504 PHYSIOLOGY OF GROWTH. 



garden soil contained in flower-pots. We place the pots in front 

 of a window exposed to very bright light. The epicotyls of the 

 young seedlings (we must only grow a few in each pot, so as to 

 avoid their shading each other) at first turn towards the light ; 

 they exhibit positive heliotropism. If the plants are allowed to 

 remain undisturbed in front of the window, exposed to the bright 

 light, the positive heliotropism passes over into negative helio- 

 tropism. They curve away from the source of light, as also do 

 the new shoots developing later. Their illuminated side (upper 

 side) hence becomes convex. The intense light, therefore, has 

 induced plagiotropism in the Tropaeolum shoots, but it is not deep- 

 seated, since the stems of our plants, both with respect to their 

 anatomical structure, and also as regards the position of the leaves 

 arising on them, always present a multilateral or radial character. 

 It is important that the illuminated or upper side of the epicotyl 

 of Tropaeolum does not gradually become convex when the objects 

 are exposed to weak light; the negative heliotropism of the 

 structures is not then exhibited ; they now curve only in a 

 positively heliotropic manner, towards the light. Moreover, any 

 flank we please of the Tropseolum stem (that, viz., illuminated 

 most strongly) may become the upper side. 1 



A striking case of local induction of dorsiventrality may be 

 proved by experiments with the horizontally growing shoots of 

 Thuja occidentalis. 2 On these shoots four rows of leaves are 

 present : one above and one below (facial leaves) ; one on each 

 flank (marginal leaves). Shoots of Thuja, grown under normal 

 conditions, are clearly dorsiventral in construction, as we can 

 readily satisfy ourselves by examination of delicate transverse 

 sections. Their mesophyll, e.g., on the upper surface is composed 

 of palisade parenchyma ; it consists on the lower side of approxi- 

 mately isodiametric cells (see Frank, Pringsheim's Jahrbiicher, Bd. 

 9, Table 16, Fig. 4). If, now, in the early spring, while vegetation 

 is still in a state of dormancy (the experiments which I made 

 began at the commencement of March), horizontal shoots of Thuja, 

 without being removed from the parent plant, are turned upside 

 down and fixed in that position, their morphological lower side 

 being thus directed upwards, it will be found that the shoot ends 

 develop quite normally during the spring, and also, as usual, 

 become dorsiventral in structure. That side of the shoot, how- 

 ever, which, without reversal, would have been morphologically 

 he under side, now becomes the upper, as is clearly indicated, 



