282 
ME. W. CEOOKES OX THE ATOMIC WEIGHT OF THALLIUM. 
Even with moderate rarefaction the iron case of the balance showed at first a certain 
amount of porosity, due to the “kish” or graphite, carbide, and silicide diffused through 
the metal like a sponge. Cast-brass and even drawn-brass tubes exhibit a similar 
porosity. This porosity in the casing of the balance admits of easy remedy, by painting 
the whole surface with two or three thin coats of white-lead paint mixed with boiled 
linseed-oil or fine copal-oil varnish, allowing each coat to dry before the next is laid on. 
The vessel should be painted when it is partially exhausted ; the multitude of small 
holes then appearing in the smooth surface of the paint as it is forced inwards by the 
pressure of the outer air should be covered carefully with thin coats of paint. When 
this effect ceases, a final thin coat should be given and allowed to dry. 
The iron flanges were first planed true with the planing-machine, and then u fined off” 
by Whitworth’s process of scraping, generally employed for such work as slides of 
engines &c. The lead wire laid between the flanges of the iron case in india-rubber, 
becoming compressed when the bolts are tightened, effectually precludes the entrance 
of air. The washers of the iron door to the left of the case are of well-greased leather, 
while the glass plates in the other parts of the apparatus are cemented into double 
frames with red lead. 
At first it was attempted to put nearly the correct weight into the pan, and then make 
the final adjustment by means of the rider. It was, how r ever, soon found that the more 
accurate method was to introduce a certain weight, and then to alter the pressure of the 
air until the balance shows equilibrium. Thus, supposing a glass vessel weighing in air 
625T200 grains has to be weighed in vacuo , calculation estimates the probable weight 
(in vacuo ) at 625-370Q. I therefore introduce rather less (625-3600) than this weight, 
and exhaust until the balance attains equilibrium, when the gauge shows an atmospheric 
pressure equal (say) to 3*75 inches of mercury. When this is obtained the weight is 
slightly increased or diminished with the rider, and the exhaustion varied until a fresh 
equilibrium is established. Two weighings at different degrees of atmospheric pressure, 
varying by a considerable interval, give data upon which to calculate with great accuracy 
what the weight would be in a perfect vacuum. 
With a rider there is some difficulty in estimating the exact point at which it rests, 
and it is necessary to note the oscillations, placing the rider as exactly as possible on 
one of the divisions of the beam. The best weighings, perhaps, will be taken when the 
arc is not very small. 
Temperature has an effect upon the air-balance, rendering it less sensitive when 
increasing. This is perhaps due to the varying expansibility of the arms and the knife- 
edges upon which the pans are hung, or the superior and inferior parts of the beam may 
expand unequally. The two arms of the balance at times expand unequally ; and in 
finding the true value of the weights employed in the determination, this cause of 
error is eliminated by following Gauss’s method of interchanges — the constant friction 
of the forceps against the weights in transferring them from one pan to another being- 
obviated by employing hooks of thin wire attached to the agate plane, upon which the 
