be brought near them. It is even desirable that the entire building, even as to its side walls and roof, should contain no iron; but it is unnecessary to be so cautious as to fear placing a clock, or a theodolite with steel pivots, at a distance of from five to six metres from the instrument. The influence of the steel parts, if they are magnetic, may be approximately deduced by calculation, and is found to be much too small to be sensible at those distances. Small pieces of iron outside the room have still less influence. If, however, there were in the neighbourhood large masses of iron, especially very long iron bars, (such as iron railings), although their influence would be very small, yet it should not be totally neglected. If they are at a distance of a hundred feet from the observatory, they offer no important impediment, at least if they are fixed. Such a locality is sufficient for measuring the Declination and Intensity, and also for observing their changes. Measurements of the Inclination may be performed in the same locality, but not, however, without interrupting the other observations. It therefore appears convenient, when circumstances permit, to assign a separate locality for measurements of the Inclination, which may be at no great distance from the first-named room. Where no absolute measurements are made, but only the changes of Declination observed at the fixed terms, such a room suffices, even should it contain much iron within and without its walls, provided that all the iron remains unmoved during the observations. The room of the Göttingen magnetic observatory is figured in Plate II., and the ground-plan in Plate III.
For the purpose of setting up the instruments, a line should be drawn on the floor representing the magnetic meridian, which line must pass nearly through the middle of the room, and terminate at the southern or northern end of it, where a firm foundation must be made for the theodolite and clock. When this foundation has been prepared, and the theodolite placed upon it, let a scale be first attached to the stand of the telescope, so that a plumb-line let fall from the object glass of the telescope passes freely before the scale. The scale must be horizontal and at right angles with the magnetic meridian; it must be capable of being raised or lowered at pleasure, and must be bisected by the magnetic meridian passing through the optical axis of the telescope. Next let fall a plumb-line from the ceiling to the floor, in such manner that the plane of the magnetic meridian passing through this plumb-line may contain the op-
