Bryn Mawr Classical Review

Bryn Mawr Classical Review 2005.07.63

David Whitehead, P.H. Blyth, Athenaeus Mechanicus, On Machines. Historia-Einzelschrift, 182.   Stuttgart:  Franz Steiner Verlag, 2004.  Pp. 236; figs. 23.  ISBN 3-515-08532-7.  €40.00 (pb).  

Reviewed by Duncan B. Campbell, Glasgow (
Word count: 2058 words

There is no doubt that the work of Athenaeus Mechanicus has been neglected by American and British scholars. It is true that the late Eric Marsden made brief, thought-provoking remarks on siegecraft, but he never made a full-scale study of Athenaeus, as his chosen field of ancient artillery was largely unaffected by the Περὶ μηχανημάτων. By contrast, our European colleagues have enjoyed over a century of debate on ancient siegecraft in general, and on Athenaeus in particular. Sparked by the publication, in 1867, of the major poliorcetic texts between the covers of one book,1 discussion quickly became the preserve of German scholars, partly owing to the compilation of the all-encompassing Real-Encyclopädie (1894-1980). Indeed, as recently as the 1980s, it seemed that the late Otto Lendle had effectively uttered the last word with his ingenious interpretations of Athenaeus and the other Poliorketiker. So this volume marks not only a welcome return to Athenaean studies, but indeed a debut for Anglophone research.2

In their introduction (13-41), W & B explore the identity of the author and the arrangement, style and substance of his work. Ever since Cichorius,3 it has been accepted that Athenaeus was a contemporary of the emperor Augustus, whose nephew and son-in-law, Marcellus, was the intended recipient of the Περὶ μηχανημάτων treatise. However, W & B rightly emphasize Cichorius's supplementary point, which has often been forgotten, that our Athenaeus and the peripatetic philosopher, Athenaeus of Seleucia, were almost certainly one and the same man. His relationship to Augustus' architect-engineer Vitruvius is also fundamental. In a nutshell, the bulk of Athenaeus' text bears a remarkable similarity to chapters 13-16 of Vitruvius' tenth book. Although plagiarism remains a possibility, the idea that 'each closely followed a common source' (14) has long been preferred, and the source in question universally agreed to have been an engineer named Agesistratus. However, it is perhaps unwise of W & B to insist that the latter never wrote anything down, but communicated his ideas orally (e.g., 73 on Ath. 7.5-6; 75 on Ath. 8.5); as a theory, this must remain unproven.

The present volume incorporates the Greek text of the Περὶ μηχανημάτων, largely as per Schneider's posthumous edition of 1912, and adds an English translation on facing pages (44-63). A lengthy commentary (65-170) provides detailed explication on the arcane subject of siege machine construction, and seven 'endnotes' (171-192) address individual topics at greater length: an alternative but erroneous date for the author; the life of his teacher, Agesistratus; the difference between a ram and a drill; the construction and function of Diades' towers; the aims of Hegetor; the helepolis of Epimachus; and how the device known as the protrochos was meant to work.

It must be stated at the outset that the entire volume is a joy to read. Both writers have prior experience of editing a poliorcetic text, so they can be trusted to have pondered the genre in some depth.4 The result is authoritative and well written, combining elegance and clarity, though the authors are occasionally over-pessimistic regarding Athenaeus' technical reliability and the purity of the received text (following B's previous observations on Apollodorus). W & B sensibly follow Wescher's pagination and line divisions, but for the commentary they helpfully arrange the work into five principal parts, with further subdivisions where appropriate. In fact, it is regrettable that this scheme is not carried into the text, where it would be truly useful. For example, the description of the χελώνη χωστρίς (ditch-filling tortoise), conventionally cited as 'Wescher 15.9-19.2', could become (following W & B's scheme) simply 'II.6'.

And so to the text itself. After explaining the rationale behind his treatise (W & B's section I, 'preface'), Athenaeus opens his discussion of siege machines (W & B's section II, 'good practice') by noting the achievements of Agesistratus (II.1), in particular the remarkable range of his catapults. W & B pick up on a suggestion of Marsden's (21), that Agesistratus had introduced a major improvement in artillery design by tweaking the so-called washers (χοινικίδες, or modioli). Marsden (and, following him, W & B) imagined that the 'un-tweaked' washers from an arrow-shooter discovered at Ampurias (Spain) provided a terminus post quem in the late second century B.C. for Agesistratus. But the case is more complex than W & B imagine. Firstly, the Ampurias evidence is not as securely dated as they imply; secondly, Vitruvius explicitly linked the improvement in washer design to the stone-projecting ballista, not to the arrow-shooting catapulta; and thirdly, archaeological evidence gathered over the last thirty years demonstrates that such an improvement was never applied to the arrow-shooter, all of which means that the Ampurias evidence is irrelevant.5 Athenaeus alleges that Agesistratus had managed to squeeze 'twelve minas of spring' into his record-breaking three-span arrow-shooter; in other words, the ligament ('sinew' in common parlance) forming each spring weighed 5.24kg (being 12 Attic minas). The point of the statement is obviously that this was an excessive amount of spring-cord for an arrow-shooter of this size. However, W & B believe that a standard three-span's spring-frame 'would have accommodated well in excess of 5kg of sinew' (76); regrettably, they omit any evidence (physical or mathematical) for this controversial statement.6

Athenaeus' brief history of the battering ram (II.2) closely parallels Vitruvius' text. Both claim that the Carthaginians invented the device, despite clear depictions of such machinery on Neo-Assyrian reliefs. Hitherto, it has been usual to postulate the transference of oriental techniques to the west via the Phoenicians and their colony at Carthage, perhaps around 500 BC.7 But W & B take a different tack. They prefer to believe that Athenaeus was literally correct, that rams 'made their first impact in the western Mediterranean around the turn of the sixth century' (81), and that the Assyrians used a drill rather than a ram. (The difference between the two devices is elaborated in endnote 3, 174-5; basically, W & B contend that the drill is a weapon designed for use against mudbrick, whereas a ram is only effective against stone walls.) They select Pericles' siege of Samos in 440 BC as the context for the new weapon's invention (citing Diodorus Siculus 12.28.3), but they fail to accommodate the evidence of the contemporary Olympia ram-head, whose vertical teeth are usually explained as a device for use against mudbrick.8 The drill theory certainly makes fascinating reading, but falls short of proof.

Next, Athenaeus proceeds to describe the siege-machines of Diades, an engineer famous for his association with Alexander the Great; he mentions the siege-tower (II.3), the ram-tortoise (II.4), the drill, the raven and the assault-bridge (II.5). W & B's main contribution here is the theory that Diades' φορητὸς πύργος was not a mobile machine (86), despite the implication of Vitruvius' parallel phrase, turris ambulatoria. W & B are not the first to appreciate the 'portable' nature of Diades' siege-tower, which was 'evidently a collapsible structure which could be conveniently transported in small units, and rapidly assembled whenever it was needed'.9 But no-one (to my knowledge) has previously suggested that the tower had no wheels. W & B appeal to Marsden's authority on this point; but the latter's theory, that this was not a 'fully mobile siege-tower mounted on a really efficient chassis fitted with wheels',10 is not quite the same as W & B's contention that it was an immovable structure. In their wider discussion of this issue (endnote 4, 176-87), W & B are unable to suggest any practical use for a wheel-less siege-tower and can cite no instances, from Alexander onwards, where mobility is not at least implied. Their suggestion, that the siege-towers which Alexander deployed on the mole at Tyre were of this design (98, 184), seems unlikely.11

Section II.6 on the 'filler tortoise' (χελώνη χωστρίς) brings us to Athenaeus' first detailed description. But, as with every other machine he describes, the instructions are sufficiently vague to permit several different interpretations. Above the wheeled undercarriage, W & B opt for a pitched roof, with two faces oriented to the front and rear, the latter at a rather steeper angle than the former (200, fig. 8). Although discussion of the crucial ridge-beam is omitted from the commentary, they have tacitly accepted Lendle's proposal that, rather than running lengthways, it is a crossbeam. But they stop short of adopting Lendle's elegant design of hipped roof with four sloping sides, claiming that 'nothing warrants such an assumption' (110). Curiously, in their later interpretation of Hegetor's ram-tortoise (their section II.8), in which Athenaeus actually specifies a transverse ridge-beam, they positively recommend Lendle's discussion of the roof (124) and adopt his hipped design (204, fig. 12).12 In the preceding section (II.7), describing the 'digger tortoise' (χελώνη ὀρυκτρίς), W & B make an unwarranted assumption of their own: namely, that this tortoise, 'which has its frontal approach-side vertical' to fit against the enemy wall, is simply the 'filler tortoise' turned through 90 degrees (119). Firstly, they assume that the latter tortoise has two vertical sides, a proposition (as already mentioned) which is not universally accepted. And secondly, if that were so, Athenaeus would have had no need to describe a second type of tortoise, but could simply have instructed the reader to turn the 'filler tortoise' around for digging duties. (Incidentally, this is surely another reason for arguing that the 'filler tortoise' has four sloping sides, just like Hegetor's version.)

The final machine in W & B's 'good practice' section is the helepolis of Epimachus (II.9), the largest siege-tower of antiquity, described by Athenaeus in rather abbreviated form.13 W & B concentrate on Athenaeus' remark that the tower could withstand blows from a three-talent stone-projector. The point of the comment is surely that this is the heaviest blow that Athenaeus could imagine, although the rather more technically-minded Philon held the opinion that the one-talent machine was far more to be feared. Of course, the helepolis itself was equipped with various catapults, prompting W & B's anxiety about vibrations (137, 190), but there is nothing, either in the ancient sources or in the practical design of the catapult, to encourage the belief that such machines recoiled violently after each shot. On the contrary, any residual energy not carried forward by the missile was dissipated along the bow-string and absorbed in the machine's substantial timberwork.14

The next section of Athenaeus' text, following W & B's scheme, covers 'bad practice': first, the σαμβύκη (III.1), which Athenaeus does not actually describe but warns that it must be of the correct height to assault walls; then Ctesibius' seesaw μηχανήματα (III.2), which he dismisses as fanciful.15 This is followed by an 'innovation' section, in which Athenaeus describes his own inventions in terms so vague and with so many hapax legomena as to defy interpretation. The first is the πιθήκιον (IV.1), a device generally translated as 'little ape' and apparently intended to keep machinery steady onboard ship. The second is the πρότροχος (IV.2), a 'forewheel' which allegedly enabled machinery to steer. And the third is the καρχήσιον (IV.3), Marsden's universal joint, here translated as the 'goblet joint', which is used somehow in conjunction with the γέρανος ('crane') to demolish walls. W & B suggest a plausible design for this, but the result is no more authoritative than Lendle's; it is simply different. Finally in this section are the τρίβολοι, giant caltrops whose deployment on a slope is intended to protect against objects rolled downhill,16 and the ἀρετή (IV.4), a kind of shelter useful when moving ladders. The work ends with a short epilogue (W & B's section V) emphasizing preparation for war as a deterrent, and defending Athenaeus' own record against unnamed critics.

Throughout, W. & B.'s commentary is sufficiently authoritative to stand on its own. They occasionally cite Paul Bentley Kern's Ancient Siege Warfare, perhaps as the only general book on the subject available to them, but this is mostly unnecessary. For example, on the question whether Alexander took a siege train over the Hindu Kush (182), they cite 'generally Kern, Siege 219-220', but this passage simply states that 'it is not clear how much, if any, of his siege train Alexander had gotten across the Hindu Kush'. In the end, W & B must be congratulated on presenting such a comprehensive piece of work, and it is to be hoped that one or other (or both) may now cast their expert eye in the direction of other poliorcetic texts.


1.   C. Wescher, Poliorcétique des Grecs (1867). As a Glasgow-based writer, I cannot resist noting that one of only two copies of this book available in the UK resides in Glasgow University Library.
2.   W & B cite all previous publications of relevance (though Dietwulf Baatz appears, inexplicably, as 'D. Baudy' on pp. 76 and 187). Note especially E.W. Marsden, Greek and Roman Artillery. Technical treatises (1971); O. Lendle, Schildkröten: antike Kriegsmaschinen in poliorketischen Texten (1975); idem, Texte und Untersuchungen zum technischen Bereich der antiken Poliorketik (1981). Marsden's remarks on siegecraft can be found (s.v.) in the Oxford Classical Dictionary (edn. 2, 1970).
3.   C. Cichorius, 'Das Werk des Athenaeus über Kriegsmaschinen', in: Römische Studien (Berlin, 1922), 271-9.
4.   D. Whitehead, Aineias the Tactician, How to Survive under Siege (1990); P.H. Blyth, 'Apollodorus of Damascus and the Poliorcetica', GRBS 33 (1992) 127-58.
5.   For the 'tweaking' of the washers, cf. D.B. Campbell, Greek and Roman Artillery 399 BC-AD 363 (2003), 34-5.
6.   The Greek three-span catapult was designed to shoot arrows measuring 69.5cm in length; its springs, being 4 dactyls in diameter (7.7cm) and 26 dactyls in height (50cm), thus measured 327 cubic dactyls (or 2.3 litres). Our only clue as to the weight of catapult springs (as Marsden realised) comes from Philon of Byzantium, who records that the springs of a stone-projector weighed about twenty-five times the shot. Applying the same logic to the arrow-shooter, and assuming that a three-span arrow weighed about 200g, Philon's rule of thumb would indeed give springs weighing roughly 5kg. However, the stone-projector was a very different machine, with springs almost 50% longer than an equivalent arrow-shooter's. In practice, the smallest stone-projector was probably the 2-mina calibre, designed to shoot stones weighing around 1kg. But in theory, if a 200g machine had existed, its springs would have been roughly 4 dactyls in diameter (just like those of the three-span arrow-shooter) but fully 38 dactyls in height (73cm), giving a volume of 477 cubic dactyls (or 3.4 litres). Logically, the springs of a regular three-span catapult, being two-thirds the size of the stone-projector's, would have been considerably lighter than 5kg.
7.   In D.B. Campbell, Greek and Roman Siege Machinery 399 BC-AD 363 (2003), 3, I have suggested that battering technology was fairly self-evident and may have developed simultaneously in different places. W & B accept Diodorus Siculus' first century BC view, that battering rams were unknown to the Assyrians (80 n.8), but ignore the fifth/fourth century BC testimony of Xenophon that they were already well-known to the Persians.
8.   W & B note the existence of the Olympia ram-head (80), but do not discuss its designer's intent. A date in the mid-fifth century B.C. has been suggested on stylistic grounds; cf. Campbell, op. cit. (note 7), 4. Mudbrick walls would have been common in Greece at this date, even (presumably) at Samos, where the ashlar walls are dated to 310-290 BC. On W & B's reasoning, if Pericles' engineer actually invented the battering ram there, he wasted his time, as he would have been better served by a drill.
9.   The quotation is from F.E. Winter, Greek Fortifications (1971), 320-1.
10.   The quotation is from E.W. Marsden, 'Macedonian military machinery and its designers under Philip and Alexander', in: Ancient Macedonia 2 (1977), 221. In Technical Treatises, 85, he refers to Diades' machines as 'mobile towers'.
11.   Mobility would seem to be a critical feature of any machinery deployed on an artificial runway which was itself being continuously extended towards the enemy. For the mole at Tyre, see now D.B. Campbell, Ancient Siege Warfare. Persians, Greeks, Carthaginians and Romans, 576-146 BC (2005), 36-7 (plate D) and 60-3.
12.   Cf. Campbell, op. cit. (note 7), 25 (plate A) and 28-9 (plate D), for reconstructions of the ditch-filling tortoise and Hegetor's ram-tortoise, following Lendle's suggestions. Incidentally, W & B are quite right to emphasize that, in the undercarriage, Athenaeus specifies four crossbars and only two lengthwise timbers; my diagram (op. cit., 14) repeats Lendle's error in reversing these numbers.
13.   For this machine, cf. Campbell, op. cit. (note 7), 9-12 and 27 (plate C). In noting the four ancient accounts of the machine, W & B pass no comment on Vitruvius' wildly inaccurate dimensions, which Marsden put down to 'faulty conversion of Greek cubits to Roman feet'. I have suggested (loc. cit.) that Vitruvius may inadvertently have been describing another of Demetrius Poliorcetes' four known helepoleis, which may well have been smaller.
14.   Cf. D.B. Campbell, 'Ballistaria in first to mid-third century Britain: a reappraisal', in Britannia 15 (1984), 75-84, at p.80 n.40. Basically, the impulse generated by shooting a 3-talent stone ball with a velocity of around 70m/s is easily offset against the catapult's massive framework, weighing (at a conservative estimate) 15 tonnes, or around three times the mass of the missile. Consequently, it is misleading for W & B to suggest that 'that impulse [i.e., the initial velocity of the missile] is a reasonable measure of the recoil which the machine [i.e., the helepolis] would have to absorb if a large catapult were discharged from it' (190).
15.   On these, cf. Campbell, op. cit. (note 7), 24 with 26 (plate B).
16.   On this general topic, cf. D.B. Campbell, 'The Roman siege of Burnswark', in: Britannia 34 (2003), 19-33, at 29. The Athenaeus passage is cited in n.64, where I translated σφόνδυλοι as 'great cylinders (perhaps tree trunks or column drums)'; W & B opt for the latter.

Read Latest
Index for 2005
Change Greek Display
Books Available for Review

HTML generated at 13:30:15, Friday, 03 April 2009