of the book it was cut out. It has been my duty to observe a number of total eclipses of the Sun, but they never last more than a few minutes, and the darkness is not so great but that one can read a watch face. However, if we landed on one of Jupiter's satellites, we could see such an eclipse as Sir Rider Haggard describes; for instead of having only our little Moon to act as a screen between us and the Sun (see diagram on p. 232), we should have great big Jupiter, and he would cut off the light for a long time. We can imitate what would happen by using a model. The Sun's light is represented by a beam from the electric lantern, which lights up one side of the model and leaves the other dark. On the bright side it is day on Jupiter, on the dark side it is night. But the darkness does not merely affect the surface; it streams away in a cylinder or cone of shadow. Now I will take a billiard ball to represent a satellite, holding it by a string. While it is anywhere outside this cone of shadow it is illuminated by the beam much as Jupiter itself is; one side of it is bright and has daytime, the other side is dark and is having night time. But as I circulate it round Jupiter, it comes into this cone of shadow; the light of the beam is cut off, and there is a total eclipse, which lasts all the time the satellite is passing through the shadow until it emerges on the other side. Some of you may not be able to see this emergence because Jupiter blocks your view; but those in a different part of the room can see it quite well, and so could the others if they changed their position. In looking at the real Jupiter, we on Earth are constantly changing our position as the
