SECTION II: VAR. ROJI 137 



method recently introduced. This may possibly point to a later 

 extended production of the annual crop. But by the date of Hove's 

 visit (1787) the roji cotton had been assigned the secondary position i n tro- 



it has since held. Dr. Hove's specimens are in the British Museum, duction of 



annual 

 and it has to be admitted that they could not be separated botanically cotton. 



from any corresponding set of more recent date. Had Hove 

 attempted to furnish particulars of the relative abundance of the 

 various plants seen by him, his report would have been to-day of 

 infinite value. But there is one circumstance that may be repeated, 

 namely, that none of Hove's samples match the higher grade 

 kahnami of to-day, and hence again a suggestion of their later in- 

 troduction (or rather production) than the others, especially the 

 roji. 



CULTIVATION. Roji cotton is met with on light soils, and mostly 

 as a mixed crop one row between ten or twelve of cereals. In the 

 hot weather it is cut down to within one foot, and in the second 

 monsoon it grows freely and yields its first crop in the incoming hot 

 months. 



Degeneration. When left to grow in the hedgerows roji becomes Short red 

 sub-scandent, the wool shortens and turns rufous-coloured, and the 

 fuzz lengthens and becomes also red-coloured. When left in the 

 fields for more than three or four seasons the wool degenerates in 

 quality, and, as stated by the early authors, is then fit for upholstery 

 purposes only. But Professor Middleton observes: 'Roji is 

 markedly different from the annual cottons and does not seem to 

 hybridise with them. I have never seen any plant that might be 

 taken as a cross. It strongly resembles G. arboreum, the chief 

 difference being a yellow flower and the absence of the marked 

 reddish tinge possessed by that species.' In G. arboreum it is the 

 rule for the twigs, leaves, and bracteoles to have a purplish tinge in 

 addition to the flowers being also purple (cf. p. 82) ; in G. Nanking 

 these structures are usually green (cf. p. 134), but a pink tinge is 

 characteristic of the nddam cottons, and in fact originated the name 

 yerra as applied to them (cf. p. 128). One more circumstance maybe 

 here added regarding roji cotton the pollen-grains appear dominant. 

 Usually only one form and size is met with (see Plate No. 53, f. 12), and 

 that so very different from the grains in either G. arboreum or most 

 other forms of G. Nanking as to convey the impression of its being 

 fully established, a dominant state, that demands independent re- 

 cognition in all breeding experiments. 



