ROTHLISBERG: VERTICAL MIGRATION OF PENAKII) SHRIMP LARVAE 



the water column by day but spread out to both 

 the surface and the bottom strata at night. Again 

 the postlarval numbers were low and little day- 

 night patterning was evident. No postlarvae 

 were caught at any time in the surface stratum 

 and were near the bottom both day and night. 



Of concern was possible day-night variation in 

 larval abundance due to avoidance of the inlet 

 hose and/or larval distribution outside the range 

 of the sampler. Of particular concern was the 

 possibility that during the day larvae were on or 

 near the bottom below the hose inlet at its low- 

 est extent. To test statistically for a temporal 

 variation in larval abundance, the larval num- 

 bers were initially combined over all depths 

 within a sampling time interval and then a 

 square root transformation was applied. A sine 

 curve was fitted to estimate gradual rather than 

 abrupt day-night changes in larval numbers. 

 Postlarval numbers were too low to include in the 

 analysis. Time alone was highly significant for 

 the total number of larvae and significant at 

 lower levels for three of six larval substages 

 (Table 1 ). The high level of significance of the sta- 

 tion-time interaction for all but the Z2 larvae in- 

 dicates that the small diel variation in abun- 

 dance was variable between sampling occasions. 

 Inspection of the data showed that the peak abun- 

 dances varied by stage, location, and time of day 

 and that there was no systematic difference in 

 catchability which would bias the interpretation 

 of the diurnal patterns shown previously. 



To add confidence to the diagrammatic inter- 

 pretation (Figs. 3-5), further analysis, using an 

 arc-sine transformation of the proportional lar- 

 val abundances by depth also, showed that the 

 abundances at the four depth strata were quite 

 variable from one date to the next. This was indi- 

 cated by the high degree of significance of the 



station effects at almost all depths (Table 2). The 

 abundances of larvae in the surface stratum 

 (level 1) showed the most consistent relationship 

 with time of day, with larvae rarely at the sur- 

 face by day, on any cruise, and increasing in 

 abundance at the surface by night. Further, the 

 station-time interaction at depth (e.g., level 3) 

 was significant for most larval stages and sub- 

 stages, indicating a high degree of station-to-sta- 

 tion variation in the depth of peak abundance. 



Table 2.— Summary of analysis of variance of proportional 

 larval abundance at four discrete depths. An arc-sine transfor- 

 mation was applied to the percentages. 



Table 1. — Summary of analysis of variance of larval abun- 

 dance. A square root transformation was applied to all data 

 pooled within each time increment over all four depths. 



— P<0.001, "P<0.01, 'P<0.05, +0 05<P<0.10. 



Consequences of Vertical Migration 



At the two sampling locations in the Gulf of 

 Carpentaria, the currents were dominated by a 

 tidal component as seen in the individual current 

 vectors, represented by the mean speed and di- 

 rection for the 15-min sampling periods (Fig. 6). 

 Detailed analyses of these records, however, 

 were complicated by three factors: 1) time lag 

 between sampling surface and bottom strata; 2) 

 short-term wind events; and 3) anomalous cur- 

 rent vectors at depth. At most sampling times, 



545 



