RÉSUMÉ
BACKGROUND: The secondary genepool of our modern cultivated potato (Solanum tuberosum L.) consists of a large number of tuber-bearing wild Solanum species under Solanum section Petota. One of the major taxonomic problems in section Petota is that the series classification (as put forward by Hawkes) is problematic and the boundaries of some series are unclear. In addition, the classification has received only partial cladistic support in all molecular studies carried out to date. The aim of the present study is to describe the structure present in section Petota. When possible, at least 5 accessions from each available species and 5 individual plants per accession (totally approx. 5000 plants) were genotyped using over 200 AFLP markers. This resulted in the largest dataset ever constructed for Solanum section Petota. The data obtained are used to evaluate the 21 series hypothesis put forward by Hawkes and the 4 clade hypothesis of Spooner and co-workers. RESULTS: We constructed a NJ tree for 4929 genotypes. For the other analyses, due to practical reasons, a condensed dataset was created consisting of one representative genotype from each available accession. We show a NJ jackknife and a MP jackknife tree. A large part of both trees consists of a polytomy. Some structure is still visible in both trees, supported by jackknife values above 69. We use these branches with >69 jackknife support in the NJ jackknife tree as a basis for informal species groups. The informal species groups recognized are: Mexican diploids, Acaulia, Iopetala, Longipedicellata, polyploid Conicibaccata, diploid Conicibaccata, Circaeifolia, diploid Piurana and tetraploid Piurana. CONCLUSION: Most of the series that Hawkes and his predecessors designated can not be accepted as natural groups, based on our study. Neither do we find proof for the 4 clades proposed by Spooner and co-workers. A few species groups have high support and their inner structure displays also supported subdivisions, while a large part of the species cannot be structured at all. We believe that the lack of structure is not due to any methodological problem but represents the real biological situation within section Petota.