Hydrobiologia 442: 89–99, 2001. 89 M. Boersma & K.H. Wiltshire (eds), Cladocera. © 2001 Kluwer Academic Publishers. Printed in the Netherlands. Biotic and abiotic preferences of the cladoceran invader Limnosida frontosa Thomas C. Jensen1, Dag O. Hessen1 & Bjørn A. Faafeng2 1Department of Biology, University of Oslo, P.O. Box 1027, Blindern, N-0316 Oslo, Norway 2Norwegian Institute for Water Research. P.O. Box 173, Kjelsås, N-0411 Oslo, Norway Key words: Limnosida frontosa, abiotic parameters, food spectrum, particle selection Abstract During the late 1980s and early 1990s, the cladoceran Limnosida frontosa invaded several lakes within its natural range in southeastern Norway. In this project, we wanted to study the types of lakes preferred by Limnosida.We also wanted to evaluate the potential competitive effects on other zooplankton species. In a survey of 65 Norwegian lakes, Limnosida showed preference for lakes of low Ca concentrations and low productivity. This is probably due to decreased competition from species with higher Ca requirements and a lower fish predation on zooplankton in these lakes compared to more productive lakes. Particle size preferences of Limnosida were studied and compared with those of the microfiltrator Daphnia magna, as there was no published information on the food preference of Limnosida. When fed monodisperse fluorescent latex beads (0.5, 1.0, 6.0 µm), Limnosida strongly selected the largest beads, while D. magna had a more nonselective feeding behaviour. Mesh sizes of Limnosida’s filtering appendages were 0.4–1.2 µm depending on the animal size, and the particle selection was correlated with the filter mesh sizes. Limnosida should thus be considered a low efficiency grazer on bacteria and µ-algae. Hence, this species probably does not interfere significantly with microfiltrators like Diaphanosoma brachyurum and most daphnids. This was supported by community analysis of lakes with and without Limnosida. In general, Limnosida commonly co-occurred with a number of filter-feeding cladocerans, and we found no sign of competitive exclusion in lakes where the species has become established. Introduction A species’ distribution depends on interactions between a number of biotic and abiotic factors, which During the late 1980s and early 1990s, the clado- together form the multidimensional niche of the spe- ceran Limnosida frontosa Sars invaded several lakes cies. Though invasion may lead to extinction of native in southeastern Norway. (Walseng & Karlsen, 1997; species through competitive exclusion (Hardin, 1960), Faafeng, unpubl. data). Invasion of species has at- competing zooplankton species may also coexist (Jac- tracted much attention, because ecosystem properties, obs, 1977; Chow-Fraser & Maly, 1992). such as productivity and community structure may Differing abiotic demands among zooplankton be influenced by invading species (Mooney & Drake, species promote coexistence (Havens et al., 1993; 1989). Zooplankton invaders may also affect the native Hessen et al., 1995a). Temporal or spatial niche dif- zooplankton communities. For example, after invad- ferentiation will also allow competitors to coexist ing Lake Naivasha, Kenya, Daphnia pulex outcom- (Leibold, 1991; Angeli et al., 1995). Furthermore, bi- peted other zooplankton species (Harper, 1984), and otic factors such as size selective predation (Brooks the invasion of Bythotrephes cederstroemii in lakes in & Dodson, 1965; Zaret, 1980), age structure (Lynch, the northeastern U.S.A., strongly affected the native 1978) and intraspecific competition (Hu & Tessier, invertebrate predator Leptodora kindti as well as the 1995) have been used to explain coexistence between herbivorous zooplankton (Lehman & Cáceres, 1993; zooplankton species. Food is also an important factor Yan & Pawson, 1997). as far as the competitive outcome between species is 90 concerned, as different species utilize different kinds al., 1995). Univariate logistic regression describes the of food (Neill, 1975; Chow-Fraser & Maly, 1992) or probability (p) of an event (presence/absence) by a food size spectra (Burns, 1968; Geller & Müller, 1981; probability ratio as a linear function of the independent Hessen, 1985). This kind of niche differentiation will variable (x): promote coexistence. p log D b C b · x Limnosida was originally described by Sars (1993 10 1 − p 0 1 [1861]). The species is planktonic (Lilljeborg, 1982 This expression describes a situation where an organ- [1901]; Sars, 1993 [1861]) found from May to Octo- ism has an optimum for either high or low values of ber, though it is mainly present in mid summer (Lillje- the independent variable. borg, 1982 [1901]; Vekhov, 1987). Pejler (1965) found Logistic regression can also be used to assess the species in lakes with relatively low productivity. species preferences along the gradient axis of the inde- Limnosida is a palearctic species (Sars, 1865, 1993 pendent variable (Jongman et al., 1995). Furthermore, [1861]; Behning, 1912; Lilljeborg , 1982 [1901]; the univariate model can be extended to describe the Vekhov, 1987). probability as a function of two independent variables The objectives of this study was to examine (Jongman et al., 1995). This has been done in previous whether the distribution of Limnosida could be related zooplankton studies (Hessen et al., 1995a). Here, the to abiotic and biotic factors such as calcium concentra- above expression is extended to a quadratic polyno- tions, lake productivity parameters and fish predation mial. The bivariate model then becomes an expression on zooplankton. Co-occurrence of Limnosida with of the form: other cladocerans was also investigated. Since there is p 2 no information on the food preferences of Limnosida log b0 + b1 · x1 C b2 · x1 10 1 − p in the literature, an assay of particle size preferences C · C · 2 C · · was undertaken in order to obtain some basic in- b3 x2 b4 x2 b5 x1 x2 formation on food size selection and filtering rates where x1 and x2 are independent variables. to evaluate the potential competitive effects on other In the bivariate regression in this study, the in- zooplankton species. teraction between Ca and total P is considered. The criterion for accepting an effect in the logistic models was p<0.05. JMP 3.2.2 (SAS Institute Inc. Cary, Materials and methods NC) was used for statistical calculations. The presence/absence data for Limnosida was A survey of nutrient levels and plankton communities mainly based on this lake survey. In a few lakes, in a large number of Norwegian lakes has been con- the species was not registered in our lake survey, ducted since 1988 (Faafeng & Hessen, 1993). For the but it presence has nevertheless been verified in later present study, only lakes within the distribution area of studies. For the statistical treatment, the species was Limnosida were chosen. This included a set of 65 lakes considered as present in these lakes as well. in southeastern Norway, geographically delimited to The co-occurrence between Limnosida and other the north and east within the area where Limnosida cladocerans was assessed from observations on pres- occurred. Only lakes with pH>5 were included in or- ence/absence of the species in the survey for the same der to avoid lakes with severe acidification. Samples lakes during the years 1988–1993. In order to examine for total phosphorus (total P), total nitrogen (total N) whether Limnosida was associated with some of the and chlorophyll a from July and August from 1988 other species the coefficient of association, Yule’s Q, to 1993 were included in this study. Samples for the was calculated (Freeman, 1987). Q varies from −1to analysis of calcium (Ca) were taken from August or 1. A value of 0 indicates no association, values >0 September from 1988 to 1993. Further description of indicate positive association, and values <0 indicate the sampling and analysis is given in Hessen et al. negative association. (1995a, b). The fish species composition of the lakes was cat- The relationship between distribution of Lim- egorised on basis of questionnaires and reports. Six nosida and Ca concentration and tentative lake pro- fish community categories were identified, ranging ductivity (total P) was assessed by logistic regression. from category 1 dominated by brown trout (Salmo This relates presence/absence of a species to con- trutta) to category 6 dominated by various cyprinids, tinuous or discrete independent variables (Jongman et also representing a supposed increase in fish predation 91 on zooplankton (Hessen et al., 1995b). The occurrence Table 1. Major characteristics for the 65 lakes. Data on total P, total N and chlorophyll a are from July and August samples from 1988 of Limnosida as related to fish community was then to 1993. Data on calcium are from August and September samples. − assessed, based on these categories. Area as km2; total P, total N and chlorophyll a as µgl 1;Caasmg − The particle size selection of Limnosida was as- l 1 sessed both by feeding experiments and examina- Average Median Max. Min. tion of the filtering appendages. In order to study particle size preferences of the species, feeding trials Area 12.6 32.1 362.8 0.2 with fluorescent latex beads were performed (Hessen, Total P 29 15 301 3 1985). The particle size preferences of Limnosida Total N 669 510 3061 237 were compared with those of the well known mi- Chlorophyll a 16.2 7.5 121.7 1.2 crofiltrator Daphnia magna (Geller & Müller, 1981; Ca 7.8 4.3 37.3 1.7 Brendelberger & Geller, 1985). Limnosida for the ex- periments was collected from Lake Gjersjøen and D. magna originated from a culture at the Department of Biology, University of Oslo. The animals from the is more than 15 min (Peters, 1984), we assumed that lake were collected by vertical net hauls (95 µm) from the incubation time of 10 min was shorter than GPT. the epilimnion at different dates during the summer of The selectivity index, Is (−1 to 1), of each bead 1997. They were transferred to the laboratory in lake size was calculated according to Ivlev (1961).