The series Variantia Ching et S. H. Wu mainly occur in China and its members are highly variable in morphology. The denomination on this group of Asplenium is very confused in the herbaria. We hop e by means of a biosystematic study to find out their genetic relationships in the reticulate evolution, and to raise a suggestion on their taxonomic treatment. Evidence from cytology, allozyme, morphology, and palynology shows that three ancestor diploids have formed Asplenium sarelii complex comprising 13 members. A. sarelii Hook. should be typified as a diploid. The so-called tetraploid 'A. sarelii' before is an allotetraploid that comes from the doubled hybrid between diploid A. sarelii and A. tenuicaule Hayata, which should be treated as a new species A. wudangense Z. R. Wang et X. Hou. A. pekinense Hance is an autotetraploid that comes from the doubled diploid ancestor A. sarelii. A. lushanense C. Chr., a diploid species and the only ancestor of A. yunnariense group, should not been sunk as a synonym of tetraploid A. yunnariense Franch. Most probably, A. varians Wall. ex Hook. et Grev. is an autotetraploid of A. tenuicaule Hayata. Three new natural tetraploid hybrids and their origins have been found out: they are A. x longmenense ( = A. pekinense x varians), A. x jingyunense ( = A. pekinense x yunnanense) and A. x kidoi ( = A. pekinense x wudangense). Three other new natural triploid hybrids have been found and their origins have been inferred: they are A. X huawuense ( = A. sarelii X wudangense), A. x luyunense ( = A. lushanense x yunnanense) and A. x teniuvaians ( = A. tenuicaule x varians). The method of allozyme comparion combined with cytological observation is employed to reveal the complicated relationships among the members of Asplenium sarelii complex in reticulate evolution and proved to be a highly effective tool to investigate the origin of polyploid and hybrid.
