FISHERY BULLETIN: VOL. 70, NO. 3 



length with a decrease in latitude. But the im- 

 portant features of interest are the unimodality 

 of most of the distributions, the decrease in var- 

 iability within schools as the fish get smaller, 

 and the comparatively narrow range of fish 

 length within schools in contrast to the wide 

 range of fish length within each area as a whole 

 (cf. Nicholson and Higham, 1964). In the Mid- 

 dle Atlantic Area, for example, the length range 

 within single-school catch samples varied from 

 3.0 to 10.1 cm, whereas the length range for the 

 combined single-school catches sampled in that 

 area was 21.5 cm (13.0 to 34.5 cm) . So the seg- 

 ment of the population inhabiting any given area 

 of the coast during the summer is evidently strat- 

 ified, with each stratum comprised of discrete 

 schools of fish grouped according to length. 



Fall school-aggregations oflf Long Island con- 

 sisted of large fish that were only represented 

 in the North Atlantic Area during the summer. 

 Length-frequency distributions of fish in sam- 

 ples from 23 single-set catches from these school- 

 aggregations tended to be unimodal and con- 

 tained relatively few fish under 25 cm (Figure 

 7) . The range in length of fish within individu- 

 al catch samples varied from 4.0 (28.0 to 32.0 



Figure 7. — Length-frequency distributions of Atlantic 

 menhaden in samples from 23 single-set purse-seine 

 catches from fall school-aggregations off Long Island, 

 1955-62. 



cm) to 11.0 cm (23.0 to 34.0 cm). The mean 

 length range within samples (7.3 cm, with a 

 standard deviation of 1.81 cm) was slightly 

 greater than that within samples from single- 

 school summer catches in the North Atlantic 

 Area (6.6 cm, with a standard deviation of 2.02 

 cm), but the diflference between the means has 

 no statistical significance. 



The fall migratory school-aggregations off 

 North Carolina consisted of length-groups rep- 

 resented farther northward during the summer, 

 and individual aggregations contained fish of si- 

 milar lengths. Fish lengths represented in sam- 

 ples from 41 single-set catches from these school- 

 aggregations ranged from 8.5 to 35.0 cm 

 (Figure 8) . Because of selective fishing by the 

 fleet on school-aggregations comprised of larger 

 fish, the catch samples are accordingly weighted. 

 The smallest length groups include young-of-the- 

 year fish which had emigrated from estuarine 

 nurseries at the end of the summer (June and 

 Nicholson, 1964; Kroger, Dryfoos, and Hunts- 

 man, 1971). The most striking feature of the 

 length-frequency curves is the relative homo- 

 geneity of fish lengths within individual school- 

 aggregations, i.e., small fish are not represented 

 in catch samples containing large fish, or vice 

 versa. The least range in length within samples 

 was 4.0 cm (9.0 to 13.0 cm) and the greatest 

 12.0 cm (22.0 to 34.0 and 22.5 to 34.5 cm) . The 

 mean of the range in length within samples was 

 8.4 cm, with a standard deviation of 6.13 cm. 

 This was the largest mean difference and the 

 greatest variation about the mean for any group 

 of samples examined in this study. The length- 

 frequency distributions of fish over about 25 cm 

 were similar to those of single-school summer 

 catches in the North Atlantic Area and to single- 

 set catches from fall school-aggregations off 

 Long Island (cf. Figure 7). 



From an analysis of length and age data for 

 1955 to 1958, June and Nicholson (1964) postu- 

 lated that these large school-aggregations that 

 migrated southward along the North Carolina 

 coast in November and December comprised por- 

 tions, if not all, of the stocks which had spent 

 the summer north of Cape Hatteras, N.C. More 

 detailed analysis of the length-age data by Nich- 

 olson (1971a) and results of tagging studies 



706 



