212 



KNOWLEDGE. 



[November, 1901. 



To obtain a fixed point, from which the displacements 

 of the stone were to be measured, two rods, one of 

 copper and the other of iron, and each .86 of an inch 

 in diameter, were driven side by side into the ground 

 to a depth of about 8| feet, and the ends were cut off 

 so that they stood at first aboiit 3 inches above the 

 ground, or not quite ah inch above the upper surface 

 of tlie stone, which was placed with the rods projecting 

 through the central hole. 



Tlie measurements were made with the instniment 

 shown in Fig. 1. A is a brass ring, containing three 



FiQ. 1, — Instrument for 



Me.isuring tlie Movement of a Stone. 

 From the Proceedings af the Royal Society. 



rounded feet, B, which, at the time of reading, were 

 placed in the V-shaped giooves mentioned above. C is 

 a vertical brass rod, to which an ai'm D, with V-shaped 

 bearings is soldered. An ordinary micrometer screw 

 gauge, F, at the upper end of which trunnions, E, were 

 fixed, is placed with the trunnions resting in the V- 

 shaped bearings, D. At the beginning of each measure- 

 ment, the lower end, K, of the micrometer screw is 

 vertically above the top of one of the rods, G, H. The 

 screw is then turned until the end. K, just touches the 

 rod, and the reading is taken. By moving the tiimnious 

 along the V-shaped bearings, D, the screw is made to 

 touch the top of the other rod, and the reading repeated. 

 The object of the double measurements with rods of 

 different metals was to eliminate the small error arising 

 from the expansion of the rods ; but, in this respect, 

 the experiment was unsuccessful, and the measurements 

 given below ai-e only those obtained from the copper rod 

 uncorrected for expansion. 



Movements of the stone were produced by other 

 agencies than the burrowing of cai-th-worms. Some, but 

 an unknown, displacement may have been caused by 

 the growth of the roots of the chcstnut^tree, under 

 which the stone lay. But this, if it occuiTed, must 

 have been far less than that due to frost and changes 

 in the dampness of the ground. The stone, indeed, was 

 found to be in a state of continual vertical oscillation. 

 When the soil was damp from recent rain, one or two 

 cans of water poured on the ground near the stone 

 made it rise jTjth of an inch in about seven hours. During 

 a severe frost in the beginning of 1879, the stone rose 



^%ths of an inch ; and, during the thaw which followed a 

 slight frost a little later in the same year, it fell nearly 

 J^th of an inch in less than five hours, even though the 

 ground below was still hard. 



The oscillations of the stone due to the varying damp- 

 ness of the giound from Feb. 19 to Oct. 9. 1880, are 

 i-epresented by the continuous line in Fig. 2, the scale 



s5 



Fig. 2. 



-Diagram showing tbe Oscillations of the St«ne. 



being magnified so that each of the divisions on the 

 left-hand side of the diagram represents ^^tlis of an inch. 

 The dotted line in the same figure indicates roughly 

 the dampness of the soil, the ordinates being propor- 

 tional to the amount of rainfall diminished by that 

 which ran off or evaporated. § With two exceptions, 

 the cause of which is unknown, it will be seen that the 

 curves follow one another very closely, especially during 

 the latter part of the time. Between Sept. 7 and 19, 

 there was heavy rain, and in the interval the stone 

 rose ^ths of an inch. At the beginning of 1881, there was 

 a severe frost, and, owing to this and the previous wet 

 season, the stone stood Jth of an inch higher on Jan. 23 

 than on Sept. 7 of the previous year. 



Until the end of 1886, the position of the stone was 

 usuallv measured each wint«r and summer, and during 

 the next ten years in the summer only. The curve 

 obtained from the summer measurements is more 

 in-egular than from those made in winter, probably on 

 account of the greater variation in the dampness of the 

 soil in summer. During the last nine yeai-s in which 

 the experiment was made (1887-1896), the total descent, 

 given by summer measurements, was less than -gth of an 

 inch. It is difficult to account completely for this very 

 slow rate, but, during part of the time, the rainfall was 

 certainl}- above the average. In the earlier half of the 

 interval, however, the rate of descent was miich gi-eater : 

 though this may haji^e been partly due to the decaying of 

 the grass, for the stone was at first placed directly on the 

 turf. Taking the winter readings, Mr. Horace Darwin 

 found that the stone sank yVths of an inch in the eight 

 years from 1878 to 1886, or at the rate of Jths of an inch 

 in ten years. This, as we might expect, is much less 

 than the rate at which worms cover up small pebbles, 

 but it is about twice as great as that at which the large 

 stone examined by his father sank into the around. 



§ The rainfall was measured at Leaves Green, a place about one 

 mi'e from Down, and nearly on the same level. In estimating the 

 deduction Tor evaporation, the soil was supposed to dry at a uniform 



rate. 



