Complex interactions link the microbial

Notes
S. J., AND R. W. GRIGG. 1980. Impact of a kaolin
clay spill on a coral reef in Hawaii. Mar. Biol. 65: 269-276.
DREW, E. A. 1972. The biology and physiology of algae-invertebrate symbioses. 2. The density of symbiotic algal cells
in a number of hermatypic hard corals and alcyonarians
from various depths. J. Exp. Mar. Biol. Ecol. 9: 71-75.
DUBINSKY, Z., P. G. FALKOWSKI, AND K. Wm.
1986. Light
harvesting and utilization
by phytoplankton.
Plant Cell
Physiol. 27: 1335-l 349.
-,
AND OTHERS. 1990. The effect of nutrient resources on
the optical properties and photosynthetic efficiency of Stylophora pistillata. Proc. R. Sot. Lond. Ser. B 239: 23 l-246.
FALKOWSKI, P. G. 1980. Light and shade adaptation in marine
phytoplankton,
p. 99-l 19. In P. G. Falkowski [ed.], Primary productivity
in the sea. Plenum.
-,
AND Z. DUBINSKY.
1980. Light-shade adaptation of
Stylophora pistillata, a hermatypic coral from the Gulf of
Elat. Nature 289: 172-174.
___
-,
L. MUSCATINE, AND J. W. PORTER. 1984. Light
and bioenergetics of a symbiotic coral. Bioscience 34: 705709.
-,
L. McCXoSKEY, L. MUSCATINE, AND Z. DUBINSKY. 1993.
Population control in symbiotic corals. Bioscience 43: 60661 1.
GLYNN, P. W. 1988. El Nino-Southern
Oscillation 1982-l 983:
Near-shore, population,
community
and ecosystem responses. Annu. Rev. Ecol. Syst. 19: 309-345.
~
1993. Coral reef bleaching: Ecological perspectives.
Coral Reefs 12: 1-17.
-,
AND L. D’CROZ.
1990. Experimental evidence for high
temperature stress as the cause of El Nino-coincident
coral
mortality. Coral Reefs 8: 18 l-l 9 1.
-,
R. I-1,
K. SAKAI, Y. N-0,
AND K. YAMAzATO.
1992. Experimental response of Okinawan (Ryuku Islands,
Japan) reef corals to high sea temperature and UV radiation,
p. 27-37. Proc 7th Int. Coral Reef Symp. Guam.
GOREAU, T. F. 1964. Mass expulsion of zooxanthellae
from
Jamaican reef communities after hurricane Flora. Science
145: 383-386.
DOLLAR,
Limnol.
Oceanogr., 40(6),
0 1995, by the American
1995, 1173-1181
Society of Limnology
and Oceanography,
1173
-,
AND A. H. MACFAFUANE. 1990. Reduced growth rate
of Montastrea annularis following the 1987-1988 coralbleaching event. Coral Reefs 8: 2 17-224.
HOEGH-GULDBERG,
O., AND G. J. SMITH. 1989. Influence of
the population density of zooxanthellae and supply of ammonium on the biomass and metabolic characteristics of
the reef corals Seriatopora hystrix and Stylophora pistillata.
Mar. Ecol. Prog. Ser. 57: 173-186.
JAPP, W. C., AND J. WHEATON.
1975. Observation on Florida
reef corals treated with fish collecting chemicals. Fla. Mar.
Res. Publ. 10: 1-17.
JEFFREY, S. W., AND G. F. HUMPHREY.
1975. New spectrophotometric equations for determining chlorophylls a, b,
c 1 and c2 in higher plants, algae and natural phytoplankton.
Biochem. Physiol. Pflanz. 167: 19 l-l 94.
Jo-,
R. E., AND W. J. WIEBE. 1970. A method for determination of coral tissue biomass and composition. Limnol. Oceanogr. 15: 822-824.
MANDEL, J. 1964. The statistical analysis of experimental data.
Interscience.
MUSCATINE, L., P. G. FALKOWSKI, Z. DUBINSKY, P. A. COOK,
AND L. MCCLOSKEY.
1989. The effect of nutrient resources
on the population dynamic of zooxanthellae in a reef coral.
Proc. R. Sot. Lond. Ser. B 236: 3 1 l-324.
PRÉZELIN, B. B. 198 1. Light reactions in photosynthesis, p. l43. In Physiological bases of phytoplankton
ecology. Can.
Bull. Fish. Aquat. Sci. 2 10.
SCHONWALD, H., Y. ACHITUV, AND Z. DUBINSKY.
1987. Differences in the symbiotic interrelation between algae and
host in light- and dark-coloured colonies of the hydrocoral
Millepora dichotoma. Symbiosis 4: 17 l-l 84.
SZMANT, A. M., AND N. J. GASSMAN. 1990. The effects of
prolonged “bleaching”
on the tissue biomass and reproduction of the reef coral Montastrea annularis. Coral Reefs
8: 2 17-224.
Submitted: 15 September I993
Accepted: 23 March 1995
Amended: I May 1995
Inc.
Trophic relations between cyclopoid copepods and ciliated
protists: Complex interactions link the microbial and
classic food webs
Abstract-Two field experiments examined the effects of
cyclopoid copepods on ciliates. The presence or absence
of Cyclops abyssorum, Cyclops kolensis, and zooplankton
~64 pm was manipulated to determine the relative importance of direct cyclopoid predation on protists vs. indirect effects mediated through cyclopoid predation on other metazooplankton.
In the second experiment, presence
or absence of C. abyssorum was cross-classified with five
concentrations of the metazooplankton
community. Cyclopoid effects on ciliates were dependent on predator and
prey species and on the abundance of alternate prey for
cyclopoids. A trophic cascade was also observed, but only
for two small ciliates, and only with the larger C. abyssorum. C. abyssorum had a stronger predation effect on oli-
gotrich ciliates when metazooplankton
had been removed,
and this effect appeared at a lower metazooplankton
concentration with a larger ciliate, compared to a smaller species of the same genus. These results suggest that for cyclopoid+iliate
interactions,
switching behavior in the
predator may be at least as important as a trophic cascade.
The concept of a trophic
cascade structuring
aquatic
food webs has become a dominant
theme in aquatic ecology. Planktivorous
fish have been shown to reduce the
abundance of herbivorous
zooplankton,
thereby reducing
1174
Notes
grazing by zooplankton on algae, which results in increased algal biomass (HrbaCek 1962; Carpenter et al.
1985; Kerfoot 1987). Sprules and Bowerman (1988)
however, have suggested that omnivory-feeding
on multiple trophic levels-is common in aquatic communities.
Omnivory
has the potential to reduce the strength of
trophic cascades, if, for example, a zooplankter consumes
both algae and other algal grazers. Increased abundance
of the omnivore could either increase or decrease algal
biomass, depending on the relative strength of the interactions between the omnivore and herbivores, omnivores
and algae, and herbivores and algae.
Cyclopoid copepods are omnivorous predators that feed
in a selective, raptorial fashion and are known to prey on
rotifers, cladocerans, Calanoid copepods, and copepod
nauplii (Brand1 and Fernando 1975; Williamson
1980;
Stemberger 1985). Cyclopoids are usually described as
selecting smaller prey items from the available prey size
spectrum (Brand1 and Fernando 1975; Gliwicz and Umana
1994). However, factors such as predator hunger, prey
shape, hardness, and behavior, and the availability
of
alternate prey are at least as important as prey size (Li
and Li 1979; Williamson
1980; Stemberger 1985). In
general, cyclopoids select for soft-bodied species that lack
defensive behaviors, and they become more selective when
they are satiated. There is also evidence that at least some
adult cyclopoids ingest algae, primarily diatoms (Adrian
1991).
Cyclopoids are also capable of preying on ciliated protists (Williamson
1980; Wiackowski et al. 1994; Wickham in press). Although maximal predation rates can be
as high as 180 ciliates copepod- l h- l, such ingestion rates
are seen only at very high ciliate densities (150-200 cells
ml-r), which are rare in nature. Total planktonic ciliate
abundance is usually ~20 cells ml-l (Pace and Orcutt
198 1; Pace 1986; Beminger et al. 1993) and at these
concentrations cyclopoid predation rates are considerably
lower (l-l 0 ciliates copepod-’ h- ’ ; Wickham in press).
The copepod nauplii, cladocerans, and some rotifers that
cyclopoids prey upon are themselves capable of preying
on ciliates (see Sanders and Wickham 1993). Given the
selective and omnivorous nature of cyclopoid predation,
it is unclear whether the direct predation impact of cyclopoids on ciliates is offset by a trophic cascade where
cyclopoids prey on other metazooplankton,
which then
releases ciliates from predation pressure.
Ciliates can be major herbivores, and metazooplankton
predation on ciliates may be an important link between
the “classic” and microbial food webs. The microbial
food web consists of bacteria and autotrophic picoplankton, preyed upon by heterotrophic flagellates and ciliates,
which then remineralize nutrients that are reutilized by
algae and bacteria (Azam et al. 1983; Stockner and Porter
1988). At times, ciliates may have a grazing impact on
phytoplankton
equivalent to that of metazooplankton,
while also being the major consumers of heterotrophic
flagellates (Weisse et al. 1990). Flagellates are often the
major consumers of bacteria (Sanders et al. 1989; Pace
et al. 1990), so strong flagellate-ciliate and ciliate-metazooplankton links have the potential to display a trophic
cascade from metazooplankton
to bacteria. In this study,
I examined the direct and indirect effects of cyclopoid
copepods on ciliates, and whether these effects would be
transmitted to the rest of the microbial food web.
Two experiments were conducted in Schohsee, a moderately eutrophic lake in northern Germany, during fall
1993. The experimental containers were 2-liter PVC bottles, incubated in situ at 2-m depth. In both experiments,
the cyclopoids used in the experiments were obtained
from the lake the day before the experiment and incubated overnight in filtered lake water at 11°C. Only gravid
females were used to ensure that animals were of the same
sex, life stage, and roughly the same physiological condition, but all egg sacs were removed to prevent reproduction during the experiment. In both experiments, cyclopoid treatments had 10 cyclopoid liter- l.
The experimental design in the first experiment was
three levels of a cyclopoid treatment (either Cyclops abyssorum, Cyclops kolensis, or no cyclopoids), cross-classified with two levels of a zooplankton > 64-pm treatment
(presence or absence). There were three replicates per
treatment combination,
giving a total of 18 bottles (3
copepod treatments x 2 zooplankton treatments x 3
replicates). Both C. abyssorum and C. kolensis are known
to consume rotifers, copepod nauplii and copepodites,
and cladocerans. C. kolensis is a small species (mean metasome length of adult females used in the experiment,
0.724 mm; SE = 0.035; n = 104). C. kolensis has been
shown to consume algae in addition to metazoan prey,
and algae may constitute as much as 57% of its diet (Adrian 199 1). C. abyssorum is a larger species (mean metasome length of adult females used in the experiment, 1.20
mm; SE = 0.06 1; n = 94) and is highly predatory, consuming rotifers, copepod nauplii and copepodites, and
cladocerans at rates as high as 20 individuals d- l, about
twice the predation rate found for C. kolensis (Adrian
199 1; van den Bosch and Santer 1993).
To begin an experiment, I pumped water from 2-m
depth into a 90-liter container with a hand-operated bilge
pump. Some of the water was then poured through a 64pm mesh into a second container in order to obtain water
for the treatments lacking zooplankton >64 pm. Bottles
were filled in random order, adding copepods into the
appropriate bottles. The first experiment ran for 4 d (l5 October 1993), with a final water temperature of 12.3”C.
Three initial samples of water with and without zooplankton >64 pm were taken at the beginning, middle, and
end of filling the experimental bottles by filling 2-liter
bottles in the same manner as the experimental bottles.
Initial and final samples were processed in the same
manner. Bacteria and flagellated protists were enumerated by fixing 5-ml samples in 2% (final concn) glutaraldehyde, filtering DAPI-stained cells onto 0.2-pm black
polycarbonate filters, and counting them with epifluorescence microscopy. Autotrophic
flagellates were differentiated from heterotrophs by their autofluorescence. Ciliates were counted by scanning all of 50-ml settled, Bouin’sfixed samples on an inverted microscope. Chlorophyll a
was determined spectrophotometrically
after filtering 250
ml of water onto GF/F filters and using ethanol extraction
Notes
(Nusch and Palme 1975). Metazoan zooplankton were
sampled by passing the remaining 1.7 liters through a 30pm mesh, fixing in sucrose Formalin, and then settling
and counting on an inverted microscope.
Data were analyzed in a 3 x 2 factorial ANOVA. Rather than testing the hypothesis that there was no difference
between the three levels of the cyclopoid treatment (the
cyclopoid main effect), the differences between the nocyclopoid control and each of the two cyclopoid treatments were tested in two contrasts. To test whether cyclopoid effects were independent of the concentration of
other metazooplankton,
I used two preplanned contrasts
which tested the cyclopoid-metazooplankton
interaction
separately for each copepod. Specifically, the contrasts
had the null hypotheses that (C. abyssorum, zooplankton
present) - (no cyclopoids, zooplankton present) = (C.
abyssorum, no zooplankton) - (no cyclopoids, no zooplankton) and (C. kolensis, zooplankton present) - (no
cyclopoids, zooplankton present) = (C. kolensis, no zooplankton) - (no cyclopoids, no zooplankton), where the
text within parentheses represents the organisms contained in the different treatment combinations.
A second experiment was designed to determine whether there was a certain threshold zooplankton concentration at which C. abyssorum’s effect on ciliate abundance
changed from direct predation to enhancement through
a trophic cascade. The experiment was also run in 2-liter
bottles in the Schijhsee, incubated at 2-m depth. Five
concentrations ofzooplankton
>64 pm (0,0.25,0.5,0.75,
or 1 times the natural concn) were cross-classified with
the presence or absence of C. abyssorum. Where it was
present, C. abyssorum was added at 10 liter-l. Two replicates were used per treatment combination,
giving a
total of 20 bottles. The experiment was run for 6 d (27
October-2 November 1994) with a final water temperature of 8.4”C. Bottles were filled in a manner similar to
the first experiment. Water was pumped from 2-m depth
into a 90-liter container. In random order, appropriate
volumes of whole water and 64-pm filtered water were
added to each bottle to obtain the correct dilutions, adding copepods to half the bottles. Two initial samples were
taken for each dilution. Initial and final samples were
processed in the same manner as in the first experiment.
The data were analyzed in a 5 x 2 factorial ANOVA (five
zooplankton dilutions vs. cyclopoid presence or absence).
Differences in treatments with and without C. abyssorum
at the different zooplankton dilutions were ascertained
by conducting five preplanned contrasts. These were essentially the same as t-tests between the two cyclopoid
levels at each zooplankton dilution, but using the pooled,
experiment-wise, measure of variance in the denominator
(the mean square error). Linear regression was used to
test whether the dilutions had any effect on ciliate and
metazooplankton
initial abundances.
The ciliate community in the C. abyssorum-C. kolensis
experiment comprised 12 species, of which six made up,
on average, 87% of initial and 94% of final total abundance. The mean initial, total ciliate abundance was 3.0
ciliates ml- l. Removing zooplankton with the 64-pm mesh
did not significantly affect the initial ciliate densities (P
1175
= 0.18). Only four species (Strobilidium sp. 1, Strobilivelox, and Urotricha sp.) had
mean abundances > 1 cell ml-l in any treatment at the
end of the experiment. The metazoan zooplankton community (metazooplankton)
was comprised primarily of a
Calanoid copepod, Eudiaptomus sp., copepod nauplii, the
rotifer Keratella cochlearis, and low numbers of the cladocerans Bosmina longirostris and small (< 1.5 mm)
Daphnia galeata (cladoceran maximum final abundance
of 2 liter-l).
Numbers of Calanoid copepods, copepod
nauplii, and cladocerans were significantly
reduced by
screening the water through the 64-pm mesh (P < 0.05).
Although the abundance of K. cochlearis was reduced
from 10.5 to 5.6 ind. liter-’ by filtering water through
the 64-pm mesh, this reduction was not significant (P =
0.12), due to high variance.
Removal of zooplankton > 64 pm generally resulted in
higher ciliate numbers, compared to metazooplanktonpresent treatments (Fig. 1). In treatments without metazooplankton
(ignoring cyclopoid effects), total ciliate
abundance was 1.6 times higher than in treatments with
metazooplankton ( 15.6 vs. 9.7 ciliates ml- 1). If only treatments without cyclopoids are examined, in five of the six
major ciliate taxa present, treatments without metazooplankton had higher numbers of ciliates than treatments
with metazooplankton.
However, significant (P < 0.05)
main effects of metazooplankton
on ciliates were found
for only three species: Strobilidium spp. 1 and 2 and
Urotricha. The paucity of significant main effects is due
to interactions between metazooplankton
and cyclopoid
treatments.
The two cyclopoid species had clear effects on the ciliate
community, but these effects depended on the ciliate species, the cyclopoid species, and whether other metazooplankton were present. C. abyssorum had a similar impact
on two of the oligotrich ciliate species (Strobilidium sp.
1 and Strombidium sp.) and to a lesser extent on a third
oligotrich, Strobilidium velox. All three ciliates are relatively large. Strobilidium sp. 1 and Strombidium sp. both
- 30-35 pm in diameter, while S. velox is somewhat largwere not
er, -45-50 pm. When other metazooplankton
present, C. abyssorum had a clear negative impact on
oligotrich abundance compared to the no-cyclopoid controls (Fig. 1). However, when other metazooplankton were
present, C. abyssorum had no impact on these three ciliate
species. The C. abyssorum-zooplankton
interaction was
significant for Strombidium sp. and Strobilidium sp. 1 (P
< 0.05) while a similar but nonsignificant trend was seen
for S. velox (P = 0.10, Table 1).
C. abyssorum had an effect on two small ciliate species,
Urotricha sp. and Strobilidium sp. 2 (both ~20 pm in
diameter), that was also dependent on the presence or
absence of other metazooplankton.
When metazooplankton were absent, there was no difference in ciliate abundance between treatments with and without C. abyssorum
(Fig. 1). When metazooplankton
were present, however,
treatments with C. abyssorum had greater numbers of
these two ciliates than did treatments without C. abyssorum. The effect was significant for Strobilidium sp. 2
(P = 0.0007), and while the effect was not significant (P
dium sp. 2, Strobilidium
Notes
lensis reduced the abundance of two species, Strombidium
sp. and S. velox, either significantly or nearly significantly
(P = 0.038 and 0.086, respectively). This reduction was
independent of whether other metazooplankton
were
present or absent (P > 0.8, Fig. 1 and Table 1). In all
other ciliate species except for the small Strobilidium sp.
2, C. kolensis had no significant impact on the ciliates,
and the presence or absence of metazooplankton
had no
effect on this (Fig. 1, Table 1). The effect of C. kolensis
on Strobilidium sp. 2 is rather puzzling: C. kolensis had
a negative effect on Strobilidium sp. 2 when metazooplankton were present, but a positive effect when metazooplankton are absent. The same, though nonsignificant,
trend (P = 0.11) was also seen with the larger Strobilidium
sp. 1.
There were also differences in the h;yclopoid impact on
metazooplankton.
The larger C. abyssorum had a much
stronger impact than did the smaller C. kolensis. C. abyssorum’s impact on copepod nauplii and Keratella was
dependent on the metazooplankton
concentration (P <
0.05, Fig. 2). The impact of C. abyssorum on nauplii and
Keratella was proportionately greater when these animals
were abundant (in the whole-water treatments) than when
they were rare (in the 64-pm-screened treatments). The
same trend was seen for Daphnia and calanoids, but numbers of these organisms were low and variance was high,
resulting in nonsignificant
effects (Fig. 2, Table 1). C.
kolensis had no effect on calanoids or copepod nauplii (P
> 0.05) and had a weak effect on Keratella and Daphnia
(P = 0.092 and P = 0.05 1, respectively; Fig. 2).
Cyclopoid effects did not extend beyond metazooplankton and ciliates. Abundance of bacteria, autotrophic
flagellates, and heterotrophic flagellates was not affected
(P >
by either cyclopoid species or by metazooplankton
0.1). Autotrophic flagellate abundance was slightly lower
in treatments with cyclopoids than in noncyclopoid controls (2.67 x lo4 ml-l vs. 3.88 x lo4 ml-l), but even in
a nonprotected posthoc test, the difference was at best
-s
g 0.15
P
2 0.10
2 0.05
El
G 0.00
-: 0.20 1":-i
L
--..... .._..-...
.-....
, .. .
3.5
E 3.0
e 2.5
%
2.0
z 1.5
-? 1.0
2
g 0.5
z 0.0
0.6
6.0
-:
5
5.0
0.5
LO.4
-F 4.0
%
2 3.0
; 0.3
2 2.0
j
$ 1.0
Ko.1
v
0.0
0.0
0.2
absent
present
Zooplankton > 64 pm
present
absent
Zooplankton > 64 pm
Fig. 1. Final ciliate abundance in treatments with the presence or absence of Cyclops abyssorum,Cyclops kolensis, and
metazooplankton.
No cyclopoids added-Cl---O;
C. abyssorum present-A- - -A; C. kolensis present-00 . . . .0. Vertical
bars represent 1 SE.
= 0.11) for Urotricha, the trend was similar and the probability of seeing a significant effect, if one were present,
was low (power = 0.25).
The effect of C. kolensis on the ciliate community was
considerably simpler than that of C. abyssorum. C. ko-
Table 1. P-values of effects in the Cyclopsabyssorum-Cyclopskolensis-metazooplankton
experiment. The C. abyssorumand C. kolensis main effects represent contrasts testing for
significant differences between the means of no-cyclopoid treatments and each of the cyclopoid
treatments. The cyclopoid-metazooplankton
interactions are from the contrasts described in
the text.
Cyclopoid-metazooplankton interactions
Main effects
Dependent
variable
Strombidium sp.
Strobilidium velox
Strobilidium sp. 1
Strobilidium sp. 2
Urotricha sp.
Cyclidium sp.
Copepod nauplii
Calanoid copepods
Daphnia
Keratella cochlearis
C.
C.
abyssorum kolensis Zooplankton
0.005
0.038
0.123
0.005
0.086
0.227
0.002
0.087
0.067
0.3 15
<o.ooo 1
0.227
0.069
< 0.000 1
0.927
0.863
0.436
0.678
0.354
0.642
0.05 1
0.092
0.000 1
0.028
0.037
0.294
< 0.000 1
0.033
0.656
0.0004
C.
c.
abyssorum
kolensis
0.036
0.103
0.005
0.0007
0.112
0.04 1
< 0.000 1
0.140
0.384
0.042
0.842
0.839
0.106
0.006
0.379
0.773
0.0683
0.182
0.484
0.46 1
Notes
z5
$4
E3
&
62
.,,.,,j
-0
51
.,..
..,>
,,.’
30
ii
7
absent
present
Zooplankton > 64 pm
,/.
,:.
,,.’
H
I
u!
7
2
G
E
2
2
70
6.0
5.0
4.0
3.0
2.0
2
z
1.0
0.0
0
absent
present
Zooplankton > 64 pm
Fig. 2. Final abundance of the major metazooplankton groups
in the Cyclops abyssorum-Cyclops kolensis-metazooplankton
experiment. Symbols as in Fig. 1.
marginally significant (P = 0.105). Chlorophyll
a levels
were also independent of the cyclopoid treatment but not
of metazooplankton
concentration. Chl a concentrations
were slightly, but significantly, higher in treatments where
metazooplankton
were present (P = 0.000 1). Treatments
with metazooplankton
had 3.7 pg Chl a liter-l; Chl a in
treatments without metazooplankton was 3.07 pg liter-l.
Ciliate abundance was somewhat lower in the second,
C. abyssorum-metazooplankton
abundance experiment,
both at the beginning and in the peak final abundance.
Initial total ciliate abundance was 2.9 ciliates ml-l, and
there was no initial difference across treatments (P =
0.48). As designed, there was a positive linear relationship
between initial metazooplankton
densities and the proportion diluted (linear regression of metazooplankton
groups vs. dilution proportion: P < 0.05, R2 = 0.5 l-0.94).
In this experiment too, the effect of C. abyssorum on
ciliates was dependent on the ciliate species and the presence or absence of metazooplankton.
With only two replicates per treatment combination, variance was relatively
high and the power to see real differences low. Nevertheless, strong effects were observed. Both Strobilidium
sp. 1 and the larger S. velox showed differences in abundance between C. abyssorum-present and C. abyssorumabsent treatments that were dependent on the metazooplankton concentration (Fig. 3, Table 2). The metazooplankton concentration at which C. abyssorum no longer
reduced ciliate abundance compared to no-cyclopoid
treatments was different for the two ciliates. C. abyssorum
had a strong effect on Strobilidium sp. 1 only when there
were no other metazooplankton
present; at 0.25 or more
times the natural metazooplankton concentration, the difference between the C. abyssorum-present and C. abyssorum-absent treatments disappeared (Fig. 3, Table 2).
In contrast, C. abyssorum depressed the abundance of the
0
0.25 0.5
0.25 0.5
0.75
0.75
1
“L.
E
d
:
s
2
0.7
0.6
0.5
0.4
0.3
0.2
g
Co
0.1
0
0.25
0.5 0.75
1
0
0.25
0.5 0.75
1
0.5 0.i5
i
0
1
0.5,
-:
-2
;
2
5
1.0
0.8
0.6
0.4
0.2
z
E
0.3
g
0.2
5
0.1
o.oA
-:
0.0
0.4 b!!skL
0 0.25
0
0.25 0.5
0.75
1
Zooplankton > 64 pm
(proportion of natural community)
Zooplankton > 64 pm
(proportion of natural community)
Fig. 3. Final ciliate abundance in the Cyclops abyssorummetazooplankton
concentration
experiment.
C. abyssorum
present-o;
C. abyssorum absent-O. The x-axis is the proportion of unfiltered water (64 pm) used in the treatment. Vertical bars represent 1 SE.
larger ciliate species, S. velox, at all concentrations of
other metazooplankton
except the highest.
The treatment effects on the other four abundant ciliate
taxa were less clear. C. abyssorum had no effect on Strombidium sp. or Cyclidium sp. (P > 0.24), although increasing metazooplankton
concentration resulted in lower ciliate numbers relative to the no-metazooplankton
treatment (P = 0.016 and P = 0.017, respectively; Fig. 3).
Urotricha and the small Strobilidium sp. 2 also had lower
abundances with increasing metazooplankton concentration (P < 0.03). However, Strobilidium sp. 2, and to a
lesser extent Urotricha, had higher abundances at the lowest two metazooplankton
concentrations when C. abyssorum was present, rather than when it was absent. This
effect was similar to that seen with Cyclidium in the first
(C. abyssorum-C. kolensis) experiment.
C. abyssorum had clear effects on K. cochlearis and
copepod nauplii, but not on Calanoid copepodites and
adults or on cladocerans (Fig. 4). When naupliar abundance exceeded 5 ind. liter-l (at 0.5 times natural abundance), C. abyssorum suppressed nauplii relative to nonC. abyssorum treatments. A similar effect was seen with
K. cochlearis. Only when K. cochlearis levels in control
Notes
1178
Table 2. P-values of effects in the Cyclops abyssorum-metaozoplankton
concentration
experiment. The cyclopoid-metazooplankton
contrasts are from the contrasts described in
the text. The proportion of natural community is the proportion of unfiltered water (64 pm)
used in the treatment.
Main effects
Dependent
variable
c.
Zooplankabyssorum
ton
Strombidium sp.
Strobilidium velox
Strobilidium sp. 1
Strobilidium sp. 2
Urotricha sp.
Cyclidium sp.
0.332
< 0.000 1
Copepod nauplii
Calanoid copepods
Daphnia
Keratella cochlearis
0.00 1
0.002
0.365
0.243
0.0004
0.809
0.541
0.260
0.016
0.279
0.002
<0.0001
0.028
0.017
<o.ooo 1
0.004
0.055
0.04 1
treatments exceeded - 5 ind. liter- 1(at 0.7 5 times natural
abundance) was there a significant difference between
control and C. abyssorum treatments (Fig. 4, Table 2). In
contrast, C. abyssorum had no affect on Calanoid copepods at any of the abundances seen in the experiment.
Calanoid abundance increased with the increase in the
proportion of unfiltered water used in the treatment, but
Calanoid abundances in treatments with and without C.
abyssorum at any given dilution were nearly identical. As
in the first experiment, C. abyssorum also had no affect
on cladocerans (Bosmina and small Daphnia galeata), but
cladoceran abundance was much lower than Calanoid
abundance, and variance was much higher. In none of
the treatments did Daphnia abundance exceed 1.5 ind.
liter - 1.
These experiments show that cyclopoids can affect cil-
0
0.25
0.5
0.75
1
2.01
Zooplankton > 64 ,um
(proportion of natural community)
7
k
121
T
Zooplankton > 64 pm
(proportion of natural community)
Fig. 4. Final abundance of the major metazooplankton groups
in the Cyclops abyssorum-metazooplankton concentration experiment. Symbols as in Fig. 3.
Cyclopoid-metazooplankton
contrasts
-Proportion of natural community
0
0.25
0.538
0.0007
0.00 1
0.300
0.0023
0.186
0.00 1
0.176
0.310
0.810
1.ooo
0.496
0.734
0.009
0.066
0.359
1.ooo
1.000
1.ooo
0.734
0.5
0.75
0.23 1 0.790
0.009 0.007
0.415 0.137
0.804 0.185
0.638 0.680
0.605 0.917
0.05 1 0.002
0.613 0.757
0.018 0.060
0.433 0.093
1.0
0.263
0.764
0.107
0.960
0.638
0.755
0.0005
0.78 1
0.188
0.134
iate communities, and that these effects depend not only
on the cyclopoid and ciliate species, but also on the presence of alternate prey for cyclopoids. Depending on these
factors, ciliate response to cyclopoid addition was consistent with suppression due to direct cyclopoid predation
(e.g. C. abyssorum-S. velox when no other metazooplankton were present), enhancement through cyclopoid predation on other predators of ciliates (e.g. C. abyssorumUrotricha), or no response at all (e.g. C. kolensis-Urotricha). For C. abyssorum, the direct predation effects were
most pronounced with the larger ciliates. If only those
treatments in which other metazooplankton
had been
screened out are considered, in the first experiment, the
reduction in ciliate numbers with the addition of C. abyssorum was greatest for Strobilidium sp. 1, S. velox, and
Strombidium (all between 30 and 50 pm in size) and small
to nonexistent for Strobilidium sp. 2 and Urotricha (both
- 15-20 pm), while Cyclidium (- 20 pm long) abundance
was somewhat higher when C. abyssorum was added.
This size-dependent predation is consistent with the
observations of Wiackowski et al. (1994) who found that
Diacyclops imposed greater mortality on larger ciliates,
rather than on smaller ones. In contrast, Wickham and
Gilbert (199 1, 1993) and Jack and Gilbert (1993) found
that filter-feeding cladoceran zooplankton had their greatest affect on small ciliates. The difference is likely to be
due to a combination of cladocerans being relatively nonselective predators, especially in comparison to cyclopoids, and the larger ciliates being large enough to be
difficult for cladocerans to ingest. Burns’ (1968) regression
of maximum particle diameter cleared vs. Daphnia size
predicts that a 2-mm Daphnia could clear a particle no
larger than 49 pm. Although this was calculated with rigid
beads, and soft-bodied ciliates may be easier to ingest, it
is nonetheless plausible that larger ciliates suffer proportionately less predation than smaller ciliates from suspension-feeding cladocerans, while smaller ciliates are less
vulnerable to cyclopoid predation. However, Carrick et
al. (199 1) found that when the natural zooplankton com-
Notes
munity was made up of Calanoid copepods, small protists
were grazed at a higher rate than large protists. Therefore,
small size may be a refuge only from cyclopoid predation
and not from copepod predation in general.
There was some evidence for a trophic cascade from
cyclopoids to ciliates through metazooplankton,
but this
was present for only two small ciliate species, Urotricha
and Strombidium sp. 2, and only when C. abyssorum was
the top predator. For these two species, presence of C.
abyssorum had no affect when other metazooplankton
were absent, but C. abyssorum enhanced ciliate abundance when other metazooplankton
were present. This
combination of effects is also consistent with C. abyssorum confining its direct predation effects to larger ciliates
and metazooplankton.
It would seem that a cyclopoidmetazooplankton-ciliate
trophic cascade only appears
when the ciliates are small enough to be relatively immune from cyclopoid predation but vulnerable to predation by other metazooplankton.
For three of the other common species (Strombidium,
S. velox, and Strobilidium sp. I), the addition of C. abyssorum when other predators on ciliates were present did
not have the positive effect expected from a trophic cascade. Instead, C. abyssorum had no net effect when other
metazooplankton
were present and a strongly negative
impact in treatments where zooplankton
>64 pm had
been largely screened out (Fig. 1). The second experiment
suggests that when this change from no impact to a negative impact occurs, the switch is at a higher metazooplankton density for larger ciliates. C. abyssorum continued to depress S. velox abundance relative to no-cyclopoid controls at all metazooplankton
concentrations less
than the natural concentration.
In contrast, it was only
at the lowest proportion of the natural metazooplankton
community (where only K. cochlearis was present at an
abundance > 1 liter-l)
that C. abyssorum significantly
depressed the abundance of the much smaller Strobilidium sp. 1.
The change in the impact of C. abyssorum on ciliates,
depending on the abundance of other metazooplankton,
was consistent with switching behavior by C. abyssorum.
Marine Calanoid copepods are known to switch between
algal and metazoan prey, depending on the relative abundance of the two (Landry 198 l), but C. abyssorum in
these experiments seems to switch between ciliate and
metazoan prey. In the first experiment, C. abyssorum had
a much greater impact on copepod nauplii, Daphnia, and
Calanoid copepods when they were abundant compared
to when they were rare. Clearance rates can be calculated
from the differences in the natural logarithms of the final
abundances in the Cyclops and no-Cyclops treatments,
and from clearance rates, electivities can be calculated
(Chesson 1983). Electivities range from - 1 to + 1 and
are a measure of whether a prey item is grazed in proportion to its relative abundance.
Although some caution must be used when interpreting
clearance rates or electivities calculated from low numbers of prey, the data do support a switching hypothesis.
In the whole-water treatments, where metazooplankton
1179
were abundant, C. abyssorum had a negative electivity
for Strombidium, Strobilidium sp. 1, and S. velox (electivity = -0.6, -0.7, and -0.3), indicating that these prey
were grazed at rates less than would be expected from
their relative abundance. In the 64-pm-screened treatments, where metazooplankton
were rare, C. abyssorum
grazed these three ciliates at rates higher than expected
by their relative abundance (electivity = 0.3, 0.2, and
0.4). In contrast, the electivity for nauplii, Daphnia, and
calanoids was negative in the 64-pm-screened treatments,
where they were rare (electivity = -0.4, -0.3, and - l),
but in the whole-water treatments, where metazooplankton were abundant, electivity for Daphnia was positive
(electivity = 0.4) and it was neutral for nauplii and calanoids (electivity = 0). Thus, it would seem that C. abyssorum captures ciliate prey when metazooplankton
are
rare, but largely ignores them when metazooplankton are
abundant.
In the second experiment, the alternate prey for C.
abyssorum appeared to be copepod nauplii and K. cochlearis. K. cochlearis has been described as being selected
against by Diacyclops (Stemberger 1985), but in that study
the alternate prey was Synchaeta, a soft-bodied rotifer
that is considerably larger than the ciliates seen in my
study. Although Keratella may not be an optimal prey
item for C. abyssorum, choices of alternate prey were
limited. Had there been higher abundances of small cladocerans and rotifers such as Synchaeta, that are known
to prey on ciliates but are also vulnerable to cyclopoid
predation, then it is possible that a trophic cascade would
have been more evident. However, the evidence from
these experiments suggests that while the metazooplankton present were capable of reducing ciliate abundance,
simply adding an invertebrate predator is not necessarily
enough to induce a trophic cascade.
Ciliate size alone is clearly not sufficient to explain all
the differences in the impact of cyclopoids on ciliates,
even when there were no differences in the metazooplankton community.
Cyclidium in the first experiment and
Urotricha and Strobilidium sp. 2 in the second experiment
are more abundant when C. abyssorum is present and
metazooplankton
are absent than when both C. abyssorum and metazooplankton are absent. All three of these
species are small (5 20 pm), but this effect is consistent
with neither direct predation nor a trophic cascade. These
three ciliates are likely to be capable of preying on bacteria
as well as flagellates (Fenchel 1980; Sherr et al. 1986) and
may have benefited from some combination of nutrient
regeneration enhancing bacterial and flagellate production and a reduction in other, larger predators on flagellates. However, the design of the experiments was such
that these simultaneous effects, if they were occurring,
could not be ascertained.
In general, the smaller C. kolensis did not have an
impact on either ciliates or other metazooplankton
that
was strongly dependent on metazooplankton
concentration. C. kolensis had moderate affects on the abundance
of Keratella and Daphnia, but these effects were independent of the relative abundance of these two species.
1180
Notes
Similarly, the effect of C. kolensis on Strombidium and
S. velox was independent of other metazooplankton,
and
there was no affect at all on Cyclidium or Urotricha. However, C. kolensis reduced the abundance of Strombidium
sp. and S. velox by about the same amount when these
two species were abundant as when they were relatively
rare (Fig. 1). Therefore, while the ingestion rates may be
equal, the clearance rate for these two ciliates was higher
when they were less abundant. Similarly, Strobilidium
spp. 1 and 2 were slightly more abundant when both C.
kolensis and metazooplankton were absent, but slightly
reduced relative to no-cyclopoid treatments when only
metazooplankton
were present (Fig. 1). In contrast to C.
abyssorum, C. kolensis seems to ingest ciliates at a rate
that is largely independent of either alternate prey or ciliate abundance.
Both cyclopoid species had strong impacts on ciliates
and metazooplankton,
but this did not cascade down as
far as flagellates or bacteria. A trophic cascade extending
as far as bacteria seems to occur only when there are high
abundances of Daphnia. Moderate changes in Daphnia
density or changes in metazooplankton
other than Daphnia does not result in changes in bacterial abundance,
even where there are changes in Protist abundance (Pace
and Funke 199 1; Wickham and Gilbert 199 1, 1993). In
my experiments, Daphnia abundance was low, and addition of cyclopoids did not produce large changes in their
abundance. The more abundant calanoids and copepod
nauplii ingest flagellates at considerably lower rates than
do Daphnia (reviewed by Sanders and Wickham 1993)
and are unlikely to be major bacterial grazers (Sanders et
al. 1989).
There was also no trophic cascade from the cyclopoids
through the other metazooplankton,
leading to an increase in chlorophyll a. Instead, in the first experiment,
chlorophyll a values were marginally (0.63 hg liter-‘, or
17%) but significantly lower when metazooplankton
had
been reduced and ciliate abundance enhanced. This suggests that the approximate 6 cells ml- 1reduction in ciliate
numbers between treatments with and without metazooplankton has a greater impact on algal biomass than the
change in cladoceran, rotifer, and copepod numbers. Ciliate clearance rates are quite variable, but 6 ciliates ml- I,
each clearing 5 ~1 h- l, would clear 3% of the water column
per hour (Jonsson 1986). This estimated impact is consistent with the findings of Weisse et al. (1990), who reported ciliates to be the major herbivore group during the
early part of the spring phytoplankton
bloom in Lake
Constance. In the second experiment, the addition of C.
abyssorum did not produce increased chlorophyll a levels,
but maximum total ciliate abundance was only 59% of
that in the first experiment, so there was less scope for
increased ciliate herbivory. In both experiments, however, metazooplankton
numbers were low, increasing the
relative importance of ciliate herbivory and reducing the
potential for a strong trophic cascade.
Although trophic cascades may structure many aquatic
communities,
Sprules and Bowerman (1988) have documented the prevalence of omnivory in aquatic systems.
In this study, the cyclopoids seemed to switch their prey
preferences among ciliates and metazooplankton,
depending on relative abundances. At least some of the
metazooplankton
were capable of preying on algae, ciliates, and flagellates. Ciliates were likely to have preyed
on algae, flagellates, and bacteria. Given that the Cyclops
in these experiments had such varied effects on different
herbivores, it is perhaps not surprising that there is no
clear cascade from carnivorous cyclopoids through other
metazooplankton
and ciliates, resulting in an increase in
algae or bacteria.
Stephen A. Wickham’
Max-Planck-Institut
Postfach 16 5
24302 Plijn
Germany
ftir Limnologie
References
199 1. The feeding behaviour of Cyclops kolensis
and C. vincinus (Crustacea, Copepoda). Int. Ver. Theor.
Angew. Limnol. Verh. 24: 2852-2863.
AZAM, F., AND OTHERS. 1983. The ecological role of watercolumn microbes in the sea. Mar. Ecol. Prog. Ser. 10: 257263.
BERNINGER, U.-G., S. A. WICKHAM, AND B. J. FINLAY.
1993.
Trophic coupling within the microbial food web: A study
with fine temporal resolution in a eutrophic freshwater ecosystem. Freshwater Biol. 30: 4 19-432.
BRANDL, Z., AND C. H. FERNANDO. 1975. Food consumption
and utilization in two freshwater cyclopoid copepods (Mesocyclops edax and Cyclops vicinus). Int. Rev. Gesamten
Hydrobiol. 60: 47 l-494.
BURNS, C. W. 1968. The relationship between body size of
filter-feeding Cladocera and the maximum size of particle
ingested. Limnol. Oceanogr. 13: 675-678.
1985.
CARPENTER, S. R.,J. F. KITCHELL, AND J. R. HODGSON.
Cascading trophic interactions
and lake productivity.
Bioscience 35: 634-639.
CARRICK, H.J.,G.L.
FAHNENSTIEL, E. F. STOERMER, AND R.G.
WETZEL.
199 1. The importance of zooplankton-protozoan trophic couplings in Lake Michigan. Limnol. Oceanogr. 36: 1335-1345.
CHESSON, J. 1983. The estimation and analysis of preference
and its relationship to foraging models. Ecology 64: 12971304.
FENCHEL, T. 1980. Suspension feeding in ciliated protozoa:
Functional response and particle size selection. Microb.
Ecol. 6: l-l 1.
GLIWICZ, Z. M., AND G. UMANA.
1994. Cladoceran body size
and vulnerability
to copepod predation. Limnol. Oceanogr.
39: 4 19-424.
HRBACEK, J. 1962. Species composition
and the amount of
ADRLAN, R.
’ Present address: Insitut fur bkologie, Universitat
Greifswald, Schwedenhagen 6, 18565 Kloster/Hiddensee,
Germany.
Acknowledgments
I thank Dietmar
Barbara Santer for
Beminger, William
ments on an earlier
Lemke for the chlorophyll measurements,
help in copepod identification,
and Uhike
DeMott, and Winfried Lampert for comversion of the manuscript.
Notes
zooplankton in relation to the fish stock. Rozpr. Cesk. Akad.
Ved Rada Mat. Prir. Ved 72: l-l 16.
JACK, J., AND J. J. GILBERT. 1993. Susceptibilities of differentsized ciliates to direct suppression by small and large cladocerans. Freshwater Biol. 29: 19-29.
JONSSON, P. R. 1986. Particle size selection, feeding rates and
growth dynamics of marine planktonic oligotrichous ciliates (Ciliophora: Oligotrichina).
Mar. Ecol. Prog. Ser. 33:
265-277.
KERFooT,
W. C. 1987. Cascading effects and indirect pathways, p. 57-70. In W. C. Kerfoot and A. Sih [eds.], Predation: Direct and indirect impacts on aquatic communities. New England.
LANDRY, M. R. 198 1. Switching between herbivory and carnivory by the planktonic marine copepod Calanuspaczjicus.
Mar. Biol. 65: 77-82.
LI, J. L., AND H. W. LI. 1979. Species-specific factors affecting
predator-prey interactions of the copepod Acanthocyclops
vernalis with its natural prey. Limnol. Oceanogr. 24: 613626.
NUSCH, E. A., AND G. PALME.
1975. Biologische Methoden
fur die Praxis der Gewasseruntersuchung.
Wasser/Abwasser 116: 562-565.
PACE, M. L. 1986. An empirical analysis of zooplankton community size structure across lake trophic gradients. Limnol.
Oceanogr. 31: 45-55.
AND E. FUNKE. 199 1. Regulation of planktonic
microbial communities by nutrients and herbivores. Ecology
72: 904-9 14.
~
G. B. MCMANUS, AND S. E. G. FINDLAY. 1990. Planktonic community structure determines the fate of bacterial
production in a temperate lake. Limnol. Oceanogr. 35: 795808.
AND J. D. ORCUTT. 198 1. The relative importance of
protozoans, rotifers, and crustaceans in a freshwater zooplankton community. Limnol. Oceanogr. 25: 822-830.
SANDERS,R. W., K. G. PORTER, S. J. BENNER, AND A. E. DEBME.
1989. Seasonal patterns of bacterivory by flagellates, ciliates, rotifers, and cladocerans in a freshwater planktonic
community. Limnol. Oceanogr. 34: 673-687.
1181
-,
AND S. A. WICKHAM. 1993. Planktonic protists and
metazoa: Predation, food quality and population control.
Mar. Microb. Food Webs 7: 197-223.
SHERR, E. B.,B. F. SHERR, R. D. FALLON, AND S.Y. NEWELL.
1986. Small, aloricate ciliates as a major component of the
marine heterotrophic nanoplankton. Limnol. Oceanogr. 31:
177-183.
SPRULES, W. G., AND J. E. BOWERMAN. 1988. Omnivory
and
food chain length in zooplankton food webs. Ecology 69:
418-426.
STEMBERGER, R. S. 1985. Prey selection by the copepod Diacyclops thomasi. Oecologia 65: 492-497.
STOCKNER, J. G., AND K. G. PORTER. 1988. Microbial food
webs in freshwater planktonic ecosystems, p. 69-84. In S.
R. Carpenter [ed.], Complex interactions in lake communities. Springer.
VAN DEN BOSCH, F., AND B. SANTER. 1993. Cannibalism
in
Cyclops abyssorum. Oikos 67: 19-28.
WEISSE, T., AND OTHERS. 1990. Response of the microbial loop
to the phytoplankton spring bloom in a large prealpine lake.
Limnol. Oceanogr. 35: 78 l-794.
WIACKOWSKI, K., M. T. BRETT, AND C. R. GOLDMAN.
1994.
Differential effects of zooplankton species on ciliate community structure. Limnol. Oceanogr. 39: 486-492.
WICKHAM, S. A. In press. Cyclops predation on ciliates: Species-specific differences and functional responses. J. Plankton Res.
-,
AND J. J. GILBERT.
1991. Relative vulnerabilities
of
natural rotifer and ciliate communities to cladocerans: Laboratory and field experiments. Freshwater Biol. 26: 77-86.
-,
AND -.
1993. The comparative importance of
competition and predation by Daphnia on ciliated protists.
Arch. Hydrobiol. 126: 289-3 13.
WILLIAMSON, C. E. 1980. The predatory behavior of Mesocyclops edax: Predator preferences, prey defenses, and starvation-induced
changes. Limnol. Oceanogr. 25: 903-909.
Submitted: 2 August 1994
Accepted: 27 March 1995
Amended: 3 May 1995