SECRETION IN THE HYPERICACER. 457 
when the section shows twenty or more cells, another process be- 
comes evident. Not only does the strand contain interspaces, 
which, as the passage still grows in diameter, extend throughout 
it irregularly (fig. 5), but the cells themselves break up and dis- 
appear. Just before this dissolution their contents have become 
more coarsely granular, and small aggregations of the resinous 
secretion may be recognized in them. The disintegration of the 
tissue begins where the intercellular spaces have first made their 
appearance, and generally the centre of the strand is the first 
part to go. Cavities thus are formed in the strand of cells, and 
these contain the secretion which was in the cells that have 
broken up. These cavities are not regular in their distribution, 
two or three sometimes appearing in the cross section (fig. 6). 
In some cases the appearance is that of a central cavity crossed 
by rows of cells of irregular form and unequal sizes. Generally 
at this stage the peripheral cells are intact, but sometimes one 
or more of these may be missing (fig. 6). The process of dis- 
integration proceeds further, advancing, speaking generally, 
from the centre to the periphery. The next stage is shown at 
fig. 7, where all the central cells have disappeared, and the 
passage appears as if lined by an epithelium of secreting-cells. 
In the resin-passages of the stem this layer appears to persist 
as long as the passages remain visible, the walls of the individual 
cell getting, however, more delicate and their outlines less 
distinet as their age advances, while the cells become more and 
more flattened. In the passages in the ovary the disintegration 
goes further, for sections through ripe fruits show some of the 
passages lined apparently only with a little débris, resulting from 
the breaking-up of this quasi-epithelium (Pl. XL. fig.20). Atthe 
stage shown in fig. 7, the resin-passage in the stem is at its 
highest point of development. Meanwhile the cells of the funda- 
mental tissue are being changed from their original condition, 
their walls thickening considerably. Instead of the passage 
continuing now to press upon these cells, the direction of the 
tension is reversed, the now yielding tube becomes smaller and 
smaller in diameter (fig. 8), till gradually it ceases to be recog- 
nizable, and, at the time when the fibrovascular tissue is well 
differentiated, no resin-passages can be identified. The formation 
of the secreting tissue hence is confined to the primary meristem, 
and the cambium does not give rise to any. 
The origin of the passages in the ovary is similar to that in the 
