138 WILD AND CULTIVATED COTTONS 



Soudan 21. Var. soudanensis, Watt. 



Mr. Broun has recently contributed a specimen (n. 693) of a plant 

 which he calls ' Soudan Country Cotton.' A further specimen of 

 what I take to be this same plant has been furnished to me by 

 Mr. W. Lawrence Balls (n. 213-2) of the Khedivial Agricultural 

 Society, and the note which accompanies the last mentioned speaks 

 of it as one of the tree-cottons of Senaar. This, while closely allied 

 to the roji of Gujarat, differs sufficiently to make me believe one 

 or two special African forms might with advantage be regarded as 

 separable from the corresponding Indian races to warrant their 

 being given distinctive names. I refrain, however, from attempting 

 to take this step at present, owing to want of sufficient material 

 and data. I have provisionally placed certain African and other 

 specimens under roji, but think it as well to isolate the present 

 Soudan plant. 



Dcscrip- Soudan indigenous cotton is a large perennial bush with the 



* lon ' leaves quite glabrous, deeply segmented and having supplementary 



lobes, also one to three glands below. Inflorescence axillary, 

 bracteoles broad ovate, two-thirds of the cordature united, and 

 margin deeply laciniate. Seeds large, coarse, ovate obtuse, with 

 copious brownish-grey fuzz and a fair quantity of harsh woolly floss. 

 In shape of leaf this comes nearer to G. arboreum than to 

 G. Nanking. But a yellow-flowered G. arboreum with deeply laciniate 

 bracteoles and three glands on the undersurface of the leaf would 

 destroy the specific isolations, while these peculiarities would not be 

 difficult of acceptance as denoting a form of G. Nanking, more 

 especially since the present plant possesses many other features 

 strongly suggestive of that species. 



General Characteristics. To conclude this account of the forms 

 of G. Nanking, it may have been observed that I have not attempted 

 to describe the races that might be mentioned as examples of each 

 of these varieties, himalayana, rubicunda, Nadam, Bani, JRoji, and 

 soudanensis. The races that can be assorted under these cannot be 

 regarded botanically as anything more than sports or hybrids, but an 

 infinite series of them exists, sufficient to almost connect G. Nanking 

 with G. arboreum on the one side, and with G. obtusifolium on the 

 other. In the localities where fairly extensively produced, some of 

 these are, however, well known, and have been so for many years, 

 because they afford recognisable trade qualities of staple. But it is 

 next to impossible to furnish descriptions that would be intelligible 



