PART 111. 
SECTION II. SUMMARY OF THE DISCUSSIONS. 
1. TrecaNIQUE OF FIELD EXPERIMENTS. 
In discussing the merits and demerits of field experimental 
methods the view was generally held that whatever method was used 
it should be so designed as to permit the degree of accuracy being 
determined. 
Attention was also called to the inaccuracy which may be 
involved in comparing varieties which are very different in habit and 
which may therefore be differently affected by the cultural conditions 
adopted; e.g., a uniform spacing for large and small types is obviously 
unsound. 
2. TecuNiQuE oF PrLaNT BREEDING. 
It was agreed that mass selection must be considered only as a 
temporary method during the time necessary to establish pure strains 
at plant breeding stations. Members of the Section compared their 
methods of management of their breeding blocks, the question of 
selfing receiving special attention. The advantages and disadvantages 
of stitching, tying, and bagging methods were discussed. The latter 
method possesses particular advantages in boll worm infested areas 
in the protection afforded against the attack of these pests. In 
Nyasaland bagging was found necessary on this account and a special 
method of converting a square of gauze into a bag by means of a 
rubber ring was described (vide Minutes of the Cotton Section, page 
148). 
The elimination of strains by means of observation rows appears 
to be the general practice in South and East Africa as a preliminary 
to full field trials of the survivors. 
3. Tne KrepiNg AND UTILISATION OF RECORDS. 
The Section accepted as an axiom the necessity for making the 
fullest possible observations of all factors affecting the plant and its 
yield, especially in the form of continuous records. Of even greater 
importance was the keeping of these records in such a way as to 
facilitate correlation. It appeared from the discussion that an 
opportunity exists in South and East Africa for the application of 
accurate knowledge obtained from correlated records embracing the 
incidence of insect attack in relation to climatic and soil factors and 
plant development. This may be of special importance in regard to 
the losses due to boll worms. For the obtaining of these records it 
was considered necessary that experimental stations be adequately 
equipped. (A list of suitable apparatus appears in the Minutes of 
the Section. See Enclosure, page 146). 
4. Tue Position witH REGARD To Cross BREEDING. 
This subject was thoroughly discussed and the Section was in 
agreement with the opinion of one of its members in saying that 
crossing ought to be valuable, but, in practice, did not appear to be 
so. If crossing is to be of any value, an early start should be made, 
for it takes a long time before any results are achieved. 
5. METHODS FOR THE EcoNOMICAL MULTIPLICATION OF SEED AND ITS 
DisTrIBUTION. 
This subject is of special importance in view of the urgency of 
supplying suitable strains with as little delay as possible. The main 
principle involved is the avoidance of all possible loss throuch thinning 
and through drought, flood or insect attack. The advantace of 
123