THERE is a passage in the “Germania” of Tacitus (chapter xlv.) which I do not think can have ever been examined by the historians of natural science, or it would have created a considerable stir amongst them. Side by side with a plain account—probably the earliest written one—of an arctic twilight, there lurks in it a description of the aurora borealis, which moreover lends countenance to the still prevailing notion that the northern lights are accompanied by sound.