6 Watson, On the Ulodcndroid Sccxr. 



scars with a central umbilicus these dots are usually- 

 continued all round the scar, but in oval scars with an 

 eccentric umbilicus they only cover the lower portion, the 

 upper being covered by longitudinal markings radially 

 arranged, which, however, gradually shorten at the sides 

 of the scar and shade off into the dots of the lower part. 

 I do not see how this arrangement, which is the normal 

 one, can be satisfactorily accounted for on the theory that 

 the dots represent crushed leaf-bases, which is that 

 adopted by the upholders of the orthodox cone view. 



If they do represent crushed leaf-bases then these must 

 vary very greatly in size and shape (wer the area of the 

 scar, which however is supposed to have been only an 

 ordinary piece of the stem surface. 



On the branch theory the dots merely represent the 

 cut ends of vascular bundles supplying leaves formerly 

 attached to the branch, and the difficulty of their varj'ing 

 distances apart does not arise ; on this theory also the 

 longitudinal markings on the upper part of the scar are 

 .accounted for as being obliquely and sometimes longi- 

 tudinally cut leaf-traces. 



It must be pointed out that in the specimen described 

 above, which shows the leaf-traces and their origin from 

 the wood, these explanations seem to work perfectly. 



The Text-figure will, I hope, make the meaning of this 

 explanation more clear ; it should be compared with a 

 somewhat similar figure given by Carruthers. 



The branch theory also explains the fact that Ulodcn- 

 droid scars are often provided with a raised rim at the 

 edge. It simply represents, I take it, a scrap of the outer 

 cortex of the branch left connected to the parent stem. 



An additional argument in favour of the branch view, 

 is that we know that some Lepidodendra, of which the 

 structure is known, did bear their branches in two 



