707 



AbStr3Ct.— We analyzed 364 spawner- 

 recruitment time series to determine 

 whether recruitment is related to 

 spawner abundance. We pose three 

 questions: 1 ) Does the highest recruit- 

 ment occur when spawner abundance 

 is high? 2) Does the lowest recruitment 

 occur when spawner abundance is low? 

 and 3) Is the mean recruitment higher 

 if spawner abundance is above rather 

 than below the median? We found that 

 when there is a sufficient range in 

 spawner abundance the answer to all 

 three questions is almost always "yes." 

 Thus, spawner abundance cannot be ig- 

 nored in the management offish popula- 

 tions. Recruitment overfishing appears 

 to be a common problem. 



Is fish recruitment related to 

 spawner abundance? 



Ransom A. Myers 

 Nicholas J. Barrowman 



Science Branch, Northwest Atlantic Fisheries Centre 

 Department of Fisheries and Oceans 

 Box 5667. St John's, Newfoundland, Canada A! C 5X1 

 e-mail address: myers@mrspocknwafc.nf.ca 



Manuscript accepted 6 June 1996. 

 Fishery Bulletin 94:707-724 ( 1996) 



Perhaps the most fundamental is- 

 sue for the study and management 

 offish populations is the relation be- 

 tween spawner abundance and sub- 

 sequent recruitment. There is sur- 

 prisingly little consensus on this 

 issue; many researchers believe 

 that there is no relevant relation- 

 ship between species abundance and 

 recruitment (reviewed by Wooster 

 and Bailey, 1989; Koslow et al., 1987) 

 whereas others believe it to be fun- 

 damental (e.g. Ricker, 1954; Bever- 

 ton and Holt, 1957; Cushing, 1971; 

 Myers et al., 1995a). The assumed 

 absence of a relationship between 

 spawner abundance and recruit- 

 ment has prompted some scientists 

 to claim that recruitment overfish- 

 ing is almost impossible (Laevastu, 

 1993). This divergence of opinion 

 has practical consequences for the 

 management of fisheries: many 

 fisheries are managed without con- 

 sideration of maintaining a sustain- 

 able abundance or biomass of 

 spawners (Smith et al., 1993). 



The purpose of this paper is to 

 provide conclusive evidence that 

 strong year classes are more likely 

 when spawner abundance is large. 

 We approach the problem using the 

 simplest possible nonparametric 

 methods in order to avoid the many 

 subtle statistical difficulties in fitting 

 spawner-recruitment functions (Wal- 

 ters, 1985, 1990; Hilborn and Walters, 

 1992). Our approach is to examine 

 systematically 364 data sets from the 

 most recent version of the database 



compiled by Myers et al. (1995b) as 

 part of an ongoing study of recruit- 

 ment variability. By analyzing many 

 populations with identical methods, 

 it is possible to arrive at conclusions 

 with greater reliability. 



The nonparametric methods we 

 used were devised in order to an- 

 swer three deliberately simple ques- 

 tions. First: Does the largest re- 

 cruitment occur when the spawner 

 abundance is high? To answer this 

 question, we examined the rank of 

 spawner abundance associated with 

 the largest recruitment. Second: 

 Does the smallest recruitment oc- 

 cur when spawner abundance is low? 

 This time we examined the rank of 

 spawner abundance associated with 

 the smallest recruitment. Third: Is 

 the mean recruitment higher if 

 spawner abundance is above rather 

 than below the median? To answer 

 this question, we examined the ratio 

 of mean recruitment when spawner 

 abundance is above the median to 

 mean recruitment when spawner 

 abundance is below the median. 



Data 



By "spawner abundance" we mean 

 any of the following metrics of the 

 size of the spawning stock: spawn- 

 ing stock biomass, the number of 

 spawners, the number of eggs, or 

 some index of spawner abundance 

 (derived from catch per unit of ef- 

 fort or research vessels). 



