of Geological Periods. 271 



We have now to apply these principles to the kind of investi- 

 gation that I wish to attempt ; and the following is the method 

 that I shall follow. The end proposed being to collect among 

 the fossil floras the indications of the temperature of the various 

 periods, I shall rely upon two groups or series of plants, of 

 which the completely opposite aptitudes alone can give rise to 

 decisive inductions. These two series are, on the one hand, that 

 of the genera now proper to the boreal zone, and, on the other, 

 that of the types of which the homologues are now met with 

 exclusively in the vicinity of the tropics. For the former, of 

 which the date of appearance is still uncertain, but which did 

 not show themselves in Europe until a comparatively recent 

 date, the starting-point to be selected for observing their pro- 

 gress is best placed in opposition to the direction of this pro- 

 gress — that is to say, in the present epoch. On departing from 

 this starting-point, we shall see this group, at first very com- 

 plete, diminishing gradually, and lastly losing itself in a past 

 state of things which is still very obscure. Must we follow the 

 same course in the observation of the tropical types ? I think 

 not. In fact when, ascending the course of ages, we meet with 

 them for the first time, these types are easily mfsunderstood ; 

 their very importance renders it necessary that they should be 

 rigorously defined. Instead of being connected, like the Euro- 

 pean types, with an order of things still existing in the land, 

 they are directly related to those of previous ages, and the latter, 

 by a train which becomes more and more visible, are themselves 

 connected with antecedent types. On the other hand, we are 

 certain a priori, notwithstanding differences of detail, of the 

 comparative elevation of the temperature during the most an- 

 cient period of vegetable life; and consequently, by starting 

 from the most ancient stages, and neglecting the genera without 

 any direct analogy with those of the present day, we are certain 

 of arriving at a moment when, the divergences becoming less, 

 genera identical with those now existing will necessarily make 

 their appearance, and furnish us with an approximate scale of 

 the temperature of these distant ages. 



Thus we obtain a double starting-point; and the courses to be 

 followed for one and for the other are in inverse directions, 

 since, in order to ascertain the existence of tropical types, we 

 shall descend the course of ages, whilst, on the contrary, we 

 shall ascend it when we have to study the progress of the 

 European types. In both courses of investigation I shall abstain 

 from mentioning any genus the existence of which does not 

 seem to me to be demonstrated, and which would not be iden- 

 tical with existing generic divisions ; and I shall always take the 

 latter in the most extended Linnean acceptation. 



