Choose a number at random from 1 to n, and then take every nth value thereafter. This means that all units have an equal chance of being selected, but it does not constitute a random sample, as not all combinations are equally likely. There is a particular danger with this method, especially for ‘production line sampling’ as any faults with a particular machine in the system will affect items regularly, and the systematic method of choosing the items to be sampled means that you are very likely either to miss them all or hit them all (or some fraction of them).