174 CARNEGIE INSTITUTION OF WASHINGTON. 



cal methods and solved sufficiently to enable us to write down approximately 

 correct batches for the types of glass which were urgently required. A few 

 experimental melts then made it possible to correct for errors arising from 

 selective volatilization and the like, and thus to set up satisfactory batch 

 compositions. 



In this paper the several factors involved are presented chiefly in graphical 

 form; the dispersion relations alone are first considered, after which the 

 chemical characteristics and the relations between chemical composition and 

 optical constants are treated in summarized diagrams to indicate the methods 

 which we adopted and which enabled us to solve the problems in a practical 

 way in a short time. No consideration is given to the more fundamental 

 problem of computing the optical constants of a glass of given chemical com- 

 position. The information at hand was not adequate for this purpose, and 

 our war-time interest was not concerned with this problem, which still awaits 

 satisfactory solution. 



(26) Two gas collections from Mauna Loa. E. S. Shepherd. Bull. Hawaiian Volcano 



Observatory, S, 65-67 (1920). 



Through the courtesy of Dr. T. A. Jaggar, jr., of the Hawaiian Volcano Obser- 

 vatory, we received two tubes of gas which he collected in November 1919 

 from the flow on Mauna Loa. These are the first gases collected at this vol- 

 cano. Dr. Jaggar had great difficulty in finding a suitable source for col- 

 lecting, and greater difficulty in approaching it. The analysis of the gases 

 accounts for part of the difficulty, since they show 2 and 8 per cent of SO3 

 respectively. (The computation is made on the assumption that SO3 acts as 

 a perfect gas at 1200° C, which is not true; but this treatment allows com- 

 parison of this constituent with the others.) The combustible gases had prac- 

 tically all disappeared and the samples may be regarded as completely burned. 

 It will be noted, however, that the amount of nitrogen present is not high and 

 could not possibly account for the amounts of water present, namely, 67 and 

 75 per cent respectively. The argon group is present in rather larger amounts 

 than at Kjlauea, but amounts to a maximum of only 0.6 per cent. Free sul- 

 phur and chlorine were absent. Compared with similarly oxidized gases from 

 Kilauea, the Loa gas seems entirely similar in composition, as was perhaps to 

 be expected. 



(27) Methods of increasing the precision of thermostats. W. P. White. J. Wash. Acad. 



Sci., 10, 429-432 (1920). 



In mercury-contact thermostat regulators, the sensitiveness of the bulb 

 depends on dimensions and materials and on a sort of lost motion in the contact 

 itself, which nearly always varies between the limits lO^t and 40/^. The con- 

 stancy attained depends on this and also on the temperature lag of the bulb. 

 When high precision is sought, this temperature lag almost inevitably takes 

 the form of phase retardation and damping of the temperature- wave which 

 penetrates the bulb. On account of this phase retardation, it is usually im- 

 possible for a bulb, no matter how sensitive, to regulate to 0.001° in a uni- 

 form bath. If the surrounding temperature is also made very constant, the 

 variable heating can be cut down and the temperature lag thus diminished; 

 a precision even exceeding 0.001° can then be obtained. Two other ways of 

 getting such precision are: to use a Gouy oscillating-wire regulator, and to 

 put the regulator bulb close to the heater. This last method, however, 

 while it gives rapid oscillations and therefore great steadiness, may show large 

 variations with change in room-temperature or heating voltage. 



