42 > COMPARATIVE EMBRYOLOGY OF THE RUBIACEAE 
In this form, on account of the large size attained by the haustoria 
it is quite easy to see that, as they penetrate the endosperm, the 
process is accompanied by mechanical readjustment of the cells of 
the latter, without any destruction. 2%. 8, fig. 5, shows a later 
stage in the same or a closely related species, from which it will 
be seen that as the embryo grows older the haustoria becomes 
filled with cytoplasm. Meanwhile their proximal extremities be- 
come compressed and drawn out as a result of the pressure exerted 
on them by the growing endosperm. The suspensor, exclusive of 
the haustoria, then presents a fibrous appearance, with little cyto- 
plasm, and is little more than an irregular tube joining the bases . 
of the haustoria and the embryo. The haustoria in some instances 
are very narrow, and extend to a considerable distance, passing 
beyond the embryo proper. Such narrow ones remind one of 
fungal hyphae, which pursue an intercellular growth, or of the 
pollen tube with its analogous behavior. 
In the other species, for example A. azurea and A. galioides, 
the suspensor develops haustoria which, instead of penetrating to 
a distance into the endosperm, confine themselves to rather narrow 
limits, while the immediately surrounding endosperm suffers his- 
tolysis. All the preparations of plants in which this behavior 
takes place show a perfectly normal structure in the haustoria, in 
which the cytoplasm is very dense and finely granular as in A. 
galioides, while in the material in which they are embedded no 
trace of structure is visible, nothing, in short, but the mucilaginous 
irregularly staining products of histolysis. 
As the embryo becomes older the haustoria become separated 
from each other but still keep a normal appearance. The study 
of a complete series, of which p/. 8, fig. 6, represents a component 
of three sections, shows no union between several of the haustoria, . 
although this fact cannot well be shown in a drawing. 
The behavior of the suspensor in the various species of Asper- 
ula is therefore so different, as well as the relation between it and 
the endosperm, that we are prevented from interpreting the facts 
in the same way. In some cases, as described, the relation in- ， 
volves histolysis of the endosperm, which indicates that, in such 
cases, the suspensor secretes an enzyme. 
The earlier cell divisions of the embryo proper are as described 
