Table 21. 



-Summary of correlation and regression analyses between the duration of attack in days, 

 and the corresponding thickness of perforated shells ^ 



lY 



5x 



lY-^ 



1x2 



IXY 



Sy/x 



Manila clam (C) 59 



Olympla oyster (0) 48 



Pacific oyster (P) 17 



Bay mussel (M) 36 



429 3.2322 3,449 



315 2.0061 2,487 



91 .5372 551 



149 



.7927 



665 



.18539 23.9134 



.09959 15.1206 



.01959 3.1482 



.01838 3.3606 



2.3292 49.5783 



1.9653 123.7722 



1.5411 103.6502 



1.1045 85.6989 



.248 (? 



.755 9.95 * 



.793 3.45 * 



.376 2.37 * 



1/ N 



z 



Y 

 X 



^y/x 



13- 

 12- 



1 1- 



Is. 



i: 



5 4- 



3- 

 2- 

 I- 



number 



sunsnation 



duration of attack in days 



thickness of drilled shells In inches 



sample standard deviation of Y on X 



Common line -^-^^ 



Boy mussels « « 



Olympia oysters* • 



Pacific oysters o—- — ^ 

 (Manila clams excluded ) 



"T 1 



.0200 



.0400 



1 



.0600 



1 1 1 



.0800 .1000 



Shell thickness in inches 



Figure 5. — Regression lines and scatter diagrams of dura- 

 tion of attack upon thickness of shells from 

 Experiment I. 



The mussel, Olympia oyster, and 

 Pacific oyster data may be repre- 

 sented by one common line (fig. 5) 

 as shown by an analysis of covariance 

 test (table ZZ). 



To determine the duration of attack. 

 (Y) for a given thickness of shell (X), 

 all available data from Experiment I 



b; estimated slope 



"r": correlation coefficient 



"t": value of "t" test on the slope 



(3: not significant at the 5 percent level 



*: significant 



were combined and analyzed. The **r" 

 value for these data was O.696 (158 

 degrees of freedom), which indicates 

 that there was a positive correlation 

 between Y and X. Also a significant 

 **t*' value of 12.23 (testing hypothesis 

 P = 0) was obtained. A regression line 

 was therefore fitted to the plotted data 

 in figure 6 to show this relationship. 

 The Yestimate for the regression line 

 in figure 6 was Y = 6.1500 + 96.4120 

 (Xo - .0411). The 95 percent confidence 

 intervals for the regression line and 

 individual points were also plotted in 

 this figure. 



DISCUSSION 



From the results of this study, itwas 

 difficult to state specifically which 

 species of food animal was preferred 

 by Ocinebra. Throughout this study, 

 Ocinebra generally preferred Manila 

 clams, Olympia oysters, or bay mus- 

 sels to Pacific oysters. 



In Experiment I, the drills showed 

 a general preference for Manila clams, 

 followed closely by bay mussels and 

 Olympia oysters. In Experiment II the 

 Olympia oysters were attacked in pref- 

 erence to bay mussels and Manila 

 clams. Chew and Eisler (1958) indi- 

 cated that Ocinebra japonica preferred 

 bay mussels and Manila clams to 

 either Pacific or Olympia oysters. 



22 



