438 Size Selectivity in Summer and Winter Diets of Great Cormorant (Phalacrocorax carbo ): Does it Reflect Season-Dependent Difference in Foraging Efficiency? MARTIN ˇ CECH 1,2,3 , PAVEL ˇ CECH 2 , JAN ˇ KUBECKA 1 , MARIE PRCHALOVÁ 1 AND VLADISLAV ˇ DRASTÍK 1 1 Biology Centre of the Academy of Sciences of the Czech Republic, Institute of Hydrobiology, Na Sádkách 7, 370 05 ˇ Ceské ˇ Budejovice, Czech Republic 2 Czech Union for Nature Conservation, 02/19 ZO ˇ CSOP Alcedo, Blanická 1299, 258 01 Vlasim, ˇ Czech Republic 3 Internet: carcharhinusleucas@yahoo.com Abstract.—Diet of the Great Cormorant (Phalacrocorax carbo) was studied using regurgitated pellets, individual fish bones and fish remains collected from below roosting trees at ˇ Zelivka and Slapy Reservoirs, Czech Republic, during summer, a warm winter and a cold winter. Both reservoirs have the same trophic status and similar fish as- semblages. Using diagnostic bones (os pharyngeum, dentale, praeoperculare) and our own linear regression equations relating dimensions of the diagnostic bone and fish total length (L T ), a total of 2,055 fish of 18 species and four families were identified in the diet of Great Cormorants and their size was reconstructed. Both fish total length and weight differed significantly between seasons being, on average, 12.0 cm and 30 g during summer, 18.3 cm and 109 g during a warm winter and 22.8 cm and 157 g during the cold winter. The average weight of fish taken by Great Cormorants significantly increased with decreasing air and water temperature. The contribution of the dominant “large growing”, torpedo-shaped fish species in the diet of Great Cormorants dramatically increased from summer to the cold winter. In contrast, the contribution of dominant “small growing”, torpedo-shaped species, or humped body-shaped species, showed completely the opposite tendency. Great Cormorants seem to consume all fish of ap- propriate size that they are able to catch in summer and select for larger fish in winter. Thus, the winter elevation of foraging efficiency described for Great Cormorants in the literature is due to capturing larger fish not to captur- ing more fish. Received 21 March 2007, Accepted 9 March 2008. Key words.—Alburnus alburnus, diagnostic bones, Leuciscus cephalus, Perca fluviatilis, foraging efficiency, prey size, regurgitated pellets, Rutilus rutilus, Slapy Reservoir, ˇ Zelivka Reservoir. Waterbirds 31(3): 438-447, 2008 The Great Cormorant (Phalacrocorax car- bo) is one of the most important avian fish predators both in marine (e.g., Barrett et al. 1990; Leopold et al. 1998; Johansen et al. 1999, 2001; Lorentsen et al. 2004) and fresh- water (e.g., Suter 1995, 1997; Keller 1998; Engström 2001; Liordos and Goutner 2007) ecosystems. Moreover, recent studies have revealed that among endotherms the Great Cormorant is also the most efficient fish predator. Prey capture rate of this bird was estimated to be 9.0 g min -1 for females and 15.2 g min -1 for males in summer (Grémillet 1997) and nearly 60 g min -1 for both sexes in winter (Grémillet et al. 2001). Such a high predatory efficiency is a necessary result of a trade-off between time spent diving into cold water (1-10°C), very limited body insulation, maintaining a high body temperature (42.3°C) and balancing of daily energy re- quirements (Grémillet et al. 2001), since high energy losses during winter do not seem to be compensated for by an increase in fish intake (Johansen et al. 2001). The win- ter elevation of foraging efficiency depends critically upon dense, highly predictable prey stocks allowing birds to gain sufficient energy during an extremely short dive ses- sion (usually one per day) lasting, on aver- age, only 9 min (Grémillet et al. 2001). In principle, wintering Great Cormo- rants could increase their catch per unit of effort (C.P.U.E.) in two different ways. Either they could catch more fish or they could catch larger fish compared with the summer situation. The former is partly possible by choice of appropriate fishing localities, such as overstocked carp ponds (Adámek and Kor- tan 2002), otherwise it is unlikely (Johansen et al. 2001) because it would require a mirac- ulous enhancement of their fishing abilities. The latter would be possible if the accessibil-