the anus is much larger in Monacanthus, in which it ex- 

 tends out far beyond the end of the encasing scales, than 

 in Stephanolepis, in which it extends out only to the dis- 

 tal end of the encasing scales. 



In Paramonacanthus the ventral dewlap is especially 

 short, extending posteriorly only to the anterior edge of 

 the encasing scale series, so that the end of the pelvis 

 protrudes prominently from the ventral contour. The 

 scales in Paramonacanthus have unbranched spinules, 

 as in Monacanthus and the vast majority of other mona- 

 canthids, but the spinules are less coarse than in Mona- 

 canthus and the skin of Paramonacanthus thus has 

 much the same velvety shagreen consistency as found in 

 Stephanolepis, rather than the rougher consistency of 

 Monacanthus. In Paramonacanthus the scale spinules 

 are unmodified, while, as described in detail by Berry 

 and Vogele (1961), mature males of Stephanolepis 

 develop a patch of bristles on the caudal peduncle and 

 mature males of Monacanthus have bristles and retrorse 

 barbs, females of Monacanthus developing smaller pos- 

 teriorly pointed barbs on the caudal peduncle but very 

 little if any bristle patch. 



In Monacanthus the dorsal lobe of the pelvis is very 

 large, the longest among monacanthids, being as- 

 sociated with support of the modified scales in the enor- 

 mous dewlap. In the various species of Stephanolepis and 

 Paramonacanthus the dorsal lobe is only moderately 

 developed, about that normal for the family when a dor- 

 sal lobe is present at all. 



In Paramonacanthus and Stephanolepis there is 

 always a free uppermost hypural, as is also the case in 

 the Indo-Pacific species of Monacanthus. In the two 

 Atlantic species of Monacanthus, one species (tuckeri) 

 usually has a free upper hypural (present in three out of 

 four cleared and stained specimens) while the other 

 species iciliatus) usually lacks the free hypural (present 

 in 2 out of 12 cleared and stained specimens). 



In Stephanolepis and Paramonacanthus epipleurals 

 usually extend back to the fifth caudal vertebra while in 

 Monacanthus they usually extend back only to the sec- 

 ond or third caudal vertebra (one specimen of M. chinen- 

 sis with epipleurals on the fourth caudal vertebra). 



In Stephanolepis, Paramonacanthus and the Indo- 

 Pacific species of Monacanthus there are 7 abdominal 

 and 12 caudal vertebrae, with 5 of the abdominal verte- 

 brae anterior to the first basal pterygiophore of the soft 

 dorsal fin, this being the most common vertebral con- 

 dition in the family. In the two Atlantic species of 

 Monacanthus the total number of vertebrae remains 19, 

 but it is the haemal spine of the seventh rather than that 

 of the eighth vertebra that supports the anterodorsal end 

 of the first basal pterygiophore of the anal fin, and the 

 neural spine of the fourth rather than that of the fifth 

 vertebra supports the anteroventral end of the first basal 

 pterygiophore of the soft dorsal fin, M. ciliatus and 

 tuckeri thereby having only 6 rather than 7 abdominal 

 vertebrae and only 4 rather than 5 predorsal vertebrae. 



With the similarities and differences discussed above 

 between the three genera here considered to be the most 

 generalized monacanthids, the question is whether one of 



them can be considered closer to the ancestral group 

 than the other two. There is little to go on. 



In balistids the ventral fan is only moderately 

 developed, with the expansible skin extending pos- 

 teriorly across most of the length of the encasing scale 

 series, ending at about the beginning of the most distal of 

 the four segments of scales. This condition is somewhat 

 intermediate between that of Paramonacanthus, with a 

 very short dewlap, and Stephanolepis, with a moderate 

 dewlap, but closer to that of Stephanolepis. 



The moderate dorsal lobe of the pelvis of Paramona- 

 canthus and Stephanolepis is closer to that of balistids 

 than is the extremely large one of Monacanthus. 



The upright spinules on small basal scale plates in 

 monacanthids is so different from the structure of the 

 scales of balistids that it is impossible to say whether the 

 branched spinules of Stephanolepis are more generalized 

 or specialized than the unbranched spinules in 

 Monacanthus and Paramonacanthus, and in any case 

 many of the scales of the head and extreme upper and 

 lower parts of the body in the Indo-Pacific species of 

 Monacanthus have branched spinules, while those of 

 most of the body are unbranched but sometimes notched 

 along their anterior edge. 



The phylogenetic implications of the presence of en- 

 larged caudal peduncular scales forming a patch of 

 bristles or spines in mature males of Stephanolepis and 

 in mature males and, to a lesser extent, in mature 

 females of Monacanthus, but in neither sex of Para- 

 monacanthus, is difficult to interpret. Among balistids 

 only two genera (Balistapus and Rhinecanthus) have 

 well-developed caudal armature, in the form of large 

 spines, but numerous genera (e.g., Sufflamen, Melich- 

 thys, Xanthichthys, Abalistes, Balistoides) have 

 horizontal ridges formed by upraised areas on the suc- 

 cessive rows of scales on the caudal peduncle and pos- 

 terior region of the body, and only a few genera are total- 

 ly lacking these ridges (e.g., Balistes and Odonus). 

 Without knowing from what type of balistids the mona- 

 canthids have been derived, it cannot be said at present 

 whether the caudal armature present in a few mona- 

 canthids is a generalized holdover from their balistid an- 

 cestors or a relatively specialized feature. 



I would guess that Stephanolepis is the most generaliz- 

 ed of the three genera, basically because of its moderate 

 dorsal lobe of the pelvis and moderate dewlap. Mona- 

 canthus would be a slightly specialized close derivative 

 of Stephanolepis whose evolutionary trend is for a great 

 increase in the size of the dewlap and of the dorsal lobe 

 supporting it, while in another direction Paramona- 

 canthus is an equally close relative of a Stephanolepis- 

 like stock, specialized by a great reduction in the size of 

 the dewlap but not of the dorsal lobe. 



Fraser-Brunner (1941b: 181) considered Pervagor the 

 most generalized monacanthid, stating that: "Pervagor, 

 in which dorsal and pelvic spines are better developed 

 than in the succeeding [i.e., all the other] genera, ex- 

 hibits two features which are not found among the other 

 forms possessing a movable pelvic spine - the forward 

 position of the dorsal spine and the presence of a deep 



