Sie sind auf Seite 1von 23

Recent Finds of Ancient Artillery Author(s): Dietwulf Baatz Reviewed work(s): Source: Britannia, Vol. 9 (1978), pp.

1-17 Published by: Society for the Promotion of Roman Studies Stable URL: http://www.jstor.org/stable/525936 . Accessed: 04/05/2012 03:43
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Society for the Promotion of Roman Studies is collaborating with JSTOR to digitize, preserve and extend access to Britannia.

http://www.jstor.org

RecentFindsof Ancient Artillery


By DIETWULF BAATZ

primarily from several ancient technical texts. The most importantamong them are the works of Philon, Heron and Vitruvius.1Taken in combination,these texts are sufficientlydetailed to allow reconstructionof functioning artillerypieces, and the first reconstructionsof catapults were made as early as the middle of the nineteenth century, long before any actual find was known.2 The second source for our knowledge of ancient artillery consists of representationsof catapults found on Greek and Roman reliefs.3 In spite of their small number these reliefs are a significant complementto the written texts, not only because they show technical detail which might otherwise be controversial,but also because most of them can be dated within narrowboundsand in this way give some hints at the direction of development of ancient artillery.But the chancesof finding new texts or new reliefs are very small. So actual finds of ancient catapults in archaeologicalexcavations constitute a third source, which will gain increasing importance,because there are fair chances of finding remains of catapults at a number of sites. The first remains of a catapult came to light in 1912 in Ampurias(Spain). A hoard of weapons was found in an arsenal near the South Gate of the Neapolis of Emporiae, the ancient site at Ampurias. The hoard was deposited in the first half of the second centuryB.C.Among its contents were the metal parts and fittings of the timber frame of a small arrow-shootingcatapult (euthytonon, catapulta, scorpio) of typical Hellenistic design. The excavators at first interpreted these as parts of an ammunition-carriage.4 Only two years later they were recognizedby W. Barthel as componentsof a catapultand were then published with the help of E. Schramm,who had just com-

derives UR knowledge ancientartillery of

pleted a programme of catapult-reconstructions after the ancient texts.5 Schrammwas amazed to see how exactly the dimensions of the Ampurias catapult correspondedto those given in the written ancient sources.6The find also yielded some details not mentioned in the texts.7 During the First World War Schramm built a full-sized reconstruction of the Ampurias catapult which functioned quite satisfactorilyand may be regarded as the best catapultreconstructionso far built because it is based on both texts and finds.8 The example of this first artillery-findof 1912 demonstratesthe difficulties an excavator has to face in identifying remains of ancient artillery. These finds are very rare: most archaeological
1The texts are collected (with English translation) in E. W. Marsden, Greek and Roman Artillery, Technical Treatises (Oxford, 1971). A disadvantage of this edition is the omission of the ancient text-diagrams, which are quite important and must be looked for in other editions, for instance: C. Wescher, Poliorcitique des Grecs (1867). 2The first reconstructions were built by Dufour and de Reffye in France by order of Napoleon III. Photos in Archaeologia Aeliana2 xxiv (1902), 72-3, figs. 3-4. '3E. Schramm, Die antiken Geschiitze der Saalburg (1918), 30-9. SJ. Puig Y Cadafalch, Annu. Inst. d'Estudis Catalans 4 (1911-12), 672, fig. 2. 5 (1r93-r4), 841-6. 6 Schramm, op. cit. (note 3), 4o-41. SFor instance the holes in the washer and in the counter-plate, through which retaining-pins could be pushed: Schramm, op. cit. (note 3), 43-4; cf. E. W. Marsden, Greek and Roman Artillery, Historical Development (Oxford, 1969), 29, fig. 15. SThe reconstruction is still preserved in the Saalburgmuseum, Bad Homburg, Germany. For a photo, see G. Webster, The Roman Imperial Army (1969), pl.
29.

'P. Bosch Gimpera, Annu. Inst. d'Estudis Catalans

DIETWULF

BAATZ

iron lever bronze washer

0.1

0.5 1,0m

frame -bz<--

string--

arm

case
trig er

slider

- arm

frame

bronze washer

windlass

claw ,
case 0

arm \(<-slider

frame

<-bronze

washer

FIG.

I. Arrow-shooter (catapulta) of Vitruvius, designed for shooting arrows of three feet (reconstruction by Schramm). Scale 1:25.

textbooks do not mention them,9 and so some of the curiously-shapedmetal objects may be left unpublishedin forgottencornersof museums.For many decades after the Ampuriasfind of 1912 no new catapult-finds were published, until in the years 1968 and 1969 N. Gudea excavated two small late Roman forts at Gorneaand Orgovanear the Iron Gate of the Danube in Roumania.In the projectingcomer-towersof these forts some metal parts of catapults were found. Another catapult

was unearthed 1972 during excavationsnear the North Gate of Hatra (Iraq), conducted by W. I. Al-Salihi of the Directorate General of Antiquities, Baghdad. The finds of Hatra, Orgova and Gornea are the subjects of the present paper. As a preliminary,the constructionand functioning of a two-armed torsion-machinemay be des9 Even in Marsden's books, op. cit. (notes I and 7) no photos and no drawings of the very important iron fittings of the Ampurias catapult-frame are given.

RECENT FINDS

OF ANCIENT ARTILLERY

cribed. FIG. 1 shows a two-armed arrow-shooter of late Hellenistic type. These engines were powered by two springs consisting of bundles of cords made of hair, or sometimes of sinew, kept under tension. The cords of each spring were wound around an iron lever at either end of the spring. The levers in turn rested on bronze washers. Both springs were held in a timber frame, which was reinforced by metal fittings. The two wooden arms of the arrow-shooterwere inserted into the springs, and were connected by a sinew-cord, acting as a bow-string. The centre of the string was engaged by an iron triggermechanism, consisting of claw and trigger. Both parts were mounted on the slider, which could move freely in a groove in the case or stock. These parts, too, were made of timber. In order to fire the catapult, the slider was first pulled back by a windlass mounted on the end of the case. By this movement the sinew bow-string was also drawn back, and with it the two wooden arms. These, being inserted in the cord-bundles,twisted them progressivelyas the windlass was wound up, thus storing the energy for the shot in the torsionsprings. Next, an arrow was placed on the slider with its end in contact with the sinew bow-string. By pulling the trigger,the bow-stringwas released

and the arrow fired. Afterwards, the slider was pushed forward again and the bow-string hooked into the trigger mechanism, to prepare the catapult for the next shot. The two-armed stone-thrower (palintonon, ballista) was of similar construction. The main differences were the bigger dimensions of these machines and the bow-stringwhich took the form of a band. In addition the engine was normally laid at a higher angle of elevation, up to 45 degrees, since it dischargedits shot in a much higher trajectory than the arrow-shooter except when the latter was being fired at extreme range.
A. THE HATRABALLISTA

Description Just behind the second tower to the west of the North Gate of Hatra the remains of a comparatively big catapult were found buried under the debris of the adjacenttower and town-wall.1' Originally the engine must have been set on top of the tower from which it will have fallen during a siege. The tower itself (Andrae XIX), one of the
10 The find was first mentioned in Sumer 27 (1971), page g. I saw and measured the remains of the catapult in December 1975 in Mosul Museum (Iraq).

----~f---- ------------------------------------------------ ---------o1 o0o

0
0

01

/0

orO o oo o
0

O o
0

O o

00

o O l o

a o O O O
0oo

oa o O o O O O O o
0

o o.

co ooo

00

1
0

ooo

0
0

0
0

Oil

LO
0 0 a 0 0 a 0 0 0 0 0 0 0

01O

0O

0
0

O 0 0

l 0 0 0

FIG. 2. Hatra ballista. Metal parts of the frame, schematically drawn. Scale I: 2o.

DIETWULF

BAATZ

B
FIG. 3. Hatra ballista. Two different types of corner-fittings (bronze). Scale 1: 75.

series of massive solid stone structures built as additions against the existing town wall, belongs to the last phase of the town's defences (PL. I)."1 As Hatra was destroyedin the middle of the third century A.D., this gives an approximatedate for the fall and burial of the catapult. Only the framewhich once held the two torsionspringswas found at Hatra (PL. II-IV A ). No trace survived of the wooden slider and the case with its windlass, or of the base on which the machine was mounted.FIG.2, which is based on the actual finds of metal fittings and on the excavation photos, gives an idea of the constructionof the frame. It was c. 2-40 m wide, 0.84 m high (without corner-fittingand washers)and 0-45 m thick. The frame proper consisted of several different pieces of wood morticed together; but, except for small fragments,nothing of this remainedand the dimensions given above are therefore only approximate. There could also have been more transverse pieces between the four long main beams. Analysis of the surviving fragments of
timber proved this to be pterocarya fraxinifolia,

sion-engines, the purpose of these was to accommodatethe two armsof the machine.Similaropenings have been noted in the frame of the Ampurias catapult,'3 and they are also mentioned in the ancient written sources.'4 All eight cornersof the wooden frame will have been fixed with elaborate joints. Each joint was protected and held together by heavy fittings of
cast bronze nailed to the frame (FIG. 2 and PL. IV A). The fittings are of two types, each roughly the mirror-imageof the other (FIG. 3). Such fit-

tings have not, so far, been recordedfor any other type of ancient artillery and are a peculiarity of the Hatra catapult. The most characteristicparts of any piece of ancient torsion-artilleryare the washers set into The Hatra catapult originallyhad torsion-springs. four washers with four levers. Only three have been found, though all four counter-plateswere recovered. The washers were made from cast was 17-5 cm at the bottom and 16 cm at the top;
bronze (PL. III and FIG. 4). Their inner diameter counter-plates, and the levers used to tighten the

a tree growing in mountain regions from Asia Minor to Northern Iran.12 The frame (main long timbers and two side stanchions)was covered on the front and sides by sheets of bronze c. 2 mm thick, fixed by nails. Especially to be noted are the two half-round openings in the side-stanchions.As in other tor-

" W. Andrae, Hatra II. Wiss. Ver6ffentl. Deutschen Orient-Gesellschaft 21 (1912), 30 ff. 1 I am indebted to Dr M. Hopf (Mainz) and to D. Eckstein (Hamburg) for the identification of species. 13Schramm, op. cit. (note 3), 44-5, figs. 16--17 4 Heron, Belopoeica 91-92 (ed. Marsden, note Vitruvius x I11,5.

d. Dr the

I);

RECENT FINDS

OF ANCIENT ARTILLERY

were arranged differently.15In the top of each washer were two notches for holding the iron levers. The notches were 3 cm wide and 1.6 cm deep. Below the notchesan internalreinforcementrib was cast in one piece with the washer. These ribs are not mentioned in the ancient sources,nor were they present in the small washers of the Ampurias catapult, but their use in the big washers of the Hatra engine is easily explicable because of the heavy strain put on the washer at this point by the lever. The iron levers were deeply corroded, so no exact dimensions could be taken (PL. IIIA). But the width of the notches on top of the washersis known (3 cm), and this gives the width of the levers. The levers seem to have been about 29 cm long and possibly 6-8 cm high. No remains of the cords (torsion-springs) running around the levers were found. The counter-plates (PL. III and FIG. 5) were square, having sides 29.0-30-5 cm long. They were made from cast bronze 6 mm thick and each plate was fixed to the frame by four nails in the 10cm

10cm 5 1

FIG. 4. Hatra ballista. Bronze washer. Scale I:4.

the outer diameter was 28 cm. Set into the counter-plate,the washer was I1 -2 cm high (from surface of counter-plate to top of washer). The total height of the washer, including the circular rim going into the central hole of the counter plate, was I2-2 cm. In the wide outerflange of the washer there are 16 holes, each hole having a diameter of 1 cm. The counter-plates have 8 holes, arranged in 4 pairs. Retaining-pins could be pushed through the holes in the flange of the washerinto the corresponding holes in the counterplate, so holding the washersin position once they had been turned sufficiently to give the springs the requisite tension. In the holes of each washer and its counter-plate were traces of up to three iron retaining pins. One of them is visible on PL. III A; also on PL.III B, foreground.The Ampurias catapult was provided with a comparable set of holes and retaining-pins,although there the holes

FIG.

5. Hatra ballista. Bronze counter-plate.Scale x1:4.

15 Schramm, op. cit. (note 3), 43-4; Marsden, op. cit. (note 7), 29-30o.

DIETWULF

BAATZ

corners.The eight holes for holding the retainingpins passing through the flanges of the washers have been mentioned already; they were arranged in four pairs. The circular rim on the bottom of the washer fits neatly into the central opening of the counter-plate, but not tightly, so that the washer can be turned easily in the opening. In order to fix the counter-platemore securely, the sides of the central opening extend downwardlike a cylindricalsleeve to fit into the hole in the timber frame. Inclusive of this sleeve the counterplate is 3-5 cm high. The diameter of the central hole is 20.9 cm on the upper flat side of the plate, upon which the washer rests. The internal diameter at the bottom of the sleeve is slightly less, being only 20.6 cm. All four counter-platesare bent a little so that the cornersare 1-2 mm higher than the edge of the central opening. This may have been caused by the pressureexerted on the washersby the torsion-springs.Counter-platesfor washers are mentioned by Heron and Philon.16 In order to pull back the arms of the catapult there must have been both a wooden case with windlass, and a slider, similar to the corresponding parts of the arrow-shooter(FIG. I). The case was always attached to the centre of the two lower long main beams joining the bottom of the frame. Nothing remained of the case, but exactly in the centre of two of the long beams of the frame (which were certainlythe lower pair), were found strong iron bolts and two heavy iron bands (PL. II A, left side of the frame). In all probability they fixed the case to the frame. If this is right, the machine will have fallen with its front in the ground and its case or stock sticking up in the air. The correctnessof this assumptionis strengthened by the observationthat only one side of the frame,certainlythe front, was protectedby bronze sheets and this was on the bottom when excavated. (PL. II A). This would also explain the absence

and forwards by means of pulleys or other devices17should pass over rollers. 2. Rectangularbronze sheet, consistingof three pieces nailed together side by side on a piece of wood now lost. The dimensions overall are 41.5 cm wide, 49-5 cm long and 0-3 cm thick. The nails were placed aroundthe periphery.As one excavation photo shows, the sheet was found under the ballista, and not in contact with the frame, so it may not be part of the machine. 3. Two smaller rectangularbronze sheets with a central rectangular hole. The length of the sides are 14-3 (14-7) and 13.0 (13-3) cm (dimensionsof

FIG.

6. Hatra ballista. Bronze roller with iron axle. Scale I:2.

of the case, and the metal parts of the windlass at its end, among the finds. There was a number of other finds made nearby and more or less certainly connected with the catapult: 1. Five sturdy bronze rollers with iron axles (FIG. 6). They were 3-4 cm long and 4-8 cm in diameter. Their use is suggested by the written sources, which state that ropes used for pulling backwards the sliders of the larger stone-throwers

second sheet in parentheses).The central hole is 4'9 by 4o0 cm. Eight nail-holes are set round the periphery of the sheets, so these sheets also were once mounted on wood. Their position on the ballista is not known. 4. Various pieces of iron, all of them heavily corroded.One, in the form of a hook, was possibly the claw of the trigger-mechanism,as it shows traces of an iron axle.18There are other bits of iron which may also have been parts of the catapult: nails, cramps, fittings, and a piece of iron
* Heron, Belopoeica 97 (ed. Marsden, note I), Philon, Belopoeica 57 (ed. Marsden). 17 Heron, Belopoeica 84-85 (ed. Marsden, note I). i i (ed. Marsden, note '8 Heron, Belopoeica 76 and I).

RECENT FINDS

OF ANCIENT ARTILLERY

sheet with two rods holding a retaining-pin.The interpretationof all these parts is much impeded by the deep corrosion. Other iron objects were definitelynot part of the engine (a shield-bossand three large bolt-heads of different sizes). Furthermore, in the Hatra room of Mosul Museum there are a number of bronze finds on display, which must be the componentsof another catapult: two cast bronze corner-fittings of a type similar to FIG. 3 and two bronze rollers resemblingthe rollershownin FIG. 6. Both cornerfittings and rollers are smaller than the Hatra finds so far discussed, and represent a smaller piece of artillery, but presumably constructedin exactly the same way. Characteristicsand Calibre As stated above, the Hatra catapult was a twoarmed torsion-engine.Because of its size the machine must have been a stone-thrower.The Latin term for a two-armedtorsion stone-thrower,used for instance by Vitruvius, was ballista.19As no other ancient stone-throwerhas so far been found, our knowledge of these machines is derived from the ancient literary sources, especially the works of Heron, Philon and Vitruvius.20 These writers describe,with minor differences,the type of stonethrowerthat was developed during the Hellenistic period. They give special calibration-rules and formulae which allow one to calculate the dimensions of the engine suitable for firing a missile of given weight, the measurementsof all parts of the ballista being given as multiples or fractions of a specific module. Converselyit is possible by using the same formula to calculate the appropriate missile-weight for a particular ballista from its dimensions. These formulae are, of course, only valid for the Hellenistic type of stone-thrower described by the authors mentioned above. The Hatra ballista, however,was built centuries after the period of these authors,and, whetherbecause of technical progress or of adaption to special conditions, represents a type differing in many respects from the machines described by them. The frame has completely differentproportions, and is both wider and lower than the frames of the Hellenistic stone-throwers.Again, details of constructionvary considerably.There was for instance no peritretos, a wooden sub-component upon which each counter-plate and washer were

mounted within the frame. The torsion-springs, too, must have had very different proportions, being much shorter but thicker than the springs of Hellenistic stone-throwers.The ratio of diameter to length of a torsion-springis 1: 87 in the Vitruvian ballista, but only 1: 67 for the Hatra stone-thrower (the length being reckoned as the distance from the top of the washer to the top of its opposite number,the diameteras being that of the hole of the washer). All these differencesprevent us from applying the calibration-formulae given by the ancient sourcesto the Hatra ballista. it Consequently is difficultto determinethe weight of missile for which this ballista was designed. One possibility of finding out the shot-weight of the Hatra ballista is to comparethe volume of to one of its torsion-springs that of a torsion-spring of a Hellenistic stone-thrower. As the torsionsprings store the energyused to propel the missile, so they determine the power of the engine to which they belong and, accordingly, the appropriate missile-weight. The volume of a torsionspring is directly proportionalto the energy it is able to store.21One torsion-springof the Hatra ballista will have had a volume of 21-7 litres (length 108 cm, diameter 16 cm). Looking throughthe list given by Vitruviuswe find the tenpound stone-thrower having torsion-springs of 22-2 litres each22 which is very close to the springvolume of the Hatra ballista. So we may conclude that the power of the Hatra ballista was approxi, mately equivalentto that of the ten-poundballista of Vitruvius (10 Roman pounds = 3-27 kg). In antiquity this was a medium rather than a large stone-throwerand machinesof this size must have been common. The ten-pound stone-throwerof Vitruvius was built accordingto a module of 8 digits (equal to half a Roman foot = 14-8 cm).23 All measure1 In later Latin (for instance in Ammianus and Vegetius) there was a change of terminology, the arrowshooters appearing now as ballistae and the stonethrowers as scorpiones or onagri. Behind the change of terminology there was also a change of construction, cf. Marsden, op. cit. (note 7), i88 f. 20 See note i. 21In comparing the torsion-springs of two different catapults we assume the quality of the hair-rope was about the same. 2"Length 8-7 modules, diameter I module; the module being I Roman foot (which is 14-8 cm), this gives a length of 128.8 cm and a diameter of 14.8 cm. " Vitruvius x. 11,3.

DIETWULF BAATZ

S1

2m

FIG.7. Hatra ballista, Tentative reconstruction.Scale I:30.

ments of the ballista are given as multiples or fractions of this module. In spite of all the differences of constructionsome measurementsof the Hatra ballista seem to be identical with the corresponding measurements of the ten-pound ballista of Vitruvius. This is the case with the lever, which should have a width of 1/5 module.24One-fifth of half a Roman foot comes to 2-96 cm. As we have noted, the lever of the Hatra ballista must have been 3 cm wide. The height of the washer of should be three-quarters the module according to Vitruvius,which gives a height of 11-1 cm. The cm high. washerof the Hatra ballista is 11.2 The similarity of these measurementsconfirms our opinion that the Hatra ballista was designed for shooting stones of ten Roman pounds. It also indicates a connexion between the artillery described by Vitruvius and the Hatra ballista. Possibly the Hatra ballista was built after a set of formulaeand instructionsdeveloped from the text of Vitruvius. Furthermorethe dimensions of the mulframe of the Hatraballistaare approximately of the Roman foot: the frame is roughly tiples 8 feet long, 3 feet high and one and a half feet deep, while the centres of the torsion-springsare

6 feet apart. The Hatra ballista, therefore, was probably a Roman machine or at least a ballista built under Roman influence.
Reconstruction of the Hatra ballista

FIG. 7 gives a side view of the ballista as it may have looked. On the front side the frame was covered by bronze sheets, possibly as a protection against fire-arrows.The cord-bundle of one torsion-springis visible as well as one of the wooden arms that has been insertedinto it. The stock with slider and windlass, and the base (none of which survive) have been reconstructedaccordingto the precepts given by Philon, Heron and Vitruvius. The dimensions of these parts are those of the ten-pound ballista of Vitruvius. The overall dimensions of the ballista are shown clearly by comparingit with the man who is using the windlass to draw back the slider.
Ancient Artillery in Hatra

Founded probablyin the first century B.C.,Hatra remained for a long time the centre of a semi24The width of the lever is given only by Philon, Belopoeica 53 (ed. Marsden, note i).

RECENT FINDS

OF ANCIENT ARTILLERY

independent desert kingdom. The city even succeeded in defending her walls (PL. I B) against two Romanemperors,Trajan(I17) and Septimius Severus (198/99), despite the use of the highly advanced siege-machinesof the Roman army, including of course artillery. The sources do not describe the first siege in any detail but we are better informed about the assault of Septimius Severus. Cassius Dio also describes the powerful artillery of the Hatreni.25Their catapults fired at very long range, even striking many of the imperial bodyguard. Some of their machines discharged two missiles at a single shot. The use of artilleryby the defendersof Hatra is both remarkable and uncommon,for there are almostno references to this sophisticatedweaponbeing employed outside the Roman army during the first centuries A.D. Possibly the Hatrene catapults were built under the supervision of a Roman engineer who had a contractwith the city or who for some reason had taken refuge there. Dio also mentions the very effective 'flamethrower' of the Hatreni, which hurled burning bituminous naphta against the Roman siegemachines, destroying nearly all of them except those built by a certain Priscus,who had previousThe Hatreni ly been an engineer in Byzantium.26 may also have used catapults to shoot burning naphta carried by fire-darts,27and Priscus may have covered his machines with bronze sheets similar to the sheeting of the Hatra ballista as a protection against fire. The weapons used in the later sieges of Hatra by the SassanidsArdashirI and Shapur I are not may have causedher to incline towardsRome, and possibly as early as during the eastern war of the Emperor Severus Alexander in 232/33 a Roman auxiliaryunit was stationedin the city. At any rate two Latin inscriptions found there give positive proof of the presenceof cohors IX Mauretanorum have been equippedwith artillery.In the first and second centuriesA.D.Roman auxiliary units were not normally thus armed, only the legions being provided with torsion-weapons;but from the first half the the third century onwards there are examples of auxiliary cohorts using artillery.29 Furthermorethere is the possibility that a legionary vexillation was sent with its artillery as a reunder Gordian III (238-244).28 This cohort may recorded. The attack of Ardashir I (A.D. 227?)

inforcementto Hatra during the late stages of her struggle against the Sassanids. In summary, the remains of torsion-artillery found at Hatra could be those of a Hatrene weapon, such as are attested by Dio; but a Roman origin cannot be excluded, especially in the present case, as the ballista discussed here clearly belongs to the last phase of the life of the city.

B. THE OR?OVA FIND

Description

During the excavation of the late Roman fort at (Roumania) N. Gudea discovered two Orsova large iron objects in the projecting south-western corner-tower.30 The two objects were found side by side in a destruction-layerof the end of the fourth century. 1. The first consisted of two heavy rings connected by two iron beams (FIG. 8 and PL. V). Both

rings show an extension where the beams were rivetedto them. One of the two beamswas strongly curved. In the curve the beam was especially reinforced. Furthermorethe beams carriedtwo iron loops each, which were of different size. Each of the heavy rings was pierced by four small round holes. Two more holes, probablyrectangular,may be conjecturedfor riveting the two beams to the ring, but because of corrosionthey were not discernible. The find was bent and twisted to a certain degree during the destructionof the fort, but in spite of this it seems possible to suggest that the iron beams and loops fastened to them were not arrangedradially to the ring as was once sup"

Cassius Dio 75,11. 26Cassius Dio 74,11 and 75,11. 27 Fire-arrows (in Latin: malleoli) filled with bitumen are mentioned by Vegetius Iv, 18; see also Ammianus Marcellinus xxIII, 4, 14. Both sources date to the fourth century. 28D. Oates, Sumer 11 (1955), 39 ff.; A. Maricq, Syria 34 (1957), 288 ff. 29 Baatz, Bonner ahrbiicher 166 (1966), 194 ff. An example of an auxiliary cohort being equipped with artillery is the cohors I fida Vardullorum at High Rochester: RIB 1280-81.
30N. Gudea, Saalburg-Jahrbuch 31 (1974), 50-59. The finds are preserved in the Institutul de Istorie pi Arheologie at Cluj Napoca (Roumania).

10

DIETWULF

BAATZ

18

4 37

27

--45

SA

IB
20cm
I

o
I

5
I

10
I

15
I

0 O 479

O
O

FIG. 8. Orpova. Field-frame (KafL3io-,rptov) of catapult, reconstruction-drawing. Below: the bottom ring seen Scale I:4. from the upper side. Dimensions given in numm.

IC

RECENT

FINDS

OF ANCIENT

ARTILLERY

II

dimensionsof the find may be seen on FIG.8. Because of corrosionthey are only approximate.The weight of the object is roughly 8 kg. 2. The second find was a long iron rod forged to an arch in the middle. The two ends of the rod
were forked (FIG. 9 and PL. II B). On either side of

posed.31 The correct arrangementis indicated in FIG. 8C, and is also visible on PL. V. The main

are termed Ka

the arch was a small round hole going through the rod. All four ends of the two forks were broken, but at one of the broken ends a single small rectangularhole was preservedintact and apparently the beginning of another. The object was
considerably bent and twisted (PL. II B). The

rowed from the Latin word campestria.Marsden's translation, which will be followed here, is field frames.35The field-frameswere to hold the cordbundles of the two torsion-springsof the engine. For this purpose each field-frame,which the text says is to be made of iron, had to carrytwo bronze washers, each with an iron lever. It follows that the two field-frames were equipped in all with four washersand four levers, and these are shown
on FIG. 10,1 to the right.

which seems to be borE(o-rTptUa,

To hold the two field-frames in position, two iron struts were needed. They are called in the
text

present overall length is 145 cm but with the broken ends restored it would have been longer.
The distance a-b is 124-5 cm (FIG. 9). Interpretation

(ladder strut). On one of the ancient diagrams


(FIG. 10,2) these are also represented. Both struts

Ka/LaptOV (arched

strut) and

KXLkLKLOCV

had forked ends. These ends were inserted into the iron loops fastened to the iron beams of the field-frames. In this way an all-metal frame was
provided for this type of artillery. On FIG. ii a

With the exception of the Gornea finds (see below), no parallel to the Oryovaobjects seems so far to have been found at any Roman site. Interpretation was stimulatedby the works of the late E. W. Marsden, who drew attention to Heron's Cheiroballistra,a comparatively dark and difficult technical text.32 Refuting R. Schneider33 Marsden was able to show that the text describes components of a small two-armed torsion-engine of a type akin to artillery on Trajan's Column which were omitted in Marsden's edition of the Cheiroballistra,decisively indicated the function of the Orsova finds.34FIG. 10 gives copies of two of the ancient diagramsillustratingthe text of the shown two objects which are very similar to the first Orqovafind. In the accompanyingtext they
Cheiroballistra. On the left side of FIG. 10,1 are (PL. IV B). But only the ancient text-illustrations,

is reconstructionof the cheiroballistra given which indicates how the components may have been connected; it is discussed below.
31 The first reconstruction-drawingof the find, in Saalburg-Jahrbuch31 (i974), 56, fig. 8, showed the iron beams and loops standing radially to the ring. After further examination of the original the reconstruction-drawing given here (FIG.8) seems to be more probable. 32 Marsden, op. cit. (note I), 206-33. 33 R. Schneider, R6mische Mitteilungen 21 (1906),
34Photos of the text-illustrations can be found in Schneider, op. cit. (note 33); see also the drawn copies in the edition of C. Wescher, op. cit. (note I), 123-34. On the question of the ancient text-diagramsgenerally: R. Schneider, 'Geschiitze auf handschriftlichen Bilzum Yahrbuchder Gesellschaft dern', Ergdinzungsheft fiir Lothringische Geschichte und Altertumskunde 2 (1907). 3"Marsden, op. cit. (note I), 222, No. I6.

142-68.

oi

0oCb
L 10 I I I I 50cm I

0 FIG.

9. Orgova. Arched strut (KaCtqaptov) of catapult, reconstruction-drawing. Scale, 1:

12"5.

12
A-1x.

DIETWULF
2? H A

BAATZ

S -s FI Z AH

Tr.

IT

2~

&

Ii

13

Vi

FIG.

IO. Medieval copies of ancient text-diagrams from Heron's Cheiroballistra (after C. Wescher, note 34).

I. Two field-frames, four washers and four levers.

2. Arched strut (above) and ladder-strut (below), the latter being shown in two projections combined in one diagram.

Clearly the arched strut with the forked ends shown on the ancient text-rillustration (FIG. 10,2) correspondsto the second Orgova find (FIG. 9). So the Orgova finds may be interpreted as one field-frame and the arched strut of a late Roman torsion arrow-shooter. All other parts of the engine are lost. The reconstructionof the cheiroballistra(FIG. 11) may also explain why two of the four iron loops of the Orgovafield-framewere larger. They must have been the bottom loops, which were to receivethe forkedends of the ladder-strut.In contrast to the two upper loops, the bottom ones had to support not only the entire weight of the torsion-springincluding the arm, but also that of the archedstrut above. In the text and illustrationsof all Heron's Cheiroballistra four loops are of equal size. Furthermorethe four round holes present in each of the two rings of the Oryovafield-frame (FIG. 8C) may be interpretedas having the same purpose as the four pairs of holes in the counter-

plates of the ballista from Hatra (FIG.5) or the 16 holes in each counter-plate of the Ampurias catapult (note 15). The function of the rings was to act as counter-platesfor the lost washersof the catapult. Through these holes, and through corresponding ones in the washers, retaining-pins could be pushed to keep the washersin place. The holes are not mentioned in the text of Heron's and Cheiroballistra not shown on the text illustration (FIG. IO.I).

For an illustrationof the Orgovatype of arrowshooter, which was considerably larger than the cheiroballistra, we may use the artillery-pieces

strut is a typical component of all of them. Both field-frames are visible in the form of cylindrical elementsat the sides of the archedstrut. The fieldframes proper seem to be covered with metal sheetingas a protectionagainstweatherand damp. The arrow-shootersappearing on Trajan's Column are light ones, being served normallyby two

shown on Trajan's Column (PL. IV B). The arched

RECENT FINDS

OF ANCIENT ARTILLERY

13

---

---__~._~ic~c'42_~C ~cc~L~c`-;~,l
=i '~_Cz--si;l`~;;~~ Q -,? 3~"~ILYphZL,'I~, ._ ~~?CI ~Cs'CsC~4~,-=~,'z-=_1

FIG. I I. Reconstruction of Heron's Cheiroballistra.

men. Most of the catapults on the Column are mounted on a carriage (PL. IV B); some others
are standing on a base similar to that on FIG. 1.36

and Calibre Characteristics The most important characteristicof the Or?ova type of artillery is the occurence of an all-metal frame holding the two torsion-springs.In this respect the weapon was superior to all Hellenistic catapults,which had timber frames only. The allmetal frame must have had a number of advantages over the timber frame. First, the metal joints of the frame were more resistant than timber joints to the shock and vibrationscaused by firing the weapon.37 Second, a metal frame is not subject to variations of humidity. Such variations may cause a timber frame to twist and crack,rendering the weapon useless. Third, the metal frame gave the opportunity to use replaceable field-frames. As the reconstructionof the cheiroballistra(FIG. 11) shows, there is no difficulty in drawing the field-framefrom the forked ends of the two transverse struts. As the Orgovafind proves, there were holes in the forked ends of the arched strut, presumably for holding retaining-pins. To remove the field-frame, one only had to draw out the retaining-pins. A catapult of this type, provided with complete spare field-frames, including the cord-bundles of the torsion spring, could be re-

paired very quickly if a torsion-springgave way. Another characteristic of the weapon is the archedstrut. In combinationwith the wide spacing of the field-framesthis provided a useful opening in the frameworkof the catapult, through which the ancient artillerymancould observe the field and spot his target.37a Perhaps the arch was also necessary in connection with a sighting device, but at present there is no positive proof that sights were used. Like the Hatra ballista the Orgovacatapultwas not built following the construction-rules and calibration-formulae laid down in the works describing late Hellenistic artillery.38So these formulae cannot be applied to the Oryovacatapultcomponentsin orderto calculatethe calibreof the engine. But the inner diameter of the two rings of the field-frame is equal to the diameter of the torsion-spring, which must have filled the ring. The diameter amountsto 7-9 cm. Strange to say, the Ampurias catapult had the same spring-diameter.39The height of the spring is given by the
3"For instance, C. Cichorius, Die Reliefs der Traianssiiule (I896-i9goo), pl. 47. "3Philon, Belopoeica 57 (ed. Marsden, note I) illustrates the difficulties caused by the use of timber frames. 37a Marsden, op. cit. (note I), 227 f. 38 See note i. 3 Schramm, op. cit. (note 3), 40.

14

DIETWULF

BAATZ

height of the field-frame (36 cm) plus the height of the two washers. As the washers are lost, the overallheight of the springsof the Or?ovacatapult is not known.The frame of the Ampuriascatapult was 41-5 cm high, the washers a little less than 4 cm each, giving the torsion-springsa height of 49 cm.40We may conclude that the power of the Oryovacatapultmay have been equal to or a little less than that of the Ampurias catapult.41This was a light arrow-shooteronly, but to cock the weapon a windlasswas needed. In this respectthe Oryovacatapult differed from the cheiroballistra, a much smallerweaponnot providedwith a windlass. The frame of the Or?ovacatapult had completely different proportionsfrom those of the Ampurias catapult or other Hellenistic arrowshooters, and is much wider but lower than the frames of the older machines. The wide, low frames are also a typical feature of the artillery on Trajan's Column. Similar proportions were already noted above in describing the frame of the Hatra ballista. So the wide, low frames seem to be a characteristicof a number of later ancient artillery-types. Chronology In comparing the Orgova finds with Heron's Cheiroballistraand the artillery on Trajan's Column one has to bear in mind the chronology of these very differentsources.Trajan's Column was erected in 113 to commemoratethe Dacian wars. The Column gives the earliest representationof this particular type of artillery.42The Or?ova finds were deposited at the end of the fourth century, nearly 300 years later. The date and authorship of Heron's Cheiroballistra are disputed. Marsden identified the author with the Heron of Alexandriawho wrote the Belopoeicaprobably in the first half of the second century A.D.43But R. Schneiderhad alreadylaid stress upon the use of different terms in the Belopoeica and in the Cheiroballistrafor the same artillery-component, for instance in describingthe claw of the triggerThis points against the same author mechanism.44 written both technical texts. Schneider having also noticed words borrowed from Latin in the for Greek text of the Cheiroballistra, instance the word for 'field-frame' (see above) or the title itself, which is clearly a translationof manuballista.

The Latin word occurs in late Roman texts only.45 Consequently Schneider concluded the Cheiroballistrato be a Byzantine text.46 To the present writer the arguments seem to support Schneider'sopinion. If this is right, the artillerytype first representedon Trajan's Column was in use over many centuries well into the Byzantine period. During this long span of time some sort of developmentand of course many variationsare to be expected. One of these variations was the cheiroballistraor manuballista (FIG. 11), a sort of torsion-crossbow. This small weapon was cocked without a windlass in a way similar to Heron's gastraphetes47and was served, as the name indicates, by one man only.
THE GORNEA FIND

Three iron objects found by N. Gudea in the late Roman fort at Gornea (Roumania)may have been components of such manuballistae.Two of them were discovered in the projecting south-eastern corner-tower, the third in the south-western The fort was of the same type as corner-tower.48 at Or?ova, and the finds here too came from a destruction-layerof the end of the fourth century
A.D.

The three objects are of similar type, so only No. 2 is representedhere on FIG.12. Fundamentally the constructionof the objects is the same as that of the Or?ova field-frame. Consequently the Gornea finds are interpreted as field-frames also. They differ only in being much smaller and
Dimensions taken from the drawing in Schramm, 0o op. cit. (note 3), 44, fig. 16, and from Schramm's reconstruction (note 8). S1 The power may have been equal if the washers of catapult were higher than the Ampurias the Orlova washers. The same assumption is made as is expressed in note 21. 42 Marsden, op. cit. (note 7), 188-90. " Marsden, op. cit. (note I), 209-10. ( Schneider, op. cit. (note 33), 164-5. See also the review of Marsden's book (note I) by A. G. Drachmann in Technology and Culture 13 (1972), 492-3. de rebus 4 Vegetius, mil. 2,15; 4,22; Anonymous, bellicis (ed. R. Schneider, Berlin, 19o8), p. 19; also manuballistarius: Vegetius, mil. 3,14; 4,21. 46 Schneider, op. cit. (note 33), 167-8. 4' Schramm, op. cit. (note 3), 16, fig. 3; 47-9; pl. I); Marsden, op. cit. (note 7), 5-12. 48 N. Gudea, op. cit. (note 30); id., 'Gornea', BanaNos. 20-22; figs. 46a, 59. The finds are now preserved in the museum at Repija (Roumania).

tica, Studii ?i Cercettln Arheologice (I977), 82-3,

RECENT FINDS

OF ANCIENT ARTILLERY

15

FIY. 12. Gornea. Field-frame (Kafi7o-rptov) of manuballista.Scale approx. 1:2.

lighter than the Oryovafield-frame. The Gornea field-frameNo. 2 had a weight of 426 g after derusting. The total height was 14-4 cm. The two rings at the top and bottom had an inner diameter of 5-9 cm and an outer diameter of 8-5 cm. The cross-sectionof the iron band of the ring was 1.3 by 0-4 cm. Each ring was pierced by four small round holes and by two larger, rectangularholes, the latter being necessary for riveting the two beams to the ring. The distance between the two beams was 6.9 cm, their cross-section roughly 2.0 by 0.7 cm. One of the beams was curved,and in the curve the cross-sectionwent up to 3-5 by 0.7 cm. The upper, smaller loops fastened to the beams had an inside width of 2-0 by 1-6 cm and the two largerbottom loops 3-1 by 2.0 cm. Some of the dimensions correspond to those given in Heron's Cheiroballistra. For instance, the distance between the beams is given in the Cheiroballistra as 31 dactyls, equal to 6-5 cm49 (Gornea No. 2: 6.9 cm). Also the cross-sections of the iron beams and rings are comparable. Therefore the dimensions of the Gornea finds may be used the other way round to clear up the question which of two different text-traditions of a dimension of Heron's Cheiroballistra may be

the more probable one. The height of the fieldframes is indicated in the text by the length assignedto the two beams.Wescherand Schneider preferredto accept 10- dactyls (19-4 cm),50Marsden followed the other reading of 20 dactyls (37 cm).51 Consequently the reconstruction of the cheiroballistrabuilt by Marsden acquired nearly double the height it would have had if built to the dimension given in the first reading.52The machine accordingly was much more powerful and needed to be tightenedby a windlass,which is not mentioned in the text. Marsden even disregarded the part of the text describing a crescent-shaped timber placed at the end of the case where the
" Heron, Cheiroballistra 128 (ed. Marsden, note I). The conversion of dactyls into cm is based on the assumption that the dactyl of Heron's Cheiroballistra is derived from the Roman uncia: I? dactyls = I uncia, cf. Schneider, op. cit. (note 33), I65, n. 2; Marsden, op. cit. (note I), 200, n. 21; 228. If one Roman foot equilk 29-6 cm and 12 unciae go to the foot, one dactyl is equal to I185 cm. SoWescher, op. cit. (note I), 128, n. 5; Schneider, op. cit. (note 33), 154, n. 2. op. cit. (note I), 222 f., n. 17. .1Marsden, cit. (note I), 224, fig. 9; 232-3; pl. "2 Marsden, op. 6-8.

16

DIETWULF

BAATZ

But windlassis mounted in bigger catapults.53 another weapon of antiquity,which had some similarity with a mediaeval cross-bow, namely Heron'sgastraphetes,was also providedwith such a crescent-shapedtimber to enable one man to tighten the machine without a windlass.54The height of the Gornea field-frame No. 2 at 14.4 cm gives the lower number of 101 dactyls the greater probability, and leads to a different reBut the differencesbetween the field-framesof Heron's Cheiroballistra and those at Gornea should not be overlooked. The Gornea fieldframes had for instance much larger loops fastened to the beams. The size of the loops points to the possibility of timber traverses being employed to hold the field-framesin position instead of the iron struts of Heron's Cheiroballistra. This may explain why only the field-frameswere found at Gornea. Clearly a variety of types of manuballistae existed. Conclusion Nearly all research on ancient artillery has been concernedwith the main written technical sources (Philon, Heron and Vitruvius). The most devel oped artillery-pieces described by these sources reflect the state of technique reached in late Hellenistic times. The Roman army had introduced artillery as early as in the third centuryB.C.Over the following centurie3 up to the late Republic the Romans remained dependent upon the technical skill of Greek engineers.Not long after the Roman army became the only outstandingmilitary power of the Old World the necessity of having its own independent artillery-production must have been felt. This is reflected by the career of Vitruvius56 and the chapter on artillery which he wrote. The text of Vitruvius includes what are perhaps the first Roman innovationsin artillery.57As has been noted by Marsden, the last representation the Hellenistic type of twoof armed torsion-artillery is to be found on the tombstone of Vedennius, which can be dated to the end of the first centuryA.D.58Only a few years later, during Trajan's Dacian wars, a completely new development of arrow-shooting artillery is
seen in action (PL. IV B). construction (FIG. I I).55

ment in ancient artillery. All new finds show influences from the type of arrow-shooterrepresented on Trajan'sColumn. In spite of all differences of constructionthe Hatra ballista (third century) and the Orjova arrow-shooter (fourth century) both, for instance, had very low but wide frames, in contrast to the narrow high frames of the Hellenistic torsion-engines. The new type of artillery had probably already been developed in the second half of the first century A.D. and la3tedfor centurieswell into the Byzantineperiod. Only few indications of these weapons are to be found in contemporarywritten sources, the most important being Heron's Cheiroballistra,a late Roman or Byzantine text.59
Heron, Cheiroballistra124 (ed. Marsden, see note "3
Technology and Culture 13
493.

I); A. G. Drachmann,
(1972),

"JIn reconstructing the cheiroballistra some points remain undecided. The text does not tell us how to assemble the components of the weapon. The two fieldframes differ in that one is the mirror-image of the other (FIG. IO,I). As we do not know which of the frames is to fit which side of the weapon, there are eight possibilities for mounting each field-frame on the forked ends of the transverse struts. Some of them may be ruled out, because the curves in the beams of the the arms of the machine. Another difficulty is connected with the inner diameter of the washers. The text

also note 47.

"5Heron, Be!opoeica 77-8 (ed. Marsden, note I). See

field-frames were certainly intended to accommodate

gives an extremely small figure (I- dactyls, equal to " Vitruvius I praef. 2; RE IX A,I, 420-89; Marsden, op. cit. (note I), 3-4.
2.5 cm), which is disputed by Marsden, op. cit. (note I), 224 f., n. 20.

"7The hole in the washer of Vitruvius's ballista is not circular as in the older machines but oval, allowing more spring-cord to be placed in the torsionsprings. Furthermore the stone-thrower of Vitruvius may have been stronger than a stone-thrower of the op. cit. (note I), 197-200, n. 21. Marsden's proposal

same size built after Philon's instructions,cf. Marsden,


and not in digiti as given in the text is hard to accept, because this seriously interferes with the text.

to read the dimensions of Vitruvius'sballista in unciae


"s Marsden, " op. cit. (note 7), 185; 190, n. 2; pl. I.

The date of Heron's Cheiroballistrais discussed above in the section on chronology CD.14). Other late sources also seem to refer to the type of artilleryhaving an all-metalframe and especiallythe arched strut mencf. tioned in Heron's Cheiroballistra, Marsden, op. cit. (note I), 247, n. 2.

The new finds illustrate some of the artillery types in use after this importantRoman develop-

Saalburg-Kastell,D-638 Bad Homburg 1, West Germany

RECENT

FINDS

OF ANCIENT

ARTILLERY

17

ACKNOWLEDGEMENTS

First I have to thank Nicolae Gudea of the Institutul de Istorie gi Arheologie, Cluj-Napoca (Roumania), who sent photos and drawings of the Orgova and Gornea finds and invited me to study the objects at Cluj. The investigation of the Hatra ballista was only made possible by the generous invitation extended to me by Isa Salman, DirectorGeneral of Antiquities, Baghdad (Iraq). My thanks are also due to W. I. Al-Salihi of the DirectorateGeneral of Antiquities, Baghdad, the excavator of the ballista, for giving all necessary information and

providing

the photos

PL. II A, III A. Without

sub-

stantial financial aid given by the Deutsche Forschungsgemeinschaft, Bonn-Bad Godesberg (Germany) I could not have accomplished the journeys necessary for investigating the finds. Very helpful and clarifying was the correspondence I had with the late E. W. Marsden (then at Liverpool), who was very enthusiastic over the Roumanian finds, with M. Hassall (London) and with G. Drachmann (Lyngby, Denmark), whom I have to thank for his critical comments and valuable suggestions.

PLATE I

(Photo: D. Baatz)

A. Hatra (p. 3). Main line of defences with tower Andrae xIx near the North Gate seen from the town side (from SE). The find-spot of the ballista is indicated by an arrow. The photo was
taken in December 1975.

114.1p

st., ' -( -. i rr
11

L'Y"~,
. .

AllII 1Ai, .....t..............

' , 11..M ",0.,-. . . . t,,,,,,,."a.t . :. , ......... .........,I,,,, ml . :?!l[,.~~ 0

"

. . ..

"

I .

(aUII . "1 - _" ... .

l... .et"

r ig~3Ul

....

1140111 Ill,-/

8V0'i.

:irr(
IN.
-z--.j~r' -. ." -.. ". _: ..

i
?

2 ~ "- . . ~ .-" "--P"" -'.. .~ ." .- .O . - --... -._

"-~." -. -_.- --?. ..;

:,,,o++++
,+ ,

. ,

-+ -,+

+;

o ?. . "p

,.

B. Hatra. Reconstruction of the North Gate and adjacent defences seen from NW. To extreme right: Tower Andrae xix having on top two ballistae. (After W. Andrae, 1912, fig. 26: see p. 4. note I I.)

PLATE II

41

r.r

(Photo:

W. 1.Al

Salihi, Direct orate-G eneral of Antiquities,

Baghdad)

A. Hatra. Frame of ballista as found behind the town-wall. Seen from the North, from the top of the main town-wall (p. 4).

(Photo: D. Baatz)

B. Orsova. Arched strut (Kaca&ptov) of catapult (pp. 11-12).

Scale approximately 1: 12.

PLATE III

(Photo: W. I. Al-Salihi,

Directorate-General

of Antiquities,

Baghdad)

A. Hatra. Iron lever, bronze washer and bronze counter-plate of ballista during excavation (p. 4).

(Photo: D. Baatz)

B. Hatra ballista. Bronze washer resting on counter-plate (after cleaning) (p. 5).

PLATE IV

(Photo: D. Baatz)

A. Hatra ballista. Bronze corner-fittings mounted on reconstructed timber frame (p. 4).

(Photo: Saalburgmuseum)

B. Trajan's Column, Rome. Two arrow-shooers mounted on carriages (after a plaster cast) (p. i6).

PLATE V

(Photo: D. Baatz)

A. Orsova. Field-frame (Kagct iE-rpttov) of catapult. Below: the bottom ring seen from the under side (p. 9). Scale approximately I: 3-5.

Das könnte Ihnen auch gefallen