Figure 225.— External features of 



other representative tetraodontid genera: 



Carinotetraodon torteti— upper left, nasal 



region as seen externally; lower left, scales 



from upper middle region of body. 



kretamensis) the course of the lateral line was found to 

 be impossible to trace posteriorly in the single alcohol- 

 preserved specimen examined. In Carinotetraodon a 

 well -developed lower lateral line is present, but, at least 

 in the few specimens examined, it does not join with the 

 upper lateral line, continuing on directly to the tail (see 

 illustration). In Ephippion a lower lateral line is present, 

 but its juncture with the upper line is highly variable 

 within the single species, as is the length of its extension 

 anterior to the point of juncture (see illustrations). None 

 of the genera with a single nostril or nasal tentacle have 

 an upraised ridge of skin ventrolaterally along the caudal 

 peduncle. 



In short, it can be said that, of those genera with two 

 nostrils, Sphoeroides and Guentheridia have only a 

 single well-developed lateral line on the body, while in 

 Lagocephalus, Amblyrhynchotes, Fugu, Torquigener, 

 and Colomesus two lateral lines usually are present, and 

 that of those genera with a single nostril or nasal ten- 

 tacle, Arothron has a single lateral line, while 

 Chelonodon, Monotreta, and Ephippion have two which 

 usually merge on the caudal peduncle, Carinotetraodon 

 usually has two which do not merge, and Tetraodon has 

 either two which merge or, in at least a few species, a 

 single lateral line. 



There is also intraspecific variation in the form of the 

 nasal apparatus, although probably less than that 

 described here subsequently for the Diodontidae. In one 

 species with two nostrils {Amblyrhynchotes richei), 9 

 (18.9-47.3 mm, ANSP 109916) out of 10 specimens ex- 

 amined had the normal arrangement of two nostrils 

 separated by a broad band of tissue, but one specimen 

 (49.1 mm, ANSP 109916) had apparently resorbed the 

 middle region of the band in both the right and left nasal 

 apparatuses to leave the nostrils confluent, and only in- 

 completely separated by two flaps of skin (see illus- 

 tration). 



In the genera with a single nostril or nasal tentacle, the 

 olfactory epithelium is usually relatively smooth, but in 

 at least a few species (see illustrations of Arothron im- 

 maculatus and Chelonodon fluviatilis) the inner surfaces 

 of the bifid tentacles are deeply pitted. 



In the genera with two nostrils, the olfactory 

 epithelium ranges from smooth to highly folded into 

 numerous successive lamellae, but in the monotypic 

 Guentheridia formosa the olfactory epithelium is deeply 

 pitted just as in some of the species with bifid tentacles 

 (all of the relatively normal species oi Arothron and one 

 of the species of Chelonodon). The degree of lamellar 

 development is variable within some genera. For exam- 

 ple, in Lagocephalus lamellae are exceptionally well 



developed in all species as folds that cover the entire in- 

 ner surface of the nasal sac, with one or two of the folds 

 on the rear wall enlarged as protruding flaps which them- 

 selves bear lamellae, except in L. scleratus, in which 

 folds are present only on the rear wall of the sac and the 

 two prominent protruding flaps are not themselves ex- 

 tensively folded. Similarly, in Sphoeroides, most species 

 have the olfactory epithelium either smooth or with only 

 one or two prominent folds or horizontal ridges on the 

 rear wall, but one species, S. pachygaster, has the folds 

 on the rear wall much better developed and more 

 numerous, between five and eight in number. While the 

 habitat of L. scleratus is similar to that of some (e.g., the 

 equally streamlined and open water L. lagocephalus) of 

 the other species of Lagocephalus with extensive olfac- 

 tory flap development, S. pachygaster, in contrast to all 

 other species of Sphoeroides, is a deepwater species, 

 which may rely more heavily on olfaction than do the 

 shallower water species. In Fugu the folds are only slight- 

 ly less well developed than in most Lagocephalus and oc- 

 cur on the entire inner surface of the nasal sac, although 

 the larger folds on the rear wall never become so exten- 

 sively folded themselves as they do in most species of 

 Lagocephalus. The numbers and size of the lamellae are 

 especially great in F. oblongus. In the species of 

 Colomesus, Torquigener, and Amblyrhynchotes ex- 

 amined, one to about four folds or ridges are developed 

 on the rear wall of the sac. 



There is variation in the size of the prickly spines and 

 in their coverage of the body. Among the materials ex- 

 amined there are four genera in which most of the species 

 have prickles but in which one or two species are com- 

 pletely spineless. In the speciose Sphoeroides most 

 species have spines on both the back and belly but two 

 species are spineless, S. angusticeps in the eastern 

 Pacific and S. pachygaster (of which Liosaccus cutaneus 

 is one on many synonyms, according to Shipp 1974:44- 

 47) in the Atlantic and Indo-Pacific to Japan. These two 

 species are not closely related within the genus, as dis- 

 cussed subsequently, and the loss of prickles undoubted- 

 ly has occurred independently in the two species. More- 

 over, Shipp and Yerger (1969a) and Shipp (1974:98) 

 reported that a minority of specimens of S. nephelus are 

 completely spineless, and nephelus does not seem closely 

 related to either pachygaster or angusticeps. 



In the equally speciose Fugu, most species have spines 

 on both the back and belly but two species are spineless, 

 F. chrysops and F. uermicularis, which on the basis of the 

 osteological data presented by Kuronuma (1943) do not 

 seem closely related within the genus and are placed in 

 different subgenera by Abe (1952). Here again, the loss of 



