626 



ONTOGENY AND SYSTEMATICS OF FISHES-AHLSTROM SYMPOSIUM 



Table 165. Matrix of Character State Codes (see Table 164 and Fig. 335 for the stromateoid genera and for the three perciform families 

 used as out-groups in the phylogenetic analysis). Dashes indicate characters that are inapplicable. Question marks indicate unknown character 



states and were designated as "missing observations" in the analysis. 



1 . Number of rows of premaxillary teeth 



2. Number of rows of dentary teeth 



3. Pharyngeal sac 



4. Shape of pharyngeal sac 



5. Arrangement of papillae in pharyngeal sac 



6. Papillae on upper portion of pharyngeal sac 



7. Position of papillae in pharyngeal sac 



8. Shape of papillae base in pharyngeal sac 



9. Condition of maxilla 

 10. Supramaxillary bone 

 1 1. Lacrimal bone 



12. Relationship of gills to isthmus 



13. Pseudobranch 



14. Scale type 



15. Opercular scalation 



16. Preopercular scalation 



17. Prominent preopercular spines 



1 8. Number of branchiostegal rays 



19. Pelvic bone and fin 



20. Number of predorsal bones 



21. Number of dorsal fins 



22. Keels on caudal peduncle 



23. Number of hypurals 



24. Procurrent spur 



25. Ray base preceding procurrent spur 



26. Juvenile pigmentation 



27. Primary juvenile association 



ing body size and shape, fin pattern, presence or absence of 

 palatal dentition, shape of the papillae in the phai7ngeal sacs 

 and the number of branchiostegal rays, vertebrae and epural 

 plus hypural elements in the caudal skeleton. 



Haedrich ( 1 969) in describing the Amarsipidae added a sixth 

 family to the stromateoid suborder. This family exhibits a mix- 

 ture of primitive and derived characters. It lacks pharyngeal 

 sacs, but the pharyngeal teeth are extraordinarily developed and 

 may perform a shredding function analogous to the sacs of other 

 stromateoids. Haedrich (1967) argued that possession of a per- 

 ciform caudal skeleton, uniserial jaw teeth, an expanded lacri- 

 mal bone, an inflated and protruding top of the head, an exten- 

 sive subdermal canal system and a bony bridge over the anterior 

 vertical canal of the ear provides the basis for placement of 

 Amarsipus in the suborder Stromateoidei. He considered the 

 new family to be distantly allied with the Nomeidae. 



In the present study, a phylogeny of the 1 6 stromateoid genera 

 was constructed using a set of 27 characters (Tables 164, 165) 

 that could be coded for all genera with little or no overlap and 

 ambiguity. Initially, a larger number of prospective characters 

 (~65) were identified and evaluated. Comparison of stroma- 

 teoids with presumed out-groups helped to generate characters 

 and to establish polarity in the transformation series. Charac- 

 ters, however, were omitted if they could not be quantified or 

 if insufficient information was available to characterize every 

 taxon. In a few cases, character states were coded as "missing 

 observations" if three or fewer genera required this coding and 

 the characters were judged important in resolving relationships 

 between the other genera. 



Three closely related perciform families— Girellidae, Ky- 



phosidae and Scorpididae— were used as potential out-groups 

 in the analysis (see below for rationale). Although these three 

 taxa are frequently classified as subfamilies of the Kyphosidae 

 (e.g.. Nelson, 1976), G. D. Johnson (pers. comm.) considers 

 them to be distinct families. 



The analysis was performed using the PHYSYS package which 

 differentiates taxa based on the presence of shared derived char- 

 acters (synapomorphies). Several phylogenetic trees were gen- 

 erated from the genus-character data matrix using the Wagner 

 distance algorithm (see Farris, 1970; Wiley, 1981). These trees 

 were diagnosed to identify the origin of each apomorphy and 

 to examine character reversals and convergences. Transfor- 

 mation Series Analysis was performed on the data to verify 

 polarities developed through out-group comparison (see Mick- 

 evich, 1982) and to resolve nonlinear series. Further optimizing 

 produced the most parsimonious tree from the data matrix (Fig. 

 335). 



This phylogenetic tree (Fig. 335) was basically similar, with 

 certain exceptions, to that proposed by Haedrich (1967). //i'- 

 peroglyphe emerged as the most plesiomorphic stromateoid tax- 

 on possessing a pharyngeal sac and Pampus as the genus with 

 the greatest number of apomorphies. Arioinma also ranked as 

 a highly derived genus in the suborder. Despite its several syn- 

 apomorphies with advanced stromateoid genera, Amarsipus 

 emerged as the sister taxon of all other stromateoid genera pri- 

 marily because it lacks a pharyngeal sac. The major differences 

 between the present analysis and Haedrich's interpretation lie 

 with the relationships of Arioinma and Tctragonurus to other 

 stromateoids and with the family limits of the suborder. Based 

 on the cladogram, Tctragonurus and Ariomma are more closely 



