makes only slight contact with the posterior edge of the 

 urostyle. The anterior end of the second hypural rests 

 against the lower posterior edge of the urostyle, while its 

 dorsal edge is slightly expanded laterally along the an- 

 terior one-third of its length into a concave articular facet 

 for contact with the anterior end of the third hypural, as 

 explained above. The ventral edge of the second hypural 

 articulates with the dorsal edge of the first hypural by fi- 

 brous tissue posteriorly but more anteriorly by slight in- 

 terdigitation. The first hypural articulates dorsally with 

 the second hypural as just described, while anteriorly it 

 rests against, and possibly slightly interdigitates with, 

 the posteroventral edge of the last centrum. The first 

 hypural articulates ventrally with the dorsal edge of the 

 parhypural by fibrous tissue posteriorly but more an- 

 teriorly by slight interdigitation. The parhypural is ob- 

 viously the haemal arch and spine of the last, or 11th, 

 caudal vertebra, but, like the haemal arches and spines 

 of the 9th and 10th caudal vertebrae, it is autogenous. 

 Dorsally the parhypural articulates with the first hypural 

 through fibrous tissue and interdigitation, as described 

 above. The portion of the dorsolateral surface of the 

 parhypural that lies below the anterior one-third of the 

 first hypural is laterally expanded into a flange to which 

 caudal fin-ray muscles are attached. A similar, but much 

 smaller, lateral flange is present from the anterior edge of 

 the first hypural. The anterior end of the flange from the 

 first hypural overlies the anterodorsal surface of the 

 flange from the parhypural. The combined flanges from 

 the parhypural and first hypural compose the hypura- 

 pophysis. The anterior half of the dorsal articular sur- 

 face of the parhypural is concave, so that where it inter- 

 digitates with the last centrum and with the anteroven- 

 tral end of the first hypural, a canal is formed in the mid- 

 line of the otherwise interdigitated surfaces between the 

 three bones. The canal opens to the exterior posteriorly 

 at a slight gap in the articulated surfaces of the parhy- 

 pural and first hypural just behind the lateral flange of 

 the parhypural. The canal is continuous anteriorly with 

 the haemal canal of the preceding vertebrae. 



Caudal fin rays. — Twelve principal (branched rays 

 plus the unbranched ray directly above and directly 

 below the series of branched rays) fin rays preceded by 

 eight (smaller specimen) or nine procurrent rays above 

 and six (smaller specimen) or seven procurrent rays 

 below. The 10 branched rays become increasingly 

 branched toward the middle 2 rays, which are branched 

 in triple dichotomies. The uppermost branched ray and 

 the lowermost branched ray are the longest of the caudal 

 fin rays, while the two middle rays are the shortest, being 

 about one-half the length of the longest branched rays. 

 The two unbranched principal caudal rays are some- 

 what shorter than the longest branched rays, and the 

 procurrent caudal rays rapidly decrease in length an- 

 teriorly in the series. The first two or three procurrent 

 rays are without cross-striations, while all the other 

 procurrent rays and principal rays are cross-striated. The 

 upper unbranched principal caudal ray articulates with 

 the second uroneural; the lower unbranched principal 



caudal ray with the parhypural. The upper three and 

 lower two branched rays of the dorsal lobe articulate, 

 respectively, with the fourth hypural and with the third 

 hypural. The upper two and lower three branched rays of 

 the ventral lobe articulate, respectively, with the second 

 hypural and with the first hypural. All the rays ar- 

 ticulate by fibrous tissue with the supporting elements. 



DORSAL AND ANAL FINS. 



Spines and pterygiophores. — The presence or ab- 

 sence of a spiny dorsal fin in T. macropterus has been 

 discussed by Tyler (1962a:794-796; 1967:92-93), it being 

 shown that most specimens from Japan to Indonesia 

 have a minute spiny dorsal fin, usually of two spines, 

 while the specimens now known from the Indian Ocean 

 lack the spiny dorsal fin. The smaller of the two 

 specimens studied for the present work, SU 13747, from 

 the Philippines, had been skinned, then cleared and 

 stained, and then dissected by the author prior to his dis- 

 covery of a spiny dorsal fin in other whole specimens, and 

 it is impossible to say, upon reexamination of the cleared 

 and stained SU 13747 specimen, whether spines had 

 been present or not. Being from the Philippines, it 

 probably did have spines which were lost during its 

 processing as an osteological preparation. The larger 

 study specimen, ANSP 98917, from the Volcano Islands 

 south of Japan, was processed with the knowledge of its 

 possession of a spiny dorsal fin. The lateral view illus- 

 tration of the entire skeleton presented here is based on 

 SU 13747, except that the dorsal spines are based on 

 ANSP 98917, as are the detailed lateral and dorsal views 

 of the spiny dorsal fin. 



When dorsal spines are present, the first is longer than 

 the second, and both can be laid back in a shallow groove 

 in the skin in the midline of the body. The bases of the 

 two spines are relatively close together, the first ar- 

 ticulating with the posterodorsal surface of the long first 

 basal pterygiophore ( = first supraneural when no spiny 

 dorsal fin is present), and the second spine articulating 

 in the middle of the much shorter second basal pterygio- 

 phore. From the anterolateral surface on either side of 

 the base of the first spine a tendon runs anteroventrally 

 along a groove in the dorsal and lateral surfaces of the 

 first basal pterygiophore to connect to a small muscle 

 whose contraction causes the erection of the spine. The 

 second spine does not have such muscles and tendons as- 

 sociated with it, but a low membrane connects the first 

 and second spines so that the erection of the first spine 

 causes the erection of the second as well. The ventral 

 ends of both spines articulate by fibrous tissue to the 

 relatively flat surfaces of their basal pterygiophores. 



There are rudiments of what is probably another dor- 

 sal spine in the form of two nubbins of bone, one to either 

 side of the midline, lying beneath the skin on the dorsal 

 surface of the first basal pterygiophore immediately in 

 front of the base of the first relatively well-developed dor- 



