( 944 ) 
first temperature stood at the highest mark on the stem which we 
may call a, sinks as soon as we proceed to lower temperatures. 
The levels of the liquid #,,%, . - %, corresponding with the 
constantly decreasing temperatures 7,, 7,...7, are read with a 
‘athetometer microscope until at last the level of the liquid sinks 
beneath the stem d,—-d, (Pl. IL. fig. 2). To turn these measurements 
to the best advantage, the dimensions of the appendix dq, must be 
so calculated that when the tap is closed for determinations of the 
densities of the vapour at the temperatures 7), 7, 7... 7) readings 
of the level ¢,,6,6,...S) in the capillary of the appendix can be 
made. In that ease one can calculate directly the corrections that 
must be applied to the rough values of Gig, @ligs +++ Olign ANA Qvap, , 
Ovaps +++ Qrayn Which are obtained by neglecting the correction for 
the small quantity of the coexisting phase, whereas otherwise these 
corrections would have to be determined by successive approximations’). 
The condition essential to the successful application of the simple 
method viz. that accurate equality may be realised between the 
eryostat temperatures at which the liquid and the vapour densities 
are determined, was fulfilled in our experiments, and so there was 
every reason to make use of this circumstance in our application 
of the method of constant mass to temperatures between 7’, and 7. 
As determinations of mass in the case of a permanent gas necessitate 
rather difficult measurements, if can be seen that the number of 
separate mass determinations necessary when the constant volume 
method (that in which the dilatometer from a mark on the appendix 
to the uppermost mark on the stem functions as a densimeter) 1s 
exclusively used, ought to be limited, to as narrow a temperature 
range as possible. For this reason the dilatometrie method (of constant 
mass) is combined with the densimetrie (or pyknometric) method of 
constant volume: the former method as indicated above gives the 
data neccesary for a series of intermediate temperatures, while the 
latter affords, as it were, the standard points in the range of tempe- 
ratures to be traversed between which intermediate points are inserted 
to correspond with temperatures occurring between two standard points. 
For oxygen, and this is in general the case with the permanent 
gases, neither the constant mass method nor the constant volume 
method can be rigorously applied over the whole range of 
temperatures. To traverse the various regions of temperature 
it is sometimes necessary to change from one temperature bath 
to another and this can only happen by exposing the apparatus 
LE. Marnias. Remarques sur le théoréme des élats correspondants. Ann. de 
Toulouse 1891. 
