remaining side a ditch has been cut, and behind this is an earthen rampart, faced with a wall 8 ft. or 9 ft. high, composed of huge stones in their natural shapes. Other intrenchments occur at Pilsbury, near Hartington, Calton, near Chatsworth, Camp Green at Hathersage; and on Fin Cop, Combs Moss, Hartle Moor, Staden Low, near Buxton, and Cronkstone, Alport, and Tapton Hills.

Pre-Roman Cave Remains.-Rains Cave, near Bras-. sington, which was excavated under the writer's supervision in 1890-1, yielded many objects of the British period anterior to the Roman occupation. The work proved that the cave had been used by man at various times; first as a dwelling-place, probably as far back as the Neolithic age, then as a burial place, then as a temporary residence, then again as a burial-place. The fragments of the pottery which accompanied the interments were of the ordinary British barrow type, hence they served to fix the period when the cave was used as a sepulchre. The animal remains of this cave furnish a very full list of the fauna of the county at the time :

Mr. Rooke Pennington obtained from two small caves in Cave Dale, Castleton, a similar series of animal remains, associated with rude pottery, flint flakes, jet, a perforated stone hammer, and a bronze axe of peculiar form and alloy, all presumably pre-Roman. The same gentleman obtained from a cave in Hartle Dale, near Bradwell, rude prehistoric pottery. A few pre-Roman objects have been obtained from the prolific cave, "Thirst House," near Buxton, by Mr. Salt, who has done excellent work there; and some of the things which he has recently found in several other caves and rock shelters around that

town, may be of like age. At Cresswell nothing appa

rently intervened between the Pleistocene and the Romano-British remains.

Pit Dwellings, Sites of Habitations, etc.-Very little has been done to elucidate this class of Derbyshire antiquities. Writer after writer has enlarged upon the "supposed site of a British town" at Linda-spring, near Crich, but no one seems to have thought of applying the evidence of the spade to the depressions. The sites of huts, presumably of pre-Roman age, have been recorded on Hartle and Abney Moors, at Over Oldhams, and Smerril Grange, near Youlgreave. The present writer excavated such a site on Harborough Rocks, near Brassington, and concluded that it was British of Roman age ; and Mr. Salt has in hand some curious and most promising pits on Ravenslow, near Buxton. It is obvious, however, that there must be many remains of this class in Derbyshire, a county so rich in prehistoric archæology.

Maenhirs, Rocking-Stones, Rock-Basins, - Altars, -Idols, etc.— At various places along the lines of the outcrop of the Millstone Grit, particularly in the vicinity of Stanton, Eyam, Hathersage and Ashover, are fantastic blocks and masses of rock, which were descanted upon by the older antiquaries as "Druidical." Some of these are rockingstones; others have been designated, according to their fancied resemblances, altars, idols, basins, etc. Most, if not all, of these are of natural origin, the result of weathering; but it is quite possible that some few were altered by man in ancient times. The well-known rockingstones and other blocks at Rowter, near Stanton, were 'improved" about a century-and-a-half ago by a gentleman who resided near. Standing-stones, which appear to have been artificially placed, are to be met with in the Peak, but as yet no antiquary has systematically investigated them. In the Buxton Gardens may be seen two stones, the one a "holed" stone, which were removed from a place in the vicinity of the town and set up in these gardens with much learned ceremony, some years ago; but they appear to have been gate-posts, and the hole has been formed naturally.

Romano-British Barrows.-About fifty barrows, interspersed among the pre-Roman series, have been found to

bear a strong family likeness, and various circumstances point to their Roman age. Their materials were wholly or mostly fine, as clays of various colours, sand, gravel, turf, and even moss, which whether used singly or several together, were usually laid in definite layers. These barrows seem to have invariably been raised over the site of the funeral pyre; and the human remains, cremated in every case, have been found remaining as left by the fire, or gathered into a little heap, and thus covered up. It was seldom that anything more than a few flint flakes and a potsherd or two rewarded the opener; and it has generally been observed that the potsherds were buried as potsherds, and not as perfect vessels. In no case did these potsherds appear to have belonged to vessels of the British sepulchral type; on the other hand, many were described as hard and wheelmade, and some as distinctly Roman.

Recently, Mr. Micah Salt laid bare two unburnt extended interments, which were associated with objects of this age, in Deepdale, immediately below the mouth of "Thirst House," and another similar interment of probably the same period has been discovered at Bradwell.

The Roman Occupation.-In 1768, Dr Pegge gave the world ("Derbeiesseira Romana," Bibl. Top. Brit., No. 24), a very complete survey of the remains of this period in Derbyshire; and in 1885 and 1886, the late Mr. W. Thompson Watkin contributed two articles to the Derbyshire Archeological and Natural History Society's Journal, which covered the same ground, only if anything with greater fulness. A comparison of these two writers shows that little progress had been made in this branch of Derbyshire archæology during the interval of 117 years, and little progress has been made since. None of the more important sites have been excavated in the interests of archæology; so that our information rests upon such superficial vestiges as have escaped obliteration, and such accidental discoveries as have chanced to come under the notice of antiquaries.

The Peak was undoubtedly an important lead-mining district during the occupation, and this is evidenced by the pigs of lead of that period found in this and the

neighbouring counties. The earliest recorded discovery of an inscribed pig in this county was on Cromford Nether Moor, in 1777; this was followed by another on Matlock Moor, in 1783; a third near the last, in 1787; and a fourth, also near the last, in 1894.

The sites of several stations are known, and their names identified with more or less certainty. Little Chester, which by a consensus of opinion was the Derbentio of Ravennas, was probably the chief Roman centre of the district. Buxton was certainly Aquæ. Two others, Melandra Castle, near Glossop, and Brough, in Hope Dale, have been ingeniously identified by Mr. Thompson Watkin as Zerdotalia and Navio. Very slight traces of the castrum of Little Chester remain, but in 1721 Dr. Stukeley could sufficiently follow up its outline to determine that it was rectangular, 500 ft. by 600 ft. Nothing is known of the castrum of Buxton; but remains of baths and other buildings were discovered there in the seventeenth and eighteenth centuries, on such a scale as to prove that the place was a fashionable resort in Roman times. Unlike these, the stations of Melandra and Brough are in fair preservation, the former being 336 ft. by 366 ft., and the latter 270 ft. by 310 ft., and at both places numerous remains have been dug up from time to time. Besides these, there was another important station, which was probably the headquarters of the lead trade, but its site is unknown. It has been identified with Chesterfield; but Mr. Watkin's theory that it was in the vicinity of Wirksworth has more in its favour. Several small camps of Roman character are pointed out: one is at Parwich, another near Pentrich; and some years ago the late Mr. Redfern, of Uttoxeter, traced what he believed to be such a camp at Ballidon, near Parwich.

The great Roman road which connected York with the Severn, known in later times as the Rykneld Street, threaded the county from the north-east to the southwest, passing Little Chester en route. Another great road, connecting Colchester and Leicester with Chester, known as the Via Devana, intersected the southern part of the county, crossing the Trent in the vicinity of Burton. Through the Peak passed a third road, which

connected Lincoln with Chester, threading in so doing the Derbyshire stations of Brough and Buxton. Another important line of road seems to have diverged from the Via Devana at Leicester, and passing through Little Chester and Buxton proceeded to Manchester (Mancunium), and the north-west generally. From Little Chester, a branch road ran due west through Rocester to Chesterton (Mediolanum), near Newcastle-under-Lyme, where it joined the Via Devana for Chester. Another took a more northerly course, and has been traced over the Chevin and in the vicinity of Belper-lane-end, pointing towards Cromford: probably it passed through the heart of the Peak country to Brough. If this be the case, the winding "Doctor's Gate," which can be followed between Brough and Melandra Castle, must be regarded as its continuation. The present highway between Derby and Tutbury seems to have been a minor Roman way. Probably it was connected with an ancient way, which has been traced on the south side of the Dove, to Uttoxeter. From Buxton there appears to have been a branch to Chesterton, passing through Leek; and another, with a course due west, making for the Mersey. A west road has been traced from Melandra Castle, which probably connected that station with Manchester.

A number of smaller vestiges of the occupation have been brought to light in various parts of the county, and serve to show how firm the Roman hold was. Hoards of coins have been found at Fenny Bentley, Alfreton (two), Crich Cliff and Moor, Pleasley, Fritchley, Culland Park, Ashbourne, Cromford and Heanor, during the last century and a half. The district around Middleton-by-Youlgreave has been prolific in remains of this period; and in less degree, also Oker Hill in Darley Dale, Chesterfield, Upper Haddon, and Eyam. Roman pottery was found upon the site of Duffield Castle in 1886, and upon that of ancient buildings at the foot of Rainster Rocks, near Brassington, by the writer in 1890. Several caves have yielded Roman objects, notably, Poole's Hole, Buxton; the upper deposits of the Cresswell Caves; the "Bat House," a fissure near Ambergate; while the array