We know that a part of the work of the root is to penetrate the soil and hold the plant firmly in place. Therefore, it needs a firm soil. We know that the part of the root which penetrates the soil is tender and easily injured. Therefore, for rapid growth the root needs a mellow soil. We know that part of the work of the root is to take moisture from the soil. Therefore, it needs a moist soil. We know that part of the work of the root is to take food from the soil. Therefore, it needs a soil well supplied with plant food. We know that roots stop their work in cold weather. Therefore, they need a warm soil. Another condition needed by roots we will find out by experiment. =Experiment.=--Take two wide-mouthed clear glass bottles (Fig. 16); fill one nearly full of water from the well or hydrant; fill the other bottle nearly full of water that has been boiled and cooled; place in each bottle a slip or cutting of Wandering Jew (called also inch plant, or tradescantia, and spiderwort), or some other plant that roots readily in water. Then pour on top of the boiled water about a quarter of an inch of oil--lard oil or cotton-seed oil or salad oil. This is to prevent the absorption of air. In a few days roots will appear on the slip in the hydrant water, while only a very few short ones, if any, will appear in the boiled water, and they will soon cease growing. Why is this? To answer this question, try another experiment. Take two bottles, filled as before, one with hydrant water and the other with boiled water; drop into each a slip of glass or a spoon or piece of metal long enough so that one end will rest on the bottom and the other against the side of the bottle, and let stand for an hour or so (Fig. 17). At the end of that time bubbles of air will be seen collecting on the glass or spoon in the hydrant water, but none in the boiled water. This shows us that water contains more or less air, and that boiling the water drives this air out. The cutting in the boiled water did not produce roots because there was no air in it and the oil kept it from absorbing any. =Experiment.=--Into some tumblers of moist sand put cuttings of several kinds of plants that root readily (Fig. 18), geranium, tradescantia, begonia, etc. Put cuttings of same plants into tumblers filled with clay that has been wet and stirred very thoroughly, until it is about the consistency of cake batter. Keep the sand and puddled clay moist; do not allow the clay to crack, which it will do if it dries. The cuttings in the sand will strike root and grow, while most, if not all, those in the clay will soon die. The reason for this is that the sand is well ventilated and there is sufficient air for root development, while the clay is very poorly ventilated, and there is not sufficient air for root growth. These experiments show us that to develop and do their work roots need air or a well-ventilated soil.