THE CARBONIFEROUS FLORA. 153 



Genus CARDIOCARPUM. 



I have found at least eight species of these fruits in the Erian 

 and Carboniferous of New Brunswick and Nova Scotia, all of which 

 are evidently fruits of gymnospermous trees. They agree in hav- 

 ing a dense coaly nucleus of appreciable thickness, even in the 

 flattened specimens, and surrounded by a thin and veinless wing or 

 margin. They have thus precisely the appearance of samaras of 

 many existing forest-trees, some of which they also resemble in the 

 outline of the margin, except that the wings of samaras are usually 

 veiny. The character of the nucleus, and the occasional appearance 

 in it of marks possibly representing cotyledons or embryos, forbids 

 the supposition that they are spore-cases. They must have been 

 fruits of phaenogams. Whether they were winged fruits or seeds, 

 or fruits with a pulpy envelope like those of cycads and some 

 conifers, may be considered less certain. The not infrequent dis- 

 tortion of the margin is an argument in favour of the latter view, 

 though this may also be supposed to have occurred in samaras par- 

 tially decayed. On the other hand, their being always apparently 

 flattened in one plane, and the nucleus being seldom, if ever, found 

 denuded of its margin, are arguments in favour of their having been 

 winged nutlets or seeds. Until recently I had regarded the latter 

 view as more probable, and so stated the matter in the second edi- 

 tion of " Acadian Geology." I have, however, lately arrived at the 

 conclusion that the Cardiocarpa of the type of C. cornutum were 

 gymnospermous seeds, having two cotyledons embedded in an albu- 

 men and covered with a strong membranous or woody tegmen sur- 

 rounded by a fleshy outer coat, and that the notch at the apex rep- 

 resents the foramen or micropyle of the ovule. The structure was 

 indeed very similar to that of the seeds of Taxus and of Salisburia. 

 With respect to some of the other species, however, especially those 

 with very broad margins, it still appears likely that they were winged. 



The Cardiocarpa were borne in racemes or groups, and it seems 

 certain that some of them at least are the seeds of Cordaites. The 

 association of some of them and of those of the next genus with 

 SigillaricB is so constant that I cannot doubt that some of them 

 belong to plants of that genus, or possibly to taxine conifers. The 

 great number of distinct species of these seeds, as compared with 

 that of known trees which could have produced them, is very re- 

 markable. 



Genus TRIGONOCARPUM. 



These are large angled nuts contained in a thick envelope, and 

 showing internal structures resembling those of the seeds of modern 



