ME. W. CROOKES ON THE ATOMIC WEIGHT OF THALLIUM. 
313 
be made with another apparatus ; but if the channel is entirely or in great part clear, 
it may be sealed up at the contraction, care being taken to apply the heat at such a 
place that no particles of metal or oxide are entangled in the fused glass. 
The apparatus lias now the form shown in fig. 12. It is hermetically sealed, entirely 
free from air, and contains a certain quantity of pure metallic thallium entirely free from 
oxide and as brilliant as mercury. 
The next operation is to ascertain the combined weight of the apparatus and metal. 
It is washed on the outside with dilute sulphuric acid to remove any particles of mag- 
nesia that might adhere to it, and after rinsing with water is dried and gently warmed. 
Its weight is then taken in the air-balance — not necessarily with extreme accuracy, but 
to enable a calculation to be made as to how much it will probably weigh in the vacuum- 
balance at a greatly reduced atmospheric pressure. As the substance weighed consists 
of thallium and glass in unknown proportions, the vacuum-weight cannot be calculated 
with any approach to accuracy ; but it is generally easy to arrive at some approximation 
to the relative proportions of thallium and glass, and in this manner the probable vacuum- 
weight may be estimated. 
The apparatus is now transferred to the vacuum-balance, and weights put which it is 
judged will balance it at an atmospheric pressure a few barometric inches short of a 
vacuum. The balance-case is then sealed up, and the exhaustion proceeded with. As 
the rarefaction proceeds, the beam is occasionally liberated until it is found that the 
apparatus and weight are in equipoise. If the barometer-gauge shows a rarefaction not 
equal to 25 inches of mercury, the air had better be let in, the requisite additional weight 
added, and the exhaustion recommenced ; but if, when the balance is in equilibrium, the 
rarefaction is above 25 inches, the weighing may be continued. 
Two sources of error have now to be guarded against: — 1. The alteration of tempe- 
rature inside the iron case consequent on the rarefaction. 2. The slow and almost un- 
avoidable leakage of air into the balance through the iron, the numerous joints, and the 
stuffing-boxes. This leakage should not exceed OT inch in an hour. 
Equilibrium having been obtained, two or three extra strokes are made with the air- 
pump, and the exhaustion raised to such a point that by about six hours’ leakage the 
balance is again in equipoise. The weights will at first appear lighter than the appa- 
ratus. The balance is allowed to remain well protected from external thermal influences, 
until the time has nearly arrived when the leakage of air into its interior has reduced 
the rarefaction to the point at which , the weights and apparatus will be exactly in equi- 
librium. The observer now enters the room, and after liberating the beam and setting 
it in oscillation, watches the movements of the index through a telescope fixed 10 feet 
off. By reason of the gradual leakage of air the inequality of the oscillations gradually 
diminishes, until at last the arcs are of the same value. At this moment the tempera- 
ture inside and outside the balance-case, the height of the barometer-gauge, and the 
reading of the standard barometer are observed. 
Six hours are generally sufficient to restore the temperature reduced by the exhaus- 
