STUDIES ON THE CHARACTERISTICS AND APPLICATION OF THE SPATIAL DISTRIBUTION PARAMETERS OF RICE SKIPPER EGGS IN RICE FIELDS

This paper deals with the spatial distribution patterns of the rampant generation eggs of the rice skipper, Parnara guttata Bremer et Grey, in hybrid rice-producing area of Northwest Sichuan.Five aggregation indices including: k, CA, lDM, M and I8 together with lwao's method were used to test the distribution pattern of rice skipper eggs and (he structure of its spatial distribu-tion. It was found that all the indicrs indicate that the distribution of the eggs is in the aggre-gative pattern. The paramcters a, (3 of Iwao's can be used to prove that the eggs in rice fields fit the negative binomial distribution. They occur in leose colonies the distribution of which is aggregative and the intra colony distribution is candom.On the basis of the M - x relationship,the spatial pattern of rice skipper eggs fets the negative binomial distribution. The reasons for aggrcgation, eggs mortality on rice plant bills, and mean colony size are analyzed with the parameters of aggregation K and mean crowding M The result shows that the aggregation of eggss seems to be due to some environmental factors as the mean density per hill is less than about 2.21, and beyond this, the aggregation may be due both to active aggregation by the adults and to certain Leterogeneity of the environment.Since the distribution of eggs follows a negative binomal pattern, the proportion of plants uncarrying eggs (Po) and the number of eggs per 100 hills (x )may be expressed by the equa-tion:i.e., the mean density per 100 hills can be estimated from the proportion of zero frequency (Po), as in the relationship shown in Fig 6. This method provides an alternative to sequential sam-pling as a basis for pest control decision. !t is particularly useful if the critical density levels are related to the value of Po bigger than 0,1 and below this, the uncertainty involved in predic-tions may be 韔o greai. The necessary sample size (n) for a fised precision level (D) is also shown in Fig 6. It can be seen that the necessary sample size increases gradually with increa-sing Po.The sequential sampling based on a binomial distribution which corresponds to an esti-mation based on the frequency of zero is made. The chart for the sequential test is shown in Fig. 8.