1 Inst. Plant Sciences, Swiss Federal Institute of Technology, 8092 Zürich, Switzerland; 2 National Plant Protection Centre, Research Extension and Irrigation Division, Ministry of Agriculture, Thimphu, Bhutan; 3International Rice Research Institute, PO Box 933, 1099 Manila, Philippines

Background and objectives
Over 90% of the total area under rice in Bhutan is planted to traditional varieties or landraces. They are given different names sometimes associated with places they are grown in. Rice breeding has to rely heavily on local materials that are uniquely adapted to the very cold conditions. Breeding for resistance to rice blast (caused by Pyricularia grisea) is hampered by a lack of information about resistance sources in adapted germplasm. Seed samples of traditional as well as some improved introduced varieties were collected from different regions of Bhutan for analysis of available resistance to blast and to determine if varieties are homogeneous for their resistance reactions.

Materials and methods
Resistance of 74 samples was tested in the Philippines against 20 isolates (with known lineages and virulence) and 70 samples were tested against 15 Bhutanese isolates. Samples that were resistant to all the isolates or whose resistances could not be distinguished were tested against 15 more Bhutanese isolates. Bhutanese isolates were selected representing different regions and putative genetic lineages if information was available and based on their reactions to five near-isogenic lines and their susceptible parent CO39. Forty to 60 plants per variety and isolate were Fourty to sixty plants per variety and isolate were inoculated at the three-leaf stage in a screenhouse and the individual plants were evaluated 7 days after inoculation in the Philippines and 9-10 days after inoculation in Bhutan. Progeny of 25 single panicles of two landraces each was multiplied and tested for their reaction to 20 selected Bhutanese isolates.

Results and conclusions
Fourty six of the seed samples were resistant to the Philippine isolates while only 23 were resistant to the Bhutanese isolates. Nine isolates (two from the Philippines and seven from Bhutan) did not attack any of the samples. The remaining 41 isolates differentiated the samples into 57 resistance types. Varieties ranged from completely resistant to all tested isolates to susceptible to more than 15 isolates. Samples of the same landrace or variety collected from different farmers often varied in their reactions. Also, reactions within samples were variable. For example, within the landrace Janam (kb) all plants were uniformly resistant or susceptible to 22 out of 35 tested isolates, but both resistant and susceptible reactions occurred with respect to 13 isolates. Varieties from lower areas were more resistant than varieties from the high altitudes presumably due to higher disease pressure at the seedling stage in the warmer lower areas and thus higher selection pressure [1].
Reactions of the single panicle progeny of the two landraces were variable within and among panicle progeny suggesting that farmers may mix seeds in the field and that some level of out-crossing is occurring.
Our data support the commonly held view that landraces are highly diverse for disease resistance and point to the great importance of maintaining the genetic resources in Bhutan. Care has to be taken in selecting parents for crosses due to their heterogeneity and more work is needed to identify a representative set of differential isolates of a manageable size that could be used for the screening of breeding materials.

1. Thinlay, Finckh MR, Bordeos AC, Zeigler RS, 1998. Agriculture Ecosystems and Environment (submitted).