HortResearch, Mount Albert Research Centre, Private Bag 92-169 Auckland, New Zealand

Background and objectives
Ascospores of Venturia inaequalis are the principal source of inoculum in New Zealand apple orchards for primary apple scab infection during spring. Ascospores are released from pseudothecia, which develop during winter in fallen apple leaves, from early September until early December, when either the supply of ascospores has been exhausted or the apple leaf litter has decayed. Rainfall during daylight hours is required for ascospore release and large variations in numbers of ascospores occur from day to day. Sometimes a large proportion (>50%) of the total season's ascospore crop can be released during a single rain event. Apple growers need to know when major ascospore releases will occur in order to apply fungicides to prevent infection. Ascospore monitoring methods have been tried but labour costs are too high for routine monitoring to be economically viable. The aim of this study was to use ascospore release data previously collected in New Zealand to develop a predictive model that will give apple growers advance warning of major ascospore releases.

Materials and methods Seasonal patterns of ascospore release in New Zealand have been examined during the 1920s to 1930s [1], the 1960s [2], the 1980s (Beresford unpublished data) and the 1990s [3]. The ascospore trapping methods used in these studies included weekly exposure of glass microscope slides over apple leaves, forced discharge of ascospores from samples of leaves using controlled wetting in the laboratory and Burkhard volumetric spore traps. From these studies, 60 data sets of seasonal ascospore release from 13 ;years and six New Zealand regions were obtained.

Results and conclusions
Regional differences in ascospore release pattern were examined by comparing day numbers on which 5,10,50,90 and 95 percentiles of cumulative release occurred, which were derived from linear regression of logit cumulative percentage release on day number. Release began 2-3 ;weeks earlier in northern regions than in southern regions. For 30 data sets where temperature data were available, logit release was examined in relation to degree day accumulations, as used by Gadoury and MacHardy [4] and Schwabe et al. [5]. Degree-day accumulations associated with cumulative ascospore release from 1% to 95% of the season's total were around 800C days. These were most similar to [5]. The release pattern showed large variations about the fitted logit line owing to variations in rainfall.

It was hypothesized that daily ascospore numbers are determined by an underlying degree-day-driven asymptotic growth curve, modified by daily rainfall. A simulation model was developed which released ascospores on days when 0.2 ;mm or more daytime rain fell and otherwise accumulated unreleased ascospores until the next suitable rainy day. This model correctly identified days when rain after several dry days caused major ascospore release events.

The underlying growth curve was found to be more accurately described by a Gompertz function than the logistic function which has been used in the past, because the former better described asymmetry of the cumulative per cent maturation pattern which was characteristic of the New Zealand data, whereby there is more rapid maturation early in the season than later in the season.

1. Parham BE, 1932. The New Zealand Journal of Science and Technology 14, 184-192.
2. Brook PJ, 1976. New Zealand Journal of Agricultural Research 19, 103-109.
3. Manktelow DWL, Beresford RM, 1995. Proceedings 48th NZ Plant Protection Conference, pp. 78-82.
4. Gadoury DM, MacHardy WE, 1982. Phytopathology 72, 901-904.
5. Schwabe WFS, Jones AL, van Blerk E, 1989. Phytophylactica 21, 13-16.