RAFFLESIA MANILLANA. £21 
root of the host and of the bud, to break through the bark is prob- 
ably due, in part, to the turgor produced by this high osmotic 
pressure. The enlargement of the tissues of the host around 
the bud is, however, not a stretching caused by an increase in the 
size of the parasite, but an active growth which would seem to 
be due to a stimulus connected with its presence. 
After the bud breaks through the bark of the host, that part 
of the bark which surrounded the flower bud becomes recurved 
and gradually disappears (fig. 7). The stalk, however, conti- 
nues to be surrounded by the tissues of the host, which project 
from the side of the root as a vase-shaped mass. The base of 
this mass is composed of both wood and bark, while the upper 
part consists entirely of bark (fig. 3). The wood and bark are 
separated by the cambium layer. The base of the stalk of the 
parasite is thus in contact with xylem of the host, and the upper 
portion with the phloem. 
After the parasite breaks through the bark all parts of the 
shoot continue to enlarge. This is accomplished by a prolifera- 
tion of cells from the cambium of the host and also in that part 
of the bark which surrounds the shoot. These processes are 
similar to those described in the earlier stages. When the 
flower is mature the stalk is still surrounded by the tissues of 
the host. There is, however, a tendency for the tissues of the 
parasite and host to become separated in places. This is due 
to phellogen-like layers in both host and parasite. In the late 
stages of the growth of the shoot there is a considerable develop- 
ment of a phellogen-like layer in the inner layers of the bark 
of the host next to the parasite. This layer has an appearance 
which is almost identical with the one frequently produced 
around young shoots, and the development of the two is prob- 
ably due to similar stimuli. When this phellogen-like layer is 
formed in the bark of the host another is sometimes developed 
opposite it in the outer layers of the parasite. The extent of 
these layers varies in different specimens. In some they are 
produced over the greater part of the region around the parasite, 
while in others they extend over only a small portion. In none 
of the old stages examined did they occupy the w^hole region and 
they were never entirely absent. Whenever these two layers 
were produced opposite each other and were developed to any 
considerable extent there was a tendency for the host and par- 
asite to become separated. 
When the shoot of Rafflesia manillana first breaks through 
the bark of the host only a few of the bracts which surround 
