Beam trawl, camera sled and submersible data from 2 000–3 300 m off central California produced similar fish faunal composition, but different density estimates. All species caught in trawls were observed in camera-sled and submersible observations. However, some rare species that were observed were not caught in trawls. The fish fauna was dominated by the families Macrouridae, Zoarcidae, Moridae, and Rajidae. Fishes both trawled and observed were the macrourids Coryphaenoides armatus, C. filifer and C. leptolepis; the zoarcids Bothrocara spp., Pachycara lepinium and Lycenchelys spp.; the morid Antimora microlepis; the rajid Bathyraja trachura, the ophidiid Spectrunculus grandis, and the liparidid Careproctus ovigerum. One unidentified liparidid ( Paraliparis sp.) and two unidentified Lycenchelys spp. were trawled and may have been seen but also could not be identified to species from photographs. Observed only in photographs were the liparidids Paraliparis rosaceus and Careproctus melanurus, synodontid Bathysaurus mollis, and notocanthid Notacanthus chemnitzii. These three techniques differed in their ability to provide specimens for accurate identification, counts, and later life history (feeding habit, age and growth, and reproduction) studies, and to provide information on dispersion, habitat utilization, behavior and interactions. Accurate density estimates were undoubtedly hampered by trawl and camera sled avoidance, escape, and uncertainties concerning the area trawled. Camera sleds produced higher (and perhaps better) estimates of density. Submersible observations from the DSV Alvin produced a similar species list but little additional, quantitative information. Both visual techniques allowed habitat characterization, but no strong faunal associations with habitat types were observed.