Had it been possible to bring off these latter blows successfully, such a debacle of the German forces would have resulted that not only would the advance of the British First, Third, Fourth, and Fifth Armies have seriously threatened Valenciennes, but the rush of German reserves to stop the gap would have withdrawn pressure from before the French about Reims, and would probably have enabled them to advance on Mézières.

A plan for an attack in the vicinity of Arras had been considered shortly before the opening of the battle of the Somme on July 1, 1916; it was then dropped, only to be revived in October, when the plan contemplated was to drive in the northern flank of the Gommecourt salient. It was hoped to employ two battalions of forty-eight tanks each in this operation; but, as the tanks promised in January did not materialise until the end of April, this plan had to be continually modified.

Meanwhile a hostile operation began to take place which bid fair to filch from us the tactical advantage we had won during the preceding summer. Towards the end of February it became apparent that the Germans intended to evacuate the Gommecourt salient; and the recent construction of the Hindenburg Line suggested a rounding off of the right angle between Arras and Craonne.

The German retirement necessitated certain changes in the British plan of operations. The Fourth Army relieved the French between the Somme and Roye; the Third Army, consisting of five Corps and three Cavalry Divisions, was now to penetrate the German defences, and by marching on Cambrai turn the Hindenburg Line from Heninel to Marcoing; the First and Fifth Armies were to operate on the left and right flanks of the Third Army.

The success of the British plan of attack depended on penetrating not only the German front-line system, but also the Drocourt-Queant line within forty-eight hours of initiating the attack; for, by so doing, so severe a wound would be inflicted that the Germans would be forced to move their reserves towards Cambrai and Douai, and away from Soissons and Reims, where the main blow was eventually to fall. Time, therefore, was, as usual, the all-important factor—could the Drocourt-Queant line be penetrated before the enemy was able to assemble his reserves?

Tanks, it was decided, should assist in gaining this time, yet on April 1, after denuding the training grounds of both England and France, only 60 Mark I and Mark II tanks could be reckoned on for the battle.

There were three ways in which these sixty tanks could be used, either by concentrating the whole against one objective such as Monchy-le-Preux, if a penetration of the centre were required, or against Bullecourt, if an envelopment of the German left flank were considered necessary, or to allot a proportion of machines to each Army or Corps for minor “mopping up” operations.

The last-mentioned course was eventually adopted and the following allotment of machines made:

(i) Eight tanks, to the First Army to operate against the Vimy Heights and the village of Thelus.

(ii) Forty tanks to the Third Army, eight to operate with the XVIIth Corps north of the river Scarpe, and thirty-two to operate with the VIth and VIIth Corps south of the river Scarpe.