Institut National de la Recherche Agronomique (INRA), URIV, B.P. 81, 33883 Villenave d'Ornon Cedex, France

Background and objectives
Bunch rot of grapes (Vitis vinifera L.) caused by Botrytis cinerea Pers.:Fr., is an economically important disease in many viticultural regions of the world. Early Botrytis rot has been defined as the outbreak of symptoms in grape clusters near the period of veraison (mid-August), the stage when grape berries beginning to ripen lose their green colour [1]. However, there is a lack of quantitative information on the relationship between these early infections and fruit rot at harvest (5 or 6 ;weeks later). In 1993, 1994, and 1995, the epidemic development of Botrytis bunch rot was studied in the same vineyard near Bordeaux, France, in the absence of fungicides to control Botrytis cinerea.

Materials and methods
Three cultivars were used: Merlot (Mrl), Cabernet sauvignon (Cs) and Cabernet franc (Cf). Control clusters of average compactness were compared with clusters treated either with the silver thiosulfate complex Ag(S2O3)2 3- for increasing compactness or with an ethylene precursor for decreasing compactness. Botrytis rot on grapes at veraison was initiated artificially as a result of controlled infestations of grape berry moth Lobesia botrana (Lepidoptera: Tortricidae). This was based on the positive correlation between the number of secon- generation larvae per cluster and the number of early infection spots [2]. A few adjacent berries showing Botrytis rot symptoms in a cluster was considered as an early infection spot. A similar relation between the number of larvae and the number of early infections was established for the three cultivars. Moreover, the number of early infection spots increased significantly as cluster compactness increased.

Results and conclusions
The connection between the number of early infection spots and disease severity at harvest was tested statistically using linear regressions. The occurrence and closeness of the connection varied with the cultivar, cluster compactness and year. The climatic conditions in August and September were particularly relevant for analysing annual variations. Positive correlations between the number of early infection spots and disease severity at harvest were significant in 1993 and 1995, but not in 1994. In 1995, the unfavourable climatic conditions (hot-dry) for the disease resulted in the persistence of dry early disease spots without any further development. In 1993 and 1994, the climatic conditions were conducive for rot development, particularly in September. In 1993, one early infection spot led to an increase of 10-20% in disease severity. In 1994, a similar trend was obtained for control clusters of Cf only. The absence of correlation in 1994 may be due to the climatic conditions conducive for rot development as early as mid-August, which caused numerous secondary cycles of infection.

Thus, the occurrence of a quantitative relationship between the incidence of early Botrytis rot and fruit rot at harvest depends on climatic conditions. Research to model the effects of climate on the relationship are in progress.

1. McCiellan WD, Hewitt WB, 1973. Phytopathology 63, 1151-1157.
2. Fermaud M, LeMenn R, 1992. Phytopathology 82, 1393-1398.