450 EXPERIMENTAL INVESTIGATIONS, 



being less rapid there than at other parts. This point would 

 then be worn away, and the stone would fall back a little ; 

 then fresh action upon new surfaces, another oscillation, and 

 so on. The effects which I have explained as taking place 

 by steps would, in fact, take place by insensible progression. 

 By assuming this process, unless there be some interfering 

 action, it becomes not improbable that the last point or line 

 worn away would be the point or line on which, from its 

 being in the line of centres of gravity, the upper stone would 

 rock. After seeing the great Logan-stone, near the Land's 

 End, I traced so many other approximations to rocking-stones 

 along that coast that I became satisfied, as far as one ought 

 to be satisfied on any subject of human enquiry, that this was 

 a correct theory. It then occurred to me, if this view be true, 

 may we not be able to hasten the operations of nature so as to 

 produce artificially (if such a term may be used) the rocking- 

 stone results ? A very little thought suggested the experiment. 

 Two parallelepipeds of iron, which had been made for keepers 

 of magnets, were taken, similar each to each, but that one was 

 twice the length of the other. The shorter was superposed on 

 the longer, and both immersed in sulphuric acid diluted with 

 three times its volume of water ; some nitric acid was added at 

 first, to hasten the corrosion. The liquid was changed from 

 time to time as it became nearly saturated, but without chang- 

 ing the position of the iron. At the end of three or four days 

 the pieces of iron were taken out, washed, and examined, when 

 the upper one was found to be a perfect analogue of a rocking- 

 stone, so delicately balanced on two points that it could be 

 made to rock by blowing on it with the mouth. (Result 

 shown.) It will be observed in this experiment that the iron 

 rocks only in one direction. Such is the case with the great 

 Logan-stone, and I believe with the greater number of rock- 

 ing-stones. It is obviously more probable that a stable equi- 

 librium should be obtained on two points than on one. I 

 have not yet got a specimen to rock or spin upon one point. 

 (Approximation to this shown by two zinc discs and explained.) 

 If the surfaces of the slabs be in such close contact that there 

 is not room for circulation of the saturated liquid, a formation 

 like those near the Cheesewring will be effected ; or if a num- 



