As things turned out, IX Army won a considerable, but irrelevant, success. Its 120,000 men took on 80,000 of the Austrian VII Army, dependent for supply on the narrow-gauge mountain railway through Körösmezö, and always hampered by Hungarian politicians’ interference—Count Tisza had forbidden the Hungarian State Railways to transport troops from Transylvania to the Bukovina because he wanted a guard kept against Romania, and feared a rising of the Romanians of Transylvania. By mid-May the Austrians had lost much of the Bukovina, and were forced to use one of III Army’s corps on this front, instead of on the Italian front. In the same way, the attacks of Südarmee on the Dniester were unsuccessful, for the most part, and losses were high. Falkenhayn disregarded these matters: if the Central Powers could succeed on the San, they would be cleared up. If the Central Powers failed there, nothing else would matter any more. Mackensen’s force was assembled for attack on the San, beginning in mid-month. On 9th May the Kaiser took up headquarters at Pless, to be near the scene of new operations in the Balkans and the east.
Mackensen’s shell-reserve was once more brought up to the standard thousand rounds per field-gun: in other words, not far short of a million rounds. The entire reserve of Ivanov’s front on 13th May amounted to just over 100,000 rounds of all types, on 20th May to 114,000 field-cannon shell, 118,000 mountain-cannon, 25,000 howitzer, 42,000 heavy. III Army could only be given trivial amounts—60,000 rounds sent from Lwów on 17th May, 15,000 to follow—even then with unpleasant observations: ‘Unless you have been throwing away your shell-boxes, there must be enough’. But as yet tactics had not been adapted to deal with this emergency. The San positions had not been prepared at all—on the contrary, Austrian wire, trenching-tools, timber-props and the like, captured in the previous autumn or in Przemyśl, had been sold off to the local populace. Przemyśl itself was still something of an obstacle, but there was nothing much from there to the north. Besides, the line had many peculiar features. On 21st Corps’s front, near Radymno, the eastern bank was completely dominated from the western bank, and the river was only a few yards broad. Its defence could only be conducted from the western bank, with all the isolation and difficulties involved.
The San battles opened, more or less, with repetition of the Gorlice pattern. Since the troops were now even weaker—III Army having lost five-sixths of its force—reserves were even fewer than before. Radko-Dmitriev’s divisions had merely been assigned bits of the river to defend: reserves, for the whole position between Przemyśl and Jaroslau, consisted of two infantry regiments. A new ‘29. Corps’ had been formed from the wreckage of 81 and 13. Siberian Divisions—together, not 2,000 men. A ‘composite division’ constructed from a new second-line division and one division of 3. Caucasus Corps, consisted barely of a ring of men round the guns. Alexeyev had, by 20th May, sent ten divisions in all: the three corps involved were now gathered on the lower San, opposite the Austrian IV Army, with a view to counter-attack. This was Radko-Dmitriev’s only hope, and yet, with shell in such short supply, not much was to be expected.
A German tactical success between 16th and 19th May led off the San battle. The Radymno salient collapsed, thousands of the defenders being drowned or machine-gunned as they tried in panic to cross the San. By 19th May the Germans had established a large salient across the river, and now sought to extend it to the south—towards the rear communications of Przemyśl. It was a mark of the Russian defence’s considerable qualities that this attack was far from being successful. As Mackensen’s group inched forward, Falkenhayn grumbled. He told his commander in the central theatre, to the north, to prepare for attack: affairs in Galicia ‘do not have good prospects’; ‘we can win only if we use great quantities of ammunition and man-power, and we have the best reasons for economising with both’.
Moreover, the Austrians, in the central theatre, suffered unmistakable reverses, and had to be given help from Mackensen’s group, itself under pressure. But a mismanaged Russian counter-attack dispelled some of these advantages now coming the Russians’ way. In the night of 19th–20th May they attacked the Austro-German lines with three corps taken from Odessa and Alexeyev’s front. Another new Corps had been set to Südarmee’s front to Stavka’s annoyance. The Austrians suffered a crisis retired, lost many thousands of prisoners, and had to be helped out by a German corps. Against the German front, this counter-offensive broke down, reducing the new troops to much the same level as the old ones.
By 25th May it had been contained; indeed, the Germans had profited from it themselves since, just before a Russian attack came, with the troops exposed, waiting to attack, the Germans had launched one of their own attacks and destroyed 5. Caucasus Corps which, in a few hours, saw its regiments sink to a few hundred rifles. Further east, the successes of Ivanov’s troops had little relevance, even when won, on the Dniester, against German troops. On the contrary, they were a waste of precious shell. Early in June, the defenders had been pushed back until only a narrow corridor of three miles connected Przemyśl with the armies to the east. On 4th June the Austrians and Germans finally entered the fortress.
On 23rd May, the Italians had finally declared war—perhaps given the final impetus to do so by the reflection, paradoxical, in view of Conrad’s strategy, that the Russian position must be saved. But their army was not at all ready for what was to come. Although, to start with, there were over 800,000 men to 100,000—armed at that with old rifles and out-of-date artillery—the Italians failed to gain more than a line of outposts on the Austrian side of the border. The tasks of supply were beyond them, and defenders, in these mountains, had an advantage still greater than elsewhere. For the whole summer, the Italians failed to divert significant German forces from either front, and the Austrians also held their new front with relatively insignificant changes—III Army from the Carpathians, some of their troops from the Balkans. The Serbian army, anxious to seize Albania before their nominal allies, the Italians did, could also give little help to Russia.
In the circumstances, Ivanov could only retire. Conrad and Falkenhayn, on their side, were not sure what to do. Conrad wanted to attack Italy. Falkenhayn would have preferred an attack, by some thirty divisions, on Serbia. They compromised on continuation of the campaign against Russia. Here great successes seemed to come, almost by themselves. Seeckt persuaded Falkenhayn, at Jaroslau station, to pursue the attack across the San.
Ivanov had now used up all of his reserves, either in the fruitless counter-offensive on the San, or in the less ineffective offensives against Südarmee in the Dniester valley. The shell-reserve sank to 240 rounds per gun, and of course much less than this at the front. Nominally, he held to the idea of ‘stubborn retention of every square foot of ground’. But the troops had lost so heavily that they had not the force to do this; increasingly, too, they had not the will. Now, he withdrew even IX Army back to the Dniester. The chief cause of defeat had been tardy movement of reserves. They had come in little-by-little, and now it was much the same. Alexeyev offered, now, 6. and 23. Corps with a Guard division, to be linked, on the western side, with the troops beaten in the San counter-offensive. Later, the whole of the Guard, 2. Siberian Corps and 31. Corps were offered from Alexeyev’s front, and Alexeyev took over defence of western Galicia.
But to concentrate these troops took time, and the Germans meanwhile brought three divisions from the Balkans and another two from the west. These were to ‘strike the enemy east of the San until we have a decision suitable for our purposes’. In mid-June the German offensive began. Soon, Mackensen was able to conclude that he faced ‘completely defeated troops’. Stavka on 17th June announced that the whole front had gone down to 500,000 men short; the shell-reserve amounted to forty per cent of complement. Moreover, the forces in Galicia had to retreat eccentrically—VIII Army to the east, III Army towards the north-east, and the new group—under the Guard general, Olokhov-to the north—such that energetic action by the Germans could always strike an exposed flank. Even within the armies, links became strained—particularly in VIII, defending the Wereszyca positions and Lwów.
Within six days, Mackensen’s offensive brought the Austro-Germans towards this city. Stavka now believed retreat must begin. The Grand Duke told the Tsar that two-thirds of Ivanov’s men had dropped out; but ‘the quality of replacements, as regards their training, is beneath criticism. Their training has been very hurried, and because rifles are short, they do not even know how to fire’. On 20th June ‘energetic evacuation of Lwów and the rest of Galicia’ was ordered; on 22nd June the Austro-Hungarian II Army entered it. The six weeks’ campaign had turned out to be one of the greatest victories of the war. For an initial investment of only eight German divisions, Falkenhayn had earned, with the forces of Mackensen’s group alone, 240,000 prisoners and 224 guns, for a loss of 90,000 men.