10 DKNSITIKS OF OXYIiKN AND HYDROOKN 



A\'itli tlic |iirseiit npiiaralus, ;(liciiialt' laisin^ jiiitl loweriiiLr "I tin* l^IhIm- \<\ incaiis 

 i>r tilt' movable jilatfoiin soon made tlie tempeiatuie iiiiifnrm. 



Wlicii the tlieiiiiometei's were used for several weeks in tlie iioii/.mital piisitioii, 

 the eolumii nf mercury in two of them sometimes sej>ai'ated. so (liat they could no 

 longer he rc.-id. Foi' some detei'minat ions, therefore, I pidcureil thernionietei-s 

 <livide<l into fiftieths of a degree, having the graduations at the end of a long tulte. 

 AVith these the temperature could l)e read, even at the l)ott<nn of the cylinder, with 

 the thermometer in the vertical position. V>\ moving them, the temperatuic of the 

 water at all depths could he determined. 



Most determinations were made at the temperatui'e of the I'oom. P)Ut a few 

 determinations were made with the temperature below 4° C, some by using ice in a 

 large tank, and some by working in an attic room in winter. 



6. EXI'ANSION OK WATKi: AN1> ol (il.ASS. 



The volume of one giaiiniie of watci- at different temperatures whiih i> rcMpiiicil 

 in reduction of the hydrostatic Aveighings was taken from the results of Marek, 

 Thiesen, Scheel, and Kreitling, as given in the Beilihutter zu den A/uuiJen, 18, o9. 

 'riif folli.\vin>' table iiives the volume of one crramme of water for each tenth of a 

 degree fiom '•" to l'l'°: 



VOI.L'MK OF ONT, OK.ANUMC OK WATER AT TEMPER ATl-RIC T. 



T. o 1234567S9 



The coetticient of expansion of the glass globes used for weighing oxygen was 

 iletermined by making three weight thermometers of the necks of three globes, 



