292 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



Table B-2. — Completed analysis of variance of differences 

 in mean growth between and within test areas Itased on 

 18 groups of clams at 5 areas 



ll'sinB values transformed by the formula: Transformed .Y = log (midpoint 

 of 2-mm. class + 1)] 



The analysis of variance by test areas was 

 recomputed without the groups of clams that 

 came from Meetinghouse Cove; the results are 

 listed in table B-3. While the differences in 

 mean growths of groups within tlie test areas 

 are still highly significant, the F value has been 

 reduced from 13.0 to 6.0 by exclusion of Meeting- 

 house Cove clams. Therefore, it appears that 

 clams from this origin were responsible for more 

 than half of the F value of 13 listed in table B-2. 

 Also, it appears likely that the clams from Saga- 

 dahoc Bay contributed a large part of the high 

 F value for this test (figs. 4 and 5). Since the 

 growth pattern of both the Meetinghouse Cove 



Table B-3. — Completed analysis of variance of differences 

 in mean growth between and within te^l areas ttased on 

 IS groups of clams (excluding those from Meetinghouse 

 Cove) at 5 areas 



(Using values transformed by the formula: Transformed X= log (midpoint 

 of 2-mm. class + 1)] 



clams and the Sagadahoc Bay clams was the 

 opposite of that which might be expected had 

 heredity been the cause of growth differences and 

 growth rate, the significance of this F value does 

 not alter the conclusions given here. 



ANALYSIS OF VARIANCE BETWEEN AND WITHIN 

 ORIGINS 



The possibility that variation between the se- 

 ries means was caused by the origin of the test 

 clams needed to be explored. Mean growths of 

 clams from the four origins planted at Bedroom 

 Cove, Sagadahoc Bay, and Robinhood Cove 

 were used in this analysis, because only at these 

 three test areas were all four groups planted. 

 Results of the analysis-of-variance tests are 

 shown in table B-4. The F value of 8.42 for a 

 comparison between origins and within origins 

 was not significant at the 5-percent level. There- 

 fore, the effect of the origin of the clams on their 

 growth rate after transplanting was not significant. 



The F value of 166.8 for a comparison within 

 origins and between individuals was highly sig- 

 nificant, indicating (as would be expected) that 

 the differences in mean growth of clams from 

 each origin planted in the three test areas were 

 highly significant. 



T.'^BLE B-4. — Completed analysis of variance of differences 

 in mean growth between and within origins of clams based 

 on 12 groups of clams from 4 origins 



[Using values transformed by the formula: Transformed X = log (midpoint 

 of 2-mm. class + D) 



U.S. GOVERNMEt.T PRINTING OFFICE 1957 O — 409441 



