PYROMETER 115 the openings by means of weighted levers, whereas the lower portions of the tubes are widened out and closed by plugs of wood, through which the electrodes in the form of pointed platinum wires penetrate to the depth of about 25 millimetres into the widened por- tions of the tubes. By a side branch the widened portion of each vertical tube com- municates by means of an India-rubber con- Fio. 6. Siemens's Pyrometer, General View. necting pipe to a little glass reservoir contain- ing acidulated water, and supported in a ver- tical slide. In raising the weighted cushions closing the upper ends of the vertical tubes, and in adjusting the position of the small reservoirs, the acidulated water will rise in both tubes to the zero line of the scale. In turning a button in front of the tubes the battery current is passed through both pairs of electrodes, the one circuit comprising the permanent resistance R and the leading wires up to the pyrometer, and the other the lead- ing wires and the pyrometer coil. If the re- sistance of the pyrometer coil should be equal to the permanent resistance R, then R, + y will be equal to R + y, and therefore V= V, ; but as the resistances differ, so will the vol- umes. Necessary conditions are, that both reservoirs are filled with the same standard solution of pure water with about 10 per cent, of sulphuric acid, that all of the electrodes are of the same form and size, and that their po- larity is reversed frequently during the progress of each observation, in order to avoid unequal polarization. With these precautions, which involve no particular skill or knowledge of electrical observation on the part of the ope- rator, very accurate results are obtained ; but in order not to incur considerable error of ob- servation, it is advisable not to continue the current, reversing the same, say twice, until at least 40 divisions of gases are produced in the least activated tube, which operation will oc- cupy from two to three minutes, if a battery of from four to six Daniell elements is employed. The volumes V and V, being noted, after hav- ing allowed half a minute for the gases to col- lect after the current has ceased, the weighted cushions upon the tubes are raised in order to allow the gases to escape, when the water levels will immediately return to their zero position, to make ready for another observa- tion. By inserting the observed values for V and V y into the expression above given, the unknown resistance R / can be easily calculated ; but in order to facilitate the use of the instru- ment, I have prepared a table which gives at a glance the resistance due to any two observed volumes, the volumes V governing the vertical, V, the horizonal columns, and the resistance read off at the point of intersection. At each point of intersection the resistance is marked in black, and the corresponding temperature in red ink. It now remains only to be shown what is the relation between the resistance and temperature in heating a platinum wire. The researches of Dr. Matthiesen, who has made the latest investigation on the effect of temperature upon electrical resistance, are restricted to the narrow range of temperatures between and 100 C., nor do they comprise platinum. He adopted the following general expression for the pure metals : R<=j T~TV which, in determining the specific values of x and y for each metal, gives a close agreement with observation between the narrow limits indicated, but is wholly inapplicable for tem- peratures exceeding 200 C., when the value t* commences to predominate and to produce absurd values for R<. It was necessary for my purpose to undertake a series of elaborate experiments with a view of finding a ratio of general application. Coils of thin wire, of platinum, iron, copper, and some other met- als, were gradually heated and cooled in metal- lic chambers containing the bulbs of mercury thermometers, and for higher temperatures of air thermometers, and the electrical resistan- ces were carefully noted. The progressive in- crease of electrical resistance was thus com- pared directly with the increasing volume of a permanent gas (carefully dried) between the limits of zero and 470 C., and a ratio estab- lished which is represented by the formula R ( =aT 2 + /rr + y, in which T signifies total temperature counting from the absolute zero (272 C.), and a, /?, and 7 specific coefficients for each metal. According to this formula, the electrical resistance is a constant at the absolute zero, and progresses in a ratio rep- resented graphically by a tipped-up parabola, approaching more and more toward a uniform ratio at elevated temperatures. Although the comparison with the air thermometer could only be carried up to 470 C., the general cor- rectness of the ratio of increase just stated has been verified by indirect means in measuring progressive heats, and by comparison with the platinum ball pyrometer. It is important to mention here that great care must be exercised in the selection of the platinum wire for the measuring spiral, one of two samples, both of which were supplied by the same eminent ma- kers, Messrs. Johnson and Mathey, having con- ducted 8'2 and the other only 4'7 times better than mercury at C. The abnormal electri- cal resistance of some platinum wire is due
