USDA-ARS National Forage Seed Production Research Centre, 3450 SW Campus Way, Corvallis, OR 97331, USA

Background and objectives
Stem rust, caused by Puccinia graminis subsp. graminicola, is the most widespread and damaging disease of perennial ryegrass and tall fescue grown for seed in the Pacific North-west of the USA [1], where most of the US crop of grass seed is produced. Rust severity varies from year to year, presumably in response to weather conditions. The disease is controlled with fungicides but a better understanding of conditions favouring disease outbreaks is needed to develop decision models for the most efficient fungicide use. The amount of inoculum overwintering on the crop could be an important determinant of spring epidemic development. Winters in the Willamette Valley, the major production area, are typically mild and may permit survival of the pathogen in the uredinial stage. During the past 38 years, low temperatures of -13, -102, -6 and -3C have been reached in 2, 15, 27 and 37 of the years, respectively. Studies with related rust fungi from wheat [2] showed that free urediniospores can be killed by a 6-h exposure to freeze temperatures as high as -4C. Our experiments were conducted to determine the quantitative effect of cold temperatures on survival of uredinia of P. ;graminis subsp. graminicola.

Materials and methods
Perennial ryegrass and tall fescue were sown in the field on 24 September, 1997. On 16 October, plugs of grass (7.5 ;cm diameter) were transferred into pots. The pots were set back into the ground for retrieval later. Plants were inoculated with stem rust urediniospores on 24 October, and symptoms were visible by 10 November. Several light freezes (-1C) occurred in December and January. Pots were taken from the field in mid-January and brought to the laboratory, where they were exposed to freezing treatments in a programmable freezer. Cooling rates, freezing duration and warming rates were selected to simulate typical freeze events in historical weather data for the locale. Low temperatures of -13, -10, -6 and -3C were used, with five replicate pots per grass species per treatment. After the freeze treatment, pots were held at 0.5C for several hours, then placed in a greenhouse kept at 18-21C. Control treatment pots were taken from the field directly to the greenhouse with no freeze treatment imposed. Pustule development was observed periodically for 2 ;weeks.

Results and conclusions
P. ;graminis subsp. graminicola responded similarly to freezing conditions whether in perennial ryegrass or tall fescue. Most leaves of plants exposed to -13C (first held overnight at 1C, then cooled at -1.5C/h, held at -13C for 12 ;h and then warmed to 0C at 1.5C/h) were killed and no active pustules had developed by 14 ;days after the freeze treatment. Freezing to 10C was less damaging to plants, but pustule development was delayed relative to that in plants exposed to warmer temperatures. Freezing to 6 or -3C had little observable effect on plant leaves and a less deleterious effect on pustule activity than the colder temperatures. By 8 ;days after the freezing treatments, perennial ryegrass exposed to -13, -10, -6, -3C or no freezing (control) had 0, 0.4, 24, 40 and 51 active pustules/pot, respectively; the respective pustules/pot 11 ;days after freezing were 0, 4.6, 40, 42 and 33. Tall fescue (which is generally less severely affected by stem rust than is perennial ryegrass) had 0, 0.6, 2, 5 and 21 active pustules/pot after 8 ;days and 0, 1, 5, 5 and 8 pustules/pot after 11 ;days in these respective temperature treatments. We conclude that the variation in low temperatures among years is sufficient to be differentially influential in the amount of overwintering inoculum available for spring epidemic development in Oregons Willamette Valley grass seed production area.

1. Welty RE, Barker RE, 1992. Plant Disease 76, 637-641.
2. Eversmeyer, MG, Kramer, CL, 1995. Phytopathology 85, 161-164.