REVISED CLASSIFICATION FOR CERTAIN CIRRHITOID GENERA 



referred to the family Cheilodactylidae (see Allen & Heem- 

 stra, 1976, for comments on the status of this genus; also p. 6 

 below). Kusaka, however, lists the species under the heading 

 Aplodactylidae. The bone depicted is certainly of the cheilo- 

 dactylid type and not of the aplodactylid type, and I presume 

 Kusaka's placing the species in the Aplodactylidae is a lapsus. 

 This author (op. cit: ) also figured and described the urohyal 

 from a specimen supposedly of Cirrhitichthys aureus Temm. 

 & Schl., 1843 (Cirrhitidae). Unfortunately I have not been 

 able to examine a specimen of this species, but the bone 

 illustrated (and described as 'shaped like a standing rat') is 

 unlike that in any cirrhitoid taxon I have examined, particu- 

 larly amongst members of the Cirrhitidae and even in a taxon 

 such as Oxychirrhites typus Bleeker, 1857, whose elongate 

 and tubular snout is an unusual morphotype within both the 

 Cirrhitidae and the cirrhitoids as a whole. If Kusaka's figure 

 and description are accurate and the specimen was correctly 

 identified, then a sixth and highly distinctive form of urohyal, 

 one far removed from that of other cirrhitids must be 

 recognised, and the higher taxonomic position of its possessor 

 or posessord be reconsidered (assuming, that is, the bone 

 Kusaka examined was not teratological or damaged during 

 preparation). 



A typical cheilodactylid urohyal (Fig. ID) occurs in all 

 members of the family (sensu Regan [1911] and subsequent 

 authors) I have examined apart from Nemadactylus and 

 members of the genus Acantholatris , viz the type species A. 

 monodactylus (Carmichael, 1818), and the species A. gayi 

 (Kner, 1869) and A. bergi (Norman, 1937). 



Parenthetically it should be noted that A. gayi and A. bergi 

 were both placed in the genus Cheilodactylus, and the family 

 Cheilodactylidae, by Norman (1937). The former species was 

 later transferred by Fowler (1945) to the genus Acantholatris , 

 with no explanation given for the change, but was retained in 

 the family Cheilodactylidae. Neither author appears to have 

 been aware, however, that Gill (1862) had included Acantho- 

 latris in his subfamily Latridinae. Mann (1954: 266) followed 

 Fowler's generic and familial placing of A. gayi, and listed the 

 species bergi under Acantholatris in the index to that publica- 

 tion. The reader is there referred to page 266 of the text. No 

 mention is made of A. bergi on that page, but on page 85 (op. 

 cit.) Acantholatris bergi Norman (the author's name not 

 enclosed in brackets) is listed amongst the Tnvasores del 

 Atlantico'. Mann (op cit.) is thus the first author to employ 

 this particular combination of names for the species. As 

 noted earlier (p. 1) Regan did not include Acantholatris in 

 any of his cirrhitiform families. 



The familial classification of Acantholatris Gill, 

 1982 and Nemadactylus 



The urohyal in all three Acantholatris species examined, and 

 in Nemadactylus macropterus, is virtually identical and differs 

 markedly from that in the cheilodactylids, cirrhitids, aplodac- 

 tylids and chironemids (see pp. 2-5 and cf Fig. 1 with Fig. 

 2A). Instead, it resembles the latrid type, both in detail and 

 in its gross morphology (c/Figs. 2A, B and C), especially in 

 its fan-like outline. This marked difference would suggest 

 that the latrid genera (as listed in Regan, 1911), together with 

 Acantholatris and Nemadactylus shared a recent common 

 ancestry distinct from that of the cheilodactylids. It also 

 suggests that the phyletic relationships of the two groups are 

 obscured by uniting the cheilodactylids with the latrids in a 



single subfamily, as did Gill (1862). 



Thus, in my view, based essentially on their urohyal 

 morphology and not negated by other characters (see, how- 

 ever, the pectoral fin character discussed below), Acanthola- 

 tris and Nemadactylus should be included in the family 

 Latridae, currently comprising species of the genera Latris, 

 Latridopsis and Mendosoma, the latter recently shown to be 

 monotypic by Gon & Heemstra (1987). In addition to the 

 urohyal characters, the genera listed above lack a suborbital 

 shelf, which in cheilodactylids is a prominent feature formed 

 from the posterior upper margin of the lachrymal bone and 

 the entire upper margins of the second and third suborbitals. 

 Also, unlike cheilodactylids, these genera have the basal 

 scaly sheath to the soft dorsal fin somewhat higher and thus 

 more prominent than that at the base of the spinous part of 

 the fin. 



As in cheilodactylids, the latrids (here taken to include 

 Acantholatris and Nemadactylus) have 35 vertebrae (14 

 abdominal + 21 caudal elements including the urostylar 

 element; data from radiographs and dry skeletons listed on 

 p. 2). To judge from the dry skeletal and dissected material 

 available to me, parapophyses are present on all precaudal 

 centra in both families, and no ribs are sessile. 



Possible lineages within the Latridae as now 

 expanded to include the genera Acantholatris and 

 Nemadactylus 



Acantholatris and Nemadactylus differ noticeably from Latris, 

 Latridopsis and Mendosoma in having one of the lower, 

 unbranched pectoral rays (i.e. the fifth, sixth or seventh ray 

 from the bottom of the series) greatly elongated, its tip, 

 which extends beyond the fin's margin, reaching to at least 

 the level of the anus and sometimes as far as the midpoint of 

 the anal fin. 

 There are also differences in the following features: 



(i) In scale size, as shown by lateral-line scale counts. In 

 Latris, Latridopsis and Mendosoma these range from 112 to 

 120 in the two former taxa, and from 68-78 in Mendosoma 

 (data from Last et. ai, 1983; Gon & Heemstra, 1987; 

 pers.obs.). In Nemadactylus macropterus the count is 59 or 

 60, and in other species 47-68 (pers obs.; Last et. ai, 1983) 

 and in Acantholatris monodactylus , A. bergi and A. gayi the 

 range is from 50-60 (Norman, 1937; pers. obs.). 

 (ii) Anal fin length. In Latris, Latridopsis and Mendosoma, 

 the number of branched anal rays ranges from 17-35 (the 

 lowest counts occurring in Mendosoma, viz. 17-21, whereas 

 in Acantholatris species and Nemadactylus macropterus the 

 range is from 12-15, and other species of the genus, 16-19 

 (sources as above). 



Pending a detailed generic revision of the various taxa 

 involved, especially the several Australian and New Zealand 

 species currently referred to the genus Nemadactylus it would 

 be premature to formally recognise the two groups as, for 

 instance, tribes or subfamilies of the Latridae, although 

 phylogenetically some split seems to have occurred within the 

 lineage. 



The condition of the pectoral fin in the Latris- Latridopsis- 

 Mendosoma group of latrids provides something of a puzzle 

 since these taxa are the only cirrhitiforms not showing any 

 marked elongation of the uppermost unbranched ray in the 

 lower section of the pectoral fin, nor, as in most other 



