(BEING CONTINUED FROM 23/05/16)
We do not have any literary evidence about the exact nature of Pronomos’ invention, but we do have two passages which testify to its effectiveness. One is from Simonides (perhaps mid-fifth century BC) and the other from Plato, writing in the 380s but pretending to report conversations in the last decades of the fifth century.22 Both of them use the word polychordos (‘many-stringed’) in connection with the aulos. Plato is seeking to restrict the musical education of his ‘Guardians of State’ to the ‘mood’ or ethos associated with two particular scales (this is discussed in Chapter 3) and he mistrusts the aulos, because it is capable of playing all the notes needed for several other scales, and can therefore skip about from one ethos to another, in a way which he found highly objectionable. The lyre and kithara are the only instruments acceptable to him, because they could not change scale without pausing and retuning. The strange thing is that both of these writers should use the word ‘many-stringed’ to describe a wind instrument, though we have already seen that virtually all the terms relating to notes, scales or intervals were derived from the structure or playing techniques of stringed instruments, with the single exception of diesis, which was derived from the aulos.
It seems probable that the ability to modulate (as we would say) had been encountered earlier than the mid-fifth century in instruments of the harp type (to be described later), which were literally ‘many-stringed’, having perhaps twenty strings as compared with the seven on a lyre or kithara; as a result the word polychordos may have been already established as a technical term simply meaning ‘versatile’ and applicable to any type of instrument. Mention has already been made of the possibility of playing harmonics on the aulos, and this whole question must now be investigated. It would be very surprising indeed if the Greeks had failed to discover that by ‘overblowing’ and tightening the embouchure it is possible to make a reed instrument play the same pattern of intervals as its normal scale, but in a higher register. Because the aulos had a cylindrical bore, this higher register must have been ‘at the twelfth’ —that is, an octave and a fifth higher, as it is for the same reason on a modern clarinet. (The recorder and the flute both have a cylindrical bore, but can produce the octave harmonic; this is because they are not reed-blown, and are open to the air at both ends of the air-column.) There is some ancient evidence, dating from the mid-fourth century BC, that aulos-players made use of this technique. The most telling evidence comes from Aristoxenos’ Harmonics, where he is talking about the upper and lower limits of pitch in practical use in music.23 Having said that the total range of the human voice from low bass to high treble is more than three octaves, he says that an even greater range of notes can be played on an aulos by the use of a device which he calls a syrinx. This word is the regular name for the panpipe, a rustic instrument of the flute type, which will be discussed in detail later. But in this context it probably signifies a device analogous to the ‘speaker’ on a modern clarinet, which makes it much easier to obtain the harmonics. The word itself can be used to mean any kind of tube or pipe, which gives some clue to its shape. The same device is also mentioned in the work on acoustics attributed to Aristotle (see Chapter 5), which was probably written at about the same time, mid- to late-fourth century BC, and by one or two other authors.24
The only real problem is that the authors disagree as to whether the syrinx was ‘drawn down’ to obtain the higher register (Aristoxenos) or ‘drawn up’ (Aristotle, Plutarch). An ingenious answer to this problem was given more than a century ago by A.A.Howard.25 He suggested that the syrinx was a small tube which fitted around the upper end of the aulos body, similar to one of the sleeves but without a hole in it. This would normally cover a small ‘speaker hole’ drilled in the body of the pipe. When the upper register was needed, the syrinx might have been moved either up or down on the pipe, so as to uncover the speaker hole—hence the apparent disagreement in the evidence. The Aristotelian work mentions ‘stopping’ (with the finger) as an alternative to drawing down the syrinx to obtain the lower register, saying that it ‘increases the volume of sound, just as with thicker strings’. Aristoxenos also uses a verb syrittein (derived from the noun syrinx) meaning to ‘play on the upper register’, as against aulein, to ‘play on the normal register’. There is also some evidence to support this interpretation to be found in the surviving fragments of auloi. One from the Athenian Agora, which is clearly a bulb (holmos), has a small hole drilled in it at about the right distance from the mouthpiece end.26 Another similar fragment from Delos has a similar hole, about 1.5 mm (0.06 in.) in diameter—obviously much too small, and in the wrong position, to have been a fingerhole. The Reading aulos also has a slightly larger hole in the mouthpiece section above the bulb. It is difficult to say with any certainty whether these holes are in the right position to act as speakers, but as a speaker hole does not determine the pitch of the notes, but merely prevents a full pressure-wave from building up in its vicinity, its position is not critical. Even if there were such a device on the ancient instruments, however, it cannot be compared with the speaker key on a modern clarinet.
Almost invariably, it would be necessary for the player to alter the setting of the sleeves when changing register (unless he were playing a musical phrase immediately followed by the same phrase a twelfth higher or lower, which seems unlikely). It is much more credible that such changes would be made during pauses in the music. If, as Aristoxenos suggested, the total range of an aulos over both registers was more than three octaves, there must presumably have been a gap between them, unless the lower register covered a twelfth, which is hardly possible. There is one other interesting passage which sheds a little light on the syrinx. In the treatise on music attributed to Plutarch27 we are told of an aulos-player called Telephanes from Megara, who disapproved of the syringes (plural) to such an extent that he would not allow the instrument-makers to fit them on his instruments (or on any instruments —the Greek is ambiguous) and for that reason he refused to take part in the musical competitions at the Pythian games. There are two possible interpretations of this. One is that he disapproved of the special earpiercing effects which could be obtained by their use (perhaps during the ‘Pythian nomos’ to imitate the dying hisses of the serpent, see p. 5), regarding them as ‘gimmicky’. The other is that he was himself able to play the harmonics without them, and regarded them as a cheap trick by which inferior players could appear more skilful than they really were. This brings us to the general question of range. The early, simple form of the instrument was bored to play in one of the old aulos scales— Dorian, Lydian or Phrygian. The compass of these scales is not known for certain, but it was probably about an octave or less (see Chapter 3). The scale system and the aulos key system presumably developed together and along the same lines, and by the time the Aristoxenian two-octave ‘complete system’ had been evolved (see p. 88), the aulos probably had a potential range of at least an octave and a half, not forgetting, of course, that at any one time only six of the possible notes would be available, all of them within the reach of the fingers and thumb. By this time the old scale-names had come to denote little more than key or pitch (again, see Chapter 3) and a new set of names for the aulos types came into use. There were apparently five of these, listed in descending order of pitch:
(1) parthenikos, or ‘girl’ aulos, presumably ‘soprano’. (2) paidikos, or ‘boy’ aulos, presumably ‘treble’ —perhaps of about the same size and pitch range. At a guess, these were probably about 25— 30 cm (10–12 in.) long, and would have a lowest note about a fourth or a fifth above middle c (about 356 Hz). (3) kitharisterios, or ‘aulos-to-go-with-the-kithara’. This was almost certainly the instrument occasionally shown being played with the kithara, and which presumably had about the same pitch range. It appears to have been about 35 cm long, and would have had a lowest note just below middle c (about 244 Hz). (4) teleios, or ‘complete’ aulos—either because it had the range of an adult male voice (teleios), or because it could play any selection of six consecutive notes from the ‘full’ system (systema teleion) of two octaves. The Pompeian instruments were probably of this type; the longest of them is about 54 cm (21 in.) which, allowing 3 cm (1.2 in.) for the reed, would have a lowest note around e flat below middle c (about 150 Hz) —roughly the lowest note of a modern B fiat clarinet. (5) hyperteleios, or ‘super-complete’. This appears only in a few illustrations, the best being a wall-painting from Herculaneum, thought to be based on an earlier Greek painting of perhaps the third century BC. (see Figure 8.14 on p. 199). The pipes appear to be about 90 cm (3 ft) long, which would give a lowest note around G, an octave and a fourth below middle c (about 95 Hz). However, to judge from the picture, the player is using a group of notes played from holes about two-thirds of the way down (i.e. from about d upwards); if he used the note from the whole pipe, which the Greeks called the bombyx, or ‘boomer’, it would have been a sort of pedal note a fifth lower. In addition to the types of aulos named from the scale they were designed to play (Dorian, Phrygian, etc.) and those named from their range of pitch, there is a bewildering list of some 35 other types. Some were apparently named from a particular feature of their construction, such as the hemiopos or ‘half-holed’ aulos.
Ancient scholars explained this as meaning ‘half-size’, or ‘with half the usual number of holes’, taking the second element of the word to refer to fingerholes. But this is to strain the sense of the Greek, and it is much more likely to have meant ‘with half-size bore’ — perhaps a pipe with a ‘throat’ or diaphragm with a small central hole at the lower end, to make the tone softer and more mellow. If this is so, the typename hypotretos (‘not fully bored’) may have meant the same instrument, and a mysterious device called the pantreton, which is referred to only once,28 may have been a key or lever by which the full bore of the aulos could be opened up, and its tone made much louder and clearer. Another interesting type-name is the magadis aulos. Until recently it was thought that the magadis was a stringed instrument of the harp type, with perhaps as many as 20 strings, and that the technical term magadizein (to ‘magadize’), which meant to double a melody in octaves, was derived from the name of the instrument. Now most scholars agree that it was the other way round, and that magadis was a descriptive term meaning ‘playing in octaves’ (something like our ‘diapason’) and not the name of an instrument.29 Some of the other names for types of aulos may in fact be names of compositions for aulos, and most of the rest are unintelligible.
The most vital and difficult question about the aulos has been left until the last. How were the two pipes of the aulos combined? Did they sound together, and if so, was it in unison, or in some kind of counterpoint, or in octaves, or twelfths? This question has been discussed at length by many scholars, but no firm or generally agreed conclusion has been reached. The question is greatly complicated by the fact that the techniques used by Roman tibia-players may not have been the same as those of the Greeks, and evidence from (say) the first century BC onwards may refer to contemporary Roman practice, or may be derived from a Greek source and refer to the Greek practice of two or three centuries earlier, and it may be very difficult indeed to decide which. A specific example of this problem arises in connection with the ‘Phrygian’ aulos. To the Greeks, this apparently meant an instrument bored to play the Phrygian scale, but otherwise similar in design and appearance to the rest, while to the Romans it seems to have been a different type, with (perhaps) unequal pipes, one or both of them fitted with a horn bell. This question is discussed later, in Chapter 8. The evidence on which we have to rely falls into four categories. The first, and by far the most valuable, is the pictorial evidence from vasepaintings. This dates from the mid-sixth century BC onwards, and covers a period of more than two centuries, over which there are a number of features which appear consistently, and must surely be significant. The literary evidence is much less satisfactory; in fact, there is no explicit statement about the piping techniques of the Classical period.
The surviving instruments and fragments of instruments are of little help, since no two of them can be said with any certainty to have formed a pair.30 So the one piece of evidence which could settle the question once and for all is denied to us. Finally, there is the evidence drawn from comparative musicology—from the way in which pairs of pipes have been played in folk-music of more recent times. This type of evidence is favoured by many historians of wind instruments, but seems to me to be very unreliable. It cannot possibly ‘prove’ anything about the ancient Greek practice. At best, it can only show that a given technique is possible; and, unless it is drawn from a musical culture which resembles that of the Greeks in the most important respects, it is of little value. This effectively rules out any non-European culture; and it would be difficult to find a folk-music tradition in present-day Europe which has not been profoundly changed over the intervening centuries. In dealing with this very complex question, it seems best to begin with what is, in my opinion, the most credible theory, discussing the evidence which supports it and that which may tell against it, and assessing its implications for Greek music.
The other theories which have been put forward will then be reviewed in turn, with discussion of the evidence offered in support of them, and of the serious objections which can be raised against them. The pictorial evidence suggests that the two pipes of the aulos sounded together, and in unison. In the vast majority of pictures the pipes appear to be of equal length. One may appear to project further than the other, but this is almost invariably the artist’s way of trying to indicate that it is nearer to the viewer—a crude attempt at perspective drawing. The player’s hands are shown in exactly corresponding positions on the two pipes, and where fingers are shown raised, it is normal for the same ones to be raised on each pipe, indicating that the fingering position is the same. And finally, in the few illustrations which show the fingerholes uncovered, they appear to correspond in position on the two pipes, so as to have sounded the same notes. However, there must have been a very good reason for using two pipes rather than one. The player had to prepare and manipulate two reeds, and most woodwind players would agree that one reed gives more than enough trouble.
He also had to blow much harder, and had to control the embouchure on two reeds, using a region of the lips on each side which is less supple and less sensitive than the central region. Why then go to all this trouble, merely in order to play the same note on another pipe? The reason is that the two pipes together produce a totally different tone quality. The two notes are very nearly, but not exactly the same pitch, and this produces a beating or tremulant effect; a similar sound is made by the vox humana stop on a modern organ, which has two metal reeds for each note, one very slightly out of tune with the other. The degree of pitch difference, and hence the speed and intensity of the ‘beats’, could be controlled by a skilful player, and no doubt contributed to the mood or ethos of the music. The technique of ‘separation’ and ‘putting together’ of the two pipes (see p. 30 above) may have been used for this purpose. By contrast, the double pipe most commonly used in the Middle East today, the zummara, has its pipes fixed together, and has single reeds which are not very effectively controlled by the embouchure. I have only encountered one illustration which is seriously inconsistent with this theory.31 It shows a scene of Dionysiac revelry, with the unusual combination of aulos and kithara, both played by satyrs. The one playing the aulos is shown in profile, but the right-hand pipe, nearer to the observer, is actually shorter than the more distant one, and the right hand seems a good deal further down the pipe than the left (Figure 2a.14 p. 44). Although the drawing is meticulous enough to show the little finger of each hand tucked under the pipe to support it while the thumb is taken off the thumbhole, it is still possible to see the hand position as a mistaken attempt at perspective, which may have been partly due to yet another odd feature. In accordance with the convention of black-figure painting, the flesh of female figures is painted white, but so also are the pipes of the aulos. Putting the hands in line would have created a confusing mess of black fingers and white pipes, which the artist may have wished to avoid. Of the other theories which have been advanced on the use of the two pipes, four deserve mention.
(1) It has been suggested that one pipe was a drone, sounding a single note throughout a piece of music. This is hardly a tenable theory. All the evidence cited above shows that both pipes had more than one fingerhole, and both were fingered at all times. It is scarcely credible that the fingers of one hand would be wasted in an unnecessary operation, limiting the ‘chanter’ or melody pipe to six notes, when other much better arrangements were possible. The word bombyx, which for the purposes of this theory was interpreted to mean ‘drone’,in fact means the lowest note obtainable on the aulos, or any other bumbling or buzzing noise. There is, however, another technique which involves a drone note, which might be reconciled more easily with the evidence. It is used on several folk-instruments with double pipes (e.g. the Sardinian launeddas), and is therefore favoured by those who rely on evidence from comparative musicology. The technical name for it is an ‘articulated drone’, and it works as follows. The two pipes have a range of (say) six notes each, the two lowest on one pipe overlapping with the two highest on the other (c’ and d’ in Figure 2a.15). When the melody (I use the term loosely) rises above these notes, it is played on the higher pipe, the other one sustaining one of the overlapping notes as a drone; when it falls on one of the overlapping notes, it is played on both pipes in unison, and when it goes below those notes it is played on the lower pipe, while the higher one sustains the drone note. Anyone listening who could not see the player’s fingering might think that he had a pipe with a range of ten notes and a drone pipe with two. The objection to this theory is that pipes designed to play in this technique are usually unequal in length, the difference being quite unmistakable, and this is not what the vasepaintings show. Even if we hold that the pipes were made equal for the sake of a symmetrical appearance, the player’s hands ought not to be level on the two pipes. Apart from the awkward example discussed on p. 43, the evidence is against this.
(2) The theory offered in the older textbooks was that one of the pipes played the ‘melody line’ and the other played an ‘accompaniment’. There is, in fact, some evidence to show that when two notes sounded together in harmony, the melody was the lower of the two, so perhaps ‘descant’ would be a better term. The main evidence in support of this theory comes from Varro, a Roman writer of the first century BC, and should, in my opinion, be disregarded as evidence for the Greek techniques. It is discussed in Chapter 8, in relation to the Roman musical practice. It is entirely reasonable to suggest that on occasion (at a cadence, for example, or at the end of a piece) the two pipes might sound a fifth apart, a fourth apart or an octave apart, these intervals being regarded as harmonious by the Greeks; but there is virtually no evidence for polyphony (in any real sense of the term) in Greek music.32 (3) It has also been suggested that the two pipes did not sound together— that the player stopped one from speaking, either by tonguing the reed, or by some key device, and that the two pipes between them were thus able to play an extended scale. This, I believe, should be discounted on the grounds of improbability.
A single pipe, fingered by both hands, is capable of a greater range of notes, because it can be held more easily, and there are more opportunities for cross-fingering. Also, as was said earlier, the difficulties of using two reeds would have been avoided if possible. (4) Finally, there is the theory that the two pipes played the same melody, an octave apart. This is open to the same basic objection as the articulated drone theory—that the pipes should not be of the same length (in fact, one should be twice the length of the other) or at any rate that the hands should not be level. This has been answered by the suggestion that the two pipes were of the same length, but that one of them was played in a higher register. However, as we have already seen, the aulos had a cylindrical bore, and the higher register, produced by the use of the syrinx, was not an octave but a twelfth above the lower register. This would have given a strange sound, which does not seem to have any parallel in other double-piping traditions. I am inclined to think that the octave effect was used at times, but that a special type of aulos, the magadis aulos,was designed for the purpose. It is possible that this is the instrument mentioned by Herodotus, and called ‘the female and male aulos’, used to accompany the soldiers of Alyattes the king of Lydia on the march.33 In Chapter 8 it will be argued that this may have been the ancestor of the ‘unequal pipes’ of Roman music, introduced via the Etruscans who, according to the ancient tradition, had Lydian connections.
(TO BE CONTINUED)
BY John G.Landels
NOTES AND READING LIST
There is a vast literature on Greek tragedy and comedy; some of the more recent works are marred by various obsessions (notably structuralism and feminism). Two will suffice here, the first being concerned with the theatrical, staging and acting aspects of Greek tragedy: Taplin, Oliver, Greek Tragedy in Action (Methuen, London 1978) and on comedy, and Dover, Kenneth, Aristophanic Comedy (London 1972).
1 E.g. Berlin black-figure amphora, cat. 1686, and on the Parthenon north frieze. The slab showing the four aulos-players and two of the kithara-players was lost in the explosion of 1687, but the numbers are known from Carrey’s drawings; they are referred to as Slab VII figures 20–25. The lower end of the front aulosplayer’s pipes, and parts of his hands, overlap on to Slab VI (which is in Athens) and badly-damaged parts of the last two kithara-players are on Slab VIII.
2 The Athenian calendar was not properly adapted to the solar year, and required a correction every so often, so that the same date by their reckoning could be some days behind or ahead of ours.
3 Athenaeus, Deipnosophistae 617b—e; Barker I pp. 273–274.
4 E.g. the Thargelia, a crop-fertility rite in honour of Apollo, and festivals in honour of the craftsmen’s deities, Prometheus and Hephaestos.
5 Pausanias X, 7.
6 Pythian 1, 1–4.
7 Isthmian 7, 3–4, fr. 61, 6–8. 8 Olympian 6, 87–
8. There has been some argument recently about the mode of performance; it has been suggested that there was a solo singer and a dancing choros. The following articles are relevant: Lefkowitz, Mary R., ‘Who sang Pindar’s victory odes?’ A.J.Phil. 109 (1988):1–11; Heath, Malcolm, ‘Receiving the komos—the context and performance of epinician’, A.J.Phil. 109 (1988): 180– 195; Carey, Christopher, ‘The victory ode in performance; the case for the chorus’, Classical Philology 86 (1991):192–200.
9 E.g. Pythian 2, 67–8.
10 Symp. 176E.
11 In the Greek text lines 1122–1264; in the Penguin Classics translation by David Barrett, pp. 80–85.
12 Lines 1351–76. The song by Simonides which the son is asked to sing may have been a ‘victory ode’, but this is no kind of evidence for the normal performance of such a piece.
13 E.g. in Euripides’ Phoenissae 791, the Argive army is (paradoxically) called ‘a band of revellers without an aulos’ and in Sophocles’ Oedipus at Colonus (line 1223) Death is described as ‘without lyre or choros’.
14 E.g. on the François Vase: see John Boardman, Athenian Black-Figure Vases (Thames and Hudson, London 1974) pl. 46 (the neck frieze).
15 Bacchylides fr. 13; when he speaks of its ‘song’ he is being ironical!
16 Athenaeus 12, 535d.
17 Boardman lists it under the Brygos painter (Athenian Red-Figure Vases—the Archaic Period, Thames and Hudson, London 1975) but says on p. 136 ‘a very late work, if his at all’.
18 See D.L.Page, Alkman, the Partheneion (O.U.P., Oxford 1951), esp. pp. 44– 69.
19 See Claude Calame, Les Choeurs de jeunes filles en Grèce archaique (Rome 1977).
20 D.A.Campbell, JHS 84 (1964): pp. 63ff.
21 For example, Euripides’ Andromache 103–116.
22 Kephisophon, in Frogs 1451, and elsewhere.
23 Poetics 1449a 10–11.
24 Oxyrhynchus Papyrus 2256, fr. 3.1b. For sensible accounts of the early stages of drama, see D.W.Lucas, The Greek Tragic Poets (2nd edn), Cohen and West, London 1959 or H.D.F.Kitto, Greek Tragedy (3rd edn), Barnes and Noble inc., New York 1961.
25 Poetics 1449a 18.
26 This is a very common problem in modern productions even when the musical accompaniment is tape-recorded and the volume can be controlled.
27 Frogs 939–944.
28 Xenophon (Symposion 6, 3) speaks of an actor reciting lines in this rhythm to an aulos accompaniment; but he may be talking about something unusual.
2(a) THE AULOS
Kathleen Schlesinger’s massive work, The Greek Aulos (Methuen, London 1939) is not recommended. It contains a vast amount of information (some of it now out of date) on woodwind instruments, but her study of the Greek aulos is vitiated by faulty method. Baines, Anthony, Woodwind Instruments and their History (W.W.Norton, New York 1957) is old-fashioned but very readable and informative.
1 They included Sir John Beazley. He described a vase-painting of an unclothed male aulos-player as ‘nude youth fluting’. I have used this memorable phrase as the caption for Figure 2c.2, where it is entirely accurate.
2 The practice seemed to be dying out in the 1980s, but it was saddening to find it back again in Looking at Greek Vases, ed. T.Rasmussen and N.J. Spivey, C.U.P., 1991; it makes a nonsense of Mary Beard’s interpretation of the symbol on p. 29 of Looking at Greek Vases—the lady is not a crypto-shepherdess!
3 In the National Museum, Athens.
4 In the Heraklion Museum, cat. 396. Illustrated in Martin Robertson, Greek Painting (Macmillan, London 1978), pp. 26–30.
5 Bk 10, 13 and Bk 18, 495.
6 In Hero’s description of the organ (see Appendix 2, Figure Appendix 2.2) the socket into which the key-block slid was called glottokomeion.
7 K.Schlesinger, in her formidable work The Greek Aulos (Methuen, London 1939, reprinted Groningen 1970) argued that they were replaced by a single reed in the late fifth century, and some other authors who accepted her modal theories (notably N.B.Bodley) have supported this view; but it is not justified by the evidence.
8 The main reed-beds in Southern France were almost destroyed in the Second World War, but have re-grown, and other sources have been found in the U.S.A. and Australia.
9 The Greek word in the MS which I have interpreted as ‘the early technique’ is not intelligible; some editors think it means ‘for purposes of accompaniment’ (see note 10).
10 I am sure that Barker (I, p. 188 n. 9) is right in interpreting kataspasmata in this way. The verb ischein need not, as Schlesinger thought, mean ‘curb’ —it can just mean ‘have’.
11 For a detailed discussion, see Barker I, pp. 186–189.
12 Aristoxenos II, 42 (Barker II, p. 158); Plutarch, Moralia 948b and 1096a.
13 801b, Barker II, p. 103 and 804a, ibid. p. 108.
14 Porphyrios’ commentary on Ptolemy, p. 20 1–2 During
15 Historia Animalium (Zoological Researches) 565a23.
16 The top end of the Reading aulos has a flare which could have served this purpose, though it is rather small.
17 See my article, ‘A newly-discovered aulos’, in the Annual of the British School at Athens 63 (1968):231–238.
18 See my article ‘Fragments of auloi found in the Athenian Agora’ in Hesperia 33:392–400.
19 The Louvre auloi show signs of shaping, particularly around the thumbholes; this suggests that they did not have keywork, and must therefore have been ‘single-pipes’ (monauloi).
20 Pausanias 9, 12, 5 and Athenaeus 631e; see also ibid., 184d (Barker I, p. 271).
21 The best surviving example of this mechanism is described and illustrated in N.B.Bodley, ‘The auloi of Meroe’, in Amer. Journal of Archaeology 50 (1946):217–239; it appears in Plates VII and VIII, numbered fragment 14.
22 Simonides fragment 56 (Poetae Melici Graeci 947); Plato, Republic 399C.
23 Bk 1, ch. 20–1, Barker II, p. 139–140.
24 Plutarch, Moralia 1096a, in an essay entitled ‘It is not possible to live happily by Epicurus’ precepts’.
25 Harvard Studies in Classical Philology IV (1893):1–60.
26 See the article quoted in note 18, Figure 1 A, p. 393, and fragment A in Plate 70.
27 1138a, Barker I, p. 226.
28 Plutarch (?), Moralia 853E, in an essay called ‘A comparison of Aristophanes and Menander’.
29 The view was first set out by Andrew Barker in a collection of essays (in Italian) —La Musica in Grecia, ed. B.Gentili and R.Pretagostini (Laterza, Rome 1988), pp. 96–107. It had been arrived at independently by West (pp. 72–73).
30 I am not able to agree with Annie Bélis in her opinion that the Louvre auloi are a pair (I have argued in note 19 that they are ‘single-pipes’). The Elgin auloi may have been a pair, but they have become so badly distorted that measurements are unreliable.
31 Black-figure hydria, British Museum cat. B 300.
32 There is a much-discussed passage in Plato’s Laws (812 d-e) in which he speaks with disapproval of lyre accompaniments which do not follow exactly the melody and rhythm of the words; but this is probably a matter of ornamentation, not true polyphony.
33 Histories I, 17.