FISHERY BULLETIN: VOL. 74, NO. 3 



the mud surface layer was very soft. No small 

 juvenile totoaba were collected over firm mud 

 sediments or sandy substrates, as is often the case 

 with the larger individuals taken in shrimp trawls. 



Guevara (1974) is presently analyzing the dis- 

 tribution of juvenile totoaba captures in shrimp 

 trawls. Most of his specimens are larger than ours, 

 implying that the fish move into deeper water as 

 their growth continues. Tidal currents in the area 

 are extreme and these may also play a significant 

 role in juvenile distribution. 



Juveniles collected in 1972 were examined for 

 stomach contents. Remains of amphipods and 

 other small crustaceans common to the habitat 

 were recognizable, in addition to remains of 

 juvenile fishes which we identified as Micropogon 

 sp., Mugil cephalus, and Leuresthes sardina. 

 Within the limits imposed by size, the diet of 

 juvenile totoaba as small as about 6 cm standard 

 length is comparable in these items with the diet 

 of the large adults. 



Decline of the Fishery 



We have traced the growth and decline of the 

 totoaba fishery and discussed its present status 

 and methodology. We have presented data on 

 aspects of totoaba life history and raised questions 

 concerning possible reproductive behaviors which 

 may have a bearing on reproductive potential. 

 Although these have significance, if we consider 

 the resource from a management perspective one 

 fact becomes clear: the annual breeding migration 

 to the mouth of the Colorado River emerges as the 

 primary source of vulnerability for this declining 

 population. It serves to concentrate adults in a 

 predictable small area where they may be fished 

 with efl^ciency during a critical phase of their life 

 cycle. To recruit, the juveniles must traverse an 

 area of intense shrimp trawling activity which 

 artificially increases juvenile mortality and leads 

 to further reduction of this already-depleted stock. 



The precise factors responsible for the decline of 

 the totoaba stock cannot be identified with cer- 

 tainty, but we can enumerate the three most 

 probable causes as: degradation of the spawning 

 grounds, degradation of the nursery grounds, and 

 overfishing. The first two are a result of re- 

 placement of brakish waters by saline waters in 

 and around the mouth of the Colorado River. Both 

 alternatives may be explored by examining 

 Colorado River flow data and annual totoaba yield 

 over the critical period of declining catch and 



significant flow reduction. We might expect a 

 relationship to exist between flow and annual yield 

 if the density of the resident breeding population 

 (as measured by catch) varies with some unknown 

 but flow-related quality of the spawning ground. 

 Relation between flow and catch n years later 

 (with n years corresponding to age at recruitment) 

 would indicate the importance of some flow-relat- 

 ed quality of the nursery ground. Although tests of 

 overfishing using these data are ambiguous, if 

 catch is statistically related to catch n years later 

 we might expect a depletion of the breeding 

 population resulting from lowered recruitment 

 levels. 



The decline in catch with declining, erratic flow 

 is evident for the 1942-58 period (Figure 7). Fol- 

 lowing 1958, the catch increased to a secondary 

 peak and then crashed to the present all-time 

 minimum, though flow varied little in the same 

 period. For reasons given earlier, we discuss 

 separately the pre-1958 and post-1958 periods. 



For the years 1942-58 we have plotted catch 

 against flow (Figure 8). Linear regression of the 

 data reveals a highly significant correlation of 

 annual flow and catch for this period (P <0.001). 

 However, the river flow data are derived from a 

 diff'erent base after 1951; analysis of these data in 

 two segments, before and after this change point, 

 shows no significant relationship between catch 

 and flow for either the 1942-50 period or the 

 1951-58 period. These results suggest that the 

 highly significant correlation of flow and catch for 

 the total 1942-58 period may be spurious and due 

 only to the artificial pairing of declining catch and 

 declining flow functions. Despite these results, we 

 cannot ignore the fact that the totoaba congregate 

 only in the Colorado River estuary (so far as 

 known), and the salient feature distinguishing 

 this from other estuaries in the northern gulf is 

 the (former) discharge of large quantities of fresh 

 water from the Colorado River. Therefore, ac- 

 cepting the tentative nature of the flow-catch 

 relationship, we explore its possible biological 

 basis. 



The mechanism could lie in olfactory cues from 

 the river system (physiological responses to either 

 fresh water or substrate "odor"). Given the pres- 

 ent agricultural scene, such cues may no longer be 

 present. The present Colorado River surface flow 

 to the Gulf of California is close to zero for all 

 practical purposes and this situation is likely to 

 continue in the future. A conspicuous bar now 

 exists across the channel upstream from the delta 



538 



