II- 



TABLES 



FOR COMPUTING DIFFERENCES OF ELEVATION FROM BAROMETRICAL OBSERVATIONS. 



BY A. GUYOT. 



Tables which, like the preceding ones by Delcros, in metrical measures, are 

 sufficiently extensive to save the necessity of interpolations, relieve the computer of 

 most of his trouble, and considerably reduce the chances of error in the computa- 

 tions. They thus render to science itself a real service, by inducing observers to 

 determine a larger number of points, and to secure the accuracy of the results by 

 repeating their observations at the same point in various atmospheric circumstances, 

 both of which they can do without fear of being overwhelmed by the labor of 

 the computation. 



Similar tables are here offered to the observers who use instruments graduated to 

 English measures. Like those of Delcros, the new tables are based on Laplace's 

 formula, with a slight modification of only one constant. They dispense with the 

 use of logarithms, and give the differences of level corresponding to every thou- 

 sandth of an inch from 12 to 31 inches by means of the simplest arithmetical opera- 

 tions, so that the data being prepared and correcte(3, the computation of an elevation 

 takes but a few minutes, and is done with scarcely any chance of error. 



Laplace's formula and constants were adopted for the computation of the tables in 

 preference to others found in the following sets for reasons which a iew words will 

 explain. 



It has been remarked, page 9, that, in consequence of Laplace's constants having 

 been retained in Gauss's, Schmidt's, and Baily's formulse, they all give snnilar re- 

 sults ; but that Bessel's formula differs in separating the correction due to the moist- 

 ure of the air from that due to its temperature, while in Laplace's, and in the for- 

 mulae just mentioned, both are united. To introduce a separate correction for the 

 expansion of aqueous vapor is, in the writer's view, a doubtful improvement. The 

 laws of the distribution and transmission of moisture through the atmosphere are too 

 little known, and its amount, especially in mountain regions, is too variable, and 

 depends too much upon local winds and local condensation, to allow a reasonable 

 hope of obtaining the mean humidity of the layer of air between the two stations 

 by means of hygrometrical observations taken at each of them. These doubts are 

 confirmed by the experience of the author and of many other observers, which shows 

 that, on an average, Laplace's method works not only as well as the other, but 

 more uniformly well. At any rate, the gain, if there is any, is not clear enough to 

 compensate for the undesirable complication of the formula. 



Though the several co-efficients of Laplace's formula need perhaps to be modified 

 according to more recent and probably more accurate determinations of the physical 

 constants on which they depend, as has been proposed by Plantamour, E. Ritter, and 

 lately by the writer himself in a paper read before the American Association for the 

 Advancement of Science at their meeting in Montreal, they have been retained in 

 preparing the following tables, partly because it was found that the errors due to 



D 33 



