RAPID REPRODUCTIVE ANALYSIS AND LENGTH – WEIGHT RELATIONS OF THREE REEF FISHES ( ACTINOPTERYGII : PERCIFORMES AND TETRAODONTIFORMES ) FROM A REMOTE SITE IN PAPUA NEW GUINEA

We present length–weight relations (LWR) and use rapid, low-cost histological methods to describe the reproductive biology of three reef fi shes from a remote area in Papua New Guinea: the striped monocle bream, Scolopsis lineata Quoy et Gaimard, 1824; the Indian goatfi sh, Parupeneus indicus (Shaw, 1803); and the blackbelly triggerfi sh, Rhinecanthus verrucosus (Linnaeus, 1758). The LWR for Scolopsis lineata is W = 0.0191FL3.02. Male L50 is 11.2 and female L50 is 13.5 cm FL. Overall sex ratio is female biased; however, for mature individuals, females dominate size-classes from 13 through 15 cm and males dominate smaller and larger size classes. This species is a batch-spawning, protogynous hermaphrodite. The LWR for Parupeneus indicus is W = 0.00904FL3.25. Male Lm is 14.7 and female L50 is 18.4 cm FL. Overall sex ratio is 1 : 1; however, for mature individuals, females dominate size-classes from 17 through 22 cm and males dominate smaller and larger size classes. This species is a batch-spawning gonochore. BF = 4.62FL2.95. The LWR for Rhinecanthus verrucosus is W = 0.0257TL2.97. Male L50 is 13.0 and female L50 is 14.1 cm TL. Overall sex ratio is female biased; however, mature individuals are male biased ≥ 16.1 cm TL and, effectively, exclusively male ≥ 17.6 cm. This species is a batch-spawning gonochore. BF = 3.35 ∙ 10–19TL19.4. * Correspondence: Dr. Ken Longenecker, Department of Natural Science, Bishop Museum, 1525 Bernice Street, Honolulu, HI, 96817, USA, phone: +1 808 847 8273, fax +1 808 847 8252, e-mail: (KL) klongenecker@bishopmuseum.org, (RL) langston@hawaii.edu, (HB) holly@bishopmuseum.org, (MM) mara@bishopmuseum.org.

To address the large knowledge gap in reproductive information for exploited fi shes (Froese and Binohlan 2000), we initiated a research program focused on the reproduction of Pacifi c coral-reef fi shes.We used recently developed methods for rapid, low-cost, on-site, histologybased reproductive analysis that requires minimal research infrastructure (Longenecker et al. 2013a(Longenecker et al. , 2013b(Longenecker et al. , 2014)).These methods use standard techniques (e.g., plastic embedding) that have been modifi ed such that work can be conducted in remote fi eld settings without electrical service.With these methods, reproductive information can be generated quickly and at low cost, thus eliminating the perceived impediments to broad-scale reproductive analysis of coral-reef fi shes (Roberts andPolunin 1993, Johannes 1998).However, in the interest of rapidly increasing the amount of available reproductive information, we analyze only one gonad subsample from the minimum number of specimens necessary to generate estimates of reproductive parameters.Therefore, results of our rapid reproductive analysis should be viewed as preliminary.
Our research program consists of annual, two-week fi eld trips to the subsistence-fi shing village of Lababia, Papua New Guinea (PNG), which holds traditional tenure over the waters of the Kamiali Wildlife Management Area (KWMA hereafter) (Fig. 1).Each year, village residents use the following criteria to choose four study species: • All are an important part of village fi sh catch; • Residents are interested in learning more about each species; • Published reproductive information is lacking or incomplete; • Each species is distinctive enough that the chance of misidentifi cation is low.
Village residents, through regular fi shing activities from 17 February-17 March 2014, caught an adequate number of specimens for the reproductive analysis of three species.These were the striped monocle bream, Scolopsis lineata Quoy et Gaimard, 1824; the Indian goatfi sh, Parupeneus indicus (Shaw, 1803); and the blackbelly triggerfi sh, Rhinecanthus verrucosus (Linnaeus, 1758).All three species occur in the Indo-West Pacifi c.The range of Scolopsis lineata stretches from the Cocos (Keeling) Islands to the Marshall Islands and Vanuatu, and from the Ryukyu Islands to northwestern Australia (Russell 1990).The range of Parupeneus indicus extends from the east coast of Africa to Samoa; in the Indian Ocean from the southern Arabian Peninsula to South Africa and in the western Pacifi c from southern Japan to the southern Great Barrier Reef (Randall 2001).The range of Rhinecanthus verrucosus spreads from the Chagos Archipelago to the Solomon Islands, and from southern Japan to Vanuatu (Matsuura 2001).Outside of KWMA, P. indicus is a highly esteemed food fi sh (Randall 2001), whereas S. lineata appears occasionally in small numbers in local markets, although there is no major fi shery for the species (Russell 1990), and R. verrucosus is described as of minor importance to fi sheries (Matsuura 2001).
Specimens were caught by hook-and-line, spear, and Derris fi shing (based on rotenone poisoning).Fish lengths, weights, and gonad samples were obtained within a few hours of being caught.All collections were made with the approval of relevant local and national authorities.We followed the guidance of Froese et al. (2011) to determine length-weight relations for fresh-caught specimens and, with the following modifi cations, used methods described in Longenecker et al. (2013b) to describe size-at-maturity, reproductive mode, sex-ratios, and length-batch fecundity relations.For batch fecundity analysis, we liberated oocytes from the stroma by vigorous shaking rather than using an ultrasonic cleaner.Length-batch fecundity relations were determined using linear regression analysis of log-transformed data rather than curvilinear regression analysis of raw data.Results are summarized in Table 1 and presented more fully in separate sections for each species, below.Scolopsis lineata.Total body weight (W) in g is an approximately cubic function of fork length (FL) in cm (Table 1, Fig. 2A).The 95% CI of regression parameters a and b are 0.0104-0.0351and 2.80-3.25,respectively (r 2 = 0.870, n = 105, FL range: 12.2-16.4,W range: 40-90).ANCOVA did not detect a signifi cant sex-based difference in the length-weight relation for this population (F = 2.26, DF = 1, P = 0.137).
We histologically examined gonads of two (2) undifferentiated, 42 male, and 72 female Scolopsis lineata.Ovaries of mature females contained several discrete stages of oocytes, indicating group-synchronous oocyte development (Wallace and Selman 1981).We therefore classify S. lineata as a batch spawner.We grouped males into 2-cm and females into 1-cm size classes to estimate size at maturity.Sexual differentiation occurred around 10 cm FL.Ovaries contained vitellogenic oocytes in females as small as 11.8 cm FL.Immature females (range 10.7-15.8)were scattered throughout the size range of mature females (range 13.2-15.9).We estimate female L 50 at 13.5 cm FL (Fig. 2B).The smallest male with spermiated testes was 11.8 cm FL.For males, the size at which 50% of individuals are mature (L 50 ) was 11.2 cm FL (Fig. 2B).This estimate assumes one-half of undifferentiated specimens were male.All males > 14.1 cm were mature.
We saw several suggestions of sex change in Scolopsis lineata.The mean length of all male S. lineata was signifi cantly larger than that of all females (t = -2.63,DF = 69, P = 0.010).Additionally, the mean length of functional males was signifi cantly larger than that of functional females (t = -3.09,DF = 56, P = 0.003).A total of 21 fi sh (18.1% of all histologically examined specimens) had presumptive ovotestes (Fig. 2C).Twenty of these were dominated by primary-growth (n = 10) or vitellogenic (n = 10) oocytes and had clusters of inactive spermatocysts within the gonad wall or along the periphery of the gonad.Conversely, the gonad of one functional male contained remnants of primary growth oocytes in its periphery.Sex change is common among nemipterids; when present, transition may be pre-or post-maturational (Russell 1990).In species with pre-maturational sex change, juvenile (i.e., inactive) gonads contain readily identifi able ovarian and testicular tissue.In contrast, the majority of the inactive S. lineata gonads were entirely ovarian (20) or testicular (4)  1, χ 2 = 7.89, DF = 1, P = 0.005).Considering only mature individuals, the observed sex ratio was not signifi cantly different from 1 : 1 (χ 2 = 0.325, DF = 1, P = 0.569).However, we did see size-specifi c sex ratios in the mature size classes; the sex ratio of mature individuals varied predictably with length.Smaller size classes are male biased, switch to a female-biased state as length increases, and the largest size classes are male biased (Fig. 2D).An equation describing the percent of mature females (%♀), throughout the size range of mature specimens (Table 1, r 2 = 0.854), predicts that the population is female biased between 13.1 and 14.6 cm FL, but male biased at smaller and larger sizes, and nearly exclusively male near minimum and maximum lengths of mature specimens.Similar patterns, with larger size classes being male-dominated, have been reported in many nemipterids (Young and Martin 1985, Russell 1990, Lau and Sadovy 2001, Mant et al. 2006, Boaden and Kingsford 2013).
We frequently misclassifi ed the status of ovaries during macroscopic examination (incorrectly judging that there were no late-vitellogenesis oocytes).This error left us with too few (n = 3) subsamples for batch fecundity analysis.Parupeneus indicus.Overall, total body weight (W) in g is an approximately cubic function of FL in cm (Table 1, Fig. 3A).The 95% CI of regression parameters a and b are 0.00724-0.0113and 3.17-3.32,respectively (r 2 = 0.982, n = 120, FL range: 13.0-29.3,W range: 40-490).ANCOVA did not detect a signifi cant sex-based difference in the length-weight relation for this population (F = 0.60, DF = 1, P = 0.442).
We histologically examined gonads of one (1) undifferentiated, 47 male, and 52 female Parupeneus indicus.Ovaries of mature females contained several discrete stages of oocytes, indicating group-synchronous oocyte development (Wallace and Selman 1981).We therefore classify P. indicus as a batch spawner.Sexual differentiation occurred around 12 cm FL.Ovaries contained vitellogenic oocytes in females as small as 13.7 cm FL.We estimate female L 50 at 18.4 cm FL (Fig. 3B).The smallest male with spermiated testes was 14.7 cm FL.All males from 24.0 to 29.3 cm FL were mature.Because all size classes contained more than 50% mature males, we could not reliably estimate male L 50 .
We saw a sex-based bimodal size distribution in Parupeneus indicus.A t-test indicated mean length of males is signifi cantly greater than that of females (t = -2.87,DF = 97, P = 0.005).There was no other evidence of sex change in P. indicus.We did not see a central membrane-lined lumen in testes, nor did any gonad contain a mixture of ovarian and spermatogenic tissue.In agreement with studies of congeners (Longenecker andLangston 2008, Longenecker et al. 2011), we classify P. indicus as a gonochore.L m = minimum mature length (the smallest mature individual observed during the presently reported study), L 50 = length at 50% maturity (50% of all individuals at this length are expected to be mature).
Overall sex ratio in this Parupeneus indicus population is not signifi cantly different from 1 : 1 (Table 1, χ 2 = 0.253, DF = 1, P = 0.615), nor does the sex ratio of mature individuals differ from 1 : 1 (χ 2 = 0.545, DF = 1, P = 0.460).However, we did see size-specifi c sex ratios in the mature size classes; the sex ratio of mature individuals varied predictably with length.Smaller size classes were male biased, switch to a female-biased state as length increased, and the largest size classes were male biased (Fig. 3C).An equation describing the percent of mature females (% ♀), throughout the size range of mature specimens (Table 1, r 2 = 0.94), predicts that the population was female biased between 17.4 and 21.7 cm FL, but male biased at smaller and larger sizes, and approximately 95% male near the maximum length of mature specimens.Similar patterns, with larger size classes being male-dominated, have been reported for congeners (Anand and Pillai 2002, Longenecker et al. 2011, Pavlov et al. 2011, Leba unpublished * ).

Rhinecanthus verrucosus. Overall, total body weight (W) in g is an approximately cubic function of TL in cm
ANCOVA did not detect a signifi cant sex-based difference in the length-weight relation for this population (F = 0.07, DF = 1, P = 0.793).
We histologically examined gonads of 90 female, and 29 male Rhinecanthus verrucosus.Ovaries of mature females contained several discrete stages of oocytes, indicating group-synchronous oocyte development (Wallace and Selman 1981).We therefore classify R. verrucosus as a batch spawner.Ovaries contained vitellogenic oocytes in females as small as 12.9 cm TL, however inactive females occurred in all size classes we examined.We estimate female L 50 at 14.1 cm TL (Fig. 4B).This estimate ignores the largest size class, represented by a single inactive female.The smallest male with spermiated testes was 13.8 cm TL.All but three larger specimens (within the 18 cm size class) were mature.For males, the size at which 50% of individuals are mature (L 50 ) is 13.0 cm TL (Fig. 4B).All males ≥ 19.4 cm were mature (n = 2).
We saw a sex-based bimodal size distribution in Rhinecanthus verrucosus.A t-test indicated mean length of males is signifi cantly greater than that of females (t =  -4.68,DF = 32, P < 0.0001).There was no other evidence of sex change in R. verrucosus.We did not see a central membrane-lined lumen in testes, nor did any gonad contain a mixture of ovarian and spermatogenic tissue.We classify R. verrucosus as a gonochore.
Overall sex ratio in this Rhinecanthus verrucosus population is female biased (Table 1, χ 2 = 31.3,DF = 1, P < 0.0001).Considering only mature individuals, the observed sex ratio is also female biased (χ 2 = 5.06, DF = 1, P = 0.0244).However, we did see size-specifi c sex ratios in the mature size classes; the sex ratio of mature individuals varied predictably with length (Fig. 4C).An equation describing the percent of mature females (% ♀), throughout the size range of mature specimens (Table 1, r 2 = 0.99), predicts that the population becomes male biased ≥ 16.1 cm TL.Moreover, the percentage of females was < 0.5%, or effectively zero ≥ 17.6 cm TL.
Batch fecundity (BF) is an unusually high exponential function of fi sh length (over a very restricted range of 1.5 cm) for Rhinecanthus verrucosus (Table 1, Fig. 4D, r 2 = 0.538).This may be related to the extremely small scaling factor; the two appear to be inversely related when fecundity is modeled as a power function of length (Ebisawa 2006).
We reiterate that the results of rapid reproductive analysis should be considered preliminary.For instance, to confi rm our hypothesis that Scolopsis lineata is a protogynous hermaphrodite we would need to examine a greater number of specimens and section the entire gonad serially rather than just obtaining a few (usually 16-20) sections for each specimen, as was done for this study.However, we think our results are a useful contribution to the currently insuffi cient body of knowledge about reproduction in Pacifi c coral-reef fi shes.
For instance, females of all species in this study become less abundant as size class increases, representing between fi ve and, effectively, zero percent of individuals in the largest size class.The impact of these size-specifi c sex ratios on estimates of population-level egg production is currently under-recognized (but see Longenecker et al. 2014).It is a common assumption that large fi sh are disproportionately responsible for population-level reproductive output because large females produce many more eggs than small females (see Roberts and Polunin 1993, Allison et al. 1998, Halpern 2003, Froese 2004, Birkeland and Dayton 2005, Sale et al. 2005).However, the assumption may not hold if females are rare in the largest size classes and cannot hold if females are absent.The curves in Fig. 5A show the relation between length and expected individual batch fecundity with observed overall sex ratios and with observed size-specifi c sex ratios.Expected individual batch fecundity at a given A length is the product of the result of the length-fecundity relation and the proportion of females at that length.Thus the curves represent the expected batch fecundity of any individual in the population, not just females.
For both examples, the small chance of an individual being a female eventually overwhelms increases in batch fecundity such that expected egg production per individual peaks well below maximum observed length (Fig. 5, solid curves).More importantly, failing to account for sizespecifi c sex ratios at the largest observed specimen size leads to a 1.4 million percent overestimate (3 013 220 vs. 21 eggs) in expected individual batch fecundity for Rhinecanthus verrucosus.Similar results should not be surprising for Scolopsis lineata, which appears to be protogynous.However, the consequence of these patterns in gonochoristic species such as Parupeneus indicus and R. verrucosus, particularly with the latter's female-biased overall sex ratio, cannot be overstated.
Despite the value of our results, life-history-based management and conservation requires more than reproductive analysis.As such, aging studies would be a valuable complement to the above results.

Fig. 2 .
Fig. 2. Scolopsis lineata from the Kamiali Wildlife Management Area, Papua New Guinea: Length-weight relation (A); L 50 for males (open circles, dashed line) and females (closed circles, solid line), undifferentiated individuals are represented by triangles (B); An ovotestis from a 15.5 cm FL individual, arrows indicate clusters of inactive spermatocysts within the gonad wall, scale bar = 100 μm (C); size-specifi c sex ratios for mature individuals (D)

Fig. 3 .
Fig. 3. Parupeneus indicus from the Kamiali Wildlife Management Area, Papua New Guinea: Length-weight relation (A); L 50 for females (closed circles, solid line), males and undifferentiated individuals are represented by open circles and triangles, respectively (B); Size-specifi c sex ratios for mature individuals (C); Length-fecundity relation (D)

Fig. 4 .
Fig. 4. Rhinecanthus verrucosus from the Kamiali Wildlife Management Area, Papua New Guinea: Length-weight relation (A); L 50 for males (open circles, dashed line) and females (closed circles, solid line), a single, inactive individual in the largest female size class is represented by a triangle (B); Size-specifi c sex ratios for mature individuals (C); Lengthfecundity relation (D)

Table 1
Summary of length-weight relations and reproductive information for striped monocle bream, Scolopsis lineata, Indian goatfi sh, Parupeneus indicus, and blackbelly triggerfi sh, Rhinecanthus verrucosus, from the Kamiali Wildlife Management Area, Papua New Guinea