2 Journal of Agricultural Research voi. xvni. No. i 



latter have been compiled in a Station bulletin; ^ the former are set forth 

 in the present paper. 



II.— METHOD OF INVESTIGATION 



Variety tests. — Several varieties of cane grown on University Farm, 

 St. Paul, were tested for several consecutive years to ascertain their be- 

 havior in this region.^ Analytical studies were made on but three varie- 

 ties, however — Minnesota Early Amber, an old, very well-established, and 

 almost universally grown variety not only in Minnesota but in the whole 

 country; Early Rose, a variety isolated from Early Amber in southern 

 Minnesota about 1 5 years ago ; and Dakota Amber, another selection from 

 Early Amber that was made in South Dakota. The data in this paper 

 are the averaged results from these three varieties. 



Samples. — The data on the progressive changes in composition of the 

 cane cover a period of three growing seasons. Not all data, however, 

 were obtained for all the years. The stages of growth at which samples 

 were taken were as follows : 



(i) When the panicles first appeared. 



(2) When the panicles were wholly emerged. 



(3) When the anthers of the blossoms appeared at the middle of the 

 panicles. 



(4) When the panicles were in full bloom. 



(5) When the seed was in milk. 



(6) When the seed was in soft dough. 



(7) When the seed was in hard dough. 



(8) When the seed was brittle and mature. 



During these stages the plants increased in height by 2 or 3 feet. 



In taking samples, 6 to 10 plants were selected which were of nearly 

 the same stage of growth and which at the same time were as representa- 

 tive as possible of the plot. Two samples were cut from each plot at 

 each stage of growth. One was weighed, sacked, and dried; the other 

 was weighed, stripped, topped, and the juice extracted by pressing with 

 a small 3-roll power mill. 



Analysis. — ^The leaves, tops, cane, juice, and bagasse were weighed 

 separately and the weights recorded, together with the losses in weight 

 during stripping and pressing. The leaves, tops, and bagasse were sacked 

 separately and, together with the sample of the whole plant, dried as 

 rapidly as possible in a steam oven to less than air-dry moisture content. 

 The samples were then exposed to the air of a well-ventilated room until 

 there was no further increase in weight due to absorption of moisture. 

 The finely ground air-dried material was resampled and the usual proxi- 

 mate determinations were made, together with an analysis of the ash for 



' WiLLAMAN, J. J., West, R. M., and Bull, C. P. sorghum and sorghum-sirup manufacture. 

 Minn. Agr. Exp. Sta. Bui. 187. 

 ' Acknowledgments are due to Prof. C. P. Bull, of this Station, for the agronomic phases of this work. 



