both experiments. Marked shrimp were recovered 

 over extended periods of time, but recovery and 

 fishing-effort information collected only within 

 the first 10 weeks following release of the marked 

 animals was used for estimating mortahty rates for 

 the Sanibel and Tortugas experiments. The period 

 was so restricted because recoveries of marked 

 shrimp reflect their relative abundance in the 

 commercial catch only as long as the interest in 

 recovering them remains constant. The reward is 

 a prime inducement for the return of marked 

 shrimp. Reasonably, when the number of marked 

 shrimp in catches drops markedly, interest wanes 

 and an increasing percentage of recaptures may 

 pass through the fishery unnoticed. 



THEORETICAL AND BIOLOGICAL 

 CONSIDERATIONS 



Fundamental tenets in computation of mor- 

 tahty estimates are that the dechne in numbers in 

 an animal population follows an exponential trend, 

 and that a constant instantaneous mortahty 

 coefficient is operative. The latter concept has been 

 applied to many animal |)opulations. Presumably, 

 it may be apphed to shrimp. PauUk (1963) noted 

 the acceptabihty of this concept in short-term 

 experiments, as are being considered here. 



Mortality estimates are derived from mark- 

 recovery experiments by measuring density 

 changes in an experimental population. In this 

 apphcation, otter trawls used by the commercial 

 shrimp fleet serve as sampUng gear, and decreases 

 in density are reflected in decreased catches of 

 marked shrimp per unit fishing intensity. The 

 average rate of loss is computed and expressed 

 numericaUy as the instantaneous total mortality 

 coefficient. 



At this point, it is pertinent to offer possible 

 explanations for the fluctuations in recoveries of 

 marked shrimp per unit fishing intensity on the 

 Sanibel grounds (table 4), but not on the Tortugas 

 grounds (table 5). It appears, superficially, that 

 decreases in the abundance of the marked shrimp 

 from one time period to the next should be con- 

 sistently reflected in decreased recaptures per unit 

 fishing intensity. Increased recaptures of marked 

 shrimp at a later time period appear to violate a 

 basic premise upon which the estimate of mortahty 

 is based. 



The reasons for this apparent anomaly are not 

 entirely clear. Possible sources of bias in the inter- 

 pretation of mark-recovery data may be intro- 

 duced by: (1) nonuniform distribution of marked 

 animals over the bottom and (2) inaccurate esti- 

 mation of fishing effort in relation to the spatial 

 distribution of marked animals. Such difficulties in 

 interpretation of marking experiments were dis- 

 cussed by Ricker (1958). 



These sources of bias, however, may not be the 

 only causes of fluctuations in recovery of shrimj) 

 per unit fishing intensity. Allen, Delacy, and 

 Gotshall (1960), Konstantinov (1964), Parrish, 

 Blaxter, and Hall (1964), and others recognized 

 that biological activities may affect the catch- 

 abihty of aquatic animals. Wathne (1963) and 

 Fuss and Ogren (1966) noted variable burrowing 

 habits of pink shrimp which may explain varia- 

 tions in shrimp availabihty to the trawl. Addi- 

 tionally, although pink shrimp are ordinarily 

 benthic, we have frequently observed them in 

 dense schools at or near the surface at night. 

 Similar observations were reported by Burkenroad 

 (1949), Higman (1952), Tabb, Dubrow, and 

 Jones (1962), Iversen and Van Meter (1964), and 

 Joyce (1965). Obviously, vertical movements of 



Table 4. — Fishing intensity and numbers of marked shrimp recovered, south Sanibel grounds, 



March 1962 to May 1962 



|No=2,49«] 



' Ter (mean) 45. 6 boat-nigbts fishing intensity. 



' Recoveries during this final period were not used in analysis. 



498 



U.S. FISH AND WILDLIFE SERVICE 



