as applied to the measurement of Heights. 345 



110 D 3^55' 61.3 29.72 —02 



111 — 4.00 61.4 29.74 +02 



From Obs. 97 to 105, the thermometer falls, and the sym*. 

 piesometer, with one exception, uniformly rises. The fall of 

 .01 at Obs. 106 is caused by a change indicated by the fall of 

 no less than 0°.6 of the thermometer at the next observation. 

 The descent of the oil between Obs. 107 and 108 is succeeded 

 in Obs. 109 by a stationary thermometer, which in the course 

 of falling is equivalent to an initial rise. The ascent of the oil 

 in Obs. 109 is followed by the descent of mercury in Obs. 110, 

 and the renewed fall of the former in Obs. 110 is produced 

 by the cause which is indicated in the contrary motion of the 

 mercury in Obs. Ill, so that this curious oscillation is preserv- 

 ed five times ; the motion of the sympiesometer always preced- 

 ing that of the thermometer, as it ought to do according to 

 theory. 



Regarding the very erroneous results of height which many 

 of the preceding tables give, I own I feel it difficult to give 

 an explanation, more especially as several trials I made in 

 spring agreed far better with the actual height, when the dif- 

 ferences of the attached thermometer were of course greater 

 My experiments are too limited to draw positive conclusions 

 on this point, which is not the one I aim at ; and I shall con- 

 tent myself with noticing two possible sources of error. That 

 the instrument (which I believe is graduated wholly experi- 

 mentally) may not have been subjected to sufficient degrees 

 of heat, which in these experiments was so considerable ; and 

 that the situation of the experiment was such as to render it 

 trying to barometric measurement, being a ravine where pro- 

 bably very various currents prevail, particularly in warm wea- 

 ther. I hope at a future time to continue my experiments on 

 the use of the sympiesometer, and in the meantime to excite 

 some attention to the subject. 



I promised, before concluding, to give some hints for the re- 

 moval of the defect I suspected, and as my paper is now longer 

 than I intended, I shall do so in a few sentences. Since the 

 the main point is, that the thermometer should indicate the 

 actual temperature ef the hydrogen gas, I should recommend 



