SECTION II. : G. OBTUSIFOLIUM 141 



between the two species. From a very extensive collection of Indian 

 Gossypia I can without difficulty, however, pick out all the examples of 

 G. obtusifolium (as here accepted) that have been recorded as found in 

 uncultivated positions. But it is not confined to India, since it is plentiful 

 in Africa. Moreover Vidal collected it in the Island of Ticao, Philippines ; 

 the label attached to his specimen (n. 2,184) describes it as ' wild.' One from 

 Madagascar in Baillon's herb. (n. 128) could not possibly be separated from 

 the wild plant of Gujarat. Eiedel found it in Letti and Lakor in the Malay 

 Archipelago, and his specimens were communicated by Dr. A. B. Meyer 

 in 1883. 



I have already alluded (pp. 117 and 127) to the fact that Mr. Burkill regards 

 certain sheets in the Calcutta Herbarium as ' type-specimens of Roxburgh's 

 G. obtusifolium. 1 Through the kindness of Captain A. T. Gage, Superinten- 

 dent of the Royal Botanic Gardens, I have just had these very specimens sent 

 to me for inspection and find myself under the necessity to dissent from 

 Mr. Burkill's opinion. On one of the sheets in question I recorded, in 1887, 

 a note to the effect that it was a hybrid ; I now endorse that view, as in 

 all probability correct. These ' type-specimens of Roxburgh's G. obtusi- 

 folium ' I regard as G. Nanking, var. rubicunda. They represent what 

 might be spoken of as one of the numerous crosses, often seen in the 

 Chinese, Japanese, Indian, and Malay cottons. And, as I have shown, 

 there is nothing to prove that these sheets were culled from even hybridised 

 descendants of Roxburgh's stock of G. obtusifolium. In the Wallichian 

 herbarium there are several sheets marked H.B.C., but none of them are 

 G. obtusifolium. In fact, so far as I have been able to discover, there is 

 no authentic type of Roxburgh's plant in existence in any herbarium, and 

 the plate and description here reproduced must be therefore accepted as the 

 only standards to go by. At all events, should the specimens in the Calcutta 

 Herbarium cited by Mr. Burkill be admitted as falling within the species 

 G. obtusifolium, they can never be upheld as Roxburghian types, since 

 they bear no such record. But their acceptance as manifestations of the 

 Roxburghian species would very possibly necessitate the suppression of 

 G. Nanking (being replaced by G. obtusifolium) and the promotion of 

 G. Wightianum to specific value, to embrace the wild plant of Gujarat and 

 all the finer cottons of Western India. From my point of view this would 

 be of no serious consequence, since to the cotton-grower it is of little moment 

 which name be used, so long as the existence of what I call the Chinese series 

 of forms be maintained as distinct from the Indian. 



Nomenclature. The above short description is abbreviated from Box- 

 Eoxburgh's original account of the species, and the more detailed t _-_ 

 particulars have been drawn up from my own material. The existing 

 information may accordingly be said to consist of a MS. coloured 

 illustration, made under Eoxburgh's supervision and named by him, 

 the original of which is in the Calcutta Herbarium and an exact 

 copy in the Eoyal Herbarium, Kew. I give, as Plate No. 20, a re- 

 production of it, photographed by the three-colour process, but reduced 

 in size. As stated above, no authentic specimen of Eoxburgh's 



