360 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



indices are termed "yields per day's trawling" to 

 distinguish them from their offshore counterparts, 

 "fishable biomass indices." 



SIZE-AGE STRUCTURE: FISHABLE BIOMASS 



Commercial landings classified according to the 

 sizes of shrimp comprising them afford some 

 insight into fishable biomass age structure only if 

 the landings represent the defined biomass with 

 reasonable accuracy. Any effects of differential 

 bias due to (1) fisherman or gear selectivity; 

 (2) nonuniform distribution of shrimping effort 

 with respect to stratification by age within the 

 fished population; (3) minimum-size restrictions; 

 and (4) varying grading practices must be assumed 

 negligible, or at least constant in time. 



Totals for the seven or eight size categories into 

 which commercial shrimp landings are separated 

 give weight frequencies whose modes, it is believed, 

 crudely delineate the age classes, or "broods", 

 making up the exploited biomass. The term 

 "brood" is used to define groups of shrimp, each 

 member of which is produced (i.e., spawned and 

 hatched) within a designated interval of time. 

 These intervals are specified as covering periods of 

 heightened spawning activity and e.xtend roughly 

 f month on either side of points in time at which 

 modal spawning occurs. Note that modal spawn- 

 ing does not necessarily recur at precisely the 

 same point in successive (corresponding) seasons. 



Monthly' weight frequency distributions for each 

 coastal area were obtained by sunmiing, within 

 each size category, the landings from each subarea 

 and depth zone. Plotted serially and fitted with 

 smooth curves, the size-distribution modes traced 

 each brood from its recruitment to its disappear- 

 ance from the fishery (fig. 11). The curve for 

 each brood is the typical sigmoid curve describing 

 population growth in weight. Its disposition with 

 respect to the ordinate is irrelevant, the midpoints 

 of each size class being arranged arbitrarily theron. 

 In the present report, the procedure of plotting 

 size-distribution modes is carried out only for those 

 population segments occupying offshore areas, 

 portions found on inshore grounds being almost 

 alwaj's composed of a single, newly produced 

 brood. 



Of the biases associated with the commercial 

 catch-by-size data used herein, that due to dif- 

 ferent grading practices is potentially the most 

 serious. Two such methods are conunonly em- 



PERCENTAOE COMPOSITION 



Figure 11. — Hypothetical e.xample showing method of 

 delineating shrimp "broods" from weight composition of 

 monthly commercial catches. (Light lines displaced 

 vertically are monthly weight-frequency curves. Squares 

 denote dominant weight-frequency modes. Shaded 

 circles indicate lesser modes. Dark lines displaced hori- 

 zontally trace progress of individual broods in fishery.) 



ployed around the Gulf, viz., "box-grading" and 

 "machine-grading." The former entails taking 

 a representative 5-pound sample of the landing, 

 separating its contents into standard size cate- 

 gories, calculating the proportion in each category, 

 and prorating the landing accordingly. In the 

 latter method, the entire landing is run through a 

 mechanical sorting device. 



Grading machines are found at relatively few 

 Gulf ports. But even if available, they may not 

 be used, each fisherman reserving the option of 

 selling his catch on a bo.x-grade basis if he so 

 desires. The problem here is that data of box- 

 graded and machine-graded landings are not com- 

 parable for areas in which both methods are used. 

 From the standpoint of getting a true picture 

 of catch size composition, machine-grading is 

 obviously far superior to box-grading. Where 

 machine-graded landings would be expected to 

 yield monthly weight frequency ciu'ves truly 

 representative of those of the fishable biomass, 

 or at least of the landings themselves, weight 

 frequencies based on box-graded landings would 

 not, the probability of their exhibiting all modes 

 being quite low. Consolidating the two types 

 of data confounds the net bias and necessitates 

 interpreting weight frequency curves resulting 

 therefrom with some reservation. 



INTEGRATION OF DATA 



Classical approaches to predicting the yields of, 

 and assessing the effects of artificial and environ- 



