The old reconstruction, widely publicized and discussed by specialists, has been on exhibition in the British Museum since 1949. It is now replaced by the new reconstruction (PL. I). The change of this very early stringed instrument, the earliest reconstructible stringed instrument in Europe, from harp to lyre, is of great archaeological and musicological interest.

The purse found at Sutton Hoo contained forty-two gold objects. Thirty-seven of them were Merovingian coins of the last decades of the 6th and the first half of the 7th century, three were unstruck circular blanks, and two were small rectangular ingots (Bruce-Mitford, 1968, 47-51; Lafaurie, 1968, 258-60, correcting Marseilles to Arles as the mint of no. 3). The coins weigh from 1.221 grams to 1.388 grams—Bruce-Mitford's text by a slip reads grains—and average 1.27 g. (= 19.6 grains), the theoretical weight of the Merovingian gold coin at this period being 1.3 grams, i.e. 20 grains. They are generally termed tremisses or trientes, this being the name of the third of the Roman solidus from which they were derived, but they were in fact somewhat lighter—1.3 g. as against 1-5 g.—and no longer represented the Roman tremissis of eight carats but the Germanic shilling of twenty grains (Grierson, 1961, 351-2). The three blanks, which are about the same diameter as the coins, are of much the same weight (1.09 g., 1.38 g., and 1.46 g.: average, 1.31 g.). The ingots are heavier, weighing respectively 5.21 g. and 4.97 g., i.e. 80 grains and 76-7 grains, and are also of slightly inferior gold. It is probable that they were each intended to represent four shillings.

This is probably the question most often asked of the metallurgist by the archaeologiist, and the object of this report is to show how far we have got in our attempts to answer it.

First, it must be clearly understood that we can only expect some correlation between ore and weapon or utensil, and therefore provenance in the early eras of metallurgy, where the bulk of the metal used emanated from primary sources, and where the amount of secondary re-used metal was negligible. This objection does not apply, however, in the case of semifinished products such as those ingots which have been made from ores and not from secondary material. This should include the plano-convex copper ingots of Late Hronze and Roman date in Britain and the ox-hide ingots of the Eastern Mediterranean.

Not long ago the theoretical literature in archaeology dealt mainly with excavation techniques and the primary analysis of archaeological data. In recent years, the successful realization of many of these empirical objectives, plus a rapidly increasing corpus of data, have motivated a younger generation of archaeologists to investigate more carefully the problems that are involved in the explanation of these data and the study of prehistory in general. This concern has produced a spate of publications which, although they often disagree radically about particular issues, are attempting (a) to investigate the theoretical structure of prehistoric archaeology, (b) to formulate a more rigorous canon for the interpretation of archaeological data, and (c) to pioneer new methods of analysis (Binford, S. R. and L. R., 1968; Chang, 1967 and 1968; Clarke, 1968; Trigger, 1968).

The first phase of development of the radiocarbon dating method covers the first two decades since its beginning in 1949. During this period it was characterized by the fact that many of the assumptions on which it had been based remained outside the effective control of the scientists. Although it was clear that its principles were basically sound and, consequently, that it could yield reliable dates within certain limits, it was not yet able to persuade all the archaeologists who should have been the first to benefit from its success. In fact, especially in Europe, most archaeologists did not give it a very favourable reception.