Salonika Campaigns

"Salonika Campaigns, 1915-1918. - Under the heading of Serbian Campaigns the conquest of Serbia in 1915 by AustroHungarian forces is narrated. The idea of reinforcing the Serbian front with Allied forces had been contemplated both in England and in France some time before it was carried out. British and French guns, in charge of naval missions, had taken some part in the campaign of 1914, and stores had been sent up from Salonika at intervals. In the winter of 1914-5 Lord Kitchener several times considered the advisability of sending a number of the British Army Divisions into Serbia via Salonika. On the part of the French, M. Briand, it is said, proposed later in 1914 to make a serious military effort in the Balkans. But the Dardanelles campaign diverted attention from this project, and it was not till in August 1915, when the failure of the Dardanelles offensive was evident, that the creation of an AngloFrench army on the Balkan front was seriously undertaken. General Sarrail, whose military reputation stood very high in France, had been suddenly deprived of his command of the III. Army by Joffre, ostensibly owing to an unsuccessful combat at Boureuilles in Argonne, but really as the result of long-continued friction between the two. Sarrail, however, stood in close relations with the political leaders of the Left, and the autocratic methods of Joffre's G.Q.G. had already raised considerable opposition in the Government and the Chamber; it suited the Government, therefore, to satisfy the Left, to snub the G.Q.G., and to remove to a distance a forceful and ambitious personality, by sending Sarrail to the Mediterranean as commander of an army yet to be created.

Appointed on Aug. 5, Sarrail was ordered to study the military situation and submit proposals. In his written projects he came to the conclusion that it was impossible to abandon ground in the Gallipoli peninsula, and had asked for both his own and the British contingents to be made up entirely from forces in France or in England. An inter-Allied conference, held at Calais early in September, had agreed to this, but with the reservation that no forces were to go till after the forthcoming Champagne and Artois offensives had taken place. But the news of the Bulgarian mobilization drove home at last the urgency of the crisis. Orders went to the Dardanelles on Sept. 26 for two British Divisions - in the sequel one - to go thence to Salonika; the French " Expeditionary Corps " was likewise to send a Division, and the Greek authorities had agreed to permit the landing. Sarrail himself was to bring a mixed brigade from France, as an earnest of the forces promised later.

On Oct. 3 advanced parties of the French landed at Salonika without difficulty, only a formal protest being made by the authorities on the spot. Next day M. Venizelos in a speech at Athens declared that Greece would come to the aid of her ally Serbia against any attack by Bulgaria, and at once a crisis arose at Athens. On the 5th King Constantine informed Venizelos that the policy indicated had not his support, and the Government fell, to give place to the neutralist Zaimis cabinet.

During the first few days instructions from Paris to Bailloud (commanding on the spot pending Sarrail's arrival) varied several times, apparently in accordance with political nuances. At first (Oct. 3) the word was to concentrate at Nish, in the heart of Serbia; next, the Greek frontier was not to be crossed (Oct. io); and then again (Oct. r2) authority was given to take over protection of the railway between Demir Kapu defile and the Greek frontier against possible attack from Strumitsa in Bulgaria, thereby releasing a small Serbian force to rejoin its own army. Meantime the Serbians demanded more direct assistance, but Sarrail (who arrived on the 12th), taking into account the size of his force only 12 divisions plus the British 10th Div. which was not under his orders - and the fact that it could only disembark and push on by driblets, determined to limit his advance to the near side of Demir Kapu. On Oct. 14 the leading French troops arrived at Strumitsa station (in Serbia) in time to aid the Serbian railway guards in repelling an inroad from Strumitsa.

In Sarrail's opinion the only service he could render was to concentrate on the routes to Strumitsa, and, by an offensive into Bulgarian territory, to draw off as many Bulgarian forces as possible from the main attack further north. General Mahon, commanding the British 10th Div., took the same view,' and formed a mixed force which began to move up to Doiran, on the right rear of the French group in the Rabrovo region. On the 17th, however, in answer to a request from the Serbian commander at Uskub, Sarrail began to push a brigade beyond the Demir Kapu defile to Krivolak, but he refused to advance it to Veles, though again pressed to do so by the Serbs, and in fact a sharp attack developed from Strumitsa on Rabrovo on the 21st and 22nd, which, till it was repulsed, threatened to isolate all French detachments N. of Strumitsa station. Meanwhile, Paris sent further instructions to the effect that all possible help should be afforded to the Serbs, subject to the limitation that the French communications with Salonika were in no case to be compromised. In reality, the French and British Governments were very uneasy about the attitude of the Greek Army, a considerable force of which lay in the region N.E. of Salonika. The fall of Venizelos had put an end to the prospect of Greek cooperation, and under the new regime the local military and civil authorities began to oppose every move of the Allies, which was not entirely covered by Serbia's treaty rights, to the use of Salonika and the railway. Thus, when Mahon's force moved forward the use of the Salonika-Kilkish (Kukush)-Doiran line was refused, and it had to use the main line, detrain in the midst of the French, and work thence outwards towards its post at Doiran.

On Nov. r the 122nd French Div. began to arrive from France, and Sarrail had already prepared to attack from Rabrovo towards Strumitsa with Bailloud's 156th Div., with Mahon in echelon behind his right, while his forces about Krivolak and Kavadar (5rst Div.) made ready to attack in flank any Bulgarian force which should advance up the Cerna (Tserna) in pursuit of the Serbians. On Nov. 3 an attack was accordingly delivered northward from a front E. of Rabrovo; weather and the difficulty of the country brought it, however, to a standstill on the 6th, though local advances were made later. At this moment (Nov. 4) Gallieni, having become War Minister in the new French cabinet, telegraphed orders for the French Army to operate towards Veles, adding that four more British divisions were to be sent, which on arrival would take over the front leftwards from Doiran. The British 22nd Div. was in fact already close to Salonika, with another under orders to follow. But Sarrail judged that it was impossible to wait for these reinforcements. Todorov's Bulgarian Army had already thrust itself between the Serbian Main Army and Krivolak, and the 1 According to Sarrail, the British Government instructed Mahon that his troops were to remain at Salonika, and it was on his own initiative that the British general formed a mobile force. Further instructions authorized Mahon to move forward but forbade him to cross the Greek frontier, until on Oct. 27 a final telegram removed this restriction.

urgent thing was to relieve pressure on that part of the Serbian forces which was retiring by the Babuna pass on Prilep, while reserving the possibility of action towards Veles if the Serbian Main Army should after all seek to break through towards its Allies. Orders were therefore given to the Krivolak-Kavadar force (57th Div. to be reinforced by the 122nd Div.) to take the offensive westward over the Cerna, so as to strike the pursuers in flank or rear. On the 6th-9th accordingly the 57th Div. crossed the Cerna and pushed an advance into the mountains towards the Babuna, still held by the Serbs. But the Bulgarians were in force, and the French retired to their Cerna bridgeheads, which the Bulgarians attacked without success on the 12th, 13th, r4th and 15th.

During these and the following days instructions came repeatedly from Paris to modify the French commander's views and dispositions, now laying emphasis upon cooperation with the Serbs, now upon dangers from the Greek Army in rear. Finally on Nov. 21 Sarrail was given a free hand to decide what aid he could give to the Serbians and at what moment he should retire on Salonika. He adopted at first a middle course. He wished neither to attack at the risk of involving two-thirds of his forces in the Serbian debacle (the Babuna had been turned by the N. on Nov. 14), nor to fall back to Salonika, where prestige counted for so much, but to hold on in the entrenched camp of Kavadar in the hope of " something turning up." On the 21st-22nd, however, the retirement of the 122nd Div. over the Cerna under some pressure, together with the general military situation and a definitive refusal of reinforcements from France, 2 decided him in favour of falling back to Salonika, a decision approved by Gallieni. Four days later Sarrail was officially informed that the Serbians were retreating in the Adriatic direction. The preparations for the Vardar retirement had already begun on the 24th with the seizure of a position on the E. bank, to prevent interference with the retreat of the KrivolakKavadar force on Demir Kapu. On Dec. 1 only rear-guards remained at Krivolak. By the night of the 3rd-4th all troops were inside Demir Kapu, and on the 6th this position also was given up. On the 8th the Bulgarians, who had from time to time attacked the rear-guards on the Vardar and the positions near Kosturino on the Strumitsa route, delivered a more concerted attack on the front Ormanli (now held by the British)- Kosturino-Gradets on the E. and Mirovcha Petrovo on the W. of the Vardar. Their evident intention was envelopment, and on the 9th, judging the centre of his line to be too pronounced a salient, Sarrail took up a position along the Petrovska stream, W. of the Vardar and the heights of Dedeli E. of it, the village of Dedeli being held by Mahon's forces, which from that point were echeloned back to Lake Doiran. From this position also the Allies retired under threat of envelopment during the night 11-12, after holding their ground against attacks on the 11th. Lastly, the French 122nd and S7th Divs., at Gyevgyeli (Gevgeli) and near Doiran, covered the evacuation of part of Mahon's force on to the Salonika railway and the reconstitution of the 156th French Div., which had been considerably split up.

Thus the drive into Serbia came to an end, with little material loss, but a sad diminution of prestige, and the forces fell back to the following positions about Salonika: advanced guards of 120th French Div. Karasuli with a detachment at Gumenye, and of 57th French Div. with cavalry, Kilkish, with a detachment at Kilindir; main body (122nd, 156th, 57th) in position on the line Doganyi-Daudli. British 10th and 22nd Divs. Salonika, with other British forces arriving. Important points on the railway had been destroyed during the retreat. Meanwhile, on Dec. 4, At that moment, according to Falkenhayn, the combatant strength of the Allies in France was to that of the Germans in the ratio of rather more than 3 to 2. Sarrail says that in his interview with Lord Kitchener on Nov. 17, the latter informed him that Joffre had declared that he would not give him (Sarrail) another man, and that the British would furnish five divisions instead. In accordance with this promise, besides the 10th and 22nd Divs., the 27th and 28th Divs. from France landed at Salonika in the last days of Nov. and first days of Dec., and the 26th Div. also from France, early in Jan. 1916.

the Serbs had evacuated Monastir in their now frankly westward retirement.

The reassembly of the Allied Salonika forces around their port of origin naturally raised the question - were they to remain there? Their locus standi had been the fact that they were Allies of Serbia using a line of communications to which Serbia was by treaty entitled. This part of the case no longer existed, Serbia being wholly in the hands of the enemy, and could only be revived if and when the Serbian Army were transferred from the Adriatic ports on which it had retreated to Salonika. Another part of the justification for the Allies' presence was the admitted fact that they had come at the request of Venizelos, and for the purposes of common action with the Greeks, but since Venizelos's fall even the Zaimis cabinet, representing " benevolent " neutrality, had given way to a cabinet representing at least strict neutrality,' which gave the Germanophil element at Salonika all the official justification it needed to pursue the policy of obstruction that it had already initiated in the Zaimis period. On the other hand the factor of prestige was one of great weight, especially in view of the pending abandonment of the Dardanelles compaign, and although Sarrail suggested that evacuation followed by a dramatic offensive at some other point would more than restore the lost prestige, it was decided that Salonika should be held. Beyond that decision, however, no clear military or political intention was at that time formed. The policies of the British, French and Russian Governments were in unison as to the problem of Greece, and it seems to have been thought that, by remaining, the Salonika force would confront the enemy with as difficult a diplomatic problem as its own. This was, indeed, the case. The policy to be followed by the Central Powers, both towards Greece and in occupied Serbia and Albania, was wholly unsettled. " While the troops of the two Imperial Armies were hastening from victory to victory," says General von Cramon, the German military commissioner at Austrian headquarters, " behind the scenes, at the two general headquarters, the clouds were gathering of that conflict which in the end brought about the reverses of 1916." Although on Nov. 6 it had been agreed that operations were to be pushed with all energy towards Salonika, Falkenhayn almost immediately began to check the further south-westward advance of German troops, and though Conrad succeeded in bringing the German command to renewed cooperation, this was obviously to be limited to a minimum, both on account of supply difficulties in the Balkans and of the pressing requirements of the two main theatres - in particular those of the forthcoming attack on Verdun, of which only a few men in the German headquarters and none in the Austrian had the secret. Falkenhayn's view was that the Bulgarians alone should undertake the campaign in southern Serbia. But, whatever the attitude of Greece towards Germany, it was so hostile towards Bulgaria that to cross the frontier in pursuit of Sarrail without a large proportion of German troops being included in the advance was politically impossible. Austria herself was absorbed in Montenegrin-Albanian enterprises, and could give no direct assistance in the advance to Salonika that her general staff advocated. Moreover, Conrad had his secret as well as Falkenhayn - he was planning to carry out his offensive of Asiago, with or without the aid of Germany.

At the end of 1915 therefore, though the Central Powers had succeeded in their purpose - Serbia being conquered and the railway to Constantinople reopened - whereas the Entente had failed, the outlook was no clearer for the former than for the latter. The pursuit was accordingly suspended at the frontier, partly perhaps in the hope that the Entente would itself take the initiative in closing down the operations. If they did not do so Falkenhayn was determined that eventually 1 The first act of the Skouloudis ministry had been to announce that any of the Allied forces in Serbia which retreated into Greece would be disarmed and interned. A prompt note from the British and French Governments closed this incident, but the indication of policy was unmistakable. About the same time Skouloudis notified the Bulgarian Government that it would not permit the latter's troops to cross the frontier.

the Bulgarians alone should remain on this front. They were, by the terms of the military agreement, unavailable for any other, and if they succeeded in containing even a smaller force of Entente troops that was not so limited, something was gained for nothing. On the other hand this idea implied a defensive position short of the Greek frontier, as a purely Bulgarian advance into Greece was impossible. Thus, at the beginning of 1915, the opposing forces stood roughly 20 m. apart, each limited against its own will to a strict defensive by political conditions and each regarded by its own superior authorities as a " commitment." At the end of the year two incidents occurred to illustrate the complexities of the Salonika front. On Dec. 30, though Bulgarian and German forces were forbidden to cross the frontier, German aircraft, by order, bombed the city of Salonika itself, where nine out of ten of their possible victims were neutrals and the tenth an agent of their own side. Sarrail promptly retaliated by arresting the German, Austrian and Bulgarian consuls, hitherto left unmolested. Another air raid took place on Feb. i 1916, to which the Allies replied by bombing the village of Petrich, just within the Bulgarian frontier, but as the village contained Greek and Serbian as well as Bulgarian inhabitants, a complaint was made, and Sarrail received orders not to repeat such raids. A few days before this another incident showed that the personal estrangements of Joffre and Sarrail were still operative. The army of the Orient had been brought under Joffre's command 2 early in December, and Joffre had taken the opportunity to send out Castelnau to report on Sarrail's management of the situation. Castelnau, however, pronounced himself satisfied with what he saw, and only issued a few instructions as to details. Nevertheless, in various ways the friends and the enemies of Sarrail alike busied themselves with accusations and counter-accusations, out of which a regular affaire was growing up to complicate an already confused situation. Relations between Sarrail and Mahon on the other hand were excellent, and although each was independent of the other, and the British general was himself under the command of General Sir C. Monro, commander-in-chief in the Mediterranean, no important divergencies of policy developed during the phase of passive defence in the precincts of Salonika.

With the Greeks, naturally, all possible causes of friction existed. Army commanders operating under war conditions are not prone to sacrifice realities to appearances, and what seemed to them plain military common sense was, from the point of view of the Greeks, high-handed conduct to be resisted by all safe means of obstruction. Amongst the major questions at issue were the disarmament of the coast defences of Salonika, the use of the Salonika-Doiran railway for the British contingent, the feeding of the Greek forces E. of Salonika who were dependent for supply upon railways seized by the Allies, and the continuance or non-continuance of the Greek garrison in Salonika city. Minor questions of an administrative character were naturally innumerable. Most of the energy of the staffs in Salonika and the legations at Athens was devoted to finding solutions for conflicts which the equivocal position of the Allies made inevitable. 3 During these conflicts the Salonika lines, 2 Joffre was Chef of the " North-Eastern group of Armies," no other formations having been contemplated before the war. On being sent to the E. Sarrail was appointed Commandant Chef also. But, in Dec., Briand placed Sarrail's forces under Joffre's supreme command On Jan. 12 1916 the bridge of Demir Hissar on the Struma was blown up by a special force sent out by Sarrail in the presence of the Greek forces stationed there - a high-handed act which could only be excused or justified by the necessity of preventing the Bulgarians and Germans from deploying heavy artillery against the N.E. part of Salonika in case of siege. On Jan. 28 1916 another problem received an enforced solution, after negotiations had failed to find an " elegant " one. Anglo-French forces by a coup de main occupied the Greek coast-defence batteries on the Gulf of Salonika. These incidents naturally intensified the hostility of the Greek officers and officials to the Allied occupation, or at least gave them tangible grievances. In particular, the feeding of the Greek forces isolated by the cutting of the Struma railways caused difficulties, and from it, in part at any rate, arose the critical question of demobilizing the Greek Army in the Spring of 1916.

with the aid of civil labour, were made defensible by the first weeks of the new year. The line selected ran from the Vardar mouth, round by Doganzi and Daudli to the neighbourhood of Langaza, whence it passed along the barrier of lakes to the head of the gulf of Orfano-80 m. of frontage for a force of nine divisions.' Of this frontage, however, nearly 45 m. was guarded by lake and swamp; and, taking into account the presence of large bodies of Greek troops in the Seres - Rupel region to the right front and in the Vodena - Florina region to the left front, Sarrail considered that danger was practically confined to the central sector between Lake Langaza and the Vardar, in the event the position was ever attacked.

During this period (Jan. - Feb. 1916), the Bulgarians were reinforced by the German XI. Army (von Gallwitz) consisting or the IV. Res. Corps (101st and 103rd Divs.) and the Alpine Corps, and by their own 1st Army, all these forces aligning themselves along the Greek border from Lake Ohrida to the point at which the Struma enters Bulgarian territory. The 1st Bulgarian Army, with flank guards at Dibra and Elbasan in Albania, had two divisions 2 on the front S. of MonastirDuditsa; the XI. German Army, with 12 Bulgarian Divisions attached, held the Vardar valley between Duditsa and the Belashitsa Planina, and Todorov's II. Bulgarian Army of three divisions that ranged from Strumitsa to Petrich, with detachments further E. at Nevrekop (Mesta valley). But in March Falkenhayn began to withdraw all the German formations save the foist Div., which continued in the Balkans and was gradually reduced to a cadre. On his side Sarrail made some slight demonstrations towards Doiran and towards Vodena, but otherwise no move occurred. Early in March 1916, however, in the crisis produced by the attack on Verdun, Joffre telegraphed orders to Sarrail to advance in order to fix the enemy's forces on his front. On the other hand General Mahon, on asking for instructions, was forbidden to move until authorized by the British Government.' The relief offensive, therefore, was limited to a skirmishing advance by the French, which began on March 13, and gradually brought the French 51st Div. to the N. of Kilkish, and the 122nd to the N. of Lake Amatovo, the 156th between them (March 31).

During April 1916, while French cavalry moved out W. of the Vardar towards Vodena and the 17th Colonial Div. came up behind the centre, the British in their turn began to move up to Kilkish, authority to participate " in an operation of a demonstrative character " having been given by the War Office about April 10 (Joffre to Sarrail, April 20). Lastly, the Serbian Army, reconstituted and partially reequipped at Corfu, was beginning to land at Salonika, and by June 1 118,000 combatants and non-combatants were present, completing their equipment and organization in the Chalcidic peninsula. These methodical proceedings, however, did not satisfy Joffre, whose instructions to Sarrail from March onwards were to prepare for an offensive in earnest. To Sarrail's demand for reinforcements for such an offensive, the French commanderin-chief replied that the French Army of the E. must prepare to attack at the moment fixed by himself, even without the British. In explanation, he hinted that when that moment came, not only would British objections be removed, and all five British divisions be equipped for mountain warfare (making Sarrail's total force, with the Serbians, 300,000 strong to the enemy's 260,000), but Rumania and Greece would be in the field as his Allies. Thus for the first time since the Serbian retreat the Salonika force was assigned to a positive purpose. It will appear in the sequel how much of reality and how much of illusion was contained in the scheme, which, in sum, was to attack at a date chosen in relation to other theatres and especially Rumania, with Sofia as the objective.

I The withdrawal from Helles freed further French troops, from whom a serviceable brigade was made up and combined with a brigade from France to form the 17th Colonial Division.

s The Bulgarian division had twice the infantry strength of a French or German division.

Sarrail's orders were issued by G.Q.G. without consultation with the British government or Lord Kitchener. (Sarrail, p. 83.) Meantime, an important incident had taken place on the Struma frontier. In accordance with their declared policy of standing aside and leaving a " lists " for the combatants, the Greeks had disarmed and evacuated their fort of Dova Tepe, situated on the watershed between the Vardar and Struma basins and commanding a knot of communications. In the course of his gradual advance to the frontier, Sarrail put a detachment into this fort on May io. But further to the right, outside his reach, lay a still more important fort, that of Rupel defile. This fort was not merely disarmed but actually handed over to the Bulgarians by the local Greek general, with or without authority from Athens (May 26).

Events had moved. Though the German forces (except the cadre of the foist Div.) had by this time been withdrawn from the Balkan front, the Greeks had apparently overcome their repugnance to a purely Bulgarian inroad, to the extent of actually facilitating it. The Allies' right was, potentially, turned, and if the occurrence were any indication of probabilities of the future, their rear also was endangered. Action was taken promptly by Sarrail. A mobile group of all arms was moved into the Struma region, and with the agreement of Gen. Milne (who in May succeeded Mahon in command of the British) and of the Entente Governments, the Greek authorities at Salonika were deprived of power by the proclamation of a state of siege (June 3). A day or two later London and Paris also acted. An economic blockade of the Greek coast was declared, and on the 18th Sarrail was ordered to send a brigade by sea to Athens. King Constantine accepted the ultimatum of the Allies (June 21), and Zaimis returned to power on the basis of friendly neutrality. A little later the Rupel incident had its last and most important sequel in the Venezelist coup d'etat of Aug. 30.

Militarily the seizure of Rupel, carried out at the suggestion of Falkenhayn, seems to have had no truly offensive intention. The Central Powers had abandoned the idea of invading Greece once for all about the end of March, and the Bulgarians acted with the idea of guarding their left, and securing connexion with any Turkish forces which might be sent to their aid by the Constantinople - Seres railway, though in view of the situation in Armenia such a reinforcement was unlikely. As for Sarrail, so also for his opponents, the Balkan front was already involved in a larger game.

As has been noted above, the idea of an Allied offensive from Salonika in cooperation with a Rumanian intervention came under discussion as early as mid-April; at that time Joffre seems to have thought that this intervention might come in a few weeks, for he overruled Sarrail's objections to commencing operations in the hot weather, and fixed the month of June for the beginning of the offensive. Under these instructions Sarrail formed his first plan (May 2) which was, in brief, to employ the Serbs on the left wing for the attack of Monastir (frontally and by envelopment), and for pressure on the .Cerna bend and the passes further E. towards the Vardar; to place a French division on the Vardar and the railway; to have three British divisions, with a fourth on their right rear, so placed as to execute a demonstrative attack on the strongest part of the enemy's front, viz. Vardar - Doiran; and to attack with three French divisions from Popovo Surlovo - Dova Tepe and Poroy northwards through the Belashitsa Planina, while the French Struma mobile group demonstrated towards the Demir Hisar angle of the Struma, and the fifth British division with cavalry watched the lower Struma front. In case the semi-offensive, semi-demonstrative, operations should develop into a real advance, the Serbs were to take Veles and Shtip, the British RadovishtaStrumitsa, and the French Jumaya as their objectives. The armies would thus condense their front as they advanced, the route Monastir - Veles marking the extreme left of the Serbian movement, and that of the upper Struma, famous in the war of 1913, taking the French into positions on the Bulgarian line of retreat. But the negotiations of the Allies with Rumania, and their internal discussions relative to their Salonika operations and their policy in Greece, dragged on. On June 6 Gen. Milne was informed by the British Government that he was not to engage in offensive operations, and was only to consider himself under Sarrail's orders in respect of the defence of the entrenched camp. He informed Sarrail accordingly, and suggested that the British should take over the Struma front, to which Sarrail agreed. The Serbs were now preparing to take up the front from Vodena to Lyumnitsa, with their centre of gravity on the right, the French held from Lyumnitsa to the Poroy road, and the British to the right of that road from Loznitsa to Orlyak. On the 12th Sarrail was ordered not to take any action that would involve the British in operations unconcerned with the defence of Salonika itself, and to limit himself to threatening the Bulgarians by a deployment close up to the frontier; and on the 14th he was notified that the French Government had agreed to the British proposal to postpone the offensive.

A few days later, on the 25th, he was informed that though the instructions of the 14th held good in general, events might rapidly make it necessary for him to attack, though with a limited objective, and using the French and the Serbians only; and on July 15 he was told that the British Government had agreed that if Rumania intervened all British troops equipped for mountain warfare should participate in Sarrail's offensive, and instructed to prepare to " pin the Bulgarians on the Greek frontier and put them out of action so far as serious operations against Rumania were concerned." Three days later, on July 17, G.Q.G. informed the French commander that the EntenteRumanian military convention would probably be signed on the basis of (a) an offensive from Salonika on Aug. 1, to cover the final preparations of the Rumanians and their initial operations against Transylvania; (b) a Russo-Rumanian offensive beginning on Aug. 8, and directed against Bulgaria; and finally (c) a combined advance of the Russo-Rumanian Army and the Salonika forces with a view to uniting and crushing the Bulgarian Army in the field. A few days later a formula agreed upon between the various Allies constituted Sarrail commanderin-chief of the French, British, and Serbian Armies, as also of the Italian and Russian contingents, 1 Gen. Cordonnier being appointed to command the Armee francaise d'Orient as a constituent part of the Allied Army.

Sarrail's new plan was to dispose Milne's available forces on the front Dova Tepe (exclusive) - Lake Ardzan (exclusive) or to the Vardar if possible, to reduce Cordonnier's troops E. of Dova Tepe and Milne's on the Struma front 2 to a minimum, and with Cordonnier's Army to attack on the front VardarDoiran, while the Serbs from above Vodena made their main attack on Huma and subsidiary advances towards the Cerna bend and possibly Monastir. This plan was approved by Joffre, who added that the British Army would receive instructions from the War Office not to limit itself to defensive or demonstrative action. But these instructions, from Gen. Robertson to Gen. Milne, introduced an important limitation in their general approval. Milne was " not to try to take the enemy's positions until an adequate equipment of heavy artillery and other conditions gave a reasonable expectation of success," and the offensive was " not to be taken till Rumania definitely came into the field," an event of which Sir W. Robertson, like Sarrail, had his doubts. Presently came the first hitch in the military convention negotiations. Rumania was not to move till Aug. 14, and Sarrail was to act ten days before that date. But on Aug. 3, the eve of the offensive, the convention was still unsigned, Rumania having expressed the intention of not declaring war on Bulgaria unless large Russian forces were added to her Danube Army; in these circumstances Sarrail's mission 1 In the case of the Italian Division the powers of the commanderin-chief were specially limited. The Russians were, however, unreservedly at the disposal of the French. In general the formula from which Sarrail derived his authority was somewhat similar to that which was agreed upon later in the case of Nivelle. It was far from being a real international command such as that of Foch in 1918.

2 The prevalence in that region of malaria, discovered by experience, had caused Sarrail to abandon the earlier project of delivering a principal attack with three French Divisions on the Belashitsa front in the summer months.

was reduced to " harassing " the Bulgarian Armies on his front. without ulterior purpose, from a date to be determined later (telegrams from Joffre to Sarrail Aug. 3 and 6). Finally, the convention was signed on Aug. 17, without any engagements on Rumania's part to declare war on Bulgaria. On that very day the Bulgarians began to push forward. Proposals for shortening and improving their line by pushing it forward on the one hand from the Monastir frontier towards Ostrovo, and on the other from Rupel to the angle of the Struma, had, in the spring, been put before Falkenhayn by Mackensen (who still commanded, under somewhat indeterminate conditions, the forces of the Central Powers in Bulgaria and Macedonia). Falkenhayn had declined at the time owing to the risk of bringing Greece into the ranks of the enemy. Now, however, it seemed safe to ignore this danger, and desirable to forestall the relief offensive that would doubtless accompany Rumania's intervention, 3 and on Aug. 17 a series of encounter-combats began between Sarrail's various groups, advancing for their deployment on the frontier, and the wing elements of the enemy. In the centre, the 17th Colonial Division, the British assisting to some extent, 4 took, lost and retook Dodzelli (Aug. 17-8). But on the left the French cavalry group, already mentioned, which was operating E. of the Struma bend, was driven in by a serious Bulgarian movement from Rupel and through the mountains from the Nevrekop region, and, had it not been that the Bulgarians used part of their forces in taking possession of the coveted coastal strip of Kavalla, the Struma line itself might have been forced. As it was both the French cavalry group and the British force further down the river were able to establish a sufficiently strong defence of the river. On the other flank the Bulgarian attack encountered the Serbians in the process of concentrating forward.

The new Serbian Army, commanded by the Prince Regent Alexander, with Boyovich as his assistant, was organized in three weak " Armies," the I. Army under Mishich, the II. under Stepanovich, and the 411I. under Yurichich-Stiirm, who was shortly afterwards succeeded by Col. Vasich; of these the I. was in touch with the left of the French 122nd Div. about Lyumnitsa, the II. on its left, and the III. formed the left wing, advancing methodically and by short stages towards Banitsa. On the 17th advanced elements of the Danube Div. (III. Army) were driven out of Florina, and on the 18th a hasty counterattack on that point failed. The Bulgarian Army developed considerable strength (6th and 8th Divs. I. Army) and on the r9th, the Danube Div., attacking again, was flung back a considerable distance to beyond Banitsa. Meantime the II. Army, working up in the Moglena district, repulsed such attacks as were made on it, and continued its deployment in front of the Moglena mountains, the left directed on Kaimakchalan, and the I. Army, between the II. and the French left at Lyumnitsa, remained undisturbed.

3 A new military convention between the four Central Powers had provided that, in case Rumania declared war on Austria, Bulgaria and Turkey would do so against her.

4 On the 18th the French divisional commander asked for British aid to secure his flanks and enable him to hold what he had won. But, Rumania having refused to declare war against Bulgaria, Milne, having regard to War Office reservations, declined and appealed to Sarrail to refrain from putting him, as a soldier, in the impossible position of being an inactive witness of enterprises that had no chance of success without his assistance. Sarrail, however, says that Milne had promised, before the movement on Doiran began on Aug. 10, that although he was not authorized to take the offensive, he would not leave the French with their flanks in the air. Taking these two pieces of evidence together, the only conclusion possible is that the formula defining Sarrail's authority as commander-in-chief was too limited to be of much practical value in ensuring military unity, yet too extended if the Governments desired to preserve their control of policy. Too much was left to interpretation, and the commander-in-chief was obviously exposed to the temptation of planning his operations so as to create the case for the promised assistance. Indeed Joffre's directions of July 15 contained a personal instruction to Sarrail in this sense (see Sarrail, p. 361, telegram 4977/ M., 4979/ M, and especially 4980/M pour le General seul). The story of the Salonika campaign can onl y be understood by bearing in mind the political and personal undercurrents affecting it.

Impressed by the attacks on his flanks and in particular that on his left, and limited by the outcome of the Rumanian negotiations to fighting without ulterior purpose - instructed, further, by Joffre (apparently on the 16th) to attack three days after the signature of the Rumanian convention, viz. on the 10th - Sarrail changed his plan. Everything E. of the Vardar was to be defensive, everything W. of it offensive. On the extreme left irregular bands (which Sarrail had formed or subsidized) were to cut communications between Florina and the S., and to work their way into southern Albania, where cooperation had been promised by the Italian troops at Valona. An improvised French brigade was to move over to support the left of the Serbians by an attack round the S. and W. sides of Lake Ostrovo. The Serbian left was to hold up the Bulgarian advance, and the four remaining divisions to attack Kaimakchalan and the range to the E. of that point, supported by part of the French 122nd Div. in front of Lyumnitsa. East of the river, the remainder of the 122nd Div. was to stand fast, the 17th Colonial to attack W. of Lake Doiran and the 57th to demonstrate E. of it, all available British artillery participating in the effort of the 17th Colonial. The British divisions in this region (22nd and 26th) were to follow, and the Italian Div. of Gen. Pettiti, in process of disembarkation, was to relieve the 57th French Div., which was then to be transferred to the left wing. East of Dova Tepe the defensive front was to be under the command of Gen. Briggs and held by the British 28th Div., the French cavalry group (which was to explore on the further bank as before) and part of the British 27th Div., the remainder of Milne's Army being in reserve.

At this date, according to a French parliamentary paper, Sarrail's total force, combatant and non-combatant, inclusive of all details, had a ration strength of about 350,000 (four French, five British, six Serbian, and one Italian division, and a Russian brigade), and a combatant strength of 145,000 rifles, 3,000 sabres, 1,300 machine-guns and 1,032 guns, of which, however, 36,000 British infantry were not available for general service, and the 11,000 Italians only began to disembark in the last days of August.

During the regrouping process the Serbian left was driven back slowly, but in order, and on the 23rd the Vardar and the Danube divisions were about the N.W. corner of Lake Ostrovo with part of Timok, and the French Provisional divisions and the French 122nd Div. were engaged in various partial combats on the line Kaimakchalan - Lyumnitsa. On the Doiran and Struma fronts small engagements were frequent, and the French cavalry group, which again attempted to operate E. of the Struma, was driven in by superior forces with somewhat heavy losses. The Bulgarian attacks, however, as has been said before, had no more serious purpose than shortening the line and sketching out a sort of preventive attack, and they died down about Sept. 1, at which date the position of the Serbian left was practically the same as it had been on the 23rd, while the Struma front was unmolested.

Meanwhile the Rumanians had come into the field, and Bulgaria had declared war upon them. In a directive of the 24th, therefore, Joffre ordered Sarrail to continue to check the Bulgarian advance, thereby fulfilling his mission of protecting the Rumanian deployment, and to prepare for a counteroffensive, the date and objective of which was practically left to Sarrail's discretion. For this offensive Sarrail relieved the 57 th Div. from the Dova Tepe front and the 122nd from the Vardar valley front, the Italians replacing the 57th, and the British divisions already on the spot taking over the whole front from Doiran to the Vardar. This enabled a group to be formed under Cordonnier consisting of the French 57th and 156th Divs. and provisional brigade, the Serbian Morava Div.' (released from the Lyumnitsa region by the 122nd), the Russians, and the various irregular bands above mentioned. As soon as this group should be ready the offensive was to be launched against the right flank of the enemy's new line. This ' In the sequel this division was taken to support the Gornichevo attack, and did not join Cordonnier.

was carried out on the whole front on Sept. ro. On the Struma front (Briggs) six British detachments, and another French cavalry group also, were sent over the river at different points to engage and hold the enemy in that quarter. No permanent foothold was gained or sought, and the troops were withdrawn on the r r th. On the Doiran - Vardar front, after heavy bombardments on the 11th, 12th and 13th, a salient in the position of the German ioist Div. was stormed on the night of the 13-14, but given up under counter-attacks on the following day. The French 17th Colonial Div. and the Italians made similar local attacks, and from time to time other coups de main took place as well as patrol activity and aerial bombing, with the object of detaining as many troops of the German XI. and Bulgarian II. Armies as possible, while the main attack was being developed W. of the Vardar. This began on the 11th, both for the Serbians and for the Cordonnier group. The former pushed up towards Kaimakchalan and Vetrenik mountains with three divisions (aided by the French 122nd Div. which carried 111ayadag) and grouped the Vardar, Danube and Morava divisions for the attack of the mountain pass of Gornichevo W. of Lake Ostrovo. The Cordonnier group, much hampered in its concentration by road difficulties, and by the piecemeal arrival of the formations composing it, advanced to Rakita and Hill 633 on the KayalarBanitsa road on the 11th, with the Russians, as flank guard at the mountain pass of Vlachoklisura. On the 12th Cordonnier's main body reached Rudnik, the Russians the mountains 1414 and 1348 N. of Vlachoklisura, and a Serbian liaison group occupied Sotir. On the 13th there was fighting along the whole front from Kaimakchalan to the Russian positions, Cordonnier gaining a line at the foot of the Malareka ridge, and on the 14th the Serbs broke through the hostile line at Gornichevo, capturing 32 guns.

Unhappily, in these operations an acute difference arose between Cordonnier and Sarrail. The former, the man on the spot, conducting his offensive on the methodical lines of the battles of France, from which he had recently come and in which he had played a brilliant role, moved too slowly to satisfy Sarrail, who, released at last from all restrictions of higher policy, was determined to signalize his name and silence his opponents with a first-class victory. As to which was in the right it will be for history to say; probably it will be found that this is no more than one of the incidents between a higher command in a central command post and a subordinate command in the field that are so common and indeed inevitable in moments of crisis. But the peculiar factors of this case, personal and international, gave the incident a lasting importance. In the event, the Bulgarians, broken by the Serbians at Gornichevo, were able to retreat across the front of Cordonnier's force and reestablish themselves on the line of the Brod, blowing up in their retreat the important viaduct of Ekshisu.

The battle now entered upon a second phase, which lasted from Sept. 15 to Oct. 30. In this period small actions on the Struma and Doiran front were continued; thus the British sent several detachments over the Struma on the 15th and 23rd, and a more serious move was made on the 30th, when Gen. Briggs initiated a methodical advance which brought him by Oct. 8 to the line Agomali - Elishan - Ormanli with advanced elements along the Belika stream facing Seres. On the Doiran front several local enterprises were carried out by the British and Italians. Meantime, on the Adriatic coast, the Italians were moving onwards from Valona along the Voyusa valley, Premeti being occupied on Oct. 9.

The practical effect of these holding attacks and demonstrative moves on the main battle-front between Vardar and Brod seems, however, to have been much smaller than it had been in the first stage. According to Sarrail twenty Bulgarian battalions were shifted over from the British front to that of the French and Serbians in the last days of September. This was only to be expected. When the real front of battle became definite, merely potential battles ceased to possess effective binding power. But the Struma front, though militarily eccentric, had possibilities in the political sense; an advance on that side threatened the Bulgarian occupation of the Kavalla region almost before it was established - for the Greek garrisons were still in being. General Milne therefore chose this front on which to exercise such holding action as was possible in the conditions.

On the main battle-front Cordonnier's advance on the left of the new Bulgarian position continued on the 15th and 16th, the Russians in the mountains, combined with a newly constituted French force 1 working round, and the main body moving directly on Florina and the line of the Brod. On the 17th Florina was taken by parts of the 57th and 156th Divs.; on the r8th Sarrail intervened again to press the advance towards the rear of the Bulgarians opposing the Serbians on the Brod and Kaimakchalan front. But, probably from supply difficulties due to the blowing up of the railway viaducts, but also because Cordonnier thought it necessary to clear the mountains on his left before swinging in, the advance came to a standstill again on the 19th. On the 10th the Serbians made an important advance in the region of Kaimakchalan, and held their gains against counter-attack. The Franco-Russian Div. on the extreme left also maintained its positions, and on the 23rd Cordonnier was ready to attack in concert with the Serbians along and W. of the Brod line. 2 But, meantime, reinforcements were arriving on the Brod from the British front, where the policy of demonstrations had at last been detected by the enemy. On the 24th and following days, while the artillery of Cordonnier's force was being got into position for the attack of the Brod line, and the mountain country W. of Florina was being cleared, the Bulgarians made fierce counter-attacks on the Serbian positions about Kaimakchalan. That point itself had been stormed by the Serbians on the loth, and their hold was not shaken, but it was the 29th before the whole area was in their possession. On Oct. 3 a concerted attack of Cordonnier and the Serbs was delivered on the whole front from Kaimakchalan westward, while the British on the Struma front developed the holding-offensive already mentioned. This general attack was initiated by a vigorous Serbian push in the mountains W. of Kaimakchalan, and, finding their line turned on its left, the Bulgarians fell back, with only rear-guard resistance, to a position defined by four geographical points: German on Lake Prespa, point 1906 in the Baba ranges, Kenali in the plain, the extremity of the Cerna bend and the ridge N. of Kaimakchalan.

This closed the second phase of the battle. The Bulgarians were being effectively held, at the least. As a relief offensive, coincident in date with a new Russian offensive on Brussilov's front, with the seventh Isonzo battle on the Italian, with renewed efforts on the Somme front, and with the battles of Hermannstadt and Dobrich on the two Rumanian fronts, the Balkan battle in its first two phases accomplished all that could be expected of it. But from the point of view of the Salonika forces it was disappointing, especially to the ardent Serbians who at Kaimakchalan had reentered their own country.

The next phase was the general advance against the KenaliCerna bend line. It occupied the period Oct. 4-1 9, and was marked by even more internal friction than the previous phase. The operations themselves consisted in artillery bombardments and local attacks, in which the Serbs advanced to the Cerna bend (7-8 Oct.) and over the river and on to the plateau N. of Velyeselo (Oct 18-19). Otherwise no important progress was made, and, indeed, a set-back occurred on the 6th, when a French and Russian attack W. of Kenali failed with somewhat heavy loss. General Cordonnier thereupon reverted to the idea of a wider turning movement, but Sarrail, who believed, from the reports of his agents and from incidents on the Serbian front, that the whole Bulgarian Army was ripe for a moral collapse, insisted vehemently on direct action. The frontal attack was carried out on the 14th, for the first time with the aid of gas shell, but it failed, with heavy losses to the French and particularly to the Russian contingent. A violent interview on the battlefield between Sarrail and Cordonnier, in the presence of many Allied officers, completed the disorganization of the 1 Henceforward called the Franco-Russian Division.

2 There was some overlapping of the Serbians and French about Boresnitsa, due to a Bulgarian counter-attack on the 1gth.

command. Cordonnier, who was suffering from almost mortal disease, was sent home a few days later, being replaced for the time being by Gen. Leblois, but for some days the Serbian attack on the Cerna bend went on unsupported, until Sarrail took the step of placing the available French forces under the superior command of the Serbian Gen. Mishich with the idea of restoring moral as well as tactical unity to his dislocated offensive. The necessity of continuing it, and the chances of a striking success if it were persevered in, were equally evident. The Bulgarian moral had suffered from stagnation and from the rude shock of meeting an enemy risen from the grave. The German elements of the XI. Army had mostly been withdrawn, 3 its commander von Winckler (who succeeded Gallwitz in July) was in early October no longer under Mackensen's orders but under those of the Bulgarian higher command; and Gen. Otto von Billow, summoned with the staff of the VIII. Army from Lithuania to take general charge of operations in Macedonia, only arrived in the middle of the battle. No higher purpose than gaining time and avoiding defeat was or could be contemplated on the part of the Germans and Bulgarians, and self-sacrifice in such a role requires either the constant moral of a professional army, such as that of the old regular divisions of Milne's command, or the crisis-moral of the citizen soldier. In these conditions, although the lateness of the season and the sharp enemy offensive in the Dobruja made a junction with the Rumanians impossible, Sarrail had reason to hope for an important victory. In any event, the turn of events in Rumania (Hermannstadt, Fogaras, Kronstadt), and on the Italian front (eighth and ninth Isonzo battles), dictated a continued offensive towards Monastir.

For some days after the Serbian success at Velyeselo weather and counter-attack prevented their further progress, and this time was utilized by Sarrail in strengthening them with French troops drawn both from the Vardar (where the British took over their line), from the western mountains (where the Italians 4 relieved them), and from the centre about Kenali, where the weather flooded the country. At the same time the artillery was established in position, the maintenance service organized to supply the battle, and touch gained on the extreme left with the Italian force advancing in Albania, 5 while Milne's Struma troops made a strong diversion by attacking and carrying Barakli Dzuma (Juma) on Oct. 31. Finally, on Nov. ro, the Serbians with French support opened a new strong attack in the Cerna bend. Between the Toth and the 13th Polog, the ridge to the E., and Iven were taken, while the artillery on the Kenali front kept up interdiction fire to prevent enemy reinforcements from crossing the Cerna; and on the 14th a general attack brought the Serbians to the ridge on which Jaratok Monastery stood, with lesser advances to right and left. The Kenali lines being now completely turned, the enemy on the night of 14-15 fell back, covered by rain and snow, to a position but little in front of the line Peristeri - Monastir - height r378 - Grunishta. But this time the advance did not halt. On the 15th, in terrible weather, the French left pushed forward to the Viro brook; in the mountains, though mud was absent, rain and snow were still more violent, but the Serbians continued to progress. On the 17th Mishich's troops, with the French aiding and conforming on their left, captured height 1212, and the 18th they stormed 1378, pursuing to Makovo in the night following, while the French centre and left forded the Viro, the Russians S. of Monastir took Kanina, Zabiani and Holeven, and the Italians, French and Russians in the western mountains developed an outflanking movement right and left of Lake Prespa.

On the morning of the 19th the town of Monastir was found evacuated, and the third phase of the Allied offensive, definitely Apart from staffs, there only remained 171 battalions and corresponding artillery and army troops.

4 A newly-disembarked Italian force relieved these in their position in the Dova Tepe region.

5 Apart from the movement of the Valona force up the Voyusa, a fresh Italian contingent had landed at Santi Quaranta and advanced by the Turkish military road to Koritsa, and thence to the region of Okhrida and Prespa Lakes. This line became a regular line of communication.

victorious, closed with the occupation of a line from Krani on Lake Prespa' - Saddle 2227 - Orizar N. of Monastir - Makovoheights S. of Staravina - N. slopes of Kairakchalan.

In view of the weather victory had come in the nick of time. The force of the attack rapidly died out, even the Serbian Army with its positive and eager moral being too weary, and all formations being too far ahead of their administrative services 2 to be able to push on. Fighting went on in the Cerna bend till Dec. 12; the French operating in this quarter captured Dobromir on the 21st and height 1050 on Nov. 27, and the Serbs Staravina on Dec. 4; but the troublesome height 1248, overlooking Monastir and the plain, remained in the hands of the Bulgarians, and the town was under shellfire throughout the winter.

By this time the evident defeat of the Rumanians and the consequences of the Venizelist coup d'etat at Salonika called for a reconsideration of policy. The offensive, considered as a relief offensive on Rumania's behalf, had completed its usefulness, and on Dec. rr, Sarrail was instructed to establish his forces on a defensive line, holding as much of the regained territory as possible and keeping in mind the resumption of the offensive at a date that was to be fixed according to the general plan of campaign for 1917, and particularly according to the date at which the Rumanian Army, reorganized by Gen. Berthelot, should be ready for the field. Meantime, the contingency of rapid action against Greece was to be provided for.

Henceforward, for some months, it is this last element in the problem which governs the action or inaction of the Salonika Armies. In answer to the Venizelist coup of Aug. 30, King Constantine began a threatening concentration of troops in Thessaly, on Sarrail's left rear. To this the Venizelist forces, hardly yet in being, were no counterpoise, and Venizelos's policy, which was also that of Sarrail, and as a rule that of the British and French Governments, was to prevent contact between the two Greek parties. 3 Sarrail therefore drew a neutral zone of some kilometres width from the Koritsa basin to the sea. Nevertheless, throughout the period of his offensive, anxieties for the flank and rear limited the play of such reserves as the Allied commander-in-chief possessed. On Dec. 1 the attempt to enforce the delivery of certain war material to the Entente representatives at Athens produced an emeute, in which the Allied landing parties suffered severe losses and the Legations were for a time practically besieged. This affair brought the confusion of the Greek problem to a climax. Though on Dec. 2 Sarrail's demand for authority to advance into Thessaly was refused, and some disjointed efforts at naval and military action led to nothing, sufficient unity of policy was achieved to enable the Allies to deliver an ultimatum on Dec. 14, demanding withdrawal of the Greek forces in Thessaly.4 The policy of the Salonika forces for 1917 was fixed at the Inter-Allied conference held in Rome in Jan., to which Sarrail was summoned. The first conclusion was that Greece was not to be attacked, both for other reasons and especially owing to the effect that any high-handed action would have on American sentiment, which was at that moment in the crisis of deciding The French on the W. side of Lake Prespa occupied Liskovats, a village on the same latitude as Krani, and then Horesovo.

2 The viaduct at Ekshisu was not repaired till Nov. 27.

The first battalions organized by the Venizelists were sent to the British front on the Struma. At one time the idea seems to have been entertained of reconquering Eastern. Macedonia (DramaKavalla region) by means of an advance of these troops, presumably in concert with a British expedition. Nothing, however, was done in this direction. The regular Greek forces on this side, cut off by the Bulgarian invasion, had accepted an offer of internment as " guests of the German Government " (Sept. 24) and been transported to Germany, except a small force which from Seres had made its way to join the Allies.

4 In the midst of these events came mutterings of trouble in the Serbian Army. Little is known as to the details, but as early as the beginning of Jan. 1917 a division commander and several brigade commanders and senior staff officers were deprived of their posts; by mid-March over twenty such officers were in prison at Bizerta charged with plotting against the Prince Regent. Finally, a number of these were brought to trial, some being condemned to death and (July 29) shot, others sentenced to 15 years' imprisonment.

for peace or war. The second was that operations on the front should be governed by a waiting policy - a compromise between Sarrail's demand for forces sufficient for a grand offensive, and Robertson's and Cadorna's demand for a withdrawal - and that the divisions sent or to be sent to Salonika should be liable to recall, in the event either of a crisis on the Italian front, or of one in Palestine. On the other hand the full subordination of the British contingent to the Allied commander-in-chief was agreed to, 6 and (according to M. Mermeix) Sarrail's personality made so favourable an impression on Mr. Lloyd George that the agitation for his recall came to an end for the time being.

In the first weeks after the end of the battle and these decisions, the Salonika front was for the first time formed as a regular and continuous system. On the right, as far as the Vardar, were the British under Milne, consisting now of the loth, 22nd, 26th, 28th, 29th, and both Divs. (the last newly arrived) with some organized units of the Venizelist forces in support. On the left of the Vardar the national contingents were more intermingled, the Lyumnitsa region being held by the French, the Kaimakchalan - Vetrenik - Duditsa front by the Serbs (6 Divisions), ? the Cerna bend by the Italians (12 Divisions), and the remainder of the front by the French main body, to the command of which Gen. Grossetti was shortly appointed in succession to Cordonnier. The French Army were now being reinforced, and included besides the old 57th, r7th Colonial, 122nd and r 56th, the r6th Colonial (arrived early Dec.), 76th (arrived end Dec.), 30th (arrived Jan.) and rrth Colonial (newly formed from odd units on the spot). In all, therefore, there were no less than 212 Divisions in this theatre at the outset of the 1917 campaign, besides two Russian brigades, and the Greek National Defence Army in embryo. The Italians in Albania are not counted, as they were not under Sarrail's command, and operated entirely as a separate force in their own theatre, interesting the Salonika force only in so far that their positions protected the opening up of the line of communication Santi Quaranta - Koritsa - Lake Prespa. This force was strong enough to hold all its line and yet to have larger reserves than in 1916, for either an offensive northward or a march into Thessaly.

The Bulgarians on their side established themselves solidly on the final battle line of 1916, and improved it on Feb. 15 1917 by the recapture of height 1050 in the Cerna bend from the Italians, who had only shortly before taken up this sector. An attempt to retake it failed on the 28th. The end of active operations against Rumania about the middle of Jan. released additional Bulgarians for the Salonika front. As a beginning Sarrail sent his newly arrived 76th Div. inland to Koritsa 8 just in time to prevent the seizure of that point by an Austrian force coming down from Pogradets (early Feb. 1917); the division then moved on to clear the Santi Quaranta road - the new alternative line of communications - of bands and make the junction with the Italians in Albania, which at present only existed by wireless and aeroplane. This mission was completed by the meeting of the French and Italian troops at Ersek, Feb. 17, of ter which the limit between Sarrail's International Army and the purely Italian force of Gen. Ferrero in Albania was fixed at a little N. of that point. The 76th Div. was then 6 The idea of a Salonika front had never had any real support in the British War Office, and the Italian point of view seems to have been that the abolition of an Inter-Allied force in the Balkans would give Italy greater freedom of action in Albania.

6 Sarrail had already been released from control of the French G.Q.G. when Joffre left it and was replaced by Nivelle (Dec. 1916). ' Owing to losses the Serbian Divisions were now reduced from the old high establishment and were equivalent in rifle strength to French Divisions, viz. 9 three-company battalions.

8 Here, in the previous summer and autumn, French cavalry had operated to disperse irregular bands that were practically bandits, and to intercept communications between Athens and the enemy. For a time Venizelist sympathizers reigned at Koritsa, and the idea of annexing the district to Greece was not far in the background; but in the face of protests from Italy and from Essad's Albanian party at Salonika, Sarrail replaced them by an autonomous government of inhabitants under French protection, a course which he considered would commit nobody and yet give no excuse for the formation of Royalist bands.

reassembled N. of Koritsa, in readiness to participate in the general offensive when this should be ordered.

Meantime the British advance beyond the Struma, originally developed out of the demonstrations made during the first stages of the Monastir offensive, had been consolidated on the line Jenimah - Osman Kamila - Hrostian Kamila - Nevolyen - Kukuluk - Elishan - Barakli Dzuma - Alipsa (Ali Pasha). This position formed a useful bridgehead in case Sarrail should be able to resume the old scheme of an offensive by way of Demir Hisar and Rupel against the Bulgarian communications, and meantime it immobilized a Bulgarian force of approximately equal strength. But the supply of the lines beyond the Struma was difficult and laborious during the winter months, and it was to be expected that the valley would be a hotbed of disease in the summer. On the Doiran front several small local attacks were made on the enemy's positions on and about Dub mountain - the Achi Baba or Hermada of this front.

The scheme which Sarrail intended to carry out was the old scheme of a principal blow from the mountain front in the direction of Demir Kapu, cutting off the defenders of the Doiran and Lyumnitsa fronts. Combined with this, local and subsidiary advances were to be made on the Cerna bend and Monastir fronts, with, to the left of all, an enveloping movement by the 76th Div. from Koritsa by Resna upon the rear of the Bulgarians N. of Monastir. Meanwhile, the British, after exercising a holding action which was to increase in intensity from left to right and culminate in a real attack on Seres, would quickly shift their centre of gravity to Demir Hisar, and thence, by Rupel and the Struma, force their way to a position on the rear of the Bulgarian line of retreat. If, as in 1915, Sofia was to be the ultimate objective - if, in short, the campaign was intended to be a decisive one - the sweep of the extreme left would pass through Uskub and Kumanovo, that of the left centre through Veles and Shtip, while the Italians bore down on Demir Kapu and the British marched on Simitli - a romantically complete success. In many ways, indeed, the scheme is identical with that eventually carried out by Franchet d'Esperey in Sept. 1918. But in both cases there were four postulates: - (a) Greece friendly or under entire control; (b) synchronization of the offensive with those in the main theatres, so that the enemy could not transfer German reinforcements in time to the Balkans; (c) enough men; and (d) enough material. Of these postulates Franchet d'Esperey was in the sequel to possess all four. Sarrail, on the other hand, though (b), the most important strategically, was within his power, had little real security as to (a) and a definite lack of means in respect of (d), and this last ranked as one of the most important factors for the tactical break-through on which all the rest depended. Several of these requisites being absent, Sarrail was under no illusions as to the chances of pushing success to the gates of Sofia. But he considered that there was a possibility, becoming a probability if he were given more heavy guns, of forcing the offensive of his left centre in the mountains so far north-eastward as to compel the evacuation of the Lyumnitsa - Doiran - Belashitsa front. In the absence of these additional heavy guns, however, the tactical break-through in the Moglena - Cerna region would be a matter of great difficulty, and Sarrail accepted a proposal of Milne that the British holding attack should be delivered on the Doiran front - where its effect on Sarrail's battle would be immediate and tactical - instead of towards Seres. This change of plan, made between Feb. 9 and 19 was due in the first instance to the difficulties of movement in the Struma valley, where exceptionally bad weather had made motor transport almost impossible. But the decision was one of grave importance since the height of Dub, now chosen for the British attack, was one of the strongest points on the whole front, and it was an attack, and not a demonstration, that was intended.

In this form the scheme was accepted by the Inter-Allied conference that met at Calais on Feb. 27 - in circumstances of great tension - and Sarrail was ordered to hold the enemy to his ground, and authorized to seize any opportunity which offered itself to inflict real damage. So far this represented XXXII.-12 merely the waiting policy of the Rome conference. But an important modification of that policy was implied in the actual instructions to Sarrail, for not only was he told that the plan of operations had been agreed to by the British War Council, but. he was instructed to be read y to launch his offensive about April 15 - a few days after the date selected for Nivelle's battle on the Aisne - in spite of the fact that no Russo-Rumanian offensive could be expected so soon. That authority should be given for an attack of any kind in the conditions of the time was of doubtful wisdom. But that the conference sanctioned an attack on the strongest parts of the enemy front in order to obtain the very limited results at which it aimed is a fact that is more than difficult to explain. The strategic results that would or might accrue from a break-through in the Moglena region were indeed so great (as 1918 showed) that special tactical risks might fairly be taken to achieve this break-through. But the conference of Calais not only had no great strategic results in mind, but had expressly agreed that the circumstances of the time excluded them. And when, in addition, we find that Sarrail was, at one and the same time, told that the decisive defeat of the Bulgarians was excluded, and instructed to " take " - according to the Calais proces-verbal - or " seek out " - according. to the French Ministry of War telegram- " every favourable opportunity for inflicting a serious check," we can only conclude that the Calais conference was so preoccupied with troubles nearer home that it could think of nothing better, as regards Salonika, than to shift the responsibility and trust to luck.

But before the offensive period arrived events had already blurred the scheme. The British were to attack the Dub, and demonstrate at other points; the French, Russians and Italians in and W. of the Cerna bend were then to move, and lastly the Serbians were to deliver the principal blow between the Cerna and the Duditsa mountain, E. of which were the French 122nd Div. and part of the Greek " Archipelago " Div. (the first of the Venizelist National Defence formations), ready to drive forward on the Huma - Vardar front so soon as the flank attack had begun to make itself felt. But the Serbians, still angered by the breakdown of joint action in the Monastir battle, and also seriously affected by their internal troubles, refused to move till after the Allied general offensive had opened in France. The Russians were deeply shaken by the revolution in their own country. Weather conditions were terrible, and the zero day of battle was several times postponed, both owing to these conditions and to the slowness with which the preliminary offensives at and W. of Monastir were developing.

As early as March 12 the French 76th Div. had begun its advance from Koritsa on Resna. But owing to snow and to the harrying tactics of enemy irregulars - who were supported by small Austrian forces from Pogradets - it made but little progress, and on March 19 Sarrail suspended the movement. Meanwhile, on the 18th, Grossetti, after several days of local attacks, delivered a larger attack on the front N. of Monastir, which by the 27th had advanced the line in this sector to height 1248, Snegovo and Rastani, with corresponding gains to the left on the spur of Peristeri called Crvena Stena, and to the right about Dobromir and height 1050. Over 2,000 prisoners were taken, and these gains eased the uncomfortable position of Monastir, but they were not pressed, as Sarrail not only held his reserves for the April offensive but sought to disengage part of the troops engaged at Monastir for the same purpose.

On the Doiran front the British attacks began about the middle of March with enterprises intended to secure advantageous positions for the general engagement. On April 22, after several postponements, the artillery preparation began, in terrible weather, and on the evening of the 24th the infantry assault was delivered on a two-division front against Dub and its under-features. The positions of the defence were strong in themselves and strongly entrenched, and, as in the trenchwarfare battles of the French front in 1916-17, the attack preparations and especially the prolonged bombardment had given full warning. Along the whole front the infantry reached the hostile front trenches, but only on the left were they able to maintain the ground won, and nowhere was progress made beyond the first enemy system. Heavy fighting, which only died down on the 28th, was necessary to consolidate the new line, and meanwhile for " climatic and other reasons " - to use the words of Milne's despatch - the rest of the front remained inactive. It was only on May 5 that the other secondary elements of the battle (W. of the Cerna, Cerna bend, LyumnitsaMayadag front) began their artillery preparations. At last, on the 9th - in the case of the British the evening of the 8th - the infantry went forward on all fronts. The British, attacking between the Petit Couronne and Lake Doiran, were repulsed, but they made good an advance further W., where Soo yards in depth were gained over a frontage of two miles; these gains were held and consolidated. On the Lyumnitsa front the French 122nd Div., aided for the first time by Greek forces, carried out methodical advances by fractions at a time, and even for a moment held the Srka di Legen. Elsewhere all holding attacks were repulsed after the usual momentary successes. But the serious feature of the situation was the fact that the Serbians, who were to have been the shock element of the attack and to have begun their advance on the 9th, like the rest, came to a standstill on the 11th after engaging only one of the three divisions between Kaimakchalan and the Cerna. More, they asked to be taken out of the line, and with this the offensive scheme practically collapsed. Sarrail did, indeed, induce them to engage again with two divisions, and on the 15th attacks in aid were delivered or threatened at other points.' This time, after some loss, they demanded that the offensive should be abandoned altogether. Already Sarrail had received authority from Paris to suspend operations and establish his defensive line (May 15), and on the 23rd, after the second Serbian disappointment, he issued orders accordingly. It could be claimed, with justice, that these operations had pinned the Bulgarian Armies to their ground, and indeed only one Bulgarian division remained in the Danube theatre, all the rest having been drawn into the Balkans by the developing threat of the offensive. But Sarrail had hoped for much more. His object had been Veles and Demir Kapu at the least, instead of which only slight advances on the Struma, Doiran, Lyumnitsa and Monastir fronts - without exception secondary elements in his scheme - had resulted. The offensive was, in short, a failure, and one, moreover, in which it was palpable that disunity had played the major part.

The reasons are to be sought partly in the Serbian breakdown and partly in the general war situation. Allusion has already been made to the fact that the Serbian Army had for some time been passing through an internal crisis, and to the demands made by the Serbian Government and higher command for tangible evidence that all the Allies would frankly engage their forces in the battle. The engagement of the French and Greeks on the Lyumnitsa front, and above all that of the British on the Doiran front, where attacks were made in earnest, local gains secured, and heavy losses incurred 2 in aid of the projected main attack, might have been supposed to be an effective answer to this last demand, even though the Italians and French in the Cerna bend showed no great vigour.' But in fact it is premature to enquire how far a sense of having been left in the lurch in the autumn battle and how far internal troubles respectively contributed to the Serbian refusal. In any case a strong motive at the back of all others was that expressed in the phrase, " What is the use of delivering Serbia if no Serbs are left to inhabit it? " But there were other elements of discouragement. Exalted hopes of a great general advance to victory on all fronts had been dashed to the ground by the Aisne battle and its tragic consequences, and they were followed by a revulsion in which the war-weariness, soldiers' grievances, 1 The British on the Struma took Kupri and Ernekeui on this day. s According to Sarrail the British losses in the two offensives were 8,000.

3 The Russian brigade in the Cerna bend on the contrary s'est fort bien conduit, according to Sarrail (p. 257).

and. the tremors of revolution produced alarming mutinies. The Russian contingent was affected not only by the revolution, 4 but by rumours of what had happened to their comrades in France, and they in turn affected the Serbs. When, on the main front and in the presence of the main enemy, the moral of a homogeneous army was shaken to its foundations, it was not to be expected that on this secondary front a patchwork of contingents, every one of which was exiled from its homeland, would fare any better.

In the midst of these conditions of exasperation the Greek question at last came to a head. M. Jonnart was sent out as Allied High Commissioner, and Sarrail was authorized to invade Thessaly (June io). He had already begun to prepare a force for this purpose immediately the offensive was abandoned, and, to make certain, other troops went to Athens, to Corinth, and to Itea in the Gulf of Amphissa, whence a line of supply for the Thessaly force was opened via Bralo. Except for a skirmish at Larisa on June 12 no fighting took place. The dethronement of Constantine and the succession of King Alexander, with Venizelos as his first Minister, were successfully accomplished and most of the troops withdrawn again during July.

But the settlement of the Greek problem came too late to have any influence on operating against Bulgaria. The events of the Spring had affected governmental policies as well as common men's passions. Thus, the British War Office began to withdraw troops for the forthcoming Syrian campaign; the Italian Government began similarly to press for the withdrawal of their 35th Div. for operations in Albania; and the French Government, reconstituted after the Aisne crisis, had made it a definite policy to economize the reduced man-power of France by avoiding battle. Meantime, the Italian advance into Epirus, made concurrently with the French operation in Thessaly, antagonized Greek opinion, the relations of Greeks and Serbians were little better, and those of the French and Italians at the margin of the respective spheres of occupation in Albania none too good. The personalities of the Venizelist, and those of the regular Greek Armies, now to be amalgamated, were inevitably opposed. Finally, Essad Pasha reappeared, with a national Albanian policy and the nucleus of an Albanian contingent.

Sarrail, nevertheless, attempted to maintain a certain military activity; in particular, an enterprise on the extreme left was carried out by a French force under Gen. Jacquemot,5 which captured Pogradets (Sept. 11) and thence advanced N. almost to Lin. In Oct. a further advance was made into the upper Skumbi valley, but this was suspended, owing to representations by the Italian Government, and the French then withdrew to Pogradets, leaving Essad's bands to operate in the Skumbi region independently. Finally, the whole front relapsed into practical stagnation.

On Dec. 10 Sarrail was relieved of his command by the new Clemenceau ministry on the renewed demand of the British and Italian Governments, and of General Foch in his capacity as chief of the staff of the French Army. This put an end to a situation which had become impossible. The proximate cause of his dismissal, so far as the French Government was concerned, had nothing to do with Salonika, but was Sarrail's relation to the parties of the Left, who, in the troubled summer of 1917, had become pacifist in character, and were suspected, rightly or wrongly, of dealings with similar elements on the enemy's side. 6 So far as the Allies were concerned, it was due partly to the personality of Sarrail, and partly to the insistent and thrusting policy of France in dealing with Greece. In effect, the over, It is only just to record, however, that to the end of the year they continued to take their share in duties in the line. The evidence of Sarrail, especially p. 289, is emphatic as to this.

The superintendence of this operation was the last service of Gen. Grossetti, who left Salonika on Sept. 24 suffering from poisoning, like his predecessor Cordonnier. He died a few months later.

M. Clemenceau in his explanation before the army commission late in Dec. 1917 seems to have laid stress principally on the fact that unless satisfaction was given to the Allies in respect of their complaints against Sarrail, they would not accept a French Commander-in-Chief on the western front. (Mermeix.) throw of the Constantine regime was the only logical outcome of maintaining a Salonika front at all. But between that logical extreme and the other logical extreme of evacuation, which was always the desire of the British War Office, diplomacy built up a series of compromises which satisfied nobody. It was to the indefinite and equivocal position created by these compromises, even more than to any military factors or personal disagreements, that the positive ineffectiveness of the Macedonian forces was due. It remains to the credit of Sarrail that he was sanguine in the most confused and difficult conditions and resolute in carrying the occupation to a common-sense issue. In spite of proposals to retreat again into the entrenched camp, he maintained the army on the front from which, nine months later, it was to deliver the great offensive blow for which he had always hoped. On the other hand he left the Allied Army in a state of badly shaken moral. The consciousness of ineffectiveness, the blight of malaria, the infrequency of leave, the sense of being a forgotten " side-show," the international causes of friction - all these factors told on the moral of the Salonika forces at the most critical period of war-psychology.' A period of remise en main was necessary before the troops could be employed either in this theatre or in another, and Sarrail had taken the line of defending his army against all criticism, fair or unfair. It was essential, therefore, if the army was to be pulled together, that new men should be put in power.

General Guillaumat, the new commander-in-chief, set himself to this task under favourable conditions. No important military operations were in prospect. The Greek danger had been liquidated. The atmosphere of exasperation was largely dispersed by his appointment alone. With Russia and Rumania in collapse, and the final military trial of strength obviously imminent in one or both the western theatres, not one of the three great Powers concerned was inclined to press its Balkan interests very closely. His military position, too, soon came to be much stronger than that of his predecessor. Although the British had withdrawn two of their six divisions (the 60th in June and the 10th in Sept. 1917, both proceeding to Syria), and one battalion per brigade in the four remaining (spring 1918), the French and Italian contingents remained practically unaltered (save for the withdrawal of the Russians), while the Serbians acquired a whole new division created from Yugoslav prisoners of war, and the Greeks were taking their place as a regular Allied contingent, the " national defence " divisions at Salonika (Archipelago, Crete, Seres) being augmented by those of the regular army, reorganized by a French military mission.

The question of the use of these forces had been put on one side by the central councils of the Entente. German pressure on the Western Front, beginning with March 21 and culminating on July 15, threw all other questions into the background, and when the tide began to recede it proved as difficult as ever to convince the directors of the War that good could come from a Salonika offensive, while, at the same time, it was impossible to offer the Central Powers the opportunity of repeating at Greece's expense their customary autumn triumph. No decision imposed itself and none was taken. The only event of the early summer was a brilliant coup de main with limited objective, which on May 30 carried the Srka di Legen on the left of the Vardar. In July Guillaumat, after practically completing the work of reorganization, was recalled to Paris, where an energetic governor was needed in case of a German breakthrough. He was succeeded by Gen. Franchet d'Esperey. At Paris he continued to act as advocate of, and so to speak agent for, the principles of a Salonika offensive. After much persuading he obtained for his successor authority to prepare one, but authority to begin was not given till a few days before the battle.

Meantime, operations in Albania, which for two years had been in the nature of post and police warfare, rose for a moment in the summer of 1918 to the level of major operations.

In the winter of 1917-18 the posts of the Italian XVI. Corps (Gen. Ferrero) in Albania ran along the Voyusa from the sea 1 Nevertheless it is clear that the phase of the mutinies passed off more easily than the corresponding phase in France.

to Memaliadz, where it turned abruptly N. and then N.E., facing Glava and Caf a Glava, Parasboar, Barguzyasi, and Cerevoda. Here it joined the French posts, which ran in a N. to S. line along the mountains to Golik in the Skumbi valley and thence nearly E. to Point 1704 S. of Lin on the shore of Lake Okhrida. Early in July 1918 a frontal advance of Ferrero's Italians from the Voyusa and combined by flank pressure by the French 57th Div. in the mountains of the Devoli regions, forced the two Austrian Divs. (47th and 81st) of Gen. KdnnenHozak's XIX. Corps to evacuate the whole Berst region in haste, with a loss of nearly 3,000 prisoners. By July 20 the Italians lay along the Semeni and the lower Devoli from the coast to Petrohaudi (with a bridgehead in front of Fieri), and thence in an E. to W. line to the foot of the Mati Siloves range which was held by the French. The positions of the latter formed a marked salient, the apex of which lay at the confluence of the Holts and Devoli rivers, and the right flank of which passed by Kumichan to Golik on the old front. The importance of this salient lay in the fact that it kept the right wing of the general line echeloned well forward, threatening the rear of the enemy's lines near Berat and ultimately the connexion between Elbasan and Lake Okhrida.

In August the Austrians, now commanded by Generaloberst ', von Pflanzer-Baltin and reinforced by the 45th Div., as well as by the fresh 12th Bulgarian Div. in the Okhrida - Skumbi sector, began a counter-offensive all along the line. Between August 20 and 24 they recaptured the line of the Semeni and the Devoli and drove back the Italians to positions only slightly in front of Fieri and Berat. On August 24 those towns fell again into their hands, and the Italians then withdrew to a line from just S. of Fieri, along the Janitsa, S. of Berst, and along the Osum to Mt. Tomor. Meantime, the French, the left rear of their Devoli salient being thus threatened, had had to fall back in the Devoli and Tomorica valleys to regain touch with the Italians at Mt. Tomor, while still holding on to their posts between the Devoli and the Skumbi. These operations are of interest as being the last military success won by forces of the Central Powers in the War. Three weeks after their conclusion the Bulgarian front was in ruins.

At the close of these operations and the eve of the final act on the Salonika front, the numbers and positions of the forces of the Central Powers were approximately as follows: In Albania, XIX. Austro-Hungarian Corps (45th, 47th, 81st Divs.) under Generaloberst von Pflanzer-Baltin; between Skumbi Valley and Lake Okhrida the 12th Bulgarian Div.; from Lake Okhrida (exclusive) to Koziak Mountain (inclusive), the XI. German Army (Gen. von Steuben), consisting of, from right to left, the Lxii. German Corps staff, with under it the Bulgarian 1st, 6th and Composite Divs., and some few German and Austrian units, the LXI. German Corps staff, with the 302nd German Div. (staff German, troops Bulgarian), the 4th Bulgarian and the 2nd Bulgarian Div., and the 3rd Bulgarian Div.; from E. of Koziak to Lake Doiran (inclusive) astride the Vardar, the I. Bulgarian Army (Gen. Nerezov), three Bulgarian Divs.; from Lake Doiran (exclusive) to Lake Tahinos (inclusive), the II. Bulgarian Army (Gen. Lukov), three Bulgarian Divs.; from Lake Tahinos to the mouth of the Struma, thence eastward in coast defence positions, the IV. Bulgarian Army (Gen. Petrov), two Bulgarian divisions. The whole of these forces were under the control of the German headquarters, which had formerly been Mackensen's and Billow's but was now the Armeegruppe Scholtz; Gen. Scholtz was, however, for certain purposes under the control of the Bulgarian higher command, at Sofia, over which Todorov presided. An exact account of the forces of Bulgaria at the crisis can hardly be given, 2 but, save for one division in Rumania and some coast 2 Some of the newer divisions had an irregular constitution whereas the older ones still retained the six-regiment organization of 1915, and some of these even the old four-battalion regiments. Moreover, at that moment movements were in progress for replacing 15 German battalions by 45 Bulgarians; the Germans had departed, all but three battalions, but few of the Bulgarians arrived in the XI. Army in time.

defence units in Thrace and E. Macedonia, the whole of Bulgaria's mobilizable force was present between the Devoli in Albania and the mouth of the Struma. In the depots were practically only the 1919 class recruits, just called up; 55,600 men had fallen in battle since 1915, some thousands were prisoners, many thousands had died or been invalided. Thus the ration strength of the Bulgarian Armies on the Macedonian front can hardly have exceeded 420,000, with some 8,000 Germans (principally Jager battalions, heavy artillery and mountain machine-gun detachments). The combatant strength may be taken roughly as 310,000 Germans and Bulgarians. The Italians of Ferrero and the Austrians of Pflanzer-Baltin, whose operations in Albania were entirely unconnected with the battle on the main front, are ignored in this calculation.

On taking up their battle grouping the larger formations of the Entente Armies were arranged thus: Devoli valley to Staravina (E. of the Cerna bend), the French Army of the East, in seven French, Italian and Greek Divs.; Staravina to Nonte (exclusive), the Serbian I. and II. Armies under Boyovich (3 Divs.) and Stepanovich (5 Divs.) respectively, the latter consisting of the Strumaja, Timok, and Yugoslav Divs., and the French 122nd and 17th Colonial Divs.; Nonte to Lyumnitsa, Gen. Anselme's group, Greek and French Divs.; Lyumnitsa to Dub (exclusive), astride Vardar, 26th and 27th British Divs.; Dub to Dova Tepe, 22nd and 28th British and two Greek Divs.; Dova Tepe to mouth of Struma, 3 Greek Divs.; Salonika, r Greek Div.; in all, 28 divisions. In addition there were a small Serbian cavalry Div., a French cavalry brigade, and other lesser formations and details. The ration strength of these forces, after making allowances for men invalided during the hot months, was about 550,000, and the combatant strength perhaps 350,000. In sum, then, there was no great numerical disparity between the opposed forces. The equipment of the Entente Army in heavy artillery was, however, considerable. Sarrail's repeated requests for an adequate equipment in this respect had been in the end met, just before his recall. As to the number of heavy batteries on the other side accounts differ, but in any case the regrouping of guns to meet an attack, in the movement conditions of Macedonia, would be difficult if not impossible for the defence, and the mere possession of a reserve of such artillery was, in the conditions, a big element in favour of an attack, provided the duration of the artillery phase of it was reduced to a minimum. On the other hand, to oppose to the powerful material of attack, the Bulgarians had all the advantages of natural and artificial strength of position. But the success or failure of an offensive would turn more on moral than on material factors. Presuming that the tactical break-through was possible, would the Bulgarians knit up again on a new line further back? Or would it be found that the third winter had completed the war-weariness on which Sarrail had twice vainly counted, so that one heavy blow would finish the matter? In the light of events the answer is easy. The new government in Bulgaria was pacifist in character. Ludendorff regarded a breakdown as almost certain and detailed several divisions from the eastern front early in Sept. to proceed to Bulgaria and Serbia, the nearest indeed being ordered to Sofia to keep order. On the front itself, according to German accounts, conditions were " indescribably pitiful," and for food the greater part of the troops were dependent on ox-transport from a base 60 m. distant, and upon what they had themselves grown behind the lines, this last fact indeed making it almost impossible to induce a unit to quit its sector for a concentration. But 1917 had for the Entente been a tragedy of disappointments founded on optimistic estimates of the enemy's broken moral and vanished man-power, and in 1918 a not unhealthy scepticism prevailed in their intelligence staffs, although it was agreed that Bulgarian moral was low. Then, too, there was the question of Serbian moral. Since the deadlock of May 1917 the Serbians had consistently followed the principle that it was useless to rescue Serbia if no Serbs were left to inhabit it, yet they alone possessed the fire and passion which would convert tactical victory into strategic triumph, who would disregard food and rest sufficiently to exploit success by leaps and bounds,. who would not stop short of the Danube. And nothing less than victory without remainder would be of any value to the Entente.

When the plans were under consideration this latter question was unexpectedly answered by the Serbs themselves. They proposed an offensive with limited objective in the Moglena sector, i.e. that lying between Kaimakchalan and the Srka di Legen. At once the situation was cleared up. If they were ready to carry out an attack on the mountain front for no more than a limited objective, it was unlikely that an offensive from that group with the objective of Belgrade would not fall to pieces as in 1917. On June 29 the plan was definitely fixed. and about July 7 the Serbian headquarters (now directed by Mishich) agreed to engage all its forces in the operation. The date of the offensive (which needed a good deal of material preparation as the Moglena sector had never been equipped for battle) was fixed for Sept. 15, though, owing to objections and preoccupations in Paris, the higher command did not give leave to carry it out till little more than a week before that date (Sept. 4).

On Sept. 14 a bombardment opened which, less overpowering than those of the western front, was far heavier than anything previously witnessed in the Balkans. Next day - the original date - the offensive was launched. From the Lechnitsa river to the Sokol mountain the I. Serbian Army (Boyovich), consisting of the Danube, Drina and Morava Divs., had a frontage of 5 kilometres per division. From Sokol to the Suchitsa brook the II. Serbian Army (Stepanovich) formed the break-through force. In front of the Dobropolye were the two French assault Divs. with the Shumaja Div. on their right, and the Exploitation Divs., Timok and Yugoslav behind them. The I. Army and the Exploitation Divs. were not to move till the French had carried Sokol, Kravitsa and Vetrenik heights. Then, passing through, the Yugoslav Divs. were to master the Koziak, and the II: Army front was to advance at the utmost speed on Gradsko, flankguarded by the I. Army on the left and the Timok Div. on the right.

The attack succeeded according to programme. It continues to be a matter of controversy whether the Bulgarians offered an earnest and fierce resistance. The impression left on the Allied infantry was that they did so; the Germans in their midst assert the contrary. In any case, it is probable that the backbone of resistance was the German mountain machinegun detachments which were dispersed in the battle zone. Be this as it may, the French Assault Divs. carried their objectives by the evening of the 15th, the Yugoslav Div. passed through them in the night of the 15-16, carried Koziak on the 16th. and on the 17th drove a deep wedge, of which the point was armed midway between Gradsko and Demir Kapu, and the lengthening left flank along the Cerna was taken up by the I. Army. On the other side there was little or no tactical handling in the ensemble. Machine-gun groups and specially resolute parties of riflemen in broken ground constituted the whole resistance. This, indeed, was the typical form of defence in all theatres of war by 1918, but here it lacked the essential element of organized counter-attack. Here and there a unit turned upon its pursuers, but in the main the rear of the enemy's position was void of reserves, although it is said that only 12,C00 Bulgarian and German infantry were in line on the front attacked. Later, all attempts at reinforcement always failed, as the troops concerned had to traverse instead of following the valleys and ridge-systems. In sum, by the 19th, the wedge had developed two horns, of which one, following the Cerna, was at the outskirts of Kavadar, and the other was engaged in rolling up the front opposed to Anselme's group.

Meanwhile, on the 18th, the British and Greeks of the Doiran front, under orders from Gen. Franchet d'Esperey, assaulted the whole enemy line from the Vardar to beyond Lake Doiran, with the general idea of cutting the communications between the Bulgarians opposed to Anselme, and the apex of their own country at Strumnitsa. Here, at any rate, the Bulgarians fought vigorously. Their positions were as strong by art as the Moglena positions were by nature, or stronger, and the attack lacked depth owing to the detachment of a division to work with Anselme's force, and to the low effective strength of units.' Thus it met with the same fate as those of 1917 - small gains of ground consolidated and held after far larger gains had been for a moment achieved. A second attack next day was no more successful. But for these critical days a large force of the enemy had been completely held.

On the 20th and 21st the exploitation of the break-through was completed. The Serbian I. Army, no longer simply guarding the flank of the II., crossed the Cerna and began a drive along the mountains of the Cerna bend to which the Italians and the French of the " A.F.O." conformed little by little. The Serbian II. Army reached the Vardar between Krivolak and Demir Kapu, and its right horn continued to push due eastward along the mountain positions in front of Anselme's force, which progressively came into action from left to right. By the afternoon of the 21st the Doiran-Vardar front was in collapse, and the British aeroplanes were bombing the intermingled troops and transport of the enemy which was seeking to make their way through Kosturino to Lyumnitsa.

From that day, though progress was sometimes slow, sometimes fast, the Allied offensive became a strategic pursuit in the full sense of the word, marked by a consistent policy of outflanking, as rapidly as possible, any solid line of resistance which the enemy managed to create. Thus, on the 24th, the resolute front offered by a hastily assembled German force on the line of the Vardar near Gradsko (the administrative centre and organized base of the enemy's centre) was turned on the N. by the steady advance of the Serbian I. Army on Veles (24th-25th), and once released thereby the Serbian II. Army marched at high speed on Shtip (25th), Nochaua (26th), and Tsarevo Selo (27th), behind the rear of the forces that were giving ground before the British, who in turn worked down the Strumnitsa basin and (in concert with the Greeks further to the right) ascended the upper Struma region, with their aeroplanes sent ahead to bomb the Kresna defile.

Thus, too, when infantry fighting threatened to become stable between Veles and Shtip, the French cavalry brigade under Gen. Jouinot-Gambetta, instead of becoming involved in the Veles fighting; made an independent dash upon Uskub, and to the astonishment of both sides seized that vital centre on the morning of the 29th. This event, which secured the communication between the (so-called) XI. Army and the remainder of the Bulgarians, forced the latter into the region of Egri Palanka, and the former into that of Kosovo. Thenceforward the I. Serbian Army, with the upper Morava as the axis of movement, moved steadily northwards with its right on the Bulgarian border, and its left following approximately the line Gilau-Kinshumlia-Kralyevo, reached Racha-Krushevats-Pirot on Oct. 15, Kralyevo-Parachin-Zayechar on Oct. 23, Pozharevats-Arangyelovats-Uzhitseon Oct. 28, Belgrade-Lyuboviya on Nov. 1, and the old barriers of the Drina, Sava, and Danube on Nov. 4. Meantime, the II. Army had reconquered Kosovo and the Sanjak of Novipazar, reaching Priboy and Plevlye on Nov. 1. By that time the strategic pursuit further E. had become a series of movements authorized by the terms of an armistice, and to the W. Pflanzer-Baltin was evacuating in turn Albania and Montenegro, with, as his only purpose, the maintenance of his divisions as formed military units. The Balkan Campaigns were at an end.

For the Salonika campaigns 1915-17, the principal authority is Sarrail's Mon Commandement en Orient, which is profusely documented; with this should be taken Gen. Milne's despatches. The German part in the Cerna bend battles of 1916 is described in K. Lubmann's monograph Herbstschlacht in Mazedonien, written from official archives. The crowning offensive of 1918 is dealt with fairly fully in C. Photiades La Victoire des Allies en Orient. (X.)