The discovery of a direct ancestor, Ipomoea trifida (6x), of the sweet potato (6x) in 1961-1962 has been followed by a further finding of its diploid (I, leucantha) and tetraploid (I. Littoralis) predecessors. On the basis of cytogenetical study it was reasonably assumed that I. Littoralis (4x), I. Trifida-(3x)-6x, I. Trifida (6x) and I. Batatas (6x), sweet potato, are autoploids derived rather from the doubling of a set of 15 chromosome pairs (genome B) of I. Leucantha (2x) than from segmental alloploidy. On the other hand an artificial hexaploid, I. Littocantha, synthesized from I. Leucantha (2x) and I. Littoralis (4x) was proved to have the same genome constitution as I. Trifida-(3x)-6x and sweet potato, their Fl hybrids having 45 bivalents or modified configurations. The hybrids were fertile. Another artificial hexaploid, I. lacunocilis, derived from I. lacunosa (2x) and I. gracilis (4x) was self-fertile, but usually did not hybridize either with I. trifida or sweet potato.
The phylogenetic relationship was fairly well supported by the pattern of the mating system in interspecific crosses.
