1897.] PUBLIC DOCUMENT — No. 33. 89 



Part I. 



We have continued to analyze, free of cost, all feed stuffs, 

 dairy products and waters sent to the station during the year. 

 Kesults have been reported as promptly as possible, together 

 with such comments as were considered necessary. There 

 have been tested 63 samples of feed stuffs, 89 samples of 

 whole milk, 11 samples of skim-milk, 9 samples of cream and 

 6 samples of butter; also 31 samples of milk and 20 samples 

 of butter for the Dairy Bureau. These results are tabulated 

 at the end of this report. 



There have also been examined 134 samples of water, of 

 which 10, or 7.5 per cent,, were pronounced excellent; 50, 

 or 37.5 per cent., fair; 39, or 29.1 per cent., suspicious; 

 and 35, or 26.1 per cent., dangerous for drinking purposes. 



In addition to the analyses above mentioned, which may 

 be regarded as control work, we have made a very large 

 number of analyses of feed stuffs, manures and milks, in con- 

 nection with various animal experiments. 



We have also spent considerable time in attempting to 

 estimate some of the various substances composing the non- 

 nitrogenous extract matter, and have compared different 

 methods for the determination of starch in different feed 

 stuffs, with a view of selecting one that will most correctly 

 ascertain the true starch, when in combination with other 

 substances of a similar nature. The results of some of 

 the work are very briefly presented under the following 

 heads : * — 



1. Remarks relative to the carbohydrates of agricultural 

 2)lants and seeds. 



2. Distribution of galactan. 



3. Tlie jphloroglucin method for the estimation of pentosans. 



* The work reporteil under these headings is of a technical character. 



