Sampling intensity and timing for estimating acremonium coenophaialum incidence in fescue pastures

Acremonium coenophialum Morgan-Jones and Gams has been associated with several animal disorders known collectively as fescue toxicosis. Reestablishment of infected (E+) pastures with seed not containing the fungus (E-) is beneficial for eliminating the symptoms of fescue toxicosis. However, since management decisions must be based on knowledge of A. coenophialum incidence, appropriate sampling methods should provide information about fungal incidence with accuracy. Two sampling studies were conducted to determine an effective sampling method. In the first, eight 4-ha research pastures that had been established for a grazing trial having E+ incidence ranging from near 0 to more than 70% were sampled in June 1986 using a transect method (TM) and a stratified random sampling design (SR) at an intensity of 23 tillers ha^{-1. In the second, four 2-ha pastures were sampled at monthly intervals from November 1985 through October 1987, using SR at 41 samples ha-1. Samples were assessed for E+ status using Protein A enzyme-linked immunosorbent assay (PAS-ELISA). In the first study, observed variability in E+ incidence of two pastures increased from 15 or 30% to about 60% during the 27-month period: six other pastures, with original 0, 45. 60, and 75% E+ incidence, had only small increases in E+ incidence. In the second study, only small fluctuations in fungal incidence were observed in four pastures with 60% E+ incidence. Significant E+ incidence could be detected at any time during the year using PAS-ELISA. The distribution of the E+ and E- plants was random in the first study because they had been seeded uniformly, but was highly aggregated in the older (>11 y) pastures of the second study. Therefore, dispersal of sampling sites across the entire field is important when sampling older pastures that may have an aggregated fungal distribution. It appears that relatively few samples are required for assessment of E+ status in pastures. Eight or more stratified random samples ha-1 might be an adequate sampling intensity for producers: for research purposes, sampling intensity should approach one sample 250 m-2 in a stratified random sampling. The TM and SR gave similar estimates, but SR could better detect spatial relationships.}