Tatle 22. — Analysis of covariance test on the data (duration of attack 



and thickness of prey's shell) for bay mussel, Olympia 



oyster and Pacific oyster i/ 



1.8407 



therefore, one common line can 

 represent the data) 



]_/ The technique employed here was to test first for a common 

 line for the data. If such a common line is found, it is 

 then unnecessary to test further for common slopes or common 

 means as a common line implies that any differences in slopes 

 or means are insignificant. 



13 

 12-1 

 II 

 10 

 9 



ten 



•v. 



^ 5 



c 



.o 



^ 3 

 2H 

 I- 

 



95 % Confidence Intcrvol for 

 Regression Line 



95% Confidence Intervol for 



Individual Points 



1 



.0200 



— I 



.0300 



-r 



"T" 



.0100 



.0400 .0500 .0600 



Shell thickness in inches 



.0700 .0800 .0900 



.1000 



.1100 



Figure 6.- -Regression Une anci scatter diagram of duration of attack upon thickness of all shells from Experiment I. 



23 



