ANALYSIS OF SORiiHOI A>D MAIZE — METHOD. 471 



which each assistant was, for the time, devoted exclusirelT to one thing, 

 e. g., determinations of density bv the balance, volumetric determin- 

 ations of glucose and sucrose, polarizations, ash determinations, total 

 solids, ash analyses, analyses of the seed, quantitative determinations 

 of acids and other proximate constituents of the juicies at seventeen 

 different stages of growth of the plant and after maturity. By this 

 svstem each co-worker became thoronghly expert as "a specialist in his 

 own dutv; and it was thus possible, by this system, to test the accuracy 

 of the work by submitting identical samples in duplicate and separate 

 numbers for analysis by the same and by different co-workers — a crucial 

 test of verification. 



The committee have critically examined the work done in this way, 

 and the details show a suprising agreement. 



Method cf Analysis. 



It is obviously of the first importance that the results of analyses 

 given should have been obtaine<i by reliable methods. 



Everv precaution has been taken to guard against error and to con- 

 tnd the results. 



In the first place, it may be remembered that each assistant, in the 

 routine work assigned him. was necessarily tree from all prejudice as 

 to what residt he was to expect, for each sample of juice, syrup, or 

 cane exanimed was kn-wn only by a number, and this was known 

 only to one who himself |>erformed no analytical work. 



Everv questionable result was at once repeated, and many dupli- 

 cate samples of juice, under different numbers, and without the know- 

 ledge of any of those engaged in the analyses, were from time to time 

 analyzed. 



E;ich new lot of either of the re-agents employed in analysis was 

 carefully tested, and indeed nothing was omitted which woidd tend to 

 accuracy in work. 



Those familiar with chemical methtxls, and considering the vast 

 amount of work actually performed in these analyses, are aware that 

 absolute accumcy is not to be expected: but whatever errors there may 

 be are certainly within very narrow limits, and the general results fur- 

 nished in the foregoing analyses may be confidently relied upon as 

 being practically near approximations to the truth. 



The Analytical Processes for the Examination <^'the Canes. 

 One or more stalks of the variety of sorghum to be examined were 

 selected in the experimental field, and, after recording the stage of de- 

 velopment and general appearance of the canes, a number was affixed 



