THE FISHES OF SAMOA. 285 



836. Abudefduf septemfasciatus (Cuvier & Valenciennes). Mittu; Alala sayii. Samoa; (iuam; 



Fate; Shortland I.; Rarotonga; Raiatea and Makatea (Seale); East Indies. 

 Of this well-marked species we have four large examples from Apia. Life colors of one of tht;se, 

 called mutu, dull olive, with black bars, six wider than interspaces; silvery streaks faint along rows 

 of scales; pale whitish streak along spinous dorsal; everythingel.se pale. 



837. Abudefduf saxatilis (Linn:tnis). Samoa; Waigiu; Tubuai (Seale); Marcus I.; East Indies; 



Japan. 

 This species, common from Japan to the Red Sea, is occasionally taken in the South Seas. We 

 have one large example, corresponding to Abudefduf waigiemis from Apia. In Hawaii it is replaced 

 by Abtidefdvf abdomhialis, and in America by the ecjually closely related Abudefduf marc/iiiatus. 



838. Abudefduf abdominalis (Cuvier & Valenciennes). Hawaii; Laysaii. 



839. Abudefduf coelestinus (Cuvier & Valenciennes). Fill muiia. Samoa; Ulietca; New Guinea; 



Waigiu; Tahiti; Kaiatea; Fat6 and Shortland Is. (Seale). (PI. XLi, fig. 1.) 

 This species is very common about Samoa, nearly a hundred specimens being taken at Apia and 



i'ago Pago. It is perfectly well distinguished from Abudefduf sa.mtilis, and rarely inhabits the same 



waters. Its life color is pale sky-blue, not olive-green, and the black stripe on each lobe of the caudal 



is constant. 



A specimen from Apia in life was vivid sky-blue with paler streaks along scales; cross-bands jet 



lilack, a Vilack stripe on each caudal lobe. 



840. Abudefduf curacao (Bloch). Xew^ Guinea (Macleay); Shortlan<l I. (Seale); East Indies; 



Japan. 



841. Abudefduf dicki (Lienard). Tu'u'u. Samoa; Ponape; Guam; East Indies. (PI. xli, tig. 2.) 

 i;l!ilihisodon unifasciatus Kner & Steindachner, Sitz. Ak. \Vis.«. Wien 1867, 37.5, Samoa. 



Of this handsome and well-marked species half a dozen specimens were taken at Apia. Life 

 colors of one of these, called tuUi' u, golden brown, dark, each scale with a vertical bar of dark blue; a 

 large jet-black crescent from front of soft dorsal across body and covering most of anal; behind this 

 abruptly pale orange- red, the color covering caudal peduncle, caudal and last rays of dorsal and anal; 

 older examples have color deeper, clear orange-red behind black crescent; pectoral bright golden 

 yellow, dusky behind; ventral blackish; head plain. 



842. Abudefduf lacrymatus (Quoy & Gaimard). Pa samusama (yellow tail). Samoa; Guam; 



Yap; HowlandL; East Indies. 



Of this well-marked species, black, with bluish white well differentiated spot", we have three 

 specimens from Apia. 



Life colors of a specimen called i'u samamma, black, with scattered round spots of deep violet-blue, 

 on head, nose, back and sides, also on dorsal; caudal peduncle, caudal and posterior part of soft 

 dorsal abruptly pale rayish yellow, the color of the ginger flower; fins otherwise black, the anal 

 washed with orange; dorsal and spinous anal with violet spots; ventral black, edged with blue; axil 

 a little dusky. 



Another specimen was black (brownish) with bright violet shade on opercle and preorbital and 

 numerous violet-blue spots on spinous dorsal, which has a narrow blue edging; axil black, dorsal 

 black, its la-it rays abruptly bright orange; caudal dusky orange-brown; anal black with pale blue 

 anterior edge; pectoral dusky; ventral blackish. 



843. Abudefduf behni (Bleeker). Samoa; Rarotonga (Seale) . 



Of this species, pale brownish black, with a black axillary spot, we have one fine specimen 

 from Apia. 



cio not wish to split up nomenclatorially. It is as if I should say the enormous genus of Eana can he subdivided for the 

 sake of convenience into (1) the temporaria 'roup; (2) the esciilenia group, etc. That Forsk&l (or his editor) uses capitals 

 and an upper case capital initial means nothing, for he also has " (C) IinntaculaUe" exactly like Abu bamrur. 



The case of the Scuenas, i. e., Abu hamnir, etc., is not exactly like that of AbudefduJ, and I believe there is just 

 difference enough to draw the line ju^t between them. I do not Ijelieve they fullill the conditions for being subgenera, 

 while Ahudcfduf AcK-f." 



