28 
A. G. Tansley. 
except by clear evidence of diversity of origin through distinct series 
of transitional types—evidence which we are very far from 
possessing. 
Meanwhile we have no direct knowledge whatever of any 
plants which suggest “ Protopteridophyta,” and in trying to picture 
to ourselves the ancestors of the vascular plants we are forced to 
take refuge in speculation based upon various considerations. 
One line of speculation has led to the well-known theory of the 
origin of the Pteridophytic sporophyte from the Bryophytic 
sporogonium by elaboration and sterilisation. 1 do not propose to 
attempt to criticise this theory, even briefly, but shall merely call 
attention to the obvious fact that it involves some tremendous 
morphological assumptions in the way of the origin of new organs, 
and that even granting these assumptions and granting that we 
might obtain by their aid something which might perhaps pass for 
a primitive Lycopod, the characteristic features of the great 
Filicinean phylum, particularly its megaphylly, which is almost 
certainly primitive so far as any forms we know are concerned, are 
in no way suggested. 
An alternative, and as it seems to me a more profitable method 
of proceeding, is to see if we cannot get some idea as to what the 
Protopteridophytes were like by a consideration of the characters 
common to the apparently more primitive members of existing and 
fossil Pteridophytic groups. 
Of these characters the “ dichotomous ” branching, both of leaf 
and stem, is very conspicuous, though not universal. It is well at 
the outset to be clear as to exactly what we mean by “ dichotomy.” 
Technically complete dichotomy only exists when the apical cell or 
the apical cell-group of an organ divides into two in a plane 
coinciding with the long axis of the organ, the two halves becoming 
the apical cells or cell-groups of two new axes, which keep pace 
with one another in growth so that they diverge from the line of the 
parent axis by equal angles. This kind of branching appears, 
however, to be extremely rare. A commoner type, often found for 
instance in the liverworts, is for the apical cell of the parent axis to 
cease dividing, two new apicals being established right and left of 
the old growing point. Strictly this is cymose monopodial branching. 
Another common type is the formation of a new apical cell from 
one of the segments of the original apical, or from one of the 
products of division of such a segment, but in any case so early 
that the branch produced by the new apical is not to be distinguished 
