1
2
3
4
Disturbance cues in freshwater prey fishes: does urea function as an ‘early warning
5
cue’ in juvenile convict cichlids and rainbow trout?
6
7
8
Grant E. Brown*, Christopher D. Jackson, Patrick H. Malka, Élisa Jacques & Marc-
9
Andre Couturier.
10
Department of Biology, Concordia University, 7141 Sherbrooke St. West, Montreal,
11
Quebec, H4B 1R6, CANADA
12
*Corresponding author: email: gbrown@alcor.concordia.ca, telephone: 514.848.2424, ext
13
4327, telefax: 514.848.2881
14
15
Urea as a potential disturbance cue in freshwater fishes. 2
16
17
Abstract:
18
Freshwater vertebrate and invertebrate prey species commonly rely on chemosensory
19
information, including non-injury released disturbance cues, to assess local predation
20
threats. We conducted laboratory studies to (1) determine if urea can function as a
21
disturbance cue in juvenile convict cichlids and rainbow trout and (2) determine if the
22
background level of urea influences the behavioural response to a subsequent pulse of
23
urea (‘background noise’ hypothesis). In the first series of trials, juvenile cichlids and
24
trout were exposed to urea at varying concentrations (0 to 0.5 mg L-1 for cichlids and 0 to
25
1.0 mg L-1 for trout). Our results suggest that both cichilds and trout exhibited
26
functionally similar responses to urea and conspecific disturbance cues and that
27
increasing the concentration of urea results in an increase intensity of antipredator
28
behaviour. In the second series of trials, we pre-exposed cichlids or trout to intermediate
29
or high concentrations of urea (or a distilled water control) and then tested for the
30
response to a second pulse of urea at at intermediate or high concentrations (versus a
31
distilled water control). Our results demonstrate that pre-exposure to urea reduces or
32
eliminates the response to a second pulse of urea, supporting the background noise
33
hypothesis. Together, our results suggest that pulses of urea, released by disturbed or
34
stressed individuals, may function as an early warning signal in freshwater prey species.
35
Keywords: Disturbance cue, metabolic wastes, chemosensory risk assessment, predator-
36
prey interactions, convict cichlids, rainbow trout.
37
38
Urea as a potential disturbance cue in freshwater fishes. 3
39
40
Introduction:
41
The ability to reliably assess local predation threats is critical for prey to balance the
42
conflicting demands of predator avoidance and energy acquisition (Lima & Dill 1990).
43
Within aquatic ecosystems, prey often rely on chemosensory cues as a source of
44
publically available information to assess the presence and level of local predation risk.
45
In addition to the well-studied damage-release chemical alarm cue system, (Chivers &
46
Smith 1998; Ferrari et al. 2010), a growing number of prey species have been shown to
47
respond to the so-called ‘disturbance cues’. Disturbance cues, unlike damage-released
48
alarm cues, do not require mechanical damage to facilitate their release. Rather,
49
disturbance cues are released, likely in the urine or across the gill epithelia, by disturbed
50
or stressed prey prior to an attack by a predator (Wisenden et al. 1995; Jordão & Volpato
51
2000). Typically, the response to disturbance cues are of a lower intensity compared to
52
the response towards other risk assessment cues (i.e. damage released alarm cues; Ferrari
53
et al. 2008a) and may increase vigilance towards secondary sources of information
54
(Wisenden et al. 1995). As such, disturbance cues are argued to serve as an ‘early
55
warning signal’ of local predation risks (Wisenden 2000).
56
Disturbance cues have been shown in a variety of invertebrate (Northern crayfish,
57
Oronectes virilis; Hazlett 1990a, hermit crabs, Calcinus laevimanus; Hazlett 1990b, red
58
sea urchins, Strongylocentrotus franciscanus; Nishizaki & Ackerman 2005) and
59
vertebrate (red-legged frogs, Rana aurora; Kiesecker et al. 1999, Iowa darters,
60
Etheostoma exile, Wisenden et al. 1995, slimy sculpins, Cottus cognatus, Bryer et al.
61
2001) prey taxa. Recent work by Vavrek and Brown (2009) has shown that the relative
Urea as a potential disturbance cue in freshwater fishes. 4
62
concentration of disturbance cue detected by prey provides sufficient information to
63
allow for threat-sensitive behavioural trade-offs in both juvenile convict cichlids
64
(Amatilania nigrofasciata) and rainbow trout (Oncorhynchus mykiss). They have shown
65
that the intensity of predator avoidance response increases with the concentration of
66
disturbance cue detected, consistent with a graded threat-sensitive response. In addition,
67
juvenile rainbow trout (Ferrari et al. 2008) and Iberian green frog tadpoles (Gonzalo et al.
68
2010) exhibit an additive response when exposed to conspecific disturbance cues and
69
damage-released alarm cues. Increased predator avoidance and vigilance behaviour in
70
response to damage-released alarm cues and disturbance cues, even at low relative
71
concentrations (Mirza & Chivers 2003), have been shown to increase survival during
72
encounters with live predators (i.e. Mathis & Smith 1993; Mirza & Chivers 2002; Leduc
73
et al. 2009). Together, these studies suggest that disturbance cues provide a valuable
74
source of risk assessment information to aquatic prey.
75
Disturbance cues have been argued to consist of nitrogenous wastes, released as
76
metabolic byproducts through the urine or across the gill epithelia (Hazlett 1990a, 1990b;
77
Kiesecker et al. 1999). As such, they should elicit a generalized, non-species specific,
78
response. The findings of Vavrek et al. (2008) which show that two phylogentically
79
distant prey species (convict cichlids and rainbow trout) respond to each others
80
disturbance cues with similar intensities supports the supposition that disturbance cues
81
are a generalized metabolic waste product. In amphibians (Kiesecker et al. 1999) and
82
crayfish (Hazlett 1990a), ammonia has been shown to function as a disturbance cue.
83
However, Vavrek et al. (2008) have demonstrated that while cichlids and rainbow trout
Urea as a potential disturbance cue in freshwater fishes. 5
84
produce and respond to a generalized disturbance cue, ammonia in ecologically relevant
85
concentrations does not elicit any increase in predator avoidance behaviour.
86
Thus, if not ammonia, then what may function as a non-specific disturbance cue
87
in freshwater prey fishes? One possible candidate could be urea. While most freshwater
88
fishes excrete the majority of nitrogenous wastes as ammonia, approximately 10 to 20%
89
of nitrogen wastes may be released as urea (Smutná et al., 2002; Wilkie 2002, Altinok &
90
Grizzle 2004, Weihrauch et al. 2009). Moreover, urea is typically released in pulses
91
(Wilkie 2002; Altinock & Grizzle 2004) rather than continuously, making it a suitable
92
candidate as an early warning signal of local predation threats.
93
Damage-released cues are said to be honest indicators of risk because they are
94
only released when the prey’s skin is mechanically damaged (Chivers & Smith 1998;
95
Chivers et al. 2007). However, given the nature of their release, disturbance cues may be
96
of lower risk assessment value, possibly due to the presence of ambient levels of
97
nitrogenous waste products. Metabolic byproducts that may function as disturbance cues
98
would be continually released, either directly prey species or indirectly through
99
decomposition, within microhabitats at low levels. As a result, under ecologically
100
realistic conditions, there should exist some level of ‘background noise’ that may limit
101
the value of pulses of disturbance cues as a source of risk assessment information. The
102
background noise hypothesis (Vavrek & Brown 2009) predicts that only when a
103
disturbance cue above the background levels of nitrogenous wastes (‘noise’), will prey
104
exhibit a behavioural response.
Urea as a potential disturbance cue in freshwater fishes. 6
105
Thus, our aim here is two fold. Initially, we test the prediction that urea elicits
106
short-term increases in antipredator behaviour, consistent with the response to naturally
107
occurring disturbance cues in two taxonomically diverse prey fishes. Juvenile convict
108
cichlids and rainbow trout were exposed to varying concentrations of urea to test if the
109
response pattern follows a threat-sensitive response (Vavrek & Brown 2009). We predict
110
that if urea does serve as an active component of the natural disturbance cue, we should
111
see a response intensity proportional to the concentration detected by focal fish.
112
Secondly, we test the prediction that background levels of urea will inhibit the presence
113
and/or intensity of behavioural responses to pulses of urea at varying concentrations.
114
Juvenile cichlids and trout were pre-exposed to three levels of urea and then tested for the
115
behavioural response to low, intermediate and high levels of urea. Given that urea is
116
relatively stable in freshwater (Park et al. 1997), we predict that as the background of
117
urea increases, the response to a secondary pulse of urea should decrease.
118
119
General methods:
120
Test fish
121
Juvenile convict cichlids originated from our laboratory stock population. These fish
122
were descendants from laboratory crosses made approximately four generations
123
previously from laboratory stock and wild caught cichlids from Costa Rica. Prior to
124
testing, cichlids were held in 110 L glass aquaria filled with continuously filtered
125
dechlorinated water. Tanks were held at approximately 26-27°C under a 12-12 hour
126
light-dark cycle. Holding tanks contained a gravel substrate and artificial vegetation.
Urea as a potential disturbance cue in freshwater fishes. 7
127
Juvenile cichlids were fed, twice daily, with commercial flake food (Nutrafin) and brine
128
shrimp (Artemia spp.). Juvenile rainbow trout were purchased from a commercial
129
hatchery (Pisciculture des Arpents Verts, Ste. Edwidge de Clifton, Quebec). Prior to
130
testing, trout were housed in 390 L recirculating tanks, with a continuous supply of
131
dechlorinated water (~750 mL min-1). Trout were housed at ~ 18°C under a 12-12 light-
132
dark cycle. Trout were fed, twice daily, with commercial trout chow (Corey Mills).
133
134
Stimulus preparation
135
We generated disturbance cues and the odour of undisturbed donors for use in
136
Experiment 1. Shoals of ten donor cichlids or trout were placed in a 20 L glass aquaria
137
(filled with 15 L of dechlorinated water) and allowed a 24 hour acclimation period prior
138
to cue collection. Cichlid donor tanks were held at ~27°C, contained a gravel substrate
139
and a charcoal filter. Trout donor tanks were held at ~18° C, contained a gravel substrate
140
and a charcoal filter. One hour prior to cue collection, we turned off the charcoal filter.
141
To collect the odour of undisturbed donors, we simply siphoned ~ 200 mL of tank water.
142
In order to collect disturbance cues, we passed a realistic model predator (15 cm) attached
143
to a glass rod through the tank 20 times. We were careful to avoid contact with the
144
donors to minimize the chance of introducing damage-released alarm cues into the tank
145
water. The model predator had an elongate body shape, typical of ambush predators and
146
was realistically painted to resemble a common Esocid predator. Mean (± SD) standard
147
length of donors was 3.82 ± 0.65 cm (cichlids) and 5.58 ± 0.72 cm (trout). This
148
procedure has successfully been used to collect disturbance cues and the odour of
Urea as a potential disturbance cue in freshwater fishes. 8
149
undisturbed donors from both cichlids and trout (Vavrek et al. 2008; Vavrek & Brown
150
2009). Urea solutions, used in Experiment 1 and 2, were generated using reagent grade
151
urea (CH4N2O, 99.5% purity, 60.06 M). We created fresh stock solutions daily (3.5 g L-
152
1
) and diluted the stock solution to the desired concentration in distilled water.
153
154
Experiment 1: Response to urea as a potential disturbance cue.
155
Here, we exposed juvenile convict cichlids and rainbow to urea at varying concentrations
156
to test the hypothesis that urea may function as a disturbance cue.
157
Test tanks used for the cichlid trials consisted of a series of 37 L glass aquaria,
158
filled with exactly 35 L of aged, dechlorinated tap water. Each test tank contained a
159
gravel substrate, and was held at ~ 26° C (cichlids) under a 12-12 light-dark cycle. Test
160
tanks were not filtered. We affixed an air stone to the back wall of the tank. In addition,
161
we attached an additional 1.5 m length of airline tubing, terminating immediately above
162
the air stone, to allow for the introduction of chemical stimuli without disrupting the test
163
fish. Tanks were wrapped on three sides with black plastic to eliminate visual
164
communication between tanks. Test tanks used for trout trials were the same as described
165
for cichlids except they were held at ~18° C.
166
We placed focal fish into test tanks and allowed a 24 hour acclimation period
167
prior to testing. Cichlids were tested in pairs as singleton cichlids are generally inactive
168
(Brown et al. 2006). Trout were tested individually. Both cichlids and trout were fed ad
169
libitum up to 1 hour before testing. Mean (± SD) standard length at time of testing was
170
2.42 ± 0.58 mm (cichlids) and 4.62 ± 0.33 cm (trout).
Urea as a potential disturbance cue in freshwater fishes. 9
171
Behavioural trials consisted of a 5 minute pre-stimulus and a 5 minute post-
172
stimulus observation period. Prior to beginning a trial, we withdrew and discarded 60
173
mL of tank water from the stimulus injection tube. We then withdrew and retained an
174
additional 60 mL. Immediately following the pre-stimulus observation period, we
175
injected 10 mL of stimuli (see below) and slowly flushed it into the tank using the
176
retained 60 mL of water. Once the stimulus was fully injected, we began the post-
177
stimulus observation trial. During both pre- and post-stimulus injection observations, we
178
quantified the time spent moving and frequency of foraging attempts. Pairs of cichlids
179
typically moved together within the test tank, so we arbitrarily followed one of the two in
180
order to quantify time moving. For cichlids, foraging was defined as pecking at the
181
substrate, with the body at an angle greater than 45° relative to the substrate (Grant et al.,
182
2002). For trout, foraging behaviour was defined as any visible snapping movement
183
directed towards the substrate or within the water column (Vavrek et al. 2008). Given
184
that focal fish were fed ad libitum prior to testing, there was sufficient food remaining to
185
allow us to quantify foraging without the addition of a food stimulus (Vavrek et al. 2008).
186
In the first series of trials, we exposed cichlids and trout to either conspecific
187
disturbance cues or the odour of undisturbed conspecifics to provide confirmation that the
188
test fish do indeed respond to conspecific disturbance cues. In the second series of trials,
189
we exposed cichlids to dilute urea such that the final concentration within test tanks (35
190
L) was 0.01. 0.1 or 0.5 mg L-1 or a distilled water control. Trout were likewise exposed
191
to urea at concentrations of 0.1, 0.5 and 1.0 mg L-1, and a distilled water control. The
192
range of concentrations were higher for trout as pilot studies revealed that they did not
193
respond to urea at very low concentrations (mean (± SE) change in time moving = 5.93 ±
Urea as a potential disturbance cue in freshwater fishes. 10
194
9.44 vs. 4.40 ± 4.31 sec (0.01 mg L-1 and DW respectively), t28 = 0.148, P = 0.88; mean
195
(± SE) change in foraging attempts = 1.20 ± 0.97 vs. 0.80 ± 1.07 (0.01 mg L-1 and DW
196
respectively) t28 = 0.276, P = 0.79). This range of concentrations was chosen because
197
they were similar to the range use by Vavrek & Brown (2009) and they represent the low
198
to intermediate range of urea previously reported for eutrophic freshwater systems (Park
199
et al. 1997; Savin et al. 2007). New urea solutions were generated immediately prior to
200
testing. The order of trials was randomized and observers were blind to treatment. For
201
the first series of trials (natural disturbance cues), we conducted 10 (cichlids) or 12 (trout)
202
replicates per treatment. For the second series (urea) we conducted 15 replicates for each
203
treatment for both cichlids and trout.
204
For time spent moving and foraging attempts, we calculated the change between
205
the pre-stimulus and post-stimulus observation periods (post – pre) and used these
206
difference scores as dependent variables in all analyses. As the change in time spent
207
moving and foraging attempts are likely highly correlated, we employed a multivariate
208
approach. Initially, we compared the change in time spent moving and foraging attempts
209
in response to disturbance cues vs. the odour of undisturbed conspecifics using a
210
multivariate ANOVA. We similarly tested for the change in response to varying
211
concentrations of urea using a multivariate ANOVA. We used a priori polynomial
212
contrasts to test the prediction that the intensity of antipredator response should increase
213
proportional to the concentration of urea (Vavrek & Brown 2009). All data met the
214
assumptions of parametric tests (Levene’s test, P > 0.05 for all).
215
Urea as a potential disturbance cue in freshwater fishes. 11
216
Experiment 2: Background noise hypothesis.
217
Here, we test the hypothesis that the background level of nitrogenous waste (urea) will
218
influence the response of prey fishes to pulses of urea. We pre-exposed cichlids and trout
219
to varying concentrations of urea and then tested their behavioural response to urea as a
220
proxy for conspecific disturbance cues.
221
Test fish were placed into test tanks as described above for Experiment 1.
222
Behavioural observations consisted of a pre-exposure phase, followed 30 minutes later by
223
a 10 minute observation phase. The 30 minute period between the pre-exposure and
224
observation phases was sufficiently long to ensure that cichlids and trout had returned to
225
baseline activity levels (i.e. were not exhibiting additive responses to multiple exposures
226
of urea). During the pre-exposure phase, we injected dilute urea such that the final
227
concentration was 0 (distilled water), 0.1 or 0.5 mg L-1 for cichlids and 0 (distilled water),
228
0.5 and 1.0 mg L-1 for trout, as described for Experiment 1. Thirty minutes following the
229
introduction of the pre-exposure cue, we began the behavioural observation. Trials
230
consisted of a five minute pre-stimulus and a five minute post-stimulus observation
231
period as described for Experiment 1. Immediately following the pre-stimulus
232
observation period, we introduced a distilled water control or an additional pulse of urea
233
at a concentration of 0.1 or 0.5 mg L-1 for cichlids or 0.5 or 1.0 mg L-1 for trout (as
234
described in Experiment 1). Thus, we have three levels of pre-exposure fully crossed
235
with three experimental stimuli. Behavioural measures were as described for Experiment
236
1. All treatments were randomized and observers were blind to treatment. Mean (± SD)
237
standard length at time of testing was 2.30 ± 0.41 cm (cichlids) and 4.93 ± 0.45 cm (trout).
238
For both cichlids and trout, we conducted 10 replicates per treatment combination.
Urea as a potential disturbance cue in freshwater fishes. 12
239
Difference scores for time spent moving and foraging attempts were calculated as in
240
Experiment 1. We tested for the effects of pre-exposure and experimental stimulus using
241
a multivariate ANOVA. All data met the assumptions of parametric tests (Levene’s test,
242
P > 0.05 for all).
243
244
Results
245
Experiment 1: Response to urea as a potential disturbance cue.
246
Juvenile cichlids exhibited a significant increase in antipredator behaviour (reduced time
247
moving and foraging attempts) when exposed to conspecific disturbance cues versus the
248
odour of undisturbed conspecifics (F2, 27 = 4.11, P = 0.028, Figure 1). Likewise, we
249
found a similar effect of urea concentration on the intensity of antipredator responses (F3,
250
56
251
univariate effects of urea concentration for both the change in time moving (F3, 56 = 3.10,
252
P = 0.034, Figure 1A) and foraging attempts (F3, 56 = 3.49, P =0.021, Figure 1B). The
253
response pattern for both behavioural measures is best described as linear versus
254
quadratic contrasts (Table 1; Figure 1), suggesting that antipredator response with
255
increasing urea concentrations.
= 8.25, P < 0.001, Figure 1). Our planned contrasts analysis revealed significant
256
We found a similar response pattern for juvenile rainbow trout. Trout exhibited a
257
significant increase in antipredator behaviour in response to conspecific disturbance cues
258
(F2, 21 = 3.72, P = 0.041, Figure 2). We also found a significant effect of urea
259
concentration on the antipredator response of trout (F3, 56 = 8.59, P < 0.001, Figure 2). As
260
with convict cichlids, our planned contrasts revealed significant univariate effects of urea
Urea as a potential disturbance cue in freshwater fishes. 13
261
for both the change in time moving (F3, 56 = 5.79, P = 0.002, Figure 2A) and foraging
262
attempts (F3, 56 = 3.01, P = 0.038, Figure 2B). The observed response for time moving
263
and foraging attempts is best described as linear (Table 1, Figure 2).
264
265
Experiment 2: Background noise hypothesis.
266
For convict cichlids, we found significant main effects of pre-exposure (F2, 81 = 4.69, P =
267
0.012) and test stimulus (F2, 81 = 23.53, P < 0.001) on the observed intensity of
268
antipredator response. We found no significant pre-exposure by test stimulus interaction
269
(F4, 81 = 1.91, P = 0.12). Overall, when pre-exposed to distilled water, cichlids showed a
270
strong response to urea at concentrations of 0.1 and 0.5 mg L-1. When pre-exposed to
271
urea at 0.5 mg L-1, cichlids showed no response to urea at concentrations of 0.1 mg L-1
272
(response was similar to the distilled water control group), and a weaker response to 0.5
273
mg L-1 of urea (Figure 3). Interestingly, when pre-exposed to 0.1 mg L-1 of urea, cichlids
274
exhibited a strong response to 0.5 mg L-1 of urea, and a weaker response to 0.1 mg L-1
275
(Figure 3).
276
Rainbow trout exhibited a similar response pattern. We found significant main
277
effects of both pre-exposure treatments (F2, 81 = 4.18, P = 0.019) and test stimulus (F2, 81 =
278
3.66, P = 0.03). There was no interaction between the two main effects (F4, 81 = 2.35, P =
279
0.06). As with cichlids, the overall intensity of response decrease with increased pre-
280
exposure urea concentrations. Trout pre-exposed to distilled water showed reduced time
281
moving and foraging attempts when tested with urea at concentrations of 0.5 and 1.0 mg
282
L-1 and no change (compared to the distilled water control) when pre-exposed to urea at
Urea as a potential disturbance cue in freshwater fishes. 14
283
1.0 mg L-1. When pre-exposed to 0.5 mg L-1 of urea, only trout exposed to the highest
284
concentrations exhibited a reduction in time moving and foraging attempts (Figure 4).
285
286
Discussion:
287
Our results suggest that pulses of urea may indeed function as a disturbance cue in
288
freshwater fishes. Both cichlids and trout exhibited significant reductions in time moving
289
and foraging rate when exposed to urea. Moreover, the intensity of these responses
290
appear similar to that elicited by conspecific disturbance cues. The results of our first
291
experiment suggest that the intensity of the predator avoidance response is proportional to
292
the concentration detected, consistent with the predictions of the threat-sensitivity model
293
(Helfman 1989). These results are consistent with previous studies by Vavrek & Brown
294
(2009), which demonstrated a graded threat-sensitive response pattern for juvenile
295
convict cichlids and rainbow trout exposed to conspecific disturbance cues. Similar
296
graded threat-sensitive response patterns have been shown for prey fishes responding to
297
damage-released chemical alarm cues (Brown et al. 2006) and visual cues (Helfman &
298
Winkleman 1997).
299
The results of our second experiment provide support for the ‘background noise
300
hypothesis’. Unlike the damage-released cues, there is likely always some background
301
level of ‘disturbance’ cues brought about by passive release or stress associated with
302
competitive interactions among conspecifics and/or heterospecifics. This may be
303
especially true for shoaling species or populations utilizing slow moving and/or littoral
304
habitats. Our data suggest that pre-exposure to urea results in the reduction or
Urea as a potential disturbance cue in freshwater fishes. 15
305
elimination of response to subsequent pulses of urea. For cichlids, pre-expose to 0.1 mg
306
L-1 of urea resulted in weaker responses to a 0.1 mg L-1 pulse of urea detected 30 minutes
307
following pre-exposure. Likewise, pre-exposure to 0.5 mg L-1 urea eliminated the
308
response to a 0.1 mg L-1 pulse and inhibited the response to a 0.5 mg L-1 pulse of urea.
309
The results of our rainbow trout trials follow a similar pattern. Trout pre-exposed to 0.5
310
mg L-1 of urea only increased their antipredator behaviour towards the highest
311
concentration of urea, while pre-exposure to 1.0 mg L-1 of urea eliminated the response to
312
all concentrations. Thus, both cichlids and trout responded to a pulse of urea only if it
313
exceeded the background (pre-exposure) levels.
314
Two possible, non-mutually exclusive mechanisms may account for our observed
315
results. Initially, the initial pre-exposure to urea may have altered the behavioural
316
decisions of focal fish following the second pulse of urea. For example, test fish pre-
317
exposed to the low dose (0.1 mg L-1 for cichlids or 0.5 mg L-1 for trout) would have likely
318
shown a short-term increase in predator avoidance. Since this initial response was not
319
paired with an actual attack, subsequent exposures at the same concentration may
320
increased vigilance towards a predator but not trigger an overt behavioural response.
321
Effectively, the prey would attend to the cue, but devalue the potential predator avoidance
322
benefits in favour of continued foraging (Foam et al. 2005; Brown et al. 2006).
323
Alternatively, the pre-exposure treatment may have increased the ambient urea
324
concentration, effectively masking the second pulse of urea. In the absence of primary
325
producers and/or bacteria, urea is highly stable in aquatic ecosystems (Park et al. 1997).
326
Given that tanks were cleaned and rinsed between trials, it is unlikely that significant
327
consumers (i.e. algae or bacteria) would be present in the test tanks. Regardless of the
Urea as a potential disturbance cue in freshwater fishes. 16
328
mechanism, our results support the hypothesis that background levels of nitrogenous
329
wastes influence the response to disturbance cues.
330
The prevailing hypothesis is that metabolic wastes released by disturbed or
331
stressed conspecifics (or heterospecifics) function as a reliable early-warning signal of
332
local predation risks. While ammonia has been shown to be a key component of the
333
disturbance cue system in some freshwater prey taxa (amphibians, Kiesecker et al. 1999;
334
invertebrates, Hazlett 1990a, 1990b), Vavrek & Brown (2008) have conclusively shown
335
that ammonia does not function as the disturbance cue in at least two phylogenetically
336
distant freshwater prey fishes. Why then should urea serve as a reliable indicator of risk
337
for freshwater prey fishes? Contrary to conventional wisdom, between 10 and 20% of
338
nitrogenous wastes are released as urea in freshwater fishes (Wilkie 2002; Weirhrauch et
339
al. 2009). Moreover, as opposed to ammonia, urea is released in pulses, either in urine or
340
across the gill epithelia (Wilkie 2002). Though not directly tested, these pulses could be
341
released in response to disturbance or stress. Recent studies have demonstrated that prey
342
fish increase their opercular beat rates upon detecting a predation threat (Barreto et al.
343
2003; Brown et al. 2005; Gibson & Mathis 2006). Either mechanism (or likely a
344
combination of the two) would result in a short-term increase in local urea concentration.
345
If urea pulses are more common among stressed or disturbed fishes (Walsh et al. 1990;
346
Wood et al. 1995), a sudden local increase in urea may be a reliable indicator of local risk.
347
Prey individuals may release a pulse of urea upon detecting a potential predator,
348
providing relevant information to neighboring conspecifics and heterospecifics.
349
350
The most parsimonious explanation for evolution of the predator avoidance
function for the damage-released chemical alarm cue system in aquatic prey is the
Urea as a potential disturbance cue in freshwater fishes. 17
351
adaptive value to cue receivers (Chivers et al. 2007; Ferrari et al. 2010). The detection of
352
a damage-released alarm cue is known to be a reliable indicator of local threat (Brown
353
2003; Ferrari et al. 2010) and responding to these cues can result in increased survival
354
benefits to cue receivers (Ferrari et al. 2010). The evolution of disturbance cue function
355
likely follows a similar adaptive mechanism. For chemosensory cues to provide reliable
356
information, they must be temporally linked to the presence of an ecologically relevant
357
predation threat. Given the mechanism of release, disturbance cues are likely less reliable
358
indicators of local risk than the damage-released chemical alarm cues within freshwater
359
prey (Chivers & Smith 1998). Consistent with this prediction, the intensity of the
360
behavioural response to disturbance cues is typically lower than seen in response to
361
damage-release alarm cues (Ferrari et al. 2008a; Gonzalo et al. 2010). While damage-
362
released chemical alarm cues are only released following mechanical damage to the skin
363
(Chivers & Smith 1989), as would occur during a predator attack, disturbance cues do not
364
require actual physical contact for release. As a result, the damage-released are more
365
tightly linked to the presence of an actual predation risk than are the disturbance cues.
366
Similarly, while damage released alarm cues likely degrade rapidly (Brown et al. 2001;
367
Ferrari et al. 2008b), disturbance cues may persist within the microhabitat, depending
368
upon local consumer communities and/or ambient temperature. The results of our second
369
experiment suggest that disturbance cues would elicit a response only when detected
370
above the background levels. Thus, while disturbance cues are capable of eliciting short-
371
term predator avoidance responses in aquatic prey, the relative benefit associated with
372
this response should be less than those associated with responding to a damage-released
Urea as a potential disturbance cue in freshwater fishes. 18
373
alarm cue. As such, a threat-sensitive trade-off model would predict the observed weaker
374
response pattern (Vavrek & Brown 2009).
375
Unlike damage-released chemical alarm cues, which show a considerable degree
376
of species specificity (Mirza & Chivers 2001; Brown et al. 2003; Kelly et al. 2006) and
377
even population specificity (Brown & Godin 1999; Brown et al. 2010), disturbance cues
378
appear to be non-specific. Vavrek et al. (2008) demonstrated that convict cichlids and
379
rainbow trout exhibit antipredator responses of similar intensities to the disturbance cues
380
of both conspecific and heterospecific donors. Likewise, Hazlett (1989; 1990a) have
381
shown similar generalized responses in crayfish. This is consistent with the hypothesis
382
that the disturbance cue originates as a generalized metabolic waste compound. While
383
our results do suggest that urea may function as an active component of the disturbance
384
cue system in freshwater prey fishes, we cannot eliminate other compounds from serving
385
similar functions. Metabolites of stress hormones such as cortisol (Olivetto et al. 2002;
386
Jordão 2004) may also be released by disturbed individuals. Olivetto et al. (2002)
387
demonstrated water from holding tanks of sea bream (Diplodus sargus) held at high
388
densities contained roughly five times the level cortisol compared to low density tanks.
389
When sea bream held at low densities were exposed to water from the high density tanks,
390
they exhibited a similar increase in cortisol levels. These results suggest that cortisol may
391
also play a role as an active component of the disturbance cue system. It is likely that
392
some combination of urea and stress hormones, released into the water column, may
393
function as a disturbance cue in freshwater prey fishes.
394
Urea as a potential disturbance cue in freshwater fishes. 19
395
Acknowledgments:
396
We with to thank Drs. James Grant and Chris Elvidge for helpful comments on earlier
397
versions of this manuscript and Dr. Bill Zerges for access to his supply of urea. All work
398
reported herein was conducted in accordance with Concordia University Animal
399
Research Ethics protocol #AREC-2008-BROW. Financial support was provided by
400
Concordia University and the Natural Sciences and Engineering Research Council of
401
Canada.
402
403
References:
404
Altinok I, Grizzle JM, 2004. Excretion of ammonia and urea by phlogenetically diverse
405
406
407
fish species in low salinities. Aquaculture 238: 499-507.
Barreto RE, Luchiari AC, Marcondes AL, 2003. Ventilatory frequency indicates visual
recognition of an allopatric predator in naïve Nile tilapia. Behav. Proc. 60: 235-239.
408
Brown C, Gardner C, Braithwaite VA, 2005. Differential stress responses in fish from
409
areas of high- and low- predation pressure. J. Comp. Physiol. B 175: 305-312.
410
411
412
413
414
415
Brown GE, 2003. Learning about danger: chemical alarm cues and local risk assessment
in prey fishes. Fish Fish. 4: 227-234.
Brown GE, Godin J-GJ, 1999. Chemical alarm signals in wild Trinidadian guppies
(Poecilia reticulata). Can. J. Zool. 778: 562-570.
Brown GE, Adrian JC Jr, Naderi NT, Harvey MC, Kelly JM, 2003. Nitrogen oxides elicit
antipredator responses in juvenile channel catfish, but not in convict cichlids or
Urea as a potential disturbance cue in freshwater fishes. 20
416
rainbow trout: conservation of the Ostariophysan alarm pheromone. J. Chem. Ecol.
417
29: 1781-1796.
418
Brown GE, Bongiorno T, DiCapua DM, Ivan LI, Roh E, 2006. Effects of group size on
419
the threat-sensitive response to varying concentrations of chemical alarm cues by
420
juvenile convict cichlids. Can. J. Zool. 84: 1-8.
421
Brown GE, Elvidge CK, Macnaughton CJ, Ramnarine I, Godin J-GJ, 2010. Cross-
422
population responses to conspecific chemical alarm cues in wild Trinidadian guppies,
423
Poecilia reticulata: evidence for local conservation of cue production. Can. J. Zool.
424
88: 139-147.
425
Brown GE, Adrian JC Jr, Erickson J, Kaufman IH, Gershaneck D, 2001. Responses to
426
nitrogen-oxides by Chariciforme fishes suggest evolutionary conservation in
427
Ostariophysan alarm pheromones. In: Chemical Signals in Vertebrates, Vol. 9 (A.
428
Marchlewska-Koj, J.J. Lepri and D. Müller-Schwarze, eds). Plenum Press, New York,
429
pp. 305-312.
430
Bryer PJ, Mirza RS, Chivers DP, 2001. Chemosensory assessment of predation risk by
431
slimy sculpins (Cottus cognatus): responses to alarm, disturbance, and predator cues.
432
J. Chem. Ecol. 27: 533-546.
433
434
435
436
Chivers DP, Smith RJF, 1998. Chemical alarm signaling in aquatic predator-prey
systems: a review and prospectus. Ecoscience 5: 338-352.
Chivers DP, Wisenden BD, Hindman CJ, Michalak TA, Kusch RC, Kaminskyj SGW,
Jack KL, Ferrari MCO, Pollock RJ, Halbgewachs CF, Pollock MS, Alemadi S, James
Urea as a potential disturbance cue in freshwater fishes. 21
437
CT, Savaloja RK, Goater CP, Corwin A, Mirza RS, Kiesecker JM, Brown GE, Adrian
438
JC Jr, Krone PH, Blaustein AR, Mathis A, 2007. Epidermal ‘alarm substance’ cells of
439
fishes maintained by non-alarm functions: possible defence against pathogens,
440
parasites and UVB radiation. Proc. R. Soc. Lond. B 274: 2611-2619.
441
Ferrari MCO, Vavrek MA, Elvidge CK, Fridman B, Chivers DP, Brown GE, 2008a.
442
Sensory complementation and the acquisition of predator recognition by salmonid
443
fishes. Behav. Ecol. Sociobiol. 63: 113-121.
444
Ferrari MCO, Messier F, Chivers DP, 2008b. Degradation of chemical alarm cues under
445
natural conditions: risk assessment by larval woodfrogs. Chemoecol. 17: 263-266.
446
Ferrari MCO, Wisenden BD, Chivers DP, 2010. Chemical ecology of predator-prey
447
interactions in aquatic ecosystems: a review and prospectus. Can. J. Zool. 88: 698-
448
724.
449
Foam PE, Harvey MC, Mirza RS, Brown GE, 2005. Heads up: juvenile convict cichlids
450
switch to threat-sensitive foraging tactics based on chemosensory information. Anim.
451
Behav. 70: 601-607.
452
Gibson AK, Mathis A, 2006. Opercular beat rate for rainbow darthers Etheostoma
453
caeruleum exposed to chemical stimuli from conspecific and heterospecific fishes. J.
454
Fish Biol. 69: 224-232.
455
Gonzalo A, López P, Martin J, 2010. Risk level of chemical dues determines retention of
456
recognition of new predators in Iberian green frog tadpoles. Behav. Ecol. Sociobiol.
457
64: 1117-1123.
Urea as a potential disturbance cue in freshwater fishes. 22
458
Grant JWA, Girard IL, Breau C, Weir LK, 2002. Influence of food abundance on
459
competitive aggression in juvenile convict cichlids. Anim. Behav. 63: 323-330.
460
Hazlett BA, 1989. Additional sources of disturbance pheromone affecting the crayfish
461
462
463
464
465
466
467
468
469
470
471
472
473
474
Orconectes virilis. J. Chem. Ecol. 15: 381-385.
Hazlett BA, 1990a. Source and nature of disturbance-chemical system in crayfish. J.
Chem. Ecol. 16: 2263-2275.
Hazlett BA, 1990b. Disturbance pheromone in the hermit crab Calcinus laevimanus
(Randall, 1840). Crustaceana 58: 314-316.
Helfman GS, 1989. Threat-sensitive predator avoidance in damselfish-trumpetfish
interactions. Behav. Ecol. Sociobiol. 24: 47-58.
Helfman GS, Winkleman DL, 1997. Threat sensitivity in bicolor damselfish: effects of
sociality and body size. Ethology 103: 369-383.
Jordão LC, 2004. Disturbance chemical cues determine changes in spatial occupation by
the convict cichlid Archocentrus nigrofasciatus. Behav. Proc. 67: 453-459.
Jordão LC, Volpato GL, 2000. Chemical transfer of warning information in non-injured
fish. Behaviour, 137: 681-690.
Kelly JM, Adrian JC Jr, Brown GE, 2006. Can the ratio of aromatic skeletons explain
475
cross-species responses within evolutionarily conserved Ostariophysan alarm cues?:
476
testing the purine-ratio hypothesis. Chemoecol, 16: 93-96.
Urea as a potential disturbance cue in freshwater fishes. 23
477
Kiesecker JM, Chivers DP, Marco A, Quilchano C, Anderson MT, Blaustein AR, 1999.
478
Identification of a disturbance signal in larval red-legged frogs, Rana aurora. Anim.
479
Behav. 57: 1295-1300.
480
Leduc AOHC, Roh E, Brown GE, 2009. Effects of acid rainfall on juvenile Atlantic
481
salmon (Salmo salar) antipredator behaviour: loss of chemical alarm funcdtion and
482
potential survival consequences during predation. Mar. Fresh. Res. 60: 1223-1230.
483
Lima SL, Dill LM, 1990. Behavioral decisions made under the risk of predation: a review
484
and prospectus. Can. J. Zool. 68: 619-640.
485
Mathis A, Smith RJF 1993. Chemical alarm signals increase the survival time of fathead
486
minnows (Pimephales promelas) during encounters with northern pike (Esox lucius).
487
Behav. Ecol. 4: 260-265.
488
489
490
Mirza RS, Chivers DP, 2001. Are chemical alarm cues conserved within salmonid fishes?
J. Chem. Ecol. 27: 1641-1655.
Mirza RS, Chivers DP, 2003. Response of juvenile rainbow trout to varying
491
concentrations of chemical alarm cues: response thresholds and survival during
492
encounters with predators. Can. J. Zool. 81: 88-95.
493
Mirza RS, Chivers DP, 2002. Behavioural responses to conspecific disturbance chemicals
494
enhance survival of juvenile brook charr, Salvelinus fontinalis, during encounters
495
with predators. Behaviour 139: 1099-1109.
496
497
Nishizaki MT, Ackerman JD, 2005. A secondary chemical cue facilitates juvenile-adult
postsettlement associations in red sea urchins. Limnol. Ocean. 50: 354-362.
Urea as a potential disturbance cue in freshwater fishes. 24
498
Olivetto I, Mosconi G, Maradonna F, Cardinali M, Carnevali O, 2002. Diplodus sargus
499
interregnal-pituitary response: chemical communication in stressed fish. Gen. Comp.
500
Endocrinology 127: 66-70.
501
502
503
Park MG, Shim JH, Cho BC, 1997. Urea decomposition activities in an ammonium
enriched freshwater pond. Aquat. Microbial. Ecol. 13: 303-311.
Savin MC, Daigh AL, Tomlinson PJ, Norman RJ, 2007. Urea persistence in ponded
504
surface water and a silt-loam soil used for rice production. In: RJ Norman, J-F
505
Meullenet & KAK Moldenhauer (eds). BR Wells Rice Research Studies, pp 202-208.
506
University of Arkansas Agricultural Experimental Research Series; Fayetteville,
507
U.S.A.
508
509
510
511
512
Smutná M, Vorlová L, Svobodová Z, 2002. Pathobiochemistry of ammonia in the
internal environment of fish. Acta Vet. Brno. 71: 169-181.
Vavrek MA, Brown GE, 2009. Threat-sensitive responses to disturbance cues in juvenile
convict cichlids and rainbow trout. Ann. Zool. Fennici 46: 171-180.
Vavrek MA, Elvidge CK, DeCaire R, Belland B, Jackson CD, Brown GE, 2008.
513
Disturbance cues in freshwater prey fishes: do juvenile convict cichlids and rainbow
514
trout respond to ammonium as an ‘early warning’ signal? Chemoecol. 18: 255-261.
515
516
517
518
Walsh PJ, Danulat E, Mommsen TP, 1990. Variation in urea excretion in the gulf
toadfish, Opsanus beta. Mar. Biol. 106: 323-328.
Weihrauch D, Wilkie MP, Walsh PJ, 2009. Ammonia and urea transporters in gills of
fish and aquatic crustaceans. J. Exp. Biol. 212: 1716-1730.
Urea as a potential disturbance cue in freshwater fishes. 25
519
520
521
Wilkie MP, 2002. Ammonia excretion and urea handling by fish gills: present
understanding and future research challenges. J. Exp. Zool. 293: 284-301.
Wisenden BD, Chivers DP, Smith RJF, 1995. Early warning in the predation sequences: a
522
disturbance pheromone in Iowa darters (Etheostoma exile). J. Chem. Ecol. 21: 1469-
523
1480.
524
525
526
Wisenden BD, 2000. Olfactory assessment of predation risk in the aquatic environment.
Phil. Trans. R. Soc. Lond. B 355: 1205-1208.
Wood CM, Pärt P, Wright PA, 1995. Ammonia and urea metablolism in relation to gill
527
function and acid-base balance in a marine elasmobranch, the spiny dogfish (Squalus
528
acanthias). J. Exp. Biol. 198: 1545-1558.
529
Urea as a potential disturbance cue in freshwater fishes. 26
530
531
Table 1: Planned contrast values for linear and quadratic estimates for convict cichlids
532
and rainbow trout exposed to varying concentrations of urea. Significant linear terms
533
suggest graded response pattern.
Convict cichlids
Time moving
Foraging attempts
Rainbow trout
Time moving
Foraging attempts
534
535
536
Contrast
difference
95% CI
P
Linear
Quadratic
Linear
Quadratic
-23.12
5.10
-1.48
0.50
-38.69, -7.55
-10.48, 20.67
-2.46, -0.49
-0.49, 1.49
0.004
0.51
0.004
0.32
Linear
Quadratic
Linear
Quadratic
27.07
7.75
2.37
-0.43
11.55, 42.59
-28.25, 2.78
0.67, 4.08
-2.14, 1.27
0.001
0.11
0.007
0.61
Urea as a potential disturbance cue in freshwater fishes. 27
537
538
Figure Captions
539
Figure 1: Mean (± SE) change in time spent moving (A) and foraging attempts (B) for
540
convict cichlids exposed to the odour of undisturbed conspecifics (UC) versus
541
disturbance cues (DC) or urea at one of three concentrations (0.01, 0.1 or 0.5 mg L-1) or a
542
distilled water (DW) control. N = 10 per treatment for disturbance cue vs. the odour of
543
undisturbed conspecifics and N = 15 per treatment for the test of response to urea at
544
varying concentrations.
545
546
Figure 2: Mean (± SE) change in time spent moving (A) and foraging attempts (B) for
547
rainbow trout exposed to the odour of undisturbed conspecifics (UC) versus disturbance
548
cues (DC) or urea at one of three concentrations (0.1, 0.5 or 1.0 mg L-1) or a distilled
549
water (DW) control N = 12 per treatment for disturbance cue vs. the odour of undisturbed
550
conspecifics and N = 15 per treatment for the test of response to urea at varying
551
concentrations.
552
Figure 3: Mean (± SE) change in time spent moving (A) and foraging attempts (B) for
553
convict cichlids exposed to distilled water (open bars), 0.1 (hatched bars) or 0.5 mg L-1
554
(solid bars) of urea, 30 minutes following pre-exposure to distilled water, or urea at 0.1 or
555
0.5 mg L-1. N = 10 per treatment combination.
556
Urea as a potential disturbance cue in freshwater fishes. 28
557
Figure 4: Mean (± SE) change in time spent moving (A) and foraging attempts (B) for
558
rainbow trout exposed to distilled water (open bars), 0.5 (hatched bars) or 1.0 mg L-1
559
(solid bars) of urea, 30 minutes following pre-exposure to distilled water, or urea at 0.5 or
560
1.0 mg L-1. N = 10 per treatment combination.