346J. N. Am. Benthol. Soc., 2005, 24(2):346–356䉷2005 by The North American Benthological SocietyEvaluating active and passive sampling methods to quantify crayfishdensity in a freshwater wetlandNATHANJ. DORN1,RAULURGELLES2,ANDJOELC. TREXLER3Department of Biological Sciences, Florida International University, Miami, Florida 33196 USAAbstract. We evaluated the sampling efficacy of 1-m2throw traps (active sampler) and baitedminnow traps (passive sampler) across an experimental density gradient (1, 2, 4, 6, 8, 10, and15/m2) of the slough crayfish (Procambarus fallax) in 2 trials with different crayfish populations. Inboth trials, throw-trap density estimates were highly correlated with actual crayfish density (r2⫽0.96). The form of the relationships between density estimates and stocked densities was consistentbetween trials, and indicated that throw traps captured a similar proportion of the stocked crayfishregardless of the stocked density. When we adjusted the relationships to account for clearing effi-ciency (proportion of captured animals actually recovered from the trap), the slopes of the regressionswere not significantly different from 1 in either trial. Size distributions and sex ratios of crayfishcollected by the throw traps accurately reflected those of the stocked populations. Baited minnowtraps performed inconsistently between the 2 trials. Catch-per-unit-effort (CPUE) and density weresignificantly correlated only in Trial 2 (r2⫽0.82). The slope of the regression in Trial 2 (0.621) wassignificantly⬍1, and the intercept was positive and nearly significant (p⫽0.074), indicating thatminnow traps captured increasingly smaller proportions of the stocked crayfish as the stocked densityincreased (i.e., differences between CPUE values underestimated actual differences between stockeddensities along the gradient). Minnow traps were biased toward capturing large male crayfish, butthe form of the relationships between CPUE and density did not improve when large-male CPUEwas used in the regressions. Our results suggest that 1-m2throw traps provide better estimates thanbaited minnow traps of crayfish densities in shallow vegetated habitats.Key words: population size, crayfish, density, throw trap, minnow trap, Procambarus fallax, wetland.Sampling aquatic animals to estimate popu-lation size (or density) and community compo-sition is an important issue that has receivedconsiderable attention in a variety of settings in-cluding lakes (Chick et al. 1992, Lamontagneand Rasmussen 1993), streams (Rabeni et al.1997, DiStefano et al. 2003), freshwater wetlands(Jordan et al. 1997), and estuaries (Rozas andMinello 1997, Kneib and Craig 2001). A majorhurdle in obtaining good population estimatesis verifying the effectiveness of the samplingmethod. Passive sampling methods usually in-volve a trap of some sort that can be left in theenvironment to catch animals that come in con-tact with it. Passive sampling methods integratedensity over an unspecified area, and they de-pend on animal abundance and animal activitylevels. Therefore, they actually measure activi-ty–density (e.g., Collins et al. 1983). In contrast,active sampling methods usually involve manycollections or counts of animals within smallunits of volume or area scattered around an en-1E-mail addresses: [email protected]@[email protected]. The overall effectiveness of a sam-pling method depends critically on the ability ofthe method to quantify differences in popula-tion density and assemblage structure in spaceand time.Monitoring crayfish populations is of partic-ular interest to researchers and resource man-agers for several reasons. First, crayfish can havehigh standing-stock biomass and secondaryproduction rates in a variety of systems (Momotet al. 1978, Rabeni et al. 1995), and they are of-ten important prey items for larger, vertebratepredators (Rabeni 1992, Robertson and Freder-ick 1994, Dorn and Mittelbach 1999). Second,crayfish strongly influence community structureand ecosystem function through a variety of tro-phic and nontrophic activities (e.g., predation,herbivory, bioturbation, macrophyte removal;Lodge et al. 1994, Nystro¨m et al. 1996, Dorn andWojdak 2004). Third, crayfish are actively har-vested for human food or fish bait (Roell andOrth 1992, Gherardi and Holdich 1999). Last,crayfish diversity in North America is threat-ened by continued introduction of nonindige-nous species from other watersheds or conti-nents (Lodge et al. 2000). Thus, monitoring2005] 347QUANTITATIVE CRAYFISH SAMPLINGcrayfish populations has many benefits for man-agement and conservation priorities.Crayfish are an integral part of the Ever-glades food web, and they provide food forwading birds, frogs, alligators, and fish (Rob-ertson and Frederick 1994). Therefore, crayfishpopulations and assemblages have been target-ed as performance measures to assess the suc-cess of the Comprehensive Everglades Restora-tion Plan (CERP) (Ogden et al. 2003). At present,no basis exists for choosing among crayfishsampling methods, particularly in wetland en-vironments like the Everglades.Many active and passive methods have beenused to sample crayfish from aquatic habitats(e.g., baited traps, electrofishing, diver collec-tions, quadrat sampling devices; Roell and Orth1992, Lamontagne and Rasmussen 1993, Rich-ards et al. 1996, Rabeni et al. 1997, DiStefano etal. 2003) and some researchers have used a com-bination of active and passive methods (Harperet al. 2002). Different methods have been com-pared to one another in a relative sense (e.g.,Capelli 1975, Collins et al. 1983, Rabeni et al.1997), but no studies have compared the efficacyof sampling methods across a known densitygradient.We evaluated the ability of active (1-m2throwtraps) and passive (baited minnow traps) sam-pling methods to estimate crayfish density anddemographic population measures (i.e., sizestructure and sex ratio) of the slough crayfish(Procambarus fallax ) in field enclosures with aknown experimental density gradient. We de-fined throw-trap accuracy as the congruence be-tween the density estimates and the actual den-sity in the environment. We also estimatedclearing efficiency of the throw traps (averageproportion of animals enclosed by the throwtrap that is collected) to determine the extent towhich clearing inefficiency can cause inaccuratedensity estimates (see Jordan et al. 1997, Rozasand Minello 1997).Study SystemOur research was conducted in emergent