Journal of Plankton Research Vol.18 na8 pp.1341-1348, 1996
Direct distributional response in Daphnia pulex to a predator
kairomone
Ole T.Kleiven, Petter Larsson and Anders Hobaek1
Institute of Zoology, Department of Animal Ecology, University of Bergen, Alligt.
41, N-5007 Bergen, Norway
1
Present address: Norwegian Institute for Water Research, Regional Office of
Bergen, Thorm0lensgt. 55, N-50O8 Bergen, Norway
Abstract. The hypothesis that zooplankton are capable of direct and differentiated behavioural
responses to different biotic factors was tested. A multichamber flow-through technique enabled predetermined food levels of the green alga Scenedesmus acutus to be maintained constant throughout
the experiments. By adding water from a culture of fourth instar Chaoborus flavicans larvae to one of
the end chambers, a concentration gradient of the chemical substance was established across the parallel chambers. Evidence is presented for a direct avoidance response in the water flea Daphnia pulex
to a substance released by the predatory midge C.flavicans. This response occurs in the absence of
visual cues and is independent of food availability. Finally, the results indicate that swarming in D.pulex
may occur in the absence of both light and predator scent, but is influenced by food concentrations.
Introduction
Predation exerts a severe selective force on many animals, and the ability to recognize predators and avoid predation are crucial elements in many animals' lives.
Consequently, they have evolved different behavioural adaptations to avoid predation. Many planktonic animals use chemical substances released by other
animals as signals to avoid predation, i.e. modify their morphology or reproductive strategy, as well as their behaviour (reviewed in Larsson and Dodson, 1993).
The behavioural responses to chemical signals have been interpreted as modifications of phototactic or other light-dependent behaviours, like diurnal vertical
migrations or daytime swarm formation (Jakobsen and Johnsen, 1988a; Ringelberg, 1991a,b). These behaviours are adaptive in reducing mortality risk by visual
predators like planktivorous fish. With the seasons, a marked succession in the
density of all types of organisms takes place (Sommer et al., 1986). This gives a
habitat with changing predation and competition pressure. In such an environment, one should expect planktonic animals to adopt inducible defence mechanisms, making them able to cope with the variation in food availability and risk of
predation.
Often, anti-predator defences are assumed to be costly for the organism that
forms the defence, whether it is a change in morphology, behaviour or life history.
However, many vertebrates and invertebrates show capabilities of adjusting their
behaviour according to hunger status and predator pressure (Milinski and Heller,
1978; Sih, 1980).
The aim of this study was to investigate the ability of zooplankton to respond
behaviourally to predator kairomones by avoiding areas of high predator densities. Further, we studied how this behaviour is influenced by food availability, and
to what extent planktonic animals can recognize and differentiate between
© Oxford University Press
1341
O.T.Kteirea, P.Larssoa and AJiobak
chemical substances from neighbouring organisms. Are individuals able to weigh
the risk of predation when they are given different food densities and thus different competition from conspecifics?
To test the hypothesis that zooplankton are capable of direct and differentiated
behavioural responses to different biotic factors, the distribution of the water flea,
Daphnia pulex, was analysed in a laboratory experiment. The animals were given
water with and without chemical substances from a predator (Chaoborus flavicans) in complete darkness and under different food concentrations. Their distributions were analysed in the different situations.
Method
A clone of D.pulex was used in the experiments which originated from Lake
Myravann, near Bergen, Norway, and which has been maintained in laboratory
cultures since 1987. All experiments were carried out in a multichamber flowthrough system (Figure 1). A multichannel peristaltic pump fed and drained water
to and from each chamber at a rate of 4 ml min-'.The flow-through technique
allowed food concentrations (pre-determined levels of the green alga
Scenedesmus acutus) to be maintained constant throughout the experiments. By
adding water from a culture of fourth instar C.flavicans larvae to one of the end
chambers, a concentration gradient of the predator scent was established across
the parallel chambers. This gradient was visualized and tested with a dye, and
found to persist for much longer than the duration of one experimental run (4 h).
The medium for all experiments was produced from purified water, to which
CaCO3 and sea salts were added (details in Hobaek and Larsson, 1990). All experiments were conducted at room temperature. Food levels were set at 0.04 mg dry
weight (DW) I"1 (i.e close to the lower limit for somatic growth; Gliwicz, 1990) or
1.0 mg DW h1 (growth unlimited by food; Lampert and Schober, 1980). Animal
stock cultures were kept in the same medium at room temperature and preconditioned to experimental food concentrations for 8 h. To produce the chemical substance (predator scent), water from a culture of 10-15 C.flavicans Yx
Fig. L An outline of the flow-through chambers used in the experiments, (a) Five inlets and outlets
driven by a peristaltic pump, (b) Theflow-throughchambers glued together are clear culture-flasks of
0.8 L (c) The openings between the chambers are 1-5 X 3 cm and closing them was possible, (d) A
plankton net preventing the animals escaping from the chambers.
1342
Response in D.pulex to a predator kairomone
(synthetic medium, 24 h incubation time) was used. The holding water was subsequently glass fibre filtered and the food concentration was established. The food
concentration was constant in allfivechambers All experiments were run in complete darkness to avoid interference with phototactic responses. After each run,
the number of animals in each chamber was counted. Between runs, the direction
of the chemical gradient was alternated. Ten adult egg-bearing D.pulex were
placed in each chamber at the start of a run. Five trials were run for each treatment.
Statistical analyses
The data consisted of 4 X 5 different trials where each trial gave an empirical
distribution of 50 animals in five chambers. The chambers were numbered from
one to five. Since the total number of animals was fixed in each experiment, the
numbers of Daphnids in the different chambers were dependent. In addition, the
animals had in general a dependent behaviour. Hence, the standard assumptions
necessary to carry out a multiway ANOVA, log-linear model or a contingency
table are not fulfilled.
To solve this problem, a specific one-dimensional parameter was chosen for the
Daphnia distribution. With estimates of this parameter, independent observations
could be achieved. The analysis consisted of four steps. The first three steps were
carried out for each trial. First, the relative frequency of the animals was calculated for each chamber. Then, an arcsine transformation (Sokal and Rohlf, 1995)
was applied. In the third step, an ordinary regression model was fitted with
chamber number as the independent variable. The one-dimensional parameter is
the slope. In the last step, a factorial ANOVA was applied to these 20 estimated
slopes. There were two factors (food and Chaoborus water), each with two levels
(low/high and without/with). A positive slope shows a directional migration away
from the chamber infused with Chaoborus water.
In order to examine the influence of the initial distribution in the controls,
another experiment with 10 additional trials was set up. These trials consisted of
five with low and five with high food concentrations without Chaoborus water
input. Each run started with all the 50 animals in the middle chamber. To analyse
these data, a one-dimensional variable that we called a score was chosen. It was
defined as:
Score = z Lc,- - x\
where x, is the number of animals in chamber number i in a given trial and x is the
average number of animals in all chambers in the trial. The maximum value of the
score variable is 80 and the minimum value is zero. A score of zero corresponds
to a distribution where all animals are equally distributed among the chambers
(no aggregation) and an increasing score represents an increasing tendency to
aggregate. Combining all data relevant to the experiment, 20 different scores were
obtained which were analysed by a factorial ANOVA. There are two factors, each
of them with two levels: the initial distribution (10 animals in each chamber or all
1343
O.T.Heivea, P.Laissoa and A.Hobick
a)
b)
30
*e3
E
20
10
2
3
4
5
1
2
3
4
5
Chamber no.
Fig. 2. Horizontal distribution patterns of D.pulex after 4 h of experimental conditions. The data are
means ± SD of Eve replicate trials. All experiments were started with 10 animals in each chamber, (a)
Avoidance reaction to a chemical released by the predator Cflavicans under low and high food conditions. In chamber no. 1, Chaoborus water was added, while the other chambers received control
water, (b) Controls with low and high food concentrations without Chaoborus water.
50 in the middle chamber) and the food concentration (high or low food levels).
Standard tests showed that the data are neither significantly heteroscedastic nor
significantly non-normally distributed.
Results
Both at high and low food concentrations, D.pulex responded significantly by
avoiding the end chamber infused with Chaoborus water, accumulating in chambers 4 and 5 (Figure 2a). The control series showed no directional migrations
(Figure 2b). From the ANOVA table, it can be concluded that Chaoborus water
has a significant impact on the Daphnia distribution, but the food factor is not significant (Table I). The variance analyses also give estimates of each group at the
same factor levels which shows that Chaoborus water gives a significant positive
Table L A factorial ANOVA testing for the influence of chemical cues from Chaoborus (Chaoborus
water versus control water) and food level (0.04 versus 1.0 mg DW Scenedesmus I"1) on horizontal
migration of D. pulex. The data are arcsin-transformed frequencies and the test is computed on the
regression coefficients calculated from each trial (five trials from each treatment, n = 20)
Source of variance
d.t
SS
MS
F
P
Chemical cue
Food level (F)
CxF
Error
Total
1
1
1
16
19
0.009
0.000
0.001
0.009
0.019
0.009
0.000
0.001
0.001
0.001
16325
0.176
1.711
0.001
0.681
0.209
1344
Response in D.pulex to a predator kairomone
30
D Low food
• High food
.§20
10
.1
1
1
1
3
2
4
5
Chamber no
Fig. 3. Controls with low and high food concentrations where all 50 animals were initially placed in the
middle chamber and without predator water. The data are means ± SD of five replicate trials. The low
food level induced a regular distribution, while the high food level induced aggregation (swarm formation).
slope (blow food = 0.0481 andfehighfood = 0.0364)."Trials without Chaoborus water do
not give slopes signiGcant from zero (b\ow food = 0.0005 and b^ food = 0.0131). All
of the estimates have a SD equal to 0.0093 (based on the pooled variance estimate
from the ANOVA). "These results imply directional horizontal migration as a
direct response to the predator signal.
The control series run at high or low food concentrations revealed an additional
pattern (Figure 2b). Daphnia pulex appeared to aggregate in the three middle
chambers when food was abundant, but became evenly distributed when food was
limited. Very similar results were obtained when these experiments were repeated
and all 50 Daphnia were initially placed in the middle chamber (Figure 3). The
ANOVA table shows that the different food regimes produced scores (distributions) that differed significantly from each other independently of the initial
position of the daphnids. The high food level gave significantly higher scores than
the low food level, while the initial position was not significantly different (Table TI).
The contrast between the two food levels is also similar to that of the original controls, i.e. independent of initial starting position. These results imply non-random
Table IL Factorial ANOVA testing the difference in aggregation in relation to high and low food levels
(1.0 versus 0.04 mg DW Scenedesmus I"1), and t w o different initial starting positions (10 animals in
each chamber versus all SO animal started in the middle chamber)
Source of variance
d.t
Start position (SP)
Food level (F)
SPXF
Error
Total
1
1
1
16
19
SS
MS
F
P
151.250
1428.050
36.450
1440.000
3055.750
151.250
1428.050
36.450
90.000
160.829
1.681
15.867
0.405
0.213
0.001
0.534
1345
O.T.KIeiven, P.Lanson and AJIotneb
aggregation in the high-food series and non-random regular distribution in the
low-food series.
Discussion
There are several studies indicating that the behaviour controlling the spatial
distribution of zooplankton is more complex than previously assumed. These
studies involve dynamic adjustment in vertical as well as horizontal positions, in
response to biological factors that directly influence their fitness (Pennak, 1973;
Gliwicz and Rykowska, 1992; Loose and Dawidowicz, 1994; Watt and Young
1994). In the continuum of habitats existing in a lake, it would be advantageous to
respond and adjust behaviourally to various neighbouring organisms. Chemical
substances released by predators, and/or the planktonic animals themselves, can
be sensed and responded to (Larsson and Dodson, 1993). This ability to detect
chemical cues supplements both vision and tactile cues.
This experiment is one of the first contributions showing that not only the vertical, but also the horizontal, distribution pattern may be induced by chemical
compounds released by a predator to the environment. It is shown that D.pulex
has a direct behavioural avoidance response to the predator kairomones. This
response occurs in the absence of visual cues and is independent of food availability. It also seems clear that Daphnia actively avoided sites where the scent of
the predator was strong, since the direction of net movement was always the same.
Whether the accumulation observed involved active swarming besides directed
movement is still unclear. Food limitation did not alter the avoidance response.
Nonetheless, Johnsen and Jakobsen (1987) found diel vertical migration in
Daphnia (a predator prevention response) to cease at low food densities.
In a series of experiments with Daphnia, Kvam and Kleiven (1995) showed a
different response to Chaoborus water when the animals were exposed to light.
In that case, well-fed Daphnia aggregated in the middle rather than the end chambers. This may indicate that in the light, aggregation overrules the directed escape
response. Their experiments were run with ample food, and so it is still uncertain
how hungry Daphnia exposed to Chaoborus water would distribute in the light.
In previous studies, chemical cues have been shown to modify phototactic
responses, and to influence horizontal or vertical migrations in zooplankters
(Ringelberg, 1991a,b; De Meester, 1993; Loose et ai, 1993). Other studies have
shown reversed diel vertical migration (ascending during daytime) in zooplankton in response to Chaoborus spp. (Dodson, 1988; Neill, 1990,1992). Our results
demonstrate directed avoidance without any concomitant changes in light conditions. Thus, the ability to detect predators chemically in the dark shows that
daphnids can respond directly to a chemical cue, in the absence of visual cues, and
also use this information to distribute in a beneficial way.
The experiments also revealed a systematic difference in spatial distribution for
both high- and low-food controls, independently of the starting position of the
experimental animals (Figure 3). The score variable used to test this experiment
underestimates the effect of dispersion to some extent, since it measures dispersion or aggregation and not the relative length of dispersion, i.e. it gives the same
1346
Response ta D.pulex to a predator kairomone
value if one animal moves from chamber no. 3 to no. 2 or 4 as from chamber no. 3
to no. 1 or 5. However, this will not affect the conclusion of the experiment. Nevertheless, considering the different initial conditions, the contagious and regular
distributions seem to result from non-random movements by the animals. When
poorly fed, the Daphnia apparently avoided each other (or were not attracted to
each other), resulting in an almost regular distribution, i.e. maximizing their
nearest-neighbour distance. In contrast, well-fed animals aggregated into swarms,
given a high variance/mean ratio.
One can argue that our experimental design may lead to aggregation/swarming
at high food concentration. Thus, another explanation for these results can be that
random movement, plus 'attraction' at high food concentration, would lead to
more or less the same result. In our experiments, high food concentrations induced
swarming in one of the three middle chambers, and this is most likely due to the
fact that the two end chambers had only one opening, while the other chambers
had two. On the other hand, low food concentration induced regular distributions,
irrespective of the starting position of the animals. Nonetheless, the observed patterns are consistent with the swarming behaviour found by Jakobsen and Johnsen
(1988a). The adaptive significance of food-dependent swarming is reasonably
explained by intensified competition for food particles with increasing animal
density (as long as food is limiting). Thus, when food levels are high, the costs
associated with swarming are probably small compared to the benefits in terms of
decreased predation success (Hamilton, 1964; Jakobsen and Johnsen, 1988b).
Therefore, the animal can aggregate as a precaution against predators. Note, also,
that no predator scent was required to elicit aggregation of Daphnia in darkness.
Hence, the animals must have detected each other's presence by means of chemical or tactile cues. Preliminary data (O.T.Kleiven, unpublished) suggest that
chemical cues are sufficient, but this aspect requires further testing. By contrast,
Jakobsen and Johnsen (1988a) found swarms of Bosmina longispina to occur only
during daytime, disintegrating as darkness fell.
The Daphnia stock used in our experiments may have adapted to intense
Chaoborus predation, since they originate from a lake where C.flavicans is the
dominant predator (Giske, 1986). Daphnids seem to be able to adjust their spatial
distribution by responding to many different cues, including light and light
changes, chemical substances from various organisms, and food levels.
Acknowledgements
We thank Mrs Inger Haugsb0 for assistance in the laboratory, Drs G. Johnsen and
P. Jakobsen, as well as Harald Saegrov and Elisabeth M. Lysebo, for comments on
previous versions of the manuscript and Hans Karlsen (Institute of Mathematics
and Statistics, University of Bergen) for statistical advice. This work was supported
by the Norwegian Research Council for Science and the Humanities.
References
De Meester.L. (1993) Genotype, fish-mediated chemicals and phototaxis in Daphnia. Ecology, 74,
1467-1474.
1347
O.T.KIeiven, P.Larsson and A.Hobsk
Dodson.&I. (1988) The ecological role of chemical stimuli for the zooplankton: Predator-avoidance
behavior in Daphnia. Limnol. Oceanogr., 33,1431-1439.
GiskeJ. (1986) Populasjonsregulerende faktorer hos to arter Daphnia i et vann uten planktivor fisk.
Master Thesis, Museum of Zoology, University of Bergen, Bergen, Norway (in Norwegian).
Gliwicz,Z.M. (1990) Food thresholds and body size in cladocerans. Nature, 320,746-748.
Gliwicz,Z.M. and Rykowska.A. (1992) 'Shore avoidance' in zooplankton: a predator-induced behavior
or predator-induced mortality? J. Plankton Res., 14,1331-1342.
Hamilton,W.D. (1964) The genetical evolution of social behaviour./ Theor. Biol.,1,1-52
Hobsek.A. and Larsson.P. (1990) Sex determination in Daphnia magna. Ecology, 71,2255-2268.
Jakobsen,PJ. and Johnsen.G.H. (1988a) The influence of food limitation on swarming behaviour in the
waterflea Bosmina longispina. Anim. Behav., 36,991-995.
Jakobsen,PJ. and Johnsen,G.H. (1988b) Size-specific protection against predation by fish in swarming
waterfleas, Bosmina longispina. Anim. Behav., 36,986-990.
Johnsen.G.H. and Jakobsen.PJ. (1987) The effect of food limitation on vertical migration in Daphnia
longispina, Limnol. Oceanogr., 32,873-880.
Kvam,O.V. and Kleiven.O.T. (1995) Diel horizontal migration and swarm formation in Daphnia in
response to Chaoborus. Hydrobiologia, 307,177-184
Lampert.W. and Schober.U (1980) The importance of'threshold' food concentration. In Kerfoot,W.C.
(ed.), Evolution and Ecology of Zooplankton Communities, University Press of New England,
Hanover, New Hampshire and London, pp. 264-267.
Larsson.P. and Dodson,S. (1993) Chemical communication in planktonic animals. Arch. Hydrobiol.,
129,129-155.
Loose.C. and Dawidowicz,P. (1994) Trade-offs in diel vertical migration by zooplankton: The costs of
predator avoidance. Ecology, 7,2225-2263.
Loose.C, von Elert.E. and Dawidowicz,P. (1993) Chemically-induced diel vertical migration in
Daphnia: a new bioassay for kairomones exuded by fish. Arch. Hydrobiol., 126,329—337.
Milinski,M. and Heller,R. (1978) Influence of a predator on the optimal foraging behaviour of sticklebacks (Gasterosteus aculeatus L.). Nature, 275,642-644.
Neill,W.E. (1990) Induced diel vertical migration in copepods as a defence against invertebrate predation. Nature, 345,524-526.
Neill,W.E. (1992) Population variation in the ontogeny of predation-induced vertical migration of
copepods. Nature, 356,54-57.
Pennak.R.W. (1973) Some evidence for aquatic macrophyes as repellents for a limnetic species of
Daphnia. Int. Rev. Ges. Hydrobiol., 58,569-576.
RingelbergJ. (1991a) Enhancement of the phototactic reaction in Daphnia hyalina by a chemical
mediated by juvenile perch (Percafluviatilis).J. Plankton Rcs.,13,17-25.
RingelbergJ. (1991b) A mechanism of predator-mediated induction of diel vertical migration in
Daphnia hyalina. J. Plankton Res., 13,83-89.
Sih.A. (1980) Optimal behavior: Can foragers balance two conflicting demands? Science, 210,
1041-1043
Sokal,R.R. and RohlfJU. (1995) Biometry, 3rd edn. W.H.Freeman and Co., New York.
Sommer.U., Gliwicz,Z.M., Lampert.W. and Duncan^A. (1986) The PEG-model of seasonal succession
of planktonic events in freshwater. Arch. Hydrobiol., 106,433-471.
WattJU. and Young,S. (1994) Effect of predator chemical cues on Daphnia behaviour in both horizontal and vertical planes. Anim. Behav., 48,861-869.
Received on September 10,1995; accepted on March 6,1996
1348