





A FORTNIGHTLY REVIEW 



OP THE 





IMPERIAL DEPARTMENT OF AGRICULTURE FOR THE WEST INDIES. 



.Ml. 



Vol. XL No. 264. 



BARBADOS. -JUNE 8, 1912. 



Pbioe Id, 



CONTENTS. 



Page. 



Atjricultiire in British 



Guiana, 1910-11 ... lf<7 



Aaricultiiro in the Lue- 



WiUil IslAiids, 1910-Jl ISO 



Carbon Assimilatinn in 



Plants 181 



Cuttle, Heniov;iI nf, fvuni 

 St. Vincent to the 

 Grenadines 186 



Characteristics (if .-( Hybrid 



Hevoa 184 



Cotton in Ceylon 189 



Cotton Notes : — 



Coiisinn|)tion iif Egyp- 

 tian Cotton in the 

 I'niied States 182 



Cotton Culliv.itiou in 

 Turkestan 182 



Cotton-Growing in 

 J.ini.iica 182 



West Indi.an Cotton ... 182 



Page. 



Department.nl Reports ... 



Field Kx])erinients, Inter- 

 pretation of the Re- 

 sults of, III 



183 



Fungus Notes : — 

 Tlu' (iiiK'ial Treatment 

 of Root Diseases of 

 Permanent Crops ... 190 



Gleanings 188 



Insect Notes : — 



A Cirixjii Bisulphitle 



Explosion 186 



The Arrowroot Worm 180 



Lime and Magne-'ia in 

 Soils, Ett'ects of Diti'er- 

 ent Proportions ... 18t 



Lime .Juice, Keeping Quali- 

 ties of 180 



Market Reports 192 



Notes and Comments ... 184 



Plant, A Haiidsonie Flow - 



ering 185 



Plant Protection in 



Papua 185 



Students' Corner 189 



Sugar Industry: — 



Sugar in the United 

 States. 1911 179 



Wood Specimens in Nature 

 Study 191 



The Interpretation of the Results of 

 Field Experiments. 



in. 



HE Last two articles on this subject were 

 written for the purpose of indicatinoj and 

 describing the means by which the extent 



of the errors due to the methods adopted, in experi- 

 mentation, may be determined. In the first, attention 

 was given more directly to some of the ways in which 

 such errors may arise; and emphasis was laid on the 

 importance of the determination of the magnitude of 

 these errors, in order that allowance may be made for 

 them in arriving at an interpretation of the results of 

 experiment. 



The second article was devoted to a description of 

 the way in which the chief kinds of information in 

 regard to experimental error are obtained. It was 

 shown, firstly, that the value of an average calculated 

 from a series of experiments, in relation to its use in 

 obtaining a definite result of the work done in con- 

 ducting these experiments, may be determined in 

 a simple way by the construction of a Frequency Curve, 

 and directions for this were given shortly; attention was 

 also paid to the kind of information with respect to the 

 results themselves that maybe obtained from the proper- 

 ties of a curve of the kind. Consideration was next giv- 

 en to the determination of the probable error of any one 

 of the results; that is, to the ascertaining of the extent 

 to which it may vary from the average, on account of 

 the method of experimentation employed. This led 

 naturally to the presentation of the method of finding 

 the probable error of the average of all, or a certain 

 number of, the results. 



The matters dealt with then proceeded to the 

 determination of the probable error of the difference 

 between the results of two sets of experiments con- 

 ducted in exactly the same way. In passing, it may 

 be useful to remind the reader that the reason for 

 making two such sets of experiments is to provide 



