1062 



Fishery Bulletin 97(4), 1999 



Observational records illustrated some differences 

 in the effects of the anesthetics. All fish anesthetized 

 with MS-222 attained deep anesthesia, whereas 407f 

 offish exposed to phenoxyethanol did not (Table 1 ). Tliis 

 resulted in twitching of the animal during treatment. 

 Occasional cramp-like convulsions were also ob- 

 served in some of these specimens during the early 

 recovery period. 



Examination of placement wounds after the ter- 

 mination of the experiment revealed consistent pat- 

 terns (Table 1). The incisions made into the body cavi- 

 ties of internal control and internal treatment fish were 

 healed (some minor inflammations did exist). The punc- 

 ture wounds through the dorsal musculature of fish 

 with externally attached transmitters were aggravated 

 and in some cases enlarged due to repeated attempts 

 to dislodge the transmitters. Wounds on external con- 

 trol fish were either healed or partially healed, and 

 there were no further signs of aggravation. 



Field evaluation of ultrasonic telemetry in the coral 

 reef environment 



Evaluation of bearing errors (true bearing-observed 

 bearing) by transect (0°, 45° vs. 90°) indicated a sig- 

 nificant difference in mean bearing error between 



transects (Ki-uskal-Wallis ^2 252=22.102, P= 0.000). 

 The 90° transect displayed the least bias (mean=- 

 1.50° ±0.7163° SE), the 0° transect had slightly 

 larger, positive bias (mean=2.09° ±1.077°SE), and the 

 45° transect showed strong, negative bias (mean=- 

 5.43°±1.021°SE). 



Owing to the differences in bias and accuracy be- 

 tween transects, observer differences in bias were 

 examined separately for each transect. Bias differed 

 between observers for the 0° transect (Kruskal-Wallis 

 //j y4=6.238, P=0.013); observer two showed positive 

 bias compared with observer one (Fig. 3). Observers 

 did not differ significantly in bias for the 45° transect 

 (Kruskal-Wallis H^ y,,=0.209, P=0.648, Fig. 3), or the 

 90° transect, (Kruskal-Wallis H, ,^=2.603, P=0.107, 

 Fig. 3). 



Paired ^-tests were used to compare precision esti- 

 mates (SD) between observers for each transect sepa- 

 rately. No significant observer difference in precision 

 was detected: 0° (^g=1.2599, ^=0.2545), 45° (fg= 

 1.5819, P=0.1648) and 90° (<g=0.6859, P=0.5184). 



Evaluation of the parameters of maximum-error 

 polygons indicated some clear patterns, both with 

 respect to distance between sound source and track- 

 ing boat and with respect to angular combination of 

 position bearings taken. The 0°-^5° transect combina- 



