136 REPORT— 1881. 



ductivity for the Trap of Calton Hill scarcely distingnistable In its value 

 from that fonnd by Sir W. Thomson.' 



Transforming these determinations to centimetres, grammes, and 

 seconds from the units of the Paris-foot and year in which they are 

 expressed, the corresponding values of the absolute thermal coefficients 

 so obtained for Calton Hill trap-rock are arranged with those obtained 

 from the Committee's experiments, in the accompanying general list. 

 Similar values in the same list denote also, in comparison with the Com- 

 mittee's observations, the absolute conductivities (by unit- weight of water 

 and by unit- volume of the rocks) found by the two independent methods 

 for the sandstone of Craigleith quarry and for the sand of the experi- 

 mental garden, near Edinburgh, where the other thermometers were 

 planted. 



The dii'ectly measured values are in general, in these cases, in such 

 very fair agreement with the thermal conductivities found by computa- 

 tions, that the verifications which they afford of those results are also a 

 sjabstantial confirmation of the theory upon which they are founded. 

 Angstrom's reductions of the 10 feet deep Upsala earth-thermometers in 

 moist clay afford the same conclusion ; and indeed the agreement pre- 

 sented by the series of Despretz's, Helmersen's, and Neumann's observa- 

 tions, and especially by Messrs. Ayrton and Perry's experiment on a 

 sphere of trachyte, made by methods based on Fourier's theoiy, with the 

 direct results of various forms of Peclet's method used in the other 

 determinations of the table, plentifully attest and abundantly substantiate 

 the same physical agreement. 



There are, indeed, as great difficulties in determining thermal 

 conductivities from earth-thermometer observations, as in successful 

 applications of the processes of direct experiment. For while in the latter 

 artificial temperature and heat-supplies are difficult to keep constant and 

 to measure accurately, in the former case the natural periodical variations 

 which furnish observational results for computation, are uncertain in 

 their separate and combined intensities and efi'ects from year to year. It is 

 only on the average of several years' records that the annual, semi-annual, 

 and other quicker oscillations of temperature at the earth's surface can be 

 regarded as sufficiently constant for a single average period to be severally 

 followed downwards by calculation, as individual heat-waves among the 

 deep- sunk thermometers. In this case the same rate of decrease per foot 

 of logarithmic amplitude, and of angular lead of the maximum or minimum 

 phase of the descending heat-wave, would be found for all the annual, semi- 

 annual, and other quicker portions of the oscillation at the surface. The 

 same value of the earth's conductivity, at any place, would be found from 

 all these periods. But the varying character, at any place, of the yearly 

 oscillations of temperature at the earth's surface from year to year, pre- 

 cludes the possibility in general of such perfectly accordant calculations, 

 for records not extending over more than one or two years. Reductions, 

 like Sir W. Thomson's and Professor Everett's, for records extending over 

 periods of seventeen or eighteen years, show very close approximations to 

 it. But in briefer periods the efiect on range and retardation near the 

 surface, produced by half-yearly and other faster oscillations. Is not always 

 found to be compatible with that attributable to the yearly one, from 

 secular changes of greater period than either being accidentally combined 



' Transactions of the Royal Society of Edmburgh, vol. xxii, p. 437, 



