FISHERY BULLETIN. VOL. 85, NO. 4 



N 



369 



• 0.42 km2 with SD = 64 and 95% C.I. 



= 125 



N = 154 ± 53 tilefish [in the surveyed area]. 



The estimates of A'^ in cases 1 and 2 could be 

 extrapolated to the entire fishing grounds using 

 an estimate of the area of the entire grounds to 

 provide an estimate of standing stock. However, 

 we believe that extrapolation to areas where no 

 density data is available is imprudent for several 

 reasons. First, the density of burrows in different 

 locations is quite variable as shown in the two 

 above examples, and density on the Middle 

 Atlantic-Southern New England ground (case 1) 

 varies over the grounds at least tenfold (Grimes 

 et al. 1986). Second, some burrows, at least in the 

 Middle Atlantic-Southern New England area, 

 may not be occupied during all seasons of the year 

 (Grimes et al. 1986). Although we do not have as 

 much background knowledge for case 2, we know 

 that burrow density at different sites off the Flor- 

 ida east coast varied at least fivefold (Able et al. 

 unpubl. data). 



Another possible source of error in using bur- 

 row density to estimate tilefish stock size is that 

 some burrows may be unoccupied. This should be 

 of particular concern in exploited fishing areas. 

 However, Twichell et al. (1985) and Able et al. 

 (unpubl. data) have shown that abandoned bur- 

 rows are filled by sedimentation relatively rapidly, 

 i.e., less than one year, somewhat ameliorating 

 the problem, at least over longer time periods. 



Perhaps the most constructive aspect of cases 1 

 and 2 is the opportunity to examine the error 

 associated with sidescan sonar estimates of A^. 

 These results show that the standard deviation 

 varied from about 5 to 20% of the mean. Hen- 

 nemuth (1976) found that the standard deviation 

 in the numbers of different demersal species 

 caught per tow within a stratum during stratified 

 bottom trawl surveys approximately equalled the 

 mean. Thus, this comparison suggests that area 

 density estimates of abundance (calculated using 

 the delta distribution) from sidescan sonar sur- 

 veys will provide abundance estimates of much 

 greater precision than trawl surveys. Reduced 

 manpower needs and rapid application are addi- 

 tional factors that favor the sidescan sonar 

 methodology. However, because the sidescan 

 methodology is only useful for certain three di- 

 mensional habitats (e.g., reefs, rocks, and bur- 



rows) that would damage or make a trawl useless, 

 application of the two techniques may usually be 

 mutually exclusive. 



Tilefish are known to occur in other habitats 

 (Grimes et al. 1986) such as boulder fields, which 

 can be detected on sidescan sonographs (Figs. 2, 

 5B). Another habitat type (pueblo habitats, 

 Warme et al. 1977; Grimes et al. 1986) occurs in 

 the clay outcrops along the walls of submarine 

 canyons (Figs. 4, 6A). Neither of these habitat 

 types lend themselves to quantification of fish 

 abundance. Recently, we have been able to con- 

 firm that the burrows of other tilefish {Caulo- 

 latilus spp.) are also detectable with sidescan 

 sonar (Able et al. 1987; Figs. 6B, 7). Subsequent 

 observations from a submersible confirmed that 

 these burrows were occupied by C. microps and C. 

 cyanops with frequent multiple occupancy. Given 

 that it has now been demonstrated that represen- 

 tatives of four of the five genera of tilefishes con- 

 struct burrows (see Able et al. 1987), it is reason- 

 able to suspect that all tilefish construct burrows. 

 Thus, those larger species of commercial interest, 

 such as red tilefish, Branchiostegus japonicus 

 japonicus (Lim and Misu 1974), also may have 

 burrows that are detectable by sidescan sonar. 



Other Examples and Possibilities 



As an outgrowth of our studies of Lopholatilus 

 we have observed other species-specific habitats 

 that can be detected with sidescan sonar. Ameri- 

 can lobster, Homarus americanus, typically oc- 

 cupy scour basins around large boulders (Cooper 

 and Uzmann 1977; Valentine et al. 1980) as do 

 cusk, Brosme brosme, and ocean pout, Macro- 

 zoarces americanus (Valentine et al. 1980; 

 Grimes et al. 1986), and these habitats also are 

 detectable with sidescan sonar. American lobster 

 (Fig. 5) and conger eels. Conger oceanicus, (Able 

 et al. 1982; Grimes et al. 1986) have been ob- 

 served in tilefish vertical burrows as well. Simi- 

 larly, it would not be surprising if the habitats of 

 other clawed lobsters are detectable with sidescan 

 sonar. For example, H. gammarus from the east- 

 ern North Atlantic is similar to H. americanus in 

 that it is shelter seeking and occurs around boul- 

 ders (Dybern 1973). In addition, recent in situ 

 observations in the Gulf of Mexico have discov- 

 ered that yellowedge grouper [Epinephelus 

 flavolimbatus ) also occupy burrows and elongate 

 trenches (R. S. Jones, E. Gutherz, and W. R. Nel- 

 son, pers. obs.) that could easily be detected by 

 sidescan sonar. 



732 



