"Ah, yes," said Harry, extinguishing the candles; "I've read, of course, that your authorities exactly reverse our own Oxford rules. You are all gated, I believe, from dawn to sunset, instead of from sunset to dawn, and have to run away helter-skelter at the first streaks of daylight, for fear of being too late for admission without a fine of twopence. But you will allow that your usual habit of showing yourselves only in the very darkest places and seasons naturally militates somewhat against the credibility of your existence. If all apparitions would only follow your sensible example by coming out before two scientific people in a well-lighted room, they would stand a much better chance of getting believed: though even in the present case I must allow that I should have felt far more confidence in your positive reality if you'd presented yourself in broad daylight, when Jim and I hadn't punished the whisky quite as fully as we've done this evening." When the candles were out, our apparition still retained its fluorescent, luminous appearance, and seemed to burn with a faint bluish light of its own. We projected a pencil through the spectroscope, and obtained, for the first time in the history of science, the spectrum of a spectre. The result was a startling one indeed. We had expected to find lines indicating the presence of sulphur or phosphorus: instead of that, we obtained a continuous band of pale luminosity, clearly pointing to the fact that the apparition had no known terrestial element in its composition. Though we felt rather surprised at this discovery, we simply noted it down on our paper, and proceeded to verify it by chemical analysis. The phantom obligingly allowed us to fill a small phial with the luminous matter, which Harry immediately proceeded to test with all the resources at our disposal. For purposes of comparison I filled a corresponding phial with air from another part of the room, which I subjected to precisely similar tests. At the end of half an hour we had completed our examination--the spectre meanwhile watching us with mingled curiosity and amusement; and we laid our written quantitative results side by side. They agreed to a decimal. The table, being interesting, deserves a place in this memoir. It ran as follows:-- _Chemical Analysis of an Apparition._ Atmospheric air 96.45 per cent. Aqueous vapour 2.31 " Carbonic acid 1.08 " Tobacco smoke 0.16 " Volatile alcohol A trace --------- 100.00 "