THE CONDUCTION OF WATER. II 69 



The tracheids are on an average about i mm. long, 0-02 mm. broad, and as 

 much in depth. In transverse section they are nearly rectangular and so 

 arranged that their walls face radially and tangentially. Bordered pits occur 

 almost exclusively on the radial walls and the tapering ends, the tangential 

 walls being almost entirely free from them. If water be forced through coni- 

 ferous wood in a longitudinal direction it will meet with transverse walls which 

 are about a millimetre apart, but whose resistance to the flow is reduced by the 

 presence of pits in them. If water be forced through in a tangential direction 

 it will meet with fifty times as many walls as it will in the longitudinal course ; 

 finally, if it be forced through in a radial direction it will have to traverse 

 about the same number of walls as in the tangential course, but it will meet with 

 no pits in its journey. After this preliminary statement let us turn to Stras- 

 burger's experiments (1891) with the fresh wood of the silver fir, which may 

 be summarized as follows : — 



1. A column of water, 50 cm. in height, filters completely through a piece 

 of wood, 8 cm. long, in a longitudinal direction in one hour : it encounters 

 no opposition worth mentioning. 



2. A similar column of water passes in a tangential direction through 

 a piece of wood 1-3-5 cm. in length, at the rate of about 4-10 cm. in 20 hours. 



3. The opposition is so great in the radial direction that passage of the 

 water cannot as a rule be noticed at all, and if the pressure be increased by 

 means of a column of mercury, it is found that 50 cm. of mercury will drive 

 water radially through a piece of wood 1-5 cm. long at the rate of only about 

 4 cm. in 24 hours, and about 6 cm. in 48 hours. 



Similar results had been previously obtained by Sachs (1879, 297"! and 

 Elfving (1882). Thus we are able to understand clearly the significance of 

 the closing membranes of the pits, and may interpret the arrangement in this 

 way, that the margin of the membrane, when in a neutral position, behaves 

 like an ordinary closing membrane, while, when the torus is pressed to one side 

 against the pit opening, this thickened portion of the membrane, to a certain 

 extent, counteracts the activity of the pit. If, therefore, the torus becomes 

 sucked against the opening, owing to extensive abstraction of water and the 

 negative pressure of the air in the vessel resulting therefrom, then any further 

 withdrawal from this vessel must come to an end until this negative pressure 

 becomes general in the vicinity, and the closing membrane of the pit again 

 takes up a neutral position. This explanation, advanced by Strasburger 

 (1891), is the most satisfactory hypothesis of the function of the bordered pit, 

 and hence must be cited here, although it has not been conclusively demon- 

 strated (compare Schwendener, 1892, Ges. Abh. i, 288). 



As this torus -formation is not found in the pits of all vessels the I- -like 

 thickening of the wall must have a special significance. That this structure has 

 a meaning may be deduced from its very general occurrence ; but before we go 

 into this question we must examine more closely into the purpose of the thicken- 

 ing itself. Since the thin regions of the wall of the vessel are especially adapted 

 for the passage of water, why is the whole wall not of this uniform thickness ? 

 We can scarcely go wrong if we recognize the advantages which the plant 

 gains by the thickening from a mechanical point of view. Ordinary parenchy- 

 matous cells with thin walls are able to acquire very considerable rigidity by 

 means of osmotic pressure, but the water in the lumina of the vessel remains 

 either at atmospheric pressure or even less ; only rarely is the pressure 

 higher than that of the air. Further, if turgid cells surround a vessel, they 

 would be able to press the walls together and so obliterate the lumen, were 

 they not prevented from doing so by the great rigidity of the membrane 

 of the vessel. The membrane acquires this rigidity by means of the 

 thickening as well as in consequence of other physical characteristics, and 



