50 



Hjælp af dette Apparat, som ved et Stykke Kautschiikrør 

 var forbundet med et lidet Reservoir, bestemt til Optagelse 

 af den ved mulig Opvarmning udskudte Del af Indholdet. 

 bestemtes nu Udvidelsescoefficienten af det anvendte Glas- 

 ror med renset Kviksolv. som nogen Tid før Forsøgenes 

 Ddførelse under Udkogning paafyldtes Pyknometret. Med 

 dette Apparat som jeg vil betegne som Pykuometer No. 3, 

 udførtes til den Ende følgende Veininger. 



Alle disse Veininger ere udforte elter Substitutions- 

 methoden ved Aflæsning af Svingninger paa en Vægt, 

 hvis Følsomhed udeu Belastning beløb sig til 1.4 Mgr. pr. 

 Delstreg og for stigende Belastning temmelig jevnt aftog 

 indtil 1.9 Mgr. ved 200 Gr. Belastning. Til disse saavel 

 som alle finere Veininger. som jeg Lar foretaget i Anled- 

 ning al' disse Arbeider, benyttedes en Platina lodsats fra 

 Deleuil i Paris, hvis (Jorrectioner jeg i Forveien Lavde be- 

 stemt ved flere vel overensstemmende Veininger paa en af 

 P. Bunge forarbeidet fortrinlig Vægt, hvis Følsomhed for 

 dr lier omliandlede Belastninger beløb sig til omkring 0.14 

 Mgr. Ved Veiningerne iagttoges altid Temperatur og Ba- 

 rometerjstand af Luften i Veieværelset, hvorimod dens Fug- 

 tigliedsgrad i Mangel af Observation passende ansattes, en 

 Mangel, der ikke bar nogen mærkbar Indflydelse paa For- 

 søgenes Paalidelighed, da selv en saa grov Feil i Ansæt- 

 telsen af Luftens relative Fugtiglied som 25 °/ under de 

 Ler omhandlede ForLold kun virker paa (h-t endelige Resultat 

 med en liden Feil i tite Decimal. Eiter disse Data er Luf- 

 tens specifiske Vægt under Veiningerne paa vanlig Maade 

 beregnet og opført i Tabellen. Pyknometret blev lor hver 

 særskilt opført Veining indstillet paany for Temperaturen 

 <i" i finstødt Is og forøvrigt i Vandbad, Lvis Temperatur 

 under stadig Omrøring holdtes constant, ligesom det raindst 

 en Gang for hver anden Veining tomtes og fyldtes igjen 

 med del Søvand, hvis Volum skulde bestemmes. Ingen 

 [ndstilling toges for god. med mindre det Lykkedes i mindst 

 h") Minutter at holde Temperaturen saa constant. at Ther- 

 inoiui-tret ingensinde viste Variationer af 0.°1 eller derover. 

 Til Brug ved Aflæsning at' Vandbadets Temperatur tjente 



absolutely unavoidable. With tliis apparatus, which was 

 connected by a caoutschoiic tube with a small receiver for 

 collecting any portion of the contents that. in the event of 

 the instrument becoming warmer might possibly be expelled. 

 the coeflicient of expansion of the glass tube was now 

 determined by means of puritied mercury, which, shortly 

 before the commencement of. the experiment. bad. when 

 boiling out the air. been introduced into the pycnometer. 

 JVitL tliis apparatus. which I will designate 'Pycnometer 

 No. 3,' were performed the following determinations of 

 weight: — 



Weight In Ail- 

 in with a Sp. 

 Grammes. tir. el' 



30. 

 3'- 



32- 

 33- 

 34- 

 35- 

 36. 

 37- 

 38. 

 39- 

 40. 



10.8654 

 ! 10.8653 

 with mercury of o" 195.9265 

 ,, o 1195.9265 

 •• 20 195.3588 

 •• 20 195.3592 



15 195-4993 



,5 o 195.9276 



empty - 10.8650 



with pure water ,, 4" ' 24.462 1 

 •■ ■■ 4 24.4634 



0.0012 14 

 0.0012 14 

 0.001215 

 0.0012 15 

 0.001205 

 0.001205 

 0.001205 

 0.001205 

 0.001203 

 0.001202 

 0.00 1 190 



These weight-determinations were all performed ac- 

 cording to the substitution method, by reading the oscilla- 

 tions of a balance, the sensibility of which. when not 

 loaded, amounted to 1.4'"'''' for every .division of the scale, 

 diininishing, on being loaded with successive weights. at a 

 comparatively uniform rate down to 1.9 '"■''''. when loaded 

 with 200*'. For the above-mentioned as for all accurate 

 weight-determinations involved in these labours, I made 

 use of a set of platina weights procured from Deleuil in 

 Paris, the corrections of which I had previous to start- 

 ing on the Expedition computed from a series of closely 

 agreeing determinations of weight performed with an ex- 

 cellent balance (made by P. Bunge). its sensibility when 

 loaded to the extent here specitied being nearly 0.14"'"'. 

 When performing the weight-determinations, the tempera- 

 ture and the atmospheric pressiire in the room were al- 

 ways observed, but the relative humidity not having been 

 found by observation bad to be roughly estimated, a source 

 of inaccuracy which however can exert no appreciable in- 

 fluence on the trustworthiness ol' the experiments, seeing 

 that an error of even 25 per cent in the computation 

 of tlie relative humidity of the atmosphere wouhl affect 

 the final result only by occasioning a. very small error in 

 the (ith decimal. Krom these data. the specific gravity 

 of the air during the process of weighing was computed 

 in the usual manner, and entered in the Table. Por each 

 successive determination, the pycnometer was plaoed in 

 linelv crushed ice. to give it the temperature ol' ( I ", and 

 for every other required, in a water-bath, which by constant 

 stining was kepl at an eipiable temperature: it was emptied. 



