Table 23.--Chl-8quare analysis of total catch by kind of bait 

 and sea condition, Huah M. Smith cruise 18 (based 

 on data In table 33) 



It was found that the differences In the catch between hooks baited with sardine 

 and those batted with herring (tables 12 and 13) were small and non-slgnlflcant. 

 Considered In relation to geographical and temporal variations (Murphy and Shomura 

 1953a, b) , and in relation to random variability (Murphy and Elliott 1954) they are 

 too small to warrant smy adjustment of the catch rate data to compensate for them. 

 The most extreme difference appeared In an experiment comparing sardine and squid 

 (table 19), where of a total catch of 331 fish, 185 (56 percent) were taken with 

 sardine baits. Though statistically significant, this difference is still small 

 relative to the differences attributable to time, space, and random variability. 

 Also these are differences of total catch of all fish, corrected for broken gear 

 and surface sharks and not actual catches of individual species for which the 

 variability from sources other than differences in baits might be still greater. 

 Adjustment to compensate for the lower catching effectiveness of squld-balted 

 longlines would appear warranted only for special types of analysis of pooled data 

 where the bias may emerge above the other sources of variability. 



25 



