March 1G, 18S3.] 



SCIENCE. 



171 



Computation of the relations of snowfall to melting and evaporation at St. Bernard, S'witsserland. 



I. Assumed general rise of temperature, in centigrade degr. — 6 



xn. 



XIII. 

 XIV. 



XV. 



XVI. 

 XVII. 



' Winter * begins 



' Winter ' ends 



Length of ' winter ' in days 



Precipitation during this period at tlie present time, in metres . . 



Mean temperature of ' winter ' 



Corresponding moan temperature over Atlantic ocean, near France 



Tension of s.ituration for temperatures VII. (millira.) 



Tension of saturation for temperatures VI. (millim.) 



Ratios of precipitation (Vlll. — IX.) 



Relative snowfall (V. X X. X .2122) 



Length of ' summer ' in days 



Mean temperature of ' summer' 



Relativeraeltng-power (XII. X XIII. X .1101.572) 



Mean annual temperature 



Corresponding tension of saturation, in mm. of barometric pressure 

 Comparative rate of evaporation (XVI. -^4.028) 



Comparative rate of dissipation (J XIV. + % XVII.) 



liatio of snowfall to snow dissipation (XI -^ XVIII.) 



Sept. 20.2 

 .June 3.9 

 235.7 



.9260 

 —7.74 

 +7.9 

 7.964 

 2.512 

 6.462 

 1.071 



Oct. 11.1 

 May 11.1 

 211.0 



164.0 

 +4.13 

 1.000 



—1.76 

 4.028 

 1.000 



Oct. 30.9 

 Apr. 19.9 

 170.0 



.5658 



—4.14 



+ 12.0 



10.457 



3.351 



7.106 



.853 



+1.24 

 5.025 

 1.247 



The air-ciirrents ■which cross the Alps, and 

 from which the precipitation at St. Bernard 

 is derived, acquire their moisture chieflj' from 

 the Atlantic ocean. The temperature over 

 the Atlantic being higher than on the Alps, the 

 air is there able to receive a larger portion of 

 moisture than it can retain in the Alps ; and 

 in a general waj- the precipitation on the Alps 

 ina3'TDe said to be due to this cause. It is true 

 that the air-currents traversing the Atlantic do 

 not become perfectly- saturated, and that on 

 'the waj' to tlie Alps the}' sometimes increase 

 their aqueous contents hy absorption from the 

 Mediterranean or from the land, and some- 

 times diminish it "by precipitation ; but the 

 only measure of Alpine precipitation available 

 for the present purpose is obtained b_y deduct- 

 ing the co-efficient of saturation corresponding 

 to the temperature on the Alps from the co- 

 efficient of saturation corresponding to the 

 temperature over the Atlantic. By ascertain- 

 ing this difference for the existing tempera- 

 tures, and again for the temperatures assumed 

 in the hypothetic cases, we are able to make a 

 comparison between the actual rate of pre- 

 cipitation and that which wouUl obtain if the 

 general temperature of the atmosphere were 

 raised or lowered. The annual procession of 

 temperature over the Atlantic ocean is not 

 accurately known ; but the tract of most im- 

 portance for the present purpose is that par- 

 HslXy surrounded hy England, France, and 

 Spain : and its temperature conditions are suffl- 

 cientlj' well determined by the observations in 

 these countries. '&y the aid of the isotherms 

 plotted for each month by the French bureau of 

 meteorology, the temperature of a definite por- 

 tion of this region has been deduced for each 



month of the j-ear. Line VT. of the table gives 

 the mean temperature of ' winter ' at St. Ber- 

 nard for each of the five cases. Line VII. 

 gives the mean temperature over the indicated 

 portion of the Atlantic for the same periods 

 and on the same assumptions. In lines VIII. 

 and IX. the maximum tension of aqueous 

 vapor in the atmosphere, expressed in millime- 

 tres of barometric pressure, is given for each 

 of these temperatures ; and the differences be- 

 tween these (X.) are taken as measures of 

 the relative rates of precipitation under the 

 various assumptions. Multiplying these rates 

 by the corresponding numbers of line V., we 

 obtain a series of numbers which measure the 

 relative snowfall under the several assump- 

 tions. (For convenience these numbers have 

 been multiplied bj- an arbitraiy constant, so 

 as to express them in terms of the present 

 precipitation as unit}-.) For example : in the 

 assumed case of a general temperature 6° 

 lower than the present, the length of ' winter ' 

 is 327.7 days. At the present time the total 

 precipitation in rain and snow during that 

 period is 1.157G metres; and in the assumed 

 case the whole of this precipitation would be 

 in the form of snow. This is notabl}^ greater 

 than the present snowfall, .7491 metres: but 

 the general rate of precipitation, affecting the 

 whole year alike, would be less than the pres- 

 ent in the ratio of 4.303 to 6.289 ; and these 

 two factors, tending in opposite directions, so 

 nearly neutralize each otlier that the total snow- 

 fall (XI.) in the assumed case differs b}' only 

 G per cent from the actual. 



The figures of line XI. show, for a thermo- 

 metric range of 12° (C), a variation of only 

 35 per cent in the snowfall, and indicate, that, 



