THRESHER ET AL.: LARVAE OF GADOID. MACRURONUS NOVAEZELANDIAE 



the southern coast of mainland Australia (see Figure 

 3). Transects 1 through 8 consisted of 2-4 stations 

 (4 on average), depending upon the vddth of the con- 

 tinental shelf. These stations were designated "near- 

 shore" (at a depth of 30-50 m), "midshelf" (70-100 

 m), "shelf edge" (immediately offshore of the shelf 

 break and usually at a bottom depth of approximate- 

 ly 200 m) and "offshore" (1 nmi offshore of the sur- 

 face temperature/salinity front between inshore and 

 offshore water masses or, if no front was evident, 

 at 10 nmi offshore from the shelf edge station). In 

 the second year of the study, occasional samples 

 were collected at sites along the west coast between 

 regular transect lines, in order to improve the spa- 

 tial resolution of analyses and to increase sample 

 sizes. 



Two samplers were used: a rectangular midwater 

 trawl (RMT) 1 + 8 (see Baker et al. 1973 for de- 

 scription) and aim diameter ring net fitted with 

 a pivoting bridle system similar to the Tranter- 

 George plankton net (Tranter and George 1972). 

 Mesh sizes for the RMT-8 was 3 mm and 1 mm for 

 the net and cod end, respectively, and 333 pm 

 throughout for the RMT-1. The ring net consisted 

 of 500 \im mesh with a 333 \^m cod end. Initially, 

 all sampling was done with the RMT 1 -i- 8 in a fixed 

 open mode. Because it was difficult to fish the net 

 in rough seas and to calibrate its fishing character- 

 istics (see Pommeranz et al. 1982), the RMT 1 -t- 8 

 system was replaced after three cruises (April- 

 August 1984) with the more manageable ring net. 

 The ring net was subsequently used on transects 1 

 through 8, while the RMT system was retained for 

 study of the vertical distribution of larvae at tran- 

 sect 9. 



Each station consisted of a stepped oblique tow 

 made to a maximum depth of 200 m— bottom depth 

 permitting— parallel to bottom contours. The net 

 was fished at 10 m depth steps for three minutes 

 each at a vessel speed of approximately 2 knots. Net 

 depth was monitored continuously by a Simrad^ 

 trawl eye. The volume of water filtered was calcu- 

 lated using Rigosha B flowmeters, calibrated in a 

 flume tank. Reported catch rates are standardized 

 to numbers per 1,000 m^ of water filtered. Except 

 where specified below, sampling was not standard- 

 ized to time of day. 



Data on larval depth distributions were obtained 

 with the 1 m ring net off the west coast of Tasmania. 

 Sampling was conducted on 20 and 21 July 1986 



'Reference to trade names does not imply endorsement by the 

 National Marine Fisheries Service, NOAA. 



between transects 5 and 6, over a bottom depth of 

 100-120 m. As this site is close to the spawning area 

 of M. novaezelandiae, the catches consisted primar- 

 ily of small larvae. On each tow, the net was sent 

 to depth quickly, allowed to stabilize at the selected 

 depth for 1-2 minutes, and then retrieved slowly on 

 a continuous oblique path. Tows were made in the 

 order of progressively deeper depths. As each tow 

 integrated larval abundance to the maximum depth 

 of the tow, it was assumed that differences in stan- 

 dardized catch rates between adjacent strata re- 

 flected larval abundance in the depth range added. 

 Twenty-four tows were made, varying from 15 to 

 90 minutes and from 10 to 90 m depth. Tows were 

 made in six sets, three during the day (0830-1330) 

 and three at night (2300-0400). Sunrise and sunset 

 were at 0730 and 1700 (Australian Eastern Stan- 

 dard Time), respectively. 



Samples were divided by hand into two portions. 

 One portion was fixed in a buffered 3.7% aqueous 

 solution of formaldehyde and the other in 95% 

 ethanol. The former were used to identify larvae; 

 larvae in the ethanol-fixed samples were used for 

 ageing and assessment of growth rates. Larval 

 abundance data are based on both portions for each 

 station. The ages of M. novaezelandiae larvae were 

 determined by examination of otolith microstruc- 

 ture, following procedures outlined in Brothers et 

 al. (1976). Whole otoliths were extracted from the 

 larvae and viewed under transmitted light at 720- 

 2500 X using a Leitz Orthoplan microscope and high 

 resolution, closed-circuit television (Ikigami Model 

 CTC-6000). Otolith features were measured with a 

 sonic digitizer (Science Accessories Corporation 

 Graf/Bar) supported by an Apple 2e microcomputer 

 and a modified version of the Basic program 

 DISBCAL (Erie 1982). Viewed laterally, the otolith 

 measured (the lapillus) was virtually circular; all 

 measurements reported are to the point on the 

 perimeter farthest from the primordium (i.e., the 

 axis of maximum growth). Rates of larval growth 

 are uncorrected for shrinkage. Preliminary results 

 suggest shrinkage (TL) due to alcohol preservation 

 averages approximately 5% and is only weakly cor- 

 related with larval size (regression of percent shrink- 

 age against preshrinkage TL, slope = -0.005, 

 R- = 0.18, n = 27). Shrinkage will affect estimates 

 of absolute growth rates, but the available litera- 

 ture (Theilacker 1980; Fowler and Smith 1983) 

 suggests it should not bias comparisons between 

 growth rates, provided the larvae being compared 

 were collected and fixed in the same manner. 

 Statistical analyses were done using Statview 512-1- , 

 Vers. 1.1. 



31 



