GEOPHYSICAL LABORATORY. 131 



fundamental hypotheses are correct. Manganese, eka- and dit;a-manganese 

 probably belong in the eighth group with iron and the rare earths, and not 

 in the seventh group with chlorine. Mercurous, gallons, and indous salts 

 are probably double complexes. So is the ferric ion. 



(7) Italian leucitic lavas as a source of potash. Henry S. Washington. Met. and Chem. 



Eng., 18, 65-71 (1918). 



This paper attempts an evaluation of the total amount of potash that is 

 present in the lavas of the six chief Italian volcanoes along the west coast that 

 have erupted leucitic lavas, which are therefore high in potash. The total 

 volume of the solid mass of each volcano was calculated, a base being assumed 

 and corrections made for craters and other features. An estimate was also 

 made of the ratio of the lava-flows to the total bulk of volcanic products, 

 largely tuffs; and the percentage of potash at each was averaged from num- 

 erous analyses made, for the most part, by the author. 



From the data it is concluded that the tot^l tonnage of potash (K2O) in 

 the lavas is about 10,000,000,000 tons; and this is regarded as a low estimate, 

 because the tuffs and ashes, which form at least 90 per cent of the volumes, 

 and which contain high i3ercentages of potash, have not been taken into 

 account. It is considered that in these volcanoes Italy possesses one of the 

 largest if not the largest of the visible suppUes of potash known to exist. 

 Some other silicate-rock sources of potash are briefly discussed, especially 

 the Leucite Hills in Wyoming and the belt of glauconite that extends from 

 New Jersey into Virginia. The latter is shown to contain about 2,034,000,000 

 metric tons of potash and to be a valuable source of supply. 



(8) Thermal leakage and calorimeter design. Walter P. White. J. Am. Chem. Soc, 40, 



379-393 (1918). 



The interchange of heat between a calorimeter and its environment (ther- 

 mal leakage) is practically proportional to their temperature difference, except 

 for the effect of evaporation and for that of convection, which is, for ordinary 

 calorimetric conditions, more nearly proportional to the square of that differ- 

 ence. 



If evaporation is suppressed the advantages of a constant thermal leakage 

 factor are obtained by preventing convection. Recent investigations upon 

 convection show how this may most advantageously be done. Diminishing 

 the width of the air-gap around the calorimeter diminishes convection very 

 rapidly, and this can be carried far enough to be effectual without too great 

 an increase of the total thermal leakage, which is then mainly due to con- 

 duction, and therefore increases about in inverse proportion to the gap width. 

 Gaps of from 1 cm. to 1.7 cm., according to circumstances, are best with 

 ordinary calorimeters. 



With large calorimeters, where the temperature change is less, freedom from 

 detrimental convection is compatible with gap widths greater than those 

 most desirable for small calorimeters. This possibility increases the relative 

 value of the large calorimeter. 



In adiabatic work there is little fear of convection, hence either very large 

 temperature intervals or very large air-gaps can be profitably employed. 



Very thin reflecting shields around the calorimeter may be used so as to 

 diminish conduction, and thus decrease the total thermal leakage without 

 increasing convection. 



Incidentally, it is pointed out that the ordinary rule, that thermometric lag 

 causes no error where only one thermometer is used, deserves careful interpreta- 

 tion, or else restatement, in the case of some thermochemical determinations. 



