The Graduate School of Experimental Plant Science, Department of Plant Breeding, Wageningen Agricultural University, PO Box 386, 6700 AJ Wageningen, The Netherlands

Background and objectives
A new powdery mildew on tomato has frequently been reported in greenhouse and field crops in the Northern Hemisphere since 1986. All cultivated tomatoes are susceptible. The causal agent has been identified as Oidium lycopersicum [1] in some countries, but provisionally designated Erysiphe sp. in many others. Comparison of the reports suggests differences in morphology and host range of the causal pathogen (unpublished data). Hence, more than one powdery mildew species may be responsible for the recent outbreaks. If this is true, this would have important consequences for tomato breeding programmes. The present investigations were conducted in order to discover the relationships among the powdery mildew fungi occurring on tomato in different countries and these fungi and other closely related powdery mildews.

Materials and methods
Four isolates of tomato powdery mildew were collected in the Netherlands (presumed to be O. lycopersicum), one each in Canada, Czech Republic, France and Hungary. The Dutch, Canadian and Czech isolates, and one isolate of E. pisi on pea from the UK were maintained and propagated in climate chambers. We also included one isolate each of the powdery mildews collected from cucumber (Cucumis sativus), potato (Solanum tuberosum), lettuce (Lactuca sativa), Begonia (Begonia sp.), rape (Brassica rapa), knot-grass (Polygonum aviculare), bear's paw (Heracleum mantegazzianum), lupin (Lupinus mutabilis), hemp nettle (Galeopsis sp.), plantain (Plantago major) and burdock (Arctium lappa). The length (L), width (W) and the L:W ratio of the spores were measured and the arrangement of the spores on conidiophores (single or in chains) was observed microscopically for all isolates except the Hungarian one. Host range tests of 46 accessions from 15 plant species were carried out with only one Dutch powdery mildew isolate. DNA was extracted by CTAB procedures. AFLP analysis was used to determine the relationship of these isolates.

Results and conclusions
Based on their hosts and spore arrangement, the isolates were classified into three groups: (1) tomato powdery mildews, spores formed singly or in chains; (2) powdery mildews not collected from tomato, spores formed singly; (3) powdery mildews not from tomato, spores in chains. In general, significant differences occurred in conidium length between the groups. For example, spores in Group 2 were much longer than those in Groups 1 and 3, except for E. heraclei on bear's paw and E. pisi on lupin. The shortest spores were produced in Group 3. However, great variation also existed within each group. In Group 1, for example, the Czech and French isolates had significantly shorter spores than most of the Dutch ones, and resembled those in Group 3. The variation in width was similar to that in length. The highest L:W ratio was found in Group 2 (>1.88) and the lowest in Group 3 (<1.70 except for lettuce powdery mildew) but there was some overlap in the L:W ratio, especially in Groups 1 and 2. Therefore, the size and shape of spores was insufficient to distinguish these powdery mildew species.

The host range tests with tobacco, cucumber, potato, eggplant and sweet pepper, among others, showed that the pathogenicity of the Dutch isolate differed from what has been reported in Canada, Czech Republic, Hungary, UK and USA. For example, the Dutch isolate was pathogenic to tobacco but almost not to cucumber, while the Czech isolate was the opposite. In order to confirm these differences, DNA fingerprinting is being applied using AFLP techniques. Preliminary data have revealed that the four Dutch O. lycopersicum isolates have rather similar AFLP fingerprints, but would be classified into two groups that are very distinct from potato powdery mildew (E. polyphaga?) or cucumber powdery mildew (S. fuligenea).

1. Noordeloos ME Loerakker WM, 1989. Persoonia 14(1), 51-60.