Remarks on the Ancient Flora of the Earth. 129 



this remarkable exception to the prevailing organization, yet the 

 influence which this discovery can have on the classification of 

 fossil vegetables, must always be of small importance. For it 

 depends^on the nature of the preservation of these remains of a 

 former age, whether we, in all our attempts to arrange them in 

 naturally'distinct groups, according to the peculiarity of their 

 forms, which arise from the organs of vegetation, will assign a 

 greater importance to these, than to the organs of propagation. 

 Their essential differences are particularly observable only in 

 those parts which can scarcely ever come under the eye of an 

 observer of the plants of a former age ; hence the characteristics 

 derived from such parts must evidently be placed in the lowest 

 order. If we now consider the single varieties of Cycadecs and 

 Coni/ercv according to their external perceptible peculiarities— 

 if we compare the formation of their stems, the nature of the 

 insertion of their leaves, &c., we will be at no loss what place to 

 assign them. A. Richard, and after him DecandoUe, has ad- 

 mitted, that the CycadecB are most closely related to palms and 

 arborescent Monocotyledoncs, while the Coniferct are imme- 

 diately connected with the more perfectly organized dicotyle- 

 dones. Indeed, the stems of the one, when found without being 

 accompanied by leaves and fruit, are subordinate to the division 

 Endogenites ; but the others are such perfect Exogenitece, that 

 we cannot, without reluctance, surrender the general validity of 

 this important distinction as the foundation of two great main 

 divisions of all remains of plants not precisely determined. M. 

 Brongniart himself, from similar grounds probably, has inserted 

 in his Tabular Survey the genera Endogenites and Exogeiiites, 

 though they here lose all signification, from his having establish- 

 ed a fourth class. 



Though, for the reasons assigned, we wished both families of 

 this new class in M. Brongniart's Tabular* Survey separated, and 

 the one associated with the Monocotyledonous Phanerogamia, 

 the other with the Dicotyledonous Phanerogamia, we have by 

 no means overlooked their very close affinity, which first of all be- 

 came remarkable from the memorable researches of Richard and 

 his son. The pecuHar embfyo with two cotyledons, which had 

 not before been observed of any known species of the Pnlviei^, 

 Liliacc(r, &c. brings the Cycadcac so very near the Dicotyle- 



OCTOBKIl — DECKMBER 1829- ^ 



