usual. In young specimens taken in the James 

 River in 1954 there is no significant difference 

 in the mean values for standard length of the 

 sample which had 11 or fewer rays, and that 

 which had 12 or more rays (table I) . 



When the frequency distributions of dor- 

 sal soft rays of the several samples taken in 

 the James River in 1955 (table 2) were tested, 

 homogeneity is indicated. The composite up- 

 stream sample for 1955 averaged slightly lower 

 than the downstream composite sample (table 3) 

 but when tested the results indicate homogeneity. 



The distribution of soft dorsal fin rays in 

 samples of young taken in the James River in 

 1954 are given in table 14. A test indicates 

 that all samples could have been drawn from 

 the same population . However, the downstream 

 samples are somewhat higher in mean value; 

 when up- and downstream composite samples 

 (table 4) are tested there is a significant differ- 

 ence at the 5 percent level which indicates that 

 they probably were not drawn from the same 

 population. A comparison of an upstream and 

 a downstream sample of young of the 1954 year 

 class from the Chickahominy River, a large 

 tributary of the James River, is shown in table 

 5, but, although the downstream sample aver- 

 ages higher, a test shows no heterogeneity. 



The frequency distributions of dorsal soft 

 rays in samples of young taken in the Hudson 

 River in 1954 are given in table 15. A test in- 

 dicates that they very probably were not drawn 

 from the same population . An examination of 

 the mean values (table 15) shows that those 

 from downstream localities average higher. 

 When the composite upstream sample is com- 

 pared to that from downstream (table 7) a test 

 indicates difference at the one percent level. 

 Similar results were obtained from the 1953 

 year class when an upstream sample (Coxsackie) 

 was compared with one from downstream 

 (Haverstraw) (table 8) . The downstream sample 

 has a higher value which is significantly differ- 

 ent at the one percent level . 



In summary it is noted that the number of 

 soft rays in the second dorsal fin is slightly, 

 but not always significantly, higher in the down- 

 stream samples from the James River. The 

 Hudson River consistently has higher values for 



downstream samples . Here a much greater 

 distance is involved than in the James River. 

 The higher downstream values perhaps are cor- 

 related with differences in ecology since the 

 lower Hudson River contrasts with the upper 

 river in several respects as noted above. 



The frequency distributions of the number 

 of soft dorsal rays in samples taken from the 

 tributaries of Chesapeake Bay are given in 

 tables 16 and 17. When the eight samples ta<en 

 from tributaries of Chesapeake Bay in 1955 are 

 compared (table 16) the low value for the James 

 sample especially when compared with the adja- 

 cent York -Rappahannock samples is noteworthy. 

 When these two (James vs. York -Rappahannock) 

 are tested, X = 98-04 which is significant at 

 the one percent level and indicates that they very 

 probably were not drawn from the same popula - 

 tion. When the James sample is compared with 

 the composite Elk to Wicomico sample (table 16), 

 X = 17. 49 which is significant at the one per- 

 cent level and indicates probable heterogeneity. 

 The data for samples of the 1954 year class 

 (table 17) shows the same general trend for these 

 three samples . When the James and York- 

 Rappahannock samples (table 17) are tested 

 X = 63 .51 and when the James is tested against 

 the composite ChoptanK-Pocomoke sample 

 X = 24.45; both of these results are significant 

 at the one percent level and indicate heterogeneity. 



The composite 1954-55 samples compar- 

 ing East Shore tributaries, West Shore tributaries 

 (excluding the James River), and the James River 

 are given in table 11 . The value for the James 

 River sample is obviously different and a test 

 indicates that the three very probably were not 

 drawn from the same population . 



The frequency distributions representing 

 composite 1954-55 samples (table 12) are arranged 

 in four geographic groups from north to south . 

 A test indicates that they very probably were not 

 drawn from the same population . The Elk to 

 Chester and Choptank to Potomac -Pocomoke com- 

 posite samples have identical means and seem 

 to represent a population which differs from the 

 Rappahannock-York sample and the James River 

 sample. Although the Pocomoke sample is in- 

 cluded with the Choptanic to Potomac sample in 

 the geographic arrangement, the data (table 17) 

 suggests that it might belong with the Rappahan- 

 nock system . 



90 



