calanoid copepods, isopods, and nematodes caught 

 (Table 2), indicating that the capture rates of the 

 two trap types varied between locations. Differences 

 in the types and numbers of animals caught by each 

 trap are more important, however. 



Reentry traps were much more effective than 

 emergence traps at capturing a greater variety of 

 animals. The most striking differences are the much 

 larger numbers of harpacticoid and cyclopoid cope- 

 pods captured in reentry traps. In addition to quan- 

 titative differences the reentry traps also caught in- 

 dividuals of six groups that were not found in the 

 emergence traps (Table 1). The six groups included 

 the lancelets and five types of demersal larvae. 



Conversely, emergence traps were more effective 

 at capturing calanoid copepods and isopods. Speci- 

 mens of five other taxa were captured only in 

 emergence traps (Table 1). 



Analysis of the capture rate of each common 

 group by the two trap types shows significantly dif- 

 ferent assemblages (x" = 881068, df = 11, P « 

 0.001) (Fig. 2a). Furthermore, each trap type caught 

 some relatively rare groups, meaning that the lower 

 capture rate of the emergence traps did not prevent 

 them from capturing groups that did not appear in 

 the reentry traps. Analysis of the capture frequen- 

 cies of these rarer groups shows that the two trap 

 types capture different assemblages of organisms 

 (x^ = 25806, df = 9, P « 0.001) (Fig. 2b). There- 

 fore, the reentry and emergence traps sampled dif- 

 ferent fauna or sampled the same fauna differently. 



Discussion 



Varying migration patterns and swimming be- 

 haviors by the various taxa and groups can explain 

 the differences between the assemblages caught by 

 the two trap types. Ascending animals would have 

 to move 1 m off the bottom in order to be captured 

 by the emergence traps. Descending animals would 

 not have been captured at all by the emergence 

 traps, but would only have needed to be a few cm 

 off the bottom to enter the reentry traps. Thus, 

 reentry traps are more likely to capture demersal 

 organisms during their migration than emergence 

 traps if many of these organisms never move very 

 far up into the water column, as Alldredge and King 

 (1985) have shown. Reentry traps also captured set- 

 tling larvae, which presumably are migrating in only 

 one direction prior to establishing a sessile mode of 

 life. Such larval forms were a small fraction of the 

 total numbers of animals caught, but could be a sig- 

 nificant portion of the reentering fauna at times. 



Both trap types may have also captured some ani- 

 mals that are holoplanktonic as noted by Robichaux 

 et al. (1981), despite our efforts to prevent this 

 during deployment and recovery of the traps. Some 

 animals may have entered the traps by crawling 

 rather than from the plankton, as Scheibel (1974) 

 observed. Finally, placement of the traps after dusk 

 may have missed animals migrating at or before 

 dusk, but the errors caused by this artifact, as well 

 as errors due to incomplete recovery of animals, are 

 not likely to alter our results significantly. 



Another possible explanation for at least some of 

 the differences between capture rates of the two 

 trap types is differential avoidance of one trap type, 

 in this case the emergence traps. Given that emer- 

 gence traps consist of materials quite unlike those 

 that demersal zooplankton would normally en- 

 counter it should not be surprising that they might 

 seek to avoid contact with them. The narrow fun- 

 nel placed in the mouth of the collection bottles, 

 while necessary to retain animals that have entered 

 the bottle, may exclude others altogether. Some 

 demersal zooplankton, such as calanoid copepods, 

 are well known to exhibit an escape response when 

 placed in contact with surfaces. Reentry traps, on 

 the other hand, work partly by replicating natural 

 sand substrate, reducing the potential for avoidance. 



The results show clearly that different sampling 

 techniques yield variable numbers of animals, even 

 within the same taxon, and collect different groups 

 of animals. Thus, evaluation of the demersal zoo- 

 plankton depends strongly on sampling techniques. 

 Adoption of a single standard sampling technique 

 might appear to be a resolution of the problem, but 

 a standard approach should sample all the organ- 

 isms that exhibit demersal behavior in a given area, 

 and neither emergence trapping nor reentry trap- 

 ing does. Furthermore, Stretch (1985) has observed 

 that not all members of a demersal population 

 migrate each night, so trapping techniques that 

 depend on animal migration must consistently 

 underestimate the actual abundance of demersal 

 organisms in association with a given substrate. 

 Tendency to migrate may vary among species, 

 within the life cycle of a given species and from day 

 to day, making accurate sampling of the demersal 

 zooplankton by trapping a logistical impossibility. 

 Collection techniques that directly sample demer- 

 sal organisms in or on the substrate, such as airlift 

 sampling (Stretch 1985) or sediment coring tech- 

 niques commonly used to sample meiofauna, should 

 give more accurate abundance estimates, but must 

 be used in conjunction with one or more trapping 



842 



