To increase precision of motion in the train of wheels, it is necessary that the mechanism be as simple as possible. For this reason all complications of date, etc., are left out. We have even abandoned the usual convenient plan of having the hour and minute hands mounted at the same centre; for this kind of mounting makes necessary a slightly more intricate form of wheelwork. The astronomer's clock usually has the centres of the second hand, minute hand, and hour hand in a straight line, and equally distant from each other. Each hand has its own dial; all drawn, of course, upon the same clock-face. Even after such a clock has been made as accurately as possible, it will, nevertheless, not give the very best performance unless it is taken care of properly. It is necessary to mount it very firmly indeed. It should not be fastened to an ordinary wall, but a strong pier of masonry or brick must be built for it on a very solid foundation. Moreover, this pier is best placed underground in a cellar, so that the temperature of the clock can be kept nearly uniform all the year round; for we find that clocks do not run quite the same in hot weather as they do in cold. Makers have, indeed, tried to guard against this effect of temperature, by ingenious mechanical contrivances. But these are never quite perfect in their action, and it is best not to test them too severely by exposing the clock to sharp changes of heat and cold. Another thing affecting the going of fine clocks, strange as it may seem, is the variation of barometric pressure. There is a slight but noticeable difference in their running when the barometer is high and when it is low. To prevent this, some of our best clocks have been enclosed in air-tight cases, so that outside barometric changes may not be felt in the least by the clock itself. But even after all this has been accomplished, and the astronomer is in possession of a clock that may be called a masterpiece of mechanical construction, he is not any better off than was the jeweller with his regulator. After all, even the astronomical clock needs to be set, and its error must be determined from time to time. A final appeal must then be had to astronomical observations. The clock must be set by the stars and sun. For this purpose the astronomer uses an instrument called a "transit." This is simply a telescope of moderate size, possibly five or six feet long, and firmly attached to an axis at right angles to the tube of the telescope.