266 Holmes . — A Contribution to the Study of 
on the plant. As explained above, it is only in the final segment of the line 
that the number of leaves supplied decreases regularly in correspondence 
with equal intervals on the base line ; consequently it is to be expected 
that in passing through the remaining segments Curve B will show less 
regularity. In Figs. 2, 6, 9, and 12, an attempt is made to correlate the 
shape of Curve B with the number of leaves supplied, as represented by the 
line G. It is necessary, of course, to fix an arbitrary vertical scale for G, as 
data for the actual areas of the leaf surfaces are not available. The line 
should be examined in connexion with Fig. 1. From observations on 
plants in leaf, it is clear that the leaves on the lower part of the leader are 
larger than those on its upper part, and on the weaker branches. Thus, 
taking as an example the line G for S4 in 1918, Fig. 9, if G were to repre- 
sent leaf area instead of leaf number, the part of the line between #3 and b 6 
would be lower and that between £7 and c$ would be higher. It is to be 
noticed that the uppermost pair of laterals are attached to the node between 
internodes 6 and 7 of segment b, so that £7 and ci serve the same number 
of leaves during 1918 ; and that the horizontal piece of Curve G is reflected 
in Curve B. The flattening from $3 towards the base of the plant is evident 
in Curve B, but less well marked. In general, Curve G seems to agree best 
with the corresponding Curve B for the wood as a whole. Of course, other 
factors, such as the length of the vessels and the proportion of wide to 
narrow vessels, ought also to be taken into account, and this might alter the 
relation. Salisbury ( 4 ) has drawn attention to the close correlation so far 
observed between leaf function and the amount and constitution of the xylem 
in the petiole, and there is at any rate reason to believe in the existence 
of such a correlation extending into the balance between the annual rings 
of wood in the stem. 
General comparison. It is in particular Curve C, for the wood as 
a whole, which should be compared with Professor Farmers figures for 
specific conductivity ( 5 ). The comparison cannot be altogether satisfactory ; 
in Professor Farmer’s experiments a standard length of shoot, 15 cm., was 
used, regardless of the number of nodes ; and the specific conductivity was 
worked out on the basis of the area of the transverse section of the wood 
at the middle of this length. In the present paper, length is disregarded in 
favour of the number of internodes, that is, the number of leaves supplied ; 
and it is shown that there is a considerable variation in the value arrived at 
for specific conductivity at different parts of the same shoot, and within 
a length of 15 cm. This variation, again, applies only to values for inter- 
nodes, and neglects conditions occurring at the nodes, while it takes into 
account only the transverse aspect. Thus it is difficult, from the figures 
obtained in this inquiry, to make general averages, which will be compar- 
able with Professor Farmer’s figures. An average can be obtained for each 
shoot at the end of each year, from the values of C for every internode along 
