Mark was smoking his cigar outside the door of his house when he saw a man, a stranger, approaching him. Suddenly he ceased to be visible! Mark, who had long desired to see a ghost, rushed into his house to record the phenomenon. There, seated on a chair in the hall, was the very man, who had come on some business. As Mark's negro footman acts, when the bell is rung, on the principle, "Perhaps they won't persevere," his master is wholly unable to account for the disappearance of the visitor, whom he never saw passing him or waiting at his door--except on the theory of an unconscious nap. Now, a disappearance is quite as mystical as an appearance, and much less common. This theory, that apparitions come in an infinitesimal moment of sleep, while a man is conscious of his surroundings and believes himself to be awake was the current explanation of ghosts in the eighteenth century. Any educated man who "saw a ghost" or "had a hallucination" called it a "dream," as Lord Brougham and Lord Lyttelton did. But, if the death of the person seen coincided with his appearance to them, they illogically argued that, out of the innumerable multitude of dreams, some _must_ coincide, accidentally, with facts. They strove to forget that though dreams in sleep are universal and countless, "dreams" in waking hours are extremely rare-- unique, for instance, in Lord Brougham's own experience. Therefore, the odds against chance coincidence are very great. Dreams only form subjects of good dream-stories when the vision coincides with and adequately represents an _unknown_ event in the past, the present, or the future. We dream, however vividly, of the murder of Rizzio. Nobody is surprised at that, the incident being familiar to most people, in history and art. But, if we dreamed of being present at an unchronicled scene in Queen Mary's life, and if, _after_ the dream was recorded, a document proving its accuracy should be for the first time recovered, then there is matter for a good dream-story. {8} Again, we dream of an event not to be naturally guessed or known by us, and our dream (which should be recorded before tidings of the fact arrive) tallies with the news of the event when it comes. Or, finally, we dream of an event (recording the dream), and that event occurs in the future. In all these cases the actual occurrence of the unknown event is the only addition to the dream's usual power of crumpling up time and space.