December 23, 1915. 
LAND AND VV A r E R 
numbers of men is no longer possible. The defence 
of such a point — in this case a Port with its stores, 
wliarves, and the rest — can only be undertaken by 
the tracing of continuous lines against any sector 
of which the assault will proceed precisely as it 
proceeds in the field work and against the field 
trenches in Flanders or Champagne. The defence 
of Salonika, should it become necessary, will be 
work of the same kind as the work proceeding 
everywhere, upon the Western front at this 
moment. . . • 
ILxperience has shown that an average of 3,000 
men a mile, counting in all one's local reserves, is in- 
sufficient to hold inch a line on normal ground, 
while perhaps 5,000 men is more than enough. 
At any rate, a, perimeter such as that we are dis- 
cussing, involves the use of at least 200,000 troops, 
adequately supported, of course, with artillery, 
given as much mobility as possible by lateral com- 
munications — that is, roads alofig the lines for 
mo\'ing men, guns and stores — which at present 
hardly exist (and in the mountain sectors would 
be very difficult to establish), and continually 
repairing wastage by fresh draftS( 
As against the difficulty of lateral communica- 
tions, however, it must be pointed out that the 
most vulnerable part of the front (froin the 
Galiko round by the north to the Vardar river) 
though ill-provided with lateral communications 
(that is with roads or light railways by which men 
and munitions can be transported rapidly from 
one part of the line to the other) is fairly well sup- 
plied with radial communications (that . is, with 
communications reaching from the centre of supply 
— in this case Salonika town — to the circumference). 
The reader will mark upon the sketch map 
one good metalled road, and no less than three 
lines of railway serving the western part of the lines. 
We have seen in a previous article that, long 
as this perimeter is, it will be impossible to shorten 
it usefully. If the Daud Baba hill is not held all 
the land to the south is at its mercy. The same 
is true of the slighter height between that hill 
and the Vardar, while the abandonment of the 
Vardar line would be extremely dangerous. Only 
those on the spot can decide whether such an 
obstacle as the Galiko in its lower, course would be 
sufficient for a stable defence. But certainly to 
give up such an opportunity as that of the Vardar 
in the defence of Salonika would be something" 
only done under the pressure of absolute necessity. 
It is further clear, quite apart from the 
nature of the ground round the town, that the 
perimeter cannot be reduced because it is neces- 
sary to have a belt securing the Port from long 
range bombardment. 
T\\e harbour of Salonika behind the break- 
water is about one-fifth of a mile in length, with 
another quarter of a mile or so extension in open 
quay to the east. It is a target which can be 
exactly established and subjected to destructive 
bombardment unless the heavy guns which the 
enemy can ultimately bring up are kept out of 
range ; and this can only be done by making sure 
l:hat they shall not be emplaced within three or 
four miles of the nearest point at which the 
mountain ridge approaches the town. 
It seems probable from the meagre reports 
reaching the West that this horse-shoe line round 
from the Vardar River to the eastern shore of the 
Gulf of Salonika will be the one choice, but there 
is an interesting alternative liiie well worth study- 
ing which mav be callerl " the line of the lakes." 
THE LINE OF THE LAKES. 
If the reader will look at the sketch map vvliich 
illustrates this article, he will see that from the 
wooded mountain mass of the Hortak Mountain 
there i'uns a chain of positions continuing due east 
until one strikes the sea at the Gulf of- Rendina. 
These positions consist in three groups of hills 
(numbered upon the sketch I, II and III 
respectively), cov^ered by two lakes and' a 
liver. 
The first hills I, with nauisu summits about 
1,500 feet above the sea, stand just east of the 
Hortak and above lake Langaza. The second 
(called the Kartal Tepe) II, running up to about 
1,000 feet, are but the spurs of much higher land, 
which runs up behind them little south, and is 
covered with wood. The third group, III, is the 
pronounced, partly isolated bulk of Sugli^i, 
wooded upon its northern slopes and part of its 
summit, and somewhat over 1,600 feet 'in 
height. 
Were these hills the only element upon the line 
in question they would offer no advantage com- 
pared with the continuation of the hor£e-shoe 
round to the Gulf of Salonika. On the contrary, 
they would suffer from three disadvantages. First, 
that they involve a longer line (it is 34 miles fr6m 
the slopes of the Hortak Mountain to the Gulf 
of -Rendina) ; next, that each' group of hills 
is divided from its neighbour by deep valleys and, 
lastly, that the communications of such a line at 
present hardly exist. The furthest end of it, the 
Sugliani, could be supplied from the sea, the middle 
parts" would be supplied with difficulty from the 
Galatista carriage road behind them, which runs 
upon an average of ten miles to the south and 
beyond high" ridges of wooded hill and mountain. 
No continuous track, even, for moving men, guns, 
stores and shell, back and forth along this line 
now runs. 
But the line has a certain feature which may 
make up for these disadvantages. This feature 
is that already mentioned, the Lakes. From the 
head of Lake Langaza (just beneath the positions 
nearest to Salonika at a range of about 12,000 
yards from that town), right away to the sea at 
the Gulf of Rendina, there is an almost continuous 
line of natural obstacle. This line begins with 
Lake Langaza itself. Lake Beshik follows ; and 
the streams and marshes belonging to the;System-of 
the two lakes complete the whole. There is only 
one very short gap in this long line of= obstacles 
and that is the drier part of the isthmus between 
thie two lakes, a drier part marked by the hamlet of 
Sarai. The Iri Brook is here no ' appreciable 
obstacle, and there is between the marsh at the 
mouth of the brook and the marshes on Lake 
Langaza a gate perhaps four miles in extent. ' 
This gate is, however, completely dominated 
b}', and observed from the high hills to the south, 
and once one gets west of it, the obstacles proceed 
without interruption. There are first the marshes 
at the mouth of the Brook Iri ; then the whole 
length of the Beshik Lake, over 16 miles along and 
more than a mile broad at its narrowest. Lastly, 
the final four ihiles between the end of Lake Beshik 
and the sea, are covered by the outlet river across 
which are but two bridges, one near the lake and 
one near the mouth. Further, the southern shore 
of the Beshik Lake is largely marshy, a point of 
no great importance unless a crossing werefo.rcid, 
in which case the landing upon the further shore 
would be impeded by this natural featnrft. 
