GEOPHYSICAL IvABORATORY. IO3 



a considerable simplification of the problem of accurate temperature meas- 

 urement, where intervals and not absolute temperatures are required, 



(22) Lag effects and other errors in calorimetry. Walter P. White. Phys. Rev. 31, 



562. 19 10. 



1. In calorimetry, the error from thermometric lag is rigorously zero for 

 the assumptions usually made as to the lag of the thermometer, as long as 

 one thermometer is used throughout. 



2. Other lags, in the water, in the metal parts of the calorimeter, and in 

 the surrounding air, are all proportional to the total temperature rise, are 

 wholly or largely independent of the rate of heating, and are, moreover, so 

 small as to be usually quite negligible in any case. 



3. The real sources of error lie in (i) temperature measurement; (2) 

 lack of uniformity in temperature, especially of the jacket; (3) variation in 

 the heat of stirring; (4) variation in evaporation. 



Of these, (i) is the most serious. An increase in the temperature in- 

 terval diminishes it proportionately, and such increase is at present one of 

 the most effective methods available for increasing accuracy. 



Complete avoidance of (2) requires absence of large projecting masses, 

 a calorimeter cover in contact with the water, thorough stirring of the jacket 

 water, and a complete inclosure by the jacket. 



Appreciable error from the heat produced by stirring (3) can always be 

 avoided. In some cases this may require a governor for the stirrer, but 

 usually it will not. The heat increases about as the cube of the speed, hence 

 an unnecessarily high speed is very detrimental. The error is probably less 

 for small calorimeters. The opinion that propellers give more heat than 

 reciprocating stirrers, for the same efficiency, needs re-examination. 



Evaporation from the calorimeter water (4) can be made regular within a 

 complete inclosure, if this is not too small. 



4. With proper precautions, the cooling correction, contrary to common 

 opinion, is not at present an appreciable source of error. Special methods 

 which aim to increase accuracy by diminishing the cooling correction are, 

 therefore, not likely to be very effective, though they may be convenient. 



(23) Some calorimetric apparatus. Walter P. White. Phys. Rev. Dec. 1910. (In press.) 



This paper deals with a calorimetric system in which all conditions affecting 

 the calorimeter temperature are, as far as possible, definite and measured and 

 in which, as a result, errors once serious are made negligible, computation is 

 often simplified, and an increase in effective thermometric accuracy can be 

 secured. The apparatus was designed for the determination of specific heats, 

 especially at high temperatures, by the mixtures method. 



1. Two calorimeters are described, designed for effectiveness in stirring. 

 In one of them the use of an inclined central tube affords a path both wide 

 and short for the circulating water, and at the same time allows the stirrer 

 itself to remain simple. Copper covers, carefully fitted and in contact with 

 the water, practically eliminate evaporation, both externally and internally. 

 When in their jacket, these calorimeters are surrounded above and on all 

 sides by a single compact body of circulating water and therefore have a 

 uniform environment. The calorimeter chamber can be opened at any time 

 without disturbing either the circulation or the connections to the calorimeter. 



2. Furnace : The internally-wound type of furnace commonly used in the 

 Geophysical Laboratory, modified to permit the dropping of bodies, was used 

 for most of the temperatures employed. At 1500° the temperature of the 

 charge was uniform to a small fraction of i per cent. 



