SCIENTIFIC SUMMARY. 
215 
covered by an insulating* material. The discs are cut so that they nearly fit 
the tubes, one being exactly double the diameter of the other, and therefore 
exposing four times the surface to the action of the liquid. On connecting 
the terminals with a galvanometer, the current will be found to increase as 
the distance between the zinc and platinum plates is diminished by lower* 
ing the zinc plate in the tube. In order to obtain the same deflection of the 
galvanometer by the narrow cell, the distance between the plates must be 
one-fourth of the distance between those of the larger one. 
Index of Refraction of Liquids . — MM. Terquem and Trannin propose a 
new method of determining the index of refraction of liquids, based on total 
reflection. A small tank with parallel sides of glass is used to contain the 
liquid. In this are placed two plates of glass cemented along the edges, and 
free to turn through a measurable angle. The whole is placed between the 
collimator and observing telescope of an optical circle. The image of the slit 
being distinctly seen, the plates are turned until total reflection takes place, 
when the slit disappears. Then turning the plates in the opposite direction, a 
second measurement is obtained, and the difference equals twice the limiting- 
angle of the liquid with regard to air. With common light the image 
turns red before disappearing, but with monochromatic light the disappear- 
ance is almost instantaneous, and can be determined within a quarter of a 
minute. With a Geissler tube containing hydrogen as a source of light, the 
image undergoes two marked changes of colour, due to the total reflection 
of the two rays ITy and TI/3. The error then is only about half a minute, 
and for H a about a quarter of a minute. The measurements of several 
liquids are given, agreeing very closely with previous determinations, the 
difference being readily accounted for by the difficulty of obtaining the 
liquid perfectly pure. — See also “ Comptes Rendus,” lxviii. p. 1843. 
Expansion of Hard Rubber . — Having several times noticed that glass 
flasks, closed by stoppers of hard rubber, burst, M. Kohlrausch concluded 
that this substance must be very dilatable. This hypothesis was fully veri- 
fied by experiment, for the expansion of this body was found to be about 
three times that of zinc. From his measures, the coefficient of dilatation 
for 1° between 16°*7 and 25°*3 = *0000770, and between 25°*3 and 35°*4 = 
•0000842. Thus, not only has hard rubber a very great coefficient of dilata- 
tion, but the latter increases very rapidly with the temperature. This 
remarkable property can be applied to the construction of very delicate 
thermometers. Thus, with a small instrument, consisting of two strips of 
rubber and ivory, 20 cms. long, glued together and fastened at one end, we 
obtain, at the other extremity, a movement of several millimetres for a 
change of temperature of one degree. The coefficient of hard rubber is 
equal, at zero, to that of mercury : above, it is greater. We can, then, as a 
curiosity, construct a mercury thermometer with a reservoir of this sub- 
stance, whose changes will be the opposite of those of a common thermo- 
meter, and which will fall with an increase of temperature. 
Singing Flames and Sympathetic Vibrations. — M. E. Gripon has studied at 
length the influence exercised on the vibrations of a column of air by neigh- 
boring sonorous bodies. He concludes that the pitch of a vibrating mass 
of air is raised by bringing near its orifice an elastic membrane or a second 
mass of air which alone would give the same note. A similar effect is pro- 
