In a 2 × 2 × 2 factorial experiment, plants of one genotype of tall fescue (Festuca arundinacea) were grown in soil in pots in a glasshouse on two occasions. The treatments were either with (E+) or without (E–) infection by the fungal endophyte Neotyphodium coenophialum; with (N+) or without (N–) inoculation by the root-knot nematode Meloidogyne marylandi; and with or without water deficit stress. In the first experiment, nematode numbers after 5 weeks had increased 10 times in the E–/N– watered pots and 22 times in the E–/N+ stressed plants. Root dry weight was decreased in all E–/N+ pots, compared with E–/N–. In contrast, nematode numbers in all E+/N+ pots decreased to nearly zero and root dry weight was unaffected. Osmotic adjustment in the growing zone of stressed plants was –0·35 MPa in E–/N– pots, but only –0·10 MPa in E–/N+ pots. Osmotic adjustment was greatest (–0·64 MPa) in E+ pots of both nematode treatments and almost no nematodes survived in the N+ pots. In the second experiment, there was complete nematode mortality in the E+ pots. Plant water relations were unaffected by treatments, however. It is concluded that endophyte-enhanced persistence of tall fescue in M. marylandi-infested soils that are prone to drought may be explained at least partly by endophyte protection of roots from nematode damage. Nematode inhibition by the endophyte may operate in addition to direct influences of the endophyte on enhancing drought tolerance of the host.