PRACTICAL EXERCISES 493 



up and down in the mixture. For very exact work the temperature 

 of the freezing mixture must not be more than a few degrees below 

 the freezing-point of the liquid which is being examined. Put on the 

 lid, and immerse tube B. Into A, which must be perfectly clean, 

 put enough pure distilled water to fully cover the bulb of the thermo- 

 meter, and introduce the latter. For ordinary purposes distilled 

 water previously boiled to expel the carbon dioxide, and then cooled 

 in a stoppered flask, is sufficiently pure. Immerse A directly 

 in the freezing mixture through the hole by which G comes out, or 

 through a separate hole (not shown in the figure) till some ice has 

 formed in the water. Take A out of the mixture, wipe it with a cloth, 

 and hold the lower part of it in the hand till nearly the whole of the 

 ice has melted. If there is a cake of ice at the bottom, see that it is 

 displaced by the platinum stirrer. A trace of ice being still left 

 floating in the water, place A in B, and allow the temperature to fall 

 to a few tenths of a degree below the freezing-point you expect to 

 get, as determined by a previous rough experiment. The freezing 

 mixture is stirred up occasionally. The meniscus of the thermometer 

 is to be carefully followed, as it goes on falling, by means of a weak 

 hand lens. Now stir the water in A briskly. Suddenly it will be 

 seen that the mercury begins to rise. Keep stirring with the 

 platinum wire, and read off the maximum height of the mercury, 

 at which it is stationary for some time. The temperature can be 

 estimated between the graduations to thousandths of a degree. 

 Take out A, and observe the fine ice crystals in the water. Heat 

 A in the hand as before till nearly all the ice has disappeared ; then 

 replace A in B, and make another freezing-point determination. A 

 third one may also be made, and the mean of the three readings taken. 



Take out the thermometer, and dry it and the platinum wire with 

 clean filter-paper, or dip them in some of the urine, which is then 

 thrown away. ' Dry A or rinse it with urine. Then make a deter- 

 mination of the freezing-point of the urine in the same way as was 

 done with the water. The freezing-point of the urine will lie much 

 lower on the scale. 



Instead of freezing the liquid first and then leaving a little ice in it 

 when A is placed in B, A may be put into B before any ice has formed. 

 Cooling is then allowed to go on with gentle stirring to a few tenths 

 of a degree below the anticipated freezing-point. A small crystal of 

 clean dry ice is then introduced through the side-piece on a clean 

 splinter of wood or the loop of a cooled platinum wire, the end of 

 which passes through a piece of cork, by which it is held to prevent 

 conduction of heat. The platinum stirrer can be raised to receive 

 the crystal. The liquid is now vigorously stirred ; freezing occurs, 

 and the observation is made as before. 



Instead of the above method, the liquid may first be cooled directly 

 in the freezing mixture, but not so much that ice forms. A is then 

 put in B, and cooling allowed to go on while it is being stirred. 

 When it has been undercooled to a certain extent i.e., cooled below 

 its freezing-point the vigour of the stirring is increased. Ice forms 

 suddenly, as before, and the temperature rises to the freezing-point. 

 With urine this method is sufficiently satisfactory, but it is not 

 usually easy to get freezing of the distilled water till the under- 

 cooling is considerable, and it has been shown that this introduces 

 some error. 



Suppose the freezing-point of the distilled water on the scale of 

 the thermometer was 5 '245 and that of the urine 3'625, the value 



