1060 



Fishery Bulletin 97(4), 1999 



Data analysis Data from experiment A were ana- 

 lyzed for differences in retention times between an- 

 esthetic agents by using a single-factor analysis of 

 covariance (ANCOVA). Fork length was used as a 

 covariate to account for any effect of size of fish on 

 the retention times. Assessment of experiment B was 

 based on the observations made during and after the 

 experiment. The change in weight during experiment 

 B was analyzed by using a two-factor analysis of co- 

 variance (ANCOVA). Occurrence of gas-bubble- 

 trauma was treated as a covariate. Underlying as- 

 sumptions were evaluated prior to analyses (Sokal 

 andRohlf, 1981). 



Field evaluation of ultrasonic telemetry in the coral 

 reef environment 



The extensive use of radio telemetry in wildlife re- 

 search has produced procedures to assess the accu- 

 racy (bias and precision) of telemetry reception 

 (White and Garrott, 1990). However, no such proce- 

 dures appear to be published for ultrasonic telemetry. 

 Given the paucity of studies with ultrasonic telemetry 

 in the coral reef environment, a standardized field 

 test using stationary transmitters was undertaken. 

 Given that wind and sea-state influence the detect- 

 ability of sound transmission in water (Jellyman et 

 al., 1996), any effect of direction of prevailing wind 

 on detectability or directional bias of observed sound 

 signals needed to be assessed. Thus, the objectives 

 of this field evaluation were 1) to establish the accu- 

 racy of directional bearings in relation to prevailing 

 wind direction; 2) to evaluate observer bias and pre- 

 cision of bearings in relation to wind direction; and 

 3) to determine the optimal angle for cross-bearings, 

 and the optimal distance between tracking boat and 

 signal in order to obtain position estimates that mini- 

 mize error polygons. 



21 positions for the three transects. Replicated com- 

 pass bearings to the transmitter were obtained for 

 each position by using the following protocol: the field 

 of view of the observer was restricted to the tracking 

 receiver for the duration of the experiment; an as- 

 sistant anchored the boat at the marker buoys in 

 random order and recorded the true compass bear- 

 ing to the transmitter. The observer determined the 

 perceived maximum directional signal strength by 

 using the directional hydrophone. The correspond- 

 ing compass bearing of the directional hydrophone 

 (observed bearing) was recorded. The direction of the 

 hydrophone was changed haphazardly by the assis- 

 tant and the procedure was repeated. Six replicate 

 bearings were taken at each of the seven distances 

 on the three transects. The procedure was repeated 

 by the second observer for a total of 252 observed 

 bearings. 



Data analysis Data were assessed for accuracy, bias, 

 and precision (Wliite and Garrott, 1990). Bearing 

 error polygons were determined by using the largest 

 and smallest directional bearing observed at each 

 distance-marker buoy for any combination of two 

 transects (0''-45°, 45°-90°, and 0°-90°). Thus, the 

 error polygon parameters obtained (polygon area and 

 maximum diagonal dimension) represented the larg- 

 est possible error polygon estimates. Only bearing 

 pairs from equidistant locations were used for poly- 

 gon determination. Statistical analyses included 

 Kruskal-Wallis nonparametric ANOVA and paired t- 

 test. All data were examined for violations of under- 

 lying assumptions prior to analysis (Sokal and Rohlf, 

 1981) and data log^Q transformed where applicable. 

 Given that the angular scale of measurement repre- 

 sented only part of a circle, and absolute direction of 

 angles was of no interest, data were treated as lin- 

 ear (Cain, 1989). 



Experimental methods An ultrasonic transmitter 

 (V8-2L, Vemco Ltd) was attached 15 cm above the 

 substratum to a surface buoy in 4-6 m water depth. 

 Thi-ee 200-m transects were run fi-om the moored trans- 

 mitter in relation to the prevailing wind direction: 



0": transmitter located directly upwind from any 

 position on the transect. 



45": transmitter located at 45° downwind from the 

 prevailing wind direction. 



90°: transmitter located at right angle to the prevail- 

 ing wind direction. 



Marker buoys were positioned at 25-m intervals 

 along each transect, starting at 50 m from the trans- 

 mitter location (i.e. from 50 to 200 m), resulting in 



Field tracking trials 



To verify the findings from the aquarium study, the 

 two most suitable long-term transmitter placement 

 techniques (external and internal placement) were 

 tested under field conditions (MS-222 used as anes- 

 thetic). Two P. leopardus (43.1 cm and 58.9 cm FL) 

 were equipped with external transmitters and re- 

 leased at their capture site after a 15-min recovery 

 period. In the second field trial, internal transmit- 

 ters were placed in two specimens (59.0 cm and 42.9 

 cm FL). Fish were released at the capture site after 

 recovery from anesthesia. 



The basic tracking technique described by Holland 

 et al. (1985, 1992) was used. Exact positions offish 

 equipped with transmitters were determined by visual 



