Agrostis castellana, Boiss. & Reuter

Tutin, T. G., Heywood, V. H., Burges, N. A., Moore, D. M., Valentine, D. H., Walters, S. M. & Webb, D. A., 1980, Flora Europaea. Volume 5. Alismataceae to Orchidaceae (Monocotyledones), Cambridge: Cambridge University Press : 235

25. A. castellana Boiss. & Reuter , Diagn. Pl. Nov. Hisp. 26 (1842).

Perennial, with short rhizomes and sometimes stolons; stems 20-80 cm, usually smooth. Leaves flat or involute, nearly always aculeolate on the ribs; ligule 1-3 mm. Panicle (6-)8-16 (-25) cm; branches patent at anthesis, rarely afterwards, aculeolate, very unequal, branched 2-3 times. Spikelets 2-5-4 mm, greenish-yellow to purplish. Glumes lanceolate, long-acute, aculeolate on the keel, at least distally, sometimes with appressed hairs. Lemma c. f as long as the glumes, almost always with few to many, usually long hairs, the lateral veins raised in the distal and excurrent for 0-2-0-5 mm, but glabrous and without excurrent veins when unawned; awn up to 5 mm, arising near the base; callus with hairs up to 0-5 mm. Palea f-f as long as lemma. Anthers 1-1-5 mm. 2« = 28+ 0-4 B, 42 + 0-3 B. 5. Europe; more or less naturalized locally in W. & C. Europe. Al Az Bu Co Ga Gr Hs It Ju Lu.

A variable species whose limits are further obscured by probable hybridization with 19, 20 and 22. A number of variants have been given specific rank, among them A. olivetorum Godron in Gren. & Godron, Fl. Fr. 3: 483 (1856), which differs from the typical variant in lacking an awn.

A. parlatorei Breistr. , Bull. Soc. Bot. Fr. 110 (Sess. Extr.): 56 (1966) ( A. frondosa Ten. ex Sprengel, non Poiret), from S. Italy and some other parts of the Mediterranean region, probably belongs here. It resembles A. moldavica Dobrescu & Beldie , Anal. §ti. Univ.Al.I. Cuza Iasi (Biol.) 16:131 (1970), from E. Romania, in having leaves 3-7 mm wide, scabrid-hairy glumes and growing

in damp habitats, but the two differ in the length of the spikelet, which is said to be 2-3 mm in the former and 1-2-1-5 mm in the latter. They may, perhaps, be derived from hybrids between 25 and Polypogon viridis .

A biosystematic investigation of the whole complex and of its relationships with 19-22 is greatly needed.