“…Despite the relevance of the c1ustering effect, many defect level models do not account for it [Williams and Brown, 1981, Agrawal et al, 1982, Sousa et al, 1996, but there are a few models that do [Seth andAgrawal, 1984, Singh andKrishna, 1996]. In [Seth and Agrawal, 1984] the c1ustering effect is implicitly taken into account by using the negative binomial distribution for the number of defects in a chip.…”