AHrs.j PORCELAIN AIR THERMOMETRY. 217 
wen when of best workmanship they .are liable to leak after some 
sage. Moreover grease or vaseline is objectionable in consequence of 
he danger incurred of choking the capillary metal tubes. 
Experimental data. — The data obtained in the experiments with the 
oustant-pressure method are fully given in Tables 58 to (»0. It is to 
ie borne in mind that the results are obtained by a single observer 
ud that the chief object is to test the availability ot methods rather 
han to reduce the results to extreme fineness. The manometer is in 
ir, screened, however, from the furnace by a thick board and at some 
istance from it. v denotes the volume of the bulb, and t is the fidu- 
ial temperature for the day. The bulbs are of the re-entrant form, 
ot glazed internally. V\ and T 2 are the volume and the temperature, 
espectively, of the air measured in the manometer tube B G (Fig. 40). 
? is the current barometric height. Tis the temperature of the bulb 
,t the time specified, all observations being made in time series. 
Thermo electric data are given on the same plan as above, t is the 
emperature of the cold junction, e 20 the electro- motive torce when the 
old junction is 20° C. at the time specified. The figures of the last 
olumn refer to quantity of gas injected, as indicated by an arbitrary 
cale. 
In Table 58 T x was read off by but a single thermometer, and T is 
herefore less accurate. In Table 59 the thermo-couple had to be taken 
ut of the bulb during the stage of cooling, owing to an accident j the 
rror thus produced is quite perceptible. The results in the remaining 
ables were obtained without accident. 
The data as a whole (Tables 58 to 66) may be divided into two groups, 
n the first of these (Tables 58 to 61) the results are obtained with the 
e- entrant bulb No. 1, and the thermo-couple No. 37. These data are 
>robably less accurate than the more complete series of the second 
roup (Tables 62 to 66) in which the results are obtained with the re- 
ntrant bulb No. 2 and the thermo-couples Nos. 37, 38, 39. The greater 
bccuracy in the latter series of data is due to the fact that corrections 
ire duly applied for the permanent contraction or dilatation of the por- 
celain bulb from measurements made after each heating. The method 
)y which this may be done with extreme accuracy will be indicated 
>elow (p. 233). 
Tables 58 to 61 contain the four series of observations classified under 
jroup I. 
(871] 
