420 UPON BOILING-POINT OBSERVATIONS. appendix. 



includes thirteen stations at greater heights than 18,000 feet, the loftiest 

 being the camp close to the summit of Cotopaxi (19,478 feet), the summit of 

 Antisana (19,335 feet), and the summit of Cayambe (19,186 feet). The 

 mean height of these three positions as deduced from observations of mer- 

 curial barometer is 19,333 feet. Five, four, and two thermometers respect- 

 ively were boiled at them, and the mean error of the final results as deduced 

 from the boiling-points at the three grouped stations is — 513 feet. 



In comparing this unsatisfactory record with the observations made by 

 Dr. Hooker, I find that the majority of his boiling-point results showed con- 

 siderable minus errors ; and that at his loftiest station (18,466 feet) the 

 error was — 600 feet. 



Upon examining the observations of the brothers Schlagintweit, * I .find 

 that their boiling-points were always higher than the simultaneously-observed 

 barometers would lead one to expect (and, so far, they are in harmony with 

 the observations by Dr. Hooker and myself). Their errors were, however, 

 small as compared with ours, and their results as a whole are, I think, more 

 harmonious than it is reasonable to expect can be obtained by a method 

 which includes so many possibilities of error. 



The information afforded by M. Wisse is not so copious as one could 

 desire. The observations which are quoted in Aniiales de Chimie et de 

 Physique were made by him at twenty-six different places between June 

 1844 and May 1849, but only five of his stations were higher than 10,000 

 feet above the sea, and only two of these reached 15,000 feet. It is not said 

 whether the recorded boiling-points were single observations or the means of 

 several ; nor is it stated whether the recorded observations are the whole 

 which were made, or are only selected instances in which the boiling-point 

 observations closely accorded with the barometric ones. From their remark- 

 able accordance with each other it seems not improbable that this was the 

 case. 



From examination and comparison of these observations as a whole it 

 would seem by no means certain that the barometric pressure corresponding 

 to any given boiling-point has been well ascertained ; and it would appear 

 desirable that more comprehensive and thorough investigations in this direc- 

 tion should be made, if it should continue to be the practice to attempt to 

 determine the heights of lofty positions by observation of the boiling-point 

 of water. 



But should such observations be made as would determine absolutely the 

 boiling-point of water corresponding to every inch of the barometer, the 

 fundamental objection to the use of this method would not be overcome, 

 namely, that it is a cumbrous procedure, a method built up upon another, 

 which must always (as is stated in the remarks already quoted from the Smith- 

 sonian Tables) be " liahle to all the chances of error which may affect measure- 

 ments by means of the barometer, besides adding to them new ones peculiar to 

 itself.^'' This objection applies (although not with equal force) to all observa- 

 tions made by this method, in every region and at any height. 



^ It should be noted that the Schlagintweits employed thermometers 21 inches long. 

 This probably permitted greater refinement of observation. 



