acanthurids and balistids "ne sauraient etre tenues pour 

 tres eloignees." What Monod did not realize was that if 

 the caudal skeleton of a triacanthodid is compared with 

 that of an acanthurid, one finds the two to be rather 

 similar. 



Just as the number of epurals, uroneurals, and hy- 

 purals becomes reduced in the more specialized plectog- 

 naths, so also the number of caudal fin rays becomes 

 reduced. In triacanthoids and balistoids there are 12 

 caudal fin rays; the uppermost ray and the lowermost ray 

 unbranched, the others branched (except Psilocephalus) . 

 In aracanids there are 11 rays, but in ostraciids the num- 

 ber is further reduced to 10. As with the other sclero- 

 derms, the uppermost ray and the lowermost ray are un- 

 branched, but the others are branched. Triodon has 12 

 principal caudal fin rays, but in addition to these it also 

 possesses a series of procurrent rays. Triodon is thus the 

 only Recent plectognath with more than 12 fin rays in the 

 caudal fin, but its number of principal caudal fin rays is 

 still the same as in triacanthoids and balistoids. In tetra- 

 odontids there are usually 11 caudal fin rays; the upper- 

 most ray and the lowermost two rays unbranched, the 

 others branched. As a part of his long and useful series of 

 papers (1942, 1944, 1949a, 1950-51, 1952, 1954, 1960), 

 Abe extensively and accurately surveyed the variability 

 of the caudal fin rays (1949b) in tetraodontids, show- 

 ing the i, 8, ii arrangement to be the normal condition 

 for all of the numerous species he studied. In diodontids 

 the number of caudal fin rays is either 9 or 10, with the 

 uppermost ray and the lowermost ray unbranched, 

 and the other rays branched. 



The complicated subject of the pseudocaudal fin of 

 molids can only be summarized here. A number of early 

 workers (Wellenbergh 1840; Goodsir 1841; Cleland 1862; 

 Wahlgren 1867; Putnam 1871) gave brief descriptions of 

 the highly modified structures of the caudal fin of 

 molids, but it was not until Ryder's (1886) work that the 

 origin of this fin was adequately discussed. Ryder 

 described some of the developmental stages of the molid 

 caudal fin and came to the conclusion that that struc- 

 ture was a gephyrocercal tail; the true caudal fin having 

 been lost and replaced by posteriorly migrated dorsal 

 and anal fin rays. Ryder said that molids and carapids 

 were the only two groups with gephyrocercal tails. 

 Ryder's opinion did not meet with unanimous accept- 

 ance, and Kaschkaroff (1914a), Grenholm (1923), and 

 Gregory and Raven (1934) continued to describe it as a 

 true caudal fin, while Whitehouse (1910) and Regan 

 (1910) both agreed that the structure was a gephyrocer- 

 cal tail. With typical thoroughness, Gudger (1935; 1936; 

 1937a, b, 1939) and Gudger and MacDonald (1935) 

 reviewed the literature on the subject and also per- 

 sonally examined a few developmental stages of molids. 

 Gudger came to the conclusion that both Mala and 

 Masturus have a gephyrocercal tail, and Raven (1939a) 

 agreed with that analysis. Raven (1939b) described the 

 tail of Ramania as also being gephyrocercal. Fraser- 

 Brunner (1951) was in basic agreement with Gudger and 

 Raven on this subject, but Fraser-Brunner stated that 

 the few fin rays that are found in the "nipple" part of the 



tail of Masturus are the remnants of the true caudal fin 

 rays, whereas the rest of the caudal structure is 

 gephyrocercal. In short, there is general agreement that 

 the entire caudal structure in Mola and in Ramania is 

 gephyrocercal, and that in Masturus at least the great 

 majority of the caudal structure is gephyrocercal. It is 

 also agreed that the bony supporting elements of the 

 pseudocaudal fin in molids are posteromedially migrated 

 dorsal and anal fin basal pterygiophores. Owen's 

 (1846:64) description of these pterygiophores as 

 "rudimental vertebrae . . . blended together at right 

 angles to the rest of the column" does not seem plausible. 

 The variability in the number of fin rays said to be pres- 

 ent in Mola (summarized by Beauregard 1893) attests to 

 the fact that it is difficult to delimit the pseudocaudal fin 

 from the dorsal and anal fins, as well as to the fact that a 

 great many of the descriptions of Mola in the literature 

 are actually based on Masturus (summarized by Gudger 

 1937a). To complicate matters even more, Fraser-Brun- 

 ner (1951) attempted to distinguish two species of Mola, 

 partially on the basis of the number of fin rays borne on 

 modified basal pterygiophores. Barnard (1935) has 

 pointed out the diagnostic value of the degree of branch- 

 ing of the fin rays of the pseudocaudal fin in molids. 



The dorsal and anal fins, and especially the dorsal fin 

 spines, being highly visible and variable between groups 

 of plectognaths, have been featured prominently in near- 

 ly all diagnoses, including here, as summarized below. 

 In triacanthodids there are six dorsal fin spines sup- 

 ported by five basal pterygiophores, with the first two 

 spines always prominent and visible externally, the third 

 to fifth either normally developed or rudiments buried 

 beneath the skin or barely protruding to the surface, and 

 the sixth spine either a short protruding element or a 

 rudiment buried beneath the skin or barely protruding to 

 the surface. There are 12 to 18 dorsal fin rays and 11 to 16 

 anal fin rays in Recent species. 



In triacanthids there are nearly always six dorsal fin 

 spines, except that the fifth and sixth spines are 

 sometimes absent in one Recent species and perhaps in 

 several fossil forms, supported by four, rarely only three, 

 basal pterygiophores, with the first four spines always 

 visible externally, the fifth spine usually very short but 

 nearly always protruding at least a short distance 

 through the skin, and the sixth spine nearly always pres- 

 ent as a buried rudiment. There are 19 to 26 dorsal fin 

 rays and 13 to 22 anal fin rays. 



In balistids there are three dorsal fin spines, the second 

 more than one-half the length of the first, supported by 

 two basal pterygiophores and a supraneural strut. In 

 monacanthids there are usually two dorsal fin spines, but 

 sometimes only one, the second spine, when present, not 

 more than one-third the length of the first, with both 

 spines supported by a single basal pterygiophore, 

 without a supraneural strut. In balistoids there are 23 to 

 52 dorsal fin rays and 20 to 66 anal fin rays. The locking 

 mechanism of the spines of triacanthoids is described by 

 Tyler (1968), and that of balistoids by the references 

 given in the description of Balistapus undulatus. 



In ostracioids there is no spiny dorsal fin, and the dor- 



