The aneroid will not answer for exact scientific purposes, as it cannot be relied upon for a length of time. Its error of indication changes slowly, and hence the necessity of its being set from time to time with the reading of a good barometer. To allow of this being done, at the back of the outer case is the head of a screw in connection with the spring attached to the vacuum box. By applying a small turnscrew to this screw, the spring of the vacuum box may be tightened or relaxed, and the index made to move correspondingly to the right or left on the dial. By this means, besides being enabled to correct the aneroid at any time, "if the measure of a height rather greater than the aneroid will commonly show be required, it may be _re-set_ thus: When at the upper station (_within its range_), and having noted the reading carefully, touch the screw behind so as to bring back the hand a few inches (if the instrument will admit), then read off and start again. _Reverse the operation when descending._ This may add some inches of measure _approximately_."--_FitzRoy._ [Illustration: Fig. 34.] =49. Small Size Aneroids.=--The patent for the Aneroid having expired, Admiral FitzRoy urged upon Messrs. Negretti & Zambra the desirability of reducing the size at which it had hitherto been made, as well as of improving its mechanical arrangement, and compensation for temperature. They accordingly engaged skilful workmen, who, under their directions, and at their expense, by a great amount of labour and experiment, succeeded in reducing its dimensions to two inches in diameter, and an inch and a quarter thick. The exact size and appearance of this aneroid are shown in fig. 34. The compensation is carefully adjusted, and the graduations of the dial ascertained under reduced pressure, so that they are not quite equal, but more accurate.