GovonI et al.: Distribution of larval Xiphias gtadius off the southeastern United States 



71 



one 3.8 mm SL (4.3 mm TL) taken off North Carolina 

 (CH9703), had recognizable yolk and oil globule rem- 

 nants. These lengths were smaller than the reported 

 length of larvae at the completion of yolk and oil glob- 

 ule absorption, about 5 mm TL for larvae (measured 

 alive, 65 h after fertilization ) from the Mediterranean 

 (Yasuda et al., 1978). 



Counts of increments on sagittae from larvae 4, 5, 

 and 6 mm SL, taken in CH9703, were 4, 3, and 6 

 (increments on swordfish otoliths have not been vali- 

 dated as daily intervals [Price et al., 1991] ). 



Discussion 



North of the Florida Straits, larval swordfish were 

 collected most frequently within the western frontal 

 zone of the Gulf Stream. This observation corroborates 

 the speculation that larvae are associated with water 

 within sharp horizontal gradients of temperature and 

 salinity (Tibbo and Lauzier, 1969). Convergence of sur- 

 face water is a possible mechanism for their accumu- 

 lation within the Gulf Stream front. The Gulf Stream 

 front south of Cape Hatteras is cyclonically sheared 

 with shelf water that directly opposes Gulf Stream 

 water (Pietrafesa et al., 1985). The retrograde, hydro- 

 graphic discontinuity between Gulf Stream and shelf 

 water and their hydrodynamic opposition results in 

 convergence of surface water within the frontal zone 

 (Garvine, 1980; Olson et al., 1994). Convergence of 

 surface water has been implicated in the accumula- 

 tion of adult fishes with depth-keeping abilities (Ol- 

 son and Backus, 1985). Positively buoyant or surface- 

 seeking larval fishes will be entrained in converging 

 water and will be advected toward the front where 

 they will accumulate as they resist downwelling along 

 the frontal axis (Govoni and Grimes, 1992). Sword- 

 fish larvae are unquestionably surface-seeking larvae. 

 Convergence of surface water within oceanic frontal 

 zones should produce accumulations of larvae on spa- 

 tial scales of 10 to 100 km (Olson et al., 1994). At a 

 coarse scale, larvae were more abundant within the 

 frontal zone; no fine-scale pattern was evident within 

 the frontal zone. Adaptive sampling (Lo et al., 1997) 

 may be a more efficient means of estimating larval 

 swordfish abundance than simple random sampling, 

 because of the rarity and the spatial aggregation of 

 larvae. 



Small swordfish larvae were collected most often 

 in the eastern Gulf of Mexico and off the east coast 

 of the United States as far north as Cape Lookout, 

 North Carolina. Swordfish apparently spawn in the 

 eastern Gulf, but the present observations corroborate 

 the speculation of spawning off the Carolinas (Squires, 

 1962; Markle, 1974) as well. Off the Carolinas, larvae 



Station number 



200 



200L 



Figure 6 



Sections of isotherms along transects occupied in September 

 1988: (A) 13 September; (B) 14 September; (C) 1.5 Septem- 

 ber; (D) 16 September; (E) 17 September; (Fl 18 September 

 (bars above station numbers indicate the Gulf Stream fron- 

 tal zone). 



5 mm SL or less occurred in 25° and 26°C water. Lar- 

 vae that were 4 to 5 mm SL had 3 and 4 apparent 

 growth increments on their sagittae. In water from 

 22° to 25°C, larvae that were 5 mm TL would be ap- 

 proximately 6 d old (Yasuda et al., 1978). Given an egg 

 incubation period of 3 d at 24°C (Yasuda et al., 1978) 

 and an additional 3 or 4 d for posthatch growth, along 

 with a average axial velocity of the Gulf Stream of 1.5 



