JoT.Y — Some JExpermenfs on Demidation. 31 



Appendix. 



A brief account of the apparatus used in the experiments on the 

 solvent denudation of basalt (coarse grain) in fresh and salt water may 

 be of value to anyone enteiing on such experiments. The arrange- 

 ment is such as to utilize the motion of a continuous water supply 

 from any source to produce a reciprocating passage of a given quantity 

 of a liquid through a U-tube containing the substance being dealt 

 with. 



In the diagram for clearness one U -tube only is shown, X, contain- 

 ing the basalt. In the actual experiments there were two U -tubes 

 attached side by side so as to be under like conditions of temperatui'e, 

 both containing basalt of same assortment of grain ; but thi'ough the 

 one salt water, thi'ough the other fresh water cii-culated. It will thus 

 be understood that the flasks containing these solvents, F^ and F2, were 

 four in number, the diagram showing those requii-ed for the one solvent 

 only. Beyond them those for the other solvent may be supposed 

 concealed. Similarly, behind X the second U-tube is concealed. The 

 tube £ biiui'cates at h, one branch ascending to the top of F-^ as 

 shown, the other ascending to the top of the flask concealed behind i^i. 



Kt S is a stop-cock controlling the city water supply. A stout 

 rubber connection, closed all but for a nearly capillary glass tube, 

 admits from this a continuous small stream of water at high pressure 

 into the tube A. It is thus conveyed in a slow continuous stream 

 into the closed "WoH's bottle placed above the tap and above the flasks. 

 If we imagine the Wolf's bottle just full of air, and water flowing 

 into it from the tap, this air will escape by the second tube B into 

 Fx and at first pass through the mineral particles in X, and escape 

 through the measured quantity of solvent in F., emerging by the 

 damping tubes T. When the Wolf's bottle gets qiiite full the siphon 

 C comes into operation and rapidly empties the bottle, Uke a cup-of- 

 Tantalus, the siphon being in fact of sufficient bore to empty the 

 bottle ia about eight minutes, although the stream is entering by tube 

 A all the time. During this emptying process evidently the solvent 

 in is is sucked up and passes through X, rising finally into F^. When 

 all is nearly drawn up the Wolf's bottle is quite empty. The siphon 

 breaks, and the cun-ent from A gradually refiUs the Wolf's bottle, 

 during which time the solvent flows back through X. This takes 

 about nine minutes. 



There is thus a tide upwards and downwards maintaiaed through 



