Even though sorghum is a well adapted crop and is quite tolerant to water stress further improvement is possible. Mostly, sorghum is grown under rainfed conditions after other, more valuable field crops or under stress conditions that other, second crops could not tolerate. However, water stressing in critical periods (booting-blooming) probably decrease grain yield by more than 50 %. To prevent yield loss, early maturity sorghum has been used. But this is not the best because both biotic and abiotic factors can affect sorghum growth and yield. It is difficult to improve sorghum lines for stress condition based on yield only. The main objectives are to select drought resistance lines and to develop selection methods.

Grain yield of 7 tested lines and 3 check varieties (UT 1, SP 60 and SP1) were evaluated under stress and non stress conditions in rainfed areas at SFCRC, Nakorn Sawan FCRC and Ban Mai Samrong FCES, using a randomized complete block design, with 3 replications, of plot size 14.4 m2, with spacings of 0.60 x 0.25 m and 2 plants/hill during August 1995-Febuary 1997. Combined analysis of data was conducted using the MSTAT Program : Experiment Model#5. The four sorghum lines ; UT 377B, UT 470-2, UT 1 and SP 1 showed good yield stability under water stress conditions. The results also showed that under water stress conditions the mean grain yield and panicle weight decreased from between 42-50 % and 6-63 % respectively, with these depending on the G x E variance more than on genetic variance.

Selection of drought resistant sorghum lines were conducted under green house conditions during March-July 1995 at SFCRC. Fourteen sorghum lines were grown. Under water stress conditions there were 2 lines, UT 522B and UT 468-1, in which the root and shoot dry weight were higher than the others. High temperatures, at peak hot times, caused the death of tested plants making it impossible to find out correlations between the characteristics related to yield and drought resistance.

A pot experiment was conducted at Suphan Buri Field Crops Research Center between August and December 1995. Three sorghum varieties were combined with five different soil salinity levels (0, 0.2, 0.4, 0.6 and 0.8 % NaCl) in the RCB design with four replicates. Non saline soil was also included as a control. The results showed that there was a negative correlation between grain yield and salinity level. Sorghum yield also decreased when the soil salinity level was higher. There were differences among the sorghum varieties in their response to the salinity level. PB 12891 was the most tolerant sorghum variety because its yield was only slightly decreased in the high soil salinity level soil. It was also found that the critical salinity level for sorghum was 9 ds/m. The experiment will be repeated again in 1996.