On the 16th, however, the fall of Ovillers prepared the way for an attack on Pozires, which was finally captured with the help of the Australians on the 26th, and the taking of Waterlot farm on our right opened up an advance on Guillemont. Much of High wood was recovered on the 20th. On that day the French pushed east of Hardecourt and seized a section of the Combles-Clery railway, while farther south they secured the German defences from Barleux to Vermandovillers. On the 27th the last German outpost in Longueval was taken, and on 4 August the Australians began their advance from Pozires to Mouquet farm and the windmill which commanded the summit of the Bapaume ridge. The ground was contested inch by inch, and it took many weary days to win. Villages and woods all along the front were only captured by fragments, and most of the fragments were lost again more than once before they finally passed into our hands. Well into September there were bits of Delville wood and High wood still in German possession, and a concerted attack of 18 August was a failure except for the seizure of Leipzig redoubt. On the 12th, however, and again on the 16th, the French improved their position north of the Somme and got close to Maurepas, of which they completed the capture on the 24th.
September was a better month for both the Allies. There was a general attack on the 3rd, when Guillemont, which had been disputed for six weeks, was carried at length, and the French rushed Le Forest, Clry, and the German lines up to the outskirts of Combles. Two days later the British got into Leuze wood between Guillemont and Combles, and captured Falfemont farm to the south, while a new French army extended the line of battle below Chaulnes and took Chilly and Soycourt; on the 6th they pushed their advance both north and south of the Somme, taking above the river L'Hpital farm and Anderlu and Marrires woods, and below it parts of Vermandovillers and Berny. The German counter-attacks were unusually unsuccessful, and on 9 September Ginchy was carried by the Irish regiments which had helped to take Guillemont. It looked as though the Allies were at least getting into their stride, or the wasting struggle was beginning to tell on the German reserves and resistance. Over two months had been spent in securing objectives marked down for the first day or two of the battle; but with the fall of Guillemont the last fragment of the German second position had fallen into our hands, their third was more or less improvised, we had a new weapon in reserve, and were half-way from our original lines to Bapaume. Farther afield Rumania had declared war, and Brussilov was still drawing German troops from West to East.
The third stage of the battle therefore opened with hopes which even the experience of the second had not been able to quench. Gough's Fifth Army had since early in July been formed as an independent command to the left of Rawlinson's Fourth, and its right comprised the 1st Canadian Corps which was to attack Courcelette. The other points of the German third line of defence were Martinpuich, Flers, Lesboeufs, and Morval. Martinpuich was the objective of a Scottish division of the New Army, Flers that of the New Zealanders, Lesboeufs and Morval those of the Guards and another division of the old Regulars. Behind the British lines were collected twenty-four "Tanks," which were to precede them in the attack and prove by this first experiment their value as a weapon of war. On the 14th a brigade of Gough's army stormed the Hohenzollern trench and a redoubt called by the Germans a wunderwerk; apart from this success, the attack diverted German attention from the real offensive, which began on the 15th with an intense bombardment. The Tanks spread terror and devastation among the German lines and the results of the day for once exceeded all expectations. Courcelette fell to the Canadians, Martinpuich to the Scots, Flers to the New Zealanders. High wood was at last enveloped in this advance, and Delville wood passed by the division of the New Army which pushed from Ginchy towards Lesboeufs. That effort on our right was, however, hampered by the Germans in the Quadrilateral and Bouleaux wood to the east of Ginchy, and the Guards were unable to carry out the most important tactical part of the day's work by carrying Lesboeufs and Morval.
The French had no such accumulation of gains on the 15th, but they conquered a larger area between the 13th and 18th. They began on the 13th with the bold capture of Bouchavesnes right across the great road from Pronne to Bapaume, and supplemented it by taking Le Priez farm on the flank of Combles. On the 17th they completed their work in Berny and Vermandovillers south of the Somme, and on the 18th added Denicourt. On that day the British at last mastered the Quadrilateral east of Ginchy, and thus prepared for the great success which attended the next general attack on the 25th. It was the best day of the whole campaign. Lesboeufs and Morval fell on the north of Combles, while the French took Rancourt on the south-east, and away to the west Gough's army made the surprising seizure of Thiepval. Further fruits were gathered on the morrow; Gueudecourt, which had been taken but abandoned on the 25th, was recovered; the French who had then failed against Frgicourt now took it; and Combles was the prize of their joint success. Then the weather broke; and the Germans, who had already begun to prepare their Hindenburg lines far away in their rear, were enabled to cling to the Bapaume salient until they had taken all the precautions for an orderly and inexpensive retreat.
The rest of the Somme campaign was an affair of local details until Gough's Fifth Army intervened on a larger scale. Eaucourt l'Abbaye was taken on 1 October, lost on the 2nd, and retaken on the 3rd. Le Sars was captured on the 7th, the Stuff and Regina redoubts, between it and Thiepval, on the 21st; and progress was made north towards the Butte de Warlencourt and north-east towards Le Transloy. The French captured Sailly and Saillisel to the east of Morval and pushed far into the St. Pierre Vaast wood and towards Moislains, while south of the Somme they took Ablaincourt, Le Pressoir, Fresnes, Villers-Carbonnel, and Barleux, and seized the west bank of the river opposite Eterpigny above Pronne. On 9 November the weather improved, and though the October rains had made transport almost impossible across the mangled soil of the battlefield on the Somme, the conditions were not so bad north of Thiepval, where our advance had been stayed on 1 July. The situation at Beaumont-Hamel was also changed for the better by the fact that the German stronghold was now a pronounced salient enfiladed by our fire from the captured Hohenzollern, Schwaben, Stuff, and Regina redoubts. But that advantage was less felt farther north at Serre, and there the left wing of our attack on 13 November was no more successful than it had been on 1 July. Better fortune attended our effort between Serre and Beaumont-Hamel, but the farthest advance of the day was that of a New Army division on the extreme right of the attacking line. St. Pierre Divion fell almost at once, and our troops advanced on the southern heights of the Ancre to the Hansa trench half-way to Grandcourt.
The task of the centre was to take the fortress of Beaumont-Hamel, including the forked ravine to the south which required a prolonged and desperate struggle. The work was done by Highland Territorials before the early November sunset; and meanwhile the Naval Division on their right drove the Germans out of their first two lines on the northern bank of the Ancre towards Beaucourt. One battalion penetrated almost to the village, but was held up in a perilous position owing to the resistance of a strong German redoubt on its flank and almost in its rear. It stood its ground throughout the day, and at night the surrender of the German redoubt to a couple of tanks opened the way for a general attack on Beaucourt on the 14th. It was stormed by the battalion which had been waiting outside it since the previous morning. German counter-attacks on the 15th were repulsed, and on the 17th a further advance was made to the Bois d'Hollande north of Grandcourt, while Canadians from the Regina trench established themselves near its western outskirts. Another avenue towards Bapaume had been opened up, but winter postponed any further advance, and the Somme campaign had come to an end.
It had proved a sort of inverted Verdun, and the comfort we had derived from that successful defence was now extracted by the Germans from their defence of Bapaume. The parallel was not exact, because while the German gains at Verdun narrowed down to a point, ours on the Somme expanded in a circle. Yet the arguments were substantially the same: the French at Verdun were willing to sell any number of acres for armies, and the Germans professed an equal content on the Somme. Each side contended in turn that the offensive was the more costly form of warfare, and then repudiated the contention when it came to attack itself; and there was not a great deal to choose between them so far as logic was concerned. It is also clear that the Germans would have been at least as successful at Verdun as we were on the Somme but for the relief afforded by counter-offensives elsewhere, and that we should have profited no more from the Somme than the Germans did from Verdun had our Somme campaign been interrupted by German offensives on other fronts. Nor was there much to choose in the way of casualties: our estimate of the German losses as approximating 600,000 was a reasonable guess, but our own casualties were well over 400,000. The French losses were lighter, but the two together cannot have been less than the German. The Germans on the Somme, like the French at Verdun, withdrew divisions to refit before they were hopelessly broken; but what was considered wisdom in the French was reckoned weakness in the Germans and the Prussian Guards, whose return to Berlin, concealed in furniture-vans to hide their pitiable plight, was graphically described in the English press by an imaginative American journalist, were really sent as a contribution to that immense effort in the East by which, in spite of the Somme campaign, Germany first closed the gaps in the crumbling Austrian front and then overran Rumania.
There was thus a good deal of justice in the German comparison between Verdun and the Somme. The fallacy lay in the facts that our offensive was not brought to a stand by a German counter-attack but by the advent of winter, that the moves elsewhere in the West were the French ripostes at Verdun in October and December and not German counter-offensives, and that their campaign in Rumania, in spite of its painful success, had no effect upon the vital situation in the West. That episode was against us, but the tendencies were in our favour; our losses might equal the German, but equal attrition would leave us paramount in the end, barring collapse on the part of a principal ally. It was the fundamental situation which led to the German proposals for peace at Christmas, and the superficial impression which provoked the simultaneous fall of the Asquith Government.
So, too, there was something superficial and unjust in the lay criticism of Sir Douglas Haig's generalship. "Tactics of the Stone Age," was Mr. Lloyd George's later comment, which should not have been made in public at the expense of a general for whose retention in the command he was himself responsible. Even Foch controlled the group of French armies which co-operated with us on the Somme without producing results of a different character; and it is idle to compare the achievements of the generalissimo of 1918 with those of the British commander on the Somme in 1916. Haig controlled the British forces in France and Flanders, but he had no jurisdiction beyond a mere fragment of the thousands of miles of front on which the war was waged. Neither he nor any other Entente general therefore enjoyed the strategical opportunities of a Falkenhayn, Hindenburg, or Ludendorff, who could direct their blows east or west as they pleased; and responsibility for the strategical conduct of the war rested not with the Entente generals but with the heterogeneous Governments which employed them. Each commander had to work in his own compartment and could not escape its limitations. Nor was the diversity merely one of military commands; there was also the Navy, upon which the whole Allied strategy hung, to be considered; and not only in the Entente, but in each of its several Governments there was, and there could be, no such unity of direction as was possible in the militarist Central Empires.
There was also something naive in the popular clamour for a general as a Deus ex machina. For, in spite of apparent exceptions, the tendency of the transition from heroic to democratic ages is to transfer both in war and in politics the decisive influence from the individual to the mass, from the protagonist to the private; and modern warfare, with its complexity and its science, has become mainly a matter of mechanics. Its hero is the mob, and its generals fight far away in the rear of the line of battle; even the telescope has given place to the telephone. Individual valour counts for little compared with accurate range-tests and spotting by waves of sound. Man has mastered nature only to become more dependent upon his servants, and the vast machinery which the modern general controls envelops him in its toils. He reaches his goal in a motor, and the race is won by the best machine. Generalship was but one of a vast number of factors which gave us control of the Bapaume Ridge but also prevented the Somme campaign from saving Rumania or spoiling the German defence against Russia.