A Comparative Analysis of Growth Zones in Four Calcified Structures ofPacific Blue , Makaim nigricans

Kevin T. Hill, Gregor M. Cailliet, and Richard L. Radtke

ABSTRACT: Sagittae, vertebrae, and anal and Various calcified structures have been utilized dorsal fin spines collected from Pacific blue marlin for age estimation of the Istiophoridae. Dorsal in Kona, Hawaii were evaluated for legibility and spine sections have provided age estimate data interpretability of growth patterns, ease of collec­ for Atlantic , Istiophorus platypterus tion and processing, and the precision of the resul­ (Jolley 1974, 1977; Hedgepeth and Jolley 1983), tant annulus counts for use in estimating age. Atlantic , Tetrapterus albidus, and Sagittae, and anal and dorsal fin spine sections contained growth zones assumed to be annual , nigricans (Prince events and there was a linear relationship between et a1. 1984). Sagittal otoliths have been described age estimates of corresponding samples. Vertebrae as potentially useful structures for ageing most had numerous minute growth increments, but con­ (Radtke 1981, 1983; Radtke and tained no marks' which could be interpreted as an­ Dean 1981; Radtke et al. 1982; Prince et al. 1984; nual. While nonparametric tests revealed no signifi­ Wilson 1984; Cyr 1987). Jolley (1974) described cant difference between age estimates from differ­ numerous circuli in the vertebrae of sailfish; ent hardparts of the same , dorsal and anal however, both scales and vertebrae have now spine counts had the best agreement. Anal and been dismissed as stmctures for age estimation dorsal fin spines were more practical in terms of in billfish (Prince et al. 1984). ease of collection, processing, legibility, and inter­ Age estimation of Pacific blue marlin is still pretation; however, age estimates of spine samples from larger fish required a statistical replacement in the developmental stages, and most data of inner growth zones that were destroyed by have focused on sagittae (Radtke 1981; Wilson matrix expansion. Although more difficult to col­ 1984), with little effort on other skeletal lect and interpret, sagittae provide more detailed structures. The objective of the present study age information. Mean length-at-estimated age was to examine, interpret, and quantitatively data based on anal spine band counts are also pre­ compare growth patterns in the sagitta, sented. vertebrae, and dorsal and anal fin spines of blue marlin from Kona, HI. Each structure was evaluated in terms of ease of collection and Increased knowledge of billfish age and growth processing, legibility of growth patterns, and is essential for sensible management of these the relative precision of the resulting age fisheries. Although there is a paucity of such estimates. information for most billfish species, the West­ ern Pacific Fisheries Management Council was MATERIALS AND METHODS forced to draft a management plan for the Pacific blue marlin, Makai-ra. nigricans, with only Pacific blue marlin were sampled at the cursory data (WPFMC 1985). This lack of infor­ Hawaiian International Billfishing TOl.uonaments mation is due to the many difficulties involved in Augusts 1982 (n = 48), 1983 (n = 113), and with studies of large species (Prince 1984 (n = 98), and at the Kona Gold Jackpot and Pulos 1983), compounded by lack of routine tournament in May 1983 (n = 20>, Kailua-Kona, sampling programs by research agencies in the HI. Additional spine samples were obtained Pacific region. from the Pacific Gamefish Research Foundation (n = 32), the Hawaii Fishing Agency (on = 2), Kevin T. Hill. University of Hawaii. Depart.ment of Zoology. and the National Marine Fisheries Service, 2538 The Mall, Honolulu. HI 96822. Southeast Fisheries Center (a specimen from Gregor M. Cailliet. Moss Landing Marine Laboratories. Kona which was shipped to Miami for taxi­ P.O. Box 450, Moss Landing. CA 95039-0450. Richard L. Radtke. University of Hawaii, Oceanic Biology. dermy). Meristic data collected for each fish Hawaii Institute ofGeophysics. Honolulu, HI 96822. included lower jaw-fork length (LJFL to 0.1 cm),

Manugcript accepted Apl·ill989. 829 Fishery Bulletin. U.S. 87: 829-843. FISHERY BULLETIN: VOL. 87. NO.4, 1989 roundweight (W to 0.5 lb converted to kg), sex, growth bands per millimeter of radius in the core and date of capture (Hill 1986). matrix, and that the first several visible bands were analogous in age to matching bands of the Anal and Dorsal Fin Spine Analyses compiled radius data. Anal and dorsal fin spines were collected and Sagitta Analyses prepared for analysis following modified meth­ ods of Prince et al. (1984) (Hill 1986). The second Sagitta were cleaned, prepared, and examined anal spine and sixth dorsal spine were selected following the methods of Radtke (1983) and Hill for age analysis. These were chosen because (986). Terminology for sagitta orientations is they were the thickest ofthe spine complex, and based on Prince et al. (1986). Sagittal otolith sections taken from spines anterior to these had weight (SW) was measured to the nearest 0.005 more prominent core matrices. mg. Age assignments were based on combined Spine length, defined as the distance from the counts of external growth features present on hole at the center of the condyle base to the spine the sagittae, which included lidges along the tip, was measured to the nearest millimeter. anterior rostrum edge on the ventral plane of Thin cross sections from anal and dorsal fin growth and ridges along the ventral surface of spines were taken at positions marked at 10% the medioventral mId medial planes of growth. and 5% (respectively) of the spine length from Previous studies of Istiophorid sagitta have the condyle hole. Spine sections were examined supplemented age estimates based on external using a compound stereoscope at 63 x and 120 x features with examination of internal growth magnification using either transmitted light or features using thin trmIsverse sections and light reflected light with a black background. The microscopy (Wilson 1984; Prince et al. 1986). focus of the spine was defined as the midpoint of However, Hill (1986) statistically compared age the distance between the anteriOl' and posterior estimates using external and internal growth portions of the spine along the midsagittal plane. features and found no significant difference be­ All growth bands were counted and their radii tween the two methods. Therefore, age esti­ measured with an ocular micrometer along the mates reported in this study were based only plane from the focus of the spine to the widest upon examination of external features. radius of the spine section. Spine radius (anal spine radius = AR; dorsal spine radius = DR) Vertebrae Analyses was defined as the distance from the focus to the outside edge of the spine along the same plane. Caudal vertebrae numbers 22 and 23 were re­ Statistical replacement of early missing anal moved from the area between the posterior por­ and dorsal spine growth bands in larger fish was tion of the second dorsal fin and the base of the accomplished by summmizing band radii statis­ caudal fin. These were the only vertebrae which tics from smaller, younger specimens in which could be removed without loweling the market these early bands were visible. Compiled band value of the fish. Vertebrae were simmered for radius statistics included spine samples which several hours in hot water to remove extraneous had at least the first or second band visible. tissues and then air dried for at least 72 hours. Unpaired t-tests were applied to compare corre­ Vertebral spines and arches were removed, sponding radii between those specimens contain­ anterior and posterior centra separated, cut ing the first and second bmId and to compare longitudinally along the dorsoventral plane, and corresponding band radii between sexes. stored in 95% isopropyl alcohol. Final corrected age estimates were assigned Vertebral cone depth (CD) (as defined by to spine samples missing early bands by compar­ Johnson 1983), referred to in this paper as cen­ ing the radii of the first four visible bands to the hum cone depth was measured to the nem'est means and 95% confidence limits of the compiled 0.05 mm. Growth rings were observed after data. When the radii of at least three successive carefully peeling away the thin layer of carti­ bands of the first four visible bands fitted well laginous tissue which covers the bony face of the within the 95% confidence limits of three or four centra. Centrum length, from focus to outside bands of the compiled data, corresponding ages edge, was divided into approximately 5 mm sec­ were assigned. The use of this technique to pro­ tions and the average number of rings per milli­ vide f'mal age estimates was based upon the as­ meter was calculated for each section by count­ sumption that there was a predictable number of ing three 1 mm portions in each section. Total

830 HILL ET AL.: ANALYSIS OF GROWTH ZONES IN PACIFIC BLUE MARLIN increment number was extrapolated from these Zar 1984). The significance of correlation coeffi­ data. cients (r) of the comparisons was tested using methods outlined by Schefler (1979). These rela­ Assessment ofAgeing Techniques tionships were also tested using a Wilcoxon Signed Ranks test

TABLE 1.-Summary of numbers and size ranges of Pacific Makaira nigricans from which skeletal hardparts and measurements .were collected. WFL = lower jaW-fork length, AR = anal spine radius, DR = dorsal spine radius, SW = sagittal otolith weight, CD = centrum cone depth.

WFL (em) Weight (kg) Hardpart size Hardpart N Min. Max. Mean Min. Max. Mean Min. Max. Anal spines AR (mm): Males 150 114.8 263.1 205.3 19.1 170.3 75.0 3.98 10.10 Females 49 163.8 363.4 260.4 35.2 555.2 180.8 5.66 18.11 Dorsal spines DR (mm): Males 66 177.0 263.1 208.7 46.3 170.3 78.5 7.65 12.46 Females 30 163.8 445.8 275.7 35.2 748.0 227.4 6.58 21.00 Otoliths SW (mm): Males 54 169.5 251.5 206.4 38.6 138.8 75.9 0.415 4.985 Females 45 163.8 445.8 265.9 35.2 748.0 197.4 0.800 10.400 Vertebrae CD (mm): Males 122 169.5 251.5 206.3 38.6 138.8 74.8 17.45 40.48 Females 77 147.0 337.7 261.1 20.9 447.7176.5 18.30 45.99 Total Males 114.8 263.1 205.1 19.1 170.3 74.2 Females 147.0 445.8 264.9 20.9 747.9190.9

831 FISHERY BULLETIN: VOL. 87. NO.4. 1989

LJFL) (Student's t-test; P < 0.05). Not all hard­ cients of determination between AR, DR. CD, parts were collected from all fish. For a break­ and W compared with those with LJFL. In addi­ down of fish size ranges and hardparts collected, tion, females generally had stronger relation­ the reader is referred to Table 1. ships between hardpart size and body size (LJFL and W) when compared to males, the exception to this being the relationship between Hardpart Growth female SW and LJFL. The relationships between AR, DR, and CD, and LJFL were best described by significant (P Growth Increments < 0.001) positive linear equations (Table 2). Coefficient of determination values ranged from Thin cross sections of dorsal and anal fin 2 r = 0.19 for male vertebrae to t = 0.72 for spines revealed a vascularized core matrix and a female anal spines (Fig. 1). The relationships cortical region containing major growth bands between SW and LJFL were exponential, with (Fig. 3). A growth band was composed of alter­ r = 0.41 for males and ,~ = 0.35 for females nations of translucent and opaque rings. Many (Table 2, Fig. 2). Logarithmic equations best growth bands were comprised of smaller rings described the relationships between AR, DR. that were most obvious at the widest lateral CD, and W for both sexes (Table 2). Coefficient pOltion of the spine section (Fig. 4). Since band of determination values were significant (P < radii were measured from the focus outward 0.001), ranging fromr = 0.30 for male verte­ along the widest portion of the spine, excessive brae to 1.2 = 0.87 for female dorsal spines. numbers ofsmaller rings along the counting path Sagitta weight had a linear relationship with W, made delineation of the outside edge of the with 'r = 0.32 for males and ,~ = 0.46 for translucent zone difficult at times, especially in females. exceptionally large spine samples. In such cases, Overall, there were generally higher coeffi- it was necessary to refer to the dorsal and ven-

TABLE 2-Modeled relationships between fish growth (in length and weight) and hardpart growth for male and female Pacific Makaira nigricans. WFL = lower jaW-fork length. W = roundweight, AR = anal spine radius, DR = dorsal spine radius. SW = sagitta weight. CD = centrum cone depth.

Comparison Equation rvalue Anal spine radius vs. LJFL Males AR = -0.5207 + 0.0394(LJFL) 0.52 Females AR = -2.5879 + 0.0528(LJFL) 0.72 Anal spine radius vs. W Males AR = 1.3061 • WO·4079 0.62 Females AR = 1.5300 • WO.3885 0.81 Dorsal spine radius vs. LJFL Males DR = -0.2700 + 0.0474(LJFL) 0.49 Females DR = 0.2440 + 0.0526(LJFL) 0.71 Dorsal spine radius vs. W Males DR = 2.0850 • WO.3513 0.53 Females DR = 2.3157. WO·3499 0.87 Sagittal weight vs. LJFL Males SW = 0.0152·10'0.0103' WFL) 0.41 Females SW = 0.3250·10(0.0038' WFL) 0.35 Sagittal weight vs. W Males SW = -0.642 + 0.0412(W) 0.32 Females SW = 1.1625 + 0.0144(W) 0.46 Vertebral cone depth vs. LJFL Males CD = 2.5616 + 0.1226(LJFL) 0.19 Females CD = 5.2871 + 0.11 06(LJFL) 0.45 Vetebral cone depth vs. W Males CD = 7.509. wO·3031 0.30 Females CD = 10.101 • WO·2377 0.50

832 HILL ET AL.: ANALYSIS OF GROWTH ZONES IN PACIFIC BLUE MARLIN

18 •- Males 00 0- Females E 15 -E -0 :::J 0 :g 12 as a: 0 CD .s..c 9 en 1; c et 6

3 100 150 200 250 300 350 400 Lower Jaw-Fork Length (em)

FIGURE I.-Relationships between lower jaw-fork length (LJFL) and anal spine radius (AR) for'male (n. = 150) and female (n. = 49) Pacific Makaira n.igricans. Males: AR = -0.5207 + 0.0394 (LJFL) r = 0.52; Females: AR = -2.5879 + 0.0528 (LJFL) r = 0.72.

FIGURE 2.-Relationships between lower jaw-fork length (LJFL) and sagitta weight (SW) for male (n. = 54) and female (n. = 45) Pacific Maka.iJY/. nigricans. Males: SW = 0.0152 * 10l0.0103* WFL) 1:? = 0.41; Females: SW = 0.3250 * IOl0.0038 * WFL, r = 0.35.

tral areas of the section where the bands were Sagittae contained external features that sug­ more compressed and clearly delineated. Anal gested changes in structural growth rate over spines were less compressed dorsoventrally than time. Growth of the rostrum was along two dorsal spines and this may have increased clarity planes. Early growth occurred ventrally to the of growth bands (Fig. 3). fish and increments were mainly comprised of

833 FISHERY BULLETIN: VOL. 87, NO.4, 1989

FIGUI(E 3.-Thin tranverse cross section of the 2nd anal spine from a 138.;) kg female Pacific M aka im II irJl'i('all,~ as viewed by a binocular dissecting; microscope at 63x magnification with transmitted light. Al'I'ows indicate translucent edges of growth bands, F == focus; M == cOI'e matrix.

ridges along the anterior rostrum edge (Fig. 5), Increment Counts and Hardpart Growth Between two and four ridges were counted along the anterior edge of the rostrum. Ridges were Statistical replacement of early missing anal also visible along the ventral portion of the and dorsal pine bands in larger !ish was accom­ medial plane of growth (Fig, 6). Most sagittae plished by ~ummarizing band radii statistics had an excess of calcium carbonate which hin­ from smaller fish in which these early incre­ dered ridge quantification to varying degrees ments were visible (Fig, 4). Twenty-one I ercent (Fig. 6). Several sagittae were difficult to inter­ of anal spines and 24% of dOl'sal spines had at pret owing to the mottled appearance of the least the first and second assumed annulus visi­ rostrum face. ble, There was no significant difference (P < Vertebrae contained numerous minute (0.05-0.1 mm) concentric gro\ovth increments which were topographical features on the centrum face (Fig. 7). There were no prominent FIGlIHI'; 5.-Right sagitta from a 35.2 kg; female Pacific Makaim lIi,ql'icCIIIS. Arrows indicate prominences quant.i­ 3-dimensional features 01' changes in ring den­ fied 1'01' estimation of early lateral rostrum g'J'll\l'th. R == sity which might be indicative of annular rostnlln; A == antirostrum; C == cOl'e I'egion: p == postel'ior; a events. == anterior,

834 HILL ET AL.: ANALYSIS OF GROWTH ZONES IN PACIFIC BLUE MARL!

FIGUHE 4.-Thin tranverse cross section of the 6th dor:;al spine from a 52.4 kg male Pacific Makaim /liyrica /IS as viewed by a binocular dissecting microscope at 63x magnification with reilected light and a black back­ ground. F = focus. Arrow indicates area of multiple rings within a single growth band incl·cment.

835 FlSHEln BULLETIN: YOLo 87, NO.4, 1989

FIGURE G.-Left sagitta from a 80. J kg male Pacific Makaira lIigricalls. Arrows indicate rostral ridges quantified for age estimation. R = rostrum; A = anti rostrum; C = core region; 0 = calcium carbonate overburden obscuring ridges.

0.01) between corresponding annuli for these two sets of data for either anal or dorsal spines; therefore, these two data sets were combined. TABLE 3-Modeled relationships between hardpart growth and There was, however, a significant difference in increment counts for Pacific Makaira nigricans. LJFL = corresponding anal and dorsal spine band radii lower jaw-fork length, W = roundweight, AR = anal spine between males and females from the sixth band radius, DR = dorsal spine radius, SW = sagitta weight, CD = centrum cone depth, AC = corrected anal spine band counts, outward; therefore data were separated by sex DC = corrected dorsal spine band counts, SC = sagittal ridge (Fig. 8A, B). counts, VC = vertebral increment counts. Increment counts (sagitta ridges, corrected 2 spine bands, and vertebral increments) in­ Comparison Equation r value creased with hardpart growth for each of the Anal fin spines 1 four structures considered. This relationship Males AC = 0.4521 • AR .4316 0.62 was logarithmic when increment counts were Females AC = 0.5878 • AR1.2729 0.78 compared to AR, DR, and SW, and lineal' when Dorsal fin spines DC = 0.9353 • DR 1.0044 0.39 compared to CD for both sexes (Table 3). Coeffi­ Males Females DC = 0.5380 • DR1.2227 0.83 cients of determination were higher for females Sagittal otoliths in all cases, and ranged from '1.2 = 0.39 for male Males SW = 7.7206 • SW02941 0.41 dorsal spines to )'2 = 0.83 for female dorsal Females SC = 7.1512' SW°.4095 0.42 spines. Vertebrae Males VC = 86.922 + 13.742(CD) 0.56 Females VC = 59.115 + 15.311(CD) 0.69 Hardpart Comparisons There was a positive lineal' relation hip be­ tween estimated counts of corrected growth in-

836 HILL ET AL.: ANALYSIS OF GROWTH 20 ES IN PACIFIC BLUE MARLIN

F'II:UHI'; 7.-Longitudinal cross :,;ecLion or the anterior centrum I'rom the 23rd vertebrae I'rom a 50.3 kg male Pacific Mnkoim l/iYI·icol/:;. a = antel'jor end; d = dOI'sal; v = ventral; I' = centrum focus.

crements in anal and dorsal spine sections and dorsal spine counts. The greatest deviation in sagitta from the same blue marlin (Fig. 9A-C). counts between corresponding hardparts was 9, The y-intercepts of these relationships were not where the anal spine count was 7 and the signiJicantly different from zero, and the slopes sagittal increment count was 16. Wilcoxon did not difIer from parity (slope = 1; P < 0.05). Signed Rank tests revealed no statistical differ­ Correlation coefficients ranged from)' = 0.84 (P ence between counts of these three structures 0.001) for anal spine and sagitta counts to )' = (P < 0.01). 0.95 (P < 0.001) for the comparison of anal and

837 FISHERY BULLETIN: VOL. 87, NO.4, 1989

12 + Males • Females 10 E f .§. 8 I Ul f f .2 i 6 a: 'g fa 4 m +i 2 •• A

O+--~-r-----'-"""'-"""T'"--r---r---r--r-----,-~-....., o 1 2 3 4 567 8 9 10 11 12 Anal Spine Band Number

14 + Males • Females 12 I I E 10 E I f -Ul I i .2 8 i a: 6 'g c as m 4 •• 2 B

O+---r-----.----.--....--...... -----.-----.----.--..--....---. o 1 2 3 4 5 6 7 8 9 10 11 Dorsal Spine Band Number

FIGURE S.-Mean (±95% confidence limits) anal spine tAl and dorsal spine

Reader Comparisons Average Percent Error (APE) values ranged from 4.88% error for anal spines counted by TABLE 4-Average percent error (APE) values calculated reader 1 to 9.68% for sagittae counted by reader from triplicate uncorrected readings of 20 randomly select­ 2 (Table 4). Reader 1 had lower APE values ed anal fin spines. dorsal fin spines, and sagittae. (more precise age estimations) than reader 2 for Average percent error results each hardpart. Both readers had lowest APE Hardpart Reader 1 Reader 2 values for anal spines and highest values for sagittae (Table 4). Anal fin spines 4.88% 8.79% Dorsal fin spines 5.50% 9.18% Comparisons of differences (D) in mean age Sagittae 8.15% 9.68% estimates between readers revealed that reader 1 had a tendency to give higher age estimates

838 HILL ET AL.: ANALYSIS OF GROWTH ZONES IN PACIFIC BLUE MARLIN

30 A than reader 2 for both anal (D = 0.85) and dorsal slope =1 (D = 0.55) counts, while reader 2 assigned C 25 higher counts to otoliths (D = 1.5) (Fig. 10). ~ None of these count differences differed signifi­ 8 20 cantly when compared with a Wilcoxon Signed CP Q Ranks test (P < 0.05). The greatest difference '0 a: 15 • between mean counts was 8 for otoliths. Overall, S there was a greater percentage of age estimates :a. 10 within ± 1 and ± 2 years for anal and dorsal enall •• spines than for otoliths. 5 • Mean Length at Estimated Age O~-""'--~-""'--~-""'--" o 5 10 15 20 25 30 Based on estimated ages from anal im spines, Anal Spine Band Count there was a pronounced difference in growth between male and female marlin (Fig. 11). Males 30 B appear to grow to an average length of 202 cm slope =1 LJFL at an estimated age of six years, after _ 25 which growth is determinate. Growth of female c ~ marlin, more variable than males. is steady and 8 20 does not level off as rapidly. CP D) '0 • a: 15 all DISCUSSION 10 :a.- • Growth patterns observed in dorsal and anal ~ fin spine sections were similar to those described 5 by Jolley (1977) for sailfish, Prince et al. (1984) for Atlantic blue and white marlin, and Wilson oL--....._-~-....._-~_....._--J o 5 10 15 20 25 30 (1984) for Pacific blue marlin. Sagittal otolith Dorsal Spine Band Count morphologies were also similar to those from previous studies of Pacific blue marlin (Radtke 1981; Wilson 1984), and rostral ridges were anal­ 30 ogous to those validated in one tag-recaptured c slope =1 C sailfish specimen (Prince et al. 1986). -~ 25 Three basic assumptions of the ageing theory 8 are that 1) the growth of a structure used is '0c 20 proportional to growth of the , 2) the all m • number of growth increments increase with the cCP 15 growth of the structure, and 3) the observed 'Q. en increments follow a discernible time scale e.g., 10 one year of life (Bagenal 1974). In this study, ai growth of each hardpart was, to some degree, eo Q 5 proportional to growth of the animal's length or weight. Increment counts increased with size of oL--....._-~_....._-~_....._-" each hardpart, providing further support for the o 5 1 0 15 20 25 30 use of these structures for age estimation Anal Spine Band Count studies. With the exception of the comparison

FIGURE 9.-Correlations between coresponding sagittal between SW and LJFL, females had higher coef­ ridge counts (SCl, corrected dorsal spine band counts

839 FISHERY BULLETIN: VOL. 87. NO.4, 1989

9 Anal Fin Spines 9 Dorsal Fin Spines 9 Sagittae 8 8 8 .. 7 CD 7 7 -ell 6 6 6 .§ 5 5 5 .. 4 4 4 W- 3 3 3~". CD 2 2 2 Q 1 ---J>~~ ---J>~ ~~ II(c ---J>

FIGURE IO.-Comparison of differences of mean age estimates between readers (I minus 2) for anal spines. dorsal spines, and sagitt.ae.

body size. It was expected that fin spines, much marlin. While the variability of our counts be­ larger in size than otoliths, would have a closer tween otoliths and their corresponding spines relationship to body size. Prince et al. (984) was greater than between corresponding found a similar relationship between female spines, much of this variation was based on and male size and hardpart growth for Atlantic several cases in which the counts of sagittae blue marlin. The relationship between spine were as much as 5-10 counts above or below radius and body size may, therefore, be more those of cOlTesponding spines. Neither the sta­ useful than sagitta weight for back calculation of tistical tests of regression slopes nor the non­ early growth in marlin. While the first two parametric tests were able to detect significant assumptions ofageing theory have been met, the deviations from parity; these three structures most important assumption, that increments ob­ may deposit increments in relation to similar served in spines and otoliths of blue marlin rep­ environmental or growth stimuli. resent one calendar year of life, has yet to be Compalisons of age estimates, within the data proven. and between the data of both readers provided While validation (or, confirmation of the information on the precision of each technique. temporal meaning of a growth increment) was Average percent error values of both reader not within the scope of this study, partial ver­ estimates were lower for fin spines than for ification was achieved. Wilson et a1. (1983) sagittae, which suggests that fin spine estimates defined verification as "the confirmation of a produce a higher degree of precision than sagit­ numerical interpretation", usually used in refer­ tae. Similarly, comparisons of differences in ence to the precision of estimated age. Precision mean age estimates between readers revealed was determined by means of comparisons of age greater variability of age estimates resulted estimates from corresponding fin spines and from sagittae compared to spines. Variability of sagittae as well as by means of measurement of sagittal age estimates may be due in part to error in age estimates within the data and problems involved with calcium overiayering, between the data of both readers. In general, and the overlapping, successive ridges, or multi­ there was good agreement in age estimates ple smaller ridges on these structures. Wilson between sagittae, anal spines, and dorsal spines (984) reported a similar individual variability in from the same fish. Prince et a1. (1984) de­ the general morphology and clarity of growth scribed similar relationships between otoliths features of Atlantic blue marlin otoliths, and it is and dorsal spines of Atlantic blue and white l'easonable to assume that the interpretation of

840 HILL ET AL.: ANALYSIS OF GROWTH ZONES IN PACIFIC BLUE MARLIN

500 g Males --5 400 g) c CD ... 300 ~ ... + 0 U. I + I: 200 :1: •• -;1;:1::1:++ ..,III ... I * + ; 100 ...0

0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 500 Females -E () • .c- 400 g) -c + + ...CD + + 300 II ~... I + + 0 ! :I: I u. I I • I ! 200 + ..,III== ... + CD 100 ...==0

0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Anal Spine Band Count

FIGURE H.-Mean WFL at age for male and female Pacific Makaira n.igricans based on anal spine corrected band counts. Vertical lines terminated by narrow horizontal lines represent 95% confidence intervals (male n = 149; female n = 48). Closed diamond indicates estimated age of largest female sanlpled <748 kg) based upon dorsal spine and sagitta counts.

sagittal growth patterns is more subjective com­ CONCLUSION pared with the interpretation of spines. Based on preliminary age estimates in this 1) Anal fin spines, dorsal fin spines, and sagit­ study, Pacific blue marlin males have a longevity tal otoliths contained growth information which of at least 18 years and females of at least 27 we assumed to be annual in nature and to hold years of age. The largest female reported in this promise for age estimation ofthis species. Incre­ study was estimated to be 22 years of age based mental patterns in caudal vertebrae, possibly upon both dorsal spine and sagitta counts. The related to some other environmental or growth largest male sampled in this study (170.3 kg) was· stimuli, were not useful for age estimation at this estimatd to be only 14 years ofage and the oldest time. male (estimated age 18) was just above mean 2) Anal and dorsal fin spines are simpler to size. Wilson's (1984) study of Pacificblue marlin collect and process than sagittae, and to provide provided similar age estimates and sizes for each more precise (although not necessarily more ac­ sex. curate) age estimates for this species. The prob-

841 FISHERY BULLETIN: VOL. 87, NO.4, 1989 lem of early growth increment destruction by land, 234 p. core matrix expansion can be partially overcome Beamish, R. J., and D. A. Fournier. through application of band radius statistics; 1981. A method for comparing the precision of a set of age determinations. Can. J. Fish. Aquat. Sci. however, this technique may introduce bias to 38:982-983. final age estimates. With the further compilation Cyr. E. C. of band radius statistics and the application of 1987. Age, growth, and reproduction of blue marlin, techniques such as discriminate function analysis MaJra.ira nigri.cans, from South Carolina billfish tour­ fin spine counts may also bring final age esti­ nament collections. M.S. Thesis, Univ. South Caro­ lina, Columbia, 41 p. mates closer to "true" age. Hedgepeth, M. Y., andJ. W. Jolley, Jr. 3) Sagittal otoliths are more difficult to collect 1983. Age and growth of sailfish, Istiophorus platy­ and process and have more variable growth pterus, using cross sections from the fourth dorsal rates and morphologies than fin spines. While spine. In Proceedings of the International Work­ age estimates based on external features of shop on Age Determination of Oceanic Pelagic : Tunas, , and Sharks, p. 131-135. U.S. Dep. sagittae are perhaps more subjective, sagittae Commer., NOAA Tech. Rep. NMFS 8. may provide more detailed information from in­ Hill, K. T. ternal features such as "daily" increments, valu­ 1986. Age and growth of the Pacific blue marlin, able for age estimation of young of the year. Makaim nigricans: A comparison of growth zones in 4) Extensive mark/tag recapture studies are otoliths, vertebrae, and dorsal and anal fin spines. M.S. Thesis, California State Univ., Stanislaus, 107 p. needed to validate the true meaning of the peri­ Johnson, A. G. odicities assumed to be annual. 1983. Comparison of dorsal spines and vertebrae as ageing structures for , E1tthynnus a.llefte/·­ ACKNOWLEDGMENT atus, from the northeast Gulf of Mexico. In Proceed­ ings of the International Workshop on Age Deter­ A number of people and organizations made it mination of Oceanic Pelagic fishes: Tunas, Billfishes, possible to collect samples for this study, includ­ and Sharks. p. 111-116. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 8. ing the fishermen and fishing teams ofthe billfish­ Jolley, J. W., Jr. ing tournaments, the Hawaiian International 1974. On the biology of Florida east coast Atlantic sail­ Billfishing Association, the Pacific Gamefish Re­ fish (lstiophorus platypte1·/UJ). U.S. Dep. Commer., search Foundation, Hawaiian Fish Distributors, NOAA Tech. Rep. NMFS SSRF-675, p. 81-88. Jerry Kinney of Volcano Isle Fish Wholesalers, 1977. The biology and fishery of Atlantic sailfish, Istio­ phor1.ls platyptems, from southeast Florida. Fla. Mar. and the Hawaiian Fishing Agency. Bruce Wel­ Res. Publ. 28, 31 p. den, Lisa Natanson, Carol Hopper, and Allison Prince, E. D.• D. W. Lee, C. A. Wilson, and J. M. Mallory provided much needed assistance in col­ Dean. lecting samples at the tournaments. Eric Ma­ 1984. Progress in estimating age of blue marlin, suda, University of Hawaii, assisted in hard­ Makaira nigricans, and white marlin, Tetrapterus albid1.ls, from the western Atlantic ocean, Caribbean part readings for the APE analyses. Eric Prince Sea, and Gulf of Mexico. Int. Comm. Conserv. Atl. and Dennis Lee of NMFS Southeast Fisheries Tunas, Coli. Vol. Sci. Pap., Madrid 20:435-447. Center, Miami, donated spine samples from the 1986. Longevity and age validation of tag-recaptured largestfemale marlin (748 kg) caught in this study Atlantic sailfish, Istiophorus platypterus, using dorsal and contributed valuable input. This work is a spines and otoliths. Fish. Bull., U.S. 84:493--502. Prince, E. D.. and L. M. Pulos (editors). result of research sponsored in part by NOAA, 1983. Preface. In Proceedings of the International the National Sea Grant College Program, and the Workshop on Age Determination of Oceanic Pelagic Department of Commerce; under grant numbers Fishes: Tunas. Billfishes, and Sharks. U.S. Dep. NA80AA-D-OOI20 and NA8IAA-D-00070, under Commer., NOAA Tech. Rep. NMFS 8. project numbers R/NP-I-llC, R/F-8I, and R/F­ Radtke, R. L. 1981. Age resolution in billfishes (lstiophoridae and 84 through the California Sea Grant College Pro­ Xiphiidae). West. Proc., 61st Annu. Conf. West. gram, under project number PM/M-2ff through Assoc. Fish Wildl. Agencies, p. 5S-74. the University of Hawaii Sea Grant College Pro­ 1983. Istiophorid otoliths: extraction, morphology, and gram; and in part by the California State Re­ possible use as ageing structures. In. Proceedings of sources Agency. The David Packard Foundation the International Workshop on Age Determination of Oceanic Pelagic fishes: Tunas, Billfishes, and Sharks, granted additional travel support. p. 12.'3-130. U.S. Dep. Commer., NOAA Tech. Rep. NMFS8. LITERATURE CITED Radtke, R. L.• M. Collins, and J. M. Dean. 1982. Morphology of the otoliths of the Atlantic blue Bagenal, T. B. (editor). marlin (Makaira nigricans) and their possible use in 1974. The ageing of fish. Unwin Brothers Ltd.• Eng- age determination. Bull. Mar. Sci. .'32:498--503. 842 HILL ET AL.: ANALYSIS OF GROWTH ZONES IN PACIFIC BLUE MARLIN

Radtke. R. L., and J. M. Dean. gion. West. Pac. Reg. Fish. Manage. Counc., April 1981. Morphological features of the otoliths of the sail­ 1985,242 p. fish, Istiophorus pla.typt.erus, and their possible use in Wilson. C. A. age estimation. Fish. Bull.• U.S. 79:360-366. 1984. Age and growth aspects of the life history of Schefler. W. C. billfishes. Ph.D. Thesis, Univ. South Carolina. 1979. Statistics for the biological sciences. Addison­ Columbia. 179 p. Wesley Pub!. Co., Reading, MA, 230 p. Wilson. C. A., E. B. Brothers, J. M. Casselman, C. L. Sokal, R. R. and F. J. Rohlf. Smith, and A. Wild. 1981. Biometry. 2nd ed. W. H. Freeman and Co., 1983. Glossary. In Proceedings of the International N.Y.• 859p. Workshop on Age Determination of Oceanic Pelagic Western Pacific Fisheries Management Council fishes: Tunas, Billfishes, and Sharks. p. 207. U.S. (WPFMC). Dep. Commer., NOAA Tech. Rep. NMFS 8. 1985. Revised draft fishery management plan for the Zar,J. H. fisheries for billflsh and associated species in the U.S. 1984. Biostatistical analysis. Prentice-Hall. Engle­ fishery conservation zone of the western Pacific re- wood Cliffs. NJ, 620 p.

843