Vol. XV. No. 366. 



THE AGRICULTURAL NEWS 



147 



•mulclied plots which enable the probable error to be 

 calculated. 



These experiments have indicated a steady gain from 

 the application of maniires, especiallj' from mulching. 

 Those who have followed these experiments know that 

 a marked feature has been the decline and death of 

 trees on the control plot. Constant supplying has 

 been necessary. Hence the recorded control yields are 

 reallv higher than they should be under strict experi- 

 . mental conditions of working. 



lu Trinidad the method adopted is similar to that 

 used in Dominica, except that two controls have been 

 employed. An important collateral series of observations 

 however, ha\e been made in Trinidad with reference to 

 the natural yields of plots similar to those manured on 

 the estates where the manurial stations are situated. 

 These have shown that the difference between the 

 yields of two untreated plots in the same year may be 

 far greater than that between a manured plot yield 

 and the yield of the average of the controls. For 

 instance, at Perseverance in 1913-14, one 'natural yield' 

 plot gave 47-28 pods per tree, while another gave only 

 9-30, a difference of 37-98. Now, in the experiments 

 •with manures on the same estate in the same year, the 

 highest yielding manured plot gave 40 pods per tree, 

 and the yield from the average of three control plots 

 was 31, a difference of only 9 pods per tree. This 

 is an extreme case, no doubt, but it shows how careful 

 investigators ought to be in using their controls as 

 standards on which to calculate monetary gain, or even 

 increased yields. These experiments have indicated 

 that but little benefit can be expected toith cer- 

 tainty from the use of manures. 



In the British Guiana experiments with cacao the 

 yields have been expressed in the form of the minimum 

 increase that can occur. The method can be expressed 

 in a formida thus: — 



Minimum increase = (v - p i ) - (c + p .j ) 

 Where 



V = average of results of manured plots similarly treated 

 in the same year and repeated over five year.s. 



Pj =the probable error of v. 



c — average of several controls repeated over five years. 



p., = probable error of c. 



This is a cautious method, and may be expected to 

 give results that are on the safe side: Detailed figures 

 have not been published, but the averages indicate 

 very considerable fluctuation. Thus two of the averag- 

 ed results for manured plots may or may not be higher 

 than the averaged controls. The results obtained 

 support in a general way those obtained in Dominica, 



except that quick-acting nitrogenous manures tend to 

 cause a decreased yield in British Guiana. 



The financial gain is obtained by substituting 

 money values for yield v.alues in the above formula. 



We have now described the chief methods of 

 experimentation that are being employed in the West 

 Indies in manurial experiments with coco-nuts and 

 cacao, and these have been commented on. Summing 

 up, it would appear that great caution must be exercised 

 in using the control results as standai-ds. The manurial 

 treatment of trees is a continuous process altogether 

 different from the manurial treatment of annual crops. 

 The behaviour of treated trees should be followed over a 

 long period of years, and this in itself can supply valu- 

 able information quite iiTespective of any control plot. 

 The determination of natural yields before starting the 

 experiment is very desirable, providing the trees are of 

 an age when productivity is constant from year to 

 year ('period b'). When enough results are available^ 

 the determination of the probable error, that is the 

 probable limit of fluctuation, is useful for arriving^ 

 at the minimum and maximum yields to be expected^ 



A great deal of disagreement in general results- 

 obtained is to be noticed in the West Indian experi- 

 ments. It has to be admitted that the effect of manures 

 will vary somewhat under different conditions of soil 

 and climate, but it is thought that if there were more- 

 uniformity of methods amongst investigators there 

 would, in all probability, be much better agree- 

 ment as regards results. There has been too strong 

 a tendency in the past to lay out manurial experiments 

 according to personal ideas rather than according to 

 scientific ideas. There is need for more concerted 

 action in these matters. Until sufficiently precise 

 methods of investigation can be devised and 

 adopted, experimenters ought to supplement their 

 statistics with a full verbal description of the pro- 

 gress and conditions of the experiment, or perhaps, 

 what would be better, supplement their descriptioa 

 with statistics, making the figures secondary in impor- 

 tance. With present methods the truth about the 

 effects of manures on trees can be more usefully 

 expressed in words than in numbers. 



The London correspondent of the Nm'th Queendand' 

 .BcjrisJer states that the area for Queensland is 429,120,000' 

 acres, and that the combined areas of Germany and Austria- 

 Hungary could be placed in Queensland and leave a margin. 

 Land is being taken up every year in thousands of acres, and 

 Queensland must be regarded as one of the most flourish- 

 ing continental tropical countries in the world. 



