Speciation by hybridization in insects has been recently recognized on the basis of isozyme and chromosome studies showing that several species, either diploid or polyploid, have genomes that combine the genes and chromosome sets of two (or more) bisexual species. Until this evidence became available, thelytokous invertebrates were all considered uniparental derivatives of bisexual species. In this paper, we review examples including the stick insects Bacillus whitei, B. atticus, B. lynceorum, Leptynia hispanica D, Clonopsis gallica, Carausius morosus; the grasshopper Warramaba virgo; some Otiorrhynchus weevils; the planthopper Muellerianella 2-fairmairei–brevipennis; and black flies of the genera Gymnopais and Prosimulium. For several species (e.g., Warramaba virgo and Bacillus whitei), both parental taxa have been recognized, and their hybrid origin has been genetically assessed. In others (e.g., B. atticus), only one of the bisexual parental species has been detected; but their hybrid origin is supported by strong evidence, at both the isozyme and chromosome levels. For other supposed hybrid species (e.g., Clonopsis gallica, Carausius morosus), no bisexual ancestors have been detected, possibly because competition with their hybrid derivatives has made them rare or extinct. Insect hybrid species may differ in their mode of reproduction (apomictic or automictic thelytokous parthenogenesis, gynogenesis), degree of ploidy, and genetic structure (level of heterozygosity, clonal variation). The parallels between insect and vertebrate hybrid species, in which this phenomenon has been recognized and widely studied in the past 50 years, are drawn. The main problems involved in the origin and evolution of hybrid species are discussed, with particular regard to (i) changes in the maturation divisions allowing the transmission of the hybrid genome to the next generation, and (ii) their successful adaptation. The "spontaneous" and "hybrid" theories for the origin of unisexual forms are compared, with regard to hybrid species. An origin of hybrid species from occasional parthenogenetic development of hybrid eggs produced in areas of extensive interspecific hybridization (e.g., disturbed habitats) is suggested. Hybridization would not itself cause changes in the maturation divisions (which are controlled by genes of tychoparthenogenetic eggs) but only favour their selection through heterosis. The role of the so-called "heterotic" advantage (resulting from high levels of heterozygosity) and "demographic" advantage (resulting from all-female reproduction) in the evolutionary success of hybrid species is discussed. It is concluded that habitat disturbance by man is favouring both the onset of hybrid species and their successful spread.