Abstract: Two fragments of mandible, one left with P4—M2 and the other right with M1-M3, of the same individual (PA 885, holotype) , assigned to a new form of adapid, Sinoadapis carnosus gen. et sp. nov. were described in this paper.
About 380 specimens of this form including 48 maxillas and mandibles, 260 isolated teeth and 75 tooth rows, were collected from Lufeng, Yunnan. 60 per cent of them were unearthed from Layer II of Section D.
Diagnosis of the new genus: An Asian Miocene adapid primate, larger than that of either Sivaladapis or Indraloris. The lower fourth premolar is highly molarised and longer than any molar of the same individual. A distinct hypoconulid twinned with the entoconid, with a deep notch between them, is found in the lower fourth premolar and the buccal cingulum is weakly developed. The lower molars are short and broad. There is no cingulum on the buccal side. The buccal cingulum is present only on the distal side of the crown. The crowns are low and round.
Sivaladapis and Indraloris are two sufficiently similar genera. Some specimens of Indraloris have been attributed to Sivaladapis by Gingerich. As a result, only a few specimens represent Indraloris. Some features of the two genera ean also be seen in our specimens. According to Chopra, it seems not easy to distinguish Indraloris from Sivaladapis, so it is not impossible that they are of the same genus.
Sinoadapis carnosus gen. et sp. nov. from Lufeng is the latest record in late Miocene. This diseovery indicates that the final extinction of Adapidae may be later than the end of Miocene (8 Myr-) and it is helpful to the study of the natural environment of Ramapithecus.

Key words: Sinoadapis carnosus; Adapidae; Prosimia; Late Miocene; Lufeng