358 M. Delcros' Barometrical Measurement [Nov, 



so little known, so little valued, so despised, may be employed 

 with advantage, and may come in place of the impracticable 

 system of geometrical levelling. 



The idea of employing the barometer for vertical topography 

 is neither mine nor is it new. All the philosophers who have 

 become practically acquainted with this instrument have recom- 

 mended it. De Saussure and De Luc may be considered as 

 its creators; after them, Ramond, Humboldt, Pictet, Prony, 

 De Buch, &c. have either applied it or brought it to the greatest, 

 degree of perfection. It is with the greatest diffidence that I 

 offer my observations after the immortal labours of these illus- 

 trious geologists. 



Notwithstanding the encouraging example of all these philo- 

 sophers, the use of the barometer is still very much neglected. 

 Mathematicians despise it without attempting to become practi- 

 cally acquainted with it. Naturalists, more disposed to adopt it, 

 are afraid of venturing to apply it, and generally neglect it. 

 I do not flatter myself that I shall prevail upon these different 

 classes of observers to make use of the barometer ; I shall merely 

 add a few facts to the mass already collected. 



Barometrical measurements may be obtained by two methods. 

 1 . By a system of corresponding observations made at the same 

 time. 2. By suites of successive observations not made at the 

 same time, but sufficiently near to answer the same purpose as if 

 they were contemporary. 



The first of these methods supposes corresponding observa- 

 tions made at points previously agreed upon, and with barometers 

 compared with each other. But it is often difficult to obtain 

 these two conditions, and they keep the observer absolutely 

 dependant on his base. In extensive measurements he is soon 

 too far from the coi responding points to be able to reckon on the 

 exactness of the comparisons. 



The second method may be of the greatest utility, in conse- 

 quence of its perfect independence, of its rapidity, and I may 

 say, likewise, in consequence of its exactness. This method, 

 like the former, supposes us to set out from a known base ; but 

 when once set out, we become independent of it. Let us figure 

 to ourselves the astonishing rapidity with which we can measure 

 the height of whole regions and of lines of immense extent, and 

 we shall be able to form a conception of this method. The want 

 of correspondence of successive observations will no doubt be 

 objected to me. I cannot deny this inconvenience ; but it 

 becomes very small if the observer abstains from observing when 

 obviously disturbing causes intervene ; and if he multiplies his 

 observations so as to have them at intervals of half an hour from 

 each ether, then his errors will be smaller than if he were 

 oblio-ed to f-mploy corresponding observations at the horizontal 

 distance of 20, 30, or 50 leagues. 



We can almost always combine these two methods, which will 



