ACADEMY OF SCIENCES] POWELL'S SURVEY 113 



there to turn eastward over the southern plateau to Moencopie on the line of the Echo Cliffs 

 monochne; another with a 4-mule supply wagon to Kanab; and a third, of which Gilbert was a 

 member, with pack train to set up signals on selected mountain tops, " the geodetic points of the 

 main chain of triangles," and thence to Kanab, Moencopie, San Francisco Mountain, and back 

 to Salt Lake City. Gilbert probably enjoyed this work, for it led him again over a region which 

 he knew fairly well from previous campaigns, and which richly repays many visits; and more- 

 over it called for an ingenious application of his mathematical capacity to field problems. But 

 it led to no published report. 



Lest some readers shoidd fail to apprehend the manner in which the practical execution of a 

 geodetic task departs from the theory of geodesy as presented in a textbook, an extract may be 

 made from one of Gilbert's later letters to Powell, written at Kanab, October 18, 1878, regarding 

 W , who was in charge of a side party 



I am afraid he is in trouble. He let go one of the men I sent with him and hired in his place a man whom 

 we know to be a horse thief just escaped from the sheriff. I have sent an Oraibe Indian to warn him, but fear 

 he will be too late. I start myself in the morning. . . . Shall write you from Moencopie. 



The promised letter from the last-named point was sent two weeks later : 



I have not yet seen W . He is not very far away and I have sent a scout after him. He lost no 



horses by the thief in his employ, but on the contrary, likedhim exceedingly and was much annoyed at my letter 

 of warning. 



The saying, "You never can tell," might have originated with this incident. 



BAROMETRIC HTPSOMETRY 



One of Gilbert's most aberrant interests concerned the principles involved in the determina- 

 tion of altitudes by means of barometric observations, a method which was much more exten- 

 sively used in the earlier years of governmental topographic surveys than it has been in later 

 years after the running of many lines of accurate levels and the geodetic determination of 

 many stations by the Coast and Geodetic Survey as well as by the Geological Survey in nearly 

 all parts of the country. Barometric measurement of altitudes was, however, an important 

 element in Gilbert's scheme of triangulation in the summer of 187S, as is shown by his mention 

 of drill in barometric observation in a letter above quoted. Indeed, the subject of barometric 

 hypsometry as a problem of atmospheric physics must have occupied his attention for several 

 years, judging from various entries in his notebooks. One of the most significant was made 

 during the first season on the Powell survey. He wrote, September 6, 1875, while in the high 

 plateaus: 



A barometric station on Musinia Plateau & another in Castle Valley would make a fine couple for an 

 hourly scientific set. At the same time a set of vertical reciprocal angles would test refraction & check the 

 determination of altitude. 



Three days later a supposed relation of the double daily curve of atmospheric pressure to 

 the expansion and compression of the lower air associated with the single daily curve of tem- 

 perature was outlined in a series of tersely stated interrogative propositions : 



While the temperature curve is convex upward, the horary [barometric] curve must be descending; while 

 it is concave, ascending (?). The horary curve is highest when the increase of rate of expansion is most rapid; 

 hence when the temperature curve is concave upward & has the smallest radius of curvature (?). Which of 

 these two propositions is correct? Is the horary curve a first, or a second differential of the temperature curve? 

 If the second proposition be true, then the radii of curvature of the two curves change signs through ra at the 

 same time. In any case the horary curve of the barometer is the true datum from which to determine the mean 

 temperature curve of the air column used in hypsometry. ... If one curve is a differential of another, then 

 the integral curve will have a max. or min. for every point in which the dif. crosses the axis & for no other 

 points. ... It is probable that the hypsometric (as distinguished from the superficial) maximum of temperature 

 occurs only 2 or 3 hours before sunset — or later than that directly indicated by thermometer. 



It is not likely that a similar series of propositions is to be found in the notebooks of any 

 other geologist on or off the survey staffs of those days or of any other days 



