26o WARD 



absent, and that its absence is always due to some external in- 

 fluence acting upon the surface which has removed it. There 

 is an abundance of proof of this, and most of the specimens 

 show parts over which the external coating still adheres and 

 other parts where it is absent. The latter usually reveal the 

 nature of the agency that has removed the coating, whether a 

 sudden and violent concussion, a gradual erosion, or a process 

 of weathering. The contact of the outer layer with the surface 

 of the armor proper is always marked b}'' a clear plane of sepa- 

 ration, and usually by an open structure or even a partially void 

 space. This becomes a natural plane of cleavage, and almost 

 any influence will cause the outer coating to scale off like the 

 outer bark of a tree. 



In the specimens of the Yale Museum this outer coating had 

 almost entirely disappeared, though not absolutely, so that the 

 phenomenon did not specially strike me, and I only noted that 

 the surfaces were obscure in places. Through the kindness 

 of Dr. C. E. Beecher, these specimens were sent to me for 

 further examination and comparison with those of the large 

 collection from Mr. Knight. Some of the important results 

 of this comparison M'ill be noted later on, but it is sufficient to 

 state here that they form no exception or anomaly, but are 

 simply part and parcel of the general lot. 



GENERIC CHARACTERS OF FOSSIL CYCADEAN TRUNKS. 



Generic characters, with the exception of Bennettites, which 

 is identical with Cycadeoidea except in the accident that seeds 

 have been discovered in the spadices, have generally been based 

 on the shape of the trunk and on the character of the armor, 

 /. ^., of the remains of the foliar organs still adhering to the 

 trunk in the fossil state. The former of these characters have 

 proved of less constancy, and in cases where the latter class of 

 characters is distinctive *authors have not hesitated to ignore va- 

 riations in the former, as, e. g.^ Cycadeoidea gigantea of Sew- 

 ard, a tall cylindrical trunk wholly different in form from other 

 species of that genus. I was obliged to do the same with C. 

 exceha and C. jenneyana. 



The second class of characters is relatively constant and diag- 



