-35' 



Total Larval 

 Captures 



Figure 3. — Numbers, size ranges, and mean lengths of swordfish larvae — collections from 

 various sources including Fisheries Research Board of Canada. 



by even the larger (50 mm) larvae, he estimated an 

 overall spawning period from the end of December 

 to the end of September over a large area — from the 

 lower Caribbean through the Yucatan Channel, the 

 Straits of Florida, and the Gulf Stream system 

 northwards to the South Carolina coast, i.e. from 

 about lat. 15°N to about lat. 32°N. 



The data presented herein for the most part sup- 

 port Arata's (19.'>4) conclusions, although they 

 cover only the period from January to March. 

 There are, however, a few discrepancies. Arata 

 (1954) suggests that the sizes of his specimens from 

 the northeast Gulf of Mexico further substantiate 

 the theory that spawning occurs in the lower Carib- 

 bean. For example, he concluded that, considering 

 the current structure, one 55.4 mm specimen in the 

 Gulf of Mexico would most likely have been 

 spawned somewhere south of Jamaica around the 

 first of March. However, sampling in the southern 

 half of the Caribbean from November to April pro- 

 duced no larvae (Ueyanagi et al., 1970). 



In his back calculations, Arata (1954) assumed 

 that the major currents moving north from the 

 Caribbean into the Gulf of Mexico do not swing 



farther west than long. 88°W. Thus, larvae would be 

 carried directly from the Caribbean into the north- 

 eastern Gulf. The pilot charts of the North Atlantic 

 and Sverdrup, Johnson, and Fleming (1942) show 

 that, while the major currents do flow directly 

 through the Straits of Florida (Fig. 4), the waters of 

 the Gulf of Mexico form independent eddies. It is 

 these secondary currents which flow into the north- 

 eastern Gulf, and which also swing farther west 

 than long. 88°W. The large larvae caught by Arnold 

 (1955) in the southwestern (mean = 38.6 mm) and 

 central (mean = 53.6) areas of the Gulf of Mexico 

 may have been spawned in the southwest part of the 

 Gulf and remained trapped there by the Gulf ed- 

 dies. On the other hand, the presence of these lar- 

 vae may indicate that secondary currents are suffi- 

 ciently strong to transport larvae from the Carib- 

 bean into the western reaches of the Gulf. Thus, 

 larvae from the Caribbean could take a longer route 

 to the northeast, initially via the more westerly cur- 

 rents. Back calculations for the large specimens 

 would then place their spawning areas somewhere 

 in the northwest Caribbean where several small lar- 

 vae have been found. 



256 



