Relationships between Otolith Size and Body Size for Hawaiian Reef Fishes1 Ken Longenecker2 Abstract: Estimating body size of fishes from remains recovered from pisci- vores, archaeological sites, and sedimentary deposits is desirable but rarely ac- complished because the relationships between the size of a fish and its durable anatomical structures are largely unknown. Regression equations to predict the size or weight of 41 common Hawaiian reef fishes from sagittae (saccular oto- liths) are presented. Data are also grouped into higher taxa to permit size pre- dictions when otoliths cannot be assigned to species. Animal remains are frequently used to re- the only identifiable remains found in geolog- construct faunal assemblages in dietary analy- ical strata (Rivaton and Bourret 1999). Oto- sis, archaeology, geology, and paleontology. liths are common fossils throughout broad This work on fishes is difficult because bones geographic and stratigraphic ranges (Hecht and scales can be hard to identify, their num- 1990), may be the only evidence that a fish bers may vary among individuals, and the re- species was present at archaeological sites lationship between the size of a bone or scale (Weisler 2002), and are often the only identi- and the size of the fish that produced it may fiable fish remains found in the stomachs or be unknown. feces of predators (Hecht 1990). Finally, the Otoliths, particularly sagittae or saccular size of otoliths can be used to estimate fish otoliths, may be used to circumvent these size, and a fair number of these relationships problems because they are taxonomically dis- have been described for fishes worldwide tinct and the number per individual does not (Frost and Lowry 1981, Echevierria 1987, vary. Further, they are harder and more dura- Gamboa 1991, Plo¨tz et al. 1991, Smale et al. ble than skeletal components and are often 1995, Granadeiro and Silva 2000, Harvey et al. 2000, Mikkelsen et al. 2002, Naya et al. 2002, Waessle et al. 2003). No large-scale analysis of otolith-fish size 1 This paper is funded in part by a grant/cooperative agreement from the National Oceanic and Atmospheric relationships has been conducted for Hawai- Administration, Project no. R/FM-8, which is sponsored ian fishes. With 1,250 species of which by the University of Hawai‘i Sea Grant College Program, 22.3% of coastal species are endemic (Mundy SOEST, under Institutional Grants No. NA86RG0041 2005), this information is needed for research and NA16RG2254 from NOAA Office of Sea Grant, Department of Commerce. The views expressed herein on resource use by early Hawaiians, dietary are those of the author and do not necessarily reflect analysis of piscivores, and reconstruction of the views of NOAA or any of its subagencies. ancient marine environments. Here I present UNIHISEAGRANT-JC-02-32. NOAA Fisheries, Pa- analyses of the relationship between otolith cific Islands Fisheries Science Center supported prepara- length and fish length or weight. These re- tion of the manuscript. This is Hawai‘i Institute of Marine Biology contribution no. 1297 and contribution sults complement a guide to the identification 2007-014 of the Hawai‘i Biological Survey. Manuscript of Hawaiian fish otoliths (Dye and Longe- accepted 28 November 2007. necker 2004). 2 Hawai‘i Institute of Marine Biology, P.O. Box 1346, Ka¯ne‘ohe, Hawai‘i 96744. Current address: Bishop Mu- seum, 1525 Bernice Street, Honolulu, Hawai‘i 96817 materials and methods (e-mail: [email protected]). Fishes were collected from May 2000 Pacific Science (2008), vol. 62, no. 4:533–539 through September 2002 from the forereef : 2008 by University of Hawai‘i Press of Ka¯ne‘ohe Bay. Collecting area boundaries All rights reserved were the 5.5 m isobath shoreward (parallel to 533 534 PACIFIC SCIENCE . October 2008 the barrier reef ), the 30.5 m isobath seaward annotation to the generic name. A significant (the bottom of a drop-off at the mouth of regression could not be constructed for either the bay), and boating channels laterally. All species alone; however, a genus-level rela- specimens were identified, then standard tionship was significant. length (in millimeters) and total body weight This phylogenetic grouping serves three (in grams) were measured. Sagittae were re- purposes. First, it can be used to predict the moved and stored dry in tissue culture plates, size of a fish from otoliths larger or smaller with one species per plate and one individual than those used in the analyses. Although per well. One otolith per fish was haphazardly linear relationships with high coefficients chosen (or the lone otolith if one was broken of determination, such as many of those in or missing) and measured with digital calipers Appendix 1, might reasonably be used for along the rostrum to postrostrum axis (no- extrapolation, doing so with curvilinear rela- menclature of Secor et al. [1992]). tionships (Appendix 2) is likely to provide un- Least squares regression analysis of otolith realistic estimates. Using a relationship for length versus standard length and otolith a higher taxon, based on a wider size range length versus total body weight was per- of individuals, may help avoid the need for formed for each species. Data were also com- extrapolation. Second, because otoliths are bined to search for the same relationships more easily assigned to higher taxa than to within all higher taxa. Equations are pre- species, the groupings provide reasonable sented only for regressions with P <:5. predictions of fish size when a species-level identification of an otolith is not feasible. results and discussion The more-general higher-taxa regressions should be used in such cases. Third, these Standard length can be modeled as a linear groups also provide predictions for species function of otolith length (Appendix 1), and not included in the analysis. With 1,250 fishes total body weight can be predicted using a known from Hawai‘i, the equations presented power function of otolith length (Appen- here represent just a fraction of the work nec- dix 2). For all but two cases, the otolith- essary for detailed reconstruction of fish as- weight relationships could be modeled with semblages. In the interim, these higher-taxa two parameters. The exceptions were three- relationships may suffice for predicting fish parameter power functions. sizes. Both appendixes are organized phyloge- netically. Species included in supraspecific analyses are either annotated at the end of acknowledgments the appendix (for space considerations) or I thank Annmarie Dehn, Ross Langston, and listed parenthetically adjacent to taxon name Joanna Philippoff for donating or helping if no significant relationship was found for collect many of the specimens; and Klahzia that species alone. Species with significant Longenecker for assisting with otolith extrac- otolith-fish size relationships are listed along tions. with their regression equations below the supraspecific taxon heading. For example, re- sults for the length regression of the sub- Literature Cited family Pomacentrinae (Appendix 1) include four species: Stegastes marginatus, which did Dye, T. S., and K. R. Longenecker. 2004. not yield a significant regression on its own, Manual of Hawaiian fish remains identifi- listed parenthetically beside the subfamily cation based on the skeletal reference col- name; Abudefduf abdominalis, which did have lection of Alan C. Ziegler and including a significant relationship, listed below the otoliths. Society for Hawaiian Archaeol- subfamily name; and Plectroglyphidodon impar- ogy, Honolulu. ipennis and P. johnstonianus, indicated by an Echevierria, T. W. 1987. Relationship of oto- Fish Size from Otolith Length . Longenecker 535 lith length to total length in rockfishes Naya, D. E., M. Arim, and R. Vargas. 2002. from northern and central California. Diet of South American fur seals (Arctoce- Fish. Bull. 85:383–387. phalus australis) in Isla de Lobos, Uruguay. Frost, K. J., and L. F. Lowry. 1981. Trophic Mar. Mamm. Sci. 18:734–745. importance of some marine gadids in Plo¨tz, J., W. Ekau, and P. J. H. Reijnders. northern Alaska and their body-otolith 1991. Diet of Weddell seals Leptonychotes size relationships. Fish. Bull. 79:187–192. weddellii at Vestkapp, eastern Weddell Sea Gamboa, D. A. 1991. Otolith size versus (Antarctica), in relation to local food sup- weight and body-length relationships for ply. Mar. Mamm. Sci. 7:136–144. eleven fish species of Baja California, Mex- Rivaton, J., and P. Bourret. 1999. Les otolithes ico. Fish. Bull. 89:701–706. des poissons de l’Indo-Pacifique. Institut de Re- Granadeiro, J. P., and M. A. Silva. 2000. cherche pour le De´veloppement, Noume´a. The use of otoliths and vertebrae in the Secor, D., J. Dean, and E. H. Laban. 1992. identification and size-estimation of fish Otolith removal and preparation for mi- in predator-prey studies. Cybium 24:383– crostructural examination. Pages 19–57 in 393. D. K. Stevenson and S. E. Campana, eds. Harvey, J. T., T. R. Loughlin, M. A. Perez, Otolith microstructure examination and and D. S. Oxman. 2000. Relationship be- analysis. Can. Spec. Publ. Fish. Aquat. Sci. tween fish size and otolith length for 63 117. species of fishes from the eastern North Smale, M. J., G. Watson, and T. Hecht. Pacific Ocean. NOAA Tech. Rep. NMFS 1995. Otolith atlas of southern African 150. marine fishes. J. L. B. Smith Institute of Hecht, T. 1990. Otoliths: An introduction to Ichthyology, Grahamstown, South Africa. their morphology and use in the identifi- Waessle, J. A., C. A. Lasta, and M. Favero. cation of Southern Ocean fishes. Pages 2003. Otolith morphology and body size 64–69 in O. Gon and P. C. Heemstra, eds. relationships for juvenile Scianidae in the Fishes of the Southern Ocean. J. L. B. Rio de la Plata estuary (35–36 S). Sci. Smith Institute of Ichthyology, Grahams- Mar. 67:233–240.