BARKLEY: SELECTIVITY OF TOWED-NET SAMPLERS 



1.000 



0.100 



0.010 



.2 



100 



"e/ CATCH/100 ml 



U WET PLANKTON 



-0.1 



0.00 



-100 



- 10 



-0.1 



Figure 5. — Day and night catches of larval anchovy by a 1-m net (left) and 

 an Isaacs-Kidd midwater trawl (right). Length-frequency data have been 

 converted to relative escape speed (abscissa) versus frequency (ordinate, 

 right hand scales), and compared with theoretical curves. Catch values 

 at P(. ^ 1.0 provide estimates of abundance for all class intervals sampled. 

 Dashed line for P^ at oCf^/R ^ 2 on Isaacs-Kidd graph shows catches which 

 could have been expected if the anchovy's reaction distance had been equal 

 to that for the 1-m net. Net retention, based on differences between theoret- 

 ical and actual catch, is shown for two class intervals on right panel: 7% 

 to 8% for 3.75 mm fish and 55% to 60% for 5.75 mm fish. (Class sizes ranged 

 from 1.75 mm to 12.75 and 13.75 mm by 1 mm increments.) Data from 

 Murphy and Clutter (1972). 



oretical lines of constant Xi)/R, for anchovy short- 

 er than 7.75 mm with the IKMT and 4.75 or 

 5.75 mm for the 1-m net. If x»/R for a given 

 species remains constant for any one net towed 

 at a single speed, extrapolation of the theoretical 

 curve toward the origin gives an indication of 

 expected catch for smaller size classes (assum- 

 ing, of course, that changes in population density 



with length are negligible) . This use of avoid- 

 ance theory is illustrated in Figure 5, right pan- 

 el, where retention is calculated at 7% to 8% 



for 3.75-mm fish and 55 



/( 



to 60% 



for 5.75-mm 



fish. These values can be easily measured on 

 the semilog graph ; the vertical distance between 

 the "theoretical" and the observed values is 

 transferred to the vertical coordinate scale (e.g.. 



809 



