see them lying on the ground, not to take them away. I have brought away a few for a certain purpose; they are valuable for the sake of illustrating my subject, but beyond that they are much more interesting on the spot where the ancient Britons, or the Romans, or the Saxons left them.

SEATON DOWN.-Returning from Hawksdown Hill across the valley of the Axe, about two miles westwards, we light upon Seaton Down. Suppose a person travelling on the road between Exeter and Lyme. On the crown of the hill just before descending to Colyford, there is a sort of spur that runs away north on the left hand side; at its furthest end, where it is in its wild state, a ditch and agger have been carried east and west across the ridge, extending to the length of 770 feet. The slope of the agger is 33 feet. The ditch is on the south side, towards the mouth of the river Axe, as if an invading enemy were expected from that quarter. As if this defence were not enough, a second of a similar nature had been begun 466 feet to the rear of it, 130 feet long, and left unfinished. These works are very similar in their nature and object to those which traverse the ground at the Three Horseshoes, presently to be mentioned, and seem to have been intended to guard the road, and to oppose the passage of an enemy coming up from the valley of the river Axe. The completion of the second vallum was relinquished, perhaps, because the makers may have been attacked and driven out, or, perhaps, because the invaders may have marched off in another direction. Possibly these things may have occurred in 937, when Athelstan successfully opposed an inroad of the Danes in the valley below.

HONEYDITCHES.--A mile south of Seaton Down lies Honeyditches, or Hannaditches. On the east side of the road there is a long, narrow, curved field leading to a square field, in which latter are the remains of an extensive Roman villa. The long, narrow field is supposed to have been the original approach to the villa. The foundations of walls, crossing each other at right angles, begin close under the hedge at the top of the field, to the width of 40 feet north and south, and run downwards toward the east 145 feet. In the field above this there are some great pits, as if they had been reservoirs of water for the use of the house. About 200 feet below the villa, connected apparently by a drain or a wall, there is a rough piece of ground, measuring 48 by 56 feet. These places had been examined before by Sir Walter Trevelyan, the owner of the land; but Mr. Heineken and myself turned up some large, thick tiles, an inch and a quarter thick, eleven

inches wide, but of uncertain length, as they were broken. The under edge had been chipped or bevelled off by the workman when he bedded them; and as they were mostly found apparently at the bottom of a cavity measuring about two feet by three, accompanied by traces of charcoal, it is supposed they had formed some portion of a furnace, oven, or hypocaust. We also found flanged roof tiles, and mortar mixed with pounded brick. Besides these evidences of Roman occupation, many evidences of much later occupation have been discovered, especially in the upper part near the hedge, such as mediæval tiles, thin pieces of lias from the cliffs towards Lyme, where the lias crops out, with holes through for the pegs by which they were fixed to the roof; also pieces of roofing slate, with holes for the pegs; and this is probably a still later evidence than the thin pieces of lias used for the same purpose. One fragment of tile is impressed with groups of parallel lines with traces of letters. It is curious that the two groups of lines on this fragment are not parallel to each other, but converge to a point; and the letters on the space between them converge to a point too; that is, they begin large and diminish towards the end. The first portion looks somewhat like the letters Mar, the rest being broken off. A friend suggests that perhaps there may have been a chapel or ecclesiastical building there during the middle ages, and that possibly the word may be intended for María.

But most of our old writers on Devonshire antiquities speak of Honeyditches as an old camp nearly circular, but unfinished on its western side, and that perhaps it was thrown up by the Danes when they landed in the memorable year 937, as before observed. From the situation of the place that now goes by that name, and from the objects exhumed there, no one can infer that this was a Danish camp, or anything of that nature. The conclusion therefore at which we may arrive is this, that the original Honeyditches (the old camp) was somewhere else in the neighbourhood, probably not far off, and that the name has been shifted or transferred from one place to another. Possibly it may have been on Coochill or Little Coochill, half a mile south-west, on the crown of which there is a peculiarly shaped field bearing traces of a fortified position. Quantities of stones were dug up and removed from this spot in or about 1862, and one of the men employed in so doing declared that the stones lay in lines as if they had been thrown into trenches and covered over, or followed the course of walls. Or it may have been on some hill nearer to the mouth of the river Axe; for some speak of

it as having been at about three quarters of a mile from that spot, whereas Coochill is nearly double that distance.

EARTHWORKS.-In my paper read before the Archæological Association at Exeter in 1861, as before observed, I mentioned the traces of a ditch and agger behind the Three Horseshoes, a wayside inn on the road from Honiton through Roncombe Gate to Colyford. It begins in a field behind the inn, and runs northward for more than 1000 feet to the declivity of the hill, where it turns eastward by a rounded corner. At that time this is all I knew of it; but since then opportunities have occurred of examining a continuation in the opposite direction for nearly another 1000 feet, until it approaches the valley on that side. On consideration this must appear a very remarkable work. If we trace it from the north end at the rounded corner, which is nearly in front of Blackbury Castle, it runs in a direction somewhat to the west of south for about 2000 feet, right through the position of the Three Horseshoes, though at this spot of course it is obliterated, but the ridge is continued in the fields below. An old man living near, who recollected the land in its wild state before it had been brought into cultivation, declared in my hearing that at that period the ridge was from twelve to fifteen feet high, and that the ditch was on the east side of it-that is, the side towards Colyford. At first this appeared very strange, because it put the ditch on the inside of the corner. On reconsideration, this vallum could not have formed any part of an ancient camp. It had been drawn across the top of the hill at right angles to the public road; and the ditch being on the east side, or the side of the enemy, may lead to the inference that this work was made for the purpose of keeping at bay or checking the advance of some force expected from the valley of the Axe. As it is just opposite Blackbury Castle, possibly it may have been thrown up by the occupiers of that camp; perhaps by the Britons to resist the Romans; perhaps by the Saxons to resist the Danes; and it might be at the same time when the similar intrenchments were drawn across Seaton Down.

I may here observe that the field opposite the Horseshoes is called "Chapel Close." A few paces from the west hedge, and at 72 from the north one, the plough had often been obstructed with stones, so an excavation was made, June 17th, 1862. I saw the south-west corner of a building laid bare. The walls were three feet thick. Perhaps some mediæval chapel may have stood there. The next field, on the west of this is known as "Chapel Meadow;" and near the

middle of this, and not far from the road, stones and traces of walls have been met with.

IRON PITS.-Several of our local writers have spoken of the existence of pits of various sizes and depths met with on the wild tops of many of the high hills in this neighbourhood, but they all seem to speak from hearsay only. I am happy to say I can speak with more confidence. Where these pits have not been obliterated in the process of cultivation, they occur on the Blackdown range of hills, Ottery East Hill (just over Lincombe Farm), on Dunkeswell Common, and other places. The nearest spot to Honiton that I know of is a short distance beyond Woolford Lodge, and of these I will speak more particularly. The way to find them from Honiton is this: Go to Coombe Rawley; then ascend the hill towards Woolford Lodge, and pass the entrance gate; a little way beyond this the four-mile stone from Honiton is seen on the right hand side, and a few score yards beyond this is a fourcross way. Go straight on. Take the second field on the left. The field is full of fern and furze, still in its wild state. The pits occur mostly along its northern side. They are of various sizes, very irregular, and mostly close together. Though their sides were perpendicular when first dug, they have fallen in by time and become sloping. Some are very large. As an instance I may mention, that being once there with a friend and a one-horse carriage, and not wishing to court the idle curiosity of passers by, we led the horse and carriage down into the bottom of one of them, whilst we made an examination, and we were all quite out of sight to any person near. In the geological maps all these hills are described as of the greensand formation; but above the greensand there is the usual stratum of flints and clay, and above this a subsoil bed in which the iron ore is found. It is what is called surface iron. It may seem rather strange that they should have sunk so many separate pits: one would have thought that it would have been better to have begun at one end, and to have dug onwards straight through. It is in these places that the ore is found: the smelting operation was performed elsewhere. Great quantities of scoria and cinders have been discovered at different spots of the Blackdown district, showing where this process was performed. There is a large heap at Clivehayes Farm, Churchstaunton : a quantity once existed at Bowerhayes Farm, near Dunkeswell Abbey; some more in a field at Tidborough, near Hemyock; and in less quantities at Kentisbeer, Culmstock, Uffculm, and

so on.

ON THE PSEUDOMORPHOUS CRYSTALS OF CHLORIDE OF SODIUM, & THEIR OCCURRENCE IN DEVONSHIRE.

BY G. WAREING ORMEROD, M.A., F.G.S.

THE Occurrence of the pseudomorphous crystals of chloride of sodium in the Trias of England was, it is believed, first noticed in a paper by myself "On the Salt Field of Cheshire," read before the Geological Society on the 8th March, 1848 (Quarterly Journal, vol. iv. p. 273), when they were shown to occur in the "Waterstones" of the Keuper in that county. The specimen then exhibited had been analyzed by Professor Grace Calvert, of Manchester, who stated that it was "silicate of protoxide of iron that had replaced the chloride of sodium." The specimen attracted considerable attention, and the late Professor Buckland mentioned various places at which he had noticed the pseudomorph, for which he had not been able to account. On December 1st, 1852, a paper on "Pseudomorphous Crystals of Chloride of Sodium in the Keuper," by the late Mr. Strickland, was read before the same society (Quarterly Journal, vol. ix. p. 5), in which he entered into an elaborate description of them and their supposed origin, regarding them apparently as a new discovery. This paper is worthy of a careful perusal; the author had paid great care to the subject, and his ardour in its investigation caused all geologists to lament the loss of a leading and powerful mind. He was killed in the autumn of 1853 by a passing train, when examining the cuttings at the mouth of a tunnel near East Retford in search of this pseudomorph. On April 20th, 1853, a paper by myself on the same subject was read at the Geological Society (Quarterly Journal, vol. ix. p. 187); it contains particulars of places in England and America where this pseudomorph has been found, and to it Professor Warrington Smyth added a note, mentioning that in Leonard and Bronn's Journal of 1847 it was mentioned as occurring