In the order Lipotyphla (Insectivora), certain reproductive features differ quite distinctly from the eutherian norms, and are of interest with regard to the evolution of mammalian gamete function and perhaps for questions of lipotyphlan phylogeny. As seen in one or more members of five lipotyphlan families (shrews, moles, hedgehogs, golden moles, tenrecs), these features can involve the configuration of the male tract including the penis, the morphology of the sperm head, the anatomy of the oviduct and the patterns of sperm transport within it, the character of the cumulus oophorus, and the way in which fertilising spermatozoa interact with the eggs. However, the picture is by no means uniform within the order. Reproductive idiosyncrasies occur variously in the different lipotyphlan families, and appear consistently and strikingly in shrews--the group studied most extensively. Compared to the patterns in most Eutheria, the most interesting anomalies in soricids include (a) the regulation of sperm transport to the site of fertilisation by oviduct crypts, whose arrangement can vary even according to species, (b) a circumscribed matrix-free cumulus oophorus that appears essential for fertilisation as the inducer of the acrosome reaction, (c) barbs on the acrosome-reacted sperm head by which it may attach to the zona pellucida. With regard to the bearing such reproductive traits might have on lipotyphlan systematics, the African mouse shrew (Myosorex varius) displays a mix of traits that characterize either crocidurine or soricine shrews, consistent with the proposal that it belongs in a more primitive tribe, Myosoricinae, or subfamily, the Crocidosoricinae, from which the crocidurine and soricine lines probably evolved. Moreover, although elephant shrews are assigned now to a separate order (Macroscelidea), they display several of the unusual reproductive features seen in lipotyphlans, particularly in chrysochlorids and tenrecs. On the other hand, if used as a phylogenetic yardstick, none of the reproductive features described serves to define the Lipotyphla as classically constituted within one order, nor necessarily all the relationships suggested by recent sequencing studies of nuclear and mitochondrial genes.