In the mature mammalian and avian central nervous systems, neuronal destructions are followed by reactive gliosis, but data on other vertebrates are rather controversial. Mammals and birds belong to different amniote groups (Synapsida and Diapsida, respectively), but exhibit common general features in their glial architecture, mainly the predominance of astrocytes. Two vertebrate groups seem to be in special positions of glial evolution: turtles (Testudiniformes) and skates and rays (Batoidea). The purely ependymoglial system of turtles seems to be the simplest one among the extant amniotes. In skates and rays, true astrocytes are preponderant glial elements, in contrast to the other "anamniotes" (and even to reptiles). We investigated stab wounds by the immunohistochemical detection of GFAP in turtles (Trachemys-formerly Pseudemys-scripta elegans), a skate (Raja clavata) and rays (Dasyatis akajei and Torpedo marmorata). Sharks (Scyliorhinus canicula) as ependymoglia-predominated chondrichthyans, and-for positive controls-rats were also studied. In the elasmobranchs, other astroglial markers: glutamine synthetase and S100 protein were also applied. Neither turtles nor elasmobranchs presented considerable astroglial reactions. Critically surveying the former reports on different vertebrates, these results complete the picture that typical post-lesion reactive gliosis is confined to mammals and birds. Analysis of the astroglial systems from phylogenetic perspective suggests that the capability of forming glial demarcation and scar formation evolved independently in mammals and birds. Predominance of astrocytes is a necessary condition but not sufficient for reactive gliosis. The intense glial reactivity of mammals and birds may be attributed to their complex cerebralization.