I now proceed to some further details, which could only be rendered intelligible in the light of the preceding preliminary remarks.
After the cambium ring is once formed the daughter-cells cut off on the inside of the cambium always become transformed into one or more of the following elements:
(1) Some cambium cells which lie on the radial continuation of a medullary ray undergo a few horizontal divisions across the long axis, and then simply pass over as constituents of a medullary ray; as the cambium ring moves outward, in consequence of the repeated formation of thickening rings, the periphery of the cylinder of course increases, and this allows of more space tangentially. One consequence of this is the occasional and gradual widening of the medullary ray in process of lengthening; this takes place to a small extent only. Another consequence of the increased space is the occasional interpolation of new medullary rays. Radial rows of cambial cells at points which lie between the planes of two gradually diverging medullary rays suddenly commence to form new medullary rays. Hence, as the wood mass increases in radial thickness, more and more of these interpolated medullary rays appear, cutting up the wood proper into partial sections. In succeeding years the cambium keeps adding to the length of these rays, as it does to that of the older rays, and again forms new ones between as space increases. In the same ring about thir-
