412 



SCIENCE 



[N. S. Vol. LII. No. 1348 



this because sea-level lies, in practically all 

 cases, below the station level, hence there is 

 no air column. The temperature argument 

 at present employed in reducing to sea-level 

 is the mean of the current surface tempera- 

 ture and the temperature as recorded at the 

 observation twelve hours before, and this 

 happens to do very well except in those regions 

 where the so-called air column is quite long. 

 The obvious solution of this difB.eulty is the 

 measurement of the mean temperature of a 

 real air-column above the station. 



The work referred to above embraces over 

 3,000 kite flights made at Mt. Weather, 

 Drexel, and Ellendale. The mean tempera- 

 ture between the surface and the levels of 1 

 and 2 kilometers above sea-level have been 

 classified by surface wind directions and by 

 months. These temperatures are not used as 

 they stand, but, to render them comparable, 

 they are each subtracted from the surface 

 temperature, thus giving a series of values 

 showing the difference between the mean tem- 

 perature of the air column and the surface 

 temperature. After doing this for the three 

 stations it was found that they all show 

 certain characteristics. For example, to 

 quote from the synopsis of the paper: 



It was found, in general, that in winter, with 

 southerly winds, the air column has a higher mean 

 temperature than the surface; that in summer, 

 with northerly winds, the air column has a tem- 

 perature below that of the surface. These effects 

 are due, primarily, to the seasonal variation of sur- 

 face temperature. The amplitude of the values 

 was much greater at the inland stations, Ellendale 

 and Drexel, than at Mount Weather, near the coast. 

 Aside from the geographical contrasts, the differ- 

 ence between the surface temperature and that of 

 the air column to a height of 1 or 2 kilometers de- 

 pends, so far as the surface factor is concerned, 

 mostly on the season; and, so far as the tempera- 

 tures aloft are involved, upon the wind direction. 



A statistical study of the data reveals the 

 fact that pressure can be reduced from the 

 station to levels one or two kilometers above 

 sea-level with a very satisfactory degree of 

 accuracy. Indeed, in reducing through an 

 air-column 2,000 meters in length, it is found 



that the probable error of the temperature 

 determination is so small that, when trans- 

 lated into terms of pressure at the upper level, 

 it amounts to only =t 1.3 mb. For the 1 kilo- 

 meter level it is only ±0:5 mb.^ 



When this study is carried further, as is 

 contemplated, all kite stations will be dis- 

 cussed and an endeavor will be made to work 

 the scheme out with such simplicity that the 

 observer can know the mean temperature of 

 the air column from the direction of the 

 surface wind, substitute this in his reduction 

 formula and obtain the pressure at the upper 

 level. 



The advantages of such charts must be 

 proved by experience. But it is obvious that 

 if they can be constructed accurately, they 

 must yield a far better basis for forecasting 

 winds aloft than do the present sea-level maps 

 — for it is in the very nature of the gTadient, 

 or geostropliic, wind to obey the pressure dis- 

 tribution at its level. The gradients on the 

 sea-level map can not be of great assistance, 

 even to the height of 500 meters above sea- 

 level. If such charts are thus productive of 

 greater accuracy in forecasting winds in the 

 upper air, their existence is quite justified; 

 and if, in the Plateau of western United 

 States, we can free our weather maps of the 

 barometric apparitions which haunt them at 

 present, it is possible that these charts of the 

 upper air may be of direct value in general 

 forecasting. 



The third paper to be mentioned is a short 

 note by Mr. Gregg-* on the program of aero- 

 logical work instituted by the Weather Bureau 

 for the hurricane season. It is believed that 

 there is a relation between the velocity of 

 motion of tropical hurricanes and the winds 

 aloft. To collect data, pilot balloon stations 

 have been placed at San Juan, P. P., and 

 Key West, Fla. These, cooperating with the 

 regular permanent pilot balloon stations in 



3 1,000 millibars is equivalent to a pressure indi- 

 cated by 750.1 millimeters, or 29.53 inches, of mer- 

 cury. 



* Gregg, W. R., ' ' Aerological Observations in 

 the West Indies, ' ' Monthly Weather Beview, May, 

 1920, p. 264. 



