296 SCIENTIFIC RECORD FOR 1884. 



is attaiiiou, even ii' slight coustaut errors remain; but in rainfall lueasiiies 

 even uniformity has never yet been obtained. Thus the new reorganiza- 

 tion of the official stations in Bavaria in 1878 has been productive- of an 

 abrupt change in the rainfall records, due to improved instruments and 

 exposures, by virtue of which some stations record twice as much rain 

 and others only I wo thirds as much as they formerly did. It is correctly 

 concluded on the basis of several investigations (Bache, Symons, Jevons, 

 &c.) that the greater part of this irregularity is due to the wind and its in- 

 fluence either on local distribution or on the instrument and its exposure. 

 To obviate this latter, rain-gauges have been established in pits or hol- 

 lows, so that their apertures are protected from severe wind, but none of 

 these are appropriate forthemeasuremeutof snow, which is easily whirled 

 out of the gauge unless the latter is furnished with a conical rim; but in 

 any case, however, the measurement of the snow is not of so great import- 

 ance, becauseithardly formsone-eighth of therainfall. [Thislatter, which 

 is true for most civilized countries, is, of course, not quite true for some 

 regions in which meteorology has an especial interest, namely, the ex- 

 treme north and south temperate and polar regions and those in which 

 glaciers have a tendency to form.] Lang's views as to the proper form 

 of a rain and snow gauge were originally about as follows: A conical ad- 

 dition as a cap to the gauge will give the wind that blows along its sur- 

 face an upward tendency, depending upon the strength of the wind, the 

 size and inclination of the conical surface; the air thus forced up will 

 carry with it a definite zone of particles, which will depend upon the 

 strength of the wind and the size of the surface. Since now, for gauges 

 whose mouihs have different sizes and are furnished with the conical 

 caps, the sloping surface of the latter increases as the diameter, but the 

 catching surface as the square of the diameter, it is evident that by 

 reason of this conical cap the ascending air currents will be of more im- 

 portance in iiroportion to the ratio between the surface of the cone and 

 the mouth, that is to say, in proportion as the gauge is of smaller dimen- 

 sions, for the current carries away from the mouth the suspended drops 

 of rain or flakes of snow. This source of error is unavoidable, but by a 

 proper selection of ratios can be made very small. Lang has, therefore, 

 chosen the following four forms for experimental comparisons: 



1. Of the same form as that used at the Bavarian stations since 1878, 

 namely, a cjliuder about 40 centimeters diameter, surmounted by a 

 cone whose vertical angle is about 60°, and so truncated that the aper- 

 ture or mouth of the gauge has a diameter 25.2 centimeters, or an area 

 of 2V of ^ square meter. 



2. A similar gauge whose cylinder has about 20 centimeters diameter, 

 and whose cap has an aperture of 10 centimeters, or ^ of a square 

 meter. 



3. A gauge similar to No. 1 and cylinder of the same size, but with a 

 conical cap tb truncated that the mouth has a diameter IG centimeters, 

 namelj% the same size as the mouth of No. 2. 



