A PREHISTORIC SITE AT KIMBLE. 437

A PREHISTORIC SITE AT KIMBLE, S. BUCKS.

BY A. E. PEAKE, M.R.C.S., L.R.C.P.

Read at Norwich, January 8th,

The site which is described in the following paper was discovered by Mr. E. Payton, B.A., of , Bucks., in May, 1913, and I had hoped to have his co-operation in describing it and the fine series of implements which he and Mrs. Payton (members of this Society) and I have accumulated from it. The war, however, has prevented Mr. Payton from carrying out his portion of the task, and I shall, therefore, include a description of some, only of my own pieces, hoping that the fine series now deposited by Mr. and Mrs. Payton in the Hambleden Museum will be described and illustrated later on, and amplify the large list of pieces now recorded. The site is on the land of Kimble Farm, on the borders of Bucks and Oxfordshire, and about midway between the villages of Fawley and Scuth End. It is on the high ground of the Chilterns, about 600 O.D. The site of the factory is a narrow neck, which falls towards Fawley, and rises towards Southend, where the ground lies at 642 ft. O.D. It is ground typical of the Chiltern country the land being here split up into dry valleys which have their starting point from the ridge mentioned above. The site is eminently suited to form a settlement though no camp can be traced, as direct access to two of the great lateral valleys which lead into the Thames is obtained from this spot. On the south-west side is the commencement of Fawley Bottom, which opens out into the wider valley which runs from I through S16nor and Assenden to Henley-on-Thames. €- / On the north-east is one of a series of dry lateral valleys which opens out to the great cleft known as the Hambleden Valley, near the present village of . A tribe settled at Kimble would, before the present roads were made, have an easily graded trackway into these main avenues, both connected with the water- way of the Thames and probably important trade routes in earlier times. The geology of this part of the country does not throw much light on the site. The whole of the high ground is, of course, part of the chalk of the Chilterns. The superficial deposits include clay with flints—plateau and pebbly gravel—and Reading beds. The whole of the high ground from Middle Assenden to Fawley, with 438 PROCEEDINGS OF THE PREHISTORIC SOCIETY. the east part of South End and Court is covered with clay with flints, spurs extending to Great Wood and Bacres, to the hills above Skirmett. It is from this deposit, made up by the decomposition of the chalky surface, and partly by the denudation of the overlying Eocene beds that the flint used on the site is derived. At Tower Woodhouse Farm, about two miles to the south, is a patch of pebbly gravel more directly derived from the Eocene beds. To the west is a large area covered with Reading beds, the nearest point to Kimble being Coxlease, about two miles away. The factory is on clay with flints, but the deposit is not many feet in thickness, as three pits lately filled in occupy parts of the field from which chalk was obtained to " marl " the surface. I cut into one of these pits to ascertain if it was prehistoric, but could not reach the bottom, and had not then got through the modern "filled in "at io ft. The implements from the Kimble site fall into various classes. Of the perfect forms some are single examples. Of others there are duplicates or many examples. It may well be called a cone culture, and in some measure falls in line with those described from the Wensum Valley in " Pro- ceedings," Vol. II., part 2, by Messrs. W. C. Clarke H. H. Halls and J. E. Sainty B.Sc. (Hellesdon, Sparham and Lyng). The cones recorded from these three sites number 183, viz. : 120, 52, 11. This falls short of those found by myself at Kimble, viz. : 130, and Mr. and Mrs. Payton have probably an equal num- ber. Another point of distinction is the fact that Kimble is a pigmy site. There are, too, some very definite tortoise cores, a few flake implements, and a number of plungers, but the flakes wi:h facetted bulbs are very few in number. An interesting group of pieces not recorded from the above sites, and which are very rare at Grime's Graves, and non-existent at Peppard, are pieces with facetting, which are clearly struck from the side of a prismatic core to obtain a new platform. Not the least interesting are a series of beak- ended implements which are reminiscent of the rostro-caiinates from under the Red Crag. There are seven varieties of flint used in making the implements. These include three of black flint, one of which is pure, the others having chert (white) in patches or entirely diffused through it. There is also a honey-coloured, a light trans- lucent flint, a grey flint with bands of chert, and a light grey flint with chert diffused through it. All these varieties are found in the gravel on the surface. From an examination of the cones it is clear that the Kimble site is closely related to Sparham, Lyng and Hellesdon. These are also surface zones. From what I have seen of Dr. Sturges' large series from a site near Icklingham, he has there the full development of this particular culture, with completed pieces in far greater numbeis, a full account of which I hope he will one A PREHISTORIC SITE AT KIMBLE. 439 day publish. Our chief difficulty on the Kimble site lies in deter- mining which of the pieces belongs to this culture, and which are earlier or later. It is easy to embrace in one period the majority of the cones, pigmies, flakes, and some of the celts, picks, prismatic tools, discs and drift forms, but hardly any of the scrapers have the same patina— a bluish white with lustre—and the same may be said of the typical tortoise core. On the other hand many of the rostro-carinate group—the plungers and the flakes with facetted platforms—cer- tainly belong to the same period as the cones. In the seven varieties of flint used on this site none has come directly from the chalk or from the large masses in the clay with flints derived from it. This is probably due to the fact that the Hambleden Valley is cut through the middle chalk, which does not contain suitable nodules. The flint is derived from the spread of plateau gravel or the neighbouring Eocene beds which contain all the varieties. The absence of mined flint is in favour of a pre- neolithic date. In determining the value of the pieces made from flint other than that patinating blue and white, it is well to bear in mind the undoubted fact that if two pieces are chipped at the same time from two different varieties of flint, though exposed to the same conditions the patina will be different. It is equally difficult to consign the flint work to one particular period by taking form, style, and finish, etc,, as the guide, as there is no known site in which all these forms occur together. The fuller knowledge we now have of the Grime's Graves cul- ture (if this is of Neolithic date), has entirely revolutionised the classification of flint implements, resting as it does largely on form. Have we any established site of Neolithic date to compare with, and what really constitutes a Neolithic date ? If the Grime's Graves implements are not to influence the date, it is perfectly clear that form is no criterion, the other factors pointing to a Neolithic date—domesticated animals, polish and pottery are all absent from this site. The fact that these implements are found on the surface, whilst favouring a post-Palaeolithic date do not preclude a palaeo- lithic one, as pieces from palaeolithic times downwards are found on the surface, where they have accumulated. At the present time it is extremely difficult to assign a piece or a group from a site to the Neolithic period, because we have no data to go upon. The Cissbury culture sites have a distinct facies which shows a very marked distinction from that at Kimble. It is probable that the prismatic core period will be found just as prolific in forms as Grime's Graves when it is fully studied, but though some of the forms may be the same on a small scale there are others which are absent from the former. I have never seen a typical Moustier racloir from the former period, such as is found at Grime's Graves. The chief difficulty in comparing two sites lies in seldom having a complete series to form the basis for comparison. Cores and lumps and flakes are often discarded, and the same fate meets the 440 PROCEEDINGS OF THE PREHISTORIC SOCIETY. rougher pieces, without which no true comparison can be made. The three sites mentioned above, viz.: Hellesdon, Sparham and Lyng, present some interesting parallels and differences. Classes 1-4 of the cones vary a good deal in their respective numbers between the Hellesdon and Kimble sites, especially in the subdivisions of Class 3. The main distinction is in Class 5, where the whole circumference of the base is used. The numbers there are 9 and 23 respectively. Another point between the two sites is the presence of pigmies. The Hellesdon site yielded only one, and Sparham one, whereas at Kimble some 10-11 examples occurred. It is probable that both sites might yield far more. These little implements are diffi- cult to find at any time and under the best conditions, and I was rarely able to visit the site under the best conditions. The most marked difference is the presence of 76 flakes worked to a square end opposite to the bulb on the Hellesdon site, and of 19 rostro-carinate implements from Kimble. The scrapers from Kimble number about 350, but this includes a large series of flakes with the edge used for scraping. At Hellesdon there were 87. The proportion (34) of long flake scrapers at the latter site is far greater than at Kimble, and there are a few, small and nearly circular, which might be classed as thumb-scrapers. The dos rabattu knives, excluding pigmies, are far fewer at Hellesdon. Most of the other forms mentioned from the latter site occur at Kimble, except a circular knife chipped on both faces, and a parallel-sided one also chipped on both faces with rounded ends. I have nothing of this kind from Kimble. They belong to a class which is usually found in barrows, or has turned up on the surface where one of these has been ploughed over. Another form from Hellesdon is a series of 68 flakes of triangular form trimmed to a central point, resem- bling a form of burin. I can find none of these in my series from Kimble. No arrowhead is recorded from Hellesdon, but a fine willow- leaf form from Sparham. The only piece of this form from Kimble is chipped on one face only, and is, I think, a knife. Mrs. Payton has, however, an arrowhead from this site. The Kimble site has yielded far more celts or axes than either Hellesdon or Sparham. The Hellesdon example is figured (17), and the peculiarity is the sharp point like an implement from the Oldbury Rock shelter. There is no parallel from Kimble. Three examples are recorded from Sparham by Mr. Sainty. The piece figured (18) is somewhat smaller than the two of similar shape from Kimble, and of about the same thickness, but the two faces are more equally convex. The second axe is combined with the cone, and of this form I have a good example; one end of this implement from Kimble has a flat platform chipped steeply up the face of the shaft of the axe. The axe end in this case has on one face a transverse facet ; trancet type. I regard it as an unfinished piece. A PREHISTORIC SITE AT KIMBLE. 441 No parallel to the polished one from Sparham has come from Kimble.

DESCRIPTION OF IMPLEMENTS. The following forms are described, and many of them figured :— Various celt-like implements, some with battering at the end showing that they were used with a hammer stone. Picks of small size, some of which belong to the class known as fabricators. Pris- matic tools, an adze plane, pigmy dos rabattu knives, No. I (Sturge) (12 examples), drift forms (3 examples described), discs merging into the degenerate hand-axe, tea-cosies and percuteurs, struck and unstruck tortoise cores and flake implements with facetted plat- form. Scrapers, both blade, medium and short-end scrapers— side and hollow scrapers, and varieties with spurs, slug planes, a series of plunging flakes, some of which are corner ones, others being failures to make small flake implements ; a few flakes with facetted platform, and a series struck from the base of the core to freshen the edge. A series of cones which can be described in six classes, a dos rabattu series (30), and some 19 rostro-carinate imple- ments. The latter have not, I believe, been describad from a surface site before

CELTS. (1.) Fig. 90 E.—A well-made implement with two convex faces, with bevelled rounded ends, one better finished than the other, and a cutting edge all round. The work is done by short flaking from the edge at right angles to the long diameter of the implement. White, iron-stained, 5f by 2J. It is a Cissbury celt, approximating most to Fig. 10, Plate XXIV., Vol. LXIII. " Archaeologia," but the sides are not so paral- lel. With this exception and two others all the eighteen other examples figuredther e have one end much more pointed. Sir J. Evans (" Ancient Stone Implements ") figures a celt of similar form from Cissbury and an example from Mildenhall, but there is no other instance given of this form either amongst those entirely ground, or with the edge partly polished. (2.) Fig. 90 C.—A neat little celt with two slightly convex faces, the broader end sharpened, and having, as is often the case, one corner squared and the other rounded. The other end has lost its point. The work is done from the edge at right angles, but it is less regular than the last with more primary work left on the faces, 3J by if, white, iron-stained. It corresponds with Fig. 8, " Archaeo- logia," Vol. LXIII., Plate XXIV., from Cissbury, but the cutting edge of this is rounded. In shape it is very similar to a small celt from Santon, found by Mr. H. H. Halls, of Norwich, but is thicker and more symmetrical. A small celt, 3J in. long, not figured, has a rounded cutting end worked on both faces, and the narrower end unworked. 442 PROCEEDINGS OF THE PREHISTORIC SOCIETY.

Fig. 90.—Implements from Kimble, S. Bucks, A PREHISTORIC SITE AT KIMBLE. 443 (3.) Fig. 90 G.—This and the next example are a good pair, and are clear evidence that implements were chipped to type. Sides parallel, one end, better finished, rounded, made with fan work. Both faces well chipped with work at right angles to the edge. One face much better than the other, which is ridged (4^ by ij in.). Unpatinated, lustrous, much iron-stained. (4.) Fig. 90 B.—Very similar to the last, but with a twist. The broader end has a transverse facet from the more convex face, with fan work on the other, and the pointed end has a symmetrical beak, slightly blue, much iron-stained, 4§ by if. Both of these celts are of a form very common from the surface, and usually called Neolithic. Sir J. Evans figures one from Oving, near Chichester (Fig. 15), and two from near Mildenhall (Figs. 12 and 13). He gives a list by sites in Wilts, Dorsetshire, Sussex, Kent, Cambs, Norfolk and Suffolk. It does not follow, because they occur on the surface, that they are Neolithic, and their place of manufacture has not yet been recorded. It is clear that they belong to the same culture as the planes or cones which are so numerous on the present site, and also the pigmy implements and other forms described, and a general consideration of all the imple- ments from this site may help to fix their horizon. (5.) Fig. 90 H.—An interesting example of a small round-ended chisel (two views) with the pointed end much battered. It is evidently a hand tool, the battered end showing where the hammer struck it. The broad end is worked on both faces, but the rest on one face only, 3J in. long. Sir J. Evans figures (348) an implement from Sawdon, very similar but smaller than this, and suggests that the bevelled end was the handle. The present example would, however, make an excellent punch. A portion of another implement came from the same site, and has the same battering of the end, and one face left plain on both. The edge near the point is much abraded, and may have been used to press off flakes. Both of these pieces are blue-white (iron-stained), and belong to the cone industry. Sir J. Evans remarks that these fabricators or flaking tools are especially common in Yorkshire and parts of Suffolk where arrowheads are common, and were used in their manufacture. The present site is not, however, an arrowhead site. Mrs. Payton found, I believe, two finely made ones at Kimble, but they are exceedingly rare. I have about a thousand pieces from the site, and have not found a single example after a minute search. It is possible, however, that they were used to press off the tiny flakes from the pigmy dos rdbattu knives presently to be described. (6.) Allied to the last are four pieces which may be regarded as picks, but which Sir J. Evans includes under fabricators. The present example (Fig. 90 F) is 3! in. long, of triangular section. The handle is thick with a patch of 444 PROCEEDINGS OF THE PREHISTORIC SOCIETY. crust, and at the extreme end is a portion £ in. square battered by frequent blows. The pointed end has flakes removed evidently from contact with a hard surface and is consequently a blunt cone. Both faces and the side edges are well worked. It is similar to Fig. 347 (Evans : " Stone Implements") from Bridlington. Lustrous, iron-stained, slightly blue, it seems to be quite a common type at Cranwich (see Halls, Vol. I., Parts 3 and 4, " Proceedings"). Mr. Lingwood has an example from Canon Greenwell's pit, Grime's Graves, Vol. I., Part 4, E.A.P.S., and Mr. Warburton other examples. Dr. Sturge has quantities of these picks or hand chisels from Norfolk and Suffolk. It is a rare form at Grime's Graves, and I cannot recall a good example either in my own work there, or in the 1914 digging. (7.) Fig. 90 D.—A form which combines the round-ended celt with the dos rabattu, 3J in. long. The rounded end is made by striking off a transverse flake from either face in the trenchet style. One side is chipped to form a wavy edge. The other is steep, 1 in. in thickness, and steeply worked, apparently to form a handle. The point is missing, but both faces are well worked. It thus resembles in all its points the implement figured from Grime's Graves, Fig. 54 B, Vol. II., Part 2 (" Proceedings "), found by Mr. H. H. Halls, white cherty flint, much iron-stained. Another piece that shows battering at the end is a long thick flake of triangular section. The bulb end is heavily battered, evidently by blows from a hammer, and the other end is broken from use as a punch. In this case the edges are steeply and regu- larly flaked, and the piece belongs to the prismatic cone series. Lustrous, slightly blue, iron-stained, 3 by J in. The side battering is well seen in a polygonal flake, 2\ by 1 in., whitened and iron- stained, which has a sharp, rather gouge-like end. The bulb end, and the edge near it is considerably battered all round, but not in the centre of the cupped platform. In this case it looks as if the margin of the platform and its adjoining edges had been used for pressure work, as hammering would certainly have abraded the centre of the platform as well. These battered pieces are thus very interesting. It is significant that not a single one from Grime's Graves shows this battering of the end of an implement. The prismatic tool is represented on this site, one example having the handle sharpened to form a steep faced plane. The edges are generally steeply battered in this class of implement, and the point broken. I have good examples from Rotherfield Greys, Peppard and Suffolk. It occurs, too, at Northfleet rarely, but is common in the French caves. At Grime's Graves we have found about forty examples, one of which is figured in colours in my Presidential address. Many are broader and bigger than is usual, as one would expect on a factory site, and shows that the edges were re-sharpened from time to time, and thus reducing the size of the implement. A PREHISTORIC SITE AT KIMBLE. 445

(8.) An exceedingly rare form also belonging to the prismatic cone culture (Fig. 90 A). It may be described as a small one-faced adze or plane. It has a rounded steep end, with angular corners, steeply fluted sides, an irregularly ridged back, and one face entirely unworked, 3J in. long. The nearest approach to this form that I have seen is Fig. D, PL LXXV., Vol. L, Part 3, " Proceedings," found by Mr. H. H. Halls at Santon Downham, whitened, bluish, some iron moulding.

PIGMY IMPLEMENTS. This site at Kimble is essentially a cone culture, and thus can be compared with those described by Messrs. Clarke, Halls and Sainty from the Wensum Valley. The most decided signs of use on these planes bear out the contention I have made in several papers, which Mr. Clarke, Mr. Reg. Smith and others support, that they are mostly implements in themselves, chipped into shape with as much intention as a celt or an arrowhead. This applies to a large number of cones where the flakes struck off are too small and irregu- lar to be of use, and especially on those sites where there is no evidence that small flakes were used, notably Grime's Graves, Peppard, Ringlands, etc. On the Kimble site the question arises whether all these cones are intentionally made planes or cores adapted or used for some such purpose after a number of serviceable flakes have been detached from some part of their face. There can be no doubt that the long narrow flakes struck off from these cones or cores were serviceable, and I have been able to collect some twelve undoubted examples of dos rabattu knives (No. 1 Sturge), indicating clearly the use to which these flakes were put. In addition to this I have some fifty flakes struck from the faces of these cones which are quite serviceable as knives or borers, or even needles. Mr. Sainty mentions that he found at Sparham a pigmy of reverse type, with a battered back and sharp crescentic edge, 11 in. long. In view of the enormous number of cones from each site, especially Kimble, it seems hardly likely that pigmies were made very extensively. However, it must be remembered that these small implements are very difficult to find, and only when the sur- face of the field is washed well with rain, and it is highly probable that a systematic digging over of the surface would yield far more. The two best examples have one end broader than the other. Both are ridged flakes which have lost the bulbs. In neither is the ridge central, and the narrower side of the ridge is battered up the face in each case for the whole length of the piece, the opposite edge being sharp for use as a knife, of which use there is evidence. The one patinated white has a rounded steep end, finely worked. The other (Fig. 90 0), faintly blue, has a diagonal end also finely 446 PROCEEDINGS OF THE PREHISTORIC SOCIETY.

chipped. In the white example the point is finely worked. In the blue one, the point has gone, but there is a small encoche finely worked near this end. The blue dos rabbatu is i\ in. long, the white one a little longer. Both are 3 4-8 in. broad. Two other examples are 1^ by -f^ in. and i^ by \ in. In both the work is only for 1 in. from the pointed end. The former is a polygonal blue-white flake of the cone period. The latter is honey-coloured translucent flint, unpatinated, but lustrous, of the same variety of flint as some of the cones. Two others seem to be the broken worked ends of dos rabattu similar to the last, both polygonal flakes. Another variety is a broken long polygonal flake with a diagonal finely worked end, 7-12 in. in length, terminating in a sharp point: This is \ in. broad. Two others are ij in. and 2 in. respectively. The latter has a diagonal worked end. A rarer form is a polygonal flake broken with a single ridge at the pointed end, from which the edge is smashed to form a well- marked variety of No. 1 (Sturge), ij in. long. The last example is lustrous, honey-coloured flint with a strongkeeled end turned to one side, and smashed for a little dis- tance along the side, 1 7-12 in. long. Fig. 90 I, J, K, L. M, and N are examples. DRIFT FORMS. These are not numerous. The best is a two-faced implement, 2f in. by 2J in., of the ovate type, thickest in the centre. It has a blunt point and a cutting edge, running all round except at one side, and sharpest at the ends. It is impossible to fit this implement into a Neolithic culture. It is of the same kind of flint as the cones, and has the same style of work, and is heavily iron-stained. Disc FORMS. The degenerate handaxe, a roughly circular implement with a blunt point, is often difficult to separate from the rough disc, thickest in the centre. Some six good examples come from this site, and of one of them Mr. Reginald Smith says, " similar to the discs of the St. Acheul period." They are all about the same size, 2 in. square, and in thickness are 1 in. or a little more. All belong to the prismatic core period, and are unpatinated but lustrous and iron-stained. One example has a patch of blue-white patina. (a) A typical piece, thickest in the centre, with a cutting edge three parts round the circumference, chipped on both faces. (b) A smaller example in which the work is done from the edge on each face at right angles to each other. This piece is square and not circular, and recalls a form which is common at Grime's Graves, an example of which was figured by Mr. Reginald Smith (Report, Fig. 28). A PREHISTORIC SITE AT KIMBLE. 447 I have a somewhat larger example to (a) from the surface at Peppard. TEA COSY SEGMENTAL TOOLS AND PERCUTEURS. The tea-cosy is an appropriate name for a type of implement first described by Dr. Alan Sturge. In form it resembles a tea- cosy, and has a flat base, a segmental cutting edge, and the work is on one or both sides. We have thus two recognised varieties— the one-faced and the two-faced cosies. Both these varieties are found at Kimble. They vary, however, in form. In the two- faced varieties the two faces are not always equally convex. In some, too, the work is mainly round the base, and the segmental edge is roughly shaped and of doubtful value for use. The base varies too. In the most typical varieties it is formed of crust, wholly or in part, and when this is the case it is obvious we have a complete implement, and not one that has been broken. Another variety has a hinge fractured base which Mr. Reginald Smith thinks could be intentionally produced. In a few examples from Grime's Graves, the edge of the base is battered to produce a false hinge. The Northfleet implement in the British Museum is a good example of hinged base. A third variety has a definitely chipped base, and when this is done with a number of small blows, it is equally clear the implement is not broken. In a fourth variety the base is a plain or natural fracture, and when the work is not done from this surface doubt may be felt whether the implement is com- plete. In a large series such as I have from the surface, it is possible to pick out the typical pieces, but it is difficult to say where the tea-cosy ends, and the chopper and slug plane begin. The series pass also into an allied form, the blunt pointed percuteur or bone-breaker, and on the other hand into the conical planes with a ridged top, and again into the rough disc and hand wedge. (a) A tea-cosy, with good cutting edge, not quite symmetrical. Both faces worked all over, but one flatter than the other. The work is done from the cutting edge, and also from the base which in this case is a hinge fracture. As the hinge base has formed the plane from which blows have detached flakes from one face, it is clearly a complete tool (blue-white patina, 2J by 2), 1 in. thick. TORTOISE CORES. The peculiar method of making a one-faced hand-axe adopted by the Northfleet knapper in the Moustier period is now fairly well known, from the descriptions of Mr. F. Spurrell and Mr. Reginald Smith, and the discovery of tortoise cores and Levallois flakes at Grime's Graves has strongly influenced the date suggested by Mr. Reginald Smith. An enormous mass of evidence has been adduced, to which I was able to add in my last Presidential address, which shows that the flint work at the Graves combines the leading characteristics of both the Moustier and Aurignac periods, pointing to the period represented by the cave of Abri Audi as the true date of this remarkably prolific site. 448 PROCEEDINGS OF THE PREHISTORIC SOCIETY.

It is, therefore, a matter of considerable interest to find in the series I am now describing a number of forms which can only be classed as tortoise cores and flake implements, one core obtained from a neighbouring field having the characteristics of an unstruck one. I shall return to this matter later on, and shall now describe some of these pieces. In the last number of the " Proceedings," I figured three examples of tortoise cores, and showed that there are two distinct varieties to be found at the Graves :— (i.) a form in which the upper face is domed, (2.) in which the upper face is flat. Both of these varieties occur on the Kimble site. (a) (Fig. 92 E) One of the best examples is 2J in. in length and if in. in breadth. It is, as the section shows, boat-shaped. The work was carried out first on the keeled face as the truncation of the facets demonstrates. The upper face was then prepared by working from the edge, the master stroke struck from one end taking with it most of the upper dressed face. The flake implement struck from this surface is if in. long and ij in. broad. The whole piece is lustrous but unpatinated. It is an example of No. 1 of the two varieties mentioned above. On comparing it with Fig. 38, Grime's Graves Report,an example from Northfleet, the resemblance is obvious. In the diagram the central flake impression appears to have been struck from a larger surface. This is wrong. In the core itself the negative impression of the bulb is seen at the lower margin of the piece, the blow being struck from the present edge. Below is a transverse section, showfng the boat-shape of the core. (b) (Fig. 92 A) TWO figures of an unstruck tortoise core (EX. I), 2§ by 2Jin., 2 in. thick, patinated white. Both faces are keeled and an edge runs all round. Compare Fig. 37, Grime's Graves Report, from Grime's Graves. A gpod example of a small unstruck variety from this site (No. 2) is very similar to a number from Peppard and Grime's Graves, and some which I have from the Desert (Egypt). Examples of small flake implements with truncated work on the upper face and facetted platforms have come from the Kimble site, but are rare, as, too, are the tortoise cores. They show that this method was used on this site, but to nothing like the same extent as at Grime's Graves and Northfleet.

SCRAPERS. These vary considerably both in the presence or absence of patina, in shape and size, and the character of the work, and it seems impossible to suggest they are all of the same date. The fact remains that they all come from a portion of the field where the rest of the pieces described m the paper occur, and there is practi- cally no chipped flint work to be found in the rest of the field, or in those on the other side of the South End road. The cores or core planes are not all of the same quality of flint. Not more than twenty-four of the scrapers out of a total of A PREHISTORIC SITE AT tciMBLE. 449 about 353, which, however, includes flakes used at some part for scraping, can be assigned to the cone culture. They may be classified conveniently into end scrapers, short thin blades, with or without side work, the end being squared, rounded, hollowed, diagonal, nosed, etc. ; the chipped face varying from a slight bevel to steep cliff-like work, sometimes overhanging. Some thirty-five are broad and short, and the same number tend to length. Some eight of these may be classed as end scrapers on blades, the rest being of medium length or external flakes. Crescent scrapers are very numerous, some fifty-five having a single crescent at the side. Five have this at the end, and two have two encoches. Side scrapers seldom occur, ten in all, two of which are massive enough to belong to the Moustier period. Another variety is shaped like the human ear, and of these there are four. Two others have a convex diagonal scraping edge, and one has both this and a used straight side meeting the end, coming thus near to the Abri Audi point which occurs at Grime's Graves. How far these scrapers help to date the site it is difficult to say. Many of the steep-faced ones could be assigned to the Aurignac culture, others might well belong to the Bronze Age, whilst the majority would be classed by most prehistorians as Neolithic. No two of them are exactly alike, and in this respect they differ from a series such as Windmill Hill, Wilts, where an unusual degree of skill was attained, and greater uniformity in size and form. Five of these are figured :—• FIG. 91 B. End scraper on blade (unpatinated). The end is finely retouched; the upper face of the bulb end also finely worked, this being carried along the sides on one of which is a well-worked encoche, i\ in. long. FIG. 91 c. Part of *&. blade scraper patinated white with and parallel sides. The end is steeply worked, somewhat undercut battered, and carried out in fan-shaped work, if in. long. Both of these are typical Cave forms. FIG. 91 K. A long round-ended blade scraper with the end undercut and engrailed. The platform is facetted in the Northfleet manner, showing that the flake was struck from a dressed core (figured below). FIG. 91 A. A long blade scraper, with a convex end and angular corners. The sides are trimmed on the under face. Patinated white and much iron-stain. FIG. 91 D. A short white patinated scraper with convex worked end and steep work carried up one side. The platform and adjoining part of the end are facetted and trimmed. The slug plane which links the racloirs with the grattoirs is seen on this site (five examples). This cave form occurs at Peppard, Grime's Graves, Cissbury, and Canterbury (gravel). All are about 2J in. long. On a site where tortoise cores and flake implements are found it is probable that the plunging flake will also occur. In the form 450 , PROCEEDINGS OF THE PREHISTORIC SOCIETY.

N Fig. 91.—Implements from Kimble, S. Bucks. -J A PREHISTORIC SITE AT KIMBLE. 451 of an angle flake struck off the edge of a core to refresh it for future work it occurs on many sites of different dates. I have many examples from Grime's Graves, Peppard, Nettlebed floor, and Kimble, and from the surface at Peppard and Hambleden. But the true " plunger " is very rare from the surface, and I have only seen it from the surface from Grime's Graves. I have two examples from Kimble, one of which has a plain platform. This latter fact is of interest. Many of my Grime's Graves and about half of my Peppard examples have a plain platform. As I shall show elsewhere the trimming of the platform occurred side by side with a plain one in the same culture, notably at Grime's Graves and Peppard. Whether this holds good at North- fleet and St. Brelade, where facetting is abundant, I do not know. In some of the Kimble facetted flakes it is clear, as at Grime's Graves, the work was done after the flake was struck off, as the bulb has been nearly chipped away, but in most cases this is not so. Quite distinct from these is a series of pieces showing facetting of the edge, which are clearly a portion of the core struck off to give a new surface to strike on. In the last " Proceedings " of this Society Messrs. W. G. Clarke and H. H. Halls contributed a paper on the cone culture of Helles- don in the Wensum Valley, in which they worked out a valuable classification of cones, based on the angle and extent of the chipped face, and the number of planes from which the flakes were struck off. This classification is materially the same as that which I have worked out from the cones from the Kimble site during the last three years (since 1913), and as, with slight modifications, I do not think it can be improved upon, I shall adopt the various classes in describing this form of implement. There are two classes in this classification which are difficult to fix very definitely. In Class 4, in which the angle of the face is not more than 450, we do not have a very well defined implement, as, though pieces corresponding to this description can be picked out, there are in a large series plenty of others which approach this angle, and shade off into the vertical and overhanging chipped face. In Class 2 we have also a feature in the undercut nose which can hardly be intentional. A few in a large series correspond with the description given, but the " nose " occurs also in various positions on other pieces which do not correspond, often in long cylindrical forms and on cones which are chipped all round the circumference. The prismatic cones, characterised by long narrow flaking generally patinated blue or white and lustrous, are very rare from the surface at Hambleden, S. Bucks, except from this one site, and equally rare in the neighbourhood of Peppard. They are, how- ever, very common in the neighbourhood of Icklingham, and Dr. Allen Sturge has a very large number. Some of the finest examples I have seen were found by Mrs. Sturge at Cavenham, Suffolk. Mr. Kendall, too, has a number from the zone at the top of Hackpen 452 PROCEEDINGS OF THE PREHISTORIC SOCIETY.

Hill, Wilts, and the late Lieut.-Col. Underhill discovered a similar zone near Ipswich. The material for dating this branch of chipped work is, then, fairly plentiful, but it is doubtful if any solution will be gained until one of these zones is excavated. One naturally asks are these cores only from which flakes have been struck, or are they implements definitely chipped for a given purpose ? They may, of course, be adapted when no longer suitable for flaking, to the latter use. If cores only they infer an enormous production of flakes both small and narrow, and it seems probable that a great many are the residue of such work. The majority of the Peppard and Grime's cones (though as both are factories they may be unused planes), show no secondary work, and were probably cores only. The majority of thost irom Kimble, however, have very delicate work at the edge, and often this is seen on a projecting spur which juts out from the edge of the base. Many, too, are so small, that not a single one of the flakes seen to be taken from the face would be of service. The pigmy implements are worked out of flakes which could have been taken from many of the larger cores, but not from the smallest. It is certain that in the dressing of skins implements of this kind with a fairly blunt edge would be of great service, and little liable to cut the hide. As the drawings show, several of the ends of the prismatic cylinders are nan owed or brought to a point, and constitute rough, but serviceable implements. It should be remembered, however, that if all the steep faced pieces from a site are collected, there are a large number which cannot be included in the cone series, owing to their not being symmetrical, and I have fully ioo from the Kimble site which can- not be classed in that way, but are cores pure and simple. The cones from the Kimble site number about 130, and Mr. and Mrs. Payton have probably a similar number. I have not been able to compare my series with theirs, and shall confine my remarks to those in my own collection. Messrs. Clarke and Halls found a nearly similar number at Hellesdon, and the numbers in the various classes from the two sites are, therefore, more interesting. The corresponding figures are as follows :— Kimble. Hellesdon Class 1 46 39 Class 2 10 14 Class 3 47 «i3 43 a. 3 b 29 61? C5 C39 Class 4 4 5 Class 5 9 23 Class 6 4 - A PREHISTORIC SITE AT KIMBLE. 453

In each case Class i single planes with the secondary work three parts round are less numerous than Class 3 double planes, in which both ends of the plane are used. Class 2, which has an undercut nose, yields nearly the same number from each site, but as I have pointed out it is difficult to decide which pieces belong to this group—as many others which have not the other characteristics of the definition have one part undercut—and the " nose " varies in its position. The main distinction is in Class 5, where the work is carried all round the circumference of the base. Only nine of these came from Kimble, whereas twenty-three are recorded from the Hellesdon site. Mr. and Mrs. Payton have, however, a number of this class which may add to the percentage. There is a class of plane which does not seem to come under any of the classes enumerated, and is yet a true cave form, as a similar implement from Les Eyzies is figured by Mr. Reginald Smith in " Archaeologia," Vol. LXIIL, and another in Lartet and Christy, Reliq. Aquitan. Messrs. Clarke and Halls, p. 198, say that, " These types of cone . . . (including in this the five classes enumerated) are, perhaps, more abundant on Cissbury type stations than elsewhere." With this I cannot agree. Those included in Class 3, especially all embraced in the cylinder form—are so rare both at Grime's Graves and Peppard, that out of some thousands from these sites I have scarcely a single example. The occurrence of the long fluted prism on these cone sites is a marked distinction between the two groups of sites, and the planes from the two sites generally show the same difference. Class 1. 46 examples.—They vary in size, the average being ij in. in height. Some have the fine work confined to one part, generally a " nose." Most of them fit the hand as if designed for use. Class 2. 4 examples.—Fig. 91 E. This is a perfect example with ridged top and two " noses " due to truncation of flakes struck from the base. The back is (white) battered and partly made by a single blow. Class 3. 52 examples.—Messrs. Clarke and Halls have divided the double-ended cones with two worked edges into three groups. The Kimble cones cannot be classed in this way, as the definitions are too rigid. (a) 1. I have no examples from Kimble with a flat top and bottom, and the work at right angles to them. 2. In this class the plane surfaces are at an angle with the chipped face, and the two chipped surfaces meet, being in the same line. Fig. 91 H. 13 examples.—The back is crushed. One end is much narrower than the other, both being fluted in the same line. 3. In this variety the plane faces are at an angle with the fluted face—the work being in one line-—but in addition, the back is worked from one side at right angles to the main fluted face. It is a very rare form. My example is 3 in. long. 454 PROCEEDINGS OF THE PREHISTORIC SOCIETY.

(b) Six varieties of this form occur at Kimble : 1. In this the planes are at right angles, and the chipped faces either meet at this angle or form a curved surface with the work always in the same line. 8 examples. 2. Two plane surfaces and these and the work at light angles to each other, n examples. 3. Very similar to the last, but the two chipped faces are on the opposite sides of the piece. The distinction is clear when the two are placed side by side. 3 good examples. 4. In these there are three flaked surfaces and three planes. 2 examples, ij and if in. long. 5. Cylinders with a plane at each end at an acute angle with the worked faces. The fluting is in the same direction, viz. : longi- tudinal, but the work at each end, owing to the direction of the planes, is at right angles. Fig. 91 j. The drawing is taken to show the fluted faces, and does not include the plane faces, which are at right angles. 6. In addition to the features of the last this variety has one side of the back chipped at right angles to the main face. 2 examples. (c) 1. Double ended planes with a slight twist, so that they are not at right angles. 4 examples. 2. Two planes and two worked surfaces in the long diameter of the piece. The planes, however, are on alternate faces, so that the fluting of one face terminates in the plane of the other. 1 example. Class 4. 4 examples.—Two single and two double planes, each with a worked face about 45°. Fig. 91 G (2 in. long). A double ended plane with a broad and a narrow fluted end, both with fine secondary work—the latter with a fine spur not shown in the illustration. The planes meet at the back at an obtuse angle. Class 5. 9 specimens.—Tfcese vary much in size, the largest being 2\ in. high, the smallest 1 in. Fig. 911. The largest specimen, base a single facet, top ridged and worked on both faces. Back slightly undercut. Front ridged, truncated or nosed at three different levels. An advocate for ani- mistic flints could very easily see in this an attempt to depict a human face. Fig. 91 F. The smallest example, very rare in this culture, but fairly common in some parts of Norfolk and Suffolk. It has a " nose " at each end, and the ridged top is finely worked, the base being a single facet. Tt is doubtful if any of the flakes from this cone were used even for pigmies. Class 6. 4 examples.—This is a true cave form as an almost identical piece is figured from Les Eyzies Cave, " Archaeologia," Vol. LXIIL, by Mr. Reginald Smith. The base is flat, top conical, with fluting carried up each end. In form it is triangular, and compressed from side to side. Ends " spurred." The best example is 2 in. high, and the base is 2§ in. long. A PREHISTORIC SITE AT KIMBLE. 455 There is a variety of Class 3 (b) which is a new type. I have 5 examples from Kimble, and three from the Peppard mine; and Miss N. Layard, one from the surface near Grime's Graves. It has two plane surfaces at right angles to each other, with fluting carried up the corresponding faces. A keel runs along the top, starting from the end of the steep plane face. This ridge splits below to form the base of the implement. They are all about the same size, viz., 3 in. long and 2| in. high. The flalces from this site, as might be expected from the corse and core planes, are mostly long and narrow, and similar to those worked up into the dos rabattu knives, and ate very similar to cave series. I have about forty which average i in. in length. These are nearly all ridged, but a few are doubly ridged. I have another series of about fifty, which average ij-if in. in length. A few are polygonal. Nearly all of the narrow long flakes are internal ones. The longest is 4 in. I have a series of 30 flakes, which can be classed with the dos rabattu knives. None of these are the true cave pigmy described above. Sixteen of these are battered from the edge, and in each of these cases it is not the bulb edge. They may be classed as dos rabattu No. 1 (Sturge). Some may be struck from the edge of a core plane. If the latter is the case, it is difficult to see what advantage was gained as the surface above would project, and would not yield a useful series of flakes when worked. Of the remainder several can only be definite implements. One (3^ in. long) is a long ridged flake with facetting confined to the platform, and work carried for about half the length of the implement from the central ridge. The end ter- minates in a sharp curved keel, vertically fluted on one side. It thus approaches the graver type. Two other keel-ended flakes came from this site. Not the least interesting of the forms from this site is a series of ten pieces which can only be classed as rostro-carinate imple- ments. With the exception of four they all have a stout keeled edge, and five of th;m have the characteristic twist to one side due to hollowing out of one lateral face near the end which is seen on the implements of this class from the detritus bed b^ieath the Red crag of Suffolk. Five of the examples might be included in the core plane series, in that one side of the ridge or keel is produced by removing 1-3 flakes, and the other is worked longitudinally from the keel edge, but in each case one side of the plane is intentionally flattened by work, and this forms the ventral surface of the rostro-carinate implement. None of the other 120 planes mentioned above has this characteristic. 456 PROCEEDINGS OF THE PREHISTORIC SOCIETY. As these forms have not hitherto been described from a sur- face site which is either Neolithic or late Palaeolithic (cave) in date, I am describing the series in detail and giving illustrations of the principal types. 1. Massive lumps which might be passed over as cores but have a very definite " beak " at one end. Fig. 92 c. This is, periiaps, one of the best formed of the series, 3| by 2 in., 2 in. high. It has a flattened base chipped mainly from one side, a steep back, and a well-defined ridge which runs up the middle of the back across the dorsum, and forms in front a constricted nose (left in diagram), which slightly overhangs the base. No crust is retained, and the whole is human work. The right lateral plane shows work from the ridge, the base and the posterior end. The " beak" curves to the right. The drawing shows the left lateral face of the implement, whitened, ironstained. Fig. 92 1. This implement is made from a cruciform shaped flint. The diagram shows the right lateral view. The keel is well made, by work from its margin. The ventral plane is formed by a single blow. The keel extends over the dorsum (4! in. long). Fig. 92 D. This is a core shaped lump with the ventral plane formed by a single blow. Without a section the diagram suggests a tea-cosy form. The implement is, however, as broad as it is long, and the keel is well formed by hollowing out the sides. The drawing gives the right lateral view. 2. Definite bird's head-like forms, with a definite keel hollowed on one side so that the beak is curved. Fig. 92 B. 3 J in. long.—The drawing shows the left lateral side of the implement. It is made from a brown stained Eolithic flint, and the work does not appear very old as the splinters have not weathered out. The ventral and dorsal planes are well chipped. The keel occupies only the front of the implement, and owing to extensive hollowing out of the left lateral face, curves to the left. Its resemblance to a bird's head and beak is remarkable. Fig. 91 L. A very small implement, i|in. long, unpatinated. The ventral plane is formed by a single blow from the side. The centre is conical, and the keel twists to the left, owing to hollowing of the left lateral face. Fig. 91 M. This example shows considerable crushing or batter- ing of the aretes. The ventral plane is flat and chipped all over. The keel curves to the right, owing to hollowing of the right lateral face. The drawing shows the right lateral aspect with the sharp keel in profile (unpatinated). 3. In the other pieces (12), the keel is central, and there is no hollowing of the lateral surfaces. Fig. 91 N. if in. long, patinated, white. It is triangular in form, the keel forming the apex. The ventral plane is flat, the bulb showing at the broad end. The dorsal plane is part of a hinge fracture, and the right lateral A PREHISTORIC SITE AT KIMBLE. 457

Fig. 92.—Implements from Kimble, S. Bucks, 458 PROCEEDINGS OF THE PREHISTORIC SOCIETY. is partly formed by blows struck from the keel. This is clearly a small rostro-carinate implement, and undoubtedly belongs to the prismatic core period. The two other examples illustrated also belong to the prismatic core period. Fig. 92 H. A crusted piece, with a flat ventral face made by a single blow. The left lateral is fluted in the long diameter of the implement like the series of planes. The right lateral, however, is not flat, and made by a series of blows struck from the base. 2J in. long, bluish patina lustrous. Fig. 92 G. 2,\ in. long. The ventral face is made by a single blow. The dorsum has work over it except at the back where crust is left, and the right and left lateral planes are well formed. The drawing shows the left lateral face. Fig. 92 F. A delicately worked knife or spear head, the bulbar face being unworked. The whole of the upper face has had small flakes pressed off it from the edge, and there are two small encoches near the broader end, unpatinated, 2 in. long. The question naturally arises, as it must always do from a surface zone, whether all the implements from this site belong to the same date. The fact remains that they were all found in a portion of the ploughed field and scarcely a flake or a flint with human work has rewarded a prolonged search of the neighbouring fields. The series described includes a large number of planes, pigmy implements, scrapers in great variety, tortoise cores and flake implements with. facetted platforms, various celts of Cissbury type, picks, etc., and a series of implements of rostro-carinate form. It is difficult to fit all these forms into one culture if we follow the existing classification. No rostro-carinate form has been recorded from any deposit later than the Crag or mid-glacial sands, though Mr. J. Reid Moir has an example exhibited in the Ipswich Museum of the prismatic core period, which he refers to the Neolithic period. The celts point to a Cissbury date, but the other pieces do not give support, especially the cones and pigmies. The tortoise cores and flake implements point to a Moustier influence whilst the blade scrapers and pigmies suggest a Madelaine date. It is difficult to fit all these forms into a Neolithic site, as we have little of that date definitely fixed to compare with. Neither the kitchen middens nor the earlier lake dwellings nor the flint mines (which are referred to the early Neo- lithic age) show many points in common. Patination helps us little, since the Grime's Graves excavations showed that black, blue and white pieces are all of the same date. At present we can only say that the Kimble site belongs to a period when the prismatic cone was the principal implement of a definite culture, and hope that excavation on one of the fairly numerous sites of this culture will throw light on its correct horizon in the Palaeolithic or Neolithic period.