THE DENSITY OF LIQUIDS 429 



(i) A modified form of the Gay-Lussac S. G. bottle 

 described below ; (ii) Wade and Merriman's form of the 

 Gay-Lussac pyknometer ; and (iii) Bousfield's pyknometer. 



(1) The ordinary specific gravity bottle, with the addition 

 of a ground-on cap covering the stopper and neck, gives very 

 satisfactory results when used with care. The ground-on cap 

 is an advantage ; it prevents a volatile liquid from losing an 

 appreciable weight in the balance case. Experiment seems to 

 show that evaporation occurs less by way of the capillary in 

 the stopper than by the ground-joint between the stopper and 

 the neck. When the determinations are made below the 

 room temperature, allowance must be made for the expansion 

 of the liquid as it is heated to the temperature of the balance- 

 case. This is provided for by a small cylinder ground to fit 

 on to the end of the stopper and small enough to go inside the 

 cap. By using the apparatus in the manner to be described, 

 no difficulty should be experienced in obtaining agreement 

 to the fifth place. 



The liquid for density determination is prepared as follows. 

 As much as possible of the dissolved air must be removed, 

 since, apart from any possible effect on the density, when the 

 liquid is put in the warm thermostat there is a tendency for 

 dissolved air to be evolved, and to cling to the glass in the 

 form of very minute bubbles. These are difficult to remove, 

 and, if allowed to remain, a low value for the weight of the 

 liquid is obtained. One method of removing the air is to 

 place a beaker containing the liquid in a vacuum desiccator 

 attached to the water-pump b}^ means of a safety device made 

 by fitting to a flask a rubber stopper through which passes 

 a vertical tube about 33 ins. long, bent down parallel to itself. 

 Mercury is put in the filter-flask, the side-tube is connected 

 to the pump, and the other to the desiccator. The latter is 

 then cut off from the moist air. At the same time, should the 

 water pressure fail, the water cannot be sucked back into the 

 desiccator. From ten to twenty minutes is sufficient to remove 

 the greater portion of the dissolved air. The liquid is then 

 poured into the weighed S.G. bottle (previously thoroughly 

 cleaned and dried — the usual precautions having been taken) 

 until it is level with the top of the neck (taking care that no 

 air-bubbles are left), and the stopper is inserted. The stopper 

 should always be in the same position relative to the neck, 



