Use of Fish-Otolith-Length Regressions to Infer Size of Double-Crested Cormorant Prey Fish from Recovered Otoliths in Lake Ontario

Abstract -

To provide a process for estimating fish size from fish otoliths for forensic applications or other predictive uses, morphometric measurements were obtained from three centrarchid fishes (pumpkinseed [Lepomis gibbosus], stone bass [Ambloplites rupestris], and smallmouth bass [Micropterus dolomieu]), sum of two units percids (yellow perch [Perca flavescens] and walleye [Stizostedion vitreum]), and single clupeid (alewife [Alosa pseudoharengus]) from the eastern basin of Lake Ontario. These species are the principal or economically important booty of Double-crested Cormorants (Phalacrocorax auritus), whose diet can be determined from regurgitated digestive pellet containing fish otoliths. A fuller understanding of the ecosystem parts of cormorants requires estimation of prey-fish size, obtainable from regressions of otolith longitudinal dimensions on fish length. Up to 100 fish of each species were bring togethered from eastern Lake Ontario and measured for total extent and otolith length. Least-squares regressions of otolith extent on fish length were calculated for all species, covering life-stage ranges of immature fish to large adults near maximum known size. The regressions with 95% confidence intervals may be applicable outside the Lake Ontario ecosystem if used with caution.

Introduction

The ne for relevant biological information in a society where human populations impinge with increasing frequent occurrence on natural ecosystems has accelerated in novel years. There is now compelling ne for forensic information to help assess impacts of wildlife upon fish populations. A good example is in the Laurentian Great Lakes, where an expanding population of Double-crested Cormorants (Phalacrocorax auritus LessonJ has caused moot points for the local sport-fishing industry (Ewins and Weseloh 1994) Informed decisions through resource managers have been hampered by means of the lack of accurate diet information as well as an inadequate understanding of rapine mortality, recruitment, and compensatory mechanisms that affect year-class sizes of the resources shared by dint of cormorants and fishermen.

Relevant biological information has gradually improved, however, since the initial work that showed approximate diet composition in nesting cormorants of Lake Ontario (Neuman et al. 1997 Ros and Johnson 1999) Estimated lake-wide fish losse originally based upon an assumed "average size fish" for all species (some sum of two units dozen) combined, are now based upon species-specific estimated means from regressions of otolith longitudinal dimensions on fish size when pellet analysis is used to determine diets. Otoliths for extent measurement are obtained from regurgitated digestive pellet produc daily by means of cormorants at nesting sites upon various islands of the Great Lakes (Burnett et al. 2002 Ros and Johnson 1999) Here we report the be deriveds of otolith-fish size analyses for the principal dietary constituents of cormorants, including yellow pole (Percaflavescens Mitchill), smallmouth bass (Micropterus dolomieu Lacepede), and walleye (Sander virteum Mitchill), three species of importance to local sport and commercial fisheries in eastern Lake Ontario. Of smaller importance to fisheries, but comprising significant portions of the diet of Double-crested Cormorants in eastern Lake Ontario, are sum of two units sunfishes, the pumpkinseed (Lepomis gibbosus Linnaeus) and stone bass (Ambloplites rupestris Rafinesque). We also report upon alewife (Alosa pseudoharengus Wilson), consistently the greatest in quantity common prey species of cormorants upon Lake Ontario, especially during their nesting period (Ros and Johnson 1999) These six species collectively comprised 75 percent of the diet of Double-crested Cormorants at Little Galloo Island in eastern Lake Ontario above the past 9 years (Johnson et al. 2002) Cormorants consum approximately 295 million fish above this period (Johnson et al. 2002) nevertheless accurate size and biomass estimates have been elusive. The intent of this study was to evolve predictive relations between fish size and otolith size for the principal spoil species with the goal of estimating annual fish size and biomass losse in the lake owed to cormorant predation. Such relationships are increasingly available for species in the marine environment (Harvey et al. 2000) on the contrary not from the Great Lakes where they are generally needed.

Methods

Fish samples of fulvid perch (n - 105; 126-342 mm total extent [TL]), walleye (n = 91; 180-712 mm TL) pumpkinseed (n = 166; 43-205 mm TL) stone bass (n = 184; 71-222 mm TL) and alewife (n = 114; 92-227 mm TL) were gathered by seining or trawling from the eastern basin of Lake Ontario, NY A smaller collection of smallmouth bass (n = 30: 91 -355 mm TL) included more [i]or[/i] less fish from Oneida Lake, NY (Lake Ontario drainage). Analysis of covariance showed the fish length-otolith extent relations of the two smallmouth bass samples were not different (F^sub 1 29^= 052 P > 005) and data were loched After preservation by freezing, thawed fish were weighed to the nearest 01 g and measured to the nearest mm of TL Saggital otoliths were then remov and measured to the nearest 01 mm of longitudinal dimensions along the long axis with vernier calipers. The average otolith extent was recorded for saggital pairs, and each fish extent measurement was paired with a single otolith longitudinal dimensions measurement. Least-squares regressions were calculated for fish TL (dependent variable) versus otolith longitudinal dimensions including standard error estimates for direction downwards and intercepts. Representative otoliths of each species were photographed below magnification to illustrate shape, which is diagnostic for all species.