difference of temperature may be calculated.

Record must

also be kept of which electrode is at the higher potential.

The above is then repeated, using electrodes of the other metal employed in the cell, and for electrolyte the liquid in which it is immersed.

198. Lastly, there is the determination of the thermo-electric effect at the junction of the two electrolytes. This is more difficult to carry out experimentally than the other, Fig. 93 showing a

form of apparatus which may be employed. Into the tubes A and B, through tightly fitting corks, pass the tubes C and D, the other ends of which terminate in the vessels E and F. The tubes A and B are filled with one electrolyte, whilst EC and DF are filled with the other. The junctions between the electrolytes are near the ends of the tubes C and D, where they dip into A and B. The tubes A and B are placed as before nto thermostats. Contact is made by means of platinum wires dipping into E and F, the temperatures at E and F being kept constant during the experiment. The tube A is then heated up, B being kept at constant temperature, and the mean E.M.F. per 1° C. rise of temperature calculated from a curve plotted as in the last case. The total thermal effect in the cell will be the algebraic sum of these separate effects.

In the case of the Daniell cell, Ca

ound that the

on (the zinc

t per 1° C., hilst that of 00073 volt the heated

thermal couples. The effect of change of temperature may determined separately in each case, the algebraic sum representing the total effect.

197. We shall first describe the method of measuring the effect of change of temperature between the electrodes and the electrolytes. For this a special form of cell must be employed, consisting of two tubes communicating with each other near their upper ends by means of a long tube of narrow bore (see Fig. 92). Both tubes and the connecting tube are filled with

B

A

FIG. 92.

one of the electrolytes; the electrodes A and B are made of the same metal that is placed in that electrolyte in the cell. The two tubes containing the electrodes are put each into a thermostat, T, so that the temperature may be regulated and kept constant at any desired value.

First the temperatures of both are kept the same, and then there should be no difference of potential between A and B if tested on a potentiometer or other instrument for measuring E.M.F.'s; then, keeping one tube at constant temperature, the other is raised to different temperatures, and the E.M.F. between A and B noted for the various differences of temperature. If the connecting tube is of sufficiently fine bore, and the apparatus is tilted so that the heated tube is at a slightly higher level than the cold one, there will be no fear of liquid convection currents being set up inside the tubes.

A curve should be plotted from the results, showing the connection between the E. M.F. and the difference of temperature between A and B ; from this curve the mean E.M.F. per

1° C.

difference of temperature may be calculated.

Record must

also be kept of which electrode is at the higher potential.

The above is then repeated, using electrodes of the other metal employed in the cell, and for electrolyte the liquid in which it is immersed.

198. Lastly, there is the determination of the thermo-electric effect at the junction of the two electrolytes. This is more difficult to carry out experimentally than the other, Fig. 93 showing a

and B, through tightly fitting corks, pass the tubes C and D, the other ends of which terminate in the vessels E and F. The tubes A and B are filled with one electrolyte, whilst EC and DF are filled with the other. The junctions between the electrolytes are near the ends of the tubes C and D, where they dip into A and B. The tubes A and B are placed as before into thermostats. Contact is made by means of platinum wires dipping into E and F, the temperatures at E and F being kept constant during the experiment. The tube A is then heated up, B being kept at constant temperature, and the mean E.M.F. per 1° C. rise of temperature calculated from a curve plotted as in the last case. The total thermal effect in the cell will be the algebraic sum of these separate effects.

199. In the case of the Daniell cell, Carhart found that the E.M.F. generated at the zinc-zinc-sulphate junction (the zinc sulphate being saturated at o° C.) was o'00079 volt per 1° C., the cold zinc rod being at the higher potential, whilst that of copper in copper sulphate (density 1'11) was o'00073 volt per 1° C., the cold plate again being positive to the heated

one; the E.M.F. generated at the junction of the two liquids was found to be negligible, being less than o'ooo03 volt per 1° C. The total temperature coefficient should therefore be

0'00079 — 0'00073 = 0'00006 volt per 1° C.

A Daniell cell was then made up with the same solutions, and its E.M.F. measured at different temperatures. This may be done by placing it in a thermostat, and steadying the temperature for some time before each reading. This was found to give a variation of 0'000073 volt per 1° C., and dividing this by 109, the E.M.F. of the cell, we get the true temperature coefficient

which agrees closely with the value as determined by the first method. The student is recommended to make the same investigation for the other standard cells.

TESTING A PRIMARY BATTERY.

200. In order to make a complete test of a primary battery so as to report upon its merits, the following data must be obtained :

:

(1) The E.M.F. of the cell on open circuit ; (2) The temperature coefficient of the cell;

(3) The internal resistance;

(4) The behaviour of the cell when sending a current ;

(5) The life and cost of working the cell.

In order to make the above measurements, any of the foregoing methods may be employed; but probably the simplest will be the potentiometer method, for the details of which see par. 219.

The methods of measuring E.M.F., temperature coefficient, and internal resistance, have already been dealt with, and need not be repeated. In (4), however, we have the most important test of all.

201. This test should be made in two parts, first taking a relatively small current from the cell, and another test taking a large current from it. In this, as in all the other tests, the

results obtained from experiments on one cell cannot be taken as conclusive; a number of cells must be tested individually, and from all the results thus obtained the general behaviour of the battery may be deduced.

In the first test, using small currents, after the E.M.F. and temperature coefficient have been determined, the cell is connected up to a resistance, which may have any value between 20 and 100 ohms-about 50 ohms will be found convenient. Time readings of the P.D. of the cell are then taken at intervals of five minutes at first, and afterwards at longer intervals; also the circuit of the cell is broken, from time to time, just long enough to allow of a measurement of its E.M.F. being made. This test may be continued for twelve hours, the circuit is then broken, and time-readings of the recovery of the E.M.F. are taken.

Curves showing the variations of E.M.F., P.D., and internal resistance with time of discharge may then be plotted from the data obtained, also a recovery of E.M.F. curve.

From these curves may be deduced the mean percentage fall of P.D. per minute and the mean rate of recovery. The approximate value of the current flowing should be stated on the curves. Several determinations of the above are made, using fresh cells each time. Another batch of cells are then tested for the high rate of discharge. In this case, the external resistance may have any value between 1 and 5 ohms, and readings similar to those made above are taken. The discharge may be kept on for 15 min. at a time, and recovery-readings taken after each discharge until the cell is completely exhausted; also, some cells may be discharged right on to exhaustion. The time-intervals between the readings must be smaller in this experiment than in the slow-discharge experiment, for obvious reasons.

From the curves plotted from the above, the percentage fall of P.D. and recovery of E.M.F. per minute may be calculated, and also the number of ampère hours supplied by the cell may be estimated from a curve showing the variation of the current with the time.

202. In the case of cells required to give out intermittent