Manchester Memoirs, Vol. Hi. (1908), No. S. 13 



between Bothrodcndron and Miadesuiia seem to disprove 

 any possibility of a connection between them. 



Except for the great radial shortness of its sporo- 

 phylls, the Bothrodcndron cone does not differ very much 

 from Lcpidostrobus. 



There is, however, no doubt that it must be regarded 

 as generically distinct from Lepidostrolms, in which the 

 radial extension of the sporophylls is a characteristic 

 feature. At the same time there is, I think, no doubt that 

 it is a member of the Lepidodendraceae. The whole 

 structure of the axis, the occurrence of a bifid parichnos 

 in the stems, and the occurrence of the Ulodendroid 

 conditions shew that it is quite closely related to the 

 " genera " Lcpidodendron, Lepidofloios, and Sigillaria, which 

 are the other members of the Lepidodendraceae as gener- 

 ally accepted. On the whole, the cone seems to differ 

 more from that of any of the three other genera than the 

 latter do amongst themselves.* 



It has recently been suggested by Dr. Scott :o6 and 

 referred to by Professor Oliver :07 that Spencerites may be 

 related to BotJirodcndroii. 



The differences between the cone here attributed to 

 iH^tJirodcndron inundinii and Spencerites are as follows : — 



1. The attachment of the sporangium to a ventral 



process at the distal end of the horizontal limb in 

 Spencerites is quite different to the simple attach- 

 ment in Bothrodcndron. 



2. It is improbable that there is a ligule in Spencerites 



(Berridge 105). Since the occurrence of a ligule 

 is regarded by many botanists as an important 



* The numerous well-preserved examples oi Sigillariostrobits which have 

 now been examined leave no doubt that that cone was in all essentials 

 identical with Lepidos/robus. 1 have reasons for believing that certain cones 



of the Lepidosirobus oldhamiitin type found petrified are Sigillarioid. 



