Age and Growth of the Leopard Grouper, Mycteroperca rosacea,in the Southern Gulf of California, Mexico1

J. Gabriel Diaz-Uribe,2 Juan F. Elorduy-Garay,3 and Ma. Teresa Gonztilez-Valdovinos+

Abstract: Growth of the leopard grouper, Myeteroperca rosacea (Streets, 1877),was analyzed in its natural habitat. Age determination was based on the readingof otoliths, and the method was validated under three main criteria: (1) pro-portionality, (2) seasonality, and (3) concordance with another method. Otolithgrowth is proportional to organism growth, with a slight degree of allometry,and the otolith registers the growth of the individual, even at advanced ages.The opaque growth zone in the otolith is deposited once a year, between Julyand October. Thus, taken together, one opaque and one hyaline mark representan annual cycle. Back-calculated lengths-at-age agreed reasonably well with ob-served lengths-at-age at the time of capture, considering that back-calculatedlengths represent an exact age (birthday), and observed lengths are taken at anintermediate age between birthdays. Fish length and otolith age data were fittedto the von Bertalanffy growth function by two methods: (1) linear regression(Ford-Walford and Beverton), using transformed data, and (2) nonlinear re-gression, by iteration. Although the nonlinear regression gave a fit with unbiasederror, parameters resulting from linear regressions had a better biologicalmeaning for the species. The resulting parameters were compared with thosereported for other species of the family Serranidae.

THE LEOPARD GROUPER, Mycteroperca rosacea(Streets, 1877), is found in subtropical shal-low coastal waters, in a restricted geographi-cal area from Magdalena Bay (Baja California

1 This study was partially funded by Direcci6n Gen-eral de Investigaci6n Cientifica y Superaci6n Academica-Secretaria de Educaci6n Publica (DIGICSA-SEP) andDirecci6n de Estudios de Posgrado e Investigaci6n(DEPI)-966509 research projects. ].F.E.-G. received fel-lowships from Comisi6n para el Fomento de ActividadesAcademicas (COFAA) and Estimulo al DesempeiioDocente (EDD) at the Instituto Politecnico Nacional.Manuscript accepted 9 August 2000.

2 Instituto de Recursos, Universidad del Mar, AP. 47,Puerto Angel, Oaxaca, 70902, Mexico.

3 Departamento de Pesquerias y Biologia Marina,Centro Interdisciplinario de Ciencias Marinas (CICI-MAR), AP. 592, La Paz, B.C.S., 23000, Mexico (tele-phone: 52 (1) 122 53 44; fax: 52 (1) 122 53 22; E-mail:jelorduy@redipn.ipn.mx).

4 Departamento de Biologia Marina, UniversidadAut6noma de Baja California Sur, AP. 19-B, La Paz,B.C.S., 23080, Mexico.

Pacific Science (2001), vol. 55, no. 2:171-182© 2001 by University of Hawai'i Press.All rights reserved

Sur) to Banderas Bay (Jalisco), including theGulf of California (Thomson et al. 1987,Heemstra 1995). The family Serranidaeforms an important part of the catches fromthe artisanal fishery of Baja California Sur,and the leopard grouper surpasses other spe-cies both in volume and frequency of catches(Rodriguez-Medrano 1990). The biology ofthe leopard grouper is poorly known. Thereare works referring to its description and tax-onomic status (Rosenblatt and Zahuranec1987), distribution and abundance in La PazBay, ReS. (Villavicencio 1983), and feedinghabits (Bermudez-Almada and Garcia-Laguna1985), but previous studies on age and growthfor this species are lacking.

Individual growth rate, age structure, andage of first maturity and recruitment are use-ful parameters for the evaluation of pop-ulations, and the confidence level at whichthey are estimated depends, in large part,on the method used for age determination(Ricker 1979, Sparre and Venema 1992). Thereading of scales and otoliths has been themost frequently used method for age deter-mination of fish. In slow-growing, long-lived

171

172

species, the growth marks are unreadablenear the edge of scales, causing an underesti-mation of age. Otoliths have proven moreuseful for age determination in this kind oforganism (Boehlert 1985, Devries and Frie1996). Considering the long life cycle andslow growth rate of species composing thefamily Serranidae (Manooch 1987), we con-sidered that otoliths would be the most reli-able structure for age and growth analysis ofthe leopard grouper. Even the reading ofotoliths does not ensure a priori that thismethod of age determination is dependable.It must be demonstrated that the marks usedto determine age are related to a specific timeinterval, and that they can be observed formost of the life cycle of the species (Beamishand McFarlane 1983, Casselman 1983). Onlyin this way can the bias in estimation ofgrowth parameters be reduced, or at leastknown, and the consequences in resourcemanagement appreciated (Beamish andMcFarlane 1983).

The von Bertalanffy growth function hasproven to be a good model in studies ofgrowth because of the applicability of its pa-rameters in more complex models describingpopulation dynamics (Sparre and Venema1992). Because of the commercial importanceof the leopard grouper, the estimation ofgrowth parameters is especially valuable be-cause of its relevance to fishery management.In this study we analyze the validity of otolithreading for age determination of Mycteropercarosacea caught in the Bay of La Paz, B.C.S.,and vicinities, and the fitting of data to theVon Bertalanffy growth function is alsopresented.

MATERIALS AND METHODS

From January 1991 to June 1992 individualsof M. rosacea were selected monthly fromboth commercial and experimental catchesmade around Espiritu Santo and CerralvoIslands, located off the western coast of theGulf of California (Figure 1). A maximumof 15 fish per length interval (50 mm), permonth, was randomly selected from thecatches. Each fish was measured in total (TL)and standard (SL) lengths and gutted weight

PACIFIC SCIENCE· April 2001

(GW). Whenever possible, total weight(TW) and sex were also determined. Aftermeasuring, both sagittae otoliths were ex-tracted by an oblique cut at the cranial basis.At the laboratory, otoliths were washed withsoap and tap water and stored dry.

Before age determination, a small sub-sample of otoliths was subjected to differentcombinations of high temperatures (300-400°C) and times (5-30 min) to improve thecontrast between marks (Elorduy-Garay andDiaz-Uribe 1994). The temperature-timecombination with best contrast was chosenand applied to all otoliths considered in thestudy.

Otolith readings and measurements weremade using a dissecting microscope, an ocularmicrometer, and reflected illumination, at10 x total magnification. Reading consistedof the counting of growth marks and thedetermination of edge type (i.e., opaque orhyaline). Each otolith was read twice in-dependently. When both readings wereidentical the otolith was declared readable,and the number of marks (rings) and edgetype were assigned to the corresponding fish.Whenever there were differences, otolithswere submitted to a third reading. If the thirdreading agreed with either of the previoustwo readings it was assigned as definitive; ifnot, the otolith was considered unreadableand discarded.

Three types of morphometric relation-ships were analyzed: TL versus SL, GW ver-sus TW, and TL versus GW. The firsttwo were evaluated using linear regression byleast squares and Student's t-test under thenull hypothesis a = 0, b = 0, and R2 = 0(ex = 0.05). The third relationship was eval-uated by the potential model Y = axb usingnonlinear regression by least squares, with theprogram FISHPARM 3.1 (Prager et al. 1989);in this case, Student's t-test was ran under thenull hypothesis b = 3 and R2 = 0 (ex = 0.05).

The use of otoliths for age determinationof M. rosacea was validated with three basiccriteria (Beamish and McFarlane 1983, Cas-selman 1983):

(1) Proportionality between otolith growthand fish growth was tested by selecting astratified subsample, as a function of fish

Age and Growth of Mycteroperca rosacea . Diaz-Uribe et al. 173

-i~ Espiritu Santo I.

1100 00'

t

San Jose I.

Bay ofLa Paz

.'.:".-,_.r.-. "'.: ".

"::.:: .;.;;:.;:r.....,..~......:-:...:.

., =~ San Fancisco I.

FIGURE 1. Fishing areas from which the samples of M. rosacea were obtained.

length and independent of the month of cap-ture. A maximum of 20 pairs of otoliths perlength interval was selected. The otolithswere measured in long diameter (OLD) andlong radius (OLR). Data on otolith growth

and fish growth were adjusted to a potentialmodel.

(2) Seasonality in the formation of growthmarks was analyzed by the monthly frequencyof otolith edge type, using the total sample.

174

This information was compared with the seasurface temperature in southern La Paz Bay.The Secretaria de Desarrollo Social, Dele-gaci6n RC.S., kindly provided temperaturedata.

(3) Concordance with another method wasanalyzed by the back-calculation of lengths atpast ages, as a reference method. The sub-sample used was the same as that used in theproportionality analysis. At this stage, theradii at each growth mark (annuli) weremeasured in each otolith. Average length-at-age was estimated using the equation:

where Lj is the length of the fish when theotolith radius was Ri; a and b are the parame-ters of the regression of fish length on otolithradius from the proportionality analysis.Back-calculated lengths were compared withobserved lengths at the time of capture.

Individual growth was calculated fittingthe age-length data to the Von Bertalanffygrowth function:

L, = L",[1 - e-k('-to)j

where L, is the average length of the fish atage t; L", is the average maximum length ofthe analyzed stock; k is the growth coefficient;and to is the parameter of origin of thegrowth curve.

Two separate methods were used to derivethe growth function. In the first method(designated LRC) we followed the Ford-Walford procedure to calculate L", and theBeverton procedure to calculate k and to(Ricker 1975, Sparre and Venema 1992). Thesecond method utilized a nonlinear regres-sion analysis (NLR) with the programFISHPARM 3.1 (Prager et al. 1989). Confi-dence intervals (a = 0.05) were calculated foreach parameter and compared with thosecalculated by the LRC method.

RESULTS

In total 769 fish were sampled during thestudy period, except for the months of June,November, and December 1991 whencatches were low, and it was impossible for us

PACIFIC SCIENCE· April 2001

to survey at sea. Individuals in the sampleranged from 283 to 975 mm TL, but 90% ofthe fish were between 350 and 700 mm TL(Figure 2). Sex was determined for 16% ofthe sample, comprising fish 286-808 mm inlength (TL). None of the fish had virginalgonads.

Because of the small dispersion of data, thelinear model fitted to the TL-SL relationshipexplains nearly 98% of the variability(P < 0.01) (Table 1). The GW-TW rela-tionship presented the same behavior, with99% (P < 0.01) of explained variance (Table1). Ordinates at the origin were not sig-nificantly different from zero (P > 0.05) inboth regressions. Therefore, the standardlength of the leopard grouper represents anaverage of 86-88% of the total length, andthe gutted weight represents an average of87-90% of the total weight.

The potential model fitted to the TL-GWrelationship explains 97% of the observedvariability (P < 0.01), so it has an importantpredictive value. The coefficient b = 2.97 isnot significantly different from 3 (P> 0.05),so the leopard grouper exhibits isometricgrowth (Table 1).

Heating the otoliths at 325°C for 25 minwas the combination that rendered best re-sults. Under these conditions, opaque ringsturned yellow, and hyaline ones turned darkbrown. The immersion of these otoliths inglycerin several minutes before the readingsnotably improved the contrast between rings.Nearly 95% of the otoliths were used in agedeterminations. The remaining were consid-ered unreadable, because of lack of agreementin the number of rings (2 %) or lack ofagreement both in the number of rings andedge type (3 %).

Otolith growth is strongly correlated tooverall fish growth. All the correlation co-efficients (r) were significant (P < 0.01), inspite of the fact that data present higher dis-persion than the previous regressions (Table2). In every correlation where TL is involvedthe exponent is significantly greater than 1(P < 0.01), whereas in every correlationwhere GW is involved the exponent is sig-nificantly different from 3 (P < 0.01). Thismeans that there is a certain degree of allom-

Age and Growth of Myeteroperea rosacea . Diaz-Uribe et at. 175

20

15

>.uc 10Q)::Jrr~u..

5

o275 375 475 575 675 775

Length class (mm)875 975

FIGURE 2. Frequency distribution of size classes (TL) for the whole sample of M. rosacea.

etry, in which the fish grows proportionallyfaster than the otolith.

In 1991 the frequency of individuals withopaque otolith edges was low in January andFebruary and highest between July and Oc-tober (Figure 3). Although there is no infor-mation for June, November, and December

in that year, the tendency in the other monthsindicates that individuals with opaque otolithedges predominated (>50%) from June toNovember. The frequency of individuals withopaque otolith edges was again low at thebeginning of 1992. This pattern follows thetemperatures registered in the Bay of La Paz.

TABLE 1

Morphometric Relationships Involving Body Length and Weight in M. rosacea

Parameters

Relationship n a SE (a) b SE (b) R2

TL-SL 762 -10.583 12.998 0.872 4.36 x 10-3 0.981GW-TW 111 5.997 120.312 1.128 8.51 x 10-3 0.994TL-GW 762 1.43 x 10-5 1.58 X 10-6 2.970 1.70 x 10-2 0.973

Note: Total length (TL) - standard length (SL) (function, y = a + bx); gutted weight (GW) - total weight (TW) (function,y = a + bx); total length (TL) - gutted weight (GW) (function, y = ax'). R', coefficient of determination; n, sample size.

176 PACIFIC SCIENCE· April 2001

TABLE 2

Morphometric Relationships between Otolith Features and Body Length and Weight in M. rosacea

Parameters

Relationship a SE (a) b SE (b) R'

OLDvs. TL 10.640 1.416 1.656 0.055 0.8360OLRvs. TL 34.640 3.276 1.639 0.055 0.8357OLDvs.GW 0.088 0.035 4.241 0.159 0.7928OLRvs. GW 1.538 0.452 4.288 0.161 0.7977

Note: All relationships were fitted to the potential model y = ax'. Linear dimensions are expressed in mm and weight in g. OLD,otolith long diameter; OLR, otolith long radius; TL, fish total length; GW, fish gutted weight, n = 187 for all cases.

In general, the temperature increase corre-lates well with the increase of opaque otolithedge development and vice versa. There is noevidence of another season with a high degreeof opaque otolith edge development in thepopulation. Thus, a set of one opaque,and

one hyaline ring represents a year in the lifeof these fish.

The back-calculation of lengths at pastages showed that the readings on otolithswere consistent, except for the age groupsof 12 or more years. In these, the back-

100 30-.~0'-' 28G80 -.z 0w 026 '-'::J \ Wa 60 q \ ~W ::J~ ", 24 ~u.. ,,\W '~ ~C) 40 w

", 0-0 \" 22 :2:w , ' W

W I-::J 20a 204:0-0

0 18JAN MAR MAY JUL SEP NOV JAN MAR MAY~ Opaque edges -+- Temperature

FIGURE 3. Monthly frequency of individuals of M. rosacea exhibiting opaque otolith edges in the total sample of 769fish, plotred with mean monthly sea surface temperature in the Bay of La Paz.

Age and Growth of Myeteroperca rosacea . Diaz-Uribe et at. 177

1000o

- Linear RegressionNon Linear Regression

Observed Lengths ± S.D.• Back-calculated Lengths ± S. D.

o 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22Age (years)

•II IJ

o

200

800

-.EE 600.........~.....ClCQ)

-.oJ

tU 400

~

FIGURE 4. Growth curves (in total length, TL) of the leopard grouper fitted by two methods: NLR (nonlinear re-gression), using FISHPARM 3.1 (Prager et al. 1989) and LRC (linear regression). Back-calculated lengths-at-age(closed circles), and observed lengths-at-age (open circles) are shown.

calculated lengths were, on average, shorterthan for the other age groups (Figure 4). Be-cause of the small number of fish older than11 yr, it is not possible to evaluate if the biasin the averages was due to a random factor orto an intrinsic factor of the otolith (Table 3).

The growth model was fitted to a set ofdata containing the back-calculated lengths-at-age, up to 11 yr, and the observed lengths-at-age (Table 4). Estimated parameters inthe LRC method were different from thosein the NLR method (Table 5). Accordingto the standard error estimated by theNLR method, Loo could vary from 1142 to1310 mm (df = 28; P = 0.05); the Loo = 1083estimated by the LRC method is outside thatrange. The growth coefficient k also showeddifferences between the two estimation

methods. The 95% confidence interval for kestimated by the NLR method is 0.0567 to0.0750 yr-l; the value of k (0.0925 yr-l) esti-mated by the LRC method is outside thisrange. Although the nonlinear regressionanalysis (NLR) explains a larger amount ofthe variance in the data set, the goodness offitshould be evaluated in view of the biologicalinformation of the species (Figure 4).

DISCUSSION

Although sex could be determined only for asmall number of fish, we can state that juve-niles were not present in the sample. Even thesmaller fish in our samples were adults, whichagrees with the reports of Heemstra (1995),who referred to juveniles of M. rosacea as

TABLE 4

Observed Mean Lengths of Individuals of M. rosacea atVarious Ages as Determined by Otolith Annual Marks

Age (yr)

2+3+4+5+6+7+8+9+

10+11+12+13+14+15+16+17+18+21+

TL (mrn)

291336385434477535582631681722747768786808860866875975

SD

7.61613.73025.23231.89930.92437.18039.17926.17932.47220.12423.629

3.53615.556

n

429

14214198

119846724

832211111

smaller than 300 mm TL. The same authorreported a maximum length of 700 mm forthe species, but Thomson et al. (1987) andAllen and Robertson (1994) mentioned thatthis species can reach up to 1000 mm and12.2 kg of weight, although they did notspecify the kind of length and weight. This

TABLE 5

Growth Parameters for M. rosacea Calculated by TwoRegression Methods

Nonlinear LinearRegression Regression

Parameter Average SE Average SE

L", 1226 41.16 1083k 0.06586 0.00447 0.09245to -1.393 0.1798 -0.2446r 2 0.99603 0.98034

Note: TL, total length; SD, standard deviation of meanlength; n, sample size. Note: L., in nun, k in year-I, and to in years.

Age and Growth of Myeteroperca rosacea . Diaz-Uribe et al. 179

divergence is shown in the length distributionfound in our study sample. Although mostindividuals caught are less than 700 mm TL,larger fish are encountered; our study regis-tered lengths to 915 mm TL, close to themaximum reported by Thomson et al. (1987)and Allen and Robertson (1994).

Heating otoliths is a useful technique toimprove contrast and ease the reading of M.rosacea otoliths; each mark acquires a differ-ential coloration.

The allometry phenomenon in the rela-tionship of otolith size to fish size has beenreported for several species, and discussedfrom the physiological point of view (Simkiss1973). The extreme case in which otolithsgrow to a maximum size and then grow onlyin thickness has been documented for severalspecies of soles (Williams and Bedford 1973).In M. rosacea the evidence indicates that thefish grows at a higher rate than the otolith.The value of b = 1.656 indicates that whenthe fish increases its length 100%, the otolithgrows approximately 53%. Other species,such as Ocyurus chrysurus (Manooch andDrennon 1987) and Lutjanus peru (Rocha-Olivares and Gomez-Munoz 1993), have bvalues near 1.6. However, otolith growthdoes not completely stop in these cases. Al-though there are no precise criteria specifyingthe critical level at which allometry wouldadversely affect age determination by exami-nation of otoliths, the values obtained for M.rosacea do not seem to cause problems, be-cause even at the oldest ages, rings could becounted and measured.

Whenever growth marks are depositedperiodically, it is possible to assess this pro-cess using the ratio of edge types (opaque orhyaline) in otoliths (Williams and Bedford1973). The prevalence of opaque edges insummer-autumn and hyaline edges in win-ter-spring represents a defined pattern in M.rosacea otoliths.

We consider that reading of otoliths is aconsistent method for determining age at thetime of capture, because observed lengthswere generally larger than back-calculatedlengths at each age. Comparison was madetaking into account that each method givescomplementary information. That is, average

lengths by age at the time of capture repre-sent the length at an intermediate age be-tween two birthdays, because at the time ofage assignment most otoliths show a fractionof edge, indicating that the birthday hasalready passed. However, back-calculatedlengths at previous ages represent the exactbirthday lengths. Therefore, the consistencyof the two methods was analyzed by the fol-lowing criteria: (1) each back-calculatedlength belongs to an exact age (1, 2, 3, ... ,etc.); (2) each length at the time of capturebelongs to an intermediate age (1.5, 2.5, ... ,etc.). In this way, both data sets intercalatereasonably well between 3 and 11 yr of age,within what could be the individual growthcurve. Because juveniles younger than 3 yrand adults older than 11 yr were scarcelyrepresented in our sample, we consider onlyages within this range to be accurately re-flected by otolith characteristics.

Individual growth is a process that impliesa change in body mass in a determined timeperiod. To express this phenomenon in termsof length, von Bertalanffy developed hismodel under the hypothesis that organismsgrow isometrically. That is, that the organismkeeps the same body form during the growthperiod. Whenever this happens, the changein body mass or respiratory surface due togrowth is proportional to the change inlength cubed or squared, respectively. Pauly(1979) found that many fish species do notcomply with this criterion, so the Von Berta-lanffy growth model (VBGM) represents onlya special case. The value of b = 2.97 in theleopard grouper is not significantly differentfrom 3, therefore the use of the VBGM canbe considered adequate in this case.

Fitting the growth curve by NLR is a sta-tistically powerful method because no trans-formation of the data is required and the,error of the parameters is estimated in anunbiased way (Prager et al. 1989). But this is acondition a priori, and the estimations mustbe evaluated as a function of the biologicalimplications of the adjusted parameters. Sev-eral authors have discussed the close rela-tionship of w;., and k with the biological cycleof species (Pauly 1979, Manooch 1987,Sparre and Venema 1992). Munro and Pauly

180

0.7

0.6

0.5

0.4K

0.3

0.2

0.1

00

PACIFIC SCIENCE· April 2001

181 Epinepheluso Mycteroperca• M. rosaceav Lutjanus

C8I

Y 'VV 0181 V' 0

JIt8lNlR

200 400 600 800 1000 1200 1400 1600 1800Loo

FIGURE 5. Growth performance (

Age and Growth of Mycteroperca rosacea . Diaz-Uribe et at. 181

classified as fully piscivorous (Hobson 1965),but Bermudez-Almada and Garcia-Laguna(1985) found that, besides fish, M. rosacea alsofeeds on crustaceans in a 2: 1 ratio (fish:crustaceans). In spite of utilizing foods withhigh energetic content, the low growth ratesof fish with this feeding regime are due to thelarge energetic cost of searching and chasingprey (Buesa 1987).

ACKNOWLEDGMENTS

We thank the fishermen of El Sargento andLa Ventana for allowing us to sample theircatches. We also are grateful to RichardRadtke and an anonymous reviewer who madeimportant contributions on early drafts of themanuscript to improve its quality and clarity.

Literature Cited

Allen, G. R, and D. R Robertson. 1994.Fishes of the tropical eastern Pacific. Uni-versity of Hawai'i Press, Honolulu.

Beamish, R]., and G. A. McFarlane. 1983.The forgotten requirement for age valida-tion in fisheries biology. Trans. Am. Fish.Soc. 112:735-743.

Bermudez-Almada, B., and G. Garcia-La-guna. 1985. H:ibitos alimenticios de lospeces de las zonas rocosas de la Bahia deLa Paz, Baja California Sur. B.S. thesis,Universidad Nacional Autonoma de Mex-ico, Mexico, D.F. [in Spanish].

BoeWert, G. W. 1985. Using objective crite-ria and multiple regression models for agedetermination in fishes. Fish. Bull. 83:103-117.

Buesa, R J. 1987. Growth rate of tropicaldemersal fishes. Mar. Ecol. Prog. Ser.36:191-199.

Casselman, J. M. 1983. Age and growth as-sessment of fishes from their calcifiedstructures-techniques and tools. In E. D.Prince and L. M. Pulos, eds. Proceedingsof the International Workshop on AgeDetermination of Oceanic Pelagic Fishes:Tunas, Billfishes and Sharks. NOAATech. Rep. NMFS 8:1-17.

Devries, D. R, and R V. Frie. 1996. Deter-mination of age and growth. Pages 483-

512 in B. R Murphy and D. W. Willis,eds. Fisheries techniques, 2nd ed. Ameri-can Fisheries Society, Bethesda, Mary-land.

Elorduy-Garay, J. F., and J. G. Diaz-Uribe.1994. Age validation of Caulolatilus affinis,Gill 1865 (Pisces: Branchiostegidae) fromthe Gulf of California using otoliths. Sci.Mar. 58:277-287.

Heemstra, P. C. 1995. Serranidae. Pages1201-1813 in W. Fisher, F. Krupp, W.Schneider, C. Sommer, K. E. Carpenter,and V. H. Niem, eds. Guia FAO para laidentificacion de especies para los fines dela pesca: Pacifico centro-oriental. Vol. 3.FAO, Rome [in Spanish].

Hobson, E. S. 1965. Diurnal-nocturnal activ-ity of some inshore fishes in the Gulf ofCalifornia. Copeia 3:291-302.

Manooch, C. S., III. 1987. Age and growth ofsnappers and groupers. Pages 329-374 inJ. J. Polovina and S. Ralston, eds. Tropicalsnappers and groupers: Biology and fish-eries management. Westview Press, Boul-der, Colorado.

Manooch, C. S., and C. L. Drennon. 1987.Age and growth of yellow-tail snapper andqueen triggerfish from the U.S. Virgin Is-lands and Puerto Rico. Fish. Res. 6:53-68.

Munro, J. L., and D. Pauly. 1983. A simplemethod for comparing the growth of fishesand invertebrates. Fishbyte, ICLARM1:5-6.

Pauly, D. 1979. Gill size and temperature asgoverning factors in fish growth: A gener-alization of von Bertalanffy's growth for-mula. Verh. Inst. Meereskunde, Univ. Kiel63:1-156.

Prager, M. H., S. B. Saila, and C. W. Reck-siek. 1989. FISHPARM: A microcomputerprogram for parameter estimation of non-linear models in fishery science, 2nd ed.Old Dominion Univ., Oceanogr. Tech.Rep. 87-10: 1-18.

Ricker, W. E. 1975. Computation and inter-pretation of biological statistics of fishpopulations. Fish. Res. Board Can. Bull.191:1-382.

---. 1979. Growth rates and models.Pages 677-743 in W. S. Hoar, D. J. Ran-

182

dall, and J. R. Brett, eds. Fish physiology.Vol. 8. Bioenergetics and growth. Aca-demic Press, Orlando, Florida.

Rocha-Olivares, A., and V. Gomez-Munoz.1993. Validacion del uso de otolitos paradeterminar la edad del huachinango delPacifico Lutjanus peru (Perciformes:Lutjanidae) en la bama de La Paz y aguasadyacentes, B.C.S., Mexico. Cienc. Mar.19:321-331.

Rodriguez-Medrano, M. C. 1990. Composi-cion especifica de la captura artesanal deescama de isla Cerralvo, B.C.S., Mexico.B.S. thesis, Universidad Autonoma de BajaCalifornia Sur, La Paz, B.C.S., Mexico [inSpanish].

Rosenblatt, R. H., and B. J. Zahuranec. 1987.The eastern Pacific groupers of the genusMycteroperca; including a new species.Calif. Fish Game 53:228-245.

Simkiss, K. 1973. Calcium metabolism of

PACIFIC SCIENCE· Apri12001

fish in relation to ageing. Pages 1-12 inT. B. Bagenal, ed. Ageing of fish. UnwinBrothers, Surrey, England.

Sparre, P., and S. C. Venema. 1992. Intro-duction to tropical fish stock assessment.FAO Fish. Tech. Pap. 306/1: 1-376.

Thomson, D. A., T F. Lloyd, and A. N.Kerstitch. 1987. Reef fishes of the Sea ofCortez: The rocky-shore fishes of the Gulfof California. University of Arizona Press,Tucson.

Villavicencio, C. 1983. Distribucion y abun-dancias relativas de la familia Serranidae enla Bahia de La Paz, B.C.S. Mem. VIICongr. Nac. Zool., Jalapa, Ver., Mexico[in Spanish].

Williams, T, and B. C. Bedford. 1973. Theuse of otoliths for age determination.Pages 114-123 in T B. Bagenal, ed. Theageing of fish. Unwin Brothers, Surrey,England.