VEGA -EXPEDITIONENS VETENSKAPLIGA APBETEN. 285 



Einfluss die Temperatur hat, dafur will ieh niir einige Data 

 aus meinen Versuchen anfiibren.» 



»Derselbe Eisencylinder sank unter demselben Drucke bei 

 einer Temperatui- zwichen — 4° und — 1"^ in 12 Stunden um 

 Vli m.m. tief ein, während bei einer zwischen — 6° und — 12^ 

 scliwankenden Temperatur in 5 Tagen bei einem Drucke von 

 5 Atmosphären das Einsinken nur 1 m.m. betrug, also in 12 

 Stunden nur 'i, o m.m.-» 



»Steigt die Temperatur der Umgebung tiber den Schmelz- 

 punkt, so wird die Weicliheit des Eises so gross, dass in einer 

 Stunde schon derselbe Eisencylinder 3 cm. tief unter dem 

 gleicli geringen Drucke sich senkte, obwohl er vollständig von 

 Schnee eingehiillt war, um die Erhöhung der Temperatur des 

 Eisencylinders iiber Null zu vermeiden.» 



I think that these observations of Person, Pfaff a. O. 

 concern the same change in the molecular constitution of 

 ihe ice, ivhich in my dilatometric experiments appears as a shrinl'- 

 ing of its volume, and I also think, that this contraction of 

 volume, which we may study by regular observations for every 

 tenth or hundredth of a centigrade degree, affords us the best 

 measurement ^ of the progress of the weakening of the ice. It 

 will also appear from the above experiments and likevvise from 

 the determinations in the next chapter on ice from salt water, 

 that this emollescence most probably is no inherent property 

 of the Chemical compound Ho O hut is caused by the influence 

 ■of slight quantities of foreign substances, salts, etc, the pres- 

 •ence of which has the power of modifying the whole melting- 

 process astoundingl3^ 



The process of thmving may thus be considered alivays to begin 

 ielotv sero. Every kind of ice, which occurs in nature, will at 

 ■a sufficiently low temperature be härd and brittle. In this 

 State, the ice, if tested, will also be found to expand regidarly. 

 The commencement of the thawing or emollescence of the ice 

 is entirely dependent upon the amount of salt etc it contains. 

 Volumetrically the softening of the ice is announced by a 

 rgradual diminution of the coefficient of expansion and finally 

 by a real contraction of its volume. The tables in this and 

 ihe following chapter, together with the curves on plate 21 will 



' As a good example of the congruent results of wiclely different meth- 

 ods we may point out the fäet, that Person could not detect any signs 

 ■of softening in pure ice below — T C. Now a glance at the extract from 

 table I on page 274 will show us, that the regular coefficient of expansion 

 of volume begins to diminish just hefore — 5' C. 



