Download JMS 70_4 353-358 eyh040 FINAL

PREDATOR AVOIDANCE IN BULINUS GLOBOSUS (MORELET, 1866)
AND B. TROPICUS (KRAUSS, 1848) (GASTROPODA: PLANORBIDAE)
EXPOSED TO PREDATORY AND NON-PREDATORY FISH
PAUL MAKONI 1 , MOSES J. CHIMBARI 2 , 4 AND HENRY MADSEN 3
1
De Beers Research Laboratory, Box 197, Chiredzi, Zimbabwe; 2Blair Research Laboratory, PO Box CY 573, Causeway, Harare, Zimbabwe;
3
Danish Bilharziasis Laboratory, Jaegersborg Allé 1d, DK2920 Charlottenlund, Denmark;
4
Present address: University of Zimbabwe Lake Kariba Research Station, PO Box 48, Kariba, Zimbabwe
(Received 25 November 2002; accepted 1 March 2004)
ABSTRACT
The cichlid fish, Sargochromis codringtonii, has been suggested for biological control of freshwater snails,
especially those serving as intermediate hosts for schistosomes. This study examined the behaviour of
two aquatic snail species, Bulinus globosus and Bulinus tropicus, when exposed to water conditioned
(defined as water inhabited by fish) by either Sargochromis codringtonii, a molluscivore, or Tilapia rendalli
(Boulenger, 1896), a herbivore, and when exposed to predation risk in the presence of a refuge. Both
snail species crawled above the waterline to a greater extent when exposed to water conditioned by
S. codringtonii than when exposed to water conditioned by T. rendalli, or unconditioned water.
Although the number of snails leaving the water tended to increase with the density of S. codringtonii,
this was not statistically significant. While Bulinus globosus elicited a greater response to water conditioned by feeding fish than to that conditioned by non-feeding fish, B. tropicus did not respond differently to the two treatments. The introduction of S. codringtonii into tanks with a refuge caused snails to
move actively into the covered areas.
INTRODUCTION
Snails may have morphological features or behavioural patterns
that make it possible for them to coexist with their predators
(Vermeij, 1974; Vermeij & Covich, 1978). The morphological
features include thick protective shells and an operculum.
These special morphological features may have evolved during
the many years of coexistence of snails and their predators
(Vermeij, 1974; Vermeij & Covich, 1978; Palmer, 1979). Snails
can escape predation if they are too large for the predator to
consume or if their shells are too thick to crush (Osenberg &
Mittelbach, 1989). Most freshwater snails are generally less well
adapted to avoid crushing from predators than most marine
forms, with shells that are generally thinner than those of
marine species. Instead, many thin-shelled pulmonates have
behavioural adaptations that aid them in avoiding predators
(Vermeij & Covich, 1978).
The predator avoidance response of snails is a behavioural
pattern that has been demonstrated in several studies (Covich,
1981; Alexander & Covich, 1991). Freshwater snails display several predator avoidance strategies that include burrowing into
the sediment, crawling into vegetation or above the waterline,
or vigorous movements of the shell upon contact (Alexander
& Covich, 1991; Turner, 1996). Alexander & Covich (1991)
observed a tendency by Physella virgata (Gould) and Planorbella
trivolis (Say) to crawl above the waterline in response to the
presence of the crayfish Procambarus simulans. Likewise, the
presence of Procambarus clarkii caused Physa acuta Draparnaud
to crawl above the waterline (Hofkin et al., 1991).
The behavioural response of snails in the presence of predators may be due to the presence of the predator, disturbances
due to attempted attacks by the predator, or fluids from crushed
conspecific snails. Chimbari (1996) found an increased tendency for Bulinus globosus to leave water that had been conditioned by Sargochromis codringtonii.
Correspondence: H. Madsen; e-mail: [email protected]
J. Moll. Stud. (2004) 70: 353–358
The aim of the present study was to investigate the behaviour
of Bulinus globosus and Bulinus tropicus in the presence of
Sargochromis codringtonii (a molluscivore) and Tilapia rendalli (a
herbivore), presence of different densities of Sargochromis
codringtonii and whether feeding by fish influenced this
behaviour. The study also investigated if snails would move into
protected areas in the presence of a predator.
MATERIAL AND METHODS
Response of snails to water conditioned by fish
The three experiments involved supplying water conditioned by
fish, or unconditioned water, from large aquaria to snails in
small test containers. Because the chemicals released by the
conditioning fish could be volatile in water, the test containers
received water continuously from the fish aquaria. Initially,
tap water was put into an outdoor concrete pond and left to
dechlorinate for 2 days. After dechlorination the water was
pumped into five large aquaria (100 cm long and 60 cm wide
and high). The aquaria were filled to 75% capacity. The water in
the big tanks would either be conditioned by the addition of fish
or left unconditioned. The water was conditioned by fish for two
days prior to conducting the experiment. Each of the large
aquaria continually supplied water to 10 500-ml plastic containers through siphoning tubes. A hole was punched into one side
of the test containers to allow for a constant volume of water. At
any given time each plastic containers held 370 ml of water. The
rate of flow of the water was 0.13 cm3 per second per test container, which meant that water in test containers was replaced
about 10 times during an 8-h period. Water was not added to the
large aquaria during the experiments.
The S. codringtonii fish used in the experiment had standard
lengths (from snout tip to base of caudal peduncle) of 13–14 cm
and total wet weights from 50 to 56 g. Tilapia rendalli, an herbivorous cichlid species, was used for comparison in experiment
1 and had standard lengths ranging from 13.0 to 14.5 cm and
Journal of Molluscan Studies Vol. 70, No. 4 © The Malacological Society of London 2004, all rights reserved.
P. MAKONI ET AL.
wet weights from 63 to 71 g. Laboratory-bred Bulinus globosus
and B. tropicus with shell heights ranging from 5.5 to 7.0 mm
were used in the experiment. Five specimens of either B. globosus
or B. tropicus were introduced into each test container. Each
large tank supplied water to five containers with B. globosus and
another five with B. tropicus. After 8 h the number of snails above
the water line was counted. Water temperature, conductivity,
dissolved oxygen and pH were measured in all the aquaria
during counting of snails. Each experiment was repeated 10
times.
Experiment 1: Water conditioned by Sargochromis codringtonii
and Tilapia rendalli, effect over time. In each of two large aquaria
a single S. codringtonii was introduced and one Tilapia rendalli
was introduced in each of another two aquaria. The fifth aquarium without fish served as a control. After the first observation,
which was done 8 h after the commencement of the experiment, the counting of snails above the waterline took place at
hourly intervals for another 8 h.
Experiment 2: Effect of different densities of S. codringtonii.
Four aquaria were used and water was conditioned by one, three
or five specimens of S. codringtonii, while in the last aquarium,
water was left unconditioned. After 8 h the number of snails
above the waterline was counted.
Experiment 3: Effect of feeding by S. codringtonii. Three
S. codringtonii were introduced into each of four aquaria, while
the fifth one was a control without fish. In two of the tanks, the
fish were fed with snails, whereas in the other two tanks the fish
were not fed. The fish were fed at the beginning of the experiment with snails and then at 3-h intervals. Therefore, the fish
were fed three times during the experiment. In order to avoid
conditioning of water by live snails in the treatment, the fish
were given snails at a rate that ensured that no snails escaped
predation. Snails were added in approximately equal numbers
of B. tropicus and B. globosus. A fish would be offered snails of the
same size range as experimental snails until it stopped feeding.
Each fish consumed 50–60 snails during the conditioning
period.
binary logistic regression (Hosmer & Lemeshow, 1989) where
repeat experiment, time (experiment 1 only), conditioning
factor and snail species were entered as categorical factors in
mentioned order. Indicator coding was used with either the first
or the last category as the indicator group. Interactions between
snail species and the conditioning factor were made as the last
step. Significance of a given factor was based on the change in
the –2 log likelihood ratios. Model fit was assessed using the
Hosmer-Lemeshow statistics (Hosmer & Lemeshow, 1989).
The log of the regression coefficient is equivalent to the odds
ratio, i.e. the likelihood of snails leaving water in a given group
relative to that in the reference group, after adjusting for the
other factors in the model. For the experiment with refuge, the
three possible positions were analysed in a similar way using
multinomial logistic regression. Comparisons of physical and
chemical variables across treatments were done using either
one-way or two-way (i.e. repeat experiment and treatment as
factors) analysis of variance. The positions of snails were
analysed using nominal logistic regression where repeat experiment, time, treatment (fish or no fish), snail species and the
interaction between treatment and time were entered as factors.
RESULTS
Experiment 1: Water conditioned by Sargochromis
codringtonii and Tilapia rendalli – effect over time
Bulinus globosus and B. tropicus responded more to S. codringtoniiconditioned water compared with T. rendalli-conditioned water
and unconditioned water (Fig. 1). The response varied significantly across repeat experiments (P 0.001), across conditioning factor (P 0.001), time (P 0.001) and between
species (P 0.001). There was no significant interaction
between conditioning factor and snail species. The number of
snails out of water increased over time, and snails exposed to
Presence of refuge
Five large aquaria as described previously were filled to 75%
capacity with water prepared as described. The refuge was built
from two pieces of 7-mm-thick glass. A rectangular glass plate
(40 45 cm) was supported by four 5-cm-tall ‘legs’ on one-half
of a second rectangular glass (80 45 cm). The space between
the top and bottom glass panes was intended to serve as a refuge
to snails after the introduction of S. codringtonii. The fish used in
the experiments ranged in standard length from 14.8 to 15.0 cm
and had wet weights from 50 g to 56 g. The fish were too large to
fit in the refuge space between the glass panes. The glass plates
were placed in the five large tanks. The tanks were aerated using
scorpion II filters during the experiment.
Laboratory-bred specimens of Bulinus globosus and B. tropicus
were used in the experiment. The snails had shell heights ranging from 5.5 to 7.0 mm. A total of 120 snails of each species were
introduced into each tank and left for 24 h. Then a single
S. codringtonii was introduced in each of three tanks whereas the
other two served as controls. The number of snails within the
refuge area, on open surfaces and out of the water was recorded
just prior to the introduction of fish. After 8 h the number of
snails occupying different positions was recorded and hourly
thereafter for a total of 9 h. Water temperature, conductivity
and pH were measured in all the aquaria. The experiment was
repeated seven times.
Statistical analysis
Position of snails (1 out of water and 0 in water) in the three
experiments with fish-conditioned water was analysed using
Figure 1. Percentage of snails above the waterline in aquaria conditioned
with Sargochromis codringtonii or Tilapia rendalli.
354
PREDATOR AVOIDANCE BEHAVIOUR IN BULINUS
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
20
1
2
3
4
5
6
7
8
9
30
1
2
3
4
5
6
7
8
9
40
1
2
3
4
5
6
7
8
9
50
1
2
3
4
5
6
7
8
9
60
1
2
3
4
5
6
7
8
Table 1. Average (± standard deviation) values of physical and chemical variables measured in tanks containing Sargochromis codringtonii
and Tilapia rendalli (n 10).
P-values
Variable
Temperature (°C)
Oxygen (%)
Conductivity (S)
Time
Control
S. codringtonii
T. rendalli
Repeat
(h)
(n 10)
(n 20)
(n 20)
experiment
Treatment
ns
0
26.07 0.69
25.78 0.88
26.16 0.79
0.01
8
25.49 0.48
25.85 0.79
26.08 0.67
0.01
0.05
16
24.96 1.11
25.93 1.37
25.81 1.24
0.01
ns
0
61.73 4.88
61.27 5.07
63.55 6.36
<0.001
ns
8
61.00 4.95
58.20 4.81
60.02 6.52
0.001
ns
16
60.63 4.89
55.93 4.89
57.88 6.44
0.001
ns
0
118 8
125 9
122 11
ns
ns
8
118 8
125 9
122 11
ns
ns
16
pH
118 8
125 9
122 11
ns
ns
0
7.69 0.26
7.67 0.22
7.79 0.19
0.001
0.05
8
7.71 0.30
7.66 0.26
7.76 0.22
0.001
ns
16
7.69 0.28
7.67 0.23
7.74 0.19
0.001
ns
ns, not significant (P 0.05).
fish were 5.548 times more likely to move out of the water than
the control snails. Those exposed to water conditioned by three
and five fish were 7.538 and 7.110 times, respectively, more
likely to leave water than the control snails. Bulinus globosus was
1.334 times more likely to move out of water than B. tropicus. The
Hosmer-Lemeshow goodness-of-fit test showed that the model
fit was good (2 15.04, df 8, P 0.05). Overall, 66.3% of the
snails were correctly classified by this model (72.0% of those in
the water and 58.8% of those out of the water). The difference
in response between the treatment with one fish and five fish was
not significant and also the difference between densities of
three and five fish was not significant. However, the response of
snails exposed to water conditioned by one fish was significantly
different from three and five fish pooled (P 0.05).
The average values of physical and chemical parameters
measured during the experiment are shown in Table 2. Oxygen
was a slightly lower in aquaria with fish than in the control
aquarium, and this difference was more pronounced at the end
of the experiment.
Tilapia-conditioned water were 2.296 times more likely to leave
water than snails exposed to unconditioned water. Those
exposed to water conditioned by S. codringtonii were 5.892 times
more likely to leave water than those exposed to unconditioned
water. Bulinus globosus was 1.274 times more likely to move out of
water than B. tropicus. The Hosmer-Lemeshow goodness-of-fit
test showed that the model of estimating whether snails were in
or out of water was good (2 0.227, df 8, P 0.05). Overall,
70.6% of the snails were correctly classified by this model
(87.0% of those in the water and 45.8% of those out of the
water). A summary of the physical and chemical parameters
measured during the experiment is shown in Table 1. Although
these variables differed significantly between repeat experiments there was only a slight indication of variation across treatments.
Experiment 2: Effect of different densities of Sargochromis
codringtonii
Bulinus globosus and B. tropicus did not respond differently to different densities of S. codringtonii (Fig. 2). Logistic regression
analysis showed significant differences in response across repeat
experiments (P 0.001), number of fish (P 0.001) and snail
species (P 0.001). Snails exposed to water conditioned by one
Experiment 3: Effect of feeding Sargochromis codringtonii
The snails demonstrated a greater avoidance response to water
conditioned by feeding fish than to water conditioned by nonfeeding S. codringtonii (Fig. 3). Logistic regression showed that
there was a significant interaction between treatment and snail
species (P 0.01) and the model reported on here includes this
interaction. There was a significant difference in the response of
snails across repeat experiments (P 0.001) and across treatments (i.e. unconditioned water, water conditioned by feeding
and non-feeding fish (P 0.001)), whereas the main effect of
snail species was not significant. The snails were 2.877 times
more likely to move out of water when exposed to water conditioned by unfed fish than the control. Snails exposed to water
conditioned by feeding fish were 3.899 times more likely to leave
water than those exposed to unconditioned water. The significant interaction between treatment and snail species is because
B. globosus elicited a greater response to water conditioned by
feeding fish than to that conditioned by non-feeding fish, while
B. tropicus did not respond differently to feeding and nonfeeding fish. The Hosmer-Lemeshow goodness-of-fit test indicated a good model (2 10.183, df 8, P 0.05). Overall,
65.0% of the snails were correctly classified by this model
(78.5% of those in the water and 46.4% of those out of the
Figure 2. Percentage of Bulinus globosus and Bulinus tropicus above the waterline in aquaria conditioned with different densities Sargochromis codringtonii.
355
P. MAKONI ET AL.
Table 2. Mean ( standard deviation) values of physical and chemical variables measured in tanks containing different densities of
Sargochromis codringtonii.
Variable
Time (h)
Control
One fish
Three fish
Five fish
P-values
Temperature
0
23.83 3.11
24.30 3.35
24.05 3.03
24.05 3.26
ns
(°C) (n = 10)
8
23.28 2.74
23.98 3.15
23.59 2.80
23.82 3.12
ns
Oxygen (%)
0
67.76 6.48
60.57 16.57
54.91 10.37
53.32 12.53
0.05
(n = 10)
8
63.72 7.27
47.82 14.31
49.07 11.00
49.06 14.81
0.05
Conductivity
0
120 10
118 12
118 4
125 12
ns
(S) (n = 6)
8
120 10
118 12
118 4
125 12
ns
pH (n = 10)
0
8.19 0.58
7.92 0.67
8.14 0.67
7.97 0.73
ns
8
6.58 0.75
6.41 0.67
6.59 0.96
6.43 0.90
ns
ns, not significant (P 0.05).
Table 3. Average (± standard deviation) values of physical and chemical variables measured in tanks containing fed and non-fed
Sargochromis codringtonii.
P-values
Variable
Temperature (°C)
Conductivity (S)
pH*
Time
No fish
Fish fed
Fish not fed
Repeat
(h)
(n = 10)
(n = 20)
(n = 20)
experiment
Treatment
0
23.57 1.21
23.64 1.34
23.61 1.18
0.001
ns
8
23.68 1.09
23.66 1.20
23.73 1.11
0.001
ns
0
144 7
149 10
148 8
0.001
ns
8
145 5
150 8
149 6
0.05
ns
0
7.68 0.82
7.47 0.72
7.73 0.59
–
ns
8
8.49 1.09
8.08 1.13
8.38 1.35
–
ns
*Done for three experiments only. ns, not significant (P 0.05).
tribution of snails remained fairly constant throughout the
study period. Multinomial logistic regression showed that
there was a significant variation across repeat experiments
(P 0.001), treatment (P 0.001), snail species (P 0.01) and
further there was a significant interaction between conditioning
factor and time (P 0.001). The model showed that snails
were 3.758 times more likely to seek refuge in the presence of
S. codringtonii than when no fish were present. Bulinus globosus
had a slightly larger tendency to seek refuge than B. tropicus.
Snails were 10.798 times more likely to leave water in the presence of S. codringtonii than when no fish were present. Bulinus
globosus had a slightly lower tendency to leave water than B. tropicus. On average at the beginning of the experiments, about 22%
of the snails were under the tray and 75% would be in the open
surfaces, after 8 h, 64% of the surviving snails would be under
the tray and 30% in the open surfaces. Averages of physical and
chemical variables measured during the experiment are presented in Table 4. There was only little variation between repeat
experiments while temperature was slightly higher in tanks with
fish than in the controls. This temperature difference was evident during the last 3 h of the experiment (data not shown).
Figure 3. Percentage of Bulinus globosus and Bulinus tropicus above the waterline in aquaria conditioned with feeding and non-feeding Sargochromis
codringtonii.
DISCUSSION
water). Although the physical and chemical variables differed
significantly between repeat experiments, there were no significant differences between treatments (Table 3).
This study demonstrated the ability of Bulinus globosus and
B. tropicus to respond behaviourally when exposed to water conditioned by S. codringtonii. Several studies have demonstrated
that the presence of predators causes snails to respond. The
addition of crayfish Procambarus simulans caused Physella virgata
and Planorbella trivolvis to crawl out of water (Alexander &
Covich, 1991). Hofkin et al. (1991) also reported a positive association between the crawling out of water behaviour of Physa
acuta and the crayfish Procambarus clarkii. The crawling out of
water behaviour of snails has been interpreted as a protective
mechanism against predation. The snails reacted more to water
Presence of refuge
Sargochromis codringtonii was observed to start feeding on the
snails once it was introduced into the tanks. Figure 4 show the
percentage of snails occupying the three available positions over
time in tanks with S. codringtonii and controls. The graphs show
that more snails moved to the refuge after the introduction of
S. codringtonii, whereas in the control aquaria, the relative dis356
PREDATOR AVOIDANCE BEHAVIOUR IN BULINUS
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
20
1
2
3
4
5
6
7
8
9
30
1
2
3
4
5
6
7
8
9
40
1
2
3
4
5
6
7
8
9
50
1
2
3
4
5
6
7
8
9
60
1
2
3
4
5
6
7
8
conditioned by S. codringtonii than to water conditioned by
Tilapia rendalli. Snails have been found to be able to distinguish
between threatening and non-threatening species of predators
(Kelly & Cory, 1987). Two freshwater snails Valvata piscinalis
(Müller) and Bithynia tentaculata (L.) were able to distinguish
between leeches, reacting only to the molluscivorous Glossiphonia complanata and not to the non-molluscivorous Erpobdella
octulata (Kelly & Cory, 1987). The recognition of harmless
species or organisms minimizes energy and time wasted due to
unnecessary escape behaviour (Townsend & McCarthy, 1980).
Chimbari (1996) investigated the behaviour of B. globosus and
Melanoides tuberculata (Müller) in the presence of S. codringtonii.
The results were inconclusive, but there was a greater tendency
for B. globosus to crawl above the waterline than M. tuberculata in
the presence of S. codringtonii.
The density of fish did not cause significant differences in
the response of snails. The observed results could also have
been affected by the size of the tanks because the water in the
tanks may have been ‘sufficiently’ conditioned by only one fish.
Marine snails have been shown to be able to detect potential
predators by distant chemoreception (Phillips, 1976; Schmitt,
1981; Watanabe, 1983). Few studies have demonstrated such
predator defence mechanisms in freshwater snails. The Florida
apple snail, Pomacea paludosa (Say) has been observed to
respond to the ‘scent’ of a predatory turtle by moving down to
the sediment and burying itself when a turtle comes into the
vicinity (Snyder & Snyder, 1969). Results from these experiments indicate that snails did not only respond to the presence
of S. codringtonii, but being exposed to water conditioned by
feeding fish induced an enhanced response. The escape
response of B. globosus to water conditioned by feeding fish
supports observations made for Physella sp. (Turner, 1996),
which increased refuge use when exposed to water conditioned
by crushed conspecifics. Snails can reduce the costs of predator
avoidance by avoiding false alarms and responding when predators pose a real threat (Phillips, 1978).
Bulinus tropicus and B. globosus were able to move into covered
areas where they were protected from predation by S. codringtonii. The high percentages (90%) of snails measured under the
tray during the study may be due to an anti-predation response
and an active refuge-seeking behaviour. However, this behaviour
could also have been assisted by disturbances from the fish causing snails to detach and sink to the bottom, where the refuge was
placed. Turner (1996) demonstrated that the behavioural
responses of Physella decreased mortality due to predation. The
snails were utilizing covered habitats more than crawling to the
surface, Physella was observed to crawl out of water when the
covered areas were absent. In the present study, few snails
Figure 4. Percentage of snails occupying different positions in an aquarium
in the presence of Sargochromis codringtonii and in control aquaria without
fish. Time 0 is 24 h after snails were introduced and when fish were introduced.
Table 4. Mean ( standard deviation) values of physical and chemical variables measured in tanks containing
Sargochromis codringtonii (n 7).
P-values
Repeat
Fish (n 21)
Control (n 14)
experiment
Treatment
0
25.09 0.79
25.50 0.72
ns
ns
8
25.39 0.81
25.30 0.59
ns
ns
17
26.04 0.80
25.29 0.62
ns
0.05
0
125 8
129 10
ns
ns
8
125 8
129 10
ns
ns
17
125 8
129 10
ns
ns
Time (h)
Temperature (°C)
Conductivity (mS)
pH
0
7.79 0.37
7.74 0.47
ns
ns
8
7.67 0.35
7.60 0.37
ns
ns
17
7.63 0.35
7.52 0.35
0.05
ns
ns, not significant (P 0.05).
357
P. MAKONI ET AL.
(about 9%) were observed to be moving out of the water in the
presence of S. codringtonii, the majority of snails were crawling
into the covered habitats.
The anti-predator responses of snails affect growth rates of
snails and delays reproduction (Yamanda, Navarrete & Needham, 1998; Crowl & Covich, 1990). During the experiments very
few snails returned to the water once they had left. The presence
of a predator will cause the snails to stay out of water, reducing
their growth and reproduction. Staying out of water for relatively longer periods might decrease the encounter rates
between miracidia and snails. However, since the snails actively
sought refuge in the presence of S. codringtonii, the use of fish to
control snails in areas with aquatic vegetation may be limited.
Macrophytes provide some cover for snails against predators
(Louda et al., 1984; Chimbari, Madsen & Ndamba, 1997), and
the distribution of intermediate snails is associated with aquatic
vegetation. For the molluscivorous fish to have an impact on
snails in ponds with aquatic vegetation, herbivorous fish have to
be introduced into the ponds (Bell-Cross & Minshull, 1988).
In conclusion, the results suggest that Bulinus globosus and
B. tropicus respond to water conditioned by S. codringtonii and
have an enhanced escape response when exposed to water conditioned by feeding fish compared with non-feeding fish. The
snails also actively sought refuge in the presence of S. codringtonii. The mechanisms that triggered the escape response of
snails were beyond the scope of this study. Therefore, there is
need to carry out further research to investigate the mechanisms that triggered the responses of the snails to water conditioned by S. codringtonii.
COVICH, A.P. 1981. Chemical refugia from predation for thin-shelled
gastropods in a sulphide-enriched stream. International Vereinigung
für Theoretische und Angwewandte Limnologie. Verhandlungen, 21:
1632–1636
CROWL, T.A. & COVICH, A.P. 1990. Predator-induced life history
shifts in a freshwater snail. Science, 247: 949–951
HOFKIN, B.H, KOECH, D.K. OUMA, J.H. & LOKER, E.S. 1991. The
North American crayfish Procambarus clarkii and the biological control of schistosome-transmitting snails in Kenya: laboratory and field
investigations. Biological Control, 1: 183–187
HOSMER, D.W. & LEMESHOW, S. 1989. Applied logistic regression. John
Wiley and Sons, New York.
KELLY, P.M. & CORY, J.S. 1987. Operculum closing as a defence against
leeches in four British freshwater prosobranch snails. Hydrobiologia,
144: 121–124.
LOUDA, S.M., MCKAYE, K.R., KOCHER, T.D. & STACKHOUSE, C.J.
(1984). Activity, dispersion, and size of Lanistes nyassanus and
L. solidus (Gastropoda, Ampullaridae) over the depth gradient at
Cape Maclear, Lake Malawi, Africa. Veliger, 26: 145–152.
OSENBERG, C.W. & MITTELBACH, G.G. 1989. Effects of body size on
predator-prey interaction between pumpkinseed sunfish and gastropods. Ecological Monographs, 59: 405–432.
PALMER, A.J. 1979. Fish predation and the evolution of gastropod shell
sculpture: experimental and geographic evidence. Evolution, 33:
697–713.
PHILLIPS, D.W. 1976. The effect of a species-species avoidance
response to predatory starfish on the intertidal distribution of two
gastropods. Oecologia, 23: 83–94.
PHILLIPS, D.W. 1978. Chemical mediation of invertebrate defensive
behaviours and the ability to distinguish between foraging and
inactive predators. Marine Biology, 49: 237–243.
SCHMITT, R.J. 1981. Contrasting anti-predator defences of sympatric
gastropods (family Trochidae). Journal of Experimental Marine Biology
and Ecology, 54: 251–263.
SNYDER, N.F.R. & SNYDER, H.A. 1969. Defences of the Florida apple
snail Pomacea paludosa. Behaviour, 40: 175–215.
TOWNSEND, C.R. & MCCARTHY, T.K. 1980. On the defence strategy
of Physa fontinalis (L.), a freshwater pulmonate snail. Oecologia, 46:
75–79.
TURNER, A.M. 1996. Freshwater snails alter habitat use in response to
predation. Animal Behaviour, 51: 747–756.
VERMEIJ, G.J. 1974. Marine faunal dominance and molluscan shell
form. Evolution, 28: 656–664.
VERMEIJ, G.J. & COVICH, A.P. 1978. Coevolution of freshwater gastropods and their predators. American Naturalist, 112: 833–843.
WATANABE, J.M. 1983. Anti-predator defences of three kelp forest
gastropods: contrasting adaptations of closely related prey species.
Journal of Experimental Marine Biology and Ecology, 71: 257–270.
YAMANDA, S.B., NAVARRETE, S.A. & NEEDHAM, C. 1998. Predation
induced changes in behaviour and growth rate in three populations
of the intertidal snail, Littorina sitkana (Philippi). Journal of Experimental Marine Biology and Ecology, 220: 213–226.
ACKNOWLEDGEMENTS
We are grateful to the Secretary for Health (Zimbabwe) for permission to conduct and publish this study. We thank D. Ndlela,
P. Sungai and S. Chadzonga for their technical support during
the course of this study. The study was funded by the Danish
International Development Agency (DANIDA) through the
Danish Bilharziasis Laboratory.
REFERENCES
ALEXANDER, J.E. & COVICH, A.P. 1991. Predator avoidance by the
freshwater snail Physella virgata in response to the crayfish Procambrus
simulans. Oecologia, 87: 435- 442.
BELL-CROSS G. and MINSHULL J. L. 1988. The fishes of Zimbabwe.
Typocrafters, Bulawayo.
CHIMBARI, M.J. 1996. Predation avoidance behaviour of Bulinus globosus
and Melanoides tuberculata (Gastropoda) in the presence of Sargochromis codringtonii (Cichlidae). PhD thesis, University of Copenhagen.
CHIMBARI, M.J., MADSEN, H. & NDAMBA, J. 1997. Simulated field
trials to evaluate the effect of Sargochromis codringtonii and Tilapia
rendalli on snails in the presence and absence of aquatic plants.
Journal of Applied Ecology, 34: 871–877
358