476 



Fishery Bulletin 89(3). 1991 



two clam dredges used. The #41 trawl catches were 

 divided by 1.79, the ratio of catch-per-tow of Cancer 

 spp. by the #41 net to the #36 net on the RV Albatross 

 IV (Sissenwine and Bowman 1978), to equate them to 

 the #36 trawl catches. The 1.22 m dredge catches were 

 multiplied by 1.56 [(1.52/1.22)(5/4)] (S.A. Murawski, 

 NMFS Woods Hole Lab., Northeast Fish. Sci. Cent., 

 Woods Hole, MA, pers. commun., Oct. 1987) to make 

 them comparable with catches of the 1.52 m dredge. 

 Carapace widths, measured between the tips of the 

 anterolateral spines, were determined at sea for all 

 specimens. When occasional collections totaled more 

 than 100 crabs of one species, a random subsample of 

 about 100 was measured and sexed, and the results ex- 

 panded to estimate total catch. Width frequencies are 

 presented from trawl collections only, because dredge 

 collections included few crabs <5cm. 



Data on sex were available for rock crabs and Jonah 

 crabs from all surveys. Sexes of northern lady crabs 

 and coarsehand lady crabs were available from winter 

 dredge surveys only. 



For analysis of the relative abundance of crabs, the 

 survey area was divided into five strata, i.e., Gulf of 

 Maine offshore and inshore, Georges Bank, and middle- 

 Atlantic offshore and inshore. The mean and variance 

 of catch-per-tow in each stratum were estimated using 

 the delta distribution, which considers the lognormally 

 transformed catch at positive tows, i.e., tows with crabs 

 (Pennington 1983). Estimates for the five strata were 

 combined by weighting by stratum area in km 2 using 

 equations for stratified mean and variance (Survey 

 Working Group, NEFSC 1988). Trawl and dredge data 

 were handled separately. Weighted estimates of abun- 

 dance were obtained by sex, time of capture, and bot- 

 tom temperature for each species of crab. Relative 

 abundances in numbers and weight were calculated by 

 year from fall surveys, because the same size net was 

 used each year. 



To examine diel variability in catch, the stations were 

 divided into four groups, i.e., dawn, day, dusk, and 

 night, using the starting time of tow. For each season, 

 two 4-hour periods of low light were defined, and the 

 remaining hours were full light and full dark. For ex- 

 ample, for dawn in winter the earliest and latest times 

 of sunrise were determined for all months and latitudes 

 of the survey, and 1 hour was added before and after, 

 to yield a 4-hour period. For winter surveys, periods 

 were, in hours: dawn 0501-0900, day 0901-1500, dusk 

 1501-1900, night 1901-0500. For spring and fall sur- 

 veys, hours: dawn 0401-0800, day 0801-1600, dusk 

 1601-2000, and night 2001-0400. For summer surveys, 

 hours: dawn 0301-0700, day 0701-1700, dusk 1701- 

 2100, and night 2101-0300. Weighted estimates of 

 abundance by time period were compared by analysis 



of variance for unequal sample size (the GT2 multiple 

 comparison method in Sokal and Rohlf 1981). 



For rock and Jonah crabs, sex ratio was calculated 

 by dividing the weighted mean abundance of males by 

 that of females. The mean abundances of each sex were 

 compared by t tests. 



We used stepwise regression analysis to fit linear 

 models of abundance (all species) and sex ratio (rock 

 and Jonah crabs) by the independent variables of depth, 

 temperature, and their interaction, using the method 

 of least squares (SAS Institute Inc. 1985). The depen- 

 dent variables were transformed to logarithms as In 

 [abundance + 1] and In [abundance of males/abundance 

 of females]. The data were fit by seasons and subareas. 



Results and discussion 



Distribution and abundance 



Rock crabs Rock crabs were distributed throughout 

 the shelf, with their center of abundance extending 

 from Georges Bank to Cape Hatteras in depths of 

 6-456 m (Figs. 3a-b, 4a-d). The principal sediments in 

 that area are sand and sand/gravel (Uchupi 1963, 

 Schlee 1973). Previous studies found rock crabs also 

 on rocky (Scarratt and Lowe 1972) and muddy (Krouse 

 1980) sediments, and on blue mussel Mytilus edulis, 

 (Reilly and Saila 1978), oyster Crassostrea virginica, 

 and shell beds (C.L. MacKenzie Jr., unpubl. data). 



Seasonal distribution patterns of rock crabs in the 

 middle-Atlantic subarea show that they migrate across 

 much of the shelf, moving inshore during cold months 

 and offshore during warm months. Winter surveys 

 show rock crabs concentrated in depths <40m, with 

 only scattered occurrences farther offshore (Fig. 3a-b). 

 In spring, their distribution remained essentially un- 

 changed (Figs. 4a, 5). In spring trawl surveys, inshore 

 stations averaged 22.3 crabs per tow whereas offshore 

 stations averaged only 1.1 crabs per tow (Table 2). 

 Later in the year, part of the population moves off- 

 shore, resulting in a more even distribution on the shelf 

 during summer and fall (Figs. 4b-d, 5). Catch-per-tow 

 during the fall trawl surveys averaged 1.2 crabs at 

 inshore stations and 0.9 crabs at offshore stations 

 (Table 2). 



In the Gulf of Maine, rock crabs occurred only in 

 coastal zones from spring through fall (Fig. 4a-d; Table 

 2). We have no data on the winter distribution of rock 

 crabs in that subarea, but Krouse (1972, 1976, 1980) 

 reported that they occupy coastal zones year-round in 

 Maine. 



Triggering cues for seasonal migrations in the 

 middle-Atlantic are unknown, but they are apparently 

 related to seasonal cooling and warming. Terretta 



