The Mechanism of Root Pressure 45 
permeability of the protoplasts releasing solutes, when the pressure 
exerted on their surfaces through the xylem sap is diminished. This 
argument might have a direct bearing on the relatively rapid way 
in which the flow of sap ceases with many plants when the stems are 
cut off, although it is probably maintained in the normal plant for 
long periods and may become manifest at any time throughout the 
growing season. Unfortunately, an analysis of Chamberlain’s data 
does not permit of any confident decision being reached as between 
these two alternative factors, both capable of accounting for an 
increased flow of sap under an increased pressure and a decreased 
flow with a diminished pressure. 
3. The Structure of the Endodermis 
The significance of the structural features of the endodermis 
will be fully considered in a later communication. As, however, the 
statements made in the original paper may evidently be seriously 
misapplied unless certain well-established facts are immediately 
pointed out, a later discussion is anticipated to this slight extent. 
Kroemer( 7 ) clearly defined the different structural stages through 
which the cells of the endodermis may pass. Originally normal meris- 
matic cells, the endodermal cells then pass into a “primary” stage 
in which tangential walls are of cellulose and permeable to water 
and solutes, but radial and transverse walls are modified and ren¬ 
dered relatively impermeable by the presence of the Casparian strip. 
Later, all or some of the cells of the cylinder may develop a suberin 
lamella which renders them relatively very impermeable, as was 
pointed out by Schwendener(i2). Such endodermal cells with suberin 
lamella Kroemer describes as in the “secondary” stage. Later, 
again, in the Angiosperm endodermis, a “tertiary” stage develops 
in which the cells contain, within the suberin lamella, later deposited 
thickening layers of cellulose. 
In roots, in the absorbing region, the endodermal cells appear 
to be without exception in the “ primary ” stage and it is to this type 
of endodermis only that the discussion in the earlier paper applies. 
This warning seems necessary, as Bower ( 3 ) has recently applied these 
conceptions of endodermal functions in an interesting discussion of 
their influence upon the evolution of the stelar structure of Ferns. 
He describes the endodermal layer as placing the exchange of 
substances between vascular strand and ground tissue “ under proto¬ 
plasmic control.” This is true for the primary stage endodermis only, 
but in the Leptosporangiate Ferns the endodermis of root, rhizome 
