Tropocyclops simplex abundance, diel distribution and the impact of the planktivore fish Stolothrissa tanganicae on the T. simplex population were studied at a deep-water site in Kigoma Bay, Lake Tanganyika during the wet period (October 2008–February 2009). The top 40 m were examined using discrete depth samples, filtered through a 40-μm mesh. Ovigerous and copepodids of Tropodiaptomus simplex exhibited a very clear diel vertical migration. Ovigerous females of Tropodiaptomus were negligibly low in the 0–30 m depth during the day, while at night they stayed above 20 m. Tropodiaptomus nauplii did not show any clear diel vertical migration. The contribution of adults, copepodite and nauplii stages to the total numbers of Tropodiaptomus simplex was 8.2, 8.6 and 83.2%, respectively. Stolothrissa tanganicae stomachs were analysed to quantify the contribution of zooplankton in their diets, which revealed a dominance of Tropodiaptomus females and the electivity indices proved that T. simplex females (with or without eggs) were highly selected by S. tanganicae, which could probably be explained by size-selective feeding. Life table analysis indicated that approximately 60% of T. simplex nauplii did not develop into copepodites, a loss which could not be explained by the results of the present study. 68% of the fish gut contents were contributed by female Tropodiaptomus individuals and the contribution of shrimps and Cyclopidae increased substantially to almost 20%. S. tanganicae consumed largely ovigerous Tropodiaptomus females. However, eggs found in the guts were significantly smaller than those found in the zooplankton discrete samples, suggesting that digestion had started. The calculated electivity indices underline that ovigerous and non ovigerous females of T. simplex are being positively selected in the feeding process. It infers existence of selective predation in L. Tanganyika by S. tanganicae on females of T. simplex (with or without eggs) and as a result distinct vertical migration as a predation escape mechanism is performed.
Lake Tanganyika is a permanently stratified meromictic lake (Plisnier et al., 1999). It relies on wind driven mixing and internal regeneration of nutrients for primary productivity (Coulter, 1991; Plisnier et al., 1999). The stratification of the lake, however, is very sensitive to small changes in temperature and wind stress (Lewis, 1996). According to Coulter (1991), the zooplankton community is dominated by protozoans, crustaceans, coelenterates and fish larvae; however, each of these taxa is represented by relatively few species.
In Lake Tanganyika, the pelagic crustacean community dominates the zooplankton in numbers and biomass (Coulter, 1991). The most common shrimp species are: Palaemon moorei (Calman), Limnocaridina parvula (Calman) and Limnocaridina tanganyicae (Calman), with the first two being more pelagic (Mathes, 1967; Kurki et al., 1999b). The copepod portion of the community consists of one calanoid species Tropodiaptomus simplex (Sars), and three cyclopoid species, Mesocyclops aequatorialis (Kiefer), Tropocyclops tenellus (Sars) and Microcyclops cunningtoni (Sars) (Coulter, 1991). Tropodiaptomus is the largest copepod in the lake, for instance, Dumont (1986) established the size range of an adult Calanoida to be between 800-1100 μm. The lengths of adult female Mesocyclops and Tropocyclops are about 1000 and 380–445 μm, respectively, with the latter being the smallest cyclopoid species in the lake (Kiefer, 1956). Cladocerans are reported to be absent due to predation (Harding, 1957). Copepods are an important link between pelagic primary production and planktivorous fish, i.e. mainly two clupeid species, Stolothrissa tanganicae (Regan) and Limnothrissa miodon (Boulenger), and juveniles of centropomid Lates stappersii (Boulenger) (Kurki et al., 1999a). The commercial fishery in Lake Tanganyika is essentially based on these species: clupeid species ca. 65% by weight and L. stappersii ca. 30% by weight. Catches in the northern and southern halves are dominated by clupeid-based and “perch” based fisheries respectively; while the central part (Kigoma) lands both species (Mölsä et al., 1999; Mannini et al., 1999; Plisnier et al., 2009). This sectional dominance is highly linked to zooplankton and phytoplankton abundance and dynamics in the lake (Kurki et al., 1999b; Mannini et al., 1999; Descy et al., 2005, 2010).
S. tanganicae juveniles feed mostly on phytoplankton (e.g. diatoms-Nitzschia and Navicula) in the inshore waters and on zooplankton (especially the calanoid Tropodiaptomus simplex and selectively on the shrimp Limnocaridina spp.) as they move into the pelagic habitat at about 50 mm (standard length, SL) (Marlier, 1957; Chéné, 1975; Coulter, 1991). The endemic sardines in the Lake Tanganyika depend on grazing copepod as a food source. However, based on characteristics of the gill–rakers and heterogeneity in the composition of copepods between stomach samples, Chéné (1975) concluded that feeding is unselective. Feeding in this species occurs at dusk and dawn following the diel vertical movement of zooplankton (Coulter, 1991). It stays in the depth range of 8–60 m, forming large schools at night at 8–15 m while scattering at 60 m during the day (Coulter, 1991; Plisnier et al., 2009). S. tanganicae reaches a maximum size of 100 mm SL, while longevity is estimated at 1.5 years (Mannini et al., 1999).
Previous zooplankton studies have emphasized crustacean zooplankton because of their importance to pelagic fishes. The information on zooplankton spatial distribution and selective feeding by planktivorous fishes is patchy. For example, calanoid copepods are regarded as the most important food items in the diet of Stolothrissa due to their large sizes (Coulter, 1991). However, sampling methods used in previous studies might have led to an underestimation of smaller groups and developmental stages of zooplankton since large mesh-size nets (≈100 μm) were used (Kurki et al., 1999a). Specifically this study aimed at determining the abundance and life history of T. simplex, and examining the impact of S. tanganicae on T. simplex.
Materials and Methods
The study was conducted at a site located 04°51.26′S and 29°35.54′E within the Kigoma Bay of Lake Tanganyika. This site is identical to the pelagic station sampled during the Food and Agriculture Organization of the United Nations (FAO)/Finnish International Development Agency Lake Tanganyika Research project (FINNIDA LTR) (Kurki et al., 1999a; Mölsä et al., 1999). The maximum depth at this site is about 500 m. The thermocline in this area oscillates between 70 and 60 m during September–early November and November–January (Chitamweba, 1999).
Zooplankton samples were collected on a biweekly basis, in the pelagic zone at 0900 and 2100 hours between October 2008 and February 2009. A single sample was collected using a 7.4-litre water sampler (Limnos Ltd., Finland) at 0, 5, 10, 15, 20, 25, 30 and 40 metres. Water samples were filtered through 40 μm plankton net. The zooplankton samples were preserved with a 4% formalin solution in the field and taken back to the laboratory for microscopic analysis. The day and night discrete samplings (i.e. experiments) were not done at the same day's cycle in order to allow enough time to process live samples. Zooplankton identification, examination and counting were done using a dissecting (Olympus model SD-ILK) and an inverted microscope (Leica DMIL) following Dumont and Maas (1988) and Vuorinen (1993) protocols. Pelagic copepods were divided into two main groups, calanoids and cyclopoids, the former being represented by Tropodiaptomus simplex only while the three species of cyclopoids occurring in the pelagic zone were pooled, as they were not identified to species level (i.e. all species and all copepodids).
Live material was also collected using 40 μm plankton net hauled at a speed of not more than 0.5 ms−1 from 30 m to the surface. Material was kept alive in a 5 litre bottle and in the laboratory adult males and females with ripe ovaries of T. simplex were sorted out and cultured at room temperature to get information on embryonic development time. For the laboratory experiments, 6 males and 6 females were placed in a Petri dish each containing 15 ml of lake water to increase the chances of mating and producing eggs. 10 ml of lake water (filtered through 40 μm) were renewed every morning and evenings in order to provide the animals with clean freshwater. Five drops of water collected from a nearby pond (Katosho), which contained high algal material, were added to each Petri dish/treatment whenever the water was renewed in the treatments. This provided a source of food to ripe females and males. Water temperature in the Petri dishes was recorded every 3 h during the experiment, and at the same time the developmental stages of the eggs were checked. Although the experiments were performed at room temperature, the temperature variation was in an acceptable range (SD ± 1.2°C). Females with newly laid eggs were picked out with a pipette and immersed in a separate Petri dish with the same conditions as mentioned above. Initial time was recorded; embryonic development was monitored every 3 h until nauplii were seen coming out of egg shells or moving in the water column; which marked the end of the experiment.
For the calculation of population dynamics the numbers of Tropodiaptomus eggs and adults derived from the day samples were corrected according to the results from the night samples. From the relationship, day results to night results a correction factor was calculated and applied to the day results.
For the calculation of population dynamics the development times of the various stages are needed. The embryonic development time (0.96 days) is derived from our own experiments (24.0°C, n = 9; 24.4°C, n = 11; 24.7°C, n = 14). For the description of the relationship between temperature and development time, the results of Hyvönen (1997) (mean experimental temperature: 26°C) were used since our own result covered a narrow temperature range (24.0°C–24.7°C). A quadratic model was used for the description of this relationship (Herzig, 1983) and the following regression equation was derived which is valid for temperatures ranging from 22–28°C:1997) for further calculations.
A graphical method (Southwood, 1978) was applied to determine the total number of individuals that pass through each instar in a generation. Successive numbers of eggs, nauplii and copepodites were plotted against time. The area under the curve was determined and divided by the mean development time. The methods assume that: there is constant mortality rate throughout a given instar; mortality during the moulting from instar n to n+1 cannot be distinguished from mortality during instar n+1 and is attributed to the latter. Population dynamic analyses of T. simplex were done using the Abacus Concepts in Stat View (Abacus Corporation).
Stolothrissa tanganicae gut analysis
S. tanganicae were caught using a lift net (8 mm mesh size) in the artisanal fishery by lowering it to ≈70 m water depth. A total of 137 freshly caught fish, with sizes ranging between 47 and 92 mm, were immediately injected with 4% formalin in their body cavity in order to arrest digestion of prey. Fish length and weight were measured. Fish guts were dissected, and food items identified and quantified. Eggs found in the guts together with other food items were counted and their diameter measured. Identification was done by looking at the undigested hard fragments such as furca of small crustaceans (Chéné, 1975) and other appendages like pereopod 5 (P5).
Plankton obtained during day and night (assuming that whenever they encounter prey item they capture it) together with gut counts were used for calculating an electivity index though it is believed that Stolothrissa feed mostly at night. Prey selectivity by fish was calculated using Ivlev's electivity index modified by Jacobs (1974):
Di is electivity index for prey item i, pi is the proportion of prey item i in the lake and ri is the proportion of that same prey item in the guts. Values of Di range from −1 to +1; 0 indicates random feeding while Di > 0 or Di < 0 indicates positive or negative selection by the fish, respectively.
Zooplankton vertical distribution (day-night)
Tropodiaptomus simplex copepodid stages showed clear vertical migration during the whole sampling period. The night time distribution of females, males and copepodids of T. simplex were significantly different (Kruskal-Wallis test, X2 = 9.86, d.f. = 2, p < 0.05). They were dominant in the 0–20 m water layer at night and 30–40 m during the day (Figure 1). Similarly, ovigerous T. simplex showed a pronounced vertical migration. They were below 40 m during the day and abundant above 20 m at night especially on 26 and 27 January 2009, respectively (Figure 2).
When all day and night data were pooled, two peaks of total zooplankton were observed (Figure 3), one in early November with 105,000 individuals m−3 and a second one in early January, with 112,500 individuals m−3. Total nauplii followed the same trend with mean density values of 75,000 and 70,000 individuals m−3, respectively. Total nauplii contributed about 69% (range: 60.4 to 77.3%) to the total number of counted zooplankton in the study period.
T. simplex population showed only one peak in early November reaching 37,500 individuals m−3 which was followed by a sharp decrease to about 15,000 individuals m−3; while in the remaining period the numbers varied between 10,000 and 15,000 individuals m−3 (Figure 4). The peak was significantly contributed by nauplii (96%), while copepodid stages and adults constituted the remaining 4%. The relative contribution of adults, copepodites, and nauplii to the total number of counted Tropodiaptomus was 8.2, 8.6 and 83.2%, respectively. The mean density of ovigerous Tropodiaptomus was 271 individuals m−3 (range: 46–574 individuals m−3) while for total adults was (mean ± SD; 1253 ± 556 adults m−3). The relative contribution of copepodids Tropodiaptomus to the total copepod zooplankton was relatively small (3.8%) throughout the study period. Three pulses of eggs (early November and December, finally mid-January), nauplii and copepodid stages were observed throughout the study period which if you divide the sampling period (103 days) by the generation time of Tropodiaptomus (Hyvönen, 1997) coincide with three generations (Figure 4). The pulses are observable at each developmental stage.
Population dynamics of Tropodiaptomus simplex
Three pulses, most likely generations, of Tropodiaptomus simplex were observed over the entire 103-day period (Figure 5; Table 1). In the three periods number passing through the naupliar stage did not differ from the number passing through the embryonic stage. However, a significant difference became obvious between the numbers passing through the copepodite stage. If one compares the total theoretical input to the adult stage (i.e. number passing through the copepodite stage) with the mean (±SD) standing stock of adults (1253 ± 556 individuals m−3), losses of about 60% might have occurred in the adult stage.
Trophic relationship between Stolothrissa tanganicae and Tropodiaptomus simplex
A total of 167 Stolothrissa tanganicae individuals were analysed for gut content, 46% of which contained food items while the remaining 54% were empty. The specimens examined had total lengths ranging from 47 mm to 92 mm. S. tanganicae guts contained mainly female Tropodiaptomus (up to >90% of the total gut content) individuals with some Cyclopidae (mainly adults) and males of Tropodiaptomus constituting their diets in mid-November (Figure 6). During early December there was reduced contribution to 68% of the total gut content of female Tropodiaptomus individuals. The same trend was also observed in the lake (Figure 5c) but shrimps and Cyclopidae increased substantially to almost 20% in the guts. In January Tropodiaptomus females regained their dominance as the prevailing food items. Tropodiaptomus eggs were also found in the stomachs of S. tanganicae. The mean (±SD) egg sizes of Tropodiaptomus found in the fish guts had a relatively small diameter (119.84 ± 6.19 μm) than those from the zooplankton discrete samples (126.89 ± 10.13 μm). The differences in egg sizes between the two samples were significant (t-test, t = −3.35, d.f. = 34, P < 0.05). The reduction in sizes of most eggs in the guts was probably caused by digestion. It follows therefore that a high number of the consumed Tropodiaptomus females were ovigerous. The calculated electivity indices underline the selective feeding of S. tanganicae (Figure 7). S. tanganicae showed a strong positive electivity towards ovigerous and non-ovigerous Tropodiaptomus females throughout the study period. Tropodiaptomus males were also positively selected although not always. At the beginning of the study period Tropodiaptomus copepodid were selected as food items while in the following months no positive electivity was observed. Lastly, at the beginning of this study, shrimps showed a negative electivity index; however, from December onwards their electivity index became positive indicating that they were selected by the fish as an important food component. No selection for Cyclopidae was observed and the electivity indices were negative throughout the investigation period.
A clear strong vertical migration was displayed by ovigerous T. simplex during the whole study period. They concentrated in the lower layers (i.e. 30–40 m) and occasionally staying at the 20–25 m water layer during day time, while at night they came up to the surface layers (5–15 m). Vuorinen et al. (1999) in their study observed that intensive visual predation by planktivore fishes which suggested distinct diel and ontogenetic vertical migration of zooplankton, egg-carrying females of Tropodiaptomushave been found in deeper water than other adult stages. The observation in previous studies (Sandström, 1980; Narita, 1983; Mulimbwa, 1988, 1991; Flinkman et al., 1992 in Vuorinen et al.,1999) have shown that zooplankton may migrate daily from 80 m up to the surface during the season of maximum stability because they are being selected by predaceous clupeids which is in agreement with this study. This work points out differences in the distribution of females, males and copepodids in the surface waters might be a good indication of the importance of predation on the T. simplex population given the selectivity factors measured and the night time distribution of fish (8–15 m). Females, males and copepodites of T. simplex revealed the same trend of vertical migration. This behaviour could also reflect the effect of light intensity on the vertical and horizontal migration of zooplankton, such that recently photosynthesizing phytoplankton (and higher oxygen and/or temperatures) is likely drawing T. simplex to the surface layers. This is also in accord with post naupliar distribution pattern observed during this study. It is plausible that T. simplex is vulnerable to predation. As a result, this species has developed night feeding strategy by performing a clear vertical migration which in turn reduces predation pressure from clupeids (Frost, 1988; Dawidowicz et al., 1990 in Lo et al., 2004) and photo damage (Haney, 1988 in Lo et al., 2004).
The present study recorded a 3.8% contribution of copepodids to the total number of observed zooplankton which was lower than in previous studies (e.g. Kigoma (17%); Kurki et al., 1999a). The average high abundance of ovigerous Tropodiaptomus (271 ind. m−3) in this study compared to the literature was due to the fact that day and night samples were pooled to get an idea of their “real” abundance considering the pronounced vertical migration. Other studies reported lower values; for example, Kurki et al. (1999a) recorded 3–36 ind. m−3 and 2–138 ind. m−3 during the wet season in 1993–1994 and 1994–1995, respectively. The difference can be attributed to sampling methods; in the present study night samples were also used in estimating the calanoid abundance as day samplings to a depth of 40 m would most likely underestimate the calanoid abundance. It is worth noting that in the previous studies quoted herein sampling was done at depths up to 100 m.
Tropodiaptomus pulses observed from field data during this study are in agreement with the proposed Tropodiaptomus generation time by Hyvönen (1997). The results of the graphical method applied for the determination of total numbers of individuals passing through each instar are in agreement with the field observation that there is a big gap between numbers of nauplii and copepodites in the T. simplex population development. Predation on the copepodite stage could be the reason for this pattern. However, S. tanganicae gut contents and electivity indices did not show a strong positive selection towards copepodides although they were always found in the guts. The difference between calanoids and cyclopoids as food for clupeids is significant: a calanoids nauplius is comparable with a small cycloploid adult in biomass (Kurki et al., 1999a). However, natural mortality and food scarcity at the copepodite stage of T. simplex, in their natural environment, could have caused the observed pattern.
It was established earlier that S. tanganicae feeds predominantly on calanoid copepods and only sporadically on Limnocaridina spp. which is a much larger prey item (Chéné, 1975). The results of the present study in general correspond with Chéné's (1975) findings, but differ in that the food item ‘Tropodiaptomus’ was split into various developmental stages. Adult females of Tropodiaptomus, especially ovigerous females, were found to be the most dominant food items during the whole period. Generally, T. simplex is the largest component of the copepod community in Lake Tanganyika (Vuorinen et al., 1999). Increasing selectivity by a vertebrate predator with increasing prey size has been demonstrated and is predicted by optimal foraging theory, which assumes that a forager maximizes net energy intake per unit time (Werner and Hall, 1974; Bohl, 1982; Pyke, 1984; Lazaro, 1987; Walton et al., 1992 in Liu and Herzig, 1996). Copepodids of Cyclopidae were the next dominant food item after female T. simplex in the Stolothrissa guts. This was probably due to the fact that adult female Mesocyclops aequatorialis were larger than adult male T. simplex, which was the least dominant prey item in the Stolothrissa guts.
Shrimps were occasionally found in the guts of Stolothrissa (Chéné, 1975). This is probably due to their larger size compared to other crustacean zooplankton found in the lake. Electivity indices do not show positive selection probably due to the fact that the biomass of shrimp taken is a significant proportion compared to other prey items. The smallest shrimp size found during the study was the same as an ovigerous Tropodiaptomus; hence one may assume that they are also suffering heavy predation by S. tanganicae although they were not as abundant as other zooplankters in the environment or in S. tanganicae guts. An interesting trend was also made during this study: when adults of T. simplex were much less abundant in the lake, they were also less abundant in the diet, even though they were positively selected.
Electivity indices confirmed that ovigerous and non-ovigerous females of T. simplex were highly selected by Stolothrissa. Ovigerous T. simplex scored high in the electivity indices followed by female Tropodiaptomus. A lot of copepod eggs (mean ± SD = 382.88 ±114.36, n = 137) were observed in the Stolothrissa guts; however, their sizes differed significantly from eggs measured from Tropodiaptomus bodies. Individual T. simplex had eggs with larger mean diameter values than those from the guts of Stolothrissa. This suggests that probably digestion process could have caused the reduction in egg sizes in the guts of Stolothrissa by almost 7.05 μm diameter. The observation that S. tanganicae had T. simplex eggs in their guts may suggest selective predation of ovigerous T. simplex by the species. It further points into the direction that some ovigerous T. simplex were consumed as selectivity data in this study indicated an existence of positive selection of ovigerous T. simplex.
Zooplankton tends to exhibit maintenance swimming (Dussart and Defaye, 1995) while carrying a darkish coloured sack of 2–12 eggs (this study). Carrying a sack of eggs makes them a more visible target than other zooplankton to the plankton feeder. In a similar case, Liu and Herzig (1996) established that cyprinid Pelecus cultratus was selectively feeding on zooplankton according to the eye size of prey rather than its length resulting in a greater preference for large sized Diaphanosoma (with large eyes) than for Leptodora which have smaller eyes. They concluded that prey visibility due to pigmentation may be an important factor determining selective feeding behaviour in visual feeding fish. S. tanganicae have been reported to have good vision as indicated by large size of the optical lobes in their brain (Coulter, 1991). Additionally, in Lake Victoria, Owili et al. (2003) established that maximum prey capture for maximum energy gain possibly due to relatively high water clarity (Secchi depth 0.5–0.9 m) recorded at their study site, as opposed to those (0.3–0.4 m) recorded at different sites (Sondu and Awach). The predation pressure on the cladocerans was enhanced since the visually feeding fishes are likely to predate on relatively bigger prey for maximum energy gain. Good vision, in combination with the highly transparent waters of the lake, is instrumental in prey selection whereby females (with or without eggs) of T. simplex are selectively preyed upon by S. tanganicae.
In conclusion, the present study found large gaps in T. simplex population development between numbers of nauplii and copepodites. This high mortality might be explained by starvation of the copepodites and/or by fish predation. However, the latter might be underestimated on the basis of the current gut analyses. Optimal temperature is a key factor in embryonic development; in the laboratory temperature condition is not stable as compared to the lake. This might affect the mating process and embryonic development which in turn affect the number of nauplii produced and developing to the copepodid stage and finally adults. This study also confirms that selective predation exists in Lake Tanganyika especially by Stolothrissa tanganicae on female T. simplex (with or without eggs), and that T. simplex performs a pronounced diel vertical migration as a strategy to reduce predation pressure by S. tanganicae.
We are very thankful to the field and laboratory personnel at TAFIRI and Neusiedler See Biological Station, Austria for their assistance in the development of this article. We also wish to thank Dr. A. I. Kimirei for his constructive comments on early versions of the article. Finally, we would like to express our gratitude to the Austrian Government through ADC for funding this study.