The observations made are included in the following Table:

No. 3. DETERMINATION OF THE COEFFICIENT OF EXPANSION OF GLASS BY MEANS OF THE WEIGHT THERMOMETER.

THE apparatus consists of a small glass flask' (Fig. 1) of from 20 cc. to 25 cc. capacity. The neck of the flask is drawn out so as to terminate in a moderately fine tubulure. A portion of the glass a at the bottom of the neck is ground, so that the tube b which surrounds the neck may fit securely.

Fig.1.

b

с

The first thing to be done is to thoroughly clean the flask, by successive rinsings with nitric acid, potash, and water. The flask is gently heated over a flame, and then the neck is quickly immersed in nitric acid. As the flask cools the nitric acid is forced into it by atmospheric pressure. In order to get the acid out again the apparatus figured below (Fig. 2) must be used. It consists of a narrow glass bell-jar, S, cylindrical in shape, and closed at the top by a well-fitting caoutchouc cork. The latter is perforated in three places, so as to allow of the passage of two bent tubes T and P, and also of a funnel F, the latter provided with a glass tap N. A thick-walled india-rubber tube connects P with a Geissler water-pump; T is also fitted

1 Supplied by Cetti, 36, Brooke Street, Holborn, E.C.

with a similar tube, which for the nonce is closed by a clamp C. Inside the glass cylinder is a small tripod, which supports the

Fig. 2.

F

N

S

flask in the manner indicated. On producing a vacuum in the cylinder S, the acid is forced out into the beaker. The flask is filled with potash by means of the same apparatus; the neck of the flask being allowed to dip under the surface of a solution of potash contained in a beaker. And so with the washing with water, which latter operation must be repeated several times. The flask must then be dried. To this intent the tube T is made to communicate with a series of U tubes filled with fused chloride of calcium, and S is surrounded by steam by placing it in a hypsometer. A vacuum is produced, the clamp Copened, and the apparatus allowed to fill with dry air1. This must be repeated many times. When the flask is dry and has

1 It is very convenient to introduce a 3-way tap between P and the pump.

been weighed accurately-call its weight W-it is so placed in S that the tube of F opens into the tube b surrounding the neck. Clean and dry mercury is poured into F. When as good a vacuum as possible is attained, mercury is allowed to flow slowly into the tube b by slightly turning the tap N. By alternately letting in a small quantity of dry air, and exhausting, the flask will be almost completely filled with mercury. The jar containing the flask should be tilted while the mercury is entering to avoid the isolation of bubbles of air against the bottom or sides of the flask. A small bubble of air will of course remain in the flask, and generally shows itself at the base of the neck of the flask. When its size has been reduced as much as possible, the flask is withdrawn from S and gently heated over a sand-bath until the mercury by its expansion has expelled the bubble, and fills the whole flask. This being the case, more mercury is poured into b, so that its surface may stand considerably above c; and the flask is allowed to cool'. It is then placed in a funnel and surrounded by wellwashed ice-care being taken the while, that the mercury never is allowed to sink below the level c. When it has been in the ice some time and been reduced to temperature zero, the tube b is lifted off for an instant, and the mercury surrounding the neck of the flask allowed to run off. The tube is then replaced, the whole apparatus lifted out of the ice, dried thoroughly and weighed accurately. This weighing gives the weight of the flask, plus the weight of mercury it contains at zero. Let it be denoted by w. The apparatus is then suspended by means of a wire cage in a hypsometer, and heated thoroughly to the temperature of the issuing steam. Call this temperature t. The apparatus is removed from the steam, the mercury which has collected in b is run off, the apparatus is thoroughly dried, allowed to cool to the temperature of the room and again weighed. Let the weight now be w'. This is evidently the weight of the flask, plus the mercury which it contains at to.

1 The appearance of mercury in contact with a perfectly clean dry glass vessel is so characteristic that it is easy to tell whether or not the flask has been properly filled.

w - W= the weight of mercury in flask at 0o,

But, expansion of glass = absolute expansion of mercury apparent expansion, and as the absolute expansion of mercury is known from Regnault's tables to be 00018153 per degree for the temperature interval 0°—100°, the expansion of glass of which the flask is made can be easily found.

Weight of thermometer + mercury at 0°......................

=

=

343-235 grs. 22.8815 grs.

...

=

338.3525 grs. 22.8815 grs.

Weight of mercury contained in thermometer at 99.5o

Apparent expansion of mercury per degree

Absolute expansion of mercury per degree

=

320 3535-315.471

315 471 × 99.5

= 00015554

=00018153 (Regnault).

Hence expansion of glass per degree, for temperature interval

0°-99.5° = '00018153

=00002599.

00015554

D. J. CARNEGIE.

1 [This equation neglects a small fraction. The following is a strict method of deducing the coefficients of expansion.

X=weight of mercury in the flask volume V at 0°,

Let

X'=

....

y' at to,

p, p' the density of mercury at 0° and to respectively,
a, y the coefficients of expansion of mercury and glass;

ρ

1+ at

In the text 1+yt is regarded as being equal to unity.]

2 Regnault's observations have been re-calculated by Wüllner (Pogg. Ann. CLIII., p. 440) with the result of giving 00018163 the mean coefficient of absolute expansion of mercury between 0° and 100o.