COMPLEX SOLUTIONS.—-ARCHIBALD. 43 
inner tube was therefore retained and an outer tube provided of 
about the same length and thickness of walls, but with an 
internal diameter of 2.88 em., thus leaving an air space of about 
0.9 mm. between the tubes. This gave complete satisfaction. 
With uniform stirring no tendency was observed for the ice to 
form on the walls of the tube or on the bulb of the thermometer, 
or to freeze in a mass. The inner tube had its lower end 
re-entrant, as recommended by Loomis. 
I should like to draw attention to the importance of having 
the air space between the two tubes of the proper size. If the 
importance of this point has been noted by former observers it 
has escaped me. 
The hammer used for tapping the thermometer was part of a 
small electric bell and was covered with a piece of thick rubber 
tubing. It was found to be very essential to drive the hammer 
so that the blows on the thermometer might be of uniform 
strenoth. Some difficulty was met with in attaining that end; 
but by careful attention to the strength of the current what 
appeared to be sufficient uniformity was attained. 
The stirrer was of the ordinary ring form, the upright rod 
passing through a glass tube, the upper end of which was con- 
stricted, and the lower so far from the solution that the wetted 
portion of the stirrer could not touch it. A stop on the upright 
rod limited the extent of the stroke so that the ring would not 
leave the solution, and ensured the equality of the strokes. It 
was worked by hand as uniformly as possible. 
The over-cooling was seldom over 0.1 degree, owing doubt- 
less to the low temperature of the room in which the observa- 
tions were made. There was consequently no need of correcting 
for over-cooling. 
The freezing point of water was determined each day before 
determining that of the solutions, and in the event of any 
appreciable change occurring in the atmospheric pressure during 
the observations on the solutions, the observation on water was 
repeated. 
