304 



Fishery Bulletin 92(2), 1994 



19°00'S 



QUEENSLAND 



10 20 30 

 kilometres 



146WE 



Figure 1 



Map of the cross-shelf transect off Townsville, Australia, showing the position of each station along 

 the transect. Station 1 = 19 km from Townsville, 2 = 24 km, 3 = 33 km, 4 = 43 km, 5 = 52 km, 6 = 61 

 km, 7 = 75 km, 8 = 92 km, 9 = 100 km, 10 = 115 km, 11 = 136 km, 12 = 145 km, 13 = 152 km, 

 14 = 163 km, and 15 = 172 km. 



and one at 20 m below the surface. In 1991/92, an 

 extra light-trap was added at 20 m. The anchored 

 light-traps had an automatic timer, enabling the 

 lights to be switched on and off automatically at 

 predetermined periods during the night. Each light- 

 trap on the reef fished for a total of three hours per 

 night; lights came on for one hour at 2200 hours, 

 2400, and 0300 hours. Light-traps at all reefs were 

 emptied the following day. 



Squid were identified in the laboratory and the 

 dorsal mantle length recorded for each individual. 

 Individuals were measured within 14 days of pres- 

 ervation in 100% ethanol. A comparison of measure- 

 ments of individuals (ranging in size from 5.3 mm 

 to 29.5 mm) before and 14 days after preservation 

 found that shrinkage was on average 0.5 mm. 



Abundance patterns of the two Photololigo species 

 during the two summers of sampling were examined 

 by using 'planned comparisons,' where specific 

 pregenerated hypotheses were examined (Day and 



Quinn, 1989). For each species we were interested 

 in differences in abundance between years, loca- 

 tions, and depths. 



To examine seasonality of juvenile Photololigo sp. 

 A, the inshore station (19 km) was sampled during 

 the austral winter months of May, June, July, and 

 August 1991. Three sites at this station were 

 sampled with four shallow and four deep (13-m) 

 light-traps. Sites were sampled during the period of 

 the new moon, on five nights in May and three 

 nights in June, July, and August. Densities in sum- 

 mer and winter months were compared by using an 

 unbalanced one-way analysis of variance (ANOVA), 

 with month as the factor analyzed. Values in each 

 light-trap for nights and sites within a month were 

 treated as replicates. 



To determine whether vertical migration might 

 influence horizontal distribution patterns we exam- 

 ined the size structure of Photololigo sp. A at two 

 depths during the night. On at least one occasion 



