Full text of DOI:10.1002/etc.5620200427

Environmental Toxicology and Chemistry, Vol. 20, No. 4, pp. 907–912, 2001
2001 SETAC
Printed in the USA
730-7268/01 $9.00 .00
᭧
0
ϩ
EFFECTS OF COPPER ON OLFACTION OF COLORADO PIKEMINNOW
D
ANIEL W. BEYERS* and MICHAEL S. FARMER
Larval Fish Laboratory, Department of Fishery and Wildlife Biology, Colorado State University, Fort Collins, Colorado 80523, USA
(
Received 16 May 2000; Accepted 29 August 2000)
Abstract—Effects of copper on olfaction of Colorado pikeminnow (Ptychocheilus lucius) were investigated by exposing fish for
4 or 96 h, then evaluating olfactory ability using a behavioral assay and observing olfactory structures using scanning electron
2
microscopy (SEM). The behavioral assay measured a response known as fright reaction. Failure of exposed fish to demonstrate a
fright reaction in the presence of skin homogenate assumed to contain fright pheromone was considered evidence of copper-induced
loss of olfactory ability. Regression analysis was used to describe the response of fish as a function of copper concentration at
each exposure duration. Olfactory ability declined with increasing copper concentration. For copper concentrations less than 66
␮
g/L, olfaction was more sensitive to exposure at 24 h than at 96 h. This result suggests that physiological adaptation and recovery
of sensory ability occurred despite continuous exposure in the 96-h treatment. Protective mechanisms induced by exposure may
have reduced sensitivity to copper by 96 h. Systematic surveys using SEM to detect presence or absence of olfactory receptors
confirmed results of behavioral assays. Copper concentrations in one river inhabited by Colorado pikeminnow were compared with
effective concentrations estimated by regression. Comparisons suggest that ambient copper concentrations may occasionally inhibit
olfaction of wild fish.
Keywords—Copper
Colorado pikeminnow
Olfaction
Behavioral assay
Scanning electron microscopy
INTRODUCTION
scanning electron microscopy (SEM). Advantages of combin-
ing the two techniques were that the behavioral assay required
study fish to detect, process, and respond to stimuli in an
ecologically relevant fashion, and SEM observations provided
evidence of related structural changes of olfactory receptors.
The behavioral assay used in this investigation measured a
response known as fright reaction in Colorado pikeminnow.
The fright reaction is a well-known response of Ostariophysan
fishes to fright pheromone, which is released from specialized
cells in the epidermis when mechanical damage occurs [10,11].
Fright pheromone is detected by olfaction [10,11]. Conse-
quently, failure of metal-exposed fish to demonstrate a fright
reaction in the presence of the pheromone was considered
evidence of a copper-induced loss of olfactory ability. Re-
sulting exposure–response relationships were compared with
copper concentrations near a Colorado pikeminnow spawning
site to evaluate potential for effects on wild fish.
The Colorado pikeminnow (Ptychocheilus lucius) is a large
cyprinid endemic to the Colorado River Basin, USA [1]. His-
torically, the Colorado pikeminnow was widespread in warm-
water streams and rivers, but the species was listed as federally
endangered in 1967 in response to declining populations [2].
The decline of Colorado pikeminnow is commonly attributed
to interactions with introduced fishes, construction of dams,
and habitat modification [3,4], but effects of chemical contam-
inants are increasingly becoming a concern. An important el-
ement of Colorado pikeminnow life history is that natural
reproduction is preceded by spawning migrations that may
exceed 100 km in length. The means of navigation used by
Colorado pikeminnow during migration are unknown, but it
has been proposed that olfaction is a critical component of the
homing process [1]. Many contaminants are known to disrupt
olfaction in fish. Short-term sublethal exposure to copper, lead,
mercury, nickel, silver, zinc, extremes of pH, and naphthalene
inhibit olfactory ability [5–9]. Olfactory receptors are not pro-
tected from contaminant exposure by external membranes but
come into direct contact with water-borne solutes. Toxic sub-
stances may alter detection of olfactory cues through several
modes of action, including damaging organelles and enzyme
systems, direct interaction with membrane receptor sites, or
masking biologically important chemical signals.
MATERIALS AND METHODS
Experimental animals
Colorado pikeminnow were obtained from Dexter National
Fish Hatchery and Technology Center (Dexter, NM, USA).
Fish were fed a mixture of live
Aquarium Products, Glen Burnie, MD, USA) and a com-
mercially prepared flake diet (TetraMin ; TetraWerke, Melle,
Germany) twice daily. Colorado pikeminnow were 185 d old
after hatching) when studies were started. Total length and
Ͻ24-h-old brine shrimp nauplii
(
௡
The purpose of this investigation was to describe the effect
of contaminant exposure on olfactory ability of Colorado
pikeminnow. This objective was achieved by exposing Col-
orado pikeminnow to sublethal concentrations of copper for
(
wet weight ranged from 25 to 44 mm and 0.14 to 0.68 g,
respectively.
2
4 or 96 h and evaluating olfactory ability using a behavioral
assay coupled with observations of olfactory structures using
Exposure conditions
The 24- and 96-h exposures were conducted following pre-
scribed methods for renewal-acute toxicity tests [12]. Six con-
*
To whom correspondence may be addressed
(
danb@lamar.colostate.edu).
Contribution 117, Larval Fish Laboratory, Colorado State Uni-
centrations were studied for the 24-h exposure (
Ͻ10, 16.6,
versity, Fort Collins, Colorado, USA.
33.3, 66.5, 133, and 266 g/L copper) and five concentrations
␮
9
07

9
08
Environ. Toxicol. Chem. 20, 2001
D.W. Beyers and M.S. Farmer
were studied for the 96-h exposure (
Ͻ
10, 15.0, 30.0, 60.0, and
pikeminnow were juveniles obtained from the same mass cul-
ture used for study fish. Homogenate preparation involved
removing skin posterior to the head and anterior to insertion
of the caudal fin. Skin was cut into pieces and homogenized
(Omni-mixer, Sorvall, Norwalk, CT, USA) in dilution water
to a final concentration of 12.5 g/L.
The acclimation period for each behavioral assay was ini-
tiated immediately after conclusion of the 24- or 96-h expo-
sures. Fish and solutions from each replicate were transferred
to flow-through observation aquaria receiving dilution water
1
20 g/L copper). The 266- g/L concentration was not used
␮
␮
in the 96-h exposure because toxic effects were severe at 24
h, and it was anticipated that too few fish would survive the
longer exposure. Hamilton and Buhl [13] reported a 96-h LC50
of 269
der similar water-quality conditions. Concentrations were as-
signed to replicate ( 10) 1-L exposure beakers using a
␮g/L for Colorado pikeminnow exposed to copper un-
n
ϭ
balanced, randomized design. Beakers were filled with test
solutions to a depth of about 8.5 cm. Exposure solutions were
renewed every 24 h. Five fish were randomly assigned to each
exposure beaker. Fish were not fed during chemical exposure.
Cool-white fluorescent lamps were the only source of illu-
mination, and a 16:8-h light:dark photoperiod was maintained.
at a rate of 50 ml/min. Observation aquaria were 10
ϫ 20 ϫ
15 cm high, and depth of water was 12 cm. Each aquarium
was enclosed within a blind to visually isolate fish. An ob-
servation port allowed access for a video camera lens. Fish
were allowed to acclimate to observation aquaria for approx-
imately 90 min before being tested for presence of a fright
reaction. After the acclimation period, movements of fish in
each aquarium were video recorded for 1 min, followed by
introduction of skin homogenate and an additional 1 min of
video recording. Aliquots (0.1 ml) of skin homogenate were
introduced into observation aquaria through injection ports in
the water-delivery system, and final concentration of the ho-
mogenate in aquaria was 0.625 mg/L. Preliminary studies in
which dilution water was injected into the delivery system
showed that the procedure did not disturb the fish or elicit a
fright reaction. Behavioral assays were conducted so that per-
sons responsible for video camera operation and fright-pher-
omone injections were unaware (blind) of the experimental
treatment assigned to fish within each aquarium. Upon con-
clusion of behavioral assays, three fish from each experimental
treatment were preserved for SEM observations and the re-
mainder were returned to culture facilities. Remaining fish
from the 96-h exposure were cultured for 14 d, then were
reassayed using the procedure described above to evaluate the
potential for recovery of olfactory ability.
Physical and chemical conditions
Dilution water for copper exposures was a mixture of well
water and deionized water prepared to match the average hard-
ness in July of the Yampa River near Maybell, Colorado, USA.
Water quality for this time and location were selected because
of correspondence with spawning activity and presence of a
spawning site approximately 95 river-km downstream. Aver-
age water quality was estimated by calculating mean hardness
of two low (1981 and 1989), medium (1980 and 1982), and
high (1984 and 1985) water years (data from U.S. Geological
Survey gage 09251000; [14–19]). Target water quality char-
acteristics were hardness, 124 mg/L as CaCO ; alkalinity, 100
3
mg/L as CaCO ; and pH, 8.3. Dilution water was vigorously
3
aerated for approximately 48 h while being maintained at the
test temperature of 20
Ϯ 2ЊC. Dissolved oxygen, hardness, pH,
alkalinity, and specific conductance were measured daily dur-
ing toxicant exposures. Water temperature was measured con-
tinuously with a temperature recorder. Means and ranges for
the measured dilution water characteristics were dissolved ox-
ygen, 6.9 (6.5–7.5) mg/L; hardness, 117 (113–120) mg/L as
CaCO ; pH, 8.3 (8.1–8.5); alkalinity, 80 (71–88) mg/L as
3
CaCO ; specific conductance, 240
␮
S/cm, and temperature,
Video interpretation
3
1
8.5 to 21.0ЊC.
The behavioral reaction to fright pheromone is facilitated
by social interactions between fish. Consequently, the com-
bined response of all fish within an observation aquarium (ex-
perimental unit) was the unit of measurement. Generally, un-
disturbed fish were dispersed and oriented in different direc-
tions within an aquarium and only occasionally changed their
position during the 1-min preexposure observation period. Cri-
teria used to identify a positive fright reaction response were
demonstration of one or more of the following behaviors: (1)
cover seeking, characterized by fish assembling into a polar-
ized school, then moving toward the bottom of the aquarium
and reducing movement; (2) agitation, characterized by rapid
movement and frequent turning; and (3) dashing, characterized
by one or more fish displaying frenzied behavior, including
jumping out of the water, repeated and rapid changes in di-
rection, and swimming against aquarium walls. The criteria
were used only to qualify presence or absence of a fright
reaction and not to quantify intensity of the response.
Video interpretation was conducted by an observer who
was unaware (blind) of the experimental treatment assigned
to fish within each aquarium. The criteria were used to interpret
behavior during the first minute of each 2-min recording in-
terval in order to identify fish that were likely to give a false-
positive response. Fish that demonstrated cover seeking, ag-
itation, or dashing behaviors before introduction of the skin
homogenate were excluded from subsequent analysis. Presence
Copper solutions and analytical procedures
Stock solutions of analytical-grade copper sulfate (Mal-
linckrodt, Paris, KY, USA) were prepared in deionized water.
Exposure concentrations were prepared by pipetting the de-
sired amount of toxicant stock into 1-L glass beakers con-
taining 0.75 L dilution water. Exposure solutions were stirred
and transferred to exposure beakers within 30 min of prepa-
ration.
Toxicant concentrations were measured at the beginning of
exposure periods. On each occasion, two 50-ml samples were
collected from separate beakers for each exposure concentra-
tion. Unfiltered samples were placed in acid-washed polyeth-
ylene bottles, acidified to pH
acid, and held at 4 C until analyzed by inductively coupled
plasma emission spectroscopy (Soil, Water, and Plant Testing
Laboratory, Colorado State University, Fort Collins, CO,
USA). Chemical analysis confirmed accuracy of exposure con-
centrations. Measured copper concentrations averaged 92%
Ͻ 2 with analytical-grade nitric
Њ
(
standard error ϭ 4.8) of nominal. Measured and nominal con-
centrations were in close agreement; consequently, statistical
analyses were based on nominal concentrations.
Behavioral assay
Skin homogenate that was assumed to contain fright-pher-
omone was prepared before each assay. Donor Colorado

Effects of copper on olfaction of Colorado pikeminnow
Environ. Toxicol. Chem. 20, 2001
909
Table 1. Maximum-likelihood parameter estimates and significance
probabilities for a regression model describing probability of fright-
reaction response to skin homogenate by Colorado pikeminnow after
exposure to copper for 24 or 96 h
of a fright reaction from remaining fish was evaluated by ap-
plying the criteria during the 1-min interval after introduction
of the skin homogenate.
Copper concentrations in Colorado pikeminnow habitat
Standard
Parameter
Estimate
2.70
error
p
Copper concentrations in the Yampa River near Maybell
were described using data from a U.S. Geological Survey gage
␤0, intercept
2.87
1.69
0.0901
0.347
0.308
0.0269
(
09251000). Data were obtained from the U.S. Geological Sur-
␤1
, log (concentration)
Ϫ1.86
10
vey, Water Resources Division, Colorado District (Denver, CO,
USA). All data were used for the period of record from Sep-
tember 1974 to September 1991. The number of observations
ranged from two to four per year with a total of 29 for total
copper and 66 for dissolved copper. Detection limits for total
␤2, exposure time
0.144
␤3, log (concentration)
ϫ
10
exposure time
Ϫ0.0792
0.0520
0.0427
a
Estimates for the regression equation logit(p) ϭ ␤0 ϩ ␤1
ϩ ␤3C T, where p probability of response to skin homogenate,
logit(p) natural log [p/(1 p)], C exposure concentration ( g/
L), and T exposure time (h).
C ϩ ␤2T
ϭ
and dissolved copper ranged from 2 to 20 and 1 to 10
␮
g/L,
ϭ
Ϫ
ϭ
␮
ϭ
respectively. Frequency distributions for each constituent were
constructed.
slopes of the 24- and 96-h concentration–response relation-
ships. A significant coefficient suggests that the regression
lines are not parallel and further analysis is not required to
demonstrate that the concentration–response relationships are
different from each other. It also suggests that the effect of
copper was not consistent at both durations of exposure, i.e.,
there was an interaction of main effects. When an interaction
is detected, statistical tests of main effects cannot be simply
interpreted [22]. Coefficient ␤1 represents the test for effects
due to concentration gradient; ␤2 represents the test for effects
due to exposure duration.
Scanning electron microscopy
Scanning electron microscopy was used to examine the
olfactory epithelium of study fish for evidence of damage from
chemical exposure. Sample preparation and analysis followed
Lee [20]. Three fish from the control and each exposure con-
centration were fixed and preserved for SEM by immersion
in 3% glutaraldehyde buffered to pH 7.2 with sodium phos-
phate. Subsequently, all fish were processed as a batch to elim-
inate potential bias from variation in preparation techniques.
Specimens were rinsed with phosphate buffer and postfixed
with 1% osmium tetroxide (Sigma Chemical, St. Louis, MO,
USA), then dehydrated in a graded acetone series and critical-
point dried in a Polaron apparatus (Bio-Rad, Cambridge, MO,
USA). Individual specimens were mounted on aluminum stubs
with colloidal silver, sputter coated with gold (20-nm thick-
ness; Hummer VII sputtercoater, Anatech, Alexandria, VA,
USA), and examined using a Philips 505 SEM (Einhovn, Hol-
land).
Scanning electron microscopy was used only to confirm
results of behavioral assays because preparation and viewing
of specimens is labor and time intensive. Observations were
made on fish from controls and from the lowest copper con-
centrations that inhibited the fright-reaction response in all
replicates. It was anticipated that olfactory receptors in ex-
posed fish would be less abundant compared with control fish.
To reduce potential for investigator bias, examinations of ol-
factory epithelium were restricted to the same general regions
of the sensory surface. Target regions were systematically sur-
Concentrations of copper estimated to inhibit olfaction in
1
and 50% of the test organisms (EC1 and EC50) were also
calculated for each exposure duration using Proc Probit (with
options d logistic inversecl lackfit; [23]). In this case, the
ϭ
EC1 was arbitrarily selected as a conservative threshold for
adverse effects on an endangered species. Transformations
(
log ) were used if they improved the fit of regression models.
10
Graphical analyses of data and residual plots were conducted
to confirm that regression models were appropriate and to
evaluate compliance with statistical assumptions.
RESULTS
Behavioral assay
Olfactory ability declined as a function of copper concen-
tration after 24- or 96-h exposure, but slopes of the two con-
centration–response relationships were not parallel (Table 1).
There was an interaction (p ϭ 0.0427) between the effects of
veyed by conducting adjacent parallel scans at
ϫ
10,000 mag-
copper exposure and exposure time. As a consequence of this
interaction, the influence of copper on olfaction at each ex-
posure time must be evaluated separately. To facilitate eval-
uation, a graph of the probability of response to skin homog-
enate as a function of exposure concentration was constructed
using the regression model (Fig. 1). At copper concentrations
nification with a viewing field of 11.5 9.0
ϫ
␮m for 15 min.
Presence or absence of ciliated receptor cells was recorded.
Statistical analysis
Logistic regression was used to analyze the binomial re-
sponse (present or absent) for fright reaction as a function of
concentration and exposure duration. Proc Genmod (with op-
below 66 ␮g/L (i.e., the point where the lines cross), olfaction
was more sensitive to exposure at 24 h than at 96 h; at higher
concentrations, the opposite relationship was observed.
Behavioral assays conducted after a 14-d recovery period
tions link
ϭ logit, dist ϭ binomial, dscale; [21]) was used to
describe the response as a function of the independent vari-
ables. The full regression model had the form
showed that Colorado pikeminnow exposed to 60.0 ␮g/L cop-
per for 96 h regained olfactory ability (Table 2). Immediately
after exposure, 5 of 10 positive responses to skin homogenate
were elicited from fish in this experimental treatment, whereas
after the recovery period, positive responses were detected in
eight of nine remaining replicates.
logit(
probability of response to skin homogenate, logit(
natural log[ /(1 )], ␤ ϭ intercept, ␤1 ␤2 ϭ coefficients
for the linear terms of main effects, exposure concentra-
tion ( g/L), exposure time (h), and ␤ ϭ coefficient of
cross products. The coefficient of cross products tests for equal
p) ϭ ␤ ϩ ␤1C ϩ ␤2T ϩ ␤3CT
0
where
ϭ
p
ϭ
p)
p
Ϫ
p
,
0
C
ϭ
␮
T
ϭ
The EC1s and 95% confidence limits for 24- and 96-h
copper exposures were 2.61 (0.331, 6.24) and 18.3 (1.58, 30.0)
3

9
10
Environ. Toxicol. Chem. 20, 2001
D.W. Beyers and M.S. Farmer
Fig. 1. Probability of response to skin homogenate for Colorado
pikeminnow exposed to copper for 24 (circles) or 96 h (squares). The
mean response for all replicates at each exposure concentration is
shown. Probability 1
ϭ response; 0 ϭ no response.
␮
(
g/L total copper; EC50s were 43.3 (28.5, 69.0) and 56.0
39.3, 86.6) g/L total copper, respectively. Comparing these
estimates with total copper concentrations in the Yampa River
Fig. 2) revealed that the 24-h EC1 was equaled or exceeded
␮
(
by about 52% of the field samples. The 96-h EC1 was equaled
or exceeded by about 14% of the samples. For dissolved cop-
per, the 24- and 96-h EC1s were equaled or exceeded by about
5
7 and 2% of field samples, respectively. The EC50 values
were approached or exceeded on one or two occasions during
the 17-year period of record.
Fig. 2. Frequency of observation of copper concentrations in the Yam-
pa River near Maybell, CO, USA, for the period of record from
September 1974 to September 1991 (n ϭ 29 for total copper, 66 for
dissolved copper). Histograms represent number of observations less
than or equal to the label value in the interval between each category.
Scanning electron microscopy
The SEM surveys always detected one or more ciliated
receptor cells in control fish but did not detect receptor cells
immediately after exposure in fish in the other copper treat-
ments examined (Table 3). Observations confirmed that Col-
orado pikeminnow regenerated olfactory receptor cells after
BD
ϭ below detection.
9
6-h exposure to 60.0 ␮g/L copper. Ciliated receptor cells were
challenging fish to detect and respond to a skin homogenate
assumed to contain fright pheromone and by determining pres-
ence or absence of ciliated receptor cells in exposed fish. In
general, olfactory ability declined with increasing copper con-
centration. However, an unexpected result was that the re-
sponse to copper varied with duration of exposure. For copper
detected in the olfactory epithelium of all three fish from this
experimental treatment after the 14-d recovery period.
DISCUSSION
Mechanism for variable sensitivity
We exposed Colorado pikeminnow to a gradient of copper
concentrations and evaluated the influence on olfaction by
concentrations less than 66 ␮g/L, olfaction was more sensitive
to exposure at 24 h than at 96 h. This response is inconsistent
with the general principle of toxicology that, for a given con-
Table 2. Results of behavioral assays after exposure to copper for
24 or 96 h or after a 14-d recovery period; values are number of
positive behavioral responses/replicates assayed
Table 3. Results of scanning electron microscopy examinations after
exposure to copper for 24 or 96 h or after a 14-d recovery period;
values are number of fish with one or more ciliated receptor cells/
fish examined
2
4-h exposure
96-h exposure
Concen-
14-d recoverya
Concen-
tration
g/L)
Concen- Positive
Positive
tration
Positive
tration
re-
2
4-h exposure
96-h exposure
Concen-
14-d recoverya
(
␮
responses
(␮
g/L) responses
(␮g/L) sponses
Concen-
tration
g/L)
Concen-
tration Receptors
(␮g/L) detected
10b
6.6
3.3
6.5
10b
15.0
30.0
60.0
10b
Receptors
detected
tration Receptors
g/L) detected
Ͻ
10/10
5/9
6/10
4/10
2/10
0/10
Ͻ
10/10
10/10
8/9
5/10
0/6
Ͻ
9/9
8/9
9/9
8/9
ND
(
␮
(␮
1
3
6
33
66
15.0
30.0
60.0
Ͻ
266
10b
3/3
0/3
Ͻ
10b
60.0
120
3/3
0/3
0/3
Ͻ
10b
3/3
3/3
ND
c
1
2
120
ND
60.0
NDc
a
a
Fish were held in control water for 14 d following 96-h exposure
Fish were held in control water for 14 d following 96-h exposure
to copper concentrations.
Control.
to copper concentrations.
Control.
b
c
b
c
ND
ϭ
not determined because few fish recovered from copper
ND
ϭ not determined because few fish recovered from copper
exposure.
exposure.

Effects of copper on olfaction of Colorado pikeminnow
Environ. Toxicol. Chem. 20, 2001
911
centration, the magnitude of toxic effect increases with ex-
posure time.
moreceptors in fish may have a neuroendocrine link to sur-
rounding cells in the epidermis. If this function exists for ol-
factory receptors of fish, it may influence mucus secretion of
the epidermis by controlling goblet and superficial epithelial
cells. Such an association would provide a basis for a feedback
mechanism in which olfactory receptor cells could optimize
local mucus production by increasing it when harmful solutes
are present and decreasing it under normal conditions to im-
prove sensitivity to information-containing odorants.
A second potential protective system is induction of de-
toxifying mechanisms. The olfactory epithelium of a number
of fishes contains high levels of cytochrome P-450 monoox-
ygenase, which can be activated by exposure to heavy metals
[9]. Presence of this enzyme system in olfactory tissues sug-
gests the ability to eliminate or sequester toxic solutes.
A third potential protective system involves a sequestering
function of melanophores in the olfactory organ. After 60-d
One explanation for the observed response is that sensitivity
of the sensory system or receptors changed with time. If this
explanation is correct, then compensatory mechanisms induced
by initial exposure must have facilitated physiological adap-
tation and the development of tolerance within 96 h. A time-
dependent response of the olfactory system has been observed
by other investigators [24] and is consistent with the general
adaptation syndrome [25,26]. The general adaptation syn-
drome describes the response of physiological systems to
stressors as a three-phase time-dependent process that includes
an initial loss of ability caused by exposure followed by a
period of physiological adaptation that ends when compen-
sating mechanisms are unable to sustain the level of activity
required to offset effects of the stressor. Our data suggest that
the initial loss of ability occurred during the first 24 h of
exposure and was reflected by the lower response rate of fish
to skin homogenate. By 96 h, induction of protective mech-
anisms allowed recovery of olfactory ability, which increased
the frequency of detection and response to skin homogenate
exposure to copper at 20 ␮g/L, Julliard et al. [38] observed a
correspondence between regeneration of olfactory structure in
rainbow trout (Oncorhynchus mykiss) and increasing metal
concentrations in melanosomes in melanophores. They suggest
that the correspondence is evidence of a mechanism that con-
tributes to recovery of olfactory ability despite continued cop-
per exposure.
in copper concentrations
Ͻ66 ␮g/L. Fish exposed to higher
copper concentrations either were not able to compensate or
needed more time for protective mechanisms to be induced
sufficiently to offset effects of exposure.
Other data support our hypothesis that loss of olfactory
ability followed by recovery can occur within 96 h. Several
investigators have reported that olfactory receptors of fish are
Implications of this investigation
The regenerative capacity and variable sensitivity of the
olfactory system of fish is probably an evolutionary adaptation
to injury during normal life [9]. These abilities may give a
false impression that there is a margin for error when assessing
potential effects of environmental contaminants on fish. How-
ever, evidence suggests that fish that are acclimated to chemical
exposure suffer reduced olfactory ability compared with un-
exposed fish [24]. In addition, intermittent loss and recovery
of olfactory ability can potentially have a high biological cost.
Olfaction in fishes facilitates a variety of ecological interac-
tions including predator avoidance, feeding in patchy envi-
ronments, mating, and migration [39]. Although the temporal
scale of these behaviors may be short (e.g., predator avoidance
may occur in seconds), they facilitate intense interactions that
have important outcomes like survival or death.
Results of this study and available water-quality data suggest
that copper concentrations in the Yampa River may occasionally
inhibit olfactory ability of resident Colorado pikeminnow. The
Yampa River was selected for comparison in this study because
data are available that describe its water quality characteristics,
adult Colorado pikeminnow inhabit the river all year, and spawn-
ing occurs in the river from late June through early August.
However, the Yampa River is considered to be one of the more
pristine large rivers in the Colorado River Basin. Conditions in
other rivers may warrant greater concern.
There are a variety of contaminant sources in the Colorado
River Basin including natural hot springs, municipal releases,
irrigation returns, and historic mining. The potential for ex-
isting sources of contamination to result in adverse ecological
effects is compounded by demand for water in the West. As
resources are diverted for municipal, industrial, and agricul-
tural uses, less water will be available for dilution. If increases
in water use outpace the process of pollution abatement within
the Colorado River Basin, frequency and magnitude of con-
taminant-induced toxic effects will increase even if contami-
nant inputs remain constant. Geographical distribution of Col-
orado pikeminnow has been greatly reduced and reproducing
damaged or destroyed in
Ͻ24 h by exposure to copper con-
centrations ranging from 0.02 to 7,600 mg/L [27–31]. In ad-
dition, investigators have reported that regeneration of olfac-
tory cells after exposure to copper occurs within 8 d to 12
weeks [7,24,30,32] and that effects and recovery of ability are
exposure dependent [24,31]. Copper concentrations in most
studies of structural damage to olfactory cells are much higher
than those in our investigation. Consequently, other copper-
specific data documenting repair of structural damage in the
olfactory organ are not available for the concentration range
and 96-h exposure duration that we used. However, Cancalon
[
9
33] observed regeneration of olfactory receptors in fish within
6 h of exposure to low concentrations (0.03–0.1%) of a de-
tergent. Thus, it is possible that protective mechanisms in Col-
orado pikeminnow were induced and regeneration of olfactory
receptors occurred within the 96-h exposure period.
Potential protective mechanisms
There are several potential protective mechanisms that may
decrease effects of long-term exposure to copper on olfactory
receptors by sequestering, eliminating, or reducing absorption
of toxicants. First, exposure to copper has been shown to in-
crease mucus production in fish [24,34]. Mucus provides a
protective coat over olfactory sensory cells. Odorants or con-
taminants must diffuse through the mucus layer before they
can stimulate or inhibit receptor cells [35]. Increased mucus
production induced by contaminant exposure increases thick-
ness of the mucus coat [34]; consequently, time required for
toxic solutes to diffuse to olfactory receptors may increase.
Miller and Mackay [36] showed that mucus is a strong copper
chelater. Thus, in addition to the physical advantages of a
thicker mucus coat, the affinity of mucus for contaminants
may prevent diffusion to olfactory receptors, and sloughing of
excess mucus may physically remove chelated toxicants from
the olfactory chamber. Whitear [37] hypothesized that che-

9
12
Environ. Toxicol. Chem. 20, 2001
D.W. Beyers and M.S. Farmer
Data Report. U.S. Geological Survey, Water Resources Division,
Lakewood, CO.
7. Duncan AC, Ugland RC, Bennett JD, Wilson EA. 1984. Water
resources data, Colorado, water year 1984, Vol 3. CO-84-3. Water-
Data Report. U.S. Geological Survey, Water Resources Division,
Lakewood, CO.
8. Ugland RC, Duncan AC, Wilson EA, Bennett JD. 1985. Water
resources data, Colorado, water year 1985, Vol 3. CO-85-3. Water-
Data Report. U.S. Geological Survey, Water Resources Division,
Lakewood, CO.
9. Ugland RC, Cochran BJ, Kretschman RG, Wilson EA, Bennett
JD. 1989. Water resources data, Colorado, water year 1989, Vol
2. CO-89-2. Water-Data Report. U.S. Geological Survey, Water
Resources Division, Lakewood, CO.
20. Lee RE. 1993. Scanning Electron Microscopy and X-Ray Mi-
croanalysis. PTR Prentice Hall, Englewood Cliffs, NJ, USA.
21. SAS Institute. 1993. SAS௡ SAS/STAT௡ Software: GENMOD Pro-
cedure, Release 6.09. Technical Report P-243. Cary, NC, USA.
2. Ott RL. 1993. An Introduction to Statistical Methods and Data
Analysis, 4th ed. Duxbury, Belmont, CA, USA.
populations currently inhabit less than 25% of their former
range [1]. There will be no refugia for Colorado pikeminnow
if remaining habitat becomes unsuitable for any life-stage re-
quirement. Given that many human-induced changes respon-
sible for decline of Colorado pikeminnow are still present in
the Colorado River Basin, it is important to prevent further
anthropogenic degradation and be proactive about predicting
how existing conditions may change in the future.
1
1
1
Acknowledgement—We thank K.R. Bestgen for statistical advice and
encouragement. J.F. Fairchild, K.E. Klima, and three anonymous re-
viewers provided comments that improved the manuscript. P.J. Sikoski
provided laboratory assistance. This study was funded by the Recov-
ery Implementation Program for Endangered Fish Species in the Up-
per Colorado River Basin.
2
2
2
REFERENCES
3. SAS Institute. 1990. SAS/STAT௡ Users Guide, Ver 6, 4th ed. Cary,
1
. Tyus HM. 1991. Ecology and management of Colorado pike-
minnow. In Minckley WL, Deacon JE, eds, Battle Against Ex-
tinction. University of Arizona Press, Tucson, AZ, USA, pp 379–
NC, USA.
4. Saucier D, Astic L. 1995. Morpho-functional alterations in the
olfactory system of rainbow trout (Oncorhynchus mykiss) and
possible acclimation in response to long-lasting exposure to low
copper levels. Comp Biochem Physiol A 112:273–284.
4
04.
2
3
. U.S. Department of Interior. 1967. Native fish and wildlife: En-
dangered species. Fed Reg 32:4001.
. Carlson CA, Muth RT. 1989. The Colorado River: Lifeline of the
American Southwest. Can Spec Pub Fish Aquat Sci 106:220–
2
2
5. Selye H. 1973. The evolution of the stress concept. Am Sci 61:
6
92–699.
6. Beyers DW, Rice JA, Clements WH, Henry CJ. 1999. Estimating
physiological cost of chemical exposure: Integrating energetics
and stress to quantify toxic effects in fish. Can J Fish Aquat Sci
2
39.
4
5
. U.S. Fish and Wildlife Service. 1990. Colorado squawfish re-
covery plan. Final Report. Denver, CO.
. Sutterlin AM. 1974. Pollutants and the chemical senses of aquatic
animals—Perspective and review. Chem Senses Flavor 1:167–
5
6:814–822.
2
2
7. Gardner GR, LaRoche G. 1973. Copper induced lesions in es-
tuarine teleosts. J Fish Res Board Can 30:363–367.
8. Cancalon P. 1982. Degeneration and regeneration of olfactory
1
78.
6
7
8
. Brown SB, Evans RE, Thompson BE, Hara TJ. 1982. Chemo-
reception and aquatic pollutants. In Hara TJ, ed, Chemoreception
in Fishes. Elsevier, New York, NY, USA, pp 363–394.
. Hara TJ, Brown SB, Evans RE. 1983. Pollutants and chemore-
ception in aquatic organisms. In Nriagu JO, ed, Aquatic Toxi-
cology. John Wiley & Sons, New York, NY, USA, pp 247–306.
. Brand JG, Bruch RC. 1992. Molecular mechanisms of chemo-
sensory transduction: Gustation and olfaction. In Hara TJ, ed,
Fish Chemoreception. Chapman & Hall, New York, NY, USA,
pp 126–149.
cells induced by ZnSO and other chemicals. Tissue Cell 14:717–
4
7
33.
2
3
3
9. Klima KE, Applehans FM. 1990. Copper exposure and the de-
generation of olfactory receptors in rainbow trout (Oncorhynchus
mykiss). Chem Speciat Bioavailab 2:149–154.
0. Julliard AK, Saucier D, Astic L. 1996. Time-course of apoptosis
in the olfactory epithelium of rainbow trout exposed to a low
copper level. Tissue Cell 28:367–377.
1. Hansen JA, Rose JD, Jenkins RA, Gerow KG, Bergman HL. 1999.
Chinook salmon (Oncorhynchus tshawytscha) and rainbow trout
9
. Klaprat DA, Evans RE, Hara TJ. 1992. Environmental contam-
inants and chemoreception in fishes. In Hara TJ, ed, Fish Che-
moreception. Chapman & Hall, New York, NY, USA, pp 321–
(
Oncorhynchus mykiss) exposed to copper: Neurophysiological
and histological effects on the olfactory system. Environ Toxicol
Chem 18:1979–1991.
3
42.
3
3
2. Moran DT, Rowley JC, Aiken G. 1987. Trout olfactory receptors
degenerate in response to water-borne ions. In Roper SD, Atama
J, eds, Olfaction and Taste, Vol 9—Annals of New York Academy
of Science. New York, NY, USA, pp 509–511.
3. Cancalon P. 1983. Influence of a detergent on the catfish olfactory
mucosa. Tissue Cell 15:245–258.
1
0. Hara TJ. 1971. Chemoreception. In Hoar WS, Randall DJ, eds,
Fish Physiology, Vol 5. Academic, New York, NY, USA, pp 79–
1
20.
1. Smith RJF. 1992. Alarm signals in fishes. Rev Fish Biol Fish 2:
3–63.
2. American Society for Testing and Materials. 1990. Standard guide
for conducting acute toxicity tests with fishes, macroinvertebrates,
and amphibians. E729-88. In Annual Book of ASTM Standards,
Vol 11.04. Philadelphia, PA, pp 360–379.
1
1
3
34. Sorensen EMB. 1991.
ton, FL, USA.
Metal Poisoning in Fish. CRC, Boca Ra-
35. Getchell TV. 1986. Functional properties of vertebrate olfactory
receptor neurons. Physiol Rev 66:772–818.
36. Miller TG, Mackay WC. 1982. Relationship of secreted mucus
to copper and acid toxicity in rainbow trout. Bull Environ Contam
Toxicol 28:68–74.
37. Whitear M. 1992. Solitary chemosensory cells. In Hara TJ, ed,
Fish Chemoreception. Chapman & Hall, New York, NY, USA,
pp 103–125.
38. Julliard AK, Saucier D, Astic L. 1995. Metal X-ray microanalysis
in the olfactory system of rainbow trout exposed to low level of
copper. Biol Cell 83:77–86.
1
3. Hamilton SJ, Buhl KJ. 1997. Hazard assessment of inorganics,
individually and in mixtures, to two endangered fish in the San
Juan River, New Mexico. Environ Toxicol Water Qual 12:195–
2
09.
1
1
1
4. U.S. Geological Survey. 1980. Water resources data for Colorado,
water year 1980, Vol 3. CO-80-3. Water-Data Report. Water Re-
sources Division, Lakewood, CO.
5. U.S. Geological Survey. 1981. Water resources data for Colorado,
water year 1981, Vol 3. CO-81-3. Water-Data Report. Water Re-
sources Division, Lakewood, CO.
39. Hara TJ. 1986. Role of olfaction in fish behavior. In Pitcher TJ,
ed, The Behavior of Teleost Fishes. Johns Hopkins University
Press, Baltimore, MD, USA, pp 152–176.
6. Steinheimer JT, Ugland RC, Burch HE, Wilson EA. 1982. Water
resources data, Colorado, water year 1982, Vol 3. CO-82-3. Water-