13 



It will be seen from this table that the percentages of the 

 constituents in the boiling stage are smaller in mogt instan- 

 ces than the corresponding percentages in the flowering 

 stage. It will be coaveuieut to refer to this decrease in 

 values in per cents of those of the earlier stage. la the 

 Drake field we find the decrease in potash in No. 6 to be 

 0.7 %, and in Nos. 8 and 9, there is an increase ot 6.1% and 

 0.6 % respectively ; while in the other two plots the decrease 

 is 38.2 "Vi a.nd 4:2.3 "o. It will be observed, also, that the 

 largest yields of cotton are in plots 6, 8, and 9. From this 

 it would seem that in the potash- fertilized plots there is a 

 sufficiency of that constituent under the circumstances here 

 existing. On the other hand, comparing the field and gar- 

 den, we find that while the latter has much higher percent- 

 ages of potash to begin with, it has at the same time larger 

 per cents of decrease than the potash-fertilized plots in the 

 field, ranging from 11.3 % in plot 8 to 53 "o in plot 7. In 

 other words, with a larger supply there is a smaller excess 

 of potash over the demands for that constituent. Little can 

 be learned from the figures relating to phosphoric acid. 

 The decrease ranges from 0.8 "o in plot 6 in the garden to 

 46.9 % in plot 8 in the field. The decrease in the values of 

 nitrogen is uniformly high, showing the great demand for 

 that valuable constituent. In the field the range is from 

 46 % in plot 5 to 53.6 % in plot 8, while in the garden it 

 runs from 36.4 % in 6 to 51.1 "o in 7. 



A few words with reference to the yield of cotton in pass- 

 ing. A reference to Table I will show that in the unfertil- 

 ized plots 1, 5, and 10 in each soil the yield is not the same, 

 but is lowest in 1 and highest in 10. This suggest that all 

 the plots are not uniformly fertile, but increase in fertility 

 from 1 to 10. This lack of uniformity in natural fertility, 

 will, of course, efiPect the results obtained by artificial fer- 

 tilization, but the effect of the latter on the yield is noticea- 

 ble, just as it was on the composition of the plant. By a 

 study of Table V we find that where we have high per- 

 centages of two or more constituents in the floweriug stage, 

 and a relatively low decrease of those percentages in passing 

 to the boiling stage, we have, generally speakiug, a large 

 yield. On the other hand, low, or even average, percent- 

 ages in the early, and a large decrease of the same in the 

 later stage, showing an insufficient supply from the soil, 

 means a relatively low yield. The application of this rule, 



