I saw at once that there was a signal difference in the two indications, and proceeded carefully to revise the earth-measurements.

On the average of thirteen measures the halo was about 87", or nearly 1-1/2' in breadth, the disc, allowing for the twilight round its edge or limb, about 2 deg.

50'.

If the refracting atmosphere were some 65 miles in depth, these proportions were correct.

Relighting the lamp, I worked out severally on paper the results indicated by the two instruments.