No label defined (Q4949790)

This paper is interesting. However, no details of all the constructions used are given here. From the summary: Suppose that the block size for a balanced incomplete block design (BIBD) for \(v\) treatments, each replicated \(r\) times, is a multiple of \(m\), and that each of the \(b\) blocks is partitioned into \(m\) sub-blocks in such a way that the \(mb\) sub-blocks themselves constitute a BIBD. We then have a nested BIBD, as defined in the statistical literature in 1967. Suppose further that the sub-block size is a multiple of \(n\), and that each of the \(mb\) sub-blocks is partitioned into \(n\) sub-sub-blocks in such a way that the \(mnb\) sub-sub-blocks constitute yet another BIBD. We then have a doubly nested BIBD. Many doubly nested BIBDs are known to exist and are not hard to construct. Such designs with \(r\leq 30\) and \(v\leq 20\) are discussed and tabulated.