We describe the spatial and temporal patterns in the abundance of nematode-trapping fungi and in suppression of nematodes in a coastal shrubland. A previous study at this location (Bodega Marine Reserve, Sonoma County, California) had documented a soil food chain in which an insect-parasitic nematode consumes and kills the soil-dwelling larva of the ghost moth, which otherwise consumes and kills the bush lupine; the patchy distribution of the nematode and lupine suggest the involvement of nematophagous organisms, including nematode-trapping fungi. To test our model (trapping fungi kill insect-parasitic nematodes, and therefore ghost moths persist and kill lupines at some sites), we hypothesized that sites with substantial lupine mortality would contain larger numbers of nematode-trapping fungi and would be more suppressive to nematodes than would sites with little lupine mortality. Soil was collected from eight sites (four with substantial lupine mortality and four with little lupine mortality) at 2-mo intervals for 1 yr and subjected to dilution plating and most probable number procedures. Nematode-trapping fungi detected were Arthrobotrys brochopaga, A. musiformis, A. oligospora, A. superba, Geniculifera paucispora, Hirsutella rhossiliensis, Monacrosporium cionopagum, M. doedycoides, M. eudermatum, M. parvicollis, Nematoctonus concurrens, and Stylopage sp. A. oligospora was the most abundant. Some soil samples contained large numbers of nematode-trapping fungi (as many as 695 propagules/g of soil), but sites with substantial lupine mortality did not contain larger numbers than did sites with little mortality. In a laboratory bioassay, suppression of the root-knot nematode Meloidogyne javanica was significant in 44% of the samples, but suppression was not correlated with fungal population density; moreover, soil from sites with substantial lupine mortality was less suppressive than was soil from sites with little mortality. The spatial and temporal patterns in fungal abundance and in nematode suppression, therefore, did not support our model that nematode-trapping fungi cause the patchy distribution of the insect-parasitic nematode and hence of lupine. Nevertheless, we must be conservative in rejecting the involvement of these fungi in the food chain, because methods for fungal quantification have important limitations.