by Norman Stone
Prittwitz let the Russian I Army advance, and attacked it on 20th August with nine divisions—half as much again as Rennenkampf deployed. He led off with a tactical success. The Russian northern group expected cavalry—of which there was a considerable quantity—at least to provide advance warning of any German movement. The cavalry failed to give this warning. In consequence, two German first-line divisions were able to march through woods, in the night of 19th–20th August, and fall on the unsuspecting Russian northern group—parts of which fled as far as Kovno, bearing regimental flags. François, commanding the German forces here, started off in pursuit, but then encountered a problem that came up subsequently, in greatly magnified form. Attackers might win a considerable tactical success; but they would be unable to follow it up, because cavalry was ineffective and supply-problems intervened. In this case, François’s two divisions were held up by the afternoon of 20th August, and François appealed to the rest of VIII Army to help him. But the German commanders to the south did not have the advantage of surprise that François had had. On the contrary, they had to launch expected, frontal attacks against an enemy that had begun to dig trenches of a sort. Mackensen, commanding the central corps of the army, lost 8,000 men in an hour or two, was then attacked in flank, and was rescued only with difficulty when the flanking Russian group was itself taken in flank by a German division to the south. By the evening of 20th August, it was clear that Prittwitz’s attack had gone wrong; Gumbinnen was a clear Russian victory. Prittwitz panicked, for he knew that not only had he been beaten in the Gumbinnen battles, but also that there was a new Russian army approaching from the south. He might not even have time to retreat. He telephoned the German high command and told them that he must retreat as far as the lower Vistula, leaving East Prussia altogether; even the river could not be held, ‘it can be waded across everywhere’. All this was an exaggeration: the Germans had not been badly defeated, and could in any case recover quickly enough. The corps commanders themselves soon recovered their nerves as they noted Russian failure to pursue. Moltke, in the west, was struck by the indecision and panic of Prittwitz, and decided to have him removed. He selected as successor, Hindenburg, and chose as Hindenburg’s chief of staff one of the best technical experts of the German army, Ludendorff. These arrived to take command on 23rd August—Prittwitz learning of his dismissal, characteristically, only when his transport-chief reported that arrangements had been made for a special train carrying his successor.
When Ludendorff arrived, Prittwitz’s staff had already recovered. Clearly, it would now not be possible to attack the Russian I Army—the attack would be expected, and in any case would be threatened from the south by three and a half Russian corps, of II Army. The obvious manoeuvre was to attack the left wing of II Army, an attack that could be carried out without fear for the rear, and the very scheme that Russians and Germans alike had foreseen before the war. Arrangements were therefore made to take one corps out of the Gumbinnen positions, by rail, together with some of the Landwehr, and collect them on the southern borders of East Prussia, where they could attack II Army’s left flank. Ludendorff sanctioned this idea. In his memoirs, he also asserts that he had the idea of taking the right wing of the Russian II Army as well—the other two corps from the Gumbinnen positions to march south-west, against this right wing, the manoeuvre eventually made, to decisive effect. But it is doubtful if Ludendorff saw this far ahead. A more difficult problem was simply to withdraw these two corps from their exposed positions, to the north-east of a Russian army. The railways were already taken up with the evacuation of one army corps, to the left of II Army. The other two corps would simply have to march southwest to join up with German troops facing II Army. If they encountered Russian troops, there would be a battle, but one at least quite close to the concentration-areas of other German troops. Ludendorff was merely keeping his options open, and was not to tell in advance on 24th August what series of miracles would give him a victory as great as Tannenberg. By 25th August he had laid out his forces: François, with two first-line divisions and some territorials, to attack the left wing of II Army; 20. Corps, with a Reserve division and some Landwehr, to hold the centre of II Army and engage its attention; 1st Reserve Corps and 17. Corps to march south-west towards this new battle area; and the entire Russian I Army to be held off by a cavalry division and a territorial brigade, with, in the north, the fortress of Königsberg and in the centre that of Lötzen.21
What followed depended to a large degree on the activity, or lack of activity, on the part of the Russian I Army; II Army was not given the help it needed. In the first place, Zhilinski, who was theoretically responsible for the armies’ co-operation, did little. By nature, he was extremely prudent—just after Gumbinnen, for instance, ordering a guard to be maintained for Grodno,22 far in the rear. He mistrusted such information as he received, and received nothing from Stavka to elucidate matters. I Army seemed to have enough to do. It would have to deal with the fortress of Königsberg, which now lay on its northern flank. Not until 23rd August was the Germans’ withdrawal noted, and even then both Zhilinski and Rennenkampf thought two of the German corps had retired towards Königsberg—the prudent, textbook behaviour to be expected from Russian generals. Organising the siege of Königsberg was difficult enough, and was thought to need ‘five to seven infantry divisions, or three-quarters of the army’s infantry, as well as siege-artillery that had to be dragged forward.23 I Army had in any case suffered from the fighting since it crossed the border on 15th August—3. Corps had lost over 6,000 men, and shell-supply was in disorder. The force had undoubtedly won a victory, and Rennenkampf supposed he had done his work. On 26th August, orders were issued for the siege of Königsberg; the advance over the Angerapp lines went very slowly, with cavalry performing its rôle, as usual, with painful inadequacy. Everyone supposed that the real task was now II Army’s: it must cut off the German’s retreat to the lower Vistula. Only on 27th August did the attack on II Army’s left begin to affect calculations; only on the 28th was there an order for I Army to help II, even then countermanded. In effect, the Germans would have until 30th August at least in order to concentrate all their forces against the Russian II Army.24
Given that II Army advanced steadily into the ring Ludendorff was setting up, a great victory was set up for him. Ill-luck and misjudgment thwarted the operations of Samsonov, commanding this army. He was told that the Germans were retreating, that II Army must cut them off; and his forces were directed even further from I Army than before—something of a gap opening between his right and his centre. On his left was 1. Corps, forbidden until 27th August to go over the border. Supply and communications were in disorder. A new supply-system had been introduced just before the war, giving a front supply-chief full powers, in place of the army ones; and no-one knew how to work this system. Even in Bialystok, bread had been lacking, and the habit of allotting the same supply-route to different corps meant that one corps would receive too much, and another not nearly enough. For 25th August, commanders appealed for a day’s rest, which Zhilinski forbade, but which they no doubt took. Communications, too, were in disorder. Telegrams to the army command went to the Warsaw post-office, and from there went in bundles, by car, to Samsonov’s headquarters.25 On the right, 6. Corps manoeuvred along sixty kilometres of front, receiving messages in a code unknown to it. Yet Samsonov was not too perturbed. He was told on 25th August that two German corps had retired into Königsberg; for 24th August he was also told by Zhilinski that there were ‘only insignificant forces’ in front of him. In any case, his centre, as it came over the border, had already done quite well against these forces, taking seven guns in one frontier-battle.26 Samsonov therefore pushed forward the two-and-a-half corps of his centre, expecting 1. Corps on his left to be sufficient guard, and 6. Corps on his right to be free from all danger.
The problem became all the greater since there were delays in German action. François, who was supposed to attack the Russian left wing, refused
to move on 25th August—his group had just arrived by train, after severe fighting, and its guns had not come in. On 26th August, François again prevaricated, received a visit from Ludendorff, and attacked feebly, with endless deliberation. Ludendorff was annoyed; but François’s delays, accidental as they were, in practice lent a dimension of success to Ludendorff’s scheme that its author had not imagined—for Samsonov, unperturbed for his left, pushed on his centre all the more. Even in the morning of 27th, François’s attack met with difficulties on the southern wing. But his two first-line divisions on that day attacked the right wing of the Russian 1. Corps, which had been told not to cross the border, and, in the area of attack, was particularly thin. François broke through between it and the left-hand group of the centre, nearly 200 German guns being opposed by only fifty; and by the end of 27th August, François had entered the frontier-town of Soldau, through which led one of the vital lines of retreat of Samsonov’s central group. He had snapped the link between left and centre; but this, with the state of communications in Samsonov’s army, was appreciated hardly at all. Artamonov, commander of 1. Corps, stood bravely with his men in the front trenches, and thought he had already done his duty, and at least saved himself.
The delay in François’s attack had allowed Samsonov’s centre to move forward, through country that gave little visibility to commanders. From 26th August two corps and a division moved forward against the Germans’ central group. They began to win some successes, on the river Drewenz and south of Allenstein, and gained misleading confidence in consequence—again, despite Ludendorff’s wishes. A series of flanking operations, in which the greater weight of the Russians told, began to go badly for the Germans; and even the intervention of a fresh Landwehr group (Goltz) brought little relief. On 28th August there was even ‘a major success’27 for the Russians at Waplitz, where they almost destroyed a German division, incontinently attacking. Ludendorff benefited from this, in the long run, because the Russian centre group persisted in its course; but he did not see it, and ordered François on the 28th to send back a division to help the threatened centre. Whether because he knew what was happening, whether because of some natural disinclination to obey the plebeian Ludendorff by sending first-class Prussian troops to the help of some third-rate Landsturm unit, François closed his eyes and continued his advance east of Soldau, thus arriving far into the rear of the Russian central group, much of which could not now retreat. Early on 29th August, François’s cavalry units entered Willenberg, on the road east of Soldau, and there encountered units of a German corps that had marched from the north-east, Mackensen’s 17. Corps: Samsonov was surrounded.
This, too, had been achieved partly by accident. Ludendorff had told the corps remaining in the Gumbinnen positions to march south-west as fast as they could. Probably he merely hoped they would come to the help of his centre, but he maybe had more ambitious ideas. On 26th August these two encountered the right-hand corps of Samsonov’s army28—the two divisions of 6. Corps (Blagoveshchenski). No-one had told Blagoveshchenski that there was any danger—on the contrary, both corps were thought to be in Königsberg. Blagoveshchenski’s front was too long, his task not at all clear. His two divisions were separated, one of them under a veteran of Plevna; and even the divisional reserves had lost touch with their command. News did reach the Russians that there were troops marching south-west—aircraft had noted this. But the movement was thought to be of Russian troops, certainly not four German divisions, two of them first-line.29 As a final twist, the Germans’ intervention, despite great superiority, was not even particularly crushing: they were tired, the first-line troops had lost heavily at Gumbinnen, and the second-line forces were insufficiently-trained; moreover, neither corps commander (Mackensen and Below) had much idea of what was happening. There was a series of actions, the greatest of them on 26th August against one of the Russian divisions, the flank of which was on a lake, and enfiladed by powerful artillery from the other shore. Its troops suffered forty per cent casualties, a third of the guns were lost, and the divisional command was in such disorder that even its report of events could not be written until a week later. The two Russian divisions retired in disorder towards the south. But they had not lost so heavily for men to believe that they had been attacked by overwhelming forces. Samsonov assumed that the matter originated in Blagoveshchenski’s misdoings, not in a German threat of any significance. Such news as reached him merely prompted him to push forward the centre.
. In this way, two German corps found themselves well into the right flank of Samsonov’s army. Accident again made this more effective than it might have been. The two corps commanders quarrelled about what to do next—each prompting the other to stay put, in order that he could use the convenient road towards Allenstein and safety. Ludendorff intervened on 27th August to tell Below, commanding the less reliable force, that he should use the road and come to the help of the German centre, around Allenstein; Mackensen was to follow. There were delays in executing this order, and in the mean time Ludendorff woke up to reality—with François well into the Russian left, Mackensen was well-placed to close the circle from their right. On the 28th Mackensen was told to move west rather than north-west, and in so doing he met up with François’s units in Willenberg. The other German corps moved towards Allenstein, and forced the Russian centre there to turn about, its guns now facing north-west and east.
Samsonov’s situation, which even at midday on 27th had seemed bright, had now become hopeless. His left wing and his right wing had been thrust away from the central body, and there was no way of getting orders through to them, since the Germans had, by 29th August, closed the roads leading to the south. The centre of his army had been involved in a complicated series of flanking manoeuvres, some of them misleadingly successful. But the arrival of a German corps in the rear, near Allenstein, now blocked even roads to the east. No doubt a command that functioned properly could have found some way out—the German cordon was thin enough. But there were over 100,000 men in this pocket, insufficiently in touch with each other, badly-supplied, and in an area that their commanders did not know. In the broken country, some units had already disintegrated, without fighting at all. There were merely confused reports of German troops blocking the passages to the south. In any case, roads to the south were few: one, through the Kurken defile, would have to be used by two corps, and already it was under German artillery-fire: and since it was only 500 yards wide, to use it would be calamitous. Tired and bewildered, the Russian soldiers began to surrender on 29th August—a whole series of captures, in which German battalions received the surrender of Russian brigades. Only ten thousand men escaped; Ludendorff, who had expected only thirty thousand prisoners, had to report nearly 100,000. Samsonov himself rode off towards Willenberg with his chief of staff, a Cossack guard, one map and a compass for the entire group. He is said to have shot himself. As a final act, the corps of the left and right—1. and 6.—stumbled forward, under Zhilinski’s prodding, on 29th August. They caused panic in the German cordon, and then prudently retired. Rennenkampf, further off still, ordered his cavalry forward on Allenstein, and then told it to stop. Confusion was such that Zhilinski himself failed to appreciate what had happened until 2nd September.30
Tannenberg—which, on rather dubious grounds, gave its name to the battle—was the most spectacular victory of the war, and generated a propaganda-myth for years to come. East Prussia had been defended, seemingly against overwhelming odds; 100,000 men and nearly 400 guns were captured. In practice, the victory was overrated at the time: the Russians recovered, and invaded East Prussia again a few weeks later. But what was dangerous to the Germans was the myth that Tannenberg launched. Men supposed—and the version produced by Ludendorff, later, buttressed their supposition—that Hindenburg and Ludendorff had made a brilliant strategic manoeuvre, leading to a new Cannae. There was something in this, but it was distorted by exaggeration. The transfer of troops to Samsonov’s left was arranged before Ludendorff arriv
ed, and formed part of a sensible scheme foreseen some years before—a coup against the Russian left by troops that also had a good line of retreat if things went wrong. Even this developed as well as it did largely because François disobeyed Ludendorff’s instructions. The manoeuvre that brought two German corps into Samsonov’s right flank was certainly Ludendorff’s; but that, too, was calculated as a form of retreat, and only chance and delay turned it into the coup against Samsonov’s right flank of the type eventually attained. In a sense, the vital part of this battle was the expensive and protracted fighting of the central groups—frontal engagements that were far from being in the Germans’ favour—for this pinned the two corps of Samsonov’s centre to the point of immobility. But the essential features were on Samsonov’s side—confused command, disorderly supply, above all a completely mistaken picture of the Germans’ dispositions. The Germans won because they were defenders, on whom sense was almost imposed by the lay-out of the land and the railways, and the nature of the task. Tannenberg none the less launched a myth of the brilliant strategic coup, a perpetuation of the Napoleonic myth in eastern Europe which, in the circumstances of 1914–1918, was dangerous enough. It was common sense, discipline, de-centralisation that won Tannenberg; just the same, Ludendorff discovered from it that he had military genius, and found a large number of Germans to agree with him.