Piper: Agropykon tenerum and its allies 545 



A. fene?'um iillaium 

 A, Novae-An^Hiie. 



Type specimen collected on Willoughby Mountain, Vermont, 

 by Grout & Eggleston, July 2, 1894. 



This plant is about 90 cm. high, has flat blades, scabrous on 

 both sides, smooth sheaths, rather slender spikes 10-17 cm. long, 

 and hairy callus and rachilla. 



The tall habit and longer leaf-blades are relied upon by the 

 author of the species to distinguish it from A. violaceuut Lange, 

 and the broad leaves and hairy rachilla from A, tenerum Va.sey. 



It will be noted that the types of two of the above plants, 

 namely A, tenerum ciliatum and A. Novae- Ayigliae, have hairy 

 rachillae ; the remamder have the rachilla merely puberulent. 



Taking these characters which have been pointed out and 

 relied upon as distinctive, the following classification is possible, 

 based partly on general interpretation, partly on types. 



Rachilla hairy. 



Sheaths pubescent. 

 Sheaths smooth. 

 Rachilla scabrous or puberulent. 



Sheaths pubescent. Unnamed. 



Sheaths smooth. 



Leaf-blades narrow or involute ; spike slender. . > 



Spike 10-15 cm, long; leaf-blades rathershort. A, tenerum. 



Spike 20-25 cm. long; leaf-blades long. A. tenerum longi/oiium. 



Leaf-blades flat, spike a little stouter. 



Stems 30- 1 DO cm. high; blades 12-20 cm. long. A. psettdorepens. 

 Stems 100-150 cm. high; blades 20-30 cm. 



long ; spike very stout, somewhat one-sided. A, pseudorepens ma^num^ 



Such a classification is quite artificial. The leaf-blade charac- 

 ter seems entirely worthless, as all possible jntergrades occur in 

 large percentage. 



To determine the value of the hairy-rachilla character all the 

 specimens with this peculiarity were selected from among the 

 specimens labelled A. tenerum ^.Vi^ A. pseudorepens, about one- 

 fourth of each being thus distinguished. Comparatively little 

 difficulty was experienced in deciding between appressed-puberu- 

 lent and hirsute rachillae, but nevertheless puzzling intermediates 

 occur. 



. The specimens thus selected showed all the variations as to 

 leaf-blades and habit shown by the remainder, and furthermore 

 the two lots were practically coextensive in distribution. 



