﻿FERN ANCESTRY AND ANGIOSPERM ANALOGIES. 243 



fact. The retention and ripening of the megaspore is then seen to be no more than 

 analogous to the maturing of pollen, though involving greater time, and eventually 

 more complex nutritive processes. 



In conclusion, it remains to suggest some of those floral features of the Cyca- 

 deoidece, as members of the proangiosperms of Saporta, which suggest at once the 

 existence of fundamental relationships between gymnosperms and aiigiosperrns as 

 well as possible main lines of angiosperm evolution. 



ULTRA-RELATIONSHIP OF THE CYCADALEAN GYMNOSPERMS (OR THE 

 CYCADEOID-ANGIOSPERM JUXTAPOSITION). 



Within the past few years evidence has been rapidly accumulating that the 

 gymnosperms have all been remotely derived from the ferns. The most general- 

 ized, but withal a much modified, gymnosperm stock combining characters of the 

 conifers and cycadophytes, the Cordaitales, are first recognized in the Devonian, 

 having by this period become highly organized and stately forest types. How- 

 ever, the primitive seed-ferns from which these conspicuous forest types of the 

 Devonian sprang doubtless persisted till a much later period before giving rise to 

 such types as the Mesozoic Cycadeoidese and, as I believe, at much the same time 

 or a little later than these the early angiospenns. It is especially to Nathorst that 

 we owe the discovery of evidence going to show the most surprising variety of 

 types of fructification among cycadophytes, such, for instance, as reproduction by 

 means of antherangia (103). This condition, however, goes far toward precluding 

 the descent of all the "cycadophytes" from Marattiaceous types, if it does not even 

 stamp this as an impossibility. It is more and more borne home to us that while 

 we ever and anon gain clearer and clearer insight into the nature of certain features 

 of the greater lines of evidence concerning the descent of the great groups of exist- 

 ing plants, we are, in the absence of the exact facts, singularly helpless when we 

 attempt to frame hypotheses of descent. The fact is also emphasized that the 

 gymnosperm plexus includes not a few score, but hundreds or even thousands of 

 modifications of generic or family value of which we know well scarcely a half 

 dozen of the fundamentally important extinct forms. In the absence of more exact 

 knowledge of such great numbers of diverse forms that gave rise at widely separated 

 periods of time to existing spermaphytes, lines of descent can not readily be picked 

 out. One too soon. becomes involved in extreme difficulties. Attempts at decipher- 

 ing these lines of descent involve morphologic conceptions which, simple though 

 they be, are not directly thinkable, or even if dimly perceived sufficiently recogniz- 

 able, because of the lack of historical perspective. As a result no two investigators 

 ever reached the same conclusion, when nearing the point to which the known 

 evidence actually carries us. But to determine this point is ever the ardent 

 ambition of the paleobiologist. It results that the sanguine worker is constantly 

 asking too much of the evidence, his patient and doubting brother too little. Yet 

 taken merely on the basis of an arithmetical probability, it would be most extraor- 

 dinary if at the present day the angiosperm line of descent could be laid down, 

 except on the broadest lines. It would be most extraordinary, we say, if a 



