the stones are placed, in order to find the entrance to the chamber, which almost always faces the east, or south. It is by the entrance, in which corpses are not unfrequently deposited, that he must arrive at the chamber, which is always filled with loose earth, and a mixture of clay and pebbles. It is as the chamber is being carefully and thoroughly cleaned out, that the enquirer generally perceives how the corpses and antiquities have been deposited. Skeletons, and in particular sculls, must be preserved; and even the bones of those animals, which have been interred with the deceased may have a value for science. It must also be observed whether there is not a second stone chamber, in connection with the first. If, on the other hand, the barrow contains no stone chambers, the grave itself must always be sought at the base of the barrow, and this is usually surrounded by a stone cist, or covered with a heap of stones. In this case we must, on no account, remove the stones at once, because the objects in the grave might easily be injured by so doing. With the view to obtain space for the investigation, the best plan is to enlarge the cavity as much as possible, and to take off the cover of the tomb as soon as the earth which lies upon it has been wholly put aside.
It is generally considered of importance, in such excavations, that the cavities should not be made too narrow, in fact they ought to be very wide at the top, because it is always necessary to make them narrower in descending, with the view to prevent their falling in.
Finally, if it is a Cromlech, with one or more chambers of stone, the examination is rendered much more easy by clearing the chambers from the earth; in which operation one proceeds exactly the same as in excavating the chambers of stone, in the interior of the barrow. The earth surrounding the Cromlechs must also be examined, to ascertain if vessels of clay with burnt bones, and ashes, are to be discovered there. For all possible contingencies in excavating barrows, of course no exact rules can be laid down. But should there be dis-
