An epidemic of the ergot fungus (Claviceps purpurea (Fr.) Tul), a non-systemic floral pathogen, appeared in populations of Spartina anglica C. E. Hubb in Poole Harbour in the 1980s. Over 70% of inflorescences were infected between 1985 and 1995. Between 1983 and 1995, there was no consistent pattern in the rank order of sites with respect to the proportion of inflorescences infected, and variation in the amount of infection was not related to inflorescence density. Among years, there was significant variation in the total biomass of ergot per inflorescence, but no significant difference in the number of ergots per inflorescence. Log–log regressions of total weight of ergot per inflorescence against the number of ergots per inflorescence showed that in all years resources for each additional ergot were limited.

There was no statistically significant difference between the number of seed set on infected and uninfected inflorescences in 1985 or 1995. Further analysis showed that, compared with uninfected inflorescences, there was higher seed output from inflorescences with fewer than 10% of spikelets infected, which balanced reduced seed output from heavily infected inflorescences. At high levels of ergot infection, a lower proportion of uninfected spikelets set seed, compared with spikelets on uninfected inflorescences. This suggests that conditions which favour ergot growth are detrimental to seed production. Compared with uninfected inflorescences, mean and total seed weights were significantly lower in inflorescences with >10% of spikelets infected.

The fungus Fusarium heterosporum Nees ex Fr. was found in association with the ergots on Spartina. There was a significant positive correlation between the average severity of Fusarium infection and the number of ergots on an inflorescence. There was a non-significant negative correlation between severity of Fusarium and mean ergot weight, when ergot number per inflorescence was held constant. These data suggest that any negative effect of Fusarium on Claviceps is small.

The high, and temporally and spatially uniform levels of ergot infection probably result from genetic uniformity of Spartina and the lack of zonation in the salt marshes of Poole Harbour. This situation is likely to persist for the foreseeable future.