by Norman Stone
Moreover, on the ground there were solid arguments against retreat: the fortresses of Poland.11 The men surrounding Grand Duke Nicholas had strongly opposed Sukhomlinov’s proposal to scrap these; they were now foist with the consequences of their own proposals to retain them. Novogeorgievsk, down-river from Warsaw, was a great fortress, with 1,680 guns, and almost a million shells. Ivangorod, though smaller, had, been restored in the course of the war, and had many partisans because of its stout defence in the previous October. These two, with Grodno, Kovno, Osowiec and Dvinsk, contained 5,200 older fortress-cannon and 3,148 modern high-trajectory guns, as well as 880 large-calibre modern cannon. These fortresses had been constructed with just the present emergency in view, and it seemed precipitate to abandon them (and thus confess the enormity of pre-war errors) at the very moment they were required. Moreover, Schwarz in Ivangorod and Sveshnikov in Osowiec could point to singular successes against the Germans, in October 1914 and February-March 1915, while in any event the evacuation of fortresses could take an unconscionable amount of rolling-stock—Prince Yengalychev, governor of Warsaw, said that he would need 2,000 trains to evacuate Warsaw, and 1,000 for Novogeorgievsk, quantities which supply of the ineffable horse obviously forbade. It would in any case be highly embarrassing to give up fortresses at the very moment Sukhomlinov was being charged with high treason for proposing the same thing years before the war. Of course, the fate of fortresses in western Europe should have shown Stavka the folly of its course: Liège, Namur, Maubeuge, Antwerp had collapsed after at most a few weeks’ defence, and the French ability to hold the town of Verdun was based on correct recognition that the forts ostensibly protecting it were just concrete traps, offering obvious targets for the Germans’ heavy artillery, and preventing the defence from behaving flexibly. But Stavka knew better: there would be an attempt to hold the fortresses.
There were of course occasional hints of realism. Palitsyn, who had been summoned to advise Stavka, thought that ‘soberly-considered, it is a mistake for us to fight for the foremost theatre’. But even Palitsyn swung the other way when news from the front seemed to improve. Stavka sent him to Ivangorod with an instruction to the commandant to prepare for retreat, but he was told to add, personally, that there should be no great hurry in such preparations. The front commands were meanwhile rapped over the knuckles for suggesting that they might retreat. Ivanov had proposed at least withdrawing the armies west of the Vistula to lines where retreat would be easier, a shortening of the front that would allow him to take some of its troops for reserve. Stavka’s response was to give responsibility of this part of the front to Alexeyev. But Alexeyev too thought retreat would be sensible. His I and XII Armies holding lines north-west of the Narev were exposed to a German attack, of which there were already signs. He proposed taking them back to the Narev, again to create reserves on a shorter line. Stavka thought that this would cause the German attack Alexeyev feared, and told him to leave the troops where they were. On 25th June Yanushkevitch recommended construction of strong lines, one after another, ‘on which our troops can successfully hold out, dragging the enemy into battle for every square foot of land’. Characteristically, nothing much was done to prepare such lines: men disagreed as to where they might be dug, lacked man-power to dig them, and in any case thought they would be bad for morale. Some desultory earth-shifting alone followed Stavka’s precepts: the lines offered obvious targets for German artillery, yet little protection for men inside them. Although many highly-placed officers knew, in their heart of hearts, that the shell-crisis, the rifle-crisis, the strategically unfavourable layout of Poland, and the moral crisis of the army could only add up to retreat, nothing much was done to prepare for it. ‘The masterly Russian retreat of summer 1915’ was a legend, invented by the Central Powers to excuse their own blunders. The Russian supreme command simply waited on events, or rather hoped there would be none.12
Stavka’s decision to retreat came piece-meal and reluctantly, as a consequence of German action. In July 1915 Falkenhayn elected to launch a double attack on Poland, from north and south, as well as to maintain pressure in Courland. The Central Powers, too, had suffered from indecision, but they needed success, and were therefore prepared to go on against Russia, where originally they had meant to stop on the San. Falkenhayn temporarily forgot his plans for an offensive against Serbia, to clear the way to Turkey, while Conrad also shelved his plans for an offensive against Italy. Even so, Falkenhayn’s aims were always very limited against Russia. He did not want to commit himself to an open-ended Russian campaign, for he feared another 1812: ‘The Russians can retreat into the vastness of their country, and we cannot go chasing after them for ever and ever’;13 there would be huge supply-problems, and the forces were in any case too weak. Falkenhayn did not believe that Germany could win a two- or three-front war. One of the major enemies must be led to a separate peace; and throughout 1915, he made attempts at a separate peace with Russia, through a Danish intermediary, Andersen. He received dusty answers. Just the same, both he and Conrad were interested in ‘building golden bridges’ to Russia—no doubt, over the Golden Horn—and their strategy was designed at least as much to show German invincibility as to bring about a complete destruction of Russia—or so Falkenhayn claimed. Ludendorff had different ideas. He was impressed at the Russians’ crumbling in Courland, and thought that a bold stroke from the north-east would work. Falkenhayn preferred his own limited strategy of attrition, and the most he would undertake was an attack over the Narev towards Warsaw. The Russians would lose Poland, and many thousands of men. They would not lose much territory—which was as well, for the Germans would not have to lengthen their lines of supply, and the Russians would not be stimulated into defence to the uttermost. By 2nd July this plan, after considerable objection from Ludendorff, was agreed to. Mackensen’s group would attack from the Austro-Hungarian border, and, from the front north of Warsaw, a strong group under Gallwitz would attack. On the map, this did not look like an ambitious pincer-movement such as Ludendorff, and, with some reservations, Conrad, hoped for. But at bottom it reflected a much more sensible view of the war.
Three German operations therefore proceeded: Galicia, the Narev, and Courland. Already, Russian reserves had been greatly dispersed, and the three-front attack proved to be an excellent way of dispersing them still more. In the Galician theatre, the Central Powers’ forces grouped as Heeresgruppe Mackensen were to move north, straight towards Brest-Litovsk. Conrad had different ideas—wanting some of this force to go to the north-east, where something of a ‘pincer-movement’ could be effective, the rear of the Russians in Poland taken. Some cavalry, and an Austrian army, were vaguely ear-marked for this purpose. Falkenhayn thought that a stroke of this type would merely end up in supply-difficulties, since it had to cross the great marshes of the Pripyat; and he had little faith in the ability of Austrian troops to carry out an ambitious scheme of this type—doubts that the event substantially confirmed. Mackensen himself was against fancy manoeuvres. He had thirty-three and a half infantry and two cavalry divisions in his own Army Group, with eight and three in the Austrian I Army on his right. He faced thirty-three infantry and six and a half cavalry divisions, many of them fresh troops, and he had no doubt that ambitious flanking-schemes in the east would simply mean a subtraction of force that he could not afford. If the flanking operation could make much speed, no doubt it would be different; but the roads were few, and mobility bound to be too little. ‘As before’, he later said, ‘I reckon that a strong centre is suitable for attack, as it will give us the best chance for successful advance.’ His method was to assemble a great weight of artillery in his centre. Hitherto, this had worked well enough, and it continued to do so. The Russians not only failed to make a strategic retreat at this time; they also had no notion of tactical retreat, and yet the trenches they built were primitive. In these circumstances, the Germans’ bombardment was always extremely destructive. A British observer wrote that at Trawniki, in late July,
the Guard and 2. Siberian Corps had been bombarded from 2 a.m. until noon, one Siberian division was ‘annihilated’. The Guard did stay put under heavy fire, ‘whereas in other corps the men run away and are destroyed by shrapnel’.14 He added, ‘It is unfair to ask any troops to stand this nerve-wracking unless they have a regular rabbit-warren of trenches.’ The Russian trenches were ‘graves’. In these circumstances, Mackensen’s bludgeon seemed a good policy to use, and Falkenhayn went ahead. Mackensen assembled troops on the Galician border, contained the attack of the four corps that Alexeyev had sent south—they won a success near Kraśnik against the Austrian IV Army, incontinently advancing—and in mid-July attacked towards Lublin and Cholm.
Both places had fallen by the end of the month, after a break-through at Krasnostaw that brought 15,000 prisoners in a day. Conrad again grumbled that more force should be sent towards the north-east, and an attempt made to trap huge Russian forces in the Polish pocket. But Falkenhayn had no time for schemes of this kind. In the days when cavalry was effective, they were all very well: now, they would simply mean infantry plodding forward through difficult country against an enemy going away by railway. It was much better to draw the enemy into a battle somewhere where he would not refuse to fight, and then knock him down by virtue of artillery-superiority. In irritation, he told Conrad—in words that have subsequently been held up as an instance of strategic bankruptcy—‘It is endlessly less important where Mackensen and the Bug-Armee break through, than that they should merely break through somewhere.’ Falkenhayn was a modern general, and had a more sensible view of the war than either Ludendorff or Conrad. He knew that great manoeuvres, as in past wars, could not fit in with present circumstances. The war in the East proved Falkenhayn to be right. What shook Stavka was not the ostensibly brilliant manoeuvring of Ludendorff—and certainly not that of Conrad—but the huge losses they suffered in set-piece soldiers’ battles such as Gorlice, or Mackensen’s bludgeoning before Lublin. They were much more costly, even than Tannenberg. Alexeyev told the Grand Duke late in July—when Stavka protested that the new army appeared to be achieving little—‘You appear not to appreciate the situation regarding III and XIII Armies (on Mackensen’s front). In numbers, they are now insignificant, exhausted in an extreme degree, incapable of further resistance.’ 10. Corps counted only 4,000 men ‘with almost no officers left’; and so few troops could come in ‘that it is all just a drop in the ocean’. The two armies together required 180,000 men to bring them up to their complement. In a world in which Napoleonic envelopment was ruled out, Mackensen’s tactics offered easily the best method of knocking Russia out of the war. It was Conrad and Ludendorff that were out of date.
Mackensen’s advance from Galicia was very slow, dependent on constant assembly of ammunition and reserve-troops. Until mid-July, the railway functioned only as far as Rozwadów on the San,15 and even thereafter only to Lwów; everything had to be brought up by cart, and even when a field-railway was constructed—rails along which horses might pull small waggons—it only went as far as Lublin, in mid-August. The attrition-battle was in itself something of an achievement, and there was certainly no room for ambitious schemes in the Pripyat.
Mackensen’s attack made sense only if there were a complementary attack over the Narev. The Germans had decided on this early in July, for a break-through here would cause the collapse of the Russian position in Poland; and Falkenhayn calculated that the Russians would make some effort to defend Warsaw and their fortresses. The Narev offered conditions for a successful battle of attrition, the more so as Russian reserves would necessarily be confused between the Kovno front, the Narev, and Mackensen’s front. These calculations—despite what Ludendorff subsequently said—were correct.
The Germans put, on forty kilometres, ten and a half divisions, one in reserve, over 1,000 guns and, for the east, ‘masses’ of shell—400,000 rounds.16 They had tried to break through in the area west of the Narev in February-March, and had failed; now they were determined to risk nothing. The operation was prepared with characteristic thoroughness, the artillery being in charge of Bruchmüller, the break-through expert. The offensive was to begin on 13th July. Success was based on much the same factors, though on a lesser scale, as at Gorlice in May. Stavka and Alexeyev had confused their reserves, partly because they were already undecided as to whether further retreat should be undertaken, partly because they were in no agreement as to priorities. Alexeyev had arranged for reserves to come to the Narev from the central theatre, but he had left it too late. When the Germans attacked, these were still many miles from the front. Tactically, the Russian position was also weaker than it should have been. The German attack fell on the inner wings of two armies—XII (Churin) and I (Litvinov). They did not co-operate efficiently. They had lost 2. Siberian Corps and the Guard for the south-western front. Yet they were deluded by their success in February. There was no inspection of the front line until a week before the Germans attacked, when a commission of 1. Siberian Corps reported that gun-positions were too open, dug-outs were constructed with weak, even rotten, timber, the trench-lines had not been planned with defence in mind, and in the second position the vegetation had not been cleared. Seven divisions with 377 guns held this line, and are said to have had only forty rounds per gun to use; although—as at almost every turn in the Russian material crisis—a great deal of the material shortage could have been compensated for by more agile behaviour on the ground. In this case, the Russians had on their flank the 1,600 guns of Novogeorgievsk, of which no use at all was made.
Litvinov and Churin also mismanaged the battle. As ought to have been foreseen, the bombardment was successfully managed. A Siberian division tried to escape from it in the middle, and lost fifty per cent in half an hour. On 13th July, the two Russian armies were forced two miles apart—the break-through, near Przasnysz. Reserves did not co-operate. Counter-attacks would be launched by the one, while pure defence was maintained by the other. Litvinov demanded ‘categorically that all troops should hold their lines’—merely a way of making sure they would be ground down by the German guns. The two central divisions were almost wiped out. Alexeyev had arranged for two corps to come to the Narev (21. and 3. Siberian), but they were too far off to help. Such reserves as did come in were mishandled—taken by one corps commander, then another.17 By 17th July, the Germans had advanced perhaps five miles, but they had inflicted seventy per cent losses on the defenders (including 24,000 prisoners, a quarter of the Russian numbers). This coincided with Mackensen’s successes at Kraśnik and Krasnostaw. Alexeyev pulled back his troops to the Narev, with corresponding withdrawals to left and right.
Maybe the Narev line could have been held, if defensive tactics had been correct. But Litvinov himself could not imagine any other system of defence than that which he had applied—every soldier to stay put until wiped out, after which reserves would arrive to be wiped out—and told Alexeyev that ‘My army will be unable to hold out for long on this new line, as it is not fortified and its right flank is altogether unsuitable for defence.’ The fortresses of Pultusk and Rozan were museums—though costly enough, before the war—and the river was not deep. On 17th July, both Falkenhayn and Gallwitz—commanding the offensive—felt that they could manage to break through grandly towards Warsaw. But Alexeyev had planned things well enough. As the Germans came forward, they stumbled against increasing numbers of Russian troops, such that the 62 battalions and 188 guns that had faced the initial German attack rose to 100 and 600 respectively once the Germans arrived on the Narev. German attacks fared badly—the guns unprepared for Russian resistance, the troops defeated by machine-gunnery; and, other than an unexpected tactical success near Rozan, there were no gains to report, and only 3,000 prisoners.
It was not the defence that snapped, but Stavka’s nerve. Alexeyev had so far fought the battle correctly enough, and Gallwitz would have had great trouble in penetrating the Narev. But there were complications on other fronts—the north-east, where a very origina
l set of blunders led the Germans towards Kovno—and even in the central theatre where, just west of the Vistula, the German IX Army was allowed to win a success with almost no superiority of numbers at all. On the southern side, the railway from Lublin to the east was cut by Mackensen, and for a time even the activity of an Austro-German cavalry force east of the Bug (Heydebreck) was taken seriously. On 19th July Yanushkevitch opined that the German break-through to the Narev meant giving up Warsaw; the Grand Duke himself turned up in Alexeyev’s headquarters and said that retreat to the Vistula, in the centre, was now required—otherwise significant forces would be cut off in Poland; and Alexeyev was now told ‘You can evacuate Warsaw, if you feel you must’. On 22nd July orders went out for a retreat of the Narev groups to the east (Lomza—Siedlce—Kock) and of the southern armies to a line north of Lublin and Cholm (Ivangorod—Opalin—Kowel). Apart from a minor mishap on the Vistula, this retreat was undertaken. The Germans entered Warsaw on 4th August.18