70 DIAL faces, for the reason that the shadow of the sun can never fall upon the south face before C. in the inuniin- or nftor in the evening; I. ut whon In- I-IM-.S K-foiv r> in the- morning liis beams will fall upon the north face. This, being the counterpart of the south face, will <lio the hour. A horizontal dial and the couth faro of a vt-rtiral one are represented
- }. The first of these cannot be used
with much accuracy in less than 20 of lati- tude because as we approach the equator the gnomon becomes more and more parallel with the horizon, so that if its length remains the same the upper end will cast but a very small shadow during the hours near midday. For instance, during the first hour before or after noon the sun in passing through 15 will cast a shadow on the horizontal plane but little more than one fourth the height of the gno- mon. If this be 6 in., the edge of the shadow will he only H in. from the perpendicular. As the height of the upper end of the gnomon above the dial is to its length as the sine of an angle is to the radius, it follows that in lat. 30 the edge of the gnomon would require to be 12 in. in order to give it an elevation of 6 in. To retain this elevation in lat. 10, the gnomon would have to be about 3J ft. long, and in lat. 5 about 7 ft. ; which, for various reasons (one of which is that the oblique shadows would be too dim to be plainly discernible), makes its use impracticable. The nearer we approach the pole the more does the gnomon approach a perpendicular position, until at the pole it becomes an extension of the earth's axis ; the hour angles, as marked upon the horizontal plane, becoming equal, as shown at E in fig. 1. Fio. 8. A glass cylinder havintr a rod for an axis and M<'- marked ith -21 equidistant lines, parallel with th- axis, was u^.-il by Ferguson in the construction of dials, and is itself a form DIALYSIS of equinoctial dial which may be used in any latitude by placing its axis parallel with the axis of the earth. It would thus be a modifi- cation of the dial of Berosus, if the hour lines FIG. 4. were marked upon the latter as meridians, the shadow of the axis falling upon the parallel lines of the cylinder precisely as they would upon the meridians of the hemispheres. Fig. 4 represents the hemisphere of Berosus sus- pended in a graduated arc, by means of which its gnomon may be adjusted to the latitude of the place. It may also be turned upon its axis and held at any degree of inclination to the east or west, so that when the days are more than 12 hours long it will indicate the time be- fore 6 in the morning and after 6 in the even- ing. The lines 5, 6, 7, 8, &c., are intended to represent meridians, and should meet at the poles A B. A magnetic needle and a pair of spirit levels facilitate its adjustment. Burt's solar compass (see COMPASS, SOLAK) contains a sun dial which is an elaboration of this plan. An hour arc is held in a plane having an in- clination corresponding to the latitude of the place. Over this arc an arm, attached to what is called the declination arc, is made to move until the sun's rays coincide with the axis of the arm, when the number of degrees trav- ersed by the hour arc will indicate the sun's apparent time. This is the most correct sun dial that can be constructed. DIALYSIS (Gr. didhvaig, a separating), or Analysis by Diffusion, names given by Prof. "Tho- mas Graham to a method proposed by him for effecting certain separations, usually of compound substances one from another, by means of the different rates at which substan- ces diffuse through moist gelatine-like films or other septa, or upward through water or viscid
