8 PPM
{tail end)
?•••••- -—-do ••••••p«r iod •
1200 hour! 2400
19 July
1200
20 July 1980
2400
1200
2400
21 July
Figure 2. Chemical Analyses of a 24-hour Dose of p-Cresol at 8 ppm in
Channel B at the Tail End of the Channel. July 19, 20, 1980.
207
-------
Dosing stopped at 1330 hours. (Recall that there is a 6 hour time lag
between the injection point and the tail end of the channel). The first
water mass which represents the termination of p-cresol injection would
reach the tail end at 1930 hours. The time lag between the end of dosing
and the onset of rapid degradation in the channel was computed to be
roughly 3.5 hours, using the 48-hour dose as a basis for this estimate.
There was no serious reduction in dissolved oxygen and no indication
of mortality in invertebrates and fish. Data from gut analyses of the four
species of resident fish indicated an inhibition of the feeding process.
The bio-accumulations of p-cresol in the four species of fish are given in
Figure 3.
24
18
12
a:
a:
P-CRESOL IN FISH
24 hr. dose, 8ppm
• Smallmouth Bass
* Largemoulh Bass
o Fat headed Minnow
a Walleyed Pike
'DQsing.'rZ....'..1IP'<:reso1 > • i pptn in water
2400 hours
20 JULY
2400
2400
2400
21 JULY
22 JULY
23 JULY 1980
Figure 3. Bioaccumulation of p-Cresol in Four Species of Fish After Being
Dosed at 8ppm for 24 Hours in Channel B. July 20-23, 1980.
The average total body load was 18 ppm, which represents a bioaccumulation
factor of 2.3. The average water ambient levels varied from 7.5 to 8.0 ppm
until approximately 2300 hours on July 20. All four species showed the
same accumulation and release pattern. By 1100 hours on July 21, the
ambient concentrations were reduced to 1 ppm and the fish contained between
2-5 ppm; 24 hours later only trace amounts were measurable in the fish;
208
-------
they had a very rapid release of p-cresol. Bacterial samples demonstrated a
change in tnicrobial flora associated with the p-cresol application. The
number, size, and color of bacterial plaques shifted dramatically during
the 24 hour period of injection. Initial attempts to demonstrate
significant increases in p-cresol degradation by these new bacteria were
not definitive. Figure 4 presents the bacterial density estimates. The
four sampling stations in Channel B were distributed uniformly along the
stream. The dosing period was from 1330 hours, July 19, to 1330 hours,
July 20. Bacterial counts increased consistently in Channel B but no
increase was observed in the control stream. The increased density was
maintained for at least five days after dosing ceased, and concentration of
p-cresol was 0.5 ppm by July 21 (Figure 2).
BACTERIAL COUNTS, eto/ml x 10 4
24 hour dose, 8ppm
40
32
"o
X24
at
S3 16
J
3
z
9
Channel B, dosed
—..Channel A, control
o-
1? 20 21 22 23
DATES- JtHY. 1980
24
25 26 27 28
29
30
Figure 4. Bacterial Density Counts (CFU/ml x 104) in Channel A (Control)
and Channel B (Experimental) During and After a 24-Hour Dose of
p-Cresol at 8 ppm in Channel B.
48-HOUR EXPOSURE
The 48 hour dose occurred,between 1000 hours, August 9, to 1000
hours, August 11. Ambient concentrations were maintained around 8 ppm
(Figure 5). Again, the p-cresol appeared to be very stable during the 6
hour retention time in the recyling mode.
Figure 6 presents the ambient concentration of the tail end of the
stream just before the water entered the sump to be recycled. The 6.25
hour retention time is illustrated by the fact that dosing started at 1000
hours and the first wave of p-cresol reached the last sampling station at
approximately 1530 hours. The oscillations were produced by periodic
problems with the dosing apparatus. By comparing Figures 5 and 6, an
estimated average of 1.7 ppm phenol was lost during the 6.25 hour retention
time. This would produce an estimated 14.7 hour half life, assuming an
exponential drop in concentration over the stream.
209
-------
10'
8.
6
* 4
48 HOUR DOSE, 8 B P.M.
Haad End
•f *
.•"ST0 . «*...• *..•*••*»* J
• I •• e .«i^F •• «
: v •
• •"
1200 2400 1200
9 AUG. 10 AUGUST 1930
TIME. HOURS
2400
1200
11 AUG.
Figure 5.
Ambient Concentration of p-Cresol at the Head End of Channel B
During a 48-Hour Dose of p-Cresol at 8 ppm. August 9-12. 1980.
-18 HOUR DOSE, 8 PPM
(loil.nd)
- *
•
1200 2400
9 Aug.
1200
10 Aoquil 1980
2400
1200
It Avguit
2400 1200
12 Aug.
Figure 6. Chemical Analyses of a 48-Hour Dose of p-Cresol at 8
ppm in Channel B. (Tail End of Channel). August 9 - 11, 1980.
The inflection point should not have occurred before 1615 on August 11
since the dosing stopped at 1000 and the retention time was 6.25 hours.
There appeared to be no time lag in degradation response, as the ambient
concentrations declined immediately. The compared half-life during the
decrease from 6 ppm to 2 ppm was approximately 5.3 hours. Obviously, the
degradation rate increased significantly after the dosing terminated.
210
-------
Examination of the differences in total phenol measurements between _
tail end of the stream and head end of the stream indicate that between
hour 14 and hour 30 of the 48 hour dosing, the differences steadily
increased, from an interpolated half-life of 69 hours to a half life of 23
hours. Also at hour 31 a decrease in concentration at the tail end of the
stream (when the head end remained relatively constant) implies 2 further
decrease in half life from 23 hours to 12 hours. This is much closer to
the half life immediately after dosing (5.3 hours). Also, immediately
before dosing ended, a difference between tail end and head end of 2.5 ppm
was apparent, which would give a half life of 11.1 hours. Thus, adaptation
by microbial populations to faster biodegradation rates was probably
occurring in the stream. During the dosing.experiments, the temperature
showed normal diurnal variation from 22 to 27°C, and the pH varied from
7.0 to 8.5. However, dissolved oxygen showed a marked change during and
after dosing which was obviously related to the effect of p-cresol as shown
in Figure 7. The 0.0. changes triggered a number of secondary biological
effects
DISSOLVED OXYGEN, Channel A control
Wink'ler -, i o j j
,oJ Channel B dosed
/"^% Channel B p-cresol
/
•• *».
101
end of dose i
Hours: 12CO 2400 1200
11 August 1980 12 August
Figure 7. Dissolved Oxygen Profiles in Channel A (control) and
Channel B (experimental) with p-Cresol Levels (48 hour dose).
211
-------
The dissolved oxygen profile of the control stream demonstrated the
normal diurnal cyclicity between 4 and 12 ppm.
It is hypothesized that the mechanism of effect is probably the increased
respiration of organisms (plants, animals and microbes) and not the
inhibition of photosynthesis. The dissolved oxygen curves for the control
and dosed streams are approximately parallel, with the major difference
being the minimal levels reached at 0500 to 0600 hours; i.e., the rate of
photosynthesis was similar and the pattern of the diurnal cycle was not
disturbed. Chemical oxidation of p-cresol is possible, yet the dissolved
oxygen increased during daylight at the same rate as the control stream and
the ambient levels of phenols were 7 to 8 ppm. The dissolved oxygen was
reduced to low levels of 1 to 1.5 ppm during the night even though the
phenol levels dropped to less than 2 ppm, suggesting that nighttime D.O.
sags can be either intensified at low p-cresol concentrations, or the
higher concentration of p-cresol the previous day had a lag effect, which
was expressable at night.
During the periods of low dissolved oxygen on both August 11 and 12,
certain fish and invertebrates experienced massive mortality. Walleye pike
died in appreciable numbers. Fish removed before death and placed in clean
water recovered. Smailmouth bass were gasping air at the surface, but none
died. Largemouth bass were not visibly stressed, but their feeding was
inhibited. Fathead minnows were not visibly stressed and, in fact,
deposited eggs on the surface of fish traps.
Large numbers of Stenonema mayflies, amphipods and some isopods were
observed dead on the surface of the algal mats. Field observations
indicated a severe impact on the macroinvertebrates.
Another estimate of nonlethal effects was obtained from the
predator-prey experiments with largemouth bass. The results are given in
Table I. Fish were tested before exposure, after 36 hours exposure to
ambient water from Channel B, (the dosed channel) and 24 hours after being
placed in clean water from Channel A. Control fish were handled in a
similar fashion, but placed in cages in Channel A for the 36 hours of
exposure.
Handling times and percent successful capture were the two parameters
that indicated an inhibitory effect of p-cresol. The fish recovered within
24 hours after being placed in control water. The results of the bacterial
sampling are presented in Figure 8. The bacterial densities showed a sharp
increase in the experimental channel B; but there was no similar increase
in the control Channel A. The timing of the increase in relation to the
dosing interval was similar to that observed during the 24-hour application
(Figure 4). The maximum densities obtained are not very different, but the
duration of the bacterial pulse was very different. The bacterial
densities declined very rapidly after 48 hours of exposure. The p-cresol
concentration declined much more slowly than the bacterial densities.
212
-------
Table I. Predation Parameters for Largemouth Bass Exposed to
p-Cresol
Ir.councer
Races (per sec.)
Mean Search
Tiaes ( sec . )
Handling
Times tsec) .
" Capture
P-cresol Exposed Fish Concrol Fish
Before During Recovery Before During
-.037
31.15^32.6
3 . 94_+ 3 . 54
1002
i
.0152 | .034
I
i
40.4_*69.5 i 33.52-35.0
1
1
13. 25-11. i i 9.5-5.0
i
I
!.-•> I 0 /. »
-» / .•• ! 7** /•
!
.0^5
23.2^23.4
.028
39.33-ii.:
1
3.32-3.971 7.00^2.5
97^
07"
40
30
CO
20
BACTERIAL COUNTS, CFU/ml x 104 (48 hour dose, 8ppm)
T. dosing period A
.Channel B. riffles (dosed)
.Channel A. riffles (control)
789
D AT E: August/ 1980
10
11
12
Figure 8.
Bacterial Counts (CFU/ml xlO^) in Channel A
(control) and Channel B (Experimental) Before, During
and After 48-Hour Dose of p-Cresol in Channel B.
August 7-12, 1980.
213
-------
Total organic carbon (TOG) analyses are presented in Figure 9.
The baseline levels were between 3..5 and 8.0 ppm in both
channels. The stations sampled were distributed uniformly along
8
6
- 4
TOTAL ORGANIC CARBON- 48hour dose at 8ppm
a:
a:
Channel A (control)
14
12
10
8
a_ 4
a:
Channel B (dosed)
T dosing--- period A
8 9 10
DAT E: August 1980
11
Figure 9. Total Organic Carbon Analysis for Channel A (control)
and Channel B (experimental) Before and During Dosing
of Channel B with p-Cresol at 8 pp for 48 hours.
August 8-11, 1980.
214
-------
the streams. The sharp increase in TOG observed in the dosed channel is
probably due to two sources. P-cresol is 77% carbon by weight. P-cresol
was added to maintain 8 ppm which produces 6 ppm additional organic carbon.
The average carbon for Channel B before dosing was about 7 ppm. The
additional carbon associated with p-cresol would raise the ambient TOC
level to 13 ppm. The observed levels peaked at 16 ppm on the night of the
initiation of dosing. P-cresol also will increase the respiration rates of
some organisms due to stress. Additional evidence of the potential impact
of increased respiration is seen in Figure 7. The reduction in dissolved
oxygen observed at 0600 hours on August 11 must have occurred during the
previous two nocturnal periods. The differences in dissolved oxygen levels
is about 4 ppm on August 11.
Another source of TOC would be the population increase of bacteria
(Figure 8); however, the timing of the TOC and bacterial population curves
are not synchronized. The TOC indicated peak values during the first 24
hours of dosing; whereas, the bacterial-peak occurred during the second 24
hours. Furthermore, the dissolved oxygen data do not indicate that the
stress factor, i.e., increased respiration in response to p-cresol, had
diminished by August 11.
' ANALYSES
BIOACCUMULATION
Table II gives the results of the bioaccumulation of p-cresol by five
species of fish that were sampled during and after the 48-hour dose.
Twenty-four hours into the 48-hour dose, walleye pike and largemouth bass
showed similar levels of p-cresol to those found near the end of the
24-hour dose. P-cresol concentrations for those species and for bluegills
and fatheads remained high throughout the dosing. Twenty-six hours after
dosing stopped, p-cresol levels in the fish dropped. The fish accumulate
the toxicant rapidly, but they also release rapidly when ambient levels o'f
p-cresol are low.
Bluegills collected 20 hours after cessation of the 48-hour dose were
dissected for organ analyses. The eye (without lenses) had the lowest
concentration (2.8 ppm) and muscle tissue the next lowest (4.7 ppm). Gill
tissue contained 7.3 ppm. The highest bioaccumulation values were in the
livers (76.1 ppm) and in the intestines (96.7 ppm). High accumulations in
the liver suggests sequestering of p-cresol from blood tissue. The very
high values in the intestines could possibly occur if the fish began eating
soon after dosing ceased. The results from fish gut analyses show that
bluegills stopped feeding during dosing but began feeding within hours
after dosing stopped. Even so, the high accumulation in the intestines was
unexpected. Skin was not analyzed but it is expected that the accumulation
probably was low, given the fact that whole bluegills taken four hours
later only contained 12 ppm p-cresol.
Various biotic components of the channel were analyzed for p-cresol
cocentration. Table III shows that only one component of those tested,
namely a crayfish, showed significant amounts of p-cresol. The crayfish,
Orconectes virilis, bioaccumulated nearly three times more p-cresol than
background concentrations in the channel. This crustacean is the largest
215
-------
TABLE II
BIOACCUMULATION* OF P-CRESOL IN FISH DURING AND AFTER 48
HOURS OF DOSING CHANNEL B AT * PPM P-CRESOL
DATE
8/10
8/10
8/11
8/11
8/11
8/11
8/12
8/12
8/15
TIME
1000
1300
0530
0745
1000
2330
1000
1200
2000
P. P.M.
Walleye
Pike
17.5
18.9
14.4
15.4
14.8
OF P-CRESOL
Largemouth
Bass
18.7
12.0
18.6
16.8
16.7
13.4
15.9
20.3
4,8
IN FIVE SPECIES OF
Bluegill Fathead
Sunf ish Minnow
5.9
13.7
12.4
17.8
14,9
24.5
29.0
14.0
12.0
0.9
FISH
Smallmouth
Bass
0.5
Gas Chromatograph Data
TABLE III
CONCENTRATION OF P-CRESOL IN SEDIMENTS, PLANTS AND INVERTEBRATES
AFTER 48 HOURS OF DOSING CHANNEL B AT * PPM P'-CRESOL
DATE TIME
P,P,M. OF P-CRESOL IN COMPONENTS
Sediment
Algae
Spirbgyra
Dragonfly
Anax
Crayfish
Orconectes
8/10. 100.0
8/10 1100
8/11 0945
8/11 1000
8/12 1000.
8/14 10.00
0.9
3,8
1.8
0.7
0.0
0.8
0,0
1,1
4.0
0,0
0.8
23.9
216
-------
and is one of the most common invertebrates in the channel. It is a
detritus feeder but also can capture other invertebrates and slow-moving "
fish; they possibly could have consumed stressed fish during dosing, which
could account for the high concentrations of p-cresol in the crayfish.
TOXICITY
Selected fish and invertebrates from the sites were tested at the
E.P.A. Laboratory at Duluth for laboratory response to p-cresol. Those
data appear in Table IV. The fish species show a distinct "all-or-none*
pattern: fatheads, for the most part, died after 48 hours in 20 ppm
p-cresol, but none died (even after 96 hours) with concentrations of 10
ppm. Largemouth bass all died after 24 hours at 32 ppm, and none, died at
16 ppm; smallmouth bass died after 24 hours at concentrations of 20 ppm,
but all survived at 10 ppm. The damselfly, Ischnura verticalis showed no
response, even at concentrations of 40 ppm. The amphipod, Hyalella azteca
showed a more gradual dose-response pattern. For short periods of time (24
hours) most of the animals survived, even at concentrations of 80 ppm, but
only 50% survived after 72 hours at concentrations of 10 ppm. Death also
occurred at concentrations of 5 ppm after 48 hours. The waterflea, Daphnia
magna, was very sensitive to the toxicant at 2.5 ppm, even after 24 hours.
During our 48-hour dose, we noted dead amphipods caught on algae.
BEHAVIORAL RESPONSES
Preliminary data show that there was a volumetric increase in
invertebrates in drift nets during dosing of p-cresol. We also observed
qualitative increases in numbers of animals during dosing. The preliminary
data bear this out. At the end of the summer, we varied velocity of
Channel B (500, 1000 and 2000 gal/min) and took drift samples during each
velocity change. There was more than a two-fold increase in drift between
1000 and 2000 gal/min, but there was little difference between 500 and 1000
gal/min.
217
-------
TABLE IV
DOSE-RESPONSE DATA ON P-CRESOL WITH SELECTED FISH
AND INVERTEBRATES FROM MONTICELLO, MINN.
Mount and Norberg
E.P.A. Lab. at Duluth
Initial
Organism Concentration
(number) of P-Cresol
Fathead 40 mg/1
minnow 20 mg/1
(10/run) 10 mg/1
5 mg/1
2 replicates 2.4 mg/1
Control
Largemouth 32 mg/1
bass 16 mg/1
(5/run) 8 mg/1
4 mg/1
1 replicate 2 mg/1
Control
Small mouth 40 mg/1
bass 20 mg/1
(5/run) 10 mg/1
5 mg/1
1 replicate 2.5 mg/1
Control
Largemouth 20 mg/1
bass 10 mg/1
Test #2 5 mg/1
(5/run) 2.5 mg/1
1 replicate 1.25 mg/1
Control
Damselfly 40 mg/1
(10/run) 20 mg/1
10 mg/1
1 replicate 5 mg/1
2.5 mg/1
Control
TIME
24-hour %
Survival
60, 30
60, 90
100,000
100,000
100,000
100,000
0
100
100
100
100
100
0
0
100
100
100
100
0
100
100
100
100
100
-
-
-
-
-
-
48-hour %
Survival
20, 0
30, 40
100,000
100,000
100,000
100,000
-
100
100
100
100
100
-
-
100
100
100
—
—
100
100
100
100
100
100
100
100
100
80
70
7 2 -hour %
Survival
0, 0
0, 10
100,000
100,000
100,000
100,000
-
100
100
100
100
100
-
-
80
100
100
—
—
100
100
100
100
100
96-hour %
Survival
0, 0
0, 0
100,000
100,000
100,000
100,000
—
100
100
100
100
100
—
-
80
100
100
—
_
100
100
100
100
100
218
-------
TABLE IV, continued
Organism
(number)
Initial
Concentration
of P-Cresol
TIME
24-hour %
Survival
48-hour %
Survival
72-hour %
Survival
96-hour %
Survival
Amphipod
Hyalella
azteca
(8/run)
2 replicates
80 mg/1
40 mg/1
20 mg/1
10 mg/1
5 mg/1
Control
Waterflea
Daphnia
magna
(5/run)
2 replicates
40 mg/1
20 mg/1
10 mg/1
5 mg/1
2.5 mg/1
Control
50,37.5
87.5,50
87.5,100
100,000
100,000
100,000
0, 0
0, 0
0, 0
60, 80
80, 80
100,000
0, 0
25, 25
37.5,50
75, 87.5
87.5,87.5
100,000
0, 0
0, 0
0, 0
60, 40
60, 20
100,000
0, 0
12.5,12.5
37.5,37.5
50, 50
87.5, 50
100,000
0, 0
12.5,12.5
12.5, 25
27.5, 50
75, 37.5
100,000
1. Flow per minute in chambers ranged from 21 to 60 ml/min.
2. Initial D.O. values ranged from 5.5 to 7.9 ppm
3. Final D.O. values ranged from 5.5 to 8.4 ppm
4. pH ranged from 7.2 to 7.5
5. Alkalinity = 40; hardness = 46
6. Temperature was 24 to 26^C.
7. Measured concentration of p-cresol was done for each run. Those data
are not presented here, but are part of the Mount and Morberg report -
to us and are available on request.
219
-------
Comments on
The Assessment of Toxic Effects
in Experimental Stream Ecosystems
Robert J. Beyers
University of South Alabama
In order to gain a complete understanding of the effects of any
toxicant on an ecological system, one must undertake studies on many
different organisms at various trophic levels within that system.
Removing organisms from their natural environment and challenging them
with pollutants always carries the possibility of missing synergistic
effects which may take place within the system. However, it is frequently
undesirable from a practical or esthetic point of view to choose a
natural system and deliberately dose it with a toxic substance. In
order to obtain better control and to protect the environment from harm
caused by the experiment, ecologists have frequently relied on some form
of microcosm experiment when engaged in this type of investigation.
This is precisely the situation in the Boling-Cooper work. In this
case they have chosen to use a large artificial stream facility located
at the Monticello Field Station. These streams are large enough to be
called "mesocosms" by the investigators. The advantage of large size is
that reasonable amounts of biota can be removed during sampling without
destroying the structure of the systems. Also larger organisms can
complete their life cycles in a normal manner. A further advantage of
this.kind of experiment is that the inputs can be carefully controlled,
in this case by a computer.
Finally, an artificial stream is basically a closed system without ;
immigration or emigration from the outside. This is effective in simplifying
the mathematical models which best describe the workings of a biological
community. It may be assumed that deviation from the model of an unstressed
community will be an indication of the reaction of an ecosystem undergoing
toxicological stress.
For these reasons, the Environmental Protection Agency Environmental
Review Panel felt that this work would be a prototype for the kinds of
experiments which must be widely done in the future to determine the
effects of various pollutants on total ecosystems, as opposed to studies
on individual "indicator" organisms. In this light the proposal was
considered highly significant, and funding was recommended.
220
-------
The preliminary results of the present work bear out these contentions.
The sheer size of the system permitted observations on the behavioral
aspects of selected organisms and the analysis of bioaccumulation in
certain members of the system.
It is to be hoped that in the future, additional work and analysis
of these and new data will bear out the expectations of the Environmental
Protection Agency. Some biomagnification was observed (algae vs. crayfish),
but sane surprises were also noted (low concentration of toxicant in
predators such as small mouth bass). v
221
-------
POLYPHOSPHATE IN NATURAL PHYTOPLANKTON ASSEMBLAGES AND
ITS FORMATION AND DEGRADATION IN CULTURED ALGAL CELLS
Linda Sicko-Goad and Diane Lazinsky
Great Lakes Research Division
The University of Michigan
Ann Arbor, Michigan 48109
INTRODUCTION
A survey of stations in Saginaw Bay of Lake Huron during the period
1974-1976 demonstrated that: 1.) Certain phytoplankton populations were ex-
iting the bay and surviving transport to southern Lake Huron (27), 2.) Many
of these populations in the bay were found to have polyphosphate accumula-
tions in them (28), and 3.) Several of the populations of blue-greens and
diatoms in the bay had lead sequestered in the polyphosphate bodies (22, 21).
Moreover, it has been demonstrated with both cultured blue-greens (3, 1) and
diatoms (22) that heavy metals may be incorporated into polyphosphate bodies
during phosphate uptake.
Polyphosphate is believed to function primarily as a phosphorus reserve
(19). It can be formed under several distinct nutritional conditions: 1.)
Restoration of phosphate following a phosphorus deficiency (19, 14), 2.) Nu-
trient imbalance other than phosphorus (16, 29, 23, 24), and 3.) Disturbance
of nucleic acid metabolism (11).
The eventual fate and ecological significance of polyphosphate is not
well known. If polyphosphate serves as a phosphorus storage form that can be
degraded under conditions of limiting phosphorus, then phosphorus limitation
should make polyphosphate (and heavy metals if they are so bound) available
to intracellular sites.
In view of the fact that Saginaw Bay is the only major basin in the
Great Lakes that has experienced a substantial reduction in phosphorus load-
ing during the last 5 years, samples were collected from Saginaw Bay to de-
termine what effect, if any, this phosphorus loading reduction had on the
distribution of polyphosphate in natural phytoplankton assemblages. In addi-
tion, laboratory experiments were conducted to determine both the time se-
quence and consequences of polyphosphate degradation in cultured algae as
well as the effects of heavy metal exposure at the time of luxury phosphorus
uptake. Preliminary results are presented in this paper.
222
-------
MATERIALS AND METHODS
Field samples were collected in conjunction with the 1980 Saginaw Bay
and southern Lake Huron Special Field Survey conducted under the auspices of
the Great Lakes National Program Office and the Grosse lie Large Lakes Re- -
search Laboratory of the U.S. E.P.A. Whole water samples were collected from
Saginaw Bay approximately every two weeks from the period April 14, 1980
through November 14, 1980 and from southern Lake Huron during five cruises
beginning in May 1980 and ending October 1980. Selected stations in the seg-
ment scheme of Saginaw Bay as well as a series of stations along the western
shore of southern Lake Huron were sampled as weather permitted.
Water samples were fixed on board ship with a mixture of paraformalde-
hyde-glutaraldehyde (17) to give final concentration of 1% of each aldehyde.
For samples to be embedded for transmission electron microscopy, cacodylate
buffer at a final concentration of 0.05M (pH 7.2) was also added.
Samples examined for polyphosphates were stained by the method of Ebel
et al. (7). Initially, samples were examined in wet mounts. However, due to
paucity of certain species and vibration, samples were filtered onto 0.8 um
Millipore filters and the filters were subsequently cleared with glutaralde-
hyde for actual counts (6). Summaries of phytoplankton counts for the
Saginaw Bay cruises were obtained from computer files and phytoplankton
species were selected for polyphosphate enumeration. The three criteria for
selection of species within a cruise were: 1.) Species present on 50% or
more slides, 2.) Species always present at > 1% of assemblage numbers, and
3.) Maximum occurrence > 5%.
LABORATORY EXPERIMENTS
Three algae were selected for study: Cyalotella aff. meneghiniana
(Bacillariophyceae), Saenedesmus quadrieauda (Chlorophyceae), and Pleetonema
boryanim (Cyanophyceae). The algae were grown to logarithmic phase at 20°C
on a 12/12 light-dark cycle in either WC (10) or modified Fitzgerald's medium
(8). Both Soenedesmus and Cyolotella were initially grown separately in WC
medium in Corning tissue culture flasks. However, for the experiment, begin-
ning with transfer to POit-free medium before luxury uptake, the two were com-
bined and treated as one culture for ease of handling. Pleotonema was main-
tained separately throughout.
The algal cells were washed three times in phosphate-free medium prior
to transfer to the P-free medium for starvation. After three days in P-free
medium, phosphate was added to a final concentration of approximately 8 mg
POit/liter (twice the normal level in the growth medium) and lead was added
with thorough mixing in the Pb treatment flasks to a final concentration of
20 ppb. The cells were incubated under these conditions in a growth chamber
for three hours, then washed in phosphate-free medium three times prior to
transfer to phosphate-free medium. The cells were then sampled according to
the schedule in Figure 1.
223
-------
EXPERIMENTAL DESIGN*
r\>
r-o
P-
a P-
I
CD
POLYPHOSPHATE
DEGRADATION
OSAUPIED
CD
1
1
1
1
1
1
t
1
1
CD d
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
D CO Ce
i
i
I
i
i
i
CD CO CD Co) CD CO
Figure 1. Experimental design of polyphosphate-lead uptake and degradation studies. This
design was used for two separate cultures—a mixed culture of Seenedesmus and
Cyclotella, and a unialgal culture of Pleotonema. *Sampling was later extended
to 3 weeks.
-------
For every sample withdrawn as indicated, cells were treated in the fol-
lowing manner:
1. Unfixed cells were washed 3 times in sterile distilled water. Drops
of cell suspension were then dried on formvar coated nylon grids for x-ray
analysis .
2 . Portions of cells fixed in paraf ormaldehyde-glutaraldehyde were
withdrawn, stored in fixative, and subsequently stained for polyphosphates .
3. The remainder of the fixed sample was then post-fixed in 1%
dehydrated in a graded ethanol-propylene oxide series and embedded in Epon
for transmission electron microscopy.
For quantitative electron microscopy, random sections were obtained
through embedded pellets of the algae. Thirty micrographs at 75,000 X were
examined for each treatment by superimposing a transparent 1.0 cm square
sampling lattice over the picture. Estimates of volume density were obtained
using the grid point -counting technique (2, 9).
RESULTS
MORPHOLOGICAL CHANGES (PLECTONEMA BORIANUM)
Detailed ultrastructural examination of controls, phosphate starved, and
phosphate uptake cells of P. boryamm have been published elsewhere (13).
The present experiment differs from that report in regard to the length of
phosphate starvation. Quantitative results of the experimental treatments
are presented in Table 1. In general, phosphate starvation results in a de-
crease in the surface area to volume ratio (Sv) of thylakoids. This low
ratio is also observed when either P + Pb uptake cells or logarithmic phase
cells are again placed in phosphate-free medium. Polyhedral body relative
volume (Vv) appears to increase slightly during P starvation. Polyphosphate
body (Vv) increases during phosphate uptake and subsequently decreases during
P starvation. The number per volume (Nv) increases substantially during up-
take. During the uptake process,-with or without lead, polyphosphate bodies "
are relatively small and scattered throughout the cell (Figures 2-3). How-
ever, fewer numbers are present during subsequent P deprivation and they are
both larger and located in areas of medium electron density (Figures 4-5).
Images of what we interpret as polyphosphate degradation are present as early
as 2 days in the P deprivation medium. Polyphosphate bodies appear to be
less electron dense; they are more prone to volatilization in the electron
beam and often only a small dense portion is left at the periphery of the
body.
Qualitative observations were also made of cells incubated for 4 and 7
days in P deprivation medium after phosphate uptake. Virtually all cells ex-
posed to lead during phosphate uptake were dead at four days. The sample al-
so contained cellular debris, indicating lysis. The P uptake cells as well
225
-------
TABLE 1. MOKPHOMETRIC RESULTS OF PHOSPHATE AND LEAD TREATMENTS. SURFACE AREA/VOLUME
RELATIVE VOLUME (Vy) : NUMBER/VOLUME (Ny) . RESULTS ARE THE MEAN + 1 S.E.
ro
IM
Treatments*
Thylakoid (Sv)
Poly P (Nv)
Poly P (Vv)
Polyhedral (Nv)
Polyhedral (Vv)
Intrathylakoidal
Space (Vv)
Other (Vy)
INOC
20.24
(1.14)
12.48
(5.78)
0.42
(0 . 10)
2.48
(0.53)
2.00
(0.38)
12.67
(1.26)
84.90
(1.26)
P STRV
14.57
(0.93).
66.38
(17.82)
0.48
(0.15)
4.37
(0.91)
3.17
(0.45)
9.08
(1.24)
87.26
(1.34)
P SUFF
17.79
(1.35)
156.81
(33.63)
0.45
(0.11)
2.83
(0.49)
2.30
(0.44)
4.91
(0.70)
92.34
(0.73)
P-UP
18.34
(1.29)
113.91
(12.66)
1.77
(0.19)
2.71
(0.44)
2.21
(0.42)
4.06
(0.74)
91.95
(0.81)
PB-UP
15.30
(1.20)
396.45
(53.87)
2.06
(0.33)
2.26
(0.49)
1.53
(0.30)
10.62
(1.29)
85.79
(1.36)
C-UP
17.19
(1.15)
201.88
(38.36)
1.46
(0.30)
2.01
(0.45)
1.50
(0.32)
13.14
(1.01)
83.90
(1.11)
P DEG2
19.64
(1.45)
79.75
(15.40)
0.51
(0.11)
2.18
(0.39)
2.44
(0.45)
9.37
(1.07)
87.68
(1.13)
PB DEG2
14.39
(1.34)
62.59
(12.90)
0.37
(.09)
3.81
(0.87)
1.91
(0.38)
7.02
(0.84)
90.70
(0.88)
C DEG2
14.32
(0.93)
197.27
(30.64)
1.25
(0.24)
4.17
(0.70)
2.44
(0.35)
10.06
(1.13)
86.25
(1.33)
*Treatment Code:
INOC = Inoculum
P STRV = P starved 3 days
P SUFF = P sufficient 3 days
P-UP = P uptake 3 hours
PB-UP = P + PB uptake 3 hours
C-UP = New transfer P sufficient cells 3 hours
P DEG2 = P uptake—2 days in deprivation medium
PB DEG2 = P + PB uptake—2 days in deprivation medium
C DEG2 = P sufficient cells—2 days in deprivation medium
-------
Figures 2-5. Plectonema boryanim. Magnification bars a 0.2 urn. Intrathyla-;
koidal space (IT), Thylakoids (T), PolyP (P) , Polyhedral bodies
(PH) .
Figure 2. Phosphate uptake cell (3 hours). Note light areas of intrathyla-
koidal space. Numerous polyphosphate bodies are scattered through-
out the cell. Thylakoids are also evident.
Figure 3. Phosphate + lead uptake (3 hours). Greater numbers of polyphosphate
bodies are present with this treatment.
Figure 4. P uptake cells incubated 2 days in P deprivation medium. Note fewer
numbers of polyP and altered sturcture (arrows).
Figure 5. Day 2 deprivation medium of cells not previously incubated in star-
vation medium. PolyP bodies are larger and more numerous than in
other P deprivation treatments.
-------
as those maintained in medium containing phosphorus began to decline after
four days in P-free medium. However, live cells were still observed after 7
days. When cells were examined by x-ray energy dispersive analysis, no de-
tectable quantity of lead was observed in cells subjected to lead exposure
during P uptake.
POLYPHOSPHATE DISTRIBUTION IN SAGINAW BAY CRUISES 1 AND 2
Species summaries were generated for the Saginaw Bay cruises where the
phytoplankton counts were complete. For cruise 1 (April 1980), and cruise 2
(May 1980), 15 and 17 species respectively met the criteria for inclusion as
key species. There were 13 species common to both cruises (Table 2). Since
many of the species present had numerous small polyphosphate bodies that were
difficult to quantitate, the cells were examined and rated according to the
following scheme: 0 = no polyP, +1 = less than 5 bodies visible in the light
microscope, and +2 = numerous (greater than 5). These numbers relate only to
light microscopic examination. Each station for the 2 cruises was examined
for the presence of polyphosphate in the key species. Polyphosphate (on a
relative scale from 0 to 2) was plotted on maps of Saginaw Bay to compare
trends in abundance and distribution with station. Several of these maps are
shown in Figures 6-11.
TABLE 2. SPECIES COMMON TO
BOTH CRUISES 1 AND 2 OF
SAGINAW BAY POLYPHOSPHATE SUR-
VEY. CRUISE 1. APRIL 14-18,
1980; CRUISE 2. MAY 5-8, 1980
Cyolotella aomensis
Chroomonas spp.
Diatoma tenue var. elongation
Dinobryan spp.
Flagellate spp.
Fragilaria aapuoina
Fragilaria cvotonensis
Qohromonas spp.
Rhodomonas minuta
Saenedesmus spp.
Stephanodisous subtilis
Stephanodiscus tenuis
Synedra filiformis
228
-------
ro
ro
RRGILRRIfl CROTONENSIS
Figures 6-11.
Figure 6.
Polyphoaphate distribution in Saginaw Bay by species and abundance. 0 to 2.5 is a rel-
ative scale of abundance with 0 representing none and 2 representing numerous bodies
per cell. Abundance and scale lines are in a north-south direction.
Poly P abundance for unidentified flagellates category, cruise 2.
Figure 7. Poly P abundance for F. arotonensis, cruise 2.
phate bodies than flagellates.
Diatoms, in general, had more polyphos-
-------
OJ
o
Figures 8-9.
Polyphosphate distribution and abundance for two flagellates, Cryptomonas and Rhodomanas.
Most species examined from cruise 1 had relatively equal distribution patterns throughout
the bay.
-------
ro
co
STEPHflNODISCUS TENUIS
Figures 10-11.
Polyphosphate distribution and abundance for two diatom species from cruise 2. In the
second cruise, relatively fewer bodies were counted in cells along the northern shore.
This is the area of entry of open Lake Huron water.
-------
Most flagellates with the exception of Khodomonas minuta possessed far
fewer polyphosphate bodies than either diatoms or greens (Figures 6-7).
Early May samples had more polyphosphate across all species than did the
April samples. For example, in cruise 1 (April) only 4 of 8 diatom species
examined had a maximum abundance for all stations of 2, whereas in cruise 2,
9 of 10 diatoms had a maximum abundance of 2 at one or more stations.
There appears to be, at present, no consistent distribution pattern or
diminution of polyphosphate with station location. For cruise 1, polyphos-
phate body abundance is random with apparently equal distribution in all seg-
ments of the bay (Figures 8-9). For cruise 2, several species seem to have
fewer polyphosphates in segment 4 (Figures 10-11).
DISCUSSION
Cells of Plectonema boryanwn exposed to 20 ppb lead for 3 hours during
phosphate uptake and subsequently placed in P-free medium are not able to
survive longer than 2 days after this short term, low dose metal exposure.
Under similar culture conditions, but with no lead treatment, cells can sur-
vive a week or longer. Although Pb was not accumulated to a concentration
detectable by x-ray energy dispersive analysis, it appears that this metal
was capable of exerting toxic effects. Whether or not the metal is associ-
ated with polyphosphate is, at this point, uncertain. Previous studies in
our lab (21) demonstrate that nutrient sufficient cultures of Pleotonema ex-
posed for seven days to Pb at a concentration of 0.1 yg-at/1 resulted in lead
accumulation in the cells but not associated with polyphosphate. This seven-
day exposure also did not result in the mass mortality of cells experienced
in the present experiment, leading us to believe that phosphate nutrient sta-
tus may play an important role in the manifestation of toxicity symptoms.
There still remains a possibility that the metal is associated with polyphos-
phate and that it is mobilized during polyphosphate degradation. Shuman et
al. (18) found that the minimal detectable mass of K in typical biological
thin sections was 10 mmol/kg. Since 1 kg of a logarithmic phase culture con-
tains roughly 5.5 x 1012 cells, it is possible that [1] we are at the limits
of detection at 20 ppb Pb assuming complete uptake and/or "trapping" and [2]
Pb diffuses across the cell wall and plasma membrane more slowly in
Pleatonema than in Diatoma (22).
Although previous reports (13, 15) have dealt with the ultrastructural
aspects of phosphate starvation and uptake in Plectonema boryanion, no quanti-
tative cytological data have been reported. The morphometric data presented
in this paper generally are in agreement with our previous studies. However,
several cellular compartments deserve further discussion. Jensen and Sicko
(13) noted increases in intrathylakoidal space during phosphate starvation
and uptake. The quantitative data show that this space actually occupies a
larger volume in cells used as inoculum. However, the vacuolization appears
to be coalesced into a larger area in starved cells, allowing us to earlier
interpret this change as an increase. Polyhedral bodies (carboxysomes) were
also observed by Jensen and Sicko (13) to elongate during starvation and up-
take. No elongated polyhedral bodies were observed in this study. The in-
232
-------
crease in relative volume (Vy) appears to be a direct result of an increase
in number per volume.
Polyphosphate relative volume for inoculum, starved, and uptake cells
correlates well with previous results (20) which demonstrated that logarith-
mic phase cells have low reserves of both acid soluble and acid insoluble
polyphosphates. No changes in cyanophycin granules were observed in this
study during phosphate starvation. Jensen and Sicko (13) reported an in-
crease in these inclusions after 5 days starvation in P. boryanum and similar
results have also been reported for Agmeneltum (25).
The cells held in medium containing phosphorus also experienced a large
increase in polyphosphate Nv with no concommitant increase in relative vol-
ume. Doonan (5) demonstrated that 14-day logarithmic phase cells produce
alkaline phosphatase, which is not constitutive in P. boryanwn nor is it re-
leased to the medium. Rapidly growing cells take up phosphate when trans-
ferred to new medium (19, 5). In the present experiment, this increase was
solely an increase in the numbers of very small (slightly larger than ribo-
somal size) polyphosphate bodies.
Cells transferred to medium devoid of phosphorus after uptake responded
differently, depending on the treatment history of the cells. Those cells
which had been subjected to a period of phosphorus starvation experienced
larger decreases in polyphosphate relative volume after 2 days in the depri-
vation medium.
The polyphosphate bodies formed during uptake change cytologically dur-
ing phosphate deprivation. In addition to there being fewer bodies present,
there appears to be a degradation sequence which resembles a reverse of the
formation sequence described by Jensen (12) and Stewart and Alexander (26).
We observed increased numbers of "porous structures" thought to be the site
of polyphosphate deposition as well as images of polyphosphate bodies where
there was only a remnant of electron dense polyphosphate at the periphery.
As the length of time in deprivation medium increases, the frequency of ob-
servation of these forms also increases.
Previous survey observations (27) demonstrated that polyphosphate bodies
were abundant in phytoplankton species that were distributed primarily along
the southern and southwestern shore of Saginaw Bay. This study also showed
that polyP was particularly abundant in blue-greens and was not found in
species of the Cryptophyceae. Although we have only compiled data for the
first two cruises, several features are already noteworthy. Polyphoshate has
been found of several members of the Cryptophyceae and is also found in a va-
riety of organisms along the northern shore, an area where previously none
was found. There is no reason to believe, at this time, that polyphosphate
abundance has declined as a result of a decreased phosphorus loading to the
bay, although preliminary phytoplankton species composition and abundance
data would indicate that shifts occurred in species composition and numbers
of algae were reduced.
233
-------
The results of the experimental uptake and degradation experiments sug-
gest that if species containing polyphosphate are transported out of Saginaw
Bay into southern Lake Huron, they would have a survival advantage unless
metals were incorporated in polyphosphate. Two additional organisms, .
Cyolote'i'la meneghini.cma and Soenedesmus quadriaauda, were also included in
the experimental treatments described earlier. Data collection at .this time ~
is incomplete. However, light microscopic examination of Scenedesmus re-
vealed that the cells survived longer than Plectonema, implicating more tol-
erance and/or exclusion of the metal. Samples have also been collected from
the western shore of southern Lake Huron. The normal water circulation of
the bay is counterclockwise with water exiting the bay along the southern
shore and being replaced by water entering the bay along the northern coast
(4). Stoermer and Kreis (27) have shown that populations which originate in
Saginaw Bay (e.g. Frag-Ltaria oapuaina) under certain conditions can survive
transport into the extreme southern part of Lake Huron. Survey data is still
being compiled from these two areas. However, we hope to determine if there
is a reduction in polyphosphate as the algae are transported into the less
nutrient rich waters of Lake Huron and whether there is any correlation with
species abundance in the lake and polyphosphate numbers in Saginaw Bay.
234
-------
REFERENCES
1. Baxter, M., and T. E. Jensen. 1980. Uptake of magnesium, strontium,
barium, and mangenese by Pteotonema. boryanwn (Cyanophyceae) with spe-
cial reference to polyphosphate bodies. Protoplasma 104: 81-89.
2. Chalkley, H. W. 1943. Methods for the quantitative morphological anal-
ysis of tissues. J. Nat. Cancer Inst. 4: 47-53.
3. Crang, R. E., and T. E. Jensen. 1975. Incorporation of titanium in
polyphosphate bodies of Anaoystis nidulans. J. Cell Biol. 67: 80a.
4. Danek, L. J., and J. H. Saylor. 1977. Measurements of summer currents
in Saginaw Bay, Michigan. J. Great Lakes Res. 3: 65-71.
5. Doonan, B. B. 197.8. Aspects of the alkaline phosphatase of Pleatonema
boryanwn and other selected cyanobacteria. Ph.D. Dissertation, The
City University of New York, New York.
6. Dozier, B. J., and P. J. Richerson. 1975. An improved membrane filter
method for the enumeration of phytoplankton. Verh. Internat. Verein.
Limnol. 19: 1524-1529.
7. Ebel, J. P., J. Colas, and S. Muller. 1958. Recherch.es cytochimiques
sur les polyphosphates. II. Mise au point de methodes de detection
cytochimiques specifiques des polyphosphates. Exptl. Cell Res. 15:
28-36.
8. Fitzgerald, G. P., G. C. Gerloff, and F. Skoog. 1952. Studies on chem-
icals with selected toxicity to blue-green algae. Sewage and Industrir
al Wastes 2: 888-896.
9. Glagoleff, A. A. 1933. On the geometrical methods of quantitative min-
eralogic analysis of rocks. TR. Inst. Econ. Min. and Metal, Moscow.
Vol. 59.
10. Guillard, R. R. L. 1975. Culture of phytoplankton for feeding marine
invertebrates, pp. 39-59. In W. L. Smith and M. H. Chanley (eds.),
Culture of marine invertebrate animals. Plenum Publishing, New York.
11. Harold, F. M. 1966. Inorganic polyphosphates in biology: structure,
metabolism, and function. Bact. Rev. 30: 772-794.
235
-------
12. Jensen, T. E. 1969. Fine structure of developing polyphosphate bodies
in a blue-green alga, Pleotonema bovyamm. Arch. Mikrobiol. 67: 328-
338.
13. Jensen, T. E., and L. M. Sicko. 1974. Phosphate metabolism in blue-
green algae. I. Fine structure of the "polyphosphate overplus" phe- ~
nomenon in Pleotonema boryanwn. Can. J. Microbiol. 9: 1235-1239.
14. Jensen, T. E., and L. Sicko-Goad. 1976. Aspects of phosphate utiliza-
tion by blue-green algae. Ecol. Res. Series, U.S. Environmental Pro-
tection Agency, Corvallis, OR. EPA Rep. No. EPA-600/3-76-103. 122 p.
15. Jensen, T. E., L. Sicko-Goad, and R. P. Ayala. 1977. Phosphate metabo-
lism in blue-green algae. III. The effect of fixation and post-stain-
ing on the morphology of polyphosphate bodies in Pleotonema boryanwn.
Cytologia 42: 357-369.
16. Lawry, N. H., and T. E. Jensen. 1979. Deposition of condensed phos-
phate as an effect of varying sulfur deficiency in the cyanobacterium
Syneohocoocus sp. (Anaaystis nidulans). Arch. Microbiol. 120: 1-7.
17. Lazinsky, D., and L. Sicko-Goad. 1979. Paraformaldehyde-glutaraldehyde
as a routine phytoplankton fixative. Micron 10: 49-50.
18. Shuman, H., A. V. Somlyo, and A. P. Somlyo. 1976. Quantitative elec-
tron probe microanalysis of biological thin sections: methods and va-
lidity. Ultramicroscopy 1: 317-339.
19. Sicko, L. M. 1974. Physiological and cytological aspects of phosphate
metabolism in the blue-green alga Pleotonema boryanum. Ph.D. Disserta-
tion, The City University of New York, New York.
20. Sicko-Goad, L., and T. E. Jensen. 1976. Phosphate metabolism in blue-
green algae. II. Changes in phosphate distribution during starvation
and the "polyphosphate overplus" phenomenon in Pleotonema boryanwn.
Amer. J. Bot. 63: 183-188.
21. Sicko-Goad, L., and D. Lazinsky. 1981. Accumulation and cellular ef-
fects of heavy metals in benthic and planktonic algae. In press,
Micron 12 (3).
22. Sicko-Goad, L., and E. F. Stoermer. 1979. A morphometric study of lead
and copper effects on Diatoma tenue v. elongation (Bacillariophyta).
J. Phycology 15: 316-321.
23. Smith, I. W., J. F. Wilkinson, and J. P. Duguid. 1954. Volutin produc-
tion in Aerobaoter aerogenes due to nutrient imbalance. J. Bacteriol.
68: 450-463.
24. Spitznagel, J. K., and D. G. Sharp. 1959. Magnesium and sulfate ions
as determinants in the growth and reproduction of Myoobaateiciim bovis.
J. Bacteriol. 78: 453-462.
236
-------
25. Stevens, S. E., D. A. M. Paone, and D. L. Balkwill. 1981. Accumulation
of cyanophycin granules as a result of phosphate limitation in
Agmenelltm quadruplicatum. Plant Physiol. 67: 716-719.
26. Stewart, W. D. P., and G. Alexander. 1971. Phosphorus availability and
nitrogenase activity in aquatic blue-green algae. Freshwater Biol. Ir
389-404.
27. Stoermer, E. F., and R. G. Kreis, Jr. 1980. Phytoplankton composition
and abundance in southern Lake Huron. U.S. Environmental Protection
Agency, Duluth, MN. EPA Rep. No. EPA-600/3-80-061. 384 p.
28. Stoermer, E. F., L. Sicko-Goad, and D. Lazinsky. 1980. Synergistic ef-
fects of phosphorus and heavy metal loadings on Great Lakes phytoplank-
ton, pp. 171-186. In W. R. Swain and V. R. Shannon (eds.), Proc. Symp.
Theoretical aspects of aquatic toxicology, Barok, Jaroslav, USSR. U.S.
Environmental Protection Agency, Duluth, MN. EPA Rep. No. EPA-600/9-
80-034.
29. Voelz, H., U. Voelz, and R. 0. Ortigoza. 1966. The "polyphosphate
overplus" phenomenon in Myxococcus xanthus and its influence on the
architecture of the cell. Arch. Mikrobiol. 53: 371-388.
237
-------
Comments on
Polyphosphate in Natural Phytoplankton Assemblages
and its Formation and Degradation
in Cultured Algal Cells
Joe M. King
Murray State University
Dr. Sicko-Goad proposed a three-year field and laboratory investigation
to: 1) determine the relationship of polyphosphate formation to dispersal
and dominance of nuisance producing phytoplanktonic populations in
aquatic systems, 2) determine if polyphosphate formation is a significant
mechanism in uptake, biological transport, and bioaccumulation of toxic
heavy metals, and 3) investigate synergistic effects of combined phosphorus
and heavy metal loadings to determine if such loadings have a greater
potential for producing undesirable modifications in the composition and
structure of phytoplanktonic communities and enhancing bioaccumulation
and dispersal through the food chain. Phytoplankton samples will be
collected from Saginaw Bay, a highly eutrophic area, and the western
shore of southern Lake Huron, a less eutrophic area of the system.
Polyphosphate distributions and abundance in the collected populations
will be determined by use of a lead sulfide precipitation procedure
which stains polyphosphate bodies for light microscopic examination and
by transmission electron microscopy. Laboratory investigations will be
conducted with natural phytoplanktonic populations or isolates of algae
from the Great Lakes. Polyphosphate formation and degradation will be
studied to determine if polyphosphate alters succession by either increasing
growth of species containing polyphosphate or by altering assemblage
composition through stimulation of other species through release of
phosphorus by the polyphosphate containing algae. Similar studies
including simultaneous additions of phosphate salts and heavy metals ;
will be conducted to determine if polyphosphate serves as a bioaccumulations
pathway for heavy metals and allows subsequent toxicant release upon
degradation. Heavy metals will be analyzed in situ by x-ray energy
dispersive analysis and scanning transmission electron microscopy. In
addition, synergestic effects of combined nutrient and heavy metal
loadings will be investigated in microcosms composed of sediments and
natural populations from western Lake Huron or Saginaw Bay.
Dr. Sicko-Goad has used her existing research data base to formulate
some intriguing questions about the ecological significance of poly-
phosphates in the phytoplanktonic populations of Lake Huron. Answers to
these questions will provide the scientific community with fundamental
information on phosphorus kinetics and mechanisms of transport of toxic
238
-------
heavy metals in aquatic systems. Field sampling will be interfaced with
the Saginaw Bay Special Field Survey, in which extensive data will be
collected on both chemical parameters and phytoplankton species composition
and abundance. Data from a previous investigation of Saginaw Bay will
be useful in determining if significant biological transport of phosphate
from a highly eutrophic area to less affected parts of the ecosystems
has occurred and if a 50% reduction in phosphorus loading to Saginaw Bay
has resulted in a decrease of stored polyphosphates, or if polyphosphate
formation is stimulated by the other factors, such as the presence of
trace or heavy metals. The laboratory investigations will provide more
specific information on mechanisms of uptake, transport, release/avail-
ability of heavy metals by algae. Dr. Sicko-Goad, the Principal Investigator,
has extensive experience in electron-microscopy and has used the proposed
methods in previous investigations. The Co-Principal Investigator,
Dr. E.F. Stoermer, has an extensive background in the taxonomy and
ecology of algae. Both investigators are prolific publishers and their
past work constitutes an excellent basis for the proposed work. Facilities
are readily available for conducting the proposed research and the
investigators have used the equipment in previous research projects.
As indicated in her report, Dr. Sicko-Goad has made significant
progress toward obtaining her stated objectives. Analysis of field data
is underway and effects of phosphate and lead treatments have been noted
at the subcellular level. Technical problems have been identified and
additional experiments have been planned.
239
-------
OIL AND RELATED CONTAMINANT EFFECTS ON WATERFOWL IMMUNE DEFENSES:
A PROGRESS REPORT
Tonie E. Rocke and Thomas M. Yuill
Department of Veterinary Science
University of Wisconsin
Madison, Wisconsin 53706
ABSTRACT
Experiments have been designed to study the effects of crude oil,
petroleum distillates, and oil spill control chemicals on resistance of
waterfowl to bacterial or viral infection and their immune responsiveness,
both antibody and cell mediated. Toxicity studies have been conducted to
determine mallard sublethal oral doses of various contaminants, including
South Louisiana crude oil, Bunker C fuel oil, a dispersant-Corexit 9527,
and oil/dispersant combinations. Additional assays for measuring immune
function in mallard ducks are being developed.
INTRODUCTION
Increased oil consumption with subsequent rises in production and
transport has inevitably resulted in the pollution of aquatic ecosystems.
Tanker groundings, collisions, blowouts, and ruptured pipelines have
spilled oil into oceans, rivers, lakes, inland canals, and estuaries, but
these accidents account for only 5% of all oil entering the marine
environment (4). Most oil is introduced via discharges from normal
transport and refining operations. Along with urban runoff and natural
seeps, these intentional discharges have induced persistent or chronic
levels of oil pollution near major ports and other areas. It has been
estimated that more than 1.5 million metric tons of oil or petroleum
products enter U.S. coastal waters each year (4).
The damaging effects of oil on local flora and fauna following major
spills have been well documented. Massive dieoffs of animals in the
vicinities of oil slicks have received considerable public attention, and
oil toxicity studies have been conducted on numerous marine organisms.
Much information has been gathered on acute, short term toxicity and
carcinogenic properties of oil, however little is known about the sublethal
effects of oil, particularly with chronic exposure. Sublethal effects are
240
-------
more subtle than direct mortality and may persist for an extended period of
time with long term repercussions.
Efforts to contain oil pollution have often included application of
chemical dispersants. Dispersants are intended to remove oil from the
water surface, prevent the formation of oil in water emulsions and reduce
the ability of oil to adhere to objects. However, chemically dispersed oil
may persist longer in the environment, penetrating sand and gravel beaches
at a higher rate than non-dispersed oil. Oil spill control chemicals may
compound the problems of oil toxicity. The effects of dispersants on
wildlife have not been fully investigated. Some combinations of oil and
dispersant have been reported to be more damaging than either contaminant
alone (1).
Waterfowl have been repeatedly subjected to oil pollution at both
acute and chronic levels. High mortality in waterfowl populations after
major spills has long concerned ornithologists, hunters, and other wildlife
enthusiasts. In general, the effects of oil on waterfowl fall into two
categories: physical effects, such as oiled feathers, and systemic
effects, which result from the ingestion of oil. Systemic effects can
develop from either acute or chronic exposure of birds to oil.
Waterfowl ingest oil by eating contaminated food or drinking water
that contain oil in suspension. Laboratory studies have shown that ducks
will voluntarily eat food that has been contaminated with petroleum or
petroleum products (7,8,9,18). Additionally, observers in the field have
noted that birds try to remove oil from their feathers by preening. Ducks
will preen up to 50% of the contaminating oil from their feathers during
the first eight days after exposure (5). Studies using radioactively
labelled crude oil have demonstrated that significant amounts of oil are
ingested by contaminated ducks while preening (5). It has been estimated
that birds in the wild could encounter and ingest 2-3 gram doses of oil per
kilogram of body weight (6).
Most studies of oil ingestion by waterfowl have concentrated on
symptoms of overt toxicity and physiological effects. Histological
examination of tissues collected from oil contaminated birds reveal
irritation of the gastrointestinal mucosa, lipid pneumonia, fatty
degeneration of the liver, acinar atrophy of the pancreas, adrenocortical
hyperplasia, and toxic nephrosis (6). In a study on the effects of mild
cold stress on ducks maintained on petroleum laden diets, adrenal
hypertrophy and-severe involution of lymphoepithelial tissues were
characteristic in all birds that died (9). Particularly interesting were
the effects on the thymus, bursa of Fabricius, and spleen weights. In
another study on the effects of chronic ingestion of crude oil on mallard
ducklings, liver hypertrophy and splenic atrophy were discovered in birds
fed 2.5 and 5.0% oil diets (18).
So far no attempts have been made to correlate the observed
pathological signs of oil toxicity with altered immunological
responsiveness and defense of birds against infectious organisms. Certain
241
-------
polycyclic aromatic hydrocarbons (PAH) are known to suppress the immune
response of some animals. Two such hydrocarbons, benzo-a-pyrene (BAP) and
derivatives of benz-a-anthracene (BAA), are present at significant levels
in many petroleum products (14). Experiments have shown that injection of
single low doses of BAP and BAA derivatives into mice prior to
administration of an antigen result in depressed levels of antibody forming
cells and impairment of graft rejection (16,17). The input of BAP into the
marine environment from petroleum spillage has been estimated at 20-30 tons
annually (14). Certain metals such as cadmium and lead, which may be
closely associated with or derived from constituents of petroleum have also
been found to be immunosuppressive (10). Interestingly, an avian cholera
outbreak followed a major oil spill in 1978 in the Chesapeake Bay area,
resulting in the death of over 10,000 birds (12). Although there was no
evidence obtained to relate the two events, the incident bears further
investigation.
Our research goal is to both develop and test an avian system for
determining the impact of oil and oil related contaminants on host immune
defenses, using the mallard duck as a model. The mallard duck (Anas
platyrhynchos) is an ideal candidate for oil iramunotoxicology studies.
Mallards are known to frequent or inhabit coastal areas, marshes, and
estuaries that are subject to periodic oil contamination. Postmortem
analysis of mallards exposed to petroleum by stomach intubation have shown
that petroleum hydrocarbons can accumulate in various tissues (11).
Furthermore, mallards are readily available and easily maintained in
captivity.
RESEARCH PLANS AND PRELIMINARY RESULTS
Several techniques have recently been developed in our laboratories
that permit rapid, economical assessment of alterations in the immune state
of mallards. Additional measures of immune function, such as blastogenesis
responses to antigens and cell mediated immune responses are currently
being adapted to mallards. Toxicity studies have been completed for five
contaminants; mallard sublethal oral doses have been established for
each. By incorporating the findings of these preliminary studies on
mallard immune function and oil toxicity, experiments will be designed to
test the hypothesis that oil and related contaminants alter host immune
defenses and resistance to infectious disease.
EXPERIMENTAL ANIMALS
All of the experiments in this study are being conducted on adult male
mallards purchased from game farms. The ducks are housed individually in
cages in the isolation building at Charraany Farms or non-isolated animal
rooms at Fred Hall, University of Wisconsin, Madison. A photoperiod of 8
hours light, 16 hours dark is maintained for the duration of every
experiment.
242
-------
ENVIRONMENTAL CONTAMINANTS
Oil samples were obtained from the American Petroleum Institute. For
biological effects studies, the institute provides a suite of four
reference oils consisting of two representative crude oils and two fuel
oils. This study will include South Louisiana crude oil (SLCO), which is
an example of a typical major Gulf of Mexico crude oil transported
extensively by tankers along the coastal waters of the eastern U.S., and
Bunker C fuel oil (BCFO), which is used for powering marine vessels.
Corexit 9527, a water soluble, self mixing dispersant, is being tested
alone and in combination with both oils. Samples of Corexit have been
provided by Exxon.
TOXICITY STUDIES
Twenty-eight day toxicity studies have been completed for five
contaminants, SLCO, BCFO, Corexit 9527 (diluted in at least 9 parts of
water), SLCO/Corexit and BCFO/Corexit mixtures (50 parts oil to 1 part
Corexit). Various amounts of each contaminant were orally administered to
adult male mallards daily by stomach tube. Mortality, behavioral
abnormalities, and weight changes were monitored. Plasma was collected and
stored for future enzyme analysis. At the termination of each experiment,
birds were sacrificed for post mortem analysis. Spleen, liver, and kidney
weights were recorded. Spleen, liver, kidney, gastrointestinal lymphoid,
heart, lung, brain, skeletal muscle, and testis tissues were collected and
saved for histological examination.
Our experiments indicate that mallards can ingest a relatively large
amount of oil with no apparent signs of illness. In this study, ingestion
of up to 15 ml of oil per kilogram of body weight per day was usually not
lethal. One out of five birds receiving 15 ml/kg BCFO per day died. None
of the birds receiving SLCO died. Some of the birds were able to
regurgitate the administered oil, particularly those receiving BCFO. At
the low levels, regurgitation was rarely observed. Although most birds
lost weight during the course of these experiments, there was no
significant difference in weight change between treated birds and controls ;
(Tables 1 and 2). Likewise, mean liver and kidney indices (calculated as
ratio of liver or kidney weight to total body weight at time of
death x 10,000) did not differ significantly between treatments in either
of the oil toxicity experiments (Tables 1 and 2). A decrease in mean
spleen weight in birds receiving _>2.5 ml/kg BCFO was the most notable
effect of oil ingestion (Tables 1 and 2). Figure 1 depicts the
relationship between mean spleen index and dose of BCFO administered. A
corresponding decline is illustrated in a plot of mean spleen index versus
dose of BCFO (Figure 2). These trends are not the result of decreasing
body weight. There was no significant difference in spleen weights between
controls and groups treated with SLCO.
243
-------
TABLE 1. MEAN BODY AND ORGAN WEIGHT INDICES OF MALLARD DUCKS SURVIVING
INGESTION OF SOUTH LOUISIANA CRUDE OIL FOR 28 DAYS*
Dose
ml /kg
0.0
1.0
2.5
4.0
7.0
12.0
Mean body
weight change
.06
.06
.12
.09
.10
.09
±
±
±
±
±
±
.05
.04
.08
.04
.06
.02
Mean
liver indext
263
248
263
255
269
259
± 41
± 18
± 21
± 36
± 53
± 19
Mean
kidney indext
50
46
54
46
41
47
± 9
± 5
± 22
± 9
± 7
± 4
Mean
spleen indext
7.4
6.0
5.8
6.7
6.3
6.5
±
±
±
±
±
±
1.2
1.3
1.8
1.8
1.6
1.9
Deaths
0/5
0/5
0/5
0/5
0/5
0/5
* None of these treatments were significantly different by one-way
analysis of variance
Organ Weight (g)
Final body weight (g)
x 10,000
TABLE 2. MEAN BODY AND ORGAN WEIGHT INDICES OF MALLARD DUCKS
SURVIVING INGESTION OF BUNKER C FUEL OIL FOR 28 DAYS
Dose
ml /kg
0.0
1.0
2.5
4.0
7.0
12.0
Mean body
weight changeT
.05
.16
.20
.15
.08
.25
±
±
±
±
±
±
.04
.11
.10
.17
.12
.11
Mean
liver indext
250 ±
323 ±
310 ±
336 ±
351 ±
354 ±
64
37
26
59
75
81
Mean Mean
kidney indext spleen indext
43
49
46
44
46
43
± 6
± 6
± 2
± 5
± 7
± 3
5.4
4.3
3.8
3.3
3.5
2.9
± 1
± 1
± 1
± 1
± 1
± 0
.5
.2
.3*
.1*
.1*
.9*
Deaths
0/4
0/5
l/5§
0/5
0/5
1/5
* Significantly different from controls at <* » .10
§ Cause of death undetermined - not directly due to oil toxicity
t
Organ Weight (g)
Final body weight (g)
x 10,000
244
-------
7.0
6.0 H
I 5.0 -I
z
Cd
J
^
CO
z
3.0 -
J
2.0 -i
1.0
0
I I
2.0
4.0
6.0
8.0
10.0
12.0
DOSE (ml/kg)
Figure 1. Mean spleen index versus dose of Bunker C fuel oil orally
administered to mallards for 28 days.
.8 ~
.7
.6
z
H
W
J
.5 H
.3 -
.2 -
.1 -
0 -j
[ I I I I I I I I I
2.0
4.0
6.0
8.0 10.0
12.0
DOSE (ml/kg)
Figure 2. Mean spleen weight versus dose of Bunker C fuel oil orally
administered to mallards for 28 days.
245
-------
In a study of the effects of ingested dispersant (Corexit 9527), a
temporary nervous system disorder was noticed in birds receiving 5.0 ml/kg
of a 1:10 Corexit/water mixture (0.5 ml Corexit). Motor uncoordination and
reduced mobility occurred within a few minutes post administration but
lasted only a few hours. By the next day the birds had returned to
normal. None of the birds died over the 28 day treatment period and no
significant weight loss was noted. Liver, kidney and spleen indices of
treated birds did not differ from controls (Table 3).
TABLE 3. MEAN BODY AND ORGAN WEIGHT INDICES OF MALLARDS SURVIVING
INGESTION OF COREXIT OIL DISPERSANT FOR 28 DAYS*
Treatment
ml Corexit/ml DDH20
0/5
.006/4.994
.012/4.988
.025/4.975
.05/4.95
Mean
liver index
207 ± 29
191 ± 22
174 ± 49
200 ± 9
184 ± 25
Mean
kidney index
33 ± 5
38 ± 7
39 ± 4
42 ± 3
40 ± 8
Mean
spleen index
3.9 ± 0.8
4.6 ± 1.3
4.4 ± 1.3
4.3 ± 1.5
3.9 ± 0.9
Deaths
0/6
0/6
0/6
0/6
0/6
* None of these treatments were significantly different by one-way analysis
of variance
t Orsan Weight (g) x 10 000
Final body weight (g)
Two oil/dispersant combinations were included in the next toxicity
study, SLCO/Corexit and BCFO/Corexit. The mixtures were prepared by adding
one part Corexit 9527 to 50 parts of either SLCO or BCFO and shaking for
ten minutes on a mechanical shaker. This ratio reflects realistic
applications of Corexit 9527 to oil spills. Treated birds received 1.0,
4.0, or 7.0 ml of one of the mixtures. The motor uncoordination observed
when Corexit alone was administered did not occur in this experiment. One
bird receiving 4.0 ml/kg body weight BCFO/Corexit died on day 24 of the
experiment. Mean spleen indices of survivors of BCFO/Corexit differed
significantly from both controls and birds treated with SLCO/Corexit
(Table 4). Figure 3 illustrates these results. Interestingly there was a
notable increase in mean liver weights of BCFO/Corexit treated birds
compared to both controls and SLCO/Corexit treated birds (Figure 4). Liver
hypertrophy had not been observed in previous toxicity studies.
246
-------
TABLE 4. MEAN BODY AND ORGAN WEIGHT INDICES OF MALLARD DUCKS SURVIVING
INGESTION OF OIL DISPERSANT MIXTURES FOR 28 DAYS*
Treatment
Control
SLCO/C
SLCO/C
SLCO/C
BCFO/C
BCFO/C
BCFO/C
Dose
ml/kg
1.0
4.0
7.0
1.0
4.0
7.0
Mean
liver index
213 ± 3la
277 ± 43ab
259 ± 16ab
268 ± 40ab
297 ± 53b
395 ± 69°
349 ± 45C
Mean
kidney index
40 ± 4a
45 ± 2a
43 ± 3a
41 ± 4a
43 ± 6a
43 ± 3a
44 ± 5a
Mean
spleen index
4.3 ± 1.5ab
4.7 ± 0.9a
4.5 ± 0.8ab
4.4 ± 0.8a
2.8 ± 0.9bc
2.8 ± 0.6C
2.7 ± 0.5C
Deaths
0/4
0/5
0/5
0/5
0/5
1/5
0/5
* Indices not showing a common letter differ significantly from one
another at a = .05
Organ Weight (g)
Final body weight (g)
x 10,000
i
g 3.0
« 2.0
1
a i.o
0
s
\
^
1 1 ( 1 t I 1
0 2.0 4.0 6.0
DOSE (ml/kg)
. /
~ .6 -
00
£ .5 -
O
M
* .4 -
z
w
si
en
< 2 -
.1 -
*
/
\
\
— ill M
— -^
•
\
J-
[ — -I
•"•
' 1 i i i i I i i
8.0 o 2.0 4.0 6.0 8.0
SLCO/COREXIT DOS£ (ml/kg)
— — -— BCFO/COREXIT
Figure 3. Mean spleen index and weight versus dose of Bunker C fuel
oil/Corexit administered to mallards for 28 days.
247
-------
600
450
300
150 J
i ii l I
2.0 4.0 6.0
DOSE (ml/kg)
6.0
5.0
8.0
w
3 3.0 H
1.0
SLCO/COREXIT
— BCFO/COREXIT
3.0
DOSE (ml/kg)
i
7.0
Figure 4. Mean liver index and weight versus dose of Bunker C fuel
oil/Corexit administered to mallards for 28 days.
KNOWN IMMUNOSUPPRESSANTS
The effects of oil and oil-related contaminants on host defenses and
the avian immune system will be assessed and compared to those of known
immunosuppressants. By definition immunosuppressants are products capable
of suppressing or depressing the development of at least one type of immune
reaction. These agents do not all share the same target cells nor the same
cytologic action. Consequently the various categories of immune responses
differ in their sensitivity to immunosuppressive agents. The effects of
two known immunosuppressants, cyclophosphamide (CY) and azothioprine (AZ)
will be used as standards for comparison.
IMMUNOTOXICOLOGY STUDIES
Since immunosuppressive chemicals may alter an animal's response to
disease agents by interfering with one or more functions of the immune
system, several tests are necessary in immunotoxicologic studies in order
to evaluate potential risks of contaminants and to discern their mode of
action. In this study the screening of potential immunosuppressants will
be based on four experiments. Several functional aspects of the avian
immune system will be tested in these experiments, including resistance to
infectious disease, antibody mediated immunity, antigen and mitogen induced
lymphocyte blastogenesis, and some measure of cell mediated immunity.
1) Protection against infectious agents depends on two components,
specific immunity, which includes cell mediated immunity and antibody
248
-------
mediated immunity, and nonspecific resistance, which is a function of
enzymes (such as lysozyme), the interferon system, innate or genetic
characteristics, phagocytes, etc. Thus, resistance to challenge with an
infectious agent measures overall response. Oil treated mallards will be
exposed to two infectious agents - avian cholera, an acute bacterial
infection, and duck plague virus (DPV), a disease that can persist in a
carrier state in apparently healthy waterfowl.
Avian cholera serotypes of known origin were obtained from the
National Fish and Wildlife Health Lab, Madison, Wisconsin. Selection of
the appropriate dose and strain of the organism to be used in these
experiments is very critical. We are currently conducting challenge
experiments on different inoculums in untreated mallards. When the
appropriate dose of bacterium has been established, it will be injected
into oil treated mallards; percent mortality will be measured.
Several ducks that survived avian cholera infection in our early
challenge experiments were included two months later in a pilot study on
the toxic effects of ingested oil/dispersant mixtures. After 10 days one
mallard receiving approximately 5.0 ml of a 1:30 BCFO/Corexit mixture
died. Upon necropsy, typical lesions of avian cholera were noted,
including an enlarged spleen and a haemorrhagic heart. Pasteurella
multocida was subsequently isolated from the spleen. Although no
conclusions can be drawn from this isolated observation, the circumstances
suggest that the BCFO/Corexit intoxication may have altered host resistance
to some extent, possibly causing an abrupt upwards shift in the level of
_P_. multocida infection resulting in septicemic death of the bird.
2) Duck plague virus will be used to assess the effects of oil and
oil-related compounds on persistent infections. Carrier birds, defined on
the basis of oral erosions under the tongue and/or shedding of virus in the
feces, will be given oral doses of each contaminant. The amount and
frequency of DPV shedding will be determined over the 28 day treatment
period and two weeks thereafter, and compared to controls. DPV shedding
will be determined by inoculation of duck embryo fibroblastic cell cultures
with eluates from periodic oral and cloacal swabs. The cells will be
tested for DPV by our standard direct fluorescent antibody test. Shedding
of virus in feces will be evaluated in terms of altered immune response.
In vitro assays of lymphocyte proliferation upon stimulation with
mitogens, such as phytohemagglutinin (PHA) or pokeweed mitogen (PWM), have
been extensively used to index lymphocyte activity. Lymphocyte stimulation
assays with various mitogens have been applied to mallards by Drs. Sauch
and Hinsdill. A similar but more specific assay is antigen-induced
lymphocyte immunostimulation (13). This assay is currently being developed
with an antigenic preparation of DPV vaccine. In this study, the compounds
under test will be administered to DPV carrier mallards. Peripheral blood
lymphocytes will be collected before and after contamination for immuno-
stimulation assays with inactivated DPV vaccine.
3) The production of antibodies to a specific antigen requires the
249
-------
action of several subpopulations of cells, including B cells and T helper
cells. In this study of general immune responsiveness, antibody mediated
immunity is measured by counting spleen plaque forming cells. For this
procedure, mallards are immunized with sheep red blood cells (SRBC). Four
days after the initial immunization, birds are sacrificed to remove their
spleens. Direct spleen plaque forming cell assays are performed according
to methods developed by Cunningham and Szenberg (3) that have been adapted
for mallards by Drs. Sauch and Hinsdill.
A preliminary study testing the effects of oil pollutants on antibody
mediated immunity was recently completed. Groups of mallards were treated
with either SLCO, BCFO, or Corexit for 28 days. On day 29 birds were
inoculated intravenously with SRBC. A known immunosuppressant, cyclo-
phosphamide (CY), was orally administered to another group two times prior
to SRBC inoculation. Four days after the injection the birds were
sacrificed. Mean liver, kidney and spleen weights are recorded in Table
5. As before, mean spleen weight of birds receiving BCFO is lower than the
control value. Mean spleen weight of CY treated birds is similarly
reduced. Figure 5 illusrates this comparison.
TABLE 5. MEAN ORGAN WEIGHT INDICES OF MALLARD DUCKS AFTER INGESTING OIL,
DISPERSANT, OR CYCLOPHOSPHAMIDE*
Treatment
Control
Corexit (1:10)
SLCO
SLCO
BCFO
BCFO
CY
Dose
ml /kg
5.0
1.0
5.0
1.0
4.0
50.0
Mean
liver indext
169 ± 25a
164 ± 18a
178 ± 22a
187 ± 22a
199 ± 26a
172 ± 31a
163 ± 36a
Mean
kidney indext
32 ± 4a
33 ± 4a
32 ± 6a
34 ± 6a
34 ± 3a
33 ± 4a
28 ± 4a
Mean
spleen indext
5.7 ± 1.0ab
6.0 ± 1.2ab
5.5 ± 0.8ab
6.2 ± 2.0a
4.8 ± 0.6bc
4.5 ± l.lc
4.4 ± 1.4C
* Indices not showing a common letter differ significantly from one
another at a = .05.
t Organ Weight (g) x 10 QOO
Final body weight (g)
The number of plaque forming cells (PFC's) to SRBC in the spleen of a
sensitized animal is an indication of the antibody forming capability of
that animal. The number of PFC's are standardized between animals by
counting the number of viable cells per gram of spleen. In Figure 6, which
illustrates data collected from this preliminary study, mean PFC's per 10
viable cells are expressed as a percentage of the mean control value.
250
-------
9.0-
X 8.0-
LU
1 7-°"
2 6.0-
UJ
a 5.0-
co
z 4.0-
UJ
2 3.0-
•r
•
M
JJ
™
I
i
2
=
i
i
E
E
IMEAN SPLEEN INDEX
= MEAN SPLEEN WT.
i
•
I
i
=
i
••
•;
~
•t.o
_q m
.9 >
z
-.8 co
13
-.7 m
m
z
~'6 H
-.5 ^
CO
-.4 p
Control Corexit SLCO SLCO BCFO BCFO CY
(1:10) Iml/kg/d 5ml/kg/d Iml/kg/d 4ml/kg/d 50mg/kg
5ml/kg/d
Figure 5. Mean spleen index and spleen weight versus toxicant administered
to mallards.
150-
140-
cn
j
g 130-
01
£g 120-
8*| HO-
IBER OF PFC's
EXPRESSED AS
00 >0 <
o o c
1 1
I 70-
1 60-
50 -
COREXIT SLCO SLCO BCFO BCFO CY
(1:10) Iml/kg/d 5ml/kg/d Iml/kg/d Aml/kg/d 50mg/kg
5ml/kg/d
Figure 6. Mean number of PFC's per 10 viable cells expressed as percent
control versus toxicant administered to mallards.
251
-------
Those receiving 4 ml/kg BCFO had the highest mean PFC value.
Unfortunately, many of the birds in this group were able to regurgitate the
administered oil and did so nearly every day. Therefore, data from this
treatment group is unreliable, and future experiments will be redesigned to
prevent regurgitation. Those receiving 1 or 5 ml/kg SLCO or 1 ml/kg BCFO
had the most reduced count; even lower than the CY treated group. This is
presumptive evidence of oil induced supression of antibody formation.
However, we hesitate to draw conclusions from this study because of high
variability within treatment groups. The effects of these chemicals on
antibody mediated immunity will be tested further.
4) Cellular mediated immune (CMI) responses include phenomena such as
rejection of grafts, destruction of tumor and virus-infected cells, and
hypersensitivity to certain organisms such as Mycobacterium tuberculosis.
These types of responses are commonly assayed in inbred mammalian hosts;
for avian species, however, knowledge regarding cellular mediated immunity
is limited. Delayed type hypersensitivity (DTK), an assay used
successfully in chickens (15), will be attempted in this study. Mallard
ducks sensitized to killed strains of II. avium by intraperitoneal injection
will be tested for the DTK response by injection of avian purified protein
derivative (PPD) into wing webs, foot webs, or the mucocutaneous junction
of the cloaca. The site which yields the most consistent DTK response and
the appropriate time of PPD injection will be determined. If feasible,
this assay will be incorporated into later experiments for assessing the
effects of oil pollutants on cell-mediated immunity of ducks.
CONCLUDING REMARKS
It is known that waterfowl encounter a wide variety of environmental
contaminants, however the influence of these compounds on the immune system
has not been completely assessed. A depressed immune state in a densely
concentrated flock may increase susceptibility to infectious agents.
Epidemics in certain wildlife populations could substantially reduce
recreational and economic benefits of an important natural resource. For
certain endangered waterfowl species, an outbreak of disease might be
devastating, especially if it is associated with environmental
contamination, such as an oil spill. Other animals, including man, are
similarly exposed to many pollutants. The findings of this study on
waterfowl might have important human health implications. We will continue
our work to develop assays for assessing the immune status of mallard
ducks. Once established, these assays will provide a system for screening
other potential immunosuppressants as well as oil contaminants.
252
-------
REFERENCES
1. Albers, P. Oil dispersants and wildlife. In: Columbus Brown (ed),
Proceedings of the 1979 Pollution Response Workshop. U. S. Fish and
Wildlife Services, 1979, pp. 67-71.
2. Bach, J.R. Immunoraanipulation. In: J.F. Bach (ed), Immunology. New
York: John Wiley and Sons, 1978, pp. 813-840.
3. Cunningham, A.J., and Szenburg, A. 1968. Further improvements in the
plaque techniques for detecting single antibody-forming cells.
Immunol. 14: 599-601.
4. Farrington, J.W. 1977. Oil Pollution in the Coastal Environment.
In: Estuarine Pollution Control and Assessment. EPA 440/1-77-007.
5. Hartung, R. 1963. Ingestion of oil by waterfowl. Papers of the
Mich. Acad. of Sci., Arts, Let. 68: 49-55.
6. Hartung, R., and Hunt, G. 1966. Toxicity of some oils to
waterfowl. J. Wildl. Manag. 30: 564-569.
7. Holmes, W.N., Cavanaugh, K.P., and Cronshaw, J. 1978. The effects of
ingested petroleum on oviposition and some aspects of reproduction in
experimental colonies of mallard ducks (Anas platyrhynchos).
J. Reprod. Fert. 54: 335-347..
8. Holmes, W.N., Cronshaw, J., and Gorsline, J. 1979. Some effects of
ingested petroleum on seawater adapted ducks (Anas platyrhynchos).
Env. Res. 17: 177-190.
9. Holmes, W.H., Gorsline, J., and Cronshaw, J. 1979. Effects of mild
cold stress on the survival of seawater-adapted ducks (Anas
platyrhynchos) maintained on food contaminated with petroleum. Env.
Res. 20: 425-444.
10. Roller, L.D. 1973. Immunosuppression produced by lead, cadmium, and
mercury. Am. J. Vet. Res. 34: 1457-1458.
11. Lawler, F.D., Loong, W., and Laseter, J.L. 1978. Accumulation of
aromatic hydrocarbons in tissues of petroleum-exposed mallard ducks
(Anas platyrhynchos) Env. Sci. Technol. 12: 51-54.
12. Montgomery, R.P., Stein, G. Jr., Scotts, V., and Setter, F. 1979.
The 1978 epornitic of avian cholera on the Chesapeake Bay. Avian Dis.
24: 966-978.
13. Maheswaran, S.K., and Thies, E.S. 1979. Pasteurella multocida
antigen induced in vitro lymphocyte immunostimulation, using whole
blood from cattle and turkeys. Res. Vet. Sci. 26: 25-31.
14. Neff, J.M. Polycyclic aromatic hydrocarbons in the aquatic
environment: sources, fates, and biological effects. American
Petroleum Institute.
15. Rose, E.M., and Bradley, J.W.A. 1977. Delayed hypersensitivity in
the fowl, turkey, and quail. Avian Pathol. 6: 313-326.
253
-------
16. Stjernsward, J. 1966. Effect of Noncarcinogenic and carcinogenic
hydrocarbons on antibody-forming cells measured at the cellular level
in vitro. J. Nat. Cancer Inst. 36: 1189-1195.
17. Stjernsward, J. 1965. Immunodepressive effect of
3-methylcholanthrene. Antibody formation at the cellular level and
reaction against weak antigenic homografts. J. Nat. Cancer Inst. 35:
885-892.
18. Szaro, R.C., Dieter, M.P., Heinz, G.H., and Ferrell, J.F. 1978.
Effects of chronic ingestion of South Louisiana crude oil on mallard
ducklings. Env. Res. 17: 426-436.
254
-------
Comments on
Oil and Related Contaminant Effects
on Waterfowl Immune Defenses
Hollie L. Collins
University of Minnesota-Duluth
The objectives of this project are to measure the resistance of
mallard ducks to bacterial or viral infections and their immune response
after pretreatment with crude oil, petroleum distillates, and selected
chemicals dispersants.
The importance of this project stems frcm the fact that offshore
drilling operations and oil transport via marine vessels have intensified
in recent decades, siginificantly increasing the input of oil into
aquatic ecosystems. Accidental spills derived from these coupled with
bilge washing, runoff frcm land based operations, ruptured pipelines,
and natural seeps have intensified oil pollution near major ports and
other areas.
In order to control oil spills, chemical dispersants are often
applied to enhance the removal of oil frcm water surface and from objects.
As a result, the dispersed oil may be more persistent in the environment
by penetrating beach materials at a higher rate than nondispersed oil.
Massive dieoffs of animals following major oil spills have been
well documented, but little is known of the sublethal effects of oil and
particularly the effects of chronic exposure. Flirthermore, less is
known of the effects of oil dispersants and the detergents used to clean
animals exposed to oil. Combinations of oil and dispersant may compound
the problem. Most animal recovery efforts have been unsuccessful,
especially when attempting to rescue and rehabilitate waterfowl by
cleaning oil with the aid of detergents and solvents. This project
does not deal with the direct physical or external effects of oil or
chemicals, but rather the systemic effects which result frcm ingestion
of oil and/or dispersants.
Many studies have detailed the symptoms of overt toxicity and
physiological effects of ingestion of oil by waterfowl but no attempts
have been made to correlate these findings with alterations in immune
responses and defense of birds against infections organisms. Since
other environmental contaminants have been shown to alter immune function
in a variety of animals, compounds in oil distillates, dispersants, and
detergents may also have effects on host defenses. At least one
255
-------
significant avian cholera outbreak followed a major oil spill in 1978
resulting in massive mortality to birds. There was no evidence obtained
to relate the two events but such rouses suspicion and support for
investigation of such an incident.
The Environmental Protection Agency's (EPA's) Environmental Biology
Review Panel reconmended approval of this project because it would be
beneficial to determine if oil and the chemicals used to clean up birds
and the environment could substantially alter an animal's resistance to
infections disease. If such effects are found, this research will
provide a system for screening other potential environmental hazards to
waterfowl. The use of dispersants or detergents that interfere with
disease resistance should then be prohibited.
Waterfowl are literally under the gun from a variety of angles,
including pollution effects. Corrective measures are being employed in
some instances, e.g. steel shot for hunting in place of lead in the case
of heavy metal effects. Here is a chance of finding solutions to another
possible chronic toxicity problem.
Lastly, I find this an interesting new avenue of investigation for
EPA. This is an area (immune-suppression) typically housed in other
agencies such as the National Science Foundation or the National Institutes
of Health. I think it refreshing that we delve into the development of
this technology and make use of the expertise of Dr. Yuill's group.
256
------- |