Abstract: in the present paper Agrostis muliensis J.L.Yang and A.eriolepis Keng ex Y.C.Yang are recognized as new synonyms of A.pilosula Trin.Ver wallichiana (Steud.)Bor;A.arisan-montana Ohwi var.megalandra Y.C.Yang is included in A.arisan-montana Ohwi.The two species.Agrostis muliensis J.L.Yang(Typus: Muli, Sichuan, Zhao Qingsheng 6514) and A.eriolepis Keng ex Y.C.Yang (Typus: Kangdin, Sichuan, Liu Shimin 1335 ), have same morphological structures, close type habitat .overlapping distribution range and an only identical paratype: Weixi, Yunnan, C.W.Wang 68430.So I think that the two new name represent a same group.The diagnosis of the two new species derectly or inderectly indicated that their identification character from A.pilosula Trin.is that the inferior 3/4 of the lemma are pilose in the two new species ,whereas in A.pilosula the lemma is hairy all over.In fact,the distinction is not present.As the authority citation of the variety var.wallichiana (Steud.) Bor used to be incorrectly cited as (Hook.f.) Bor, so that some people did not know the origin of the name of var.wallichiana.In fact, Steudel (1854)did not say in the original description of A.wallichiana that the lemma was hairy all over,but said"valvulis plus minus dense pilosis, apice hyalinis acutis integris ".The figure of the report (Rajbhandari, 1985)from the type habitat of var.wallichiana also showed that the tops of the lemmas in the group are glabrous.The descriptions of var.walichiana that some authors (Hook.f., Bor, Hara.Rajbhandari)given in the past: "lemma hairing" or "lemma hairy all over" are just rough.They did not emphasize that the top of the lemma is naked.After the research on the type specimen of Agrostis arisan-montana Ohwi, J.F.Veldkamp (1982)indicated" Ohwi (1933)described the anthers as 0.3mm long, but in the type specimen all see were 0.6～0.7mm".So I include A.arisan-montana Ohwi var.megalandra Y.C.Yang which has 0.7～0.8mm long anthers in A.arisan-montana Ohwi.