290 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



Table A-4. — Falls Cove lest area 

 [Samples collected on dates marked by asterisks were used in statistical analysis and to obtain mean growth of 3.28 mm. based on 160 clams] 



May 2 



June 9- 



July 6 



August 7 



September 13 - 

 October 15- -.. 

 November 16* . 

 December 11*. . 



March 19. 



Date sampled 



wee 



Clams transplanted from — 



Western Beach 



Number 

 recovered 



87 



A verage 

 Rrowth 



Mm. 



0.0 

 1.4 

 1.4 

 3.1 

 2.0 

 2.3 

 2.5 

 2.1 



2.6 



Meetinghouse Cove 



Number 

 recovered 



62 



Average 

 growth 



Mm. 



0.1 

 1.9 

 3.1 

 2.7 

 3.2 

 3.1 

 3.7 

 3.9 



Falls Cove 



Number 



recovered 



Average 

 growth 



0.3 



1.2 



.8 



3.8 



1.0 



2.8 



Table A-5. — PUim Island Sotmd test area 



(Samples collected on date marked by asterisk were all recovered dead but were used in statistical analysis and to obtain mean growth of 19.3.'j ram. based on 



321 clamsj 



APPENDIX B.— STATISTICAL ANALYSIS 



ANALYSIS OF VARIANCE BETWEEN AND WITHIN 

 TEST AREAS 



Standard deviations for the 18 groups of clams 

 plotted against their means follow a straight line 

 having the formula E= 1 .15 + 0.284A' (fig. B-1 

 and table B-1). The slope of this line indicates 

 the need for transformation to make variances 

 independent of the means in order that methods 

 for analysis of variance shall become applicable. 

 The fact that standard deviations plotted against 

 means follow a straight line indicates that the log 

 transformation is the one to be used (Quenouille 

 1950). 



Figure B-2 shows tiie variance plotted against 

 the mean for each of the 18 groups of clams after 

 the values had been transformed by taking the 

 log of the midpoint of each '2-mm. class plus 1 

 (table B-1). The very sliglit slope of the least- 

 squares line, as indicated by the formula E— 

 0.077 — 0.0304A', indicates that the variances 

 have been made virtually independent of the 



means. Analysis of variance can therefore be 

 completed, using the transformed values. 



Table B-2 shows the completed analysis of 

 variance of differences in mean growth between 

 and within test areas using transformed values. 

 The F value for a comparison of between and 

 within test areas was 43.9, which is highly signifi- 

 cant. This indicates that differences between 

 growth rates in the various test areas are liighly 

 significant. 



A comparison of the differences within test 

 areas and between individuals yielded an F value 

 of 13.0, which is also highly significant. This 

 indicates that there is a considerable amount of 

 variation among the groups of clams that were 

 used in the various test areas. It appears likely 

 from examination of the untransformed mean 

 growths in table 1 (in text) and from the growth 

 curves in Hgiires 4 and 5 (in text) that Meeting- 

 house Cove clams are largely responsible for the 

 high F value in this test. 



