OF ARTS AND SCIENCES. 435 



((llamcter .065"). The former had a total resistance of 29. G8 ohms, 

 the latter of 9.18 ohms, or the single loops .99 and .1G7 ohms. To 

 allow for accidental variations in the wire or in the connection, each 

 loop was measured separately, and a table of resistances thus formed. 

 To the end of each loop was attached a short piece of stout copper 

 wire, wound in a helix and sunk in a hole, so that any two could be 

 connected by a w'ire terminating in copper plugs. By this system, any 

 resistance from .17 ohms to 370 ohms was easily thrown into the 

 circuit, and the connections were few in number and of very small 

 resistance. This plan also had the advantage of extreme cheapness. 

 The heating of the wire was independent of its length, except so far as 

 the current altered. Practically, the three sizes of wire employed 

 woidd convey 1.5, 5, and 10 vebers, without undue heating. A change 

 of temperature of 100° C. increases the resistance of German silver wire 

 about 4 per cent ; and, to allow for this, a so-called thermometer-board 

 was employed, on which pieces of the three wires wound in a helix 

 were stretched. To determine the heating of either size of wire, the 

 proper helix was inserted in the circuit, and a thermometer placed in 

 it. On trial, it was found that the readings were much too high, the 

 radiation prevented by the adjacent spires of the helix much more 

 than compensating for the imperfect connection with the thermometer. 

 This difficulty might be avoided by stretching the helix until these two 

 errors should compensate, which might be tested by covering the helix 

 and a straight wire with iodide of mercury and copper, and altering 

 the form until the color of the iodide changed with the same current 

 in both. As, however, the correction is small at ordinary temperatures, 

 it was deemed best to neglect it, taking care to touch them occasionally 

 when very powerful currents were passing, to make sure that the wires 

 did not become very hot. 



Current. A special device is also needed for the determination of 

 the current produced in absolute measure. If an ordinary tangent 

 galvanometer with a single coil of thick wire was employed, the stronger 

 currents could be well comjiared ; but it would be difficult to reduce 

 them to vebers, since a feeble current suitable for depositing copper 

 would not appreciably deflect the needle. Accordinglv, a cosine gal- 

 vanometer shunted was employed, or rather, as here used, a tangent 

 galvanometer, since the coils were kept vertical. The coils consisted 

 of about 50 turns of No. IG copper wire, G inches in diameter, and 

 3 inches apart. The needle had a length of but | ', and was made of 

 a piece of watch-spring. An index, 3 inches long, was attached ; and 

 the magnet, being suspended by a filament of silk, swung over a gradu- 



