HOLE : SOME INDIAN GRASSES AND THEIR CECOLOGY. 53 



Whatever characters are taken, however, whether the habit of growth, dimensions of culms, 

 width of leaves, length of spikelets, or of the callus-hairs, numerous intermediates connect- 

 ing the above forms can be found and observations in the field indicate that the characters 

 in question vary directly according to the quality and stability of the soil, liability to the 

 ri laying " action of wind and water, and the quantity of available moisture in the habii.it 



Plants of the above three forms of this species are now being cultivated in the Dehra 

 Dun Experimental Garden with the object of testing the effect of soil and varying quantities 

 of available moisture on the characteristics of the plants. Some tufted plants taken from 

 dry sand and planted in loam have already, in a single growing season, produced some wide- 

 spreading robust culms which resemble those of the decumbent loam-form. 



The African forms placed by Hackel under his sub-species cegyptiacum, variety 

 cegyptiacum, differ from the Indian plants examined by the writer chiefly by their slightlv 

 larger spikelets. This difference however is very slight and fails in the case of some African 

 specimens. Considering the great variability of the species in India it seems possible that 

 a more complete knowledge of the African plant will prove cegyptiacum to be merely one of 

 several cecological forms which are defined by inconstant characters and which are con- 

 nected by numerous intermediates. 1 



Hackel himself notes that intermediates do occur between his sub-species indicum and 

 cegyptiacum (I.e., p. 116). 



Roxburgh's species Saccharum spontaneum, S. semidecumbens and S. canaliculatunr 

 are undoubtedly all referable to the single species described as S. spontaneum in this paper. 

 It is however important to note that in his descriptions of all these three plants Roxburgh 

 states that the corol is one-valved. Now glume IV is frequently present in this species, 

 although as a rule it is minute and very narrow, while the pale is also present, although 

 the latter is minute and often shorter than the lodicules. Hence when Roxburgh speaks 

 of a " one-valved corol " in his descriptions it must not be assumed that glume IV and the 

 pale are absent, but rather that these parts are small and inconspicuous. This is im- 

 portant, since Hackel (I.e., p. 119) rejects Roxburgh's name of Saccharum Munja on the 

 ground that the description of that species as having a " two-valved corol " necessarily 

 implies the absence of the pale. 



Roxburgh's Saccharum eooaltatum is a doubtful plant. The sheet from Roxburgh's 

 collection representing this species in the Herbarium of the British Museum is only named 

 in pencil and is believed to have been named by Solander. The plant on this sheet is 

 undoubtedly that described as Saccharum Munja in this paper. 



1 The view taken in this paper regarding the limitation and definition of plant groups in descriptive botany is that it is neither advis - 

 able on the one hand to combine together distinct species or sub-species between which only isolated intermediates of probably hybrid 

 origin exist and which can therefore be distinguished in the field as distinct groups, nor. on the other hand, to separately describe and name 

 o-roups which are denned by inconstant characters varying with the cecological factors of the habitat (available moisture, soil, light, tem- 

 perature, etc.) and which are connected in the field by numerous intermediates. Groups of the latter description can he sufficiently indicated 

 by such terms as swamp-form, sand-form, shade-form, etc. 



2 Flora Indica, Ed. Carey, 1S32, Vol. I, pp. 235, 236, 2-16. 



r 53 i 



