444 



ONTOGENY AND SYSTEMATICS OF FISHES-AHLSTROM SYMPOSIUM 



Table 112. Continued. 



* Taxonomic status is not agreed upon by current workers 

 •* No. of anal spines recognized vanes among workers. 



' Howe. K. Compilation of menstic data from published and unpublished sources, available from Northwest and Alaska Fishenes Center. NOAA-NMFS, Seattle, WA. 

 - Howe. K.. unpublished data. 



bladder is huge, occupying much of the body cavity with anterior 

 lobes reaching near the rear of the cranium. A more thorough 

 study of these gasbladders is needed. Dactylopterids. in relation 

 to triglids, have: 1) a hinged bony connection with the pre- 

 opercle; 2) much heavier and more elaborate ossification of the 

 cranium; 3) first three vertebrae elongate and modified; and, 4) 

 reduced opercular and gill openings. Given the extreme osteo- 

 logical modifications of these fishes, a current working hypoth- 

 esis (WNE) is that the Apistinae. triglids, peristediids and dac- 

 tylopterids share a common ancestry. However, more 

 information is needed before any formal changes are proposed. 



Suborder Cottoidei 



We include the following families in this group: Hexagram- 

 midae, Zaniolepidae. Anoplopomatidae. Cottidae (broad sense). 

 Agonidae. Cyclopteridae, and Liparididae. The limits of these 

 families and subfamilies are not well-defined and there is con- 

 siderable lack of understanding among workers in defining both 

 family limits and those of higher categories (see Washington 

 and Richardson, MS for review). We treat these diverse groups 

 together in order to facilitate discussion of past classifications, 

 not because we believe they necessarily form a monophyletic 

 assemblage. 



Meristic features and approximate number of species for in- 

 cluded groups are provided in Table 1 1 2. Data have been com- 

 piled from many sources and may not be complete for some 

 groups and may omit extremes found in abnormal individuals. 



Matsubara ( 1 955). in a thorough treatment of Japanese species, 

 recognized: 1) a superfamily "Hexagrammicae" (including An- 

 oplopomatidae and Hexagrammidae); 2) a superfamily "Cot- 

 ticae" (including Cottidae with subfamilies and Psychrolutidae); 

 and, 3) a superfamily "Agonicae" (including Agonidae and As- 

 pidophoridae). He placed the cyclopterids and liparidids in a 

 larger division, Cyclopterina. 



Quast (1965), in a treatment which focused on relationships 

 of hexagrammid fishes, followed Regan (1913a) and Berg ( 1 940) 

 in recognizing a superfamily Hexagrammoidae and a superfam- 

 ily Cottoidae. He separated the Cottoidae from the Hexagram- 

 moidae on the basis of four characters: 1 ) lack of a basisphenoid; 



2) dentigerous upper pharyngeals restricted to one or two pairs; 



3) pleural ribs absent or developed on only a few posterior 

 abdominal vertebrae; and, 4) pectoral interradial foramina small 

 or absent. However, Quast proposed that the hexagrammids 



and cottoids form a single evolutionary lineage within the Scor- 

 paeniformes and that the Anoplopomatidae are significantly 

 distinct from both the hexagrammid-cottid lineage and the scor- 

 paenid lineage to warrant separate superfamily status. He further 

 suggested that the zaniolepids are intermediate between the hex- 

 agrammids and cottids. 



Other workers (Greenwood et al,, 1966; Nelson, 1976) have 

 placed the hexagrammids. anoplopomatids. and zaniolepids to- 

 gether in the suborder Hexagrammoidei. and the cottids (broad 

 sense of Washington and Richardson, MS), agonids, and cy- 

 clopterids in the suborder Cottoidei. Hallacher (1974) found a 

 cranioclavical (gasbladder) muscle present in the zaniolepids, 

 cottids (broad sense), agonids, and cyclopterids. In contrast, 

 Hexagrammos was found to have the scorpaenoid condition. 



In the following discussion, we present information about 

 recent studies which have helped resolve relationships within 

 cottoid subgroups and outline groups where problems remain. 



The systematic status of the Hexagrammidae is the subject 

 of disagreement at the specific through family levels. Quast (1965) 

 and Nelson (1976) include four genera in the Hexagrammidae— 

 Oxylehiits. Ophiodon, Hexagrammos (inc\u(i\n% Agramnms), and 

 Pleurogramfuiis. Quast considered Oxylehiits to be the most 

 primitive genus because of low numbers of meristic elements 

 and the "lack of specializations." Hexagrammos and Pleiiro- 

 grammus were considered to be closely related, relatively spe- 

 cialized genera because of the reduction in head spination. dorsal 

 and anal fin spines, etc. 



Quast (1965) and Nelson (1976) included the two species of 

 the genus Zaniolepis in the family Zaniolepidae. Other workers 

 (Rutenberg, 1962) have included Zaniolepis in the family Hex- 

 agrammidae. while others (Hart, 1973) have combined Zani- 

 olepis and Oxylehius in the family Zaniolepidae. 



The Anoplopomatidae contains two monotypic genera, An- 

 oplopoma and Erilepis (Quast, 1965; Nelson. 1976), however 

 some workers have placed Erilepis in its own family, the Eri- 

 lepidae. 



Those families that traditionally have been placed in or near 

 the Cottidae are not clearly defined. Previous workers have 

 proposed between 1 and 1 7 families of "cottids." Greenwood 

 et al. (1966) and Nelson (1976) recognize 7 cottid families: 

 Cottidae, Icelidae. Cottocomephoridae. Comephoridae. Cot- 

 tunculidae, Psychrolutidae and Normanichthy idae. Other work- 

 ers have chosen to combine these 7 families in the single family 



