THE LNSER TISSUES OF PLANTS. 263 



elongated. It does this for two reasons. That narrowing oi 

 the cells and intercellular spaces which accompanies their 

 elongation, facilitates capillarity ; and at the same time fewer 

 of the septa formed by the joined ends of the cells have to be 

 passed through in a given distance. Hence the 



general fact that the establishment of a rudimentary vascular 

 system, is the formation of bundles of cells lengthened in the 

 direction which the liquid is to take. This we see very 

 obviously among the lower Acrogens. In one of the lichen- 

 like Liverworts, the veins which, branching through its 

 frond, serve as communications with its scattered rootlets, are 

 formed of cells longer than those composing the general tissue 

 of the frond : the lengths of these cells corresponding in their 

 directions with the lengths of the veins. So, too, is it 

 with the midribs of such fronds as assume more definite 

 shapes ; and so, too, is it with the creeping stems which 

 unite many such fronds. That is to sa}^ the current which 

 sets towards the growing part from the part which supplies' 

 the materials for growth, sets through a portion of the tissues 

 composed of units that are longer in the line of the current 

 tha,n at right angles to that line. The like is true 



of Phsenogams. Omitting all other characteristics of those 

 parts of them through which chiefly the currents of sap 

 flow, we find the uniform fact to be that they consist of cells 

 and intercellular spaces distinguished from others by their 

 lengths. It is thus with veins, and midribs, and petioles; 

 and if we wish proof that it is thus with stems, we have but 

 to observe the course taken by a coloured solution into which 

 a stem is inserted. 



What is the original cause of this differentiation ? Is it 

 possible that this modification of cell-structure which favours 

 the transfer of liquid towards each place of demand, is itself 

 caused by the current which the demand sets up ? Does the 

 stream make its own channel ? There are various reasons 

 for thinking that it does. In the first place, the simplest and 

 earliest channels, such as we see in the Liverworts, do not 

 Vol. II. 12 



