170 



SCIENCE. 



[Vol. I., No. 6. 



plain in that way is the Scandinavian. So far 

 as local conditions are concerned, he practicalh' 

 leaves the phenomena of England, Spain, Swit- 

 zerland, India, New Zealand, and the Atlantic 

 and Pacific coasts of North America, without 

 a plausible suggestion. His analj'sis of the 

 subject is, moreover, conspicuoush' incomplete 

 in that it omits all but the most casual mention 

 of ocean-currents. These great distributers 

 of climate are in continual conflict with the 

 elements dependent on latitude ; and an3' re- 

 modelling of coast-lines or sea-bottoms which 

 facilitates or impedes their circulation must 

 influence the local distribution and local mag- 

 nitude of glacial ice. While, therefore, his 

 presentation of the subject is interesting and 

 valuable, it is unsatisfaetor3^ It suggests a 

 line of inquir_y of great promise, but it falls 

 far short of a solution of the problem. 



The idea that a general elevation of atmos- 

 pheric temperature is more favorable to gla- 

 ciation than a general lowering, is one which 

 arises from an exaggerated appreciation of the 

 importance of precipitation as a condition of 

 glacier- formation. The existence of a glacier 

 shows that the local precipitation in the form 

 of snow exceeds the local ability of the pro- 

 cesses of evaporation and melting to dissipate 

 that snow in the course of the year : it shows 

 an excess of solid precipitation over dissipa- 

 tion. All will admit, that, if the local tempera- 

 ture be lowered without a concomitant change 

 in other conditions, the ice will increase ; and 

 vice versa. All will admit, too, that, if the 

 local precipitation be increased without modi- 

 fication of the other conditions, the ice will be 

 augmented ; and vice versa. That is to saj', 

 the amount of the ice depends on local tem- 

 , perature and local precipitation. If the gen- 

 eral temperature of the atmosphere be elevated 

 bj' a change in solar radiation, the local effect is 

 twofold : on one hand the local temperature 

 is raised, and on the other the local precipi- 

 tation is increased. The first change tends to 

 diminish the volume of ice ; the second, to in- 

 crease it. Whitne3''s proposition is, that the 

 latter tendency outweighs the former, and 

 the glacier grows : the majority of investigators 

 assume that the change of local temperature 

 is the more important, and that the glacier 

 shrinks. Considering the importance of this 

 question to his discussion, aud the all but uni- 

 versal prejudice against his view, it is surpris- 

 ing that he suffered the matter to rest with a 

 mere declaration of opinion, without attempt- 

 ing" a quantitative comparison. Let us en- 

 deavor to supplj' his omission. 



There is no comprehensive knowledge of the 



climate of auj- point where glacial ice now 

 actualh' accumulates ; but we fortunately' have 

 an excellent meteorologic record of a station 

 high in the Alps, where the conditions are 

 presumably' on the verge of glacier- formation, 

 and where the climate cannot be far different 

 from that of the surrounding ice-fields. More- 

 over, the observations at St. Bernard have 

 been so thoroughly- discussed bj- Plantamour, 

 AVolf, and others, that the material is in the 

 most available shape. Having for data a 

 mathematically' deduced annual curve of tem- 

 perature, and an annual curve of precipitation, 

 each based on the record for a long series of 

 j'ears, it is not difficult to introduce the h}'- 

 pothesis of a variation in general temperature, 

 and obtain an approximate quantitative indica- 

 tion of the effect of this variation on glacia- 

 tion. The mean temperature at St. Bernard 

 .is — 1.76° (C). Let us first assume that 

 through a variation in solar radiation this tem- 

 perature is raised 3°, and again that it is raised 

 6°; then that it is lowered 3°, and again 6°; 

 and let us inquire what effect theee variations 

 will have upon the snowfall. Evidently there 

 are two waj's iu which the snowfall is affected 

 by a general rise of temperature : first, the 

 fraction of the year during which precipitation 

 takes the solid form is diminished, so that the 

 snow forms a smaller percentage of the total 

 precipitation ; second, the change in tempera- 

 ture being general and not local, the power of 

 the atmosphere to receive and transport moist- 

 ure is increased, and the local precipitation 

 is therefore increased. If we note the day in 

 the spring when the curve of the annual oscil- 

 lation of temperature passes upward through 

 the freezing-point, and again the day in the 

 fall when it passes the same point in descend- 

 ing, we have the limits of the portion of the 

 3'ear during which all the precipitation is theo- 

 retically' fluid. (We are, of course, speakiug 

 of the ideal average j'ear : in anj' individual 

 3'ear there is a time of transition, with more 

 or less alternation of rain and snow.) Let us 

 call this period ' summer,' and the remainder 

 of the year, when precipitation takes tlie form 

 of snow, ' winter. ' Assuming that the form and 

 amplitude of the temperature curve remain 

 unchanged, while the mean temperature is va- 

 ried as by lij'pothesis, we can readily ascertain 

 the lengths of ' winter ' and ' summer ' for each 

 of the assumed cases. These have been com- 

 puted, and will be found in the subjoined table, 

 lines IV. and XII. We next ascertain, by 

 the aid of the precipitation curve, the amount 

 of precipitation during each of these periods 

 (V). 



