Point quadrat (PQ) and video transect (VT) techniques were compared to determine the feasibility of using the latter to sample subtidal red algal assemblages. More taxa were found by PQs than by VTs, both overall and on a per quadrat basis. VTs tended to overestimate the percent cover of rock and articulated coralline algae. VTs were incapable of resolving many of the less common red algae. Video sampling required less time in the field than point quadrats but laboratory analysis was very time consunling. Lack of resolution and extensive laboratory analysis make this technique insufficient for sampling benthic red algal populations. However, VTs may be useful in documenting large-scale differences between regions, such as the percent cover of understory kelps or the density of large mobile invertebrates in central California kelp forests. Video sampling has been successfully used in the intertidal but traditional, random point quadrats s t d appear to be the most efficient and accurate way to sample benthlc communities in kelp forests. The advent of quantitative techniques to sample subtidal benthic communities has aided the scientific description of pattern and helped transform marine ecology from a wholly observational to a more experimental science. Techniques to measure percent cover were first developed by terrestrial botanists (e.g. Greig-Smith 1983) and have been successfully adapted to subtidal communities by marine ecologists (e.g. Foster 1975, Cowen et al. 1982, Kennelly 1987). Photography has been used quantitatively in intertidal, subtidal and deep-sea habitats (Bohnsack 1979, Budd 1982, Witman & Sebens 1985, Jaap 1986, Rosman & Boland 1986, Thouzeau & Hily 1986, Meulstee et al. 1988, Edmunds & Witman 1991). Due to recent advances in video resolution and ease of use, it has begun to replace traditional photography as both a qualitative (Potts et al. 1987, Semple & Sharp 1988) and quantita' Present address: Graduate Program in Ecology and Evolutionary Biology, Brown University, Providence, Rhode Island 02912, USA tive (Jaap et al. 1990, Maney et al. 1990, Berklemans 1992) sampling tool. The strength of video is its ability to rapidly acquire data and permit analysis in the laboratory at a later date. Video quadrats and transects are now being used routinely to measure the density and percent cover of subtidal marine organisms (e.g. Edmunds & Witman 1991). As part of a study on the effect of light on understory algae in central California (USA) kelp forests, we have been using point quadrats to estimate algal cover. These involve scoring a small number of random points (usually 10 to 20) along a bar randomly placed on the substrate. This technique is excellent for relatively abundant organisms but, theoretically, undersamples rare species because of the few points used. Whorff & Griffing (1992) recently suggested that point quadrats undersample the substrata while Foster et al. (1991) concluded that point quadrats provide better estimates of the total number of taxa and their total and individual percent cover. An alternative to point quadrat methods for rare species are search techniques (e.g. Nybakken 1978, McClanahan 1990, McClanahan & Muthiga 1992). These have been designed for extremely rare species (i.e. approximately 0.1 ind. m-2; McClanahan & Muthiga 1992) but understory algae in many benthic habitats range from 1 to 30% cover (Foster 1975). In this paper, we present results of a direct comparison of point quadrat (PQ) and video transect (VT) sampling for benthic algal assemblages. Video was used over traditional photo-quadrats because video film is significantly cheaper than still film and more samples can be taken between film changes. To test for and alleviate the potential PQ bias of using too few points, we sampled the video image using 5 times as many random points as the PQ. This should have increased the likelihood of detecting rare species. The number of 0 Inter-Research 1993 204 Mar. Ecol. Prog. Ser. 101: 203-208. 1993