tips of the middle caudal fin rays. This measure- 

 ment involves fin rays that are fragile and fre- 

 quently damaged on small specimens. The base 

 length that I selected for my comparisons is the 

 trunk length, defined by de Sylva (1957) as the 

 distance between the posterior margin of the orbit 

 and the anterior insertion of the caudal keels. This 

 length involves neither the snout nor the caudal 

 rays.* 



* See definition for trunk lengtli in section on methods and data 

 for determination of posterior point for this measurement of 

 sailfish without caudal Iteels. 



Initially, individual measurements of selected 

 body parts of the western Atlantic sailfish were 

 plotted against trunk length on arithmetic paper 

 (figs. 6-l;2). The distribution of tlie data suggested 

 that the relatively fewer specimens with trunk 

 length greater tJian (iO mm. should l>e treated sepa- 

 rately. Two regression lines were calculated — one 

 for those with trunk length less than 62 mm. and 

 one for longer specimens. The regression equation 

 is Y=a + bX (table 5). Inspection of the graphs 

 with tlie calculated regression lines added suggests 

 tliat tlie location of point of inflection varies some- 



100 - 



80 



t. 



60 - 



< 



40 - 



20 



20 



40 60 80 



TRUNK LENGTH (MM.) 



100 



120 



Figure 6. — Relation of head length to trunk length in sailfish. Small dots represent western Atlantic sijecimens ; open 

 circles, eastern Atlantic specimens. Regression lines represent western Atlantic data only. 



184 



U.S. FISH ANB WILDLIFE SERVICE 



