NOTES ON GHOSTS AND GOBLINS. 431 



monly appearing, without reference to the deaths of 

 individuals, cases would happen pretty frequently 

 where a spectre appeared which was not that of a person 

 then dying, but of some near relative. I feel by no 

 means sure, however, that I have rightly caught De 

 Morgan's meaning.) ' But this,' he proceeds, 6 is, we 

 know, almost without example. It remains then, for 

 all who speculate at all, to look upon the asserted 

 phenomenon, think what they may of it, the thing 

 which is to be explained, as a connection in time of the 

 death, and the simultaneous appearance of the dead. 

 Any person the least used to the theory of probabilities 

 will see that purely casual coincidence, the wrong 

 spectre being comparatively so rare that it may be 

 said never to occur, is not within the rational field of 

 possibility. 



I have quoted this argument because it applies 

 equally well to the case of spectral appearances after 

 death. The right spectre is always seen, so far as is 

 known, and it appears always on a suitable occasion 

 (at least, an occasion as nearly suitable as the case 

 permits). 



It must be admitted, however, that this explanation 

 does not cover the facts of all ghost-stories. There are 

 some narratives which, if accepted in all their details, 

 appear to admit of no explanation other than that 

 which refers the events described to supernatural 

 causes. But it must not be forgotten that these 

 narratives have come in every instance from believers 

 in ghosts and spirits; and without questioning the 



