FISHERY BULLETIN: VOL. 80, NO. 2 



bones in Xiphias. Hypurals 1-4 and the urostyle 

 fuse to one posteriorly notched hypural plate 

 during development (Fig. 23); the three epurals, 

 the uroneural pair, hypural 5, and the par- 

 hypural remain autogenous, whereas in Thun- 

 nini and Sardini only one epural remains 

 autogenous and the paired uroneural fuses to the 

 urostyle (Collette and Chao 1975; Potthoff 1975; 

 Collette and Russo 1978). In Xiphias, hypurals 1- 

 4 develop initially from distinctly separate 

 pieces of cartilage and fusion of the hypurals into 

 the notched hypural plate occurs. In Scombridae 

 a similar yet different development takes place, 

 because in Thunnini hypurals 1 and 2 originate 

 from one distinctly larger piece of cartilage, 

 whereas in Scomber (Pneumatophorus), hypurals 

 1 and 2 originate from separate pieces of 

 cartilage as in Xiphias (Kramer 1960). 



The caudal rays in adult scombrids, except 

 Scombrini, cover the whole hypural plate 

 (Collette and Chao 1975; Collette and Russo 

 1978), whereas in Xiphias a smaller area is 

 covered by the rays (Figs. 17, 22, 23). When the 

 rays are disarticulated from the hypural plate in 

 adult Xiphias, long vertical depressions caused 

 by the rays can be observed on the hypural plate 

 (Fig. 23). 



Xiphias has a greater number of precaudal 

 than caudal vertebrae (Fig. 6) (Leim and Scott 

 1966; Ovchinnikov 1970). The same tencency was 

 observed in the gempylids (Matsubara and Iwai 

 1958; Potthoff et al. 1980) and the opposite 

 tendency in the scombrids (Conrad 1938; de 

 Sylva 1955; Mago Leccia 1958; Kramer 1960; 

 Gibbs and Collette 1967; Matsui 1967; Potthoff 

 and Richards 1970; Collette and Chao 1975). 

 Generally, the tendency in the perciform fishes is 

 to have a higher caudal vertebral count; the most 

 typical count being 10+14=24 vertebrae (John- 

 son 1981). 



The neural and haemal arches in Xiphias first 

 develop distally opened (split) (Fig. 27). Dur- 

 ing development the neural and haemal arches 

 fuse forming spines. Fusion of the neural and 

 haemal spines proceeds from posterior in an ante- 

 rior direction (Table 13). In other perciforms 

 studied by Potthoff, split arches were sometimes 

 observed on small larvae on the anteriormost 

 first and second centra only, but these two arches 

 fused to spines during development. Adult 

 Xiphias retain three to six anteriormost split 

 neural arches (Bruce B. Collette footnote 3). 



Rib development and position is unique in 

 Xiphias. Commonly, perciforms have pairs of 



dorsal (epipleural) ribs on the precaudal verte- 

 brae starting on the first centrum and pleural 

 ribs starting on the third centrum (Houde and 

 Potthoff 1976; Potthoff et al. 1980). These ribs 

 develop from anterior in a posterior direction. 

 Xiphias, however, has lost many of its ribs. 

 Generally, there are only one pair of ribs on each 

 of the first four centra, which develop from 

 anterior in a posterior direction and one pair on 

 the last two precaudal vertebrae which develop 

 from posterior in an anterior direction. We do not 

 know if the ribs in Xiphias were originally 

 epipleural, pleural, or a combination of epi- 

 pleural and pleural. We were able to determine, 

 however, the cartilage origin of ribs in Xiphias. 

 Tibbo et al. (1961) stated that ribs in adult 

 Xiphias are short and poorly developed, but no 

 details on rib position were given. 



An account of rib development in lower and 

 higher fishes is given by Emelianov (1935). He 

 found that some bony fish develop ribs from 

 cartilage, in others rib development from 

 cartilage is bypassed and ribs develop directly 

 from bone cells, and still in others, parts of the 

 ribs develop from cartilage and other parts of the 

 same rib develop directly from bone. In Xiphias 

 the proximal portions of each rib originate from 

 cartilage, the distal portions develop directly as 

 bone. 



The branchiostegal ray count in Xiphias may 

 vary by one ray from specimen to specimen or it 

 may vary between left and right sides in a speci- 

 men. Usually, branchiostegal ray counts are 

 conservative and characterize fish families and 

 sometimes genera (Kishinouye 1923; McAllister 

 1968; Fraser 1972; Ahlstrom etal. 1976; Kendall 

 1979; Matsuura 1979), however variability has 

 been reported in some groups such as 

 Carangidae (McAllister 1968). 



We cannot make firm conclusions about the 

 phylogenetic status of Xiphias. From our study 

 we conclude that Xiph ias is a perciform fish that 

 differs from other perciforms to warrant the 

 separate suborder Xiphiioidei. We were unable 

 to determine relationship with the scombroids 

 (gempylids, scombrids). A comparison with 

 istiophorids remains to be done, and we believe 

 we furnished sufficient material to facilitate 

 such a comparison. 



ACKNOWLEDGMENTS 



We thank Bruce B. Collette, Edward D. 

 Houde, G. David Johnson, Izumi Nakamura, 



184 



