Catapults, a Case Against Sliders (Heresy)

[P. Clodius Secundus]

The Abstract to a recent article by Duncan Campbell, “Ancient Catapults, Some Hypotheses Reexamined” (Hesperia 80, 2011 Pp.677-700), points out that “Recent summaries and overviews of the development of ancient catapults have mistaken working hypotheses for established fact”. The Author wisely warns us that “A critical reexamination of these questions, setting them within the framework of the known facts, reveals the fragility of the accepted history of the catapult, as currently presented in general handbooks”. Taking inspiration from that admonition, I would like to challenge the universally accepted belief that all wood framed arrow-shooters had a grooved moving component called a diostra, or slider, as was formerly found on the belly-cocked gastraphetes and other early weapons.
A critical re-examination of the evidence actually indicates that instead of a slider, such engines featured a trigger block (chelonium), which travelled back and forth shuttle-like along a fixed rail embedded in the stock of the weapon. John Anstee had hinted at this arrangement in his ground-breaking treatise on iron-framed ballistae. He was mocked by critics for daring to question that article of faith. I suggest his real error was not in issuing that modest challenge, but in failing to ask the bolder question, “Did other Roman catapults all have sliders?”
I propose that with the exception of the Polybolos and hand-held belly cocked weapons, only the very earliest catapults had them. This of course, means that nearly every modern reconstruction of a Roman catapult is fundamentally (pun intended) flawed.
Anyone care to take up the gauntlet for the traditional view?

[P. Clodius Secundus]

Well, since nearly a month has elapsed and my challenge has been viewed over 100 times without any comment.
I can only see three possible reasons for this...

First, but most unlikely, everyone agrees with me. Having accepted my point as the new paradigm they are all too busy removing those anachronistic "sliders" from their stand-mounted catapults to reply.

Second, everyone has lost all interest in figuring out how Roman catapults were actually built and how they worked. (also doubtful)

Last and most likely, folks are treating this like an encounter with a raving lunatic on the subway. They are hoping that if nobody engages the madman he'll move on and stop disturbing their little corner of the world.

To stimulate discussion, I'll ask one simple question. On the Cupid Gem, the only known side view of a contemporary wooden-framed arrow-shooter, the string is drawn only about half the way back and yet none of the presumed "slider" extends beyond the sping frame. Can anyone explain this without resorting to the arrogantly presumptuous suggestion that the aritst was either ignorant of his
subject matter or engaging in artistic license?

[Epictetus]

I'll offer a fourth possibility why there have been no replies. If people are like me, they believe this is fascinating but don't think they know enough about catapults to say anything intelligent.

[Robert]

[attachment=5305]100_0732.jpg[/attachment]

[attachment=5306]100_0734.jpg[/attachment]

You may very well have a point. Here are the replica and the original of the Xanten manuballista. You will notice the differences in the way the slot in the opening is interpreted. When you look at the drawings of the find, there is no real evidence for a slider passing through in the way it was reconstructed. It would be possible to interpret the recess as a slot for the wooden slat on which the arrow runs.

Having shot crossbows and pondering on the idee, a simple triggerblock running in a groove the same way as the slider would work just as well BUT the groove for this block would have to be less wide then that of a slider. Why? Well, you do not want the arrow to fall into or get stuck in the groove in which the triggerblok runs. A narrow groove could be concievable, as there is little upward or lateral force on a triggerblock, the force in line with the rope for the pulleysystem to cock it and attachmentpoint/clamp holding the bowstring. The slider does solve that problem, as it combines the triggerblock with the arrowrest, no risk of the arrow snagging.

So the only way to solve this debate or at least prove ballista could well work without a slider is to build one along the lines you describe, with a triggerblock which has just a narrow guidance groove.
Aftertought: the triggerblock could also utilise two narrow grooves, the hollow recess for the arrowrest being between these two. This would give good stablity to the triggerblock, too, running in two grooves and passing over the arrowgroove. It would certainly be far easier to construct then a slider, too!!

[Macedon]

I'll offer a fourth possibility why there have been no replies. If people are like me, they believe this is fascinating but don't think they know enough about catapults to say anything intelligent.

Very true... I for one follow this thread but have no knowledge whatsoever on the issue...

[D B Campbell]

Anyone care to take up the gauntlet for the traditional view?

Okay, you've flushed me out. I was waiting for some of our other RAT catapultists to step up, before offering my own three-ha'penny-worth.

I would like to challenge the universally accepted belief that all wood framed arrow-shooters had a grooved moving component called a diostra, or slider, as was formerly found on the belly-cocked gastraphetes and other early weapons.

I am as guilty as anyone of building upon Eric Marsden's work without necessarily subjecting it to critical analysis. However, his interpretation of the arrow-shooting catapulta (written without the benefit of the archaeological finds) seems to have been largely correct.

When Heron of Alexandria describes the first torsion-frame, he writes that "they made the remainder [viz. of the machine] the same as the previously-mentioned ones." (Marsden has: "The remainder they made just as described before." 1971, p. 25.) As Heron has just described the gastraphetes, I take this to mean that the torsion-frame was tacked onto the gastraphetes design in place of the bow, and no other components were altered.

Confirmation of this comes a little later, when Heron writes about using the catapult. This is worth quoting in full, as it's a wonderful step-by-step description of catapult-shooting: "Whenever there is need to pull back the bowstring, they place the syrinx onto the anapausteria, having tilted it up and propped it against the underside of the syrinx, which has a cavity; then, having drawn the diostra back, they lift the syrinx from the anapausteria and, traversing it by means of the karchesion and inclining or tilting back by means of the pin, and having sighted the target, and having placed the missile on, they open the trigger."

Notes: The syrinx ("pipe") is Marsden's "case"; I prefer to call it the "stock" of the catapult. The anapausteria ("rest") is the prop that holds the stock in a horizontal position, while winching the arms back; it is clearly hinged on the stand, although (iirc) some modern reconstructions have it hinged on the underside of the stock so that it hangs down. The diostra ("slider") is the component that we're taking about. The karchesion ("cup") is a tilt-and-swivel joint that Marsden calls the "universal joint". The trigger is described as being "opened", because its hook has previously been "shut" over the bowstring; "opening" it releases the bowstring.

[Robert]

In a bellycocked weapon, the slider would have a function of cocking the weapon. Pressing down on the slider would propel the triggerblock fixed to the slider backwards to cock the weapon BUT this would mean the slider or triggerblock would have to be fixed/pinned in a cocked position to allow for aiming and firing. There is no mention of "fixing the slider" in that text. Without fixing/pinning the slider, it would slide back the moment forward pressure is lessened. I agree that in the quoted text, pinning the slider would not be called for, as it would have been held back by the prawl of the crankdrum with which the cord to cock the weapon is wound.

In a ballista with a crank at the rear, the slider would have no other function than as a moving arrowrest. Now, it could be argued on basis of the text this feature was incorporated in the first of the torsionspring ballista as described, but this in no way assures it would have remained in use in later models. And it is the later models we all use in re-enactment. So I do think it is a valid point being raised, intreaging enough to be explored for its validity, preferably on the basis of archeology. As the example I posted shows, even in reconstructing we tend to "rearrange" the evidence before us to fit a "known" patern. That reconstruction could be seriously flawed through just such a mechanism. The drawings, which unfortunatly I only viewed and have not stored on the computer, do suggest this could be the case. I did notice just this discrepancy at the time and wondered about it.

[P. Clodius Secundus]

Thanks to all who have replied. It's difficult to learn anything new when talking to yourself. Epicteus and Macedon, please don't sell yourselves and other RAT members short just because you may not be total artillery geeks. One of the biggest problems with the prevailing theories is that they contain too many non-sequiturs, confusing terms, and unexplainable contradictions. My basic principle has always been that the best answer is usually the simplest. Sometimes you just have to look past your assumptions to see the obvious truth.

Robert, given the size of its washers, I believe that the Xanten weapon was most likely a hand-held belly cocker like the gastraphetes before it and the cheiroballistra that followed. As such they all had sliders and lacking a winch they needed some form of hold-back mechanism. Once the springs got beyond that 35-40mm range a winch and stand were needed and the “sliders” were not. I’m glad that you agree with my point that just because the earliest versions of something had a particular feature, that does not mean that all the rest did. A perfect example is the steering tiller on early automobiles, which was soon replaced by the steering wheel.

That brings me to you Duncan and your citations of Heron’s texts. Since his writings are either descriptions of early engines or small hand held ballistae, neither is really relevant to the discussion of Roman era wooden-framed and winch-powered catapults of the Late Republic or Early Imperial periods most reconstructionists favor. An examination of what contemporary writers such as Philon and Vitruvius had to say about contemporary weapons is far more useful and revealing. If you abandon your preconceived notions based on Heron’s writings and objectively examine what they said and didn’t say a much different picture emerges.

For example…

Philon’s descriptions of various engines;



1. Standard Euthytone Arrow-Shooter

“The length of the case must be symmetrical if it is about 16D”.



2. Philon’s own “Wedge” Engine

“We make the case, mounting, block, windlass, and handspikes, like the standard ones, that is to say, the old ones, except we slightly increase the size of the pull-back system”… “The question of the case, mounting, and the block has already been dealt with, because you must simply use the same old type; but we alter the frame, and in this of course lies the whole secret of long range.



3. Philon’s revision of Ctesibius’ “Bronze-Spring” Engine

“The mounting, case, block, claw (which draws the bowstring), the windlass, ratchet, hand-spikes, and all other parts in this machine, are constructed exactly as in other arrow firers; but the frame takes another form.”



4. The Polybolos

“A certain Dionysius of Alexandria constructed in Rhodes what is called a repeating catapult, which has a unique and very intricate arrangement.

The case’s length conformed to the usual; but its thickness was 6 dactyls, its breadth 5 dactyls, and depth 3 dactyls. It had on it a wooden structure of appropriate breadth and height, which covered the whole case, but was just a little longer than the distance over which the bowstring was to be pulled back. This structure was made as a slider so that it could be pushed through the case. On this slider was fitted a bronze claw, made with a double prong. It was so contrived that, automatically, it first grasped the bowstring and locked the trigger and, then, when drawn back, released the trigger, all in the following way.

The claw was fitted to the slider, just as to the blocks in other catapults, except that it was low down, while the locking trigger projected a little at one side (as usual), but did not project at all on the other. When the bowstring had to be pulled back, the slider was in the forward projecting position and the excess part of it was pushed out at the front through the frame….”

Philon discusses four variations on the euthytone arrow-shooting catapulta, he describes three of them as having cases which conform to the conventional or “standard” method of operation. As you can see above, the word slider (diostra) is never mentioned in their descriptions. However, the term block (chelonium) is associated with two of the first three. Marsden explained this confusing use of terms with the assumption that “Philon prefers chelonium to Heron’s diostra for the slider of a catapult”. I submit that the difference in terminology had nothing to do with any preference on Philon’s part. It is far more likely, that he was just using the correct names for the components based on how they functioned in the machines of his era. As Marsden himself noted, “Philon mentions no machinery earlier than the fairly advanced torsion variety”. It is not difficult to imagine that in the years between the invention of the Gastraphetes and the time of Ctesibius and Philon someone had hit on the simple idea that there was really no need for the whole grooved slider to move. As long as the block containing the claw and trigger could move back and forth like a shuttle along the raised rail with a groove to guide the projectile, the complexities of the slider, and pulling-forwards mechanisms could be eliminated. In fact, the only instances where a slider was needed were belly-cocked weapons like as the gastraphetes or ones with “a unique and very intricate arrangement” such as the repeating catapult.

This is why the polybolos is the only one of the four catapults Philon describes, which has a slider, and it is also the only one of the group to have a pulling-forwards mechanism. According to Ctesibius, such a mechanism was often required when a slider was used. Moving only the block of even the heaviest arrow-shooter would be a simple task, requiring no mechanical assistance. This could help explain why the pulling-forward mechanism never appears again in the historical record.

Perhaps the most telling passage in Philon’s coverage of the repeating catapult is his rather involved description of how prior to drawing the bowstring, the excess part of the slider was pushed out and extended beyond the frame. If this were standard practice on all catapults of the day it would have gone without saying. There would have been no reason to mention it. He saw no need to explain this on any of the three preceding engines, but then again, they all had cases and blocks which functioned in the usual way.

Also missing from Philon's text is any mention of dovetail grooves in the case which would be needed to keep a slider from cantilevering downwards when extended out the front of the frame.

So, according to Philon, only the Polybolos had a slider and one of the ways in which it differed from other more conventional engines was that it extended forward when being cocked.

With me so far?

[P. Clodius Secundus]

Next installment, sorry to be so long winded, but I really think this is important stuff to get right.

So Heron/Ctsebius says sliders on hand-held belly-cockers and early torsion engines. Philon puts a slider only on the polybolos, which needs one to function. No dovetails on any of them, but he uses the term in describing the protective cover for his wedge engine. The lack of dovetails on the polybolos is understandable if you consider that the magazine upper structure would keep the slider from cantilevering out.

Moving on to the Roman era we see what Vitruvius has to say about cases, blocks, sliders, and dovetails.

Let the length of the case, which is called the syrinx in Greek, be 19f; the length of the side-pieces which some call check-pieces and which are fastened to the case on the right and left side, is to be 19f; their height and thickness 1f.

The length of the claw is 3/4f, its thickness 1/4f, the thickness of the block is the same. The length of the trigger is 3f, its breadth and thickness 1/4f. The length of the canalis fundi is 16f, its thickness 1/4f, its height 3/4f.

Notice that despite his demonstrated mastery of the traditional Greek terminology, Vitruvius chooses a Latin term canalis fundi in lieu of the Greek term for slider. This new term, often translated as “grooved base” does not describe or even imply a sliding motion. Using the terms interchangeably is a seemingly innocent but flawed assumption.

Like Philon who only used the term dovetail when describing the cosmetic cover on his proposed “wedge” engine, Vitruvius also fails to mention any dovetails in the case or canalis fundus even though he uses the term to describe the manner in which the winch box is assembled. From one engineer such an omission might have been an oversight, but when your two best sources fail to mention something it’s probably because it wasn’t there to begin with.

Traditionalists may point to the carvings on Trajan’s Column as evidence of sliders, but their argument does not take into account the in-swinger theory, originally suggested by Victor Prou for wooden-framed stone-throwing ballistae and revived by John Anstee for iron-framed arrow shooters. Many now accept the Trajanic engines as in-swinging ballistas, and as such their spring frame would be mounted much further back on a shorter case and only a grooved rail (quadratus stylus) would need to project forward of the frame to allow the block to be run forward to capture the string. This explains why none of the weapons on Trajan’s is shown with its slider drawn back behind the spring frame, even when the arrow has been placed in the groove and is ready to shoot.

To sum things up, my theory of non-sliding “sliders” works just as well as the traditional view with the known artifacts. It follows precisely the actual texts of Philon and Vitruvius without assuming, embellishing, or inventing any supposedly missing components/features. Last but not least, it explains without contradiction why the arrow grooves on Trajan’s Column always project forward, but the one on the Cupid Gem doesn’t.

If anyone can point out contemporary evidence or contradictory sources that I’ve missed, please help relieve my ignorance. Likewise, if anyone’s interested in how my theory relates to the wood-framed stone-throwers of the period, I think I’ve got that pretty well figured out as well.

[P. Clodius Secundus]

Once more a month has passed and aside from Duncan's original citation of Heron/Ctsebius and a few encouraging comments from others there has been no reply to my challenge. I think I have demonstrated that when critically examined, the works of Philon and Vitruvius do not support the use of a traditional diostra/slider on winched and stand-mounted scorpions from the Late Greek through Imperial Roman eras. Thanks to Duncan's mention of Heron's edition of Ctsebius' history of early catapults we can all
see precisely where and why the assumption could be made that sliders remained in use. Just because a component was once used on a particular machine
doesn't mean it was never improved or replaced. To those who might be tempted to say that a catapult with only a sliding block/rail set-up wouldn't work, I cite in my defense the fact that two of the most powerful arrow-shooters in the world today Carnifex and Nick Watts' Firefly use this system to great effect. Although both are iron-framed ballistas from a slightly later period, there is no practical reason why it would not work equally as well on these machines.