Sequence A103505 is a duplicate of A002378 except for the first two terms. Is there a point to keep A103505 as a separate sequence? A deviation from A002378 looks inessential to me. At the same time A103505 misses a lot of references and useful comments from A002378. Max