The date of Hero of Alexandria, and another translation of some extracts of the “Mechanics”

When did Hero of Alexandria live?  The truth is that we know little other than what can be inferred from his works.

Karin Tybjerg[1] tells us that Hero quotes Archimedes, who lived ca. 287-211 or 212 BC, and is quoted by Pappus who flourished around 320 AD.  But it seems that in his Dioptra Hero refers to a lunar eclipse visible at Alexandria and Rome.  The only one that fits these criteria happened in 62 AD, around the time that St. Paul was released from house arrest in Rome, and also around the time that Mark’s gospel was being written.  Ptolemy (fl. ca. 127-158 AD) does not make use of Hero, which is perhaps an upper limit.

It has been speculated that a new model of water-organ, demonstrated to Nero in 68 AD, was probably Hero’s invention; which means that he might even have been in Rome at the same time as the apostles.  At his lavish new palace, the Domus Aurea, a new and ingenious technology entertains an emperor and his court.  Meanwhile, across the city the apostle Peter is addressing a humble congregation.  It is a reminder that the most important events of an era are not always the most heralded.

Hero’s Mechanics is about how to lift heavy weights.  The first two books go through a number of principles for doing so with limited power, while the third book describes designs for weights and presses.

I’ve come across another handbook which contains an English translation of some extracts.  This is G. Irby-Massie & P.T. Keyser, Greek Science of the Hellenistic Era : A Sourcebook, 2002, starting on p.168.  They are taken from Drachmann’s handbook which I have yet to see.[2]  Here they are:

6.11. Heron of Alexandria

Mechanics 1.20–21: weights; 2.1.1: simple machines; 2.3: the pulley; 3.2.1–2: the crane

There are many who think that weights lying on the ground are only moved by an equal force [contrast Aristotle, Physics 7.5 (250a11–19)], wherein they hold wrong opinions. So let us prove that weights placed in the way described are moved by an arbitrarily small force, and let us make clear the reason why this is not evident in fact. Let us imagine a weight lying on the ground, and let it be regular, smooth and with its parts coherent with each other. And let the surface on which the weight lies be flat, smooth and completely joined, and able to be inclined to both sides, i.e., to the right and the left. And let it be inclined first towards the right. It is then evident to us that the given weight must incline towards the right side, because the nature of weights is to move downwards, if nothing holds them and hinders them from movement; and again if the inclined side is lifted to a horizontal position and will be level [i.e., in equilibrium], the weight will come to rest in this position. And if it is inclined to the other side, i.e., to the left side, the weight will again sink towards the inclined side, even if the inclination is very small, and so the weight will need no force to move it, but will need a force to hold it so that it does not move. And if the weight again becomes level without inclination to either side, then it will stay there without a force holding it, and it will not cease being at rest until the surface inclines to one side or another, and then it will incline towards that side. Thus the weight that is ready to incline to whichever side, does it not require only a small force to move it, namely as much force as causes the inclination? And so isn’t the weight moved by any small force?

21.  Now, water on a surface that is not inclined will not flow, but remains without inclining to either side. But if the slightest inclination occurs, then all of it will flow towards that side, till not the smallest part of the water remains thereon, unless there are hollows in the surface, and small amounts stay in the bottom of the hollows, as happens often in vessels. Now water inclines like this because its parts lack cohesion and are very soluble.
As for the bodies that are coherent, since by their nature they are not smooth on their surfaces and not easily made smooth, it happens through the roughness of the bodies that they strengthen each other, and it happens that they lean upon each other like teeth, and they are strengthened thus, for if the teeth are numerous and closely joined, they require a strong and coherent force [to separate them]. And so from experiment people gained understanding: under tortoises [war machines: see Athenaios above, Section 6.9] they placed pieces of wood whose surfaces were cylindrical and so did not touch more than a small part of the surface, and so only very little rubbing occurred. And they use poles to move the weight on them easily, even though the weight is increased by the weight of the tools. And some people put on the ground cut boards (because of their smoothness) and smear them with grease, because thus their surface roughness is made smooth, and so they move the weight with smaller force. As for the cylinders, if they are heavy and lie on the ground, so that the ground does not touch more than one line of them, then they are moved easily, and so also balls; and we have already talked about that [1.2–7].

2.1.1 [simple machines (surviving in Greek)]

Since the powers by which a given weight is moved by a given force are five, it is necessary to present their form and their use and their names, because these powers are all derived from one natural principle, though they are very different in form. Their names are as follows: the axle-in-wheel [windlass], the lever [mochlos], the pulley [trochilos], the wedge [sphên], and what is called the “endless” screw [kochlia].

[construction of the axle-in-wheel]

[2.2 the lever]

2.3 [pulley (surviving in Greek)]

The third power [pulley] is also called the ‘‘multi-lifter” [poluspaston].

Whenever we want to move some weight, if we tie a rope to this weight we pull with as much force as is equal to the burden. But if we untie the rope from the weight, and tie one of its ends to a stationary point and pass its other end over a pulley fastened to the burden and draw on the rope, we will more easily move the weight. And again if we fasten on the stationary point another pulley and run the end of the rope through that and pull it, we will still more easily move the weight. And again if we fasten on this weight another pulley and run the end of the cord over it, we will much more easily move the weight. And in this way, each time we add pulleys to the stationary point and to the burden, and run one end of the rope through the pulleys in turn, we will more easily move the weight. And every time the number of pulleys through which the rope runs is increased, it will be easier to lift that weight. The more “limbs” [kôla] the rope is bent into, the easier the weight will be moved.

And one end of the rope must be securely tied to the stationary point, and the rope must go from there to the weight (Figure 6.8). As for the pulleys that are on the stationary point they must be fastened to one piece of wood, turning on an axle, and this axle is called manganon; and it is tied to the stationary point with another rope. And as for the pulleys that are fastened to the burden, they are on another manganon like the first, tied to the burden. The pulleys should be so arranged on the axles that the “limbs” do not get entangled and unwieldy. And why the ease of lifting follows from the number of “limbs,” and why the end of the rope is tied to the stationary point, we shall explain in the following.

3.2.1–2 [the crane (surviving in Greek): compare Vitruvius 10.2.8]

1. For the lifting of burdens upwards there are certain machines: some have one mast, and some have two masts, and some have three, and some have four masts. As for the one that has a single mast it is made in this way. We take a long piece of wood, longer than the distance to which we want to raise the burden, and even if this pole is strong in itself, we take a rope and coil it round, winding it equally spaced, and draw it tight. The space between the single windings should not be greater than four palms [ca. 30 cm], and the windings of the rope are like steps for the workers and they are useful for anyone wanting to work on the upper section. And if the pole is not elastic, we must estimate the burdens to be lifted, lest the mast be too weak.

2 This mast is erected upright on a piece of wood, and three or four ropes are  fastened to its top, stretched and tied to fixed points, so that the beam, however it is forced, will not give way (being held by the ropes). Then they attach to its top pulleys tied to the burden. Then they pull on the rope either by hand or with another engine, until the burden is raised [Figure 6.9].

[3.3–5: two-, three-, and four-mast cranes]


  1. [1]Karin Tybjerg, “Hero of Alexandria’s Mechanical Treatises”, in: Astrid Schurmann (ed), Physik / Mechanik, Stuttgart (2005), 204-226.  Preview here.
  2. [2]Drachmann [1963] 46–47, 50, 53–55, 98–99; Drachmann, A.G. (1963) Mechanical Technology of Greek and Roman Antiquity, Copenhagen: Munskgaard.

Extracts from Hero of Alexandria’s “Mechanics”

Earlier this week I saw a reference online to a work by Hero of Alexandria, the ancient constructor of machines who lived at an uncertain time, possibly even in the late 1st century AD.[1]  The reference was to his Mechanics.

In the Mechanica, I am told, Hero explored the parallelograms of velocities, determined certain simple centers of gravity, analyzed the intricate mechanical powers by which small forces are used to move large weights, discussed the problems of the two mean proportions, and estimated the forces of motion on an inclined plane.

The original Greek of the Mechanica is lost.  Fragments are quoted by the 3rd century AD author Pappus in his Mathematical Collection; but the complete work is preserved in an Arabic translation by Qusta Ibn Luqa (9th century).  (Remarkably, I found a manuscript of the Arabic online here, and a discussion of Greek mechanical texts here; an interesting page with bibliography on Hero here.).[2]  The work was edited with a German translation by L. Nix in Heronis Alexandrini Opera quae supersunt, Leipzig: Teubner (1899-1914), vol. II.2 (1900), which may be found online at Wilbour Hall, here.  There are extracts in English in A.G. Drachmann, The Mechanical Technology of Greek and Roman Antiquity, Copenhagen: Munksgaard, 1963, some of which are quoted by Papadopoulos at this article preview here.  I wish I had access to Drachmann.  UPDATE:  I had forgotten the complete French translation by Barob Carra de Vaux.[3]

The reference that I saw online was in the form of a blog post. The blogger posted a couple of pages of the Arabic (p.177 and 179 of the Nix edition, book II chapter 34, questions  f-i) – which relate to topics like why breaking a stick is easier if you put your knee in the middle, bending of planks, and so on- [4] and asked if someone would make a translation for him; someone familiar with medieval Arabic, and the relevant technical terms, which is quite a  request. He also said that he was familiar with the standard English translation, which annoyingly he did not name, but he wanted specialist interpretation of specific words, with a view to scholarly publication.  I found that, oddly, comments on the blog are only permitted from those who sent him money.[5]

Anyway, all this made me search for whatever English translation that I could find.  The work has not been translated as a whole, as far as I could tell.

However I found that extracts have been turned into English in Morris R. Cohen and Israel E. Drabkin, A Source Book in Greek Science (1958), which is itself a rather remarkable and useful volume.  Few will have access to it, however, so I thought that I would give a couple of these extracts here.  The footnotes are those of Cohen and Drabkin.

Book I, chapters 20-23:[6]

20. Many people have the erroneous belief that weights placed on the ground may be moved only by forces equivalent to these weights. Let us demonstrate that weights placed as described may [theoretically] be moved by a force less than any given force, and let us explain the reason why this is not the case in practice.[7] Suppose that a weight, symmetrical, smooth, and quite solid, rests on a plane surface, and that this plane is capable of inclining toward both sides, that is, toward the right and the left. Suppose it inclines first toward the right. In that case we see that the given weight moves down toward the right, since it is in the nature of weights to move downward unless something supports them and hinders their motion. If, now, the side sloping downward is again lifted to the horizontal plane and restored to equilibrium, the weight will remain fixed in this position.

Again, if the plane is inclined toward the other side, that is, toward the left, the weight, too, will tend toward the lowered side, even if the slope is extremely small. The weight, in this case, does not require a force to set it m motion but rather a force to keep it from moving. Now when the weight again returns to equilibrium and does not tend in either direction, it remains in position without any force to support it. It continues to be at rest until the plane is made to slope towards either side, in which case the weight, too, tends in that direction Thus it follows that the weight, which is prone to move in any desired direction, requires, for its motion, only a very small force equal to the force which inclines it.[8] Therefore the weight will be moved by any small force.

21. Pools of water that lie on non-sloping planes do not flow but remain still, not tending toward either side. But if the slightest inclination is imparted to them they flow completely toward that side, until not the least particle of water remains in its original position (unless there are declivities in the plane in the recesses of which small parts of water remain,as sometimes happens in the case of vessels).

Now this is the case with water because its parts are not strongly cohesive but are easily separable. Since, however, bodies that cohere strongly do not, naturally, have smooth surfaces and are not easily smoothed down, the result is that because of their roughness they support one another. That is, they are engaged like cogged wheels in a machine, and are consequently prevented [from rolling].[9]

For when the parts are numerous and closely bound to one another by reason of mutual cohesion, a large coordinated force is required [to produce motion of one body made up of such parts over another]. Experience has taught men to lay logs with cylindrical surfaces under tortoises,[10] so that these logs touch only a small part of the plane, whence only the smallest amount of friction results. Logs are thus used to move weights easily, but the weight of the moving apparatus must exceed that of the load to be moved. Others plane down boards to render them smooth, fasten them together on the ground, and coat them with grease, so that whatever roughness there is may be smoothed out. Thus they move the load with little force. Columns [cylinders], even if they are heavy, may be moved easily if they lie upon the ground in such a way that only one line is in contact with the ground. This is true also of the sphere,[11] which we have already discussed.

22. Now if it is desired to raise a weight to a higher place, a force equal to the weight is needed. Consider a rotating pulley suspended perpendicular to the plane and turning about an axis at its midpoint. Let a cord be passed around the pulley and let one end be fastened to the weight and the other be operated by the moving force. I say that this weight may be moved by a force equal to it. For suppose that, instead of a force, there is, at the other end of the cord, a second weight. It will be seen that if the two weights are equal the pulley will not turn toward either side. The first weight is not strong enough to overbalance the second, and the second is not strong enough to overbalance the first, since both are equal. But if a slight addition is made to one weight, the other will be drawn up. Therefore, if the force that is to move the load is greater than the load, it will be strong enough to move the latter, unless friction in the turning of the pulley or the stiffness of the cords interferes with the motion.

23. Weights on an inclined plane have a tendency to move downward, as is the case with all bodies. If such movement does not take place we must invoke the explanation given above.[12]Suppose we wish to draw a weight up an inclined plane the surface of which is smooth and even, as is also the surface of that part of the weight which rests on the plane. For our purpose we must have a force or weight operating on the other side and just balancing the given weight, that is, conserving the equilibrium so that any addition of force will be sufficient to move the weight up the plane.

To prove our contention, let us demonstrate it in the case of a given cylinder. The cylinder has a natural tendency to roll downward because no large part of it touches the surface of the plane. Consider a plane perpendicular to the inclined plane and passing through the line of tangency between the cylinder and the inclined plane.

Clearly, the new plane will pass through the axis of the cylinder and divide the cylinder into two halves. For, given a circle and a tangent, a line drawn from the point of contact at right angles to the tangent will pass through the centre of the circle. Now pass a second plane through the same line (i.e., the line at which the cylinder touches the inclined plane) perpendicular to the horizon. This plane will not coincide with the plane previously constructed, but will divide the cylinder into two unequal parts, of which the smaller lies above and the larger below. The larger part, because it is larger, will outweigh the smaller, and the cylinder will roll down. If, now, we suppose that from the larger [of the two parts into which this plane perpendicular to the horizon divides the cylinder] that amount be removed by which the larger exceeds the smaller portion, the two parts will then be in equilibrium and their joint weight will remain unmoved on the line of tangency to the inclined plane, tending neither upward nor downward. We need, therefore, a force equivalent to this difference to preserve equilibrium.[13] But if the slightest addition be made to this force, it will overbalance the weight.

The next fragment is from Book II, chapter 34d:

d. Why do heavier bodies fall to the ground in shorter time than lighter bodies?

The reason is that, just as heavy bodies move more readily the larger is the externaI force by which they are set in motion, so they move more swiftly the larger is the internal force within themselves. And in natural motion this internal force and downward tendency are greater in the case of heavier bodies than in the case of lighter.

These kinds of works all need reliable translation.  It is telling that Cohen and Drabkin plainly just translated the German translation of Nix.

It would be interesting to find if there is a real translation of Mechanics II, 34!

  1. [1]Evangelos Papadopoulos, here: “The chronology of Heron’s works is disputed and not absolutely certain to date. Many contradictory references on Heron exist, partly because the name was quite common. However, historians cite that he came after Apollonius, whom he quotes, and before Pappos, who cites him. This suggests that he must have lived between 150 BC and 250 AD (Thomas, 2005). In 1938, Neugebauer, based on a reference in Heron’s Dipotra book of a moon eclipse, he found that this must have happened on March 13, 62 AD. (Neugebauer. 1938). Since the reference was made to readers who could easily remember the eclipse, this suggests that Heron flourished in the late first century AD. According to Lewis (2001), and assuming that Cheirobalistra, a powerful catapult, is genuinely his, Heron should have been alive at least till 84 AD, the year in which the Cheirobalistra, was introduced.”
  2. [2]The shelfmark is British Library Additional Ms. 23390, fol. 3r-50r; 17th century.
  3. [3]Heron d’Alexandrie, Les Mechaniques ou l’elevateur de Heron d’Alexandrie, 1894 (Google books).  In: Journal asiatique, IXe serie, tome II, 1893, 152-289 and 420-514.  Online at here:
  4. [4]The German translation is:

    “f. Warum treibt ein Schufs von der Mitte der Sehne den Pfeil auf eine grolse Entfernung hinaus?

    Weil die Spannung daselbst am stärksten und die treibende Kraft am gröfsten ist. Deshalb macht man auch die Bogen aus Hörnern, weil hierbei das Biegen möglich ist. Wenn sie stark gebogen sind, ist auch die Sehne mit dem Pfeil stärker gespannt, so dais eine gröfsere Kraft in ihn kommt und er deshalb eine weitere Strecke durchdringt. Deshalb treiben harte Bogen, deren Enden sich nicht biegen lassen, den Pfeil nur auf kurze Strecken.

    g. Warum läfst sich Holz schneller brechen, wenn man das Knie bei demselben in die Mitte bringt?

    Weil, wenn man das Knie dabei in geringere Entfernung (vom einen Ende) als die Mitte bringt, so dafs der eine der beiden Teile kürzer ist als der andre, es eine in zwei ungleiche Teile geteilte Wage ist, weshalb die von dem Knie entferntere Hand das Übergewicht über die ihm nähere hat. Die eine erreicht aber die Kraft der anderen nur, wenn beide an dem Ende des Holzes (gleichweit von der Mitte) sind.

    h. Warum ist ein Stück Holz, je länger es ist, desto schwächer und warum nimmt seine Biegung zu, wenn es in einem seiner beiden Enden aufgerichtet wird?

    Weil im langen Holze grofse Kraft auf seine Teile verteilt ist, so dafs das Ganze das Übergewicht hat über den festen Teil desselben, auf welchem es sich erhebt. Daher tritt hierbei dieselbe Erscheinung ein wie hei kurzem Holz, wenn an dessen Enden etwas hängt, das es niederdrückt. Der Zuwachs an Länge des Holzes entspricht also dem Gewichte, welches das kürzere Holz herabzieht. Deshalb begegnet dem langen Holze durch sich selbst wegen seiner Länge dasselbe, wie dem kurzen Holz, wenn an seinem Ende etwas Schweres angebunden wird.

    i. Warum benutzt man beim Zahnausziehen Zangen und nicht die Hand?

    Weil wir den Zahn mit der ganzen Hand nicht packen können, sondern nur mit einem Teil derselben; und wie es uns schwerer fällt, ein Gewicht mit nur zwei Fingern zu heben, als mit der ganzen Hand, so ist es auch schwerer für uns, den Zahn mit zwei Fingern zu packen und zu drücken, als mit der ganzen Hand. In beiden Fällen ist die Kraft dieselbe, aber die Teilung der Zange bei ihrem Nagel bewirkt dazu, dafs die Hand die Übermacht über den Zahn hat; denn es ist ein Hebel, an dessen gröfserem Teil die Hand ist, und der Abstand der Zange erleichtert das Bewegen des Zahnes. Denn die Zahnwurzel ist das, um was sich der Hebel bewegt. Weil aber der Abstand der Zange gröfser ist als die Zahnwurzel, um die sich etwas Grofses bewegt, so überwiegt die Hand über die in der Zahnwurzel liegende Kraft. Es ist nämlich kein Unterschied zwischen dem Bewegen eines Gewichtes und dem Bewegen einer Kraft, die jenem Gewichte gleichkommt. Denn wenn wir die Hand schliefsen, nachdem sie ausgebreitet war, so entsteht ein Widerstand, nicht wegen…”.  This Google Translate renders as:

    f. Why does a creature from the middle of the tendon push the arrow out to a great distance?

    Because the tension is greatest there, and the driving force is greatest. That is why the bow is made of horns, because bending is possible. When they are strongly bent, the tendon is tightened with the arrow, so that a greater force comes into it and therefore it penetrates a further distance. Therefore, hard bows, whose ends can not be bent, drive the arrow only for short distances.

    g. Why can wood break faster when the knee is brought into the middle?

    If the knee is placed at a distance (from the one end) as the middle, so that one of the two parts is shorter than the other, it is a divided into two unequal parts, which is why the hand distant from the knee The overweight has got closer to him. But the one reaches the power of the other only when both are at the end of the wood (equidistant from the center).

    h. Why is a piece of wood the longer it is, the weaker and why does its bending increase when it is raised in one of its two ends?

    Because great force is distributed over its parts in the long wood, so that the whole has the preponderance over the fixed part on which it rises. Hence the appearance of the wood is like that of a short wood, when there is something hanging on the ends of it, which it presses down. The increase in the length of the wood therefore corresponds to the weight which the shorter wood draws. Therefore the long wood meets with itself by its length, as the short wood, when at its end something heavy is tied.

    i. Why use forceps when pulling teeth and not the hand?

    Because we can not grasp the tooth with the whole hand, but only with a part of it; And as it is harder for us to lift a weight with only two fingers than with the whole hand, it is also harder for us to grasp and press the tooth with two fingers than with the whole hand. In both cases the force is the same, but the division of the forceps with its nail causes the hand to have over the tooth; For it is a lever, on the greater part of which the hand is, and the distance of the forceps facilitates the movement of the tooth. Because the tooth root is what the lever moves. But because the distance between the tongs is greater than that of the tooth-root, around which a great extent moves, the hand overcomes the force in the tooth-tooth. Indeed, there is no difference between moving a weight and moving a force equal to that weight. For when we close the hand after it has spread, there arises a resistance, not because of …”.  This gives the idea at least.

  5. [5]The appeal may be found here.
  6. [6]p.197, n.1. “The Mechanics is extant in an Arabic version in three books. Of the original Greek only fragments are extant, the longest of which is a passage, which Pappus quotes (p 224, below), on the five simple machines. The first book deals with various problems in statics, dynamics, and kinematics, the second discusses the simple machines and includes a collection of problems similar to that in the Aristotelian Mechanics, while the third deals with the construction and operation of machines, especially those for lifting large weights For the selections given here the German translation of L. Nix has been used as a basis. Professor A S. Halkin has helped interpret the Arabic text.”
  7. [7]p.197 n.2. “The general point of this and the following section (I.21) is clear. The passage constitutes an important step in the development of the principle of inertia. Particularly to be noted is the treatment of friction and the means of reducing it. There is some difficulty in the precise interpretation of certain points; see the following note.”
  8. [8]p.198 n.1 “The idea seems to be that the force required to set the weight in motion on the horizontal plane is equal to the resistance that enables the weight to continue at rest while the plane is inclined until the critical angle is reached beyond which the weight moves down the plane. If a plane be inclined at this critical angle, a, we should say that the force required to draw weight W up the plane is W sin a + R, where R is the component due to friction. Hero’s point seems to be, but it is not put very clearly, that the same force would be necessary to set the weight in motion on the horizontal plane. Of course, where R approached zero, a would approach zero, as would also the force necessary to set the weight in motion on the horizontal plane.”
  9. [9]p.198 n.2. “I.e., even in the case where the surface is somewhat inclined. The text is not clear but the reference may be to the interaction of a body and the surface on which it rests.”
  10. [10]p.199 n.1: “The reference is to sheds moved by rollers and used in military operations.”
  11. [11]p.199 n.2: “Here the tangency approaches a single point.”
  12. [12]p199 n.3 “I.e., the effect of friction.”
  13. [13]p.200 n.1 “The method of Hero amounts to holding that (in the case of a cylinder of radius r, height A, and density d, rolling on a plane inclined to the horizontal at angle a) the force (F) required to draw the weight (W) up the plane is r = W[2(a + cos a sin a) /r] — 2r2hd (a + cos a sin a). The modern formulation is F = W sin a. Hero’s formula is obtained by noting that he takes the part of the cylinder standing over LMN as the part to be balanced by force F, so that F/W = area LMN/(πr2)\ and area LMN = sector MON + 2 ΔNOL = r2 (2a + sin 2a). Compare the problem of the inclined plane as formulated and treated by Pappus (p. 194). The modern solution was not discovered until the days of Stevinus and Galileo.”

Hero of Alexandria, on the making of automata

The technical works of antiquity are not well known, not least because modern technical knowledge is often necessary to understand them.  For instance a reading of an alchemical work may well baffle anyone without a Chemistry degree!  So … they go untranslated and unread.

Four years ago I listed the works of the engineer, Hero of Alexandria, here.  In this I included a reference to a translation of his work Peri automatopoietikes, on making automata: Susan Murphy, “Heron of Alexandria’s On Automaton-making“, in: History of Technology 17 (1995), 1-46.  At the time I was quite unable to locate this journal, or the translation, so the matter went no further.

A couple of weeks ago I obtained further details.  The information came from Sydney University library, from whose catalogue I learned that it was published in London by Mansell.  It seems that the “journal” is actually a series of books, published under the imprint of Mansell of London, by none other than Bloomsbury Press (2023 update: who have a webpage about volume 17, with table of contents here.  ISBN 9781350018747).  It is a series aimed at engineers, and so naturally shelved away from the sort of material with which we are familiar.  In fact the series seems to be widely held, and it is merely the rather generic title which makes searching difficult.  Each volume contains a number of articles.

Dr Murphy’s article is full of interest.  I learn that the work is illustrated in the manuscripts, with diagrams that may go back to the author but are supposedly corrupt.  The critical edition of the text does not trouble to reproduce them – no doubt because of the difficulties of printing coloured photographs – but instead has drawings by a modern author, based upon them.

The work describes the construction of  two automata, as an example of two types of automaton.

The first is a mobile shrine of Dionysus, complete with little figurines of the god and his worshippers.  This rolls of its own accord on a wheeled base to a specified point, at which the figurines enact a scene of sacrifice and pouring libations.  It then returns to the original point.

The second is a minature theatre, which stages a complete tragedy when activated.

Both types of automaton rely on a descending counterweight and various cords and axles – essentially upon clockwork.

The opening section of the work, before the technical receipes, is itself rather interesting.


Book I

1. 1. The study of automaton-making has been considered by our predecessors worthy of acceptance, both because of the complexity of the craftsmanship involved and because of the striking nature of the spectacle. For, to speak briefly, every facet of mechanics is encompassed within automaton-making, in the completion of its several parts.

2. These are the topics to be discussed: shrines or altars of appropriate size are constructed, which move forward of themselves and stop at specified locations; and each of the figures inside them moves independently according to the argument of the arrangement or story; and then they move back to their original position. Thus such realizations of automata are called mobile.

3. But there is another kind, which is called stationary, and its function is as follows: a toy stage with open doors stands on a pillar, and inside it an arrangement of figures has been set up in line with some story.

4. To begin with, the stage is closed, and then the doors open by themselves, and the painted representation of the figures is displayed. After a little while the doors close and open again of their own accord, and another arrangement of figures, sequential to the first one, appears. Again the doors are closed and opened and yet another arrangement, which logically follows the one before it, appears; and either this completes the planned story, or yet another display appears after this one, until the story finally is finished.

5. And when the figures which have been described are shown in the theatre each one can be shown in motion, if the story demands; for instance, some sawing, some chopping with the adze, some working with hammers or axes – making a noise with each blow, just as they would in real life.

6. Other movements can be effected below the stage; for instance, lighting fires or making figures which were not visible at first appear and then disappear again. Simply, anyone can move the figures as he chooses, without anybody being near them.

7. But the mechanism of the stationary automata is safer and less risky and more adaptable to every requirement than that of the moving ones. Older generations called such feats of craftsmanship miraculous because they offered an amazing spectacle.

8. Therefore, in this book I am going to write about moving automata, and set out my own complex scenario, which is adaptable to every other scenario, so that someone who wanted to offer a different presentation would not lack anything for the implementation of his own scenario. In the following book I talk about stationary automata.

Well worth hunting out this volume, if you have even the slightest interest in ancient technology.

UPDATE (Sept 2023): Some additions and corrections made after locating the book on the Bloomsbury website.


The text tradition of the Greek artillery manuals

Few of us know much about the technical treatises of antiquity.  My last post, on Hero of Alexandria, inevitably mentioned his two works of this kind.  When I went to look at the volume in which a translation exists,[1] I was drawn into the question of how these works reach us.

As long ago as 1867, Wescher addressed this question, and, since his work is online, we can inform ourselves what he thought.[2]  The following notes are taken from Marsden’s volume, however.

The collection of works in Greek consists of Biton’s Construction of War Machines and Artillery; Hero’s Belopoeica and Cheiroballistra, and Philon’s Belopoeica.

Our knowledge of this collection depends in the main on four manuscripts.  All contain Biton and Hero, but only P and V contain Philon.

  • M — Codex Parisinus inter supplementa Graeca 607.  This is a collection of several manuscripts, bound together in the 15-16th century.  At one time it was in the library of Matthias Corvinus (1457-1490), and later belonged to the library of the Abbey of Vatopedi on Mount Athos.  It was acquired by a French government agent, Minoidas Minas, who was paid to search for and acquire manuscripts in the East.  He brought it to Paris in 1843, and transcribed a few sections which he made available to scholars, but concealed the manuscript itself, which was only discovered among his papers at his death in 1863.  It was naturally claimed as government property.  The central portion of the ms — folios 16-104 — contains Biton, and then the two works of Hero, and was copied by dictation from a manuscript in uncials, and may be 9-10 or 10-11th century.
  • F — Fragmenta Vindobonensia 120.  A rather carelessly copied ms. of the 16th century, contains lengthy excerpts of Biton and Hero which derive, not from M, but from a sister manuscript of equal antiquity, and that ancestor was somewhat better written than M.
  • P — Codex Parisinus 2442 (part of the manuscript is also in Codex Barberinianus 276).  It was carefully copied in the 11th century, and contains Biton, Hero and, at the end, Philon.
  • V — Codex Vaticanus 1164.  A sister manuscript of P, copied at the same time from the same exemplar.

There is also information to be had in the following:

  • C — Fragmentum in Codice Coisliniano 101.  This contains in its front binding two pages from an 11th century ms. very like P and V, and the text is part of Biton, plus some of Athenaeus Mechanicus.
  • V1 — Codex Vaticanus 219.  This is early 15th century, from the same sources as C, P and V.  But several later mss derive from it.
  • P2 — Codex Parisinus 2435.  A late ms., which was the original of the 1693 edition of Mathematici Veteres by Thevenot.
  • E — Codex Escorialensis Υ-111-11.  E seems to be a copy of V, made not long after V was written.  Some missing leaves (from Philon) are in Codex Borbonico-Neapolitanus.

There are other and more recent manuscripts, but all of them are copies or descendants of these eight.

It seems that a single uncial manuscript of Biton and Hero survived into the 9th century, when two copies were made from it.  One of these, M, survives, and contains traces of the Ionian dialect in Biton and Athenaeus Mechanicus.   The other is lost, but was the ancestor of F.  However at some point the dialect in this family was normalised.

Some time later, someone in the Byzantine world decided to create a collection of seige and tactical works.  This collection used the second family as a source; added three works of Philon, the Belopoeica, Parasceuastica, and Poliorcetica; and also added tactical works from a third source, creating a compendium of works.  P and V are copies of this collection.

In the process of compilation, however, the compiler managed to lose some of the diagrams.  Spaces are left in the text for three diagrams in Biton, which do appear in M, and one of them in F (the portions of the text where the others would be is not preserved in F).  There are likewise spaces in the text of Philon, where illustrations should be, which the compilation did not preserve.

  1. [1]E. W. Marsden, Greek and Roman artillery: technical treatises, Oxford, 1971.
  2. [2]C. Wescher, La poliorcetique des Grecs, Paris, 1887:, plus other copies on etc.  It contains details of the manuscripts, and then the texts with diagrams.

Translations of the works of Hero of Alexandria

The appearance of a manuscript of works by the ancient engineer, Hero of Alexandria (ca. 62 AD) online at the British Library led me to look online for an English translation for his Automata.  I had no luck, but I thought that some notes on what he wrote and how we got it might be useful.

The Greek texts, with German translations, are all available in W. Schmidt, Heronis Alexandrini Opera Quae Supersunt Omnia, in 5 volumes, Teubner, 1899 etc.  These are online here.

The mechanical works include: [1]

  •  The Pneumatica in 2 books, covering devices powered by compressed air, steam and water. An old English translation exists and is online: The pneumatics of Hero of Alexandria, 1851; as does an old French translation.[2]
  •  Peri automatopoietikes or Automata, on ways to fake miracles in temples.  An old partial French translation exists: Victor Prou, Les théâtres d’automates en Grèce au IIe siècle avant l’ère chrétienne d’après les Automata d’Héron d’Alexandrie, Paris, 1881: Greek text with French translation of section 2 of the Automata: p. 94-136.  This is online at[3]  An English translation supposedly exists: Susan Murphy, “Heron of Alexandria’s ‘On Automaton-making,'” History of Technology 17, 1995, pp 1-44;[4] and of selections in Finlay McCourt, “An examination of the mechanisms of movement in Heron of Alexandria’s ‘On Automaton-making'”, in: T. Koetsier &c, Explorations in the history of machines and mechanisms, 2012 (preview). Update (2020): F. Grillo’s Hero of Alexandria’s Automata: A critical edition and translation, diss (Glasgow), 2019 (online here).
  •  The Mechanica in 3 books survives only in Arabic, in a translation made by Qosta ibn Luka in the 9th century.  In the 17th century Grolius brought back a 16th century manuscript of it from the Orient, thereby making it accessible.  The first full edition and a French translation of this was by the baron Carra de Vaux in 1893.[5]  It covers weight-moving machines.
  •  The Dioptra covers instruments for sighting and other purposes.  A partial English translation supposedly exists from 1963[6]
  •  The Catoptrica, on mirrors.  Preserved only in Latin.

I also found a mistake in the literature: Drachmann’s Mechanical technology does NOT contain translations of any of his works, but is rather a commentary on the Mechanica (only), albeit with excerpts embedded.[9]  Unfortunately it does not specify which works.

There are also two artillery manuals covering different types of catapult.

  •  The Belopoeica.
  •  The Cheiroballistra, (=De constructione et mensura manubalistae)

These are both translated into English with facing Greek text and useful notes following in Marsden, E. W.: Greek and Roman artillery: technical treatises. Oxford, 1969, which also includes a useful introduction[10].  It seems that a bunch of these manuals travelled down the centuries together, and I will post on the manuscript tradition.  An 1883 French translation of the second work exists and is online.[11]

Hero also wrote a number of mathematical works.

  • The Metrica in 3 books, on the measurement and division of surfaces and bodies.  There is an English translation of all three books in “Codex Constantinopolitanus Palatii Veteris, edited by E.M.Bruins”, volume 3, Brill (1964), starting at page 182.

  • The Definitiones, on geometrical terms.

  • The Geometrica, the  Stereometrica, and the Peri metron (or On Measures); all on measurement, all revised by later editors.

  • The Geodaesia and Geoponica (=Liber geoponicus) on measurement of land.

  •  A commentary on Euclid is extant in substantial quotations in the 10th c. Arabic writer an-Nairizi (=Anaritius), which was composed in Arabic and then translated into Latin.[12]

Some extracts in French are online at[13]

Hero’s writings, apart from the Belopoeika, were published with a German translations in Heronis Alexandrini Opera.[14]  This seems to be Herons von Alexandria Mechanik und Katoptrik, herausgegeben und ubersetzt von L. Nix und W. Schmidt (Heronis Alexandrini, Opera quae supersunt omnia, Leipzig, 1900 f.).[15].  I was able to locate a few volumes online: vol.1, 1899; vol.1, supplementum, 1899; vol.2, fasc.1 (backwards!) 1900; vol.3, 1903.  But I think there are five volumes, and obviously we’re missing a lot here.

There is also an article on why Hero thought that automata could be used in temples: Karin Tybjerg, “Wonder-making and philosophical wonder in Hero of Alexandria”, Studies in History and Philosophy of Science, 34 (2003) 443-466 (abstract).

So the sum total of all of this is rather disappointing!  Few of his works have been translated into English, and even the Greek-German critical text, which is out of copyright, is not easily accessible.

It’s worth considering that Hero may have been a contemporary of the apostles, or at least late first century, although his dates are vague.

UPDATE: I find that an excellent source for these works is here, at  All the volumes of the edition are here, often based on versions at Gallica.  In addition an-Nairizi / Anaritius is also there.  The site is somewhat slow, however, but the author has done a great deal of work to make these writers accessible — well done!

UPDATE 7th April 2016:  I have learned this week of a volume containing significant English translations from the works of Hero of Alexandria.  It is Morris R. Cohen and Israel E. Drabkin, A Source Book in Greek Science, Harvard, 1958.

  1. [1]F. N. Magill and Christian J. Moose, Dictionary of World Biography: The ancient world, 2003. There is a Google books preview of the section on Hero, starting on p.514.
  2. [2]
  3. [3]
  4. [4]So Koetiser, p.198History of Technology is not a journal, but a book series. It was and is published by Mansell of London, an imprint of Bloomsbury Publishing.  A list of volumes is here.  A copy of this volume exists at Sydney University Library.
  5. [5]W. R. Laird, S. Roux: Mechanics and natural philosophy before the scientific revolution, 2008. p.197; Heron d’Alexandrie, Les Mechaniques, 1894 (Google books).  Online at here:
  6. [7]
  7. [8]
  8. [6]This annoyingly vague information from F. N. Magill, Dictionary of World Biography 1, p.514; p.517 refers to A.G. Drachmann, The mechanical technology of Greek and Roman antiquity: a study of the literary sources, Madison: University of Wisconsin Press, 1963, as containing “translations of Hero’s mechanical writings”.  This is probably what Magill was referring to.[7]; a translation of a portion in M.J. Taunton Lewis, Surveying instruments of Greece and Rome, Cambridge, 2001, p.259-62;  and a French translation by Victor Prou is online.[8]1888: online at
  9. [9]Magill and Moose,  p.517: Drachmann, Aage Gerhardt, The mechanical technology of Greek and Roman antiquity: a study of the literary sources. Madison: University of Wisconsin Press, 1963: “Contains translations of Hero’s mechanical writings, with useful running commentary.”
  10. [10]I had access to the book on Friday 25/11/11.
  11. [11]Victor Prou, La chirobaliste d’Héron d’Alexandrie, Paris, 1883: online at
  12. [12]Eleanor Dickey, Ancient Greek scholarship, p.60: “For editions and translations see Mansfield 1998, 26 n.90”
  13. [13]
  14. [14]Pamela O. Long, Openness, secrecy, authorship: technical arts and the culture of knowledge, 2001, p.258, n.56.
  15. [15]A. I. Sabra, Theories of light from Descartes to Newton, 1981, p.70, n.4