August Q, 1917 
LAND & WATER 
to the south of it, yet so long as CzemowUz held there was 
still ample road facility for using that town as a terminus, and 
southward from it working the road and railway of the Sereth 
valley in support of the defensive line which was holding the 
Carpathians from 40 to 50 miles westward. With Czernowitz 
gone this road and railway lead nowhere. They would still 
be of full service if munitionment and supply came up from 
the south, but they do not, they come down from the north. 
The armies in Roumania are dependent upon the bases in 
Southern Russia and Kiev, and after the loss of Czernowitz 
they can only be supplied through Jassy by a very round- 
about way. 
The only favourable element in the situation is the obvious 
numerical weakness of the enemy. If he had been at the 
strength at which he was a year ago, he would, as we saw last 
week, have certainly destroyed the 7th and 8th Russian 
armies, for the breakdown of the second army had put him right 
upon their flank and even behind it, so that it seemed a matter 
of not more than two or three days before he should appear 
upon their communications and cut them off. That he was 
not able to do this was certainly not due to any strength in 
the Russian second army, which was in rout, and had not yet 
begun its stand upon the frontier. It was due simply to lack of 
means in the pursuit. The same weakness may prevent the 
enemy from taking advantage of the present situation in 
the corner of Dorna Watra. 
Not a bad test of the future situation will be the point of 
Piatra. If the line has to fall back through northern Roumania 
so as to uncover Piatra, the northern valleys issuing from the; 
Carpathians are in possession of the enemy. If Piatra con- 
tinues to be held it means that the enemy is not in sufficient; 
strength to reap the fruits qf his opponents' breakdown. 
The Importance of C^culation— I 
Amid a great deal of correspondence which I have received 
upon the figures of enemy losses and man-power published 
from time to time in these columns, I find a certain proportion 
the tone of which is somewhat as follows : 
■' Of what use in this stage of the war is the publication of 
such estimates ? They are only calculations upon very im- 
perfect evidence, while the war itself and the progress of 
events do not bear out what the estimates promise." 
In other words, a certain proportion of my readers believe 
this sort of calculation to be at the best of little service to our 
judgment of the situation, and at the worst fallacious. 
J say " ^ certain proportion " only, for naturally the greater 
part of those who write to one about these figures are serious 
students — who know the fundamental importance of numbers 
in war and who have carefully followed the responsible studies 
appearing in the Continental Press. These are usually con- 
cerned with details — as, for in.stance, the evidence for and 
against the presence of class 1919 in the depots as early as the 
first week of May (everyone knew that the first batch would 
not be actually in the depots earUer than the end of April 
or later than the beginning of June). 
This " certain proportion," however, has its importance, 
and I would like to address myself to those composing it, 
both the comparati\-cly few who have written to me per- 
sonally and the much larger number who have acquired the 
same impression from the popular Press — I mean the im- 
pression that calculations of this sort are cither so imperfect 
as to be negligible, or, even if fairly accurate, are unimportant. 
I must, of course, as often before, beg the reader's inclulgence 
for a certain amount of repetition which is inevitable, because 
previous statements will have been missed by some and 
forgotten by many more. 
"There are two distinct elements in this feeling of doubt. 
The first concerns the accuracy of the^tatements made ; the 
second the value of such statements to our judgment even 
when they are proved to be fairly accurate. 
I will deal with these elements separately. First, then 
as to the accuracy of such statements : 
The various sorts of statistics published, whether in these 
columns or in any other survey of the war, are based upon 
different kinds of evidence and enjoy different degrees of 
certitude. They are of two kinds : Statistics of the enemy's 
positive strength — his total army, his fighting force, his man- 
power at any moment, his recruitment — ai^d the negative 
element, his losses. 
In the first of these there are four categories of different 
degrees of accuracy. 
(i) Estimates of the whole " ration strengths" of the enemy 
at any moment are based upon knowledge that is very full 
and checked in numerous ways. When we say, for instance, 
that the German Empire (which provides about half the total 
belligerent force opposed to us) is maintaining a ration 
strength of somewhat over 5I millions, we are saying some- 
thing which is certainly not wrong by 10 per cent. It is a 
figure that is taken as a sort of common knowledge by every- 
one dealing with the subject. 
(2) When we say that the strength in organised figblinc; 
units, that is, the strength of the field army properly so-called 
(as distinguished from the services auxiliary to the fighting 
bodies) is ^i millions, we are again upon quite firm ground. 
The enemy is possessed of much the same knowledge witli 
regard to ourselves, and we certainly of full knowledge with 
regard to him in this respect. The average strength of full 
units is known, and by far the greater part of such units at 
any moment have been identified, and thejr positions can 
easily be dotted down upon a map. All sorts of information 
from the indication of divisions in contact with our forces by 
the interrogation of prisoners, the consultation of captured 
documents and the many other forms of intelligence avail- 
able, concur to this result. There are moments when the 
margin of error is greater, as, for instance, when the enemy 
is moving considerable bodies of men, or when he is in process 
of making new formations. But his fighting strength is 
known upon the average within a very small margin of error. 
This second category just mentioned is, of course, one of 
the two bases of all such estimates, for, according to the 
lighting force employed and the nature of its activity will be 
the proportion of losses and the size of drafts required to 
repair them. 
(3) The annual power of recruitment, the " crops of men," 
is known still more accurately as to its maximum. All the 
belligerent powers are able to make the test for themselves 
and to reason without danger of error, that any other belli- 
gerent power, enemy or Ally, will be able to find not more 
than such and such a number of youths capable of service 
at such and such an age. At any rate, this calculation is 
always easy in the case of nations which keep accurate census 
tables, and which are know^ to be making the fullest effort 
possible. In such estimates, then, one always supposes the 
enemy to be able to get all the young men available — all that 
proportion which will pass the doctor. In point of fact, the 
real number raised is less, but one remains on the safe side by 
.standing to one's maximum. For instance, one sa}^ of the 
German class 1919, most of which is already called up and 
the rest of which will be called up in the present summer, that 
it can yield some 500,000. It may yield considerably less, 
but it will not yield more, and we set it down at that figure. 
(4) The only remaining class of recruitment is the men, 
" combed out " from auxiliary or civiUan employment. 
This is much the most doubtful figure. The estimates 
made of it have to allow for a considerable mjirgin of error, 
and the political and other conditions governing such " comb- 
ing out " change from month to month with the situation of 
tlie enemy. It is difficult to ascertain with accuracy how 
far the enslavement of occupied territories relieves him, and 
he is himself, like every other belligerent, compelled to con- 
tinual changes of estimate in the number of valid men he can 
afford to take away from civil or auxiliary work. Evidence 
is frequently forthcoming of men having been taken away 
from such and such work behind the army, of the numbers so 
taken having proved excessive and dangerous, and of the 
consequent necessity'of returning some of them back from the 
army to their former work. 
But though the category of " combing out " is the least 
certain and the one subject to most fluctuations and errors, 
there are these two things to be said about it at the present 
moment. First : that it is and has been for now more than a 
year, a very small category of the whole, and that therefore the 
errors in it do not greatly affect the errors of our total estimates ; 
and secondly : that the power of further combing out men 
has for many months past ceased. Perhaps the rrioment of 
greatest error in this respect was in the central period of the 
war, say from November 15th to August i6th. It was then 
that the novel and detestable poUcy of enslavement first began 
to be generally practised and the Allies naturally under- 
estimated at first the length to which the enemy would push 
such an atrocity. The first serious " combing out " for the 
German forcesjtook place, it will be remembered, in the month of 
October, 1915. The process continued with various experiments 
and shiftings from the army and back to the army until 
the end of last year. The present year was not far advanced 
when a point had been reached after which no further recruit- 
ment from this source was possible. 
It must be remembered, of course, that any considerable 
political change at once affects this category. The occupation 
or abandonment of any considerable piece of territory, help 
direct or indirect from a neutral, etc., changes the balance, 
but as things now are and have been for many months, 
