C03IPAPJ:S0^■ of analysis with POLAEIZATIOy, ETC. 



483. 



TESTS OF ACCURACY. 



Tests of Vie Accuracy of iJxe Chemical Work. 



In this investigation of the author, at the Department of Agri- 

 culture, there was taken every reasonable precaution to secure accu- 

 rate results in the analytical work. 



The several methods of checks and control have been such, that, 

 although errors may exist in individual analyses, it is impossible that 

 the general results recorded should be other than a very close approxi- 

 mati(m to the truth. 



It will be seen in another place, that the average results in 1882 of 

 the analyses of 855 sorghum juices gave 10.998 per cent of sucrose by 

 precipitation, while the same juices gave an average of 10.9G6 per 

 cent by the polariscope; or. as 100 : 100.255. practically identical 

 results. 



During the season of 1882 there were made, in all, analyses of 24 juices 

 in duplicate ; and, as has been said, none of those engaged upon these 

 analyses had any knowledge of the fact that duplicate juices were 

 present in the twenty or more samples under analysis daily. 



The average results of these 24 analyses, in duplicate, are as 

 follows : 



