88 THE PHENOMENON OF 



middle, simply a protuberance, while the lateral portions 

 growing together on the side opposite to the middle, 

 acquire very considerable development. Hence arises an 

 appearance as though the subtending leaf stood opposite 

 to the branch. It often attains a considerable size and a 

 more or less foliaceous expansion, exhibiting a distinction 

 into stalk and blade. The sheath usually reaches a length 

 of li to 3 inches; the blade seated upon this, about of 

 equal length, is double, on account of the absence of the 

 middle, the two halves diverging at an obtuse angle. 

 Among the grasses in which the disappearance of the 

 leaves in the inflorescence is most complete, so that even 

 the lowermost often does not even leave a protuberant 

 process, are Catapodiimi, many species of Eragrostis, 

 Eleusine, Bigitaria, and Sorghum. In the spikelets, 

 finally, the leaf-formation comes to light again, frequently 

 gradually, as in Oryza* and all other grasses in which 

 the first glume is very small {Vulpia, Airochloa, Anthox- 

 anihu7n) ; frequently suddenly, as in all grasses with two 

 large glumes (IIolcus, PJialaris). Most of the many- 

 flowered grasses exhibit, further, a distinct increase and 

 falling off in the successive hypsophyllary leaves of the 

 spikelet, since the first sterile palete are shorter than the 

 succeeding fertile ones, which themselves again decrease 

 •ill size towards the point of the spikelet.f It is rarer for 

 the first palese to be the largest, so that a simply decreasing 

 condition exists, j The reverse, a merely ascending con- 

 dition of the palese of the spikelet, is exhibited by many 

 one-flowered grasses, in particular the already cited Rice, 

 and most of the genera allied to Panicum.% 



* In Oryza the spikelet begins witli four sterile palc83, followed by only 



one fertile. The first two or throe sterile palese appear only as small teeth. 

 Leersia is distinguished from Ori/za merely through the still more imperfect 

 development of the first four palese of the spikelet. 



f Thus, for instance, in Bromus, Festttca, Poa, Bactylis, Secale, Melica, 

 Mollnia, Phraijmiies, Clitoris. 



X 'fhus in Triodia, Agrostis, Calamagrostis. 



§ A decrease may take place in these in an abnormal manner, when, 

 namely, the axis or rachis of the spikelet develops additional palea; (and 

 flowers in llii'ir axils). An intcicstiug case of this kind occurs almost 



