In his great study of practical ways and means of daily life in prehistoric Europe, our former Hon. Editor, and President, wrote of the archaeology of horse driving and riding as being ‘slight and difficult to interpret’ (Clark, 1952, 301), and these words have remained very true twenty years on. The study of horse drawn vehicles, and of items of bridling and harness, has, nevertheless, received a good deal of attention, and the leading references have been usefully brought together by Piggott (1968, 313–18), and by Littauer (1969, 300). Much more difficult is the matter of the ridden horse in temperate Europe not only for the scarcity of direct skeletal material, and the usually deplorable circumstances of its excavation, but on account of confusion in the application of terminology, zoological and veterinary, no less than in the existence of many traditional assumptions about the identity and characteristics of the earliest horse domesticators, not to speak of the animal itself. From an archaeological point of view, it has been too easy to forget the animal for the trappings, and to appear to assume that any horse could be decked out regardless of shape or size to perform specialized tasks.

In this short contribution, most gladly dedicated to Grahame Clark, an attempt is made to bring together material for an interim statement, and to consider what guide lines in interpretation can be followed in the light of information on horsemanship in adjacent regions of the Ancient World that are considerably better documented.

C-14 datings can not only provide us with estimates of the absolute age of objects or occupation layers, but also, when available in sufficient numbers, with initial and terminal datings for cultural phases, thus defining their duration. The value of this is obvious: differential duration in different areas can at last provide definitive answers to long-disputed questions concerning the direction of cultural movement.

Working with large numbers of C-14 dates is, however, not entirely free of problems. We are, for example, regularly confronted with larger differences between datings expected to be of similar age than can be accounted for by mere statistical error (Vogel, 1969a) or which can be explained by contamination or other simple causes. One can stop at this point and accept a limited testimonial value for C-14 dates (e.g. Steuer and Tempel, 1968), or one can try to go further by calculating average dates, assuming (for the most part incorrectly) that the chance of a date being too young is equal to its chance of being too old (Neustupný, 1968). The danger in this procedure is that one loses sight of the individual character of each determination: in fact one sample is much more securely associated and more closely contemporary with finds of a particular cultural phase than another, and the chance of contamination or admixture is different for each sample.

Another problem is that the number of C-14 dates that one must take into consideration is often so large that they cannot be digested without some form of graphic presentation, and for this there is as yet no uniformity of practice.

Current belief ascribes the origins of agriculture in the Nile to diffusion in the fifth millennium B.C. from southwest Asia of plants and animals the domestication of which had first begun there in the eighth millennium or earlier. A growing body of evidence is now becoming available which shows the cultural pattern in the Nile Valley at the termination of the Pleistocene to have been appreciably more complex than was previously thought and necessitates a re-examination of the evidence on which the belief for the late appearance of domestication in Egypt is based. This pre-agricultural complexity, when examined in the light of the abundant historical and ethnographic evidence for distinctive man/animal, if not also man/plant relationships in north and northeast Africa, suggests that a process of pre-adaption using the local animal and plant resources is most likely to have preceded the introduction of the Asian domesticates. Many of the indigenous practices relating to local animals and plants persisted as late as Middle Kingdom times (2052–1786 B.C.). The appearance of the Asian food plants and animals in the fifth millennium may reflect, therefore, not the beginnings of domestication in Egypt, but the replacement of genetically less suitable local species by more satisfactory, exotic forms, a situation made possible by the beginnings of regular communication between the Nile and the Levant from that time onwards.

Eight years ago Grahame Clark published in these Proceedings a fundamental study of the earlier prehistory of archery in north-western Europe (Clark, 1963), and last year Roger Mercer developed the theme in a discussion of the evidence provided by metal arrow-heads from Reinecke Bronze B to Hallstatt D (Mercer, 1970). Between these two papers there may be thought to be little opportunity for useful additional comment, but within the ambit of the Beaker cultures of Europe (discussed only incidentally by Clark, and antecedent to Mercer's survey) some evidence seems worth critical examination, and such an examination is offered here as a footnote to Clark's wide-ranging paper. The primary suggestion to be made is that certain distinctive forms of arc-shaped or bow-shaped pendants from a restricted group of Beaker graves are in fact miniature representations, in some detail, of actual archers' bows of uniform type.

The investigation of settlements as functional units is one of the cornerstones of the approach to prehistory that Grahame Clark has done so much to foster over the years. I hope therefore that the following account of the earliest phases in the development of one which played a role of major importance in the prehistory of Crete, and so of Europe, may be an appropriate contribution to the present collection of essays in his honour.

The beginnings and subsequent expansion of the Neolithic community of Knossos has become fully intelligible for the first time as a result of the two seasons of excavation carried out in 1969 and 1970. The early soundings of Evans and Mackenzie, though they indicated that the Neolithic deposit had covered a large area, threw little or no light on the growth of the settlement, and, apart from the Late Neolithic houses in the Central Court, which were cleared in 1923–4 (Evans, 1928, 1–21), none at all on its nature.

In der Fachliteratur wird oft darauf hingewiesen, dass die Schlüsse über die Dichte der urzeitlichen Besiedlung in bestimmten Gebieten, die hauptsächlich aus kartographischen Darstellungen der Fundplätze auf Grundlage des der-zeitigen Forschungsstandes abgeleitet werden, in der Regel zu früh gezogen werden und dem tatsächlichen, bisher noch weitgehend unbekanntem Stande nicht entsprechen. Manchmal erbringen weite Territorien, die als von manchen Kulturen unberührt gelten, erst unter dem Drucke konzentrierten Interesses und systematischer Forschung Beweise darüber, dass sie keine Ausnahmen darstellen, sondern gleichfalls den vorgeschichtlichen Gesellschaften in den jeweiligen Phasen ihrer anthropogeographischen Entwicklung Unterkunft boten. So war dies auch bei der postpaläolithischen und mesolithischen Besiedlung in Mitteleuropa und namentlich auf dem Gebiete der Tschechoslowakei, von welchem es noch knapp vor dem zweiten Weltkrieg (J. Skutil, 1936) nicht möglich war, verlässliches und beweiskräftiges Fundmaterial jener Periode zu erbringen, für deren allseitige Erkenntnis sich J. G. D. Clark schon damals besondere Verdienste erwarb.

During the last fifty years, research in the Late Palaeolithic and Mesolithic periods in Denmark (to use the now generally accepted terms) was decisively influenced from outside. The interesting paper by Gustav Schwantes (1928) was unfortunately published when there was little activity in Denmark regarding ‘Early Stone Age’ problems. When J. G. D. Clark a few years later wrote his classic work The Mesolithic Settlement of Northern Europe (1936), the situation was different, and the book came just in time to stimulate one of the real active periods in the study of Danish Mesolithic. Since then many new finds and excavations have been known from this country but without altering the main lines and the general picture of the hunters and fishers during the Boreal and Atlantic phases.

It would not be quite correct to say that our idea of Clark's first period in question has remained unchanged. The following lines are meant as a short survey of the time of the Northern reindeer-hunters or, according to later suggestions (e.g. by J. G. D. Clark (1950, 91)) the Late Palaeolithic. In 1936 it was only possible to register stray finds of implements here (mainly from reindeer-antler), and following Schwantes most of them were ascribed the still rather theoretical Lyngby culture (Schwantes, 1923; Ekholm, 1926). During the late thirties Alfred Rust began his revolutionary excavations in Holstein (1937; 1943), and they were, of course, followed with keen interest also by Danish archaeologists.

In the first of his three chapters forming Aspects of Prehistory (1970), Grahame Clark discussed one important aspect of prehistoric archaeology about which he has not hitherto written much, namely the historical development of that branch of historical science which, by his own researches in the last forty-five years, he has done so much to further. This chapter, the first of three lectures he delivered at Berkeley in the spring of 1970, is entitled ‘The Relevance of World Prehistory’, and is mainly concerned with listing the story of the discovery of archaeological facts. The interpretation of archaeological facts is of equal importance and in this paper I am concerned briefly with some aspects of the history of archaeological theory and interpretation, and more particularly with the changing models of thought used by prehistorians.

Here, we need not go further back than the three-age technological model of C. J. Thomsen. Before that we note the mythological models of the medieval and later antiquaries who floated Trojans and Phoenicians and the sons of Noah across their invented prehistory, and the literary and theological models of the later antiquaries of the seventeenth and eighteenth centuries who, eschewing invention, sought prehistorical information in the Bible and in Classical writers, and made over our prehistoric monuments to the Druids, and our gravels to an Universal Deluge.

The great family of Carp's-Tongue Swords clearly embraces a number of varieties. The anatomization of this considerable mass of material has not yet received the attention it deserves; but the subject is a slippery one, and it is not easy to isolate a starting-point for analysis.

One assemblage, however, of mutually related, idiosyncratic, and uncommon forms can readily be identified, and their position examined. The material consists of two interlocking groups together comprising fourteen swords (or hilts) from ten finds; and a third, closely akin to the second group, represented by three daggers, distinguished alike by a highly specialized form and an uncommonly scattered distribution.

All but one of the swords come from hoards, but I cannot detect that this material significantly differs from the general run of hoards in which Carp's-Tongue swords are commonly found all over Atlantic Europe. In broad terms they may on current views be assigned to Late Bronze 2 or 3; for the strikingly devolved forms found at Monte Sa Idda need be no more than provincialisms. So far as is known all the daggers are unassociated finds.

The following is a preliminary report on excavations undertaken at this site under the aegis of the Department of Archaeology and Anthropology of the University of Cambridge. The work was carried out during a number of seasons over the past ten years, partly as a research project and partly as field training for third year and research students under the direction of C.B.M.McB. To P.C. fell the eventual task of collating and summarizing the extensive stratigraphical observations made by us and by previous excavators, and doing the same for the pollen samples and palaeontological data. Responsibility for the report as a whole is shared, but many others too numerous to thank separately at this stage have contributed basically to the collection and analysis of field and laboratory data. It is hoped that the full results after further field work will provide the material for a detailed monograph. The work would of course have been impossible but for the kind permission of the Société Jersiaise and the active assistance of many of its members on numerous occasions.

The site, the largest and most productive cave or rock-shelter site in the British Isles, was originally made famous by the discovery of a rich Mousterian industry, fauna, and eventually fossil traces of Neanderthal man at the turn of the century.

From their onset, the first radiocarbon dates gave a range of the absolute chronology to come, but in their detail, they opened more problems than they settled, chiefly because of the possible or unsuspected questions in relation to the reliability of the samples themselves (Delibrias and Giot, 1970). It is only with the experience of great numbers of dates, and the possibility of considering them so to speak statistically, that one can evaluate the real implications of the time scales provided by the method.

In Brittany, beginning with a few dates provided by the Groningen Laboratory under H. de Vries, we have been afterwards nearly totally supplied by the Centre des Faibles Radioactivités at Gif-sur-Yvette (Giot, 1960, 1961, 1962, 1963, 1965, 1966, 1967, 1968, 1969, 1970, 1971; Coursaget and Le Run, 1968; Delibrias, Guillier and Labeyrie, 1964, 1965, 1966, 1969, 1970; Coppens, Durand and Guillet, 1968; Vogel and Waterbolk, 1963).

We now benefit with more than 200 radiocarbon dates for Brittany alone. We shall consider here about 140 of them, disregarding some duplicates, dates pertaining to periods older than the Neolithic cultures or on the contrary later than the Iron Age, and dates only concerning geological natural sites, though these can be full of interest by their information about the botanical scenery and the effects of cultivation or pasture.

Zu den eindrucksvollsten Geländedenkmälern der keltischen Zeit zählen nebst den grossen Stadtanlagen (oppida) die sog. Viereckschanzen. Ihre Hauptverbreitung (fig. 1) liegt im süddeutschen Raum, zwischen Main im Norden, dem Alpenrand im Süden, dem östlichen Schwarzwaldrand im Westen und dem Inn im Osten (Schwarz, 1959). Einige Schanzen finden sich auch in Frankreich (Schwarz, 1962, 74). Die Geschichte der Erforschung dieser Denkmälergruppe geht bis in das Jahr 1885 zurück und die im Laufe der folgenden Jahrzehnte ausgesprochenen Deutungen bewegen sich zwischen Befestigung, Gutshof, Viehpferch und Kultplatz (Schwarz, 1960, 7 ff., vgl. auch Berger, 1962, 26 ff.).

Eine dieser Schanzen konnte in den Jahren 1958 und 1959, dank der Unterstützung durch die Deutsche Forschungsgemeinschaft, zu einem grossen Teil untersucht werden. Sie liegt im östlichen Teil der Schwäbischen Alb bei dem Ort Tomerdingen (fig. 1), nordwestlich von Ulm an der Donau. Die bis jetzt noch nicht publizierten Ergebnisse sollen hier in einem Vorbericht bekannt gegeben werden. Sie gewinnen insofern an Bedeutung, als sie die aufsehenerregenden Grabungen von K. Schwarz (1958, 1959, 1960, 1962, 1963, [1968]) in der Schanze von Holzhausen südlich von München zu bestätigen vermögen und aufzeigen, dass die Situation in solchen Schanzen, auch wenn sie weit voneinander liegen (von Holzhausen bis Tomerdingen sind es rund 140 km) etwa dieselbe ist.

My initial Australian fieldwork was still a year ahead when, surprisingly, I featured anonymously on a verbal distribution map confidently sketched by the Disney Professor. I was the ‘Australia’ reference of his Inaugural lecture, thereby illustrating his proposition (Clark, 1954, 33), that ‘the geographical field of Cambridge prehistorians has been … the whole world’. My Cambridge contemporaries, Sieveking and Golson, respectively represented ‘Malaya’ and ‘New Zealand’.

We were indeed the forerunners of a sizeable band of hunter-diggers which ventured into the Pacific world and which numbers Cambridge amongst its totemic centres. Such incipient, though benevolent imperialism, has not gone unchallenged. A participant in an Australian archaeological congress recently deplored what he judged to be the errors and dogmatism of the ‘Cambridge school’. He erred in affixing this label, because there is no such implicit cohesion within the ranks; diversity is already a healthy characteristic of Pacific research.

What was relevant about Cambridge training, but no more so than of any sound methodological discipline, was that it encouraged respect for, and critical evaluation of, evidence. What was intellectually satisfying to the student were the concepts of universality and interdependence in past human affairs. Such is the message from Miles Burkitt's Prehistory to Grahame Clark's World Prehistory.

The attributes to be defined in this paper have been designed to characterize the main areas of variation of those scrapers which have been termed ‘Aurignacian’ (de Sonneville-Bordes, 1960, 29), owing to their numerical pre-eminence in Aurignacian assemblages. These attributes were worked out primarily on the Aurignacian assemblages (Levels 6–14) from the Abri Pataud, Les Eyzies (Dordogne), although to test the general applicability of the method, samples from the Grand Abri de la Ferrassie and from the Abri Castanet (Musée des Eyzies) were also considered.

As a group, Aurignacian scrapers may be distinguished from end-of-blade scrapers in that the majority of Aurignacian scrapers are manufactured on blanks produced by a non-blade technology. The term ‘Aurignacian scraper’ includes: (1) carinate scrapers (fig. 1), (2) nose-ended scrapers, (3) some shouldered scrapers (see p. 260) and (4) miscellaneous steep scrapers on flakes (fig. 2).

‘There thus appears to be good reason at least to believe that the Larnian is best known in its para-Neolithic tradition associated with the coastal phenomena of the Raised Beaches and their middens, and perhaps even to question the Mesolithic status of the whole industry’ (Herity, 1970).

The Mesolithic settlement at Morton lies some six miles north-north-west of St. Andrews in Fife, Scotland (fig. 1). The area around the site, north-eastern Fife, is surrounded on three sides by water, and projects as a blunt peninsula into the North Sea. The river Eden and the Firth of Tay serve as southern and northern boundaries of this peninsula, which consists today principally of the wind-deposited sands of Tentsmuir. The area of Tentsmuir, now afforested, is one of the relatively few regions in Britain where land is building up into the North Sea, and this process has been in action for many centuries (Sissons, 1967).

The Tentsmuir Sands are a prolific source of later prehistoric and early historic finds, the earliest yet known being of the late third millennium B.C. In February 1957, while searching for such material on Tentsmuir, Mr R. Candow of Tayport collected some flints from molehills and other exposures on the high ground of the ‘Old Quarry’ field at Morton Farm (National Grid Reference NO 467257). Surface collections continued to be made until 1963 when excavations of the site were undertaken by Mr Candow in collaboration with Dundee Museum and Art Gallery (Candow, 1966). A total of 41 trenches were excavated between May 1963 and April 1967, and these are shown on the plan of site A (fig. 4, no. 1–41). In November 1967, site B in the same field was discovered and partially excavated (fig. 29, no. 51), before work was suspended and the writer was invited to continue the investigations.

The purpose of this article is to bring together further information from Epirus which is relevant to the consideration of man's environment in Palaeolithic times in Greece. Some aspects of this study have already been published elsewhere (Dakaris et al., 1964; Higgs and Vita-Finzi, 1966; Higgs et al., 1968) and it is sufficient here to give only a brief summary of earlier conclusions.

It has been thought that there are good grounds for believing that during the Last Glaciation mobile human groups occupied annual territories which ranged from the western coast of Greece to the central Pindhos mountain range (Higgs et al., 1968). Within these annual territories there were two groups of sites. One group, represented by Kastritsa, was at the base of the Pindhos and was occupied in summer, and another group of sites such as Asprochaliko was occupied in winter in warmer regions nearer to the coast. Together these groups exploited seasonally complementary resources which were so distributed that such a pattern was necessary.

The question arises, however, was Asprochaliko in fact a more suitable site for occupation during the winter than Kastritsa? An attempt was therefore made to assess the living environment of the two caves themselves. While such an attempt was immediately relevant to the problem with which we are concerned, it has wider implications with regard to the choice of cave sites for human habitation (Legge, 1972).

This preliminary report is the result of an archaeological expedition to Uganda, which took place from February to April 1968, and was sponsored by the Musée royal de l'Afrique centrale at Tervuren (Belgium) and the Comité des Fouilles beiges en Afrique. The expedition staff consisted of Dr F. Van Noten and Mr E. Vertriest of the department of Prehistory at the Tervuren Museum, and the author. The voluntary help of Miss J. Renard during the excavations is gratefully acknowledged here, as well as the advice of Professor M. Posnansky and the administrative assistance of Mr P. Bulenzi of the Uganda Ministry of Culture and Community Development. Much useful information was given by Mrs M. McFarlane of the Uganda Geological Survey, who has elsewhere pointed out the archaeological importance of the Buvuma Island group (McFarlane, 1968). Mrs McFarlane also deposited a series of Sangoan implements in the Uganda Museum. All drawings in this paper are by Miss N. Nypels.

The islands of Buvuma and Bugaia form part of the archipelago situated near the northern shore of Lake Victoria southeast of Jinja and about halfway between Entebbe and the Kavirondo Gulf. Both islands lie immediately to the North of the Equator, approx. between 0°00′–0°15′ North and 33°15′–33°20′ East, Bugaia being to the southwest of Buvuma (see the 1: 50.000 Dept. of Lands and Surveys map, Magyo, series Y 732, sheet 72/4, edition I-U.S.D.).

As will be apparent from this report, several interesting sites were encountered on both islands, two of which were excavated, one on Buvuma and one on Bugaia.