A summary of the history, construction and effects in warfare of the projectile-throwing engines of the ancients, with a treatise on the structure, power and management of Turkish and other Oriental bows of mediæval and later times

by Sir Ralph Payne-Gallwey

PREFACE

Since my recent book on mediæval archery and ancient weapons was issued,[1] I have obtained a considerable amount of information concerning the projectile engines of the Greeks and Romans. I now print a concise account of the history, construction and effects in warfare of these engines.

In this summary the additional notes I have acquired are included.

I also append a treatise fully describing that remarkable weapon the Turkish composite bow, which I only cursorily dealt with in the work referred to.

R. P. G.

Thirkleby Park,
Thirsk:
Dec. 1906.

[1] The Crossbow, Mediæval and Modern, Military and Sporting: its Construction, History, and Management. With a Treatise on the Balista and Catapult of the Ancients. 220 illustrations. Messrs. Longmans & Co., 39 Paternoster Row, London.

PART I
INTRODUCTORY NOTES ON ANCIENT PROJECTILE ENGINES

Of ancient Greek authors who have left us accounts of these engines, Heron (284–221 B.C.) and Philo (about 200 B.C.) are the most trustworthy.

Both these mechanicians give plans and dimensions with an accuracy that enables us to reconstruct the machines, if not with exactitude at any rate with sufficient correctness for practical application.

Though in the books of Athenæus, Biton, Apollodorus, Diodorus, Procopius, Polybius and Josephus we find incomplete descriptions, these authors, especially Josephus, frequently allude to the effects of the engines in warfare; and scanty as is the knowledge they impart, it is useful and explanatory when read in conjunction with the writings of Heron and Philo.

Among the Roman historians and military engineers, Vitruvius and Ammianus are the best authorities.

Vitruvius copied his descriptions from the Greek writers, which shows us that the Romans adopted the engines from the Greeks.

Of all the old authors who have described the engines, we have but copies of the original writings. It is therefore natural that we should come across many phrases and drawings which are evidently incorrect, as a result of repeated transcription, and which we know to be at fault though we cannot actually prove them to be so.

With few exceptions, all the authors named simply present us with their own ideas when they are in doubt respecting the mechanical details and performances of the engines they wish to describe.

All such spurious information is, of course, more detrimental than helpful to our elucidation of their construction and capabilities.

It frequently happens that in a mediæval picture of one of these machines some important mechanical detail is omitted, or, from the difficulty of portraying it correctly, is purposely concealed by figures of soldiers, an omission that may be supplied by reference to other representations of the same weapon.

Fig. 1.—Besieging a fortified Town with a Battery of Catapults and Balistas.

Criticism.—In this picture the balistas are fairly correct, but the catapults are too small.

From Polybius. Edition 1727.

It is, indeed, impossible to find a complete working plan of any one of these old weapons, a perfect design being only obtainable by consulting many ancient authorities, and, it may be said, piecing together the details of construction they individually give.

* * * * *

We have no direct evidence as to when the engines for throwing projectiles were invented.

It does not appear that King Shalmaneser II. of Assyria (859–825 B.C.) had any, for none are depicted on the bronze doors of the palace of Balâwat, now in the British Museum, on which his campaigns are represented, though his other weapons of attack and defence are clearly shown.

The earliest allusion is the one in the Bible, where we read of Uzziah, who reigned from B.C. 808–9 to B.C. 756–7. ‘Uzziah made in Jerusalem engines invented by cunning men, to be on the towers and upon the bulwarks, to shoot arrows and great stones withal.’ (2 Chronicles xxvi. 15.)

Diodorus tells us that the engines were first seen about 400 B.C., and that when Dionysius of Syracuse organised his great expedition against the Carthaginians (397 B.C.) there was a genius among the experts collected from all over the world, and that this man designed the engines that cast stones and javelins.

From the reign of Dionysius and for many subsequent centuries, or till near the close of the fourteenth, projectile-throwing engines are constantly mentioned by military historians.

But it was not till the reign of Philip of Macedon (360–336 B.C.) and that of his son Alexander the Great (336–323 B.C.) that their improvement was carefully attended to and their value in warfare fully recognised.

As before stated, the Romans adopted the engines from the Greeks.

Vitruvius and other historians tell us this, and even copy their descriptions of them from the Greek authors, though too often with palpable inaccuracy.

To ascertain the power and mechanism of these ancient engines a very close study of all the old authors who wrote about them is essential, with a view to extracting here and there useful facts amid what are generally verbose and confused references.

There is no doubt that the engines made and used by the Romans after their conquest of Greece (B.C. 146), in the course of two or three centuries became inferior to the original machines previously constructed by the Greek artificers.

Their efficiency chiefly suffered because the art of manufacturing their important parts was gradually neglected and allowed to become lost.

Fig. 2.—A Siege.

Criticism.—The picture is open to the spectator in order that he may see both defenders and besiegers at work.

The besieged have just cast a stone from a catapult. The stone is falling on the movable tower belonging to the attacking side.

From Polybius. Edition 1727.

For instance, how to make the skein of sinew that bestowed the very life and existence on every projectile-casting engine of the ancients.

The tendons of which the sinew was composed, the animals from which it was taken, and the manner in which it was prepared, we can never learn now.

Every kind of sinew, or hair or rope, with which I have experimented, either breaks or loses its elasticity in a comparatively short time, if great pressure is applied. It has then to be renewed at no small outlay of expense and trouble. Rope skeins, with which we are obliged to fit our models, cannot possibly equal in strength and above all in elasticity, skeins of animal sinew or even of hair.

The formation of the arm or arms of an engine, whether it is a catapult with its single upright arm or a balista with its pair of lateral ones, is another difficulty which cannot now be overcome, for we have no idea how these arms were made to sustain the great strain they had to endure.

We know that the arm of a large engine was composed of several spars of wood and lengths of thick sinew fitted longitudinally, and then bound round with broad strips of raw hide which would afterwards set nearly as hard and tight as a sheath of metal.

We know this, but we do not know the secret of making a light and flexible arm of sufficient strength to bear such a strain as was formerly applied to it in a catapult or a balista.

Certainly, by shaping an arm of great thickness we can produce one that will not fracture, but substance implies weight, and undue weight prevents the arm from acting with the speed requisite to cast its projectile with good effect.

A heavy and ponderous arm of solid wood cannot, of course, rival in lightness and effectiveness a composite one of wood, sinew and hide.

The former is necessarily inert and slow in its action of slinging a stone, while the latter would, in comparison, be as quick and lively as a steel spring.

When the art of producing the perfected machines of the Greeks was lost, they were replaced by less effective contrivances.

If the knowledge of constructing the great catapult of the ancients in its original perfection had been retained, such a clumsy engine as the mediæval trebuchet would never have gained popularity. The trebuchet derived its power from the gravity of an immense weight at one end of its pivoted arm tipping up the other end, to which a sling was attached for throwing a stone.

As regards range, there could be no comparison between the efficiency of a trebuchet, however large, as worked merely by a counterpoise, and that of an engine deriving its power from the elasticity of an immense coil of tightly twisted sinew.

It is certain that if the latter kind of engine had survived in its perfect state the introduction of cannon would have been considerably delayed, for the effects in warfare of the early cannon were for a long period decidedly inferior to those of the best projectile engines of the ancients.

Notwithstanding many difficulties, I have succeeded in reconstructing, though of course on a considerably smaller scale, the chief projectile throwing engines of the ancients, and with a success that enables them to compare favourably, as regards their range, with the Greek and Roman weapons they represent.

Still, my engines are by no means perfect in their mechanism, and are, besides, always liable to give way under the strain of working.

One reason of this is that all modern engines of the kind require to be worked to their utmost capacity, i.e. to the verge of their breaking point, to obtain from them results that at all equal those of their prototypes.

A marked difference between the ancient engines and their modern imitations, however excellent the latter may be, is, that the former did their work easily, and well within their strength, and thus without any excessive strain which might cause their collapse after a short length of service.[2]

[2] Again, though my largest catapult will throw a stone to a great distance it cannot throw one of nearly the weight it should be able to do, considering the size of its frame, skein of cord and mechanism. In this respect it is decidedly inferior to the ancient engine.

The oft-disputed question as to the distance to which catapults and balistas shot their projectiles can be solved with approximate accuracy by comparing their performances—as given by ancient military writers—with the results obtainable from modern reproductions.

While treating of this matter we should carefully consider the position and surroundings of the engines when engaged in a siege, and especially the work for which they were designed.

As an example, archers, with the advantage of being stationed on high towers and battlements, would be well able to shoot arrows from 270 to 280 yards. For this reason it was necessary for the safe manipulation of the attacking engines that they should be placed at about 300 yards from the outer walls of any fortress they were assailing.

As a catapult or a balista was required not only to cast its missile among the soldiers on the ramparts of a fortified place, but also to send it clear over the walls amid the houses and people within the defences, it is evident that the engines must have had a range of from 400 to 500 yards, or more, to be as serviceable and destructive as they undoubtedly were.

Josephus tells us that at the siege of Jerusalem, A.D. 70 (‘Wars of the Jews,’ Book V. Chapter VI.), stones weighing a talent (57¾ lbs. avoirdupois) were thrown by the catapults to a distance of two or more ‘stades.’

This statement may be taken as trustworthy, for Josephus relates what he personally witnessed and his comments are those of a commander of high rank and intelligence.

Fig. 3.—A Fortified Town being Bombarded by a Catapult.

Criticism.—The stones thrown by the besieged may be seen falling in the trenches of the besiegers. The catapult depicted is drawn on much too small a scale.

From Polybius. Edition 1727.

Two or more ‘stades,’ or let us say 2 to 2¼ ‘stades,’ represent 400 to 450 yards. Remarkable and conclusive testimony confirming the truth of what we read in Josephus is the fact that my largest catapult—though doubtless much smaller and less powerful than those referred to by the historian—throws a stone ball of 8 lbs. in weight to a range of from 450 to nearly 500 yards.

It is easy to realise that the ancients, with their great and perfect engines fitted with skeins of sinew, could cast a far heavier stone than one of 8 lbs., and to a longer distance than 500 yards.

Agesistratus,[3] a Greek writer who flourished B.C. 200, and who wrote a treatise on making arms for war, estimated that some of the engines shot from 3½ to 4 ‘stades’ (700 to 800 yards).

[3] The writings of Agesistratus are non-extant but are quoted by Athenæus.

Though such a very long flight as this appears almost incredible, I can adduce no sound reason for doubting its possibility. From recent experiments I am confident I could now build an engine of a size and power to accomplish such a feat if light missiles were used, and if its cost were not a consideration.

Fig. 4.—A Siege Catapult (without a sling).

From Polybius. Edition 1727.

PART II
THE CATAPULT (WITH A SLING)

Fig. 5.—A Siege Catapult (without a sling).

Criticism.—This engine was moved into position on rollers and then props were placed under its sides to adjust the range of the projectile.

The end of the arm was secured by the notch of the large iron catch and was released by striking down the handle of the catch with a heavy mallet.

The arm is, however, too long for the height of the cross-bar against which it strikes and would probably break off at its centre.

The hollow for the stone is much too large, as a stone big enough to fit it could not be cast by a weapon of the dimensions shown in the picture.

From an Illustrated Manuscript, Fifteenth Century (No. 7239), Bibl. Nat. Paris.

The mediæval catapult was usually fitted with an arm that had a hollow or cup at its upper end in which was placed the stone it projected, as shown above in [fig. 5].[4] I find, however, that the original and more perfect form of this engine, as employed by the Greeks and ancient Romans, had a sling, made of rope and leather, attached to its arm.[5] ([Fig. 6], following page.)

[4] See also The Crossbow, etc., Chapters LV., LVI., illustrations 193 to 202.

[5] In mediæval times catapults which had not slings cast great stones, but only to a short distance in comparison with the earlier weapons of the same kind that were equipped with slings. I can find no allusions or pictures to show that during this period any engine was used with a sling except the trebuchet, a post-Roman invention. All evidence goes to prove that the secret of making the skein and other important parts of a catapult was in a great measure lost within a couple of centuries after the Romans copied the weapon from their conquered enemies the Greeks, with the result that the trebuchet was introduced for throwing stones.

The catapult was gradually superseded as the art of its construction was neglected, and its efficiency in sieges was therefrom decreased.

The catapults of the fifth and sixth centuries were very inferior to those described by Josephus as being used at the sieges of Jerusalem and Jotapata (A.D. 70, A.D. 67), [p. 37].

Fig. 6.—Sketch plan of a Catapult for slinging Stones its Arm being partly wound down.

Approximate scale: ¼ in. = 1 ft.

The addition of a sling to the arm of a catapult increases its power by at least a third. For example, the catapult described in Chapters LV., LVI., of my book,[6] will throw a round stone 8 lbs. in weight, from 350 to 360 yards, but the same engine with the advantage of a sling to its arm will cast the 8-lb. stone from 450 to 460 yards, and when its skein is twisted to its limit of tension to nearly 500 yards.

[6] The Crossbow, etc.

If the upper end of the arm of a catapult is shaped into a cup to receive the stone, as shown in [fig. 5], p. 11, the arm is, of necessity, large and heavy at this part.

If, on the other hand, the arm is equipped with a sling, as shown in [fig. 6], opposite page, it can be tapered from its butt-end upwards, and is then much lighter and recoils with far more speed than an arm that has an enlarged extremity for holding its missile.

When the arm is fitted with a sling, it is practically lengthened by as much as the length of the sling attached to it, and this, too, without any appreciable increase in its weight.

The longer the arm of a catapult, the longer is its sweep through the air, and thus the farther will it cast its projectile, provided it is not of undue weight.

The difference in this respect is as between the range of a short sling and that of a long one, when both are used by a school-boy for slinging pebbles.

The increase of power conferred by the addition of a sling to the arm of a catapult is surprising.

A small model I constructed for throwing a stone ball, one pound in weight, will attain a distance of 200 yards when used with an arm that has a cup for holding the ball, though when a sling is fitted to the arm the range of the engine is at once increased to 300 yards.

The only historian who distinctly tells us that the catapult of the Greeks and Romans had a sling to its arm, is Ammianus Marcellinus. This author flourished about 380 A.D., and a closer study of his writings, and of those of his contemporaries, led me to carry out experiments with catapults and balistas which I had not contemplated when my work dealing with the projectile engines of the Ancients was published.

Fig. 7.—Catapult (with a Sling). Side view of frame and mechanism.

Scale: ½ in. = 1 ft.

Ammianus writes of the catapult[7]:

‘In the middle of the ropes[8] rises a wooden arm like a chariot pole ... to the top of the arm hangs a sling ... when battle is commenced a round stone is set in the sling ... four soldiers on each side of the engine wind the arm down till it is almost level with the ground ... when the arm is set free it springs up and hurls forth from its sling the stone, which is certain to crush whatever it strikes. This engine was formerly called the “scorpion,” because it has its sting erect,[9] but later ages have given it the name of Onager, or wild ass, for when wild asses are chased they kick the stones behind them.’

[7] Roman History, Book XXIII., Chapter IV.

[8] i.e. in the middle of the twisted skein formed of ropes of sinew or hair.

[9] The upright and tapering arm of a catapult, with the iron pin on its top for the loop of the sling, is here fancifully likened to the erected tail of an angry scorpion with its sting protruding.

[Fig. 7].—Catapult (with a sling), see opposite page.

A. The arm at rest, ready to be wound down by the rope attached to it and also to the wooden roller of the windlass. The stone may be seen in the sling.

The upper end of the pulley rope is hitched by a metal slip-hook ([fig. 6], p. 12) to a ring-bolt secured to the arm just below the sling.

B. The position of the arm when fully wound down by means of the windlass and rope. See also EE, [fig. 8], page 16.

C. The position of the arm at the moment the stone D leaves the sling, which it does at an angle of about 45 degrees.

E. By pulling the cord E the arm B is at once released from the slip-hook and, taking an upward sweep of 90 degrees, returns to its original position at A.

The Sling (open).

[F. Its fixed end which passes through a hole near the top of the arm.

G. The leather pocket for the stone.

H. The loop which is hitched over the iron pin at the top of the arm when the stone is in position in the sling, as shown at A and B, [fig. 7].]

Fig. 8.—Catapult (with a Sling). Surface view of frame and mechanism. Scale: ½ in. = 1 foot. The arm EE is here shown wound down to its full extent. (Compare with B, [fig. 7], page 14.)

I.	I.	}	The side-pieces.
II.	II.	}
III.	IV.		The large cross-pieces.
	V.		The small cross-piece.

The ends of the cross-piece beams are stepped into the side-pieces.

AA. The skein of twisted cord.

BB. The large winding wheels. The skein is stretched between these wheels, its ends passing through the sides of the frame, and then through the wheels and over their cross-bars. ([Fig. 12], p. 19.)

By turning with a long spanner ([fig. 6], p. 12) the squared ends of the spindles DD, the pinion wheels CC rotate the large wheels BB and cause the latter to twist the skein AA, between the halves of which the arm EE is placed.

FF. The wooden roller which winds down the arm EE. ([Fig. 6], p. 12.)

The roller is revolved by four men (two on each side of the engine) who fit long spanners on the squared ends of the iron spindle GG.

This spindle passes through the centre of the roller and through the sides of the frame.

The small cogged wheels, with their checks, which are fitted to the ends of the spindle GG, prevent the roller from reversing as the arm is being wound down. ([Fig. 6], p. 12.)

HH. The hollows in the sides of the frame which receive the lower tenons of the two uprights. Between the tops of these uprights the cross-beam is fixed against which the arm of the catapult strikes when it is released. ([Fig. 6], p. 12.)

KK. The hollows for the lower tenons of the two sloping supports which prevent the uprights, and the cross-beam between them, from giving way when the arm recoils. ([Fig. 6], p. 12.)

Fig. 9.—One of the Pair of Winches of a Catapult. Scale: 1/16 in. = 1 in.

I. Surface view of one of the winches and of the thick iron plate in which the socket of the large winding wheel of the winch revolves.

II. View of a winch (from above) as fitted into one of the sides of the frame of the catapult. One end of the twisted skein may be seen turned round the cross-bar of the large wheel.

III. Side view of the large wheel of a winch.

IV. The cross-bar of one of the large wheels. These pieces fit like wedges into tapering slots cut down the barrels, or inside surfaces, of their respective wheels.

V. Perspective view of the wheels of a winch.

The winches are the vital parts of the catapult as they generate its projectile power.

They are employed to twist tightly the skein of cord between which the butt-end of the arm of the engine is placed.

The cord composing the skein is stretched to and fro across and through the sides of the catapult, and alternately through the insides of the large wheels and over their cross-bars; as shown in [fig. 8], p. 16.

[Fig. 10]. The Iron Slip-hook.

Fig. 10.

This simple contrivance not only pulled down the arm of a catapult but was also the means of setting it free. However great the strain on the slip-hook, it will, if properly shaped, easily effect the release of the arm.

The trajectory of the missile can be regulated by this form of release, as the longer the distance the arm is pulled down the higher the angle at which the projectile is thrown.

On the other hand, the shorter the distance the arm is drawn back, the lower the trajectory of its missile.

The slip-hook will release the arm of the engine at any moment, whether it is fully or only partially wound down by the windlass.

The slip-hook of the large catapult shown in [fig. 6], p. 12, has a handle, i.e. lever, 10 inches long, the point of the hook, which passes through the eye-bolt secured to the arm, being one inch in diameter.

Fig. 11.—A Spring Engine with a Sling attached to its Arm, which cast Two Stones at the Same Time.

From ‘Il Codice Atlantico,’ Leonardo da Vinci. 1445–1520.

Fig. 12.—The Skein of Cord.

A. The skein as first wound over the cross-bars of the large wheels (shown in section) of the winches.

B. The skein with the butt-end of the arm (shown in section) placed between its halves.

C. The skein as it appears when tightly twisted up by the winches. Compare with AA, [fig. 8], p. 16.

Cord of Italian hemp, about ¼ in. thick, is excellent for small catapults. For large ones, horsehair rope, ½ in. thick, is the best and most elastic. Whatever is used, the material of the skein must be thoroughly soaked in neats-foot oil for some days previously, or it is sure to fray and cut under the friction of being very tightly twisted. Oil will also preserve the skein from damp and decay for many years.