70 



Fishery Bulletin 93(1). 1995 



Figure 2 



Baited video camera assembly with bait container positioned 60 cm in 

 front of the camera lens. 



for each video sequence, three indices 

 of abundance were scored for each spe- 

 cies taped: maximum number 

 (MAXNO); time to first appearance 

 (TFAP); and total duration in sequence 

 (TOTTM). The MAXNO index was de- 

 termined as the peak number of a spe- 

 cies visible at any one time (maximum 

 interval one second) during a deploy- 

 ment. Fork length (FL) to the nearest 

 0.1 centimeter (cm) was recorded for 

 fish caught on the longline. The FL of 

 opakapaka observed on video was esti- 

 mated and rounded to the nearest 5 cm 

 by comparing fish swimming in the 

 plane of the bait container with the 

 known size of the container. 



Statistics 



An average maximum number offish re- 

 corded for data collected in 1992 was cal- 

 culated for each of nine sequences (three 

 video stations) by using a mean weighted 

 by the duration of each occurrence: 



lowered to the bottom and marked by a buoy; later 

 they were raised to the surface by outboard engine 

 power. Cameras were allowed to rest on the bottom 

 for a standard interval of 10 minutes before retrieval. 

 The duration and number of video camera deploy- 

 ments to be used on the ship cruise were estimated 

 on the basis of three earlier pilot deployments of the 

 video assembly from small craft. These prior tests 

 indicated that about 10 minutes were required to 

 deploy and retrieve the camera assembly. The time 

 to first appearance (TFAP) of opakapaka from the 

 three pilot stations was 227 ± 300 sec (mean ± 1 stan- 

 dard deviation of the data [SD]) after bottom con- 

 tact. A bottom time of 10 minutes was chosen to ac- 

 commodate likely extremes and also to allow 6 deploy- 

 ments per 2-h tape (20 min per deployment x 6 deploy- 

 ments). With two cameras, 12 deployments per day 

 could be made without changing tapes. The maximum 

 number of longline sets was limited to four per day, 

 based on three camera deployments per longline set. 



Types of data 



Species presence, total number of individuals per 

 species, and the number of hooks lost were recorded 

 for each longline set. Species presence and duration 

 of squid bait attachment to the bait container (BTM) 

 were recorded for each video sequence. In addition, 



£*„ N h 



A w 



(1) 



N 



where X w =the weighted average maximum num- 

 ber of fish, rc=total number of occurrences, X h = maxi- 

 mum number of fish seen in the h th occurrence, 

 N /i =duration (s) of the h th occurrence, and N=YN h = 

 600 s. 



Video indexes were calculated as means (of up to 3 

 deployments) to standardize for multiple deploy- 

 ments per station. Video indices were derived in two 

 logarithmic forms — mean of logs (ML), 



]Tln(x,+l) 

 ML = -^ 



(2) 



and log of means (LM), 



LM = In 



+ 1 



(3) 



where ac- = individual datum for a variable (i.e., the 

 value for the variable MAXNO, TFAP, or TOTTM for 

 each deployment at a station) and /i=number of de- 

 ployments per station. The longline index consisted 

 of log-transformed individual set data [In (catch + 



