Abstract: From April to June, 1975,1977 and 1978,a randomized cloud seeding experiment with metaldahyde(MA) and silver iodide (AgI) from the ground by the small rockets was carried out in Gutian reservoir area, Fujian province. In this experiment, the design of randomized regression of area was used and the experimental effect was evaluated with the two-sample regression analysis with unequal residusl variances. The experimental objects were mainly the stratiform and cumuliform clouds under the influence of precipitating weather system, and the period of three hours was taken as experimental time unit. The experimental units were selected totaled 101,50 of them were seeded and the others unseeded as controlled by randomized procedure. The statistical analysis indicated that: (1) the nor-mality of the transformation of areal rainfall was very good, which the goodness of fit reached 0.94,and the correlation coefficient of this rainfall index between the two experimental areas exceeded 0.82;(2) the difference between the residual variances of seeding and of control samples in the regression analysis might be significent, hence it was necessary to use the two-sample regression analysis method with unequal residual variances for testing the seeding effect; (3) the average rainfall in the target area(1500 km2) was increased aboud 20.3% by seeding, with a significant level of 0.05 (one side test); (4) the seeding effect under frontal weather was 70.6% in average (a-1),the effect was non-significant; (6) the area of effect exceeding 2000 extended 10-50 km downwind of the operation set, and it's width was about 20 km; (7) the difference between seeding effects of MA and AgI was non-significant, although the amount released once of MA was about ten times larger than AgI.