THE CAMBIUM AND ITS DERIVATIVE TISSUES 365 
fusions, and changes from the diploid to the tetraploid or polyploid con- 
dition are of frequent occurrence, many cells depart widely from the in- 
herited, specific cell size of the plant. Therefore he suggests that there is a 
close correlation between cell size and chromosomal mass in both meriste- 
matic and non-meristematic somatic tissues. 
Reconnaissance surveys of the higher plants indicate that the cambium 
should provide a favorable medium for testihg the validity of these and 
similar generalizations concerning cell size, the working sphere of the 
nucleus, and the nucleo-cytoplasmic relation. Not only does the average 
size of the cambial initials fluctuate greatly in different groups of the 
Siphonogama, in different parts of a given individual, and in plants grown 
under different environmental conditions, but adjacent elements of the 
lateral meristem vary considerably in length, cross-sectional area, and 
volume. The cambial initials are of two distinct shapes and sizes: (i) 
numerous large, elongated cells, whose size variations have been described 
on preceding pages, and (2) scattered aggregations of small, more or less 
isodiametric elements which divide to form the horizontal sheets of radially 
disposed parenchyma, so-called medullary rays. The bulk of the divisions 
in both types of initials is periclinal, or parallel to tangents to the circum- 
ference of the stem or root. In other words, the large cells divide in a 
tangential, longitudinal plane which is a division plane of maximal area, 
whereas the ray initials form partition membranes that commonly are 
surfaces of minimal area. In gymnosperms and less highly differentiated 
dicotyledons, the cambium does not increase its periphery by radial, 
longitudinal divisions of the elongated initials and lateral enlargement of 
the products of such divisions. Instead, the cells elongate, sliding by one 
another, until they have attained a certain length. They then divide, by 
means of a pseudo-transverse partition, into two short halves which in turn 
elongate and divide.^ Owing to the fact that the initials do not elongate 
and divide (transversely) in unison, there is usually a very considerable 
variability in the length and pari passu in the volume of adjacent fusiform 
elements. However, the volume of the more or less isodiametric ray initials 
is very much less than that of even the smallest fusiform initials, and is of 
the same general order of magnitude as that of the undifferentiated cells of 
the embryo or terminal meristem. Therefore, in any particular portion of 
the cambium of these plants it is possible not only to study cell division 
and the nucleo-cytoplasmic relation in adjacent fusiform initials of very 
different lengths and volumes, but to contrast them with similar phenomena 
in adjoining ray initials, which resemble the cells of the terminal meristem 
in size and shape. Furthermore, by proper experimental methods, the 
fusiform initials may be induced to divide into small isodiametric units of 
the general order of magnitude of the ray initials or embryonic cells, and 
subsequently to regenerate elongated elements of normal dimensions. 
^ During this process of elongation, between successive transverse divisions, the cells 
continue to divide in the tangential, longitudinal plane. 
