120 



In the case of the former, the value of the results seems 

 dependent on the uniformity of the yields of the two control plots 

 and their agreement with the yield of the plots now treated in the 

 years prior to their treatment, while in the case of the latter the use 

 of the control is quite subordinated, one only being used, and no 

 indication of the range of variation or the causes of variations in the 

 plots under uniform treatment is obtainable. 



As far as published results go, and certainly as far as theoretical 

 considerations direat experiment, the only rational way where 

 experiment is framed on yield estimation (there are numerous cases 

 where growth-record experiments may be used to advantage) seems 

 to be the utilization of the results of an initial investigation of yield- 

 diiferences over a large number of blocks of palms similarly treated, 

 and the value of controls as a check certainly seem to lie in the 

 detection of the range between their maximum and minimum values. 



It is difficult in a paper of this nature and length, to indicate 

 clearly tlie interpretation which the experimenter will put to his 

 records, but the citation of the points above mentioned may not be 

 without interest in the light of what follows. 



Taken together they lie at the basis of the problem which every 

 estate manager is confronted with, viz., estimates — the one concern- 

 ing which he thanks his good fortune if his figures of crops for any 

 period in advance approaches the figures of the actual crops 

 produced. 



It has been mentioned previously in this paper, that experiments 

 are in hand to investigate the question of the incremental-crop produced 

 in young palms. Whatever values may be ultimately determined 

 by experimentation as being theoretically possible, there is the ever 

 present fact that the arrears in the number of ripe nuts obtained 

 under present conditions and methods are disproportionately great, 

 and here we approach the question of " falling nuts." So far, the 

 actual numbers have not been obtained for matui'e palms but it 

 is extremely probable that they are of high magnitude. In general 

 the nuts which fall may be placed into the following categories : 



(a) Those which are unpollinated and therefore unfertilized ; 



(b) Those which are pollinated but in which fertilization 



does not eventuate ; 



(c) Those which are pollinated and fertilized but which are 



thrown off in the run of the physiology of the tree 

 for reasons of the trees' inability to carry them. 

 Nutritional factors — water supply in particular — will 

 here play an impor-tant part, and may possibly largely 

 determine periodicity in crop production. 



In young palms category (c) will probably account for a larger 

 number than has been previously supposed and similarly in the case 



