254 Mr. H. Darwin. On the Small Vertical Movements 



form of geometrical bearings for the three rounded feet of the stand 

 which carried the micrometer used for measuring the relative positions 

 of the stone and the top of rod. 



The action of the earthworms would cause the stone to sink rela- 

 tively to the top of the rod, but the following other causes should also 

 be considered : — 



1. The Groioth of the Roots of the Tree. — The copper rod passed 

 through about 2*63 metres of slightly sandy red clay which overlies 

 the chalk, and contains many flints ; some of these were broken or 

 displaced by the passage of the iron rod. Great force was required to 

 draw the rod out of the ground, and in doing so its sides became 

 scored by the flints. It is, therefore, safe to assume that the flints 

 were pressed with considerable force against the rod, and that their 

 sharp edges gripped it tightly. The point where the rod was gripped, 

 and where there was no relative movement between it and the clay, was 

 unknown ; probably, however, it was well below the level of the roots 

 of the tree. The roots growing larger in diameter would raise the 

 stone relatively to the top of the rod. The amount of this movement 

 is quite uncertain. 



2. Dampness of the Ground. — The clay and the surface soil both, no 

 doubt, swell with increase of moisture. The swelling of the clay 

 above the unknown point at which the rod is gripped will raise the 

 stone, and the swelling of the surface soil will have the same effect. 



3. Expansion of the Rod from Change of Temperature. — The effect of 

 this is very small and is quite negligible when measurements, taken at 

 the same time of year, are compared. If we take a high estimate 

 and assume that the summer and winter temperature of the rod 

 differed by 10° C, the relative movement of the stone and the top 

 of the rod would be about 0'4 mm. : this is on the assumption that 

 the rod is only gripped close to its lower end, and that the expansion 

 practically of its whole length is taken into account. An attempt 

 was made to eliminate this error by sinking two rods alongside of 

 each other, one being of iron and one of copper, and by taking 

 measurements from both rods. This attempt failed, and the results 

 now given are the measurements from the copper rod only.* 



The measuring apparatus is shown in fig. 1. It consists of a brass 

 ring A, with three short rounded feet B, which rest in the radial 

 grooves before mentioned. This annular base carries a vertical brass 

 rod C, to which is soldered an arm with V-bearings D. Trunnions E 

 were fixed to the usual form of micrometer screw gauge as shown in 

 the figure, the trunnions were supported by the V-bearings in the arm, 



* Professor Judd pointed out that the clay with flints through which the rod 

 passed probably contained small quantities of calcium carbonate which would be 

 slowly dissolved by rain, and that this would produce a small error. — May 2, 1901. 

 H.D. 



