LAND es? WATER 
September 5, 1918 
after the very lieat'y casualty lists of the great attacks between 
March and July, the enemy was able to keep up his full estab- 
lishment by drafts from depots, he has not been able to do so 
since the counter-offensive began upon July iSth. That is 
certain. We know, in point of fact, that he has recently 
reduced the number of divisions by eight. That is, he has 
had to break up eight divisions in order to form drafts for 
the remainder. And although he has summoned two Austrian 
divisions to his aid, he now has only 198 divisions upon the 
West. But these divisions cannot be at full strength. He 
has lost in prisoners alone, since the surprise at -goissons, 
at least 130,000 men. In the normal proportion this would 
mean that his total casualties in the last forty-five days had 
come to the,.tremendous total of over 600,000, for to reckon 
the total casualties as five times the number of prisoners is 
a normal calculation. His real losses have not been nearly 
so high, for the proportion of prisoners taken has been 
abnormal. Most of these prisoners have been captured 
during the early period of each blow of the counter-offensive, 
when a belt of the enemy defences, varying from eight 
miles to two, has been ovemm. These blows have been 
successful. Those of Julj' i8th, between Soissons and 
Chateau-Thierry, of August 8th in front of Amiens, of August 
2ist north of the Somme, and of August 26th upon the 
Scarpe. In each case the number of prisoners taken on the 
first day has largely exceeded those taken in the remainder 
of the battle, with the exception of the third blow, that of 
August 2ist, when the first day had very limited objectives, 
giving only 3,000 prisoners, while the next five days gave 
almost as many prisoners each day as the first. But although 
the number of prisoners is here no exact guide to the total 
casualties, it is difficult to believe that those total casualties 
are much less than 6,400. There arfe only two sources from 
which such losses can be recruited. The first is the hospital 
returns, the second is the incorporation of class 1920 — that is, 
of the boys just under or just over 18 years of age — who 
were born in igoo, and who have been under training during 
the last few months. The total number of the latter before 
the end of the year will probably come to about 450,000 ; 
but they can only gradually be incorporated, for the quality 
is necessarily poor, and the proportion which you can allow 
in each unit therefore necessarily restricted. The -hospital 
retu'rns are calculated upon an average of four months. 
There has not yet been time, therefore, for any considerable 
number to have come back to reinforce the German line. 
We may sum up and say that we haye on front of us pro- 
bably less than 1,400,000 bayonets — certainly not more than 
i^ millions — and of these less than half a million are at the 
present moment south and easf of the Argonne Forest. The 
remainder are held in the desperate fighting between Rheims 
and Arras, or are holding the line between Arras and the 
sea or the trenches of Champagne between Rheims and 
Argonne : the two wings upon either side of the great action 
which have hitherto not come into play. 
As against this curiously ill-distributed line, the Allies 
have now — quite apart from a tactical superiority, of which 
I will speak in a moment — a definite superiority in numbers, 
which is due, of course, to the continued advent of fresh 
American troops trained and ready for the line. This 
superiority is not yet very great, but it is already apparent 
and it is rapidly growing. 
The possession of the initiative enables this superiority 
to be used in such a fashion that the congestion of the enemy 
between Arras and Rheims is maintained without a corre- 
sponding congestion upon the Allied side. The enemy is 
"held"— that is, compellecj to stand and concentrate — 
between Arras and Rheims by forces which are not superior 
in number to his total force in the region, and meanwhile 
Allied strength is released for action elsewhere. 
The reason that the Allied Higher Command can thus 
compel the enemy to concentrate on one section of his line 
(which is, as- the crow flies, less than 100 miles in length out 
of a total of between 400 and 500), and that without any 
undue counter concentration of the Allies against him — is 
due to two reasons. First, the fact tnat the enemy is here 
not only on the defensive, but anxiously trying to retreat 
in order to redistribute his strength to shorten his line in the 
north (thereby saving men), and to reinforce the south. 
Secondly, the possession by the Allies of two new tactical 
instruments which the enemy cannot rival : the new tanks 
and the American contingents — such small proportion of 
them as have been summoned to this northern battle, for the 
growing mass of them are elsewhere. 
As to the first of these conditions adverse to the enemy 
— the fact that he is trying to retreat — nay, that he must 
retreat if he is to have any hope of redistributing his strength 
and relieving the congestion in the north : 
We must appreciate why this necessity of retreat is adverse 
to him, and why, under present conditions, it compels him 
to continued congestion upon the sector where the retreat 
is taking place. 
When armies act independently in a war of movement 
— move, that is, as "great isolated units— a retreat is an 
operation which, short of bad blundering or of very bad 
obstacles upon its route, can usually be conducted without 
great loss and with dispatch. The retiring force, however 
closely watched by the pursuing force, has the advantage of 
knowing the moment when it chooses to move ; it can move 
first under cover of darkness, it can cover its movement 
with a screen of men — that is, with rearguards ; and though 
it is true that observation from the air has weakened the 
power to retreat unobserved, yet it must be remembered 
that observation from the air is not possible cvcrj' day, that 
it fails at night, and that another modern condition has 
increased the facilit}' of retreat, which is the value of the 
machine gun on the defensive. The rearguard screen armed 
^vith machine guns can check the pursuing force after a 
fashion that the old infantry could not. An army in move- 
ment, thus isolated, is one organism in the midst of con- 
siderable territory ; falling back upon its own line of com- 
munications perpendicular to its front, it can normally 
retreat with success — that is, upon its own plan and without 
great loss. It creates obstacles behind it as it goes, blowing 
up bridges and cross roads, damming streams, and so forth. 
PROBLEMS OF RETREAT 
But the retreat of an army under what may be called siege 
conditions — the retirement, that is, of a section which has 
continuous flanks upon either side — is another matter. 
In any retreat the mass of your force is, during the process 
of retreat, out of action. They are in column streaming 
back along the roads. If upon either flank of the sector from 
which retreat is taking place attacks are developed against 
you, you must meet them. You must concentrate men to 
stop them, or you will be turned. If the enem}' gets through 
either flank and comes down behind you on your lines of 
retreat, your army is ruined. With each such development 
on the flanks you must halt your retreat, for you have to 
send men up to meet the pressure. The moment you halt 
your retreat, your pursuer attacks violently not only on the 
wings on which he has just engaged, but upon your centre, 
too. And you must meet him there also. In general, this 
attempt to retire a portion of a fixed line is a much more 
difficult business than the retirement of an independent 
army, and a vigorous action against it on the part of the, 
pursuer can always throw out its plans. Vigorous pressure 
maintained by the pursuer has a paradoxical effect : it 
detains a would-be fugitive. Its object is one of actually 
checking withdrawals, of preventing the retreat from going 
as quickly as the retreating commander would desire ; for 
when a retreat is checked against its will, the losses involved 
by it are increased. This is exactly what has been going on 
against the German retirement since August 17th or iSth. 
It was about then that the German Higher Command 
determined"* to retire between the Somme and the Oise, 
Hardly had they made that decision when Mangin, on the 
southern of the retreating sector, attacked with great vigour. 
We know how the seven German divisions in front of him 
■ appealed for help to the centre, and how the centre sent back 
word that nd-help could be given for the moment. Mangin 
got right up to Aillette, and the retirement of the Germans 
between the Somme and the Oise was correspondingly ham- 
pered and expensive. But that was not all. In the second 
step of that retirement another blow was given on the other 
wing, to the north, and the British Third Army struck upon 
August 2ist. Once more the process of retirement was 
halted. German divisions had to be rushed up northward ; 
for if the Third British Army had advanced a few miles more 
the whole German retirement would have been jeopardised 
by their getting round and behind it. These German divi- 
sions having been rushed up northward, they contained the 
British pressure at great expense to themselves. Once more 
the retreat was organised, and began a fresh step, when, 
upon the 26th, the remainder of the British Third Army 
and the right of the British First Army struck again, still 
more to the northward, upon the valley of the Scarpe. .Here 
three German divisions — not of the best — had suddenly to 
be reinforced, and once more, therefore, the process of retreat 
was checked. At the moment of writing, the pressure is 
still so vigorously maintained that the retreat may be said to 
crawl even where it exists, and tobe halted altogether elsewhere. 
The total effect is that the enemy not only cannot retire 
as he wills, but, above all, cannot retire at the pace he would 
have chosen. He is condemned to continue a heavy expendi- 
ture in men in order even to prepare each step jn his retire- 
