y^m^.: 



Figure 275.— Dorsal view of skull of Carinotetra- 

 odon lorteti, 33.1 mm SL, locality unknown. 



the palatines than in any other tetraodontids, i.e., the 

 medial edges of the premaxillaries in leiurus, cutcutia, 

 and palembangensis have a much longer concave region 

 than in any other tetraodontids. 



The four species of Monotreta studied are alike in that 

 there are small to minute teeth on the first pharyngo- 

 branchial, no trituration teeth in either jaw, no evidence 

 of prootic prongs representing remnants of the dorsal roof 

 of the myodome, no interhyal, a single hypohyal (one 

 specimen of leiurus with two on one side and one on the 

 other), and no supraneural element. The only other 

 tetraodontids in which the supraneural is absent are 

 members of the closely related genera Tetraodon (in 

 which it is absent in mbu but present in Uneatus), 

 Chelonodon (absent in fluuiatilis but present in patoca), 

 and Carinotetraodon (absent in the single species 

 studied). In one of the species oi Monotreta (leiurus) the 

 first five abdominal vertebrae have bifid neural spines, 

 and the neural spine of the sixth abdominal vertebra is 

 bifid anteriorly but single posteriorly, while in the other 

 three species the neural spines of only the first four verte- 

 brae are bifid. In most other tetraodontids the first three 

 abdominal vertebrae have bifid neural spines, and the 

 neural spine of the fourth vertebra is bifid anteriorly but 

 single posteriorly. In Arothron, however, the neural 



spines of the first four abdominal vertebrae are bifid, and 

 at least a short portion of the anterior end of the neural 

 spine of the fifth vertebra is bifid. Thus, Arothron and 

 Monotreta tend to have a slightly greater development of 

 bifid neural spines than do the other genera of tetraodon- 

 tids, a specialization. 



In Monotreta the vertebrae are 18 in cutcutia and 

 gularis, but 20 or 21 in leiurus and palembangensis. 



Even though the ensemble of subsidiary characters, in- 

 ternal and external, does not serve to separate gularis 

 from the other three species of Monotreta studied, the 

 differential shapes and sizes of the frontal, prefrontal, 

 ethmoid, vomer, palatine, and, especially, premaxillary 

 clearly indicate that there are two lines of diversification 

 within the species of Monotreta studied. If additional 

 species of Monotreta not studied here are eventually 

 shown not to bridge the gap between the two groups, they 

 should be recognized as at least subgenerically distinct 

 (the identification of the specimen here studied as 

 gularis needs verification). One suspects that the ex- 

 amination of additional species will bridge the gap. For 

 example, Carinotetraodon lorteti, otherwise assignable 

 to Monotreta because of the nasal tube with a single nos- 

 tril (but with the upper and lower lateral lines not joining 

 and both reaching the tail), has the length of the eth- 

 moid about intermediate between gularis and the other 

 three species of Monotreta, the palatine-vomerine shaft 

 about as long as in the three specialized species, the 

 length of the premaxillary pedicel intermediate between 

 gularis and the other three species, and the size of the 

 space between the concave posteromedial regions of the 

 two premaxillaries very similar to that of gularis. The 

 anterolateral wing of the sphenotic in Carinotetraodon is 

 about like that of one of the three specialized species 

 (palembangensis) of Monotreta, but there is no struc- 

 ture in any of the four species of Monotreta similar to the 

 posterolateral wing of Carinotetraodon. 



It seems obvious that Carinotetraodon and thus 

 Canthigaster, as discussed earlier in this section, have 

 their closest relationships among the Afom)tre(a-like 

 tetraodontids, but that a clearer understanding of the 

 relationships of these fishes awaits the time when the 

 critical internal and external features can be compared 

 for a larger number of species than it has been possible to 

 do here. Such a further study must also obviously in- 

 clude Chelonodon and Tetraodon, the generic limits of 

 neither of which are presently clear. 



For example, in Tetraodon the lateral line system is 

 highly variable, but it is probably basically represented 

 by a lower line which joins the upper and continues to the 

 tail, this being modified in some species by the loss of the 

 connection between the upper and lower lines, or even 

 the loss of the lower line, and perhaps of much of the up- 

 per line as well. The bifid nasal tentacle tends to be deep- 

 ly split, almost to the base, and the inner epithelium 

 smooth. The skulls of the two species studied are 

 relatively similar, differing mainly that in Uneatus the 

 sphenotics are more anteriorly expanded than in mbu 

 (probably partially attributable to the much greater size 

 of the study individual of Uneatus) and that the postero- 



