72 METABOLISM 



rings. Indeed, in many plants the youngest annual ring normally exhibits long, 

 continuous water columns entirely destitute of air-bubbles, while, further in, 

 the wood contains so much air that its conducting power is completely destroyed. 



The physical apparatus a transpiring osmotic cell with a glass tube attached 

 filled with water which we used as a model of what is found in a tree, 

 differs in many essentials from the reality. The glass tube has a wall which is 

 impermeable to water and air, it forms a continuous channel, and is also full 

 of water throughout. The vessel in the plant is permeable both to water and air, 

 it has always here and there transverse walls which interpose a certain amount 

 of resistance to the movements of water, and, furthermore, it is generally not con- 

 tinuously filled with water but with a chain of alternating bubbles of water and 

 air. Now if, as DIXON and ASKENASY affirm to be the case, the cohesion of water 

 particles plays an important part in the ascent of sap, the vessels ought to con- 

 tain no such chains. It cannot be denied that, as NAGELI (1866), DIXON and 

 JOLY (1895), and ASKENASY (1895, 16) have shown, the entrance of air-bubbles 

 into the vessel may be prevented by certain factors, but in reality everything 

 goes to show that, as a rule, the entrance of air into a vessel transporting water 

 can no longer be doubted. For this reason alone the cohesion hypothesis must 

 be given up, or, at least, it must be admitted that it can play no very important 

 part in the process. [Compare STEINBRINCK on the cohesion hypothesis (Ber. 

 d. bot. Gesell. 22, 526, 1904).] Further, even though the columns of water in the 

 vessels were continuous, the cohesion hypothesis could not be accepted as proved. 

 What we require in any theory of water movement is proof not only that the 

 water can be raised to a certain height, but also that it can rise to this level in 

 sufficient quantity. 



The knowledge which we have previously acquired on the subject of the 

 ascent of water as a result of capillarity is instructive. It is well known that 

 the concave meniscus, which forms in a glass tube immersed in water, has less 

 surface tension than a corresponding level surface of water, and that, in con- 

 sequence, water rises in a capillary tube above the level of the surrounding 

 fluid. The height to which the water ascends depends upon the concavity of 

 the meniscus, and that, in turn, on the diameter of the tube, and it is easy to see 

 that, given a sufficiently narrow tube, any height may be attained. If we con- 

 sider the cavities of the vessels filled with water as capillary tubes in a microscopic 

 sense and not much can be urged against this assumption then it is con- 

 ceivable that very tall trees might be supplied with water by capillary attraction. 

 NAGELI (1866) and STRASBURGER (1891) have, however, shown that a capillary 

 ascent of this kind is quite insufficient to replace the water lost in transpiration. 

 Although the principle under discussion is, from a purely physical point of view, 

 perfectly correct, it does not come into play under the conditions existing in the 

 plant. The same might be affirmed of the cohesion hypothesis. 



Taking into account all the facts which have been observed, we may 

 formulate our problem in this way : How can the ascent of sap be effected in the 

 plant if chains of water and air-bubbles ( JAMIN'S chains) occur in the vessels and 

 if transpiration is always effecting a suction at their upper ends ? On this question 

 we have to thank, more recently, SCHWENDENER (1893) and STEINBRINCK (1894) 

 for valuable information. It is possible that the ascent of water in a JAMIN'S 

 chain may take place in one of two ways, either the whole chain, or at least its 

 upper segments (including both air-bubbles and water drops), moves upwards, or 

 the water alone moves while the air-bubbles are stationary. Let us look first 

 at the movement of the whole chain. Let us imagine a long vessel or a glass 

 tube filled with air and water segments, each i mm. in length, with a suction 

 pump acting at the upper end ; by this means the air-bubbles will be stretched 

 out and the water columns will be pulled upwards. It is also obvious that the 

 topmost air-bubble will be extended most, and, in the long run, will show a pres- 



