The Roach Population in the Hypertrophic Bautzen Reservoir: Structure, Diet and Impact on Daphnia Galeata

Limnologica 31 (2001) 61-68 LIMNOLOGICA http://www.urbanfischer.de/j ournals/limno © by Urban & Fischer Verlag

Dresden University of Technology, Institute of Hydrobiology, Dresden, Germany

The Roach Population in the Hypertrophic Bautzen Reservoir: Structure, Diet and Impact on Daphnia galeata

UWE KAHL, HENDRIK DORNER, ROBERTJ. RADKE~ANNEKATRIN WAGNER • JURGENBENNDORF

With 7 Figures and 4 Tables

Key words: Roach, population structure, diet, biomanipulation, Daphnia galeata

Abstract

The structure and diet of the roach (Rutilus rutilus) population in the hypertrophic Bautzen Reservoir (BENNDORF 1990, 1995). hypertrophic Bautzen Reservoir was examined from April to Novem- The enhancement of the piscivorous fish stock by stocking ber 1998. Under the long-term impact of high predation pressure by with pikeperch (Sander lucioperca), pike (Esox lucius), eel piscivorous fish, a very heterogeneous population structure of roach (Anguilla anguilla) and wels (Silurus glanis) in combination had developed. Only a few age classes were dominant while other age with catch restrictions, led to strong top-down control of the classes were nearly absent. The proportion of males decreased with food web. As a consequence of this food web manipulation, increasing age to 4% of the total abundance of one age class, which nevertheless seemed to have no negative effect on reproductive suc- the proportion of piscivorous fish in the total fish biomass in- cess. Food analysis revealed that the diet consisted of a high propor- creased to 68% in 1997 (BENNDORF et al. 1998), while the tion of algae and macrophytes. The collapse of the Daphnia galeata total fish biomass decreased to less than 50% of the initial population in early summer 1998 forced the roach to switch to benthic biomass. Consequently, the fish stock in Bautzen Reservoir food resources [macroinvertebrates and fish: chironomids, molluscs in biomass terms is not dominated by cyprinids, but by per- and ruffe (Gymnocephalus cernuus)] in early June. Total consumption cids, in contrast to the general expectation for eutrophic or of age-2 and age-4 roach, the two most dominant year classes, was hypertrophic lakes without food web manipulation (HART- calculated by a bioenergetics model. Additionally, consumption of MANN 1977a, b; KUBBCKA 1993). age-0 roach was estimated by assuming a fixed daily food consump- The aim of this study was: i) to describe the age and size tion rate. The results suggest that daily consumption by these age structure of a roach population which had developed under a groups, which never exceeded 0.2% of total biomass of the D. galeata population, had a negligible impact on the population of daphnids in long-term high predation pressure of mainly pike and Bautzen Reservoir during the period studied. pikeperch (ScHULTZ et al. 1992), ii) to detect whether selec- tive predation pressure on male roach could be the cause for the drastic decline of the proportion of males in the roach Introduction population of Bautzen Reservoir as described by SCHULTZ (l 996) and iii) to evaluate the predation pressure on daphnids Cyprinids affect water quality by consuming zooplankton, by by estimating the consumption of the remaining roach popu- their excretory products and by releasing nutrients from the lation. sediment as an effect of their feeding activities (HoRPt'mA 1994). Roach (Rutilus rutilus) play a key role, particularly in eutrophic lakes, because of their ability to utilize almost any Methods kind of food sources, especially in situations of strong compe- Age-1 and older roach were caught with gill-nets (mesh sizes 12, 15, tition (PERsSON 1983a; BRABRAND 1985). As a consequence of 18, 22, 25, 32, 40, 50, 60 and 70 mm) set overnight in the littoral this ability roach frequently dominate in eutrophic and hyper- (area of sediment resuspension, 1-6 m) and pelagic zone (7-11 m) on trophic waters (HARTMANN 1977a, b; KUBECKA 1993). April 20, August 19 and November 15 in 1998. Each fish was With the aim of water quality improvement, a whole-lake weighed to the nearest gram and measured [total length (TL) and biomanipulation experiment was initiated in 1981 in the maximum body depth] to the nearest millimeter. Sex was determined

0075-9511/01/31/01-061 $ 15.00/0 6l and the opercular bone was removed for age determination. A ran- the nearest 0.1 mm. Individual biomasses of the food items measured dom sub-sample of 100 roach per gillnet were taken on April 20 and were backcalculated by using the equations given by MEHNER et al. of 50 on August 19 were taken for age determination. Nominal catch- (1995) and H. VoIGT (Inst. of Hydrobiology, unpubl, results). The pro- es were standardized to catch per unit effort (CPUE) according to portion of the non-measurable food items was estimated visually. The equation 1: consumption estimate was calculated with the help of a bioenergetics model (KITcH~LL et al. 1977; HANSON et al. 1997). The underlying pa- As C~ A~ rameters were taken from HORPPILA & PELTONEN (1997). The con- CPUE : -- (1) sumption estimate was made only for the period between May 11 and t June 8, before and during the collapse of the daphnid population (S. HOLSMANN, Inst. of Hydrobiology, unpubl, data). Consumption rates with: CPUE - catch per unit effort; C N - nominal catch; A s - area of standard net (100 mS); A N - area of net used (mZ); t - time of exposi- tion (h). The CPUE of the pelagic zone was weighted by factor 7.33 ac- Table 2. Abundance, mean individual wet body mass and proportion cording to the proportion of pelagic zone volume versus littoral zone of daphnids in the diet of age-0 roach in the littoral and pelagic area volume. Additionally, roach were caught weekly from May 11 to in May and June 1998 in Bautzen Reservoir. L - littoral; P - pelagial; June 22 and on July 6, July 20 and August 10 with a bottom trawl standard error in parenthesis; *) no SE because only one trawl was (10 mm mesh size in the cod end, three horizontal hauls per date and performed. zone) at night in the littoral and pelagic zone and by beach seining in the evening. A flowmeter was used to calculate the water volume Date Sampling Abundance Mean wet Proportion fished by the trawl. To estimate the abundance of roach in Bautzen area (Ind m -3) body mass of daphnids Reservoir the number of fish caught per unit volume was multiplied (mg Ind-1) in the diet (%) by the total volume of the respective area (littoral or pelagic zone). The nominal catch and the abundance estimate for every sampling 11 May L 0 date are given in Table 1. The abundance of each age group was cal- P 0 culated by applying its proportion in the total population, determined 18 May L 5.33 (0.12) 3.18 (0.36) 7.2 from the gill-net catches on April 20 and August 19, to the respective P 1.11 (0.09) 1.40 (0.07) 0 abundance estimate. In the case of age group 1 only the August data 25 May L 2.18 (0.07) 9.25 (1.49) 19.1 were used as this age group was not caught representatively in April P 0.40 (0.08) 5.38 (0.59) 95.1 with the mesh sizes used. Age-0 roach were sampled once a week be- 02 June L 3.31 (0.37) 36.79 (2.04) 15.0 tween May 11 and June 2 at night by using two bongo-nets (MEHNER P 0.45 (0.08) 10.99 (1.61) 63.6 et al. 1997, 1998a). On June 8, age-0 roach were sampled by using a 08 June L 1.01 *~ 108.78 (2.60) 0 small otter trawl as described by MEHN~a et al. (1998a). Fish abun- P 0.09 (0.04) 113.42 (3.54) 0 dance was calculated in the same way as for trawls (Table 2). Diet of roach caught by trawl and beach seine was pooled for food analysis. The content of the anterior part of the gut (pharynx - first Table 3. Mean water temperatures over the whole water column in bend) was filtered through 100 gm gauze and analysed under a stereo May and June 1998 and respective days of simulation used for the microscope. The food items were counted and if possible measured to calculation of roach consumption.

Date Day of simulation Water temperature (°C) Table 1. Nominal catches of roach in the littoral and pelagic zone by trawl fishery in Bautzen Reservoir with abundance estimates 11 May 1 15.7 (number of individuals) in 1998. 18 May 8 16.4 25 May 15 15.0 Date Number of roach Number of roach Estimated 02 June 23 17.4 caught in the caught in the total 08 June 29 19.4 littoral pelagial abundance llMay 28 61 298135 Table 4. Mean individual initial and end wet weight (ww) of roach 18May 118 193 787881 used for the calculation of roach consumption in the period from 25May 76 46 281080 May 11 to June 15. Number of roach per sampling date and age 2June 51 19 258766 group (n). Standard error in parenthesis. 8June 112 4 168482 15June 90 13 172701 Age Initialmass (g ww) End mass (g ww) 22June 94 71 300182 6July 94 28 195532 May 11 May 18 n June 8 June 15 n 20 July 51 39 166684 10August 89 25 247893 2 20.7 (1.4) 6 26.3 (2.2) 9 2 19.3 [back- 24.9 [back Mean 80 50 287733 calculated] calculated] Standard error 9 17 58070 4 104.5 (5.5) 35 141.9 (7.4) 11

62 Limnologica 31 (2001) 1 per individual and day were calculated for age-2 and age-4 roach - Results which formed 70% of roach population biomass - and then multiplied by the estimated abundance of the age group and the proportion of Population structure daphnids in the diet of the respective age group. Mean water temperatures (total water column) used in the model calculations are shown in Length-frequency distribution of age-1 and older roach was Table 3. Initial and final wet body mass of roach are given in Table 4. very heterogeneous (Fig. 1). In April, roach with a total length Because of low numbers caught on May 11 and June 8, initial and final between 10 and 12 cm were dominant, whereas in August wet body mass for age-2 roach were backcalculated assuming a linear roach of a length between 12 and 16 cm dominated, and in increase in body mass of 0.2 g d-1. This daily growth rate was derived November roach between 24 and 28 cm total length were dom- from the mean mass increase of this age class between May 18 and June 15. Since the physiological parameters given by HORPPlLA & inant. The length-frequency distribution is reflected in the age PELTONEN (1997) are not applicable to age-0 roach, consumption of distribution (Fig. 2), showing that the roach population (with- this group was estimated by making a few simple assumptions: First, out young-of-the-year) of Bautzen Reservoir was numerically total biomass of the age-0 roach population was calculated by using dominated by only three age groups (age-l, 2 and 4). In April the abundance estimates and the mean individual biomass values of the proportion of age-2 roach was 83% but decreased to 17% of the respective dates (Table 2). Second, assuming that roach of this size the total catch in November, while the age-4 roach dominated have a daily ration of 100% of their own body mass (MARMULLA with a proportion of 62% at that time. The distribution of ROSCH 1990) and knowing the proportion of daphnids in the gut (Table biomass showed dominance by four year old roach (42% to 2) the consumption of age-0 roach was calculated. With the total daily consumption rates of these three age classes the predation pressure on 69%) over the whole study period (Fig. 3). The proportion of D. galeata was then estimated for the period between May 11 and June two year old roach, which were dominant in abundance in April 8 with reference to the standing crop of the daphnids (S. H1]LSMANN, and August, was 23% and 30%, respectively. The total biomass Inst. of Hydrobiology, unpubl, data). proportion of the remaining age groups never exceeded 10%.

20 Apr 98 20 Apr 98 4C 100-

n = 844 n = 565 i;!

2£ 50 '/.i

, , I'-] r-~ R I7 I-1 = ,1, , ,_ t- ..c 0 19 Aug 98 ~ 19 Aug 98 ',3 100 0 40 n = 303 ,~ n = 797 "5

v 5o >, 20 0 O C C 19 19 23 0" °12r Nn 19 19 ~" 01--, .... ~ F-t , ,F],~, ,rm,m,__ <0"1 ¢- 15 Nov 98 1 O0 ._]19 15 Nov 98 40 n = 136 n = 137

50 20

• 1 2 3 4 5 6 7 8 9 10 11 12 13 14 <8 <10 <12 <14 <16 <18 <20 <22 <24 <26 <28 <30 <32 <34 <36 <38 Age TL (cm)

Fig. 1. Length-frequency distribution of roach caught by gill-netting Fig. 2. Age distribution of roach caught by gill-netting in Bautzen in Bantzen Reservoir in April, August and November 1998. Reservoir in April, August and November 1998.

Limnologica 31 (2001) 1 63 20 Apr 98 Sex ratio was determined for the age groups 1-4 and pooled for the older age groups (Fig. 4, data for each age 751 n = 565 group pooled over the sampling period). Age-2 roach showed 50 a sex ratio of 1. The proportion of males in the age group 1 was 36% and in the age group 3 it was 65%. However, both proportions were not significantly different from 50% (chi- o .... ~,~, , rl ,t-i ,_ , ~ , F1 t- squared-Test, c~ = 0.05) because of small sample size, while O age-4 roach with a proportion of males of 35% showed a sig- O 19 Aug 98 nificant difference from 50%. The proportion of five year old -8 n = 303 and older male roach was only 4% and also significantly dif-

"5 50 ferent from 50%. Fig. 5 shows the shift of the sex ratio of the £ age-4 roach during the period of study in 1998. The propor- oO dl 25 tion of males was 43% in April, 19% in August and 22% in November. The latter two values were significantly different E , ,rq,rq,FI ...... ,_o 0 ..ID from the April value (chi-squared-Test, c~ = 0.05). Thus, .~: O the decline of the proportion of males must have taken place O 15 Nov 98 between April and August. rr n = 136 Significant differences in body depth between males and females were only found in the four year old roach in April 50 75t H and August (t-Test, p < 0.05), with female body depth (56.7 25 , mm _+ 5.5 mm SD and 73.9 +_ 5.1 mm SD) deeper than that of the males (51.1 -+ 5.0 mm SD and 55.7 _+ 4.8 mm SD). 0 ~ ,I-1 , ~ R R ..... 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Age Diet and impact on Daphnia galeata Fig. 3. Biomass distribution of roach caught by gill-netting in The diets of a total of 441 age-1 and older roach were ana- Bautzen Reservoir in April, August and November 1998. lysed. Because of the small sample size of some age groups the data were pooled in three groups (group 1: age- 1 + age- n = 25 143 15 508 82 2, group 2:age-3 + age-4, group 3:age-5 and older). In 100 general, diet of roach in the Bautzen Reservoir contained plant material such as coccal and filamentous algae and parts of macrophytes in addition to a wide spectrum of ani- v 50 CC mal food. Apart from daphnids, other crustaceans such as CO Aloha, Bosmina, Chydorus, Leptodora and, in a few cases, copepods (Cyclops) were found. The benthic diet compo- nent mainly consisted of chironomids, but molluscs, mites 1 2 3 4 5-14 and the eggs and larvae of other insects were also found. Age Older age groups were partly piscivorous. Specifically, • male [] female roach of the age group 1 and 2 showed a wide spectrum of Fig. 4. Sex ratio (SR) of several age groups of roach in Bautzen food (Fig. 6) which was dominated by algae and macro- Reservoir in 1998. phytes (up to 87% of prey biomass). Daphnia galeata was rarely ingested by this age group (not exceeding 10%). The n = 305 120 82 diet of the three and four year old roach was dominated by 100 algae and macrophytes (47% to 95%). Daphnids were only ! found in the diet in May at proportions of 9% to 34%. From o~ June onwards, benthic organisms were ingested predomi- v 50 ¸ rr nantly (up to 53%) amongst algae and macrophytes. On O3 I June 15, age-0 ruffe (Gymnocephalus cernuus, TL = 20 mm) was found in the diet of this age group. The diet of five year old and older roach contained mainly algae and 20 Apr 98 19 Aug 98 15 Nov 98 macrophytes (up to 100%) and daphnids were found only Date in May. In June and July, ruffe (TL = 12-56 mm) was an • male [] female important part of the diet (up to 96%). Benthic organisms Fig. 5. Sex ratio (SR) of age group 4 roach in Bautzen Reservoir in and zooplankton were of little importance in the diet of this April, August and November 1998. age group,

64 Limnologica 31 (2001) 1 Age group 1 and 2 The estimated consumption of D. galeata by age groups 0,

n = 12 4 6 13 7 10 13 10 10 2 and 4 is shown in Fig. 7. The maximum biomass (wet weight) of D. galeata consumed by age-0 roach was 194 kg day-1 on June 2, by age-2 roach it was 26 kg day i on May 11. Age-4 roach had consumed the highest amount of daphnid biomass (up to 212 kg day-1 on May 25). The biomass consumed by these age groups on May 25 corresponds to 0.15% d I of the standing crop of D. galeata in Bautzen Reservoir 03 and reflects the maximum consumption detected during the Age group a and 4 E time of our study (Fig. 7). On all other days of the study the O :5 n=26 29 19 10 7 11 12 16 8 10 proportion of biomass of the total D. galeata population con- £ sumed by these roach age groups was less than 0.1%.

50 :.:Ji ::i:: i::il

£ Discussion "5

O Population structure "-E Age group 5 and older O o.. While roach populations usually consist of a mixture of age o n= 6 5 3 3 6 11 5 3 0- classes with many fish up to 10 years or more in age (TowNSEND & PERROW 1989), the roach population in Bautzen Reservoir exhibited a very heterogeneous structure. The age distribution shows that only few year classes (1994, 1996) were dominant while other year classes were nearly 0 absent. An underestimation of certain year classes due to gill net selectivity seems unlikely as the shift of the size classes as a result of growth during the time of the study is well docu- Date mented (Fig. 1). It can be generally stated that the growth of [] fish • Daphnia ~ other zooplankton roach in Bautzen Reservoir is very fast compared to other [] zoobenthos D algae and macrophytes lakes (e.g. WYATT 1988; SCHULTZ 1992; RADKE 1998). Ac- cordingly, age group 4, which forms the major part of the roach stock by biomass, had a mean total length of 261 mm Fig. 6, Relative proportion (by biomass) of food components in the diet of roach caught by trawl net fishing in Bautzen Reservoir in and a mean body depth of 65 mm at the time of our study. t998. °'Other zooplankton" includes other cladocerans and copepods, Roach of this size are probably only vulnerable to predation "zoobenthos" includes chironomids (larvae and pupae), molluscs, by pike and were seldom found in the diet of pikeperch, the mites, eggs and larvae of other insects. Data of littoral and pelagic most important piscivore in Bautzen Reservoir (ScHULTZ et catches were pooled. Numbers of guts analysed are indicated above al. 1992). Consequently, the predation pressure of piscivores bars. is restricted to younger age groups (0 to 3), and it can be pre- dicted that only age groups reaching the size refuge with a sufficient number of individuals will be dominant over the following years.

~250 0.16 4=-b 200 0.12 ~- E 15o ..Q 008 g 100 ~25 E 004 #~ 50 Fig. 7. Consumption of Daphnia galeata (kg wet weight) by roach age group 0, 2 and 4 (vertical bars) 0 0 o i w o NS- and daily uptake of daphnids by roach age-0, 2 and 4 11 May 18May 25May 2Jun 8Jun (expressed as a proportion of the total population Date biomass of the daphnids, circles) in Bautzen Reservoir from May 11 through June 8 in 1998. Values of con- [] age-O [] age-2 [] age-4 sumption estimates refer to the total volume of the • Percentage of consumed D. galeata biomass water body of the reservoir.

Limnologica 31 (2001) 1 65 Our results concerning sex ratios corroborate those of an theless, the roach showed a positive energy budget even with earlier study by SCHULTZ(1996) in Bautzen Reservoir. Figs. a diet based entirely on plant material. The lower energy 4 and 5 indicate a decline of the proportion of males. It yield from plant material is compensated partially by an in- should be taken into account though that it shows the instan- creased ingestion of this food resource. However, growth is taneous state of the sex ratio at a small time scale. A study reduced compared to a diet based on animal food (HoFER et following the fate of the proportion of males within a dis- al. 1985). tinct year class over a longer time period could give clearer In general, the proportion of zooplankton in the diet of evidence of a change of the sex ratio. Body size of prey is a roach decreases with increasing age and size (e.g. JAMETet al. decisive factor for prey selection of gape-limited piscivo- 1990; HORPPILA 1994). While age-0 roach caught in the rous fish such as pikeperch (HAMBRIGHTet al. 1991; VAN pelagic zone followed this general pattern, those caught in DENSEN 1988). From this it could be deduced that male the littoral zone and age-1 to age-2 roach had a smaller pro- roach should be selectively preyed upon by piscivorous portion of daphnids in their diet in May than age-4 roach. predators if they had a smaller body depth than the females This finding might partly be explained by the lower density of the same age group. For younger roach this hypothesis is of daphnids in the littoral zone, a consequence of higher pre- refuted because of a lack of significant differences in the dation pressure of age-0 perch on daphnids, compared to that body depth between males and females within the age in the pelagic zone of Bautzen Reservoir (H{JLSMANNet al. groups 1 to 3. It is unlikely that predation was the main rea- 1999). In this context it has to be noted that the majority son for the decline of the 4 year old males because a signifi- (63%) of age-1 to age-2 roach included in the diet analysis cant decline could only be detected in the period between had been caught in the littoral zone with the beach seine. The April and August and not thereafter (Fig. 5). Though we preference for structured habitats in the shallower part of the have no direct evidence for the increased mortality of four littoral zone is interpreted as a predator avoidance behaviour years old males during or after the spawning period in early of juvenile roach (CHRISTENSEN & PERSSON 1993; BEAN & May, the catch data make this hypothesis seem reasonable. WlNnELD 1995) and very likely accounts for the observed An effect of estrogenic substances can be excluded, as visu- distribution pattern of young roach. ally detectable hermaphrodites only made up 0.02% According to TOWNSENDet al. (1986), roach switch to ben- (SCHULTZ 1996) and 0% (this study) of the adult population. thic food resources at a zooplankton density of fewer than 40 Furthermore, the very low proportion of males older than individuals per litre, or use other food resources such as four years probably does not negatively affect reproduction macrophytes or detritus (HoRPPILA& PELTONEN 1997). Such and consequently does not lead to a restriction in recruit- a predictable shift in the diet composition of the roach oc- ment success as postulated by MEHNER et al. (1994), be- curred in Bautzen Reservoir at the beginning of June 1998, cause reproduction of the small roach population in the when the population of daphnids collapsed. After this col- years 1994 and 1996 was sufficiently high to produce very lapse roach fed mainly on chironomids and molluscs. Even large year classes. ruffe up to 56 mm total length were frequently ingested by older roach, which has rarely been documented (MICHEL & Diet and impact on Daphnia galeata OBERDOP,FF 1995) and might partly account for the good growth of roach in Bautzen Reservoir. Results of the diet analysis were pooled for all roach caught The estimation of fish abundance represents a greater by trawl and beach seine, as diel migration is typical for source of error than calculation of consumption rates on an cyprinids (BOHL 1980; BRABRANDet al. 1990). This is sup- individual basis (HEWETT 1989). Our abundance estimates, ported by HORPPILA(1994) who found that roach caught in though, show little variation (apart from value on 18 May) the pelagic area had macrophytes and invertebrate larvae in and no seasonal trend, making a systematic underestimation their diet, which had been consumed in littoral areas. In addi- of the impact of the roach population on the Daphnia population, Bautzen Reservoir is not distinctly structured by macro- tion unlikely. The low proportion (< 0.2% d-1) of the standing phytes, so it can be assumed that roach feed on the zooplank- stock biomass of the daphnid population consumed by the ton of the whole lake. The diet of roach in Bantzen Reservoir age-0 roach and the two dominant older age groups is clear contained a high proportion of algae and macrophytes. This evidence that predation by roach was not the reason for the corresponds well with many other studies of roach in eu- collapse of the D. galeata population in 1998 in Bautzen trophic lakes (PERSSON 1983a, b; BRABRAND 1985; JAMET et Reservoir. Apart from roach, age-0 percids are the most im- al. 1990; HORPPILA 1994). Such feeding behaviour is not portant zooplanktivores (MEHNER et al. 1997, 1998a), while only a consequence of the availability of algae and macro- older percids are mainly benthivorous or piscivorous (DOR- phytes in the environment, but also a consequence of a lack HER et al. 1999 and in this volume). During early summer of animal food and resulting food competition (PERsSON (1995 and 1996) consumption of Daphnia by age-0 percids 1983a; BRABRAND 1985; PERSSON & GREENBERG 1990). never exceeded 1% d-1 of the standing stock of Daphnia Algae and macrophytes have little energy that can be used by biomass and was in the same range as that of roach in 1998 roach compared to animal food (HOFER et al. 1985). Never- (MENNER et al. 1997, 1998a).

66 Limnologica 31 (2001) 1 Conclusions DORNER, H., WAGNER,A. & BENNDORF,J. (1999): Predation by piscivorous fish on age-0 fish: spatial and temporal variability in a biomanipulated lake (Bautzen Reservoir, Germany). Hydrobiolo- The aim of biomanipulationis a high density of large filtering gia 408/409: 39-46. zooplankton. An important prerequisite for the maximum ef- - SCHULTZ,H., MEHNER, T. & BENNDORF, J. (2001): Interaction be- fect of biomanipulation is an optimal stock of planktivorous tween prey availability and feeding behaviour of age-1 and age-2 fish (BENNDORF 1990; WISSEL et al. 2000), which should be perch (Percafluviatilis) in a biomanipulated lake (Bautzen Reser- high enough to exert a sufficient predation pressure on inver- voir, Germany). Limnologica 31:11-16. tebrate predators (Chaoborus, Leptodora), but low enough to HAMBRIGH%K. D., DRENNER, R. W., McComas, S. R. & HAIRSTONE, allow daphnids to flourish. In the case of Bautzen Reservoir N. G. Jr. (1991): Gape-limited piscivores, planktivore size refuges, the stock of piscivorous fish is very high and has led to a and the trophic cascade hypothesis. Arch. Hydrobiol. 121: strong reduction of the planktivorous fish stock and thus to a 389-404. HANSON, E C., JOHNSON, T. B., SCHINDLER,D. E. & KITCHELL,J. E low impact of planktivores on the Daphnia galeata popula- (1997): Fish Bioenergetics 3.0 for Windows. University of Wis- tion. The existence of older roach, in a size refuge, with a re- consin Sea Grant Institute, Report WISCU-T-97-001, Madison, productive capacity much higher than that of the younger in- WI. dividuals (TowNSEND & PERROW 1989), may counteract the HARTMANN,J. (1977a): Sukzession der Fischertr~tge in kulturbedingt effects of biomanipulation. A solution to this problem is to eutrophierenden Seen. Fischwirt 27: 35-37. enhance further the size refuge threshold by even higher - (1977b): Fischereiliche Verfinderungen in kulturbedingt eutro- stocks of piscivores such as pike and wels, which are less phierenden Seen. Schweizer Z. Hydrol. 39: 243-254. mouth gape limited than pikeperch of the same body size. HEWETT, S. H. (1989): Ecological applications of bioenergetics mod- els. Am. Fish. Soc. Syrup. 6: 113-120. HOFER, R., KREWEDL,G. & KOCH, F. (1985): An energy budget for an omnivorous cyprinid: Rutilus rutilus (L.). Hydrobiologia 122: Acknowledgements: We thank S. HI]LSMANN for supplying un- 53-59. published data of Daphnia biomass, S. WORISCHKA,E WIELANDand HORPPILA, J. (1994): The diet and growth of roach (Rutilus rutilus all other members of our working team for support. We appreciate (L.)) in Lake Vesij~irvi and possible changes in the course of the comments of M. WETZEL, T. MEHNER and two anonymous refer- biomanipulation. Hydrobiologia 294: 35-41. ees. The study was financed by the German Research Council (DFG, - & PELTONEN, H. (11997): A bioenergetic approach on food con- Germany, project number BE 1671/2-2). sumption of roach (Rutilus rutilus (L.)) in a eutrophic lake. Arch. Hydrobiol. 139: 207-222. HOLSMANN, S., MEHNER, T., WORISCHKA, S. & PLEWA, M. (1999): Is References the difference in population dynamics of Daphnia galeata in littoral and pelagic areas of a long-term biomanipulated reservoir af- BEAN, C. W. & WINFIELD, I. J. (1995): Habitat use and activity pat- fected by age-0 fish predation? Hydrobiologia 4081409: 57-63. terns of roach (Rutilus rutilus (L.)), rudd (Scardinius erythrophtal- JAMET, J. L., GRES, E, LAIR, N. & LASSERRE, G. (1990): Diel feeding mus (L.)), perch (Perca fluviatilis L.) and pike (Esox lucius L.) in cycle of roach (Rutilus rutilus, L.) in eutrophic Lake Aydat (Massif the laboratory: the role of predation threat and structural complex- Central, France). Arch. Hydrobiol. 118:371-382. ity. Ecol. of Freshw. Fish 4: 37-46. KITCHEL, J. E, STEWART,D. J. & WEININGER,D. (1977): Applications BENNDORF, J. (1990): Conditions for effective biomanipulation; con- of a bioenergetics model to yellow perch (Perca flavescens) and clusions derived from whole-lake experiments in Europe. Hydro- walleye (Stizostedion vitreum vitreum). J. Fish. Res. Bd. Can. 34: biologia 200/201: 187-203. 1922-1935. - (1995): Possibilities and limits for controlling eutrophication by KUBECKA,J. (1993): Succession of fish communities in reservoirs of biomanipulation. Internat. Revue ges. Hydrobiol. 80:519-534. Central and Eastern Europe. In: STRA~KRABA,M., TUNDISI,J. G. & - KOOP, J. H. E., SCHULTZ,H., WAGNER,A., DEPPE, T., HERBST, R., DUNCAN, A. (eds,), Comparative reservoir limnology and water HfJLSMANN, S., KAMJUNKE,N. & VOIGT, H. (1998): Biomanipula- quality management, pp. 153-168. Kluwer, Amsterdam. tion zur Sanierung eutrophierter Seen - Erweiterung der Anwen- MARMULLA, G. & R0scn, R. (1990): Maximum daily ration of juve- dungsmOglichkeiten. Bundesministerium far Bildung, Wissen- nile fish fed on living natural zooplankton. J. Fish Biol. 36: schaft, Forschung und Technologie; Schlugbericht zum 789-801. Forschungsvorhaben FKZ 0339549; 65 pp. MEHNER, T., SCHULTZ,H., BENNDORF,J., RITTER, P. & KNESCHKE, H. BOHL, E. (1980): Diel pattern of pelagic distribution and feeding in (1994): Gew~ssersanierung durch Nahrungskettensteuerung? Die planktivorous fish. Oecologia 44: 368-375. Talsperre Bautzen (Deutschland). Osterreichs Fischerei 47: BRABRAND, A. (1985): Food of roach (Rutilus rutilus) and ide (Leu- 204-21 I. ciscus idus): significance of diet shift for interspecific competition - SCHULTZ,H. & HERBST, R. (1995): Interaction of zooplankton dy- in omnivorous fishes. Oecologia 66: 461-467. namics and diet of 0+ perch (Percafluviatilis L.) in the top-down - FAAFENO,B. A. & NILSSEN,J. P. M. (1990): Relative importance of manipulated Bautzen reservoir (Saxony, Germany) during sum- phosphorus supply to phytoplankton production: fish excretion mer. Limnologica 25: 1-9. versus external loading. Can. J. Fish. Aquat. Sci. 47: 364-372. -- PLEWA,M., HI]LSMANN,S. & VOIGT, H. (1997): Age-0 fish preda- CHRISTENSEN, B. & PERSSON, L. (1993): Species-specific antipreda- tion on daphnids - spatial and temporal variability in the top-down tory behaviours: effects on prey choice in different habitats. manipulated Bautzen Reservoir, Germany. Arch. Hydrobiol., Behav. Ecol. Soeiobiol. 32: 1-9. Spec. Issues Advanc. Limnol. 49:13-25.

Limnologica 31 (2001) 1 67 -HfJLSMANN, S., WORISCHKA, S., PLEWA, M. & BENNDORF, J. many): Results of direct and indirect effects of biomanipulation. (1998a): Is the midsummer decline of Daphnia really induced by Limnologica 26: 153-164.

age-0 fish predation? Comparison of fish consumption and Daph- - WIELAND,E ~% BENNDORF,J. (1992): Raubfischbesatz zur Regula- nia mortality and life history parameters in a biomanipulated tion des Fischbestandes in einer hypertrophen Talsperre. In: M. reservoir. J. Plankton Res. 20: 1797-1811. VON LUKOWlTZ (Hrsg.), Die Bedeutung der flschereiwirt- - MATTUKAT,E, BAUER, D., VOIGT, H. & BENNDORF, J. (1998b): schaftlichen Bewirtschaftnng ftir die aquatischen Lebensr~iume. Influence of diet shifts in underyearling fish on phosphorus Arbeiten des Deutschen Fischerei-Verbandes 55: 57-92. recycling in a hypertrophic biomanipulated reservoir. Freshw. TOWNSEND, C. R. & PERROW, M. R. (1989): Eutrophication may pro- Biol. 40: 759-769. duce population cycles in roach, Rutilus rutiIus (L.), by two con- MICHEL, R & OBERDORFF, T. (1995): Feeding habits of fourteen Eu- trasting mechanisms. J. Fish Biol. 34: 161-164. ropean freshwater fish species. Cybium 19: 5-46. -- WINFIELD,1. J., PEIRSON, G. & CRYER, M. (1986): The response of PERSSON, L. (1983a): Food consumption and the significance of de- young roach (Rutilus rutilus) to seasonal changes in abundance of tritus and algae to intraspecific competition in roach Rutilus ru- microcrustacean prey: a field demonstration of switching. Oikos tilus in a shallow eutrophic lake. Oikos 41: 118-125. 46: 372-378. - (1983b): Effects of intra- and interspecific competition on dynam- VAN DENSEN, W. L. T. & GRIMM, N. P. (1988): Possibilities for stock ic and size structure of a perch PercafluviatiIis and a roach Rutilus enhancement of pikeperch (Stizostedion lucioperca) in order to in- rutilus population. Oikos 41: 126-132. crease predation on planktivores. Limnologica 19: 45-49. - & GREENBER6,L. A. (1990): Juvenile competitive bottlenecks: the WISSEL, B., FREIER, K., MIFLLER,B., KooP, J. & BENNDORF,J. (2000): perch (Perca fluviatilis) - roach (Rutilus rutilus) interaction. Ecol- Moderate planktivorous fish biomass stabilizes biomanipulation ogy 71: 44-56. by suppressing large invertebrate predators of Daphnia. Arch. Hy- RADKE, R. J. (1998): Strukturbitdende Prozesse in Fischartengemein- drobiol. 149: 177-192. schaften mesotropher Seen des norddeutschen Tieflandes. Disser- WYATT, R. J. (1988): The cause of extreme year class variation in a tation, Universit~it Konstanz. 139 pp. population of roach, Rutilus rutilus L., from a eutrophic lake in SCHULTZ, H. (1992): Bestandsgr63e, Wachstum und Zooplank- southern England. J. Fish. Biol. 32: 409-421. tonkonsum der Kleinen Marine (Coregonus albula) und anderer Fischarten im Arendsee. Limnologica 22: 355-373. Author's address: UWE KAnL, Dresden University of Technology, - (1996): Drastic decline of the proportion of males in the roach Institute of Hydrobiology, D - 01062 Dresden, Germany; e-mail: (Rutilus rutilus L.) population of Bautzen Reservoir (Saxony, Ger- [email protected]

68 Limnologica 31 (2001) 1