A NEW TYPE OF ACCURATE SUNDIAL OR SOLAR 



CLOCK. 



By James Moir, M.A., D.Sc. 

 {Plate 26.) 



(Read July 10, 1918.) 



As is well known, the time as shown by a common sundial is 

 usually incorrect, the error* being, as a maximum, as much as 

 22 minutes, whilst it can also vary to the other side of correctness 

 to the extent of 9 minutes, giving a total variation during the year 

 of 31 minutes. This error, it may be mentioned for the sake of 

 amateurs, is due to the non-circular form of the earth's orbit 

 round the sun, and to the eccentric position of the sun in that 

 orbit. 



Now there are several types of sun-clock on the market, in 

 which, by means of a graduated mechanism, the sun's irregular 

 motion can be compensated so that the instrument gives the 

 correct time. Examples of these can be seen in Johannesburg 

 at the Union Observatory and at the Country Club : their only 

 fault is their grossly excessive cost, which is quite beyond the 

 pocket of most private persons — particularly scientists. 



About 15 years ago I had the idea that this problem could be 

 solved by means of the equatorial sundial. The latter is merely 

 a circular plate of metal, with a central axis of stout wire at 

 right angles to the plane of the plate. If such an arrangement is 

 supported at the ends of the wire so that the plate is parallel 

 to the equator (i.e., the wire points to the visible celestial 

 pole), then the sun in its apparent daily motion moves uniformly 

 round the wire, and the shadow of the latter on the plate also 

 moves uniformly, so that all that is necessary is to subdivide the 

 perimeter of the plate into equal divisions at the rate of 1° of 

 arc for every four minutes of time. 



A simple dial of this kind was constructed for me in 1903 in 

 the engineering shops of the Kleinfontein Mine, and stood for 

 several years in my late brother's garden there. In this case, the 

 permanent sundial-error of eight minutes — due to its being set up 

 in longitude 28 E., whereas the clocks go by 30 E. — was compen- 

 sated for by setting it up so that the axial wire did not point 

 exactly at the south pole, but 2° off. This dial did not, of course, 

 give correct time, but the latter could always be obtained by 

 applying the correction called the " equation of time " (as given 

 in " Whitaker's Almanac," for example) to the observed time. 



Within the past few years I have made two further improve- 

 ments on this type of dial, and* have now an instrument which, 

 made at small cost, is constantly reliable within one minute of 

 time, i.e., is much more correct than any of the clocks in town 



'*/.^., in the Central Transvaal. 



