The difficulty of obtaining really accurate readings where thermometers are being quickly transported through varying temperatures is generally not duly appreciated. In the case of instruments carried m a balloon it should be remembered that the balloon itself conveys, clinging about it, no inconsiderable quantity of air, brought from other levels, while the temperature of its own mass will be liable to affect any thermometer in close neighbourhood. Moreover, any ordinary form of thermometer is necessarily sluggish in action, as may be readily noticed. If, for example, one be carried from a warm room to a cold passage, or vice versa it will be seen that the column moves very deliberately, and quite a long interval will elapse before it reaches its final position, the cause being that the entire instrument, with any stand or mounting that it may have, will have to adapt itself to the change of temperature before a true record will be obtained. This difficulty applies unavoidably to all thermometers in some degree, and the skill of instrument makers has been taxed to reduce the errors to a minimum. It is necessary, in any case, that a constant stream of surrounding air should play upon the instrument, and though this is most readily effected when instruments are carried aloft by kites, yet even thus it is thought that an interval of some minutes has to elapse before any form of thermometer will faithfully record any definite change of temperature. It is on this account that some allowance must be made for observations which will, in due place, be recorded of scientific explorers; the point to be borne in mind being that, as was mentioned in a former chapter, such observations will have to be regarded as giving readings which are somewhat too high in ascents and too low in descents. Two forms of thermometers at extremely simple construction, yet possessed of great sensibility, will be discussed in later chapters.