148 D. H. Scott. 



course exclusivel}'. coal-plants. Though the conditions favourable to 

 the formation of coal-fields must have been extraordinarily wide- 

 spread during the later Palaeozoic times, yet these conditions, invol- 

 ving the presence of a luxuriant swamp-vegetation, must after all have 

 been local. As has often been said, we probably know next to nothing 

 of the flora of the Palaeozoic uplands. As an illustration of the 

 special nature of the particular floras witli which we are concerned, 

 it may be mentioned that the petrified plants of the roof-nodules in 

 certain Lancashire collieries, are usually of a different tj^îe from those 

 of the nodules in the seam immediately beloAv them. The difference 

 in age must be unimportant, but the roof-nodules probablj* represent 

 remains drifted from a distance, after the coal-bed was sunk below 

 the water-level, while those of the seam itself contain the plants 

 which grew more or less in situ. If a trifling difference such as this 

 perceptibly changes the character of the flora, we may conjecture 

 what much greater contrasts would reveal themselves if the plants of 

 localities remote from the coal-area were open to investigation. 



The petrified remains, which are so essential to the right inter- 

 pretation of the fossil evidence, are especially limited in their distri- 

 bution, compared with the more common form of preservation as 

 casts and impressions. For example, scarcely anything is known as 

 to the structure of the plants composing the "Glossopteris Flora" ot 

 India and the Southern Hemisphere, and hence we are left much in 

 the dark as to their affinities. These limitations to our knowledge, 

 while they need not discourage us from making the fullest use of the 

 available data, may warn us of the danger of framing too narrow a 

 conception of the vegetation of the ages Ave are considering. 



In discussing the affinities of Palaeozoic Vascular Plants there 

 are certain advantages in beginnino- with the .Sphenophyllales, a class, 

 which, though not extensive, is important from its synthetic character, 

 and probably represents an extremely ancient stock. 



A. LYCOPSIDA. 



I. Sphenophyllales. 



The plants definitely referable to this class are at present placed 

 in two genera, SphenophyUum and CJieirostrohus, each of which must 

 be taken as representing a distinct family. 



SplienophyJlnm, which embraces a number of species, ranging from 

 the Middle Devonian to the Permian, or perhaps the base of the 



