Properties of Soils on Vegetation. 59 



brass plate, of the diameter of the eprouvette, and the same height, 

 provided with lids. Five or six of these could be brought into 

 the eprouvette. 



By this arrangement two sources of error could be eliminated : 

 the first was the absorption of heat by the solution of the salts 

 contained in water; and the other, and more important, the 

 heat produced when these porous substances are moistened. By 

 using turpentine instead of water, the former could be avoided 

 but not determined ; and it was thought that the other could be 

 got rid of by the boxes. Both errors were eliminated by a coun- 

 terexperiment with an equal weight of the same substance at 

 the same temperature as the water. With a weight of 20 grms. 

 the heating amounted to from o, 01 to O- 16 C, which in all 

 cases was deducted from the maximum calculated. 



In order to introduce the correction for the heat which the 

 calorimeter loses by contact with the surrounding air when the 

 substance has been placed in it, Pfaundler uses a method devised 

 by Regnault, which he details at length. 



He finally discusses the influence of various sources of error. 

 These consist (1) in the uncertainty of the weighings, (2) in the 

 measurement of the quantity of water, (3) in obtaining equality 

 of temperature in the interior of the eprouvette and in the 

 steam, (4) in reading off the barometric height (the sum of 

 these errors does not exceed O00083 of the total result) ; while 

 the fifth and most important source consists in the uncertainty 

 in reading off the thermometer in the calorimeter (this may 

 amount to 0*004 of the total result). But in a substance whose 

 specific heat is =0*2, the sum of these errors amounts, in the 

 most unfavourable case, to not quite a unit in the third decimal 

 place. 



There are still two other sources: these are, first those 

 arising from an inaccurate calibration of the thermometer, and 

 in the determination of the loss of heat in the calorimeter to the 

 surrounding medium. These were ascertained by a compara- 

 tive experiment with Iceland spar and water, by which numbers 

 were obtained quite agreeing with those of Regnault and of Kopp. 

 This agreement shows at the same time the accuracy of the 

 method, and its delicacy and convenience. 



From the results obtained, which are detailed in the paper, 

 Pfaundler draws the following conclusions : — 



The limits of the thermal capacity of soils lie between 0*19 

 and 0*50; and we may say that in round numbers it varies 

 between one-half and one-fifth of the specific heat of water. 

 Soils free from humus have the lowest specific heats ; and it is 

 immaterial whether they consist chiefly of silicates or of lime. 

 This is intelligible, considering that calcareous spar and rock- 



