All our forest trees are of the exogenous type; that is, their stems increase in thickness by the addition of new wood formed outside the older wood and underneath the bark. Thus the bark, which is composed of a layer or mass of dead, sapless cells, is gradually pushed outward as the stem thickens. The result is that the bark is either more or less fractured, as in the Elm and the Oak, or it flakes off and falls to the ground, as is the case with the Plane and the Birch. A new layer of bark is always formed during each summer, and this, in turn, either cracks or peels away; but while, in the former instance, the accumulated bark presents a very rugged appearance, and becomes very thick, in the latter case it remains smooth, and is always thin. Then again, how are we to account for the great variety in the general forms of our different trees--the irregular, crooked nature of the Oak; the slender, but denser branching of the airy Birch; and the tall, pyramidal form of the Lombardy Poplar? All this is easily understood if we carefully observe the positions of the buds as seen during the winter months; and watch the development of these buds during early spring. [Illustration: TREES IN WINTER OR EARLY SPRING 1. Hazel, with catkins. 2. Ash. 3. Oak. 4. Lime, with remains of the last season's fruits.] If the buds are irregularly scattered on the twigs, the lateral buds being as strongly developed as the terminal ones, while, in the spring, as is often the case, certain only of the buds develop into new twigs, the others remaining dormant, then the branches assume that irregular, crooked appearance so characteristic of the Oak. If, on the other hand, all the terminal buds are well developed, and the lateral buds are weaker and more regularly distributed, but farther apart, then the tree grows more rapidly in height than in breadth, and assumes more nearly the character of the Pyramidal Poplar. It will thus be seen that the study of trees in their winter condition is not altogether lacking in interest. Referring once more, but briefly, to the matter of dormant buds, we recommend the reader not only to observe that some buds do not expand with the others during the spring, but to make them the subject of experiment. Thus, when the Horsechestnut is well in leaf, dormant buds will usually be seen on the sides of the twigs, sheltered by the spreading leaves produced at the tips. Now remove the whole cluster of leaves formed by the terminal bud, together with the bud itself, and the hitherto dormant laterals, under the influence of increased light and warmth, and supplied with sap that is now directed into new channels, will speedily show signs of growth. Similarly, the fruit-gardener will remove the tips of the branches of his fruit trees, which often bear buds that are destined to produce leafy twigs only, and thus encourage the growth of the fruiting buds that are situated lower on the twigs.