Why did Zama go wrong for Hannibal?

[Justin Swanton]

This in response to Patrick's sensible request.

I agree with his hypothesis that Hannibal planned a double envelopment of the Roman infantry using his veterans once his cavalry, which he believed superior in numbers to their Roman counterparts, had sent the Roman horse off the field. Caught with his pants down by the surprise appearance of Masinissa's Numidians, he made a change of plan and went for an Ottoman-style frontal assault with all three lines, in the hopes of wearing the Roman foot down to breaking point.

For me the central problem is this: once the Carthaginian cavalry, by design or accident, had drawn the Roman cavalry off the field, why did not Hannibal execute his outflanking manoeuvre anyway? He had the time to do so and it would have won him the battle.

My suggestions:

1. Scipio, a well-versed student of Hannibal, kept back a flank guard once the bulk of his cavalry had gone off in pursuit of the Carthaginians. It wouldn't need much to stop two infantry columns marching around the flanks of the legions, even if the columns were made up of veterans. It would surprise me if Scipio, possibly the most brilliant general in Antiquity, had not made some sort of provision against outflanking.

2. Hannibal did not know the situation with the cavalry once they had quit the field. The Romans, manifestly superior, could return at any moment. He judged it safer to go for a frontal assault, with the third line handy as a defence against a rear attack by the Roman horse. When faced with the unknown, Hannibal became cautious, taking no risks. He was a methodical planner, not a brilliant improviser like Caesar.

3. Hannibal knew that even if he did perform a successful envelopment, he would need more time to destroy the Roman infantry than he believed he had. It took him several hours to kill off the surrounded Roman foot at Cannae. He may have calculated that he did not have that time at Zama. Should the Roman cavalry return before he had reduced the Roman infantry, he would be cut to pieces.

Any ideas?

[Mark G]

what are the actual sources to suggest that Hannibal intended to deploy is veterans as a flanking force?

what are the actual sources to indicate that he then changed this plan on the field?

I ask because changing battle plans on the field after seeing the enemy army is not at all something which is typical of either general or the period. - especially when 'shown' that army in camp the night before.

Which makes me doubt the entre premise.

[Justin Taylor]

My view, the attack with the elephants failed. In view of how close it was at the end with the failed elephant attack, then a successful charge of the elephants should have brought victory.

I have (generally) seen two views on why the Italian campaign veterans were kept back as a third line; to prevent an out-flanking move by Scipio or (as they were used) to take on the Roman infantry after it had fought two lines of the Carthaginian infantry . Mind you those two options are not mutually exclusive.

Having fought it as a game, the elephant attack does work provided you go in hard with the mercenaries. Stampeding elephants also hurt the mercenaries but the Romans find it very tough to deal with infantry and elephants at the same time.

[Justin Swanton]

what are the actual sources to suggest that Hannibal intended to deploy is veterans as a flanking force?

what are the actual sources to indicate that he then changed this plan on the field?

I ask because changing battle plans on the field after seeing the enemy army is not at all something which is typical of either general or the period. - especially when 'shown' that army in camp the night before.

Which makes me doubt the entre premise.

The point though is that what Hannibal's spies were shown - before being allowed to escape - was the Roman cavalry only. No Numidians in the camp until the following day. Which would explain why Hannibal arranged to meet Scipio and size him up, that escape being a little too easy.... Scipio acts the part of the pompous and overconfident fool, the sort of general Hannibal was used to encountering, and Hannibal falls for it.

But I'm just paraphrasing. Ut loquatur Patricius ipse!

[Mark G]

nothing in that episode of the spies indicates Hannibal planned to use his veterans to repeat Cannae.

remember, there is absolutely no evidence that his first lines attempted to give ground, as they did at Cannae, which was central to the plan working then, quite the contrary in fact, he refuses to allow them to give pull back through the centre, and instead forces them around his flanks.

and hence, what are the source evidences for this plan - cause to me, the whole scenario you outline seems incredibly complicated for something which is then changed on the spot that morning when the armies are deployed and he sees more cavalry on the Roman side than he thought when he drew up the battle layout and issued his orders/

much more likely is that he simply tried to repeat the same thing which had won victory in almost exactly the same circumstances at Bagradas -
which is, an elephant charge leading successive waves of attacks by expendable troops which wear down the Romans until his main troops, the African Veteran, move in to make the culminating attack which destroys the Roman army.

a simple plan which had worked against this army before and which conforms pretty exactly to the moves which his army makes on the day.

[Duncan Head]

The mechanics for extending Hannibal's third line to outflank the Romans, and some reasons to believe that was his plan, were set out in Steven James' article in Slingshot 241 (I remember the number because by chance I was looking at that issue for something else just yesterday). I don't think his reconstruction included Hannibal being misled about cavalry numbers though -  that's an interesting idea I don't recall seeing before.

[Patrick Waterson]

Bagradas does seem to have been Hannibal's basic template for Zama, but it is worth remembering that at Bagradas Xanthippus had one line of infantry and no 'expendable' troops.  His predominantly Carthaginian citizen infantry (cf. Hannibal's second line at Zama) followed the elephants in directly and cut up the Romans who filtered through between the beasts before they could form up again, with the elephants doing most of the execution (Polybius I.34 and I.39).  The cavalry were also a vital part of the Bagradas battle-plan, driving their Roman opponents off the field and surrounding the Roman infantry (no 'off-board pursuit', incidentally).

Where Hannibal's plan differed was the availability of extra infantry and the unavailability of sufficient elephants.  Xanthippus' 12,000 infantry (total), less detachments supporting the cavalry wings, would have covered, at 16 deep, a frontage of no more than 750 yards.  Regulus, with 15,000 infantry less losses in earlier successful engagements, deploying at standard depth (21 men in lines of 8, 8 and 5 prior to detaching velites), would cover just over 700 yards.  Polybius indicates he deployed in double depth ('great depth', degree not specified, but double works), like Varro at Cannae, which would roughly halve his frontage (just over 350 yards).  Xanthippus presumably deepened his line to 32 deep to match.  The frontages would have been very similar, allowing for some of Xanthippus' infantry doing their cavalry support on the wings.

Of this 350 yard-ish frontage, about 13/15 was fronted by elephants (2,000 Roman infantry were left not facing elephants, beat their opponents and marched off the field, later being picked up by ships and making their way back to Sicily).

This gives c.100 elephants fronting c.300 yards.  Xanthippus seems either to have placed them in a very dense line more reminiscent of Indian than Hellenistic practice, or to have deployed them in files.  Either way they crunched into the Romans en masse, velites and all, and the Romans had no real counter except to filter through between the elephants and try to attack the formed Carthaginian infantry following up behind.  This led to each Roman being speared by multiple opponents as the meat-grinder pushed forwards.  The cavalry, surrounding the Romans on flanks and rear, prevented any effective redeployment.

Hannibal, by contrast, had something like 80 elephants and a longer frontage to cover: Scipio's OB is unclear, but a guess that he had around 24,000 infantry (exclusive of Masinissa's Numidians) is probably not too far off the mark.  A good rule of thumb seems to be 200 yards frontage per legion, and Scipio's troops amount to roughly four (large) legion equivalents.  The standard infantry frontage for a consular army appears to have been c.800 yards, so we give him that and look at the arithmetic.   His immediate opponents, the c.12,000 Celts and Balearics assumed to be Mago's old command could draw up on a 750-yard frontage 16 deep.  The approximately 12,000 Carthaginian citizenry would line up behind them in similar depth on a similar frontage.   Hannibal thus has about 80 elephants on a front of about 800 yards, or 1 elephant per 10 yards rather than Xanthippus' 1 elephant per 3 yards.  Xanthippus evidently felt this elephant density essential, because he left part of his frontage pachyderm-free and, as the event demonstrated, vulnerable to Roman infantry frontally - given his general competence he would not have done this by choice.

Hannibal cannot muster sufficient elephants to achieve the same density: he has only one per ten yards, and if he wishes to achieve the same density he can only achieve it on about one third of his frontage.  He therefore needs ersatz elephantry to help out and spread the effect: enter the Celts.

How the Celts were deployed can be surmised from the passage in Polybius that indicates that the Carthaginian second line ended up fighting Romans and mercenaries at the same time.  Were the Celts deployed in a solid line 16 deep, this would be patently impossible: the second line would have to fight aggrieved Celts first and tiring Romans later, not both simultaneously.  Furthermore, Polybius refers to their debut as a 'counter-attack' or 'charge' [trope ... epistestontes, a movement ... falling upon/assailing], so how do they reach Romans through the ranks of Celts, and reach them with enough organisation and impetus to cause the hastati to falter and the principes to be brought up ready to relieve them?

To me, the only arrangement that makes sense is that the Celts were deployed part as light infantry and part as discrete 'attack columns' intended to follow up the elephants and plough into any gaps they made.  This would give the Carthaginian citizen troops the exact role they had at Bagradas - spearing any Romans who managed to push forwards through gaps between elephants and Gauls.

So what went wrong?  Polybius records the Celts advancing at a measured pace, whereas a howling charge would have been more traditional and perhaps what Hannibal wanted - he may have underestimated Mago's training of these men, though normally he would have paid attention to such details.  Conceivably the Celts, who were all experienced warriors, preferred to let the elephants go on ahead and then pitch in themselves.  If so, it seems to have been a departure from any intent to send them in on the heels of the elephantry in order to maximise disruption.  Their slowness may have inhibited the advance of the second line, or the fact that the Carthaginian cavalry were going backwards pressed by Romans rather than the expected other way round (on the basis that Hannibal was surprised by Masinissa's presence) could have caused the Carthaginian citizenry to wonder if their orders would change, and the stillness of Hannibal's veterans in the rear could have induced them to halt, too.  And with Scipio's men blowing every trumpet they had, it might have been impossible to distinguish signals anyway.

Scipio's unusual deployment, with centuries one behind the other rather than side-to-side, made it easy for the elephants to take the course of least resistance and charge down the lanes.  It is quite possible the Celts also made for these gaps, expecting the Carthaginian citizenry to be marching up behind them to give them cover and support Bagradas-style.  Their measured progress may even have been with the aim of letting the citizenry keep up.  Having pitched in against the Romans and inserted themselves between the doubled-up hastati maniples, they were less than pleased to discover the citizenry had not followed them and were still hanging around in their initial position or close to it.  At this point two things seem to have happened: the citizen troops finally received, or were able to distinguish, the signal to advance, and the Celts concluded that the citizen troops had let the side down (the Celts, being experienced, could probably distinguish signals from background noise much better than the inexperienced citizen troops could).  Result: a tardy grand charge by the citizen troops which slammed into the hastati maniples which were already trying to cope with Celts in the gaps, and some simultaneous blue-in-blue strife as Celts in the gaps between the hastati conveyed with their blades what they thought of this late arrival.

So far Hannibal's new Bagradas had more or less achieved one of its aims: the Roman line had been disrupted and the Carthaginian citizens were able to present a coherent and cohesive line to a disrupted Roman one.  Not part of the plan was that the elephants had run out of control as a result of Scipio's trumpet and horn concerto and that the Celts had in part turned on their supports, but the Roman army was effectively pinned even if it was not being methodically masticated.  The second part of the arrangement, encirclement of the Roman infantry, had however come wildly unstuck.

Even with Tychaeus' Numidians, Hannibal did not have enough cavalry to wrap up Scipio's infantry Bagradas-style.  He did however have a substantial force of veterans who he told his other troops to rely upon as the 'sure foundation of victory' - in other words, they had a decisive role.  I shall not repeat Steven James' observations and conclusions here, but shall just observe that Hannibal had, or rather thought he had, enough cavalry to chase off Laelius's men (Masinissa being assumed absent) and more than enough spare infantry to encircle Scipio's.  The obvious conclusion is that the Carthaginian cavalry were tasked with removing their Roman counterparts and the veterans with gift-wrapping Scipio's infantry.

This would have worked had Hannibal indeed been superior in cavalry - and would have worked if he had known Masinissa was present and hence waited for Vermina.  The episode of the spies as told by Polybius makes it seem that Scipio really wanted Hannibal's spies to convey some specific information to their master, and the only variable that changed between their report and the battle was the arrival of Masinissa and his troops.  Ergo, it is hard to escape the conclusion that Scipio deliberately misled Hannibal in order to bring on a battle where Scipio would have a cavalry advantage - and before Vermina could reinforce the Carthaginians and make Roman success impossible.

[Justin Swanton]

Any ideas though on what would stop Hannibal's veterans encircling the Romans once both sides' cavalry had quit the field? Or to put it another way, how would the veterans have executed the encircling manoeuvre? Something like Scipio's own manoeuvre at Ilipa? For me it's still a toss between Hannibal being unable to send the veterans around the flanks and him judging he didn't have the time for it.

[Mark G]

I have to say I find Todd Cary far more convincing that James as an analyst of this battle.

[Justin Taylor]

On answer to Justin S. to my mind the key to not sending the veterans round Scipios flanks, is timing.

At the time that the Carthaginian cavalry had left the field, the elephants and the mercenaries had been defeated. The second line either were about to or were already fighting the Roman infantry.

So now would not be a good time to remove the 'back-stop' and leave a big gaping hole in the middle of his army.

[Patrick Waterson]

The classic method for moving the veterans would be to form a 'column' of subunits (essentially the subunits on the left wheel left, those on the right wheel right, and the resultant serpentine chains of subunits snake out round the flanks).  These would trudge past the embattled battle lines and swing in behind the Romans, Cannae-style, hopefully meeting up behind the Roman army, at which point the subunits all wheel to face the target and close in.

Such an arrangement is naturally somewhat sensitive to cavalry thundering down their projected course track, not to mention stray uncontrolled elephants crossing the intended line of march.  A massive dust could is also not conducive either to seeing where one is going or to finding out exactly what is going on amid all that noise and general background of incessantly drumming hoofbeats.  One can envisage Hannibal cancelling the projected movement when he saw his own cavalry begin to fall back, and declining to reinstate it both when the respective cavalry contingents thundered past and subsequently while he feared the possible return of the Roman horse.

I would think that the architect of Cannae would have been able to send troops around the Roman flanks, and that the veterans who had accompanied him back to Africa were more than capable of achieving the manoeuvre.  Why, then, did he not set it in motion once the cavalry had passed?

The most likely answer is that he did not know when the Roman cavalry would be coming back, and so delayed until the manoeuvre no longer seemed viable.  There is however another dimension: for the first time in his life he had been outwitted, outfoxed and out-thought.  He would also be considering Scipio's past record, how at Ilipa the Roman general had deployed his legions unconventionally, the three lines not behind each other, but side to side.  If Scipio did the same here, at Zama, Hannibal would find his encircling manoeuvre intercepted, with an infantry meeting engagement on the flanks.  Without the Carthaginian cavalry to inhibit the Roman infantry, Scipio might stop Hannibal's manoeuvre half-way, turning it into a mere extended slog rather than a brilliant and deadly embrace of a helpless foe.  Yet how could this be worse than what eventually occurred, a mere extended slog without even the benefit of his first two lines occupying the hastati?

My feeling is that Hannibal was a prey to indecision, and it was indecision that finished him at Zama.  Had he initiated the encircling manoeuvre as soon as it was obvious that the cavalry had gone past and were not about to fall on the rear of his line, he would have got it moving while Scipio's principes were still up behind the hastati ready to relieve them, while mercenaries and citizens were still fighting the Romans (in addition to each other).  Scipio would thus have had only the not-very-numerous triarii to counter the manoeuvre, and Hannibal's veterans would have beaten these in fairly short order, outnumbering them perhaps six to one.  Scipio would have had to take his principes from behind the hastati just as the latter were wavering in the face of the Carthaginian citizenry's determined charge, with potentially dire consequences for the entire Roman army.

And yet Hannibal shrank from trying.  He performed no bold strokes, indeed he fails even to sort out the mutual slaughter among the citizens and mercenaries.  His one clear action is to order his last line not to admit the fugitives from the first two - was his veterans' morale that shaky?  Were they so poor at resuming order after opening files?  Or was it Hannibal who had his presumably one and only attack of nerves, having seen his battle plan crumble before his eyes as Masinissa's troops appeared on the battlefield?  For a general who made his career out-thinking and out-foxing opponents, to realise he had been out-foxed himself would have been a shattering blow.

To me, the irony of Zama is that for all Scipio's cunning deception, Hannibal could still have won had he carried through his original plan.  He came so close in the event that had he started his veterans into action earlier he could have caught Scipio at a disadvantage and given him a thorough drubbing.  With the Roman infantry gone, the return of the Roman cavalry would serve no purpose.

On the question of timing, Polybius' account suggests that the mercenaries hung on for much longer than the elephants and cavalry.  The time was there had Hannibal chosen to make use of it.

[Justin Swanton]

Interesting analysis, Patrick.

I suspect too, that Scipio would have anticipated an encircling manoeuvre - he knew Hannibal and he remembered Cannae. Hannibal outnumbered him more than 2:1 in infantry. What else did he think the great Barca would do? I can't imagine him being content with hoping Hannibal would just stand and gibber once his lovely plan fell to pieces. There must have been a plan B in Scipio's mind.

Unless he knew Hannibal so well that he counted on him not doing anything once his best-laid plans had gang agley. Really good stuff for a novel...

[Mark G]

So I take it that these proponents of the 'abandoned flanking march' theory also have it that the accepted deployment at Cannae was false.

that is

that Hannibal at Cannae did NOT deploy is veterans in columns on the flanks (hidden from view by other troops),

and that they in fact deployed in a fighting line,
then redeployed to a flank march position,
allowed the gauls and Spanish to withdraw fighting and tempt the romans forward,
and then proceeded to march even further again in to the full flanking position.

since that is what you are saying was the expected plan at zama - multiple redeployments by the same troops in the middle of the battle.

initially into the fighting line (surely a compete waste of time if this is not your plan()
redeployment into columns to march to the flanks
redeployment again to a flank aligned position, awaiting their opportunity to march
and then again, the march itself past the side of the legions over ground which the cavalry had been fighting over (and as we know, troops were routing out past)
and again redeployment into a fighting position to envelop.

Cannae required only one redeployment
they begin in flanking columns in position.
the romans march INTO the trap as the Spanish and gauls withdraw.
the veterans redeploy once into the fighting position to envelop.

And his veterans at Zama not only have to be capable of doing this in the battle, and prepared for it the night before (remember, Hannibal did not chose the field, and seems to have had battle forced on him, so such a complicated manoeuvre plan seems improbable at first sight)
- but more still - the whole thing is cancelled at very short notice after the cavalry combat finishes, and those same troops are able to beat off superior numbers of Romans and appear to be winning the main fight right up until they are hit by cavalry in the rear.

It really does strike me that this debate was entirely correct to be under rules and reserves thread, since this whole plan looks to me like the sort of thing a wargamer would come up with after seeing all those troops hanging about at the rear doing nothing.

[Paul Innes]

Patrick's most recent post is interesting, and it sparked off a minor aside for me regarding Scipio's redeployment of the Triarii to the flanks of what is about to become a single line of legionaries.  If I'm reading it correctly, Patrick is suggesting that Hannibal was wary of sending his veteran infantry around the flanks of the main engagement due to a number of factors.  One of which is the possibility of an interception by the Triarii, the kind of thing Hannibal knew Scipio had pulled off at Ilipa by deploying the Triarii wide - which is indeed where they end up at Zama.  This would explain why they are not used centrally when Scipio gains the time to reorganise his troops for the final confrontation, something that was queried earlier on this thread.

In our large recreation of the battle, we placed the Bruttian heavy foot in the centre of Hannibal's third line, with his veteran spearmen to both sides of them.  This deploys the best troops in position to turn the flanks of the Roman centre, while the Bruttians are still fresh and could stand up to tired Hastati and the already blooded Principes.  It could also produce a stand-off between the veterans and the Triarii at the ends of the line; the result is a very close infantry combat which recreates the situation at the end of the day as the cavalry return.

To be more precise, I suppose I'm wondering if Hannibal chose to hold back instead of committing to the flanking manoeuvre because he suspected that the Triarii would be sent wide to counter.  He was right, and this explains why the Triarii are deployed at the ends of the infantry line instead of in its centre.

Just a thought!

Good discussion, by the way!
Paul

[Mark G]

I've found the James article now.

I also passed over Pats one and even Stuart Hay's more recent ones on my way back in time.

None offers any evidence for a flanking march in the course of this battle.

James cites Aelian twice - once to indicate that it was theoretically possible to form the veterans in two lines (which he takes to mean that they therefore could have had one to go left and one to go right), once to indicate that there was a march column formation in operation 400 years after Zama, which he takes to mean that this was the likely formation to use for such a redeployment.

the evidence is less than thin however.

the notion that the triarii were put on the flanks of the exhausted remnants of the principes and hastatii to block or prevent a (delayed) flanking marching route column by the veterans is ridiculous.

these same veterans were deployed initially in a standard third line, no one disputes that.

they were still present in that formation and in that place when the Carthaginian citizens initially refused to fight and retreated 'in cowardly fashion' when the mercenaries and elephants had failed.

They were still present front and centre when the citizens and the principes and what was left of the hastatii had battered themselves to exhaustion.

They did not take advantage of the pause which Scipio introduced after the first and second lines had fought to march down this (exposed) flank, but rather waited and allowed the triarii to fill the space in front of them

That Scipio needed to rest and reform the principes and then combine them again with the triarii is more than adequately explained by the simple fact that there were more veterans that there were triarii - Polybius even states that when these third lines met they were equally matched in numbers, enthusiasm, experience and equipment.

That is, Scipio needed these men to simply match the numbers and frontage he faced from the veterans.

putting the triarii on the flanks rather than putting the reformed remnants of hastatii and principes on the flanks is not some revisiting of Ilipa, nor is it a blocking attempt for a flanking march.

Its just a 50 50 choice - move the reformed men again to the wings, or leave them in the middle and redeploy the triari there - it has to be one or the other - but the point is all  it achieved was to match frontages.

Something I think we over look in this discussion - this was the only battle which Hannibal fought against Romans where he did not have to think about the next Roman army. 
There was no reinforcing army on the way, so he had no need to take risks and come up with a cunning plan - he just had to kill the men in front of him - and he sought to do this by replicating the same tactic which the Romans used themselves.

successive lines of fighting men, with the most experienced and well equipped forming the final line which wins the decisive victory.  A deployment based around the maximum killing possible rather than one based around a swift dramatic victory - something which he undoubtedly learned from fighting Romans for a dozen years.