BARKLEY: SELECTIVITY OF TOWED-NET SAMPLERS 



until the net has almost caught up with them. 

 We do not expect to catch any animals which 

 can swim faster than the net; we wish to cal- 

 culate minimum probabilities for certain cap- 

 ture, which implies that animals which could 

 possibly escape do so. That is, the size of the 

 "lethal cone" is minimized by assuming that 

 animals are capable of optimum avoidance be- 

 havior. 



However, probabilities of certain capture can 

 only be calculated for animals which react indi- 

 vidually. Animals which school or form clus- 

 ters can "beat the odds" by reacting to each 

 other's behavior instead of reacting only to the 

 oncoming net. The effect of this is illustrated in 

 Figure Ic, which represents a net approaching 

 a group of animals whose reaction distances and 

 escape velocities are assumed to be similar. If 

 the animal shown nearest the net begins to re- 

 act at the distance labeled Xo, it can escape be- 

 cause it is outside of the "lethal cone." Those 

 animals already within that cone will be cap- 

 tured in any case. If the other animals respond 

 immediately to the actions of the one nearest the 

 net, they can all escape. If, instead, they re- 

 spond as individuals, as the net moves through 

 the group, more will enter the "lethal cone" and 

 then be captured. 



Whether an organism, once captured (i.e., en- 

 closed by the net) will be retained depends on 

 the characteristics of the net's meshes and the 

 size, shape, behavior, and fragility of the or- 

 ganism. Losses through the mesh have been 

 reviewed by Heron, 1968; Tranter and Smith, 

 1968; and Vanucci, 1968. Lenarz (1972) pre- 

 sents results of more recent work. Losses can 

 also occur due to faulty handling of the sampler, 

 particularly sudden decreases in towing speed. 

 Our primary interest here is in avoidance prior 

 to capture, although the theory to be developed 

 will throw some light on the problem of mesh 

 losses. 



The catch obtained from a towed-net sampler 

 can be calculated, in principle, from equations 

 such as the following: 



Catch = Captures — losses 



= (volume sampled) x ("Q- of organisms) 



'^ ^ unit volume 



So that 



X (probability of capture) — (losses) 



(2) 



no. of organisms 

 unit volume 



(catch + losses) 



(volume sampled) x (probability of capture) 



(3) 



The central problem of selectivity theory is 

 to evaluate the unknown factors in equations (2) 

 and (3), so as to permit solution of these equa- 

 tions with a minimum of empirical work. The 

 most important unknown factors are those gov- 

 erning probability of capture and losses after 

 capture— primarily losses through the meshes. 



Since both probability of capture and degree 

 of mesh loss must vary widely with species, age, 

 and condition, the above equations apply sep- 

 arately to each component of the plankton and 

 nekton community. 



For present purposes, a component must be 

 defined as a set of organisms having the same 

 probability of capture, and the same degree of 

 mesh retention, under given circumstances. 

 That is, they must be similar in their reaction 

 distance, escape speed, shape, size, and con- 

 dition. 



The above operational definition of a compo- 

 nent of the plankton or nekton may often coin- 

 cide with the more usual biological definitions. 

 For single species, one can reasonably expect 

 that fish of similar size or crustaceans at the 

 same stage of development should have similar 

 reaction distances and escape speeds and the 

 same percentage mesh retention. However, the 

 distinction between the two ways of defining a 

 component must be carefully borne in mind. The 

 operational definition may lump several biolo- 

 gically defined components; to take a trivial ex- 

 ample, bacteria, phytoplankton, and fish eggs 

 small enough to pass through the mesh all have 

 the same probability of capture, 1.0, and all have 

 0% retention. On the other hand, the opera- 

 tional definition may split one biological compo- 

 nent into two or more parts; e.g., healthy and 

 moribund animals of the same size and species 

 should difi^er in their ability to avoid capture. 

 In principle, empirical data on reaction distance, 



801 



