How Cleon issued his fatal order to retreat from before Amphipolis does not seem to have been discussed by the commentators. But it would appear that he violated correct military procedure, and that this violation was in itself the cause of much of the subsequent confusion in the Athenian ranks. Thucydides says that both a signal and a verbal command were given—the signal conveying the definite meaning ‘Retire’, while the verbal command explained how it was to be done; and signal and command were issued together.

Signals are divided by Arrian into three groups. Verbal orders, being most readily intelligible, are preferable when they can be heard above the din of battle. Visual signals may be obscured in the dust and confusion. Finally, the trumpet is useful in overcoming ‘atmospheric disturbances’ (τὰ ἐκ τοῦ ἀέρος ἐμπόδια).

Visual signals are found in Greek warfare from at least the fifth century B.C. onwards, but they are generally prearranged, either to convey the news that some foreseeable event has actually occurred or to coordinate the operations of two bodies of troops who are widely separated but in sight of each other. Everybody concerned must be instructed in advance that after certain developments a certain signal will be displayed, upon which certain movements will be carried out. In this way Croesus in the Cyropaedia directs the evolutions of his vast army, Antigonus plans to coordinate the separate attacks at Sellasia, and Gorgidas ends the feigned retreat of the Sacred Band. But the signals used—the display of a red cloak or a white; the raising of a helmet on a spear—are all obviously arranged for the occasion. There does not seem to have been any code by which the commander of a Greek army, at any rate in the classical period, could convey orders to his men on the spur of the moment by visual signals. In the army described by Arrian each battalion (σύνταξις) had its own standard bearer, and the rank and file could at least have conformed to the movements of the standard.

‘The most interesting excavated site, after Al Mina—Posideion, is Tarsus. This area may yet hold the key to many important problems, and provide firm associations between East and West which will give fixed points for early Greek history and chronology’ (Dunbabin, The Greeks and their Eastern Neighbours 33).

The publication of Tarsus iii (reviewed later in this volume) offers scholars the opportunity to judge how far the high hopes entertained for the absolute dating of Greek pottery found in the town destroyed by Sennacherib in 696 B.C. have been fulfilled. Hanfmann, who publishes the pottery, had already given some indication of the results in The Aegean and the Near East, 165 ff., a volume dedicated to Hetty Goldman, excavator of Tarsus. Some of the results seemed a little disturbing, like the appearance of East Greek bird bowls with rays before 696 B.C. With the publication we can see that the dating for Protocorinthian pottery of the end of the eighth century is moved back at least a quarter century (pp. 115, 129, 308), while the disturbance to the bird bowl series suggests even more radical changes, leaving something of a vacuum in the first half of the seventh century, so far as the usually accepted dates of Protocorinthian and East Greek pottery are concerned. Hanfmann does not pursue all the implications of this, nor was it his task to in this book. Fortunately the quality of the publication makes it possible to study these problems in some detail, to evaluate the evidence of the pottery and stratigraphy, and even to suggest possible accounts, different from those of the publishers, for the years around 696 B.C. In what follows I have taken Payne's chronology for Protocorinthian as the standard since none of the detailed attempts to upset it seem to me to have been at all successful, while much new evidence has appeared to confirm it. Dunbabin's.remarks on the possible margins of error have also to be remembered (in AE 1953–54 247 ff.).

The use of preliminary sketch in Attic red-figure is so widespread and so familiar that even in a detailed publication its presence often passes unmentioned, yet illustrations of it are not always easy to find. It may therefore be helpful to bring together some examples on which the sketch-work is of particular interest and at the same time to include instances of what is basically the same procedure, but applied to other techniques of vase-painting.

As is well known, the sketch is composed of shallow grooves made in the surface of the clay before firing; the lines are most obvious in the reserved areas, but it can sometimes be seen that they extend into the black background, and when they do, the shininess of the black in the grooves shows clearly that they were made before the black was applied. The exact nature of the instrument with which the sketch was drawn is not known; the grooves generally look as if they had been made with a small, blunt tool, though whether it was wood or metal or some other material cannot be determined. Each artist no doubt had his own favourite implement, but it is worth noting that occasionally, and above all on Apulian red-figure, the sketch-lines are narrow slits cut into the clay; this kind of line, at least, must have been made with something sharp, presumably a metal graver. The amount of detail in the sketch varies from man to man, and there may even be differences between works that can be attributed to the same hand, or between the two sides of the same pot. Sometimes the artist does no more than block out the general masses and arrangement of the figures; on other vases the preliminary work is very exact and on occasion may be more detailed than the final drawing.

Hellenists still debate whether the ‘wild she-mules’ (ἡμιόνων γένος ἀγροτεράων) of the Iliad (ii 852) are domestic mules or wild equidae.

The problem is further complicated by the fact that the ancients did not seem to differentiate adequately between the true wild ass and the wild hemione, although, according to Hančar's authoritative book—which incorporates the findings of modern zoology—the two are definitely distinct. There is, furthermore, a great deal of plausibility in Keller's statement that the ancients did not always differentiate properly even between wild horses and wild hemiones. In fact, the only author who seems to have had more than an inkling that there is a difference between wild asses and hemiones is Aristotle (Historia Animalium vi 36=580b). Even the distinction between wild species and the domestic mule is sometimes so blurred that only the report that a ‘mule’ had a foal enables one to decide that the animal is a domesticated hemione or wild ass. Under these conditions little help can be derived from nomenclature, though, for the sake of completeness, we will indicate whether the animal is called hemionos, onos agrios, onagros, asinus ferus, or mula.

Although imagery from music and song is not uncommon in Greek poetry as a whole, it is usually of no more than superficial significance. In Aeschylus, however, its roots strike deeper, and for that reason I have chosen to concentrate on him here. For the sake of comparison, a briefer survey of its uses in Sophocles and Euripides will be added.

Aeschylus' method of using key images to sustain and develop a dramatic theme has for some time now been recognised as an important feature of his style. Whether he expected the subtleties of his technique to be appreciated by his audience—even by a perceptive minority—is another question. His painstaking craftsmanship would tend rather to suggest that he wrote with more in view than the immediate appeal of the spoken word, deliberately shaping his work as a κτῆμα ἐς αἰεί. A number of dominant images in his drama, such as the yoke in the Persae, the ship in the Septem and the alternating light and dark in the Oresteia, have already received their fair share of attention. But this has not been so with his musical symbolism which, although less apparent, is employed with greater consistency. In no drama is it entirely absent, and it permeates the substance of the Septem and the Oresteia.

There has been a tendency in recent studies of Aeschylus to exalt Zeus or Fate into a position of supremacy from which they dictate and determine the actions and the conditions of men. The argument of this paper is that Aeschylus believed men to be free in taking some actions and at the same time recognised the limitations which circumscribe the conditions of men. This argument is developed through a study of the issues which Aeschylus set forth in the Oresteia, and it leads on to an analysis of the meaning of Moira and of the extent of human responsibility.

I take as a starting point Professor H. Lloyd-Jones' interpretation of the guilt of Agamemnon. It expresses the exaltation of Zeus and the powerlessness of man in a precise and striking manner. In his view Agamemnon had no choice when he was faced with the demand for the sacrifice of his daughter at Aulis; and even if he had had a choice he could not have exercised it, because his power of judgement was taken away by Zeus. As Lloyd-Jones puts it, ‘Zeus is indeed determined that the fleet must sail; Agamemnon has indeed no choice. But how has Zeus chosen to enforce his will?…by sending Ate to take away his judgement so that he cannot do otherwise.’ Lloyd-Jones sees the same thing happen when Agamemnon is asked by Clytemnestra to walk on the purple carpet. ‘Zeus has taken away his wits. But why has Zeus done so? For the same reason as at Aulis; because of the curse.’ Agamemnon is seen as a puppet, of which the strings are pulled by Zeus. But Agamemnon is only one figure in what Lloyd-Jones describes as ‘the grand design of Zeus’.

On the basis of Labarbe's new arguments in favour of July 21st rather than August 19th as the date of the full moon occurring about the time of the Thermopylae campaign, which involves our acceptance of a gap of something like five weeks between the Persian arrival in Attica and the battle of Salamis, R. Sealey has recently revived (Hermes xci (1963) 376–7) Munro's view that the Greeks had placed a regular garrison on the Acropolis in 480, so enabling it to hold out against the Persian assault for a considerable period and delaying any further military or naval operations. The validity of Labarbe's arguments has now been challenged by C. Hignett, Xerxes' Invasion of Greece, 449–51 (cf. A. R. Burn, Persia and the Greeks, 403–5), who also meets, 212 f., most of the arguments originally put forward in support of the same thesis by Bury (CR × 1896). Hignett himself explains the delay before Salamis, on his view a delay of some three weeks, in terms of the unwillingness of the Persians to enter the Salamis strait, coupled with their recognition that it was impossible to by-pass it. The Acropolis is held to have resisted, rightly I believe, for only a few days, a resistance to be explained by its natural strength rather than because of any garrison placed upon it. The dismay attributed to the Greeks by Herodotus (viii 56) upon the news of the fall of the Acropolis, however, still appears to Hignett a difficulty needing explanation and he suggests that Herodotus' emphasis upon it may be no more than a transitional device leading up to the next act of the drama (203).

The primary purpose of this article is to publish two terracotta representations of lute-players in the British Museum. The subject is rare, but not quite so rare as might be supposed from the scarcity of literature about it. It has, therefore, seemed worth while to add a Ust of the examples known to us—a list which does not claim to be exhaustive—and to discuss briefly some of the problems which they raise. We do this in the hope that it may stimulate further investigation of a neglected theme.

Between lutes and lyres there is a difference of principle which could hardly be more fundamental. The strings of the lyre are relatively numerous, but, in default of a fingerboard, fret-board, or neck, against which they could be firmly pressed (or ‘stopped’), the possibilities of obtaining more than one note from each string, in so far as they existed, must have been limited as to the number and quality of notes obtainable. The lute has few strings, but they are stretched over a solid neck, or a prolongation of the sound-box, against which they can be pressed so as to shorten the string-length and produce notes of higher pitch than those of the open strings; each string can thus provide a number of notes of approximately equal quality. Lutes and lyres were both common in Asia and in Egypt. In Greek lands the lyre predominated, and no examples of the lute are found in art before the fourth century B.C. The examples known to us are mostly terracottas.

The curious art of mosaic had one period of splendour: the Byzantine empire; and that art was assuredly Greek. The Greek mosaics of my title, however, are the much more humdrum productions of an earlier age. In mosaic, even more than in other arts, it is hard to draw the line between Greek and Roman, but this provisional survey of early development hardly reaches a point where that comes in question. I am not concerned with Pompeii, and little with Delos, Pergamon or even Sicily; a trifle more with Alexandria; but primarily with the pebble-mosaics of the fourth century, though looking forward as well as back. The earliest pebble-mosaics are of a simple, decorative character; later they become much more pictorial and sophisticated. The later pieces are of the greatest interest to the art-historian, but may trouble the art-critic. Mosaic by its nature is essentially an art of decoration, and can only achieve real greatness in an aesthetic ambience where purely decorative values are dominant, as they were in Byzantium; not in the Greco-Roman world, where, as in the Renaissance, half the artist's excitement comes from wrestling with the representation of nature.

Technically the pebble-mosaics of which I shall be speaking are all set in approximately the same way: a layer of coarse cement or plaster as a foundation, and above that one of finer quality into which the pebbles are pressed. Presumably the mosaicist, like the fresco-painter with his top coat of plaster, laid only so much of the fine layer at a time as he could adorn before it hardened; and as the fresco-painter normally worked down the wall, the mosaicist, I suppose, worked from one edge, moving backwards across the already hard surface of the coarse cement, laying the fine and setting the pebbles in it. In Renaissance fresco-painting the artist often drew a sinopia or outline of the composition on the under-plaster, to guide him as he painted the wet plaster with which he gradually covered it. Traces of such guides have been observed by the excavators under mosaic-floors at Pella. One would guess that it was normal practice, and that the mosaicist working on any but the very simplest pattern-design must also have followed a cartoon, but of what nature we have no idea.

Rather less than a century ago Morelli began a revolution in art-historical method by demonstrating that every painter has his formulae for rendering details—ears, eyes, hands, drapery-folds and so on—amounting to a personal system; and that, for attribution, a study of these minutiae affords a valuable check on, if not a sounder basis than, a general sense of style; or rather that the two together form the only sound basis. There is no rule of thumb. Formulae are the artist's servants, not his masters. They appear and vanish, change and merge, according to the development of his technique and style, the influences he undergoes, the speed at which he is working, all the circumstances of his art; but in much of any painter's work they will be found recurring; rarely, as a system, in another's. Morellian method can only be effectively used by one who, like Morelli himself, is sensitive to works of art not only as aesthetic achievements but as expressions of personality; but without the tools of his forging it is impossible for sensitivity alone to make much headway. The study of Italian painting before Morelli was a chaos of unchecked traditions and conflicting hunches. Despite fine work by Hartwig, Furtwängler and others, the study of red-figure vase-painting remained much the same (without the traditions) until Beazley brought to it a rare combination of sensitivity to personal artistic style with Morellian discipline.

A small Corinthian phiale mesomphalos of which a few joining sherds were found at Perachora, had been decorated outside with animals in a rather perfunctory style, and inside with a figure-scene, a frieze with the figures' feet towards the centre of the bowl (plate XXIIIa). The surface was finished differently inside and outside, a striking instance, as J. K. Brock points out, of variation in technique without chronological implications. Surface-finish and painted decoration together also exemplify how a change in the entire approach of Corinthian artists—both potter and painter, in this case—will often accompany a change of subject-matter. The filling-ornaments enhance the difference: outside, a rather dense filling of the usual solid shapes; inside, a dot-rosette only.

The subject of the inside picture is described as ‘three padded dancers, the one in the middle holding a horn’. Padded dancers are undoubtedly present, the best-preserved figure (on the right) is wholly typical, except for his excited gesture indicating that something unusual is afoot. But his neighbour is different. Nude, slim-waisted, strong-limbed, he comes striding in from the left, turning his head towards the quarter whence he came. He too has lively arms, their length expressively exaggerated; in his right hand he holds what could certainly be a misdrawn horn, though it should be said that padded dancers take better care of their drink.

I have tried to analyse elsewhere the archaeological evidence for Greek armour and weapons, and their possible effects on tactics, in the critical period of the eighth and seventh centuries B.C. There, I was of necessity concerned with the monumental evidence, and did not look far beyond it. But there are historical implications which should be faced and also, I think, some further historical support for the conclusions there reached.

The conclusions were briefly these. The equipment of arms and armour, which modern writers tend to group together as the ‘hoplite panoply’, was originally a motley assemblage. Certain of its components—the long iron sword and spear—were part of the equipment of most warriors of the era, and of many periods before and since. Other items resemble those used by Mycenaean warriors some five centuries earlier: these include the bronze plate-corslet, the greave and (an optional accessory) the ankle-guard. I cannot believe, with some scholars, that such advanced and costly products of the bronze-smith had been produced continuously throughout the Dark Age that followed the fall of the Mycenaean civilisation; and indeed for at least 400 years there is no evidence of any kind that they were. Rather, they were revived or readopted: the corslet apparently under the influence of the metal-working cultures of Central Europe and Italy, the greave and ankle-guard spontaneously, although the Epic tradition had never forgotten their earlier use. Other items again, the closed helmet of the type that the Greeks called Corinthian, and the large round shield with arm-band and hand-grip, were Greek variants devised as an improvement on foreign models, principally the metal open-faced helmets and round single-grip shields used by the Assyrians, Urartians and other Eastern peoples. The combination of all these elements together was an original Greek notion; as was their later association with a novel form of massed infantry tactics, the phalanx.

Nearly ten years ago the first specimens of a new group of painted Mycenaean sarcophagi, or larnakes, began to become known in Greece. They attracted immediate attention, and some disbelief. The group has not yet been studied as a whole, or evaluated as a relic of Aegean art, for the circumstances of discovery and dispersal have made close examination difficult. The scenes of mourning figures painted on them have considerable interest, however, and it seems timely to put together what is known about them in spite of the incomplete evidence. Once the larnakes become better known it will be a pleasurable task for scholars to relate them as harmoniously as possible to neighbouring monuments of Aegean painting and to the late Mycenaean environment which produced them.

Until the discovery of these Greek larnakes, scholars rightly believed that the practice of using clay coffins for burial was essentially a Minoan one, not Mycenaean. The great number of larnakes on display in the Herakleion Museum in Crete demonstrates how widespread larnax-burial was in the Late Minoan period, apparently gathering momentum after the destruction of the Cretan palaces. While the earliest terracotta larnakes known in Crete are as old as the latter part of the Early Minoan period, they were not used extensively among the middle classes until the fourteenth and thirteenth centuries B.C. The first preserved wooden coffins of Crete are found in chamber tombs of sea-captains and soldiers who died in the late fifteenth century. Perhaps they adapted the custom from Egypt in an age when relations between coastal Crete and the Nile Valley were particularly active. One of the coffins in the harbour cemetery of Katsaba near Knossos was painted blue; otherwise there are no traces of rich surface elaboration in the Egyptian fashion. As far as one can tell from the rotted condition of wood in the Aegean climate, the Minoans did not use external face masks, or gilding, but made simply carpentered containers for simple inhumations.

K. Latte, De saltationibus Graecorum 44, notes the possibility that κορε was originally written, representing κῶρε. There are instances in the hymn of ε and ο representing secondary ε̄ and ο̄ (cf. on 5, 30, 38, 50), though we also find κατῆχε in 38. This orthography must go back to the original written text, and helps to date the composition, being found in a few Cretan inscriptions of the third century B.C. but not later (Bechtel, Gr. Dial, ii 680 ff.). Before the fourth century, η and ω were not used at all in Crete, so that if the hymn were as early as the fifth century, as Wilamowitz asserts (Griech. Verskunst 502) without giving his reasons, we should expect either ε and ο throughout or a uniform transliteration, not the distinction that is actually apparent between η, ω for original η, ω (also for η < αε and εα, ω < εο) and ε, ο for the contraction of εε, οο, and for ε, ο lengthened by compensation. The usual dating of the hymn to the fourth or third century thus receives confirmation.

This note is meant to comment on a point arising out of P. A. Brunt's and G. T. Griffith's articles on Alexander's cavalry (JHS lxxxiii (1963) 27 fr. (at 42 ff.) and 68 ff.). In the course of their arguments, they examine two passages in which Arrian lists the grievances which, by 324, the Macedonian soldiers felt against Alexander. Brunt examines vii 6.2 f. and comes to the conclusion that the grievances there listed were mainly recent, and that the reorganisation of the cavalry into four (and soon five) hipparchies, as well as the admission of Orientals to them, took place after the Gedrosian disaster. Griffith examines vii 8.2 and concludes from this passage that some Orientals served inside the hipparchies during the Indian campaign. I here wish to point out that the two passages concerned need closer scrutiny.

It is important to notice that they must be taken closely together. In 6.2 f. we have (i) the epigoni; (ii) the King's ‘Medic’ dress; (iii) the Susa marriages; (iv) Peucestas' Orientalistn; (v) the integration of Orientals in the Companions (at length). In 8.2 we have (i) the King's ‘Persian’ dress; (ii) the epigoni; (iii) the integration of Orientals in the Companions (briefly). Chapter 6 is a λεγόμενον, its scene Susa; chapter 8 is from the main source (probably Ptolemy), its scene Opis. Chapter 8 leads on to the famous mutiny; chapter 6—obtrusively—fails to lead on to anything: in its place, it is a pointless interruption of the narrative.

In JHS lxxxiii (1963) 108 B. D. Meritt and H. T. Wade-Gery write as follows about P. Strassburg 84 verso line 8: ‘In earlier studies of this papyrus (we include our own studies, in ATL i and ii and elsewhere) it has been assumed that these words start a new clause and are preceded by a heavy stop. On this assumption, no credible restoration has yet been found: the difficulties disappear when μετ' ἐκεῑνο is taken to qualify the preceding verb [ὰναφέρ] ειν εἰς τὴν πόλιν μετ̕ ἐκεῑνο. This, we believe, restores μετ' ἐκεῑνο to a natural place in its own sentence and gets rid of a long-standing obstacle: and if we are right, this ties line 8 firmly to what precedes.’

Since the publication of ‘The Greek Kitchen’ in JHS lxxxii (1962) 121–137, fresh material for the Appendix (pp. 132–137) has come to light, mainly in the shape of terracottas which were unknown to me at the time, and I take this opportunity of correcting and adding to the original list. A number of pieces listed below were brought to my notice by Dietrich von Bothmer, of the Metropolitan Museum, New York (nos. 14, 15A, 30A, 33A, 33B), with whom I had two valuable discussions; for knowledge of no. 53B, I am indebted to Mademoiselle V. Verhoogen, of the Musées d'Art et Histoire, Brussels. Others to whom I am indebted for information and help with photographs, are: Madame Lilly Ginouvès, of the Louvre; Miss Elaine Loeffler, of the Museum of Art, Rhode Island School of Design, Providence; Signorina Paola Pelagatti, of the National Museum, Syracuse; Jean Balty, of the Brussels Museum; Bernard von Bothmer, of the Brooklyn Museum; P. Devambez, of the Louvre; R. Noll, of the Kunsthistorisches Museum, Vienna; A. Oliver, of the Metropolitan Museum, New York; N. Raumschüssel, of the Staatliche Kunstsammlungen, Dresden. Miss Alison Frantz took the photographs of nos. 45 and 74, here plate XXIX.2, for which many thanks are given, Mrs A. D. Ure was assiduous in lending rare books, and Miss Lucy Talcott made valuable suggestions on the subject of the Vienna lekythos.