THE ANGiOSPERMAE : STEMS 893 



increasing the number of cells in the ring, as it expands with growth. In 

 radial section the long cells are very narrow, with chisel edges at both ends, 

 which overlap those of the next cells above and below. This overlapping 

 of the cell ends has been attributed to " sliding growth," in which the walls 

 of neighbouring cells actually slip along each other, but this is doubtful, for 

 many reasons, and it seems more probable that it is produced by local growth 

 of the young, plastic cell-walls, above and below each new transverse division 

 wall, at opposite ends respectively, in such a way that the transverse wall is 

 tipped into the vertical without any mutual slipping of the cells. The 

 cambial ray cells, on the contrary, divide in all three planes. 



Bailey has shown that in Dicotyledons there are three main types of 

 cambium, corresponding to three types of wood structure. 



1. Initial cells averaging 890 microns in length, producing long, narrow 



vessel segments with scalariform pitting of the end walls. 



2. Initial cells averaging 410 microns, producing short, broad vessel 



segments with cross walls which are simply pitted or else 

 absent. 



3. Initial cells averaging 250 microns, producing vessels similar to type 2, 



but the cambial cells are arranged in regular horizontal rows 

 (" Etagen cambium ") and give rise to a corresponding " storied " 

 structure in the wood. 



These three types are regarded as marking advancement in specialization 

 of the vascular structure. 



The cambium in the resting state, in winter, is sharply marked off from 

 the vascular tissues, with fully formed vessels and sieve tubes lying in 

 immediate contact with the cambial cells from which they have been formed. 

 During the summer activity, however, the cambial zone seems much broader 

 and less clearly delimited, because tangential divisions follow one another so 

 rapidly that the true cambial zone is flanked on both sides by a layer of still 

 undifl'erentiated cells, and the transition to the mature vascular elements is 

 thus spread over a broader zone of development. 



It has been shown that the cambial growth in spring begins in the swelling 

 buds and spreads downwards from them. It begins simultaneously all over 

 the tree, but the rate of spread is much slower in the older branches and it 

 stops sooner in them than in the young shoots. There is good evidence that 

 the stimulus to cambial growth is the diffusion of hormones from the leaf 

 initials in the bud. Auxin-a and heterauxin (see Volume III) can both 

 start strong cambial growth when they are supplied to a shoot, and auxin 

 has been found in naturally active cambium and is known to be produced by 

 leaf initials. This direct relation of leaf development to cambial activity is 

 probably the reason for the proportionality between the extent of leaf surface 

 and the amount of vascular tissue formed in the stem, which observation has 

 revealed. The hormone explanation may hold good not only in the growth 

 of the normal cambium but in the differentiation of new vascular tissue from 



