216 BROWN, 
In some cases the rows of cells of the parasite appear to be 
wedged in between the cells of the root, and in others they 
apparently replace them as by a dissolving action. In either case 
there is very little distortion of the cells of the root, which would 
indicate that the growth of the parasite through the tissues of 
the host is made possible by a dissolving action of the parasite 
rather than by pressure exerted by its cells. In the first case 
the dissolving action would be exerted on the middle lamella and 
in the second on the whole cells. The expansion of the cells of 
the parasite between those of the host must, however, be due in 
part to pressure exerted by the turgescence of the cells of the 
parasite. 
The bulk of the vegetative cells of Rafflesia makillana is small 
compared with that of the root of Cissus, and since they do not 
cause a distortion of the root or its tissues and probably take 
only enough nourishment to carry on their growth, it is likely 
that these vegetative cells, in themselves, do comparatively little 
harm to the host. As will be shown later, it is not probable that 
the flowers are nourished by these vegetative cells to any great 
extent. 
The flowers of Rafflesia manillana are not produced progress- 
ively in the direction of growth of the root of the host, but young 
buds are frequently found between the scars of old flowers. This 
would indicate that the cells in one region retain their activity 
for a considerable length of time, although some of the latter 
buds may, of course, be due to more recent infections of the 
root. 
DEVELOPMENT OF THE FLOWER. 
The flower shoots of Rafflesia manillana have their origin in 
rows of cells which usually cross the cambium of the host. When 
such a row is to give rise to a shoot, the cells in the region of the 
cambium proliferate and form a globular mass of meristematic 
cells (fig. 11) , which is embedded partly in the woody cylinder and 
partly in the bark. The first proliferation may take place 
equally in the two regions or almost entirely in either. The 
cambium is destroyed by the developing parasite. These rudi- 
ments of flower shoots have been found in roots less than 2 mm 
in diameter. The method of the origin of the flowers of Rafflesia 
manillana is quite similar to that of Brugmansia zippelii.^^ The 
formation of the flower shoot usually begins in a medullary ray, 
although it would appear that they sometimes originate between 
II 
Solms-Laubach, H. Graf zu, Bot. ZeiL 34 (1876) 449. 
