286 



Fishery Bulletin 105(2) 



0.04 



o 



Q. 

 O 



0,12 

 10 

 008 

 0.06 

 0.04 

 0.02 

 00 



Arrowtooth flounder 



Rex sole 



— I 1 1 — 



20 30 40 



12 

 0.10 

 0.08 

 0.06 

 004 

 0-02 

 0.0 



Length (cm) 



20 30 40 50 60 



Dover sole 



40 50 



Figure 4 



Length-frequency distributions, expressed as the proportion caught at each length, for each of the 

 four flatfish species (arrowtooth flounder (Atheresthes stomias), flathead sole {Hippoglossoides 

 elassodon), rex sole iGlyptocephalus zachirus), and Dover sole i Microstomas pacificus)) averaged 

 over all hauls taken in the herding experiment are shown for three bridle lengths: short (thin line), 

 standard (medium line), and long (thick line). 



the slope statistically significant and this significance 

 was dependent on the relatively low value of ^^ in the 

 smallest size interval. In previous studies, evidence 

 for length-dependent herding has been inconsistent. 

 Although Engas and God0 (1989a) observed increases 

 in bridle efficiency with fish length for Atlantic cod 

 and haddock, Somerton and Munro (2001) found either 

 no change or decreases in bridle efficiency for seven 

 species of flatfish, and Ramm and Xiao (1995) found 

 no evidence of length dependent herding for any of a 

 variety of species. 



An alternative way of detecting length-dependent 

 herding is through the changes in mean length of the 

 catch as bridle length is changed, because the processes 

 leading to the differences in herding ability with fish 

 size would be intensified with increases in bridle length. 

 Although Engas and Godo (1989a) and Andrew et al. 

 (1991) found an increase in fish length with increas- 

 ing bridle length, Somerton and Munro (2001) found a 

 decrease in fish length. Because, in the present study, 

 we observed no significant changes in mean size with 

 bridle length (Fig. 4), even for arrowtooth flounder. 



which had a significant increase in /z,, with length, 

 length-dependency in the herding process is, at best, 

 weak for flatfish sampled with the Poly Nor'eastern 

 trawl and is therefore unlikely to contribute substan- 

 tially to size selectivity. 



Although estimates of kf^ for the four species were 

 pairwise not significantly different from each other, the 

 estimates for flathead sole, rex sole, and Dover sole were 

 very similar (Table 3), but considerably greater than the 

 estimate for arrowtooth flounder. From an ecological 

 perspective, similarity in the herding coefficients for the 

 three sole species makes sense because all eat sessile or 

 slow moving prey taken from the bottom and are likely 

 relatively slow swimmers that stay close to the bottom 

 when herded, whereas arrowtooth flounder eats rela- 

 tively large pelagic fish (Yang and Nelson, 2000) and 

 is likely a stronger, more agile swimmer that readily 

 leaves the bottom. 



The kfj estimates for the three sole species are also 

 quite similar to the estimates that we obtained previ- 

 ously for the 83-112 Eastern trawl (flathead sole, 0.24; 

 rex sole, 0.22; Dover sole, 0.27; Somerton and Munro, 



