FISHERY BULLETIN: VOL. 85, NO. 2 



Table 1 .—Catch statistics for Decapterus punctatus in the South 

 Atlantic Bight, n = number of trawls; x.,, = stratified mean 

 catch tow"\ X|n = stratified mean catch tow"' for ln(x + 1) 

 transformed data; Xgi.ss = Bliss (1967) estimate of the stratified 

 mean catch tow"'; CV = coefficient of variation for untrans- 

 formed data; CV^ = coefficient of variation for ln(x +1) trans- 

 formed data. 



Cruise 



n 



'Bliss 



CV cv„ 



Fall 1973 



(23 Oct. -16 Nov.) 

 Spring 1974 



(1 Apr. -9 May) 

 Summer 1974 



(13 Aug.-19 Sept.; 

 Winter 1975 



(16 Jan.-2 Feb.) 

 Spring 1975 



(31 Mar.-IO Apr.) 

 Summer 1975 



(30 Aug.-19 Sept.; 

 Winter 1976 



(12 Jan.-7 Feb.) 

 Summer 1976 



(28 Aug.-21 Sept.; 

 Winter 1977 



(17 Jan.-9 Mar.) 



67 2.83 0.59 1.86 31.6 1.68 

 89 1.21 0.25 0.58 35.8 1.70 

 69 3.13 0.59 1.97 25.5 1.69 

 52 0.24 0.11 0.79 0.8 2.02 

 40 0.08 0.06 0.07 0.4 1.31 



68 1.17 0.37 0.89 7.9 1.47 



69 1.73 0.17 0.52 62.4 2.81 

 69 1.51 0.34 0.87 25.7 1.70 

 72 3.30 0.14 1.18 197.0 9.01 



abundance and distribution (Wenner et al. 1979a). 

 All trawling was conducted from the RV Dolphin, 

 a 32.6 m converted tug, for approximately 30 

 minutes at 6.5 km/hour. Weight and fork length 

 (FL) were measured for each individual, except 

 for large catches which were subsampled. Surface 

 and bottom temperatures were taken after each 

 tow. 



Length-frequency distributions were compared by 

 season and by depth zone. An index of relative abun- 

 dance (IRA = lln X In (x -i- 1); n = # trawls in each 

 depth zone, x = weight of fish for each tow) was 

 calculated from the catches for each depth zone 

 (Musick and McEachran 1972). The stratified mean 

 catch per tow and the estimated variance of the 

 stratified mean catch per tow (Cochran 1977) were 

 calculated from untransformed and In (x + 1) trans- 

 formed data to reduce the effects of contagion 

 (Elliott 1971). The coefficient of variation was used 

 to compare variation in catches (Clark and Brown 

 1977). The Bliss (1967) approximation retrans- 

 formed the data from logarithmic to arithmetic 

 units. The Wilcoxon rank sum statistic (Hollander 

 and Wolfe 1973) was used to compare catches, 

 depths, and temperatures of sponge-coral (Wenner 

 1983) and sandy open-shelf (Struhsaker 1969) habi- 

 tats. Habitat designations of Wenner et al. 

 (1979a) were used. Catches collected north and south 

 of lat. 31°30'N were compared for winter and sum- 

 mer cruises to determine if seasonal migration 

 occurred. 



Reproductive Biology 



Specimens used for reproductive analyses were 

 collected in 1980 by several research and commer- 

 cial vessels, frozen, and examined in the laboratory. 

 Specimens were measured (nearest mm) and 

 weighed (nearest 0.1 g). Ripe ovaries were fixed in 

 Gilson's solution (Bagenal and Braum 1978) for 

 fecundity determination. All other gonads were 

 fixed in formol-alcohol, stained with a modified 

 Harris hematoxylin and counterstained with eosin 

 (Humason 1972). Maturity stages for testes follow 

 Hyder (1969); maturity stages for ovaries were 

 based on Wallace and Selman (1981) and the fre- 

 quency distributions of oocyte diameters. Frequen- 

 cy distributions of oocyte sizes were determined 

 from randomly selected ovaries in each maturity 

 stage. Analysis of variance revealed no differences 

 in means and variances of oocyte diameters taken 

 from different regions of ovaries in any stage. 

 Therefore, oocyte distributions for each ovary were 

 determined from two or three randomly selected 

 sections. 



Proportions of fish in each maturity stage were 

 determined bimonthly, and gonadosomatic indices 

 [GSI = gonad wt/(body wt - gonad wt)] were deter- 

 mined for each maturity stage (except stage 1 when 

 gonadal tissues weighed <0.1 g). 



Ova numbers were estimated by modifying the 

 methods of Macer (1974). Both ripe ova and devel- 

 oping oocytes with diameters >0.115 mm were 

 counted. Developing oocytes were included in fecun- 

 dity estimation because they exhibited characteris- 

 tics of secondary growth phase (Wallace and Selman 

 1981) and were atretic in spent ovaries. 



Feeding Habits 



Stomach contents of 457 fish collected in 1980 

 were fixed in 20% formalin, and stored in 50% iso- 

 propyl alcohol. Frequency of occurrence (%F0) and 

 percentage composition by number (%A'^) were com- 

 puted for major prey categories. Volumetric 

 displacement (%VOL) of prey categories from a 

 representative subsample of 30 stomachs were 

 determined by using a 0.1 cm- grid (Windell 1971). 

 Seasonal and ontogenetic change in diets were com- 

 pared with an index of relative importance [IRI = 

 (%A^ + %VOL)(%FO)], computed from the sums of 

 each prey category (Pinkas et al. 1971). 



Feeding periodicity was determined by plotting 

 the percentage of empty stomachs collected per time 

 period, using 377 stomachs with known collection 

 times. Distributions of fish lengths collected in dif- 



252 



