EPIDEMICS IN BIRDS AND
MAMMALS CAUSED. BY
CHEMICALS IN THE GREAT
Fish Habitat Management Branch, Department of Fisheries and Oceans,
Ottawa, Ontario, Canada KIA OE6*
1. Epidemics and causes
2. Reproductive failure in ranch mink
2.1. Initial report
2.2. Symptoms and hypotheses
3. Reproductive failure and mortality in birds
3.1. Initial reports
3.1.1. Lake Michigan and Lake Huron.
3.1.2. Lake Ontario
3.2. Upper lakes
3.2.1. Symptoms and hypotheses
3.2.4. New symptoms and hypotheses
3.2.5. Etiology of adult mortality
*Present address: Commercial Chemicals Branch, Environment Canada, Ottawa, Ontario. Canada KIA
I~- I .,(i
134 Epidemics in Birds and Mammals Caused by Chemicals in the Great Lakes
3.3. Lake Ontario
3.3.2. Signs and symptoms
1. EPIDEMICS AND CAUSES
During the past twenty years there has been a series of anomalies involving repro-
ductive failure and mortality in populations of mammals and birds in the Great
Lakes basin. The common factor between these observed anomalies has been that
the organisms affected were primarily fish eaters. Subsequent etiological research
has shown that the toxic agents in the fish were industrial and agricultural chemi-
cals and that these were organochlorine compounds.
This review is mostly concerned with case histories of anomalies that were
noticed in the mid 1960s and early 1970s and it is valid to ask why these old spoors
should be reworked. First, after a period when the levels of organochlorine chem-
icals declined, recent data show that the levels in fish from some Great Lakes are
increasing (International Joint Commission, 1983). Second, researchers from Green
Bay, Wisconsin, have shown that there is a new outbreak of reproductive dys-
function and abnormalities in fish-eating birds. Third, various interest groups, such
as mink (Mustela vision) farmers, need to know whether Great Lakes fish are yet
safe to be used as a feedstuff for their mink-ranching operations. Fourth, these
case histories could show principles of how to structure research on severely con-
taminated populations to obtain credible results. Finally, these case histories could
be important in negotiating cleanup pursuant to the Great Lakes Water Quality
Classically (Table 6.1), epidemiological cases begin with the observation of an
anomaly at some level of biological organization. For instance, the observed
anomaly may be a missing year class, reduced hatchability, a tumor, an enlarged
liver, a deformity, or biochemical or histological lesion. Subsequent to these initial
signals, detailed data on the species affected and geographic and temporal variation
in the incidence of the anomaly yields information on the scale, severity, and
trends of the epidemic. The pathological signs and symptoms are investigated and
described. Finally, the question is posed as to what is causing the anomaly and
epidemiological approaches give way to etiological methods using experimentation
based on the formulation of testable hypotheses of causation. After a causal agent
2. Reproductive Failure In Ranch Mink
6.1 Flow Diagram of Activities for Formal
Demonstration of Causation of a Biological
Anomaly by a Suspected Agent
Observation of a biological anomaly.
Field studies of species affected and of geographic and temporal
incidence and severity of the biological anomaly
Description of pathological signs and symptoms
Formulation of hypotheses of causation
Experimentation with possible causal agent
Experimentation with other causal agents with similar mode of
Conclusions about causation of the biological anomaly
Source: After Oilhelson (1914).
has been demonstrated, alternative causal agents must also be considered. At some
point in the procedure the anomaly may acquire a name.
The International Joint Commission has already sponsored a literature review
of the effects of chemicals on Great Lakes aquatic populations (Fitchko, 1986).
This present review aims at a transformation of the literature review into a series
of selected case histories based on the most complete examples of investigations
to relate biological anomalies observed in the field to a cause or causes.
2. REPRODUCTIVE FAILURE IN RANCH MINK
2.1. Initial Report
Twenty years ago. Dr. G. R. Hartsough (1965), Director of Ranch Service. Great
Lakes Mink Alsociation, published a short, carefully worded news article impli-
cating Great Lakes fish in a large number of reproductive failures in ranch mink
operations in North Central states and in Ontario. The report noted that the evi-
dence was based only on clinical evidence and that it required substantiation. He
also noted the dilemma for the decision maker. On the one hand, it was important
that no mistake be made in attributing the cause of the problem to Great Lakes fish
because removal of this feed source would strain alternative feed sources and in-
crease feed costs. On the other hand, he recognized his obligation to the industry
to prevent a widespread problem. Chemical analyses had shown that pesticide
residues were present in significant quantities in female mink and in fish, but he
II~X:IL~C I --
136 ,pllks In Birds and Mammals Caused by Chemicals in the Great Lakes
also noted that the offending chemical might not be a pesticide but some entirely
different compound. One aspect that is not clear from the article is the exact source
of the fish, because little was recorded about which of the Great Lakes contained
the fish with the toxic factors at that time.
The details of the epidemiology and etiology of the mink reproductive failure
have been published in several review articles (Aulerich and Ringer, 1977: Ringer,
1982). There are, however, few references to the geographic extent of the failures
or to the economic losses incurred. Without these data it is difficult to gauge the
significance of the outbreaks. The disease in the mink was associated with fish
taught in Lakes Michigan and Erie fed to mink during the early 1960s. However,
all Great Lakes fish became suspect and use was rapidly discontinued. The industry
turned to sources of ocean fish. However, no estimate has yet been made of the
socioeconomic impact on the Great Lakes fishery or of the increased costs of these
2.2. Symptoms and Hypotheses
In 1967 mink ranch operations, adult animals were not adversely affected by feed-
ing up to 15% of Lake Michigan coho salmon (Oncorhynchus kisutch) in their diet
(Aulerich et al., 1971). This level of feed even supported normal breeding and
whelping. However, depending on the percentage of coho salmon in the diet and
the duration of feeding, mortality of the kits reached 80%. Based on these studies,
several hypotheses were put forward as possible explanations for the observed
reproductive failure including rancidity, mercury contamination, organochlorine
pesticides and the possibility of other chemical contaminants (Hartsough, 1965;
Aulerich and Ringer, 1970; Aulerich et al., 1971).
In 1968, the first of a series of experiments was undertaken to verify the indus-
try's observations associated with feeding Great Lakes fish to mink and to identify
the etiological agent. The experiments not only confirmed (Table 6.2) the presence
of a toxic agent in Lake Michigan coho salmon but also showed that the agent was
present in other Lake Michigan species and in Lake Erie fish. In addition to sig-
nificant increases in kit mortality and general failure to whelp when fed Lake Mich-
igan fish species, kits weighed less at whelping and after 4 weeks.
Adult mortality occurred in all experimental mink fed by-products from canning
Lake Michigan coho salmon. Symptoms at death were nonspecific but included
anorexia, bloody stools, fatty infiltration and degeneration of the livers and kid-
neys, and hemorrhagic gastric ulcers (Au)erich et al., 1973).
No relationship was found between rancidity or mercury content and any of the
biological parameters of whelping success or kit mortality. Based on the results of
organochlorine pesticide residue analysis of Lake Michigan coho salmon sampled
2. Reproductive Failure In Ranch Mink
Table 6.2 Reproductive Performance of Mink Fed Various Fish
Number of Kits
Fish Potion of Number of Females Birth
Fish Portion of Alive at
Diet Mated Whelped Alive Dead 4 weeks
Ocean Whiting 12 10 45 10 29
(8.2 g) (125 g)
Coho Salmon 10 I 0 I 0
Bloater Chub 12 7 10 12 .5
(7.0 g) (91 g)"
Lake Erie 12 11 32 17 7
Coho Salmon (7.6 g) (86 g)**
oP < 0.05.
**' < 0.01.
Sown: After Aulertih el al. (1971).
in 1969, there was an apparent relationship between the content of organochlorine
pesticides (DDT and related isomers or dieldrin) and reproductive failure (Aulerich
et al., 1971). However, previous work (Duby. 1970; Aulerich and Ringer. 1970)
had shown that mink were able to reproduce nonnally when exposed to an intake
of DDT or dieldrin greater than the calculated DDT or dieldrin intake in these
experiments in which fish were fed to mink. In addition, there was no indication
of clinical signs of organochlorine pesticide toxicity such as tremors, convulsions.
or loss of coordination (Aulerich et al.. 1973). Another reproductive experiment
in which the acetone hexane extract of Lake Michigan coho salmon was fed to
.mink showed that the toxic factor was in the lipids (Aulerich and Ringer, 1977).
The next clue to the possible etiology came from analytical chemistry. In ad-
dition to DDT and dieldrin found in the Lake Michigan coho salmon sampled in
1969, polychlorinated biphenyls were detected at concentrations between 10
and 15 ppm (Aulerich and Ringer, 1977). Samples of three PCB products (Aroclor
1242, 1248, and 1254) were obtained from the manufacturer (Monsanto) and their
toxicity was tested in a subsequent mink-breeding experiment (Aulerich et al..
1973). Mink were fed a basic diet with either (I) ocean fish, (2) Lake Michigan
coho salmon, or (3) ocean fish with 30 ppm PCB (10 ppm of each of the three
above Aroclors). The reproductive success (Table 6.3) and PCB content of the
adults were determined and the gross pathology was determined after death. None
of the mated females fed Lake Michigan coho salmon or 30 ppm PCB whelped,
whereas the controls produced 35 live kits. There was a striking similarity between
the clinical signs and lesions of the mink that died when fed diets containing coho
salmon from Lake Michigan and those fed PCB. The levels of PCB in the tissues
of mink that died after eating Lake Michigan coho salmon were remarkably similar
to those of mink that were fed the PCB (Table 6.3).
138 L,..7 s In Birds and Mammals Caused by Chemicals in the Great Lakes
Table 6.3 Reproductive Performance and Average PCB Residues in
Tissues of Mink Fed Various Dietary Treatments
Control 30% 30% Ocean Fish
30% Ocean Lake Michigan Mix plus
Fish Mix Coho Salmon 30 ppm PCB'
No. females mated 12 12 11
No. females whelped I 0 0
No. kits born
Alive 35 -
Dead 19 -
No. live kits at 4
weeks 29 -
PCB residue in ppm in
Brain 0.1 11.07 1 .00
Liver 0.01 5.21 4.18
Kidney. 0.01 6.37 4.47
"PCB consisted of 10 ppm each of Aroclor 1242. 1248. and 1254.
Source: After Aulerich and Ringer (1977).
Aulerich and Ringer (1977) used five criteria to "strongly implicate" PCB as
the etiological agent responsible for the observed reproductive failure and lesions:
1. The PCB residue content of Lake Michigan coho salmon
2. The sensitivity of mink to PCB
3. The similarity of the reproductive complications in mink fed coho salmon
4. The similarity of the clinical signs and lesions in mink that died after being
fed Lake Michigan coho salmon or PCB
5. Accumulation of comparable quantities of PCB in mink that died after being
fed Lake Michigan coho salmon or PCB.
There is no doubt that this series of observations and experiments was an excellent
piece of epidemiological and etiological sleuthing. It represents one of the very
few cases in which low-level exposure to organic chemicals has been shown to be
the cause of a subtle epidemic in a mrammal population, other than humans. Per-
haps the most important aspect of this work is that the initial signals of the presence
of a toxic factor in Lake Michigan fish were detected 20 years before epidemio-
logical evidence showed effects on human reproduction (Fein et al., 1984: Jacob-
son et al., 1984).
2. Reproductive Failure in Ranch Mink )
How has this causal explanation stood up to time? For the mink industry the
avoidance of fish products from the Great Lakes has resulted in a cessation of
losses attributable to this source of feedstuff. During the past decade research has
demonstrated other organic compounds that have a similar mode of action to PCB
(Goldstein, 1980) including various extremely toxic isomer of the polychlorinated
dibenzo-p-dioxins and polychlorinated dihenzofurans. Their concentrations in tis-
sues may also correlate with the levels of PCB and other organochlorine com-
pounds. Could these compounds have contributed significantly to the mink repro-
ductive failure? Recent analytical data on Great Lakes fish have shown the presence
of these compounds (Stalling et al., 1983). Various polychlorinated dibenzofurans
are contaminants of PCB fluids (Bowes et al., 1975). There are, however, several
other sources of these compounds to the Great Lakes environment. Incinerators
.are sources of both polychlorinated dibenzo-p-dioxins and dibenzofurans (Rappe,
1984). The extremely toxic 2.3,7,8-tetrachlordibenzo-p-dioxin is leaching from
chemical dumps where wastes from production of 2,4,5-trichlorophenol were bur-
The relative toxicities of many of these compounds have been determined using
rapid screening bioassays based on induction of aryl hydrocarbon hydroxylase ac-
tivity (Bradlaw and Casterline, 1979) or production of a keratinization response in
cell cultures (Knutson and Poland, 1980). If the mode of action of these com-
pounds is the same between mammal species, detailed isomer specific analysis of
Great Lakes fish and of mink experimentally exposed to these fish could provide
*an indication of the relative significance of the PCB content compared to other
microcontaminants in causing the mink reproductive failure. These data may not
be important to the mink ranchers that are no longer dependent on Great Lakes
fish as a feedstuff. They are, however, important for administrators of the Great
Lakes Water Quality Agreement so that priorities can be decided on a scientific
basis for cleanup of sources of the most dangerous pollutants, in order to reduce
human exposure through eating contaminated fish.
In addition, recent reports have shown that population densities of wild mink
and otters (Lutra canadensis) are very low in areas close to Lake Ontario (Robert
Foley, New York Department of Environmental Conservation, Gloverville, NY,
personal communication). Similarly, trapping data from the Wisconsin Depart-
ment of Natural Resources (Kohn and Ashbrenner, 1984) showed that no otters
were caught in four counties near the Door Peninsula, which divides Green Bay
from Lake Michigan. In both instances the presence of organochlorine pollutants
in fish was the suspected cause, though historical trapping data, the possible role
of human disturbance and the contaminant levels must be considered before a causal
connection can be made. Further evidence of the probable role of PCB in the
ecology of wild piscivorous mammals was provided by the analysis of a dead mink
found in a marsh near Green Bay. The liver of this animal contained 5.7 ppm PCB
(T. J. Kubiak, U.S. Fish and Wildlife Service. Green Bay, WI, personal com-
munication) and this corresponds with the levels in mink liver associated with adult
mortality in feeding experiments (Aulerich and Ringer, 1977).
140 Epidemics In Birds and Mammals Caused by Chemicals In the Great Lakes
Because mink and otter are piscivorous and extremely sensitive to PCB and,
likely, to compounds that have a similar mode of action, these species have a high
potential to be critically exposed. Though data on mink and otter abundance may
be difficult to collect, there is a priority requirement to undertake population sur-
veys to document areas where the species have been extirpated. Without these data
it will continue to be difficult to make a scientific case for stringent controls on
sources of organochlorine pollutants to the Great Lakes Ecosystem.
3. REPRODUCTIVE FAILURE AND MORTALITY IN BIRDS
3.1. Initial Reports
Birds, in contrast to mammals, fish, and humans, have given clear, easily observed
signals when chemically induced epizootics occurred. These signals were observed
in colonies of fish-feeding birds on Lakes Michigan and Huron and on Lake On-
tario between 1964 and 1970.
3.1.1. Lake Michigan and Lake Huron
In 1964, bird banders visiting herring gull (Larus argentatus) colonies in Door
(\ County, Green Bay, Wisconsin, noted a drastic reduction in the number of band-
able chicks compared with previous years (Keith, 1966); the colonies contained a
O great many disused nests with cold eggs. In 1963, adult herring gulls had been
r shot for pesticide residue analysis and found to contain more than 2000 ppm in the
S body fat. In 1964, Keith (1966) undertook his detailed pioneering work on the
herring gull colony on Little Sister Islaid, Green Bay, Wisconsin, to investigate
whether reproductive success was affected by this pesticide residue. In the same
year. Ludwig and Tomoff (1966) noted unusually low reproductive success of her-
ring gulls nesting on Bellow Island in Grand Traverse'Bay, Lake Michigan, and a
decrease in the number of nesting pairs at the colony. Subsequent annual obser-
vation of colony size showed a precipitous decline in the breeding population (Lud-
wig and Ludwig, 1969).
In the late 1960s, Postupalsky documented reproductive failure and 30% egg-
shell thinning among double-crested cormorant (Phalacrocorax auritus) popula-
tions in Georgian Bay, Lake Huron (cited in Weseloh et al., 1983). During the
past four years, 1981-1984, there has been an anomalous incidence of congenital
abnormalities among Green Bay, Wisconsin, colonies of Forster's terns (Sterna
forsteri) black-crowned night herons (Nycticorax nycticorax), and double-crested
cormorants (T. J. Kubiak, personal communication).
3.1.2. Lake Ontario
There is anecdotal evidence from two sources that severe reproductive failure in
herring gulls occurred in Lake Ontario colonies in 1966 (Gilbertson, 1983) that
had been apparently normal in 1965. Several other bird species exhibited repro-
ductive failure including double-crested cormorants in Lake Ontario in 1963 (cited
3. Reproductive Vallure and Mortality In Birds 141
in Gilbertson, 1975) and common tears (Sterna hirundo) in Hamilton Harbour in
1970 (Gilbertson and Reynolds. 1972). Bald eagles (Hallaretus leucocephalus)
had completely disappeared from this lower Great Lake by the late 1950s.
3.2. Upper Lakes
3.2.1. Symptoms and Hypothesis
The detailed surveys of reproductive success of Lake Michigan herring gulls
showed (Table 6.4) that eggs in the colonies from Green Bay .(Keith, 1966) and
Grand Traverse Bay (Ludwig and Tomoff, 1966) had a low hatching success of
41 and 28%, respectively. Twenty percent of the eggs at the Grand Traverse Bay
colony were cracked and shell damage was noted in the Green Bay colony. At
these two colonies only three or four fledged young were produced per 10 nests.
In contrast, at a midlake colony in the Beaver Islands, over 80% of the eggs hatched
and an estimated 12 young were produced per 10 nests.
. Adult birds at the colony at Grand Traverse Bay were exceptionally.aggressive
and it was suggested that this behavioral anomaly could account for the high pro-
portion of cracked eggs. Both authors suggested that disturbances in adult behavior
might adversely affect incubation of the eggs (Keith, 1966; Ludwig and Tomoff.
1966). The presence of organochlorine chemicals in the eggs and in adults was
hypothesized to be the cause of the reproductive failure at the Green Bay and Grand
Traverse Bay colonies (Ludwig and Tomoff. 1966).
Herring gull eggs from the three Lake Michigan study colonies were analyzed for
organochlorine residues. The results (Table 6.5) indicate that eggs from the Green
Bay colony were the most contaminated and contained twice the level of DDE
compared to the least contaminated eggs from the colony at the Beaver Islands.
There was an apparent inverse relationship between fledged young per pair and
the residue levels in the eggs. Both papers concluded, however, that there were
difficulties relating the observed productivee failure to organochlorine residues.
Keith (1966) had reservations based on the discrepancy between his findings of
low hatchability and experimental feeding studies with quail (Colinus virginianus)
and pheasant (Phasianus colchicus) in which eggs of contaminated hens hatched
as frequently as eggs of controls. Ludwig and Tomoff (1966) concluded that be-
cause the residues in samples of live and dead eggs from the Grand Traverse Bay
colony were similar to the levels in live and dead eggs from the Beaver Islands the
residue content could not be a. factor in the reproductive failure.
These studies were among the first in the Great Lakes to try to investigate the role
of organochlorine insecticides on avian reproduction. In the Green Bay studies,
there were no controls but reliance was placed on comparisons with other pub-
lished studies of herring gull reproduction. Even in the mid 1960s it was likely
AN 00 00 2
*J GO '
3. Reproductive Failure and Mortality in Birds 3
Table 6.5 Organochlorine Residue Levels in Lake Michigan Herring Gull
Eggs, Collected Alive, in Wet Weight Parts per Million
Location Year DDT TDE DDT + TDE DDE Dieldrin
Oeen Bay 1964 19 6 200 -
Grand Traverse Bay 1965 18 140 1.5
Beaver Islands 1965 16 100 1.0
Source: After Keith (1966): Ludwlg and Tomoff (1966).
difficult to find relatively uncontaminated colonies. There were serious problems
in the egg-sampling regime since eggs were not sampled randomly and some eggs
were taken at the end of incubation. Thus no generalizations can be drawn with
certainty concerning the relative contamination of the colonies. These two studies
were undertaken and published before the phenomenon of eggshell.thinning had
S been documented (Ratcliffe, 1967) and before the relationship between DDE and
eggshell thinning had been demonstrated on a correlative basis (Hickey and An-
demon, 1968; Gilbertson, 1974) or on an experimental basis (Wiemeyer and Por-
Organochlorine residue analysis in the mid 1960s was still in its infancy.. As
S Ludwig and Ludwig (1969) later stated, the chemist who analyzed the Lake Mich-
igan eggs suspected that part of the DDT and DDD reported represented interfer-
ence from polychlorinated biphenyls. Estimation of the levels of PCBs was not
included in these analyses and thus their role in the reproductive failure cannot be
assessed. More particularly, the subject of ultratrace contaminants such as poly-
chlorinated dibenzo-p-dioxins and dibenzofurans was not even being discussed by
analytical chemists involved in analysis of environmental samples. The absence
of any stored specimens precludes the possibility of reanalyzing the samples.
3.2.4. .New Symptoms and Hypotheses
The research (Ludwig and Tomoff. 1966) did. however, provide a new hypothesis
(Ludwig and Ludwig, 1969): the presence of organochlorine insecticides in herring
gulls might cause significant mortality in adults, particularly during times of stress
such as migration, storms, and moulting. Long-term observation of the colony at
SGrand Traverse Bay had shown that the population increased during the 20-year
period 1944-1964 from 1000 to 2500 breeding pairs. However, in 1964 a slight
decline to 2200 was noted. In subsequent annual visits to this island, a rapid pop-
ulation decline was recorded until only 850 breeding pairs remained in 1968 (Lud-
wig and Ludwig, 1969).
3.2.5. Etiology of Adult Mortality
To test this new hypothesis, the Ludwigs in 1967 captured 15 nesting adult herring
gulls from Bellow Island, Grand Traverse Bay, and transported them to Port Huron
to investigate the effects of starvation on the gulls. Even as they were capturing
144 Epidemics In Birds and Mammals Caused by Chemicals in the Great IakLes
the gulls, they observed an adult bird with symptoms of pesticide poisoning, in-
cluding uncontrollable eye tremors and violent body spasms (Ludwig and Ludwig,
1969). Previous research with presumably uncontaminated herring gulls captured
at Rogers City, Lake Huron, had shown that this species could withstand 30% loss
of body weight in 17 days without losing vigor or flight. In contrast, within 3 days
the first of the gulls started to exhibit signs of insecticide poisoning, a second bird
died in tremors on day 6, and four others died by the tenth day. Chemical analysis
of the brains indicated that combined levels of insecticides associated with mor-
tality in ths population were about 12 ppm DDT, 6 ppm DDD, 180 ppm DDE.
and 2 ppm dieldrin. The authors concluded that the 20% annual decline in the
breeding population of herring gulls on Bellow Island at Grand Traverse Bay re-
suited from increased adult mortality caused by insecticide residues.
Further support for the conclusion of increased adult mortality caused by or-
ganochlorine chemicals was forthcoming from subsequent investigations. In
Southern Ontario, an anomalous number of emaciated dead or dying ring-billed
gulls (Larus delawarensis) were found during the 1969 and 1973 fall migrations.
Sileo et al. (1977) undertook autopsies and residue analysis on specimens submit-
ted to the University of Guelph. The investigations resulted in an improved method
for determining the significance of multiple residues by calculation of an organo-
chlorine index for residues in the brain. Based on residue analysis of wild birds
and experimental poisoning of a variety of bird species, Sileo et al. (1977) con-
cluded that the following individual levels of organochlorine compounds in the
brain were associated with death: 5 ppm dicldrin, 150 ppm DDE, or 300 ppm
PCB. The organochlorine index was calculated by adding the ratios of the levels
determined by residue analysis compared with the above levels associated with
death. The investigators showed that there was an inverse relationship between
body weight and brain organochlorine indices. They concluded that stresses such
as post nuptial or post juvenal molt likely contributed to the observed emaciation
3.3. Lake Ontario
The epidemiology and etiology of the reproductive failures in the lower Great
Lakes have been reviewed in detail by Mineau et al. (1984) and Gilbertson (1983).
In the 1960s and 1970s Lake Ontario and Lake Michigan were two freshwater
locations with high levels of organochlorine pollutants. After the initial findings
of the 1970 reproductive failure of common terns in Hamilton Harbour (Gilbertson
and Reynolds, 1972), fieldwork was undertaken to determine the geographic extent
of the reproductive failure and the species affected. Seven species were studied
during the next two years: herring gulls, ring-bill gulls, common teams, caspian
teams (Hydroprogne caspia), black terns (Chlidonias niger), black-crowned night
herons, and double-crested cormorants. In addition, contaminated eggs of these
species were collected and incubated artificially to determine hatchability under
3. Reproductive Failure and Mortality In Bird. )45
controlled environmental conditions and to investigate signs and symptoms of the
3.3.2. Signs and Symptoms
These preliminary studies revealed serious reproductive failure among double-
crested cormorants in Lake Ontario. Several of the Lake Ontario species, partic-
ularly the common tern, exhibited a high incidence of congenital abnormalities
such as crossed beaks, small eyes, or duplicate limbs (Gilbertson et al., 1976).
Similarly, low hatchability and fledging rates were evident in several of these spe-
cies. The artificial incubation procedures indicated that the low hatching success
was due, at least in part, to a factor inside the egg. Characteristically, chicks that
died after pipping the shell were undersize and had large unresorbed yolk sacs.
After the 1972 field season, the herring gull was chosen (Gilbertson, 1974) as
the species for routine monitoring of trends of reproductive success and of levels
of organochlorine compounds and for detailed etiological research. Among the
most important reasons for the choice was the relatively nonmigratory habits of
the adult breeding population in the Great Lakes. Thus the adult gulls tended to
lay eggs that reflected the organochlorine contamination near the colony. This
routine monitoring subsequently became part of the surveillance plan of the Inter-
national Joint Commission in 1977.
The data for the number of fledged young per pair are shown in Table 6.6 and
indicated that severe reproductive failure occurred in the Scotch Bonnet Island
colony in Lake Ontario between 1972 and 1975 but that reproduction was normal
in the year 1977. Similarly the Port Colbome colony in eastern Lake Erie exhibited
reduced reproductive success between 1974 and 1976 but was normal in 1978.
Reproductive success in Lakes Huron and Superior appeared normal between 1975
and 1978. At the same time organochlorine residues in herring gull eggs collected
from colonies in the lower Great Lakes declined (Weseloh et al., 1979). Thus
interpretation of cause and effect was made more complex because of the shifting
baseline of fledging success and residue levels.
In 1972, fledging success in five Lake Ontario colonies was between 0.06 and
Table 6.6 Fledged Young per Pair of Great Lakes Herring Gulls
Location 1972 1973 1974 1975 1976 1977 1978
Lake Ontario 0.12 0.06 0.15 1.10 1.01
Lake Erie 0.48 0.65 0.79 1.45
Lake Huron 1.48 1.12 1.40
Lake Superior 1.32 1.55 1.12
Source: After Weseloh et al. (1979).
146 Epidemics In Birds and Mammals Caused by Chemicals In the great t lakes
0.21 fledged young per pair (Gilbertson, 1974). which represents only about a
tenth of the normal fledging success. These studies also demonstrated a relation-
ship between eggshell thinning and DDE: eggshell breakage and flaking was ob-
served in the most contaminated colonies.
Detailed studies of the colony on Scotch Bonnet Island in 1973 showed that the
reproductive failure was characterized by failure to lay eggs in 27% of the nests,
early embryonic mortality in 20% of the eggs laid. low hatching success (17%),
low fledging success (0.06 fledged young per pair), and high chick mortality (74%)
(Gilbertson and Hale, 1974a,b). Chemical analysis of eggs from the Scotch Bonnet
Island colony showed the presence of about 140 ppm DDE and 550 ppm PCB on
a dry weight basis. In addition, the eggs contained dieldrin, hcptachlor epoxide,
1. Based on the small size of chicks that pipped the shell but failed to hatch
and the similarity of these findings to those caused by antithyroid compounds, it
was hypothesized (Gilbenson and Fox, 1977) that an antithyroid compound was
acting on the embryos. No definitive study on this aspect has yet been completed
2. Based on the high levels of PCB present it was postulated that a chick edema
active compound such as polychlorinated dibenzofurans could be causing the high
Embryonic mortality and abnormalities. A study (Gilbertson and Fox, 1977) was
then undertaken to look for signs of chick edema disease in herring gull chicks at
3. Based on the poor hatchability and high rate of disappearance of eggs it was
postulated (Fox et al., 1978; Peakall et al., 1980) that adult breeding behavior had
been disturbed as had also been noted in the two Lake.Michigan studies (Keith,
1966; Ludwig and Tomoff, 1966).
These hypotheses concerned both intrinsic and extrinsic factors affecting the hatch-
ability of the eggs. To separate the relative importance of these factors, a series
of egg exchange experiments was conducted between clean and contaminated col-
onies in 1975, 1976, and 1977 (Peakall et al., 1980). These three years were the
period in which herring gull reproduction in the Scotch Bonnet Island colony on
Lake Ontario improved from 0.15 to 1.1 fledged young per pair. The design of
the egg exchange experiment is shown in Table 6.7.
"Clean" maritime and "dirty" Lake Ontario eggs were incubated by clean and
dirty adults and by artificial incubation. The results showed that in 1975 there were
both marked intrinsic and extrinsic (adult behavioral) factors affecting hatchability:
in 1976 the extrinsic was less evident and in 1977 there was no apparent effect.
To investigate the behavioral aspect, a study was undertaken in 1976. using
telemetered eggs. The study showed that Lake Ontario herring gulls were less
attentive to their nests and applied less heat to their eggs than uncontaminated
; r< e
I i. 8
148 Epidemics In Birds and Mammals Caused by Chemicals In the Great Lakes
controls on Kent Island, New Brunswick (Fox et al.,1978). In the Lake Ontario
colony there was a significant 1 C* decrease in the average temperature of the
telemetered .eggs in unsuccessful nests compared with successful nests. Unsuc-
cessful nests were left unattended for three times longer than successful nests. The
level of organochlorine residues in eggs removed for analysis from telemetered
nests was correlated with the length of time that nests were left unattended (Fox
et al., 1978).
To investigate whether there was a chick edema factor present in the Lake On-
tario and eastern Lake Erie herring gull eggs a bioassay was undertaken in 1974
(reviewed in Gilbertson, 1983). It was postulated that if a chick edema factor were
present in these eggs then a suite of signs and symptoms would be evident, in-
cluding embryonic mortality, growth retardation, abnormalities, subcutaneous
edema, pericardial edema, peritonial edema, porphyria, and liver enlargement and
necrosis with fatty degeneration (Gilbertson and Fox, 1977).
Eggs were collected from two colonies on Lake Ontario, at the eastern end of
Lake Erie, and from control colonies in New Brunswick and northern Alberta, and
were incubated artificially. Only 26% of the eggs from Lake Ontario colonies
hatched compared with 53% from Lake Erie colonies and 69% for the controls.
Chicks from the Great Lakes colonies, particularly those that failed to hatch,
showed extensive subcutaneous edema and fragility of the skin and pericardial
edema. Lake Ontario chicks exhibited a milky edema in the peritoneal cavity. In
Addition, there was evidence of growth retardation: Lake Ontario chicks that
hatched had tarsus lengths that were 7% smaller than controls, and those that pipped
S and failed to hatch were 20% smaller and had large unresorbed yolk sacs. Com-
Sparison of liver weights to tarsus lengths indicated that Lake Ontario chicks were
19% greater than controls and Lake Eric chicks were 24% larger, indicating liver
enlargement. Lake Ontario chicks that pipped but failed to hatch had markedly
elevated porphyrin levels and there was evidence of liver pathology in chicks from
Lake Erie and Lake Ontario colonies including fatty degeneration and liver necro-
sis (Gilbertson and Fox, 1977). The rate of abnormalities in several colonies of
fish-feeding birds on the lower Great Lakes in the year 1971 to 1973 was about
1%. indicating the presence of a teratogenic compound (Gilbertson et al., 1976).
In the incubation experiment one abnormal herring gull embryo was found among
the 47 eggs from Lake Ontario.
Based on these findings it was concluded that there was a chick edema active
compound in the herring gull eggs from the lower Great Lakes. Subsequent chem-
ical analysis of aliquots of frozen egg samples revealed the presence of 2,3,7,8-
tetrachlorodibenzo-p-dioxin at a concentration of 500 ppt in 1974 (Canadian Wild-
life Service, unpublished data). Other retrospective analyses showed that levels in
eggs had been as high was 1200 ppt in the early 1970s.
The detailed review of the chick edema studies (Gilbertson, 1983) concluded with
two questions. How important was chick edema disease in the reproductive failures
in fish-feeding colonial birds in the lower Great Lakes? What other chick edema
S4. Concluslom s 9
compounds wee present that might have contributed to the signs of chick edema
Disease? The review also concluded that those questions could not be answered
without isomer-specific analysis of the chick edema compounds present in the
eggs.* These questions might seem academic a decade after the end of the epi-
demic. However, a new outbreak of chick edema disease has been reported (T. J.
Kubiak, personal communication). Investigations at Green Bay have had several
advantages over the previous work and this is likely to result in the first compre-
hensive statement of the causes of an epidemic. These advantages include knowl-
edge of the workable methodologies used in the studies on the lower Great Lakes
and access to reliable ultratrace analysis which will yield congener specific anal-
These case histories, concerning reproductive dysfunction in populations of fish-
feeding mammals and birds, formally demonstrate the causal interrelatioitship be-
tween the observed epidemics and contamination of Great Lakes fish with organo-
chlorine chemicals. This formal demonstration is based on a consistency of the
evidence relating the characteristics of the observed reproductive pathology to the
presence of specific chemicals in the fish and in the adults and their progeny and
to the experimentally determined toxicological properties of the chemicals. Be-
cause these case histories are the essence of the complaints concerning organo-
chlorine pollution of the Great Lakes, it is vital that an unassailable scientific case,,
based on sound epidemiology and etiology, should be prepared as the basis for
cleanup negotiations or litigation (Mortimer. 1984; Gilbertson, 1984). To be un-
assailable, cases have to be prepared using the methodology and rigor of forensic
*Note Added in Press: Haris. Kubisk. and Trick (1985) have published a detailed account of their
studies on Forster's tems nesting at green iay. Lake Michigan. and at an inland location. At the Gren
Bay colony, egg hatchability was significantly reduced and al hatching chicks weighed less and had
engaged livers. Embryonic mortality was associated with subcutaneous edema. Egg-swap experiments
and controlled incubation studies showed evidence or aberrant parental behavior in Green Bay items
resulting in an increase of 3.7 days in the incuhation period. In addition. Green Bay tem ggs took
4.6 days longer to incubate. Thus eggs hatchability at the Green Bay colony was affected by factors
intrinsic nd extrinsic to the eggs. Fledgin success in the Green Bay colony was markedly reduced.
Organochklrine chemical residue analysis of eggs from the two colonies revealed significantly higher
residues of various chick edema active compounds including polychlorinated biphenyls. dibenzo-p-
dioxins, and dibenzofurans at the Green Bay colony. Transformation of the results of isomer-specific
analyses to 2.3,7.8-TCDD equivalents revealed that more than 80% of the chick edema activity was
contributed by 2,3,4.3'.4'.penachlorobiphenyl.
This combination of biological and chemical analysis has produced a sophisticated epidemiological
case history. Specific regulatory controls can now be formulated for the industries on the Fox River
based on a knowledge of the relative contribution of the various etiological agents present. The caue
history reemphasizes the powerful versatility of studies of avian reproduction for investigating envi-
150 Epidermis in Birds and Mammals Caused by Chemicals in the Great Lakes
The case histories of research on the effects of chemicals on fish-eating birds
and mammals in the Great Lakes show the effectiveness of using these organisms
as indicators of the health of the aquatic environment. In contrast, epidemiological
and etiological research on fish and human populations have proved much more
difficult to conduct and, to date, much less successful. It is recommended that
research and monitoring on birds and mammals should continue to be fostered to
provide this essential part of the scientific case for restoring ecosystem health.
The model depicted in Table 6.1 has been used, sometimes perhaps unknow-
ingly, to investigate the incidence and cause of outbreaks of chemically induced
diseases. The model has application when the preventive approach to management
of chemicals has not succeeded and contamination by persistent, toxic, and bio-
concentrated chemicals has occurred on a widespread geographic scale. The suc-
cessful implementation of the model demands teamwork and careful scientific
management because no individual has the requisite skills to solve the toxicolog-
ical riddles posed by releases of chemicals to the natural environment.
I am indebted to Tim Kubiak and Robert Foley for information on the situation in
Wisconsin and New York, respectively, and to Dr. David Peakall, Dr. Chip
SWeseloh, and Pierre Mineau for reviewing the manuscript. Dr. Marlene Evans
S kindly asked me to undertake this assignment. Ms. Brenda Black patiently pro-
cessed the words in the various drafts.
Aulerich. R. J., and Ringer. R. K. (1970). Some effects of chlorinated hydmrcarbon pesticides on
mink. Am. Fur Breeder 43, 10-11.
Aulerich. R. J.. and Ringer. R. K. (1977). Current status of PCB toxicity to mink. and effect on their
reproduction. Arch. Envirom. Contam. Toxicol. 6. 279-292.
Aulerich. R. J.. Ringer. R. K.. Seagann H. L.. and Youats. W. 0. (1971). Effects of feeding coho
salmon and other Great Lakes fish on mink reproduction. Can. J. Zool. 49, 611-616.
Aulerich. R. J.. Ringer. R. K.. and Iwamoto. S. (1973). Reproductive failure and mortality in mink
fed on Great Lakes fish. J. Reprod. Feril.. Suppl. 19, 365-376.
Bowes. G. W., Mulvihill. M. J.. Simoneit. B. R. T., Burlingame. A. L., and Risebrough. R. W.
(1975). Identification of chlorinated dibenzqfurans in American polychlorinated biphenyls. Nature
(London). 256, 305-307:
Bradlaw. J. A.. and Casterline, J. L. (1979). Induction of enzyme activity in cell culture: A rapid
screen for detection of planar polychlorinated organic compounds. J. Assoc. Of. Anal. Chem. 62.
Duby. R. T. (1970). Pesticides vs reproduction still a puzzle. Am. Fur Breeder 43, 15.
Fein. 0. G.. Jacobson. J. L.. Jacobson. S. W., Schwartz. P. M.. and Dowler. J. K. (1984). Prenatal
exposure to polychlorinated biphenyls: Effects on birth size and gestational age. J. Pedlatr. 105,
Fitchko. J. (1986). Literature review of the effects of persistent toxic substances on Great Lakes Biota.
Report of the Health of Aquatic Communities Task Force to the Great Lakes Science Advisory
Board. International Joint Commission. 256 pp.
Fox, G. A.. Gilman. A. P., Peakall. D. B.. and Anderkl. F. W.(1978). Behavioural abnormalities of
nesting Lake Ontario herring gulls. Wildly. Manage. 42. 477-483.
Gilbertson. M. (1974). Pollutants in breeding herring gulls in the lower Great Lakes. Can. Field Nat.
Gilbertson. M. (1975). A Great Lakes tragedy. Nature (Canada) 4, 22-24.
Gilbertson. M. (1983). Eiology of chick-edema disease in herring gulls in the lower Great Lakes.
Chemosphere 12, 357-370.
Gilbertson M. (1984). Need for development of epidemiology for chemically induced diseases in fish
in Canada. Can. J. Fish. Aquat. Se. 41. 153.-1540.
Gilhenson. M.. and Fox. 0. A. (1977). Polluatnt.asociated embryonic mortality of Great Lakes her-
ring gulls. Enmiron. Pallut. 12, 211-216.
Gilbertson. M.. and Hale. R. (1974a). Early embryonic mortality in a herring gull colony in Lake
Ontario. Can. Field Noa. 8, 354-356.
Gilbertson. M.. and Hale. .(1974b). Characteristics of the breeding failure of a colony of herring
gulls on Lake Ontario. Can. Field NUt. 8. 356-358.
Gilbertson. M.. and Reynolds, L. M. (1972). Hexachtorobenzene (HCB) in the eggs of common tears
in Hamilton Harbour, Ontario. Bull. Environ. Contam. Toicol. 7, 371-373.
Gilbertson. M.. Morris. R. D.. and Hunter. R. A. (1976). Abnormal chicks and PCB residue levels
in eggs of colonial birds on the lower Great Lakes (1971-73). Auk 93, 434-442.
Goldstein J. A. (1980). Strcture-activity relationship for the biochemical effects of halogenated aro-
matic hydrocarbons and the relationship toxicity. In R. Kimbrough (Ed.). Halogenated Biphe.
nyls. Terphenyls. Nahthalenes. Dihentodixlns. and Related Products. ElsevierlNonh-Holland.
Hanrtough. G. R. (1965). Great Lakes fish now suspect as mink food. Am. Fur Breeder 38. 35-27.
Hickey, J. J.. and Andeson. D. W. (1968). Chlorinated hydrocarboas and eggshell changes in rap-
torial and fish-eating birds. Science 162, 271-273.
International Joint Commission (1983). Report on Great Lakes Water Quality. Appendix. Great Lakes
Surveillance. IJC. Windsor. Ontario.
Jacobson. J. I... Jacohson. S. W.. Schwartz. P. M.. Fein (. 0. and Dowler. J. K. (1984). Prenatal
exposure to an environmental toxic: A test of the multiple eflecct model. Dev. Pswycho. 20, 523-
Keith. J. A. (1966). Reproduction in a population of herring gulls (imrus argeqtatanl ) contaminated by
DDT. J. Apple. Eco. 3 (Suppl.). 57-70.
Knutson, J. C., and Poland. A. (1980). Keratinization of mouse teratoma cell line XB produced by
22.214.171.124-tetrachlorodibenso-p-dioxin: An in vitro model of toxicity. Cell (Cambridge. Mass.) 22,
Kohn, B. E.. and Ashbrenner. J, E. (1984). Harvest and Population Status of River Otter in Wisconsin.
Rep. No. 129. Department of Natural Resources. Madison. Wis.
Ludwig. J. P., and Ludwig. C. E. (1969). The effect of starvation on insecticide contaminated herring
gulls removed from a Lake Michigan colony. Proc. 12th Conf. Great Lakes Res.. 53-60.
Harris. H. J.. Kubiak. T. J. and Trick. T. A. (1985). Microaontaminants and Reproductive Impairment
of Forster's Tens on Green Bay. Final Report to U.S. Fish and Wildlife Service. University of
Wisconsin, Sea Grant Inst.. Wis. Dept. of Nat. Resour. and Green Bay Metropolitan Sewerage
Dist.. Sea Grant Office. ES-105. University of Wisconsin. Green Bay.
Ludwig. J. P.. and Tomoff. C. S. (1966). Reproductive success and insecticide residues in Lake
Michigan herring gulls. Jack-Pine Warbler 44. 77-84.