Ferm et al.: Foraging habits and dietary overlap of Limanda aspera and Lepidopsetta polyxystra in the Bering Sea 3 
Coat 
\aer 
a nis E 
Nee 
Gulf of Alaska 
* Herendeen Bay 
160°40'W 
oO & 
© 
~, 
Ae rt Moller @ 
Mi 
sexs _Inner Port Molter 
160°40'W 
Figure 1 
Map of the Port Moller-Herendeen Bay (PM-HB) system, with the inset showing the location of the 
PM-HB system along the Alaska Peninsula. Pie symbols indicate sampling locations, and the colors 
in these symbols indicate proportions of catch at each sampling station composed of yellowfin sole 
(Limanda aspera) (black) and northern rock sole (Lepidopsetta polyxystra) (white). Gray x symbols 
indicate stations where neither species was captured. Regions within the sampling area are Bering 
Sea coastal waters (Bering Coast), Herendeen Bay, the Outer Port Moller, and the Inner Port Moller. 
Dashed lines indicate borders between regions. Areas in a light gray shade are intertidal habitats. 
Gray lines indicate depth contours. 
and off-bottom activity of juvenile flatfish. The endoben- 
thic category is for organisms, such as polychaetes and 
bivalves, that live within the sediment. Epibenthic organ- 
isms, such as barnacles and most harpacticoids, are 
directly associated with the sediment surface. Hyperbenthic 
organisms are those that spend part or all of their time 
associated with the sediment but are more mobile than 
epibenthic organisms (Mees and Jones, 1997). Pelagic 
organisms inhabit the surface or water column and do not 
come in regular contact with the bottom. 
Prey taxa richness and sampling coverage 
Total species richness of the prey in diets of yellowfin sole 
and northern rock sole in the sampling area was esti- 
mated from extrapolation of the relationship of species 
richness and sample size by using the function iNEXT in 
the package iNEXT (vers. 2.0.20; Chao et al., 2014; Hsieh 
et al., 2016, 2020) in statistical software R (vers. 4.0.5; R 
Core Team, 2021), with 95% confidence intervals calcu- 
lated with 1000 bootstrap iterations. Non-prey (e.g., sand) 
and unidentifiable items were removed from all analyses. 
Practical limitations to prey identification resulted in prey 
types reflecting different levels of taxonomic aggregation. 
Sampling coverage provided an estimate of diet complete- 
ness, where the probability of encountering a new prey 
taxa in the next independent observation is 1 minus the 
sampling coverage from the last observed sample (Hsieh 
et al., 2016). 
Diet descriptions 
To correct diet compositions for uneven sampling cov- 
erage (stations with high catch rates represented by 
