In a previous paper an account of certain investigations into the physics of the flow of water through wood has been given. These experiments showed that in the wood of actively transpiring trees the total pressure required to maintain the transpiration rate of flow might amount to a head of water of from five to seven times the height of the tree. The factors which produce this resistance are the friction against the sides of the vessels, the passage through the end walls, and the special highly variable resistance due to the presence of air bubbles in the vessels. Other things being equal, the resistance to flow is strictly proportional to the length of the stem, provided that the conducting tract is approximately similar in character and in sectional area throughout the length of the piece of stem tested. Hence it is difficult to see any reason for Curtis’s statement that the resistance offered by a stem to the flow of water is not proportional to the length of the stem, unless this observer worked with partially blocked stems. This author also states that the same force is required to overcome the resistance to a definite rate of flow, whether applied as a pressure or as a tension. This is, however, only the case when the conducting elements are completely filled with water, and even then the resistance to flow is two or three times greater than would be expected from a viscosity calculation. The latter fact is partly due to the internal thickenings on the vessels, which reduce their effective radius, and partly to the existence of transverse partitions in long pieces of stem, which add to the resistance to longitudinal flow. The average resistance to flow through the side wall of a vessel or through that of a tracheide of a Conifer may be from 2 to 10 or more times the resistance to flow through the entire length of the same vessel when filled with water, although when large numbers of air bubbles are present the tendency to lateral exudation from a given longitudinal path is much increased. In this way water can be transmitted rapidly and with but little lateral exudation through tracts where the supply is abundant and the local demand small, so that the vessels are filled, whereas in a region where the loss by transpiration is excessive, the appearance of air bubbles in the vessels partially blocks the upward flow and increases the lateral exudation until the local needs are supplied. Strasburger attempted to show that the flow of fluids through the vessels of wood was dependent upon their viscosity, by driving such liquids as water, turpentine, alcohol, ether, and benzole through short lengths of stem under similar heads and noting the time of formation of a drop in each case. Strasburger’s method is vitiated by his assumption that the drops of the different liquids were of the same size, which is not the case, and he also forgets that the different densities of the liquids cause the weight of the columns in the vessels to vary. Nevertheless he found that the number of drops passing through in a given time was approximately proportional to the viscosity of the liquids used. As a matter of fact, the values obtained for the viscosity by this method depend largely upon the order in which the different liquids are passed through the stem, even when the actual volume escaping is measured. Thus water preceding turpentine or benzole gives absurdly different results to water following these liquids, and to a less extent the same applies with alcohol and ether. The relation between flow and viscosity can be best shown by comparing the flow of water at different temperatures, as has already been done.
VIII. The ascent of water in trees. (Second paper.)
