POTTHOFF ET AL.: DEVELOPMENT OF SCOMBROID FISHES 



neural spaces 13-16. The anal pterygiophores first 

 developed in a group above the future anterior part 

 of the anal fin below interhaemal spaces 16-18. Fur- 

 ther addition of cartilaginous dorsal and anal pte- 

 rygiophores was in an anterior and posterior direc- 

 tion. The posterior pterygiophore additions dorsad 

 and ventrad were completed before the anterior 

 additions ceased. The full complement of anal pte- 

 rygiophores was reached before the full dorsal com- 

 plement. Dorsal and anal fin rays first originated 

 in the same areas as the pterygiophores, but at 

 larger sizes with addition of rays in the same direc- 

 tions (Table 2). 



Xiphias did not have predorsal bones. The first 

 dorsal pterygiophore originated from one or two 

 pieces of cartilage and inserted in the second inter- 

 neural space, supporting from one to three fin 

 spines. The first anal pterygiophore developed from 

 one or two pieces of cartilage, supporting from one 

 to three fin spines. Xiphias had no middle radials 

 in the dorsal or anal pterygiophores, but a double 

 ray and a nonbifurcated stay were associated with 

 the posteriormost dorsal and anal pterygiophores 

 (Potthoff and Kelley 1982) (Table 1). 



In Xiphias, cartilaginous hypurals were first seen 

 before precaudal neural spine development was com- 

 plete, but after dorsal and anal pterygiophore devel- 

 opment had started. The hypural complex develop- 

 ment was described by Potthoff and Kelley (1982). 

 Hypurals 1-5 and the parhypural developed from 

 separate cartilages, and there was no cartilage fu- 

 sion. There were three epurals and one uroneural. 

 Only one autogenous haemal spine was present on 

 preural centrum 2. In adults the three epurals, the 

 uroneural, hypural 5, and the parhypural remained 

 autogenous, but hypurals 1-4 fused with each other 

 and the urostyle forming a notched hypural plate 

 (Gregory and Conrad 1937). The neural and haemal 

 elements of only preural centrum 2 supported the 

 procurrent caudal rays. A procurrent spur and basal- 

 ly foreshortened ray were absent in Xiphias (John- 

 son 1975) (Table 1). 



DISCUSSION AND CONCLUSION 



Developmental features observed in our study are 

 illustrated in Figures 4-5 and 9-15. These features 

 along with meristic and osteological characters are 

 compared among the six scombroid families and the 

 primitive percoid Morone in Tables 1 and 2. Al- 

 though our conclusions are still preliminary because 

 of lack of adequate developmental series for many 

 genera, some comparisons, based largely on devel- 

 opment, are worth noticing. 



There are three major kinds of early development 

 and addition of the cartilaginous neural and haemal 

 arches and spines along the notochord. Each kind 

 may differ slightly between taxa. Scombrolabrax, 

 Scomber (Scombrini), Scomberomorus (Scombero- 

 morini), and Thunnini have one kind in which there 

 are four initial developments on the notochord, but 

 not necessarily in the given order, e.g., anteriorly 

 dorsad, centrally dorsad, centrally ventrad, and pos- 

 teriorly ventrad with a subsequent merger of the 

 initial areas. Gempylidae, Sarda (Sardini), Istio- 

 phoridae, and Xiphiidae have a second kind in which 

 there are three initial developments, e.g., anterior- 

 ly dorsad, centrally ventrad, and posteriorly ven- 

 trad; then the addition is from anterior in a poste- 

 rior direction with a merger in the posterior, near 

 the hypural complex. Trichiurus, which lacks hypu- 

 rals, has the third kind in which there are two ini- 

 tial developments, e.g., anteriorly dorsad and cen- 

 trally ventrad with addition in a posterior direction. 

 We could not fully determine the cartilaginous 

 development for Acanthocybium, because of an in- 

 complete series, and for trichiurids with tails, 

 because a series was lacking. 



In the Scombrolabracidae, Gempylidae, and Scom- 

 bridae, the vertebrae first develop by coalescence 

 of saddle-shaped ossifications positioned dorsad and 

 ventrad. We were not able to observe saddle-shaped 

 ossification in Acanthocybium because we lacked 

 specimens. The other scombroid families, Trichiu- 

 ridae (Trichiurus), Istiophoridae, and Xiphiidae, and 

 the primitive percoid Morone did not have these 

 saddle-shaped ossifications. Saddle-shaped ossifica- 

 tions have been observed during ontogeny in other 

 perciform fish such as Enchelyurus brunneolus 

 (Blenniidae) by Watson 5 and Lutjanus campechanus 

 (Lutjanidae) by Potthoff and Kelley 6 . We are unable 

 to comment at this time on the significance of these 

 saddle-shaped ossifications until the ontogeny of 

 many more taxa is studied. 



In the Scombrinae two species belonging to two 

 different tribes share a peculiar ossification se- 

 quence not observed by us in any other scombroids. 

 Both in Scomber (Scombrini) and Acanthocybium 

 (Sardini), initial ossification of the neural and haemal 

 arches and spines and the hypural complex started 

 at four locations on the vertebral column (Kramer 



6 Watson, W. Larval development of Enchelyurus brunneolus 

 from Hawaiian waters (Pisces: Blennidae: Omobranchini). Un- 

 publ. manuscr. Marine Ecological Consultants of Southern 

 California, 533 Stevens Avenue, Soloma Beach, CA 92075. 



6 Research on the development of Lutjanus campechanus is in 

 progress at the Southeast Fisheries Center Miami Laboratory, Na- 

 tional Marine Fisheries Service, NOAA, 75 Virginia Beach Drive, 

 Miami, FL 33149. 



675 



