The illustration (Fig. 33) shows the instrument with the telescopic tube drawn out for use, and the thermometer surrounded by the vapour of boiling water. The lamp is protected from wind by a perforated japanned tin case covered with wire gauze. When the boiler is charged and the lamp ignited the mercury ascends, and the point at which it becomes stationary shows the temperature, which will give the elevation in feet above the sea-level on reference to the table supplied by the optician from whom the instrument is purchased. [Illustration: 34. Barograph. Scale about 1/6.] A highly-refined automatic arrangement is adopted at some observatories called a Barograph, which, by the aid of photography, becomes a self-recording mercurial barometer. It is simpler in its arrangement than the thermograph, and includes a clock of superior construction, causing a cylinder bearing photographic paper to make one complete revolution in forty-eight hours. A double combination of achromatic lenses brings to a focus rays passing through a slit placed in front of the mercurial column, behind which is a strong gaslight or paraffin lamp, the rays of which are condensed upon the slit by a combination of two plano-convex lenses. Although a barometer is an instrument _artificially_ constructed by man, it should not be forgotten that when once made the column of mercury is placed in a passive or quiescent state in direct relation with the great forces of nature, so that its indications become to some extent _natural_ phenomena. This is aptly illustrated by what is called the “daily fluctuation” of the barometer which occurs in all countries, though the hours and extent vary with the latitude, diminishing as the latitude increases, according to a definite law. The phenomena does not admit of a satisfactory explanation, but is doubtless connected with the daily variations of temperature and of vapour in the air. The mercury falls _naturally_ (so to speak) from nine or ten to between three and four p.m.; it then rises till between nine and ten p.m. It falls again about four a.m., and rises again about ten a.m. It is usually highest at nine a.m. and nine p.m., and lowest at three a.m. and three p.m. These natural elevations and depressions of the mercury should be allowed for in reading the barometer, as any rise or fall in opposition to the natural rise and fall possesses for that reason increased importance. For instance, fine weather may be expected if the mercury rises between nine a.m. and three p.m.; in like manner rain may be expected should a fall take place between three p.m. and nine p.m. It will be inferred from the preceding facts that there are certain hours better suited for “taking a reading” than others. When one observation only is made daily, noon is the best time, two observations should be made at nine a.m. and nine p.m., and for three the best hours are nine a.m. (maximum), noon (mean), and three p.m. (minimum).