The eastern areas of the South Island of New Zealand have marked rainfall variability. There has been, from time to time, interest in augmenting rainfall by artificial methods. Using recorded rainfall, numerical experiments are conducted to determine by Monte Carlo methods the probability of statistically detecting particular rainfall increases in experiments lasting 5 or 10 years and conducted over the November-April period. A rainfall increase of 30% on seeded occasions would probably be detected but smaller increases might not be identifiable even in 10 years. Restriction of seeding to occasions when there is a large water need has little effect on the probability of detecting effects of seeding.