Sources of bias in the use of shell fragments to estimate the size of zebra and quagga mussels (Dreissena polymorpha and Dreissena bugensis)

Several researchers have examined size-selective predation on dreissenid mu ssels by first measuring septa from crushed mussel shells found in predator s' gastrointestinal tracts and then using a regression of septum length on shell length to infer the size of consumed mussels. We examine three assump tions made when using this approach: (1) that the shell length : septum len gth relationship is site-independent within the study area, (2) where both zebra mussels (Dreissena polymorpha) and quagga mussels (Dreissena bugensis ) are present, that the shell length : septum length relationship is the sa me for both mussel species, and (3) that the predator foraged exclusively a t the site of collection. We collected mussels at 6 sites along an 8-km str etch of Lake Erie shoreline and found that the shell length : septum length relationship varied significantly both among sites and between zebra musse ls and quagga mussels. We then compared the regression for quagga mussels a t one of these sites with that for intact valves of mussels taken from scau p (Aythya marila, Aythya affinis) collected at the site. Although ice cover at the time of collection restricted scaup to the site while foraging with in the study area, regressions were again significantly different, i.e., sc aup had been foraging elsewhere. Our results indicate that for at least som e study sites, the use of septa to estimate dreissenid mussel size is not a ppropriate. However, when intact valves are found in a predator, variation in shell morphology can help to confirm or exclude possible foraging locale s.