FISHERY BULLETIN; VOL. 74, NO. 3 



islands, and flow measurements at the southern- 

 most Mexican hydrographic station known as El 

 Mari'timo, formerly considered the best single 

 index of actual surface input to the gulf (Schreiber 

 1969), were discontinued in 1968 for lack of mean- 

 ingful data.'' Further, the extensive use of all 

 available water from the lower Colorado River 

 drainage system for irrigation has resulted in 

 hypersalinity of return flows and is a major prob- 

 lem on both sides of the international boundary. 

 Water returned to the river channel which may 

 reach the gulf is now likely to be at least as saline 

 as the marine water it joins. Thompson (1968:8), 

 summarizing the history of Colorado River flow 

 and effects of exploitation on detrital loads, con- 

 cluded, "Probably little river detritus has reached 

 the northwestern Gulf of California in the last 

 55-60 years." Thus, if odor is not carried beyond 

 upstream dams and fields where the detrital load 

 stops, it must originate from reworking of the 

 massive deltaic deposits by the strong tidal cur- 

 rents of the uppermost gulf. If Thompson's es- 

 timate is correct, this may have been occurring 

 during the developing years of the fishery; the rate 

 of decay of such a process is unknown. 



The post-1958 flow and catch data (Figure 7) 

 contrast with those of the previous period. We feel 

 that the secondary peak in totoaba production may 

 be attributable to extraneous factors such as 

 changes in eff"ort or efficiency (availability of nylon 

 gill nets?) which produced a temporary increase in 

 catch. Another possible reason may have been the 

 enforcement of the 1955 breeding preserve 

 regulations which offered some temporary relief 

 from exploitation. If fishing in the sanctuary were 

 to resume after a period of time, the yield might 

 recover and fall in the observed manner. 



We now consider the second hypothesis, that the 

 cause of stock depletion is degradation of the 

 nursery ground. When annual totoaba yield is 

 compared with river flow in earlier years (e.g., the 

 1951 totoaba catch compared with the 1942 river 

 flow, etc.), lag times ranging from 6 to 10 yr all 

 give significant negative correlations (P<0.05) 

 using standard linear regression techniques. The 

 relationship is most distinct (Figure 9) when the 



lag time is 9 yr (P<0.01). The 6- to 10-yr periods 

 correspond with estimated ages of recruitment 

 employed below. ^ We find this negative relation- 

 ship of flow and (lagged) totoaba yield highly 

 interesting, though puzzling. The relation could be 

 taken to imply that survival of young stages is a 

 critical factor, since it couples increased river flow 

 in any one year with reduced recruitment of that 

 year class to the population. This interpretation 

 discounts hypotheses of larval and juvenile phys- 

 iological dependence on waters of lowered salinity 

 (Berdegue 1955, 1956; Cannon 1966; Cause 1969; 

 Sotomayor 1970). An alternate analysis using flow 

 data only for the March-July period over the years 

 of catch decline would be a better test of the effect 

 of flow on larvae and small juveniles. 



We know that successful reproduction still 

 continues in the northern gulf as demonstrated by 

 our ability to find juvenile fish on the nursery 

 grounds. Despite searching, we have found no 

 conspicuous subsurface freshwater seeps which 

 might have provided local areas for limited suc- 

 cessful spawning. We believe that reproduction 

 occurs over the entire ancestral spawning 

 grounds. Thus, we conclude that adverse effects of 

 salinity changes must operate in a relative and not 

 an absolute manner. The advantages realized by 

 potential recruits on the nursery ground may be 

 those of reduced predation and abundant food 



•^Nishikawa-Kinomura, K. A. 1973. Flow of the Colorado 

 River into the Gulf of California. In S. Alvarez-Borrego et al., 

 Preliminary report to the Secretariat of Hydraulic Resources on 

 the second stage of the chemical study on insecticide contami- 

 nation at the mouth of the Colorado River, p. 15-19. Unpubl. rep. 

 Mar. Sci. Unit, Inst. Oceanol. Res., Univ. Baja Calif., Ensenada, 

 Mex. 



^The senior author has reviewed the published estimates of 

 growth curves and ages of recruitment (see Arvizu and Chavez 

 1972, for a summary of this literature). Apparent discrepancies 

 between reported lengths at different ages and serious disa- 

 greement between Berdegue's (1955) growth estimates and the 

 distribution of lengths in observed commercial catches in 1963 

 (Arvizu and Chavez 1972) encouraged closer scrutiny of these 

 data. The variation in lengths at particular ages and in maximum 

 lengths reported by different authors and summarized by Arvizu 

 and Chavez appear to derive from use of both standard length 

 and total length measurements without discriminating between 

 the two. The senior author calculated von Bertalanfty growth 

 curves using a resolved maximum standard length of 1,600 mm 

 and the intermediate lengths reported by Berdegue (1955). The 

 new growth curves indicate that the best estimate of recruit- 

 ment age is 6 or 7 yr; they also produce a length series which 

 corresponds well with that observed in commercial catches. Male 

 and female totoaba may vary significantly in growth rates and 

 therefore may recruit at different ages. This variation allows 

 extension of the possible recruitment age to 10 yr. J. E. Fitch 

 (pers. commun.) has examined totoaba otoliths and concluded 

 that totoaba first spawn at age 8. If totoaba do not accompany 

 the migrant population until reproductively mature, his results 

 are consistent with the ages of recruitment used here. However, 

 his overall ages as read from otoliths indicate that these new 

 growth curves may contain a wide margin of error in terms of 

 predicted age at length observed. Fitch has also found that 

 totoaba scales are of little use for growth studies after about age 

 8; this may explain the maximum lengths at age 8 or 9 reported 

 by Nakashima (Jordan 1916), which we now believe to be 

 erroneous. It also may account for errors in Berdegue's (1955) 

 estimates, since he relied heavily on age determinations from 

 scales. 



540 



