The width of the photographic sheet which receives the spot of light reflected from the mirrors in the above instruments is generally so great that in the case of ordinary changes the curve does not go off the paper. Occasionally, however, during a disturbance such is not the case, and hence a portion of the trace would be lost. To overcome this difficulty Eschenhagen in his earlier type of instruments attached to each magnet two mirrors, their planes being inclined at a small angle so that when the spot reflected from one mirror goes off the paper, that corresponding to the other comes on. In the later pattern a third mirror is added of which the plane is inclined at about 30° to the horizontal. The light from the slit is reflected on to this mirror by an inclined fixed mirror, and after reflection at the movable mirror is again reflected at the fixed mirror and so reaches the recording drum. By this arrangement the angular rotation of the reflected beam is less than that of the magnet, and hence the spot of light reflected from this mirror yields a trace on a much smaller scale than that given by the ordinary mirror and serves to give a complete record of even the most energetic disturbance. See also Balfour Stewart, _Report of the British Association_, Aberdeen, 1859, 200, a description of the type of instrument used in the older observatories; E. Mascart, _Traité de magnétisme terrestre_, p. 191; W. Watson, _Terrestrial Magnetism_, 1901, 6, 187, describing magnetographs used in India; M. Eschenhagen, _Verhandlungen der deutschen physikalischen Gesellschaft_, 1899, 1, 147; _Terrestrial Magnetism_, 1900, 5, 59; and 1901, 6, 59; _Zeits. für Instrumentenkunde_, 1907, 27, 137; W. G. Cady, _Terrestrial Magnetism_, 1904, 9, 69, describing a declination magnetograph in which the record is obtained by means of a pen acting on a moving strip of paper, so that the curve can be consulted at all times to see whether a disturbance is in progress. The effects of temperature being so marked on the readings of the horizontal and vertical force magnetographs, it is usual to place the instruments either in an underground room or in a room which, by means of double walls and similar devices, is protected as much as possible from temperature changes. For descriptions of the arrangements adopted in some observatories see the following: U.S. observatories, _Terrestrial Magnetism_, 1903, 8, 11; Utrecht, _Terrestrial Magnetism_, 1900, 5, 49; St Maur, _Terrestrial Magnetism_, 1898, 3, 1; Potsdam, _Veröffentlichungen des k. preuss. meteorol. Instituts_, "Ergebnisse der magnetischen Beobachtungen in Potsdam in den Jahren 1890 und 1891;" Pavlovsk, "Das Konstantinow'sche meteorologische und magnetische Observatorium in Pavlovsk," _Ausgabe der kaiserl. Akad. der Wissenschaften zu St Petersburg_, 1895. (W. Wn.) FOOTNOTES: