Recent analyses of nucleotide sequence data suggest that living placental m
ammals belong to one of four superorders. The early divergence of these gro
ups was followed by long periods of geographical isolation, due to the brea
k up of continental land masses, allowing for convergent evolution of simil
ar traits in different superorders. As an example, the transition from epit
heliochorial to haemochorial placentation occurred independently in bats, r
odents, anthropoid primates, armadillos and others. A group of ancient Afri
can mammals is suggested by the molecular data, but is not fully supported
by morphological evidence. The hypothesis is, however, consistent with some
of the data on fetal membranes, suggesting that it would be worthwhile to
study the early development of tenrecs, golden moles and elephant shrews. A
nalyses of fetal membrane traits that group the tarsiers with anthropoid pr
imates, and separate them from the lemurs, are challenged by the molecular
data. Other relatives of the primates seem to include tree shrews and flyin
g lemurs, and little is known about the fetal membranes of the latter group
. Comparative studies of placental function normally are confined to primat
es, rodents, lagomorphs and domestic animals: the biological diversity repr
esented by mammals that evolved in ancient Africa and South America is not
represented. Therefore, future comparative studies should strive to include
species such as the rock hyrax and the armadillo. (C) 2001 Harcourt Publis
hers Ltd.