Praetorium straight to the N. gate. Again, the arrangement of the wooden barracks in the retentura undoubtedly points to the existence of a now vanished street that had passed from the S. gate to the back of the Praetorium. Besides these four, the only other which we can identify with certainty was one that ran round the interior margin of the southern defences, occupying (so far as that side of the fort was concerned) very much the position held by the Via sagularis in the temporary camp of Hyginus. Its remains are shown on the right in fig. 21, at the spot where they were most extensive. The post-holes on the left of the illustration belong to the S. front of Barrack Block No. V., the view being taken from the W. The street itself was 7 to 8 feet wide. Towards the W. its outer kerb was about 17 feet behind the inner kerb of the stone base of the rampart. Further E. the corresponding interval was only 15 feet. At its western end, if not also at its eastern one, there were faint indications that the street may have rounded the corner with an easy curve. Was it continued along the line of the defences on the other three sides of the fort? To this question no positive answer can be given. If it was, then we can see that, after crossing in front of the N. gateway, it must have swung slightly southwards, so as to leave the Baths and Latrines in what Hyginus calls the intervallum.

(h) Fireplaces.-Remains of rude hearths or fireplaces were found in various directions throughout the fort. Many of these must mark the site of the camp-fires that warmed the soldiers' quarters. Probably there was one sunk in the floor of each of the compartments into which the wooden barracks would be divided. But they could not all have been of this nature. Among the most notable exceptions were three that lay in a row, close to the rampart on the W. side of the fort, about midway between the gate and the S.W. corner. Built of stone, they were circular in shape, 7 or 8 feet in diameter, and about 3 feet high. They had evidently been much used, although nothing survived to suggest their real purpose. Whatever that purpose may have been

and it might be anything from the kindling of signal-fires to the consumption of rubbish-it is odd that they should have been placed just where a westerly wind-the prevailing wind in the district-would drive the smoke straight on to the Praetorium and the blocks of barracks in the retentura.

The most remarkable of the fireplaces was, however, a circular recess, cut into the W. side of the outer ditch, 21 feet N. of the W. gateway, and walled with solid masonry. A good idea of its appearance is conveyed by fig. 22. The floor was of boulder clay, and on the same level as the bottom of the ditch. The dimensions were as follows:

Beneath the line of the scarcement the recess contained a large quantity of red ashes, above which were loose stones and soil. Ashes of a similar character were lying in abundance in the ditch outside; they had evidently been drawn from the fireplace, when it was in use. It should be added that the mass of loose stones was so considerable as to show that the building had originally been much higher.

One's first impulse is to regard the circular recess as an oven. Four ovens were found in the body of the rampart at Birrens,1 and a like number at Inchtuthil. Or the red ashes might suggest that it had been a kiln for drying bricks or tiles. There are serious difficulties in the way of both of these views. It would be strange indeed if the garrison had been dependent for the baking of their daily bread on an

1 Proceedings, 1896, vol. xxx. p. 101.

2 Proceedings, 1902, vol. xxxvi. pp. 209 ff.

3

3 For kilns near the S. gate at Amboglanna, Aesica, and Housesteads, see Arch. Ael., xxv. pp. 282 ff.

oven to which no access could be had except by going outside the defences. On the other hand, had it been a kiln, one would have expected to find fragments of bricks or tiles or pottery among the debris. Such fragments were conspicuously absent, and the redness of the ashes may well have been due to the dye-stuffs from the wood that served as

fuel. Again, the ruins, considerable as they were, furnished no evidence that flagging had been laid across the scarcement, no sign of the usual provision having been made for draught and for the smoke from the fire. These last objections are not perhaps insuperable. Flagging and the stones of a dome might quite well have been carried away by those who plundered the site for building material. Or the kiln might have been

used for drying corn, in which case the fire would be a smouldering one, so that the top would be left open, and planking might suffice as a bridge.

Against all these suppositions there is one most powerful argument the intense degree of heat to which the whole recess had been subjected. The actual furnace was large out of all proportion to what would be required for a kiln or oven of similar dimensions, and the stones that formed the wall were burned red, deep below the surface. In these circumstances, a suggestion that originated with Mr M'Intosh deserves to be carefully weighed. His view is that the recess was employed for cremation that it was, in fact, the ustrinum of the fort. He supposes that the fireplace below the scarcement would be piled up with dry fuel, and that the body would then be lowered on a strong hurdle of green wood, the ends of which would rest upon the ledge provided by the scarcement. It will be noted that the diameter (7 feet) suits this hypothesis exactly. The fire would be kindled and fed from beneath, and the hurdle would support the body until it was wholly consumed. It is a matter for regret that it was not found possible to test this theory by having specimens of the ash microscopically examined. An opportunity for that may occur elsewhere. In the meantime it is worth pointing out that, if cremation was to be effective, some such system as has been described would be essential. A body would not be consumed by being merely thrown upon a blazing fire. Nor is there much force in the objection that a ustrinum in such close proximity to the fort would be offensive. The walls, be it remembered, were high. If the fire were fierce and the consumption rapid-as it would be in such circumstances-no smell would be observable. Except when the fire was freshly lit, there would be little or no smoke. Besides, have already found great hearths much nearer to the soldiers' quarters.

we

(i) Refuse-Holes.-Within the ramparts of the fort the excavators discovered nine rubbish-pits or refuse-holes, all of which were thoroughly examined. The position of each will be found marked upon the Plan