FRICTION OF BORDERED PITS 319 



Similarly, it is not yet possible to indicate the precise function 

 performed by the bordered pits which constitute such a characteristic 

 feature of the walls of water-conducting elements. As Schwendener 

 has pointed out, the construction of a bordered pit tends to increase the 

 surface available for diffusion without reducing the strength of the wall 

 more than is absolutely unavoidable (cf. Ch. I.) The development of 

 the border which overlaps the pit-cavity, thus in a sense represents a 

 compromise, between the requirements of the mechanical principle on 

 the one hand, and the demands of translocation on the other. This 

 explanation, however, does not account for the presence of the most 

 characteristic part of the limiting membrane, namely, the torus. Direct 

 observation has shown, that the torus may become closely appressed to 

 the orifice of the pit, so that the latter is tightly closed. The bordered 

 pits are thus undoubtedly adapted for a very special purpose. Russow 

 states that they represent minute valves, a view which is certainly 

 borne out by every feature of their construction. The torus must be 

 assumed to be relatively impervious to water, and, when in a state of 

 imbibition, also to air ; if, therefore, the pressure becomes unequal on the 

 two sides of the limiting membrane, the torus is forced against the 

 mouth of the pit, which thus becomes more or less effectually closed. 

 Where the border is thin, the margin of the aperture bulges slightly 

 inwards, an arrangement which helps to ensure a good contact with the 

 limiting membrane ; this point was first noted by Russow. In the closed 

 condition of the pit the thin marginal area of the limiting membrane 

 stretches elastically so as to fit the border exactly, if it is not already 

 wide enough for this purpose. When the pressures on the two sides of the 

 wall are equal or not greatly different, the torus presses quite loosely, 

 or not at all, against the orifice ; in this condition of the pit, diffusion 

 can proceed freely through the thin marginal area of the limiting mem- 

 brane, which is highly pervious, at any rate to water. 



What difference of pressure whether of water or of air is required 

 in order to induce the bordered pits to close, is not known with any 

 certainty. 1 ' 32 Russow, however, points out, that when fresh Coniferous 

 wood is allowed to dry, the pits invariably close, all the limiting mem- 

 branes shifting towards the low-pressure side. A difference of pressure 

 of less than one atmosphere is thus certainly effective. In this par- 

 ticular experiment the limiting membranes become pressed so tightly 

 against the border, that the tori assume the form of hemispherical pro- 

 tuberances projecting through the pit orifices. It has been calculated 

 by Pappenheim, on the basis of certain filtration experiments, that most 

 of the bordered pits in a piece of Coniferous wood will close under a 

 difference of pressure of 5 cm. of mercury (about -^r-.th of an atmo- 

 sphere). But for various reasons estimates of this kind cannot be 



