OF THE DIPTEROCARPACEE. 21 
remarks I must refer the reader.  Burck and Heim have 
attempted to establish complete diagnostic generic characters 
taken from the internal structure. Some of the conclusions 
arrived at by these authors do not commend themselves to me, 
and, where it appeared necessary, I have explained this in its 
proper place. In my opinion the internal structure of a much 
larger number of species must be studied before a satisfactory 
diagnostic table of the genera, based upon anatomical characters, 
can be established. 
It must not be forgotten that the application of anatomical 
characters to systematic botany is as yet in its infancy. We 
have still to learn which characters are essential in a particular 
genus or species. In Dryobalanops, for instance, the great length 
of the cortical leaf-traces seems to be an essential character; in 
Anisoptera their length seems to vary between one-half and an 
entire internode; Vatica, with few exceptions, has hardly any 
cortical leaf-traces, the lateral traces separating from the central 
cylinder close under the node. In Doona and Hopea they are 
short; but in Shorea there is great variety, the cortical leaf- 
traces being very short in some species, while in others they run 
through one internode and a half, I have already drawn atten- 
tion to the raised lines visible on the outside of an internode, 
which in Dryobalanops readily indicate the length of the cortical 
leaf-traces. In other genera, particularly in Shorea, I have 
often found these raised longitudinal lines deceptive. It may 
be that in these species the central cylinder bulges out before the 
lateral leaf-traces have entered the bark, thereby producing raised 
lines even where the traces have not yet separated from tne central 
cylinder. 
As regards the number of resin-ducts in the pith at the base 
of the internode, Dryobalanops, as already mentioned, occupies 
an exceptional position ; and the same holds good for a number 
of Hopeas of the section Dryobalanoides. In Shorea, as justly 
observed by Burek and Heim, there is great variation (in S. ro- 
busta from 3 to 20). In other genera the number seems to 
fluctuate within narrow limits. So far as known at present, in 
Balanocarpus and Vatica the number varies from 10 to 20, in 
Anisoptera from 18 to 24. 
In several genera the ducts are unequal in width, the larger 
ducts having 10 to 15 times the diameter of average cells of the 
pith. In Vatica they are generally small, only 1-3 times the 
