GANKETT.] 



GEOGRAPHICAL WORK. 



459 



ject to but very small errors of gradient. That the results are as good 

 as could be expected from barometric work, the following table, com- 

 paring measurements made in 1878 and in 1872, when the conditions 

 were not as favorable, will show : 



1872. 



1878. 



Electric Peak, west summit . 



Mount Washburn 



Camp on Elk Creek 



Camp at Mud Geysora 



Tellowstone Lake 



Camp in Lower Geyser Basin 



Slioshone Lake . . . .' 



Lewis Lake 



Ked Mountain 



Mount Sheridan 



Porks of Snake River , 



North Twin Butte 



Feet. 



Feet. 



10, 997 



11, 100 



10, 388 



10, 346 



6,236 



6,436 



7,712 



7,725 



7,788 



7,738 



7,252 



7,236 



7,870 



7,830 



7,757 



7,800 



9,806 



9,777 



10, 343 



10,385 



6,892 



6,786 



7,962 



7,976 



These discrepancies, which, to one accustomed to deal only with re- 

 sults by level, would appear enormous, are, in reality, quite within the 

 range of ordinary barometric work in a mountainous region. Discrep- 

 ancies of 200 feet between single results are not unfrequently met with, 

 and a. close accordance among three, four, or five results is a most rare 

 accident, and suggests to one acquainted with such work anything but 

 great accuracy in manipulation and computation. The best baromet- 

 rical observer finds it impossible to set and read his instrument with 

 certainty closer than .006 or .008 of an inch. Add to this the defects in 

 the barometric formula and constants, which produce periodic errors of 

 great magnitude, and which, though they can be to some extent avoided, 

 cannot be wholly eliminated, and the great accidental errors arising 

 from differences in barometric gradient, and it is easy to see that the 

 barometer has no claims and can have none to being an accurate instru- 

 ment. All that can be claimed is that, when used to the best possible 

 advantage, it will give results approximating to the truth. To illus- 

 trate the range of results given by this instrument, I add, in the follow- 

 ing lists, the measurements of a few points in the region under consid- 

 eration, as made by several different observers, at different times and 

 under different conditions. 



The measurements made by Dr. Hayden's first expedition, in 1871, 

 were made with the siphon barometer, and were computed by the aid of 

 an assumed base of 30 inches for the barometer at the sea-level. The 

 discrepancies between these results and my own, amounting, in some 

 cases, to several hundreds of feet, are due to the faulty method of com- 

 putation, as I demonstrated in my report on the work of 1872. 



The large differences between my own results and those of Captain 

 Jones, which, as will be noticed, average about 200 feet, and are quite 

 constant, his results being the lowest, are not as easily accounted for. 

 But as my results during the two years 1872 and 1878 are, within fair 

 limits of error, the same, and as they remain the same, whether com- 

 puted from Virginia <Jity, from Salt Lake City directly, or from Chej''- 

 enne, I cannot believe that I have a constant errov. Moreover, the re- 

 sults of Messrs. F. A. Clark, George B. Chittenden, and G. R. Bechler, 

 who have, under this organization, during the past two years, carried 

 on topograpliical surveys embracing parts of the region traversed by 

 Captain Jones, show the same discrepancies between elevations of 

 points determined by them and by Captain Jones. It seems fair to 

 assume, therefore, that the burden of proof lies with the latter. 



