Open up constructive details by the images of bolt-shooters on the Trajan's Column in Rome


No comprehensive explanation of ancient artillery.

Look up literature- and internet register under reference place.


What is visible?

How can it be interpreted?


Images of the Trajan's Column and knowledge of Roman arrow throwing artillery.


Publications and internet research.

How can the images on the Trajan's Column be interpreted?

By interpreting the images on the Trajan's Column You must observe, that the sculptor who creates the relief's performed no realistic photo illustrations. It is rather a matter of of strongly personally oriented representations of the Dakien Wars of the emperor Trajan, which took place in the years 101 to 105 A.C.. Persons stand in the foreground, the emperor appears approx. 60 times on the column. The perspectives are frequently distorted and the persons are represented over sized in the surrounding area and background.

Illustration 1
Legionaries cross a ship bridge.
© German Archaeological Institute Rome Anal

Despite these restrictions the relief's supply an abundance of details over Roman military equipment and warfare. Regarding the shown torsion catapults  there is the problem that at present there are no findings known to assign to the type of weapons which are represented on Trajan's Column.

However it is taken for granted, that there is a connection between the represented and the found late ancient arrow-shooters.

 It's common to all representations that the arrow firerers are reduced to 3 substantial elements.

1. The frames with the torsion sockets.

2. The pedestal

3. The slider with the arrow gutter

Despite these restrictions the relief's supply an abundance of details over Roman military equipment and warfare. Regarding the shown torsion catapult  there is the problem that at present there are no findings known to assign to the type of weapons which are represented on Trajan's Column.

However it is taken for granted, that there is a connection between the represented and the found late ancient arrow-shooters.

It's common to all representations that the arrow-firerers are reduced to 3 substantial elements.

1. The frames with the torsion sockets.

2. The pedestal

3. The slider with the arrow gutter

Illustration 2
Trajan's Column
One ballista fires from one with trunks fastened position.
In the background are located arrow- shooters in a bricked
© German Archaeological Institute Rome A

Furthermore all representations do not show catapult-arms. Further construction units like the hoist gear for stretching the fibre, the catch and the mechanism for adjusting the height setting are not available, or not recognizable. From the view of the ancient contemplator this was not necessary because the appearance of these arrow- throwers must have been well known. Simply the suggestion of the three above mentioned elements was sufficient to identify the description as arrow-shooter.

Illustration 3
Excerpt from illustration 2.
Legionaries fire with a ballista from a
mounted position.
© German Archaeological Institute Rome A

Analysis illustration 3

1. It is assumed that it is an arrow- firing catapult (scorpio, tormenta, ballista). Size and flat firing angle argue against a stone-thrower and rather for a bolt-shooter.

2. It shows two soldiers who operate the catapult.
The arrow-firer is operated by two man. Practical firing tests confirm this as an ideal manning. A man adjusts and stretches the fibre. Another man applies the projectiles and activates the shot. The two man crew is also shown in other illustrations and thereby might be authentic.

3. The bolt-shooter is operated from a fortification of trunks.
The operating crew had to protect itself – probably illustration 2 represents a siege – by a fortification. This permits statements about the effective firing range. Obviously the catapult stood in reach for opposing long-range weapons. For this reason an arrow bombardment is more probable than a spear throwing. A covered position against attacks (losses) of the enemy is a lower ranking possibility. For the enemy the dangerous catapult positions must have been an profitable aim.

4. The construction is defined as torsion catapult.
The arrow- firer lets itself assign to later findings by its construction style. Strut and lateral pedestal for the torsion bundles also appear in late ancient models. The catapult works according to the torsion principle.

5. The two torsion systems stand - compared to the Hellenistic type - very far apart.
An explanation would be: The catapult had an arm system swivelling inward. This design requires a large spacing of the torsion systems since the catapult arms swing inwards.

Illustration 4
Graphic reconstruction. Catapult with interior arm system.
The strained condition is represented. Perspective similar to picture 3.

6. Size of the distance of the torsion sockets.
This is difficult to determine on base of the illustration . The representation is not a technical one, the catapult which is seen by the side is represented diagonally. Fact is, the distance is clearly larger as at traditional bolt- shooters. Orienting at the findings in Orsova (Romania), where the distance amounts 1.25 m, You can assume that the distance might lie between more than 300 mm until maximally 1.25 m. 

7. Size of the torsion systems.
Here the finding of a torsion stands with washers in the French Lyon  (current dated to 197 A. C.) can help. The stands have a height of 320 mm and the inside diameters of the washers lies at approx. 80-85 mm. The two vertical struts of the stands would
have, with a lining by a sheet metal cap, an outside diameter of approx. 140 mm. With these technical defaults - distance and height of the torsion sockets - the illustration 4 was provided. 

8. There are no catapult arms to be recognized.
These circumstances are common to all representations. There are the following interpretation possibilities:

a. The arms were made of metal and were put into the stone relief. It is considered as secure that there were such metal parts e.g. spears on the Trajan’s Column, but in the course of the centuries they have been stolen by metal collecting people. However there is no indication for anchor points for such metal parts. 

b.Furthermore the possibility exists that this detail of the catapult arms was completely omitted. The artist would have suggested the arrow- firer just by its substantial construction units - the characteristic clamping frames, tripod and stock -. But this would disagree with the otherwise extremely exact representations, which represent even little aspects of the legionary equipment.

c. An alternative possibility would be that the catapult had a catapult arm system swivelling inward. The bolt-shooter could be represented in the strained condition. In this case the catapult arms would not be visible, because they disappear behind the right torsion socket and the stock. (See illustration 4) 

d. A remarkable idea is, that these arrow-throwers had no catapult arms at all and contained a still completely unknown mechanism. That is pure speculation, but would be of value for a theoretical examination.  (See to menu option different “)

Illustration 5
The far distance of the torsion systems results from the interior movement of the catapult arms.

9. The pedestal (tripod) is locates directly under the clamping frame.

At the traditional arrow-firer the pedestal was located behind the clamping frame and was installed on the stock. This might be a further indication for an internal catapult system. Stock and slider poke out of the clamping frames construction dependently. Assuming that the system should be in equilibrium, a support on the clamping frame would be more favourable.

10. The tripod is implemented very short and stands on a base of trunks.

a. The tripod might have been longer in the original version and might have stood not visibly confessed behind the attachment. Then the catapult would have been better to take cover without effect restrictions. See also illustration 10.

A review of the illustration 2 on page 2 could confirm a volitional falsification.
The catapult operation itself is located in an exposed position, but it is not using the covering possibility of the attachment. The situation is in fact a combat, though no helmets were worn. The primary aim was obviously to individualize the catapult crew. Very distinct features can be recognized. Two persons should have been pictured clearly to recognize. Maybe an outstanding catapult crew which should be particularly honoured by this way.

Illustration 6
Other interpretation of the illustration 2 and 4.  The arrow thrower
has a tripod to the ground  and stands well protected behind an
attachment of wood trunks.

In order to achieve the goal mentioned above, the artist had „to pull “ the catapult out of the attachment and so to falsify the represented situation. In a realistic representation catapult and crew would nearly disappear behind the covering and the desired effect couldn't be reached. The optical effect of the solid defiled remains, because the bolt- firer was now represented with a shortened  tripod on a wooden pile.

 b. There were technical reasons to perform the stand so briefly that taking shelter was necessary. A short tripod limits the applications of the arrow- thrower, because a podium to set up is always necessary. It is although conceivable, that the carriages offered this possibility, but during a set off enterprise, e.g. on walls always a „catapult bank “ needed to be build. A pedestal, which allows the use on the carriage without further expenditure, in free deployment in the area and on walls, remains the most favourable and thus the most probable design for a catapult under body.

11.. The clamping frame seems to be connected with the pedestal by a hinge. It is recognizable that the lower strut to the tripod in the area of the binding is omitted. This was obviously necessary to create place for a slew able device. How this connection was constructed is difficult to recognize.

12. A possibility for side direction is not recognizable. At the tripod it is not recognizable that the catapult have a possibility for side direction. This would cause a pivot bearing with a vertical axial direction. Probably the catapult  were positioned completely into the desired firing direction.

Note: With the torsion catapults only static targets were combated. To hit a moving target (e.g. a trooper) was perhaps rather luck than plan able. The projectile speeds were much too small, the deviations by wind, uneven projectile weight and aerodynamics much too large as this would have been possible. If a catapult was once aligned to a fixed target, troop accumulation, an excursion gate or the wall of an opposing fortification, essentially no further corrections were necessary. It is quite conceivable that for these reasons a technical possibility for the side direction was abdicated.

13. The torsion bundles are cylindrical totally enclosed. The caps with conical covers are locked above and below. The torsion ropes consisted of sensitive, organic materials, chords or the hair of horses. These had to be protected by packaging against external influences such as humidity, sun exposure or mechanical impairments (enemy fire).

Note: Starting from the 1st century B.C. ancient authors only speak about chord bundles and no longer about horse hair. Conceivably that hair was unsuitable for internal working catapult systems.

14. The lower crossbeam seems to be made of metal from the construction. A construction like that made of wood would not have beard up the developing force influences by stretching and shooting. One has the impression that both struts have in-forged or cast in gutters. Probably a measure for weight decrease. This detail occurs only on the Trajan's Column and is not proved at later catapult findings.

llustration 7
Cross section by the upper tie bar.  The groove
would result in the characteristic appearance.

15. The upper arched-strut is implemented strongly arc shaped.

An internal catapult system provided, in relaxed condition the catapult arms stand beyond the clamping frames. For stretching the operator has to grip forward under the upper strut to the chord to engage it into the stop claw of the slider. For this reason the necessity consists to construct the upper strut accordingly space giving.

Illustration 8
Grasp under the arched-strut forward to the string (photo author) :

At the Trajan catapults the upper arched-strut is distinctness strongly and goes nearly over the entire internal width of the weapon. At the late ancient finding from Orsova (Romania) the arc shaped projection concentrates on the strut centre and a cannelure is abdicated.

Illustration 9 
Upper arched-strut (Kamarion) of a late ancient torsion catapult  (4th century A. C.)
Drawing after N. Gudea and D.  Baatz „Parts of late Roman ballista  from Gornea
and Orsava (Romania) “

Note: At the "Orsova strive" on the left and on the right beside the arc shaped projection drillings are to be recognized. Because this strut - compared with the representations on the Trajan's Column - was out forged very thinly, it is conceivable that these drillings were serving the buttress as points of anchor for an additional stabilizing spanning or bracing. The strut is out forged plainly. A cannelure is missing. How the binding - by means of the forked ends - on to the torsion stands took place still unexplained.

16. The clamping frame

a. At the catapult parts found in Romania and in the French Lyon the struts were tied to the torsion pedestals by ends formed like a fork, which were fixed on the pedestals by u-shaped, clinched fasteners.
How the attachment was done isn't clarified. Possibly the adjustment was made of hit wood wedges. A variant would also be a well removable connection by attached iron wedges. A removable connection between torsion pedestal and struts had an enormous advantage.

One could change the torsion systems including the covering very fast and relative uncomplicatedly with a prepared replacement in the case of damage.  This possibility wasn´t given at the traditional Hellenistic systems with wooden pedestal construction. Here the catapult was useless after damage and the covering had to be exchanged time-consumingly.
Well conceivable that in the case of application ready covered torsion pedestals was available to reconstitute the readiness of the bolt-shooters in the case of damage.

Illustration 10
Associated to arched-sturt illustration 9. Torsion pedestal (Kambestrion) of a late ancient
catapult from the Romanian Orsova. The shaping would not permit a cylindrical
packaging. (4th Century a. C.). Drawing after N. Gudea and D. Baatz „Parts of
late Roman ballists from Gornea and Orsava (Romania) “

b. Also an independent solution would be conceivable, which is clearly different from the recent models. The torsion pedestals could have been riveted. However this would have brought the possibility to change for the torsion systems to the level of the predecessor models.

Remarkable at the Trajan catapults  is „the clean “cylindrical lining which requires a suitable constructional design of the torsion pedestals.

The torsion stands of the Orsova catapult are designed in a way that a cylindrical lining isn´t possible. Even a protective coating for the torsion bars with another design would have been hardly to implement.
Probably the reason for this is very simple:  this late ancient catapult parts were found in fortresses. They were set up stationary under a roof, thus a covering of the chord bundles from weather influence or secondary enemy effect could be abdicated. Differently with the catapults on the Trajan's column: these were carried on campaigns. They practically always stood outdoor and were exposed to hostile weapon effect in a combat.


Illustration 11        
Torsion stand find from the French Lyon.
Dated into the 2nd half of the 2nd century A. C.
On the right bird view.

Illustration 11 shows a finding from Lyon. The torsion pedestal is worked very roughly. Probably manufactured in a crisis period with insufficient capital. The findings of the associated clamping chucks also argue for this. These are forged of iron instead poured of expensive bronze and also manufactured in very bad quality. With such a structure - possible in better quality - a cylindrical coating would be possible.

Illustration 12
Left: Suggested clip system in detail. On the right torsion pedestals
without struts. (See menu option reconstruction)

17. Stock and slider.
Both in illustration 3 and 13 a slider with a dove tail like channel are recognizable. The stock into which the slider has to be insert isn´t recognizable. Illustration 9 shows a graphic reconstruction. It was assumed that at a chord standing in front a projectile length for arrow guidance has to remain.

 Fig. 13.a.
Stretching the string. Situation under the suggestion
that the catch has to be put forward to the cord. During the shot and at a cord standing in
front the arrow channel is overlapping an arrow length.


Fig. 13.b.
Firing position.




Illustration 14
Cut out of illustration 2.

Analysis illustration 14

1. The torsion sockets with covers are good to recognize.

2. The arched-strut set more deeply set as in fig. 3 The question here arises whether the buttress was really so strong set off of the lower edge of the upper cover. A reconstruction as drawn in illustration 5 would be questionable.

3. The arched-strive seems again however formed represented strongly distorted. If the buttress is strongly simplified in the two following illustrations,  the grooves are again good to recognize.

4. The arched-strut is clearly more deeply set than in illustration 3 The starting point for the strive seems to lie with the represented perspective close upper edge. This representation would mean that the prop had its own starting point at the torsion stand and not at the upper flange of the stand attached or was riveted. 5. The stock appears again as a construction unit.  One meant down a dove tail like prepared slider with arrow gutter to recognize these would have to stand out however to the longitudinal guidance from the larger stock.

Illustration 15 
Transport catapults by means of cars.
© German Archaeological Institute Rome Analysis illustration 15


Analysis illustration 15

1. The catapults were posed or installed on a car. An indispensable condition to the field trains catapults from campaign.

2. There are two course animals recognizable. The weight of cars and catapult caused the use of course animals.

3. The car is with one axle and has wheels with 8 spokes. Cars with one axle with wheels with spokes are occupied for the Roman time

4. The wagon seems to have a box like structure. This is difficult to interpret. Regarding the representation on the Marc Aurel Column is
more  informative.

 Illustration 16 
 Transportation situation. Catapult arms, the crank and the two
 lower caps of the torsion stands are distant. a
 (Representation without course animals).  Comparably Illustration 14

5. In driving direction right a stowing ring is to be recognizable.

a. The eye could be a reference on the fact that the cannon was only loosely posed on the car and was fixed by means of ropes at the car.

b. The eye served to fasten a tarpaulin for taking the catapult, the tarpaulin off could from waxed lines or greased leather have been over as protection before influences of the weather.

6. The lower covers of the torsion sockets are missing. A simplified representation would be conceivable. Possible also that in the case of transport the lower covers, there they were removed were only attached and by vibrations in the case of transport would have dropped. Illustration 15

7. The upper covers have a steeper angle as during illustration 3. Here it seems to keep itself simple around artistic freedom. The covers must be trained so that clamping chuck and pins find under it's place below.

8. The tripod is still more briefly represented as in illustration 3 The representation is simplified or it can also apply to the reasoning in illustration 2.

9. The stock is not recognizable. It would be possible that the stock was dismantled for transport. It is close lying that the sculptor concentrated itself towards with this motive more on the wagon and the legionary strong into the spokes of the wheels reaches. The catapult is also represented as secondary matter in the background.

10. The buttress is represented without the characteristic recesses. The representation is highly simplified in contrast to illustration 2.

11. The lower prop is much more broadly represented than in illustration 3 and also had no recesses. Which creates the sculptors had these strive so strengthened to represent, does not let itself not explained. One has formally the impression a board is pushed by the torsion sockets.


Illustrationl 17
 Legionnaire fire from one car over the course animals away.
 © German archaeological Institut Rome

Analysis illustration 17

1. It shot with one from a car installed catapult out over the linked up course animals.

It could be thus shot even in the transportation condition. Firing direction was over the course animals in driving direction. Even if it is not visible here is shot over the heads of the own troops. A reference for it that these catapult. systems were also used in the open battle field.

2. Two catapults are represented. The battery form - always several cannons - might have been the usual system. As single catapult the target effect might have been only insignificant.

3. The tripod is not recognizable. During this illustration one is inclined to the opinion that the catapult in the box like structure of the car stands and only the parts, clamping frame and stock necessary for shooting and slider over the edge of car, exceed.

Illustration 18
Shoot from car.  According to Illustration 17

4. The car seems in the back open for catapult operation, the upper edge of daring is strengthened represented. The car could have been closed in front, laterally was present and from the rear two bordwalls was the car openly over to the control elements such as hoist to come elevator direction mechanics and the catch.

5. The rear soldier holds on to a grasp or embraces the lever arm of the clamping hoist. The rear legionary seems to serve the winch.

6. The front legionary seizes by the arched- strive and locks sees by means of catch. This could be interpreted as indication for an interior catapult system, the clamping grasp must take place via the buttress to the chord lying before the clamping frame.

 Illustration 19
 Marc Aurel Column. A catapult is to be seen also here,
 that is mounted on a car.

Analysis illustration 19

1. Does it concern an arrow-catapult or a stone-thrower? Similar to the representations on the Trajan's Column speaks alone the size itself much for one ballista (arrow- firer). For a stone- thrower the steep angle of incidence could speak itself.

2. The stand is longer implemented than in the illustrations of the Trajan's Column which obliviously rests on the front edge of the wagon. If assumed the stand on the front edge of the car is fastened would make, a short stand execution would make sense, since the front car wall has a kind of landing function stand. It catapult be excluded again naturally that the catapult is located in the car and the representation is therefore falsified.


 Illustration 20   High angle firing, according to Illust. 17  
 Marc Aurel Column  (Representation without course animals)

3. The side parts of the car are chamfered in the back. This is necessarily served  around the catapult. Particularly the hoist is with high angles of incidence close because of the ground.

4. The angle of incidence of the catapult is to 45° with 40° -.
Under this angle the catapults shoot furthest. This type of catapult, contrary to which far from stone- thrower, gladly it assumes that one shot flat, this representation occupies the opposite. There are thus no practical bases against  „bow firing “speaks. Aiming accuracy decreases strongly, however the target is crucial. A surface moderately large target, troop accumulations or urban ranges behind town walls, internal areas of fortresses etc. offered determine hit possibilities. This is particularly valid to the employment in battery form, with which a multiplicity of catapults one shot simultaneous.

5. The catapult makes an incomplete impression. The torsion sockets are completely missing. Duncan B. Campbell expresses the assumption of the catapult in his publication - „Greek at novel Artillery 399 BC-AD 363 “- in the divided condition was transported. For that however those stand much concentrate of working floats of the catapult operators in firing position. It would be conceivable that the torsion sockets rising up out of the relief, broke off - in the range where the torsion stands must have been, one believes abort pure to be able to recognize - or the sculptor has for reasons of the treatment of stones completely left it out.

6. The wheel shown has only four spokes. Very common a strong simplification. 8 spokes would be appropriate with security reasons
for stability.

7. Compared with illustrations on the Trajan's Column.
The similarity between the illustrations on the Trajan's- and the Aurel Column are unmistakable. The Aurel Column probably gives the more realistic impression regarding the construction of the car, since the wagon construction place let's to serve the catapult. The representations on the Trajan's Column are less detailed here. Also remarkable is the appearance that the stand of the arrow-firer is turned off on the front edge of car. Here a connection with the too briefly working representations of the stands would let itself be represented on the Trajan's Column.


Illustration 21
Hypothetical drawing of one catapult
on a car. Track width 1,2m.

Illustration 22
Reconstruction a car with one axle.


* The representations on the Trajan's Column show bolt-firing systems with catapult arms
swivelling inward (internal catapult system).

* The large distance of the torsion sockets is necessary for the catapult arms with stretching from
the front by the area between the sockets for them to be swivelled there.

* The characteristic arched-strut is necessarily because the catch under the prop through, to which
before the torsion sockets lying string, must be pushed and latched.

*The high power density of such systems permitted compact dimensions, additionally did not give it,
obstructive outstanding catapult arms more. Ideally for the employment from close areas and transport
on war courses

* During conventional arrangement the far distance of the torsion sockets did not make a sense,
because the efficiency of the catapult  worsens were clearly opposite by close together lying
sockets. (See to menu option „simulation “)

* The endangerment of the service personnel by moving catapult arms is smaller, since this move in the
interior area clamping frames.

* The stand is fastened over a hinge directly to the clamping frame. This became necessary, with an internal
catapult system of the slider far forward to be pushed must and the system - at a point of hinge behind the
clamping frame - from the equilibrium, speaks would have tilted forward.

*These catapult types were from the beginning transportable designed. The arrow- throwers should be able to fire in
transportation condition. Tactically reasonably, because the field gentleman could shift its artillery fast and react
with it - by again referred firing positions -  to a changed combat situation.

*In addition, possible the catapults were to be used immobile. This is valid for the storage war or with the
employment for the defence of structures.

*During the execution of the stand both a long variant possibility for the ground and truck list and one - as on the
column represented - would be a  short execution. The short design however required an under body. During the
truck list this could have been one „catapult bank “at  the front of  the car, with immobile list as a bricked or a
timber construction.

 Designs and text: Author

Source references

Used publications and Internet sites: See menu option "references".