THE UREDINIUM IN PUCCINIASTRUM AGRIMONEAE 57 
the chain. The peridial cells at this stage are much longer in a vertical 
direction than they are in the mature sorus, and their side walls, 
instead of being oblique, are perpendicular (fig. i). The later change 
in shape is probably due to the pressure of the spores as they are 
produced. Such a pressure coupled with the decreasing cell turgor 
would tend to collapse the cells somewhat and to press their bases 
laterally. 
Sections at a later stage show that as the sorus develops the 
epidermis of the host is hfted from the mesophyll; and this separation 
of the surface layer from the underlying tissues often extends for some 
distance beyond the very definite lenticular sorus. The mesophyll 
tissues are apparently not at all or only slightly injured. Often the 
cells directly under the center of a mature sorus retain their shape 
perfectly. Only a little mycelium can be seen under the sorus and 
none beside or over it; in fact the amount of vegetative mycelium 
visible at any stage is very small. The mature sorus (fig. 3), as may 
be inferred from what has just been said, is a very clearly differentiated 
structure which is sharply set off from the host tissue and from the 
vegetative mycelium. It is bounded above and at the sides by a 
peridium of somewhat overlapping, thin-walled cells in which the 
overlapping above the center is in a manner opposite to that of the 
shingles on a roof. At the side of the sorus, the direction of the over- 
lapping becomes indefinite, so that from this point to the hymenial 
layer the cross walls may be oblique in either direction or some in one 
and some in another. The cells are usually so much collapsed that it 
is impossible to make out details accurately even in stained paraffin 
sections; but in spite of its fragile appearance it is evident- that the 
peridium has considerable tensile strength, for the sorus maintains 
its shape and, until it is quite aged, furnishes no means of escape for 
the spores except a central ostiole. At this point, the peridium later 
begins to disintegrate. The ostiole is bordered by a ring of cells 
(fig. 2) which are larger and thicker walled than those of the rest of 
the peridium. 
The base of the sorus is characterized by a plate-like hymenial 
layer of hyphal cells which at its margin often appears almost or quite 
continuous with the peridium and which separates the sorus from the 
underlying structures with a definiteness unusual in rust sori. The 
spore chains rise from a layer of basal cells just above the hyphal 
plate layer. In the young sori, where the chains are clearly visible, 
