1ORSTOM, BP A5, 98848 Nouma, New Caledonia; 2ORSTOM, BP 5045, Montpellier, France; 3CIFC, 2790 Oeiras, Portugal

Background and objectives
The objectives of the modelling investigations performed in New Caledonia on the Coffea arabica pathosystem are to understand its overall functioning, and to identify the environmental conditions which influence the dynamics of the diseases (cf. paper No. 1). To enhance this holistic approach and to increase the accuracy of disease risk forecasts [1], it was essential to evaluate the variability caused by the diversity of the main pathogens (Hemileia vastatrix and Colletotrichum gloeosporioides) and to the diversity of the host. Consequently, in addition to the epidemiological survey, four complementary biological studies were carried out.

Results and conclusions

Analysis of host genetic diversity. The existence of rust resistant coffee hybrids (canephora X arabica) growing in some coffee plots in New Caledonia may be a source of genetic heterogeneity that must be assessed in terms of susceptibility of the trees. For each of the 15 plots in the epidemiological survey, biomolecular analyses (RAPD) of the genome of the coffee trees surveyed were performed in Montpellier using the methodology developed by Lashermes et al.. [2]. Among the trees studied, no canephora character was found, thus excluding any effect on their behaviour vs. diseases. However, a molecular polymorphism indicates the existence of two cultivated coffee types typica and bourbon as well as intraspecific hybrids, with respective ratios varying within the plots. The sorting of the initial data files by 'plot' into subfiles by 'coffee cultivated types' shows differences in the respective susceptibility of these categories of trees.
Analyses of Hemileia vastatrix population. After the sampling of spores from every tree in every survey plot, a rust race identification was done in Oeiras using the CIFC method [3]. Three different races (I, II, III) were identified, with race II most common in some plots, the coexistence of the II and I or II and III races was observed. Complementary studies were developed to detect possible genetic diversity between the 25 rust samples collected (20 from New Caledonia, three from Papua New Guinea and two from Vanuatu). Among the ITS, REP, ERIC and RAPD methodologies tried, only RAPD (with severe constraints for spore collection and DNA extraction) performed with OPA9 and OPC8 primers revealed a polymorphism pattern between the rust sources. Trials to link these results with the characteristics of the races identified and with the dynamics of the diseases are in progress.
Analysis of Colletotrichum gloeosporioides population. Using biomolecular tools (RAPDs), 125 isolates collected in six surveyed plots (four in New Caledonia, one in Vanuatu and one in Papua New Guinea) were analysed [4]. A high level of genetic diversity was found in all populations and each isolate exhibited a unique RAPD haplotype. Genetic differentiation was found between three out of the six populations (Wright's index, Fst=0.27), but no relationships could be shown between the genetic variability of the populations and other traits such as environmental or epidemiological characteristics of each plot. A high recombination rate could be hypothesized to explain this situation.

Once included in the statistical model, the results of these biological investigations constitute important complementary information for devising a decision-making tool for integrated management of coffee diseases.

1. Nandris D, Kohler F, Monimeau L, Pellegrin F, 1997. Proceedings 17th International Conference on Coffee Science, ASIC, Nairobi, Kenya, in press.
2. Lashermes P, Trouslot 0, Anthony F, Combes MC, Charrier A, 1996. Euphytica 87, 59-64.
3. Rodrigues Jr CJ, Bettencourt AJ, Rijo L, 1975. Annual Review of Phytopathology 13, 49-70.
4. Faugeron S, 1996. Memoire de DEA, USTL-ENSA, Montpellier, No. 21.