FISHERY BULLETIN; VOL. 74, NO. 1 



The foregoing results assume that the physical 

 efficiency of harvesting is not increased by the 

 power drum. The study by Hester et al. (1972) 

 revealed that the use of a power drum and fish 

 pumps to unload the nets often enabled the ex- 

 perimental vessel to get in an extra set during the 

 brief time fish were available before dawn. This 

 circumstance depended on the size of catches 

 being made since use of the equipment actually 

 increased the set time for very small catches. No 

 data were presented, however, as to the average 

 number of sets or the frequency of catch size for 

 evaluation of efficiencies. 



The above analysis points up the importance 

 of union work rules permitting the use of new 

 technology. The application of the power drum to 

 vessels apparently would improve the economic 

 viability of the fishery, permitting its operation 

 even with old hulls or at fish meal prices below 

 $250/MT. Although use of the drum reduces crew 

 size on an individual vessel, its general adoption 

 apparently would provide considerable economic 

 incentive for fleet expansion, leading to an in- 

 crease in overall employment beyond its 1968 

 level. 



To make this inference, however, we have as- 

 sumed away the real problem, which is not the 

 adoption of new technology but the alteration of 

 traditional union share agreements which pay 

 the crew a percentage of net revenues. Unless 

 new technology resulted in increased gross reve- 

 nue as well as a reduction in crew size, the same 

 share of the net revenue would simply be divided 

 among fewer crewmen, and the investor would 

 gain nothing to compensate him for the addi- 

 tional costs of the technological change. Con- 

 sequently, the present system does not allow the 

 investor a sufficient return, and the fishery suf- 

 fers in terms of employment levels as well as with 

 respect to economic efficiency. 



DISCUSSION 



In discussing his model of the ecological bio- 

 energetics of isopods, Hubbell (1971) indicates 

 that there is a twofold utility in modelling a given 

 system. First, the model can be regarded as a tool 

 to guide and orient future research on that sys- 

 tem. Second, once the model exhibits satisfactory 

 performance, it can be put to predictive use, 

 answering hypothetical questions about the con- 

 sequences of different input conditions upon sys- 

 tem behavior. As demonstrated by the preceding 



applications, we feel that the northern anchovy 

 model definitely has the potential for fulfilling 

 both of these purposes. 



However, in its current state of development 

 the model is admittedly speculative in some of its 

 content. Several of its shortcomings have already 

 been discussed, but perhaps its greatest failing is 

 that its behavior has not yet been adequately val- 

 idated. To do so would currently require the circu- 

 lar logic of testing the model against the very 

 data from which its assumptions and estimates 

 derive. Consequently we have been forced to rely 

 upon our own subjective view of what constitutes 

 well-behavedness in the model and have applied 

 this criterion in evaluating its performance. 



According to Patten (1972), we probably could 

 do little more to validate the model since there 

 currently exists no theoretical base for approach- 

 ing this fundamental modelling problem. In any 

 regard, the predictive use of this model should 

 therefore be treated in only the most general of 

 terms, i.e., with the aim of gaining insight into 

 the structure and behavior of the anchovy fishery. 

 In this sense, it presently is a conceptual rather 

 than an analytical model. 



This leaves its use as a tool for guiding and 

 planning research as the model's primary reason 

 for being. To that end it has proven quite useful, 

 providing a systematic means by which extant 

 data might be organized and pinpointing areas 

 characterized by a glaring lack of data. For 

 example, our approach to modelling stock-recruit 

 behavior was necessitated by a lack of appropri- 

 ate indices measuring recent stock and recruit- 

 ment sizes. 



Additionally, we feel that the model provides 

 the capability for identifying and ranking critical 

 research areas. Management decisions must be 

 timely and as correct as possible, yet the cost of 

 collecting and analyzing relevant data is very 

 high both in money and time. Given budgetary 

 constraints, all research needs cannot possibly be 

 satisfied. Therefore, decision makers should be 

 asking themselves whether the cost of better in- 

 formation will be justified by a better choice of 

 management policy. 



The model could play an important role here by 

 allowing the decision maker to test the sensitiv- 

 ity of his information upon policy alternatives. 

 Some policy sets will not be affected by slight 

 changes in estimates resulting from fuller infor- 

 mation: a somewhat higher growth rate than ini- 

 tially believed, for example, may not occasion any 



128 



