We present DNA–hybridization data on 21 amniotes and two anurans showing that discrimination is obtained among most of these at the class and lower levels. Trees generated from these data largely agree with conventional views, for example in not associating birds and mammals. However, the sister relationships found here of the monotremes to marsupials, and of turtles to the alligator, are surprising results which are nonetheless consistent with the results of some other studies. The Marsupionta hypothesis of Gregory is reviewed, as are opinions about the placement of chelonians. Anatomical and reproductive data considered by Gregory do not unequivocally preclude a marsupial–monotreme special relationship, and there is other recent evidence for placing turtles within the Diapsida. We conclude that the evidential meaning of the molecular data is as shown in the trees, but that the topologies may be influenced by a base–compositional bias producing a seemingly slow evolutionary rate in monotremes, or by algorithmic artefacts (in the case of turtles as well).