242 



Fishery Bulletin 104(2) 



jaws, throat, ventral portion of cheeks, interopercle, 

 extreme ventral tip of subopercle, and extreme dorsal 

 margin of gill membrane naked; remaining part of head 

 and dorsal half of body everywhere covered with rough, 

 granular scales; approximately 10-16 close-set rows of 

 tiny granular scales on anterior and dorsal periphery 

 of eyeball, rows usually extending nearly to dorsalmost 

 margin of eye. Oblique dermal folds terminating on 

 caudal peduncle well before reaching ventral midline 

 but sometimes crossing space between anus and base of 

 anal fin. Dorsal-fin spines 10-11, anal-fin rays 23-25; 

 vertebrae 45-47. 



Color in preservation: dark brown dorsally, light 

 brown, tan, to cream ventrally, with four dark saddles 

 along dorsal length of body extending ventrally to, and 

 sometimes slightly below, lateral line, anteriormost 

 saddle situated beneath spinous dorsal fin, two be- 

 neath soft dorsal fin, posteriormost saddle on caudal 

 peduncle; most specimens with a small saddle or nar- 

 row band of pigment crossing soft-dorsal fin at mid- 

 length (homologous to fifth saddle of many congeners). 

 A series of dark, more-or-less equally spaced spots or 

 streaks extending along length of body just ventral 

 to lateral line, spots nearly always interconnected 

 in males to form a conspicuous dark stripe across 

 length of body, but not extending onto caudal-fin base. 

 Anterior portion of lower lip darkly pigmented, some- 

 times interrupted at symphysis of jaw; upper lip with 

 a dark streak across symphysis and two or three simi- 

 lar, more-or-less equally spaced streaks on each side, 

 posteriormost streak contiguous with similar spot or 

 streak on maxilla (pigmentation of lips more intense 

 in males); males with a streak of pigment extending 

 along dorsal margin of upper jaw from ventral margin 

 of lacrimal to beneath and slightly past eye; a large, 

 diffuse spot posterior to distal tip of maxilla (more 

 heavily pigmented in males), sometimes forming an 

 oblique streak extending from posterodorsal margin 

 of maxilla posteroventrally to preopercle. Dorsomedial 

 surface of opercle and posterodorsal margin of gill 

 membrane densely covered with small melanophores, 

 pigment extending ventrally along medial surface of 

 branchiostegal rays (pigmentation often bleached away 

 in old, poorly preserved material); lateral surface of 

 distal portion of four or five posteriormost branchioste- 

 gal rays, as well as gill membrane in between, heavily 

 peppered with small melanophores. Each spine and 

 ray of dorsal fin (but not membrane in between) with 

 two to four, dark, more-or-less equally spaced spots. 

 Four to seven narrow, more-or-less equally spaced 

 bands of dark pigment across dorsalmost or nonex- 

 erted rays of pectoral fin; a small, dark spot at dorsal 

 insertion of pectoral fin often continuous with pig- 

 ment extending ventrally to mid-base of fin. Anal fin 

 usually unpigmented, but some specimens with rays 

 darkly pigmented or lined with rows of melanophores, 

 pigmentation especially intense in males; pelvic fins 

 usually unpigmented, but some males with distal tips 

 of first and second rays lightly pigmented. Base of 

 caudal fin with dark dorsal and ventral spots, ventral 



spot usually more distinct than dorsal; males with a 

 dark spot near distal end of dorsal and ventral lobes 

 of caudal fin, sometimes faint and difficult to discern 

 in preserved specimens. 



Sexual dimorphism 



Females with pectoral-fin rays slightly shorter (22.6- 

 23.4% SL) than those of males (20.9-25.0% SL). Females 

 with pelvic-fin rays slightly shorter (12.6-14.7% SL) 

 than those of males (12.7-17.7% SL); tissue surrounding 

 pelvic-fin spine and first ray expanded in males. 



Statistical analyses 



Univariate analyses of morphometric characters indi- 

 cated that only interorbital width and the length of the 

 first infraorbital pore exhibited significant differences 

 between T. dorothy and T. pingeli (Table 1). The analy- 

 ses of meristic characters, on the other hand, revealed 

 significant differences in several characters, including 

 fewer dorsolateral scales, lateral-line scales, and pecto- 

 ral-fin rays, and a greater number of dermal folds and 

 gill rakers in T. dorothy than in T. pingeli (Table 1). 

 Three-way plots of dorsolateral scales, dermal folds, gill 

 rakers, and pectoral-fin rays exhibit a clear separation 

 between T. pingeli and T. dorothy (Fig. 2B). 



In plots of scores from the PCA of meristic characters, 

 two distinct clusters, one for each of the two species, 

 were evident (Fig. 3A). The primary axis of separation 

 was PCI (30.2%: of total variance), which was heavily 

 loaded positively on dorsolateral scales and gill rakers 

 and negatively on lateral-line scales; loadings along the 

 PC2 axis (22.4% of total variance) were heavily loaded 

 positively for breast-scale rows and negatively for anal- 

 fin rays and dorsal-fin rays (Table 3). In contrast, broad 

 overlap was displayed in the analysis of morphometric 

 characters, with individuals of T. dorothy clustering 

 among the larger group of T. pingeli from the Sea of 

 Okhotsk (Fig. 3B). The first principle component (PC) 

 was interpreted as a size component accounting for 

 95.5% of the total variance. In plots of the scores of 

 the sheared second and third shape components (1.5%c 

 and 1.1% of total variance, respectively), the great- 

 est dispersion was along the PC2 axis (Fig. 3B, C), 

 and loadings along this axis were highest for interor- 

 bital width, snout length, and caudal-peduncle depth 

 (Table 3). Loadings along the PC3 axis were highest for 

 pectoral-fin ray length, ventral caudal-peduncle length, 

 and orbit length (Table 3). 



The single linear discriminant function equa- 

 tion produced was highly significant (x~ = 96.5, 6 df, 

 P<0.0001): 



D = 0.894(rfs) - 0.691(g) -i- 0.530(p) - 0.508(c?/') 

 -0.284(/o) + 0.098(//), 



where D = the discriminant score of an individual; 

 ds = number of dorsolateral scales; 

 g = number of gill rakers; 



