Greek and Roman Musical Studies 10 (2022) 15–77
brill.com/grms
Two Auloi from Megara
Chrḗstos Terzḗs | ORCID: 0000-0001-7988-1389
Austrian Archaeological Institute, Vienna, Austria
christos.terzis@oeaw.ac.at
Stefan Hagel | ORCID: 0000-0001-5654-4013
Austrian Academy of Sciences, Vienna, Austria;
University of Vienna, Vienna, Austria
stefan.hagel@oeaw.ac.at; stefan.hagel@univie.ac.at
Abstract
Two aulos pairs (Δ1965 and Δ1964), unearthed in 1980 and 2005 respectively at Megara
(Attica), are exhibited in the city’s Archaeological Museum. Both are associated with
metal sliding keys (resembling the keys on the well-known Pergamon aulos model),
either mounted on the pipes (Δ1964) or displayed next to them (Δ1965). The present
paper describes the parts, proposes a meaningful re-assemblage of the bone sections,
and gives a detailed account of the sliding mechanism, which is here for the first time
attested on finds of actual musical instruments, including an entirely new technology of speaker hole keys. A musical analysis, based on determining plausible reed
configurations using software modelling, suggests that the finds represent a partially
standardised design of professional modulating auloi, playing attested harmoníai
while hovering between the enharmonic and chromatic.
Keywords
ancient musical instruments – Greek aulos – aulos syrinx – sliding keys – Pergamon
model – ancient musical notation – Aristoxenus
© Chrḗstos Terzḗs and Stefan Hagel, 2022 | doi:10.1163/22129758-bja10040
This is an open access article distributed under the terms of the CC BY 4.0 license.
16
1
Terzḗs and Hagel
Introduction
After a salvage excavation carried out in 1980 at the private property located
at 6 Cheimáras Street (Megara, Attica), the archaeologist on duty, Pantelḗs
Zorídēs, proclaimed the discovery of a certain number of burial gifts found in
Grave xxxvii, which held a sarcophagus sealed with three covering plaques.1
Besides the skeleton of the deceased, the components of an aulos made out of
bone and its set of keys were found, safely preserved in the stone-made coffin.
The burial finds also comprise three female clay figurines and an unspecified
number of broken pieces formed from molten alabaster. Zorídēs dated the
burial to the earlier Hellenistic period and announced a complete publication.
Subsequently, the fate of the Megara aulos remained obscure for more than
three decades. In November 2015, Chrḗstos Terzḗs visited the Archaeological
Museum of Megara for an autopsy of the exhibited instrument, and then
applied for access to the relevant entries of the excavation inventory, hoping to
be granted permission to examine and publish this important find. The aulos
pair was displayed in a suitably tailored showcase under the inventory number
Δ1965αβ, accompanied by nine bronze keys lying next to the bone sections, in
four pairs of roughly similar items plus one. Surprisingly, another aulos pair,
in many respects a sibling, was also exhibited there, under the inventory number Δ1964αβ, but with its bronze keys situated in what was presumed to be
their original positions along the sidewalls of the pipes. In the following, we
will refer to the older find, Δ1965, as ‘Meg1’, and to the younger find, Δ1964, as
‘Meg2’. The Museum’s sigla α and β refer to the original reassemblage of supposed pipes, even though we will need to modify it in the course of our study.
The second pair had been unearthed in 2005 in the course of a salvage excavation of fifteen graves in the northern part of the ancient city cemetery, located
along today’s Thēvṓn street. The segments were recovered and re-assembled,
together with the bronze keys, in 2012–13 by Geōrgίa Theodṓrou, conservator
of the ephorate. On July 13th 2016, at the 9th international MOISA conference,
held in Athens, the archaeological context of the Megara Museum auloi was
presented by Panagiṓta Avgerinoú, while Geōrgía Theodṓrou described the
restoration of Meg2; moreover, a provisional re-assemblage and acoustical
demonstration of the pair were presented by the present authors.
Avgerinoú (forthcoming) dated both burials by the typology of the other
grave goods, inferring that the individuals would probably have flourished
in the first quarter of the 3rd century BC. She accepted the results of an
osteological study carried out by Ylva Bäckström, suggesting that Meg1 had
1 Zorides 1980, 45.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
17
accompanied a young female adult aged 20–25 years, whilst Meg2 belonged
to an older man, probably aged 50–55 years. Regarding the excavation conditions of the latter, she admits that unconventional methods had to be adopted,
given the inclination of the field and the narrowness of the grave, which forced
workers to crawl within and pull out clay deposits using their arms as shovels,
thus revealing ancient material. Of the other pair, apart from a b/w image of
the offerings in situ, the archaeological registry, diaries and restoration data are
officially lost. Hence, unravelling the riddle of the Megara auloi amounts to a
risky venture, not only because of the absence of crucial metadata but, more
importantly, due to the undocumented modern intervention on the archaeological material in the course of restoration and assemblage, which, although
carried out with diligence and ingenuity, lacked indispensable archaeomusicological support. However, as we appear to be dealing with sophisticated
instruments built for professional musicians, these have hardly exposed even
small errors of design or construction. Consequently, by construing and disentangling the intricate design of the Megara musical finds – intricate, that is, in
comparison with the known earlier instruments which comprised merely four
sections with no mechanism – valuable archaeomusicological insights may
emerge from these so far unique instruments, especially if a feasible reconstruction can be proposed.
2
Other Evidence for Slider Keys
The evolution of ancient Greek music, particularly towards the end of the Classical period, doubtless brought about organological innovations. The invention
of a modulating aulos, ascribed to the renowned Theban aulete Pronomus
(late 5th century BC), was said to have enabled him to play Dorian, Phrygian
and Lydian tunes on a single doublepipe.2 Unquestionably, such an expansion of the instrument’s potential implies technical modifications regarding
the production of alternative pitch sets. The archaeological record attests to
two distinct kinds of mechanisms.3 One consists of rotating sleeves, which are
mostly known from Roman-period examples,4 where two or three metal layers
covered the core of the pipe. Rotating against each other, these sleeves enabled
2 Cf. Ath. 14.631e, Paus. 9.12.5.
3 On auloi with sleeves, see Howard 1893, 7f.; for an extensive list of the published finds carrying sleeves and sliders, Hagel (2009, 337 n. 28) and Sutkowska (2015), who also offers a brief
description of both types of aulos mechanisms.
4 For a detailed account of the rotating mechanisms of some auloi from Pompeii, dating from
the Roman Imperial Period, see Hagel 2008, 53–5.
Greek and Roman Musical Studies 10 (2022) 15–77
18
Terzḗs and Hagel
selective covering and uncovering of one or more side holes, making different
sets of holes accessible to the fingers at one time. Thus, aulos players became
able to adjust both the register and the key (tónos) during their performances.
A less sophisticated and probably older type of mechanism served an altogether different task: operating holes that were out of the reach of the player’s
hands. Sliders, attached to the surface of the instruments, were designed for
motion along their longitudinal axis, extending from the bass region up to the
playing position of the hand. In the case of the Megara auloi, each pipe was
equipped with three of them, accessing three distinct bass notes. Each ended
in a plate, skillfully forged in a curved shape so that it fitted perfectly over
the cylindrical bone surface. Rods of various lengths connected these plates
to knobs that were evidently situated near the lowest two fingerholes of the
pipe. By pulling down or drawing back a knob, the respective sliding plate was
remotely set into functional motion.
This kind of slider mechanism was originally discovered and described in the
early twentieth century, on the basis of the cast of an aulos, or rather the lower
half of one of its pipes, most probably once part of a votive statue. The fragment
was found near the northwest gates of the Eumenic walls of Pergamon and
seems to have been lost during the Second World War. According to Alexander
Conze, who offered a comprehensive description, a detailed drawing and convincing functional interpretation of the sliders, the artefact must have been
produced before the end of the Attalid Kingdom of Pergamon in 133 BC.5
The relative positions of the pipe’s metal keys can be specified as follows:
while the two longer keys (numbered 1 and 3 in Conze) are mounted at opposite sides, the shortest key (2) sits right in between them, at their left side, if
viewed from the blowing end. The long slider close to the upper side of the
instrument, in turn, sits about 45 degrees left to the extant real fingerholes at
the upper end of the fragment. This suggests that it formed part of the righthand pipe of the pair, assuming that the mechanism was operated by the
playing hand. Conze correctly states that 1 and 3 shared the same fixing band
(x); these bands, which appear wrapped around the pipe sidewalls, would have
prevented any deviation from the slider’s predefined path along the pipe surface. The plates of the long sliders 1 and 3 operated the two downmost side
holes; slider 3, the longest, is captured at its upper position, whereas the plate
of slider 1 is shown covering its respective hole. Towards their upstream end,
both pass over the fixing band (y) of key 2. The knobs at the upper ends of the
5 Conze 1903, 7f.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
19
keys are placed in such a way as to allow one or the other of the player’s fingers
to switch their states from closed to open and back.
Conze also pointed out that the plate which we would expect at the downstream end of key 2 is missing, though there is also no trace of the respective
hole that it would have covered. Note also the (incomplete) representation of
a wrapping (z) beneath the sliders, which doubtless depicts the typical reinforcement of the socket-spigot junction that we frequently find on bone auloi.
Conze did not comment on the utility of the oblong slots close to the
upstream end of the metal keys, right beneath the knobs. These obviously
formed the primary means of restraining the longitudinal motion of the
sliders. However, such a slot would have been useless without a pin or nail
that went through it and was fastened in the sidewalls, thus at the same time
keeping the key movement parallel to the longitudinal axis of the pipe and
establishing its boundaries. Without such a limiter, the forces exerted from the
player’s finger would inescapably increase torque levels at the upper end of
the band below, which might easily cause the rod to block, and in consequence
bend or even damage its trail. Assuming the original pipe had been equipped
with pins, those would have been fastened in the pipe sidewalls through the
lower end of the slot of key 3, thus allowing remote covering of the respective
side hole. On the other hand, the pins interacting with keys 1 and 2 should have
passed through the upper ends of the respective slots, so that the holes would
be uncovered by pulling the knobs back.
However, while the presence of such limiters in the original pipe appears
beyond doubt, the precise technical implementation of the two bands (x, y)
remains to be explained: had they been made as a single bronze piece, it would
have been next to impossible to assemble them around the rods; on the other
hand, simple wrappings of leather or thread would necessarily be too tight to
allow for unrestrained movement of the rod beneath them.
Conze knew very well that efforts at a musical interpretation of the artefact
are hampered by the incomplete evidence in combination with our limited
knowledge about ancient music in general. While unmistakably describing the Pergamon artefact as a physical model of an actual instrument, part
of a votive offering, he implicitly allowed for inadvertent deviation from the
unknown original on the part of the artist when disregarding inconsistencies
arising from the operation of the sliders, notwithstanding the seductive precision of the elaborate artefact. Virtually a century later Maurice Byrne glossed
over the fact that the object is solid (i.e. it does not have the internal cavity that
defines any wind instrument) as well as the missing plate of slider 2 together
with its corresponding side hole, and wanted to promote Conze’s model to a
Greek and Roman Musical Studies 10 (2022) 15–77
20
Terzḗs and Hagel
part of an actual instrument.6 He calculated the presumed length of its missing part, accommodated further tone-holes, suggested a playing technique and
thus interpreted it as an actual representative of a newly-introduced type of
instrument, the “extended drone aulos”. Certainly, a unique surviving piece
of evidence for pipes with sliding keys justified the publication of novel hypotheses concerning its living past; however, the Pergamon find was definitely no
instrument, and a perfect level of faithfulness regarding its proportions cannot
be presumed without further argument.7
In any case, few finds of actual slider mechanisms have so far come to light.
Byrnes’ list comprised the aulos fragments from Meroë at the Boston Museum
of Fine Arts,8 those from the Oxus Temple exhibited at two Archaeological
Museums in Dushanbe,9 an isolated slider at the Berlin Pergamon Museum
(Inv. nr. 30259) and the aulos Meg1. Apart from the second Megara aulos pair
(Meg2), two separate sets of bronze sliders, each comprising three pairs of similar length (long/middle/short), are exhibited in the Archaeological Museum
of Lefkada. Both were unearthed from graves in the southern cemetery of the
ancient town, located at the northern end of the modern village Karyṓtes.
The better-preserved group, originally misidentified as a set of “goldsmiths’
tools”,10 was dated approximately to the end of the 4th century BC, whilst an
excessively corroded and still unpublished set (Inv. nr. 1758α-στ) was retrieved
from a sarcophagus dated to the late classical era. Although the grave comprised
disturbed remains of consecutive burials, the first quarter of the 3rd century BC,
after which the cemetery was no longer used, constitutes the terminus ante
quem for both groups of Lefkadas’ sliders.
On balance, the technology of remote-controlled wind-instrument side
holes outside the reach of the player’s hand is already attested from the second
half of the 4th century BC at the latest. Interestingly, the available evidence
comes from a broad cultural environment, extending from (Late classical)
Western Greece and the Meroitic Empire (in the early Roman Imperial age)
to the easternmost borders of the (Hellenistic) Middle East, though, while the
various instances are almost certainly connected by a common technological
history, the associated musical practices need not always have been similar.
6
7
8
9
10
Byrne 2002, 367–9.
For a positive assessment, cf. Hagel 2009, 347–9.
Bodley 1946. For a detailed account of the conservation and technical examination of the
Meroë fragments, see Gänsicke and Hagel 2017.
Litvinsky 1999; Litvinsky 2010.
Kostoglou 1970, 331.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
3
21
Technical Description and Reconstruction
3.1
Original Restoration by the Museum
As is universal among published bone auloi, the instruments from Megara
are assembled from multiple bone sections using socket-spigot junctions
(Plate 1). The Museum currently presents them on either side of a bronze sistrum (1st BC–1st AD) devoted to the Egyptian deity Isis, whose image appears
on the handle. The pair of Meg1 is situated to the left, while on the other side,
Meg2, with its slider keys mounted along its sidewalls, is easily identified.
Below the sistrum, a bronze ring and nine sliding keys purportedly associated
with the left aulos pair are arranged as follows: three lengthwise sorted pairs
(i.e. of long, middle and short length) are accompanied by a single key sibling
to the middle-sized pair, while a slightly different group of two keys each consisting of two parts connected by a ring complete the series. At the bottom
corners of the case, two clusters of six and three broken bone fragments have
respectively been placed. Conservators were unable to join them either among
themselves or to the assembled pipes. Apparently they belonged to aulos pipes
that did not survive.11 Auloi and keys are lying on a slanted surface, facilitating
inspection by visitors. The fact that the pipes are all aligned at their bottom
ends betrays an intended symmetrical presentation of the finds on behalf of
the designer of the exhibit. This, however, precludes perceiving the musically
important arrangement of corresponding tone holes that would yield (near-)
identical pitches, which would require an alignment at the upper ends.
At the left side of the showcase, aulos Meg1 stands with its higher pipe comprising the typical bulbous uppermost section ending in a socket for the reed,
an extension part followed by two sections with five fingerholes in total, and,
finally, the bottom section with the three side holes that were controlled by
slider-keys. Next to it, the lower pipe of the pair is presented. Here the bulb and
the cone with the reed socket are made as two separate parts; the long extension part is supplemented by a short extra section, followed by the long middle
part with four fingerholes, a notably shorter one with only a single side hole,
and finally the bottommost segment of the pipe with two side holes that were
equipped with slider keys. The ends of both pipes appear slightly flared, retaining the physical expansion of bones towards the junctions. Hammered bronze
11
The external diameter of some of these fragments matches the internal of others, suggesting layers of bone that could rotate within each other. Metal handles show that for two of
them it must have been the external part that moved, allowing the closing and opening
of tone holes in a perfectly adjusted inner tube. They resemble fragments from Taranto
and are better discussed together with these in a separate paper.
Greek and Roman Musical Studies 10 (2022) 15–77
22
Terzḗs and Hagel
rings, about 6.5 mm in width, are nailed around the exits, while slightly narrower rings embrace the main tubes close to their upper ends. Surprisingly,
what we would regard as the ‘low’ pipe, based on a comparison of treble
notes – the pipe whose highest fingerhole is further removed from the mouthpiece – appears to represent the overall shorter pipe in the pair. In fact, the
designation of ‘low pipe’ here appears to lose its sense, since its overall ambitus
would have been encompassed within the pitch range of the allegedly ‘higher’
pipe. Though not a priori an impossible design,12 this apparent inconsistency
alerts us to a potential problem.
At the right side of the showcase, the higher component of aulos Meg2
appears to the left, comprising six bone sections ordered similarly: bulb with
reed socket, extension, one part with four fingerholes, and finally a second
extension followed by two separate bone parts sporting one and two slider
holes respectively. In the lower pipe next to it, the arrangement of its five bone
sections and the distribution of its tone-holes attest to a component of similar
sort, with four plus one fingerholes accommodated across two adjacent parts
and three further tone holes controlled by sliders located on a singular longer
downmost section. Both pipes are resting backwards, with their thumb-hole in
direct sight, highlighting the sets of slider keys that are mounted on the tubes,
whose function can thus be easily gleaned by visitors. Unexpectedly, the side
hole that is supposed to be covered by the little finger is situated at the underside
of the pipe between two adjacent slider keys, in a position where the player cannot possibly cover it.
Despite the confusing assemblage of both auloi, a striking resemblance of
the lengthwise sorted key pairs either exhibited separately (Meg1) or mounted
on the instrument (Meg2) with the orphans unearthed in Lefkada and those
observed on the Pergamon model leaps to the eye in respect to their relative
sizes and the nature and distribution of their constituting components (knobs,
slots, rods and plates). Moreover, the distribution of the keys on Meg2 corroborates the design of the Pergamon model: in both cases, sliders were responsible
for remotely operating the three bass tone-holes of the instrument.
3.2
Detailed Description of the Bone Segments
Both pairs include the typical bulbous mouth-end pieces (M; cf. Figure 1), the
uppermost parts, which, however, did not actually touch the mouth of the player
(Plate 2; Plate 3; Plate 6; Plate 7). These typically display an external shape of a
truncated cone followed by a bulb, which tapers down to a neck, about as long
as the bulb, ending in a spigot, which goes into the next component of the pipe.
12
It is found on the wooden Berlin aulos, which, however, belongs to a later date and a different type; cf. Hagel 2010.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
figure 1
23
Abbreviations for bone sections and finger holes
Internally, its cylindrical longitudinal bore widens slightly but conspicuously at
its upper end, forming a concentric socket which would hold the reed. Material
loss of limited (Meg2αM, βM, Meg1αM) or more considerable (Meg1βM)
extent, observed at the transition point from bulb to the neck, hardly reflects
traces of intentional drilling; this kind of damage may be expected given the
sensitivity of the thin-shelled narrow neck. Meg1βM comprises two distinct
parts joined by spigot and socket: the conical reed socket and its bulb + neck
complement. An oblong region of the bulbous sidewalls extending from its
upper end towards the point of its maximal diameter is missing. Similarly, a
little semi-circular part is missing from the tip of the spigot of the conical reed
socket, too. By conjoining both parts and rotating the cone around its axis until
material losses are getting aligned, an elliptic hole (8.5 × 6 mm) emerges, but
its shape does not in any way suggest an original hole. None of the upper
rims of the three M parts have survived; however, the overall length of their
upper conical part can satisfactorily be inferred from the preserved dimensions of Meg1βM.
The mouth-end pieces connect to cylindrical extensions of varying dimensions, without any side holes (X). Their relative length would typically make
a pipe either the high or the low component of the pair. Here these extensions preserve their rims at both ends, providing their overall as well as their
effective dimensions (without the spigots). The external surface of their upper
end exhibits a very slight recess, 8–9 mm wide and only 0.15 mm deep. Such
recesses evidently held some kind of enforcement; sometimes rings of copper
alloy are still found in place. However, since the present items show no signs of
the green coloration that is so typical for bone parts that were in contact with
copper, we must rather reckon with windings of thread as would easily serve
the same purpose of keeping the thin bone socket from exploding. Both X sections of Meg1, however, are equipped with a bronze ring in an unexpected
position, about 5 mm below the shallow recess. This ring is 5.8 mm wide and
nailed to the bone surface.13 Notably, Meg1βX comprises two sections: the
upper long one displays extensive damage, resulting in an irregular oblong gap.
13
At the time of the second visit to the Museum in May 2017, in order to inspect Meg1, the
ring, which had formerly been placed around the upper end of αFi, had been moved to its
correct position, indicated by conspicuous traces on the tube surface.
Greek and Roman Musical Studies 10 (2022) 15–77
24
Terzḗs and Hagel
The following shorter part holds a small hole (3 × 4 mm) in the upper end of
its spigot, reflecting an even smaller, perhaps deliberately drilled original hole,
which has probably expanded in the course of ongoing material disintegration.
Two adjacent sections of conspicuously different length (Fi, Fii), with side
holes controlled by the player’s hands, follow segment X at Meg2αβ and Meg1α.
The little-finger hole lies on the lower, shorter part (Fii), while the upper, longer part (Fi) hosts the index, thumb, middle and ring finger holes. Shallow
recesses 8 mm in width around their upper ends resemble those found on the
X sections and suggest that some sort of bands also reinforced the X-Fi and
Fi-Fii joins. On Meg1β, section Fii, which held the little-finger hole, is missing. The external rims of all fingerholes are rounded, ensuring a smooth fit
of the player’s fingers. The upper side holes are not perfectly aligned, but lie
at slightly different azimuths in the range of about ±30 degrees in Meg1, and
±35 degrees in Meg2; with the thumb hole having been placed opposite to the
middle-finger hole, effective handling is ensured and the assignment of each
pipe to its proper hand is straightforward.
A feature of utmost importance for questions of reconstruction is the presence of two pin-holes close to the downstream end of Fi, though at different
distances from it. The one closer to the lower rim of the bone section is aligned
with the centre point of the thumbhole, relative to which the other is rotated
by about 90 degrees towards the side that faces the other pipe of the pair when
played. A similar pin-hole lies on the spigot of each surviving Fii bone sections, lined up with the centre of its side hole. Interestingly, the pipes of Meg2
preserve five out of originally six elegant solid bone pins with semi-spherical
heads within their respective pin-holes. A single similar small side hole is
located close to the upper end of Meg1αFi, aligned with the thumb-hole centre, preserving a headless pin body within it. A corresponding hole may have
existed on the other pipe of the pair, but here, at the upper part of Meg1fiB,
we encounter a substantial loss of material; during conservation the extensive
gaps were filled with plaster. However, the undersides of both Meg2Fi sections
are preserved intact towards their upper ends, showing that no pin was located
in this position.
The sections Fi and Fii have lost different amounts of material, including
parts of their spigots, due to disintegration and subsequent crumbling of the
bone. However, thanks to at least partially preserved upper and lower rims,
we are able to restore their effective dimensions. Perhaps even the action of
the sliding keys during performance, in combination with the presence of corroding metal, had already caused some fatigue of the bone surface beneath
the rods and knobs. The unfriendly environment of the grave may have transformed such initial lesions into larger and smaller gaps, none of which necessarily
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
25
reflects original deliberate structural modifications of the tube surface. Extensive
material loss has now been filled in with plaster, including the expected positions
of the ring-finger and small-finger holes on Meg2βFi and Meg2αFii respectively,
holes whose presence was not anticipated during conservation.
Below the F sections, the end of each pipe, with its three slider side holes,
is formed either by a single long section (Meg1αS, Meg2βS) or a couple of
two successive shorter parts (Meg1βSi + ii, Meg2αSi + ii). We cannot detect
recesses at their upper ends. Surprisingly, Meg2αSi does not feature a spigot
at its downstream end like all other parts, but bears two sockets of dissimilar
depth at both ends, into which the spigots of the preceding and subsequent
segments fit perfectly. Unlike the fingerholes, all slider holes maintain a sharp
edge. They are allocated to the side walls of the single long sections Meg1αS
and Meg2βS as follows: in respect to the upper hole, the intermediate one is
located 90 degrees clockwise, while the lowest hole, close to the exit, stands
opposite to the latter, and therefore 90 degrees counterclockwise from the former. Within the two shorter corresponding sections on the other pipes, the
lower one (Sii), which terminates the pipe, features two side holes at different distances from the pipe exit, lying at opposite sides of the tube, while the
preceding section (Si) features a single hole. Both Meg2αSi and βS preserve a
pin-hole close to their upper rims, which fit perfectly with those on the Fii spigots, thereby determining the rotational alignment between sections Fii and
S. Only Meg2βS features another small hole close to its end, at the rim of the
bronze collar that encloses all pipe ends.14
Meg2αSi + ii are preserved in excellent condition. Overall, all sections with
slider-operated holes maintain their upper and lower rims, so we are able to
obtain their effective lengths, which are so crucial for our research. However,
significant loss of material, now substituted with plaster, occurs especially in
regions over which a slider key must once have moved. Interestingly, thorough inspection reveals a few pairs of deliberately engraved spots or short
and shallow transversal niches. Sporadically these are accompanied by barely
detectable lines, presumably defining certain distances or points of particular
14
The purpose of this hole is doubtful. It might have held a pin, but not of any discernable
purpose. It is too small to serve as a tuning hole, as it would raise the bass note only by
3 cents while not effecting any other pitches. Had it been drilled to regulate the sound of
the bass note, we would expect similar holes on all pipes. Lying at the outermost usable
part of the bone, which was usually discarded except when becoming an exit section, it
might have been drilled for pinning the piece to the lathe at an early stage of production.
The hole would then have been filled, for instance with a mixture of glue and bone meal,
like the grooves in the bone under the lowest slider plates needed to be filled in order to
make these functional.
Greek and Roman Musical Studies 10 (2022) 15–77
26
Terzḗs and Hagel
interest for the aulos makers. Their significance will be revealed in the course
of the following detailed account of the construction and function of the
slider keys.
3.3
Slider Keys: A Descriptive Analysis
A typical slider key of the Megara type comprises four distinct structural parts
in the following order: knob, slot, long rod, and plate. A knob consists of two
concentric, closely spaced, half-lenticular compact shapes of different diameter, standing in upright position, with the larger one forming the upper end
of the key and both ending in flat bottoms. Its shape enabled the aulētḗs (or
aulētrís, as regards Meg1) to control the slider-key with the fingertips. The slots
provide a linear path within which the key is allowed to travel, ensuring its
moving direction along the longitudinal axis of the pipe. The pin-holes that
are found drilled on all the tubes and the surviving pins on both pipes of Meg2
suggest that the latter were firmly secured in their holes through the key slots.
Consequently, the slot function was twofold: to direct the slider movement
along the right path and to confine the motion of the plate within a limited
range. The knobs are attached to the slots by means of a short rod of rectangular cross-section.
The longer intermediate part of the key is formed by a flat-based rod of initially rectangular cross section which gradually transforms into a triangular
cross-section. Interestingly, at some point within its lower half, the rod, in a
sharp step, narrows down markedly, to a mere 1 mm in both height and width,
until it either joins the plate or returns to its primary dimensions before reaching the plate. Thus, a precisely confined region is generated, bounded by the
two points that define the transition from the initial rod cross-section dimensions to the smaller ones and the back. These points are further marked by a
bump at either side on top of the rods, terminating their wider parts towards
the constriction. Within the narrow region, a separate roof-shaped guide is
accommodated, which terminates in sharp protruding edges at its four corners,
calling to mind the pairs of engraved spots or short and shallow transversal
niches that we have observed on the S sections. Apparently the roof-shaped
guides were firmly pressed onto the tube, so that their sharp edges created those
niches, which fit their dimensions accurately. Within the roof-shaped guide,
the triangular portion of the rod can travel freely. At the same time, its movement is once more restricted within precise limits defined by the respective
widths of the guide and the narrow section of the rod. The finds do not tell us
how the guides were so firmly attached to the tubes. However, with the roofshaped guides the Megara finds provide the evidence to substantiate Conze’s
bands covering both rods and tubes on the Pergamon artefact. Whether these
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
27
were made of leather, hide or tight windings of thread, the ‘band’ would likely
obscure the guide underneath from view to such an extent that the maker of
the Pergamon item would have been content to show only the former. At any
rate, the material must have been very thin, because it could pass below other
rods, as can also be observed on Conze’s object. At such places, slight recesses
are worked into the underside of the rods or slot parts, obviously in order to
prevent any friction between the band material and the metal.
The plate, finally, is carefully adjusted to the cylindrical curvature of the
tube, so that, in the closed position, it can form an air-tight seal. One might
surmise that this was additionally ensured by some soft, organic material fixed
to the underside of the plate, as is generally the case on modern woodwind; at
any rate, inspection by the naked eye could not corroborate any such traces.
The width of the plates approximately equals a quarter of the tube circumference, while their length varies between 13–17 mm.
We have thus gained a comprehensive picture of the functional design of
the sliders: the motion of the keys was not only efficiently restrained to a linear
path along the pipe but was also confined between precisely determined limits
that were defined by the length of the slot, on the one hand, and the extent of
the narrow part beneath the guide, on the other. Clearly, both these means
of limitation would optimally prescribe precisely the same motion range with
identical extreme positions. Such a simultaneous constraint of key movement
at two different places, one near the upper, one close to the lower end of the
rod, creates two significant technical advantages. Firstly, being fixed in its longitudinal position at two places, the rod cannot act as a lever when undesired
lateral forces are applied during operation, preventing wear and damage that
would be associated with the high torques exerted on a single guide. Secondly,
the pushing and pulling forces exerted by the finger are split between the pin
in the slot on the one hand and the ends of the guide on the other, preventing damage of the more delicate of these components. The minimal distance
by which a slider would need to move in order to cover and uncover its side
hole completely, is obviously equal to the diameter of the latter; the maximal
distance is determined by the length of the plate. On the reasonable assumption that optimal sealing is achieved when the centre of the plate is aligned with
the centre of the hole, the optimal movement range, given a mean slider hole
diameter of 9 mm, varies between 12 and 14 mm, depending on the size of the
respective plate. As a consequence, the length of the slot as well as the size of
the plate form a solid basis for assessing the original length of the guide, where
it has perished, or the length of the narrow rod part below it, where corrosion
has agglomerated guide and rod in a way that obscures one of the ends of the
narrow section.
Greek and Roman Musical Studies 10 (2022) 15–77
28
Terzḗs and Hagel
As has become clear, each pipe of the instrument Meg2 is equipped with
three slider-keys, which, despite variation in detail, come in three general sizes,
which we will address as long, intermediate, and short (Plate 8; Plate 9). Since
no more than three pin holes per pipe are detected on this particular instrument, we have good reason to assume that three similar pairs of sliders, one
in each length category, constitute a complete set of keys. This assumption is
corroborated by the similarly arranged three sliders found on the Pergamon
model, and the two identical sets of keys at the Museum of Lefkada, which
were doubtless retrieved from independent burials. On the other hand, Meg1
was not only found together with such a set of six, but with an additional key
of the intermediate kind, and, moreover, with the specially shaped pair of keys
that include a middle ring, which end in a significantly smaller plate than all
the others, suggesting that these were dedicated to plugging a sýrinx (speaker)
hole, located above the highest fingerhole (Plate 4; Plate 5). Whether a similar
key would also have been present on the Pergamon model, whose upper part
is missing, remains obscure. At any rate, it is highly improbable that the aulos
Meg2 was equipped with sýrinx keys, though we cannot exclude the possibility
of simple sýrinx holes as observed on other fragments that must normally have
been plugged by other means.
Knob, slot, rod and plate are all found on the sýrinx keys, too. However, the
rod, following the cluster of knob and slot, at about half its length splits in two
strands, which bend outwards and upwards to form a ring that stands perpendicular to the original part of the rod. Where the two strands meet again, they
bend once more, uniting again to a single rod that extends further in the same
direction, running parallel to the first part, though on the other side of the
tube, and terminates in the small curved plate. However, neither sýrinx key has
survived intact. Apart from various flaws along their body, one part of the rod
was apparently detached from the ring in both cases, affecting the connection
between the knob-side rod part and the ring on the shorter of the two keys,
and that between the plate-side rod part on the longer. Both were re-attached
during restoration. Surprisingly, on a first glance, the key with the longer knobside rod part ends in a somewhat shorter plate-side part. Although there is no
trace of roof-shaped guides, sharply set-off sections of reduced diameter close
to the plate suggest that such parts once existed; their length can be retrieved
by subtracting the effective range of the slot (i.e. the slot length minus the pin
diameter) from the length of the narrow region. A pin body surviving in its hole,
aligned with the thumb-hole towards the upper end of section Meg1αFi, must
have run in the slot of one of the sýrinx keys and thus furnishes the position of
one of them. With the slot and the knob at the underside of the instrument,
and the ring, which obviously went around the tube, switching the pathway
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
29
of the rod to the opposite side, the sýrinx hole was therefore situated at the
upper side.
The question of the apparent sýrinx keys is further complicated by an older
drawing by our late colleague Maurice Byrne, a copy of which Olga Sutkowska
kindly provided. This drawing presents the slider keys in their natural size, and
it may reflect an earlier restoration phase. Here the longer of the two keys is
sketched with its plate-side segment flipped around: instead of continuing its
way towards the top of the instrument, the rod thus doubles back, so that the
plate is ultimately not very far removed from the knob, albeit on the opposite
side of the tube. In addition, the knob-side part of the shorter sýrinx key does
not appear on the drawing at all. Similarly, one of the two short keys is not represented, whereas one of the keys of the intermediate size appears both in top
and side view. Otherwise, only small discrepancies can be observed between
the shape and size of the rest of the slider-keys as displayed in the drawing
and in their present condition. On the other hand, the in-situ photograph
flatly contradicts the drawing, since it clearly shows the plate-side part of the
longer slider key pointing towards the upper end of the pipe, while still being
attached to (parts of) the ring that engulfs the tubing. We shall return to this
matter after dealing with the arrangement of the bone segments in an effort to
reconstruct both auloi pairs with reasonable certainty.
3.4
Towards a Meaningful Assemblage of the Megara Pipes
Despite our limited understanding of the technical intricacies associated with
ancient music culture, some ubiquitous observations on the surviving aulos
finds of various types ranging from the Classical up to the Roman Imperial eras
allow us to state two specific rules that are borne out by all the evidence that
has surfaced so far. Firstly, an aulos consists of a ‘low’ and a ‘high’ pipe, in the
sense that their highest fingerholes do not coincide, even though the ranges
covered by each hand always overlap; within this overlap, the respective fingerholes on both pipes are seemingly intended to produce identical pitches.
The shift between the pipes varies between one and three fingerholes, though
the evidence so far favours offsets of merely a single hole. Secondly, a ‘rule of the
4L’ has been established,15 postulating that the lower-pitched pipe is simultaneously the longer, played by the left hand, with its thumb-hole lying slightly
leftwards from the position opposite to the other fingerholes.
15
The “4L rule” was introduced by Stélios Psaroudákēs (2008, 202) on the basis of his measurements of surviving aulos sections of the early type. The 4L rule is still applicable to
those parts of the Megara finds that reflect the design of such instruments, but cannot be
applied to the lengths of their downmost slider sections.
Greek and Roman Musical Studies 10 (2022) 15–77
30
Terzḗs and Hagel
On Meg2, the little-finger hole of the left-hand pipe as well as the ring-finger
side hole of the right-hand pipe are lost due to substantial damage of the tube
wall at their expected positions (now indiscriminately filled in with plaster).
The centre of the little-finger hole on Meg2αFii, as far as can be determined, may
have sat at any distance up to about 26 mm from the upper rim of the section,
likely close to the lower end of that range, where the downmost part of its round
edge may just be discernible. But then, any effort to align corresponding tone
holes among the pipes is doomed to fail, because the wider-spaced upper three
tone holes (1–3) on βFi cannot possibly be aligned with the more closely packed
(2–4) on αFi that ought to correspond with them. Moreover, the Museum’s original assemblage of Meg2 violates the 4L rule by assigning the longer low-pitched
pipe to the player’s right hand. However, conservator Theodṓrou has convincingly put together the appropriate segments with tone holes, i.e. the F and
S sections of both pipes, on the basis of the only possible solution that aligns
slider pins with slider holes while at the same time accounting for the lengths
of the existing slider keys. Given the problematic distribution of tone-holes
between the alleged pipes and notably, the already mentioned unconventional
methods applied in the excavation, an alternative distribution of bone sections
between the pipes suggests itself. The only possible re-arrangement merely
involves exchanging their extensions or the entire upper parts, for instance
resulting in a high pipe of the scheme α(MX)β(FS), and a low pipe of β(MX)
α(FS). In this way, the 4L rule is preserved, and the corresponding side holes
align themselves within the customary offset of one hole. Moreover, the centre
of the lost ring-finger hole on the higher pipe can now be placed in a natural
position, 6 mm above the lower rim of βFi (without spigot), corresponding to
the middle-finger hole on the lower pipe’s αFi. These suggestions have now
also been accepted by the conservator, and the exhibits have been readjusted
accordingly, without any violation of the basic archaeological criteria for
potential joins, such as surface texture and coloration.
A closer inspection of the remains of Meg1, in turn, leads to a discomforting though inescapable realisation. Taken as it is, its longer pipe would generate
the higher treble note, again in violation of the 4L rule. A comparison with the
longer pipe of Meg2 suggests that Meg1β is missing one of its segments; this
is corroborated by the extant sliders, as we shall see. Pipe Meg1α, however,
resembles the higher member of Meg2, comprising a similar number of bone
sections, equivalent in form and size, with their tone and pin holes at similar
positions and almost identical azimuths. Only one major deviation leaps to the
eye: the presence of the aforementioned pin body in Meg1α, aligned with
the thumb hole, but positioned at some distance from it towards the upper end
of the pipe. As we have seen, this pin probably constrained the movement of
the sýrinx hole key.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
31
The single surviving photograph of the find in situ, which Panagiṓta
Avgerinoú generously provided, not only enables us to trace the missing section, but also to determine its relation to the adjacent parts, and finally to
estimate its length. The instrument was found resting between the leg bones
of the deceased, with the right-hand pipe placed to the right, resembling the
playing position. The disintegration of the junction between Xii and Fi seems
to have resulted in a slight displacement of the part of the pipe below it,
whose finger holes were facing the ground, probably as the result of a counterclockwise rotation (from the viewpoint of the deceased) by approximately
180 degrees. The parts of the higher pipe of the pair at the right side displayed
a similar shift of the MX complex, presumably also resulting from its collapsing
in the course of the body’s decay.
Although βFi and βSi have consecutively been arranged during conservation, close observation of the series of bone fragments in the excavation image
reveals an extensive gap between βFi and βSi, reflecting severe damage of the
find at this particular point. The image clearly suggests that this gap was occupied by an additional bone section, then highly fragmented, equipped with a
small hole. How this part became lost remains a mystery – after all, a significant number of other little fragments were successfully recovered. At any rate,
precise measurements of the preserved pipe components enable us to determine the total effective length of the missing βFii section as 71.1 mm.
The rest of the available lengths can be verified on the image, with one
exception. Meg1αS exhibits an almost transparent short cut at its otherwise
entirely broken upper end, giving the impression of a surviving rim. However,
close inspection of the in-situ image suggests that the upper end had actually
been longer by about 3.5 mm (resulting in an unusually deep socket of no less
than 21 mm), so that it accommodates the extant slider keys nicely.
3.5
Assigning Slider Keys to Their Original Positions
Since conservator Theodṓrou has assigned the three sizes of keys to their
proper positions – the long one (L) to hole S3, the intermediate (M) to S1, and
the short (S) to S2 – one may wonder why the sliders were not accordingly
placed on the tubes of Meg1. Of course, devoid of its Fii part, Meg1β could
not have accommodated any of the available triads of keys. On the other
hand, given the striking structural similarities between the F-S complexes of
Meg1α and Meg2β, the restoration of its keys to the former would have been
straightforward. Also, by matching the remnants of Meg1α with the available
keys, the absence of an Fii part as found in its counterpart Meg1β would have
promptly emerged.
Greek and Roman Musical Studies 10 (2022) 15–77
32
Terzḗs and Hagel
However, detailed measurements of the keys reveal mismatches, at least at
first glance: some of them do not seem to fit in their proper positions perfectly.
Disconcertingly, this does not only concern Meg1, but also Meg2. However,
since we can hardly doubt the functionality of the mechanism, unless we misunderstand something very basic about its function, we must assume that the
mismatches are the result of small dimensional distortions (possibly accumulative) caused by the required restoration work. We would therefore like
to conclude this preliminary study of the Megara instruments by suggesting
probable corrections, while at the same time providing the raw data that will
allow a critical assessment of our suggestions (Table 1–Table 4).
Our search for a feasible solution can be based on some robust parameters,
most of all the fixed lengths of the bone segments of Meg2, which are safeguarded by surviving rims, as well as those joints whose rotational alignment
is precisely determined by a pin hole through socket and spigot. Furthermore,
the sliders of Meg1 appear to be in excellent condition, so that we can exclude
any modification of their lengths in the conservation process. No solution
ought to be proposed that tampers with these data.
Interestingly, all twelve slider holes on the two instruments can be assigned
well-fitting sliders, as long as only the lower part of these, consisting of guide and
plate, is considered, establishing matches between the guide niches, the movement range and the edges of the side holes. The niches thus match the lengths
of the roof-shaped guides, and this leads to configurations where the holes are
nicely covered and uncovered. Thus it emerges that six out of the seven surviving
sliders associated with the find of Meg1, and doubtless the entire set assigned
to Meg2 indeed belonged to the instruments under investigation. The discrepancy concerning Meg1α is apparently systematic: compared with the distance
from pin to hole, all of its sliders appear too long by 3.5 mm. This is easily
accounted for by increasing the length of its S section by the same amount: the
preserved end of this fragment does not present its original upper rim, after all.
The case is different with Meg2, whose section lengths are undisputed,
while the poorly preserved sliders also appear slightly too long, but by different amounts. This can probably be explained by the nature of the conservation
process, in the course of which the rod lengths became misrepresented by the
frequent necessity to apply larger portions of glue between the corroded edges
of the broken and distorted parts. Unquestionably, the conservator has wisely
refrained from jeopardising the physical preservation of fragmented rods for
the sake of an ultimately unrealistic functional restoration.
Therefore the nature of the evidence suggests estimating the lengths of the
poorly preserved Meg2 sliders from the intact bone parts; conversely, in the case
of Meg1 we needed to assess the dimensions of a missing and a broken bone secGreek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
33
tion from the well-preserved sliders. The proposed reconstruction therefore
includes the lost section Meg1βFii of 71.1 mm length, adding 3.5 mm to the
upper end of Meg1αS, as well as reducing the rod lengths of the Meg2 keys, by
up to 5 mm in one extreme case. In this way, we obtain twelve perfectly functional slider keys (Table 5). Of the three preserved slider keys of intermediate
size, it emerges that I2 does not fit on either pipe.16 It must have been part
of another instrument, perhaps the one otherwise represented only by the
aforementioned small bone fragments that cannot have belonged to any of
the pipes discussed here.
Finally we need to assess the original shape and mechanism of the supposed
sýrinx keys. Their function appears to include the otherwise mysterious nailed
bronze rings on the tube, which occupy corresponding positions on both pipes.
When adjusting a model of the short key (Sy1) to the surviving pinhole (p.s.)
on Meg1αFi, the ring of the key comes to lie above the nailed ring and consequently moves between it and the first few millimetres of the recess around
the upper rim of αX, situated about 5 mm above. The same seems to be true
on the other pipe: when the slot of the longer key (Sy2) is aligned with the
probable pinhole position (p.s.) on the spigot of Meg1βXii, its ring similarly
moves from above the corresponding nailed ring towards the recessed area
near the upper rim of Meg1βXi.
By butting up against these fixed rings, the moving rings on the sliders thus
restricted the downward movement of the key, assisting the corresponding
functionality of pin and (probably) guide. In addition, they would be forced
into a precise position regarding all three spatial dimensions, which guaranteed the accurate placement of the plate, irrespective of otherwise possible
wobbling of the construction. This detail suggests that the plates covered
their holes when the sýrinx keys were in the downward position. Notably,
such a construction, however, conflicts with the notion of ‘pulling downward’
(κατασπᾶν) for activating the sýrinx, which is found in Aristoxenus as well as
the Peripatetic De audibilibus.17
As described above, an early drawing appears to suggest that the part of the
sliders, or at least of the long one, that connected ring and plate did not extend
further towards the upper end of the instrument, but turned back towards the
knob, though on the other side of the tube. When we implement such a design,
16
17
One of our reviewers wondered whether a slider hole may have been lost. This is impossible on the higher pipes because the bone at the respective position is undamaged, as
is the surface where the corresponding slider pinhole would be expected. On the lower
pipe, the absence of a pinhole similarly precludes a fourth slider key.
Aristox. Harm. 1.21 (27.1 Da Rios), [Arist.] Aud. 804a.
Greek and Roman Musical Studies 10 (2022) 15–77
34
Terzḗs and Hagel
reversing the direction of the plate parts in regard to their present state, the
slider plate of the short key (Sy1) settles on the upper-left side of the indexfinger hole of Meg1α, in a region where the original tube wall has not survived.
On Meg1β, in turn, the key plate of the longer key (Sy2) would extend to the
lower end of a large gap on the top surface of the bone section, corresponding
with the assumption that extensive damage may be correlated with the corrosion of a bronze rod on the bone surface. The resulting distances of the sýrinx
holes from the upper ends of the pipes would thus differ by less than 6 mm.
Interestingly, this difference would disappear by exchanging the rod parts
containing the plates among both keys; but this would presume that the rings
themselves are assembled wrongly from fragmentary portions – an assumption that cannot be explored at present, but appears a priori not all too likely.
This kind of design has one obvious advantage: consisting of facing parts on
opposite sides of the tube, the keys would easily maintain an even grip on it
and consequently guarantee the sealing of their holes exceptionally well. Also,
if the key did not run along the bulb, it would impact less on the aesthetic
appearance of the instrument. However, it is hard to see why such a large and
complex mechanism would have been implemented for bridging such a small
distance between operating finger and sýrinx hole, and why it would have been
designed to go around the tube. After all, there was no reason to place those
holes on the upper side of the instrument.
More importantly, such a design would have compromised the functionality
of the instrument. On the one hand, on a respective model of the high pipe, the
plate of the sýrinx key sits so close to the index finger hole that it encumbers its
operation significantly, since the finger can hardly seal the hole without touching the plate. On the other, a small hole so close to the finger holes is ill placed
for producing the intended effect of serving as a speaker hole, which is meant
to switch the pitch produced on the instrument to a higher partial by weakening the oscillatory regime of the fundamental, at least, or even of several of the
lowest partials.18 For this purpose, a speaker hole needs optimally to be placed
at about one third of the distance between mouthpiece tip and the highest
open finger hole. In practice, a single speaker hole may work well for several or
all finger holes, but it still needs to be situated high up the pipe. At the position
effected by the hypothesis of downward-reaching sýrinx key plates, it might
work only for the lowest tone holes, if at all. Indeed such a position is not paralleled on any of the known eleven certain or possible extant sýrinx holes.19
Especially the earliest examples are generally situated on the uppermost parts,
18
19
Cf. Hagel 2012a, 498f., with 515 fig. 1.
Hagel 2012a, 517 fig. 5.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
35
either in the narrow section between reed socket and bulb (Delos, B5168;
Taranto, National Museum 12528/7 and s.n. 2; apparently also the Reading
aulos in the Ure Museum), on the bulb (Athens, Agora BI 593) or on the neck
below it (Paestum Archaeological Museum 23068, both pipes); on later instruments they may also be found on the bulb (Berlin, Egyptian Museum 12462),
or on the tube immediately below it (Berlin, Egyptian Museum 12461; Naples
National Museum 76892; one of the Meroë fragments, Bodley (1946) pl. 3.2).
Would the sýrinx keys in the shape suggested by their present state work out
better? When the slot of the longer of them is aligned with the pin in Meg1αFi,
it turns out that the plate of the longer would extend beyond the expected
upper end of the pipe, which is impossible. The shorter, in contrast, would
reach up between bulb and cone, ending about in the region of the socket
where the reed stem is inserted. The longer sýrinx key would thus need to be
assigned to the lower pipe, just as one might expect, where it would terminate
in about the same region, when its slot is aligned with the possible pin location
at the spigot of Meg1βXii.
It is difficult to establish the positions of the expected sýrinx holes with precision, due to small uncertainties regarding the repairs of the rods as well as
the placement of the nailed rings. Figure 2 displays the results of three different approaches, one working from scaled photographs of the individual
parts, starting at the fixed rings, another calculating the measured distances
from the placement of the fixed rings, and the third one starting from the
pin holes. Each of them predicts a specific range within which the slider plate
would move; in both cases, these are somewhat higher when the last method
is used. Judging from the placement of the fixed rings, therefore, the measurements for the lower rod lengths appear slightly too long: when the sliders are
drawn downwards until stopped by the pin, the moving rings would still
not touch the fixed. The discrepancy is in the range of less than 2.5 mm,
however, and larger on the higher pipe, where the slider is badly bent and corroded and the correct alignment of the rings difficult to establish (Table 6).
Each moving plate establishes two possible regions for the hole, one on
either side of it. Above we have concluded from general consideration concerning the alignment of the rings that the sýrinx hole was more probably
closed by moving the slider key downwards. This appears corroborated by the
ranges. As can be gleaned from Figure 2, the sýrinx plate on the higher pipe
came to lie above the reed socket when pushed upwards. Here a hole would
be useless, because it would only extend down to the wall of the reed, once
one was inserted. One might of course drill a matching hole into each reed;
but that would be extremely impractical for several reasons. Apart from the
troubles involved in aligning the holes, such an approach is hardly compatible
Greek and Roman Musical Studies 10 (2022) 15–77
36
figure 2
Terzḗs and Hagel
Possible sýrinx hole positions on Meg1 for straight sýrinx keys
with the practice of ensuring an airtight connection by a winding of (waxed)
thread around the foot of the reed. Even if the thread would initially be placed
so as not to cover the hole, it would likely move when the reed is pushed into
its socket. Secondly, one would lose the paramount possibility of fine-tuning
the instrument by changing the position of the reed in its socket.
On the lower pipe, the plate, when pushed upwards, would come to lie below
the reed socket, where it might cover a functional speaker hole. However, as far
as we see, the wall of the cone is intact throughout the respective region without showing a trace of a hole. Consequently, both sýrinx holes must have been
situated at the downward-side of the slider plate.
Indeed the higher pipe preserves a tiny opening about 64.5 mm above the
lower effective end of the bulb section (Meg1βM), resembling other extant
sýrinx holes, in the range predicted by the ring-based approaches. Its position
on the slope of the bulb close to the narrowest point resembles that on one of the
mouth-end pieces in Taranto. Given the nature of the slope, the plate would need
to be inclined; consequently, a hole at this position can only be closed when the
slider is pulled downwards. Being situated within the predicted areas of two
out of three methods and no more than 3 mm removed from the lowest usable
position of the third, this small opening therefore plausibly formed the sýrinx
hole of its pipe.
On the low pipe, no similar hole is preserved. The bulb has lost substantial material around where we would expect it, and once more this is likely
related to the proximity of the copper alloy. However, the calculated ranges fall
within that length of the bulb where the spigot of the cone part was inserted,
which has survived largely intact, with only one larger part missing from the
rim. If we have not overlooked traces of restoration elsewhere, it is apparently
here that we have to search for the original hole position. The innermost edge
of the gap seems consistent with the assumption of a small drilled hole and
lies just within the calculated boundaries, about 2 mm from their centre, at
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
37
about 59.2 mm above the effective lower end of the bulb section (Meg1αMii).
For optimal functionality, the upper part of the rod ought to be at least 1.3 mm
shorter than our data suggest; this small mismatch may once more result from
uncertainties associated with the restoration and measurement of the bent
slider rod. Interestingly, at 58 mm from the effective lower end of the bulb piece,
a transversal line can be observed that might be related to the placement of the
slider, perhaps indicating the position of the slider’s roof-shaped guide.
On balance, the sýrinx keys in their present state seem to make much more
sense than they would in the way they appear on the early drawing: for the
shorter key, an existing hole strongly suggests that it extended beyond the bulb;
the longer, in turn, would appear dysfunctional in the reverted form. On top
of this, the precise curvature of its longest intact fragment appears to reflect
the shape of the bulb and the neck below it (cf. Figure 2); aligned backwards
onto the cylindrical surface of the tube, its shape would be incomprehensible.
The stretched-out version of the keys also explains other details. First of all,
the sýrinx hole of the higher pipe is also situated a bit higher – about 5.5 mm –
just as has been observed on the Paestum and the Berlin auloi, the only extant
pairs with such holes on both pipes. Since the rings sit at similar heights on
both pipes, doubtless for aesthetic reasons, it is consequently required that
the plate-side rod of the higher pipe and therefore of the overall shorter key
is longer: what appeared a curiosity, thus becomes a necessity. The additional
enforcement of this particular type of key with a system of two rings was likely
prompted by the reduced stability of the part of rod that needed to move in
the air above the curved surface, in contrast to the tone-hole keys that could
rest safely on a straight underground. Finally, the precise positioning of the
holes at asymmetric distances from the respective pipe tops shows that their
location was carefully chosen. The higher one, on the higher pipe, appears precisely at the highest possible position. As we have seen, it could not be placed
within the range of the reed socket in the cone. Consequently it needed to be
drilled on the slope of the bulb, just far enough from the cone so that the latter
would not obstruct the upwards movement of the key plate when the sýrinx
was opened. This is no coincidence. As the comparison with Meg2 shows, the
two instruments differ almost exclusively in two respects: Meg1 features sýrinx
keys, and its upper part is longer. The latter now emerges as a corollary of the
former: in order to put the sýrinx holes at the desired positions, the tubes had
to be extended, shifting the bulbs closer to the player’s mouth. However, it
appears that this prolongation involved hardly a millimetre more than was
absolutely necessary – perhaps an indication that suitable bone was not an
unlimited resource, after all, or that long reeds were still favoured for aesthetic reasons.
Greek and Roman Musical Studies 10 (2022) 15–77
38
Terzḗs and Hagel
Sealing a hole on a three-dimensionally curved surface by means of a metal
‘plate’ posed a significantly greater technical challenge in comparison with the
cylinder segments that covered the lower holes. Possibly the plates were cast
from a wax form that had been moulded right onto the pre-fabricated bulbs.
Or the lower surface of the plates was simply padded with some softer material such as leather. How the slider guides, whose existence is suggested by the
narrower lengths of rod close to the plates, may have been shaped and fitted
on the bulbs needs to be explored by further experimental research. Have they
disappeared because they were made from a different material?
It has been proposed that aulos makers of the Roman period may have placed
the sýrinx holes at one third of the distance between the upper pipe end and
the highest finger hole.20 This rule of thumb would have disregarded the reed
cavity; but Roman-period reeds were much shorter than those of early instruments. Clearly, at least on those which had their sýrinx holes at the mouth side
of the bulb, a very different strategy must have been followed. For instruments
with holes throughout down to the end, obvious points of reference might
have been the upper and lower end of the pipes as well as the centre or the
upper edge of the highest finger hole. Of all possible combinations of these,
the design of Meg1 gives a meaningful result only for that of upper pipe ends
and index hole centres, analogously to the later examples.21 However, instead
of using a factor of three, the sýrinx holes appear to have been drilled at only a
seventh of the distance to the highest finger hole. At least, this model predicts
the positions we have suggested above by error margins of 0.4 and 0.6 mm.
Unlike the later rule of thumb, this practice would not have been optimal for
producing the first available harmonics, (approximately) a twelfth above the
fundamental scale, throughout. Was the sýrinx of these early instruments
meant to destabilise this oscillatory mode as well, in order to produce even
higher notes, even more suitable for depicting a dying dragon’s hisses? After all,
on the later instruments, the second mode would have extended the available
scale upwards, which, it has been argued, was an important function of the
sýrinx at least on the Berlin aulos.22 On the instrument from Megara, in contrast, the different vibration modes inevitably create well-separated scales, in
accord with Aristoxenus’ take on the matter, which appears to presuppose at
least the second available harmonic, two octaves and a major third above the
20
21
22
Hagel 2012a, 505–9.
For this purpose, the length of the mouth-end part of the higher pipe was extrapolated
by scaling its lost cone section by the same tiny ratio by which its bulb section is smaller,
shrinking the extant cone of the lower pipe by 1.16% or a quarter of a millimetre.
Hagel 2010.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
39
fundamentals.23 With the mitigating effect of the comparatively long reeds, at
any rate, the effective ratios would amount to 3.28 and 3.55 for the index holes,
so that overblowing to a twelfth may have been possible from the highest fingerholes, while lower tone holes would likely have produced higher harmonics.
The effects cannot be predicted, though – too much depends on the softness
of the reed blades, on lip control and playing pressure.
3.6
Conclusions
A close study of the Megara pairs thus corroborates the first impression that
they belong to an otherwise unattested aulos type. Each pipe had five fingerholes plus three holes towards its downstream exit that were operated
remotely by bronze sliders. While the arrangement of these holes follows
the same general principles on both pairs, their precise positions nevertheless differ so markedly as to suggest intentionally different musical designs.
Most conspicuous is the divergence between the upper two key holes on the
higher pipes, as well as the wider spacing of the fingerholes on the higher pipe
of Meg2, leading to an enormous distance between the centres of index and
small finger of 13.5 cm, compared with the already straining 13.2 cm on the
higher pipe of Meg1.
Generally, the comparatively even spacing of the fingerholes resembles
the design of earlier auloi, which have been interpreted in terms of tones and
¾-tone intervals or equally divided tetrachords.24 The small steps between
adjacent slider holes at the lower end, however, suggest smaller intervals.
Regarding the function of the sliders, it is paramount to appreciate that the
finger that operated any of them could not do so without releasing its own
fingerhole. As a consequence, it would never have been possible to play two
bass notes in immediate succession. Between these, the player would either
have had to pause, or some higher note would have sounded, the one associated with the fingerhole of the operating finger or any higher fingerhole note
by lifting another finger as well. In this way, the sliders appear more suitable for
changing the available bass note, and thus effecting some kind of modulation,
than for melodic use. Nonetheless, the considerable technical effort associated
with the production of remote keys, as opposed to other means of changing
the available pitches, such as plugging of holes, certifies that they must have
been used for swift action right within a performance, not merely between
different pieces.
23
24
Aristox. Harm. 1.20–1 (26.8–27.3 Da Rios); cf. Hagel 2012a, 497; 511–13; 518 fig. 8.
For a detailed discussion of potential tetrachord divisions of surviving auloi of the early
type, see Hagel 2021a, 436–47.
Greek and Roman Musical Studies 10 (2022) 15–77
40
Terzḗs and Hagel
The astounding parallels in the arrangements of the bone sections and the
placement of the keys might suggest that the two instruments originated in
the same workshop. On the other hand, the two similar sets of slider-keys
unearthed in Lefkada, in combination with the fragment from Pergamon,
testify to standardised features far beyond the region of Megara. Common
characteristics would have included the number of finger and slider holes, the
latter constrained by the available space around a tube, and the arrangement
of the sliders, guided by practical necessities – most importantly, the rods must
not encumber the fingers. Within these general parameters, individual designs
seem to have differed widely; the much larger spacing of the Pergamon slider
holes does not resemble the small intervals found on the Megara instruments.
These differences set aside, we may well be dealing with the most important
general type of professional modulating aulos of the Hellenistic period, flourishing across a large geographic region.
4
Musical Evaluation
4.1
General Considerations
The music-archaeological evaluation of an aulos cannot be considered complete without embedding its design within our wider knowledge of Greek
music in terms of scales, pitches, notation, ‘modes’ and their evolution. This
endeavour can hardly become more exciting than in the case of the instruments from Megara, the lifetimes of whose owners likely overlapped with
those of Aristoxenus, to whom we owe most of our knowledge of ancient
musical structures, and Theophrastus, our primary literary source for the construction of aulos reeds.
As we have seen, the Megara auloi were almost certainly modulating instruments, and in spite of all their structural similarities, also different enough not
to be addressed as two different specimens of precisely the same design. Their
modulating nature emerges from the number of consecutive small intervals at
their ends, operated by sliders (though hardly available in immediate melodic
succession). Such a row of microtones does not fit any single-scale system of
ancient theory, from its earliest attestation in Aristoxenus up to late antiquity.
On the other hand, similarly small intervals are not, and for obvious reasons
cannot, be implemented between the five finger holes per pipe. Nevertheless,
if the instruments modulated between scales where the small intervals in the
bass region played a role, similar intervallic shifts are also to be expected in
the higher pitches. Without a mechanism for changing the position or size
of the fingerholes, or switching between different sets of them, and without
the option to use cross-fingering (which is useless on instruments where the
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
41
fingerhole diameters approach that of the bore), such modifications of pitch
would have to be effected by partially covering fingerholes and/or perhaps
modifying the effective lengths of the double reeds by moving their respective
positions relative to the player’s lips.
As a consequence, we must anticipate the possibility that the physical positions of many, if not most fingerholes result from a compromise. Moreover,
the nature of this compromise is difficult to determine a priori. On the one
hand, it would be straightforward to drill each hole so that it emits the highest
pitch that may be required from it. On the other hand, the nature of Greek
music ascribes the status of ‘fixed’ notes to the lowest of a triplet separated
by small intervals (called a pyknón), and these lowest notes were evidently of
higher modal importance than the ‘moving’ notes inside the tetrachord. An
instrument where many of these ‘fixed’ notes would need to be played by
the unstable means of partial covering or reed manipulation, making them
in practice more mobile than their ‘moving’ peers, would appear particularly
awkward. We ought therefore to reckon with instrument designs that tried to
achieve modulating capabilities while maintaining stable ‘fixed’ notes as far
as possible – and as far as anything is stable on a double-reed instrument.
However, the probable nature of the expected compromises deprives us of
one of the mightiest tools of aulos research. Since vital consonances may have
been played between open and partially covered holes, the automated search
for an optimal reed configuration in terms of maximised harmonicity may
yield misleading results. Nevertheless the technique of modelling the pitches
expected for various reed lengths by means of dedicated software must form
the basis of the following considerations, in the hope that the physical design
of the instrument reflects at least a sufficient amount of structurally primary
notes – and that consistent results will eventually justify the method. After all,
the uneven distribution of fingerholes betrays that they were carefully placed
precisely for determining their pitch relations;25 any musical interpretation of
the instruments must therefore account for all these positions.
Since, as we have seen, the key lengths and bone sections of our instruments
control each other sufficiently to establish most of the musically decisive
data – though the precise position of three fingerholes is unfortunately lost –
the paramount question, as usual, concerns the effective length of the required
reeds. On early auloi, these may have extended up to a palm from their socket,
while they typically measured only a few centimetres on later instruments. At
25
Note the existence of at least one treatise ‘On the drilling of auloi’, in more than one
book (Aristox. ap. Ath. 14.634e). Such an approach stands in contrast to instruments with
evenly spaced fingerholes (or fingerholes with evenly increasing distances in either direction) in many musical cultures, which require pitch adjustment throughout.
Greek and Roman Musical Studies 10 (2022) 15–77
42
Terzḗs and Hagel
any rate, the reed sockets from Megara preserve the characteristic step which
ensures that the internal diameter of the reeds continues that of the main
bore as smoothly as possible. This is important, because it allows us to work
from the assumption that the effective length of the reed can, in very good
approximation, be treated as a constant for all notes on a pipe. With the given
wide diameter of the bore and consequently the reed, the blades of the latter, which resulted from flattening one end of a length of cane,26 must have
been quite substantial: starting from a tube of 13 mm diameter (i.e. that of
the reed socket), the resulting blades would be 20 mm wide and proportionately longer than those of narrower auloi. Their tips must have extended beyond
the pipe end by more than 35 mm. On the other hand, their maximal extension
cannot have exceeded 70 mm by much.27 Notably, three of the four fragile reed
sockets were destroyed beyond repair, so we only know the total dimensions
of the uppermost sections of Meg1β; for Meg2 we need to estimate the length of
the missing reed socket from its extant termination. If the depths of the reed
sockets were similar, and identical on both pipes, Meg2αM would have been
9.1 mm longer, and Meg2βM, 11.6 mm. On a typical aulos pair, the effective reed
lengths for the two pipes differ only marginally. It may, however, be expected
that the reed of the right-hand (higher) pipe, even if cut to the same physical
length, would be effectively shorter by a slight amount, up to a few millimetres,
if the pair of reeds was retrieved from the same internodium of a single plant,
as described by Theophrastus.28
4.2
Bass Notes
While the actual pitches of each side hole note as well as the actual intervals
between these notes depend on the absolute effective reed lengths, this is
not similarly true when we confine ourselves to the question of the identity
between respective notes on the two pipes of a pair, as long as it is assumed
that the pipes were equipped with identical reeds (of any reasonable length).
By computing the expected pitch relations, we may thus be able to determine,
26
27
28
Cf. Hagel 2021a, 430–5.
Cf. Hagel 2021a, 428f.
Hist. Plant. 4.11.7 συμφωνεῖν δὲ τὰς γλώττας τὰς ἐκ τοῦ αὐτοῦ μεσογονατίου, τὰς δὲ ἄλλας οὐ
συμφωνεῖν· καὶ τὴν μὲν πρὸς τῇ ῥίζῃ ἀριστερὰν εἶναι, τὴν δὲ πρὸς τοὺς βλαστοὺς δεξιάν. τμηθέντος δὲ δίχα τοῦ μεσογονατίου τὸ στόμα τῆς γλώττης ἑκατέρας γίνεσθαι κατὰ τὴν τοῦ καλάμου
τομήν· ἐὰν δὲ ἄλλον τρόπον ἐργασθῶσιν αἱ γλῶτται, ταύτας οὐ πάνυ συμφωνεῖν. ‘The tongues
from one and the same internodium, they say, are in concord, but not the others; and
the one closer to the root is the left-hand, the one closer to the blossom the right-hand
[tongue]. When the internodium is cut in two, the mouth of each tongue is fabricated at
the cut of the cane. If the tongues are manufactured in another way, they are not in good
concord, it is said’. Cf. the 3–4 mm predicted difference in effective length for the Pydna
and Paestum auloi in Hagel 2021a, 429.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
43
tentatively at first, whether it is likely that two fingerholes at roughly the same
distance from the upper end of each pipe were intended to play in unison or
not. Or, alternatively, assuming that a certain pair of notes between two pipes
sounded the same pitch, we would become able to establish for which of the
other potential note pairs this would have been true as well and for which not.
When the upper ends of the pipes of Meg1 are aligned in this way, the pitch
of the second-lowest slider hole of the lower pipe is only slightly higher (within
less than 20 cents) than that from the exit of the higher pipe. The pitches
become identical when we make the reed of the lower pipe effectively longer
by 5 mm, in the way Theophrastus leads us to expect, but perhaps by just a bit
too much.
Meg2 is slightly different. With the proposed reconstruction that switched
the entire complexes of mouth-end plus extension sections between the
pipes, a similar 4 mm difference between the reeds yields 20 cents difference
between the respective pitches, but in the opposite direction; similar effective
reeds therefore give an even larger difference of 32 cents, and in order to get
the respective pitches from the higher pipe exit and the lower pipe secondlowest slider hole identical, the effective reed lengths would need to differ by
an unrealistic 10 mm.
However, the mouth-end parts of higher and lower pipe are not completely
identical; between the termination of the reed socket and the effective end
they measure 70.85 mm and 69.5 mm respectively. Since their spigots, however,
appear identical, it should be possible to exchange them between the reconstructed pipes, so that the slightly shorter item becomes part of the shorter pipe.
In this way, the required effective length differences between the reeds drop by
2.7 mm. Correspondingly, the pitch difference with equal reeds is diminished
to 25 cents, and with a more realistic assumption of 4 mm effective difference
between the reeds, to 13 cents. Therefore we propose revising the original
Museum restoration of Meg2 only by exchanging the extension sections (X),
leaving the uppermost sections (M) in place.
4.3
Coincidence of Pitches: Higher Tone Holes
When the discussed pair of pitches is aligned, those of the highest slider hole
(s1) of the lower and the middle slider hole (s2) of the higher pipe also become
very close – though apparently not identical; once more with reeds differing by
4 mm, these holes are predicted to differ by 20 cents on Meg2, and by 16 cents
on Meg1, though in opposite directions: the hole on the lower pipe is lower on
Meg1, but higher on Meg2.
But what about the fingerholes? On the known early auloi, four of them
generally coincide: with an offset of a single interval, index to ring finger on the
lower pipe can play the same notes as do thumb to small finger on the higher.
Greek and Roman Musical Studies 10 (2022) 15–77
44
Terzḗs and Hagel
On the Megara instruments, in contrast, the respective hole positions coincide
only very roughly, and the distances are incompatible. As a result, even though
the pitches of these holes, which form the higher range of the pipe, are much
more sensitive to changes in reed length, no plausible configuration can make
them coincide. On Meg1, the central two of the four hole pairs in question may
play identical notes, when the reed of the shorter pipe is effectively longer by
7 mm, contrary to expectation and Theophrastus’ statement. But even in this
way, the outer notes differ significantly, by almost a sixth of a tone. Since those
of the higher pipe span a larger distance, its lower note in question is lower
than its counterpart, and the higher note is higher. On Meg2, a much better
accord can be achieved, but here also the reed on the shorter pipe would need
to be longer, by 7 mm. Nonetheless, had we only this instrument, we would
almost certainly try to enforce a musical interpretation on that basis. The obvious structural parallels with Meg1 encourage us, however, to take into account
the possibility that non-matching fingerholes might form a structural feature,
after all.
4.4
Intervals: Bass Notes
Independently from the precise reed lengths, the two lowest intervals on the
long pipes as well as the consecutive lowest on the short pipes must have been
very close to actual quartertones. We can specify the boundaries with some precision on Meg1β, where the depth of the reed socket is known. For the shortest
possible reeds, the intervals in question would measure 47, 63, and 61 cents
respectively; for reeds extending rather extreme 80 mm from the instrument,
they would instead amount to 43, 58, and 56 cents. Such differences are barely
noticeable; all the values are close to the true quartertone (50–51 cents29) or
lying between a quartertone and a third of a tone (67–68 cents). Assuming
similar reed socket depths, the respective ranges on Meg2 are 55–60, 53–58,
and 60–64 cents.
These figures, it must be noted, refer to plausible reed-length ranges on the
two instruments individually. However, although the upper ends of Meg2 are
not preserved, it is clear that the distances between the highest fingerholes
and the top of the instrument without reed were obviously smaller on Meg2.
Normally this would indicate that the instrument was higher pitched overall.
29
The double values are necessitated by the unclear relationship between Aristoxenus’ definitions and the modern notion of an equally tempered tone. The smaller value follows
from an equation of Aristoxenus’ tónos with the latter, according to his stance that the
circle of fifths can be closed for practical purposes, the larger from his definition of a tone
as the difference between perfect fifth and fourth, intervals that the master would not
have experimentally ‘tempered’ in real-life music-making.
Greek and Roman Musical Studies 10 (2022) 15–77
45
Two Auloi from Megara
But in the present case such an interpretation is hardly possible. The general
hole disposition is very similar on both pairs, pointing to very similar ranges or
even standardised pitch. Moreover, where we observe significant variation at
all, in the finger spans required for the two higher pipes, Meg2 has the larger
distance, demolishing the hypothesis of its representing a higher-pitched
variant of the same design. Consequently we must assume that the difference in upper lengths was compensated by a difference in reed length. As we
have seen, the longer tubing of Meg1 was almost certainly adopted in order
to accommodate the sýrinx slider keys, which inevitably shifted the boundary
between bone instrument and cane reed upwards.
Taking the preserved lower ends of the reed sockets as points of reference,
we would therefore expect that reeds suitable for Meg2 were about 14–18 mm
longer than those for Meg1. This in turn narrows down the plausible ranges for
the effective reed lengths, because the shortest possible overall length on Meg2
would produce reeds which are too short on Meg1, while the longest plausible
reeds on the latter would demand improbably long ones on the former. Set aside
the differences between the pairs for the moment, and the plausible ranges thus
emerge as 50 mm to about 80 mm maximum for Meg2, and 35–65 mm for
Meg1. This affects the ranges for the lowest intervals only marginally. Their
sizes in cents are now predicted as follows:
low E – S3
Meg1:
Meg2:
low S3 – S2
44–47
59–63
103–109
55–59
53–56
109–115
high E – S3
58–61
60–63
Irrespective of the originally intended reed sizes and ensuing absolute pitches,
we may thus state, as a first certain result, that on the Megara instruments
a crucial structural function was attributed to intervals that ancient theory
clearly assigned to the enharmonic genus. Still, the presence of any sort of
‘quartertones’ in the slider notes as such does not a priori prove the melodic
use of these intervals; after all, we have seen that slider-operated notes could
hardly be played in immediate succession. At any rate, the lowest couple of
tone holes on the longer pipes together with its exit form a veritable enharmonic pyknón.
Greek and Roman Musical Studies 10 (2022) 15–77
46
Terzḗs and Hagel
On Meg1 the quartertones of this pyknón add up to a very precise semitone, which, as we have seen above, is practically identical with the difference
between the exits of the pipes. If the pipes were in fact intended to differ
by a semitone (either the abstract ‘tempered’ semitone of Aristoxenus and
his predecessor harmonicists of 100–102 cents, or the slightly smaller one of
92 cents obtained by tuning or modulating in perfect fifths and fourths), the
coincidence of this semitone with the size of the enharmonic pyknón would
be reduced to a corollary of the strictly Aristoxenian nature of the latter. In this
case, we should not expect a similarly precise coincidence on Meg2 with its
slightly wider pyknón.
4.5
Intervals: Higher Tone Holes
In spite of the closely-packed pyknón-like structures at the bottom of the instruments, the distances between the fingerholes on the upper parts by no means
reflect the distinction between smaller semitone-like and wider intervals of
which the respective enharmonic scales or their related semitonic diatonic and
chromatic counterparts would consist. Instead, the wide and comparatively
regular – though not equidistant – spacing appears to proliferate the raw tonality of older auloi with neutral thirds and near-equally divided tetrachords,30
where the smaller intervals which appear in all regular scales would mostly
have been realised by partial covering. On the one hand, such a design may
certainly be explained by purely physical demands: with the required extreme
finger span, a distinctly unequal spacing of fingerholes becomes impossible to
manage. On the other hand, a wider spacing also opened up richer choices of
playable microtonal shades, given a sufficiently expert performer.
4.6
Octave Harmoníai
On balance, we must therefore expect that the makers of the Megara auloi
were inspired by a musical system that worked with quartertones, perhaps
aligning its scales along an ideal grid of quartertones, while at the same time
encouraging playing techniques which reflected the microtonal variability that
plays such an important role in Aristoxenus’ theory of tetrachord divisions.
Aristoxenus mentions two models that endeavoured to incorporate existing scales within a comprehensive system of potentially modulating musical
keys (tónoi or trópoi).31 He explicitly associates one of them, which places
some scales at distances of three quartertones, with the making of auloi.
30
31
Cf. Hagel 2021a, 436–43.
Aristox. Harm. 2.37–8 (47.1–16 Da Rios). For the reconstruction of these scales see Hagel
2000, 165–89; Hagel 2009, 371–90.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
47
The other one is entirely ‘commensurable’ on a quartertone grid. Although it
appears more modern, Aristoxenus mentions it first, so that we do not know
whether the author’s final criticism of an approach to which he refers by the
term katapýknōsis “condensing [the scheme]”32 as unsuitable and violating
the harmonic rules would apply to both or only the one described second. At
any rate, it is difficult to find any principal fault with the first-mentioned system on the grounds of Aristoxenus’ own tenets, so his presentation may have
inverted the historical sequence precisely in order to obfuscate the fact that
the more modern system did not actually merit his condemnation.
As is well known, the interval of the octave had always featured prominently
in ancient music theory. Apart from systematic endeavours to describe all
modal scales as ‘species of the octave’,33 it is telling that the dignified term harmonía itself had acquired the sense of ‘harmonically structured octave’ as early
as Philolaus.34 Even the overall pitch range of early auloi, in spite of their dearth
of tone holes, was evidently defined by an octave between the exit of the lower
pipe and the highest fingerhole on the higher.35 Later the lower pipe alone on
an instrument such as the Louvre aulos would form an octave scale.36 It does
not therefore come as a great surprise that the lower pipes of the Megara auloi
also appear to comprise an octave when equipped with reeds whose lengths
realise the mean of the discussed plausible ranges, approximately 65 mm for
Meg2 and 50 mm for Meg1 (resulting in intervals of 1219 cents and 1186 cents,
respectively; precise octaves are predicted for 71 and 45 mm). Moreover, on
Meg1, this octave is structured into a low fifth and a high fourth by the ringfinger hole (with a perfect octave, these are predicted to measure 705 cents and
495 cents, respectively, indistinguishable from perfect consonance). The same
is, however, not true on Meg2, where the hole in question sits a quartertone
lower – so far the strongest indication that the instruments are not intended
to be identical.
A reed of corresponding length on the higher pipes also creates acceptable
octaves between their respective bass notes and thumb holes. Fascinatingly,
when the reeds are chosen so that the respective octaves on each pipe of Meg1
are perfect (45.5 and 40 mm), the coincidence between the lower pipe middle
slider hole and the higher pipe bass note also becomes exact.
32
33
34
35
36
Cf. Aristox. Harm. 1.7 (12.15 Da Rios) καταπυκνῶσαι βουλομένοις τὸ διάγραμμα.
Associated with the name of Eratocles, cf. West 1992, 126–8.
Cf. Hagel 2018, 447–50.
Hagel 2021a, 424–7.
Hagel 2014.
Greek and Roman Musical Studies 10 (2022) 15–77
48
Terzḗs and Hagel
4.7
Absolute Pitch
Starting from these plausible octaves, we may finally assess possible absolute
pitches. Based on a perfect octave between the bass and treble notes of the
lower pipe of Meg1, a frequency of 294Hz is calculated for the latter. According
to the established absolute pitch of ancient notation (I I ≈ 245Hz),37 this corresponds to the note written E E in diatonic/chromatic Lydian and Hypolydian,
but G G in the Phrygian and Dorian tónoi. This is precisely the note that, according to the reconstructions of the pre-Aristoxenian systems, typically formed
their higher boundary, representing the characteristic highest melodic pitch of
a modulating aulos – apart from the Mixolydian key in what has been called
the ‘commensurable’ system, which reached a semitone higher – and a typical
treble note of melodies from the Hellenistic period.38 That reconstruction of
an ancient paradigm, on the one hand, and the present reconstruction and
preliminary evaluation of the Megara finds, on the other, thus corroborate
each other.
4.8
Dorian and Mixolydian
In the said ancient mappings of tonal space, the octave down from G G was
universally shared between Dorian, Phrygian, Lydian and possibly other scales
whose structure eludes us (only those happen to have been transmitted in
the extant literature whose names associated them with the modes famously
mentioned in Plato’s Republic). The pyknón that we find on the lower end of
the Megara auloi would, however, only appear as part of the Dorian, but not the
Phrygian or the Lydian scale (and hardly the Hypodorian or Hypophrygian).
We have already established that the bass notes of the higher pipes stood a
semitone above those of the lower pipes. The octave extending upwards from
this note thus coincides with the Mixolydian scale of the ‘commensurable’
system. Like the Dorian octave, the Mixolydian also featured a pyknón at its
lower end. However, although the higher pipes also start with a sequence of
small intervallic steps, only the lower of these can be addressed as a quartertone, while the second one, between the two lowest slider-key holes, is larger,
equating three eighths of a tone (78 cents) on Meg2, and a full semitone
(about 106 cents) on Meg1. The sum of the two lowest intervals thus amounts
to 139 cents on Meg2, and 166 cents on Meg1. The former is practically identical with what Aristoxenus defines as the boundary between enharmonic
37
38
For the ancient pitch standard, cf. West 1992, 273–6; Hagel 2009, 68–95; Hagel 2012b; Hagel
2014; Hagel 2019, 185; Hagel 2020, 302. Precise data are only retrieved from Roman-period
sources, but all available earlier evidence is consistent with it.
Hagel 2009, 343–51.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
49
and chromatic (133–136 cents), while the same author would have accepted
the latter only as some narrow variant of the chromatic. When discussing the
practical demise of the narrow quartertone enharmonic in contemporary
music, Aristoxenus interestingly associates the ‘chromaticising’ enharmonic
with the notion of greater ‘sweetness’, obviously transferred from a culturally
agreed evaluation of the chromatic as sweet and mournful, which later musical writings still echo.39 The notion of mournfulness in turn fits the Mixolydian
exceedingly well, which was also consistently associated with such an ethos
during the period from which the Megara instruments date.40 The enharmonic, in contrast, was said to reflect typically male qualities,41 as was the
Dorian.42 It appears, therefore, that the literary sources can explain the different sizes of the pykná on the higher and lower pipes of the Megara pairs as the
reflex of a twofold harmonic dichotomy. While the austere ‘real’ enharmonic
quartertones on the lower pipes emphasised the manly character of its Dorian
scale, the near-chromatic or chromatic on the higher Mixolydian pipes underlined the sweet sorrow of a scale that Aristoxenus traced back to none other
than Sappho.
The respective passage is of the highest relevance for our subject:
Ἀριστόξενος δέ φησι Σαπφὼ πρώτην εὕρασθαι τὴν Μιξολυδιστί, παρ᾿ ἧς τοὺς
τραγῳδοποιοὺς μαθεῖν· λαβόντας γοῦν αὐτὴν συζεῦξαι τῇ Δωριστί, ἐπεὶ ἡ μὲν
τὸ μεγαλοπρεπὲς καὶ ἀξιωματικὸν ἀποδίδωσιν, ἡ δὲ τὸ παθητικόν, μέμικται δὲ
39
40
41
42
Aristox. Harm. 1.23 (29.14–30.8 Da Rios) γλυκαίνειν ‘to sweeten’; Anon. Bell. § 26 ἥδιστόν
τε καὶ γοερώτατον, ‘the sweetest and most mournful’; gloss to Aristid. Quint. 2.18 (92.25f.
Winnington-Ingram) ἥδιστόν τε καὶ γοερόν, ‘sweetest and mournful’; Theon 55.6f. Hiller
γοερώτερόν τε καὶ παθητικώτερον ἦθος, ‘most mournful and emotional character’. For the
combination of emotions cf. already Od. 22.500f. τὸν δὲ γλυκὺς ἵμερος ᾔρει / κλαυθμοῦ καὶ
στοναχῆς, ‘but he was overcome by sweet longing for weeping and groaning’.
Plat. Resp. 398e τίνες οὖν θρηνώδεις ἁρμονίαι; … Μειξολυδιστί, ἔφη, καὶ συντονολυδιστὶ καὶ
τοιαῦταί τινες, ‘So what are the mournful harmoniai? […] Mixolydian, he said, and Taut
Lydian and similar ones’, Arist. Pol. 1340b ἀλλὰ πρὸς μὲν ἐνίας ὀδυρτικωτέρως καὶ συνεστηκότως μᾶλλον, οἷον πρὸς τὴν μιξολυδιστὶ καλουμένην, ‘but to some [harmoníai the listeners
would become] more tearful and constrained, as for example to the so-called Mixolydian’;
Aristox. ap. [Plut.] Mus. 1136d καὶ ἡ Μιξολύδιος δὲ παθητική τίς ἐστι, τραγῳδίαις ἁρμόζουσα,
‘And the Mixolydian is emotional, fitting for tragedies’; Plut. Recta rat. aud. 46b “εἰ μή τις
ἦς ἀναίσθητος” εἶπε “καὶ ἀμαθής, οὐκ ἂν ἐγέλασας ἐμοῦ μιξολυδιστὶ ᾄδοντος, ‘If you wouldn’t
lack taste and education’”, [Euripides] said, ‘you wouldn’t have laughed when I sang in
Mixolydian’.
PHibeh 13.15–22 κακῶς εἰδότες ὅτι οὔτε χρῶμα δειλοὺς οὔτε ἁρμονία ἀνδρείους ποιήσειεν τοὺς
αὐτῇ χρωμένους, ‘without realising that it is not true that the chromatic would make those
who use it cowards or the enharmonic, manly’.
Plat. Resp. 399b, Arist. Pol. 1342a: ἠθικοῖς, ‘moral’.
Greek and Roman Musical Studies 10 (2022) 15–77
50
Terzḗs and Hagel
διὰ τούτων τραγῳδία. ἐν δὲ τοῖς ῾Ιστορικοῖς τοῖς Ἁρμονικοῖς Πυθοκλείδην φησὶ
τὸν αὐλητὴν εὑρετὴν αὐτῆς γεγονέναι.
[Plut.] Mus. 1136d
Aristoxenus says Sappho has invented the Mixolydian and the tragic composers learned it from her. At any rate, they took it and coupled it with the
Dorian, since the latter produces magnificence and dignity, but the former
emotions, while tragedy constitutes a blend of all that. However, in the
History of Harmonics he says that the aulete Pythocleides was its inventor.
Not only do we find the Mixolydian associated with the aulos, which is no surprise given that this was the typical instrument to accompany laments, we also
learn explicitly of its being paired with the Dorian in tragic music. The term
συζεῦξαι, normally used for the physical, for example sexual, pairing of two
objects, conveys a curiously technical flavour. Since the aulos was the instrument of drama, it is tempting to read the remark as a reference to the creation
of an instrument that played precisely these two harmoníai. If each of its two
pipes was mainly associated with one of them, as appears to have been the
case with the Megara instruments, the idea of coupling two modes gains a
much more physical notion.43 It is therefore not implausible that the quotation in ps.-Plutarch preserves an echo of Aristoxenus’ familiarity with auloi
that may have looked and functioned very much like those under scrutiny,
which, after all, may well have been produced during the author’s lifetime,
and which ended up only a good day’s march from the Lyceum.
4.9
Pyknón Shapes
While the size of the Mixolydian pyknón may be comprehensible from our
sources from the fourth century BC, and its division into two roughly equal
intervals of about a third of a tone on Meg2 is entirely inconspicuous in terms
of Aristoxenian tetrachord divisions, its internal structure on Meg1 calls for
further comment. As we have seen, the respective intervals amount to roughly
60 and 106 cents, about a third of a tone and a semitone. This is not a division
Aristoxenus acknowledges expressly, though it falls straightforwardly within
the boundaries he delineates. Once he even quotes the possible combination
of a third of a tone with a larger second interval, but (implicitly) exemplifies
43
For the idea of an aulos forming a harmony of two pipes with qualities that are opposed
on a gendered axis, cf. Poll. 4.80 καὶ τὸ μὲν γαμήλιον αὔλημα δύο ἦσαν αὐλοί, μείζων ἅτερος,
συμφωνίαν μὲν ὑποδηλοῦντες, μείζω δ’ εἶναι χρῆναι τὸν ἄνδρα, ‘And the nuptial aulos performance consisted of two pipes, one of them larger, indicating concord, but with necessary
pre-eminence of the male’.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
51
the latter with two thirds of a tone.44 Many centuries later, however, we learn
from Ptolemy that within his musical horizon any pyknón would need to consist of a smaller and a much larger interval.45 In Ptolemy’s numerical account
of their sizes, his music-mathematical axioms forced him to make the larger
almost twice as large as the smaller, so it is difficult to assess how well this
would have reflected actual musical practice;46 but there can be no doubt that
the difference as such was conspicuous at least in the Eastern Mediterranean
of the second century AD, and probably in a much wider region, since touring
star musicians would doubtless have spread musical fashions quickly. Meg1
might therefore be taken as the hitherto earliest evidence for an unequal pyknón in actual practical use, more than 400 years before Ptolemy. Its structure
is practically undistinguishable from Ptolemy’s ‘soft chromatic’, whose mathematical description in terms of the superparticular intervals 28:27 × 15:14
evaluates to 63 + 119 cents.
4.10
Realising Harmoníai
So far the proposed interpretation of primarily Dorian and Mixolydian instruments is based on a couple of octaves, a fourth and two groups of microtonal
intervals per pair, plus the perfect pitch coincidence with a theoretically reconstructed system. It is therefore time to turn to the remaining tone holes and
the instruments as a whole. Figure 3 displays the pitches of Meg1 as predicted
by modelling software,47 on the basis of effective reed lengths that are not
very different from those lengths that produced the perfect octaves discussed
above, but which establish a good compromise between various intervals.
Fortunately, this instrument preserves all tone holes except the small-finger
hole on the lower pipe, whose position needs to be estimated. This could be
done, however, without too much uncertainty on the basis of the slider key
that was operated by the same small finger, and for that purpose ran up to the
small-finger hole. Even in order only to cover its hole, the player’s finger was
already stretched to a span that is only available to trained players. Nonetheless,
it needed to reach out even further in order to push the key downwards, and
again further to catch the knob and pull it back. Clearly, there was no space to
be wasted: every millimetre more or less would have crucially encumbered or
facilitated the handling of the instrument. Thus we may safely assume that
the knob, in its upwards position, came very close to the small-finger hole,
stopping short of its rim only by about a millimetre, so as not to get in the
44
45
46
47
Aristox. Harm. 2.52 (65.4–20 Da Rios).
Ptol. Harm. 1.14 (32.23–5) and 1.15 (34.11–35.12 Düring).
For discussion of Ptolemy’s more doubtful semitones see Hagel 2009, 217–19.
Cf. Hagel 2021b.
Greek and Roman Musical Studies 10 (2022) 15–77
52
Terzḗs and Hagel
figure 3
Predicted pitches of Meg1, in comparison with the ‘commensurable’ trópoi as
reconstructed in Hagel 2000
way of the finger when it uncovered and covered its hole during normal playing. From the position of the key hole, the length of the rod and the length of
the intervening sections, which can also be established within very small margins, the probable position of the small-finger hole can thus be retrieved. In the
figure, its centre is placed 44 mm above the lower end of the missing section,
which in turn must have been about 71.1 mm long. The diameter of the hole
is set to 9.5 × 9.8 mm, by extrapolating the minute difference between the
middle-finger and ring-finger holes.
As can be gleaned from Figure 3, the hole thus plays a pitch about 525 cents
above the bass note, an eighth of a tone higher than the fourth we would
expect at this place because it would complement the internal structure of
the pipe’s octave to the fundamental tetrad that underlay the central octave
of theoretical systems no less than all known ancient lyre tunings, and which
had been recognised as implementing two mathematical means long before the
Megara instruments were interred.48 Conceivably the required lower position
for that fingerhole in order for it to play a true fourth would have exceeded the
physical capabilities of the player; in this case, one would have needed to correct the pitch during playing. Alternatively, the lost hole might also have been
made slightly smaller; with a diameter of 8.5 mm, for instance, the divergence
shrinks to 17 cents. At any rate, in the figure the corresponding pitch is distinguished by a fainter line, so that it is not mistaken for part of the evidence.
48
Lyre tunings: Hagel 2009, 133f.; means: cf. Arist. ap. [Plut.] Mus. 1139b.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
53
On the right hand of the figure, the lines indicating the pitches of all tone
holes are overlaid with the reconstruction of the ‘commensurable’ system as
it was published more than twenty years ago.49 The black circles there indicate relative pitches that are attested in Aristides Quintilianus, our only source
for the structure of some harmoníai of the Classical Period, presumably taken
from a lost work of Aristoxenus. The lines without such circles represent mere
educated guesses about how the Hypophrygian and Hypodorian might have
fitted into the picture, two scales whose positions in the system we learn from
Aristoxenus, without ever being told about their structure.
We see how the bass notes of the lower pipe coincide nicely with the pyknón at the lower end of the Dorian octave. Below, Aristides’ Dorian expands
the ‘Dorian octave’ by another tone, which is absent on the pipes. This may
confirm earlier speculation that the extra bass note had its original place as the
hyperypátē of lyre music.50
Next would come mésē P P, which would supposedly have been sounded
from the problematic lost fingerhole, as we have just discussed. The following
paramésē M M can be played on both pipes. It forms the basis of another pyknón, which must be realised by half-covering the next higher fingerhole. Being
spaced in the traditional manner at a distance of about three quartertones, this
hole supported not only a narrow enharmonic but also wider variants up to a
soft chromatic. The index-finger hole, finally, provided the treble note nḗtē. The
thumb hole below it could supply a diatonic paranḗtē, if required (Aristides’
harmoníai come only in enharmonic guise, even though the age and the primary role of the diatonic were recognised among ancient theorists).51
In the lower tetrachord, at any rate, only a very soft version of diatonic might
have been available, using the upper slider hole. Interestingly, Aristoxenus, if
taken at face value, testifies to a universal use of precisely such a soft diatonic
variant by the adherents of a modern musical style:
χρώμενοι γὰρ αὐτοὶ τοιαύταις τετραχόρδων μάλιστα φαίνονται διαιρέσεσιν, ἐν
αἷς τὰ πολλὰ τῶν διαστημάτων ἤτοι περιττά ἐστιν ἢ ἄλογα· μαλάττουσι γὰρ
αἰεὶ τάς τε λιχανοὺς καὶ τὰς παρανήτας.
[Plut.] Mus. 1145cd
They themselves seem to use precisely those types of tetrachordal divisions in which the majority of the intervals is either odd [i.e. comprise
49
50
51
Hagel 2000, 172, Abb. 23.
Hagel 2009, 408–11.
Cf. e.g. [Plut.] Mus. 1134f–1135c for a pre-Aristoxenian speculative derivation of enharmonic from diatonic; in general, cf. Franklin 2002.
Greek and Roman Musical Studies 10 (2022) 15–77
54
Terzḗs and Hagel
an odd number of quartertones] or irrational [i.e. do not form a multiple
of quartertones at all], since they soften the likhanoí and the paranêtai
all the time.
The notion of softening the notes in question, which are the higher ‘moving’
notes in the tetrachord, in a way that results in odd multiples of quartertones
can only apply to diatonic or chromatic intervals. The ensuing soft chromatic
features a pyknón of about 150 cents, like we find on the higher Megara pipes.
The higher ‘moving’ note of a respective soft diatonic tetrachord, in turn,
would lie about 250 cents above its bottom note, which coincides with the
highest slider note on the lower pipe. Is it conceivable that the Megara auloi
testify in such detail to the material background of the music that Aristoxenus
had perceived as modern? The combination of some of these apparently fashionable traits, on the one hand, with some old-style quartertone enharmonic,
on the other, might locate them within a cultural environment that embraced
innovation without necessarily abolishing all characteristics of an earlier
style wholesale.
The higher pipe also starts with a pyknón, albeit of the wider, chromaticising type. The next hole, operated by the highest slider key, plays S S, precisely
the irregular ‘diatonic’ note that characterises the lower part of the Mixolydian
scale. There, another pyknón would now follow, the lowest note of which is,
however, beyond the possible reach of the small finger. As a consequence, this
pyknón would only have been playable starting from its highest note, reaching its ‘fixed’ bottom component by the precarious technique of half-covering.
Here, more than anywhere else on our instruments, Plato’s remarks would
seem perfectly appropriate:
οὐκοῦν μεστὴ μέν που μουσικὴ πρῶτον, τὸ σύμφωνον ἁρμόττουσα οὐ μέτρῳ
ἀλλὰ μελέτης στοχασμῷ, καὶ σύμπασα αὐτῆς αὐλητική, τὸ μέτρον ἑκάστης
χορδῆς τῷ στοχάζεσθαι φερομένης θηρεύουσα, ὥστε πολὺ μεμειγμένον ἔχειν τὸ
μὴ σαφές, σμικρὸν δὲ τὸ βέβαιον.
Plat. Phlb. 56a
Isn’t music, firstly, somehow full of that, as it tunes the consonances not
with the help of some measure but by practical aiming, and entirely the
art of the aulos, whose hunt for the right measure of each note amounts
to aiming at a moving target, so that it has a great deal of uncertainty
attached to it, but little that is definite.
In Aristides’ version, the top half of the Mixolydian consists merely of an empty
interval of a tritone, which leads directly to the thumbhole of our pipe. At the
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
55
period in question, however, theorists had long realised that the structurally decisive note, mésē, from which its relation with other scales would be determined,
would sit within that empty interval, one whole tone below the treble note. On
our pipe, this Mixolydian mésē Z Z is also available, as the middle-finger hole.
As we have seen, both the Dorian and the Mixolydian harmonía appear
indeed provided for, as far as this was possible given the physical restriction of
the human hand. There is perhaps one shortcoming. Being played downwards
from the available fingerhole, the Mixolydian central pyknón would apparently
need to remain closer to a quartertone enharmonic than its lower pyknón is.
The lower pitch of the two produced by partial hole-covering would belong
to the category of ‘fixed’ notes and ought to stand a perfect fifth below the treble note of the scale, at least in theory. The size of the central pyknón between
that note and the one produced from the open fingerhole would thus be restricted
to about 130 cents. Without any information about how the Mixolydian used to
be accompanied on the other pipe, we cannot, however, know whether the
possible melodic fifth needed to be ensured harmonically, or whether the apparent restriction might have been less relevant in practice, and a larger pyknón
might therefore have been practicable by bending its lowest note below its
theoretical position.
Above the highest Mixolydian note, the instrument still provides a higher
hole, belonging to the index finger of the right hand. When fully opened, it
would have sounded a whole tone higher than its lower neighbour, and 290 cents
higher than the treble note of the left-hand pipe. This hole must typically have
been used for accompanying various notes from the left-hand pipe, presumably by modifying its pitch in various ways by partial closing or manipulation
of the reed.52 We will leave further exploration of this point to future experimental players.
4.11
More Harmoníai
The modulating capabilities of Meg1 are not exhausted with Dorian and
Mixolydian. Using pitches from the two pipes alternately, all notes of the
(Syntono-)Lydian harmonía are accessible. The bounding notes of its pyknón
reuse the holes for S S and P P; the mesopyknon in between can be produced
by partial covering of the latter. Its mésē I I is provided by the right ring finger, and its highest note is identical with the Dorian nḗtē by definition. Using
52
For notes belonging to the accompaniment (kroûsis) lying above the perceived melody
(mélos) in a certain style of aulos music, see Aristox. ap. [Plut.] Mus. 1137b–d; cf. Hagel
2009, 407–11. In modern literature it is sometimes asserted that in ancient Greek music
the accompaniment was invariably above the melody, which may be an unwarranted
generalisation.
Greek and Roman Musical Studies 10 (2022) 15–77
56
Terzḗs and Hagel
the thumb hole of the left-hand pipe, one could further create another pyknón
above I I, which would modulate to Phrygian and presumably Hypophrygian;
the Phrygian mésē is in turn supplied by Dorian paramésē M M. The lower
Hypophrygian pyknón might have been identical with the Lydian. However, at
first glance, no provision seems to be made for the lowest part of the Phrygian
scale. Only the middle note of its pyknón can be played from slider key holes;
in order to access its lowest note, one would have to shut one of the respective sliders partially, while it cannot easily be seen how the highest might be
provided at all. However, if the instrument, as we had reason to speculate,
reflected the design of a theatrical aulos suitable for tragedy, we might expect
some support for both Hypophrygian, which may have been employed for solo
arias, and for Phrygian, which was allegedly used already by Sophocles.53
We have so far ignored a special constructional conundrum related to the
‘intermediate’ slider key of the lower pipe of Meg1 (I3). While its plate is comparatively short, measuring no more than 14 mm, the range within which it was
designed to move is unusually large: its slot length and the difference between
the length of the guide and the constricted part that moved in it agree on a
motion range of 16.8–17 mm. This seems to make no sense; in order to close its
hole most efficiently, the plate ought to travel no more than the total length
of the hole diameter plus half the difference between this diameter and the
length of the plate, which in this case would amount to 11.35 mm. But here,
assuming that plate end and hole rim just coincide when the slider key is
in the open position, pushing the slider downwards the whole way moves
the plate beyond the position where it covers its hole. As a consequence,
when the plate has reached the lower end of its path, a 3 mm-wide segment
of the hole has re-emerged behind it. On the surface, the problem might be
mended by moving the initial position of the plate from the hole edge by about
5 mm. But this has the consequence that the slider needs to travel an extra
half centimetre across the closed tube wall without achieving anything, which
is not only entirely pointless but positively annoying: bent over the pipe in
order to reach the upper part of the key, the player’s middle finger is considerably restricted regarding its lateral movement and will find it rather difficult to
switch the key between its extreme states in one swift motion. Since nothing
would have impeded the makers from manufacturing this key precisely as all
the others, whose proportions appear technically perfected, we must infer that
this singular design fulfilled a special function.
It is difficult to see what such a function might have been other than precisely the effect described above, pushing the plate beyond the hole. The
slider would in this way acquire three possible states: open (knob in highest
53
Hypophrygian: [Arist.] Pr. 19.48; Phrygian: Aristox. ap. Vit. Soph. 23.
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
57
position), closed (knob pushed downwards as far as on a normal slider key),
and partially open (knob pushed downwards completely). From the available
dimensions it is fortunately possible to predict the effect of the third state with
some precision.54 The resulting opening has the shape of an ellipse segment
with an area of somewhat less than 17 mm2. Part of this opening is, however,
obstructed by the rod that extends across it at a slight elevation, leaving an
effective opening that can be estimated to about 11–16 mm2. It functions as
a small tone hole, producing a significantly lower pitch than the same hole
does in its fully open state. The predicted configuration for an effective opening of 13.5 mm2 is shown in Figure 4; it does not differ much across the possible
range. It emerges that the hole fills the gap in what is now a row of five quartertones in the bass region, providing the missing Phrygian hypátē F F (a fourth
below its mésē M M), which probably also formed the Hypophrygian mésē. This
would complete the Hypophrygian harmonía, as far as we are able to guess its
structure. The low part of the Phrygian, however, would not be available in the
enharmonic genus without resorting to the much more difficult trick of halfshutting the highest slider hole on the high pipe. Otherwise, only a chromatic
pyknón is available on the high pipe, once more of the ‘Ptolemaic’ soft variant
(64 + 131 cent), as well as a diatonic with a very small ‘semitone’, if the higher
‘moving’ note is obtained from the low pipe. This ‘semitone’ is identical with
that of Ptolemy’s standard diatonic, where it is described as the ratio of 28:27
figure 4
54
Predicted pitches of Meg1 with slider key I3 in partially-shut state over hole βS1
Notably, the strict requirements of this slider key also contribute to determining the
length of the lost pipe section within very small margins.
Greek and Roman Musical Studies 10 (2022) 15–77
58
Terzḗs and Hagel
(63 cents), which he also quotes from the much earlier account of Archytas, and
which also resembles the ‘third of a tone’ that Aristoxenus acknowledges as a
possible lowest diatonic interval.55 The virtually missing enharmonic Phrygian
may in turn be related to another statement of Aristoxenus, where he asserts
that the enharmonic is suited for the Dorian harmonía, and the diatonic for
the Phrygian.56 Once more, a seemingly impressionistic tenet of ancient music
theory might thus assume material cloth.
4.12
Notation
The melodic range of the pair (i.e. excluding the highest note, which was very
probably used only in the accompaniment) therefore extended from the highest note of the Mixolydian octave harmonía to the lowest of the Dorian. In
terms of ancient notation, the Mixolydian of the ‘commensurable’ system
was Aristoxenus’ ‘Higher Mixolydian’, being situated a semitone higher than
the Mixolydian of the rivalling scheme. Much later, in the revised nomenclature of the fifteen-tónoi system known from the Roman period, it would
appear under the name of ‘Hyperiastian’.57 The note signs, however, remained
the same. In the older, ‘instrumental’ notation, the common highest note of the
Dorian, Phrygian and Lydian harmoníai, as far as they are present on our instrument, was written as G, presumably abbreviating ‘nḗtē’ as the traditional term
for the treble pitch. Here the highest Mixolydian note would require a secondary sign, of the form A; apparently the ‘instrumental’ notation was older than
the design of an instrument with a higher variant of this scale. In contrast, the
younger ‘vocal’ notation appears perfectly suited for such an instrument.
Here the Ionian alphabet was distributed precisely across its melodic range.
The Mixolydian treble note thus appears as A, and the Dorian bass note as
W.58 If the Megara items are indeed the remains of professional instruments
55
56
57
58
Aristox. Harm. 1.27 (35.3–7 Da Rios).
Aristox. ap. Clem. Strom. 6.88.1.
Cf. e.g. Hagel 2020, 299, Figure 21.1.
Regarding a terminus ante quem for this innovation that would predate the Megara finds,
a much discussed passage in Aristotle’s Metaphysics (1093b) may be relevant. There he
exemplifies numerical pseudo-philosophy by the notions ὅτι ἴσον τὸ διάστημα ἔν τε τοῖς
γράμμασιν ἀπὸ τοῦ Α πρὸς τὸ Ω, καὶ ἀπὸ τοῦ βόμβυκος ἐπὶ τὴν ὀξυτάτην νεάτην ἐν αὐλοῖς, ἧς ὁ
ἀριθμὸς ἴσος τῇ οὐλομελείᾳ τοῦ οὐρανοῦ, ‘that the distance is the same in the letters between
Α and Ω, and from the bómbyx to the highest nḗtē in auloi, the number of which [nḗtē]
is equal to the totality of the heavens’. If this refers to the association of ancient notation
with an instrument design similar to the Megara auloi, it would reduce the alleged arguments in Hagel (2005) to nothing more than the excessive, though deplorably baseless,
speculation for which that author is all too well known. However, two problems remain:
firstly, the notes are named in reverse sequence in respect to the letters (perhaps Aristotle
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
59
as were used in prominent civic events, for which the most celebrated poetcomposers would provide the music, it would be no wonder to find crucial
innovations regarding the notation of melodies associated with their design.59
The notes emitted from the index-finger hole of the higher pipe, by the way,
would not necessarily require corresponding notation, if this hole played
only what was perceived as the accompaniment, which was indeed never
written down.
4.13
Further Differences
What about Meg2, whose predicted pitches are shown in Figure 5? Here, as
well, the placement of the left-hand small-finger hole needs to be estimated,
and the positions of the hole for the ring finger of the right hand cannot be
established with precision either, while the exact diameter of the small-finger
hole is also difficult to assess.60 Notably, an optimal configuration requires
assuming a comparatively large divergence between the effective lengths of
the reeds. In the diagram, it is set to 4 mm (for Meg1, 2 mm proved sufficient).
In many respects Meg2 appears to have functioned much like Meg1. Apart
from the aforementioned differences in the size of the Mixolydian pyknón,
the following divergences must, however, be noted. Most importantly, the
ring finger hole of the lower pipe appears too low. The resulting interval of
only 166 cents below it is clearly too small for the expected Dorian disjunctive whole tone (even if minor adjustments of fixed notes, as mentioned by
Aristoxenus,61 are granted). The hole in question might instead be suited for
supporting a Hypodorian pyknón. The Dorian paramésē M M, in turn, would
need to be supplied from the right-hand pipe.
Secondly, the highest slider-key hole on the higher pipe appears to sit too
high for either the Lydian bass note or an irregular Mixolydian note at the
‘expected’ position. At any rate, the difference to Meg1 is so eye-catching that
it must be intentional, unless we assume a gross production error. Whether it
represented a different flavour of Mixolydian (while disregarding the Lydian
option), must for the present remain a matter of speculation. At any rate, the
59
60
61
was not really acquainted with musical notation?); secondly, how does a treble note
notated A possibly reflect some number of cosmological significance?
On might wonder whether the following facts are somehow related: (a) the Megara instruments implement no strictly enharmonic pyknón in the bass region of the Mixolydian and
(b) the same pyknón cannot be written as a sign triplet in the ‘instrumental’ notation.
The figure is based on the following assumptions: Meg2αFii, small-finger hole: 39.5 mm
above rim, 9.5 mm effective diameter; Meg2βFi, ring-finger hole: 6.5 mm above rim,
10 mm effective diameter; Meg2βFii, small-finger hole: diameter 10.7 × 10 mm.
Aristox. ap. [Plut.] Mus. 1145d; cf. Hagel 2009, 140–2.
Greek and Roman Musical Studies 10 (2022) 15–77
60
figure 5
Terzḗs and Hagel
Predicted pitches of Meg2, in comparison with the ‘commensurable’ trópoi as
reconstructed in Hagel 2000
theoretical optimal size of the central Mixolydian pyknón, which follows from
the requirement of a perfect fifth between its base note and the highest note
of the scale, precisely matches that on Meg1.
The treble note of the whole instrument is significantly higher. Fully opened,
it produces a good fourth with the thumb hole of the left-hand pipe, and a
good fifth with its middle-finger hole, crucially augmenting the harmonic
capabilities. The required increased finger span, however, may not have been
accessible to the owner of Meg1.
Finally, Meg2 has no three-state slider key for the Phrygian F F, but a perfectly normal one. To achieve the same effect, one would need to push it to just
the right amount of ‘almost shut’, hitting the correct position within a millimetre or less without any guidance – at least none that has left detectable traces,
as far as we see. The advantages of the fancy solution realised on Meg1, which
sets the important pitch with precision, stand out all the more clearly. Here
the motional guesswork concerned the closed state, which required much less
precision: with a slider plate of 14 mm length and a hole of merely 8.7 mm
diameter, any position within a range of about 5 mm would have served
the purpose.
4.14
Conclusions
Our findings, in combination with literary sources going back to the period in
question, suggest that the two doublepipe instruments held in the Museum
of Megara represented professional instruments of the highest cultural level,
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
61
probably similar to those accompanying theatrical performances in the late
Classical period. They share traits of a standardised design, adhere to a standard
pitch, which would remain stable for some centuries, but also exhibit small
idiosyncrasies that probably reflect both the musical tastes and the physical
capabilities of the artists who had originally commissioned them. Their tonality is informed by theoretical efforts that are reflected in the literature. The
reconstruction of the respective scalar systems in recent years has predicted
many aspects of their design; as far as we know, these are the first published
instruments that confirm those predictions. Unlike the known simpler auloi
from earlier periods as well as those from the Roman era with an altogether
different kind of metal mechanism, the pairs from Megara are above all characterised by a microtonal mismatch between their rows of fingerholes, which
probably called for a stupendous degree of virtuosity, complementing the primary pitches, which were built right into the instrument and can therefore
be evaluated, with fitting intervals from the other pipe that almost certainly
involved a good deal of pitch-bending.62
Acknowledgements
We would like to express our sincere thanks to Panagiṓta Avgerinoú, director of the Archaeological Museum of Megara, to the Ephorate of Antiquities
of West Attica for granting permission to examine and publish the finds, and
especially to Geōrgίa Theodṓrou, conservator of the Ephorate, for the invaluable information she has so kindly and generously shared on the restoration
process of the aulos Meg2. Two expert reviewers invested substantial energy to
supply a treasure trove of suggestions; our gratitude to both must not eclipse
the fact that we can only hope to satisfy one of them fully.
Parts of this publication are based on research funded by the Austrian
Science Fund (FWF) through grant P32816-G, and by the European Research
Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 787522), respectively. The views
presented here, however, reflect only those of the authors; neither the ERCEA
nor the FWF are responsible for any use that may be made of the information
contained.
62
As one of the reviewers points out, such a procedure is familiar even to players of to modern Western reed instruments, in environments where they need to achieve pure thirds
or a dominant seventh that reflects the harmonic series – the latter involving a deviation
from the instrument’s designed pitch by about a sixth of a tone.
Greek and Roman Musical Studies 10 (2022) 15–77
62
Terzḗs and Hagel
Bibliography
Avgerinou, P. (forthcoming). The Auloi of the Megara Museum and Their Archaeological Context. In: S. Psaroudakēs and S. Perrot, eds, Transactions of the 9th Moisa
Conference, 11–14 July 2016, Athens, Athens.
Bodley, N.B. (1946). The Auloi of Meroë: A Study of the Greek-Egyptian Auloi Found at
Meroë, Egypt. AJA 50, pp. 217–240.
Byrne, M. (2002). Understanding the Aulos II. Extended Pipes and Drone. In:
E. Hickmann, R. Eichmann, A.D. Kilmer, eds, The Archaeology of Sound: Origin and
Organisation, Rahden, Westf.: Marie Leidorf, pp. 367–373.
Conze, A. (1903). Die Kleinfunde aus Pergamon. Berlin: Königl. Akademie der
Wissenschaften.
Franklin, J.C. (2002). Diatonic Music in Greece: A Reassessment of Its Antiquity.
Mnemosyne 55, pp. 669–702.
Gänsicke, S., and Hagel, S. (2017). The Auloi from Meroë: Preliminary Notes on the
Conservation, Technical Examination, and Interpretation of a Cache of Ancient
Musical Instruments. In: J.M. Daehner, K. Lapatin, A. Spinelli, eds, Artistry in Bronze:
The Greeks and Their Legacy, XIXth International Congress on Ancient Bronzes,
Los Angeles: Paul Getty Museum, pp. 381–388.
Hagel, S. (2000). Modulation in altgriechischer Musik: Antike Melodien im Licht antiker
Musiktheorie. Frankfurt a.M.: Lang.
Hagel, S. (2005). Twenty-four in Auloi. Aristotle, Met. 1093b, the Harmony of the
Spheres, and the Formation of the Perfect System. In: S. Hagel and C. Harrauer,
eds, Ancient Greek Music in Performance, Wien: Österreichische Akademie der
Wissenschaften, pp. 51–91.
Hagel, S. (2008). Re-evaluating the Pompeii Auloi. JHS 128, pp. 52–71.
Hagel, S. (2009). Ancient Greek Music: A New Technical History. Cambridge: Cambridge
University Press.
Hagel, S. (2010). Understanding the Aulos Berlin Egyptian Museum 12461/12462. In:
R. Eichmann, E. Hickmann, L.-Ch. Koch, eds, Musical Perceptions – Past and Present.
On Ethnographic Analogy in Music Archaeology, Rahden, Westf.: Marie Leidorf,
pp. 67–87.
Hagel, S. (2012a). The Aulos Syrinx. Rudiae 22–23, pp. 491–518.
Hagel, S. (2012b). The Pompeii Auloi: Improved Data and a Hitherto Unknown Mechanism. In: R. Eichmann, F. Jianjun, L.-Ch. Koch, eds, Sound from the Past. The
Interpretation of Musical Artifacts in an Archaeological Context, Rahden, Westf.:
Marie Leidorf, pp. 103–114.
Hagel, S. (2014). Better Understanding the Louvre Aulos. In: R. Eichmann, F. Jianjun,
L.-Ch. Koch, eds, Papers from the 8th Symposium of the International Study Group on
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
63
Music Archaeology in Suzhou and Beijing, China, 20–25 October, 2012, Rahden, Westf.:
Marie Leidorf, pp. 131–142.
Hagel, S. (2018). Music and Harmonic Theory. In: P.T. Kayser and J. Scarborough, eds,
Oxford Handbook of Science and Medicine in the Classical World, Oxford: Oxford
University Press, pp. 445–461.
Hagel, S. (2019). Reconstructing the Auloi from Queen Amanishakheto’s Pyramid. In:
R. Eichmann, L.-Ch. Koch, F. Jianjun, eds, Music Archaeology from the Perspective of
Anthropology, Rahden: Marie Leidorf, pp. 177–197.
Hagel, S. (2020). Notation. In: T.A.C. Lynch and E. Rocconi, eds, A Companion to Ancient
Greek and Roman Music, Hoboken: Wiley, pp. 297–310.
Hagel, S. (2021a). Understanding Early Auloi: Instruments from Paestum, Pydna and
Elsewhere. In: A. Meriani and G. Zuchtriegel, eds, La tomba del Tuffatore: rito, arte
e poesia a Paestum e nel Mediterraneo d’epoca tardo-arcaica, Pisa: ETS, pp. 421–459.
Hagel, S. (2021b). Assessing Unknown Parameters of Instrument Finds by Writing Software. Archeologia e Calcolatori 32.1, pp. 403–421. DOI: 10.19282/ac.32.1.2021.22.
Howard, A.A. (1893). The Αὐλός or Tibia. HSPh 4, pp. 1–60.
Kostoglou, K. (1970). Ανασκαφικαὶ Εργασίαι. Καρυῶτες. Αρχαιολογικὸν Δελτίον, 25
(Χρονικά), pp. 329–332.
Litvinsky, B.A. (1999) греческие флейты (авлосы) в глубинной азии. Greek Flutes
(Auloi) in Central Asia. In: J. Duchesne-Guillemin, ed., Monumentum Marcelle
Duchesne-Guillemin, Leuven: Peeters, pp. 517–543.
Litvinsky, B.A. (2010) Храм Окса в Бактрии, 3: Искусство, Художественное ремесло,
Музыкальньие инструменты (The Temple of Oxus in Bactria. 3: Art, Fine Art,
Musical Instruments). Moscow: Vostochnaya Literatura.
Psaroudakēs, S. (2008). The Auloi of Pydna. In: A.A. Both, R. Eichmann, E. Hickmann,
eds, Challenges and Objectives in Music Archaeology, Rahden, Westf.: Marie Leidorf,
pp. 197–216.
Sutkowska, O. (2015). Archäologische Auloi/Tibia Funde mit Mechanismen. Tonkunst 9,
pp. 412–422.
West, M.L. (1992). Ancient Greek Music. Oxford: Clarendon Press.
Zorides, P. (1980). Μέγαρα. Αρχαιολογικὸν Δελτίον, 35 (Χρονικά), pp. 41–55.
Greek and Roman Musical Studies 10 (2022) 15–77
64
Terzḗs and Hagel
table 1
Megara auloi; bone segment dimensions
Bone
section
effective
length
(mm)
Meg1αM
78.2
socket depth /
⌀ (mm)
spigot
length / ⌀
(mm)
⌀ (int)
(mm)
(10)/(12)
14/(14.5)
11.7
Meg1αX
Meg1αFi
Meg1αFii
Meg1αS
62.8
141.0
54.15
133.8 + 3.5
14.2/(14.5)
15.5/(14.6)
13.5/(14.5)
18.1 + 3.5/(14.5)
14.8/(14.5)
13.5/(14.5)
17.5/(14.5)
–
(11.7)
(11.7)
11.7
11.7
Meg1βMi
Meg1βMii
Meg1βXi
Meg1βXii
Meg1βFi
Meg1βFii
Meg1βSi
Meg1βSii
22.7
68.7
94.5
20.4
126.6
71.1
42.8
68.8
19.6/(13.3)
(13)/(14.2)
15.9/(14.8)
12.5/(14.5)
13/(14.5)
14.1/(14.5)
13.2/(14.2)
18.2/(14.2)
13.3/(14.2)
15.8/(14.4)
12.5/(14.5)
12.7/(14.5)
14.1/(14.5)
13.2/(14.2)
16.7/(14.2)
–
11.7
11.7
11.7
11.6
11.7
11.7
11.7
11.7
Meg2αM
Meg2αX
Meg2αFi
Meg2αFii
Meg2αSi
(81.35)
(47.5)
(153.15)
65.3
52.8
16.85/15.2
18.3/15.3
(17)/(15.5)
(18.5)/(15.5)
–
11.7
11.7
(11.5)
(11.5)
(11.5)
Meg2αSii
63.0
(10.5)/(13)
(>15)/(15)
(>17.5)/(15.5)
(17)/(15.5)
18.5/(15.5)
14(15.5)
–
14/(15.5)
Meg2βM
Meg2βX
Meg2βFi
Meg2βFii
Meg2βS
(77.5)
82.7
149.0
59.2
134.2
(8)/(12.5)
(17)/15.5
(18)/(15.5)
17/(15.5)
17.5/(15.5)
(17)/15.5
17.8/15.5
17/(15.5)
17.5/(15.5)
–
(11.5)
13.1 × 12.8
(11.5)
(11.5)
<(11.7)
(11.5)
11.5
12.5
⌀ (ext)
max/min
↔ × ↕ (mm)
upper
rim
lower
rim
×
✓
(18) × (17)
(19) × (17.5)
(18.5) × (17.5)
(18) × (17.5)
(20.5) × (16.5)
18.3/15.8
19.4/14.25
(18) × (17)
17.5
(19) × (17.5)
(18.5) × (17.5)
18.2
17.5
(20.1) × (18)
19.25/14.5
18.0
18.3
18.3
18.2
✓
✓
✓
?
×
✓
×
✓
✓
✓
✓
✓
✓
✓
✓
×
×
✓
?
✓
?
✓
×
×
×
✓
✓
✓
✓
✓
✓
✓
18.2
21 × 16.7
19.25/14.5
18.2
(18)
(18)
(18) × (17.5)
(19) × (17.5)
✓
✓
×
✓
✓
✓
✓
✓
×
✓
✓
✓
18.5/14.0
In brackets: Approximate values; in bold: proposed dimensions
Greek and Roman Musical Studies 10 (2022) 15–77
65
Two Auloi from Megara
table 2
Aulos Meg1: side hole/pinhole centre distances from segment upper rims
⌀↔×↕
(mm)
azimuth
(degs)
transveral
niches,
lower pair
transveral
niches,
upper pair
<1
2.2
9.8 × 10.2
9.8 × 10.0
9.75 × 10.0
9.65 × 9.8
2.2
2.2
0
180
20
180
0
−5
−105
180
n/a
×
×
×
×
×
×
×
n/a
×
×
×
×
×
×
×
9.75 × 10.0
2.2
25
25
×
×
×
×
2.2
25
×
×
n/a
(60) + 3.5
61.3 + 3.5
n/a
×
×
×
×
×
×
×
×
n/a
n/a
49.5 + 3.5
10.7
×
×
×
×
×
×
×
×
×
×
n/a
n/a
n/a
×
×
n/a
−2.4
−11.5
Bone
section
Hole
distance from
upper end (mm)
Meg1αM
Meg1αFi
»
»
»
»
»
»
s.h.
p.s.h.
f.1
f.2
f.3
f.4
p.1
p.3
Meg1αFii
»
f.5
p.2
Meg1αS
p.2
»
»
»
Meg1βMi
Meg1βXii
Meg1βFi
»
»
»
»
»
»
s.1
s.2
s.3
s.h.
p.s.
p.s.
f.1
f.2
f.3
f.4
p.1
p.3
Meg1βFii
»
Meg1βSi
Meg1βSii
»
f.5
p.2
s.1
s.2
s.3
13.7
26.45
32.7
63.6
93.13
129.28
123.4
130.5 < d < 138
130.7
23.23
56.6 < d < 57.5
<57.05>
(2.5) < d < (3.4)
2.9
42.8 + 3.5
78.3 + 3.5
110.1 + 3.5
(9.5)
22.9
(2.5)
21.75
51.7
84.95
115.55
112.4
117.7 < d < 123
120.6
27.1
63.0
25.95
30.15
49.75
s.h.:
p.:
f.:
s.:
in brackets:
sýrinx hole
pin hole
finger hole
slider hole
values drawn on indirect measurements; in bold: proposed positions
9.2
9.2
9.2
(<1)
2.2
2.2
9.9 × 9.8
9.8 × 10.0
9.9 × 10.2
9.7 × 10.0
2.2
2.2
−105
25
180
0
180
180
−20
180
0
5
105
180
9.5×9.8
2.2
8.7 × 8.8
9.5
9.5
−25
−25
105
−25
180
Greek and Roman Musical Studies 10 (2022) 15–77
66
table 3
Terzḗs and Hagel
Aulos Meg2: side hole/pinhole centre distances from segment upper rims
Bone
section
Hole
distance
from upper
end (mm)
⌀↔×↕
(mm)
Meg2αFi
»
»
»
»
»
Meg2αFii
»
Meg2αSi
»
Meg2αSii
»
Meg2βFi
»
»
»
»
»
Meg2βFii
»
Meg2βS
»
»
»
»
f.1
f.2
f.3
f.4
p.1
p.3
f.5
p.2
p.2
s.1
s.2
s.3
f.1
f.2
f.3
f.4
p.1
p.3
f.5
p.2
p.2
s.1
s.2
s.3
p.x
41.88
78.15
115.1
145.1
133.15
146.65
26.0
67.5
2.2
28.05
22.8
40.5
37.05
72.25
108.34
142.5
132.0
140 < d < 146
26.05
66.2
7.0
33.9
84.35
110.15
127.2
9.55 × 10.6
10.2 × 10.3
10.2 × 11
9.6 × 10
2
2
9.6 × 10
2
2
9.5
9.0
9.0
10.3 × 10
9.3 × 10.6
10.4 × 9.3
10 × 10
2
2
10.7 × 10
2
2
10.6
9.3
9.5
2
p.:
f.:
s.:
In bold:
azimuth
(degs)
−20
180
0
5
90
180
−10
0
0
90
0
180
20
180
0
−5
−90
180
10
0
0
−90
0
180
−35
transveral
niches,
lower pair
transveral
niches,
upper pair
×
×
×
×
×
×
×
×
×
×
×
×
×
×
×
×
×
×
7
2
4
×
×
×
×
×
×
×
×
×
12
66
n/a
×
−7
−8.5
−10
×
×
×
×
×
×
×
×
×
n/a
54
n/a
×
pin hole
finger hole
slider hole
proposed hole positions
Greek and Roman Musical Studies 10 (2022) 15–77
67
Two Auloi from Megara
table 4
Megara slider key dimensions
Slider
O.L.
SlU – E
SlL – E
CL – E
Sl.L
C.L.
G.L.
Pl.L.
Meg1L1
Meg1L2
Meg1I1
Meg1I2
Meg1I3
Meg1S1
Meg1S2
Meg1Sy1
Meg1Sy2
Meg2αI
Meg2αS
Meg2αL
Meg2βI
Meg2βS
Meg2βL
207.8
198.5
145.5
150.0
140.5
102.5
106.5
187.6
211.2
143.5/142.5
106/103
201/195.5
140.5/136.5
105.0
206
188.3
178.5
124.5
131.5
124.0
89.0
90.0
171.1
195.7
124.5/123.5
85.0/82.0
179.0/174.5
124.5/119.5
87.5
183.5
171.3
162.5
109.5
115.5
105.0
73.0
75.0
164.5
186.7
108.5/107.5
71.0/68.0
162.5/158
108.5/103.5
71.5
167
42.5
39.2
35.5
33.0
17.5
17.5
18.0
= Pl.L
= Pl.L
28
16.5
30.0
= Pl.L
= Pl.L
33.5
14.8
13.8
12.8
13.8
16.8
13.8
12.8
4.8
6.8
14.0
13.0
15.5
14.0
13.5
15.0
25.8
28.3
27
24.5
25
26.3
22.5
14
13.2
18
23
28
26
24.5
36
11.8
14.5
(14.8)
11
8
12.5
(9.7)
(9.2)
(6.4)
14
10
14
12
11
22
16.3
16
16
16.5
14
14
15.5
9.3
9.5
16
15.5
17
16
16
17
O.L.:
SlU – E:
SlL – E:
CL – E:
Sl.L:
C.L.:
G.L.:
Pl.L.:
in bold:
Overall length
slot internal upper edge – plate end
slot internal lower edge – plate end
rod constriction lower end – plate end
slot effective length
rod constriction length
roof-shaped guide length
plate length
corrected dimensions
Greek and Roman Musical Studies 10 (2022) 15–77
68
Terzḗs and Hagel
table 5
Megara Auloi: adjusting sliders to their side holes (mm)
Slider
hole
Pin lower edge –
hole upper rim
± from Pin upper edge –
SlL – E hole lower rim
Meg1αS1
Meg1αS2
Meg1αS3
Meg1βS1
Meg1βS2
Meg1βS3
Meg2αS1
Meg2αS2
Meg2αS3
Meg2βS1
Meg2βS2
Meg2βS3
108.85 + 3.5 =
112.35
69.7 + 3.5 =
73.2
168.85 + 3.5 =
172.35
(βFii = 71.1)
105.8
[8.1 − 1.1] + 68.2 =
75.2
[6 − 1.1] + [71.1] + 87.8 = 164.8
107.5
68.1
159.5
103.8
74.5
167.6
2.85
0.2
0.95
0.9
0.2
2.3
0.0
0.1
1.5
0.3
3.0
0.6
table 6
Aulos Meg1: Distances in mm related to sýrinx sliders. Discrepancies assuming that the hole,
when closed, should be covered by more than 1 mm of plate.
Meg1α/Sy1
Meg1β/Sy2
p.s.:
f.R.:
SlU:
s.h.:
PlL:
Plc:
s.R.:
± from Slider
SlU – E
120.25 + 3.5 =
123.75
81.1 + 3.5 =
84.6
180.25 + 3.5 =
183.75
(βFii = 71.1)
116.7
[8.1 + 1.1] + 77.7 =
86.9
[6 + 1.1] + [71.1] + 97 = 176.5
119.2
78.3
170.7
116.6
84.4
178.2
−0.75
−4.4
−3.55
−7.3
−3.1
−2.0
−4.3
−3.7
−3.8
−2.9
−3.1
−5.3
I1
S1
L1
I3
S2
L2
αI
αS
αL
βI
βS
βL
Pipe
p.s. – f.R.
Slider
SlU – s.R.
Pipe /Slider
discrepancy
Pipe
f.R. – s.h.
Slider
s.R. – PlL
s.R. – PlC
Pipe/Slider
discrepancy
76.45
103.8
78.8
105.5
2.35
1.7
76.2
71.7
76.0
71.7
80.65
76.45
1.3–4.45
1.5–4.75
sýrinx pin (upper edge)
fixed Ring (upper rim)
slot internal (upper edge)
sýrinx hole
slider plate lower end
slider plate centre
slider Ring (lower edge)
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
plate 1
Auloi Meg1 (Δ1965αβ) and Meg2 (Δ1964αβ); original restoration by the Museum
Greek and Roman Musical Studies 10 (2022) 15–77
69
70
plate 2
Terzḗs and Hagel
Meg1α bone sections; scale: 2/3
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
plate 3
Meg1β bone sections; scale: 2/3
Greek and Roman Musical Studies 10 (2022) 15–77
71
72
plate 4
Terzḗs and Hagel
Meg1 slider keys: long (L.1–2), intermediate (I.1–3); scale: 2/3
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
plate 5
Meg1 slider keys: short (S.1–2), sýrinx (Sy.1–2); scale: 2/3
Greek and Roman Musical Studies 10 (2022) 15–77
73
74
plate 6
Terzḗs and Hagel
Meg2α bone sections; scale: 2/3
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
plate 7
Meg2β bone sections; scale: 2/3
Greek and Roman Musical Studies 10 (2022) 15–77
75
76
plate 8
Terzḗs and Hagel
Meg2α slider keys: long (L), intermediate (I), short (S); scale: 2/3
Greek and Roman Musical Studies 10 (2022) 15–77
Two Auloi from Megara
plate 9
Meg2β slider keys: long (L), intermediate (I), short (S); scale: 2/3
Greek and Roman Musical Studies 10 (2022) 15–77
77