VEGA -EXPEDITIONENS VETENSKAPLIGA AP BETEN. 273 



at — 0°.25 c. Still more easily this contraction may be followed 

 by the deflectioa of the curve III (marked with red) on plate 22. 

 I next proceeded to the examination of entirely pure 

 water, prepared b}^ repeated distillation of water from the well 

 of St. Eric in Upsala, which according to the analyses of Prof. 

 Almén gives the purest drinking water in Sweden. Two sepa- 

 rate experiments were made with such water. The volumes 

 of the ice are found in tables I & II and are graphically re- 

 presented in plate ' 22 by the red ciirves I & II. These meas- 

 "urements were taken with two different dilatometers, biit this 

 circumstance will hardly suffice to account for the marked 

 discrepanc}' of the curves of volumes. The real cause seems 

 to be, that in experiment II the water was boiled for more 

 than an hour by the operation of filling the instrument and 

 therefore has dissolved a trace of the substance of the glass 

 reservoir. In experiment I, which I undertook last of all, I 

 purposel}" restricted the boiling operation to the least possible 

 time. Still the inflection of curve I on plate 21 & 22 clearly 

 shows, that the purest water, which can be employed in dila- 

 tometric experiments, also contracts its volume before melting. 



». . . . Bei Abnahme der Temperatur fand stets eine Ausdehnung und 

 bei Zunahme der Temperatur stets eine Zusammenziehung statt, und zwar 

 ergab sich, wenn die Längenausdehnung bei — 1° R = 1 gesetzt wird, 



die lineare Ausdehnung im mittel fiir Erkältung um 1' R 

 in d. 2ten Versachsreilie, bei Erlcält. v. — 2' R bis — 7\5 R zu O.ooosc . . 

 » » Sten » » » » » — 2° » » — 7° j> » 0.ooo4o . . 



» » 4ten » » » » » — 2^ » » — 8' » » O.ooos?.. 



etc. . .» [Beiträge zur Geognosie von T_yrol, Leipz. 1843]. 



If Petzholdt had employed ice from salt water in his experiments, 

 we would find no difficulty in explaining the negative coefticient of expansion. 

 Tn chapter 4 we will meet with some specimens of sea ice, which show 

 verv considerable contractions, when cooled. But as he asserts, that the ice 

 of his experiments was formed from pure water (»luftfreies destillirtes 

 Wasser»), which does not show any abnormal contraction l»elow — 0^3 C, 

 we must admit the justness of the remark of Brun ner, who in 1845 re- 

 peated the experiments of Petzholdt and found the coefticient of expan- 

 .sion of ice to be 2^osit.ive and greater than that of any other solid : 



»TI résulte donc de mes expériences, que la glace se contracfe j)ar le 

 froid, ainsi que les autres corps solides, ce qui est directement opposé aux 

 observations de Mr. Petzholdt» (Bibliothéque universelle de Geneve T. LVI). 



Notwithstanding this, some of the observations of Petzholdt on gla- 

 ciers highly merit our attention. If we confine the application of his theory 

 to those limits, where a contraction bj^ heat really takes place in the glacier- 

 ice, i. e. to femperafni-es in the i)iimediafe vicinity of zero, we certainly can 

 not deny, that this phenomonon must have something to do with the move- 

 ment of the glaciers. 



' For the sake of distinctness the curve II is left out in plate 21. 



18 



