On War - title image

Carl von Clausewitz

[Table of Contents]

 • Mobile Compatible • 

NOTE: This version of Carl von Clausewitz's On War is the long-obsolete J.J. Graham translation of Clausewitz's Vom Kriege (1832) published in London in 1873. The 1976/84 Howard/Paret version is the  standard translation today; for the most accurate text one should always consult the 1943 Jolles translation. Consider the more modern versions and other relevant books shown below.

Book Cover, ON WATERLOOOn Waterloo: Clausewitz, Wellington, and the Campaign of 1815. Ed./trans. Christopher Bassford, Daniel Moran, and Gregory W. Pedlow (, 2010). ISBN: 1453701508. This book is built around a new and complete translation of Clausewitz's study of the Waterloo campaign [Berlin: 1835], which is a strategic analysis of the entire campaign (not just the Battle of Waterloo), and the Duke of Wellington's detailed 1842 response to it.

Jolles translation, book coverBuy the best translation—recommended for serious readers. The Book of War (The Modern Library, February 2000). ISBN: 0375754776. Clausewitz's On War and Sun Tzu's Art of War in one volume. The translation of Clausewitz's On War is the 1943 version done by German literary scholar O.J. Matthijs Jolles at the University of Chicago during World War II—not today's standard translation, but certainly the most accurate.

On War, Princeton ed.Buy the standard English translation of Clausewitz's On War, by Michael Howard and Peter Paret  (Princeton University Press, 1976/84). ISBN: 0691018545 (paperback). Kindle edition. This quite readable translation appeared at the close of the Vietnam War and—principally for marketing and copyright reasons—has become the modern standard.

Book coverVanya Eftimova Bellinger, Marie von Clausewitz: The Woman Behind the Making of On War (Oxford University Press, 2015), ISBN: 0190225432. A rich biography of Countess Marie von Clausewitz that also sheds enormous light on the life, ideas, influences upon, and character of the great military thinker himself.




IN the modern system of war cantonments have become again indispensable, because neither tents nor a complete military train make an army independent of them. Huts and open-air camps (bivouacs as they are called), however far such arrangements may be carried, can still never become the usual way of locating troops without sickness gaining the upper hand, and prematurely exhausting their strength, sooner or later, according to the state of the weather or climate. The campaign in Russia in 1812 is one of the few in which, in a very severe climate, the troops, during the six months that it lasted hardly ever lay in cantonments. But what was the consequence of this extreme effort, which should be called an extravagance, if that term was not much more applicable to the political conception of the enterprise!

Two things interfere with the occupation of cantonments—the proximity of the enemy, and the rapidity of movement. For these reasons they are quitted as soon as the decision approaches, and cannot be again taken up until the decision is over.

In modern wars, that's, in all campaigns during the last twenty-five years which occur to us at this moment, the military element has acted with full energy. Nearly all that was possible has generally been done in them, as far as regards activity and the utmost effort of force; but all these campaigns have been of short duration, they have seldom exceeded half a year; in most of them a few months sufficed to bring matters to a crisis, that is, to a point where the vanquished enemy saw himself compelled to sue for an armistice or at once for peace, or to a point where, on the conqueror's part, the impetus of victory had exhausted itself. During this period of extreme effort there could be little question of cantonments, for even in the victorious march of the pursuer, if there was no longer any danger, the rapidity of movement made that kind of relief impossible.

But when from any cause the course of events is less impetuous, when a more even oscillation and balancing of forces takes place, then the housing of troops must again become a foremost subject for attention. This want has some influence even on the conduct of war itself, partly in this way, that we seek to gain more time and security by a stronger system of outposts, by a more considerable advanced guard thrown further forward; and partly in this way, that our measures are governed more by the richness and fertility of the country than by the tactical advantages which the ground affords in the geometrical relations of lines and points. A commercial town of twenty or thirty thousand inhabitants, a road thickly studded with large villages or flourishing towns give such facilities for the assembling in one position large bodies of troops, and this concentration gives such a freedom and such a latitude for movement as fully compensate for the advantages which the better situation of some point may otherwise present.

On the form to be followed in arranging cantonments we have only a few observations to make, as this subject belongs for the most part to tactics.

The housing of troops comes under two heads, inasmuch as it can either be the main point or only a secondary consideration. If the disposition of the troops in the course of a campaign is regulated by grounds purely tactical and strategical, and if, as is done more especially with cavalry, they are directed for their comfort to occupy the quarters available in the vicinity of the point of concentration of the army, then the quarters are subordinate considerations and substitutes for camps; they must, therefore, be chosen within such a radius that the troops can reach the point of assembly in good time. But if an army takes up quarters to rest and refresh, then the housing of the troops is the main point, and other measures, consequently also the selection of the particular point of assembly, will be influenced by that object.

The first question for examination here is as to the general form of the cantonments as a whole. The usual form is that of a very long oval, a mere widening as it were of the tactical order of battle. The point of assembly for the army is in front, the head-quarters in rear. Now these three arrangements are, in point of fact, adverse, indeed almost opposed, to the safe assembly of the army on the approach of the enemy.

The more the cantonments form a square, or rather a circle, the quicker the troops can concentrate at one point, that is the centre. The further the place of assembly is placed in rear, the longer the enemy will be in reaching it, and, therefore, the more time is left us to assemble. A point of assembly in rear of the cantonments can never be in danger. And, on the other hand, the farther the head-quarters are in advance, so much the sooner reports arrive, therefore so much the better is the commander informed of everything. At the same time, the first named arrangements are not devoid of points which deserve some attention.

By the extension of cantonments in width, we have in view the protection of the country which would otherwise be laid under contributions by the enemy. But this motive is neither thoroughly sound, nor is it very important. It is only sound as far as regards the country on the extremity of the wings, but does not apply at all to intermediate spaces existing between separate divisions of the army, if the quarters of those divisions are drawn closer round their point of assembly, for no enemy will then venture into those intervals of space. And it is not very important, because there are simpler means of shielding the districts in our vicinity from the enemy's requisitions than scattering the army itself.

The placing of the point of assembly in front is with a view to covering the quarters, for the following reasons:—In the first place, a body of troops, suddenly called to arms, always leaves behind it in cantonments a tail of stragglers—sick, baggage, provisions, etc., etc.—which may easily fall into the enemy's hands if the point of assembly is placed in rear. In the second place, we have to apprehend that if the enemy with some bodies of cavalry passes by the advanced guard, or if it is defeated in any way, he may fall upon scattered regiments or battalions. If he encounters a force drawn up in good order, although it is weak, and in the end must be overpowered, still he is brought to a stop, and in that way time is gained.

As respects the position of the head-quarters, it is generally supposed that it cannot be made too secure.

According to these different considerations, we may conclude that the best arrangement for districts of cantonments is where they take an oblong form, approaching the square or circle, have the point of assembly in the centre, and the head-quarters placed on the front line, well protected by considerable masses of troops.

What we have said as to covering of the wings in treating of the disposition of the army in general, applies here also; therefore corps detached from the main body, right and left, although intended to fight in conjunction with the rest, will have particular points of assembly of their own in the same line with the main body.

Now, if we reflect that the nature of a country, on the one hand, by favourable features in the ground determines the most natural point of assembly, and on the other hand, by the positions of towns and villages determines the most suitable situation for cantonments, then we must perceive how very rarely any geometrical form can be decisive in our present subject. But yet it was necessary to direct attention to it, because, like all general laws, it affects the generality of cases in a greater or less degree.

What now remains to be said as to an advantageous position for cantonments is that they should be taken up behind some natural obstacle of ground affording cover, whilst the sides next the enemy can be watched by small but numerous detached parties; or they may be taken up behind fortresses, which, when circumstances prevent any estimate being formed of the strength of their garrisons, impose upon the enemy a greater feeling of respect and and caution.

We reserve the subject of winter quarters, covered by defensive works for a separate article.

The quarters taken up by troops on a march differ from those called standing cantonments in this way, that, in order to save the troops from unnecessary marching, cantonments on a march are taken up as much as possible along the lines of march, and are not at any considerable distance on either side of these roads; if their extension in this sense does not exceed a short day's march, the arrangement is not one at all unfavourable to the quick concentration of the army.

In all cases in presence of the enemy, according to the technical phrase in use, that is in all cases where there is no considerable interval between the advance guards of the two armies respectively, the extent of the cantonments and the time required to assemble the army determine the strength and position of the advanced guard and outposts; but when these must be suited to the enemy and circumstances, then, on the contrary, the extent of the cantonments must depend on the time which we can count upon by the resistance of the advance guard.

In the third*1 chapter of this book, we have stated how this resistance, in the case of an advanced corps, may be estimated. From the time of that resistance we must deduct the time required for transmission of reports and getting the men under arms, and the remainder only is the time available for assembling at the point of concentration.

We shall conclude here also by establishing our ideas in the form of a result, such as is usual under ordinary circumstances. If the distance at which the advanced guard is detached is the same as the radius of the cantonments, and the point of assembly is fixed in the centre of the cantonments, the time which is gained by checking the enemy's advance would be available for the transmission of intelligence and getting under arms, and would in most cases be sufficient, even although the communication is not made by means of signals, cannon-shots, etc., but simply by relays of orderlies, the only really sure method.

With an advanced guard pushed forward three miles in front, our cantonments might therefore cover a space of thirty square miles. In a moderately-peopled country there would be 10,000 houses in this space, which for an army of 50,000, after deducting the advanced guard, would be four men to a billet, therefore very comfortable quarters; and for an army of twice the strength nine men to a billet, therefore still not very close quarters. On the other hand, if the advanced guard is only one mile in front, we could only occupy a space of four square miles; for although the time gained does not diminish exactly in proportion as the distance of the advanced guard diminishes, and even with a distance of one mile we may still calculate on a gain of six hours, yet the necessity for caution increases when the enemy is so close. But in such a space an army of 50,000 men could only find partial accommodation, even in a very thickly populated country.

From all this we see what an important part is played here by great or at least considerable towns, which afford convenience for sheltering 10,000 or even 20,000 men almost at one point.

From this result it follows that, if we are not very close to the enemy, and have a suitable advanced guard we might remain in cantonments, even if the enemy is concentrated, as Frederick the Great did at Breslau in the beginning of the year 1762, and Buonaparte at Witebsk in 1812. But although by preserving a right distance and by suitable arrangements we have no reason to fear not being able to assemble in time, even opposite an enemy who is concentrated, yet we must not forget that an army engaged in assembling itself in all haste can do nothing else in that time; that it is therefore, for a time at least, not in a condition to avail itself in an instant of fortuitous opportunities, which deprives it of the greater part of its really efficient power. The consequence of this is, that an army should only break itself up completely in cantonments under some one or other of the three following cases:

1. If the enemy does the same.

2. If the condition of the troops makes it unavoidable.

3. If the more immediate object with the army is completely limited to the maintenance of a strong position, and therefore the only point of importance is concentrating the troops at that point in good time.

The campaign of 1815 gives a very remarkable example of the assembly of an army from cantonments. General Ziethen, with Blucher's advanced guard, 30,000 men, was posted at Charleroi, only two miles from Sombreff, the place appointed for the assembly of the army. The farthest cantonments of the army were about eight miles from Sombreff, that is, on the one side beyond Ciney, and on the other near Liége. Notwithstanding this, the troops cantoned about Ciney were assembled at Ligny several hours before the battle began, and those near Liége (Bulow's Corps) would have been also, had it not been for accident and faulty arrangements in the communication of orders and intelligence.

Unquestionably, proper care for the security of the Prussian army was not taken; but in explanation we must say that the arrangements were made at a time when the French army was still dispersed over widely extended cantonments, and that the real fault consisted in not altering them the moment the first news was received that the enemy's troops were in movement, and that Buonaparte had joined the army.

Still it remains noteworthy that the Prussian army was able in any way to concentrate at Sombreff before the attack of the enemy. Certainly, on the night of the 14th, that is, twelve hours before Ziethen was actually attacked, Blucher received information of the advance of the enemy, and began to assemble his army; but on the 15th at nine in the morning, Ziethen was already hotly engaged, and it was not until the same moment that General Thielman at Ciney first received orders to march to Namur. He had therefore then to assemble his divisions, and to march six and a half miles to Sombreff, which he did in 24 hours. General Bulow would also have been able to arrive about the same time, if the order had reached him as it should have done.

But Buonaparte did not resolve to make his attack on Ligny until two in the afternoon of the 16th. The apprehension of having Wellington on the one side of him, and Blucher on the other, in other words, the disproportion in the relative forces, contributed to this slowness; still we see how the most resolute commander may be detained by the cautious feeling of the way which is always unavoidable in cases which are to a certain degree complicated.

Some of the considerations here raised are plainly more tactical than strategic in their nature; but we have preferred rather to encroach a little than to run the risk of not being sufficiently explicit.

*1. Actually, this refers to Chapter 8.

Return to the top of this page

Go to next chapter

Go to Table of Contents

Visit logo Visit the
Clausewitz Bookstores!
US UK France Germany