06 Progress in Science. [July, 
brought up round the vertical sides of the zinc, so as to form a kind of bag 
round it. The parchment paper is thus a separator between the mass of 
liquid in the cell and that immediately surrounding the zinc. There is a 
circular hole in the middle of the zinc, and the tube of a glass or earthenware 
funnel passes through this and through a hole in the parchment paper, the 
edges of which are tied round the tube, down to the bottom of the cell. The 
cell is then filled up with a saturated solution of sulphate of zinc till the level 
of the liquid is higher than that of the top of the zinc; and on the top of this 
a layer of pure water 2 or 3 inches deep is poured carefully, so as to avoid 
mixing. The pure water forms an atmosphere into which the sulphate of zinc 
formed during the action of the battery may diffuse, and thus crystallisation is 
avoided. To set the cell in action crystals of sulphate of copper are put into 
the funnel. The dense solution flows down over the copper plate at the 
bottom, and in less than five minutes the cell is in full work. The average 
resistance of each cell is o-19 of a B.H. unit. 
Professor G. Carey Foster, F.R.S., has recently brought forward some 
valuable observations and improvements on the Wheatstone bridge. The 
fault of this instrument in its usual form is, that the resistance of the copper 
bands at the two angles is not taken into account. When large resistances 
have to be measured the importance of this omission is very small; but when 
resistances are to be measured smaller than the resistance of the apparatus, it 
is at once apparent that the mistakes incurred become serious. The essence 
of the fault lies in this, that the same amount will be added for resistance 
of the band whether the measurement be one or a thousand; and to make it 
quite flagrant, the nature of the fault is of the kind that would make (1+a) 
a thousand times less than (Iooo+a). Suppose a to be 1 only; then 2 would 
be made to answer for the thousandth part of roor. For the purpose of 
making accurate measurements of very small resistances, such as those of 
short lengths of wires, it is essential that we should either have a standard of 
comparison, the value of which can be changed continuously, or, in other 
words, by indefinitely small amounts ; or, that the method adopted for compa- 
ring the unknown resistance to be measured with an invariable standard, 
should be such as to enable us to vary continually, or by indefinitely small 
amounts, the indicated ratio between the unknown resistance and the 
standard. The former of these alternative means of measurement, namely a 
continuously adjustable standard, is afforded by the rheostat of Wheatstone 
and Jacobi; and the second, or a continuously adjustable method of compa- 
rison, is afforded by those forms of Wheatstone’s ‘ Differential Resistance 
Measurer”’ in which the ratio between the resistances to be compared is 
established by moving one of the galvanometer contacts backwards and for- 
wards along a graduated wire, until a point is found which causes the current 
in the galvanometer to vanish. But neither the indications of the rheostat 
nor those of any of the adjustable forms of Wheatstone’s bridge can be 
accepted immediately as trustworthy, where anything like minute accuracy is 
required. The direc indications of both instruments are liable to require cor- 
rections, the most important of which are those required on account of 
the resistance of connections not represented in the resistance of the scale, 
and those required on account of irregularities in the diameter or conducting 
power of the wire by which the adjustments are made. Professor Foster pro- 
poses to employ a graduated wire as an adjustable standard for the measure- 
ment of any resistance less than its own, in such a way as to eliminate 
entirely from the result the resistance of the connections by which it is united 
with the rest of the circuit. A German silver wire, EF, one metre in length, 
Ages k Clk Dp 
E F 
graduated in 1000 millimetres, is connected at the ends with a copper band, 
which passes at right angles to each end of the wire, and runs parallel to 
it on the opposite side, where it is interrupted at A, B, c, and D, for the intro- 
