36 Dr. W. C. Williamson on 



this series must have belonged to "branches of different 

 orders." In like manner they disbelieved that Fig. 20 of 

 Plate XLIII. ever was in the condition of Fig. 19, only- 

 altered by age and growth. Until now no tabulated col- 

 lection of the facts by which this controversy must be 

 decided has been made by anyone ; and I further doubt if 

 any cabinets exist, save my own, large enough to supply 

 such facts. But I have long been utilising the contents of 

 those cabinets in bringing together the facts, the full 

 knowledge of which is necessary for the settlement of the 

 question. 



But before proceeding to lay before the Society the 

 various measurements and other results of my investiga- 

 tions, I would call attention to what takes place sooner or 

 later, when each hitherto undivided stem dichotomises, pro- 

 ducing, at a higher level, two younger and smaller branches. 

 It is to the ascending series of these dichotomies that the 

 Lepidodendra owe their characteristic structure and modes 

 of development. The normal form of a transverse section 

 of an ordinary branch is cylindrical, as is also that of 

 several of the tissue systems that occupy its interior. The 

 first internal change that takes place affects the central 

 vascular cylinder of Primary Xylem, and its contained 

 cellular medulla. These can only be well seen on making 

 transverse sections of the branch. This cylinder becomes 

 split vertically for a very short distance, dividing it into two 

 crescentic diverging halves. As soon as this process has 

 fairly occurred, the external form of the branch begins to 

 alter. The two leafy surfaces parallel to a plane through 

 the two diverging crescents become flattened, and trans- 

 verse sections of the branch, at this point, gradually assume 

 an oval in place of a circular form. The higher the point 

 at which such sections are formed within the dichotomising 

 area, the greater becomes the difference between the longer 

 and shorter axes of the dividing cortex. Still no separation 



