396 ALTITUDES DETERMINED IN ECUADOR, appendix. 



554 upon seventeen occasions, and the mean of the extreme differences 

 between the three instruments was now found to amount to 0*022 of an 

 inch. From comparison of the readings as a whole, I concluded that No. 554 

 had taken in some air, but I was unable to determine which of the other two 

 was most in error. 



It seemed desirable, upon return to London, to have all three baro- 

 meters verified by reference to some acknowledged standard other than Kew ; 

 and by the kind permission of Mr. Robert H. Scott, F.R.S., they were com- 

 pared, in September 1880, against the standard of the Meteorological Office. 

 The certificates given by Mr. Strachan ^ stated that at 30 inches the error of 

 No. 550 was + 0-017, of No. 554 + 0-030, and of No. 558 - 0*004 of an 

 inch. Hence it appeared that, since the verification at Kew, No. 554 had 

 acquired an error of 0*030 of an inch, that No. 550 had increased its error by 

 0*013 of an inch, and that No. 558 had decreased its error by 0*001 of an 

 inch. But as one can read on the verniers of these barometers only to 5^0 

 of an inch, the certificate for No. 558 really amounted to saying that there 

 was no alteration in the error ; and this was the more satisfactory inasmuch 

 as No. 558 was the barometer which had travelled everywhere, which haii 

 been employed upon all the summits, and against which the comparisons had 

 been made that are recorded in the subsequent sections of this appendix. 

 My indefatigable assistant, Jean-Antoine Carrel, carried this instrument 

 throughout the entire journey in the interior ; and to have preserved it 

 intact, without alteration of error, whilst executing the severe labour incident 

 upon our ascents, is, I believe, an unprecedented performance.^ I give here- 

 with reduced facsimiles of the certificate of Kew Observatory and of the 

 Meteorological Office relating to No. 558. 



The altitudes which have been deduced by Mr. Ellis from my obser- 

 vations of mercurial barometer have been calculated by Guyot's Tables 

 {Smithsonian Meteorological and Physical Tables) ; and, as he has used the 

 corresponding observations made by Mr. Chambers at Guayaquil, they gener- 

 ally differ to some extent from the results of my rough computations on the 

 spot, where the ' corresponding observations ' were unknown to me. 



The determinations which are most deserving of confidence are those 

 which are obtained from the means of a number of observations, and those 

 which are least reliable are naturally those which depend upon one observa- 

 tion only. In several instances, the exceptionally high temperature which 

 was observed at great elevations has caused me some perplexity. The most 

 extreme case was upon the summit of Antisana. At this great elevation, 

 surrounded by snow and ice in every direction for several miles, the tempera- 

 ture fluctuated during our stay from 44° to 60" Faht., though the sun was con- 



'' Who was neither acquainted with my recent comparisons nor with the certificates 

 previously given at Kew. 



2 The manner of packing, transporting, and suspending the two barometers which 

 were used in the field has been already mentioned. It may be added here that through- 

 out the whole time they were employed, whilst being moved from one place to another, 

 they were invariably carried reversed, i.e. with cisterns uppermost. Before they were 

 reversed, and packed in their cases, the mercury was driven to the tops of the tubes by 

 means of the cistern screws ; and, after reversal, the cistern screws were turned back 

 about half a turn, to allow a little play to the mercury. 



