In 'practical work thermometry fairly bris tles with errors. For several months after a thermometer is made the bulb gradually shrinks, probably owing to some molecular instability in the glass caused by the excessive heating em ployed in the process of blowing thetulb. This causes the thermometer to read too high. After each time a thermometer is used • for a very high temperature the bulb on cooling fails to contract promptly to the volume proper to the new temperature, and so now the thermometer for a while reads too low; however, prolonged heating at the temperature of boiling mercury tends to put the glass into a more stable state. Also such troubles are much reduced by the use of hard glass instead of soft glass for the bulbs. Errors also arise from the following causes: non-uniformity of the bore; variations of atmospheric pressure, which cause a yielding of the bulb; failure to have the stem of the thermometer at the same temperature as the bulb; the hydrostatic pressure on the bulb due to the liquid being tested, especially when the thermometer is sunk to great depths; a varia tion in the internal pressure of the mercury itself on the bulb when the thermometer is inclined from the vertical position to the hori zontal; a peculiar jerking motion of the mer cury when it ascends a very fine bore; the fact that equal volumes of the bore marked off on the tube do not represent equal expansions of the mercury, since at high temperatures the volume of the bore indicating a degree has in creased (this is quite distinct from the matter of the relative expansion of glass and mer cury) ; irregularities in the expansion of the glass of the thermometer; and lastly irregular ities in the expansion of the fluid itself, be it mercury, alcohol, air or any other substance. This last source of error is worth much con sideration because two thermometers otherwise perfect but containing different liquids, as alco hol and mercury, fail to agree in their indica tions. Further, we have no right arbitrarily to select any particular fluid as a standard and yet feel that our temperature scale has anything more than an empirical value. It will, how ever, be explained in the last section how a theoretical definition for temperature measure ment can be formulated (the thermodynamic scale), agreeing fairly with ordinary thermom eters, very closely with the hydrogen or nitro thermometer, and perfectly free from am iguity. In the hydrogen thermometer advantage is taken of the increase of pressure of a gas at tending an increase of temperature, the volume of the gas being kept constant. The hydrogen is confined in a glass bulb about two inches in diameter which is connected by a thick-walled capillary tube with the top of one side of a U-shaped apparatus consisting of two vertical glass tubes connected by a rubber hose at their lower ends and partly filled with mercury. When the hydrogen in the bulb is warmed it tends to expand and push the mercury down its side of the U and to cause it to rise on the other side, which is open to the atmosphere. This effect is counteracted by raising the glass tube op the open side, the rubber tubing allow ing this to be done. The extra back pressure of the mercury forces the hydrogen back to its former volume. In measuring the pressure to which the hydrogen at any time is subjected, the difference in level of • the mercury columns must have added to it the length of the baro metric colt= measured at the time. For each degree Centigrade added to the temperature, the hydrogen is found to increase in pressure about 1/273 of its pressure measured at 0° C.
Similarly for each degree subtracted, the pres sure decreases 1/273 of the pressure at 0° C. If this law held to the limit, we would conclude that at — 273° C. the hydrogen would lose all its pressure, thus indicating the cessation of all molecular motion — a veritable absolute zero of temperature. However, at extremely low tem peratures the perfect working of this law is interfered with through the dominance of cohe sion which reduces unduly the pressure of the hydrogen, and may cause it to assume the liquid or even the solid state. Nevertheless, this limit ing temperature as predicted by the hydrogen thermometer agrees almost exactly with the true absolute zero of the thermodynamic scale referred to above. On this absolute scale the temperature of freezing water is approximately +273° Abs., and temperature of boiling, +373° Abs.
Other methods of measuring temperature depend upon change in the electrical resistance of platinum, and upon the electromotive force created when the juncture of two dissimilar metals, as platinum and rhodium, is heated. Very high and very low temperatures may be measured by such methods.
Conduction of Heat.— When a sterling silver spoon is placed in a cup of hot tea, the handle of the spoon soon becomes'uncomfort ably warm to the hand. Heat has been con ducted through the silver. The molecules in the bowl of the spoon are the first to have their motion accelerated by contact with the tea. This extra motion is communicated to their neighbors which in turn pass it on until, step by step, the motion reaches the molecules in the handle. It appears that in some sub stances the character of connection between the molecules is more favorable to conduction than in others. As we might have expected, from the close crowding of the molecules found in solids, that class of bodies furnishes the best conductors; but in gases, in which the molecules are very loosely distributed, we nat urally find the poorest conductors. Liquids as conductors occupy a position intermediate between solids and gases. Metals surpass all other materials in conducting power, silver standing at the very head of the list, while near the foot of the list of solids are found organic materials and mineral substances, es pecially when in the porous or fibrous state, such as horn, leather, magnesia brick, asbestos fibre, sand, cotton wool, cowhair felt and down. Great value is attached to poor conductors of heat. They are called insulators. Bone is used in joining the handles to silver teapots. Our clothes are made of organic material woven so as to leave a multitude of fine pores, a condi tion favorable to insulation and met with in the fur of animals and in the feathers of birds. Sawdust and mineral wool for the same reason are made to serve as insulators of heat in the outer casing of ice boxes.
In the following table' of conductivities the better conductors have the higher numbers. These numbers, called the coefficient of conduc tivity, indicate the amount of heat energy meas ured in calories (a calorie is the amount of heat energy required to raise the temperature of a gram of water 1° C) conducted from one face to the opposite face of a centimetre cube of the substance when one of the faces is maintained one degree hotter than the other. The amount of heat energy conducted is pro portional to the difference in temperature be tween the opposite faces.