This library is built in the open.
If you spot an error, have a suggestion, or just want to say hello — we’d love to hear from you.

light of any kind, there is an adjacent dark place with an opening between them, we find the dark place illuminated by the light opposite to it.
I 28a
[25] This state of affairs can be tested at all times. If an experimenter wishes to do so, let him employ a dark chamber with a door facing an adjacent wall on which sunlight may shine. The chamber door should not be exposed to the sky. Thus, the light should reach the facing wall through an opening or door at the top of the wall of the dark chamber, assuming this wall to be higher than the chamber’s roof. The space between the two walls, namely the wall with the chamber door and the wall facing it, should be roofed above the opening | or shaded by an opaque body. And let the back of the chamber face east. The experimenter should observe the place when morning light shines on that wall through the opening opposite, which should be fairly wide. He will find the chamber illuminated by that light, and the light in the chamber weaker than the light on that wall. Then, as the light on the wall grows stronger, so will the light in the chamber. And when the sun’s light radiates on the wall, the light in the chamber will gain in intensity and strength. Further, that place inside the chamber facing both the door and the irradiated wall will be found more intensely illuminated than the rest of the chamber. Then when the sun moves away from that wall, the light in the chamber will weaken.
I 28b
[26] Suppose now that inside the chamber there is another dark chamber whose door is exposed to the wall facing the first door. When the light shines upon the outside wall, and consequently appears on the wall inside the first chamber opposite the door, then, assuming this light to be strong and the door of the second chamber open and exposed to this wall, | this second chamber will also be illuminated by the light of this wall, especially if the irradiated outside wall is pure white. And those parts of the second chamber that face this wall and are close to it will be found to be more intensely lit than the rest of the chamber. And if the illuminated wall inside the first chamber is white, the light appearing in the second chamber will be more manifest. Similarly, if moonlight and daylight are tested in this manner, the dark place will be found to be illuminated by them if it faces them.
I 29a
[27] It is therefore shown by this experiment that from the light that shines on any body, light radiates in every opposite direction. This being the case, the light that appears on the surface of the earth at the beginning of day before sunrise and at the end of day after sunset is but a light that comes to it from the light that is manifest in the opposite atmosphere. The atmosphere is lit before sunrise by sunlight | because it is facing the sun and the sun is not at this time hidden from it but only from the surface of the earth. And this light extends in straight lines from the body of the sun into the illuminated atmosphere. Then from the atmosphere illuminated by sunlight there also radiates, again in straight lines, some light on those places facing it on the earth’s surface. And as
long as the light is scanty and weak, what radiates of it on the earth is not apparent, but as it gains in strength and intensity, the light radiating upon the earth becomes stronger and visible.
I 29b
[28] And so it is with evening light. After sunset, the part of the atmosphere facing the sun is illuminated with its light; this light extending in straight lines. From the illuminated atmosphere, light radiates on those places facing it on the earth’s surface. And as long as the light in the atmosphere is strong, the earth will be lit with a visible light, but when it weakens, the light on the earth’s surface weakens and the earth’s surface becomes dark. It is similarly the case with the lights found in dwellings and walled courtyards shaded from the sun | and with all shadows illuminated in day-time before and after sunrise and after sunset and during the rest of the day: these lights reach them only from the luminous atmosphere facing them and also from the illuminated walls and nearby surface of the earth illuminated by daylight. Such, then, is the case with daylight and the lights in places that are lit at night by moonlight when they are concealed from the moon, and also the case with the light of fire.
I 30a
I 30b
[29] This property, I mean the radiation of lights from accidental lights in straight lines, can be examined by an accurate experiment that leads to certainty. Morning light can be examined as I shall describe. The experimenter should make use of a chamber inside which there is another chamber; let them be situated between east and west, and let no light enter them except through the door. Let the eastern wall in the eastern chamber be exposed, and a hole be pierced at the top of this wall similar to those made in the walls of chambers to let in the light. Let the hole be circular with a diameter not less than | a foot. Let it take the shape of a cone, wider inside than outside. This hole then faces the eastern side of the sky. In the opposite wall common to both chambers let the experimenter pierce two holes facing the first hole. These two holes then lead to the western chamber. Let them be nearer to the ground than the first hole so that an observer looking into either of them will see the sky through the first, higher hole. Let each of the two holes be equal and similar to the first. It should be ensured that the extension of each of these two holes through the thickness of the wall should be along the straight lines imagined to proceed from the outer end of the first hole to the facing end of the second hole. This can be assured by making the wall so thick as to allow the holes a fairly deep extension in its body, so that the light coming out of these holes will not diverge too much. The extension of each of these holes in the body of the wall should be | of equal width and not conical in shape.
[30] Now stretch a thread from the first, higher hole to one of the two holes, and make sure that the thread passes through a point on the outer end of the