334 BOTANICAL GAZETTE ' [November 



decreased free-reducing substances, sucrose, polysaccharides, and 

 total dry matter." 



In the work with Amaranthus plants I have found a similar . 

 situation so far as the relation between nitrogen and carbohydrate 

 is concerned; that is, low nitrogen is accompanied by high carbo- 

 hydrate and high nitrogen by low carbohydrate. Upon computing 

 the reciprocal condition in the different fractions I find that the 

 product of carbohydrate by nitrogen is not a mathematical con- 

 stant, but that it varies considerably, sometimes decreasing as the 

 development of the plant progresses. The product varies least in 

 the stem and roots. 



Let Ci, C2, and C3 be the carbohydrates and Ni, N2, and N3 

 denote the nitrogen, the sub-numbers representing the time of 

 collection. If the carbohydrate and nitrogen hold a reciprocal 



relation, then 7^ = aF ^ p" = r^ » ^^^ (^ ^ f^ ' ^ clearmg the frac- 

 tions, CiXNi = C2XN2 = C3XN3, etc., or carbohydrate X nitrogen = 

 constant K. Applying this principle, the following constants are 



obtained. 



Insoluble fraction (F3) 



June 3 June 20 July 8 Av. K. 



Roots 15.7 13 06 10.92 13-23 



Stems 16.20 10.62 9.20 12.01 



Leaves 28.00 20.4 27.6 25.30 



Soluble fractions (F1+F2) 



Roots 10.10 11.00 6.27 9.12 



Stems 9 . 45 8.97 7 . 88 8 . 43 



Leaves 2.39 2.97 3.15 2.84 



These data show that the carbohydrate-nitrogen ratio is not a 

 constant as we think of a constant in mathematics or physics. In 

 plants where great fluctuation occurs in their substratum through- 

 out different parts of the day and different times in the season, this 

 disparity is no positive evidence that such a ratio does not exist. 

 Secondly, regardless of the exactness of the ratio, this much is true, 

 when the carbohydrates are high the nitrogen compounds are 

 relatively low, and vice versa. Figs. 9-1 1 show this reciprocal 



