Vol. V, No. 1.] Eaperimental Breeding of Indian Cottons. 29 
[V.S.] 
importance to form the basis of a primary division of at least 
that group of cottons which includes the Indian forms and 
which are capable of being cross-bred, and with this possi- 
bility in view observation on the branching habit is being 
extended to other types. The method, however, involves 
of relatively small importance. For example, a tertiary 
b 
are not wanting indications that such a classification is a natu- 
ral one. These cannot be taken in detail here: it be 
No such difficulty has occurred in such Crosses as Nurma (G. 
indicum, Lamk.) x Bengals (G. neglectum, Tod.).' _ 
4. Economic im mce.—The manner in which the date 
inferio: 
Sympodial types are consequently more suited for the plains of 
Northern India. 
Sympodial branches are, however, usually pendent, es- 
pecially in the later stages, when the weight of fruit is consider- 
at ne Se OT DTI rs eae oar WOOO OVE Ca 
1 Owing to the doubt here thrown on the classifications hitherto 
adopted, specific names have, as far as ible, been omitted. Wh 
they have been used they bear the same interpretation as in Pt. 1 of this 
note—Journ, As. Soc. Bengal (new series), iv (1908), p. 13. Considerable 
confusion already exists with regard to the terminology, and, in the light 
of the above doubt, it has been considered advisable, for clearness sake 
to retain for the time being the names previously used. rather cage 
those employed in the important work recently published by Sir George 
Watt, but which fall under the same doubt. 
