THE MECHANISM OF CURVATURE 241 



to its relative position, as is especially well shown when a stimulated node 

 of grass is cut into a series of parallel horizontal slices, and the growth of 

 each followed. In this case the splitting releases no disturbing tissue- 

 strains l , but even when these come into play positive results may be 

 obtained. Thus Sachs 2 found that when a root, split into two equal 

 longitudinal halves which remained in contact, performed a positive geo- 

 tropic curvature, the upper half elongated more than the lower. When an 

 erect stem is split, the two halves curve apart owing to the released tissue- 

 strains, and as the result of their changed position each performs a nega- 

 tively geotropic curvature during which the under side of each half grows 

 more rapidly than the upper inner one 3 . Similar experiments have been 

 performed by Hofmeister 4 with the stalks of Agaricineae and by Copeland 5 

 with the stems of seedlings. The latter found that a horizontally placed 

 segment in which a negatively geotropic curvature was produced grew 

 more rapidly than a vertical one when the cut surface was upwards, and 

 less rapidly when it faced downwards. Further investigation appears, 

 however, to be needed in this direction. 



In any case longitudinal halves of stem and roots are capable of 

 geotropic curvature when placed horizontally, and the curvatures always 

 take place in the same direction independently of which side is placed 

 downwards, so that the curvature may either take place towards or away 

 from the cut surface. It follows, therefore, that in the intact organ as in 

 unicellular ones, correlative relationships determine the relative rate of 

 growth of the different parts, and these must even influence the growth of 

 the collenchyma strands in the nodes of grasses, since the tensions brought 

 into play are incapable of directly stretching them 6 . 



The conditions are naturally rendered more complicated by the fact 

 that the cells in a tissue are not all equally active and responsive, and 

 that inactive elements may be present which, when comparatively rigid, 

 may partially arrest or completely prevent an attempted curvature. Even 

 a realized curvature may involve the compression of cells which strive to 

 expand, as well as the regulation of the growth of some and the plastic or 

 elastic stretching of others. Actions of this character, although they may 

 influence curvature, do not induce it. Kohl 7 supposed that geotropic 

 curvature was due to an active contraction of the tissues on the concave 

 side, but Rothert and Noll 8 have shown the incorrectness of this supposition, 



1 H. Miiller-Thurgau, Flora, 1876, pp. 69, 92. 



2 De Vries, Landw. Jahrb., 1880, Bd. IX, p. 483 ; Pfeffer, Druck- und Arbeitsleistungen, 1893, 

 pp. 394, 408 ; Sachs, Arb. d. hot. Inst. in Wiirzburg, 1873, Bd. I, p. 470. 



3 Sachs, Flora, 1873, p. 330, 



4 Hofmeister, Jahrb. f. wiss. Bot., 1863, Bd. ill, p. 93. 



3 Copeland, Botanical Gazette, 1900, Vol. XXIX, p. 189. 6 Pfeffer, 1. c , pp. 401, 426. 



7 Kohl, Mechanik der Reizkriimmungen, 1894, pp. 4, 40, 87. 



8 Rothert, Biol. Centralbl., 1895, Bd. 15, p. 593 ; Noll, Flora, 1895, Ergzbd., p. 44. 



PFEFFER. Ill R 



