This OHL commentary on Ludendorff’s inspection of the Western Front in January 1917, downbeat though it is, nevertheless understates the poor condition of the German army following the attritional battles the year before. By the end of 1916, 60 per cent of all German divisions had fought at Verdun or the Somme or both. German casualties in the two battles reached some 750,000, giving a total for 1916 of 1.2 million (and 4 million since the beginning of the war).Footnote ²

Despite these terrible figures, manpower in the German field army was in fact still increasing at this period, particularly through calling up younger year groups and combing out fit soldiers from non-combat units. The real damage to the German army was more to its quality and morale than its size. The German official history commented that what still remained of the old, peace-trained German infantry bled to death on the Somme. For the first time there was doubt whether Germany would win the war.Footnote ³ In an important memorandum, written in October 1916 after two tours as a corps commander on the Somme, General der Infanterie Max von Boehn discussed why most German attacks and counter-attacks during the battle had failed. Six different divisions had passed through his corps on the Somme; according to him none were fit to carry out a successful attack, and this was common to all corps on the Western Front. He saw the root cause as the lower quality of units, which he ascribed ultimately to inadequate training both in Germany and in the field. He believed that the army could have coped with inferior manpower if it had had good junior officers, but by the end of 1916 they too were lacking.Footnote ⁴

Though Boehn may have been exaggerating to make a point, there is no doubt that the quality of many formations had suffered at this period. Hermann von Kuhl recorded in October 1916 that 15th Infantry Division, in Boehn’s corps, had completely failed, serious indiscipline had occurred and the men had not moved forward into their positions. Boehn said he had never come across such an appalling unit.Footnote ⁵ Boehn himself told Crown Prince Rupprecht, the army group commander, that men had shot at their own officers. Following these events the divisional commander was sacked and there was a series of court martials.Footnote ⁶ Disciplinary problems and failures of performance continued well into 1917. The next divisional commander was also removed, as were other senior officers, including the chief staff officer. When the third commander, Generalmajor Gerhard Tappen, arrived in September 1917, he was horrified by the division’s condition and the number of pending court martials.Footnote ⁷

Rupprecht piously thanked God in his diary that a case such as 15th Infantry Division was rare.Footnote ⁸ It may indeed have been unusually bad, but the division was not the only one of concern. Rupprecht himself referred to problems in ten other divisions in autumn 1916. Of these, 12th and 16th Infantry and 5th Ersatz Divisions had failed in action; 183rd, 212th and 221st Infantry and 22nd, 44th and 53rd Reserve Divisions were shaky or under-performing; 9th Infantry Division had been badly battered at Verdun.Footnote ⁹ Including the divisions Boehn mentioned, there are indications of similar problems or concerns about performance in 27 of the 118 divisions that fought at Verdun or the Somme or both, over one-fifth of them. These divisions are of course the ones for which we have evidence: there may well have been others.

In OHL’s judgement, by the end of 1916, the army was exhausted but morale was still good. A post-war Reichsarchiv analysis concluded that in general this was correct, as shown by the army’s stubborn resistance on the Somme, but that there were warning signals, such as the large number of officers and men surrendering at Verdun in December. OHL was aware of these signals but because of its faith in the German soldier failed to appreciate the gravity of the morale problem.Footnote ¹⁰ Whether or not OHL’s judgement on morale was right, it was clearly worried by the quality of its divisions and decided, for the first time, to assess formally the capabilities of German troops. In early November 1916, it therefore ordered formations to report weekly on the combat value, readiness for deployment or need for relief of every division.Footnote ¹¹

As ‘Weltkrieg’ commented, despite efforts to make all divisions as homogeneous as possible, nowhere near all were suitable for major battle and the combat value of even those that were suitable constantly changed during action; the highest demands were made of the best divisions. This increased the risk that reliable divisions would suffer long-term damage from excessive deployment in major battle. We can identify twenty-four that fought and suffered appreciable casualties at both Verdun and the Somme: OHL apparently saw them as the workhorses of the army during 1916.Footnote ¹² Strikingly, six of the divisions just described as of concern were from the workhorse group, implying that overuse had indeed caused them serious damage.Footnote ¹³

When Ludendorff attempted to convince Chancellor Theobald von Bethmann Hollweg in January 1917 that Germany should launch unrestricted submarine warfare, his strongest argument was the need to relieve pressure on the army: ‘We must spare the troops a second Somme battle.’ The poor state of the army in late 1916 therefore directly contributed to what has been called ‘the worst decision of the war’, which, by instigating US entry, ultimately cost Germany victory.Footnote ¹⁴

Efforts to Strengthen the German Army

From autumn 1916, OHL had been warning that the enemy offensive which was expected next spring would decide the very existence of the German people. This may have been partly propaganda aimed at the ‘remobilisation’ of the German war effort, but OHL undoubtedly did assess that the Entente would launch even heavier assaults than in 1916. The overall balance of forces – both Entente strength as well as German weakness – precluded German offensive action on land. As described in the Introduction, the army’s role in grand strategy at this period was therefore to hold out for the five months believed necessary for unrestricted submarine warfare to knock Britain out of the war. On 6 January, OHL issued an order stating that the task on the Western Front was now to organise the defence in every detail, as well as to create and train reserves.Footnote ¹⁶ It took various steps to improve the German army’s capabilities from the low point of late 1916. These included promulgating new defensive tactics, improving training at all levels, increasing the number of divisions available (including by withdrawing to the Hindenburg Line), reorganising the army on the Western Front and bringing new equipment into service.

The new tactics were known variously as mobile defence, elastic defence or defence in depth. Formally, they were introduced on 1 December 1916 with the issuing of the ‘Defensive Battle’ manual, but in fact they had grown up in stages since 1915, in particular during the battle of the Somme. Tactical control of the combined arms battle would be delegated to the divisional commander. The defence should be mobile, in depth and offensive; it was mainly to be conducted by machines not men. ‘Mobile’ meant that the defence would be conducted around and not in the forward position; ground could be given up temporarily but must be recaptured by counter-attack. ‘In depth’ meant that there should be several successive positions which the enemy could not engage in one operation; that the ground must be properly prepared with fixed defences such as trenches, machine gun positions, shelters and wire obstacles; and that units should be deployed throughout the defensive zone. ‘Offensive’ involved both the counter-attacks and offensive use of the artillery. ‘Machines not men’ required thinner deployment of manpower, especially in the forward positions, and greater use of firepower from machine guns, mortars and artillery. The division would be the main battle unit and would be allocated important extra resources to conduct the new form of defence.Footnote ¹⁷

It will be helpful to explain the nature of a defensive zone at this time in more detail. A position [Stellung] comprised several lines of trenches with barbed wire obstacles, as well as shelters – some made of reinforced concrete [Mannschafts-Eisenbetonunterstände] – and emplacements for machine guns and other infantry weapons constructed throughout a deep ‘intermediate area’ [Zwischengelände] to the rear.Footnote ¹⁸ A complete defensive zone now comprised a number of such positions one behind the other, to a depth of several kilometres (Figure 1.1).Footnote ¹⁹

Figure 1.1 German defences at Arras in early 1917 (not to scale)

There were risks in introducing these new tactics in the middle of a war, and Hindenburg stressed two. Conservatism and misunderstandings complicated change even in peacetime. More seriously, in abandoning a rigid defence and stressing the need for independent action, the new system made greater demands on the troops. Given the concerns about the quality of the officers and men, could the army of 1917 actually carry out the new tactics?Footnote ²⁰ In view of such doubts, it is not surprising that there was considerable opposition to certain aspects of the new tactics, especially the authorisation to give ground if necessary when attacked and the level at which artillery should be controlled. These concerns were addressed in a second edition of ‘Defensive Battle’, published in March 1917 and therefore the version in effect at the time of the spring offensive.

OHL knew that doctrine had to be converted into reality by training, and it tackled this task at two levels. In February 1917, it introduced courses on the new tactics for division and brigade commanders and staffs. The courses were clearly important to OHL, considerable resources were devoted to them and throughput of students was impressive. OHL was pleased with the result, but as Chapter 7 explains, there are questions about just how effective the courses were. The second level was training for the field formations of the army. In early 1917, OHL tried to arrange at least three weeks’ training for all divisions, but again Chapter 7 shows that how much divisions actually received varied widely.

OHL also took action to increase the number of divisions available on the Western Front. First, the withdrawal to the Hindenburg Line [Siegfried-Stellung] in February–March 1917 shortened the front, thus saving thirteen to fifteen divisions as well as considerable amounts of field and heavy artillery.Footnote ²¹ A further benefit was that the withdrawal disrupted a substantial part of the plan for the Nivelle offensive. All in all, the Germans considered it a great operational success and boost to morale. But as a major operation – thirty-six divisions from four of Army Group Rupprecht’s Armies took part – it occupied much of the army group’s attention in a crucial period before the Entente offensive. The French saved almost as many divisions from the shorter line, which they committed to another part of the offensive. The withdrawal also handed the Entente a propaganda victory by creating the impression that the Germans were retreating after being defeated in the battle of the Somme.Footnote ²²

Second, in November 1916, orders were given for the establishment of thirteen divisions, to be ready for action by the next March. They were a mixture of experienced soldiers and new recruits. The training of these units suffered badly from turnover of officers, late delivery of equipment, geographical separation of infantry and artillery – which affected practice of all-arms co-operation – and the harsh winter.Footnote ²³ The generals who inspected the new divisions in March noted other problems such as the insufficient skill of the mainly very young officers and NCOs. The physical and disciplinary state of the young recruits also left something to be desired. The inspecting officers’ reports were nevertheless, surprisingly perhaps, generally satisfactory; but they did recommend more training in quiet sectors.Footnote ²⁴ In keeping with the priority now given to the Western Front, ten of these divisions were deployed there. They were not fit to take part in the opening phases of the spring fighting, though they did start to appear later on. Meanwhile they could free up experienced divisions.

Another thirteen divisions were raised between November 1916 and the opening of the spring offensive. Formed largely from existing units, they were in principle immediately deployable. However, only four of them were fit for major battle on the Western Front: the others were a mixture of Landwehr and lower-quality formations that could only be used in quiet sectors in the west or on the Eastern Front.Footnote ²⁵ This ratio suggests that even though the German army’s manpower continued to increase at this period, it was scraping the barrel in terms of fully combat-capable divisions. Indeed, a draft for ‘Weltkrieg’ commented that the increase in the number of divisions only partly resulted in an increase in fighting power.Footnote ²⁶ But in terms of number alone, with the last of these creations, the total of infantry divisions in the field when the Entente offensive opened on 9 April 1917 was 228.Footnote ²⁷

The third step was to move divisions from the Eastern to the Western Front. Between mid-November 1916 and the opening of the spring offensive, twelve infantry divisions arrived in the west. Of these, eleven were in principle fit for major battle on the Western Front but their actual combat capability there could not be taken for granted.Footnote ²⁸ Experience in 1916 had shown that divisions which performed well on the Eastern Front could fail in the west. Eight of the arriving divisions which had been in the east for a long time had to familiarise themselves with new Western Front techniques, but not all of them received the necessary training.Footnote ²⁹ As we saw in the Introduction, OHL was reluctant to move more divisions from the east in early 1917. The Russian Revolution in March occurred too late to affect the initial deployment of divisions facing the Entente attack in France. Even in the second half of May, OHL’s Operations Department was recommending caution in stripping further divisions from an already thinly garrisoned Eastern Front, though it was possible to exchange fought-out divisions from the west with fresh ones from the east.Footnote ³⁰

Hindenburg and Ludendorff also intended to increase further the ratio of firepower to manpower, in an effort to counterbalance Entente superiority in matériel. There were three aspects to this – production, introduction of new weapons and organisation of the army to use them. Immediately after taking over OHL they launched the so-called Hindenburg Programme. The objective was to secure a dramatic increase in the production of weapons and equipment by May 1917 when the Entente offensive could be expected. Modern research, however, stresses that production had already greatly increased before they took over, casts doubts on their ambitious targets and demonstrates the chaos which ensued. In the long term, the strain on German society inflicted by the programme’s fantasy targets paved the way for revolution. In the short term, steel production actually fell in the six months to February 1917. In addition, at a time when military requirements for manpower continued to grow, the programme needed increasing numbers of workers who were in principle liable and fit for active military service. By early 1917, there were about 1 million men in this category. They could not, however, be called up because of the adverse effect on production for the programme. In fact, the army agreed to give up 125,000 skilled workers to help it.Footnote ³¹

Under the Third OHL, new weaponry was introduced into each of the three main components of combined arms battle – infantry, artillery and aviation – though given the lead time Hindenburg and Ludendorff could not take the credit. This new equipment had the greatest effect in the air: Albatros fighters brought into service from September 1916 outclassed anything the Entente had and made a major contribution to regaining air superiority.Footnote ³² New equipment for ground troops had a less dramatic effect. The most important new infantry weapon was the MG 08/15 light machine gun that would equip infantry companies. It was an unloved wartime compromise and in any case production and distribution were so delayed that few divisions had it by the beginning of the spring fighting. The artillery were to receive model 1916 field guns and howitzers; both had increased range but at the cost of added weight, which impeded changing position, a key tactic to defeat enemy counter-battery fire. In addition, as with the light machine gun, wartime compromises caused problems of reliability. It also took time to equip all batteries, some only receiving the new equipment in early 1918.Footnote ³³

The Third OHL could take credit for the reorganisation of some of the units and formations employing this equipment. Once again, the most significant change was in aviation, which was established as a separate arm of service. A new post of ‘commanding general of air forces’ [Kommandierender General der Luftstreitkräfte or Kogenluft] took control of all air-related assets – including aircraft, airships, balloons, anti-aircraft guns and searchlights – in the field and in Germany. The number of squadrons continued to grow, and in particular the formation of specialist fighter units [Jagdstaffel] contributed to German predominance in the air at this period.Footnote ³⁴ Field artillery batteries increased from 1,929 in autumn 1916 to 2,709 in summer 1917; the 7,130 heavy guns at the front in February 1917 represented the highest number reached during the war.Footnote ³⁵ The bulk of this weaponry was controlled by divisional artillery commanders, a new post created in February 1917 to integrate handling of field and heavy artillery. Finally, from summer 1916 the original establishment of one medium machine gun company per infantry regiment had tripled to one company per battalion; and guns per company had doubled from six to twelve. A range of specialist units was also available for allocation to divisions as required, including extra machine gun units, flamethrowers, gas troops and especially storm troops. OHL ordered the creation of more storm troop units, stressing that their main role was to improve the assault capabilities of the ordinary infantry through training.Footnote ³⁶

In the longer term, the various steps taken to strengthen the German army would produce benefits, but not all the changes had actually been implemented by the time the Entente attacked in April 1917. As Boehn said in his memorandum the previous October, ‘It therefore seems at least questionable whether the army will be up to the task of countering the great deployment of strength by our enemies in the West which OHL expects next spring.’Footnote ³⁷

Entente Plans, German Preparations

The Entente were well aware of the low state of the German army in late 1916. In November, Joffre organised an inter-allied conference at his headquarters in Chantilly, with the aim of repeating the agreement on joint action reached the previous November which had put the Central Powers under such pressure in 1916. The conference decided that in order to exploit the damage done to the enemy, offensive action would be maintained as far as possible over the winter. All the allied armies would be ready to launch general offensives from early February 1917. These would begin as soon as they could be synchronised on the different fronts, meaning that they should start within three weeks of each other.Footnote ³⁹

Events soon nullified these intentions. In Russia’s case, it was agreed that the Chantilly obligations would not come into force until ratified at a follow-up conference in Petrograd. When this finally began in February, it became clear that weather conditions and the deficiencies of the Russian and Rumanian armies precluded a major offensive before 1 May. By the end of March, the Russians were stating that the revolution and continuing transport problems would force a further delay to at least mid-May and possibly June or July. Nivelle, who had now replaced Joffre, tried but failed throughout this period to make the Russians adhere to the Chantilly programme.Footnote ⁴⁰ Ideally, the Italian offensive would take place at the same time as the Russian in order to put maximum pressure on Austria-Hungary. However, the Italians feared, with some justification, that they themselves would be attacked. This delayed preparation of their offensive until an agreement was reached in March on what help their allies would offer if needed.Footnote ⁴¹

As a result of these problems, only the French and British were in a position to launch an early offensive. Joffre’s original agreement with Haig called for British and French offensives in the Arras-Bapaume and Somme-Oise areas, followed by subsidiary French attacks on the Aisne and in Alsace. All would take place on a broad front and consist of a sequence of assaults to the maximum depth that could be supported by artillery, following each other as quickly as possible and aimed at destroying the enemy’s capacity for resistance.Footnote ⁴² Nivelle retained the idea of a number of attacks on a broad front to be carried out as quickly as possible. But he changed the overall aim as well as the respective importance of the attacks and greatly increased their tempo.

In his new scheme, the Anglo-French forces would break through and destroy the main enemy forces in open battle. To achieve this, the Arras-Bapaume and Somme-Oise attacks, though large-scale, would be preliminary operations to draw in German reserves. The French would then launch the main offensive on the Aisne. In choosing these sectors, Nivelle reasoned that the region between the coast and Flanders was too wet to allow an attack before summer. Arras-Bapaume was suitable for an early offensive, the British had already begun preparations and the German salient in the area was vulnerable to a converging attack. The Somme area to the south as far as the river Avre was unsuitable for operations because of the devastation wrought by the 1916 fighting. Next to it, the narrowness of the Oise sector made only a limited offensive possible. Eastwards, the Aisne-Champagne sector was the nearest area suitable for a major offensive. Its advantages included that an attack from here could converge with the British advance from Arras-Bapaume and that there was a favourable starting point for capturing the vital Chemin des Dames position. Nivelle apparently only formulated this reasoning after the offensive, but the British official historian at least found it ‘certainly cogent’.Footnote ⁴³

For the main offensive on the Aisne front, Nivelle believed that the technique he had successfully used at Verdun, overwhelming artillery bombardment followed by a rapid and massed infantry assault, would enable him to break through the entire German defensive system in one day. He stressed to his subordinates that all operations were to be properly prepared but must be conducted with decisiveness, audacity and speed; the watchword was ‘when lightning can be used it should be preferred even to cannon’.Footnote ⁴⁴ This triggered an informal competition between generals as to whose infantry would advance furthest and fastest, easily won by General Charles Mangin, commanding French Sixth Army. No doubt the earlier sacking of two corps commanders for expressing doubt in a breakthrough encouraged the others.Footnote ⁴⁵

Nivelle’s plan quickly ran into difficulties. Although Haig recorded a positive impression when they first met, he soon began to make various objections. He wanted the French to take over a greater length of the front; he was concerned that the offensive might be prolonged, thereby endangering the possibility of a later British attack in Flanders; and finally he wanted to delay the attack till May. Slightly later, Lloyd George’s plot to subordinate the British army in France to Nivelle led to an outraged reaction from Haig and Sir William Robertson, the Chief of the Imperial General Staff. Lloyd George backed down, Nivelle’s authority over the British was limited to the duration of the offensive, and even then there were exceptions to it.Footnote ⁴⁶

Still more seriously, senior French commanders expressed increasing doubts about the prospects for the offensive the closer it came. As early as December 1916, Pétain, commanding Army Group Centre [Groupe d’armées du centre] commented that the terrain chosen for the main attack was extremely difficult and suggested an alternative. Nivelle therefore established a new Reserve Army Group [Groupe d’armées de réserve] to command the attack. However, by late March 1917 even its commander, General Alfred Micheler, was expressing severe doubts. Such concerns culminated in an extraordinary politico-military council of war at Compiègne on 6 April, chaired by President Poincaré. Nivelle effectively only secured continued governmental backing for the offensive at this late hour by threatening to resign.Footnote ⁴⁷

Nivelle’s final orders, issued on 4 April after the German withdrawal to the Hindenburg Line, set the overall aim of the offensive as the destruction of most of the German army on the Western Front. He would achieve this in two stages: a series of separate attacks to break through followed by exploitation. In order of implementation, the breakthrough operations were a powerful but subsidiary British attack at Arras; a comparatively minor French operation on the Oise; the main element of the offensive, the French attack on the Aisne northwest of Reims including on the Chemin des Dames position; and closely linked with that, a smaller but still important attack in Champagne just to the east of Reims. A decoy attack in Alsace would distract German attention from the real assault front. During the exploitation stage, French, Belgian and British forces from the coast to Ypres would join the attack on the weakened enemy. In this stage, Nivelle hoped to reach Bruges, Ghent and Mons in Belgium and Mézières in France.Footnote ⁴⁸ An advance of this extent would have liberated most of France and much of Belgium (and was similar to the line actually reached in November 1918).

The planned operational role of the Arras attack was to draw in German reserves before the French offensive. The British would make a strong thrust to break through and advance towards Cambrai (some thirty-five kilometres from Arras). As Figure 1.1 shows, German defences in this area consisted of three and in some places four positions, with another being constructed further back (the Wotan-Stellung). This gave a depth of up to seven kilometres for the main positions, and fourteen kilometres including the incomplete Wotan-Stellung.Footnote ⁴⁹ Following an unprecedentedly heavy bombardment, the British intended to penetrate the main German positions on the first day. The northern and southern flanks would be covered by capture of the commanding Vimy Ridge and the village of Bullecourt respectively. Cavalry would exploit any breakthrough by advancing to the Wotan-Stellung, then preparing to move towards Cambrai and if possible Douai.Footnote ⁵⁰ Twenty-six infantry and three cavalry divisions would participate in these operations.Footnote ⁵¹

The subsidiary French offensive on the Oise had been intended to employ twenty infantry and two cavalry divisions.Footnote ⁵² The German withdrawal to the Hindenburg Line in March completely disrupted this plan, and most of the French formations in the sector were moved to take part in the Champagne attack. Under the revised plan, French Army Group North [Groupe d’armées du Nord] would attack two days after the British, on an eleven kilometre front between St Quentin and the Oise. How the attack developed would depend on how strongly the Germans resisted. Two divisions would launch an initial assault to determine this. If there was little resistance, three more would join in. If the Germans resisted strongly, artillery preparation would continue and an attack by all five divisions would be mounted three days later. Should this lead to a breakthrough, Army Group North would initially advance to Hirson (some fifty-five kilometres east of St Quentin).Footnote ⁵³

In the main offensive, French Fifth, Sixth and Tenth Armies, comprising forty-seven infantry and seven cavalry divisions in Reserve Army Group, would attack on a front of forty kilometres.Footnote ⁵⁴ The breakthrough would take place on the first day. As an example of what the plan called for, II Colonial Corps in Mangin’s Sixth Army was to advance up to ten kilometres (a depth never achieved on the Western Front since the start of trench warfare). This entailed penetrating four German positions in the Chemin des Dames sector, as well as crossing the difficult Ailette valley and three successive ridge lines. The exploitation stage would begin the same evening and the corps would advance to just east of Laon, fifteen kilometres from the start line. By the evening of the second day, the advance guard was to be thirty-three kilometres from the start line; cavalry would push twenty kilometres further.Footnote ⁵⁵ Fifth and Tenth Armies would make similar progress. At about the same date, French Fourth Army east of Reims would attack with eight divisions on a front of ten kilometres. Its immediate objectives were to reach and cross a line of commanding hills to the north, advancing four kilometres and penetrating three strong German positions. Next day Fourth Army would link up with the Fifth Army attack and also begin exploitation towards the Aisne, forty kilometres from its start line.Footnote ⁵⁶

What did the Germans know of these plans? At the turn of the year, OHL assessed that the German army faced multiple possible attacks, including in Italy, several areas on the Western Front, Rumania and the Eastern Front.Footnote ⁵⁷ It could not permanently provide all threatened areas with forces and fixed defences sufficient for a long battle. Timely intelligence was therefore crucial to deploy reinforcements and complete defensive preparations; it would also prevent the French achieving surprise, an essential precondition if they were to reach their ambitious tactical objectives.

As Chapter 5 shows, OHL was uncertain about British intentions until shortly before the offensive at Arras opened. In particular, Army Group Rupprecht and Sixth Army did not submit their finalised request for extra forces till the end of March. By the time of the British assault on 9 April, reinforcements were on their way but too late to avoid a heavy defeat. German intelligence work was more effective against the French, allowing OHL to identify enemy intentions with increasing confidence from the end of February and to convert intelligence assessments into practical countermeasures. The defending forces were substantially increased, and much labour put into both forward and rearward defences. Preparations against the French were therefore well advanced by the time of the infantry assault on 16 April.

These preparations also involved a complex reorganisation to achieve the most effective chain of command, described in Chapter 3. From early March, three German army groups covered the Western Front. Army Group Rupprecht faced the British in the north and Army Group Crown Prince the French. The newly formed Army Group Albrecht controlled the sector from east of Verdun to the Swiss border, a quiet front where formations came to recover from battle and consequently a pool from which reserves could be drawn. Table 1.1 sets out these command arrangements in more detail, focusing on the formations facing the offensive in mid-April 1917. To explain two possibly unfamiliar terms, an Army Detachment [Armee-Abteilung] was a small Army with its own headquarters; and a Gruppe [group] was a corps-level formation comprising two or more divisions.Footnote ⁵⁸ Table 1.2 shows the opposing forces, in the front and in reserve, once the two sides had concentrated before the opening of the Entente attack.

The Battles: Arras

The battle of Arras divided into four main phases. In the opening phase, 4 to 14 April, a preparatory bombardment of unprecedented weight enabled a successful infantry assault by twenty-three divisions. The first day of the attack in particular, 9 April, saw appreciable British gains, including the capture of Vimy Ridge; units further south made the longest single advance on the Western Front to date, six kilometres, and broke through all German prepared defences in the area. In addition the Germans suffered heavy casualties and loss of matériel. The British made some further gains over the next few days but were seriously hampered by bad weather – in 34th Division some men died of exposureFootnote ⁶⁵ – and the difficulty of getting forward over the shelled area. An even more significant factor was British inability to improvise and co-ordinate in the semi-open warfare which now began, especially when faced with increasingly effective German resistance.Footnote ⁶⁶ The attack was halted on 14 April. During this period one Australian and one British division made a failed assault at Bullecourt.Footnote ⁶⁷

The Introduction described the panic that the successful attack of 9 April caused in the German command. The reaction included sacking Sixth Army’s chief of staff and operations officer; the new chief was Oberst Fritz von Loßberg, the defensive expert. Later in the month the opportunity arose to move the Army commander, Generaloberst Ludwig Freiherr von Falkenhausen, who was replaced by General der Infanterie Otto von Below. The Germans made a planned withdrawal to their original third position, and brought up sizeable reinforcements.Footnote ⁶⁸

Despite the decreasing progress by the end of this first phase, Haig had agreed to support the French – whose own offensive had not yet opened – so there was no question of stopping the battle. However, its subsequent three phases were much less successful for the British. Haig insisted on a pause in the attack in order to ensure proper co-ordination. His original plan was for parts of three Armies to attack on 20 April. In the event the second phase did not open till 23 April, was confined almost entirely to one Army and ended the next day. Conditions were now much more difficult for the British, especially because their artillery no longer had the advantage of engaging long-identified targets. By contrast, the German artillery had greatly increased and were more effectively employed. In addition, the pause in operations enabled the Germans to deepen their defensive positions and deploy reserves properly for counter-attacks. The British therefore made only limited progress at the cost of quite heavy casualties.Footnote ⁶⁹

By now Haig was aware that the French offensive – which was launched on 16 April and which was the motive for the battle of Arras in the first place – had generally failed. He also knew that the French government was considering stopping it and replacing Nivelle. A conference in Paris on 20 April attended by Lloyd George had agreed that operations should continue, but that progress should be reviewed two weeks later. However, a discussion with the French Minister of War, Painlevé, on 26 April revived Haig’s concerns.Footnote ⁷⁰ His thinking now was to pursue Arras operations while the French offensive continued; but if or when it stopped, to mount an attack in Flanders to clear the U-boat bases there. Because of this possibility he was not willing to use troops in Flanders to relieve those at Arras. So the next phases of the offensive were fought mainly by British divisions that had already been engaged and were ‘tired and depleted’.Footnote ⁷¹ The description earlier of 34th Division’s experiences shows what this carefully chosen phrasing meant in human terms.

Phase 3 of the battle, a secondary British attack on 28–29 April to improve positions by gaining ground, was an almost complete failure.Footnote ⁷² Soon after, it became obvious that Nivelle was likely to be removed and the French offensive stopped. The original British objectives were therefore no longer relevant. Haig now intended to move the focus of operations to Flanders, where he would attack as soon as possible after the expected Italian and Russian offensives. At Arras, he merely aimed to improve British positions locally and pin down German forces.Footnote ⁷³ Accordingly, the fourth phase of the battle there opened on 3 May; it lasted little over twenty-four hours. The equivalent of ten British divisions attacked seven German. The Canadian Corps captured the tactically important village of Fresnoy, but otherwise the attack was poorly co-ordinated and failed almost completely in the face of German artillery fire, unsuppressed by British counter-battery work, and small-scale but vigorous local counter-attacks. Notably, the Germans did not have to commit the divisions held back for counter-attack. For the British, Phase 4 ‘was nothing less than a disaster’, and the Germans even retook Fresnoy a few days later.Footnote ⁷⁴

Although this was the last major engagement in the battle of Arras proper, related fighting continued into June. The largest-scale clash was the second battle of Bullecourt, lasting from 3 to 17 May, with three British and three Australian divisions engaged against the equivalent of four German. The British and Australians made very limited gains in a battle that had ‘the reputation of a killing match, typifying trench warfare at its most murderous’.Footnote ⁷⁵ Fighting continued in May over various localities, with both sides having some success. As late as 28 June, five British and Canadian divisions made a limited attack to gain tactically useful ground. At a strategic level, these operations were intended to distract German attention in the gap between the battle of Messines in early June and the opening of the Ypres offensive at the end of July; and to contribute to the wearing down of the German army agreed at Anglo-French conferences at the beginning of May.Footnote ⁷⁶

The Battles: The Nivelle Offensive

The story of the first French action, the subsidiary attack on the Oise, can be quickly told. An initial attack by three divisions on 13 April met fierce German resistance and made very limited progress. It was clear that the Germans were not planning to withdraw from the Hindenburg Line. The second phase was therefore cancelled, and French artillery resumed the preparatory bombardment for an assault provisionally set for 19 or 20 April. However, this depended on the arrival of artillery reinforcements and was never mounted.Footnote ⁷⁹

As Pétain had commented when he first saw it, the ground chosen for the main offensive was extremely difficult. The western sector of the attack faced the steep and rugged Chemin des Dames ridge terminating in the locally dominating Plateau de Californie (Winterberg to the Germans); the valley of the little river Ailette, which had become a nearly impassable bog; and one or more further ridges. A particular feature of this area was the many caverns formed by underground quarries, which could be used to shelter defending troops. To the east of the Chemin des Dames position was flat, low country that offered better going, including for tanks, but could be commanded by artillery fire from both flanks. The eastern sector of the main assault and the subsidiary attack by French Fourth Army in Champagne both faced naturally strong lines of hills.

The assault phase of the French offensive began on 16 April and made some progress at various points of the line. The best results came from the flat terrain at Berry au Bac on the Aisne, where French infantry, for the first time supported by tanks, advanced four-and-a-half kilometres. In the central sector of the Chemin des Dames ridge, the advance was about two kilometres. Over 10,500 German prisoners were captured during the day. These were respectable achievements compared with some earlier offensives but they fell far short of Nivelle’s objectives. II Colonial Corps, which was to have advanced fifteen kilometres on the first day, at furthest penetrated one-and-a-half kilometres; most of its units fared worse. The gains that were made came at the cost of heavy casualties. The tanks gave some useful support to the infantry but in general did not fulfil the hopes placed in them and nearly half were knocked out. By the evening of 16 April, the Germans realised that the long-awaited and feared French breakthrough attempt had failed.Footnote ⁸⁰

But Nivelle had no thought of abandoning the offensive and gave orders for the main effort to be shifted to the apparently promising Berry area. However, the assault there made only minor progress over the next few days. In contrast, a concentric attack on the hinge of the German line on the western flank of the battle persuaded the Germans to withdraw on 18 April. This led to the biggest French territorial gains of the offensive, to a depth of seven kilometres on a twelve kilometre front. But it proved impossible to exploit this success and the first phase of the offensive ended on 20 April.Footnote ⁸¹

Over this same period, French Fourth Army carried out its supporting attack in Champagne, east of Reims. On the first day, 17 April, it advanced up to two-and-a-half kilometres, penetrating two German positions at some points and taking over 2,000 prisoners. However, it did not capture the line of hills and their northern slopes which were the objective. In the next few days it made limited progress, including occupying some of the hills. But the attack had run out of impetus and, like the main offensive, it was halted on 20 April.Footnote ⁸²

Nivelle was now subject to conflicting pressures. His initial decision was to pursue the original aims of the offensive. On 21 April he asked the British to continue and even expand their operations.Footnote ⁸³ However, the same day Micheler, the commander in charge of the main French effort, told Nivelle that a breakthrough was no longer possible. He cited the serious casualties, fatigue of the troops, bad weather and lack of ammunition; and there was no question of surprise. Micheler proposed instead two powerful but local attacks aimed at wearing down the enemy and improving French positions. In broadly accepting these ideas, Nivelle changed the nature of the offensive from breakthrough and rapid exploitation to much more limited objectives. By this stage too, the French government was increasingly concerned with the progress of operations and intervened to limit their scale. The 29 April appointment of Pétain as chief of the general staff with expanded powers signalled its loss of confidence in Nivelle.Footnote ⁸⁴

The new series of operations began on 30 April with a six-division attack by Fourth Army aimed at firmly capturing the line of hills to its front. The attack made little progress, at the cost of heavy casualties, and renewed efforts over the next few days suffered the same fate or worse.Footnote ⁸⁵ Nor was the four-division assault by Fifth Army south of Berry on 4 May any more successful. In view of the German resistance, the Army commander recommended against continuing and this was agreed the same day.Footnote ⁸⁶ On 4–5 May, a three-division attack by Tenth Army had better luck. In particular, it captured the Plateau de Californie, took almost 1,200 prisoners and beat off a series of counter-attacks. To its left, ten divisions of Sixth Army attacked with tank support on 5 and 6 May. They made a limited advance and captured 3,800 prisoners but failed to achieve their objective of taking the whole of the Chemin des Dames ridge; the attack was called off. On 15 May, Nivelle was sacked and Pétain became French commander-in-chief.Footnote ⁸⁷

In what may be seen as a third phase of the offensive from 20 to 25 May, the French captured the locally commanding Mont Cornillet in Champagne in a four-division attack; however, owing to poor co-ordination, an attack by elements of three divisions on the Aisne made limited progress.Footnote ⁸⁸ By this stage mutiny had broken out in the French army following the high casualties and disappointing results of the battles since mid-April. The French therefore now ended full-scale offensive action. But this did not mean the end of fighting in the area: the German army recorded fifteen local actions from the beginning of June to mid-July, mainly German initiatives to recapture tactically important ground lost in April–May. Despite the mutinies, German attackers generally met fierce French resistance and this fighting drew in German formations that could better have been used against the British in Flanders. Operations gradually declined after mid-July, but nevertheless there were a further eight local actions up till the battle of Malmaison in late October, when a successful French attack finally captured the whole Chemin des Dames ridge.Footnote ⁸⁹

The Battles: Messines

While the fighting continued, Anglo-French conferences in Paris on 4–5 May agreed that a decisive breakthrough was no longer feasible in 1917. However, it was essential to prevent a German attack and to wear down German strength until American forces had arrived in sufficient numbers to enable a final offensive in 1918. These aims would be achieved by launching a series of powerful attacks with limited objectives; the British would take on the main operations.Footnote ⁹¹ Haig envisaged that these operations would be in three stages. First, the Arras battle would continue in order to wear down and mislead the Germans. Next, the Messines Ridge would be captured as a preliminary operation to the third stage. This final stage, the Flanders operation, would be launched some weeks later and would aim to secure the Belgian coast.Footnote ⁹²

By mid-May the British were beginning to doubt that the French would carry out their side of the Paris agreement, and the War Cabinet told Haig it would not authorise his offensive plans if the French did not co-operate. To clarify their intentions, Haig met Pétain as French commander-in-chief for the first time on 18 May. Pétain told him about the unrest in the French army, but said he nevertheless planned to make four attacks including at Malmaison on 10 June and Verdun in late July; in addition French First Army would take part in the Flanders offensive.Footnote ⁹³ These French plans were badly disrupted by the mutinies, which were at their most serious between 15 May and 30 June. Pétain informed Haig on 2 June, five days before the opening of the battle of Messines, that he would be unable to launch the Malmaison attack; the earliest French attack would now be the Verdun operation in late July. Haig decided not to tell the War Cabinet that the French could no longer co-operate fully as originally agreed, and to go ahead with the battle of Messines anyway.Footnote ⁹⁴

The Messines plan called for an attack by twelve divisions on a front of sixteen kilometres, penetrating the defences to a depth of up to four kilometres; the assaulting infantry would be supported by the explosion of nineteen long-prepared mines and by seventy-two tanks. The battle lasted from 7 to 14 June, with the main fighting taking place on the first day. The British took all their objectives, and the fifteen German divisions involved in the defence lost 7,400 men captured.Footnote ⁹⁵ Hermann von Kuhl described Messines as one of the worst tragedies suffered by the German army during the war because the defeat could have been avoided by a local withdrawal, which had been considered but was rejected. A major reason for this decision was that intelligence on the scale of British mining activity was conflicting and that the German planning process therefore did not take full account of the severity of the threat.Footnote ⁹⁶

The Results of the Entente Offensives in Early 1917

In defeating the Entente breakthrough attempt and inflicting many more casualties than it suffered, the German army won a strategically important victory. True, the French made some useful tactical gains as well as capturing almost 29,000 prisoners and 187 guns.Footnote ⁹⁸ At an earlier period of the war, this would have been seen as success, but it bore no relation to Nivelle’s objectives, to the hopes of the French soldiers or to the casualties. Of particular significance was the collapse of the French army into mutiny. Although the mutinies were comparatively short-lived, they and the steps needed to restore military effectiveness left the army in a weakened state for the rest of the year.Footnote ⁹⁹ The British too made considerable tactical gains and captures at both Arras and Messines. The latter was a prelude to a much bigger operation, but the former had been intended as a breakthrough and in this light it failed completely. Nor did it achieve its operational role of attracting reserves from the French front.Footnote ¹⁰⁰

However, as Chapter 8 on performance shows, the German army itself was badly damaged by these battles and for the first time its field strength began to fall. Ludendorff believed that what saved the German army in early 1917 was the Entente’s failure to co-ordinate assaults in different theatres. He may well have been right given the scale of the Italian and Russian offensives that finally took place in May–June and July, respectively. The Italians deployed twenty-eight divisions for the tenth battle of the Isonzo in May, the largest yet, and a further ten for the battle of Asiago in June. In both cases, they suffered severe losses for almost no gains. They did draw in six-and-a-half Austro-Hungarian divisions from the Eastern Front, but the delay in launching the Kerensky offensive cancelled out the effect. The Russians and Rumanians attacked with 134 infantry and 27 cavalry divisions. They made good initial progress against the Austro-Hungarians, but were then pushed back by a powerful counter-offensive.Footnote ¹⁰¹

All this came too late to support the Anglo-French offensives. Nevertheless, as the Introduction described, the Italian and especially Russian attacks did have a significant effect on the Western Front: knowledge of the coming threats was one of the reasons why OHL refused to release forces to attack the French army at the moment of its maximum weakness during the mutinies. The British minor attacks at Arras, continuing into June, had the same effect.Footnote ¹⁰² Joffre’s and Nivelle’s grand strategy may not have worked for the offensive, but it contributed to saving France later.