INVESTIGATION OF FOSSIL PLAJSTS. 169 



in fact, are woody fibres, some larger ones are interspersed 

 in a definite manner, in that case it belongs to some other 

 tribe of dicotyledons. It is indispensable that the arrange- 

 ment of the larger tubes should be definite, for appearances 

 of the same kind exist in most coniferous wood ; but in the 

 latter they are scattered in an indefinite manner, among the 

 smaller tubes, and are not vessels, but cylindrical cavities for 

 the collection of the resinous secretions peculiar to the fir- 

 tribe. If the walls of the longitudinal tubes of any fossil 

 specimen are found to exhibit appearances of little warts, 

 or excrescences growing from their sides, such a specimen 

 is to be referred to some coniferous or cycadeous plant, as 

 no other tribes present such a structure. If a trace of pith 

 can be discovered, that circumstance alone proves the plant 

 'to be dicotyledonous, because all other classes are destitute 

 of that central, cellular column; it being remembered, at the 

 same time, that as the roots of dicotyledonous plants are 

 destitute of pith, the absence of pith does not prove the plant 

 not to have been of that order, as the part examined may 

 have been a portion of the root. 



If a stem is in such a state that nothing can be deter- 

 mined respecting its anatomy, it will be necessary to judge 

 of it by another set of characters. In the first place, it 

 should *be inquired whether it had a distinctly separable 

 bark ; in the second, if it had a cortical integument that 

 differed, in its organization, from the wood, without being 

 separable from it; or, thirdly, if it possessed neither the 

 one nor the other. In the first place, it would have been 

 dicotyledonous ; in the second, monocotyledonous ; in the 

 third, acotyledonous, supposing it had been a trunk which 

 many successive years had contributed to form. The dis- 

 tinction, as applied to the two latter classes, is not, how- 

 ever, so positive as could be wished, because tree-ferns 

 have a cortical integument ; but they are easily known by 

 the long ragged scars left by their leaves ; and no other 

 cryptogamic plants possess the character of having a spu- 

 rious bark. 



A third, and very important kind of evidence, is to be 

 collected from the scars left upon stems by the fall of the 

 leaves. Although these will neither inform us of the shape 

 nor other characters of the leaves themselves, yet they indi- 



