424 
Rushton . — The Development of 
it is too small to demonstrate. Frequently this dot-like structure is com- 
pletely enveloped in protoplasm (Fig. 3, A and b), or has small masses of 
protoplasm attached to it laterally. The thread-like bar continues through 
the cambium radially and often widens out again in the phloem region. 
In all these tangential sections the bar rarely occurs near the transverse 
end walls of a tracheide, but most frequently about half-way between the 
end walls. From the evidence obtained by the presence of protoplasmic 
masses around the bars in the cambium region, it would appear that once 
started in P. Inops the bars continue their existence by means of these 
protoplasmic masses, but exactly what starts them could not be demon- 
strated. That they are formed by partial reabsorption of the end walls of 
tracheides, as suggested by Raatz and Muller, does not seem to hold for 
this species, as the end walls of the tracheides where they are flattened 
frequently have bars present a short distance above or below this end wall. 
The view put forth that these bars are formed by tangential walls of the 
cambium cells bending in, until they come in contact, did not appear to be 
the case in P. Inops , as throughout the whole radial width of the cambial 
cells they did not vary in diameter, which would be the case if a folding of 
the wall took place, as they would appear smallest in diameter in the 
middle of the cell and gradually increasing towards the outer or inner 
tangential walls. The following would appear to be the probable explana- 
tion of the observations recorded in this paper : 
In its first stage as yet revealed a ‘ Sanio’s bar ’ is a very delicate 
cylindrical rod stretching radially and freely across a cambium cell in 
a transverse plane. As the cambium cell grows in a radial direction, the 
rod keeps pace with it by stretching along its main axis. The cambium 
cell eventually cuts off a daughter-cell, which in turn grows radially and 
directly or indirectly, after a further division, gives rise to a tracheide (or 
sieve tube), with whose radial increase of width the bar keeps pace. At the 
same time surface growth of the tangential walls of the growing daughter- 
cell or cells takes place in longitudinal transverse and intermediate directions. 
The bar, firmly fixed into the primary walls, is at its ends exposed to 
tension at right angles to its long axis in all directions, and the natural 
result is a rupture in the centre of the rod, which now becomes hollow. At 
first it would appear that there is no disparity in rate of growth in surface 
of the tangential walls in the longitudinal and transverse directions, so that 
the hollow in the centre of the axis itself is cylindrical in form ; but when, 
as is the case later, growth in length of the cell greatly predominates, the 
rod is exposed at its ends mainly to tensions at right angles to its main 
axis in the radial longitudinal plane of the cambium cell, with the con- 
sequence that the bar and its central cavity are stretched in a -corresponding 
manner, and the appearance in cross-section is that shown in Fig. 3, F. The 
bar is thus from a very early stage a solid elastic body, since the tension in 
