Decomposition is usually accompanied by pronounced odors. Decaying poplar emits a disagreeable odor, while red oak often becomes fragrant, its smell resembling that of heliotrope. WEIGHT OF WOOD A small cross-section of wood (as in Fig. 19) dropped into water sinks, showing that the substance of which wood fibre or wood is built up is heavier than water. By immersing the wood successively in heavier liquids, until we find a liquid in which it does not sink, and comparing the weight of the same with water, we find that wood substance is about 1.6 times as heavy as water, and that this is as true of poplar as of oak or pine. [Illustration: Fig. 19. Cross-section of a Group of Wood Fibres (Highly Magnified.)] Separating a single cell (as shown in Fig. 20, _a_), drying and then dropping it into water, it floats. The air-filled cell cavity or interior reduces its weight, and, like an empty corked bottle, it weighs less than the water. Soon, however, water soaks into the cell, when it fills up and sinks. Many such cells grown together, as in a block of wood, when all or most of them are filled with water, will float as long as the majority of them are empty or only partially filled. This is why a green, sappy pine pole soon sinks in "driving" (floating down stream). Its cells are largely filled before it is thrown in, and but little additional water suffices to make its weight greater than that of the water. In a good-sized white pine log, composed chiefly of empty cells (heartwood), the water requires a very long time to fill up the cells (five years would not suffice to fill them all), and therefore the log may float for many months. When the wall of the wood fibre is very thick (five eighths or more of the volume, as in Fig. 20, _b_), the fibre sinks whether empty or filled. This applies to most of the fibres of the dark summer-wood bands in pines, and to the compact fibres of oak or hickory, and many, especially tropical woods, have such thick-walled cells and so little empty or air space that they never float. [Illustration: Fig. 20. Isolated Fibres of Wood.] Here, then, are the two main factors of weight in wood; the amount of cell wall or wood substance constant for any given piece, and the amount of water contained in the wood, variable even in the standing tree, and only in part eliminated in drying. The weight of the green wood of any species varies chiefly as a second factor, and is entirely misleading, if the relative weight of different kinds is sought. Thus some green sticks of the otherwise lighter cypress and gum sink more readily than fresh oak.