Collections: Logistics, How Did They Do It, Part I: The Problem

In this three-part series (I, II, III) we’re going to be bowing to reader demand and taking a close look at the nuts and bolts of maintaining an army in the field.  In our last series, after all, we noted that before gunpowder the ability of a general to affect the course of a battle after it had begun was relatively limited; yet the best generals clearly had substantial ‘value over replacement.’  Often that value is reflected by their bold operational maneuvers – swift marches over difficult terrain to surprise enemies in places they were not expecting.  But if ‘march fast’ is the secret to spectacular victories, what is so hard about ‘march fast?’

That’s what we’re going to look at in this series.  Now we have discussed logistics before, at some lengthI am, after all, the orc logistics guy!  But when we’ve discussed logistics we’ve done so in terms of what an army can plausibly do and how fast it can plausibly move, all in rules of thumb.  As I’ve said before, such rules of thumb are common for military historians (and indeed, militaries as well though modern staff work is much more precise), but many of you have asked me to elaborate on what lies beneath those rules of thumb.  What are the essential tasks the army needs to do as it moves and what demands do they place on it? How do you actually do pre-railroad logistics for an army on the move?

Now I should note that I am going to learn fairly heavily here on my knowledge of the logistics system I know best (the Roman one, particularly in the Middle Republic), though the general rules will be applicable more broadly, as the nutritional logic that drives the system is the same for everyone with an agrarian subsistence system (on nomads with non-agrarian subsistence systems, we discussed them here and also note W. Lee’s recent article on the logistics of grass).  Of course the Roman system was famously effective, so where relevant I will note where Roman capabilities were exceptional and what armies might do if they lacked those capabilities (lost to the Romans, mostly).  Finally, I should note that the basics of these logistics systems hold mostly true through the ancient and medieval worlds and into the pre-modern, until finally disrupted by the development of railway logistics in the 1800s.  Consequently, some of my notes here will be to gunpowder armies (particularly early modern European armies); I’ll be noting where gunpowder changes things, but in terms of logistics those changes are modest.

The plan will be for this series to run in three parts. This first part outlines ‘the problem’ – the actual size of an army with a given number of combatants and the demands that creates. Then next week we’ll look at the regular tasks armies relied on to get the necessary supplies (food, wood and water, mostly) and how different armies handled those tasks, with some implications for what they could do. And then in the last week we’ll look at how armies move, with a particular focus on how a skilled general might move an army fast but also the potential dangers of doing so.

And before we march onward, if you want to support my logistics, you can do so via Patreon. If you want updates whenever a new post appears, you can click below for email updates or follow me on twitter (@BretDevereaux) for updates as to new posts as well as my occasional ancient history, foreign policy or military history musings.

(Bibliography note: The tendency on this topic is to write detailed studies of the logistics of specific armies rather than general guides to the concept. Nevertheless, perhaps the best starting point for understanding pre-railroad agrarian logistics is J. Landers, The Field and the Forge: Population, Production and Power in the Pre-Industrial West (2003), which discusses many of the general features of organic economies, including logistics from the ancient world to the pre-industrial gunpowder age. There is also a very handy logistics introduction to K. Chase, Firearms: A Global History to 1700 (2003), though most of the book is, of course, about firearms. For ancient Greek logistics, the best starting point is J.W. Lee, A Greek Army on the March: Soldiers and Survival in Xenophon’s Anabasis (2008), though one must note that the ten thousand were a professional mercenary army and most polis armies do not seem as well organized (a grim statement, as the ten thousand aren’t particularly good at logistics either). For the logistics of Alexander’s campaigns, D.W. Engels, Alexander the Great and the Logistics of the Macedonian Army (1978) remains, to my knowledge, the standard reference though it is dated and not without problems. For the Roman Army, the two standard books (both quite good) in English are J.P. Roth, The Logistics of the Roman Army at War (1999) and P. Erdkamp, Hunger and the Sword: Warfare and Food Supply in Roman Republican Wars (1998). For early modern armies, G. Parker, The Army of Flanders and the Spanish Road, 1567-1659 (2nd ed. 2004) is a good focused study. On the considerable role of women in all of this, there’s the older B.C. Hacker, “Women and Military Institutions in Early Modern Europe: A Reconnaissance” Signs 6.4 (1981): 643-671, and more recently J.A. Lynn, Women, Armies and Warfare in Early Modern Europe (2008); note also this article by Lynn over at HistoryNet on “Women in War” (2018), as it is openly available.)

The Backpack and the Belly

We’ve introduced this problem before but we should do so again in more depth.  Logistics in modern armies is rather unlike logistics in pre-modern armies; to be exact the break-point here is the development of the railroad.  Once armies can be supplied with railroads, their needs shift substantially.  In particular, modern armies with rail (or later, truck and air) supply can receive massively more supplies over long distance than pre-railroad armies.  That doesn’t make modern logistics trivial, rather armies ‘consumed’ that additional supply by adopting material intensive modes of warfare: machine guns and artillery fire a lot of rounds that need to be shipped from factories to the front while tanks and trucks require a lot of fuel and spare parts.  Basics like food and water were no less necessary but became a smaller share of much, much larger logistics chains that are dominated by ammunition and fuel.

But in the pre-railroad era (note: including the early gunpowder era well into the 1800s) that wasn’t the case.  Soldiers could carry their own weapons and often their own ammunition (which in turn put significant limits on both). For handheld weapons, the difference gunpowder made here was fairly limited, since muskets were fairly slow firing and soldiers had to carry the ammunition they’d have for a battle in any event. The major difference with gunpowder came with artillery (that is, cannon), which needed the cannon, their powder and shot all moved. The result was a substantial expansion of the ‘siege train’ of the army, which did not change the structure of logistics but did place new and heavy demands on it, because the animals and humans moving all of that needed to be fed.1 But overwhelming all of that was food and, if necessary, water.

Adult men need anywhere from 2,000 to 3,200 calories per day in order to support their activity; soldiers marching under heavy load will naturally tend towards the higher end of this range.  Now, these requirements can be fudged; as John Landers notes, soldiers who are underfed do not immediately shut off.  On the other hand, they cannot be ignored for long: no matter the morale an undernourished army will struggle to perform.  Starvation is real and does not care how many reps you could do or how motivated you were when the campaign started (in practice, armies that are not fed sufficiently dissolve away as men desert rather than starve).

Different armies and different cultures will meet that nutritional demand in different ways, but staple grains (wheat, barley, corn, rice) dominate rations in part because they also dominated the diet of the peasantry (being the highest calories-per-acre-farmed-and-labor-added foods) and because they were easy to move and store.  Fruits and vegetables were, by contrast, always subject to local availability, since without refrigeration they were difficult to keep or move; meat at least could be smoked, salted or made into jerky, but its expense made it an optional bonus to the diet rather than the core of it. So the diet here is mostly bread; many armies reliant on wheat and barley agriculture came up with a fairly similar idea here: a dense but simple flour-and-water (and maybe salt) biscuit or cracker which if kept dry could keep for long periods and be easy to move. The Romans called this buccelatum; today we refer to a very similar modern idea as ‘hardtack.’ However, because these biscuits aren’t very tasty, for morale reasons armies try to acquire actual bread where possible.

In practice the combination of calorie demands with calorie-dense grain-based foods is going to mean that rations tend to cluster in terms of weight, even from different armies. Spartan rations on Sphacteria were 2 choenikes of barley alphita (a course barley flour) per man per day (Thuc. 4.16.1) which comes out to roughly 1.4kg; Spartan grain contributions to the syssitia (Plut. Lyc. 12.2) were 1 medimnos of barley alphita per month, which comes out to almost exactly 1kg per day (but supplemented with meat and such).2 Both Roth and Erdkamp (op. cit. for both) try to calculate the weight of Roman rations based on reported grain rations and interpolations for other foodstuffs; Roth suggests a range of 1.1-1.327kg (of which .85kg was grain or bread), while Erdkamp simply notes that they must have been somewhat more than the .85kg grain ration minimum.3 The Army of Flanders was given pan de munición (‘munition’ or ‘ration’ bread) made of a mix of wheat and rye in loaves of standard size; the absolute minimum ration was 1.5lbs (.68kg) per day (Parker, op. cit. 136), somewhat less than the more logistically capable (as we’ll see) Roman legions, but in the ballpark, especially when we remember that soldiers in the Army of Flanders often supplemented that with purchased or pillaged food. Daily U.S. Army rations during the American Civil War were around 3lbs (1.36kg; statistic via Engels (op. cit.) who inexplicably thinks this is a useful reference for Macedonian rations), but some of the things included (particularly the 1.6oz of coffee) were hardly minimum necessities; the United States much like the Romans has a well-earned reputation for better than average rations, though this is admittedly a low bar.

So we can see a pretty tight grouping here around 1kg, especially when we account for some of these ration-packages being supplemented by irregular but meaningful amounts of other foods (especially in the case of the Army of Flanders, where we know this happened). There is some wiggle room here, of course; marching rations like hardtack are going to be lighter per-day than raw grains or good bread (or other, even tastier foods). But once meat, vegetables and fruits – and the diet must be at least sometimes supplemented with non-grain foods for nutritional reasons – are accounted for, you can see how the rule of thumb around 3lbs or 1.36kg forms out of the evidence. Soldiers also need around 3 liters of water (which is 3kg, God bless the metric system) per day but we are going to operate on the hopeful assumption that water is generally available on the route of our march. If it isn’t our daily load jumps from 1.36kg to 4.36kg and our operational range collapses into basically nothing; in practice this meant that if local water wasn’t available an army simply couldn’t go there.4

Marching loads vary by army and period but generally within a range of 40 to 55kg or so (60 at the absolute upper-end).  As you may well imagine, convincing soldiers to carry heavier loads demands a greater degree of discipline and command control, so while a general may well want to push soldier’s marching load up, the soldiers will want to push it down (and of course overloading soldiers is going to eventually have a negative impact on marching speed and movement capabilities).  But you may well be thinking that 40-55kg (which is 90-120lbs or so) sounds more than ample – that’s a lot of food!

Except of course they need to carry everything and weapons, armor and (for gunpowder armies) shot are heavy.  Roman soldiers were and are famous for having marched heavy, carrying as much of their equipment and supplies as possible in their packs, which the Romans called the sarcina (we’ll see why this could improve an army’s capabilities).  This practice is often attributed to Gaius Marius in the last decade of the second century (Plut. Marius 13.1) but care is necessary as this sort of ‘reform’ was a trope of Roman generalship and is used of even earlier generals than Marius (e.g. Plut. Mor. 201C on Scipio Aemilianus).  Various estimates for the marching load of Roman troops exist but the best is probably Marcus Junkelmann’s physical reconstruction (in Die Legionen des Augustus (1986); highly recommended if you can read German; alas for the lack of an English translation!) which recreated all of the Roman kit and measured a marching load of 54.8kg (120.8lbs), with ~43 of the 54.8kg reserved for weapons, armor, entrenching kit and personal equipment, leaving just 11.8kg for food (about ten days worth).  Other estimates are somewhat less, but never much less than 40kg for a Roman soldier’s equipment before rations, leaving precious little weight in which to fit a lot of food.

Via Wikipedia, Roman soldiers from the Column of Trajan (c. 113) marching with their sarcinae suspended from forked staves called furca.

The same exercise can be run for almost any kind of infantryman: while their load is often heavy, after one accounts for weapons, armor and equipment (and for later armies, powder and shot) there is typically little space left for rations, usually amounting to not more than a week or two (ten days is a normal rule of thumb).  Since the army obviously has more than two weeks of work to do (and remember it needs to be able to march back to wherever it started at the end), it is going to need to get a lot more food.  And this leads us to:

The Problem

The problem in a nutshell is that anything available to these armies prior to the advent of the railroad that can carry food, also eats food (except for boats, but rivers and coastlines may well not go where you want to go).  We may call this problem the ‘tyranny of the wagon equation’ as a number of readers have noticed the similarity to the tyranny of the rocket equation.

The first option is, of course, to just bring more people carrying food. A porter (that is a human non-combatant) can carry more supplies because they don’t have their own arms and armor, but is eating a bit more than a kilogram of it a day, just like your soldier, and also may have to carry a bunch of personal possessions (or in many armies, the superfluous personal possessions of the soldier).  A pack-mule can carry a fair bit more, around 130kg (figure via Landers, op. cit.).  But mules have to eat and they have to eat quite a bit, around 2.25kg of barley per day in addition to grass or hay, which again one hopes is largely available roadside (if it isn’t, you need another 4.5kg of hay or straw per day to feed the mule, in addition to the barley).  So in the best case, the mule eats its entire load in 57 days; worst case it does so in just 20 days. The good news with mules (which explains their use over wagons in many cases) is that they can handle moderately rough terrain, which wagons cannot. Porters and pack animals thus offer only a limited increase in operational range; the ratio of food carried to food eaten is just too low, not useless, but limited.

Wagons are more promising.  A big wagon pulled by two horses can carry perhaps a ton (1000kg) at maximum (in practice many medieval wagons capped out well below this and were pulled by four horses), but now we have two horses and a driver to consider.5   Now the small native ponies of the Steppe can subsist entirely off of grass, but the sort of horses available in the agrarian world are bred too big and strong to eat entirely grass. Their nutrition requirements are too high and so they require feed, at least some 4.5kg of it per day assuming local grass is available along with time to let the horses graze it (during which the wagon is, of course, stopped).  The Romans seem to have allocated around 7kg of barley per day per cavalryman for their cavalry, though its possible this also provided for a servant or groom for the cavalryman.  So under best conditions the two horses and one driver are going to eat about 10kg out of the wagon each day (a figure one may safely double if there is no grazing available), giving the wagon on its own at 100 days of range.

From the British Museum an etching (1591) of a pike formation on the move, with supplies and in tow, including wagons, pack mules as well as cows and sheep being moved ‘on the hoof.’ I am unsure if the presence of the elephant at right means we should understand this scene to take place in India and alas the museum description offers no clues.

Which sounds great except remember that the goal here isn’t to deliver wagons, but to deliver an army.  Factoring in the food demands of the soldiers, the number of wagons quickly spirals out of control.  Rather than reinvent the wheel here, I’ll note that K. Chase (op. cit.) ran these numbers assuming two-horse 1400lb wagons and found that assuming the army acquired no local food (but could get grass for the horses), for a group of thirty infantryman the first wagon doubles their range from 120 to 240 miles (less really, horses cannot be worked so many days consecutively).  Doubling again to around 400 (accounting for horse rest time) requires not two but six wagons for thirty men.  To double the range again would require more wagons than men.  And that assumes ample water and grazing; remove either and these figures collapse. So there are sharp limits to how much extra range an army can get by adding carrying capacity; in practice about a month and a half’s logistics buffer is the maximum that’s possible and even doing that is expensive.

From the British Museum, a print (1855) showing French cavalry alongside a wagon along with a vivandière with baskets of provisions.

Adding animals, especially draft animals pulling wagons, have other impacts too, particularly to speed and freedom to maneuver.  Pack mules may be taken off roads, but in order to keep up with marching infantry wagons need good roads.  Meanwhile wagons are prone to breakdowns and most importantly eat up a lot of road space which is also a key limiter to how fast an army can move.  The more wagons you add, the slower the army gets. We’ll run through this problem – moving the army – later in this series. Of course that isn’t to say the army won’t use wagons and pack mules – it absolutely will because it needs to carry lots of things, including spare food. Instead the point is the army cannot solely rely on the food it carries.

The tyranny of the wagon equation is thus inescapable: an army that plans to be in the field for more than just a few days cannot bring all of the food it needs with it; it must find most of its food locally. This is the significant of the famous but apocryphal quip (attributed to both Napoleon and Frederick the Great but probably said by neither) that ‘an army marches on its stomach.’  The solution to that problem is foraging, which we’ll come to a bit later, but before we go there we need to talk about the rest of the army.  And you may well be asking, ‘the rest of the army?  We have our animals, we have our soldiers, who is left?’

The Rest of the Army

It is worth keeping in mind that an army of 10,000 or 20,000 men was, by ancient or medieval standards, a mid-sized town or city moving across the landscape.  Just as towns and cities created demand for goods that shaped life around them, so did armies (although they’d have to stay put to create new patterns of agriculture, though armies that did stay put did create new patterns of agriculture, e.g. the Roman limes). Thousands of soldiers demand all sorts of services and often have the money to pay for them and that’s in addition to what the army as an army needs. That in turn is going to mean that the army is followed by a host of non-combatants, be they attached to the soldiers, looking to turn a profit, or compelled to be there.

We can start with sutlers, merchants buying or selling from the soldiers themselves (the Romans called these fellows lixae, but also called other non-soldiers in the camp lixae as well, see Roth (2012), 93-4; they also call them mercatores or negotiatores, merchants).  Sutlers could be dealing in a wide array of goods.  Even for armies where ration distribution was regular (e.g. the Roman army), sutlers might offer for sale tastier and fancier rations: meat, better alcohol and so on.  They might also sell clothing and other goods to soldiers, even military equipment: finding ‘custom’ weapons and armor in the archaeology of military forts and camps is not uncommon.  For less regularly rationed armies, sutlers might act as a supplement to irregular systems of food and pay, providing credit to soldiers who purchased rations to make up for logistics shortfalls, to collect when those soldiers were paid.  By way of example, the regulations of the Army of Flanders issued in 1596 allowed for three sutlers per 200-man company of troops (Parker, op. cit.), but the actual number was often much higher and of course those sutlers might also have their own assistants, porters, wagons and so on which moved with the army’s camp. Women who performed this role in the modern period are often referred to by the French vivandière.

Via Wikipedia, a painting of a vivandière with soldiers from the Napoleonic Wars by Adrian Moreau (1843-1906). By this point historically, the greater degree of logistics centralization had allowed European generals to mostly exclude women from the camp (back right), so this vivandière operates an impromptu eatery on the roadside.

For some armies there would have been an additional class of sutlers: slave dealers.  Enslaved captives were a major component of loot in ancient warfare and Mediterranean military operations into and through the Middle Ages.  Armies would abduct locals caught in hostile lands they moved through or enemies captured in battles or sieges; naturally generals did not want to have to manage these poor folks in the long term and so it was convenient if slave-dealer ‘wholesalers’ were present with the army to quickly buy the large numbers of enslaved persons the army might generate (and then handle their transport – which is to say traffic them – to market).  In Roman armies this was a regularized process, overseen by the quaestor (an elected treasury official who handled the army’s finances) assigned to each army, who conducted regular auctions in the camp.  That of course means that these slave dealers are not only following the army, but are doing so with the necessary apparatus to transport hundreds or even thousands of captives (guards, wagons, porters, etc.).

And then there is the general category of ‘camp follower,’ which covers a wide range of individuals (mostly women) who might move with the camp.  The same 1596 regulations that provided for just three sutlers per 200-man Spanish company also provided that there could be three femmes publiques (prostitutes), another ‘maximum’ which must often have been exceeded.  But prostitutes were not the only women who might be with an army as it moved; indeed the very same regulations specify that, for propriety’s sake, the femmes publiques would have to work under the ‘disguise of being washerwomen or something similar’ which of course implies a population of actual washerwomen and such who also moved with the army. Depending on training and social norms, soldiers may or may not have been expected to mend their own clothes or cook their own food.  Soldiers might also have wives or girlfriends with them (who might in turn have those soldier’s children with them); this was more common with professional long-service armies where the army was home, but must have happened with all armies to one degree or another. Roman soldiers in the imperial period were formally, legally forbidden from marrying, but the evidence for ‘soldier’s families’ in the permanent forts and camps of the Roman Empire is overwhelming.

Via Wikipedia, the Landsknecht with his Wife, etching by Daniel Hopfer (c. 1470-1536) showing a professional infantryman of the period, a Landsknecht, on campaign with his wife. Women are often omitted from military artwork of this period but we know from reports from the period that they would have been extremely common.

The tasks women attached to these armies have have performed varied by gender norms and the organization of the logistics system. Early modern gunpowder armies represent some of the broadest range of activities and some of the armies that most relied on women in the camp to do the essential work of maintaining the camp; John Lynn (op. cit., 118-163) refers to the soldiers and their women (a mix of wives, girlfriends and unattached women) collectively as ‘the campaign community’ and it is an apt label when thinking about the army on the march.6 As Lynn documents, women in the camp washed and mended clothes, nursed the sick and cooked meals, all tasks that were considered at the time inappropriate for men. Those same women might also be engaged in small crafts or in small-scale trade (that is, they might also be sutlers). Finally, as Lynn notes, women who were managing food and clothing seem often to have become logistics managers for their soldiers, guarding moveable property during battles and participating in pillaging in order to scrounge enough food and loot for they and their men to survive. I want to stress that for armies that had large numbers of women in the camp, it was because they were essential to the continued function of the army.

From the British Museum, a satirical print (c. 1828) showing a vivandière working in a French military camp, tending to a wounded soldier while also offering another soldier a glass of wine.

And finally, you have the general category of ‘servants.’  The range of individuals captured by this label is vast.  Officers and high status figures often brought either their hired servants or enslaved workers with them.  Captains in the aforementioned Army of Flanders seem generally to have had at least four of five servants (called mozos) with them, for instance; higher officers more.  But it wasn’t just officers who did this.  Indeed, the average company in the Army of Flanders, Parker notes, would have had 20-30 individual soldiers who also had mozos with them; one force of 5,300 Spanish veterans leaving Flanders brought 2,000 such mozos as they left (Parker, op. cit. 151).

Looking at the ancient world, many – possibly most – Greek hoplites in citizen armies seem to have very often brought enslaved servants with them to carry their arms and armor; such enslaved servants are a regular feature of their armies in the sources.  The Romans called these enslaved servants in their armies calones; it was a common trope of good generalship to sharply restrict their number, often with limited success.  At Arausio we are told there were half as many servants (calonum et lixarum) as soldiers (Liv. Per. 67, on this note Roth (2013), 105), though excessive numbers of calones et lixae was a standard marker of bad general and the Romans did lose badly at Arausio so we ought to take those figures with a grain of salt, as Livy (and his sources) may just be communicating that the generals there were bad. That said, the notion that a very badly led army might have as many non-combatants following it as soldiers is a common one in the ancient sources. And while Roman armies were considered notable in the ancient world for how few camp servants they relied on and thus how much labor and portage was instead done by the soldiers, getting Roman aristocrats to leave their vast enslaved household staff at home was notoriously difficult (e.g. Ps.Caes. BAfr. 54; Dio Cass. 50.11.6). Much like the early modern ‘campaign community,’ our sources frequently treat these calones as part of the army they belonged to, even though they were not soldiers.

Kick Them Out!

All of this contributes to the ‘tooth to tail’ ratio of the army on the move, which is to say the ratio of effective combat troops to the folks who are just supporting the army.  Now I should note (and this will come up a bit later), many of these non-soldiers in the camp might be armed (but most certainly are not always or even generally so), particularly camp servants and camp wives.  They might thus participate in various ways both in the defense of the camp and in foraging operations.  But they do not contribute to the offensive power of the army, so they are squarely in the ‘tail’ of the army.

Given the tyranny of the wagon equation it is easy to see the problem.  Each camp servant, sutler, family member and so on that the army adds is another mouth that had to be fed somehow through the army’s logistics system (even if they are, for instance, not fed directly by the army but rather through soldier’s pay or rations).  And, as we’ll get to, the larger the army train is (including these folks) the slower it moves, so non-combatants not only increase the logistics demand on the army but also slow it down.  In essence an army overburdened with non-combatants experiences all of the problems of a big army without any of the combat power benefits of a big army.

And as a reminder, big armies have a lot of problems!  They’re harder to coordinate, they move much slower, they are harder to secure on the march (because the line of march is longer), they’re more expensive and eat more food.  And because they eat more food, as we’ll see as we move forward, there are places they cannot go: routes available to a small, lean army which cannot support a large one.

Naturally then, commanders try to restrict the number of non-combatants with the army.  The basic motions of doing this occur so frequently in the Roman sources they are almost rote: generals demand soldiers carry all of their equipment, restrict rations to essentials, limit carried goods and equipment down to the bare minimum (often with the standard ‘good general’ first ostentatiously disposing of his own bits of private property) and then banish women, slaves and most of the merchants from the camp, allowing only those who remain to conduct business directly with the army command (in Roman sources, through the quaestor alone).7  Now we should be cautious with these reports: these are literary tropes meant to signify a broad sense of ‘good generalship.’  At the same time, actually doing this would have had an effect and it seems safe to suppose many Roman generals did this, though of course the fact that Roman generals have to keep doing this suggests the intransigence of the problem.

Why don’t all generals do this?  In practice the reasons fit under two big headings.  The first big heading is that, well, the soldiers don’t want the general to do that, so doing so damages morale.  They like having all of these folks around (especially, one assumes, their families for professionals who have no other home to go back to).  And that is not only true of the regular soldiers but also of the officers, whose rank often entitles them to bring more attendants of various kinds than regular soldiers; they are jealous of that privilege of rank.  So there is a tug-of-war between the interest of the general in having a lean, effective army and the interests of everyone else in creature comforts on campaign.  The second big heading is logistics: if these non-combatants are providing essential services to the army (like markets for food or replacement clothes or cooking) and you kick them out, the army ceases to function and the general will not look very clever for his fastidiousness.

Consequently the ability of generals to tamp down on this sort of thing tends to come down to a mix of their own personal willingness to indulge but also the leverage their soldiers and officers have.  In early modern gunpowder armies (like the Army of Flanders, for instance), the soldiers are at best weakly attached to the state; they are mercenaries with highly valuable skills who could as easily fight for someone else or just mutiny.  They are also long-service professionals who, if they have families, are bringing those with them.  Moreover, and we’ll come back to this, many of those camp attendants are filling vital logistical roles that the army’s primitive logistics apparatus is simply incapable of doing.  In that context, generals have little leverage to force those creature comforts out of the camp and so regulations on the permissible number of attendants, prostitutes, merchants and so on were just ignored.

Over the course of the late 1600s and 1700s, European states and monarchs worked to fix these problem, steadily centralizing supply, pay, command and logistics into national bureaucracies. That in turn allowed for more and more non-combatants, especially the women of the campaign community, to be pushed out of the camp and excluded from the army as their logistical functions were subsumed by these new bureaucracies (there is a good basic introduction to these processes in Lee, Waging War, 301-315). This in turn leads to the leaner field armies of the 1800s, although they are by no means devoid of women following the army doing essential services.

Via the Library of Congress, a photograph from 1862 showing a woman along with a soldier (likely her husband) and three children in the camp of the 31st Pennsylvania Infantry Regiment, encamped outside of Washington, D.C..

Or to take another negative example, a polis hoplite army consisted of most – if not essentially all – of the politically influential, voting citizens.  The (often elected) general’s ability to enforce any kind of uncomfortable discipline on such armies was extremely limited because the entire voting body was there; what the hoplites wanted (collectively), they largely got.  Consequently, hoplite armies seem generally to have had lots of non-combat attendants, despite generally operating at very limited range and for only limited parts of the year.

On the other hand, you have Roman armies: the soldiers are citizens, but represent only a small slice of the total voting body.  Their generals as either consuls (and thus members of the Senate) or legates of the emperor (also members of the Senate) are effectively insulated from a fair bit of that political pushback, while Roman law and custom give generals wide latitude to discipline their soldiers (military service was, for instance, the only context in which a citizen could lawfully be beaten by a magistrate or his subordinates).  Meanwhile as we’ll see, the Roman logistics system is well developed, with the legion able to handle most of the required tasks internally, making it feasible from both a logistics and a political standpoint to simply kick all of the non-combatants out.  And so Roman generals do, at least some of the time – though doing so could cause a crisis of morale.  Remember, the soldiers like having these folks around.


So now we have our entire ‘campaign community’ of men, women and animals. And so it might be worth doing some quick calculations to close out this section to get a sense now of exactly what a community of this size is going to require. For a general sense of scale, we’ll consider the demands of a standard Roman army of the Middle Republic: two legions plus matching allied detachments, totaling around 19,200 soldiers (16,800 infantry, 2,400 cavalry).8

Let’s deal with animals next. Each contubernium (‘tent group’) of 6 soldiers likely had its own mule (for reasons we’ll get to later), so that’s 3,200 mules for the army, plus some additional number for the siege train and any army supplies; perhaps around 5,000 total (see Roth, op. cit. on this). On top of this we have horses for the cavalry; this will be rather more than 2,400 since spare horses will have been a necessity on campaign. Judging by Roman barley rations for cavalrymen (presumably intended to feed the horse) it seems a good guess that each cavalryman had one spare; for later medieval armies the number of spares would be substantially higher (at least three per rider). But for our lean army of Romans, that’s just 4,800 horses. An early modern army might require quite a few less mules (replacing them with wagons), but at the same time it is also probably hauling both field artillery and siege guns which demand a tremendous number of draft animals (mostly horses). My sense is that in the end this tends to leave the early modern army needing more animals overall.

Next the non-combatants. The mules will need drivers9 and the cavalrymen likely also have grooms to handle their horses, which suggests something like 3,400 calones as an absolute minimum simply to handle the animals. Roth (op. cit., 114) figures 1 non-combatant per four combatants in a Roman army, while Erdkamp (op. cit. 42) figures 1:5. Those figures would include not merely enslaved calones but also sutlers, slave-dealers, and women in the campaign community. Taking the lower estimate we might then figure something like 4,000 non-combatants for a ‘lean’ Roman army, with many armies being more loaded up on non-combatants than even this. And while estimating the number of non-combatants for Roman armies is tricky, we actually have some figures for pre-modern armies to give a reference. Parker (op. cit. 252) notes units of the Army of Flanders (between 1577 and 1620) as high as 53% non-combatants, including women in the campaign community; one Walloon tercio in 1629 was 28% camp women on the march. It is tempting to compare these but caution is necessary here – both Roth’s and Erdkamp’s estimates are heavily informed by more modern armies so the argument would be circular: the estimates for the Romans look like later armies because later armies were used to calibrate estimates for the Romans.

That gives us an army now of 19,200 soldiers, 4,000 non-combatants, 5,000 mules and 4,800 horses. Roman rations were pretty ample and it seems likely that many of the calones did not eat so well but the ranges are fairly narrow; we can work with an average 1.25kg daily ration per person normally, with the absolute minimum being the 0.83kg daily grain ration following Polybius (Plb. 6.39.12-14, on this note Erdkamp op. cit. 33-42) if the army was short on supplies or needed to move fast eating only those buccelatum biscuits. That’s a normal consumption of 29,000kg per day for the humans, with the minimum restricted diet of 19,256kg for short periods. Then we need about 2.25kg of feed for each mule and about 4.5kg of feed for each horse (we’re assuming grazing and water are easily available), which adds up to 11,250kg for the mules and 21,600kg for the horses.

And at last we now have the scale of our problem: our lean army of 19,200 fighting men consumes an astounding 61,850kg (68.18 US tons) of food daily. It also consumes staggering amounts of water and firewood (we’ll talk about foraging these later). In order to move this army or sustain it in place it is thus necessary to ensure a massive and relatively continuous supply of food to the army. Failure to do that will result in the army falling apart long before it comes anywhere close to the enemy.

And that’s where we’re going to go next week: how does the army get the food it needs and then distribute it to the soldiers, non-combatants and animals.

  1. At the same time, armies tend to be expanding in Europe in this period, which also strains logistics.
  2. Calorie values for ground alphita are tricky, for reasons discussed in Foxhall and Forbes, “Sitometria” Chiron 12 (1982), but Spartan rations are certainly on the high calorie side, leading in some cases to suggestions that the Spartiates rations may have also covered enslaved helot servants. Alternately, it may just be a product of the Spartiates all being elites, who ate like elites and not like common soldiers. It is also possible, as Foxhall and Forbes note, that ancient alphita may have ended up with a lower proportion of the edible parts of the barley seed, leading to a lower calorie value per unit-weight than what they estimate based on modern computations using efficient, mechanical grinders.
  3. To be clear, we know with some certainty that Roman rations were supplemented, but not by how much. If you read much older scholarship, you will find the notion that Roman soldier’s diet lacked regular meat; both Erdkamp and Roth reject this view decisively and for good reason.
  4. I may return to the logistics of water later, but some range can be extended here by taking advantage of the fact that pack animals, while they need a lot of water per day over a long period, can be marched short periods with basically no water and still function, whereas water deprived humans die very quickly. Consequently an army can do a low-water ‘lunge’ over short distances by loading its pack animals with water, not watering them, having the soldiers drink the water and then abandoning the pack animals as they die (the water they carried having been consumed). This is, to say it least, a very expensive thing to do – animals are not cheap! – but there is some evidence the Romans did this, on this see G. Moss, “Watering the Roman Legion” M.A. Thesis, UNC Chapel Hill (2015).
  5. For the Romans, wagons of these sort were unavailable. Roman carts were typically two-wheelers (lacking an independently turning front axle) which limited the weight they could carry by quite a lot. Consequently, the Romans only used carts for moving the siege train and all other logistics was done with pack animals, almost always mules.
  6. And of course many of those women had children by their soldiers, who would also be moving with the camp.
  7. Moralizing examples in the sources of generals either doing this or failing to do this are numerous. E.g. for the Romans Plut. Mor. 201C; Plut. Mar. 13.1; App. Hisp. 14.85; Sall. Iug. 44.5-45.2; Tac. Hist. 2.88; Hdn. 4.7.4-6.
  8. That’s two legions of Roman citizens, each with 4,200 infantry (1,200 velites, hastati and principes each, plus 600 triarii) and 300 equites, plus two alae of socii (‘wings of allies’), each with roughly the same number (4,200) of infantry but three times the cavalry (so 900 each), Plb. 6.21.7-9, 6.25-26.
  9. One muleteer can manage around five mules though.

311 thoughts on “Collections: Logistics, How Did They Do It, Part I: The Problem

  1. Aside: Camp followers are also a rarity in fantasy, but Turtledove treated them as serious business (and lampshaded the difference between the Roman Early Empire/Late Republic and Late Roman/Byzantine attitudes towards camp followers in his Videssos Cycle)..

  2. The wagon-rocket analogy is better than the math here shows. Much like a rocket takes less thrust to accelerate the same amount as it gets lighter, these armies could cut loose teamsters (and their animals!) as the wagons depleted.

    There would be no reason to eat from all the wagons at the same rate. If a Roman army is eating 19,000 kg per day and a wagon carries 1000 kg then each day there’s 19 fewer wagons eating.

    This would work to the extent that the wagons aren’t expected to be refilled enroute, and the math would help extend the range of the army a whole lot.

    1. I see two problems with doing that, which may or may not apply in all situations.

      Unless things go absolutely disastrously you’re probably expecting to bring at least seventy percent of your army back the same way even if you lose, and if you end up withdrawing after a few indecisive skirmishes or secure a surrender you’ll need to bring the whole army back. There’s good odds you’re planning to at least partially resupply, either by raiding or because you’re coming to the aid of a friendly settlement, so you’ll want empty wagons for that.

      Second, whether or not you need all your wagons for this trip, you probably don’t want them destroyed and their teamsters slaughtered. The enemy probably can’t move their entire main army around behind you because they’ve got the same logistics to deal with as you do, but fifty light cavalry have a much smaller logistics footprint, so if you’re regularly sending back a bunch of wagons the enemy is probably going to detach some guys to attack and burn them. So you’d need to detach a reasonably large escort for each set of wagons, and then you’re weakening your main army.

      If you’ve got safe places along the route to leave them, odds are you could resupply at them instead and not bring as many wagons in the first place. It might have its own agricultural production, or you might be able to send food ahead while the army is assembling, especially if you’ve only got your vassals for forty days out of the year but usable roads for six months.

  3. Something to consider regarding sutlers is how efficient is it to put strain on a “parallel” logistics train by carrying currency and buying food and supplies from them rather than carrying those supplies. It seems like that might have been a factor in the Flandes war.

    1. This runs into a few problems.

      1) The parallel logistics train still has to physically move supplies the same way the regular logistics train does; if your army is carrying all its supplies on the backs of pack mules or enslaved porters or in wagons, then so are your sutlers. If your army cannot march more than ten miles a day because it is too encumbered by the sheer number of wagons required to feed it, then outsourcing half your supply needs to a “parallel” system of very similar wagon trains run by independent contractors does not make your problem easier. It may even make your problem worse, because the contractors are not as easy to keep under strict discipline.

      2) Soldiers have weapons and, in the normal course of their activities, occasionally use these weapons to deadly effect. Selling food to hungry soldiers can be a risky undertaking, because as a brutally practical matter, it is hard to be sure they won’t simply seize the food at swordpoint and send you on your way with no money. To ensure that this doesn’t happen (which would collapse your parallel supply train) you need very tight discipline and to make sure the soldiers don’t ever truly need to rob their own sutlers, which is a lot easier if you can ensure that their basic subsistence needs are met without depending on the sutlers. Furthermore…

      3) Paying armies regularly is more of a problem than you’d think in most pre-modern societies. Kings routinely underestimated expenses, had to deal with subordinates embezzling some of the funds, or otherwise couldn’t or didn’t maintain the cash flow, even though naively you’d think “failure to pay thousands of heavily armed men” would be the kind of mistake no one would be willing to risk making. The thing is, if your soldiers haven’t been paid in a month but the food keeps coming in, they can sit around waiting for back pay to show up… because they don’t need the money in order to eat. If you’re relying on a market economy to supply them, then within a very short time, interruptions of pay become an interruption of food… At which point you’re back to point (2), the soldiers break discipline and rob everyone in the vicinity of everything that isn’t nailed down, then pry up the nails and take the rest, putting your sutlers out of business.

    1. Yeah, I was wondering about how often hardtack-equivalents would be fed to the pack animals and especially ridden animals ?

      (And in postmodern times, with bioethanol fuels ! Though that one must be out of desperation…)

      1. I’m glad you liked it! My fiancée and I have been giggling about “harmonic donkey transfer” and the “minimum donkey rate” for the last two days. I look forward to extending the donkey equation to relativistic donkeys.

  4. I think you greatly over estimate camp followers, local and while they provide categorically “essential work” such as laundry, cooking and selling of equipment that does not mean it is essential to the running of the army. One, camped by rivers or bodies of water (as armies often are) can easily do their own sundry or at least part of it. But how often has it been seen since the soldiers had a good clean wash? Furthermore same could be said with cooking with hot meals. Or maybe the blanket one brought is not good enough for the north. This is the camp followers function. To provide extra services to the army, but these are not essential to the armies function. Furthermore, these camp followers (disregarding family) are an enterprising bunch on an enterprising endeavors, they will not follow the army from home to foreign land and back (yes, you have have a source that states there is notable exception) but most would be locals who supply the army as that gives the highest profit margin. Breaking from history for the next point, I think one can can find parallels (though weak) in modern tourism sites and along population hiking trails (especially in poorer and remote areas) that one could argue is similar to how a local population would support a standing arm in yesteryear.

  5. After a little cross-referencing, it looks like the etching numbered 1591 which you’re using above is part of a collection of etchings “all of scenes representing the Habsburg campaign against the Turks and Hungarians in 1591, including battle formations, foot-soldiers and cavalry, munition and supply waggons”.

  6. Great read, but…p

    How does the wagon equation change if you replace horse-drawn wagons by ox-drawn carts? Wagons could be pulled by horses only rather late (China maybe 400 BC, first arrival of the collared harness in Europe maybe 1000 AD, use in Central and Western Europe since the 16th century).

    The Romans had the Chianina, a huge cattle breed for pulling ploughs and carts “inherited” from the Etruscans, rivalling the aurochs in size and possibly greater in power, which was available for heavy haulage.

    1. Oxen are less demanding in not needing much in the way of supplementary grain, but slower. They would be good for routinely moving supplies and equipment to a forward outpost or depot but would seriously hamper the speed of a marching army. This is not only bad for military reasons but because soldiers need to eat every day regardless of how much progress they’re making or not.

    1. 54.8kg is heavy, but it’s not impossible for a fit man to march with it. Infantry in Afghanistan were patrolling with an average weight of 50kg. Patrolling, not marching; so you have to cross ditches, climb walls, run, crawl, fight… that was too much, and it harmed combat effectiveness immensely, but it was still physically possible.

      Remember that the Romans wouldn’t be going very far each day – probably 8-10 miles along a road. A fit man can cover that distance in a day with a 55kg pack over rough ground and hills.

      And, as the food gets eaten, you free up some space on the carts or mules, allowing you to put your kit on it instead (assuming it isn’t taken up with the sick and injured instead).

  7. I know I’m months behind, but “what armies might do if they lacked those capabilities (lost to the Romans, mostly)” is the level of historical snark that I come here for.

  8. Been a horsewoman for 40 mumble years. Modern horses require 3 percent of their desired bodyweight in some kind of food per day on average. So, let’s make the math easy on me, the horse is 1000 pounds. That’s a pretty good size horse, my Arabian horses tend to clock out at 800 pounds. My breeding foundation (read: heavy built) Quarter Horse mare was 1100 at peak show condition. So, for a 1000 pound horse, that comes out to 30 pounds of feed from SOMEWHERE. For horses that aren’t working, all of that today can come from grass and high quality hay, but that’s only if they’re totally chilling in a pasture. That’s not doing any kind of work (my horses are lazy bums right now because it’s too hot to do anything). When putting that horse into work, the amount of grain to be fed varies up to 30% of the total, depending on how hard the work is. So, if I was riding that 1000 pound horse to do heavy ranch work 10 hours a day 6 days a week, that comes out to 9 pounds (4.09 kg) of grain and 21 lbs (9.55 kg) of grass or hay. Breeding and lactation changes that equation yet again. When my Quarter Horse mare was nursing her foal, it took 21 lbs (9.55kg) of hay AND 18 lbs (8.1 kg) of grain split into four feedings throughout the day just to keep weight on her with no work burden at all. She just poured it all into her foal for four months, and I had to keep it up for another month to get her back up to pre-pregnancy condition.

    1. Some good figures, note that historical horses range significant in size and weight and many are not as heavy as modern horses. Steppe ponies, in particular, are substantially smaller and thus better able to subsist off of a purely grass-based diet.

  9. I suppose you know and use “Supplying War: Logistics From Wallenstein To Patton” by Martin Van Creveld. Masterful by a really pardigm changing author.

  10. Hannibal’s invasion of Italy at the outset of the Second Punic War might be worth its own post on logistics. It was considered logistically so difficult, that acquiring and eating slaves on the march was considered. In the end, they settled for cattle and probably had several voluntary and involuntary resupplies along the way, resetting, and making the curve less exponential to reach Roman territory.

Leave a Reply