182 



Abstract — Larval and juvenile rock- 

 fishes iSebastes spp.) are difficult to 

 identify using morphological charac- 

 ters. We developed a key based on 

 sizes of restriction endonuclease frag- 

 ments of the NADH dehydrogenase-3 

 and -4 (ND3/ND4) and 12S and 16S 

 ribosomal RNA (12S/16SI mitochon- 

 drial regions. The key makes use of 

 variation in the ND3/ND4 region. 

 Restriction endonuclease Dde I varia- 

 tion can corroborate identifications, as 

 can 12S/16S variation. The key, based 

 on 71 species, includes most North 

 American taxa, several Asian species, 

 and Sehastolnbus alascanus and Heli- 

 colenus hilgendnrfi that are closely 

 related to rockfishes. Fifty-eight of 

 71 rockfish species in our database 

 can be distinguished unequivocally, 

 using one to five restriction enzymes; 

 identities of the remaining species 

 are narrowed to small groups: 11 S. 

 polyspinis. S. crameri, and S. ciliatus 

 or variabilis (the two species could 

 not be distinguished and were consid- 

 ered as a single species) ; 2) S. chlo- 

 rostictus, S. eos, and S. rosenblatti: 

 3) S. entomelas and S. mystinus; 4) 

 S. emphaeus, S. variegatus, and S. 

 wilsoni; and 5) S. carnatus and S. 

 chrysomelas. 



A key to selected rockfishes iSebastes spp.) 

 based on mitochondrial DNA restriction 

 fragment analysis 



Zhuozhuo Li' 

 Andrew K. Gray' 

 Milton S, Love' 

 Akira Goto^ 

 Takashi Asahida^ 

 Anthony J. Gharrett' 



' Fisheries Division, School of Fisheries and Ocean Sciences 

 University of Alaska Fairbanks 

 11120 Glacier Highway 

 Juneau, Alaska 99801 



Email address (for A J Gharrett, contact author) ffaigia'uafedu 



^Marine Science Institute 

 University of California 

 Santa Barbara, California 93106-6150 



3 Graduate School of Fisheries Sciences 

 Hokkaido University 

 3-1-1 Minato-cho, Hakodate 

 Hokkaido 041-8611, Japan 



'' School of Fisheries Sciences 

 Kitasato University 

 Sanriku, Ofunato 

 Iwate 022-0101, Japan 



Manuscript submitted 11 October 2004 

 to the Scientific Editor's Office. 



Manuscript accepted for publication 

 1 August 2005 by the Scientific Editor. 



Fish. Bull. 104:182-196 (2006). 



More than 100 species of rockfishes 

 (genus Sebastes) are found worldwide 

 (Kendall, 2000), and the majority are 

 distributed along the Pacific coast of 

 North America and in the northwest 

 Pacific from the western Bering Sea 

 south to Japan and Korea (Love et 

 al., 2002). Sixty-five rockfish species 

 are found along the California coast 

 (Moser, 1996). Within the genus, there 

 is a high degree of similarity of the 

 morphological characters among many 

 species. These similarities are in part 

 due to recent divergence but also may 

 have resulted from convergence of con- 

 geners occupying similar habitats. 



Historically, species identification of 

 Sebastes has been based on morphol- 

 ogy; however, this approach is often 

 insufficient, especially for identify- 

 ing sympatric species, where there 

 is considerable similarity in size and 

 physical features and in pigmenta- 

 tion, as well as overlap in meristic 

 characters. The difficulty in identify- 

 ing Sebastes to species level exists for 



all developmental stages, but larvae 

 are especially difficult to identify be- 

 cause they lack diagnostic characters 

 (Kendall, 1991), Currently, only 15 

 species at the larval stage can be sep- 

 arated by physical characters such as 

 body shape, pigmentation patterns, 

 and head spine development (Love et 

 al., 2002). 



Rockfishes are important ecologi- 

 cally, and most species are economi- 

 cally valuable. Sebastes larvae form 

 a large portion of ichthyoplankton 

 collections and rank third or fourth 

 in abundance among all fish larvae 

 taken during California Cooperative 

 Fisheries Investigations (CalCOFI) 

 surveys that have covered the entire 

 length of the California and Baja Cal- 

 ifornia coast and now cover waters 

 off southern California. The ability 

 to identify Sebastes accurately and 

 efficiently at all developmental stages 

 will greatly improve both our ability 

 to learn about their life histories and 

 our management and conservation ef- 



