OBSERVATIONS ON THE COMMERCIAL FISHERY AND REPRODUCTIVE BIOLOGY OF THE TOTOABA, CYNOSCION MACDONALDI, IN THE NORTHERN GULF OF CALIFORNIA

CHRISTINE A. FLANAGAN AND JOHN R. HENDRICKSON'

ABSTRACT

Information gathered from fishers and records of the failing totoaba, Cyno8cion macdonaldi, commercial fishery demonstrate the ability of the three principal ports to fully exploit the dwindling population during its annual breeding migration to the mouth of the Colorado River. Gonadal maturation, daily catch, and capture incidence data document the timing and route of the migration, provide evidence for a tendency toward unisexual schooling in its early phase, and point to the possibility that totoaba may form large aggregations before spawning is initiated. A trend toward reduction in the length of the migratory and spawning period, from 5 or 6 mo in 1965 to 1 mo in 1972 is documented with data from the port of Golfo de Santa Clara. In surveys of the hypothesized nursery area, 28 juvenile totoaba (6-12 cm standard length) were collected at 4 of 14 sampling sites. The four collection sites were commonly characterized only by depth «1 m) and substrate type (soft clay-silt sediments). Three hypothesized causes of the decline of this commercial fishery are examined by statistical analyses of Colorado River flow and annual totoaba catch data: overfishing,loss of spawning grounds, and loss of nursery grounds. Overfishing was found to be the most likely cause of the decline. Recent trends of catch data among the principal commercial fleets, and evidence that regulatory measures may have resulted in temporary recovery of totoaba production, provide further support for the overfishing hypothesis. The journey of the migrant population along a known route and its concentration into a predictable small area, its hypothesized requirement for dense aggregations prior to spawning, and the added mortality of juveniles taken by shrimp trawls in the near-delta waters are important points of vulnerability that render this endemic species particularly susceptible to fishing pressure. The possibility of the extinction of Cynoscion macdonaldi, without continuation of the newly decreed prohibition of fishing, is reiterated.

The totoaba; Cynoscion macdonaldi Gilbert 1891, the totoaba, declared by the Government of Mex­ 'is the largest species of the family Sciaenidae, ico on 2 August 1975, prohibits all capture of this with maximum reported lengths of almost 2 m species by both commercial and sport fisheries (H. (Berdegue 1956) and weights exceeding 135 kg Chavez, pers. commun.). (Cannon 1966); the larger females in present-day Although the species has been heavily exploited, commercial catches approximate 1.5 m and 35 kg its life history, population dynamics, and general (Arvizu and Chavez 1972). The species is endemic ecology are poorly known. Species accounts are to the Gulf of California, where it used to support a given in Jordan and Evermann (1898, 1902), Jor­ fishing industry and popular sport fishery based on dan et al. (1930), Gabrielson and Lamonte (1954), its annual spring breeding migration to the shal­ and Lanham (1962). The totoaba was included in low, formerly brackish waters of the Colorado accounts of commercial sciaenid species by Croker River Delta region at the extreme northern end of (1932) and Fitch (1949). Aside from these refer­ the gulf. The major portion of the catch was ences and others cited here, little has been pub­ exported from Mexico to the United States (prin­ lished on the totoaba; remaining incidental refer­ cipally San Diego) and brought a high price per ences may be found in Arvizu and Chavez (1972), pound under the influence of apparently unlimited the most recent summary of all available infor­ demand. Presently an indefinite closed season on mation on this species. Although notes on the ecology of the totoaba were first published in 1916 by Jordan, most of the presently accepted life 'Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, AZ 85721. history information is based on fisher's lore. These 2The common name is often spelled "totuava" by writers from beliefs were first documented by Berdegue in his the United States for no known reason. The spelling used here is that preferred and used by Mexicans; it should become the 1955 study of the fishery in which he also examined established spelling. scale annuli series and published the only derived

Manuscript accepted January 1976. FISHERY BULLETIN: VOL. 74, NO.3 531 FISHERY BULLETIN: VOL. 74, NO.3 growth estimates for this species. His work con­ cluded with a warning that the totoaba is a declining species, in danger of extinction from a combination of overfishing and the disappearance of brackish water spawning grounds due to diver­ sion of Colorado River waters for agricultural and other purposes. Gause (1969) and Sotomayor (1970) later echoed this view. ~ In this paper we present a short history of the o commercial fishery and report new information on ... totoaba life history. We summarize what is known

about the ecology of the species and speculate on YEAR consequences of the present small population size FIGURE 1.-Yield of commercial totoaba fishery, northern Gulf of and the intense fishing effort to which the fish have California for the 1929-75 period. Figure modified from Arvizu been exposed. We discuss the three most probable and Chavez (1972). Data for 1971-75 were obtained from H. causes for the decline in the fishery: degradation of Chavez (pers. commun.). spawning grounds, degradation of nursery grounds, and overfishing. We examine Colorado (Figure 1). After 1942, despite intensified fishing River flow data and annual catch data in the light effort and increased gear efficiency, the annual of these hypotheses, and discuss our results. In yield exhibited erratic fluctuation to the all-time conclusion, we draw together all these elements in minimum of approximately 58 metric tons in 1975 an attempt to assess the present and future status (H. Chavez, pers. commun.4). of this commercial population. Fishing methods evolved from spearing out of dugout canoes and primitive handlining in the HISTORY OF THE FISHERY early years, through dynamiting and primitive gill netting, to the use of efficient nylon gill nets. The Until about 1920, commercial exploitation of the usual modern net has a stretched mesh size of totoaba was confined to export of dried air blad­ approximately 25 cm and measures 100-200 x 4-5 ders to the Orient as an ingredient of a gourmet m. Gill nets were managed from diesel-powered soup (Chute 1930). Craig (1926) reported the first shrimp trawlers (12-18 m, some temporarily export of totoaba flesh to the United States. In diverted from shrimping during prime season these early, developing years, the totoaba fishery totoaba fishing), and from 4.5- to 7.5-m wooden or was directly responsible for the establishment of fiber glass "pangas" (launches) fitted with out­ three northern gulf fishing villages: Golfo de board motors. The activities of commercial fishers Santa Clara and Puerto Penasco in the State of have been largely limited to the prime breeding Sonora, Mexico, and San Felipe in the State of season (January-March) when the spawning Baja California Norte (Berdegue 1955). Analysis adults are in the shallow waters of the extreme of registered catches by all Mexican ports for the northern gulf. Prior to the 1975 total protection of 1966-70 period shows that these three ports totoaba, the prime fishing season ended with the produced from 94.9 to 97.7% of the total catch (H. advent of an official closed season, 1 April-15 May Chavez, pers. commun.). (Arvizu and Chavez 1972), a protective measure From 1929 (when Mexican Government statis­ enacted by the Mexican Government in about 1955 tics were first collected) onward, the fishery re­ (Berdegue 1955)." At the same time, a sanctuary sponded to a growing U.S. market by developing was designated at the mouth of the Colorado transportation and refrigeration capabilities and by improving fishing gear and boat facilities. 4The 1975 yield reported here is based on catch from principal Annual yield began to increase rapidly in 1934 and ports for the prime season only (through the month of March). the catch peaked at 2,261 metric tons3 in 1942 The final figures may be as much as 10% higher. 5According to Berdegue (1955), before 1955 there was a closed season extending from 20 March to 1 May; the prohibited period was changed to the later dates because active spawning was 3We follow the example of Arvizu and Chavez (1972) in giving observed after 1 May. In 1969 and 1970 the beginning of the yields as weights of cleaned fish lacking heads and viscera unless closed season was delayed 15 days in response to the fishers' specifically designated otherwise. To convert to whole weights, petitions when breeding schools had not appeared by the end of multiply by l.l (H, Chavez, pers. commun.). March (H. Chavez. pers. commun.).

532 FLANAGAN and HENDRICKSON: FISHERY AND REPRODUCTIVE BIOLOGY OF TOTOABA

River. All fishing was prohibited north of an sexual composition of the catch, and reproductive imaginary line extending from Bahia Ometepec state of the individual. on the Baja California coast to the mouth of the All fish examined by us were breeding adults. Rio Santa Clara on the Sonora coast. They were classified according to three mutually In addition to the standard commercial fishery, exclusive categories of gonadal development: If the Seri Indians of the Bahia Kino and Punta not running eggs or milt at the time of capture (or Chueca areas of Sonora were alleged to capture within 24 h of capture in the case of several totoaba in coastal waters during the fall and individuals kept alive for a period of hours), they winter, but we have been unable to confirm this by . were classified as "unripe"; if milt or hydrated personal observation. Further pressure was ex­ eggs ("applesauce" color and texture) could be erted on the stocks by an enthusiastic sport expressed with light pressure, they were classified fishery, based largely on the Baja California side as "ripe"; females with flaccid ovaries and running of the northern gulf, which took unknown ripe males taken in the same catch with such numbers of breeding adults during the prime females were classified as "spent." season. In recent years when diminution of the Effort data are reported as the number of stocks caused the success rate to drop, sport fishing operational net hours rather than total time virtually disappeared. At the peak of the sport (man-hours or boat-hours) spent fishing because fishery, large numbers of immature fish resident many of the large boats "hunt" for schools suitable in the upper gulf waters were also reportedly for encircling with their nets during the day and taken, usually unrecognized as totoaba. For a time, then set their gill nets in the usual manner to fish a deepwater handline commercial fishery and overnight. We believe that recent daytime hunt­ accompanying sport fishery continued out of San ing for schools to encircle was practiced more in Felipe during the summer after the adult fish had memory of times past than as a practical matter of left the spawning grounds, but this activity also probability. In approximately 50 days aboard such declined in recent years. Craig (1926), Chute vessels, we have never seen a school located, (1928), and Berdegue (1955) provided further although one heard of such catches each season. information on the history of the fishery and The method persisted because, if successful, it can contain most of the documented information on yield very high tonnage. The larger, diesel­ the sport fishery. powered trawlers with ice-filled holds frequently stayed at sea for more than a week and commonly METHODS AND MATERIALS traveled considerable distances back to their home ports to land the catch. This is in marked contrast The junior author began field studies on the to the methods of the fishers of Golfo de Santa species in 1970 with the primary objective of Clara, who fished primarily from pangas and who gathering life history information for conserva­ customarily inspected their nets each day by tion purposes. The results reported here derive passing the net over the boat, leaving the weight­ primarily from data collected by the senior author ed ends in place. Such nets "fished" continually, during three cruises aboard commercial fishing except for occasions when they were taken up to be vessels from Puerto Penasco in March and April moved to alternate spots. Lacking storage and 1972. Fishing patterns during these cruises in­ refrigeration facilities, the pangas had to return cluded most of the Gulf of California north of lat. to port each day with their catch from one or two 31°N; with few exceptions, the locations were gill nets. Catch in kilograms was recorded by a selected by the fishing captain. Mexican government fisheries inspector for each The data were gathered by direct observation of panga, each day. Although we attempted to cal­ catch and, in a few cases, by reports from fishers culate catch per unit effort, we were unable to on "companion" vessels (as many as five other resolve its heterogeneous nature. Here we present boats in the cooperating group, in one instance). only our analysis of effort from the panga fishery During 22-24 March 1972 we also collected data of Golfo de Santa Clara. from the panga fleet at Golfo de Santa Clara as the In late May and early June 1972, a number of catch was landed and cleaned at the port. In both Sonora and Baja California sites around the pe­ circumstances, our data consisted of information rimeter of the extreme northern gulf were sur­ on location, time and size of catch, number of veyed for juvenile totoaba, using both commercial operational net hours, time and state of tide, trawl nets and beach seines. Many of these sites

533 FISHERY BULLETIN: VOL. 74, NO.3 were revisited in June 1973. Observations were RESULTS AND DISCUSSION made of water temperature, salinity, turbidity, and substrate character; associated faunas at each Breeding Migration site were sampled. A few juveniles were trans­ ported alive back to Tucson, Ariz., and maintained The fishers believe that the annual migration of there for about 80 days. Information on distribu­ totoaba is prompted by the urge to reproduce and tion and habitat of the juveniles is presented here; is guided by the search for a suitable estuarine notes on behavior of the juveniles in captivity will spawning environment. According to their beliefs be reported elsewhere (C. A. Flanagan in prep.). the breeding population, seeking areas of reduced In our discussion of the hypotheses for the salinity, leaves deep water in the mid-gulf and decline of the totoaba fishery, we present statistics follows the Sonora coastline northward; eventually of Colorado River flow and annual totoaba yield. the schools reach the mouth of the Colorado River, The annual yield data are those already presented where they spawn. Following spawning, the to­ (Figure 1). For flow, we have attempted to es­ toaba supposedly seek out the clearer, deeper timate the amount of water delivered to Mexico in waters to which they are more accustomed and the main river channel at the southerly interna­ follow the Baja California coastline on their return tional boundary on the assumption that it will bear migration southward. These beliefs are based some regular relationship to the volume of fresh upon commercial catch experience dating back to water entering the Gulf of California. This as­ the late 1920's. sumption becomes tenuous with the development Localities and dates of capture observed by the of lowland agriculture in Mexico and with sig­ senior author in 1972 (Figure 2 and Table 1) appear nificant groundwater pumping in the United to document a pattern consonant with the above States, both in evidence since about 1960. Suitable hypothesis, as do observations by the junior author effort data for the totoaba fishery are unavailable in earlier years. The regular port statistics also but we have assumed that, following the peak implicitly support the hypothesis, with catches catch in 1942, effort was constant or increasing. each year reported chronologically first by Puerto This assumption is probably warranted given the Penasco, then by Golfo de Santa Clara, and last by demand and high price paid for totoaba flesh. The San Felipe fleets. The data in Figure 2 represent limitations imposed by our assumptions are that but a small fraction of the total 1972 fishing effort, no catch datum before 1942 and no flow datum however, and in the most conservative interpreta­ after 1960 may be considered in these analyses. tion demonstrate only that experienced fishers

11S' 114'

.....::.:::.-. 'RUDING Plun.... '

OBSlRVID CA1CH AUAS

• Golfo de Sente Cle,. FIGURE 2.-Locations and dates of observed commercial catches of to­ toaba during the 1972 prime fishing season. Catch information in terms of tonnage by day, boat, and area were also obtained from Fisheries Inspec­ tors. These latter data are reflected in MAI.16·11 the early capture date of 12 February and the extended capture period of MAI.5fEI.12- -(,••.•..••...... •... 11-29 March in areas I and III, re- <.:.:: 3" spectively. Chart shows Gulf of ,,' California north of lat. 31 0 N (see locater in upper right-hand corner).

534 FLANAGAN and HENDRICKSON: FISHERY AND REPRODUCTIVE BIOLOGY OF TOTOABA TABLE 1.-Totoaba captures observed in 1972 (see Figure 2 for mid-February, but may take place as early as areas and timing). The figures shown here are personal observa­ December or as late as the end of March. tions of the senior author. Three references exist in the literature regard­ No. of observed No. of successful Total no. of Area net sets net sets totoaba caught ing the spawning period. Nakashima (in Jordan

I 5 1 1 1916) said that the main spawning period was in II 1 1 1 early May, while Berdegue (1955) reported the III (no datal '149 {'l IV g 8 42 reproductive season as extending from the end of V 6 5 18 February or early March until early June. Obser­ VI 11 1 1 vations by D. A. Thomson (1969) and the junior 'These data are displayed in Figure 5. Although catch data are author over the last four seasons indicate peak available for area III only in terms of tonnage, not head count of fish taken, use of the 35-kg average per fish would give a con­ spawning as late as April and early May but, more servative estimate of at least 2,500 individual fish taken In area III during 1972. commonly, in mid- to late March. Historical data and existing statistics confirm the fisher's claims know where and when to find fish. Ideally, Figure 2 that the period of concentrated catch (which should reflect the results of an even pattern of apparently coincides with peak spawning) has standard net sets through the February-June become progressively abbreviated during the past period. 20 yr. The monthly catch data for Golfo de Santa To our knowledge, no one has investigated the Clara for the 1964-72 period show a clear reduction salinity preferences or tolerances of spawning in length of season from 5 or 6 mo to an ab­ adults, but this raises the question of totoaba breviated period in March-April at present spawning sites in estuarine areas of other major (Figure 3). We believe (see below) that the catch of gulf rivers. Spawning totoaba have never been the Golfo de Santa Clara fleet is a good reflection reported from locations other than the Colorado of spawning activity and suggest that a pattern of River mouth. While further investigation is clear­ repeated spawnings formerly extending from ly warranted, at present we accept the fisher's January-February to May and June has collapsed hypothesis as an adequate predictor of population to a single event which coincides with the old migratory patterns. temporal mode. A small remnant population might be expected to react more uniformly to environ­ Spawning Concentration mental cues than would a large one, a factor leading to progressively shorter migratory and Because the annual breeding migration results spawning periods. This is consistent with our in a high density of fish within a limited area, it observation of breeding population residence time has become the single most important aspect of of only 18 days on the spawning grounds in 1972 the fishery: total prime season catch is a function (from 11 March to 29 March, see discussion below). of the number of fish arriving in the spawning A limited amount of qualitative data on gonadal area before 31 March in an average year. The maturation, collected during the 1972 prime appearance of migrant schools of totoaba in shal­ fishing season (Figure 4), indicates that males low coastal waters, as signaled by catches from ripen before females and retain spawning readi­ exploratory boats which have ventured out in ness for longer periods of time-a common occur­ anticipation of their arrival, usually occurs in rence among fishes. It also provides evidence for a

FIGURE S.-Monthly yield in metric tons of totoaba, port of Golfo de Santa Clara. Data for 1966-70 from Arvizu and CMvez (1972); H. CMvez (pers. commun.) supplied data for 1964-65 and 1971. The 1972 catch data were obtained from F. Aguilera. Fisheries Inspector (15 additional metric tons recorded in 1972 are not shown because month of capture was uncertain).

YEA R

535 FISHERY BULLETIN: VOL. 74, NO.3

SPENT a a a •• '0

RIPE a a a '0 ------1• •• • •• • UNRIPE a o. a a • } J.8 w 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 u 204 ~ MARCH 1972 12 170 10 U Z FIGURE 4.-Degree of gonadal maturation observed on specific INCIDENCE

days during the 1972 prime fishing season. See text for explana­ 10.2 tion of maturation categories. Each symbol represents one or more individual fish. Open squares represent females; solid 60 squares represent males. J.' c;:o II =12 17 '8 19 20 21 22 23 24 29 tendency toward sexually segregated schooling, at MARCH 19 7 2 least in the case of male fish. It will be noted that FIGURE 5.-Golfo de Santa Clara catch and capture incidence all the records for the early portion of the period plotted against days of March for 1972 prime fishing season. See portrayed in Figure 4 show a single sex per catch, text for explanation of capture incidence unit. Between 15 and 22 while the records for the later portion of the period March the number of individually successful nets remained show both sexes in all but one instance. It must be comparatively constant, despite the peak catch on 22 March. clearly understood that some of the catch records Official statistics indicate that 15 metric tons in addition to the total of 71 metric tons shown in Figure 3 were recorded for this portrayed in Figure 4 represent single fish, mak­ port in 1972 (H. Chavez, pers. commun.). These additional data ing those data points meaningless in this context, cannot be traced to daily catch for inclusion in this figure. but a majority of the data points represent multi­ ple individuals. This apparent sexual separation of readiness each year for the arrival of the migrant prespawning schools conforms with general ob­ population and, within a few days of the first servations by Hendrickson in years before 1972 catches by exploratory nets, virtually all available and with the caption for figure 84 in Chute's (1928) gill nets were deployed for fishing of totoaba. paper describing the earlier hook-and-line fishery: Despite daily success or failure, fishing effort "Practically all of the fish in this picture were continued at this level until the season closed on 1 males ...." (the figure, depicting the butchering April. Most of the panga fishers worked a definable process, shows about 15 large fish caught by three area of the delta where "canals" (extensions of men in 3 h). Colorado River channels) deep enough to accom­ Success of the Golfo de Santa Clara panga modate the large totoaba gill nets are separated by fishery is related to the siz€ of the migrant totoaba shallow mud bars (see area III in Figure 2). population, the length of the period in residence on In 1972 the first catch off Golfo de Santa Clara the fishing grounds, and population behavioral occurred on 11 March and was followed by a period patterns. Because the fishing grounds are identical of increasing catch and capture incidence due to with, near to, or in the path to the spawning increasing effort until 15 March (Figure 5). During grounds, analysis of the panga fishery catch sta­ the March 15-22 period, capture incidence was tistics can yield valuable insight into the breeding relatively constant and high. During this same biology of this species. We have used "capture period, catch varied somewhat erratically and incidence" as a measure of fishing success, em­ peaked on 22 March, after which both catch and ployed here to give a quantitative indicator of the capture incidence fell off drastically despite no presence of breeding adults on or near the sus­ reduction in fishing effort. The 22 March catch pected spawning grounds (Figure 5). One capture amounted to 27% of Golfo de Santa Clara's yield incident is defined as the catch of at least one for that year and represented 9% of the total totoaba per panga per day; the daily total reflects recorded yield from all ports for 1972. The average the number of individually successful net sets. We Golfo de Santa Clara net must have contained over assume that: 1) fishing effort is constant after a twice as many fish on 22 March as on 21 March and given date within the prime season and 2) fishers 3-5 times as many as on other "good" days in the individually and collectively fish in the same area prime season. each day throughout the period. These assump­ What factors in totoaba reproductive biology tions are in keeping with the nature of the Golfo might explain these results? Catch per net may be de Santa Clara fishery. This vitlage waited in considered an index of migrant arrivals if we

536 FLANAGAN and HENDRICKSON: FISHERY AND REPRODUCTIVE BIOLOGY OF TOTOABA suppose that, as breeding adults reach the north­ from about 15 cm to about 45 cm. The holotype in ern end of the gulf, they immediately move up into the U.S. National Museum is approximately 25 cm the channels at the mouth of the Colorado River. long and was taken in 20 fathoms of water (Gilbert The arrival of the largest population segment 1891). To our knowledge, the first collection of would then be indicated by the peak catch. Alter­ really small juveniles (6-12 cm size range) which natively, the peak catch may have signaled a peak were positively identified as totoaba was made in of spawning activity by an already-resident 1970 near San Felipe, B.C., and described by breeding population, becoming more vulnerable to CMvez (1973). We surveyed probable northern the nets by virtue of spatial concentration and/or gulf sites for the presence of such small juveniles behavior. While the data do not allow firm conclu­ during May and June 1972 and 1973 (Figure 6). sions, we favor the second alternative. Substrate and depth appear to be more impor­ A period of behavioral stimulation in schools to tant than either temperature or salinity in char­ induce the spawning act is suggested by the fact acterizing the habitat of the captured juveniles. that enormous numbers of individuals allegedly For all sites, surface water temperatures ranged used to gather in this relatively small area to from 25° to 29°C and salinities were recorded spawn (Jordan 1916; Berdegue 1955). Although the between 35 and 40 %0. Sites where we collected population has been drastically reduced, the fish juveniles were shallow as compared to the other apparently continue this habit. If the release of sampling locations, and none were collected from reproductive behavior patterns depends upon depths greater than 1 m. Substrates were com­ mutual stimulation within large aggregations posed of fine clay-silt sediments, devoid of sand; (consistent with their sound-producing air blad­ der; see Breder and Rosen 1966), the present small population might be experiencing some break­ down in the behavioral sequence with consequent SI Ie No. lowered reproductive success. , 872: • Gallo de Although the significant yields of 21 and 22 Santa Clara March may indicate a peak in spawning activity, this does not preclude the possibility thatother fish 2 EI 10rnlll.1 Chlnero" t.- were later in arrival and that spawning also occurred in April (during the closed season). The aahla de Golfo de Santa Clara fleet's near-failure to catch Adal r fish during the 25-31 March period, and our failure to find adult fish during the April cruise lend doubt Puerto Pen••co to this possibility, but the existence of more than one breeding population should not be ruled out. ". Juvenile Totoaba Distribution, 1973 SI te No. !! Habitat, and Diet ..1) 12. The microhabitat and residence time of juvenile totoaba on the nursery grounds are largely un­ known. Berdegue (1955) reported that juveniles remain in the shallow waters near the Colorado • Golla de River mouth until they begin a southward migra­ Sante Clar. tion to join the parent population. The Colorado River Delta is heavily exploited by the shrimp fishery during parts of the year (effort was ob­ served to be especially intense during April, May, and June), and Berdegue first called attention to FIGURE 6.-Sites sampled in northern Gulf of California for the increased mortality of juvenile totoaba due to presence of juvenile totoaba in 1972 and 1973. Circled numbers indicate offshore areas sampled by otter trawl. All other locations shrimp trawling activity. are shore stations sampled by seine. Sites where juvenile totoaba In our experience, the juveniles captured in were found are indicated by stars. Numbers captured are shown shrimp trawls are individuals ranging in length at upper right of each map.

537 FISHERY BULLETIN: VOL. 74, NO.3 the mud surface layer was very soft. No small significant flow reduction. We might expect a juvenile totoaba were collected over firm mud relationship to exist between flow and annual yield sediments or sandy substrates, as is often the case if the density of the resident breeding population with the larger individuals taken in shrimp trawls. (as measured by catch) varies with some unknown Guevara (1974) is presently analyzing the dis­ but flow-related quality of the spawning ground. tribution of juvenile totoaba captures in shrimp Relation between flow and catch n years later trawls. Most of his specimens are larger than ours, (with n years corresponding to age at recruitment) implying that the fish move into deeper water as would indicate the importance of some flow-relat­ their growth continues. Tidal currents in the area ed quality of the nursery ground. Although tests of are extreme and these may also playa significant overfishing using these data are ambiguous, if role in juvenile distribution. catch is statistically related to catch n years later Juveniles collected in 1972 were examined for we might expect a depletion of the breeding stomach contents. Remains of amphipods and population resulting from lowered recruitment other small crustaceans common to the habitat levels. were recognizable, in addition to remains of The decline in catch with declining, erratic flow juvenile fishes which we identified as Micropogon is evident for the 1942-58 period (Figure 7). Fol­ sp., Mugil cephalus, and Leuresthes sardina. lowing 1958, the catch increased to a secondary Within the limits imposed by size, the diet of peak and then crashed to the present all-time juvenile totoaba as small as about 6 cm standard minimum, though flow varied little in the same length is comparable in these items with the diet period. For reasons given earlier, we discuss of the large adults. separately the pre-1958 and post-1958 periods. For the years 1942-58 we have plotted catch Decline of the Fishery against flow (Figure 8). Linear regression of the data reveals a highly significant correlation of We have traced the growth and decline of the annual flow and catch for this period (P <0.001). totoaba fishery and discussed its present status However, the river flow data are derived from a and methodology. We have presented data on different base after 1951; analysis of these data in aspects of totoaba life history and raised questions two segments, before and after this change point, concerning possible reproductive behaviors which shows no significant relationship between catch may have a bearing on reproductive potential. and flow for either the 1942-50 period or the Although these have significance, if we consider 1951-58 period. These results suggest that the the resource from a management perspective one highly significant correlation of flow and catch for fact becomes clear: the annual breeding migration the total 1942-58 period may be spurious and due to the mouth of the Colorado River emerges as the only to the artificial pairing of declining catch and primary source of vulnerability for this declining declining flow functions. Despite these results, we population. It serves to concentrate adults in a cannot ignore the fact that the totoaba congregate predictable small area where they may be fished only in the Colorado River estuary (so far as with efficiency during a critical phase of their life known), and the salient feature distinguishing cycle. To recruit, the juveniles must traverse an this from other estuaries in the northern gulf is area of intense shrimp trawling activity which the (former) discharge of large quantities of fresh artificially increases juvenile mortality and leads water from the Colorado River. Therefore, ac­ to further reduction of this already-depleted stock. cepting the tentative nature of the flow-catch The precise factors responsible for the decline of relationship, we explore its possible biological the totoaba stock cannot be identified with cer­ basis. tainty, but we can enumerate the three most The mechanism could lie in olfactory cues from probable causes as: degradation of the spawning the river system (physiological responses to either grounds, degradation of the nursery grounds, and fresh water or substrate "odor"). Given the pres­ overfishing. The first two are a result of re­ ent agricultural scene, such cues may no longer be placement of brakish waters by saline waters in present. The present Colorado River surface flow and around the mouth of the Colorado River. Both to the Gulf of California is close to zero for all alternatives may be explored by examining practical purposes and this situation is likely to Colorado River flow data and annual totoaba yield continue in the future. A conspicuous bar now over the critical period of declining catch and exists across the channel upstream from the delta

538 FLANAGAN and HENDRICKSON: FISHERY AND REPRODUCTIVE BIOLOGY OF TOTOABA

13

12

11

2,000 10 ..... til jjj C I 9 0 .. iii ..... ',600 8 ;; U m -.:: m CI I:: m ... .. m m , CI !I! .. , jjj CI 111 E m Hi m 1"1,,,1 !Ii .. w ". ;il iF U 1,200 Hi m 6 III H= iii m i.".!. m ~ m m ' III "'0 il! m (,) m III ... .. ; lit III !!i m 5 ...... f- m i!~ m :'; m ct ill m Rl iii 111 ~ (,) 800 m ii! m .,',I"l",l o 11] Hi !!! !II iH ::: ..J m III I~ Hi II. ill III l!1 m l1t m Iii iii ii' !Ii IIi III 111 III 400 - ill i" j:: :i! !1i ,g !Ii !If i:l !H III F' ill ill :;: i!! iii ill m ;,. !II Iii 11,.( ~,.,':l",',!,. ::: 11)m ill m ni m !'I m W m H

YE AR FIGURE 7.-Colorado River flow in thousands of acre-feet and totoaba fishery annual yield in metric tons for the 1930-75 period. Totoaba annual yield data are those of Figure 1. Flow data were calculated from records published in Water Supply Papers 1313, 1733, and 1926 (U.S. Geological Survey, respectively, 1954, 1964, and 1970). Flow data for 1966-75 are not shown but we do not expect them to deviate beyond the 1960-65 variation above. We have calculated the flow delivered to Mexico at the southerly international boundary (near San Luis, Ariz.) as follows (data sources are cited only on first mention): 1930-36: Colorado River at Yuma (1954:710) + Yuma Main Canal Wasteway (1954:717) + Calif. Drainage Canal (1954:723) - Alamo Canal (1954:724) + Eleven-mile Wasteway (1954:726) + Cooper Wasteway (1954:726); 1937-50: Colorado River at Rockwood Gate, Calif. (1954:712) - Alamo Canal + Eleven-mile Wasteway + Twenty-one Mile Wasteway (1954:727) + Cooper Wasteway; 1951-65: Colorado River at southerly boundary, near San Luis (1964:563; 1970:519-521). 21° 0

2POO

.. 1,600 C 0

U 2 FIGURE 8.-Plot of annual totoaba YEARS r d f p ~ 1,200 yield and annual Colorado River flow 1942-1950 .414 .643 7 <.10 II for the 1942-58 period. Data are those - -.112 6 >.10 E 1951·1958 .012 displayed in Figure 7. The points "42'1958 .574 .758 15 <.001 :x: below the dashed line represent the u 800 Y=178.15 +.153 X 1951-58 flow years. Though the rela­ l- tionship between catch and flow for e the 1942-58 period are highly sig­ u , . nificant, the disparity between the r2 400 levels for the component periods 1942-50 and 1951-58 invites caution in interpretation of these results.

6 7 8 9 10 11 12 fLO W (lOll. ere· feet) 539 FISHERY BULLETIN: VOL. 74, NO.3 islands, and flow measurements at the southern­ lag time is 9 yr (P

540 FLANAGAN and HENDRICKSON: FISHERY AND REPRODUCTIVE BIOLOGY OF TOTOABA

2,000 FLO W YE II RS 1936 ·1960 Cate h lag p I - 5 <.396 ",05 6 .449 <.05 7 459 <,05

.502 <,05 A E Y ,1033 -.606X _9 .550 <.01 \0 .454 <.05

~ FIGURE 9.-Plot of annual totoaba III .. yield (metric tons) 9 yr following the >- 800 annual recorded flow, for the 1936-60 '" period. This plot displays the sug- J: gested relationship between flow and U >­ recruitment level. Linear regression « 4 00 . calculations employing lag times from u 1to 10 yr are significant only for those years corresponding to estimated ages of recruitment (6-10 yr).

o I , I I ~ 7 8 9 '0 11 12 13 FLOW (106acre.feell resources, both of which are directly related to gear efficiency, effecting a realized increase in substrate and shallowness, and indirectly related effort, serves to confound the analysis; although to flow. catch would increase, this factor alone could not The final cause suggested for the decline in explain recovery of catch to such high levels. We totoaba stock is overfishing of the breeding can visualize a combination of factors giving rise population. We have examined the relationship of to the significant second peak in totoaba produc­ catch with catch n years later. If catch is a good tion (increase in gear efficiency acting on an indicator of population size, then we would expect increased population size following the period of a linear, positive relationship between population regulation) but lacking effort data throughout the size and the size one recruitment age later. Alter­ period, our hypothesis must remain speculative. natively, if catch is partially a function of Support for the overfishing hypothesis may lie sociopolitical constraints (e.g., enforcement of a in the recent trend of the relative catches of the preserve area and closed season resulting in a three main totoaba fishing fleets. Historically, the catch which significantly underrepresents the population size), we might expect a more com­ '" 1,200 ..~ plicated plot with a distinct cluster of years c ~ corresponding to periods of fishing regulations. y: 182:.9· ,194X '" / p •• 800 / .497, 05 c'" / . We have analyzed plots of catch against catch for .. / .. ~ recruitment periods ranging from 6 to 10 yr and .. //// ,. //~" have found significant relationships which satisfy / . U 400 . / . both of the above predictions. Graphs for all :; .--; . estimated ages of recruitment from 6 to 10 yr U showed essentially the same pattern (Figure 10). ~-r-,--r---r---r--.,.-'T'-,-,-'---'-1"I-rl" ---,-," -rTi.~ However, we note inconsistencies which advise 400 800 'j200 ',600 2pOO against drawing strong conclusions of either C" TCH overfishing or the demonstrated worth of enforced FIGURE 10.-Totoaba catch plotted against catch 8 yr previous. regulatory measures. For example, increases in This examines for evidence of reduction in breeding stock by catch occur 2-3 yr earlier than is consistent with overfishing. All data are from Figure 1, in metric tons. Points above dashed line are for the 1952-59 period. The breeding our assumption that regulatory measures were not sanctuary was established in 1955 and the points corresponding enforced until 1955, given that the minimum age to the years 1955-59 may reflect enforcement of this regulation of recruitment is 6 yr. The question of a change in (see text).

541 FISHERY BULLETIN: VOL. 74, NO.3

TABLE 2.-Prime season catch,' in metric tons, of the three The totoaba breeding behavior we describe principal totoaba fishing ports, 1964-75. Data for 1965-70 period are from Arvizu and Chavez (1972). Data for 1971-75 were serves to render the resident spawning population provided by H. Chavez (pers. commun.). The 1974-75 data are especially vulnerable to fishing effort. Frenzied preliminary but are not expected to increase by more than 10% spawning in dense aggregations following a pe­ from these figures. riod of behavioral stimulation insures that when a Year Puerto Penasco Golfo de Santa Clara San Felipe net is encountered, the capture rate will be par­ 1964 72.7 128.4 ticularly high. Our capture incidence, daily catch, 1965 97.0 561.2 and gonadal maturation data confirm the con­ 1966 177.5 388.6 1967 188.3 173.4 sequences of these attributes. The bathymetry of 1968 37.9 385.7 the delta restricts the spawning schools to highly 1969 60.2 213.8 1970 27.2 248.7 limited areas. These areas, or channels, are the 1971 69.0 46.0 prime fishing sites for the Golfo de Santa Clara 1972 52.0 86.0 1973 88.0 21.0 fleet. They appear to lie (by our estimate) partially 1974 51.0 17.0 within the breeding sanctuary established by the 1975 49.0 4.0 Mexican Government. 'Catch is calculated by adding January-April yields as recorded in the official statistics. We reiterate here the artificial fishing mortality suffered by juveniles in their forced crossing of the near-delta waters as they make their way south Puerto Penasco fleet catches fewer totoaba than from the nursery grounds. We have documented either the San Felipe or Golfo de Santa Clara fleets some known nursery sites and have suggested (Table 2). 'This is a logical result of the fishing characteristics of the juvenile habitat which may methods and areas worked by the three fleets. have predictive value in future surveys of the area. However during the last 3 yr, Puerto Penasco has We have examined the three most probable equalled or exceeded the other ports in recorded factors responsible for the decline in totoaba stock. totoaba yield despite no apparent increase in Subject to the limitations of our catch and flow effort. Our interpretation of this new trend is that data, our results suggest that overfishing has the migrant population, encountered first by the played the most significant role during the pre­ Puerto Penasco fleet along the Sonora shore, is 1958 catch period. We speculate that the low yields being decimated before reaching the spawning of the 1956-59 period may have been due to en­ grounds. forcement of the breeding sanctuary regulation andthatthispartialternporaryrelieffrom exploita­ CONCLUSIONS tion may, together with increased gear efficiency, have been responsible for the second peak in Our review of the history of exploitation of the totoaba production. Ifthis is true, then the power stock, our data on spawning concentrations, of regulatory measures for recovery of this com­ breeding migration, and juvenile habitat, and our mercial stock has been demonstrated. The cor­ analyses of proposed hypotheses for the decline of relation of annual yield with annual Colorado the fishery have emphasized points of population River flow, though weakened by statistical ir­ vulnerability. Fleets of the three major ports are regularities, attests to the importance of some highly skilled at finding the migrant schools of flow-related quality of the spawning grounds. totoaba. They have, in a sense, specialized to Degradation of the spawning grounds, possibly in exploit the ascent, resident, and descent phases of the ability to provide olfactory cues, also may have the breeding migration, both by nature of their resulted in a decline of the commercial population. vessels and their port facilities, and by con­ According to our results, degradation of the nur­ sequence of their geographic locations. This level sery grounds, through deterioration of some un­ of exploitation is possible only because the fishers known flow-related quality, has probably not are able to predict with accuracy the migration played a significant role in the fishery's decline. pattern. In the past, the commercial population Although it may be possible to ignore statistical level was high and the temporal nature of the port analyses and the conclusions therefrom, one can­ specializations was not a factor in the ranking of not deny that the annual yield in 1975 was the port yields. Now, when the population level has minimum recorded in the history of the fishery, a reached an all-time low, the Puerto Penasco fleet mere 2.5% of the highest recorded catch. The area seems to have some new advantage. of the fishery has shrunk to a small fraction of its

542 FLANAGAN and HENDRICKSON: FISHERY AND REPRODUCTIVE BIOLOGY OF TOTOABA former size, and catch events have become spo­ ACKNOWLEDGMENTS radic and undependable. The span of the breeding period has been reduced from several months to a Various research grants and contracts con­ period of only 18 days within the open season of tributed to the support of this work and are here 1972. Of those 18 days, a majority of the catch gratefully acknowledged: International Union for occurred on 21 and 22 March. Although data are the Conservation of Nature and Natural Re­ scarce, the average size of adult fish is reduced and sources/World Wildlife Fund Project No. 623; in recent years most commercially caught in­ National Science Foundation Grants GB29101 and dividuals have probably been first- or second-year . GB34675; National Aeronautics and Space Ad­ spawners. These harsh facts are indications of a ministration Contract NAS5-21777; and a special fish population struggling unsucessfully for sur­ University of Arizona Foundation Grant. We vival under pressure. acknowledge with sincere appreciation the sup­ The future of the species is uncertain. Until the port and professional participation of Mexican recent action of the Mexican Government in authorities and scientists in this work. Luis Ka­ establishing a total closed season, the outlook was suga Osaka, Director of the National Institute of bleak, indeed. While the commercial fishery was Fisheries, Mexico, and Pedro Mercado Sanchez of ready to crash before its legal cancellation (a the Subsecretariat of Fisheries, Ministry of In­ number of financial failures were reported. to us), dustry and Commerce, helped us obtain neces­ and would presumably never have hunted down sary permits (3977, 5344, 8202, and 12183) for work and eliminated the last reproductive pair of these in the field and in other ways provided important magnificent animals, the continued rising prices support and encouragement. Among the many for totoaba in a seller's market would have guar­ members of the National Institute of Fisheries anteed continued maximum pressure. If there are who gave freely of their time and efforts, special behavioral elements in the reproductive pattern of mention must be made of Joaquin Arvizu and the species which require mutual stimulation in Humberto Chavez, who cooperated in field work large schools for reproductive success, a threshold and data analysis. Weare particularly indebted to may have already been crossed which will drive the Biologist Chavez, Head of the Institute's Depart­ totoaba the way of the passenger pigeon. The ment of Fisheries Resources, for important con­ trends produced by irreversible change of the tributions and professional advice from inception spawning ground may prove more important than of field work to manuscript preparation. The we have speculated. In either of the last two Fishery Inspectors of San Felipe, Baja California circumstances, the ability of the stocks to rebound Norte, and Puerto Penasco, Sonora, rendered upon release of fishing pressure may be critically valuable assistance in making contacts with impaired. fishers and made date available. Francisco We suggest three meaningful measures at this Aguilera Grijalva, Fishery Inspector at Golfo de stage: 1) Continuation of the total closed season Santa Clara, Sonora, made particular efforts on our which has been imposed, until intensive studies behalf and was of invaluable assistance in collec­ document a strong and vigorously increasing tion of detailed catch data. We thank all the population. We suggest an enforcement period of students of the University of Arizona who ren­ about 1Ih times our estimated 6-yr minimum dered assistance on field trips, and all the staff and recruitment age, or 10 yr. 2) Action by the U.S. students of the Institute of Oceanologic Inves­ Government (the major market area) com­ tigations, Autonomous University of Baja plementing the Mexican action by declaring the California, who have been such productive totoaba an endangered species, to facilitate en­ partners in most of our work in the northern Gulf forcement of the neighbor country laws by of California. Special recognition is due L. T. removing much of the stimulus for poaching and Findley for his field and museum contributions smuggling. 3) Intensive scientific investigation to and enthusiastic interest. We greatly appreciate provide knowledge of the species' autecology and the time spent by D. A. Thomson, J. Tash, and G. behavior with potential application to all facets of Pyke who read the manuscript critically and management, ranging from environmental ma­ consulted on matters of data analysis and inter­ nipulation to hatchery techniques. Failing these, pretation (all responsibility for errors is ours). It is we conclude that the probability of extinction of impossible to overvalue the contribution of Lupe Cynoscion macdonaldi by the year 2000 is high. P. Hendrickson in clerical, translation, and edi-

543 FISHERY BULLETIN: VOL. 74, NO.3 torial functions throughout the entire course of Commission steamer Albatross. No. XII-A preliminary the project. Lastly, a particular expression of report on the fishes collected by the steamer Albatross on the Pacific coast of North America during the year 1889, gratitude is due to Javier RamIrez of Golfo de with descriptions of twelve new genera and ninety-two Santa Clara, master fisher and astute student of new species. Proc. U.S. Nat!. Mus. 13:49-126. nature, who was frequently the key to success in GUEVARA, S. field projects. 1974. Sobre Ie eclogia de los juveniles de totoabe Cynoscion macdonaldi Gilbert. (Abstr.) In Restimenes Quinto Congreso Nacional de OceanografIa, 22 Oct. 1974, p. 7. LITERATURE CITED Guaymas, Sonora, Mexico. JORDAN, D. S. ARVIZU,J., AND H. CHAVEZ. 1916. Notes on the totuava (Cynoscion macdonaldi Gilbert). 1972. Sinopsis sobre la biologfa de la totoaba, Cynoscion Copeia 1916:85. macdonaldi Gilbert, 1890. FAO (Food Agric. Organ. U.N.) JORDAN, D. S., AND B. W. EVERMANN. Fish. Synop. 108, 21 p. 1898. The fishes of North and Middle America. Bull. U.S. BERDEGUE, A. J. Nat!. Mus. 47, Part 2:1241-2183. 1955. La pesqueria de la totoaba (Cynoscion macdonaldi 1902. American food and game fishes. Doubleday, Page & Gilbert) en San Felipe, Baja California. Rev. Soc. Mex. Co., N.Y., 573 p. Hist. Nat. 16:45-78. JORDAN, D. S., B. W. EVERMANN, AND H. W. CLARK. 1956. Peces de importancia comercial en la costa nor­ 1930. Check list of the fishes and fishlike vertebrates of occidental de Mexico. Secre. Mar., Dir. Gen. Pesca Ind. North and Middle America north of the northern boundary Conexas, 345 p. of Venezuela and Colombia. U.S. Bur. Fish., Rep. U.S. BREDER, C. M., JR., AND D. E. ROSEN. Comm. Fish. 1928. Part II. Append. 10,670 p. (Doc. 1055.) 1966. Modes of reproduction in fishes. Natural History LANHAM,U. Press, Garden City, N.Y., 941 p. 1962. The fishes. Columbia Univ. Press, N.Y., 116 p. CANNON, R. SCHREIBER, J. F., JR. 1966. The Sea of Cortez. Lane Magazine and Book Co., 1969. Changes in Colorado River flow. In D. A. Thomson et Menlo Park, Calif., 283 p. a!. (editors), Environmental impact of brine effluents on CHAVEZ,H. Gulf of California, p. 83-87. U.S. Dep. Int., Off. Saline 1973. Descripci6n de los ejemplares juveniles de totoaba, Water, Res. Dev. Prog. Rep. 387. Cynoscion macdonaldi Gilbert. Rev. Soc. Mex. Hist. Nat. SOTOMAYOR, C. 34:293-300. 1970. La totoaba, una especia que se extingue lentamen­ CHUTE,G. R. teo Tee. Pesq. 3:22-25. 1928. The totuliva fishery of the California Gulf. Calif. Fish THOMPSON, R. W. Game 14:275-281. 1968. Tidaillat sedimentation on the Colorado River Delta, 1930. Seen Kow, a regal soup-stock. Calif. Fish Game northwestern Gulf of California. Geol. Soc. Am., Mem. 16:23-35. 107,133 p. CRAIG, J. A. THOMSON, D. A. 1926. A new fishery in Mexico. Calif. Fish Game 12:166-169. 1969. The commercial fisheries industry. In D. A. Thomson CROCKER, R. S. et a!. (editors), Environmental impact of brine effluents on 1932. The white sea-bass and related species that are sold in Gulf of California, p. 100-103. U.S. Dep. Int., Off. Saline California fish markets. Calif. Fish Game 18:318-327. Water, Res. Dev. Prog. Rep. 387. FITCH,J. E. U.S. GEOLOGICAL SURVEY. 1949. Mexican corbina and totuava. In The commercial fish 1954. Compilation of records of surface waters of the United catch of California for the year 1947 with an historical States through September 1950. Part 9. Colorado River review 1916-1947, p. 83-84. Calif. Dep. Fish Game, Fish Basin. [U.S.] Geol. Surv. Water-Supply Pap. 1313, 749 p. Bull. 74. 1964. Compilation of records of surface waters of the United GABRIELSON, I. N., AND F. R. LAMONTE. States, October 1950 to September 1960. Part 9. Colorado 1954. The fisherman's encyclopedia. Stackpole Co., Harris­ River Basin. [U.S.] Geo!. Surv. Water-Supply Pap. 1733, burg, Pa., 730 p. 586 p. GAUSE, C. I. 1970. Surface water supply of the United States 1961­ 1969. A fish threatened. Underwater Nat. 6:28-31. 65. Part 9. Colorado River Basin. Vo!. 3. Lower Colorado GILBERT, C. H. River Basin. [U.S.] Geol. Surv. Water-Supply Pap. 1926, 1891. Scientific results of the explorations by the U.S. Fish 571 p.

544