K.— BOTANY 197 



sequent elongation of the cambium initial but will remain nearly transverse. 

 As a result, when the cambial initials are short, it is impossible for their 

 number to increase at any one level in the cambium cylinder in the manner 

 described for the softwood cambium. Instead, the cambium cells increase 

 in mass in a tangential direction and are then divided by a radial longi- 

 tudinal wall. When this process has been in progress for some time 

 the cambium naturally has a ' stratified ' appearance, when viewed in 

 longitudinal tangential section, as the ends of the cells derived from 

 one another will all be at the same level. Bailey has pointed out 

 that such stratified cambia are characteristic of hardwoods with 

 relatively short cambial initials, and that they are associated with a 

 marked contrast between the length of the vessel segments and the 

 fibres in the wood differentiating from the products of cambial 

 activity. 



In many hardwoods the cambium initials are not so short, new division 

 walls which are originally nearly transverse do subsequently become 

 oblique, and so the number of initials in the cambial cylinder multiplies 

 without radial longitudinal division and the cambium does not become 

 stratified. None the less, even these initials are shorter on the average 

 than those found in the softwoods, and, in the products formed from the 

 cambium on the inner and outer surfaces, as the result of the usual 

 tangential divisions, the differentiating elements begin with walls at top 

 and bottom which are relatively transverse. There are a few rare 

 exceptions. For example, in Drimys and Trochodendron , genera of the 

 Southern Hemisphere, the fusiform initials average more than four 

 millimetres in length, no walls appear to be specially transverse and no 

 vessels are subsequently differentiated. 



In almost all other hardwood genera the original fusiform initials are 

 comparatively short, the elements differentiating from them have walls 

 that, even if oblique, are relatively transverse, and when these elements 

 begin to vacuolate, the collapse of this wall leads to the formation of a 

 vessel . For when vacuolation commences , as the cell expands transversely , 

 expansion of a cell just below follows immediately and the transverse 

 walls between them, thus violently stretched, suddenly perforate, so that 

 the contents of the two cells coalesce. This happens with great rapidity 

 throughout a long chain of cells, and the common tracheal element 

 thus formed, which may be very long indeed, ultimately thickens and 

 lignifies its wall, loses most of its protoplasmic contents and becomes a 

 vessel. 



It is very frequently assumed that these perforating cross walls are 

 gradually digested, but this assumption becomes untenable when the 

 differentiating tissues are examined by the new method. Very long strips 

 can readily be peeled off the surface of the old wood in spring in which these 

 differentiating vessels can be traced for long distances. It then becomes clear 

 that the vacuolation which expands a series of vessel segments takes place 

 with extraordinary rapidity in cells whose walls are in a very thin ' primary ' 

 stage. Even in these preparations, in which the course of the differ- 

 entiating vessel can be followed, stages in perforation are very difficult 

 to find. There can be no doubt that the process occurs absolutely 



