The Battle of Pharsalus: The Battle

With the pieces in place it is now possible to try to model the battle. It has to be noted that this model should not be construed as an accurate description of the real battle of Pharsalus. Rather it is a kind of test of the models of Roman fighting. How well can the model that has been presented be used to describe an actual battle? The goal is not so much to understand Pharsalus better but to use it to assess the value of the models.

That said, there are some insights into the Battle of Pharsalus to be gained from the exercise of creating the models. Some have already been noted -- the space available constrains the way in which the cohorts of the legion could have been deployed and the easy-to-say description of the role of the 6 cohorts is not so easy to model. Other insights may develop as the model works through the battle.

The drawings of the battle will be presented with map tilted so that the armies appear horizontal to the viewer. This is a more customary way of seeing battle diagrams.

The starting position

The starting position below, with the armies 600 feet apart, is the same starting position used in the other earlier models.

This was to have been the true starting position for the battle sequence. That is, until modeling actually began. Almost immediately a major problem became apparent.

This starting position has the armies 600 feet apart. When the cavalry models were done, the model had them each starting their charge from this distance. The reason was to give enough time and space for one side to turn around and run before contact was made. Two groups of trotting horses could close a distance of 600 feet in 22 seconds. If one side is to begin forward movement, realize it is overmatched, and then spontaneously turn around and flee before the two sides actually come in contact then anything much less than 20 seconds would not seem like enough time.

The infantry, on the other hand, were modeled starting their running charge at a separation of 300 feet. Each side runs 120 feet in about 14 seconds, throws its pila when the separation is 60 feet and then each runs 30 more feet in 3.5 seconds. Each side runs 150 feet. Any longer distance would seem too far for heavily armored soldiers to run.

This scenario seemed fine when the models were built. Applying the model to Pharsalus uncovered a flaw.

For reasons already elaborated, the infantry are shown starting their running charge at a separation of 300 feet. At Pharsalus Pompey had his men hold their positions so that Caesar's men would do all the running. The only way this works is for the two armies to have completed their pre-battle maneuvers and to face each other at the charging distance, 300 feet of separation. If they were 600 feet apart when the final approach was begun the men would have to walk about 150 feet first, then start their running charge. Suppose that Caesar's men walked the first 150 feet and Pompey's men did not move. At the end of the 150 foot march they would still be 450 feet from Pompey's army, half again the normal distance that would trigger a run. Even if they were not suspicious of what Pompey's army was up to, they would not start running at that distance. By the time they marched yet another 150 feet they would be quite aware that Pompey's army was just standing and waiting. The whole scene Caesar describes, in which they started running, got half way, realized that Pompey's army was not also charging, and then stopped -- that sequence of events would be improbable if Caesar's men had marched 300 feet first and knew full well that Pompey's men were not moving.

The account of the battle, though told for dramatic effect and not for strict historical accuracy, does seem to indicate that the armies got to the "jumping-off"point and paused there. It seems that Caesar had time to address his troops just before they started their charge. It also seems that he might have moved the 6 cohorts to the 4th line at the last minute.

This would make some sense from a command perspective. Both Pompey and Caesar are described as looking over the opposing army, assessing its disposition and strengths, even making adjustments. They would have been able to get a much better idea of enemy strength and deployment from 100 yards than from 200, especially if they were positioned behind the first line of cohorts -- some 75 or 80 feet behind the front rank.

If the armies complete their pre-battle maneuvers and hold at the "jumping-off" distance of 300 feet for the general to assess the enemy, make last minute changes and/or give the traditional speech to the men, then then the cavalry has to be in a holding pattern as well. If the cavalry were lined up beside the infantry then its charging distance would be limited to the 8 to 11 seconds it would take them each to close the 300 foot gap. This would be insufficient time for one side to start a charge and then turn and flee. The other alternative is that the cavalry lined up well back of the infantry so that they would be 600 feet apart while the infantry was only 300 feet apart. This seems unlikely, at least for the stronger more confident cavalry. Even for the weaker, less confident side, lining up well back so that there is lots of time and space to turn tail and run for it would surely be a morale breaker. What general would allow that?

For all of these reasons the model of Pharsalus has to change. The "jumping-off" distance is 300 feet. The armies hold at that distance until one or the other initiates an attack. Both cavalry and infantry units are at roughly the same distances. The new starting position is shown below as Frame 1.

As these models progress there will be several different types of units and speeds involved. This chart may help keep track of just how fast each group of soldiers moves at any given time. All of the drawings are not only accurate in their dimensions, they are also accurate in the times and distances involved.

Frame 1: 0:00:00 Starting position (revised)

The revised starting position with a separation between the armies of 300 feet is shown below.

The curved blue line indicating the bank of the river is just visible on the left.

The infantry and cavalry are about 300 feet apart. Caesar's light infantry were in line with the cavalry in the earlier drawing. Because of the reduced distance they have been moved back so that they can not be so quickly overrun by Pompey's cavalry. To fill out the space, Caesar's cavalry has been reduced to a single line of turmae. That means the cavalry wing is only three ranks deep but has the light infantry behind it for support.

The title of the section "Frame 1: 0:00:00" indicates the elapsed time (h:mm:ss) with Frame 1 being time 0. Most of the frames will be 5 minutes apart. the first several are an exception, being just 1/3 of a minute apart to capture some of the action during the initial charge. Between frames 5-6 and 8-9 there are a longer time gaps because little of interest was happening then.

Frame 2: 0:00:20 The cavalry attack

Frame 2 is 20 seconds later. The attack has been begun by Pompey's cavalry. When both armies are holding at the jumping-off point, one side has to take the initiative in launching an attack. The other will respond but it will surely take some 10 or 15 seconds for them to react. In this frame Pompey's cavalry is trotting forward, the individual squadrons are no longer in a nice straight line since some started slightly earlier than others or move faster. The foremost squadrons have covered 264 feet, almost the entire distance between the armies. The rear lines of Pompey's cavalry wing have only moved a little. Caesar's infantry is just beginning its running charge, the furthest having run just a few seconds and covering 30 feet. Most of the cohorts have not yet even moved. Caesar's cavalry has almost no time to react. It has already begun its retreat, turning and trotting away to the rear, the light infantry running, trying to keep ahead of them. Two units move, respectively, 198 feet and 165 feet.

At Pharsalus the texts are unclear about if the cavalry charged first, charged at the same time as the infantry or charged later. In the model I have elected to show it charging first. Pompey's tactic was to have his cavalry take the infantry on the flank. He is said to have promised that Caesar's army would be defeated before his own ever had to throw their weapons. That must have meant that he thought the cavalry would be able to rout Caesar's cavalry and then turn to attack the flank before the infantry charge was completed. It also helps explain why he held his own line stationery. That tactic not only caused Caesar's line to run twice the distance but it also gave his cavalry more time to complete its assignment.

Frame 3: 0:00:40 The infantry pause in their attack

Frame 3 is just another 20 seconds later, the time it takes Caesar's infantry to cross half the distance separating the two armies. According to Caesar, the charging men realized on their own that Pompey's men were not charging. They stopped about half way, rested awhile, then resumed the charge. In this frame the some units have already stopped, some are still running but will shortly stop. The purple line is 60 feet from Pompey's front rank, the approximate range of the pila volley. All of Caesar's cohorts are well back of that line, safely beyond pilum range of Pompey's men, able to resume their run before they need to throw their pila.

Caesar's second line has moved forward too, but at the slower pace of the quick march. The third line holds its position. The cavalry and light infantry have moved roughly the same distances as they did in the last frame except that now they begin to angle to the sides.

Caesar's retreating cavalry has already overtaken the light infantry units. On the left the two light infantry units can run toward the infantry formation for protection. Those in the center have no where to hide from Pompey's cavalry. Already the model is awkward. Less than a minute after Pompey's cavalry starts its charge the light infantry are in difficulty. Although the cavalry turmae are moving only at a medium trotting pace, the whole action still seems to be happening too fast. However, the distances and speed of horses trotting would seem to dictate these kinds of movements, even in such a short time.

Frame 4: 0:05:00 The pilum volley

Caesar's cohorts could not all have stopped at the same time, nor would they be able to resume the attack without some rest, without re-dressing their ranks and files, and without some communication and coordination between the centurions. 5 minutes is estimated for the pause and renewal of the charge. On average the cohorts had about 50 feet to run to get within range of their pila.

The light infantry on the left has run to the safety of the infantry formation. The light infantry in the center and the cavalry have worked together to slow down Pompey's charge by making a fighting retreat of it. In this way they have turned the charge into a series of pauses, feints and short advances. Without any real justification the model will simply propose that the whole cavalry action has not moved more than another 150 yards during the time between frames. The blue arrows give a general sense of the action imagined for this period. The ending positions are shown in the drawing below.

The 1st line of Caesar's legions is at the distance from which they could throw their pila. The 2nd and 3rd lines have moved forward 75 feet. The 6 cohorts at the back have begun to shift positions, facing somewhat toward the cavalry action on their flank. Caesar's cavalry is retreating slowly, skirmishing with Pompey's force but not closing. This slows down the charge. Pompey's whole cavalry force has moved forward, followed by the slingers and archers.

Pompey is said to have promised that the battle would be over by this time -- before the pila were thrown. The cavalry should have already attacked Caesar's flank and caused a rout.

This is clearly impossible. Certainly the cavalry could by now be attacking the infantry. Perhaps they actually were. But even if the attack had been launched at the 4 minute mark, about the earliest it could reasonably have happened, the attack on one flank would not have prevented the left side of Caesar's line from carrying through the initial charge. At best the assurances said to have been given by Pompey are exaggerations.

Frame 5: 0:10:00 The retreat of Caesar's cavalry

Caesar says his cavalry fell back a little. Pompey's cavalry then turned to attack the flank Caesar's army. In his reconstruction, Delbruck has Pompey's cavalry chase Caesar's from the field, then turn to the flank attack, only to be themselves attacked on their own flank by the 6 cohorts and the returning cavalry. This is as good a general scenario as any. For it to happen, Caesar's cavalry should probably be seen to flee the field far enough that Pompey's quits the chase and turns back to the flank attack. It is that retreat of Caesar's cavalry that is shown in this frame.

The two infantry lines are fully engaged. At this point in the battle both sides have already thrown their pila, Caesar's veterans have closed the remaining distance and the two sides have been engaged in sword fighting for almost the entire 5 minutes that have elapsed since frame 4. That means that the front rankers are already starting to reach the time when they will be exhausted from the rigors of close combat.

There continue to be problems with modeling the cavalry. The three squadrons of Caesar's cavalry at the bottom are 3,450 feet from the line of battle. This seems a reasonable guess as to how far away they would have to retreat for Pompey's men to give up the chase and turn back to attack the flank. With those squadrons so far back there would be a wide open avenue for Pompey's cavalry to swing around the back of Caesar's whole army unless some of the troops offered resistance in that direction. This frame has all of the light infantry and one of the cavalry squadrons in position to protect against that move. Another reason for putting the light infantry to the side was that if it had retreated right along with the cavalry it would pretty much be out of the action. One could hardly expect even fit men to go into battle, make a fighting retreat for 2/3 of a mile, re-group, then run right back and attack.

Frame 6: 0:25:00 Pompey's cavalry prepared for the flank attack

In Frame 5 drawing Pompey's cavalry were scattered over 1/3 of a square mile. The time it would take to recall this force, re-assemble the scattered units and prepare for the attack on the flank is estimated at 15 minutes -- 5 minutes for all of the units to make their way back to an assembly point, another 10 minutes for the large group to get itself organized. That puts Frame 6 at 25 minutes into the battle.

On Pompey's side there is an embarrassment of wealth. There are so many cavalry squadrons it is hard to imagine what to do with them. They are shown scattered across the plain in Frame 5, but that is really just for looks. They serve no purpose doing that. The few squadrons at the front are all that are needed to chase Caesar's cavalry, the rest could be sitting on their horses back at the line of battle, just waiting. There is a problem with the skirmishers too. Where should they be placed?

An Alternative

The illustration below shows an alternative.

Is this a better reconstruction? It has just a few squadrons chasing Caesar; the rest have re-grouped for the flank attack. The skirmishers are arranged behind the cavalry. The problem with this is that there are enough cavalry forces to both pursue Caesar's cavalry and to launch the flank attack. Those chasing Caesar have no need to turn back and join the others. But if they do not it is hard to see quite how the battle turns around. If they keep pressing Caesar's cavalry and continue to outnumber it by at least 2 to 1, what would enable Caesar's cavalry to suddenly turn the tide?

Massing the cavalry for this attack is a challenge to model. One single cavalry squadron could attack the entire depth of Caesar's legion formation. As drawn, each squadron is two lines deep, some 6 ranks. For all of them to attack the flank of the legion the squadrons in this drawing would have to be arranged 6 or 7 deep. And this does not include the 5 squadrons chasing Caesar's cavalry.

Another Alternative

The skirmishers and cavalry squadrons could have been used for a massive attack along the rear of the army. In that scenario they would ride right around the back and swing in to charge the back of Caesar's lines. The cavalry could spread out and attack all along the army. Although this did not happen at the actual battle of Pharsalus, it would have made a reasonable plan. One small part of the cavalry could be designated to attack the flank, the rest would be ordered to stay wide, ride all the way around and come up from the rear.

If this were the tactic, then, when the small force designated to attack the infantry flank was surprised by the 6 cohorts and quickly routed, the larger body of cavalry and light infantry further out would see the rout. Not knowing what had caused it, they would also turn and flee, leaving the slower light infantry behind to be slaughtered by Caesar's men.

The drawing on the right shows this scenario and some of the troop movements that would result.

First Caesar's cavalry would turn on Pompey's pursuing force [1] and rout it [2]. Then the cavalry, light infantry and 6 cohorts would attack[3] the flank of Pompey's squadrons which had been in position to attack the flank of Caesar's own infantry. When they are routed and gallop away to the hills [4] the rest of Pompey's cavalry, the large massed unit on the right of the illustration, follows suit and also flees [5]. This leaves the cohorts of slingers and archers (the open squares at the rear of Pompey's cavalry block) unprotected. They are attacked and killed [6] by Caesar's forces which then turn to attack the flank and rear of Pompey's army [7].

Fatal Problems with the 7,000 Cavalry Model

The 7,000 cavalry simply do not work. The source texts say that they attacked Caesar's army on the flank, where the blue arrow points. The drawing has the 6 cohorts and light infantry spread out so that it looks as if the cavalry should just attack straight ahead from the position shown. But this should not have been the case. The attack by the 6 cohorts presumably took the cavalry by surprise and came on its flank. Neither Caesar's light infantry nor his cavalry are even mentioned as participating in the rest of the fighting; it is Delbruck who argues for their continuing and important role. The 7,000 cavalry, if the model is to follow the sources, should be seen as attacking the flank of the normal 3 lines of the legion.

The drawing does not show this because it is impossible to do so in anything like a credible manner. The depth of the 10th legion, the one that was placed on the right flank, from the front of the first cohort to the back of the cohort in the third line is 626 feet. If cavalry line up knee-to-knee the horses cannot get closer than 3 feet to each other. Therefore, if there were an absolutely solid line of cavalry it would not be possible to fit more than 209 men and horses in the 626 feet of front. 7,000 cavalry arranged 209 files wide would be 33 ranks deep. This formation would be foolish, at best.

A More Credible Model

The only way to make a marginally credible model is either to imagine a wholly different role for the 7,000 cavalry (for example, as suggested above, most could be intended to circle and attack the rear) or to follow Delbruck after all. The lesser of two evils seems to be to follow Delbruck, it involves less free-floating invention and does have the merit of following one of the great military historians.

His claim that Pompey may only have had 3,000 cavalry begins to look more attractive. The following drawing scales down Pompey's cavalry to Delbruck's number. It also re-arranged the skirmishers to show them in two long lines each close behind the cavalry.

Compare the cavalry positions above with those in Frame 1. If Delbruck were correct, it would certainly make modeling the battle easier. However, as was indicated earlier, since no one since Delbruck has seemed to follow his numbers for the cavalry the model will follow what appears to be current thinking and stay with the 7,000. One value of the models, however, is to point out problems with standard descriptions. It is easy to say there were 7,000 to 1,000 but much more difficult to translate that to accurate models.

The next illustration is a revised Frame 5 showing the position with 3,000 instead of 7,000 cavalry.

This shows the same scene as in the first Frame 5 illustration. Caesar's cavalry have drawn back a considerable distance from the line of battle. This time, though, they are only being chased by two small squadrons of Pompey's cavalry. The premise is that, with Caesar's cavalry apparently in full rout, the main body of Pompey's cavalry holds at the line of battle in preparation for the flank attack, leaving it just a couple of squadrons to keep up the chase.

The next scene is the revised Frame 6, still at the 25 minute time mark. In it Pompey's cavalry and skirmishers have re-deployed for the attack on the flank.

At the bottom of the illustration Caesar's cavalry, aided by some of the light infantry, have turned to attack the two squadrons of Pompey's army. This "event" is purely speculative; it is an attempt to imagine some logical sequence of events that may have happened. The rationale is that Caesar's cavalry would have had to be at some distance for Pompey's to ignore them and turn to the flank attack. They would hardly have ridden off the field if no one was pursuing them. If Pompey's whole cavalry did the chasing then it is difficult to imagine how Caesar's retreating cavalry could turn the tables on the larger pursuing force.

Meanwhile, the two infantry forces are heavily engaged. Both of their second lines are involved in the fighting. The scene shows most of the line at some distance, according to the lull and stand-off model of fighting. The third lines of both armies are held in reserve.

The model of the battle now continues with Frame 7, using 3,000 instead of 7,000 cavalry for Pompey.

Frame 7: 0:30:00 The attack of the 6 cohorts

This frame comes another 5 minutes later. The frame projects a couple minutes for fighting between Pompey's cavalry and Caesar's, for Pompey's various sub-units to decide to flee and for panic to spread. It would take about two minutes for Pompey's cavalry to ride the 1900 feet back to the main body.

This frame shows 4 squadrons of Pompey's cavalry charging the flank of Caesar's infantry. Simultaneously the two squadrons fleeing from the bottom ride into the formation, chased by all of Caesar's cavalry and half of the light infantry. Caesar's 6 cohorts are starting their attack on the flank of Pompey's cavalry; they are supported by the other half of the light infantry. These events will trigger panic among all of Pompey's cavalry, resulting in them turning and fleeing at top speed.

Frame 8: 0:35:00 The flight of Pompey's cavalry

Caesar says that Pompey's cavalry fled to the highest mountains [montes altissimos], not to their camp. Fromthis battleground high mountains that are accessible to mounted men could only mean two locations labeled A and B in the illustration on the right. Location A is not a likely choice because it is beyond the camp. If the cavalry rode that far they would have just gone into the fortified camp not the high ground around it. Morgan, whom I am following for the location of the battlefield and camps, prefers location B, the hill closest to the battle itself. This makes obvious good sense; when fleeing for their lives they went to the closest high ground. This hill is about 3,000 yards from the battlefield, a distance a horse could gallop in 5 to 7 minutes.

In Frame 8 the scale is the same but the size of the frame is bigger to show Pompey's cavalry fleeing to the hills away from the field of battle.

Pompey's cavalry are breaking in to small groups riding for the protection of the hill. They are widely scattered, reflecting the different positions they started from, their proximity to the fleeing squadrons being chased by Caesar's men, and their reaction time. The skirmish units are shown trying to follow suit but they are not able to keep up. At two points the squadrons of Caesar's cavalry are seen overriding skirmishers. Most casualties in ancient battles came when troops broke and ran; pursuing and killing them was one of the deadliest uses of cavalry. The light infantry follow close behind Caesar's cavalry. When they catch the skirmishers it will be an uneven battle. The light infantry, with their spears, swords and shields would have the advantage. The only chance for the archers and slingers would be if they could stand and launch dense barrages of missiles, keeping both cavalry and light infantry at bay. However, once the formations broke and ran there would be no chance for them to do that on open ground. If they were to find shelter of some kind (fortifications, a natural barrier, even a solid formation of friendly infantry) they might re-assemble around that point and then have a chance. But, running over open ground, they would have no chance to save themselves.

The main infantry battle continues unabated. In this scene, unlike the last, the two lines are shown in close combat. The second lines of both armies are fully involved in the fighting. Caesar's third line continues to be held in reserve.

The 6 cohorts are re-grouping and resting.

Frame 9: 0:55:00 The killing of the archers and slingers

Caesar says that the archers and slingers were cut down, then his forces turned on Pompey's flank. He does not mention any role for his own cavalry or light infantry but, as Delbruck correctly points out, the heavy infantry of the 6 cohorts could not chase down the archers and slingers so the other forces must have done the deed.

How long does it take to kill 4,200 archers and slingers? Since the model followed Delbruck, it credits Caesar with 2,000 cavalry and 2,000 light infantry to do the job. Not enough men to surround Pompey's men, it would have been done by chasing them down or using the cavalry to cut off their retreat while the light infantry came up behind. In either case, the process could not have been quick. Frame 9 allows 20 minutes for the process. This may or may not be a good guess.

The drawing shows the whole battlefield with Pompey's cavalry out of the action, taking refuge on the hill to the right.

The close-up below shows that most of Pompey's skirmishers have been killed; they are shaded a solid blue. (For simplicity's sake, all of the men within these units are considered to have been killed. In reality many would only be wounded at this time, probably to have been killed later. A few would have escaped.) They all tried to run, the furthest any unit is shown to have gotten is just under 1,000 feet from the previous position. The model has them being cut down in small groups wherever Caesar's forces could catch them. Three small groups survived and are fleeing back toward the camp. Caesar's cavalry and light infantry are scattered all over the field. The 6 cohorts have re-grouped, rested and are prepared to launch an attack against the flank of Pompey's infantry.

At the time of Frame 9 in the battle the two infantry armies continue the fight. At the scale of the drawings above it appears that the entire front lines are in contact with each other. However, this seems unlikely. As was shown on "The Infantry Fighting Model" page, after the initial clash the cohorts would have fallen back and fighting along the line would be sporadic and localized. The third line of Caesar's cohorts has moved up to maintain a constant distance as the right side of the line pushes Pompey back.

The illustration below is an extreme closeup of the a small section of the battleline. The vertical red line is 83 feet long. In this section of the battleline there is actual sword fighting at only two places, indicated by the jagged red lines. All of the other cohorts would be either resting, harassing the enemy by throwing pila at them, reinforcing their ranks, or preparing for a sortie.

The red circle indicates a problem area for the model. Most illustrations of Roman tactics simply indicate that one side was pushed back. But when one attempts to model this based on the centuries, the cohesive small units from which the legion was built, then problems arise. As the battleline becomes jagged just how can the centuries be visualized. This problem is addressed in detail in the page "Battleline".

Frame 10: 1:00:00 Preparations for the flank attack

The 6 cohorts have moved about 600 feet and are about 100 ft from Pompey's flank, the position from which they would begin a running charge. The cavalry and light infantry have begun to re-assemble. Some light infantry units are assigned to dispatch the wounded, they are scattered throughout the battlefield.

Caesar, positioned on the right side of his line, should have been able to observe the victory of the cavalry. He says he committed the third line at the same time the 6 cohorts attacked.

Delbruck maintains that the 6 cohorts were not enough to turn the flank and that the real credit goes to the Xth legion. He says that Caesar said as much in his letters but by the time he wrote the commentaries the Xth had mutinied and was out of favor. Public opinion in Rome was frowning on the notion of him using barbarian Gallic cavalry against Roman soldiers so he gave almost no credit to his own cavalry either. Therefore, all the credit in the commentaries went to the 6 cohorts. As with question of the sizes of the armies, later authors do not seem to have picked up on Delbruck's thinking. The model does reflect the larger role of the cavalry based on the impossibility of the cohorts chasing down the skirmishers but it does not include Delbruck's notion of the role of the Xth legion.

For a detailed study of the dynamics of the flank attack please see the Addendum at the bottom of the page. Click here to view the Addendum now.

Frame 11: 1:05:00 The 6 cohorts attack

The 6 cohorts are at pilum range. The cavalry is preparing to attack the rear of Pompey's infantry. The solid blue boxes represent the skirmishers killed or wounded.

Frame 12: 1:10:00 The cavalry attacks

This frame shows the cavalry joining in the attack. They way this attack is envisioned is that the cavalry would sweep across the rear of the army rather than charge directly into it. The mere presence of the enemy horse behind them would create panic in the infantry, even if there were no physical contact. The cavalry is in small units, one following another in the sweep. This style of attack seems reasonable but is purely my own invention and should be taken as such.

This is not attested to in the literature except in Delbruck's analysis, and even he does not specifically say that the cavalry attacked the rear of Pompey's army. At this point in the development of the model the dynamics seem to almost dictate it. The cavalry has dispatched the skirmishers, rested awhile, and is within easy striking distance of Pompey's army. Furthermore, the slaughter of 22,000 men would normally be viewed as happening during the rout when the cavalry chased down the running footsoldiers.

Arguing against this cavalry action could be Pompey's actions. It is said that he saw the dust of his cavalry, deduced that it had been defeated, gave up the fight and fled to the shelter of the camp. He is reported as being on his left flank by Caesar but on his right by Plutarch. If he were on his own left the cavalry action shown above would have already overrun his position. There is no indication that he himself was the subject of attack or that he was chased from the field. However, if he had been on his left flank it would seem that his knowledge of the cavalry battle would have been much more direct than watching for dust clouds.

These models are not trying to settle questions like this; they remain unanswered. The model is simply trying to use what we know about Pharsalus as the basis for attempting a more detailed representation than is normally seen. The discrepancies between what the model shows and what the texts say is raised here so that the model itself is not misleading. It is a reasonable but not a certain interpretation of what might have happened.

Frame 13: 01:15:00 The rout begins

This frame is set 5 minutes after the previous frame. The time here is just a guess. How long it would take an army to turn the flank of an army, for it to break and run or for the cavalry to sweep through the fleeing men, killing them as it goes? A few minutes? Many minutes? It is, literally, anyone's guess. Five minutes is just a convenient interval. The only actions that are closely tied to the time interval is the distance the fleeing cohorts cover.

The model shows the infantry retreating roughly 1,000 feet in 5 minutes. This is not a fast pace; it is actually slower than a march. There are a number of factors that would slow them down. The distance from the furthest cohorts to the camp is about 2 miles. Heavily burdened men could not run that far, certainly not after fighting a battle. Even though Caesar says that they threw away their arms and standards, they were still burdened with the armor they were wearing and, surely, some would have retained swords or even pila. Many men would have been wounded. During the flight there would have been frequent need to stop to defend against attack, if only to throw up a shield to parry a lance-thrust. Some individuals may have run faster but, overall, a relatively slow retreat seems reasonable. The chart to the right gives speeds, distances and times for the various units.

Caesar claims that Pompey lost 15,000 men. If the skirmishers lost 3,000 of their 4,200 that leaves 12,000 casualties among the infantry cohorts. Using the standard casualty figure of 5% during active fighting, Pompey's force may have lost as many as 2,000 during the battle up to this time. Surely some were lost in subsequent fighting during the day, say another 2,000. That would leave 8,000 to have been killed during the pursuit phase of the battle. 8,000 casualties would equal 17.7 of the 450 man cohorts the model uses. For the sake of simplicity, those casualties are shown as 18 complete cohorts wiped out. As with the skirmishers, cohorts considered casualties are colored solid blue. In the frame above there are 4 cohort casualties already.

Frame 14: 1:20:00 The rout continues

This and the following frames will illustrate many of the same events. The flight of the cohorts toward their camp continues. Some cohorts in close contact with Caesar's infantry or cavalry move slowly because they are defending from attack. Other cohorts move more quickly, some even at a running pace. The cavalry and infantry chase down cohorts which become casualties.

Frame 15: 1:25:00 The cavalry rests

The most prominent feature of this illustration is the grouping of the cavalry on the left. The cavalry have already fought their own battle, they have then ridden the entire length of the battlefield. From their last resting and re-grouping point to this location is about 1.8 miles. Their horses would need a rest by this time. The model shows the cavalry making a single sweep across the back of Pompey's army and then resting and regrouping on the far flank -- this is pure conjecture. Many other scenarios are possible. They could simply charge directly into the rear of the army, as mentioned earlier. They could hang back from the rear and make quick sorties into it here and there. They could split up and attack several places at once. They could attack one location and then, when those cohorts turned to flee, they could chase them down and ignore the rest. With no real evidence from the sources, it is a matter of choosing a scenario that is reasonable.

All of Pompey's infantry are shown as now in flight. It is difficult to imagine just how this happened. The model shows the flight starting with the rear cohorts as they become aware of the threat of cavalry behind them. As they begin to run away the cohorts in front of them also panic and flee. Those cohorts in contact with the enemy are more difficult to imagine. They may just turn their backs and run, hoping they can outrun the enemy, or they may try to make a fighting withdrawal. In most cases they are shown with some separation between them and Caesar's pursuing cohorts, indicating that they ran away and are maintaining their distance. In a few cases Caesar's cohorts are touching them, chasing them down, causing casualties. The presumption is that as the men at the back are caught and killed those at the front simply try to run faster; they do not stop, turn around, and help out. It seems likely that Pompey's soldiers would have been able to outrun Caesar's infantry simply because they would be more willing to lighten their loads by discarding their shields, pila and even swords. Caesar's men would most certainly have held on to their weapons, making them much slower.

The three squadrons of light infantry on the right flank are shown harassing the retreating cohorts. This may be problematical. These units are represented as armed with several light javelins. If they used up their javelins in the previous fighting they would have nothing but hand weapons left at this time. Being so lightly armed they would not present much of a threat to the heavy infantry. If some of them did still have a few javelins left, or had reclaimed some from the battlefield, they could harass the heavy infantry but still would probably not choose to come into close contact. And once those few javelins had been thrown the light infantry would be back to hand-held weapons.

The pursuit by Caesar's cohorts is widely scattered. Without the extra boost of adrenaline that Pompey's men would have (literally running for their lives) the tired cohorts of Caesar's infantry might not have mustered the same level of energy. Some cohorts are shown making the chase, others are shown not pursuing as actively. On the far right Caesar's men have caught and surrounded two cohorts of Pompey's army. Cohorts, if they stood and fought in a defensive formation, could hold out against larger numbers rather well. These two are still fighting, still holding their own.

Frame 16: 1:30:00 The cavalry re-groups for a second attack

Caesar's cavalry has re-grouped and is ready to mount another charge into the retreating cohorts. The 3 light infantry squadrons on the right have broken off contact. At the bottom right the cohorts of Caesar have surrounded two of Pompey's cohorts which are still holding out.

Frame 17: 1:35:00 The cavalry makes a second attack

The cavalry has attacked again. Three more cohorts are shaded blue, casualties of the cavalry attacks. Otherwise the retreat continues. The pursuing infantry have fallen off the chase, leaving the pursuit to the cavalry.

Frame 18: 1:40:00 The retreat continues

The cavalry has killed another cohort or two. 13 cohorts are now casualties.

Frame 19: 1:45:00 The legions begin to re-group

Caesar noted that it was mid-day when Pompey's army fled to its camp. By that time his soldiers were exhausted from the fighting and from the heat of the day. Still, he exhorted them to attack the camp.

Since the legions did not pursue Pompey's men right to the gates of the camp and continue on with an immediate assault it would seem that the whole army may have stopped the pursuit at some place. This would have been well out of bow range of the enemy on the camp ramparts. But it may well have also been only part way to the camp. Once Caesar's men saw that they would not chase down Pompey's men on foot there would be little reason for them to exhaust themselves running. The cavalry could carry out the pursuit much easier.

The model has the legions slowing and stopping their pursuit at about this place on the battlefield.

Frame 20: 1:50:00 Caesar's army stops

The pursuit is all but finished. A few cavalry squadrons continue the chase. In this model 17 cohorts have become casualties, leaving only one more to fall. Caesar's army has now stopped and begun to re-assemble into legions.

Frame 21: 1:50:00 The end of the battle

The final cohort has been caught and killed by the cavalry. The rest have reached the safety of the camp.

The army continues to re-form into its legions. The cavalry and light infantry rejoin them.

Frame 22: Subsequent events in the day

Caesar says that he exhorted his men to attack the camp; this was at about midday. The 7 cohorts left to defend it could not hold out, the fleeing army had dropped its arms and was of no help. Pompey's force fled to a high mountain which did not have water on it. Caesar began to invest it to cut them off from water (this action is not illustrated in the drawing below). Seeing this, Pompey's force fled from the mountain over the high ground. Caesar gave chase with 4 legions, cut them off, and again began to dig entrenchment's to cut them off from water, completing the job by nightfall.

According to Morgan the location of the final confrontation was back up the road to Larissa. This location is off the detailed portion of the map. It would be somewhere further up the road (brown line) in the direction of the two arrows.

The duration of the battle

The battle model has taken 1 hour and 55 minutes. Is this a credible time frame? Caesar says that they were breaking camp when scouts reported that Pompey had offered battle on more favorable terms. He then led his army out. In other descriptions it seems clear that he was willing to break camp quite early in the morning -- as early as 4 or 5 AM. In this case he does not say. There is no particular reason to assume an extraordinarily early start. As a tentative time frame, assume that they began to break camp at 6:00 AM and were an hour into the process when the scouting report came in, making it 7:00 AM. Leading the army to the site of the battle and deploying it was almost certainly a matter of several hours. The battle may have begun around 9:00 AM. Midday, when the camp was assaulted, could be as early as 11:00 AM or as late as 1:00 PM. An earlier finish to the main battle would leave more time for the attack on the camp, initial investment of the hill, 6 mile pursuit, and final investment -- all of which happened before nightfall. Two hours for the battle may be a bit short if it began earlier than 9:00. Still, for a model, 2 hours is within a generally reasonable timeframe. This is not to claim that the battle happened in time increments similar to those used for the model. Not at all. It is merely to suggest that the overall time frames used in the model are not obviously unreasonable.

Evaluation of the model

Was the effort to model Pharsalus a success? If the goal was to create a definitive image of what actually happened then the answer would, of course, have to be "no." We simply do not know enough to generate such a model. But that was not the goal. The aim was to apply the modeling technique to a real battle to present a more realistic and more interesting view of Roman fighting, one based on an actual battle. A secondary aim was to use the modeling process to identify aspects of the battle that may be poorly understood.

Is the model helpful and interesting? That is something each reader will have to decide. Can the model identify aspects of the narrative that require additional clarification. Yes, I think it can.

The page "Pharsalus: Pre-Battle" discussed the location of the camps and battlefield and showed that careful scaling of the army formations can shed light on potential locations. The models can help understand where Pompey might have lined up his army for battle and what the advantages of the various positions might have been.

The model demonstrates the difficulty in accounting for the 7,000 cavalry claimed by Caesar. Though the mere fact that it is hard to model that size a force is not proof that Caesar's numbers are wrong, the model does support Delbruck's arguments.

Descriptions of battles tend to gloss over the details. It is easy to say that 7,000 cavalrymen attacked the flank of the army. The models force clarity to those kinds of statements. In a similar vein, statements that an army was routed and pursued to the gates of its camp are made much clearer when modeled.

Another description easily glossed over is the exact role and function of the skirmishers: both those lined up behind Pompey's cavalry and the light infantry interspersed among Caesar's cavalry. When confronted with the necessity of getting the archers within bow range while allowing room for the cavalry to maneuver in front of them the problems become much clearer. Similarily, working light infantry units into cavalry formations in a way that does not have them simply ridden down in the first rush is much more difficult than simply saying they fought alongside the cavalry.

The location, formation and use of the 6 cohorts comes into sharper relief when presented in the model.

Placing the army models on an accurate map forces decisions about the formation of the Roman legion. The plane of Pharsalus is confined by the creek and the hills. This places limits on the width of the army formations. The model uses a triplex acies formation with 4 cohorts side by side in the front lines. There is some room between all of the units, but no place are there wide gaps. With this formation the army of Pompey fits, but not with a lot of extra room. Were the legion arranged in Sabin's dogtooth formation, with gaps between each of the 4 front line cohorts and the cohorts of the second line filling the gaps but just back from the actual battle line -- were that formation used the legions would be 3/4ths again as wide as they are and would not fit on the plane. Nor can one argue that Pharsalus is just an exception. Neither Pompey nor Caesar was forced to fight at this place. Both chose to do so. That must mean that each considered the terrain suitable for battle as they knew it. Each felt he could deploy his troops in the way he wanted to and in a manner which would ensure victory. When considering the various schemes for arranging centuries and cohorts within the legion, the space limitations of Pharsalus argue in favor of the narrower formations.

Limitations of the model

In the earlier pages of the update section the model of infantry fighting attempted to break the action into century-sized pieces. This conveyed the realization that actual fighting was a matter of nice straight lines and tidy drawings. The models used for Pharsalus reverted to cohort sized elements, simply because of the numbers involved.

Because most descriptions are so vague, the models had to be built on a large number of assumptions, estimates, and outright guesses. Even though most of these were indicated in the text, the result is still misleading. The power of the image is such that the fact that the image is based on guesswork is too easily forgotten. A few of the areas in which guesswork may result in erroneous images are noted below.

The location of the casualties to Pompey's army do not look right. Even allowing for the convention of counting casulaties in terms of whole cohorts at a time, the location of the casualties seems off. There are many at or near the original line of battle and relatively few over the course of the pursuit. A better model would probably show a much higher percent of the casualties coming in the mid to latter stages of the pursuit. The cavalry could be shown chasing Pompey's men right to the gates of their camp.

Caesar's infantry may well have continued the chase further as well. It does seem that they stopped short of the camp itself, but they may have been much closer than shown in the drawings.

The rout itself may be poorly imagined. Panic may have spread among the cohorts much more quickly than depicted. The entire army may have bolted and run almost as one. The model shows it taking about 10 minutes for the panic to spread from Pompey's left to his right flank. At one point Pompey's army is spread out over nearly a mile. If the panic spread quickly the army may have retreated in a much tighter grouping.

The model indicates the various time increments between frames. But so much of that is guesswork that it could be utterly misleading. The battle may have taken much longer than the model suggests or the duration of the various stages of battle could have been much different.


Addendum to Frame 10 at 1:00:00 -- Detailed Study of the Flank Attack

At just about every scale there is a temptation to gloss over the details by simply moving colored rectangles around, forgetting or ignoring that those rectangles represent individuals. And those individuals have both physical and psychological factors that determine what they could and would do in an actual battle.

The flank attack is an excellent opportunity to take a closer look at the difference between moving rectangles and viewing the scene at the scale of the individual soldier. .

This frame shows the general situation in the model at the point where Caesar's 6 cohorts prepare to make the flank attack. They will pivot around to attack the sides and rear of the flank of Pompey's blue cohorts. It looks like a simple maneuver when viewed at this scale and when the individuals are reduced to nice tidy rectangles.

But cosider the same scene at the level of the individual soldier. The next illustration will focus on just the individuals inside the purple rectangle.

At this scale the individual soldiers making up the ranks and files can just be distinguished. Yet, even this illustration has serious problems. The century formations have been drawn with all of the ranks and files in perfect alignment with each other. According to the model, the battle has now been raging for about one hour. After that length of time the century formations would be ragged, not rectangular. There would be casualties, centuries would have pulled apart in one area and crowded together in another.

Nevertheless, for the purposes of this addendum, it suffices to represent the centuries as tidy rectangles.

The two blue lines indicate some significant distances. The vertical blue line on the right hand edge of the drawing is 150 feet from the side of the century -- as is indicated by the lower connecting horizontal blue line. This is "normal" charging distance for the infantry. Meaning, when the infantry reaches that distance it can and probably will break into a run to generate energy for the pilum throw and subsequent clash of arms. According the the chart above, the infantry can close a 150' distance in less than 18 seconds.

The vertical blue line midway between the formations is at 60 feet from the flank of the rear century. This is the effective range for the pila. This line could be reached by Caesar's centuries in 8 to 10 seconds.

In this battle both sides were veteran Roman legionaries who would feel these distances and know the times involved in their very bones. It is a certaintity that the soldiers on the flank of Pompey's century would monitor the approach of Caesar's six cohorts and would be exquisitely aware of the threat to their own safety. They would be compelled to take defensive measures.

The next illustration zooms the view in to the purple rectangle.

At this scale the individual soldier can be seen clearly. The drawing shows 8 files and 4 ranks of the rear century plus one file of the adjacent centruy. Below, or in front of these men, are the Optio and officers of the lead century and 6 files and 3 ranks of the back corner of that century.

Some changes from the tidy rectangles and even from the mid-zoom view just above are apparent. The soldiers on the flanks have recognized their imminent danger and have taken up new positions to defend themselves against the flank attack. Initially they must defend against a pilum volley. But just seconds later they may expect to face a fierce charge from relatively fresh troops. They have no choice but to turn to face the threat.

But as soon as they turn the formation becomes immobile. As illustrated, two files now face sideways and all of their concentration is on the threat to that flank. Were the formation to move either forward or backward it would necessarily leave these men behind for they cannot keep pace by sidestepping.


They have to turn, but as soon as they do, the formation becomes immobile. It cannot advance or retreat because movement in any direction would require that some ranks and files step sideways or backwards -- and Delbruck has argued persuasively that these types of maneuvers in combat situations are impossible to do without disrupting the formation's integrity.

Click here to return to Frame 10

The Attack on the Skirmishers

The model depicts Pompey's skirmishers as running away when their own cavalry panics and quits the field. But quite quickly Caesar's cavalry catchs and surrounds the skirmishers, bringing them to bay, then killing them. In fact, this is quite possibly not how it would happen at all. Skirmishers armed with either a sling or a bow cannot effectively defend themselves while running away. To make us of either weapon the man has to stop, turn around and stand still for at least a few seconds. As long as the skirmishers can be kept running they are defenseless and can be cut down by the pursuing cavalry at will. But were they to be surrounded so that flight was no longer possible they would then be in a position to re-form their ranks and begin to make effective use of their slings and bows. Knowing this, the cavalry would want them to run, not stop and so would not wish to surround them at all.

Site Map

© 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.