to the far more numerous tangential walls. It is also easy to see that the cork-cells must be arranged in radial rows, and this arrangement is very conspicuous (Fig. 18). The earlier cork-cells have very thin walls, later ones have the walls thicker.
After the development of the first layer of cork the stretched epidermis dies, and forms simply a dead membrane outside the thin cork. In succeeding years layers of phellogen are formed annually beneath the older ones, and thus the cork layers increase. Moreover, since the successive layers cut out thin, scale-like areas of cortex, trapping them, as it were, between the present and the preceding cork, the thickening corky covering is stratified — consists of successive and obliquely overlying thin sheets of dead cortex and cork proper (Fig. 30). Again, since the increase in thickness of the stem or branch is continually driving these corky and dead structures outwards, they at length crack, and form the fissured bark found on older parts. Bark is thus seen to be something more than cork, or even periderm, and it is defined to be all the dead tissues cut out by the phellogen.
It is also to be noticed that the successive phellogen layers of different years are not concentric, but the new ones cut the old ones at acute angles (Fig. 30), thus cutting out scale-like areas of cortex; the consequence of this is the formation of the very irregular scales of bark thrown off from the older stems and branches of the oak. It follows from what has been said that in older