64 Journal of Agricultural Research voi. xviii. No. 3 



We have explained heretofore our reasons for omitting a study of the 

 roots of plants growing in natural soils. For the same reasons we have 

 little confidence in the exactness of the findings of other investigators 

 from similar experiments. It is interesting to note, however, that 

 computations from the well-known work of Wilfarth, Romer, and Wim- 

 mer (9) show that of the elements in question not more than 10 per cent 

 of the total nitrogen absorbed was contained in the roots at any stage 

 of growth, and only 3 per cent of the total potassium. Data from the 

 same authors, including the roots and stubble wit^ stems to a height of 

 5 cm. above the surface of the soil, show a maximum of 22 per cent of 

 the total nitrogen and 1 8 per cent of the total potassium at what appears 

 to be about the same stage of development as our plants when the 

 observed losses took place. Data obtained in this laboratory (j) from 

 the same strain of barley grown in sand cultures at a little later stage 

 of growth showed the roots to contain 9.6 per cent of the total nitrogen 

 and 7.3 per cent of the total potassium content of the plant. In our 

 experiment in 191 7 the proportions of the total nitrogen and potassium 

 lost from the upper part of the plant, including all of the stems and 

 most of the stubble (see method of cutting described on page 53), 

 were 38 and 34 per cent, respectively. 



The magnitude of these losses as compared with the amounts of the 

 elements found in the roots in the cases cited seems sufficiently great 

 to justify the opinion that there is an actual movement of potassium 

 and nitrogen from the plant into the soil at this stage of development. 

 The losses, it will be observed, occur at the time the heads are beginning 

 to form and to draw upon the remainder of the plant for these same 

 constituents (fig. 5-8). There is evidently a concurrent migration of 

 important constituents from the stems and leaves in two directions 

 into the heads and into or through the roots into the soil. 



The losses to which we have called attention are not to be confounded 

 with those which apparently take place in numerous plants at the extreme 

 end of the growing season. Here the losses occurred comparatively early 

 in the growth cycle of the plant and were by no means final, being fol- 

 lowed by appreciable gains in 191 6 and by very substantial increases in 

 1 91 7. However, the other kind of loss is also to be noted here, there 

 being appreciable losses of potassium, calcium, and magnesium after the 

 fifteenth week when the grain was ripening, in the 191 7 experiment, and 

 some indications of the same sort of thing in the work of the preceding 

 year. 



The experiments reported by Wilfarth, Romer, and Wimmer (9) show 

 evidence of similar losses of potassium and nitrogen and also sodium, 

 at what appears to be about the same early stage ^ of development of 



1 An exact comparison of the two studies is not possible because of differences of soil and climate. In the 

 work quoted, the intervals between cuttings were so much greater (about four weeks as against two weeks) 

 that a closer analogy may have been obscured. 



