FISHERY BULLETIN: VOL. 85, NO 4 



the along-track direction and is due to the ship's 

 speed. Normally the records are compressed in 

 the direction the ship travelled relative to the 

 true geometry. A circular feature will look ellipti- 

 cal on the sonograph, with the long axis of the 

 ellipse perpendicular to the ship's track, and a 

 linear feature will look more perpendicular to the 

 track than it really is. 



RESULTS AND DISCUSSION 



The Tilefish Example 



Our initial discovery (Able et al. 1982) that tile- 

 fish, Lopholatilus chamaeleonticeps , construct 

 large (up to 4-5 m diameter and 2-3 m deep), ver- 

 tical burrows in the substrate (Fig. 5A) suggested 

 that we might be able to detect these burrows 

 with sidescan sonar. Since then we have success- 

 fully determined tilefish occurrence, distribution 

 patterns, and relative abundance based on side- 

 scan sonar observations at the edge of the conti- 

 nental shelf in the Mid-Atlantic Bight (Twichell 

 et al. 1985; Grimes et al. 1986) and the upper 

 slope off Florida (Able et al. 1987) (Table 1). Ver- 

 ification of sonograph images as tilefish burrows 

 (Figs. 2, 5A) was accomplished by in situ observa- 

 tions fi-om the Johnson-Sea-Link submersibles 

 (Askew 1985) (Table 1). 



As a result of these studies, we have demon- 

 strated that sidescan sonar can consistently de- 

 tect tilefish burrows both where the substrate 

 consists of semilithified clay (Mid-Atlantic Bight: 

 Twichell et al. 1985; Grimes et al. 1986) and 

 softer carbonate muds (off east coast of Florida; 

 Able 1987). During sidescan and submersible op- 

 erations near Veatch Submarine Canyon, we de- 

 termined that sidescan sonar could detect tilefish 



burrows as small as 0.5 m in diameter. Detection 

 of small burrows on sidescan was confirmed when 

 burrows in the area were measured in situ and 

 found to be 35-65 cm in diameter (mean = 48 cm, 

 sample number = 8). 



Under certain situations, tilefish abundance 

 could be estimated directly from sonographs. 

 Usually one tilefish is associated with each bur- 

 row (Able et al. 1982), therefore sidescan sono- 

 grams providing burrow counts could be used to 

 estimate standing stocks in areas surveyed, with 

 a modification of the area-density method (Ever- 

 hardt and Youngs 1981). Frequency distributions 

 of burrow density per unit area were log-normal, 

 and there were considerable numbers of zero ob- 

 servations (i.e., about 14-24% zero observations). 

 Therefore we log,, transformed the burrow density 

 data, and calculated the sample mean and vari- 

 ance of the delta distribution according to Pen- 

 nington (1983). We present sample estimates 

 from our data for two different locations: 



Case 1: Middle Atlantic Bight in the vicinity 

 of Hudson Submarine Canyon (number of obser- 

 vations = 407, number of nonzero observations 

 = 316, data from Twichell et al. 1985) based on 

 the formula 



a 



where A^ = total number offish (burrows) in sur- 



delta-distribution mean number 



veyed area, — 



of burrows observed per unit area surveyed, 

 A = total area surveyed, SD = standard devia- 

 tion, and C.I. (confidence interval) = 95% (1.96 

 X SD) calculated from the delta-distribution vari- 

 ance, thus 



Table 1.— Sidescan sonar observations of tilefish, Lopholatilus 

 and Caulolatilus, burrows on the seafloor off the east coast of the 

 United States. 



N = ^^^ • 0.407 km2 with SD = 123 and 95% 

 km2 



C.I. = 241 



N = 1,041 ± 98 tilefish [in the surveyed 

 area]. 



Case 2: South Atlantic Bight off Ft. Pierce, FL 

 (number of observations = 46, number of nonzero 

 observations = 40). The data was obtained with 

 167 kHz sidescan sonar from the research sub- 

 mersible NR-1 (Able et al. unpubl. data). In this 

 instance 



730 



