years of high river flow. Turner and 

 Chadwick (1972) demonstrated that in 

 the Sacramento-San Joaquin System 

 survival of young striped bass (up to 

 3.8 cm TLJ was related to summer 

 river flow through the delta, which 

 controls the transport of young bass 

 to suitable nursery areas. Stevens 

 (1977) and Chadwick et al. (1977) 

 have shown that these river outflows 

 and diversion of river water to the 

 California aqueduct system impact 

 recruitment to the sport fishery 

 several years later and play a major 

 role in controlling the size of the 

 striped bass population. Although 

 years of higher river flow in the 

 California Delta have resulted in 

 large year-classes, virtually all the 

 eggs produced in the early and mid- 

 portions of the spawning season in 

 these high flow years are swept into 

 the lower bays of the delta where 

 survival is extremely low. The mid- 

 summer size distribution of the 

 juvenile fish indicates that they 

 were produced from a small fraction 

 of late spawning fish (Chadwick 

 19 74). Likewise in the Potomac- 

 Estuary, striped bass eggs and larvae 

 apparently experience a differential 

 mortality with a greater probability 

 of survival toward the end of the 

 spawning season at the up-river 

 transects (Pol gar et al. 1976, 

 Ulanowicz and Polgar 1980, Setzler- 

 Hamilton et al. 1981. Such results 

 would seem to indicate that the 

 production of a successful year-class 

 is largely a dens lty- independent 

 phenomenon, a conclusion first 

 alluded to by Vladykov and Wallace 

 (1952). 



available and several functions of 

 these were used singly and in combi- 

 nation as predictors of recruitment 

 success. Summer surveys of juvenile 

 striped bass relative abundance have 

 been made in the Potomac since 1958. 

 Recently, this data set has been 

 shown to be a good indicator of both 

 recruitment success (Polgar 1977, 

 Ulanowicz and Polgar 1980) and com- 

 mercial catch (Boynton et al. 1977) 

 and was used here as the dependent 

 variable in regression analyses. 

 Results of a single factor and 

 multiple factor analyses are sum- 

 marized in Table 2. In general, 

 statistically significant relation- 

 ships were indicated for several 

 functions of river flow and air 

 temperature although the percent of 

 the variability explained by the 

 regressions was quite low (about 

 25%) . The percentage of the vari- 

 ability explained using multiple 

 linear regressions was considerably 

 better (about 70%) and in all cases 

 the 5-day maximum flow in April was 

 the strongest predictor. A three- 

 dimensional plot of this regression 

 is shown in Figure 4. Note that all 

 dominant year-classes are clustered 

 in the quadrant bounded by colder 

 than normal winters and greater than 

 normal spring river flows. Addi- 

 tional plots were made using the same 

 temperature function but the highest 

 five-day mean flow occurring in 

 either March or May. Interestingly 

 enough, the previous pattern was not 

 observed suggesting that the timing 

 as well as the quantity of river flow 

 is an important factor in determining 

 recruitment success. 



We reviewed several climatic 

 data sets in an attempt to identify 

 extrinsic factors which may play a 

 strong role in regulating recruitment 

 success. Such analyses are obviously 

 constrained by the types of data 

 available; in our case air tempera- 

 ture and river flow were readily 



As in all statistical models, 

 significant results or interesting 

 patterns are, per se, incomplete; 

 causation is certainly not demon- 

 strated and for the model to be help- 

 ful we need to be able to suggest 

 mechanisms responsible for the 

 statistical results. In our case, we 



157 



