The Tineoidea are the earliest-originating extant superfamily of the enormous clade Ditrysia, whose 152 000+ species make up 98% of the insect order Lepidoptera. Though more diverse than all non-ditrysian superfamilies put together (3719 vs 2604 species), the tineoids are not especially species-rich among ditrysian superfamilies. Their phylogenetic position, however, makes tineoids potentially important for understanding the causes of ditrysian hyperdiversity, through their effect on inferences about the traits of ancestral ditrysians. To reconstruct early ditrysian evolution, we need a firmly established ground plan for tineoids themselves, which in turn requires a robust knowledge of their biodiversity and phylogeny. Tineoid systematics is under-studied. The description of the world fauna remains very patchy, especially in the largest family, Tineidae, and phylogenetic studies within and among families have been few. Recently, molecular analyses have shown strong promise for advancing tineoid systematics. Here we present the largest tineoid molecular study to date, sampling five to 19 nuclear gene regions (6.6?14.7 kb) in 62 species, representing all tineoid groups ever assigned family rank, 25 of the 31 subfamilies recognized in recent classifications, and 40 genera spanning the morphological diversity of Tineidae, for which monophyly has not been established. Phylogenetic analysis used maximum likelihood, with synonymous substitutions alternatively included and excluded. The main findings confirm and extend those of other recent studies, as follows: (i) monophyly is strongly supported for Psychidae subsuming Arrhenophanidae, for Eriocottidae, and for Tineidae subsuming Acrolophidae but excluding Dryadaulinae and two genera previously assigned to Meessiinae; (ii) two new families are described, Dryadaulidae stat. rev. and Meessiidae stat. rev., based on subfamilies previously included in Tineidae but strongly excluded from this and all other families by our molecular results; (iii) Doleromorpha, formerly placed in Meessiinae sensu lato, is likewise here excluded from Tineidae, but left incertae sedis pending better characterization of what is potentially another new family; (iv) basal division of Tineidae sensu novo into ?tineine? and ?acrolophine? lineages is moderately to strongly supported, but most subfamily relationships within these lineages are very weakly supported, and polyphyly is confirmed for Meessiinae and Myrmecozelinae as previously defined; (v) basal division of Psychidae sensu novo into ?arrhenophanine? and ?psychine? lineages is moderately to strongly supported, as are most subfamily relationships within these lineages; (vi) Tineoidea are paraphyletic with respect to all other Ditrysia when synonymous substitutions are eliminated, with branching order (Meessiidae stat. rev. (Psychidae sensu novo ((Eriocottidae (Dryadaulidae stat. rev.+ Doleromorpha)) (Tineidae sensu novo+ all other Ditrysia)))). Support for tineoid non-monophyly varies, among the relevant nodes and among analyses, from weak to moderate to strong; and (vii) paraphyly of Tineoidea, coupled with parsimony mapping of feeding habits on the molecular phylogeny, suggests that the earliest ditrysians may typically have been detritivores and/or fungivores as larvae, like most extant tineoids, rather than host-specific feeders on higher plants, as in most non-ditrysians and most non-tineoid Ditrysia, i.e., the great majority of Lepidoptera. Thus, radiation of Ditrysia, a leading example of insect diversification linked to that of higher plants, may have started with reversion to feeding habits more like those of ancestral amphiesmenopterans.