Ellis and DeMartini: Video camera sampling of Pristipomoides filamentosus abundance 



75 



43 



e 



3 



E 



3 

 fi 



X 



10 



5 10 15 20 



Station 



Figure 5 



Scatterplot of maximum number of opakapaka observed (MAXNO) for shallow 

 and deep positions and their mean by stations for data collected during May 

 1993. Stations are ordered in geographic sequence from farthest southeast (sta. 

 1) to farthest northwest (sta. 18). Means with component data hidden represent 

 coincident shallow and deep values. 



1992) observed that, although time of arrival of the 

 first fish (TFAP) was strongly related to estimated 

 fish densities, the maximum number offish seen at 

 a station (MAXNO) either was unrelated or inversely 

 related to densities. Our observation that MAXNO 

 was highly correlated with an abundance estimate 

 (CPUE ) may at first seem contradictory to these prior 

 findings. However, there are important differences 

 between our methods and those of previous studies: 

 previous deep-sea work operated in unproductive 

 depths >2,000 m and cameras recorded data for at 

 least 11 hours per station, whereas our study was 

 limited to productive depths <100 m for which a rela- 

 tively short soak time ( 10 min) was sufficient. In the 

 deep-sea studies, all bait was open to consumption. 

 The partly internal bait of our system created a res- 

 ervoir of odor that persisted for the soak duration in 

 most cases; puffers removed all bait in only 3 out of 

 75 deployments. Differences in rates of bait consump- 

 tion between the two deep-sea stations and result- 

 ing variations in bait attractiveness may have con- 

 tributed to the disparity between MAXNO and fish 

 density in the deep-sea studies. 



The MAXNO and TFAP indices in our study were 

 highly correlated (Table 2). This correlation suggests 



that the greater the density, the faster the fish ar- 

 rive at the bait. These data agree with the observa- 

 tions of Priede et al. ( 1990), where fish arrived at 

 the camera faster at the station with presumed 

 higher densities. Since the MAXNO and TFAP indi- 

 ces were both significantly correlated with CPUE in 

 our study (Table 2), the MAXNO index was chosen 

 as the best index of abundance because it had the 

 better correlation. A persistent bait source and short 

 soak time may have contributed to this stronger cor- 

 relation. In the future, the use of MAXNO as an in- 

 dex of abundance should be reevaluated separately 

 for each species and application. 



Conclusions 



Video cameras provide an accurate tool for sampling 

 juvenile opakapaka, and the video MAXNO variable 

 provides a relatively precise and accurate index of 

 abundance. Based on 1993 data for a series of two 

 camera deployments per station, minima of 17 to 22 

 pairs of deployments (34-44 sets) per study area 

 would be necessary to detect a twofold change in ju- 

 venile opakapaka numbers (at <x 2 =0.1, 0.05, respec- 



