FISHERY BULLETIN: VOL. 79, NO. 1 



seems to be related to a neustonic habitat in some 

 (e.g., Scorpaenichthys , Hemilepidotus) but not 

 others (e.g., Radulinus). Relative constancy of 

 pigment pattern (such as presence or absence of 

 lateral pigment posterior to the anus) within a 

 group used in conjunction with other characters, 

 however, may provide additional evidence for 

 within-group relationships. 



If this line of reasoning and these assumptions 

 are valid, then certain trends seem apparent 

 which may be indicative of relationships. Group 1 

 (Artedius et al.) appears to be a natural group 

 sharing a number of derived characters not pres- 

 ent in any other group or genus (i.e., multiple 

 preopercular spines, somewhat trailing gut, un- 

 usual gut diverticula, or at least bulging guts). 

 A preanal fin fold is apparently absent and pig- 

 ment pattern is relatively constant. The grouping 

 agrees with findings of Taranets (1941), in part, 

 and Bolin (1947), who considered the genera to be 

 closely related (Table 3). It seems to be a rather 

 specialized group as Bolin (1947) implied, and, 



based on the distinctiveness of larval characters, 

 may warrant consideration at possibly the sub- 

 familial level. 



Group 2 iParicelinus et al.) shares the derived 

 slender body form with pointed snout, and also 

 possesses relative constancy of pigmentation, i.e., 

 no lateral pigment. Relationships among at least 

 some of the genera in this group have been implied 

 previously (Table 3). The distinctiveness of larval 

 form within this group suggests a separate line- 

 age; this group may warrant possible subfamilial 

 status. 



In Group 3, all but Dasycottus share a highly 

 modified larval form tending in degrees toward 

 globose. The constancy of the pigmented pectoral 

 fin is unique among all groups or genera con- 

 sidered. With the possible exception of Dasycottus, 

 the genera appear to bear at least some relation- 

 ship to each other. 



Group 4 is the most generalized in that a num- 

 ber of primitive character states are exhibited and 

 relationships cannot be assessed on given present 



Table 3. — Intergeneric relationships of cottids as interpreted by A = Regan (1913), B = Taranets (1941), C = Bolin (1947), and 

 D = Watanabe (1960). Included are only those 25 northeast Pacific genera for which larvae are known and discussed in this paper. 

 Parentheses indicate a more distant relationship. 



3 



<1> 



Genus 



a 

 <i> 



S 



3 

 O 

 C 

 g 



o 



o 

 o 

 o 

 c 



o 



o 

 o 



3 

 .C 



O 



«5 



o 

 o 

 o 



o 

 u 

 o 

 o 



CD 



3 

 '(5 



a 



o 

 o 



i^ i; -J 



to 

 Q. 

 O 



Artedius 



Blepsias ' 



Chitonotus 



Clinocottus 



Cottus 



Dasycottus 



Enophrys 



Gilbertidia' 



Gymnocanthus^ 



Hemilepidotus 



Hemitripterus^ 



Icelinus 



Icelus 



Leptocottus 



Malacocottus 



Myoxocephalus 



Nautichttiys 



Oligocottus 



Orthonopias 



Paricelinus 



Psyclirolutes 



Radulinus 



Rhamphocottus ' 



Scorpaenichthys^ 



Triglops^ 



B,'(C) 



(C) (C) 



(A),D 



B,'(C) 

 B^ 



B,C 

 B 



B,D 



A,C,D 



B,C 

 B,^C 



(C) 



B,C 

 B,C 



(C) 

 D 



A,B 



(A).D 



(C) 



c 



A A 



B,C 



(C) 



Considered in family Blepsiidae by Taranets (1941). 

 Artedius as Ruscarius. Ruscariops of Taranets (1941). 

 ^Artedius including Aryias , Allartedius , Astrolytes , 

 Parastrolytes of Taranets ( 1 94 1 ) , 

 "Considered in subfamily Gilbertinae by Watanabe (1960). 



^Considered In subfamily Gymnocanthinae by Watanabe ( 1 960) . 

 'Considered in family Hemitripteridae by Taranets (1941). 

 'Considered in family Rhampfiocottidae by Taranets (1941). 

 'Considered in order Nototfieniiformes by Taranets (1941). 

 'Considered in subfamily Triglopsinae by Watanabe (1960). 



118 



