A sampling scheme consisting of three stages. The first step is to set the sampling interval. The second stage is a random selection of units from the first lottery interval. In the third step, we select the remaining units of the sample as all those units that are distant from the drawn unit by multiples of the sampling interval. The key feature of this sampling scheme is that if the population units are ordered according to features that have a direct bearing on the feature being investigated, this pattern increases the representativeness of the sample.