The concentration of prehistoric monuments, broadly of the later third and early second millennium B.C., in the upper Tay valley, has been noted by more than one archaeologist in recent years (Stewart, 1958–9; Coles, J. M. and Simpson, 1965). The multiple stone circle to be described lies no more than 2 miles (3·2 km) from the north-eastern end of Loch Tay at Kenmore, and 4 miles (6·2 km) west of Aberfeldy, at Nat. Grid Ref. NN/797472 (fig. 1). It stands some 400 feet (122 m) above sea level, near and to the south of the main Aberfeldy–Kenmore road near the farm of Croft Moraig or Morag, and has long been known as a prehistoric monument: ‘Yon's the Druid Stones!’ shouted the coach drivers to their passengers sixty years ago, and only ‘the begoggled motorist’ of the day was thought likely to miss it (Coles, F.R., 1909–10, 139). Probably the earliest record of the site is that made by Robert Burns on his Highland journey in 1787 (quoted in Gillies, 1938, 14):

This is the first report by C.B.A. Group IV (Yorkshire) on the petrological identification of stone axes since 1962 when the group undertook to help members, in part financially and in part with organisation, to obtain sections and petrological identifications of implements in their possession. Some of the sectioned artifacts are in the possession of individual members but the bulk belongs to public museum collections.

The programme was initiated by Mr D. P. Dymond, a former group secretary, and since 1964 has been continued by Mr Laurence Keen. A considerable quantity of axes remains to be sectioned. The rate of sectioning at present is about one hundred and twenty a year. Attempts by the writers to increase the rate of sectioning have been unsuccessful because it has proved impossible to interest capable technicians in the project. However, grateful acknowledgment must be made here to Mr E. D. Evens who has willingly arranged for the sectioning of virtually all the axes in the following Register. The majority of thin sections was made by Mr E. W. Seavill. The writers would also like to thank Dr F. S. Wallis, who has also kindly checked the geological part of this report, Dr D. W. Humphries, Professor W. F. Grimes, Mr J. Bartlett who has kindly made available for publication material in his keeping, Mr H. Coope, and Dr K. C. Dunham, who have made the petrological identifications.

Jacquetta Hawkes in a paper called ‘The Proper Study of Mankind’ (Hawkes, 1968) has made an impassioned plea against what she considers to be the ‘dehumanization’ of prehistory at the hands of those who have sought to apply to it numerical methods and the techniques of the natural sciences. There can be few who do not sympathize with her denunciation of aimless manipulation of data and the use of techniques for their own sakes; and many must share Clark's fear that prehistoric archaeology will serve merely ‘… to provide intellectual games for the meritocracy’ (Clark, 1967, 472). But such comments as these only prompt the question ‘what are we studying prehistory for?’ or ‘what is prehistory about?’

The well-being of any academic discipline depends upon the posing of questions like these, for if a subject is to develop at all, it must undergo periods of self-analysis when some of its practitioners take stock of what they themselves and others are doing. The analysis may be concerned with methods and techniques and may either tacitly or openly accept the existing philosophy. Or it may consider the philosophy of the subject directly, by asking if the priorities of a past age are still acceptable or whether in the different climate of thought of the present, they must be changed. Self-analysis of this kind may however produce a state of unease and disquiet.

The purpose of this paper is to give an account of the prehistoric fortifications (pa) of New Zealand, firstly by describing the cultural background as it is reconstructed by ethnographers for the period immediately preceding European contact, and then by presenting new information from two excavations which have yielded valuable results on the form and function of these sites.

The New Zealand fortifications, which are mainly of the earthwork type with timber superstructures, have long been on record, and were first described by James Cook for the year 1769 (the initial discovery of New Zealand, by Tasman in 1642, was not accompanied by a landing). Recent surveys indicate that there are about 4,000 pa in New Zealand, most distributed in coastal situations in the North Island and northern South Island, and this distribution correlates with that of prehistoric populations living by simple horticulture and the exploitation of marine and forest resources. In the southerly parts of the South Island, where climate was not favourable for horticulture and where population density was slight, there appear to be no fully prehistoric fortifications. From 1769 onwards increasing European contact introduced pigs, the white potato, muskets, metals and other items which, in combination, gave rise to radically different technological and economic patterns. This paper is concerned solely with prehistoric Maori culture.

Morphologically, New Zealand pa resemble the Late Bronze Age and Early Iron Age earthwork hillforts of north-western Europe, and many, by their size and strength, show clear evidence of engineering skill and the ability to organize large labour forces.

A number of distinctive physiographic regions fall wholely or partly within the county of Lincolnshire: the limestone and chalk uplands of the Lincoln Edge and the Wolds, the valleys of the Trent, Ancholme and Witham, and the low lying areas of the Isle of Axholme, the Fens and the Marsh (fig. 1). The scatter of finds of Later Bronze Age metalwork shows considerable local variation (fig. 2).

These differences in distribution can be analysed statistically using the Chi Squared test (Gregory, 1963, 151–66). A comparison is made between the observed distribution of objects (o), and the pattern expected if the distribution was the product of random factors alone (e). A null hypothesis is tested by applying the formula:

.

The probability that the null hypothesis is true can then be shown graphically (Lindley and Miller, 1953). A high reading (e.g., more than 95 per cent) would indicate that the observed distribution was consonant with a random scatter, while a low reading (e.g. less than 5 per cent) would suggest that some factor other than chance governed the distribution.

Within recent years pollen-analytical studies carried out in lake sediments and raised bog peats around the head of Morecambe Bay, Lancashire, have disclosed extensive evidence for anthropogenic change at the British Pollen Zone VIIa–VIIb (Atlantic–Sub-Boreal) boundary, including characteristic ‘Ulmus Decline’, and Landnam phenomena (summarized by Oldfield, 1963). The question immediately arose as to the archaeological identity of the prehistoric communities responsible. The beginnings of Neolithic settlement in north-western England were quite obscure, and have largely so remained. The principal archaeological evidence for any Neolithic penetration of the region derives from the accidental discovery, nearly a century ago, of a bog-site at Ehenside Tarn, close to the west Cumberland coast (Piggott, 1954, 295–9; Walker, 1966), and to the presence of numerous stone axes occurring as stray finds throughout the region. More recently, the discovery of ‘stone axe factories’ on the high screes of Great Langdale and Scafell, and the wide distribution of Group VI axes therefrom, have emphasized the potentialities for further investigation (Fell, 1966, and refs. therein).

The work presented in this paper arose from an attempt to study society and economy in England during the period conventionally referred to as the Early and Middle Bronze Ages. The discussion of questions of this sort is traditionally supposed to be less reliable and important than taxonomic studies aimed at the construction of a relative chronology; this is largely because the basic assumptions of typology have not been questioned very much recently. The establishment of an absolute chronology has also been a major object of research, but now that the Mycenaean import horizon has finally collapsed (Newton and Renfrew, 1970) there is a vacuum which is hardly filled by the handful of radiocarbon dates available. We have the vaguest of ideas about what was actually happening in terms of people, but our understanding of when it was happening, and in which order, is only marginally better.

Most information has come from the contents of burial mounds, but the topographical aspects of the barrows have been neglected, despite the immensely valuable information collected by Grinsell for much of southern England, including lists of all known barrows and what has been found in them. Without this information, this kind of study would have been impossible. Grinsell's maps of round barrow distributions in Dorset (Grinsell, 1959, maps 2 and 4) and Wiltshire (Grinsell, 1957, maps IV and V) suggest that even on one geological solid (in this case chalk) there are great differences in the densities of these monuments, that bell- and disc-barrows are not distributed at random, and that a study of the sizes and types of cemetery might prove rewarding, although maps at this scale can only provide general indications.

In descriptions in the archaeological literature of the marks of workmanship characteristic of certain bronze objects of the insular Iron Age, confusion exists concerning terminology and possibly (though the imprecision of wording makes it difficult to be sure without re-examination of individual pieces) concerning technology also. Macrophotographs are seldom provided to illustrate discussion of Iron Age bronzework, even when marks of workmanship or supposed techniques of manufacture are cited as chronological or cultural symptoms, so that the confusion has been able to multiply and is now endemic in work in English in the field; individual publications need not be listed in this connection. In the present paper we illustrate, describe and as far as possible distinguish one series of tool marks experimentally produced, with notes on the methods used to produce it, in the hope of making clear the need for further experimental studies and an agreed descriptive terminology. We have chosen as the subject of the study a set of marks which illustrate some ideas associated with the so-called ‘rocked graver’ technique which have been specially liable to confusion. Since our intention is not to offer any new analysis of Iron Age bronzework, but to point out in a few details criteria which may be found useful and to offer a terminology of them, the notes are mostly concerned with the experimentally produced type series; few ancient objects are cited (as examples only, not necessarily insular, and with no implication that other ancient examples of the use of the same technique do not exist) and only two are illustrated, though references to accessible illustrations of others are given.

From ethnography and social anthropology ‘the prehistorian learns how particular peoples adapt themselves to their environments, and shape their resources to the ways of life demanded by their own cultures: he thus gains a knowledge of alternate methods of solving problems and often of alternative ways of explaining artefacts resembling those he recovers in antiquity. Study of ethnography will not as a rule … give him straight answers to his queries. What it will do is to provide him with hypotheses in the light of which he can resume his attack on the raw materials of his study’ (Clark, 1957, 172). Such a controlled use of ethnographic parallels has recently been applied successfully in the spheres of art and burial practices (e.g. Ucko, 1969, 262 and references there cited) but not as yet to the study of prehistoric settlement patterns or economy. In this paper it is hoped to show how the consideration of ethnographic parallels can help us to reach some possible alternative interpretations of two classes of excavated evidence: the pits and the two-, four-, five- and six-post-hole structures found mainly on Lowland Zone Iron Age settlements in Britain. These, usually interpreted in the literature as storage pits, ‘drying-racks’ and ‘granaries’ have been taken to be characteristic features of the ‘Woodbury Type’ economy of the earlier pre-Roman Iron Age in the Lowland Zone (Piggott, 1958, 3–4 and Bowen, 1969, 13–5), bearing in mind that ‘the type site must be clearly distinguished from the economy, and the economy itself seems to have been as variable as possible within the rather narrow Iron Age technical limits, (Bowen, 1969, 13).

Chevaux-de-frise is a term used to describe the (normally stone) stakes placed upright in the ground outside the walls of early fortifications with the intention of making access more difficult for an approaching enemy, be he on foot or on horse back. The existence of this defensive technique outside prehistoric forts in Britain or Ireland was first mentioned in 1684 when Roderick O'Flaherty described the Aran Island fort of Dun Aenghus in his Ogygia (O'Flaherty, 1684, 175), and it has often been discussed since, among others by Christison (1898), Westropp (1901, 661), Hogg (1957) and most recently and judiciously by Simpson (1969a, 26). Some writers, for instance Raftery (1951, 214) and Hogg (1957, 33) have suggested that the origins of chevaux-de-frise in Britain and Ireland should be sought in the Iberian Peninsula, where they occur in greater numbers (Hogg, 1957 and Harbison, 1968, a), and chevaux-de-frise are often taken as one of the most important pieces of evidence of close ties between Spain–Portugal and Britain–Ireland during the Early Iron Age. The purpose of this paper is to put forward a hypothesis that the Spanish–Portuguese examples on the one hand, and the Scottish–Welsh–Irish–Manx ones on the other, are not so closely related to one another as has hitherto been thought, but that both are merely distant cousins in so far as both are descended from a common ancestral wooden prototype which originated probably in Central or Eastern Europe.