109 — REPORT III THE INFLUENCE OF ENYIRONIENT AND AaEICULTURAL TREATMENT ON THE COTTON CROP.

E. E. MASSEY, Economic Botanist.

In the present report it is proposed to consider briefly the effect of the various environmental factors already described on the development of the cotton plant.

Owing to the impossibility of obtaining an artificial approximation to field conditions by laboratory methods, precise data on the physiology of the plant are lacking. Moreover, the operating factors influencing and limiting development are so interwoven that an attempt to isolate any one for comparative development has always resulted in failure. A scheme of observations has therefore been elaborated by which the most salient features of the life history of the plan,t may be correlated to some extent with the measurable factors of the environment.

Methods.—The general arrangement of the experimental plots has been described elsewhere, and it only remains to add that in each plot rows wpre marked off, from which all observations were taken throughout the season. In many instances the experimental area consisted of half a feddan, the other half being employed as a control or check plot. To ensure uniformity, an observation row was chosen in each half, so that every feddan contained two such rows. Weekly measurements and counts were made of the following features : (a) height, (6) number of flowers, (c) the number of bolls per plant, (d) number of open boUs per plant. The minimum number of plants employed was a hundred, except in the case of the height, where the average of twenty-five was found sufficient.

It will be fairly obvious that when conducting a number of experiments each season with a limited supply of labour, uniformity of agricultural treatment could not be ensured if the laborious chessboard system of plots were adopted.

It was felt that sufficient accuracy could be obtained and inequalities due to the soil eliminated, if observations were taken from a row extending the whole length of the plot.

A comparison was made of the two methods by dividing a plot into a number of equal areas and counting groups of thirty-five plants in each.