As long as the units formed one continuous battle line they presented an unbroken front, but anything
that broke the integrity of the line resulted in units separated from each other. Roman military formations were
based on smaller units, centuries, which were rectangular arrays of ranks and files. In so far as the soldiers were able to maintain
those formations the rectangles were preserved. Once the integrity of the line was breached and gaps appeared there
were vulnerable corners at the exposed flank of the centuries. The illustration on the right shows an unbroken
line of centuries (red) and a broken one (blue). The vulnerable corners are marked with circles. The portions of
the red line that are unopposed are indicated with purple rectangles.
The traditional method of illustrating battles uses large rectangles that may represent anywhere from 500 to 10,000 men. At that level of abstraction it is easy to ignore the problem these corners represent. Even in the illustration above, the smallish gap between the blue centuries and the slight overlap on the far right do not appear to be significant problems.
If instead of boxes symbols representing the individual soldiers are used the problems begin to become more apparent. The illustration on the left is exactly the same as that above but this time the individual soldiers are indicated.
One more enlargement is necessary. In the two illustrations below the purple square is enlarged, first as symbols and then with human figures drawn in.
What appeared to be a minor issue when viewed as large boxes now becomes much more significant. Consider the lone soldier on the corner of the bottom century. He is about to be attacked by the man directly in front of him and the three men to his left who are unopposed. The fact that the gap is small when viewed from a distance means little to the individual soldier about to be attacked. He simply cannot survive in this vulnerable position. And, if he is killed, like dominoes, the enemy can roll up the flank of the century.
When illustrated as simple boxes it seems rather simple. The red and blue boxes meet and stop there. The minor gap in the blue line is ignored. But this cannot be an accurate view.
The blue century cannot maintain its rectangular formation with one man on the vulnerable corner.
The rex century cannot be held in place in a nice straight line.
No. The unopposed men from the red century will curl around to attack the lone individual on the corner. That man, in turn, must pull back into the formation for shelter. His comrades from the rear ranks must come forward to protect the open flank.
In short, the tidy rectangular formation disintegrates at its edges.
But there are different situations in which corners and edges arise. Each may have different repercussions. The next section explores some of the situations that could lead to exposed corners and flanks.
Imagine an ideal battle. There are two armies of the same size, arranged opposite each other in identical formations. The opposing lines line up perfectly. On paper. On the battlefield, no. For there are no field surveyors to precisely line up the two armies. So even under ideal circumstances, it would be impossible for two armies to line up so exactly that their two flanks meet perfectly. And, if they do not meet, there is an overlap at one or the other end of the line. Whether it is large or small, still it is an overlap. And to the soldier on that corner it doesn't matter a lot if there are 5 enemy soldiers out on the exposed flank or 50 because it is the 2 or 3 nearest soldiers who can kill him.
As the centuries charge gaps open along the line because units either don't keep up with each other or they naturally spread apart a bit. Gaps might be half a century wide to a full century wide, especially between adjacent legions. Even if centuries from the second line eventually come up to fill in the gaps, there are still gaps at the onset. And gaps mean corners. The end man in the formation is faced with multiple enemies and no help on the flank.
There is a growing consensus that there were many lulls in the fighting. Not only have men like Zhmodikov
and Sabin introduced arguments based on ancient sources, pure logic simply demands that there be some kind of pause
in the fighting. The simplest model is to imagine the Roman army charging, closing to hand-to-hand combat. As fatigue
sets in the two sides fall back from each other. The lull. At this point one can imagine the two armies some distance
apart, the front line men heaving with exhaustion, tending wounds, getting water, resting, taunting the enemy,
perhaps throwing darts or pila back and forth. The lull is followed by renewed fighting as either the front
line men rest up or fresh men filter forward to take their places.
It is the next phase that demands corners. It is simply not reasonable to imagine that either Roman or enemy army simultaneously decides to renew the fighting. Rather, small groups here and there along the line build up the resolve to attack. As these small groups move forward they of necessity have a limited front and, therefore, corners and edges to their formation.
Units were separated from each other either by design or by accident. At the Sabiw (Sambre) the two legions on either flank pushed the Nervii back and pursued them accross a small river, leaving the two legions in the center to fight on their own. In this example, at a minimum, there had to be exposed flanks on both sides of each set of two legions.
The army having been marshaled, rather as the nature of the ground and the declivity of
the hill and the exigency of the time, than as the method and order of military matters required; while the legions
in the different places were withstanding the enemy, some in one quarter, some in another, and the view was obstructed
by the very thick hedges intervening, as we have before remarked, neither could proper reserves be posted, nor
could the necessary measures be taken in each part, nor could all the commands be issued by one person. Therefore,
in such an unfavorable state of affairs, various events of fortune followed.
The soldiers of the ninth and tenth legions, as they had been stationed on the left part of the army, casting their weapons, speedily drove the Atrebates (for that division had been opposed to them,) who were breathless with running and fatigue, and worn out with wounds, from the higher ground into the river; and following them as they were endeavoring to pass it, slew with their swords a great part of them while impeded (therein). They themselves did not hesitate to pass the river; and having advanced to a disadvantageous place, when the battle was renewed, they [nevertheless] again put to flight the enemy, who had returned and were opposing them. In like manner, in another quarter two different legions, the eleventh and the eighth, having routed the Veromandui, with whom they had engaged, were fighting from the higher ground upon the very banks of the river. [Caesar: The Gallic Wars: 2.22-23]
There are examples of Roman armies being overwhelmed and surrounded but within the overall rout some soldiers formed defensive squares and fought their way out to safety. These formations must have had exposed flanks on both sides as well as being vulnerable both front and back. Not only were they able to defend themselves, they were able to cut their way through larger enemy forces and fight their way to safety.
21.56] They now stood fighting in a square, and about 10,000 of them, unable to escape in any other direction, forced their way through the centre of the African troops and the auxiliary Gauls who supported them and inflicted an immense loss on the enemy. They were prevented by the river from returning to their camp, and the rain made it impossible for them to judge where they could best go to the assistance of their comrades, so they marched away straight to Placentia. [Livy: 21.56 ]
Some experts assert that the famous exchange of lines in which the second line replaced the first was accomplished by using formations with gaps in the line of battle. These gaps would be at least the width of a maniple.
In virtually every ancient battle at some point the losing side is subject to an attack on its flank. This is frequently the beginning of a general rout. The attack may come from cavalry, from light infantry or from heavy infantry. Less frequently it comes against the exposed rear of the army.
In this section each of the seven scenarios described above will be explored. The way in which both the attacker and defender responded to each scenario was different.
Duriing a lull in the hand-to-hand fighting the two armies would be separated by some distance. Just what distance
would depend on the local circumstances and the fighting character and abilities of the men opposite each other.
The page Elements of Roman Fighting argued for about 60 feet as a psychological safe distance. After some period
of rest and perhaps of missile warfare one side or the other renews the hand-to-hand phase by charging. The key
factor is that the charge is localized. The charge could consist of a singlel century or could be a coordinated
effort of several centuries. In either case there will be exposed flanks of both sides of the charging units. The
illustration below shows several scenarios.
The top frame shows two lines approximately 60 feet apart during the lull period. The arrows indicate the advances of the various centuries. On the left side two red centuries engage two blue centuries. On the right one blue century charges the red line.
The precise number of units involved is unimportant. Because the centuries of the two armies do not precisely line up with each other it would be virtually impossible for any number of them to come together without some overlap on one side or the other.
What happens at the corners in situations like these.
To explore what happens at the corners and flanks a small part of the above frames will be isolated. The purple box at the left shows the edge of the blue century and a bit of the red century that overlaps the blue. This purple box is blown up on the right and filled in with soldiers.
If the blue formation were to keep a rectangular formation it would isolate the soldier on the corner. Since the red soldiers who are out on the left have no one in front of them they can easily curl around the flank and attack the corner man from several sides at once. The corner man would have no chance of defending himself against such an attack. Once he fell, the same tactic could be used on the next man, and so forth.
The blue century simply cannot adopt this tactic. It must somehow defend its exposed flank.
The following frames show one way this may have been done.
In the first frame the men in the third rank of the blue century begin to turn to the exposed flan. One man is in the process of throwing his pilum at the oncoming red soldiers.
In the next frames the process of rounding out the corner continues. More and more soldiers from the rear rank move into the open space to protect against the flank attack. The red army soldier curl around the open flank. The result in the final frame is a rounded corner.
The frame at the far right is the final image in the sequence. It shows how the blue formation must spread out into the open space to protect its own flank. The corner is rounded off so that no single soldiers are exposed to attacks from multiple directions.
The effect of this natural and necessary reaction by the red and blue centuries is to stretch and distort the rectangles into something like the amorphous shapes shown below.
Both the red and the blue centuries have now lost all semblance of ranks and files and have flowed outside of the theoretical rectangles that they should have occupied. This means that any scheme to replace front line men that depends on ranks and files or on careful spacing of centuries or even on certain spaces between men would probably not work. Once this phase of hand-to-hand is over and both sides fall back into another lull period, however, it would be possible for the centurion and optio to restore order.
The illustration below shows two armies ready for battle as they are usually drawn up.
All of the centuries are perfectly aligned with each other. Of course, only paper people and marching bands can achieve such perfect formations.
The illustration below shows a more realistic front line. The centuries are a bit out of alignment as would be normal. The formation does not look all that different from the perfect one above.
But when the two lines come together in the illustration below the effects of even small deviations in the alignment of the two armies becomes apparent. There are numerous gaps and overlaps at both ends and all along the lines.
This scenario is primarily concerned with what happens at the two flanks. On the left is a close-up of the right side of the blue line. It shows that the minor misalignments lead to a slight overlap -- slight when viewed at the scale of the entire battlefield. But when viewed at the scale of the individuals the overlap is significant. There are 3 and 1/2 centuries of blue soldiers on the flank unopposed.
It is beyond one's imagination to think that they would simply stop and stand in place. Nor is it reasonable to believe that the right-hand red century would placidly hold its position while it is being flanked by three blue centuries.
Something different has to happen at this flank. And since this situation would arise in virtually every ancient battle there had to have been a method of solving it that did not disrupt the battleline.
The illustration on the right shows the far right hand side of the line and three options that the red
The first, purple arrow, is that the end red century simply run straight at the end blue century so that it is not outflanked.
The second, green arrow, is that some centuries from the second line move up to fill in the overlap.
The third, dark red arrow, is that the front line red centuries spread themselves out thinner to cover the overlap.
Each of these alternatives will be explored below.
One can imagine that the centuries were trained to run directly at the end of the enemy line and eliminate the possibility of being overlapped. In the middle frame the red centuries are shown at the eventual battleline. The end century has run directly toward the end of the blue line. The other centuries have spread out to fill in. This creates small gaps but centuries from the second line move up and will eventually fill them in.
The problem with this scheme is that the blue army does not run toward the red along that same slanting line. In fact, it cannot. The blue centuries can only run straight ahead. When the two armies meet at the battleline there is still an overlap.
The only way "running right at them" could work is if the red centuries follow a continuously changing curved path. This does not seem practical.
The next alternative is to have centuries from the second line fill in the space. This seems like an almost obvious solution. The only difficulty is that it would be very difficult to estimate just exactly how much of an overlap there will eventually be. At location in the first two frames the red and blue armies are about 600 feet apart. At that distance it would be very difficult to make the necessary judgment.
The third frame shows the result if two centures are brought up from the second line. There is still an overlap but additional centuries cover the flank so that the blue army cannot freely attack it.
Overall this seems like a viable solution. It does require the close suppport of the second line.
In this system the centuries of the front line simply spread themselves thinner to cover the open space.
There are two problems with this solution. The first is the same as was noted above, at 600 feet it is difficult to estimate the eventual overlap and to know just how far out the line needs to be spread. The second is that the maneuver to change three or four centuries from a 10 file 6 rank formation to a 15 file 4 rank formation is not that easy to execute. However, given that the armies are far enough apart that there is no immediate threat, it could probably be done.
In the final frame the centuries of the second line have come up to reinforce the thinned out first line. This would be necessary to maintain sufficient strength on the wing.
In any of the scenarios what is clear is that the second line must come up to support or fill in at the onset. Battlefield drawings that do not provide for this may be inaccurate.
All three of the scenarios reduce but do not eliminate overlap. But they do limit it to a manageable problem. Second line centuries protecting the exposed flank prevent wholesale flank attacks. The remaining overlaps can be handled by rounding off the corners (Scenario 3).
Scenario 1 illustrated several ways in which gaps along the battleline might develop. Delbruck argued
that gaps were necessary to allow room for the centuries to maneuver during the initial charge. But, even if gaps
were not intentional, they were inevitable. These gaps between are relatively small. Wider gaps would
be temporary in that men from the second line would come up to fill them in in very short order.
The method of handling them can be essentiall the same as in Scenario 3 below.
This is a difficult situation to envision. There is ample evidence that it happened but almost no evidence about
just how the conditions the independent units faced.
The key element to understand is that having exposed flanks did not hinder these units from being quite effective. Take the case of the separated legions at the Sambre. The aggressive legions on either side had great success, too much, even, so that they pushed the enemy far back. This left the two legions in the center exposed. Those two had difficulties protecting their flanks, the aggressive legions did not.
Why were the aggressive legions able to operate with exposed flanks? The easy answer is that the enemy were on the run and there was no one to attack their flanks. This may, in fact be a sufficient answer.
What should be noted is that simply having exposed flanks was not a deterrent to these legions. Apparently it was not a concern, indicating that they had confidence in their ability to defend their flanks.
The difficulty with defensive squares is not so much how they defended themselves but how they retained
the ability to move. It is fairly easy to imagine a body of men forming a square and presenting an impenetrable
shield-wall in all directions. But how does such a formation manage to move?
In the quotation from Livy cited above he says that some 10,000 men formed a square and fought their way right through the enemey center, escaping to safety. Ten thousand men is about two legions. Two possible formations are shown below. Each square represents a century of about 60 men. The arrow indicates the direction the square would move.
Livy as not there so his information may be suspect. Almost certainly the defensive formation was not a nice rectangular formation as is illustrated. But just what it might have been we do not know.
The illustrations are meant solely to give an idea of the size of a square of 10,000 men. The smallest would fill two football fields.
The obvious question is how the formation could move.
Assume that the 10,000 compressed themselves into the smaller square. It is still more than 100 yards wide and deep. It would be impossible to signal from one side to the other. The only way men at the sides or back could know what was happening would be to feel the movement of the men next to them. And, like a line of cars at a stop light when it turns green, there is a considerable delay.
One presumes the men faced out to present a shield wall in all directions. How should we imagine movement. Do the centuries at the rear walk backward? Delbruck convincingly argues that this is impossible.
At this time there is no resolution to the problems. Simply raising them points out the danger of lightly passing over statements like Livy's and assuming that we understand what he was talking about.
There have been any number of attempts to explain the Roman line relief system described by Polybius. Many have suggested that the Romans actually fought with wide gaps in their battleline. Today most authorities would no longer support that idea. The possibility of intentional gaps is simply mentioned here without further exploration.
As noted above, virtually every ancient battle appears to have been decided by a flank attack of some variety or other. If, as we have seen, some units could defend against a flank attack, why could they do that while the army could not?
There would seem to be only two major differences. The flank attack against an army often came from the cavalry or the light infantry. Sometimes a heavy infantry formation might be able to get on the flank but this was not that common. Fast moving forces attacking the side or rear of the army could more easily cause panic. The reaction time would be greatly reduced the men might not have had enough time to take the necessary defensive measures.
The other major difference may be in the number of attackers. The first 6 scenarios deal with fairly small scale gaps and overlaps. When there are fewer than 50 men trying to exploit a gap the defensive measures may be effective. But in a wholesale flank attack by hundreds or even thousands those same defensive measures may not be effective at all.
Probably the most famous defense against a cavalry flank attack is Caesar's placement of the 6 cohorts as a fourth line at Pharsalus. There are any number of questions about Caesar's story and one ought not believe everything Caesar wrote. Nevertheless, the six cohorts do seem to have been effective at preventing a flank attack. This particular countermeasure was taken before the battle commenced so that there was no question of reaction time. And it involved a significant number of men, enough to counter several thousand cavalry.
The key to understanding flank attacks would, therefore, seem to be time and numbers. Rapid attacks did not allow time for countermeasures. Tactics useful to defend against small scale attacks were not effective against larger numbers.
© 2003, Gary Brueggeman. All rights reserved world wide. No part of this work may be reproduced in part or whole, in any form or by any means, without permission from the author.