EPA/600/J-94/283
Fish Intersexuality as Indicator
of Environmental Stress
Monitoring fish reproductive systems can serve to alert humans
to potential harm
Stephen A. Eortone and William P. Davis
Mike Ho well and Ann Black
examined each diminutive
fish in the minnow seine
haul they had made in the tea-col-
ored waters of a small coastal stream,
Elevcnmile Creek, in the western-
most, panhandle area of Florida in
1 978. All the mosquitofish (Gambu-
sia af/'inis) appeared to be males. At
first, the researchers, then both at
Sntnford University in Birmingham,
Alabama, were not too surprised
because they were aware that among
mosquitofish populations there oc-
curs some microhabitat segregation
by sex (Martin 1975). Moreover,
they knew differential mortality
could also lead to unequal sex dis-
tributions (Snelson 1989). On closer
inspection, however, the scientists
noted an astonishing oddity.
Many of the apparently male fish
(as distinguished by the presence of
a gonopodium, an elongation of the
anterior anal-fin rays that serves as
a copulatory organ to inseminate
females) bore what appeared to be a
gravid spot, which in these live-bear-
ing fish normally indicates females
with internally developing young
(Figure 1). Later, a laboratory analy-
sis confirmed that many of the fish
thought to be male were actually
masculinized females (Howell et al.
1980). These masculinized female
Stephen A. Bortone is a professor in the
Department of Ecology and Evolution-
ary Biology at the University of West
Florida, Pcnsacola, FL 32514. William
P. Davis is a research ecologist at the
Environmental Research Laboratory,
Gulf Breeze, FL 32561.
Intersexuality presents
an opportunity for
biologists to design
testing procedures to
evaluate detection of
endocrine-disrupting
agents
mosquitofish showed, in addition
to a gonopodium, more subtle male
secondary sex characters, including
behavioral characteristics such as
pursuit and gonopodial swinging.
Speculation on the potential
causes of this masculinization ranged
from genetic phenomena to envi-
ronmental induction. The environ-
mental hypothesis gained strength a
few months later when a second
population of masculinized fish was
discovered in the Fenholloway River
(Bortone and Drysdale 1981), an-
other coastal Florida stream, ap-
proximately 300 kilometers east of
Elevenmile Creek. The second group
included masculinized females from
two other live-bearing species in the
family Poeciliidae (the least killi-
fish, Heterandria formosa [Figure
2], and the sailfin molly, Poecilia
latipinna) as well as masculinized
female mosquitofish.
In both streams, the masculinized
females occurred downstream of
paper mills discharging kraft-mill
effluent (KME), but they were not
upstream of the paper mill discharge
nor in any tributaries to the efflu-
ent-receiving streams. We observed
in least killifish a higher degree of
masculinization among females cap-
tured closest to a KME discharge
point, and we saw proportionately
reduced mascuiinization in samples
collected downstream from the point
of KME discharge (Figure 3). The
phenomenon was not observed in
nonpoecilid fishes inhabiting the
streams, however.
Fish have long been known to
display a wide range of variation in
the expression of their sexuality (Atz
1964). Moreover, investigators spe-
cializing in the reproductive biology
of fishes have noted a morphologi-
cal plasticity of their expressed sex
in response to changes in environ-
mental factors (Reinboth 1980).
These factors include temperature,
pH, light intensity, and social con-
ditions, anong others. Careful ex-
amination of sexuality in fishes has
the potential to allow us to measure
the nature and extent of some types
of environmental stress (Davis and
Bortone 1992).
In this article, we examine how
human modification of the environ-
ment through waste discharge has
altered the sexual condition of fishes.
We also examine the degree and
nature of fish intersexuality relative
to environmental stress. In addition,
we show how these sexual alter-
ations can serve as sentinels to alert
resource managers to potential prob-
lems. Last, we indicate how this
phenomenon can be used to direct
future research.
March 1994
165
-------
l-igurc 1. Normal female (top) and male
mnsquitofish (bottom). A masculinized
tvnv.ilc is pictured in the center, sho%v-
ing the gonopodial elongation of the
anal fin.
A few words about sex
Fish species arc extremely broad in
the expression of their reproductive
modes. These modes can vary be-
tween phylogenctic lineages and
among members of the same family.
The tcrmintersexuality encompasses
a broad range of sexual expression
in which both male and female char-
acters, cither primary or secondary,
are found in a single individual at
some time during its life. Hertnaph-
roditisnt describes a subcategory of
intcrsexuality where the primary re-
productive characters (testes and
ovaries) are found in a single indi-
vidual. The male and female charac-
ters can occur at the same time (as in
the small sea basses; e.g., Fischer
and Percrsen 1987) or at different
ages (i.e., sex reversal). Individuals
can initially be male (described as
protandry, as in the anemone fishes;
e.g., Fricke and Fricke 1977) or fe-
male (termed protogyny. as in the
groupers; e.g., Moe 1969) . Mascu-
linization, called arrhenoidy, is a
subcategory of intersexuality where
the presence of male secondary sex
characters is observable in females.
It should not, however, be concep-
tually confused with hermaphrodit-
ism, sex reversal, or "adaptive am-
biscxuality" (Reinboth 1988), where
sex change is part of a species' nor-
mal, reproductive mode.
Induced alterations of sexuality
in fishes have been well documented
(see review by Reinboth 1980). For
example, salmon are known to at-
tain sexual maturity earlier in their
development when treated with go-
nadotropins (Funk and Donaldson
1972). In laboratory experiments
pregnant-mare serum, chorionic
gonadotropin, X-rays, and incom-
plete hypophysectomy have all been
used to produce arrhenoidy in fe-
male swordtails. Additionally, more
naturally occurring phenomena such
as old age or parasitism have also
been associated with masculiniza-
tion among fishes, including gup-
pies (Atz 1964).
The complex nature of paper
mill effluent
The preparation of wood pulp for
paper and cellulose manufacture
separates cellulose fibers and ligntHfl
from the sugars, saps, and other
components (including animal in-
habitants) of tree stems. Different
paper mills use variations of the
pulping process depending on the
materials being processed (e.g., pine
or hardwoods) and the end product
(e.g., bulk cellulose, brown kraft
paper, and white paper).
Basic to all the pulping processes
is separation and discharge of sug-
ars, lipids, resins, and fatty acids
that are the digestion by-products
of kraft-mill operations. Typically,
these waste products are entrained
in heated waters that are directed to
settling and aeration ponds. There
they receive bacteriological treat-
ment analogous to sewage treatment
processes. Plants, especially pine
trees, are potentially rich sources of
phytosterols (Conner et al. 1976).
However, owing to the structure of
the complex resins and sugars, there
is an inherent resistance to a rapid
breakdown into humic and fulvic
acids. The treatment processes also
appear not to destroy phytosterols,
some of which masculinize poeciliid
fishes in the laboratory.
The complexity of paper-mill ef-
fluent presents a considerable chal-
lenge to specific identification of
androgenic factors. On one hand,
there is the presence of various phy-
tosterols, which have been shown to
induce masculinization under con-
trolled conditions (Denton et al.
1985). Phytosterol action may be
influenced by a variety of environ-
mfntal conditions, including the sea-
sons, tree species being pulped, de-
gree of effluent treatment, and
effluent dilution. We have chosen to
assess phytosterol activity.
On the other hand, the chemical
processes within a mill (e.g., chlo-
rine, chlorine dioxide, or oxygen
bleaching) may add to the complex-
ity of the effluent and thus may
influence the production of dioxin,
which is known to be an endocrine
disrupter. However, the literature
has attributed feminization, rather
than masculinization, to dioxin
(Peterson et al. 1992). For this rea-
son, and because dioxin exposure
presents serious hazards, we have
chosen not to assess dioxin effects.
Quantification of pulp-mill ef-
fluent is typically done through com-
parative effects of whole and dilut-
ed effluent samples. Because the
.masculinization response in test fish-
es commences in microbially degrad-
ed phytosterols after 15-20 days of
exposure, experiments using efflu-
ent samples are cumbersome and
highly variable. To carefully assess
the androgenic effects of paper-mill
effluents, it is preferable to have a
direct connection to the effluent with
continuous, flow-through exposure.
Selectively filtered samples would
allow fractionation analyses so that
identification of components can be
determined by their relative activi-
ty. No research support that we
know of, to date, has been made
available to conduct such experi-
ments.
Intersexuality and
paper-mill effluent
Fishes and other organisms occur-
ring in aquatic habitats receiving
166
BiaScience Vol. 44 No. 3
-------
Figure 2. An array of anal fins from the least killifish. a. Normal female, b. Fish exposed to kraft-mill effluent; note the
incursvcl segmentation at the tips of the rays. c. Another example of a fish exposed to kraft-mill effluent with some
elongation of the tin rays. d. Normal male gonopodium resembles this highly masculinized female fin. '
pulp and paper-mill effluents ex-
hibit a variety of biochemical, physi-
ological, metabolic, and behavioral
responses (Mel.cay et al. 1987,
Owens 1991). Recent reports indi-
cate that paper-mill effluents in natu-
ral waters can produce modifica-
tions in the life-history features of
fislu-s including adverse schooling
behavior in whitefish (Myllyvirta
and Vuorinen 1989), changes in
maturity and serum steroid levels in
while suckers and lake whitefish
(Munkittrick et al. 1992), and in-
creased growth, biochemical
changes, and higher levels of stress
in juvenile coho salmon (McLcay
1979). Juvenile American eels
(A»}*nilla rnstrata) from Elevcnmile
Creek exhibited precocious male
secondary sex characters (enlarged
eyes and precocious testicular de-
velopment; Caruso ct al. 1988).
KMK-induccd male secondary sex
characters in female mosquitofish
occur under both field and labora-
tory conditions (Bortone ct al. 1989,
Davis 1989, Drysdale 1984, Drys-
dale and Bortone 1989, Rosa-Mol-
inar and Williams 1984). We ex-
posed female mosquitofish to stream
water that had received naturally
degraded KME for three weeks both
in the field and in the laboratory.
Several morphological characters
were affected including anal-fin
length, pre-anal length, dorsal-fin
height, pelvic-fin height, intcrorbital
width, eye diameter, and body depth.
Each character changed to a more
masci'line state. These modified
characters were statistically signifi-
cant indicators of KME exposure
(Bortone et al. 1989). After three
weeks of exposure, the morphologi-
cal characters of KME-treated adult
females were intermediate to the
normal male and female conditions.
Masculinization in female mos-
quitofish and other poeciliids is most
readily observed in the increased
segmentation and elongation of the
third, fourth, and fifth anal rays
(Howcll and Denton 1989, Howcll
et al. 1980). These changes can be
observed in fish exposed to such
steroids as methyltestosterone
(Turner 1960), androstenedione,
androstanol, and spironolactone
(Hunsinger and Howell 1991).
A hypothesis has been offered that
bacterial processes associated with
KME may be responsible for the
arrhenoid condition among natural
populations of mosquitofish. Den-
ton et al. (1985) and Howell and
Denton (1989) exposed mosquito-
fish to phytosterols (i.e., sitosterol
and stigmastanol) mixed with a bac-
terium, Mycobacteriitm smegmatis.
The masculinization response they
observed (principally the elongation
March
167
-------
of the anal fin) was nearly identical
to that we observed in a similar
laboratory experiment exposing the
le.'sr killifish to KME that had also
been microbially degraded.
Phytosterols in tall (pine) oil can
be microbially converted to C-19
sterols (Conner et at. 1976). These
sterols include steroid compounds
known to influence sex and repro-
duction. During 1974, for example,
the more than 800,000 tons of tall
oil produced in the United States
could have yielded more than 20,000
tons of phytosterols (Conner et al.
1976). The amount discharged to
the environment cannot be calcu-
lated, but even a small fraction could
significantly harm fishes and other
aquatic organisms.
Our studies indicate that repro-
ductive systems in fishes are useful
bioindicators for detection of sub-
stances that interfere with endocrine
modulated processes. More specifi-
cally, live-bearing poeciliid fishes
such as mosquitofish and least killi-
fish may be useful to detect the pres-
ence of endocrine disrupters.
Poeciliid masculinization by KME
may be due to a few, many, or spe-
cific combinations of endocrine-dis-
rupter compounds that mimic or
trigger an organism's receptors to
its own steroidal androgens. It is
likely that the masculinization re-
sponse varies with exposure time,
concentration, degree of microbial
activity, fish species, and water con-
ditions (e.g., temperature, pH, and
conductivity).
Bioindicators for the duration
and intensity of KME exposure
The field and laboratory studies
conducted thus far represent only
short-term observations of the mas-
culinization response to KME; the
results may not be indicative of what
one might observe under long-term,
continuous exposure. Rosa-Molinar
and Williams (1984) noted the po-
tential for diminished fecundity
among arrhenoid mosquitofish that
had been captured from streams re-
ceiving KME. However, most previ-
ous studies have been concerned with
the morphological features of KME-
exposed individuals and have not
considered life-history traits that
could ultimately affect an indivi-
7 -
6 -
01
c
o
_J
I 3
10 12 U 16 18 20
Stondord Length (mm)
22
Figure 3. Scatter diagram of anal-fin
length versus standard length from the
least killifish, indicating the degree of
anal-fin modification relative to the
point of kraft-mill effluent discharge:
fish collected immediately below the
discharge point of kraft-mill effluent
(dark triangles), fish from the farthest
distance downstream (open triangles),
and fish from mid-distance along the
stream (asterisks).
dual's fitness.
Fish life-history traits, including
the reproductive characteristics of
poeciliid fishes, can vary relative to
environmental conditions (e.g.,
Trexler 1988). Therefore, continu-
ous exposure to KME might affect
not only the sexual characteristics
of female fishes but also their
offspring's fitness and development.
Long-term exposure may also have
an impact on other population pa-
rameters such as mortality, growth,
reproductive rates, and abundance.
We strongly suspect that females
masculinized by long-term (perhaps
eight months to a year) exposure to
water receiving KME suffer from
impaired reproductive function.
Studies that have investigated the
response of fish to continuous KME
exposure have not reported any overt
androgenic effects in other species.
However, Munkittrick et al. (1992)
noted testicular atrophy in males
and abnormal oocytes in the ova-
rian tissue of females among KME-
exposed whitefish. We believe the
long-term effects of KME exposure
in fishes will be found to include
reduced embryo viability, develop-
mental modifications, and neuter-
ing of female reproductive function.
The effects may include responses to
other potential KME components
(e.g., dioxins, furans, and chlori-
nated lignins).
Hermaphroditism among KME-
masculinized females appears to he
rare (Bortone and Drysdalc 1981).
However, we may be observing in-
terscxuality that could evenrunlly
lead toward facultative hermnphro-
ditism. If fishes exposed to KME
have reduced fitness and shortened
life spans, they may not live long
enough to achieve full hermaphro-
ditism. Notably, the observation that
highly masculinized females (as wit-
nessed by the extreme elongation of
the anal fin rays) are rare in our field
collections might indicate higher
mortality levels among modified
fish.
At present, we have no way of
measuring the exact amount of KME
oritscomponents presentin a stream
at any given site or time. We have
noted, however, that the degree of
masculinization varies with differ-
ent conditions. Interestingly, large-
size mosquitofish and least killifish
were generally absent where the most
masculined female fish were found—•
from streams having the highest rela-
tive amount of KME and from sites
proximate to the KME discharge
point. However, in streams with rela-
tively low concentrations of KME,
where poeciliid fish displayed less
masculinization, the fish were larger.
During prolonged drought (three
months or more), both mosquitofish
and least killifish were smaller and
less abundant in streams receiving
KME than in other streams. The fe-
males exposed to KME during
drought displayed more prominent
anal-fin elongation compared with
females exposed under other condi-
tions. Drought conditions may con-
centrate the masculinization factor
present in KME, or the process gen-
erating the masculinizing factor may
be more effective under these condi-
tions.
The degree of masculinization
among female fish (when compared
with normal male behavioral traits)
decreased when fish were captured
from the field and placed in aquaria
free of KME. Furthermore, gono-
podial and other morphological fea-
tures showed no further develop-
ment. Our unpublished observations
and those of Hunsinger et al. (1988)
confirm that KME-masculinized
168
BioScience Vol. 44 No. 3
-------
mosquitofish and least killifish can
produce viable offspring after hav-
ing been pluced in aquaria free of
KME. Although Larkin (1986) ob-
served male reproductive behavior
in female mosquitofish continuing
as long as they were exposed to
water containing microbially de-
graded phytosterols, we found that
male reproductive behavior was not
detectable amongKME-masculinized
female mosquitofish after they had
been acclimated for two months to
water lacking KME (Bortone et al.
1939).
Assessment of KME effluents re-
quires long-term strategies to effec-
tively define and evaluate effects on
reproductive cycles and life-history
alterations. To date, such in-depth
assessments and evaluations have
not been conducted. Therefore, we
are not yet able to definitely state
whether the observed masculiniza-
tion is due to androgenic or, alter-
natively, antestrogenic stimuli.
(Antestrogeris are compounds that
inhibit estrogenic processes and ova-
rian function.) We have described
the effects as an androgenic stimu-
lus because we observed production
of embryos in females after removal
from laboratory exposure to the
microbially transformed phytoster-
ols and because male juvenile Ameri-
can eels demonstrate accelerated tes-
tes development (Caruso etal. 1988)
at a time in their life history when
sex determination is not normally
possible (Helfman et al. 1987).
Also, the hypothesis that mascu-
linization is due to the presence of
an androgenic stimulus in paper-
mill effluent is supported even
though there are often dioxins
present in the effluent. Dioxins are
antestrogens. Laboratory simula-
tions using plant phytosterols in the
absence of dioxins induced mascu-
linized morphological and behav-
ioral effects comparable to those
observed in the poeciliid species ex-
posed to KME in field studies.
These poeciliid fishes are dedi-
cated invertebrate foragers. Piscivo-
rous fish species feeding at higher
trophic levels may be more vulner-
able to uptake and accumulation of
halorganics from their prey species.
This masculinization may reflect the
waterborne mode of exposun; con-
trasted with the acquisition of a
c, 3
o>
:§ 2
o
< t
' molt control
female control
I I I I I I I I I I I I I I I
30 60 90 120 150
Age In Days
Figure 4. Results of continuous expo-
sure of female mosquitofish to kraft-
mill effluent (beginning one day after
birth). The male and female controls
were fish not exposed to kraft-mill ef-
fluent. (Redrawn from Drysdale and
Bortone 1989.)
halorganic chemical body burden
through the food web.
Degrees of masculinization as
a bioindicator
Until other methods of der?'. .ion
and measurement other thai; the
fish's morphological and behavioral
response per se are developed, one is
faced with an experimental tautol-
ogy: heavily exposed fishes may be
so severely modified that the female
fish appears as a normal male and,
therefore, undetectable by gross in-
spection in the field. Moreover, fe-
male fish may proceed through a
complete series of intersexual steps
in response to continued androgen
exposure. This progression may be
significant if exposure occurs dur-
ing ontogenetic development. It
could eventually lead to hermaph-
roditism. Clearly, both laboratory
and controlled field exposures are
needed, the former focusing on KME
component extraction and the latter
on multigeneration and long-term
exposure effects.
There is additional experimental
evidence indicating that different
androgenic compounds produce dif-
fering degrees of morphological
modification among fish species
(Asahina et al. 1989), further con-
founding attempts to predict a fish's
response to the complex chemistry
of KME exposure. Males of the vari-
ous poeciliid species mature at dif-
ferent sizes and age as a genetic or
adaptive trait (Travis et al. 1989,
Trexler and Travis 1990), which
further underscores the need for a
carefully conducted series of obser-
vations on fishes exposed to KME.
Although the mechanisms for sex
determination among fishes are di-
verse (Angus 1989), sex expression
may be adaptive. Genetically, sex
determination has been described as
polygenic in many fishes (Kosswig
1964). If the polygenic sex determi-
nation hypothesis is correct, there
could exist multiple, genetically
based and variable, responses to the
exposure to endocrine disrupters.
Regardless of whether or not the
endocrine disrupters are acting
through a genetic or environmental
operand, we are alarmed at the po-
tential problems they may cause to
aquatic and other organisms.
To date, masculinization of
poeciliid fish has been considered a
scientific oddity. Concomitantly, its
importance as a bioindicator has
been unappreciated. Few studies
have elucidated the potentially dis-
ruptive effects of hormones or hor-
monelike substances on natural
populations in the field. Gibbs et al.
(1991) found that populations of
the American oyster drill (Urosal-
pinx cinerea) declined when the fe-
males apparently become masculin-
ized (i.e.Jmposex) when exposed to
tributyltin in nature.
After the initial observations of
arrhenoidy in mosquitofish, we con-
ducted different bioassays to deter-
mine if an individual had been ex-
posed to KME in the coastal stream
environment. The bioassays made
use of the morphological and be-
havioral responses of females ex-
posed to paper-mill effluent (Bortone
et al. 1989). Drysdale and Bortone
(1989) conducted a series of experi-
ments that indicated that the induc-
tion of masculinization can occur
early in the life history of these fishes
(Figure 4). In another study, we
noted the relationship between the
degree of masculinization among
female mosquitofish to the proxim-
ity of the paper-mill discharge point
(Bortone and Drysdale 1981).
In the Fenholloway River, poe-
ciliid fishes are virtually the only
species present during times of
March 1994
169
-------
drought, when KME is presumably
more concentrated (judging by the
dark-colored water). In a biotic re-
gion where high aquatic biodiver-
sity is the norm, this stream is dis-
tinctly depauperate of fish species.
Tributary streams, peripheral pools,
and entering spring runs are rich in
fish species absent from the con-
taminated portions. Additionally,
the female poeciliids in these adja-
cent habitats show no sign of mas-
culinization.
Poeciliids have the potential to
serve as indicators of water quality
to determine if and how fishes and
other aquatic organisms (especially
other vertebrates) are affected by
effluent containing endocrine dis-
rupters. Fishes in the family Poe-
ciliidae are naturally distributed in
the lowland and coastal areas of the
temperate areas of the New World.
Most species are tolerant of a wide
range of environmental conditions
and, therefore, are found in a broad
range of habitats.
The mosquitofish is naturally dis-
tributed in a variety of habitats as
well (Krumhplz 1948), either as its
eastern Atlantic slope form, Gani-
busia affinis holbrooki, or its Gulf
of Mexico and Mississippi embay-
ment (orm,Gambusia affinis affinis.
Moreover, it has been introduced
throughout the world as an ill-con-
ceived aid for mosquito control. Its
abundance, small size, broad distri-
bution, and tolerance of a range of
environmental conditions indicate
that the mosquitofish is an excellent
candidate to serve as a natural sen-
tinel for detecting environmental
stress. It may be useful at detecting
stress caused by the addition of en-
docrine disrupters to the aquatic
environment. Moreover, the fish can
be especially useful when stress
causes them to react through ob-
servable changes in reproduction,
development, and other life-history
features.
In the future, it is likely that long-
term monitoring of secondary sex
characters and life-history traits in
poeciliid and other fishes will serve
to detect the presence of endocrine
disrupters. The advantage of such a
detection system is that it recog-
nizes the ecological impact these
compounds may cause. Similarly,
the life-history sentinel may prove
to be a meaningful indicator of stress
owing to disruption of endocrine
modulations in the life cycle of many
organisms.
Application of the model to
other research areas
Endocrine disrupter has emerged as
a generalized term to denote factors
responsible for inducing in the criti-
cal timing of events in ontogenetic
developmental sequences (Colborn
et al. 1993). These alterations can
affect morphology, physiology, and
life-history traits. This concept has
been brought together in the presen-
tations edited by Colborn and Clem-
ent (1992) from a 1991 workshop
(called "Chemically Induced Alter-
ations in Sexual Development: The
Wildlife/Human Connection") held
in Racine, Wisconsin. Participants
in the conference examined various
examples of developmental alter-
ations induced by environmental
exposure to synthetic compounds.
The focus of the workshop was on
compounds that, after assimilation
by parents,caused transgenerational
exposure to offspring. The second-
generation exposures were often at
dose concentrations much higher
than those typically occurring in the
environment and produced various
induced morphological or neurologi-
cal modifications, reduced immune
function, and altered reproductive
function and behavior among all
classes of vertebrates.
As more field and laboratory stud-
ies are reported, it is becoming in-
creasingly clear that effects of endo-
crine disrupters are more common
than one might at first suppose.
Halorganic compounds and plasti-
cisers are among 50 or more com-
pounds that can induce similar re-
sponses. Their collective biological
significance has been neglected and,
thus far, notably absent from risk
assessment analysis.
Fortunately, a considerable body
of information has recently become
available on the life history and ba-
sic biology of poeciliid fishes, espe-
cially mosquitofish (Larkin 1986,
Meffe and Snelson 1989, Snelson
1989). The potential thus now ex-
ists for these fishes to serve as
bioindicators for some types of en-
docrine disrupters. These baseline
data can serve to help assess the
impact of environmental stress on
the varied life-history features of
these coastal live-bearing fishes.
Through the efforts of Farr
(1989), Travis ct al. (1987), and
Henrich (1988), information exists
on the effectiveness of several ex-
perimental designs to measure the
impact of environmental factors on
the fitness of these poeciliid fishes.
Moreover, there are several studies
that have established the normal
behavior of these fishes and make
effective use of reproductive behav-
ior as a way of assessing the impact
of environmental stress on them
(Itzkowitz 1971, Martin 1975,
Schroder and Peters 1988). These
studies on behavioral response, when
coupled with the well-documented
morphological response of live-bear-
ing fishes to environmental stress
(Bortone et al. 1989, Howell et al.
1980, Okada and Yamashita 1944,
Riehl 1991, and Turner 1960), can
provide reliable endpoint responses
to potential endocrine disrupters.
Environmental masculinization of
live-bearing fishes (including mos-
quitofish, least killifish, and sailfin
mollies) may represent a unique form
of intersexuality. Typically, andro-
gens are aromatized in vertebrates,
resulting in the feminization of males
or antestrogen-induced female dys-
function. Wester et al. (1985) re-
ported aromatization in the poeciliid
guppy.
The masculinization of poeciliid
females by androgens in KME repre-
sents only one example of a potent
environmental sentinel among wild-
life species. Concomitantly, it rep-
resents an opportunity to design test-
ing procedures to evaluate the
general application of the masculin-
ization bioassay to the detection of
endocrine-disrupting agents.
There is currently strong advo-
cacy for initiating additional tests
to detect potential endocrine dis-
rupters in the procedures used to
evaluate pharmaceuticals and regu-
late chemical discharge. Endocrine
disruption of life history has a
broader significance than traditional
concepts of birth defects or terato-
genesis. With the potential of trans-
generational exposure and delayed
responses, as in the impairment of
immune systems, the effects of these
170
BioScience Vol. 44 No. 3
-------
compounds are particularly sinis-
ter.
Synthetically produced chlorinat-
ed and brominated organic com-
pounds often demonstrate high
levels of activity as endocrine dis-
rupters. Perhaps the ability of some
compounds to cause a high level of
endocrine disruption was a factor
contributing to their original empir-
ical selection as effective pesticides.
Toxicological testing has progressed
from a focus on acute and chronic
lethal effects to cancer induction on
test species. Literature references of
effects on offspring or adult life his-
tory and fitness have not typically
been included in ecological risk-as-
sessment procedures. However, re-
productive toxicology is rapidly
growing and moving from the labo-
ratory toward field assessments that
include life-history impact.
Developing a well-documented
and well-referenced system to rec-
ognize the subtle changes in inter-
sexuality in fishes represents a step
in the right direction to better moni-
tor the health of aquatic ecosys-
tems. It may prevent unpleasant sur-
prises, such as the recent reports of
a global decline in reproductive suc-
cess among amphibians (Phillips
1990, Vitt et al. 1990). Fish inter-
sexuality, as a sentinel, may prove a
reliable bioindicator to some forms
of environmental stress.
Acknowledgments
We are grateful to the many indi-
viduals who aided us with their sage
advice and counsel on the matters of
intersexuality and environmental
stress: Alani Davis, Tom Denton,
Dale Drysdale, Leroy Folmar,
Roxanna Hinzman, Mike Howell,
and Tibor Kovacs. We also thank
John Blackie, Shelby Curry, Jenni-
fer Kensler, and Dan McLeod for
help in preparing the manuscript
and figures.
References cited
Angus, R. A. 1989. A genetic overview of
poeciliid fishes. Pages 51-68 in G. K.
Meffe and F. F. Snelson, eds. Ecology and
Evolution of Livebearing Fishes. Prentice
Hall. Englewood Cliffs, NJ.
Asahina, K., A. Urabe, T. Sakai, H. Hirose,
and T. Hibiya. 1939. Effects of various
androgens on the formation of papillary
processes on the anal fin rays in the fe-
male medaka Oryzias latipet. Nippon
Suisan Gakkaishi 55: 1871.
Atz, J. W. 1964. Intersexuality in fishes.
Pages 145-232 in C. N. Armstrong, and
A. J. Marshall, eds. Intersexuality in Ver-
tebrates Including Man. Academic Press,
New York.
Bortone, S. A., W. P. Davis, and C. M.
Bundrick. 1989. Morphological and be-
havioral characters in mosquitofish as
potential bioindication of exposure to
kraft mill effluent. Bull. Environ. Contam.
Toxicol. 43: 370-377.
Bortone, S. A., and D. T. Drysdale. 1981.
Additional evidence for environmentally
induced intersexuality in poeciliid fishes.
Association of Southeastern Biologists
Bulletin 28: 67.
Caruso, J. H., R. D. Suttkus, and G. E.
Gunning. 1988. Abnormal expression of
secondary sex characteristics in a popu-
lation of Anguilla rostrata (Pisces:
Anguillidae) from a dark colored Florida
Stream. Copeia 1988: 1077-1079.
Colborn, T., and C. Clement. 1992. Chemi-
cally-Induced Alterations in Sexual and
Functional Development; The Wildlife/
Human Connection. Princeton Scientific,
Princeton, NJ.
Colborn, T., F. S. vom Saal, and A. M. Soto.
1993. Developmental effects of endocrine-
disrupting chemicals in wildlife and hu-
mans. Environ. Health Persp. 101:
378-384.
Conner, A. H., M. Nagaoka, J. W. Rowe,
and D. Perlman. 1976. Microbial conver-
sion of tall oil sterols to C-19 steroids.
Appl. Environ. Microbiol. 32: 310-311.
Davis, W. P. 1989. Review of the androgenic
effects on poeciliid and other fishes living
in streams receiving kraft mil) effluents.
EPA/600/X-89/184, Gulf Breeze, FL.
Davis, W. P., and S. A. Bortone. 1992. Ef-
fects of kraft mill effluent on the sexual-
ity of fishes: an environmental early warn-
ing? Pages 113-127 in T. Colborn and C.
Clement, eds. Advances in Modern Envi-
ronmental Toxicology: The Human/Wild-
life Connection. Princeton Scientific, Prin-
ceton, NJ.
Denton, T. E., W. M. Howell, J. J. Allison, J.
McCollum, and B. Marks. 1985. Mascu-
linization of female mosquitofish by ex-
posure to plant sterols and Mycohacte-
rium smegmatis. Bull. Environ. Contam.
Toxicol. 35: 627-632.
Drysdale, D.T. 1984. Intersexuality in the
mosquitofish, Gambusia affinis. Masters
dissertation. University of West Florida,
Pensacola.
Drysdale, D. T., and S. A. Bortone. 1989.
Laboratory induction of inter-sexuality
in the mosquitofish, Gambusia affinis,
using paper mill effluent. Bull. Environ.
Contam. Toxicol. 43: 611-617.
Farr, J. A. 1989. Sexual selection and sec-
ondary sexual differentiation in poeciliid
fishes: determinants of male mating suc-
cess and the evolution of female choice.
Pages 91-123 in G. K. Meffe and F. F.
Snelson, eds. Ecology and Evolution of
Livebearing Fishes. Prentice Hall,
Englewood Cliffs, NJ.
Fischer, E. A., and C. W. Petersen. 1987. The
evolution of sexual patterns in the
seabfisses. BioScience 37: 482-489.
Fricke, H., and S. Fricke. 1977. Monogamy
and sex change by aggressive dominance
in coral reef fish. Nature 7.66: 830-832.
Funk, J. D., and E. M. Donaldson. 1972.
Induction of precocious sexual maturity
in male pink salmon (Oncorhynchut
gorbuscha). Can.). Zoo/. 50:1413-1419.
Gibbs, P. E., B. E. Spencer, and P. L. Pascoe.
1991. The American oyster drill,
Urosalpinx cinerea (Gastropoda): evi-
dence of decline in an imposex-affected
population (R. Blackwater, Essex)../. Mar.
Biol. Assoc. UK 71: 82.7-838.
Hclfman, G. S., D. E. Facey, L. S. Hales Jr.,
and E. L. Bozeman. 1987. Reproductive
ecology of the American eel. American
Fisheries Society Symposium 1: 42-56.
Henrich.S. 1988. Variation in offspring sizes
of the poeciliid fish Heterandria formosa
in relation to fitness. Oikos 51: 13-18.
Howelt, W. M., D. A. Black, and S. A.
Bortone. 1980. Abnormal expression of
secondary sex characters in a population
of mosquitofish, Gambusia affinis
holbrooki: evidence for environmentally
induced masculinization. Copeia 1980:
676-681.
Howell, W. M., and T. E. Denton. 1989.
Gonopodial morphogenesis in female
mosquitofish, Gambusia affinis affinis
masculinized by exposure to degradation
products from plant sterols. Environ. Biol.
Fish. 24:43-51.
Hunsinger, R. N., B. R. Byram, and W. M.
Howell. 1988. Unchanged gonndal mor-
phology of mosquitofish masculinized by
exposure to degraded plant sterols. ].
Fish. Biol. 32: 795-796.
Hunsinger, R. N., and W. M. Howell. 1991.
Treatment of fish with hormones: solubir
lization and direct administration of ste-
roids into aquaria water using acetone as
a carrier solvent. Bull. Environ. Contam.
Toxicol. 47: 272-277.
Itzkowitz, M. 1971. Preliminary study of the
social behavior of male Gambusia affinis
(Baird and Girard) (Pisces: Pocciliidac) in
aquaria. Chesapeake Sci. 12: 219-224.
Kosswig, C. 1964. Polygenic sex determina-
tion. Experientia 20: 190-199.
Krumholz, L. A. 1948. Repro'i'.ction in the
western mosquitofish Gambusia affinis
affinis (Baird and Girard) and its use in
mosquito control. Ecol. Monogr. -18:
1-43.
Larkin, M. J. 1986. The effects of induced
masculinization on reproductive and ag-
gressive behaviors of the female
mosquitofish, Gambusia affinis affinis:
evidence for environmentally induced
masculinization. Masters dissertation,
Samford University, Birmingham, AL.
Martin, R. G. 1975. Sexual and aggressive
behavior, density and social structure in
a natural population of mosquitofish,
Gambufia affinis holbrooki. Copeia 1975:
445-454.
McLeay, D. J. 1979. Stress and chronic ef-
fects of untreated and treated bleached
kraft mill effluent on the biochemistry
and stamina of juvenile coho salmon
(Oncorhynchtu kistitch).}. Fish Ret. Bd.
Can. 36: 1049-1059.
McLeay, D. J., and Associates Ltd. 1987.
Aquatic toxicology of pulp and paper
mill effluent: a review. EPS Report 4/PF/
March 1994
171
-------
1. Environmental Canada, Ottawa, ON,
Canada.
Meffe, G. K., and F. F. Snelson. 1989. An
ecological overview of poeciliid fishes.
Pages 13-32 in G. K. Meffe and F. F.
Snelson, eds. Ecology and Evolution of
Livebearing Fishes. Prentice Hall,
Englewood Cliffs, NJ.
Moe, M. A. Jr. 1969. Biology of the red
grouper Epinephelus morio Valenciennes
from the eastern Gulf of Mexico. Profes-
sional Papers Series No. 10, Florida De-
partment of Natural Resources, St. Pe-
tersburg.
Munkittrick, K. R., M. E. McMaster, C. B.
Pom, G. J. Van Der Kraak, I. R. Smith,
and D. G. Dixon. 1992. Changes in matu-
rity, plasma sex steroid levels, hepatic
mixed-function oxygenase activity, and
the presence of external lesions in lake
whitefish exposed to bleached kraft mill
effluent. Can. J. Fish. Aquat. Sci. 49:
1S60-1569.
Myllyvirta, T., and P. J. Vuorinen. 1989.
Effects of bleached kraft mill effluent
(BKME) on the schooling behavior of
vendace (Coregonus alhula L.) Bull.
Environ. Contam. Toxicol. 42: 262-269.
Okada, Y. K., and H. Yamashita. 1944.
Experimental investigation of the sexual
characters of poeciliid fishes. Japanese
Faculty of Science Imperial University of
Tokyo Section IV Zoology 6: 589-633.
Owens, J. W, 1991. The hazard assessment
of pulp and paper effluents in the aquatic
environment: a.review..Environ. Toxicol.
cnWonmcnu a rvVic\. i:.»rir
------- |