199 
Still another reason for supposing a functional difference be- 
tween bordered and simple pores is founded on the fact, that when- 
ever the libriform cells are used as reservoirs of starch, which is 
often the case, it has hitherto been found only in those with simple 
pores. In all the woods examined in the work described in the 
following pages only two exceptions to this statement were found, 
and these did not materially weaken the argument used. 
An attempt has been made, as here described, to examine 
the wood of the principal dicotyledonous families, in reference 
to this characteristic of the libriform tissue and to show; first, 
to what extent it is possible for it to aid in the process of water- 
carrying, as shown by the presence or absence of these borders ; 
second, how far this anatomical structure coincides with the mor- 
phological characteristics which are made the basis of the present 
system of classification and serve to determine the limits of genera, 
families, groups, etc. 
It will be necessary at the outset to explain somewhat in de- 
tail what principle or definition has been followed in deciding to 
which class a pore belongs; for though in general, it is very easy 
to distinguish a bordered from a simple pore, many cases occur 
where the necessity for a clear and fixed definition is seen. Inall 
thick-walled cells, where simple pores are found, there is, or may 
be, a tendency to a slight funnel-formed enlargement of the canal 
at the extremity where the two pores of the adjacent 
cells meet. The question is, what must be the form of 
this enlargement to give it the name border? As a 
test of this, examples were taken from the pores of bast 
and sclerenchyma cells, where there is no reason for 
supposing any adaptation to water transport. In 
many of these pores the canal widens sufficiently to a 
present the appearance of a very small border. If we 
take, however, the so-called profile view, made either 
by a cross or longitudinal section through the border, 
and compare it with the same view of that of the pores 
of Coniferze (see figures a and 0) it will be seen that 
in the former the wall of the canal passes by a curve 
into this opening, while in the latter, it makes always 
a distinct angle. This was taken for a test, and in all 
