Kirchhellen and Weeze , Lower Rhine Bay / Germany , with an Elder Acheulean : starting points for research into the problems of recognizing stone-artefacts in pebble-accumulations of fluviatil terraces

The continuing critic hy part of the prehistorians, including the principal rejection of pieces found in and on fluvial terraces without additional archaeological docu­ ments, which does include Kirchhellen and Weeze, caused the author to continue research and he began to study si­ tuations, where geofacts are produced, which could be mistaken for archaic artefacts. Results up to now seem to indicate a better suited approach to this problem, as ex­ plained in the following part. To postulate criteria, in this case enabling to distinguish between arteand geofacts.an extensive fundamental study of the origin of geofacts and the context should have been undertaken, to form a basis from which to draw conclusions and then to establish cri­ teria. This is missing in our case. The author's report de­ scribes observations, which seem to point in a direction en­ abling a better understanding of the problem. The state­ ments, that the existence of artefacts in fluvial terraces is higly improbable and that it is not possible, if they exist, to separate them from geofacts in pebble accumulations, is contrasted by numerous archaeological sites with a wealth of artefacts, the latter even excavated in thousands from fluvial terraces in the mediterranean area. The production of geofacts as a natural process is much to complex to po­ stulate simple criteria, as they are used now. The pattern of natural damages differs from rock to rock, frequently very strong. Experience gained with flint/silex, in Gennany or elsewhere, cannot be generalized and used on other rocks: on one site archaic looking geofacts, made perhaps on li­ mestone, may exist in hundreds together with isolated an­ thropogene similar artefacts on quartzite. The research in­ to a possible archaeological site in an elder tluvial terrace demands to study the flow of material for a longer distance of the river, as well as the tributaries and to take into ac­ count numerous complex aspects, which influence the production of geofacts in the case involved. As many collections as possible of artefacts from the Elder Paleolit­ hic have to be studied, besides extensive field work on ter­ races in different areas plus experimental knapping; these are basic conditions to gain the necessary knowledge, and this expressly over long periods, preferably many years. The final aim is the possession of a wealth of detailed, per­ sonal knowledge of boths: the artefacts involved as well as the geofact-forms in general: then as a next step the speciSCHMUDK, Habichtstr. 71, *) Anschrift des Verfassers: K. 45134 Essen al geofact-forms of the area involved in new research. The pieces found in Kirchhellen and Weeze are reconsidered in view of the above mentioned observations and groups are created: a. one group of artefacts; b. one group which com­ prises pieces from a zone of passage. There are still que­ stions open: for instance is the geology of Kirchhellen ob­ ject of new studies, but also questions with respect to the general problem artefacts/geofacts might be further clarified and this may help with the classification of certain pieces. [Die Fundstellen Kirchhellen und Weeze mit einem älteren Acheulean: Ausgangspunkte für weitere Untersuchungen des Problems, Stein-Artefakte in Geröllansammlungen von Flussterrassen zu erkennen.] Kurzfassung: Die anhaltende Kritik eines Teiles der Prähi­ storiker einschliesslich der prinzipiellen Ablehnung von Fundstellen in und auf Flussterrassen ohne zusätzliche ar­ chäologische Dokumente, was ja auch Kirchhellen und Weeze einschliesst, veranlasste den Author zu weiteren Untersuchungen; er begann daher, sich mit Stellen zu be­ schäftigen, an denen Geofakte produziert werden, die für archaische Artefakte gehalten werden können. Die bisher vorliegenden Resultate scheinen eine besser geeignete Annähening an dieses Problem zu zeigen und werden im folgenden Teil erklärt. Zur Erstellung von Kriterien, in die­ sem Fall solcher zur Ermöglichung der Trennung zwischen Artefakten und Geofakten, hätte eine umfangreiche grundsätzliche Studie des generellen Ursprunges von Geo­ fakten und ihres Kontextes vorgelegt werden sollen, als Basis für allfällige Schlüsse und der Erstellung von Kriteri­ en. Diese fehlt jedoch. Der hier vorgelegte Bericht be­ schreibt einige Beobachtungen, die in eine Richtung deu­ ten, die ein besseres Verständnis für das Problem gestattet. Die Behauptung, verrollte Fundstücke seien keine Artefak­ te mehr, ebenso sei es nicht möglich, in Geröllansammlun­ gen Arte von Geofakten zu trennen, kontrastiert mit zahl­ reichen Fundstellen mit verrollten Artefakten, sogar zu Lau­ senden aus Flussterrassen ergraben, im Mittelmeer-Raum. Die Entstehung von Geofakten aus natürlichen Prozessen ist viel zu komplex, um einfache Kriterien, wie sie im Ge­ brauch sind, dafür aufzustellen. Die Erforschung einer möglichen archäologischen Fundstelle in einer älteren Flussterrasse erfordert die Untersuchung des MaterialThe Sites of Kirchhellen and Weeze, Lower Rhine Bay/Germany, with an Elder Acheulean: starting points for 121 research into the problems of recognizing stone-artefacts in pebble-accumulations of fluviatil terraces trusses über eine längere Distanz des Flusses sowie seiner Zuläufe und dabei die Beachtung zahlreicher komplexer Aspekte, die den jeweiligen Fall beeinflussen. Dazu sollte die Kenntnis sovieler Sammlungen wie möglich aus dem älteren Paläolithikum erworben werden, daneben sind ausgedehnte Feldarbeit auf anderen Terrassen und experi­ mentelles Schlagen eine grundsätzliche Bedingung, um das notwendige Wissen zu erlangen und dies ausdrücklich über eine längere Periode, vorzugsweise eine Reihe von Jahren. Das Endziel ist der Besitz umfangreichen detaillier­ ten Wissens sowohl über die involvierten Artefakte als auch über die generellen Geofakt-Formen und dann als nächster Schritt die der Region, in der sich die abzuklären­ de Fundstelle befindet. Die Fundstellen von Kirchhellen und Weeze werden mit Blick auf die hier beschriebenen Beobachten überprüft; sie enthalten eine Serie von Arte­ fakten neben einer Reihe von Stücken, die in eine Über­ gangszone gehören. Es sind noch Fragen offen: so ist, als Beispiel, die Geologie von Kirchhellen Objekt einer neuen Untersuchung, aber ebenso könnten Fragen des generel­ len Problems Artefakte/Geofakte weiter geklärt werden und dies könnte bei der Ansprache bestimmter Stücke helfen. 1 I n t r o d u c t i o n In 1992 the description o f the above ment ioned sites and their artefacts was published (SCHMUDE 1992) ; the recognit ion and accep tance of these w a s based on the personal exper ience o f the author with arte­ facts in France, Germany and Spain, and in addition the confirmation of a n u m b e r o f prehistorians expe­ rienced with this complex. In the following years the artefacts o f Kirchhellen and Weeze met with critic and non-acceptance as well as acceptance. A certain picture developed during the years following the publication o f 1992. T h e persons accepting theses pieces as artefacts based their opinion on long ex­ per ience and intimate knowledge of artefacts, found on surfaces o f tenaces as well as from gravelpits and from inclined, eroded terrace-slopes in agricultural zones. Most o f these persons, further named group "A", had, in addition, a lso gathered exper ience with geofacts as well as with experimental production o f stone-tools. This exper i ence was collected in coun­ tries with longstanding research into the numerous fluvial terraces and their palaeolithic sites with inventaries made mainly on quartzite, quartz and others, but only to a low degree on silex. S o the re­ cognition o f artefacts is, by this group, based on a wealth o f empirical knowledge , including excava­ tions with 2ooo to 6 o o o artefacts on quartzite/ quartz, showing a wide spectrum of types o f tools as well as a wide range o f conservation covering all as­ pects: eolised, leached, patinated, rolled and mixtu­ res o f these. Experimental artefact-production is an additional tool. (TAVOSO 1978; QUEROL & SANTONJA 1979). T h e oppositional group,"B", bases its critical position on several statements listed in the following. a) P ieces , which are rolled, are only acceptable as ar­ tefacts, if additional archaeological documentat ion as rests o f fire, worked b o n e s etc. exist. (HAHN 1991: 52) ; b ) only, if pieces, made on rocks not originating from this location and which are not in a secondary position, exist, are such p ieces acceptable as artefac­ ts from secondary sites (HAHN 1991: 51) ; c. it is not possible to recognize elder, sparsely work­ ed artefacts in great accumulations o f pebbles , as there are to many pebbles damaged by nature and alike to artefacts, especially if collected in lengthy periods o f time. (ROEBROEKS 1993: 12+13); d. if artefact-character is claimed, a number o f simple criteria should be fullfilled: unifacial regular and low angled flaking indicates an artefact, bifacial flaking therefore a geofact; an anthropogene flake should have negatives on its dorsal face etc. and others on details (HAHN 1991: 5 3 & 5 4 ) . These are also the criteria used as arguments against the finds from Kirchhellen and Weeze . So the author found himself faced with two opinions. At present there s e e m s to be a stalemate situation, where one can only join one of these schools: an unsatisfactory situation and the author searched for possibilities to open ways for new movement. The approach is as follows: to b e able to distinguish be tween arteand geofacts, it is a basic condition to know both in de­ tail. T h e r e are numerous sites and artefacts from an Elder Paleolithic, which can be studied. This is, ho­ wever , different, when it comes to geofacts. Isolated observat ions and a few publications attempting to establish criteria (PATTERSON 1983,also in HAHN 1991) have b e e n published. As the separation o f artefrom geofacts is one o f the very fundamental questions in prehistory, the es


Introduction
In 1992 the description of the above mentioned sites and their artefacts was published (SCHMUDE 1992); the recognition and acceptance of these was based on the personal experience of the author with arte facts in France, Germany and Spain, and in addition the confirmation of a number of prehistorians expe rienced with this complex.In the following years the artefacts of Kirchhellen and Weeze met with critic and non-acceptance as well as acceptance.A certain picture developed during the years following the publication of 1992.The persons accepting theses pieces as artefacts based their opinion on long ex perience and intimate knowledge of artefacts, found on surfaces of tenaces as well as from gravelpits and from inclined, eroded terrace-slopes in agricultural zones.Most of these persons, further named group "A", had, in addition, also gathered experience with geofacts as well as with experimental production of stone-tools.This experience was collected in coun tries with longstanding research into the numerous fluvial terraces and their palaeolithic sites with inventaries made mainly on quartzite, quartz and others, but only to a low degree on silex.So the re cognition of artefacts is, by this group, based on a wealth of empirical knowledge, including excava tions with 2ooo to 6ooo artefacts on quartzite/ quartz, showing a wide spectrum of types of tools as well as a wide range of conservation covering all as pects: eolised, leached, patinated, rolled and mixtu res of these.Experimental artefact-production is an additional tool.(TAVOSO 1978;QUEROL & SANTONJA 1979).The oppositional group,"B", bases its critical position on several statements listed in the following.a) Pieces, which are rolled, are only acceptable as ar tefacts, if additional archaeological documentation as rests of fire, worked bones etc. exist.(HAHN 1991: 52); b) only, if pieces, made on rocks not originating from this location and which are not in a secondary position, exist, are such pieces acceptable as artefac ts from secondary sites (HAHN 1991: 51); c. it is not possible to recognize elder, sparsely work ed artefacts in great accumulations of pebbles, as there are to many pebbles damaged by nature and alike to artefacts, especially if collected in lengthy periods of time.(ROEBROEKS 1993: 12+13); d. if artefact-character is claimed, a number of simple criteria should be fullfilled: unifacial regular and low angled flaking indicates an artefact, bifacial flaking therefore a geofact; an anthropogene flake should have negatives on its dorsal face etc. and others on details (HAHN 1991: 53&54).
These are also the criteria used as arguments against the finds from Kirchhellen and Weeze.So the author found himself faced with two opinions.At present there seems to be a stalemate situation, where one can only join one of these schools: an unsatisfactory situation and the author searched for possibilities to open ways for new movement.The approach is as follows: to be able to distinguish between arte-and geofacts, it is a basic condition to know both in de tail.There are numerous sites and artefacts from an Elder Paleolithic, which can be studied.This is, ho wever, different, when it comes to geofacts.Isolated observations and a few publications attempting to establish criteria (PATTERSON 1983,also in HAHN 1991) have been published.As the separation of arte-from geofacts is one of the very fundamental questions in prehistory, the establishment of criteria should be the result of an extensive, detailed study of the ori gin of geofacts and the context.Then, following an analysis, as a next step conclusions could have been drawn and, finally, criteria established.For the group "B"-arguments no such analysis seems to ha ve been published.The criteria postulated and used seem to be the result of isolated personal observati ons, generalised regardless of the context.Here an example for illustration: one of the statements pu blished by HAHN (1991:52) and widely used, de mands, that a flake found, especially in secondary si tuation, should have negatives or scars on its dorsal side, indicating human work, as natural forces can also produce flakes, but then these would be with out traces of further work, i.e. completely cortex-co vered.However, nature produces also flakes with scars on the dorsal side (personal observation of the author: Bay of Biscay-beach,limestone; Pyrenees ri ver: Gave d'Oloron, limestone).More important: the Acheulean in the Mediterranean Area contains a sig nificant percentage of cortex-flakes, separated from the numerous great cobbles; they comprise 2o-5o % of the total of flakes.Many times a small series just contains a few such pieces.This criterium did in one discussion lead to the doubtfull situation, that out of a series with hand-axes, choppers and cores the on ly flake found, a typical great piece for the produc tion of tools, completely cortex-covered, was decla red a geofact (TAVOSO 1978: 80, 258, 268, 275, 288, 291, 296, 3ol, 356).It must, however, be mentioned, that HAHN (1991) explains repeatedly, that the last decision should always be based on the geological situation.The author felt, that one way to progress could be a more detailed study of the creation of ge ofacts and the context, on the limited scale available to him.He began to observe locations,where geofa cts are produced.The results up to now are highly interesting and seem to indicate certain tendencies.These shall therefore be explained here and may perhaps show a different approach to the problem.

Working Method
The basic principal is a change of the presently used manner of approach, that is to imagine hypothetically, what could (and partly may!) happen in nature and use this speculations to explain in which man ner a disputed piece may have been created; fre quently publications with observations of isolated cases are added to this hypothetical approach and are generalized.The author returned instead to the empirical approach: see for himself, what rivers real ly produce, note the observations, collect typical pieces and only thereafter try to recognize trends and, if possible, draw conclusions.This also means the end of discussions with partners, which cannot denominate locations and contextes, on which they can base their argument or which do not dispose of the necessary general experience.In other sciences it is an accepted method to form a working-hypo thesis and then test, if facts fit into the theoretical frame; this testing is, in our case, missing.Following example may illustrate this: in the river Sieg, in a cer tain pebble-and-cobble-accumulation, flat slabs of quartzitic sandstone do, under the peculiar condi tions of this spot, form numerous long, slender pic like geofacts of rectangular diameter.Theoretically they should not exist: it is a physical and technical experience, that any material will first break at its weakest point and a very long and slender point should, in line with this, break off somewhere in the first third.However, nature here shows, that facts are different from a theoretical approach.As a practical method in the field to study accumulations of peb bles, the same manner of search as for artefacts was used, that is: slow, regular walking, adapting speed to circumstances and using geometrical patterns en suring observation covering the complete area chosen, focusing now on natural damages.For each site notices of the observations are taken, typical pieces collected.A worker, amateur or professional, studying the geofacts of a certain location and/or area will in most cases detect a pattern of damages typical for the same and the conditions mling there.He may encounter repetitively the same form of damage, but also a design composed of many, sometimes difficult to describe, facettes or multi tudes of such damages.lt is therefore necessary for him to store in his memory the picture of the design of the significantly damaged pieces, to be later able to recognize such a pattern.
3 Observations and Tendencies recognizable

Tendencies recognizable and conclusions
The study of the material of these locations did show certain typical, repetitive natural damages in connection with certain geomorphological condi tions.This enabled the recognition of tendencies and therewith conclusions.a) It is not possible to establish simple criteria to distinguish artefacts from geofacts as: unifacial, re gular, flat flaking indicates anthropogene origin, bi facial the contrary, a natural origin.Natural forces produce both and other geofacts, for instance, flakes with dorsal negatives, indistinguishable from arte-facts, in fact complete series of geofacts looking like "archaic tools".Furthermore, discussions with qua ternary geologists confirm, that the processes in na ture are much to complex to allow the establishment of simple rules.b) Rocks differ in their behaveour against forces more or less, in many cases extremely.Silex, the greatest deceiver, produces easily geofacts, which can be mistakenly collected as artefacts; limestone/dolomite also does, but to a lesser degree.Sandstone, if brittle, shows typical breakage, while tough quartzi te is much less liable to break at all.It is of utmost im portance to differentiate: at the same site one may find hundreds of artefact-like looking geofacts, perhaps from limestone, while tough quartzite pro duces nearly no geofacts at all and if so, separation follows crevices and generally natural zones of weak ness.And so there can be quartzite-artefacts at the same site, for instance choppers, which would, with out differentiation, be mistaken for geofacts.A per son experienced only with silex cannot utilize the knowledge gained to critizise disputed assemblies made on other rocks.To be able to do so, one has to pass an intensive period of learning to understand the respective rock and its peculiarities, which should by no means be underrated: it should in clude practical field-work, the study of collections of artefacts made on this rock, experimental knapping and then the same for geofacts.andall this in a prolonged period: we speak about years.To be able to recognize, in addition, altered (eolised, rolled, patinated, leached etc.) artefacts and to dis tinguish them from geofacts, which -contrary to the position frequently taken -can be learned, one has to have a certain talent, namely to see the original forms of the surface through the veil of alteration.One sees occasionally in collections of amateurs be tween numerous well-conserved artefacts a few altered pieces: the collector "has the eye".There are, on the other side, professionals and collectors with great collections of good standard, but which do not recognize altered pieces.A student, who has learn ed, additionally, that altered pieces are "unpersons" and cannot be artefacts, will have the greatest diffi culty in ever learning it, even if he has the talent.It is quite astonishing and a contrast, for a worker from northern or middle Europe to cooperate with prehistorians in the Mediterranean area, which have due to their lifelong contact with these type of artefacts and sites, the necessary exerience.From TAVOSO (1978: 255, 256) the following is cited as an example: "... les 29 outils et eclats qui composent cette serie ne representent ... qu'un echantillon -assez pauvre ... ä cause de l'intensite de l'usure fluviatile qui, effacant les aretes, polissant les facettes de taille et emoussant leurs contours les rends si semblables aux galets qui les entourent, que leur decouverte est beaucoup 123 plus une affaire de chance, de patience ou de "flair" que de 1'observation.La reconnaissance de ces quartzites tallies est rendue encore plus difficile par le fait qu'il n'y a aucune difference de coloration entre le cortex des galets et les facettes de taille ..." (... the 29 tools and flakes, which compose this series, represent ... a rather poor sample ... caused by the intensity of the rivers usage, which by erasing the ridges, polishing the scars of flaking and blunting/rounding their contours makes them simi lar to the pebbles by which they are sunounded, so that their discovery is much more an affair of chan ce, patience or "flair" then of observation.Recogni tion of these worked quartzites is made even more difficult by the fact that there is no difference of colouring between the pebbles and surfaces of the scars."(End of citation.)This is an excellent descrip tion of the problem.(See also DE L'UMLEY, 1971: 184, 196, 197 andMACRAE 1988: 92).Another aggravating circumstance is the dominant silex-tradition of North ern/Middle-Europe; the introduction of quartziteand quartz-artefacts in Germany has been a con tinuous, tenaceous discussion through decades, against rigid traditions, while in other countries the work with them was already well established rou tine for long periods.The discussion of A. Rust's eo liths has also left traces, which still today aggravate the problems.
c) The creation of geofacts depends on many factors, which intermingle and influence each mutually, such as the original form of rock, the length of the transport-distance, the geomorphology in general and specifically.(Fig. 1) According to the observations geofacts are in rivers very fast rolled to such a degree that the damage cannot be recognized anymore.This differs from rock to rock, but after 20-40 km transport most damage seems to be unrecognizable.This is partly different in moraines, where damage increases with the length of transport, but scars are thereafter still well conserved.Pebbles transported in a river would, after 500 km, be completely roun ded; this is not so in the nordic moraines.As mentioned before, all rocks react differently to forces and processes.Therefore geofacts of one rock may in several places show similarities, in others they may differ; pieces found may, of one rock, be geofacts, but -in the same place -of another rock be artefacts, while both look similar, e. g. like "choppers".In a steep valley, still in the mountains, where masses of very differently sized rocks, sometimes very big ones, are transported, geofact -production is high.In a wide valley, further down, where the pebbles/cobbles have already been sorted to a cer tain degree and are on average smaller and more evenly sized, geofact-production diminishes and dif fers from the above mentioned case.The same ap-The Sites of Kirchhellen and Weeze, Lower Rhine Bay/Germany, with an Elder Acheulean: starting points for research into the problems of recognizing stone-artefacts in pebble-accumulations of fluviatil terraces All this is part of very basic circumstances; they should be studied first.Thereafter one may ap proach the more complicated situations: the influen ce of periglacial conditions and the mixture of pro ducts of many conditions in a great stream, which transports over a long distance and accepts tributa ries entering with a diversity of material, loaden with different transport-effects.But to be able to under stand these, one has first to learn the simpler, the ba sic forms.Here a negative example: in a discussion a participant, asked how he knew, that certain pieces from a terrace were geofacts, replied that he had studied the pebbles of this terrace.This is, in the aut hors opin-ion, a doubtful approach, because not ha ving stud-ied the basic forms in simple cases, he is not in a position to recognize these under the much more complicated conditions on an elder terrace from a great stream, with a mixture of geofacts and possible artefacts.This can be compared with so meone trying to translate a Latin author as Caesar without having learned the grammar.And it brings us back to an additional problem: not being able to call on the necessary knowledge, succour is sought from hypothetical, speculative construction.

3-3
The Site of Sorde l'Abbaye As an illustration of the above mentioned points, the approach to a site in SW-France is described.
The Site.In the valley of the Gave d'Oloron, near the above mentioned village and close to the rivers end, a grav el-pit produces sands and small gravels from the re mains of a terrace conserved in the form of an island, with altitudes between 37 and 44 m a.s.l.The river bed is at 0 m a.s.l., under tide-influence, ca.35 km from the sea.The gravel contains an enormous amount of pebbles and cobbles, the great majority being of tough palaeozoic quarzite, up to 50 -60 cm length and nearly equal width and thickness.This terrace is dated as "Mindel", the age is therefore between 0,4 and 1,0 mio y., as defined in this area.
The Industry.Since September 1991 on 23 visits 64 artefacts were research into the problems of recognizing stone-artefacts in pebble-accumulations of fluviatil terraces found.They contain an Acheulean with pics, De faces and cleavers, together with choppers/chop ping tools, 1 scraper, flakes and numerous cores, all of quartzite.Remarkable are the extraordinary size and weight of many of the tools and cores, e.g.pics with up to 2,7 kg, cores up to 18,9 kg.The larger tools and cores are formed with very few, but wellplaced, bold, very strong blows.All pieces are more or less rolled, some to the limit, many are eolised, so me patinated.A few of these pieces could, on the first look, be geofacts; it is therefore necessary to stu dy this aspect.
The General Geology.In the valley and the area surrounding this site a number of terraces can be distinguished, but their remains are only seldom clearly recognizable; over wide stretches they have been destroyed, and we find only their remains in the form of pebble-fields, the highest on hill-tops of ca.140 and 170 m a.s.l.There are good-conserved terrace-rests at ca. 70-80 m a.s.l.near Sorde.The composition of the gravels is similar to the one on the site: in the majority the tough palaeozoic green and bluish quartzite, then quartzitic sandstone, sandstone, decomposed grani te and shale, ophyte etc.The surface of the quartzite-pebbles and cobbles is altered, on bigger pieces abt.2-5 mm deep, but smaller pieces may be com pletely altered and show inside an olive-brown tin ge of colour.
Industry on Terraces and Hills.Collectors, the author included, have found isolated pieces as well as concentrations on the 70/80 m-terrace and higher up on the slopes.Further south, con centrations are found on top of the hills, covered not by terrace-remains, but a sand-clay mixture.The in dustry on the terrace is a broad Middle Palaeolithic, with pieces from a Middle Acheulean down to Mousterian, plus Neolithic.On top of the hills one finds a rather young, evolved Acheulean, with flat bifaces, cleavers, rare polyedres, cores and many flakes.The private collections comprise from 500 to 800 quartzite-artefacts plus a few thousands of pieces of worked, untypical silex, which cannot be classified.From the terrace-surface at 70/80 m originate a few bigger cores with 5-6 kg and scars up to 18 cm length.Corresponding flakes have been found there.These artefacts are all more or less altered: coloured, encrusted, patinated.Everybody handling these ar tefacts has a very clear picture of what an quartziteartefact from this area looks like.

Flow of Material.
In the Pyrenees, between 1300 to 1800 m altitude, a Cretaceous conglomerate crops out, forming moun tain-ranges; it contains already well-rounded peb bles of quartzite and quartzitic sandstone, which erosion transports via creeks into the valleys.These creeks form small rivers, localy named "Gaves", which flow in a northerly direction for ca.50-60 km, where they join larger rivers, here the Adour, which enters the Atlantic at Bayonne.The author has, in a number of cases, studied peb ble-accumulations in such rivers, as well as a gravelpit producing sands from the lowest/youngest ter race.The picture he has gained shows that todays rivers transport mainly limestone, ophyte, sandstone, some silex and others: higher up in the river the peb bles/cobbles are frequently and to a considerable degree damaged, but further down-river the damage diminishes and in addition, natural damage beco mes less and less recognizable as the result of rolling and polishing.Newer damage in these lower reaches is rare and easy to undertand.The tenaces higher then 30 m contain, in contrast, a great majority of the tough quartzite/quartzitic sandstone, whose pebbles are well rounded and do not offer points/areas for attack.Damage is therefore rare; if at all, pieces se parate along zones of weakness and here, on the sharp edges, splintering causes occasionally a series of small scars.There is, especially if one is familiar with the quartzite-industry from the terrace-surfaces, a very clear distinction possible between arte-and these geofacts.With the 64 artefacts only two or three questionable pieces were found, clear geofacts, broken along crevices and with consequent splinter ing along the newly formed sharp crests.The greater problem is the degree of rolling: one finds relatively frequent pieces, which have most probably been ar tefacts destroyed by the river; to draw the line, what to collect is frequently difficult.The criterium used is the condition, that there must still be marks of the blows recognizable, either the point of impact, ra diating striae, bulbus or scar, or an unmistakable configuration of polished scar -forms of a typical piece, a biface or a discoid core, for instance.The nearest recent, steeper, cut-in valley-stretch, which might favour geofact-production is ca.30 km distant, too far, to deliver freshlooking geofacts to Sorde l'Abbaye, but this stretch did, anyhow, not exist in the period of formation of the 37-44 m ter race as it is now situated at a lower altitude.It must be emphasized, that this is a simple case.But even here, a considerable amount of knowledge and practical experience has to be combined with re search on the local peculiarities to enable one to arrive at a well-founded, realistic position.

Additional Observations.
The very specialized experience which the person working on a site on or in a terrace, gains, cannot be highly enough rated: he has seen thousands of peb-bles, geofacts and artefacts, learned their peculiari ties and has become familiar with them to a degree which an occasional visitor cannot acquire in a short time.To this visitor, a certain number of pieces may look acceptable, but others doubtful or even suspi cious: his reaction will be to refuse these latter.If he has enough practical experience and has gone through the same process, he will rather state that he has difficulties to follow the opinion of the worker, but will realize his own disadvantage and leave these pieces open, until he can spend more time on an intensive study.
An imperfect method frequently used is the refusal of pieces which the critic does not accept, with the state ment "this is nothing".In 90% of all cases this is, even combined with an explanation, unconvincing.It is a basic necessity to show the professional, student or amateur a way to learn and gain his own, personal ex perience.Examples are: showing the respective pie ces of ones personal collection of artefacts/geofacts or indicating places where one can see them, or collections of museums/individuals with this type of discussed artefacts.The authoritative statement "this is not an artefact" without exact explanation and not combined with an example in stone or the indication, where to study it in nature, is unconvincing.Another point of discussion is the length of time in volved to find isolated artefacts: a long period is sup posed to underline the natural origin of the putative pieces.Roebroeks states "It took Würges about two years (!) to assemble this set, which is very clearly a selection of pieces, whose number is infinitesimally small compared to the whole" (1993: 16).For any one having worked on terrace-surfaces or in gravelpits producing from thebeds, this is the usual daily routine.Numerous collections in the Garonne-area, the Nahe, the Mosel, the Wetterau or also Sorde 1'Abbaye near Bayonne would not exist, if this would be used as a criterium.Many a concentration was found lateron as a consequence of earlier isolated doubtful finds.On todays terrace-surface as well as on palaeosurfaces artefacts were loosely spread, normally iso lated; the concentration is the exception.If isolated pieces or small-concentrations are mingled with the pebbles of accumulations, they become nearly "infinitesimal", but with the necessary patience (and luck!) they can be found.Even in English or German gravel-pits producing silex-artefacts patience is re quired: MACRAE states 8 hours on average per artefact found (1988: 129) in England, a country with an ex traordinary wealth of artefacts.
Another good example is the collector Plasse, who found in the Leine-river-gravels near Hannover the first silex-flakes relatively fast, but the first handaxe after 13 years!0ACOB-FRIESEN 1949: 15).The above mentioned criticism (and other similar points) shows a lack of communication between some of the professional prehistorians and ama teurs, to the detriment of prehistory.One of the rigid rules of prehistory when consider ing the possible artefact-nature of a piece found, is: "from a collection or series every piece has, also when considered isolated, to be clearly an artefact."At least with respect to collections of altered pieces made on quartzite, quartz etc. this rule is not sensi ble.In an inventory consisting of a number of piec es, which are altered in varying degrees, from near ly fresh to nearly unrecognizable, one has, based on the part of better conserved and clear artefacts, to consider accepting as well those pieces, which one might not accept as an isolated find.There is no log ic in rejecting a piece which is well known from ex cavated or otherwise secured inventories or which constitues a regular part of many series together with bifaces, choppers, flakes, cores etc. in the gravels of the terrace, only because it is little worked and diffi cult to recognize for the inexperienced person.One should never tear a piece out of its system, its con text, into which it belongs.One might ask for a de scription of the pattern of geofact-creation for com parisons, but it must be underlined, that artificial sep aration of pieces from their context will lead to mis-

JLtdgement.
Frequently one encounters statements, that differing alterations of scars and/or parts of a piece are an in dication of differing periods of creation and that therefore these are geofacts.This argumentation is not applicable on quartzite and quartz, as on each site with artefacts from gravel-pits these contain numerous pieces with differing states of alteration, sometimes more or less limited to certain scars, sometimes to parts of the surface.The configuration of such pieces, bifaces, cleavers or chopping tools, shows however that they have been produced in one act.The differences in alteration are the results of variations of the position on or in the ground and the respectively varying conservation of parts of such a piece.

Kirchhellen and Weeze
If we apply the above explained observations and conclusions to the sites of Kirchhellen and Weeze and their inventaries, a number of points call for at tention: a) frequent experimental knapping showed, that the main rocks, of which the pebbles/cobbles consist, had certain limitations in their usefullness as tools.b) silex/flint: the numerous globular pieces (in Kirchhellen one of 11 kg), are full of fractures; if hit, they burst into many small pieces, which could after retouch be used for scraping and cutting.Such tools were not found, but have probably been dest royed by the river and the climate.Only in Kirchhel-research into the problems of recognizing stone-artefacts in pebble-accumulations of fluviatil terraces len one piece with a scraper-like retouch was found, but it is impossible to assure its anthropogene origin.c) quartz: is freqLtently full of crevices and frac tures and very coarsely structured.If hit with strong blows, it sometimes bursts into pieces, which can be used for scraping and perhaps cutting; again it is not possible to identify such specimens.According to JU STUS (1988: 49) occasionally compact blocs from the Alps have survived transport and provide better ma terial.Apart from a few artefacts on mediocre quartz, one great flake of 3,25 kgs, made of a better variety, was found.d) Palaeozoic quartzites, comprising such varieties as the Taunus-quartzite, the Revin-quartzite from the Ardennes/Eifel-region and others, many of Devonian origin (with Spirifer sp.) offer frequently magnificent-looking oval, flat pebbles or slabs, invit ing experimental knapping.If hit, they mostly separate along surfaces of schistous inner structures and other zones of weakness, offering irregular, smaller pieces of rock, useful only for small tools.They have sharp cutting edges, which blunt relatively fast, as experiments on hard wood and bone show.e) Tertiary-q ua rtzite is also difficult to work, but has important advantages over the other materi als.There exist great pebbles or blocks, well round ed, occasionally ca.60 x 40 x 30 cm, normally in the range closer to 30 x 20 x 25 cm.Frequently one finds remains of the greater blocks/pebbles, in form of prismatic, jagged or irregular blocks as well as flat pieces.If hit with strong blows, flukes will separa te from the pebbles or blocks, but they often follow zones of weakness in an irregular manner and only a limited amount of real flakes with bulbus and conchoidal forms are produced.However, in experi mental knapping flakes in the range of 10-12 cm length were produced, which might be converted into tools.The result of this work is a very resistant tool, which stands up well against use and in fact, as trials show, much better than the other rocks.This seems the ideal material for heavy work, but can also cut very efficiently.
Considering the great refuse-heaps in the pits, one should bear in mind that they offer a false impres sion: in relation to the volume of gravel and sand produced they are very small, as observation shows.So in the past these pebbles/cobbles were sparsely and widely distribttted, there was no surplus as for instance in Sorde or the Garonne-area.Men had to search and make use of what was within reach.In addition to being rare, larger cobblesAboulders in Tertiary and other quartzites have frequently strong ly rounded forms and the angle of the planes is un favorable for knapping.Having nevertheless found a cobble with favorable surface-forms and having produced at best one or two medium-sized flakes (4-8 cm), the rocks burst into irregular pieces along crevizes.Such pieces may show the form of a flake or triedre and have certainly been used, but to recog nize them as artefacts is quite difficult.The above picture is based on experimental knapping.This means, that the rocks from the Rhine-gravels do offer material for tool-production, but of irregular and mediocre quality; the resulting tools or artefacts will have their peculiar traits and a picture much more dif ficult to recognize than inventories of better material as for instance in Middle Spain, where excellent and SiO-reach quartzite is available.The quartzite from Sorde l'Abbaye is in-between the two mentioned ma terials.Sites with rocks offering even more difficult pictures are the coarse quartz of the Roussillon-terraces, very irregular limestones from North Africa (Terra Amata-Museum, Nizza) or the quartzites from Olduvai Beds III and IV, on which scar-boundaries are hardly visible and not countable; this is the ma terial used for bifaces/cleavers!(LEAKEY 1994: 265).
Here an example is used to explain the problems created by rash critizism on a difficult material.In Weeze one convex chopper on quartz has been fo und, on which 5 to 6 long parallel scars on one side can be recognized, at a rather steep angle of 70-80°.Visiting prehistorians all accepted this piece as a convincing chopper; however, one found on the lo wer, flat surface of this plano-convex piece, partly in parallel with the working-edge, fine crevizes and felt, that the postulated scars were in fact planes of some of the crevizes, where the quartz broke off un der pressure by other cobbles or boulders.As this does of course happen, the argument seemed at first sight valid; a detailed study showed, however, how dangerous these rash judgements are.On closer and more detained look it became clear, that the next parallel crevize (7 mm distant) entered the cobble at an angle of ca.45°, as do other visible rests of such: these crevizes cannot have caused the scar-like sur faces with an angle of 70-80°.It is still thinkable, that the beginning of the crevize in the lower surface could have served as a starting point.The author collected quartz-cobbles in the pit and partly work ed them, producing chopper-like pieces.Satdying these collection at home he found, that only rather marked, deep crevizes serve as a starting zone or "point" for a break-off, while the thinner fissures have apparently no influence on the internal compact ness of the quartz.If quartz breaks off at a crevize, it follows as a rule the flat, smooth surface which forms the crevize.All this shows, that the recogni tion of such a piece demands very detailled, intensi ve study and an intimate knowledge of the material, gained over a longer period.With difficult material such as this, it is not possible for a visitor to form a definite opinion just by letting the pieces run a few hours through the hand.
To be able to recognize artefacts in the a.m.sites, it is necessary to study at least the basic formation of ge ofacts in the more important tributaries with represen tative character, close to the sites.As a start, accumula tions in the river Sieg have been studied and the nu merous geofacts found show a very clear pattern of geofact-production.In quartzite, these are easy to di stinguish from artefacts; the basalt-geofacts form more frequently surprising pseudo-artefacts ("choppers"), but the general context and their proportion in the pebble-content of the river confinn the geofact-origin.In the near fuaire other rivers in the proximity will be studied as well, but as their pebbles originate mainly from the Devonian, their contents should be the same or very similar to the Sieg.It will be interesting to stu dy other rivers such as the Maas or Mosel, which have a longer/greater area of material inflow.The conclusions based on the above general and also the specific part, concerning Kirchhellen and Weeze (leaving aside local differences), are: a) A certain part of the pieces are artefacts.They dif fer strongly from the geofacts and conform with ar tefacts well-known to the author.
b) The other part contains artefacts, but of which a geofact-character of some cannot be excluded.c) Flakes are, due to their inherent problems, treated as a separate group.The flakes from Weeze do contain in the authors opinion a few, which are definitely artefacts.Having seen large amounts of flakes, including the typical large Acheulean specimens, one can discern typical traits, which mark the intentionality.As these traits are composed of a multitude of facettes, they are dif ficult to describe, but a result of experience.There is, e. g. a large flake of quartz with 3,25 kg, showing well-conserved and strongly espressed marks of blow, which differs completey from everything seen and/or found in quartz in Weeze, but compares well with large Acheulean flakes, e. g. those from the authors collections from SW-France.Observations of indirect character can also be im portant and underline the arguments connected with the lithology.One such argument is the regu larity of finds on certain sites only: if we would deal with geofacts, we should find them everywhere.Out of 15 pits visited by the author, only four contain in dustry.This includes four pits in the area of the Weezeterrace (ca. 4 x 1 km), in only two of which artefacts are constantly encountered, in the other two not; this is a mirror of the situation on the sur faces of many terraces with industry.Again: if here geofacts, presented as artefacts, are collected, we should find them in all places.

A recently found typical artefact: a cleaver
Here follows the description of this cleaver, mention ed in connection with the Tertiary-quartzite-flakes, on one of which it is made.The dimensions are: L = 172, B = 122, Th = 68 (mm); W = 1,51 kg.The tool is made on a flake of the a. m.Tertiary-quartzite, from which the bulbus has teen removed; the striae of radiation are, however, strongly expressed and easy to recognize.The form of the piece is paral lel-sided with a pointed butt.The edge is slightly obli que.The ventral face shows lateral retouch, partly in vasive.The dorsal face is worked around the butt.The working edge is battered (use; river; refreshening?).This cleaver fits well in the line of cleavers known from the Spanish Meseta, as in Pinedo and El Sartalejo (QUE-ROL 1979;SANTONJA 1985).It does not show the elegan ce known from many north-african cleavers; here the difficult lithic material intervenes unfavourably, but for an artefact of over 783 000 years it is well worked.

Kirchhellen and Weeze: conclusions
The research of the period after 1992 does produce a certain shift in accent, but the first statement is still fully valid, even strengthened: Kirchhellen and Wee ze are paleolithic sites with industry, in Weeze older then 783 000 y., in Kirchhellen either somewhat younger then the Matuyama/Bntnhes-limit at 783 000 y., or perhaps elder, with further clarification hope fully in the future.One has to accept, that in this type of site -fluvial terraces with rolled/eolized/ patinated artefacts -the recognition and acceptance of the anthropogene origin of the specimens in ques tion, which are part of a complex of pieces difficult to interprete, demand differentiation as well as an unusual degree of specialised experience.

Summary
The essential points in the afore mentioned text are: 1.It is possible to find altered, sparsely retouched ar tefacts in fluvial terraces and other pebble-accummulations: it can be learned, but needs intensive train ing and a certain talent.2. It is not possible to establish rigid general "criteria" to distinguish between arte-and geofacts.
3. Each worker, engaged with such sites, has: -to gather a wide knowledge of such artefacts, -to gather a broad knowledge of the general geofact-creation, -to start at each site anew to study conditions and context, which can successfully only be done by someone intensively experienced in these matters.
4. These problems can only be approached in an empirical way and therefore the years of experience as well as the amount of material seen count.To obtain a closer approximation to the solution of this vital question for prehistory, in our case: how to distinguish artefacts from geofacts, several ways ap pear open.There are numerous questions of the processes and their possible influences on geofactproduction still open, which could be explored; the ideal aim would be to find objective, measurable cri teria, which seems at least at present very unlikely.A first possibility would be the extension and enlar gement of the studies described in this paper.To ob serve more pebble-accumulations should increase knowledge and assist in closer limiting the area of the gray zone between artefacts and geofacts, apart from increasing and spreading personal experience.In addition, one may, even if to the author this does at least at present not seem a good prospect, attempt the study and description of the complete surface of selected limited areas on pebble-accumulations, in sq.-metres or complete bancs, and see, if this pro duces new impressions.
Another approach would be to form a conception of the forces necessary to move coarser pebbles in the rivers and to damage each other, producing geo facts, which may occasionally mimic artefacts.This would mean to enter the complex of sedimentology concerned with the dynamics and forces in rivers, their calculation and the analysis of the morphome To the author it seems a basic necessity to study as many aspects as possible of the processes in rivers and to recognize the framework of conditions and factors in the role of geofact-production, which plays only a secondary part in the total process and of this limited part pseudo-artefact-production in its turn plays again only a less important role.But de tailed knowledge of the context may assist in achie ving better understanding of the facts, important to prehistory, still accepting that a "gray" zone of inde terminable pieces will always exist under the highly complex conditions.
A different line of approach would be to systematisize technical characteristics on postulated artefacts.The above described cleaver (4.1.)could serve as an example.Traces of damages, which are postulated to be anthropogene, are visible; listing all the scars plus the modifications of the pebble-body reaches a point where, in comparison with cleavers from collections from similar circumstances as well as the geofacts typical for the site concerned, a decision should be facilitated.In our case, the cleaver, it be comes apparent, that the possibility of a conformity by hazard with a geofact can be overwhelmingly excluded, especially if such piece is found on a site where other artefacts have already been found.The following could be another possibility: in a con versation this remarke was made: the experience, the numerous artefacts of the type with which we are concerned and as well the geofacts, which you have both seen and memorized, can never be stored and programmed in a computer.At least to the au thors knowledge this has not been tried.On a very different level inquiries might be started: how far are possibilities explored, that in other disciplins of science, for instance cristallography, petrography, mineralogy, egineering, physics or others, in this period of fast developments new knowledge, technics and experiences exist, which could lead to advances in the solution of our pro blem, perhaps even a first step to objective criteria.The great majority of recent new scientific and tech nical developments came from interdisciplinary cooperation: this problem is a typical case for such.

6
Possibilities for future research Concluding this paper, possible directions for furt her research are suggested.research into the problems of recognizing stone-artefacts in pebble-accumulations of fluviatil terraces try of pebbles.(See REINECK & SINGH 1980, numerous references).