372 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 51 



If in such a case of general uniform temperature, we draw a system 

 of lines in such a manner that the logarithms of the atmospheric 

 pressure proceed by equal differences, then to a high degree of 

 approximation, the accelerations will be proportional to the dis- 

 tances apart of the neighboring lines. The lines would then also 

 be truly comparable with isohypsen, at least so long as the inclina- 

 tion of the surface in question is so slight that we may consider the 

 sine and tangent of the inclination angle as equivalent. 



From the preceding considerations it follows, that any conclu- 

 sions as to the acceleration effective at various points can only be 

 drawn with great care, even in the case of the ordinary charts of 

 isobars at the earth's surface. 



"The gradient is always inversely proportional to the distance 

 apart of the neighboring isobars; the gradient-acceleration is in 

 general greater in proportion as the pressure is smaller. If there- 

 fore we would adhere to the meaning ordinarily attached to the chart 

 of isobars, we must think of the isobars in the neighborhood of 

 the barometric pressure as being closer together than they really 

 are." 



If we draw isobaric charts for higher levels we incur danger of 

 drawing still more erroneous conclusions, unless we give special 

 attention. For instance, the chart for the altitude 5500 meters 5 

 should have its isobars drawn for every 2.5 millimeters, since only 

 then will its appearance justify general conclusions such as those 

 suggested by the chart drawn for the sea-level (on which the iso- 

 bars are drawn for every five millimeters). 



If from this point of view we consider the isobaric charts for alti- 

 tudes 5000 and 10,000 meters communicated by H. Hergesell to 

 the Met. Zeit. for January, 1900, we are surprised to find how enor- 

 mous the gradient acceleration is at these altitudes on these dates. 



Indeed Hergesell himself intended to indicate this point in that 

 on page 27 he said that at an altitude of 5000 meters the same differ- 

 ence of atmospheric pressure corresponds to about twice the gradi- 

 ent, whereas he evidently should have said that the same gradients 

 at sea-level and at this altitude produce a gradient acceleration at 

 the highest level twice that at the sea-level. 



It is evident from what precedes that the representation of the 

 distribution of atmospheric pressure by isobars on a level surface 

 of gravity has the important advantage of being easy to prepare and 

 of allowing of a comprehensive view of an unlimited extent of area 



5 See the table No. 1, Memoir XIV, p. 309 of this collection of translations. 



