• Dendrobatidae
  • Aromobates
  • Colostethus
  • Dendrobates
  • Epipedobates
  • Mannophryne
  • Minyobates
  • Phyllobates

Ford and Cannatella (1993) defined Dendrobatidae to be the node-based name for the most recent ancestor of Allobates, Aromobates, Colostethus, Dendrobates, Epipedobates, Minyobates, Phobobates, and Phyllobates, and all of its descendants. Synapomorphies include the presence of a retroarticular process of mandible, and the arrangement of the superficial slip of the m. depressor mandibulae (Ford, 1989b; Myers and Ford, 1986). Other features that have been argued to support monophyly include the tendon of the semitendinosus piercing the mm. gracilis major and minor (Noble, 1922), egg attendance culminating in the transport of aquatic larvae on the dorsum of the parent (Weygoldt, 1987), divided dermal scutes on dorsal surfaces of fingers, and cephalic amplexus (Duellman and Trueb, 1986).

The retroarticular process and conformation of the depressor mandibulae muscle are unique to dendrobatids. The use of other features as synapomorphies assumes that the similar conditions observed in other neobatrachians are convergent. The thigh muscle pattern has been reported in some myobatrachids, some Physalaemus, and Crossodactylus (Lynch, 1971, 1973). However, L. Ford has observed the "bufonoid" condition in Crossodactylus. Dorsal transport of the larva is reported in Rana microdisca, Cyclorhamphus stejnegeri, and Sooglossus seychellensis (Duellman and Trueb, 1986) but Weygoldt (1987) emphasized that dendrobatid parental care was unique in its behavioral complexity. Divided finger scutes have been reported in other frogs, such as hylodine leptodactylids and some myobatrachids (Lynch, 1971).

Data for the amplexus character are meager, but its presence in putative basal (Colostethus) and more deeply embedded taxa (Epipedobates and Phyllobates) suggests that cephalic amplexus is ancestral for Dendrobatidae; its condition in Aromobates nocturnus, the putative sister-group of all other dendrobatids, is unknown (Myers et al., 1991). Two closely related species of Dendrobates lack amplexus (Duellman and Trueb, 1986), but this is probably derived within Dendrobatidae. The less parsimonious alternative is that the cephalic amplexus of dendrobatids is independently evolved from amplexus in other frogs. These and other diagnostic characters of Dendrobatidae were discussed by Ford (1989b).

Relationships of Dendrobatidae to other neobatrachians have been controversial. Griffiths (1959b) allied Dendrobatidae to petropedetine ranids. Noble (1926, 1931) and Lynch (1971, 1973) argued that dendrobatids were derived from hylodine leptodactylids. Following Griffiths, Duellman and Trueb (1986) placed Dendrobatidae with ranoids, although not explicitly with petropedetines. Laurent (1979, 1986) included Dendrobatidae among hyloids, rather than in ranoids. Dendrobatids have a cartilaginous sternum, horizontal pupil, fused second distal carpal, fused second distal tarsal, and an unnotched tongue, all of which are plesiomorphic at the level of Ranoidea. These data do not refute the placement of Dendrobatidae in Ranoidea, but rather suggests that, like Microhylidae, dendrobatids are not nested within Laurent's Ranidae or Hyperoliidae.

Ford (1993) reviewed the controversy concerning the placement of dendrobatids and analyzed the relevant characters in support of each hypothesis. The position of Dendrobatidae was addressed by Ford (1989b) in a phylogenetic analysis of neobatrachians, with an emphasis on ranids and leptodactylids. She concluded that dendrobatids were nested within Ranoidea, close to arthroleptid and petropedetine ranoids.