THRESHER ET AL.: LARVAE OF GADOID, MACRURONUS NOVAEZELANDIAE 



•a 



3 



T3 



o -p 



'55 ? 



« o 



OJ u 



C 



o H 

 t/) c 



5 10 15 20 25 30 35 



Mean Age of Larvae (days postfirst-feeding) 



40 



Figure 11.— Relationship between age olMaeruronus rwvaezelandiae larvae examined (pooled 

 into 3 d increments) and the slope of the regression between relative growth rate (residual 

 from semilog regression of In total length against age) and transect number. The correlation 

 between slope of the regression line and age class of larvae is significant at P < 0.01. 



negative; relative growth rates were highest on the 

 west coast, near the spawning area, and declined 

 to the south and east (Fjioe = 20.25, P < 0.001) 

 (Fig. 12). In contrast, for larvae older than approx- 

 imately 25 d postfirst-feeding, the slope was 

 significantly positive; relative growth rates were 

 lowest on the west coast, increased towards the east 

 coast (Fi 124 = 25.58, P < 0.001), and were highest 

 farthest from the spawning area (Fig. 12). The tran- 

 sition from a negative (west coast fastest) to a 

 positive (east coast fastest) slope occurs at a larval 

 age of approximately 17-22 d postfirst-feeding. 



The second factor confounding the correlation 

 between distance from the spawning ground and 

 relative growth rates is an apparent seasonal change 

 in the strength of the correlation, particularly for 

 older larvae. A correlation between distance from 

 the west coast and relative growth rates of larvae 

 accounts for 27% of the variance in growth residuals 

 for larvae aged more than 25 d postfirst-feeding in 

 the early, slow-growth portion of the 1985 spawn- 

 ing season. By August, however, during the period 

 when larval growth rates were uniformly high, the 

 correlation accounts for only 10% of the variance 

 in growth rates and, by the end of the spawning 

 season, for larvae hatched after 25 August, the rela- 

 tionship between location and growth rates for these 

 older larvae disappears altogether (i?^ ■= 0.02). 

 There are insufficient data for a comparable anal- 

 ysis of seasonal changes in growth rates of older 

 larvae in 1984. 



DISCUSSION 



The increase in mean age and size of larvae with 

 increasing distance from the west coast, the pattern 

 of drift card returns, and the distribution of surface 

 isotherms on the west coast of Tasmania during 

 winter all support the hypothesis that larval M. 

 novaezelandiae are transported by longshore cur- 

 rents from a spawning ground on the west coast to 

 the southeastern and eastern coasts. This hypothe- 

 sis is also supported by independent studies of the 

 physical oceanography of the west coast. A south- 

 ward flowing, longshore current off the west coast 

 in winter was first suggested by Newell (1961); drift 

 bottles he deployed off the coast moved in a similar 

 pattern to our drift cards. Subsequently, Baines et 

 al. (1983) inferred the presence of this current from 

 a shelfward depression of isotherms and confirmed 

 it by the drift pattern of a drogue released off the 

 northwestern coast. Baines et al. (1983) reported the 

 Zeehan Current, as they named it, to be relatively 

 narrow (approximately 40 km wide) and restricted 

 largely to the edge of the continental shelf. It moves 

 southwards at a depth averaged flow in the order 

 of about 20 km/d (C. Fandry, pers. commun."). This 

 figure is reasonably consistent with our data on lar- 

 val ages at different points along the advection 

 route. The distance between the spawning ground 



'C. Fandry, CSIRO Division of Oceanography, GPO Box 1538, 

 Hobart, Tasmania 7001, Australia, pers, commun. June 1987. 



41 



