No. 505] KELPS AND RECAPITULATION THEORY 



25 



ah appearance similar to Lessonia, it is brought about by 

 a fundamentally different process, as lias already been 

 noted by Setchell ('93) and by Keinke ('03), win. dgure 

 several stages in the development of a branch. Some of 

 the earlier proliferations, as stated above, soon develop 

 roughenings on their stalks like those of the main stipe 

 and take on the appearance of younger specimens of the 

 species (Fig. 31, b). This is the first external indication 

 of an important difference in the constitution of these out- 

 growths from the ordinary proliferations. For in them 

 has become differentiated a meristem independent of that 

 of the primary branch. They develop exactly as did the 

 main axis and soon become indistinguishable from it ex- 

 cept in the manner of attachment to the holdfast, possess- 

 ing all the structures which have been described for a 

 primary branch including other branches which in turn 

 go through the same process. After several such 

 branches have been formed there is a modification of the 

 process. The laminae are dwarfed, while their margins 

 become conspicuously puckered and ruffled (Fig. 35, a). 

 Sometimes the ruffles are so pronounced as to completely 

 enfold the meristem. In such a branch proliferations 

 from the lamina appear very late, but the ruffle gives it a 

 similar aspect. The dwarfed condition of the lamime per- 

 sists until the stipes become several centimeters in length, 

 when the usual relations of stipe and lamina become 

 manifest. Though roughening may appear on other 

 parts of the plant, the development of meristematic pro- 

 liferations is confined to the basal portion; branches do 

 not develop at a distance much exceeding L'O cm. from the 

 holdfast. Around the base of any old plant there is al- 

 ways a large number of short branches in all stages of de- 

 velopment, but there are not often more than a dozen 

 long branches at any one time. The general appearance 

 of the numerous dwarf branches suggests that they may 

 not have a rapid development like the first branches, but 

 rather grow very slowly or lie dormant for a time like the 

 dormant buds of trees. 



This method of branching is peculiar to Egregia and, as 



