Govoni et al,: Early life history o\ Xiphias gladius in the western North Atlantic 



779 



Arata (1954) inferred the age and growth of larvae from 6 

 to 192 mm preserved total length (PTT). Aside from this 

 effort, the age of larval swordfish has been undetermined 

 and growth has not been described. Diets of lai-vae have 

 been reported (Gorbunova, 1969), but the apparent transi- 

 tion in diet has neither been detailed nor reconciled with 

 changes in physical features and growth. Similarly, the 

 vertical distribution of larvae has not been reconciled with 

 their diet or growth. Most larvae are collected near the 

 surface of the ocean, typically in neuston nets, but some 

 larvae are collected in nets that sample below the surface 

 (Govoni et al., 2000). 



In the present study, we resolve and summarize the early 

 life history of swordfish in the western North Atlantic. We 

 report estimated age, describe growth, relate growth to 

 feeding, morphological features, and vertical distribution, 

 and infer spawning time and location and the sources and 

 fates of swordfish larvae. This study supplements that of 

 Govoni et al. (2000) by providing the dimension of time, 

 i.e. age of larvae, to the spatial distribution and possible 

 transport of these larvae. 



Methods 



Collections of larvae 



Ichthyoplankton collections from cruises in 1989 (in the 

 northeastern Caribbean about the Lesser Antilles), 1991 

 and 1997 (off the southeastern United States), and 2000 

 (in the Straits of Florida and off the southeastern United 

 States), produced 63 larvae that were preserved in 95% 

 ethanol (for examination of otolith microstructure). Sam- 

 ples were collected either from the neuston (i.e. taken with 

 a 1.0x0.5 m neuston net) or from depth intervals (i.e. taken 

 with a 1-m MOCNESS [multiple opening and closing net 

 and environmental sampling system] [Wiebe et al., 1976]). 

 Larvae were measured for preserved standard length 

 (PSL), the conventional length measure for larval fishes 

 (Kendall et al, 1984), and lower-jaw-fork-length (LJFL), 

 the measure in common use for juvenile and adult sword- 

 fish (Megalofonou et al., 1995). 



Otolith excision and examination 



Of the 63 larvae collected, sagittae were found and success- 

 fully excised from 37 larvae, lapilli from 32, and asterisci 

 from six. Otoliths were mounted on glass slides and dried 

 before examination. Broken sagittae and lapilli, and some 

 large sagittae, were embedded in plastic, sectioned with a 

 saw, and polished (Secor et al., 1991). 



Otolith growth increments were counted along the lon- 

 gest axis of each sagitta and lapillus by using a compound 

 light-transmission microscope; increments in asterisci 

 were fewer and less defined and were not counted. Three 

 blind counts were made by the same observer. Although 

 increments were consistently visible on both sagittae 

 and lapilli, counts from individual larvae were greater on 

 sagittae (Student's Mest; P<0.001). Standardized counts 

 (the standard deviate of each repeated count) on the right 



and left sagittae were not significantly different (nested 

 ANOVA; P<0.05). Increment counts from either the left or 

 right sagitta, decided by coin toss, were used for age and 

 growth rate determination. The mean of three replicate 

 counts was rounded to a whole number. 



Increment counts from sagittae were used to estimate 

 larval age. Increments were assumed to form daily (Cam- 

 pana, 2001). The core increment was assumed to form at 

 hatching (Jones, 1986). The first increment outside of the 

 core increment was counted as one. Age from hatching 

 (AFH) was the number of increments counted from the 

 core increment on sagittae. This definition differs from 

 that of Prince et al. (1991) who counted the core increment 

 as increment one for the istiophorid blue marlin (Makaira 

 nigricans) and for a single larval swordfish that was 8.5 

 PSL. The radius of each otolith was measured by image 

 analysis. 



Growth model 



The best empirical fit among a suite of regressions of 

 estimated age (AFH) and length (PSL) — linear, polyno- 

 mial, and piece-wise, and moving — was chosen to describe 

 somatic growth (Forbes and Lopez, 1989; Hare and Cowen, 

 1995; Rogers et al., 2001). Criteria for best fit were the 

 following: the interpretation of fit from graphical display; 

 regression coefficients (r^), and dispersion or convergence 

 of regression residuals. The model of best fit was also 

 applied to the estimated age and lower jaw fork length 

 (LJFL) to allow comparison with published accounts of 

 juvenile swordfish growth. 



Diet 



Sixty-eight specimens from the present collections and 

 from Govoni et al. (2000), all with undamaged alimentary 

 canals, were examined for gut contents. These specimens 

 were taken in the northeastern Caribbean (3 specimens), 

 Gulf of Mexico (5), and off the southeastern coast of the 

 United States (60). Food was identified to the lowest taxon 

 possible following Govoni et al. (1983). 



Physical features of larvae 



Histological sections of three larvae, 21.5, 30.0, and 

 52.0 mm PSL, were cut as a preliminary aid to the loca- 

 tion of otoliths within the cranium and to determine the 

 histological constitution of the larval alimentary canal. 



Time and location of spawning 



Spawning dates were estimated from the ages of larvae 

 (estimated from the growth model (AFH)), plus 3 days 

 (the incubation period at 25°C for swordfish eggs given 

 by Yasuda et al. [1978]). Spawning dates are thus days 

 from fertilization (DFF). Spawning location was inferred 

 by applying DFF to larval swordfish lengths reported in 

 the present study, as well as lengths given in Govoni et al. 

 (2000), taking into consideration the time that eggs and 

 larvae were at large and adrift (DFF) and the location where 



