TBANSACTIONS OP SECTION D. 671 



Elfving concludes that the visible changes are the result and not the cause of the 

 curvature. Elfving also produced curvature in Phaseolus by bending the apex of 

 the plant towards its base and tying in that position. Under these conditions the 

 convex side of the shoot showed the changes described by Wortmann in geotropic 

 plants. Here again Elfving gives reason to believe that the thickening of the 

 cell-walls is a result, not of curvature, but of strain mechanically produced. 

 When a plant is prevented from executing an apogeotropic movement it is clear 

 that a longitudinal strain is put on the upper (concave) side. But the longitudinal 

 strain in Elfving's plants is on the convex side. Therefore, if, as Elfving believes, 

 the visible changes are due to strain, they should, as they do, occur on the convex 

 side in his experiments, on the concave in Wortmann's. 



AVortmann replied in the ' Bot. Zeitung,' 1888, p. 469, and attempted to explain 

 how Elfving's results might be explained and yet his own theory hold good. The 

 reply is by no means so strong as the criticism, and it must be allowed that 

 Elfving has seriously shaken Wortmann's argument. 



Somewhat similar criticisms have been made by Noll.^ In the acellular plants, 

 Derbesia and Bryopsis, Noll studied growth-curvatures, and was quite unable to 

 detect any thickening of the concave cell-walls, except when the curvatures were 

 very sudden, and in these cases the result could equally well be produced by 

 mechanical bending. 



Noll further points out what is undoubtedly a fault in Wortmann's theory, 

 namely, that he explains the retardation on the concave rather than acceleration 

 on the convex side. This criticism is only partially just, for though Wortmann's 

 description only shows a relative thinness of the walls on the convex side, yet it ia 

 clear he believed there to be an absolute diminution of resisting power on that 

 side. 



Noll's experiments with grass-halms show clearly that acceleration of growth 

 on the convex side is the primary change, rather than retardation along the concave 

 lialf. Wlien the halms are fixed in horizontal glass tubes, so that they are stimu- 

 lated but unable to bend, the lower half of the pulvinus forms an irregular out- 

 growth, increasing radially since it is not able to increase longitudinally. 



A similar argument may be drawn from Elfving's experiments. He found that 

 the pulvini of grass-halms placed on the klinostat increase in length. This experi- 

 ment shows incidentally that the klinostat does not remove but merely distribute 

 equally the geotropic stimulus : also that geotropic stimulus leads to increased, not 

 to diminished growth. The same thing is proved by the simple fact that a grass 

 halm shows no growth in its pulvinus while it is vertical, so that when curvature 

 begins (on its being placed horizontal) it must be due to acceleration on the convex, 

 since there is no growth on the concave side in which retardation could occur. 

 Noll's view is that the primarj' change is an increase in extensibility of the tissues 

 on the convex side. This view he proceeded to test experimentally. A growing 

 shoot was fixed in a vertical position, and a certain bending force was applied to 

 make it curve out of the vertical, first to the right and then to the left. If the 

 cortical tissues are, at the beginning of the experiment, equally resisting all round, 

 it is clear that the excursions from the vertical to the right and left will be equal. 

 As a matter of fact the excursions to the right and left were nearly the same, and 

 the difference was applied as a correction to the subsequent result. The shoot was 

 then placed horizontally until geotropic or other curvature was just beginning, 

 when the above bending experiment was repeated. It was then found that when 

 it was bent so that the lower side was made convex, the excursion was greater 

 than it had been. In the few experiments given by Noll the excursion in the 

 opposite direction (stretching of the concave side) was less than it had been, and 

 he states that all the other experiments showed a similar result. The increased 

 extensibility of the convex side is clearly the most striking part of the phenomenon, 

 but I fail to see why Noll takes so little notice of the diminution in the extensi- 

 bility of the concave side, which is only mentioned towards the end of his paper.* 

 Yet such a diminution is a necessary factor in the mechanism of curvature. It 

 «hould be noted that results like Noll's might be obtained under other conditions 



' Sichs' Arbeiten, 1888, p. 496. - Loc. cit. p. 529. 



