FISHERY BULLETIN: VOL. 84, NO. 3 



Two to five cartilaginous first dorsal fin pteryg- 

 iophores developed first above interneural spaces 

 3-5 at the time of ossification onset of the anterior- 

 most neural arches and spines. The addition of car- 

 tilaginous first dorsal fin pterygiophores was in a 

 posterior direction. Five cartilaginous second dor- 

 sal and anal fin pterygiophores developed second 

 simultaneously in the anterior portions of the future 

 second dorsal and anal fins. Some addition of carti- 

 laginous second dorsal and anal fin pterygiophores 

 occurred in an anterior direction, but most of the 

 addition was posteriorly. Dorsal and anal fin rays 

 and spines developed in the same sequence as their 

 corresponding pterygiophores, but a little later 

 (Table 2). 



Scomberomorus does not develop predorsal bones. 

 The first dorsal pterygiophore originated from one 

 piece of cartilage and inserted in the third inter- 

 neural space supporting two fin spines (one super- 

 numerary spine). The first anal pterygiophore 

 developed from two pieces of cartilage and sup- 

 ported three spines (two supernumerary spines). The 

 posteriormost nine dorsal and anal pterygiophores 

 had middle radials. The last dorsal and anal pteryg- 

 iophore supported a double finlet and had a non- 

 bifurcated stay (Table 1). 



In Scomberomorus, first caudal development of 

 the cartilaginous parhypural and hypurals 1 and 2 

 was concurrent with the anterior development of the 

 neural spines and the central appearance of haemal 

 spines. Hypurals 3-5 were added posteriorly, the two 

 autogenous haemal spines anteriorly. Hypurals 1 

 and 2 and hypurals 3 and 4 fused before ossifica- 

 tion onset to a cartilaginous ventral and dorsal 

 hypural plate. The dorsal and ventral plates fused 

 after ossification to a single hypural plate with a cen- 

 tral notch (Collette and Russo 1984). Hypural 5 

 gradually fused with the paired uroneural forming 

 an autogenous bone resembling a third epural and 

 mistaken as such by Leccia (1958). Two epurals 

 developed anterior to the uroneural-hypural 5. These 

 epurals remained autogenous. The neural and hae- 

 mal elements of preural centra 2, 3, 4, and 5 sup- 

 ported the procurrent caudal rays. A procurrent 

 spur and basally foreshortened ray were absent in 

 Scomberomorus (Johnson 1975) (Table 1). 



Only six Acanthocybium solanderi were available. 

 We were therefore unable to ascertain a complete 

 developmental sequence. Our smallest 6.2 mm NL 

 specimen had two cartilaginous development centers 

 along the notochord: some neural spines and arches 

 anteriorly and the parhypural, hypural 1-3 poste- 

 riorly. The next larger specimen 9.2 mm SL had all 

 neural and haemal arches and spines developed, thus 



we were unable to tell if in Acanthocybium four ini- 

 tial centers (as in Scomberomorus) or only three 

 centers (as in Xiphias and Sarda) of cartilaginous 

 development along the notochord were present. In 

 all our Acanthocybium specimens, hypurals 1 and 

 2 gradually fused before ossification onset to a ven- 

 tral cartilaginous hypural plate. In the 8.5 mm SL 

 Acanthocybium, hypurals 3 and 4 were fusing before 

 ossification onset; in the larger 9.5 and 10.4 mm SL 

 specimens hypurals 3 and 4 were ossifying while still 

 separate. The dorsal and ventral hypural plates were 

 fused in adults to one plate with a notch (Conrad 

 1938; Collette and Russo 1984) (Table 1). Ossifica- 

 tion of the vertebral column initially started in four 

 places and was similar to the ossification in Scomber. 

 The development of the dorsal and anal fins and 

 their supporting pterygiophores in Acanthocybium 

 was similar to that described in Scomberomorus. 



Tribe Sardini 

 Figure 12 



Ninety-nine Sarda sarda (2.4-9.0 mm NL or SL, 

 59-102 mm SL) were available. Of the larval speci- 

 mens (2.4-9.0 mm NL or SL) only 32 were larger 

 than 5 mm NL, and of these 10 were between 6.0 

 and 6.9 mm NL or SL, 6 were between 7.0 and 7.9 

 mm NL or SL, and 3 were larger than 8 mm SL. 

 Thus, since development of the vertebral column in 

 Sarda begins around 5 mm NL, only 32 specimens 

 were useful to our study and they were too few to 

 yield a complete developmental series. Our conclu- 

 sions on Sardini development are not as well sup- 

 ported as for most other scombroids. 



Development of the vertebral column in Sarda ini- 

 tially started in three places on the notochord: 1) 

 anterodorsad (neural arch and spine of future cen- 

 trum 1), 2) posteroventrad (parhypural, hypurals 1 

 and 2), and 3) ventrad at center (haemal arches and 

 spines, parapophyses). The anterior neural spines 

 were added in a posterior direction and the haemal 

 spines probably first appeared when the correspond- 

 ing neural spines developed above them at the 

 center of the notochord. Our evidence, however, is 

 only indirect, because one 7.5 mm NL specimen had 

 21 neural spines and no haemal spines, but our 8.1 

 mm SL specimen had all neural and haemal spines 

 developed. The cartilaginous hypurals were added 

 posteriorly, but we could not observe the anterior 

 addition of the autogenous haemal spines, although 

 we assume that it happens in Sarda as in other 

 scombroids with tails. Ossification of the vertebral 

 column in Sarda initially started in two places: ante- 

 riorly (neural arches and spines) and posteriorly 



668 



