The role of subtropical zooplankton as grazers of phytoplankton under different predation levels

The role of subtropical zooplankton as grazers ofphytoplankton under different predation levels

GISSELL LACEROT* , †, ‡, CARLA KRUK ‡, § , MIQUEL LURLING † AND MARTEN SCHEFFER †

*Ecologıa Funcional de Sistemas Acuaticos, Centro Universitario de la Region Este, Universidad de la Republica, Rocha, Uruguay†Department of Aquatic Ecology and Water Quality Management, Wageningen University, Wageningen, The Netherlands‡Ecologıa Funcional de Sistemas Acuaticos, Limnologıa, IECA, Facultad de Ciencias, Universidad de la Republica, Montevideo, Uruguay§Laboratorio de Etologıa, Ecologıa y Evolucion, Instituto de Investigaciones Biologicas Clemente Estable, Montevideo, Uruguay

SUMMARY

1. Large zooplankton such as Daphnia play a fundamental role as consumers of phytoplankton in

temperate lakes. These organisms are scarce in subtropical lakes where smaller cladocerans or

copepods take this niche. However, such smaller grazers appear to be less able to exert an effective

top–down control on the phytoplankton community.

2. We experimentally analysed the ability of zooplankton typical of subtropical, nutrient-rich lakes

to graze effectively on the phytoplankton community. We conducted two outdoor mesocosm

experiments in a hypertrophic lake, with combinations of three different zooplankton densities

and three different omnivorous fish densities. In the first experiment, the zooplankton community

was dominated by a small-sized cladoceran (Moina micrura) and in the second by a calanoid

copepod (Notodiaptomus incompositus). The phytoplankton community also differed between

experiments, with dominance of large size classes and less palatable species in the first experiment

and edible sizes in the second.

3. In both experiments, the effect of fish on the largest zooplankton was strong and negative, and

low fish densities were sufficient to eliminate the larger zooplankton. Fish presence had positive

effects on the biovolume of the largest phytoplankton size fraction (30–100 lm) in the first

experiment. This effect was more pronounced in combination with high zooplankton biomass,

suggesting that nutrient recycling by both fish and zooplankton may have been an important

mechanism promoting phytoplankton growth.

4. None of the zooplankton communities tested had significant top–down effects on the

phytoplankton community. In view of the phytoplankton species that dominated the communities

at the end of both experiments, inedibility, toxicity and antigrazer defences may explain the

absence of significant effects of zooplankton grazing.

5. Our results support the idea that in subtropical nutrient-rich lakes, drastic removal of small

omnivorous fish may be needed to allow an increase in zooplankton biomass. In addition, our

results imply that for such a change to result in effective top–down control of phytoplankton, a

shift in zooplankton community composition is essential too, as the experimental increase in

small-sized grazers had little effect on the phytoplankton communities.

Keywords: biomanipulation, grazing, subtropical, top–down control, trophic cascade

Introduction

In temperate lakes, the key role of large Daphnia in trophic

cascades is well known (Carpenter, Kitchell & Hodgson,

1985). Increasing their abundance is an important goal of

the biomanipulation techniques meant to improve trans-

parency in eutrophic lakes (Perrow et al., 1997). Even

though Daphnia is present in subtropical regions, they

often are smaller than in comparable temperate lakes

(Gillooly & Dodson, 2000; Lacerot, 2010). Rather, typical

Correspondence: Gissell Lacerot, Centro Universitario de la Region Este, Universidad de la Republica, Rincon esq. Florencio Sanchez, Rocha,

Uruguay. E-mail: [email protected]

Freshwater Biology (2013) 58, 494–503 doi:10.1111/fwb.12075

494 � 2012 Blackwell Publishing Ltd

representatives of the pelagic mesozooplankton in sub-

tropical lakes are smaller Cladocera (e.g. Moina, Cerio-

daphnia and Bosmina, and calanoid copepods; Crisman &

Beaver, 1990; Jeppesen et al., 2007; Havens & Beaver,

2011). The effect of grazing by zooplankton on phyto-

plankton is related to their body size as well as their

taxonomic composition (Cyr & Curtis, 1999). Copepods

can eat larger particles than some cladocerans (Peters &

Downing, 1984) and use a mixture of passive and active

strategies to collect small and large particles, respectively

(Vanderploeg, 1981). Small cladocerans and copepods

feed on a narrower size range of algae and have lower

grazing rates than large Daphnia on edible algae (Cyr &

Curtis, 1999). In view of these differences, the effective-

ness of the smaller subtropical zooplankton as grazers of

phytoplankton is believed to be limited compared to

temperate zooplankton. However, grazing in subtropical

communities has also been much less studied than in

temperate systems.

Fish predation is a major factor controlling crustacean

zooplankton in subtropical lakes (Jeppesen et al., 2007;

Havens et al., 2009), as fish communities are dominated

numerically by small omnivores (Sunaga & Verani, 1997;

Meerhoff et al., 2007; Fernandes et al., 2009; Teixeira-de

Mello et al., 2009) with short lifespan, early maturation,

high growth rates and high reproductive frequencies

(Lowe-McConnell, 1999; Blanck & Lamouroux, 2007; Van

Leeuwen et al., 2007), and prolonged spawning season

(Lappalainen & Tarkan, 2007). Hence, it seems logical that a

sufficient reduction in the predation pressure on zooplank-

ton by removal of fish could promote top–down control of

phytoplankton biomass. In principle, release from fish

predation should increase mesozooplankton abundance in

subtropical lakes, and indeed, some examples confirm this

possibility (Iglesias et al., 2008, 2011). However, such an

increase in mesozooplankton body size, biomass or shift in

taxonomic composition towards large Daphnia (>1.5 mm)

is not always observed (Crisman & Beaver, 1990).

Therefore, controlled experiments are needed to explore

whether such biomanipulation methods applied in tem-

perate systems might work in shallow (sub)tropical lakes

too (Rondel et al., 2008). For instance, it might be that the

absence of large cladocerans in (sub)tropical lakes is due

not only to fish predation, but also to the presence of

inedible cyanobacteria (Havens et al., 2000) or physiolog-

ical effects of higher water temperature (Crisman, Philips

& Beaver, 1995).

Our aim in this study was to explore the effectiveness of

zooplankton typical of subtropical, nutrient-rich lakes to

suppress the phytoplankton community by manipulating

the level of zooplankton and fish abundance. To this end,

we analysed two different zooplankton communities at

three density levels in two outdoor mesocosm experi-

ments, one dominated by a small-sized cladoceran (Moina

micrura) and the other by a calanoid copepod (Notodia-

ptomus incompositus), reflecting a spring and summer

community. The underlying hypothesis is that in the

absence of fish predation, high density of zooplankton

will reduce phytoplankton biomass, while fish presence

will hamper such phytoplankton control.

Methods

Our research was conducted in Lake Rodo, a hypertrophic

lake under restoration in Montevideo, Uruguay (35�55¢S56�10¢W). Lake characteristics and restoration techniques

applied during the period 1997–2001 are described else-

where (Scasso et al., 2001; Kruk et al., 2002; Rodrıguez-

Gallego et al., 2004). We conducted two experiments

(A and B) in 80-L transparent plastic mesocosms (mouth

diameter of 32 cm and 1 m depth). The bags were kept

open to the atmosphere and hung from a floating frame

(Table 1). Bags were not open to the sediment. In both

experiments, we first filled the bags with equal amounts

of lake water filtered through a 50-lm-sieve. This mesh

size was used to remove larger zooplankton, while

maintaining the phytoplankton size fractions as in the

lake. We then constructed three different zooplankton

densities: (i) a control with no large zooplankton (Z0),

(ii) densities similar to what was found in the lake at the

moment of the experiment (Z1) and (iii) high densities of

10· those in the lake (Z2; Table 1). For Z2, based on

Table 1 Schematic representation of the experimental design, indi-

cating the different combinations of fish and zooplankton in each

experiment (A and B). All treatments were replicated three times and

randomly assigned to the enclosures. Z0 = no large zooplankton,

Z1 = zooplankton in densities similar to those of the lake,

Z2 = zooplankton in densities 10· those of the lake, P0 = no fish,

P1 = low fish densities, P2 = high fish densities, D = treatment with

only Daphnia obtusa specimens.

Z0

P0 A&B A&B A&B

A&B A&B A&B

A&B A&B A&B only A

only A

only A

P1

P2

Z1 Z2 D

Zooplankton grazing in subtropical lakes 495

� 2012 Blackwell Publishing Ltd, Freshwater Biology, 58, 494–503

filtration rates for several zooplankton species reported in

Reynolds (1984), we calculated the community abundance

needed to obtain a community filtration rate high enough

to control phytoplankton under our experimental condi-

tions. The zooplankton added to the enclosures came from

a concentrate obtained after repeated 68-lm-net tows in

the lake. First, we took a sample from this concentrate to

estimate the initial density of each taxonomic group (i.e.

rotifers, copepods, nauplii and cladocerans) and then

calculated the volume of concentrate needed to construct

each treatment. We took the different volumes of the

concentrate in duplicates, of which one was counted to

confirm whether the zooplankton density was similar to

our calculations. The second was added to the designated

enclosure. In the case of fish, we added Cnesterodon

decemmaculatus in three densities: (i) no fish (P0), (ii) low

densities (=four fish, P1) and (iii) high densities (=10 fish,

P2). Cnesterodon decemmaculatus is a small-bodied poeciliid

with a broad distribution in subtropical South America

(Rosa & Costa, 1993). It is a visual-feeding omnivore with

a high preference for large zooplankton (Quintans et al.,

2009; Quintans, Scasso & Defeo, 2010) and can reach

extremely high abundances (Scasso et al., 2001). All fish

were acclimated in separate bags before addition. We

found similar light conditions in the lake and the different

enclosures in both experiments (light attenuation coeffi-

cient, Kd). Thus, we expected that visual predation by fish

was not affected by differences in transparency caused by

our experimental design (experiment A: Kdlake = 8.3,

Kdenclosures 7.7–11.5; experiment B: Kdlake = 4.3, while

Kdenclosures 3.2–5.5).

All treatments were randomly assigned to the enclo-

sures and replicated three times. In total, we had nine

different combinations of fish and zooplankton abun-

dances in each experiment, with their corresponding

replicates. Experiment A had three additional combina-

tions (see below, and Table 1). Occasionally, replicates

were lost due to problems in the mesoscosms. However,

except for one case (see results), we always counted with

at least two replicates for analysis. Both experiments (A

and B) were run for 5 days. The zooplankton and

phytoplankton communities were different in the two

experiments (Table 2; Fig. 1).

Experiment A

This experiment was run in spring 2000. The phytoplank-

ton community was dominated by filamentous cyanobac-

teria, particularly Aphanizomenon gracile (maximum linear

dimension, MLD = 90.7 lm), which constituted more

than 95% of total phytoplankton community biovolume

(Fig. 1, Table 2). Cladocerans were the dominant grazers

in the natural zooplankton community, and Moina micrura

(average size = 0.60 mm) was the dominant species

(Table 2). Since cladocerans in this experiment were

small, we included an extra treatment where only Daphnia

obtusa was present (D; average size = 1.5 mm) in order to

allow comparison with a larger grazer (Table 1). Daphnia

obtusa specimens for this extra treatment were obtained

from cultures. Daphnia obtusa in the cultures were fed with

yeast and starved for 1 day prior to the beginning of the

experiments.

Experiment B

This experiment was run in summer 2002. Phytoplankton

biovolume was higher than in the previous experiment,

Table 2 Characteristics of the lake and the phytoplankton and zoo-

plankton communities at the beginning of each experiment. MLD,

maximum linear dimension

Experiment A Experiment B

Phytoplankton

Number of species 27 27

Dominant species

(MLD lm) and

relative biovolume (%)

Aphanizomenon

gracile (90.7)

(95.4%)

Monoraphidium

griffithii (28.0)

(76.3%)

Co-dominant

species (relative

biovolume %)

Cryptomonas

sp. (1.6%)

Synedra acus (5.1%),

Desmodesmus

quadricaudata

(5.0%)

Average and standard

deviation of total

phytoplankton

biovolume (mm3 L)1)

29.29 ± 9.55 110.8 ± 9.38

Dominant size

fraction (lm)

30–100 10–30

Zooplankton

Number of species 18 –

Dominant species

(average length lm)

Moina

micrura (600)

Notodiaptomus

incompositus

(800)

Total biomass

(lgDW L)1)

49.5 44.0

Dominant size

fraction (biomass

lgDW L)1)

20.9 33.7

Lake

Maximum depth (m) 0.80 0.80

Secchi disk (m) 0.20 0.40

Temperature (�C) 19.9 20.2

Dissolved

oxygen (mg L)1)

12.4 17.2

pH 8.13 8.13

Conductivity (lS cm)1) 812 939

NH4-N (lg L)1) 101.2 28.9

PO4-P (lg L)1) 16.7 <10

496 G. Lacerot et al.


and dominated by the smaller chlorophyte Monoraphidium

griffithii (MLD = 28 lm), that reached 76% of the total

phytoplankton biovolume (Table 2; Fig. 1). Zooplankton

community biomass was also lower than in experiment A,

but in this case the calanoid copepod Notodiaptomus

incompositus (average size = 0.80 mm) was the dominant

grazer (Table 2).

Lake and enclosures’ sampling

The lake and each enclosure were sampled at the begin-

ning and end of each experiment. We measured temper-

ature (T, �C), dissolved oxygen (DO, mg L)1), conductivity

(K, lS cm)1), pH and Secchi disk depth (Secchi, cm). We

also took water samples with a 1-L Ruttner bottle, to

estimate soluble reactive phosphorus (PO4-P, lg L)1),

nitrate (NO3-N, lg L)1) and ammonium (NH4-N, lg L)1)

following standard methodology (Murphy & Riley, 1962;

Koroleff, 1970). N : P was estimated as the sum of NH4-N

and NO3-N divided by PO4-P in lg L)1. We estimated

chlorophyll-a (Chla, lg L)1) following the method of

Nusch (1980).

Plankton sampling and enumeration

At the end of both experiments, we took water samples

from each mesocosm for phytoplankton and zooplankton

analysis. Phytoplankton samples were taken with a 1-L

Ruttner bottle and preserved in Lugol’s solution. The

remaining water in each enclosure (70–80 L) was filtered

through a 50-lm sieve and preserved in 4% neutralised

formaldehyde for zooplankton analysis. Phytoplankton

units (cells and colonies mL)1) were counted in random

fields using the settling technique (Utermohl, 1958) in

1-mL Sedgewick-Rafter chambers, as recommended for

phytoplankton samples with high concentration (Lund,

Kipling & Le Cren, 1958; McAlice, 1971; Legresley &

McDermott, 2010). The samples were counted at 400· until

we reached at least 100 individuals of the most frequent

species (Lund et al., 1958; McAlice, 1971). Organism

dimensions, including MLD (lm), were estimated for

volume (V, lm3) and surface (S, lm2) calculations.

Organisms were measured at 400· during counting and

at 1000· using concentrated samples. Phytoplankton

biovolume was approximated according to Hillebrand

et al. (1999). We calculated population biovolume

(mm3 L)1) as the individual volume of the species mul-

tiplied by the abundance of individuals. Then, we calcu-

lated the relative percentage of each species biovolume to

the total biovolume, and we classified a species as

dominant if it reached at least 30% of the total biovolume.

We classified phytoplankton species into size classes

according to their MLD. The size classes were selected to

represent the main growth strategies of phytoplankton.

Following Reynolds (1988), we plotted the mean value per

species of log S ⁄V versus log MLD for all treatments and

replicates and selected four MLD classes (<3, 3–10, 10–30

and 30–100 lm). The biovolumes for each size class were

summed per sample.

All zooplankton samples were counted using 2- to 5-mL

Sedgewick-Rafter chambers following Paggi & Jose de

Paggi (1974) criteria. Counting stopped when 100 speci-

mens of the most abundant species of each taxonomic

group (cladocerans, copepods and rotifers) were reached.

If necessary, the entire sample was counted. Zooplankton

biovolume (lm3) was estimated by measuring at least 20

individuals of each rotifer species or nauplii and using

volume formulas described in Ruttner-Kolisko (1977).

Biomass (lgDW L)1) was estimated assuming a density of

1.0 (Ruttner-Kolisko, 1977). In the case of copepods and

cladocerans, we measured 50 specimens of each species

Fig. 1 Phytoplankton biovolume in Lake Rodo at the beginning of

each experiment (A and B). Biovolume is divided in size classes

according to the maximum linear dimension (MLD) of each species.

Error bars indicate the standard deviation.



and estimated their biomass using length ⁄weight regres-

sions available in the literature (Bottrell et al., 1976;

McCauley, 1984; Culver et al., 1985).

Data analysis

Normality of all variables was tested with a Kolmogorov–

Smirnov test and homogeneity of variances with a Levene

test. Variables tested included all chemical and physico-

chemical variables, as well as zooplankton biomass (total

and taxonomic groups), and phytoplankton biovolume

(total and size classes). If variables were not normally

distributed, we transformed them using log10(x) and

log10(x + 1). We used two-way ANOVAANOVA with fish densities

and zooplankton biomass as fixed factors to test for

differences among treatments in all the measured vari-

ables. If variables remained not normal after transforma-

tion, we used nonparametric chi-square analysis. Linear

regressions were used to study the effect of fish on

cladocera biomass in all experiments. Results were con-

sidered significant at P < 0.05, unless noted otherwise.

Statistical analysis was performed with STATISTICA VSTATISTICA V7,

Statsoft, Tulsa, OK, U.S.A.

Results

Experiment A

In the absence of fish, manipulation of the zooplankton

community had no effect on phytoplankton biomass

(F3,6 = 0.137; P = 0.934) or community composition

(Fig. 2, upper panel; Fig. S1). The distribution of the

different size classes was similar among treatments

(P = 0.404), and the 30–100-lm size class dominated in

all of them (Fig. 2, upper panel). This size class

comprised cyanobacteria, mainly the filamentous A. grac-

ile with an average length of 91 lm. Zooplankton was,

as expected, absent in the zooplankton-free enclosures

(Fig. 3, upper panel), and in all treatments with fish,

zooplankton biomass was lower than at the start of the

experiment or in the fish-free enclosures (Fig. 3, upper

panel; Fig. S1). Also, larger zooplankton (cladocerans

and copepods) was more affected by fish than the

smaller rotifers (Fig. 3, upper panel). Although fish

appeared to have a negative effect on zooplankton

biomass, the two-way ANOVAANOVA revealed neither a statis-

tically significant fish effect (F2,14 = 2.21; P = 0.147) nor a

zooplankton effect (F3,14 = 2.67; P = 0.088) or an interac-

tion between the two factors (F6,14 = 0.98; P = 0.475) on

zooplankton biomass (Fig. 3, upper panel; Fig. S1).

However, it should be noted that loss of two Z1–P1

replicates influenced the statistical analysis. The overall

trend of the impact of fish on zooplankton is more

pronounced when all Z1 and Z2 treatments are used in

a linear regression (F1,17 = 4.95; r2 = 0.475; P = 0.040)

against fish densities. The regression model: Zooplank-

ton biomass = 33.5–2.88*Fish density clearly revealed the

negative relation between C. decemmaculatus and

zooplankton. More specifically, the presence of fish

had a strong negative effect on cladoceran biomass

(v2 = 10.53; P = 0.005), particularly on Moina micrura

(v2 = 9.42; P = 0.009), which was the dominant cladocera

Fig. 2 Phytoplankton biovolumes (stacked bars) in each treatment at

the end of experiment A (upper panel) and experiment B (lower

panel). Phytoplankton biovolume is divided in size classes according

to the maximum linear dimension (MLD) for each species. The

phytoplankton biovolume (±1 SD) at the start of the experiment in

each treatment is given as reference (open symbols). Z0 = no zoo-

plankton, Z1 = similar to the lake, Z2 = 10 times the lake, P0 = no

fish, P1 = 4 fish, P2 = 10 fish. D, Daphnia obtusa. In the lower panel,

ND, no data, as there was no treatment with Daphnia obtusa in

experiment B. Note the differences in the y-axis scale between upper

and lower panels.



species. As a result, at the end of the experiment, no

cladocera were found at all in Z1 and Z2 treatments that

also contained fish (Fig. 3, upper panel). The effect of

fish predation was also observed in the D mesocosms

containing only D. obtusa, although in this case, statisti-

cal differences were marginally significant (v2 = 5.58,

P = 0.061). Enclosures with fish had higher phytoplank-

ton biovolume (v2 = 6.68, P = 0.035), due to an increase

in their largest size fraction (30–100 lm), composed of

the filamentous cyanobacterium A. gracile (v2 = 6.80;

P = 0.033; Table 2, Fig. 2, upper panel). This effect was

more pronounced in the D and Z2 mesocosms (Fig. 2,

upper panel). Dissolved nutrients among treatments

were all in the same order of magnitude (NO3-N

62.9 ± 39.6 lg L)1, NH4-N 27.8 ± 15.0 lg L)1, PO4-P

32.2 ± 14.9 lg L)1). Nitrogen concentrations were lower

than in the lake, while PO4-P was higher (Table 2) The

mass N : P ratio in all enclosures was low and on

average 2–3. Nutrients were not significantly different

among treatments, except for the mesocosms containing

only D. obtusa (D), which tended towards higher NH4-N

concentrations than the other treatments (v2 = 9.60;

P = 0.022), particularly in the presence of fish.

Experiment B

As in experiment A, manipulation of the zooplankton

community had no effect on the phytoplankton biomass

(F2,5 = 1.08; P = 0.407), even in the absence of fish (Fig. 2,

lower panel; Fig. S2). At the end of the experiment, the

Fig. 3 Zooplankton biomass in each treatment at the end of experiment A (upper panel) and experiment B (lower panel). Zooplankton biomass

in the lake, and in each treatment, at the beginning of the experiment is on the right side of both panels. Z0 = no zooplankton, Z1 = similar to the

lake, Z2 = 10 times the lake, P0 = no fish, P1 = 4 fish, P2 = 10 fish. D, Daphnia obtusa. In the lower panel, ND, no data, as there was no treatment

with Daphnia obtusa in experiment B. Note the differences in the y-axis scale between upper and lower panels.



distribution of the different phytoplankton size classes

was similar among treatments (two-way ANOVAANOVAs;

P = 0.163), although the 30- to 100-lm size class became

dominant in all of them (Fig. 2, lower panel). Interest-

ingly, Monoraphidium griffithi, which dominated at the

beginning of the experiment, remained the dominant

species, as its cell size increased in all treatments from an

average of 29 lm at the start to 48 lm at the end of the

experiment. Unlike experiment A, rotifer populations

grew considerably in the zooplankton-free enclosures

(Z0) compared to the starting conditions (Fig. 3, lower

panel; Fig. S2). Rotifers were not retained in the 50-lm

sieve used to remove zooplankton at the start of the

experiment and probably benefited from the absence of

predators (e.g. cyclopoid copepods). In the presence of

fish, however, population growth in this treatment was

suppressed substantially (Fig. 3, lower panel). Overall,

there was no fish effect (F2,17 = 1.66; P = 0.220) on

total zooplankton biomass (Fig. S2). However, copepod

biomass in the high zooplankton treatments (Z2) was

marginally lower in mesocosms containing fish (v2 = 2.54;

P = 0.051), while cladocerans disappeared completely

(Fig. 3, lower panel).

Dissolved nutrients were similar among treatments

(NO3-N 276.4 ± 68.7 lg L)1, NH4-N 47.8 ± 18.3 lg L)1

and PO4-P 11.4 ± 4.7 lg L)1) and showed higher concen-

trations compared to the lake. The mass N : P ratio in all

enclosures varied between 19 and 39.

Discussion

Our experiments show that small herbivorous zooplank-

ton, typical of subtropical, nutrient-rich lakes, had limited

ability to impose top–down control on the phytoplankton

community. This occurred even in the presence of edible

phytoplankton size classes and in the absence of fish

predation. Increased densities of the natural zooplankton

communities did not significantly affect the phytoplankton

community, although theoretically, grazing rates could

have cleared the entire volume of the enclosures during the

experiments. Such apparent uncoupling of the zooplank-

ton–phytoplankton interaction has been described for other

subtropical regions (Crisman & Beaver, 1990; Havens, 2002;

Hunt & Matveev, 2005; Malthus & Mitchell, 2006; Von

Ruckert & Giani, 2008).

Our results support the variance-inedibility hypothesis

(Holt & Loreau, 2002), which states that trophic cascades

only occur when trophic levels are dominated by species

edible to the next trophic level (Polis et al., 2000). Interest-

ingly, in experiment B, the phytoplankton community at

the beginning of the experiment was considered to be

edible to the zooplankton, but no grazing down of the

phytoplankton community occurred. By contrast, an

increase in the size of the dominant species M. griffithii

was observed, which might indicate an inducible defence

against the populations of rotifers that occurred during the

experiment. Such morphological changes in chlorophytes

are well-known anti-predator strategies (Lurling, 2003).

Further support for the absence of phytoplankton top–

down control was found in the fish-free treatments with

Daphnia obtusa addition (experiment A). This might be

explained by the dominance of filamentous cyanobacteria

(A. gracile) in the phytoplankton community during this

experiment. Filaments can negatively affect the clearance

rate even for large-bodied grazers such as Daphnia, with

longer filaments having a stronger effect than shorter ones

(Gliwicz & Lampert, 1990; DeMott, Gulati & Van Donk,

2001). However, the length of A. gracile in experiment A

was c. 91 lm, and earlier studies suggest that this size of

filaments can be consumed. For example, Fulton (1988)

showed that D. pulex and D. parvula consumed Anabaena

flos-aquae filaments with a length of 111 (±18) lm, while

longer filaments of Aphanizomenon flos-aquae

(210 ± 24 lm) and other Anabaena species (from 233 to

423 lm) were not consumed. Similarly, Planktothrix rubes-

cens measuring <100 lm were preferably ingested by

adult Daphnia pulicaria over longer filaments up to 984 lm

(Oberhaus et al., 2007). Although no feeding experiments

have been performed, the overall negative effect on the

relatively large-bodied Daphnia obtusa might also point to

toxicity effects of this cyanobacterial species (Pereira et al.,

2004). Finally, similar to observations in a tropical lake by

Rondel et al. (2008), our results clearly demonstrate that

manipulation of the fish stock, or even elimination of it,

may not be enough to control a cyanobacterial bloom.

Strong size-selective predation by fish on zooplankton

was evident in the two experiments we conducted. These

results are consistent with findings from other mesocosm

experiments in the region (Boveri & Quiros, 2007; Iglesias

et al., 2008; Mazzeo et al., 2010), as well as from field data

(Scasso et al., 2001; Havens, 2002; Mazzeo et al., 2003;

Lacerot, 2010). Moreover, similar experiments in tropical

regions Okun et al. (2007) inferred that the mere presence

(rather than particular densities) of omnivorous fish

appears to guarantee a major top–down control in warm-

lake food webs. In line with this idea, our results show that

even low fish densities are sufficient to nearly eliminate

the largest zooplankton size fraction in all zooplankton

densities and community compositions tested.

Fish had a positive effect on phytoplankton during

experiment A, where Aphanizomenon gracile was the domi-

nant species. This effect was more pronounced when



combined with high densities of natural zooplankton or

Daphnia obtusa treatments. In this situation, NH4-N concen-

trations were higher too, while elevated phosphate concen-

trations were found at the higher fish densities. Nitrogen

may have been important as a limiting nutrient as N : P

ratios were <3. Nutrient excretion by fish and zooplankton

may therefore have favoured the observed phytoplankton

development (Attayde & Hansson, 1999).

The response of phytoplankton to different nutrient

levels is similar across mesocosms of different sizes,

although it may vary with the duration of the experi-

ments (Spivak, Vanni & Mette, 2011). However, it is less

clear how enclosure size or length of the experiment may

affect trophic interactions, and we did not test these

effects in our study. In previous years in Lake Rodo,

Scasso et al. (2001) observed a brief increase in mesozoo-

plankton abundance and a coincidental decrease in the

phytoplankton community, resulting in higher lake water

transparency. The mesozooplankton increment occurred

at the beginning of spring, following fish removal

procedures in the lake, and was correlated with an

increase in small cladocerans (M. micrura and Daphnia

pulex; Scasso et al., 2001). However, the cause of the

subsequent clear-water phase was not unequivocal, as

silica depletion may also have played a role in driving

the collapse of the dominant diatom species (C. Kruk,

unpubl. data). Nonetheless, other studies have shown

that relatively large cladocerans can occasionally develop

and graze down phytoplankton in subtropical and

tropical lakes in response to drastic fish removal (Boveri

& Quiros, 2007).

In conclusion, while longer-term absence of fish may in

principle allow zooplankton communities to develop and

control phytoplankton in warm lakes, our results illustrate

that a mere increase in densities of the existing zooplank-

ton community may not be enough to cause such a top–

down effect. In practise, reduction in fish stock as a tool to

control phytoplankton may therefore be of little use as fish

populations in these systems typically recover very

quickly, and a situation with very low fish densities will

be difficult to maintain long enough to allow the

zooplankton community to truly restructure and control

phytoplankton (Jeppesen et al., 2007, 2010; Meerhoff et al.,

2007; Van Leeuwen et al., 2007; Iglesias et al., 2011).

Acknowledgments

We wish to thank Federico Quintans for invaluable field

assistance and Instituto de Investigaciones Pesqueras

(Facultad de Veterinaria) for the D. obtusa cultures. This

study was financed by Consejo Sectorial de Investigacion

Cientıfica (CSIC) and Intendencia Municipal de Montevi-

deo (IMM), Uruguay. CK and GL were supported by SNI

(ANII). We thank two anonymous reviewers for their

helpful comments to improve this manuscript.

References

Attayde J.L. & Hansson L.-A. (1999) Effects of nutrient

recycling by zooplankton and fish on phytoplankton

communities. Oecologia, 121, 47–54.

Blanck A. & Lamouroux N. (2007) Large-scale interspecific

variation in life-history traits of European freshwater fish.

Journal of Biogeography, 34, 862–875.

Bottrell H.H., Duncan A., Gliwicz Z.M., Grygierek E., Herzig

A., Hillbricht-Ilkowska A. et al. (1976) A review of some

problems in zooplankton production studies. Norwegian

Journal of Zoology, 24, 419–456.

Boveri M.B. & Quiros R. (2007) Cascading trophic effects in

pampean shallow lakes: results of a mesocosm experiment

using two coexisting fish species with different feeding

strategies. Hydrobiologia, 584, 215–222.

Carpenter S.R., Kitchell J.F. & Hodgson J.R. (1985) Cascading

trophic interactions and lake productivity. BioScience, 35,

634–639.

Crisman T.L. & Beaver J.R. (1990) Applicability of planktonic

biomanipulation for managing eutrophication in the sub-

tropics. Hydrobiologia, 200 ⁄201, 177–185.

Crisman T.L., Philips E.J. & Beaver J.R. (1995) Zooplankton

seasonality and state relationships in Lake Okeechobee,

Florida. Advances in Limnology, 45, 213–232.

Culver D.A., Boucherle M.M., Bean D.J. & Fletcher J.W. (1985)

Biomass of freshwater crustacean zooplankton from

length-weight regressions. Canadian Journal of Fisheries and

Aquatic Sciences, 42, 1380–1390.

Cyr H. & Curtis J.M. (1999) Zooplankton community size

structure and taxonomic composition affects size-selec-

tive grazing in natural communities. Oecologia, 118, 306–

315.

DeMott W.R., Gulati R. & Van Donk E. (2001) Daphnia food

limitation in three hypereutrophic Dutch lakes: evidence

for exclusion of large-bodied species by interfering fila-

ments of cyanobacteria. Limnology and Oceanography, 46,

2054–2060.

Fernandes R., Gomes L.C., Pelicice F.M. & Agostinho A.A.

(2009) Temporal organization of fish assemblages in

floodplain lagoons: the role of hydrological connectivity.

Environmental Biology of Fishes, 85, 99–108.

Fulton R.S. (1988) Grazing on filamentous algae by herbiv-

orous zooplankton. Freshwater Biology, 20, 263–271.

Gillooly J.S. & Dodson S.I. (2000) Latitudinal patterns in the size

distributionandseasonaldynamicsofnewworld, freshwater

cladocerans. Limnology and Oceanography, 45, 22–30.

Gliwicz Z.M. & Lampert W. (1990) Why do cladocerans fail to

control algal blooms? Hydrobiologia, 200 ⁄201, 83–98.



Havens K.E. (2002) Zooplankton structure and potential food

web interactions in the plankton of a subtropical chain-of-

lakes. The Scientific World Journal, 2, 926–942.

Havens K.E. & Beaver J. (2011) Composition, size, and

biomass of zooplankton in large productive Florida lakes.

Hydrobiologia, 668, 49–60.

Havens K.E., East T.L., Marcus J., Essex P., Bolan B.,

Raymond S. et al. (2000) Dynamics of the exotic Daphnia

lumholtzii and native macro-zooplankton in a subtropical

chain-of-lakes in Florida, USA. Freshwater Biology, 45,

21–32.

Havens K.E., Elia A.C., Taticchi M.I. & Fulton R.S. (2009)

Zooplankton-phytoplankton relationships in shallow

subtropical versus temperate lakes Apopka (Florida,

USA) and Trasimeno (Umbria, Italy). Hydrobiologia, 628,

165–175.

Hillebrand H., Durselen C.-D., Kirschtel D., Pollingher U. &

Zohary T. (1999) Biovolume calculation for pelagic and

benthic microalgae. Journal of Phycology, 35, 403–424.

Holt R.D. & Loreau M. (2002) Biodiversity and ecosystem

functioning: the role of trophic interactions and the

importance of system openness. In: The Functional Conse-

quences of Biodiversity: Empirical Progress and Theoretical

Extensions (Ed. D. Tilman), pp. 246–262. Princeton Univer-

sity Press, Princeton.

Hunt R.J. & Matveev V.F. (2005) The effects of nutrients and

zooplankton community structure on phytoplankton

growth in a subtropical Australian Reservoir: an enclosure

study. Limnologica, 35, 90–101.

Iglesias C., Mazzeo N., Goyenola G., Fosalba C., Teixeira-de

Mello F., Garcıa S. et al. (2008) Field and experimental

evidence of the effect of Jenynsia, a small omnivorous–

planktivorous fish, on the size distribution of zooplankton

in subtropical lakes. Freshwater Biology, 53, 1797–1807.

Iglesias C., Mazzeo N., Meerhoff M., Lacerot G., Clemente

J.M., Scasso F. et al. (2011) High predation is of key

importance for dominance of small-bodied zooplankton

in warm shallow lakes: evidence from lakes, fish exclosures

and surface sediments. Hydrobiologia, 667, 133–147.

Jeppesen E., Meerhoff M., Holmgren K., Gonzalez-Bergonz-

oni I., Teixeira-de Mello F., Declerck S.A.J. et al. (2010)

Impacts of climate warming on lake fish community

structure and potential effects on ecosystem function.


Jeppesen E., Meerhoff M., Jacobsen B.A., Hansen R.S.,

Søndergaard M., Jensen J.P. et al. (2007) Restoration of

shallow lakes by nutrient control and biomanipulation –

the successful strategy varies with lake size and climate.


Koroleff F. (1970) Direct determination of ammonia in natural

water as indophenol-blue. In: ICES, Information on Techni-

ques and Methods for Seawater Analysis. An Interlaboratory

Report, 3, 19–22.

Kruk C., Mazzeo N., Lacerot G. & Reynolds C.S. (2002)

Classification schemes for phytoplankton: a local valida-

tion of a functional approach to the analysis of species

temporal replacement. Journal of Plankton Research, 25,

901–912.

Lacerot G. (2010) Effects of Climate on Size Structure and

Functioning of Aquatic Food Webs. PhD thesis, Wageningen

University, Wageningen.

Lappalainen J. & Tarkan A.S. (2007) Latitudinal gradients in

onset date, onset temperature and duration of spawning of

roach. Journal of Fish Biology, 70, 441–450.

Legresley M. & McDermott G. (2010) Counting chamber

methods-Haemocytometer, Palmer-Maloney cell and

Sedgewick-Rafter cell. In: Microscopic and Molecular Methods

for Quantitative Phytoplankton Analysis (Eds B. Karlson,

K. Cusack & E. Bresnan), pp. 25–30. IOC Manuals and

Guides. Intergovernmental Oceanographic Commission of

UNESCO, Paris.

Lowe-McConnell R.H. (1999) Estudos Ecologicos de Comunid-

ades de Peixes Tropicais. EDUSP, Sao Paulo.

Lund J.W.G., Kipling C. & Le Cren E.D. (1958) The inverted

microscope method of estimating algal numbers and the

statistical basis of estimations by. Hydrobiologia, 11, 143–

170.

Lurling M. (2003) Phenotypic plasticity in the green algae

Desmodesmus and Scenedesmus with special reference to the

induction of defensive morphology. International Journal of

Limnology, 39, 85–101.

Malthus T.J. & Mitchell S.F. (2006) On the occurrence, causes

and potential consequences of low zooplankton to phyto-

plankton ratios in New Zealand lakes. Freshwater Biology,

22, 383–394.

Mazzeo N., Iglesias C., Teixeira-de Mello F., Borthagaray A.,

Fosalba C., Ballabio R. et al. (2010) Trophic cascade effects

of Hoplias malabaricus (Characiformes, Erythrinidae) in

subtropical lakes food webs: a mesocosm approach. Hyd-

robiologia, 644, 325–335.

Mazzeo N., Rodrıguez-Gallego L., Kruk C., Meerhoff M.,

Gorga J., Lacerot G. et al. (2003) Effects of Egeria densa

Planch. beds on a shallow lakes without piscivorous fish.


McAlice B.J. (1971) Phytoplankton sampling with the

Sedgwick-Rafter Cell. Limnology and Oceanography, 16,

19–28.

McCauley E. (1984) The estimation of the abundance and

biomass of zooplankton in samples. In: A Manual on

Methods for the Assessment of Secondary Productivity in Fresh

Waters (Eds J.A. Downing & F.H. Rigler), pp. 228–265.

Blackwell Scientific Publications, Oxford.

Meerhoff M., Clemente J.M., Teixeira-de Mello F., Iglesias C.,

Pedersen A.R. & Jeppesen E. (2007) Can warm climate-

related structure of littoral predator assemblies weaken

clear water state in shallow lakes? Global Change Biology, 13,

1888–1897.

Murphy J. & Riley J.P. (1962) A modified single solution

method for the determination of phosphate in natural

waters. Analytica Chimica Acta, 27, 31–36.



Nusch E.A. (1980) Comparison of different methods for

chlorophyll and phaeopigments determination. Archiv fur

Hydrobiologie, Beiheft Ergebnisse der Limnologie, 17, 14–36.

Oberhaus L., Malorie G., Pinel-Alloul B., Ghadouani A. &

Humbert J.F. (2007) Grazing of two toxic Planktothrix

species by Daphnia pulicaria: potential for bloom control

and transfer of microcystins. Journal of Plankton Research,

29, 827–838.

Okun N., Brasil J., Attayde J.L. & Costa I.A.S. (2007)

Omnivory does not prevent trophic cascades in pelagic

food webs. Freshwater Biology, 53, 129–138.

Paggi J.C. & Jose de Paggi S. (1974) Primeros estudios sobre el

zooplancton de las aguas loticas del Parana medio. Physis,

33, 91–114.

Pereira P., Li R., Carmichael W., Dias E. & Franca S. (2004)

Taxonomy and production of paralytic shellfish toxins by

the freshwater cyanobacterium Aphanizomenon gracile

LMECYA40. European Journal of Phycology, 39, 361–368.

Perrow M.R., Meijer M.-L., Dawidowicz P. & Coops H. (1997)

Biomanipulation in shallow lakes: state of the art. Hydro-

biologia, 342–343, 355–365.

Peters R.H. & Downing J.A. (1984) Empirical analysis of

zooplankton filtering and feeding rates. Limnology and

Oceanography, 29, 763–784.

Polis G.A., Sears A.L.W., Huxel G.R., Strong D.R. & Maron J.

(2000) When is a trophic cascade a trophic cascade? Trends

in Ecology and Evolution, 15, 473–475.

Quintans F., Scasso F. & Defeo O. (2010) Unsuitability of

Cnesterodon decemmaculatus (Jenyns, 1842) for mosquito

control in Uruguay: evidence from food-preference exper-

iments. Journal of Vector Ecology, 35, 1–6.

Quintans F., Scasso F., Loureiro M. & Yaffe A. (2009) Diet of

Cnesterodon decemmaculatus (Poeciliidae) and Jenynsia mult-

identata (Anablepidae) in a hypertrophic shallow lake of

Uruguay. Iheringia, 99, 105–111.

Reynolds C.S. (1984) The Ecology of Freshwater Phytoplankton.

Cambridge University Press, Cambridge.

Reynolds C.S. (1988) Functional morphology and the adap-

tive strategies of freshwater phytoplankton. In: Growth and

Reproductive Strategies of Freshwater Phytoplankton (Ed. C.D.

Sandgren), pp. 388–433. Cambridge University Press,

Cambridge.

Rodrıguez-Gallego L.R., Mazzeo N., Gorga J., Meerhoff M.,

Clemente J., Kruk C. et al. (2004) The effects of an artificial

wetland dominated by free-floating plants on the restora-

tion of a subtropical, hypertrophic lake. Lakes & Reservoirs:

Research & Management, 9, 203–215.

Rondel C., Arfi R., Corbin D., Le Bihan F., Ndour E.H. &

Lazzaro X. (2008) A cyanobacterial bloom prevents fish

trophic cascades. Freshwater Biology, 53, 637–651.

Rosa R.S. & Costa W.J.E.M. (1993) Systematic revision of the

genus Cnesterodon (Cyprinodontiformes: Poecilidae) with

the description of two new species from Brazil. Copeia, 3,

696–708.

Ruttner-Kolisko A. (1977) Suggestions for biomass calcula-

tion of plankton rotifers. Archiv fur Hydrobiologie, Beiheft

Ergebnisse der Limnologie, 8, 71–76.

Scasso F., Mazzeo N., Gorga J., Kruk C., Lacerot G., Clemente

J. et al. (2001) Limnological changes in a subtropical

shallow hypertrophic lake during its restoration: two years

of a whole-lake experiment. Aquatic Conservation: Marine

and Freshwater Ecosystems, 11, 31–44.

Spivak A.C., Vanni M.J. & Mette E.M. (2011) Moving on up:

can results from simple aquatic mesocosms experiments be

applied across broad spatial scales? Freshwater Biology, 56,

279–291.

Sunaga T. & Verani J.R. (1997) The fish communities of four

lakes. In: Limnological Studies on the Rio Doce Valley Lakes,

Brazil (Eds J.G. Tundisi & Y. Saijo), pp. 359–369. Brazilian

Academy of Sciences and University of Sao Paulo, Sao

Paulo.

Teixeira-de Mello F., Meerhoff M., Pekcan-Hekim Z. &

Jeppesen E. (2009) Substantial differences in littoral fish

community structure and dynamics in subtropical and

temperate shallow lakes. Freshwater Biology, 54, 1202–1215.

Utermohl H. (1958) Zur Vervollkomnung der quantitativen

Phytoplankton-Methodik.Mitteilungender InternationaleVere-

iningung fur Theoretische und Angewandte Limnologie, 9, 1–38.

Van Leeuwen E., Lacerot G., Van Nes E.-H., Hemerik L. &

Scheffer M. (2007) Reduced top–down control of

phytoplankton in warmer climates can be explained by

continuous fish reproduction. Ecological Modelling, 206,

205–212.

Vanderploeg H.A. (1981) Seasonal particle-size selection by

Diaptomus sicilis in offshore Lake Michigan. Canadian

Journal of Fisheries and Aquatic Sciences, 38, 504–517.

Von Ruckert G. & Giani A. (2008) Biological interactions in

the plankton community of a tropical eutrophic reservoir:

is the phytoplankton controlled by zooplankton? Journal of

Plankton Research, 30, 1157–1168.

Supporting Information

Additional Supporting Information may be found in the

online version of this article:

Figure S1. Total phytoplankton biovolume (upper panel)

and total zooplankton biomass (lower panel) in the

different treatments, at the end of experiment A.

Figure S2. Total phytoplankton biovolume (upper panel)

and total zooplankton biomass (lower panel) in the

different treatments, at the end of experiment B.

(Manuscript accepted 7 November 2012)



Documents

The role of subtropical zooplankton as grazers of phytoplankton under different predation levels