82 



Two thousand six hundred and seventy-two analyses of the juice 

 from average samples of cane were made this season. It is, of course, 

 necessary that the samples of cane taken for analysis should be so 

 taken as to lead to correct conclusions in regard to the value of the 

 lot from which they were taken. There are difficulties in taking such 

 samples. The canes in one lot may not be uniformly mature. One 

 part of the lot may have been injured by drought, by frost, by wind, or 

 by hail. One part of a lot may be on good soil, while a part may be 

 upon unsuitable soil. A sample may do injustice to a variety or it may 

 promise too much for the variety. In taking samples of cane for 

 analysis at this station the effort has been to have the samples cor- 

 rectly represent the normal canes of the variety from which they were 

 taken. When canes have been accidentally injured, as by stock or 

 frost or other cause, they have not been included in the sample. In 

 such cases the intention has been to have the sample represent the 

 normal canes. 



The object of the work is simply to compare the varieties under as 

 similar conditions as possible; obviously it would tend to defeat this 

 purpose of the comparison if accidentally injured canes were compared 

 with ordinary canes. In order to check the methods of taking samples 

 many hundred large samples were taken during the season and were 

 ground in a two-horse mill, and comparative analyses of the larger and 

 smaller samples were made. To test the method of taking samples 

 still further, from three to five large samples were taken at one time 

 from as many different places in one lot of each of the best varieties. 

 The mean of these samples from the same lots represented as nearly 

 as may be the value of the canes in those lots. The analyses of the 

 juices of single canes for seed selection also confirm the analyses of 

 average samples, low averages giving low selections and high percent- 

 ages in average samples giving high percentages in seed selections, as 

 will be seen in the analyses of each variety. 



Twenty-six thousand six hundred and thirty-five polarizations of juice 

 from single canes were made for the purpose of selecting seed from the 

 best canes. A first selection was made by rejecting juices whose 

 density was not unusually high. About 100,000 canes were thus sepa- 

 rately tested, 26,635 samples of juice being accepted for analysis. From 

 these a third selection, based on high percentage of sugar, was made, 

 and from these a final selection, based on high purity, was made, whose 

 juice had high density, high percentage of sugar, and high purity. In 

 studying these seed selections it will be observed that the purity, with 

 few exceptions, follows the percentage of sugar. Perhaps the readiest 

 way of increasing the purity of sorghum juice consists in increasing the 

 percentage of sugar. There appears to be a double gain in sugar by 

 selecting varieties and by improving varieties; a gain in percentage of 

 sugar and a gain in purity, in effect a double gain in yield of sugar. 

 Twenty-six thousand six hundred and thirty-five seed heads were se- 



