122 EEPORT— 1881. 



At present the errors introduced by unknown inequalities of level are 

 probably nearly eliminated by tbe number of observations taken ; but it 

 could not fail to diminish the probable error of each observation, if a 

 correction were applied for this cause of disturbance from hour to hour, 

 or even from minute to minute. If the changes noted by M. Plantamour 

 are not entirely abnormal in amount, such corrections are certainly suffi- 

 cient to merit attention. 



In our first set of experiments we found that stone piers are exceed- 

 ingly sensitive to changes of temperature and to small stresses. Might it 

 not be worth while to plate the piers of astronomical instruments with 

 copper, and to swathe them with flannel ? We are not aware as to the 

 extent to which care is taken as to the drainage of the soil round the 

 piers, or as to the effect of the weight of the observer's body ; but we draw 

 attention to the effect produced by the percolation of water round the 

 basement, and to the impossibility we have found of taking our obser- 

 vations in the same room with the instrument. 



In connection with this subject we may notice an experiment which 

 was begun 3^ years ago by my brother Horace. The experiment was 

 undertaken in connection with my father's investigation of the geological 

 activity of earthworms, and the object was to determine the rate at which 

 stones are being buried in the ground in consequence of the excavations 

 of worms. 



The experiment is going on at Down, in Kent. The soil is stiff red 

 clay, containing many flints lying over the chalk. There are two stout 

 metal rods, one of iron and the other of copper. The ends were 

 sharpened and they were hammered down vertically into the soil of an 

 old grass field, and they are in contact with one another, or nearly so. 

 When they had penetrated 8 feet 6 inches it was found very difficult to 

 force them deeper, and it is probable that the ends are resting on a flint. 

 The ends were then cut off about three inches above the ground. 



A stone was obtained like a small grindstone, with a circular hole in 

 the middle. This stone was laid on the ground with the two metal rods 

 appearing through the hole. Three brass V grooves are leaded into the 

 upper surface of the stone, and a moveable tripod-stand with three 

 rounded legs can be placed on the stone, and is, of course, geometrically 

 fixed by the nature of its contact with the Vs. An arrangement with a 

 micrometer screw enables the observer to take contact measurements of 

 the position of the upper surface of the stone with regard to the rods. 

 The stone has always continued to fall, but during the first few months 

 the rate of fall was probably influenced by the decaying of the grass 

 underneath it. The general falling of the stone can only be gathered 

 from observations taken at many months apart, for it is found to be in a 

 state of continual vertical oscillation. 



The measurements are so delicate that the raising of the stone pro- 

 duced by one or two cans full of water poured on the ground can easily 

 be perceived. Between September 7 and 19, 1880, there was heavy rain, 

 and the stone stood 1*91 mm. higher at the latter date than at the former. 

 The effect of frost and the wet season combined is still more marked, for 

 on January 23, 1881, the stone was 4'12 mm. higher than it had been on 

 September 7, 1880. 



The prolonged drought of the present summer has had a great effect, 

 for between May 8 and June 29 the stone sank through 5-79 mm. The 

 opposite effects of drought and frost are well shown by the fact that on 



