[Sidenote: Skirting a lake.]

Toward the middle of the short winter afternoon the gorge we had been following opened out into a narrow valley, and straight over across the little lake which the road skirted, reflected in the shimmering sheet of steaming water that the thaw was throwing out across the ice, was a vivid white triangle of towering mountain. A true granite Alp among the splintered Dolomites—a fortress among cathedrals—it was the outstanding, the dominating feature in a panorama which I knew from my map was made up of the mountain chain along which wriggled the interlocked lines of the Austro-Italian battle front.

"Plainly a peak with a personality," I said to the officer at my side. "What is it called?"

[Sidenote: The Col di Lana an important position.]

"It's the Col di Lana," was the reply; "the mountain Colonel 'Peppino' Garibaldi took in a first attempt and Gelasio Caetani, the Italo-American mining engineer, afterward blew up and captured completely. It is one of the most important positions on our whole front, for whichever side holds it not only effectually blocks the enemy's advance, but has also an invaluable sally-port from which to launch his own. We simply had to have it, and it was taken in what was probably the only way humanly possible. It's Colonel Garibaldi's headquarters, by the way, where we put up to-night and to-morrow; perhaps you can get him to tell you the story." . . .

[Sidenote: The story of the Col di Lana.]

By the light of a little spirit lamp and to the accompaniment of a steady drip of eaves and the rumble of distant avalanches of falling snow, Colonel Garibaldi, that evening, told me "the story:"

[Sidenote: Legion Italienne withdrawn]

"The fighting that fell to the lot of the Legion Italienne in January, 1915, reduced its numbers to such an extent that it had to be withdrawn to rest and reform. Before it was in condition to take the field again, our country had taken the great decision and we were disbanded to go home and fight for Italy. Here—principally because it was thought best to incorporate the men in the units to which they (by training or residence) really belonged—it was found impracticable to maintain the integrity of the fourteen battalions—about 14,000 men in all—we had formed in France, and, as a consequence, the Legion Italienne ceased to exist except as a glorious memory. We five surviving Garibaldi were given commissions in a brigade of Alpini that is a 'lineal descendant' of the famous Cacciatore formed by my grandfather in 1859, and led by him against the Austrians in the war in which, with the aid of the French, we redeemed Lombardy for Italy.

[Sidenote: Defensive and offensive advantages of the peak.]

[Sidenote: Bitter struggle for the Col di Lana.]

"In July I was given command of a battalion occupying a position at the foot of the Col di Lana. Perhaps you saw from the lake, as you came up, the commanding position of this mountain. If so, you will understand its supreme importance to us, whether for defensive or offensive purposes. Looking straight down the Cordevole Valley toward the plains of Italy, it not only furnished the Austrians an incomparable observation post, but also stood as an effectual barrier against any advance of our own toward the Livinallongo Valley and the important Pordoi Pass. We needed it imperatively for the safety of any line we established in this region; and just as imperatively would we need it when we were ready to push the Austrians back. Since it was just as important for the Austrians to maintain possession of this great natural fortress as it was for us to take it away from them, you will understand how it came about that the struggle for the Col di Lana was perhaps the bitterest that has yet been waged for any one point on the Alpine front.

[Sidenote: The Alpini get a foothold.]

[Sidenote: Col. Garibaldi takes command.]

"Early in July, under cover of our guns to the south and east, the Alpini streamed down from the Cima di Falzarego and Sasso di Stria, which they had occupied shortly before, and secured what was at first but a precarious foothold on the stony lower eastern slope of the Col di Lana. Indeed, it was little more than a toe-hold at first; but the never-resting Alpini soon dug themselves in and became firmly established. It was to the command of this battalion of Alpini that I came on the 12th of July, after being given to understand that my work was to be the taking of the Col di Lana regardless of cost.

[Sidenote: Scientific man-saving needed.]

"This was the first time that I—or any other Garibaldi, for that matter (my grandfather, with his 'Thousand,' took Sicily from fifty times that number of Bourbon soldiers) had ever had enough, or even the promise of enough, men to make that 'regardless of cost' formula much more than a hollow mockery. But it is not in a Garibaldi to sacrifice men for any object whatever if there is any possible way of avoiding it. The period of indiscriminate frontal attacks had passed even before I left France, and ways were already being devised—mostly mining and better artillery protection—to make assaults less costly. Scientific 'man-saving,' in which my country has since made so much progress, was then in its infancy on the Italian front.

[Sidenote: Out-gunned by the Austrians.]

[Sidenote: First time of gallery-barracks.]

"I found many difficulties in the way of putting into practice on the Col di Lana the man-saving theories I had seen in process of development in the Argonne. At that time the Austrians—who had appreciated the great importance of that mountain from the outset—had us heavily out-gunned while mining in the hard rock was too slow to make it worth while until some single position of crucial value hung in the balance. So—well, I simply did the best I could under the circumstances. The most I could do was to give my men as complete protection as possible while they were not fighting, and this end was accomplished by establishing them in galleries cut out of the solid rock. This was, I believe, the first time the 'gallery-barracks'—now quite the rule at all exposed points—were used on the Italian front.

[Sidenote: Working under heavy fire.]

"There was no other way in the beginning but to drive the enemy off the Col di Lana trench by trench, and this was the task I set myself to toward the end of July. What made the task an almost prohibitive one was the fact that the Austrian guns from Corte and Cherz—which we were in no position to reduce to silence—were able to rake us unmercifully. Every move we made during the next nine months was carried out under their fire, and there is no use in denying that we suffered heavily. I used no more men than I could possibly help using, and the Higher Command was very generous in the matter of reserves, and even in increasing the strength of the force at my disposal as we gradually got more room to work in. By the end of October my original command of a battalion had been increased largely.

[Sidenote: Austrians hold one side and summit.]

[Sidenote: Austrian position seems impregnable.]

"The Austrians made a brave and skilful defense, but the steady pressure we were bringing to bear on them gradually forced them back up the mountain. By the first week in November we were in possession of three sides of the mountain, while the Austrians held the fourth side and—but most important of all—the summit. The latter presented a sheer wall of rock, more than 200 metres high, to us from any direction we were able to approach it, and on the crest of this cliff—the only point exposed to our artillery fire—the enemy had a cunningly concealed machine-gun post served by fourteen men. Back and behind, under shelter in a rock gallery, was a reserve of 200 men, who were expected to remain safely under cover during a bombardment and then sally forth to any infantry attack that might follow it. The handful in the machine-gun post, it was calculated, would be sufficient, and more than sufficient, to keep us from scaling the cliff before their reserves came up to support them; and so they would have been if there had been only an infantry attack to reckon with. It failed to allow sufficiently, however, for the weight of the artillery we were bringing up, and the skill of our gunners. The apparent impregnability of the position was really its undoing.

[Sidenote: Machine-gun post key position.]

"This cunningly conceived plan of defense I had managed to get a pretty accurate idea of—no matter how—and I laid my own plans accordingly. All the guns I could get hold of I had emplaced in positions most favorable for concentrating on the real key to the summit—the exposed machine-gun post on the crown of the cliff—with the idea, if possible, of destroying men and guns completely, or, failing in that, at least to render it untenable for the reserves who would try to rally to its defense.

[Sidenote: The Alpino thoroughly dependable.]

"We had the position ranged to an inch, and so, fortunately, lost no time in 'feeling' for it. This, with the surprise incident to it, was perhaps the principal element in our success; for the plan—at least so far as taking the summit was concerned—worked out quite as perfectly in action as upon paper. That is the great satisfaction of working with the Alpino, by the way: he is so sure, so dependable, that the 'human fallibility' element in a plan (always the most uncertain quantity) is practically eliminated.

[Sidenote: Alpini scale the cliff.]

"It is almost certain that our sudden gust of concentrated gunfire snuffed out the lives of all the men in the machine-gun post before they had time to send word of our developing infantry attack to the reserves in the gallery below. At any rate, these latter made no attempt whatever to swarm up to the defense of the crest, even after our artillery fire ceased. The consequence was that the 120 Alpini I sent to scale the cliff reached the top with only three casualties, these probably caused by rolling rocks or flying rock fragments. The Austrians in their big 'funk-hole' were taken completely by surprise, and 130 of them fell prisoners to considerably less than that number of Italians. The rest of the 200 escaped or were killed in their flight.

[Sidenote: Difficulties of holding the summit.]

[Sidenote: An Austrian counter-attack.]

"So far it was so good; but, unfortunately, taking the summit and holding it were two entirely different matters. No sooner did the Austrians discover what had happened than they opened on the summit with all their available artillery. We have since ascertained that the fire of 120 guns was concentrated upon a space of 100 by 150 metres which offered the only approach to cover that the barren summit afforded. Fifty of my men, finding shelter in the lee of rocky ledges, remained right out on the summit; the others crept over the edge of the cliff and held on by their fingers and toes. Not a man of them sought safety by flight, though a retirement would have been quite justified, considering what a hell the Austrians' guns were making of the summit. The enemy counter-attacked at nightfall, but despite superior numbers and the almost complete exhaustion of that little band of Alpini heroes, they were able to retake only a half of the summit. Here, at a ten-metres-high ridge which roughly bisects the cima, the Alpini held the Austrians, and here, in turn, the latter held the reinforcements which I was finally able to send to the Alpini's aid. There, exposed to the fire of the guns of either side (and so comparatively safe from both), a line was established from which there seemed little probability that one combatant could drive the other, at least without a radical change from the methods so far employed.

[Sidenote: Idea of blowing up positions.]

"The idea of blowing up positions that cannot be taken otherwise is by no means a new one. Probably it dates back almost as far as the invention of gunpowder itself. Doubtless, if we only knew of them, there have been attempts to mine the Great Wall of China. It was, therefore, only natural that, when the Austrians had us held up before a position it was vitally necessary we should have, we should begin to consider the possibility of mining it as the only alternative. The conception of the plan did not necessarily originate in the mind of any one individual, however many have laid claim to it. It was the inevitable thing if we were not going to abandon striving for our objective.

[Sidenote: Engineering operation of great magnitude.]

"But while there was nothing new in the idea of the mine itself, in carrying out an engineering operation of such magnitude at so great an altitude and from a position constantly exposed to intense artillery fire there were presented many problems quite without precedent. It was these problems which gave us pause; but finally, despite the prospect of difficulties which we fully realized might at any time become prohibitive, it was decided to make the attempt to blow up that portion of the summit of the Col di Lana still held by the enemy.

[Sidenote: Gelasio Caetani the engineer.]

"The choice of the engineer for the work was a singularly fortunate one. Gelasio Caetani—he is a son of the Duke of Sermoneta—had operated as a mining engineer in the American West for a number of years previous to the war, and the practical experience gained in California and Alaska was invaluable preparation for the great task now set for him. His ready resource and great personal courage were also incalculable assets.

[Sidenote: Miners from North America.]

"Well, the tunnel was started about the middle of January, 1916. Some of my men—Italians who had hurried home to fight for their country when the war started—had had some previous experience with hand and machine drills in the mines of Colorado and British Columbia, but the most of our labor had to gain its experience as the work progressed. Considering this, as well as the difficulty of bringing up material (to say nothing of food and munitions), we made very good progress.

[Sidenote: Mining under constant fire.]

[Sidenote: Thirty-eight shells a minute.]

"The worst thing about it all was the fact that it had to be done under the incessant fire of the Austrian artillery. I provided for the men as best as I could by putting them in galleries, where they were at least able to get their rest. When the enemy finally found out what we were up to they celebrated their discovery by a steady bombardment which lasted for fourteen days without interruption. During a certain forty-two hours of that fortnight there was, by actual count, an average of thirty-eight shells a minute exploding on our little position.

[Sidenote: Silencing an Austrian battery.]

"We were constantly confronted with new and perplexing problems—things which no one had ever been called upon to solve before—most of them in connection with transportation. How we contrived to surmount one of these I shall never forget. The Austrians had performed a brave and audacious feat in emplacing one of their batteries at a certain point, the fire from which threatened to make our position absolutely untenable. The location of this battery was so cunningly chosen that not one of our guns could reach it; and yet we had to silence it—and for good—if we were going to go on with our work. The only point from which we could fire upon these destructive guns was so exposed that any artillery we might be able to mount there could only count on the shortest shrift under the fire of the hundred or more 'heavies' that the Austrians would be able to concentrate upon it. And yet (I figured), well employed, these few minutes might prove enough to do the work in.

[Sidenote: A young giant endeavors to climb with a gun.]

"And then there arose another difficulty. The smallest gun that would stand a chance of doing the job cut out for it weighed 120 kilos—about 265 pounds; this just for the gun alone, with all detachable parts removed. But the point where the gun was to be mounted was so exposed that there was no chance of rigging up a cable-way, while the incline was so steep and rough that it was out of the question to try to drag it up with ropes. Just as we were on the verge of giving up in despair, one of the Alpini—a man of Herculean frame who had made his living in peace-time by breaking chains on his chest and performing other feats of strength—came and suggested that he be allowed to carry the gun up on his shoulders. Grasping at a straw, I let him indulge in a few 'practice man[oe]uvres'; but these only showed that, while the young Samson could shoulder and trot off with the gun without great effort, the task of lifting himself and his burden from foothold to foothold in the crumbling rock of the seventy-degree slope was too much for him.

[Sidenote: Men pull man and gun to position.]

"But out of this failure there came a new idea. Why not let my strong man simply support the weight of the gun on his shoulder—acting as a sort of ambulant gun-carriage, so to speak—while a line of men pulled him along with a rope?

We rigged up a harness to equalize the pull on the broad back, and, with the aid of sixteen ordinary men, the feat was accomplished without a hitch. I am sorry to say, however, that poor Samson was laid up for a spell with racked muscles.

"The gun—with the necessary parts and munition—was taken up in the night, and at daybreak it was set up and ready for action. It fired just forty shots before the Austrian 'heavies' blew it—and all but one or two of its brave crew—to pieces with a rain of high-explosive. But the troublesome Austrian battery was put so completely out of action that the enemy never thought it worth while to re-emplace it.

[Sidenote: Italians mine and Austrians countermine.]

[Sidenote: The final explosion.]

"That is just a sample of the fantastic things we were doing all of the three months that we drove the tunnel under the summit of the Col di Lana. The last few weeks were further enlivened by the knowledge that the Austrians were countermining against us. Once they drove so near that we could feel the jar of their drills, but they exploded their mine just a few metres short of where it would have upset us for good and for all. All the time work went on until, on the 17th of April, the mine was finished, charged, and 'tamped.' That night, while every gun we could bring to bear rained shell upon the Austrian position, it was exploded. A crater 150 feet in diameter and sixty feet deep engulfed the ridge the enemy had occupied, and this our waiting Alpini rushed and firmly held. Austrian counterattacks were easily repulsed, and the Col di Lana was at last completely in Italian hands."

* * * * *

During the late spring and summer of 1916, there was waged in France that great series of battles participated in by both British and French armies known as the battles of the Somme. Next to the defense of Verdun, they formed the most important military operations on the western front during that year. These battles are described in the narrative which follows.



THE BATTLE OF THE SOMME

SIR DOUGLAS HAIG

[Sidenote: An offensive summer campaign planned.]

The principle of an offensive campaign during the Summer of 1916 had already been decided on by all the Allies. The various possible alternatives on the western front had been studied and discussed by General Joffre and myself, and we were in complete agreement as to the front to be attacked by the combined French and British armies. Preparations for our offensive had made considerable progress; but as the date on which the attack should begin was dependent on many doubtful factors, a final decision on that point was deferred until the general situation should become clearer.

[Sidenote: British armies and supplies increasing.]

Subject to the necessity of commencing operations before the Summer was too far advanced, and with due regard to the general situation, I desired to postpone my attack as long as possible. The British armies were growing in numbers and the supply of munitions was steadily increasing. Moreover, a very large proportion of the officers and men under my command were still far from being fully trained, and the longer the attack could be deferred the more efficient they would become. On the other hand, the Germans were continuing to press their attacks at Verdun, and both there and on the Italian front, where the Austrian offensive was gaining ground, it was evident that the strain might become too great to be borne unless timely action were taken to relieve it. Accordingly, while maintaining constant touch with General Joffre in regard to all these considerations, my preparations were pushed on, and I agreed, with the consent of his Majesty's Government, that my attack should be launched, whenever the general situation required it, with as great a force as I might then be able to make available.

[Sidenote: Pressure on Italian front.]

[Sidenote: Heroic French defense at Verdun.]

By the end of May, 1916, the pressure of the enemy on the Italian front had assumed such serious proportions that the Russian campaign was opened early in June, and the brilliant successes gained by our allies against the Austrians at once caused a movement of German troops from the western to the eastern front. This, however, did not lessen the pressure on Verdun. The heroic defense of our French allies had already gained many weeks of inestimable value and had caused the enemy very heavy losses; but the strain continued to increase. In view, therefore, of the situation in the various theatres of war, it was eventually agreed between General Joffre and myself that the combined French and British offensive should not be postponed beyond the end of June.

[Sidenote: Objects of new offensive.]

The object of that offensive was threefold:

(i.) To relieve the pressure on Verdun.

(ii.) To assist our allies in the other theatres of war by stopping any further transfer of German troops from the western front.

(iii.) To wear down the strength of the forces opposed to us.

[Sidenote: Enemy attempts at interference.]

While my final preparations were in progress the enemy made two unsuccessful attempts to interfere with my arrangements. The first, directed on May 21, 1916, against our positions on the Vimy Ridge, south and southeast of Souchez, resulted in a small enemy gain of no strategic or tactical importance; and rather than weaken my offensive by involving additional troops in the task of recovering the lost ground, I decided to consolidate a position in rear of our original line.

[Sidenote: A position lost and retaken.]

The second enemy attack was delivered on June 2, 1916, on a front of over one and a half miles from Mount Sorrell to Hooge, and succeeded in penetrating to a maximum depth of 700 yards. As the southern part of the lost position commanded our trenches, I judged it necessary to recover it, and by an attack launched on June 13, 1916, carefully prepared and well executed, this was successfully accomplished by the troops on the spot.

Neither of these enemy attacks succeeded in delaying the preparations for the major operations which I had in view.

These preparations were necessarily very elaborate and took considerable time.

[Sidenote: Vast stores accumulated.]

[Sidenote: Shelter and communication facilities prepared.]

Vast stocks of ammunition and stores of all kinds had to be accumulated beforehand within a convenient distance of our front. To deal with these many miles of new railways—both standard and narrow gauge—and trench tramways were laid. All available roads were improved, many others were made, and long causeways were built over marshy valleys. Many additional dugouts had to be provided as shelter for the troops, for use as dressing stations for the wounded, and as magazines for storing ammunition, food, water, and engineering material. Scores of miles of deep communication trenches had to be dug, as well as trenches for telephone wires, assembly and assault trenches, and numerous gun emplacements and observation posts.

[Sidenote: Mining operations.]

Important mining operations were undertaken, and charges were laid at various points beneath the enemy's lines.

[Sidenote: Water supply insured.]

Except in the river valleys, the existing supplies of water were hopelessly insufficient to meet the requirements of the numbers of men and horses to be concentrated in this area as the preparations for our offensive proceeded. To meet this difficulty many wells and borings were sunk, and over one hundred pumping plants were installed. More than one hundred and twenty miles of water mains were laid, and everything was got ready to insure an adequate water supply as our troops advanced.

[Sidenote: Spirit of the troops.]

Much of this preparatory work had to be done under very trying conditions, and was liable to constant interruption from the enemy's fire. The weather, on the whole, was bad, and the local accommodations totally insufficient for housing the troops employed, who consequently had to content themselves with such rough shelter as could be provided in the circumstances. All this labor, too, had to be carried out in addition to fighting and to the everyday work of maintaining existing defenses. It threw a very heavy strain on the troops, which was borne by them with a cheerfulness beyond all praise.

[Sidenote: Formidable enemy position on the Somme and the Ancre.]

The enemy's position to be attacked was of a very formidable character, situated on a high, undulating tract of ground, which rises to more than 500 feet above sea level, and forms the watershed between the Somme on the one side and the rivers of Southwestern Belgium on the other. On the southern face of this watershed, the general trend of which is from east-southeast to west-northwest, the ground falls in a series of long irregular spurs and deep depressions to the valley of the Somme. Well down the forward slopes of this face the enemy's first system of defense, starting from the Somme near Curlu, ran at first northward for 3,000 yards, then westward for 7,000 yards to near Fricourt, where it turned nearly due north, forming a great salient angle in the enemy's lines.

Some 10,000 yards north of Fricourt the trenches crossed the River Ancre, a tributary of the Somme, and, still running northward, passed over the summit of the watershed, about Hebuterne and Gommecourt, and then down its northern spurs to Arras.

On the 20,000-yard front between the Somme and the Ancre the enemy had a strong second system of defense, sited generally on or near the southern crest of the highest part of the watershed, at an average distance of from 3,000 to 5,000 yards behind his first system of trenches.

[Sidenote: German methods of making position impregnable.]

During nearly two years' preparation he had spared no pains to render these defenses impregnable. The first and second systems each consisted of several lines of deep trenches, well provided with bomb-proof shelters and with numerous communication trenches connecting them. The front of the trenches in each system was protected by wire entanglements, many of them in two belts forty yards broad, built of iron stakes interlaced with barbed wire, often almost as thick as a man's finger.

[Sidenote: Veritable fortresses.]

[Sidenote: Machine-gun emplacements.]

The numerous woods and villages in and between these systems of defense had been turned into veritable fortresses. The deep cellars, usually to be found in the villages, and the numerous pits and quarries common to a chalk country were used to provide cover for machine guns and trench mortars. The existing cellars were supplemented by elaborate dugouts, sometimes in two stories, and these were connected up by passages as much as thirty feet below the surface of the ground. The salients in the enemy's lines, from which he could bring enfilade fire across his front, were made into self-contained forts, and often protected by mine fields, while strong redoubts and concrete machine-gun emplacements had been constructed in positions from which he could sweep his own trenches should these be taken. The ground lent itself to good artillery observation on the enemy's part, and he had skillfully arranged for cross-fire by his guns.

[Sidenote: A composite system of great strength.]

These various systems of defense, with the fortified localities and other supporting points between them, were cunningly sited to afford each other mutual assistance and to admit of the utmost possible development of enfilade and flanking fire by machine guns and artillery. They formed, in short, not merely a series of successive lines, but one composite system of enormous depth and strength.

[Sidenote: Many lines prepared in the rear.]

Behind this second system of trenches, in addition to woods, villages, and other strong points prepared for defense, the enemy had several other lines already completed; and we had learned from aeroplane reconnoisance that he was hard at work improving and strengthening these and digging fresh ones between them and still further back.

In the area above described, between the Somme and the Ancre, our front-line trenches ran parallel and close to those of the enemy, but below them. We had good direct observation on his front system of trenches and on the various defenses sited on the slopes above us between his first and second systems; but the second system itself, in many places, could not be observed from the ground in our possession, while, except from the air, nothing could be seen of his more distant defenses.

[Sidenote: The lines of the Allies.]

North of the Ancre, where the opposing trenches ran transversely across the main ridge, the enemy's defenses were equally elaborate and formidable. So far as command of ground was concerned we were here practically on level terms, but, partly as a result of this, our direct observation over the ground held by the enemy was not so good as it was further south. On portions of this front the opposing first-line trenches were more widely separated from each other, while in the valleys to the north were many hidden gun positions from which the enemy could develop flanking fire on our troops as they advanced across the open.

[Sidenote: Period of active operations.]

The period of active operations dealt with in this dispatch divides itself roughly into three phases. The first phase opened with the attack of July 1, 1916, the success of which evidently came as a surprise to the enemy and caused considerable confusion and disorganization in his ranks.

The advantages gained on that date and developed during the first half of July may be regarded as having been rounded off by the operations of July 14, 1916, and three following days, which gave us possession of the southern crest of the main plateau between Delville Wood and Bazentin-le-Petit.

[Sidenote: The enemy's efforts to hold the ridge.]

We then entered upon a contest lasting for many weeks, during which the enemy, having found his strongest defenses unavailing, and now fully alive to his danger, put forth his utmost efforts to keep his hold of the main ridge. This stage of the battle constituted a prolonged and severe struggle for mastery between the contending armies, in which, although progress was slow and difficult, the confidence of our troops in their ability to win was never shaken. Their tenacity and determination proved more than equal to their task, and by the first week in September they had established a fighting superiority that has left its mark on the enemy, of which possession of the ridge was merely the visible proof.

[Sidenote: The plateau gained.]

[Sidenote: Successes of the French south of the Somme]

The way was then opened for the third phase, in which our advance was pushed down the forward slopes of the ridge and further extended on both flanks until, from Morval to Thiepval, the whole plateau and a good deal of ground beyond were in our possession. Meanwhile our gallant allies, in addition to great successes south of the Somme, had pushed their advance, against equally determined opposition and under most difficult tactical conditions, up the long slopes on our immediate right, and were now preparing to drive the enemy from the summit of the narrow and difficult portion of the main ridge which lies between the Combles Valley and the River Tortille, a stream flowing from the north into the Somme just below Peronne.

[Sidenote: Careful artillery preparation.]

Defenses of the nature described could only be attacked with any prospect of success after careful artillery preparation. It was accordingly decided that our bombardment should begin on June 24, 1916 and a large force of artillery was brought into action for the purpose.

[Sidenote: Gas discharges.]

Artillery bombardments were also carried out daily at different points on the rest of our front, and during the period from June 24 to July 1, 1916, gas was discharged with good effect at more than forty places along our line upon a frontage which in total amounted to over fifteen miles. Some seventy raids, too, were undertaken by our infantry between Gommecourt and our extreme left north of Ypres during the week preceding the attack, and these kept me well informed as to the enemy's dispositions, besides serving other useful purposes.

[Sidenote: Attack by the Royal Flying Corps.]

On June 25, 1916, the Royal Flying Corps carried out a general attack on the enemy's observation balloons, destroying nine of them, and depriving the enemy for the time being of this form of observation.

[Sidenote: British and French co-operate in attack.]

On July 1, 1916, at 7.30 a. m., after a final hour of exceptionally violent bombardment, our infantry assault was launched. Simultaneously the French attacked on both sides of the Somme, co-operating closely with us.

The British main front of attack extended from Maricourt on our right, round the salient at Fricourt, to the Ancre in front of St. Pierre Divion. To assist this main attack by holding the enemy's reserves and occupying his artillery, the enemy's trenches north of the Ancre, as far as Serre, inclusive, were to be assaulted simultaneously, while further north a subsidiary attack was to be made on both sides of the salient at Gommecourt.

[Sidenote: Rawlinson and Allenby.]

I had intrusted the attack on the front from Maricourt to Serre to the Fourth Army, under the command of General Sir Henry S. Rawlinson, Bart., K. C. B., K. C. V. O., with five army corps at his disposal. The subsidiary attack at Gommecourt was carried out by troops from the army commanded by General Sir E. H. H. Allenby, K. C. B.

[Sidenote: Mines exploded under enemy lines.]

[Sidenote: Advance over open ground.]

[Sidenote: Trenches taken near Fricourt.]

Just prior to the attack the mines which had been prepared under the enemy's lines were exploded, and smoke was discharged at many places along our front. Through this smoke our infantry advanced to the attack with the utmost steadiness in spite of the very heavy barrage of the enemy's guns. On our right our troops met with immediate success, and rapid progress was made. Before midday Montauban had been carried, and shortly afterward the Briqueterie, to the east, and the whole of the ridge to the west of the village were in our hands. Opposite Mametz part of our assembly trenches had been practically leveled by the enemy artillery, making it necessary for our infantry to advance to the attack across 400 yards of open ground. None the less they forced their way into Mametz, and reached their objective in the valley beyond, first throwing out a defensive flank toward Fricourt on their left. At the same time the enemy's trenches were entered north of Fricourt, so that the enemy's garrison in that village was pressed on three sides. Further north, though the village of La Boisselle and Ovillers for the time being resisted our attack, our troops drove deeply into the German lines on the flanks of these strongholds, and so paved the way for their capture later.

[Sidenote: Fight for the Leipsic Salient.]

On the spur running south from Thiepval the work known as the Leipsic Salient was stormed, and severe fighting took place for the possession of the village and its defenses. Here and north of the valley of the Ancre, as far as Serre, on the left flank of our attack, our initial successes were not sustained. Striking progress was made at many points, and parties of troops penetrated the enemy's positions to the outer defenses of Grandcourt, and also to Pendant Copse and Serre; but the enemy's continued resistance at Thiepval and Beaumont Hamel made it impossible to forward reinforcements and ammunition, and in spite of their gallant efforts our troops were forced to withdraw during the night to their own lines.

[Sidenote: The attack at Gommecourt.]

The subsidiary attack at Gommecourt also forced its way into the enemy's positions, but there met with such vigorous opposition that as soon as it was considered that the attack had fulfilled its object our troops were withdrawn.

[Sidenote: Instructions to General Gough.]

In view of the general situation at the end of the first day's operations I decided that the best course was to press forward on a front extending from our junction with the French to a point half way between La Boisselle and Contalmaison, and to limit the offensive on our left for the present to a slow and methodical advance. North of the Ancre such preparations were to be made as would hold the enemy to his positions and enable the attack to be resumed there later if desirable. In order that General Sir Henry Rawlinson might be left free to concentrate his attention on the portion of the front where the attack was to be pushed home, I also decided to place the operations against the front, La Boisselle to Serre, under the command of General Sir Hubert de la P. Gough, K. C. B., to whom I accordingly allotted the two northern corps of Sir Henry Rawlinson's army. My instructions to Sir Hubert Gough were that his army was to maintain a steady pressure on the front from La Boisselle to the Serre road and to act as a pivot on which our line could swing as our attacks on his right made progress toward the north.

[Sidenote: Fricourt to Contalmaison.]

During the succeeding days the attack was continued on these lines. In spite of strong counter-attacks on the Briqueterie and Montauban, by midday on July 2 our troops had captured Fricourt, and in the afternoon and evening stormed Fricourt Wood and the farm to the north. During July 3 and 4 Bernajay and Caterpillar woods were also captured, and our troops pushed forward to the railway north of Mametz. On these days the reduction of La Boisselle was completed after hard fighting, while the outskirts of Contalmaison were reached on July 5. North of La Boisselle also the enemy's forces opposite us were kept constantly engaged, and our holding in the Leipsic Salient was gradually increased.

[Sidenote: Result of five days' fighting.]

[Sidenote: Prisoners taken.]

To sum up the results of the fighting of these five days, on a front of over six miles, from the Briqueterie to La Boisselle, our troops had swept over the whole of the enemy's first and strongest system of defense, which he had done his utmost to render impregnable. They had driven him back over a distance of more than a mile, and had carried four elaborately fortified villages. The number of prisoners passed back at the close of July 5, 1916, had already reached the total of ninety-four officers and 5,724 other ranks.

[Sidenote: Readjustments and reliefs.]

[Sidenote: Contalmaison and Mametz Wood.]

After the five days' heavy and continuous fighting just described it was essential to carry out certain readjustments and reliefs of the forces engaged. In normal conditions of enemy resistance the amount of progress that can be made at any time without a pause in the general advance is necessarily limited. Apart from the physical exhaustion of the attacking troops and the considerable distance separating the enemy's successive main systems of defense, special artillery preparation was required before a successful assault could be delivered. Meanwhile, however, local operations were continued in spite of much unfavorable weather. The attack on Contalmaison and Mametz Wood was undertaken on July 7, 1916, and after three days' obstinate fighting, in the course of which the enemy delivered several powerful counterattacks, the village and the whole of the wood, except its northern border, were finally secured. On July 7 also a footing was gained in the other defenses of Ovillers, while on July 9, 1916, on our extreme right, Maltz Horn Farm—an important point on the spur north of Hardecourt—was secured.

[Sidenote: British troops in Trones Wood.]

A thousand yards north of this farm our troops had succeeded at the second attempt in establishing themselves on July 8, 1916, in the southern end of Trones Wood. The enemy's positions in the northern and eastern parts of this wood were very strong, and no less than eight powerful German counterattacks were made here during the next five days. In the course of this struggle portions of the wood changed hands several times; but we were left eventually, on July 13, 1916, in possession of the southern part of it.

[Sidenote: Assault on the German second system of defense.]

Meanwhile Mametz Wood had been entirely cleared of the enemy, and with Trones Wood also practically in our possession we were in a position to undertake an assault upon the enemy's second system of defense. Arrangements were accordingly made for an attack to be delivered at daybreak on the morning of July 14, 1916, against a front extending from Longueval to Bazentin-le-Petit Wood, both inclusive. Contalmaison Villa, on a spur 1,000 yards west of Bazentin-le-Petit Wood, had already been captured to secure the left flank of the attack, and advantage had been taken of the progress made by our infantry to move our artillery forward into new positions. The preliminary bombardment had opened on July 11, 1916. The opportunities offered by the ground for enfilading the enemy's lines were fully utilized, and did much to secure the success of our attack.

[Sidenote: A night operation of magnitude.]

In the early hours of July 4, 1916, the attacking troops moved out over the open for a distance of from about 1,000 to 1,400 yards, and lined up in the darkness just below the crest and some 300 to 500 yards from the enemy's trenches. Their advance was covered by strong patrols, and their correct deployment had been insured by careful previous preparations. The whole movement was carried out unobserved and without touch being lost in any case. The decision to attempt a night operation of this magnitude with an army, the bulk of which had been raised since the beginning of the war, was perhaps the highest tribute that could be paid to the quality of our troops. It would not have been possible but for the most careful preparation and forethought, as well as thorough reconnoissance of the ground, which was, in many cases, made personally by divisional, brigade, and battalion commanders and their staffs before framing their detailed orders for the advance.

[Sidenote: The assault on July 14.]

The actual assault was delivered at 3.25 a.m. on July 14, 1916, when there was just sufficient light to be able to distinguish friend from foe at short ranges, and along the whole front attacked our troops, preceded by a very effective artillery barrage, swept over the enemy's first trenches and on into the defenses beyond.

[Sidenote: Trones Wood cleared of the enemy.]

[Sidenote: Longueval occupied.]

On our right the enemy was driven from his last foothold in Trones Wood, and by 8 a.m. we had cleared the whole of it, relieving a body of 170 men who had maintained themselves all night in the northern corner of the wood, although completely surrounded by the enemy. Our position in the wood was finally consolidated, and strong patrols were sent out from it in the direction of Guillemont and Longueval. The southern half of this latter village was already in the hands of the troops who had advanced west of Trones Wood. The northern half, with the exception of two strong points, was captured by 4 p.m. after a severe struggle.

[Sidenote: The enemy counterattacks.]

In the centre of our attack Bazentin-le-Grand village and wood were also gained, and our troops pushing northward captured Bazentin-le-Petit village and the cemetery to the east. Here the enemy counterattacked twice about midday without success, and again in the afternoon, on the latter occasion momentarily reoccupying the northern half of the village as far as the church. Our troops immediately returned to the attack and drove him out again with heavy losses. To the left of the village Bazentin-le-Petit Wood was cleared, in spite of the considerable resistance of the enemy along its western edge, where we successfully repulsed a counterattack. In the afternoon further ground was gained to the west of the wood, and posts were established immediately south of Pozieres.

[Sidenote: General Rawlinson employs cavalry.]

The enemy's troops, who had been severely handled in these attacks and counterattacks, began to show signs of disorganization, and it was reported early in the afternoon that it was possible to advance to High Wood. General Rawlinson, who had held a force of cavalry in readiness for such an eventuality, decided to employ a part of it. As the fight progressed small bodies of this force had pushed forward gradually, keeping in close touch with the development of the action, and prepared to seize quickly any opportunity that might occur. A squadron now came up on the flanks of our infantry, who entered High Wood at about 8 p.m., and, after some hand-to-hand fighting, cleared the whole of the wood with the exception of the northern apex. Acting mounted in co-operation with the infantry, the cavalry came into action with good effect, killing several of the enemy and capturing some prisoners.

[Sidenote: British withdrawn from High Wood.]

On July 15, 1916, the battle still continued, though on a reduced scale. Arrow Head Copse, between the southern edge of Trones Wood and Guillemont, and Waterlot Farm on the Longueval-Guillemont road, were seized, and Delville Wood was captured and held against several hostile counterattacks. In Longueval fierce fighting continued until dusk for the possession of the two strong points and the orchards to the north of the village. The situation in this area made the position of our troops in High Wood somewhat precarious, and they now began to suffer numerous casualties from the enemy's heavy shelling. Accordingly orders were given for their withdrawal, and this was effected during the night of July 15-16, 1916, without interference by the enemy. All the wounded were brought in.

[Sidenote: Progress toward Pozieres.]

In spite of repeated enemy counterattacks further progress was made on the night of July 16, 1916, along the enemy's main second-line trenches northwest of Bazentin-le-Petit Wood to within 500 yards of the northeast corner of the village of Pozieres, which our troops were already approaching from the south.

[Sidenote: Ovillers captured.]

Meanwhile the operations further north had also made progress. Since the attack of July 7, 1916, the enemy in and about Ovillers had been pressed relentlessly and gradually driven back by incessant bombing attacks and local assaults, in accordance with the general instructions I had given to General Sir Hubert Gough. On July 16, 1916, a large body of the garrison of Ovillers surrendered, and that night and during the following day, by a direct advance from the west across No Man's Land, our troops carried the remainder of the village and pushed out along the spur to the north and eastward toward Pozieres.

[Sidenote: A new line definitely established.]

The results of the operations of July 4, 1916, and subsequent days were of considerable importance. The enemy's second main system of defense had been captured on a front of over three miles. We had again forced him back more than a mile, and had gained possession of the southern crest of the main ridge on a front of 6,000 yards. Four more of his fortified villages and three woods had been wrested from him by determined fighting, and our advanced troops had penetrated as far as his third line of defense. In spite of a resolute resistance and many counterattacks, in which the enemy had suffered severely, our line was definitely established from Maltz Horn Farm, where we met the French left, northward along the eastern edge of Trones Wood to Longueval, then westward past Bazentin-le-Grand to the northern corner of Bazentin-le-Petit and Bazentin-le-Petit Wood, and then westward again past the southern face of Pozieres to the north of Ovillers. Posts were established at Arrow Head Copse and Waterlot Farm, while we had troops thrown forward in Delville Wood and toward High Wood, though their position was not yet secure.

[Sidenote: Sir Henry Rawlinson commended.]

I cannot speak too highly of the skill, daring endurance, and determination by which these results had been achieved. Great credit is due to Sir Henry Rawlinson for the thoroughness and care with which this difficult undertaking was planned; while the advance and deployment made by night without confusion, and the complete success of the subsequent attack, constitute a striking tribute to the discipline and spirit of the troops engaged, as well as to the powers of leadership and organization of their commanders and staffs.

[Sidenote: Guns and prisoners taken.]

During these operations and their development on the 15th a number of enemy guns were taken, making a total capture since July 1, 1916, of eight heavy howitzers, four heavy guns, forty-two field and light guns and field howitzers, thirty trench mortars, and fifty-two machine guns. Very considerable losses had been inflicted on the enemy, and the prisoners captured amounted to over 2,000, bringing the total since July 1, 1916, to over 10,000.

[Sidenote: Enemy able to bring up fresh troops.]

There was strong evidence that the enemy forces engaged on the battle front had been severely shaken by the repeated successes gained by ourselves and our allies; but the great strength and depth of his defenses had secured for him sufficient time to bring up fresh troops, and he had still many powerful fortifications, both trenches, villages, and woods, to which he could cling in our front and on our flanks.

We had, indeed, secured a footing on the main ridge, but only on a front of 6,000 yards, and desirous though I was to follow up quickly the successes we had won, it was necessary first to widen this front.

[Sidenote: Pozieres and Thiepval still to be carried.]

West of Bazentin-le-Petit the villages of Pozieres and Thiepval, together with the whole elaborate system of trenches around, between and on the main ridge behind them, had still to be carried. An advance further east would, however, eventually turn these defenses, and all that was for the present required on the left flank of our attack was a steady, methodical, step by step advance as already ordered.

[Sidenote: Salient at Delville, Wood and Longueval.]

On our right flank the situation called for stronger measures. At Delville Wood and Longueval our lines formed a sharp salient, from which our front ran on the one side westward to Pozieres, and on the other southward to Maltz Horn Farm. At Maltz Horn Farm our lines joined the French, and the allied front continued still southward to the village of Hem, on the Somme.

[Sidenote: Enemy's advantages.]

This pronounced salient invited counterattacks by the enemy. He possessed direct observation on it all around from Guillemont on the southeast to High Wood on the northwest. He could bring a concentric fire of artillery, to bear not only on the wood and village, but also on the confined space behind, through which ran the French communications as well as ours, where great numbers of guns, besides ammunition and impediments of all sorts, had necessarily to be crowded together. Having been in occupation of this ground for nearly two years, he knew every foot of it, and could not fail to appreciate the possibilities of causing us heavy loss there by indirect artillery fire; while it was evident that, if he could drive in the salient in our line and so gain direct observation on the ground behind, our position in that area would become very uncomfortable.

[Sidenote: Confidence in the troops]

If there had not been good grounds for confidence that the enemy was not capable of driving from this position troops who had shown themselves able to wrest it from him, the situation would have been an anxious one. In any case it was clear that the first requirement at the moment was that our right flank, and the French troops in extension of it, should swing up into line with our centre. To effect this, however, strong enemy positions had to be captured both by ourselves and by our allies.

[Sidenote: Plateau from Delville Wood to Morval]

[Sidenote: New enemy defenses.]

From Delville Wood the main plateau extends for 4,000 yards east-northeast to Les Boeufs and Morval, and for about the same distance southeastward to Leuze and Bouleau Woods, which stand above and about 1,000 yards to the west of Combles. To bring my right up into line with the rest of my front it was necessary to capture Guillemont, Falfemont Farm, and Leuze Wood, and then Ginchy and Bouleau Woods. These localities were naturally very strong, and they had been elaborately fortified. The enemy's main second-line system of defense ran in front of them from Waterlot Farm, which was already in our hands, southeastward to Falfemont Farm, and thence southward to the Somme. The importance of holding us back in this area could not escape the enemy's notice, and he had dug and wired many new trenches, both in front of and behind his original lines. He had also brought up fresh troops, and there was no possibility of taking him by surprise.

[Sidenote: Rain and unfavorable ground.]

[Sidenote: Constant haze.]

The task before us was, therefore, a very difficult one and entailed a real trial of strength between the opposing forces. At this juncture its difficulties were increased by unfavorable weather. The nature of the ground limited the possibility of direct observation of our artillery fire, and we were consequently much dependent on observation from the air. As in that element we had attained almost complete superiority, all that we required was a clear atmosphere; but with this we were not favored for several weeks. We had rather more rain than is usual in July and August, and even when no rain fell there was an almost constant haze and frequent low clouds.

[Sidenote: British and French must advance together.]

[Sidenote: Positions the French must capture.]

In swinging up my own right it was very important that the French line north of the Somme should be advanced at the same time in close combination with the movement of the British troops. The line of demarkation agreed on between the French commander and myself ran from Maltz Horn Farm due eastward to the Combles Valley and then northeastward up that valley to a point midway between Sailly-Saillisel and Morval. These two villages had been fixed upon as objectives, respectively, of the French left and of my right. In order to advance in co-operation with my right, and eventually to reach Sailly-Saillisel, our allies had still to fight their way up that portion of the main ridge which lies between the Combles Valley on the west and the River Tortille on the east. To do so they had to capture, in the first place, the strongly fortified villages of Maurepas, Le Forest, Rancourt, and Fregicourt, besides many woods and strong systems of trenches. As the high ground on each side of the Combles Valley commands the slopes of the ridge on the opposite side, it was essential that the advance of the two armies should be simultaneous and made in the closest co-operation. This was fully recognized by both armies, and our plans were made accordingly.

[Sidenote: A pause necessary.]

To carry out the necessary preparations to deal with the difficult situation outlined above a short pause was necessary, to enable tired troops to be relieved and guns to be moved forward; while at the same time old communications had to be improved and new ones made. Intrenchments against probable counterattacks could not be neglected, and fresh dispositions of troops were required for the new attacks to be directed eastward.

[Sidenote: Pressure on whole front.]

It was also necessary to continue such pressure on the rest of our front, not only on the Ancre, but further south, as would make it impossible for the enemy to devote himself entirely to resisting the advance between Delville Wood and the Somme. In addition, it was desirable further to secure our hold on the main ridge west of Delville Wood by gaining more ground to our front in that direction. Orders were therefore issued in accordance with the general considerations explained above, and, without relaxing pressure along the enemy's front from Delville Wood to the west, preparations for an attack on Guillemont were pushed on.

[Sidenote: Enemy counterattack on Delville Wood.]

During the afternoon of July 18, 1916, the enemy developed his expected counterattack against Delville Wood, after heavy preliminary shelling. By sheer weight of numbers, and at very heavy cost, he forced his way through the northern and northeastern portion of the wood and into the northern half of Longueval, which our troops had cleared only that morning. In the southeast corner of the wood he was held up by a gallant defense, and further south three attacks on our positions in Waterlot Farm failed.

[Sidenote: Progress bought by hard fighting.]

[Sidenote: Enemy in great strength.]

This enemy attack on Delville Wood marked the commencement of the long, closely contested struggle which was not finally decided in our favor till the fall of Guillemont on September 3, 1916, a decision which was confirmed by the capture of Ginchy six days later. Considerable gains were indeed made during this period, but progress was slow, and bought only by hard fighting. A footing was established in High Wood on July 20, 1916, and our line linked up thence with Longueval. A subsequent advance by the Fourth Army on July 23, 1916, on a wide front from Guillemont to Pozieres found the enemy in great strength all along the line, with machine guns and forward troops in shell holes and newly constructed trenches well in front of his main defenses. Although ground was won, the strength of the resistance experienced showed that the hostile troops had recovered from their previous confusion sufficiently to necessitate long and careful preparation before further successes on any great scale could be secured.

[Sidenote: Two powerful counterattacks.]

An assault delivered simultaneously on this date by General Gough's army against Pozieres gained considerable results, and by the morning of July 25, 1916, the whole of that village was carried, including the cemetery, and important progress was made along the enemy's trenches to the northeast. That evening, after heavy artillery preparation, the enemy launched two more powerful counterattacks, the one directed against our new position in and around High Wood and the other delivered from the northwest of Delville Wood. Both attacks were completely broken up with very heavy losses to the enemy.

[Sidenote: Delville Wood recovered.]

On July 27, 1916, the remainder of Delville Wood was recovered, and two days later the northern portion of Longueval and the orchards were cleared of the enemy, after severe fighting, in which our own and the enemy's artillery were very active.

[Sidenote: Fighting at Guillemont.]

On July 30, 1916, the village of Guillemont and Falfemont Farm to the southeast were attacked, in conjunction with a French attack north of the Somme. A battalion entered Guillemont, and part of it passed through to the far side; but as the battalions on either flank did not reach their objectives, it was obliged to fall back, after holding out for some hours on the western edge of the village. In a subsequent local attack on August 7, 1916, our troops again entered Guillemont, but were again compelled to fall back owing to the failure of a simultaneous effort against the enemy's trenches on the flanks of the village.

[Sidenote: Dominating enemy positions.]

[Sidenote: Series of French and British attacks.]

The ground to the south of Guillemont was dominated by the enemy's positions in and about that village. It was therefore hoped that these positions might be captured first, before an advance to the south of them in the direction of Falfemont Farm was pushed further forward. It had now become evident, however, that Guillemont could not be captured as an isolated enterprise without very heavy loss, and, accordingly, arrangements were made with the French Army on our immediate right for a series of combined attacks, to be delivered in progressive stages, which should embrace Maurepas, Falfemont Farm, Guillemont, Leuze Wood, and Ginchy.

[Sidenote: Attacks and counterattacks.]

An attempt on August 16, 1916, to carry out the first stage of the prearranged scheme met with only partial success, and two days later, after a preliminary bombardment lasting thirty-six hours, a larger combined attack was undertaken. In spite of a number of enemy counterattacks the most violent of which leveled at the point of junction of the British with the French, succeeded in forcing our allies and ourselves back from a part of the ground won—very valuable progress was made, and our troops established themselves in the outskirts of Guillemont village and occupied Guillemont Station. A violent counterattack on Guillemont Station was repulsed on August 23, 1916, and next day further important progress was made on a wide front north and east of Delville Wood.

[Sidenote: Advance by bombing and sapping.]

[Sidenote: Progress near Thiepval.]

Apart from the operations already described, others of a minor character, yet involving much fierce and obstinate fighting, continued during this period on the fronts of both the British armies. Our lines were pushed forward wherever possible by means of local attacks and by bombing and sapping, and the enemy was driven out of various forward positions from which he might hamper our progress. By these means many gains were made which, though small in themselves, in the aggregate represented very considerable advances. In this way our line was brought to the crest of the ridge above Martinpuich, and Pozieres Windmill and the high ground north of the village were secured, and with them observation over Martinpuich and Courcelette and the enemy's gun positions in their neighborhood and around Le Sars. At a later date our troops reached the defenses of Mouquet Farm, northwest of Pozieres, and made progress in the enemy's trenches south of Thiepval. The enemy's counter-attacks were incessant and frequently of great violence, but they were made in vain and at heavy cost to him. The fierceness of the fighting can be gathered from the fact that one regiment of the German Guards Reserve Corps which had been in the Thiepval salient opposite Mouquet Farm is known to have lost 1,400 men in fifteen days.

[Sidenote: A general attack.]

The first two days of September, 1916, on both army fronts were spent in preparation for a more general attack, which the gradual progress made during the preceding month had placed us in a position to undertake. Our assault was delivered at 12 noon on September 3, 1916, on a front extending from our extreme right to the third enemy trenches on the right bank of the Ancre, north of Hamel. Our allies attacked simultaneously on our right.

[Sidenote: Guillemont stormed.]

[Sidenote: Counterattacks on Guillemont.]

Guillemont was stormed and at once consolidated, and our troops pushed on unchecked to Ginchy and the line of the road running south to Wedge Wood. Ginchy was also seized, but here, in the afternoon, we were very strongly counterattacked. For three days the tide of attack and counterattack swayed backward and forward among the ruined houses of the village, till, in the end, for three days more the greater part of it remained in the enemy's possession. Three counterattacks made on the evening of September 3, 1916, against our troops in Guillemont all failed, with considerable loss to the enemy. We also gained ground north of Delville Wood and in High Wood, though here an enemy counterattack recovered part of the ground won.

On the front of General Gough's army, though the enemy suffered heavy losses in personnel, our gain in ground was slight.

[Sidenote: British assault on Falfemont Farm.]

In order to keep touch with the French who were attacking on our right the assault on Falfemont Farm on September 3, 1916, was delivered three hours before the opening of the main assault. In the impetus of their first rush our troops reached the farm, but could not hold it. Nevertheless, they pushed on to the north of it, and on September 4, 1916, delivered a series of fresh assaults upon it from the west and north.

[Sidenote: Leuze Wood cleared.]

Ultimately this strongly fortified position was occupied piece by piece, and by the morning of September 5, 1916, the whole of it was in our possession. Meanwhile further progress had been made to the northeast of the farm, where considerable initiative was shown by the local commanders. By the evening of the same day our troops were established strongly in Leuze Wood, which on the following day was finally cleared of the enemy.

[Sidenote: Advance on the right.]

[Sidenote: Enemy's barrier broken.]

In spite of the fact that most of Ginchy and of High Wood remained in the enemy's hands, very noteworthy progress had been made in the course of these four days' operations, exceeding anything that had been achieved since July 14, 1916. Our right was advanced on a front of nearly two miles to an average depth of nearly one mile, penetrating the enemy's original second line of defense on this front, and capturing strongly fortified positions at Falfemont Farm, Leuze Wood, Guillemont, and southeast of Delville Wood, where reached the western outskirts of Ginchy. More important than this gain in territory was the fact that the barrier which for seven weeks the enemy had maintained against our further advance had at last been broken. Over 1,000 prisoners were taken and many machine guns captured or destroyed in the course of the fighting.

Preparations for a further attack upon Ginchy continued without intermission, and at 4.45 p.m. on September 9, 1916, the attack was reopened on the whole of the Fourth Army front. At Ginchy and to the north of Leuze Wood it met with almost immediate success. On the right the enemy's line was seized over a front of more than 1,000 yards from the southwest corner of Bouleau Woods, in a northwesterly direction, to a point just south of the Guillemont-Morval tramway. Our troops again forced their way into Ginchy, and passing beyond it carried the line of enemy trenches to the east. Further progress was made east of Delville Wood and south and east of High Wood.

[Sidenote: German prisoners taken.]

Over 500 prisoners were taken in the operations of September 9, 1916, and following days, making the total since July 1, 1916, over 17,000.

[Sidenote: French progress.]

Meanwhile the French had made great progress on our right, bringing their line forward to Louage Wood (just south of Combles)—Le Forest-Clery-sur-Somme, all three inclusive. The weak salient in the allied line had therefore disappeared and we had gained the front required for further operations.

[Sidenote: Ability of new armies.]

[Sidenote: Depth of enemy fortifications.]

[Sidenote: Failure of counterattacks.]

Still more importance, however, lay in the proof afforded by the results described of the ability of our new armies, not only to rush the enemy's strongest defenses, as had been accomplished on July 1 and 14, 1916, but also to wear down and break his power of resistance by a steady, relentless pressure, as they had done during the weeks of this fierce and protracted struggle. As has already been recounted, the preparations made for our assault on July 1, 1916, had been long and elaborate; but though the enemy knew that an attack was coming, it would seem that he considered the troops already on the spot, secure in their apparently impregnable defenses, would suffice to deal with it. The success of that assault, combined with the vigor and determination with which our troops pressed their advantage, and followed by the successful night attack of July 14, 1916, all served to awaken him to a fuller realization of his danger. The great depth of his system of fortification, to which reference has been made, gave him time to reorganize his defeated troops, and to hurry up numerous fresh divisions and more guns. Yet in spite of this, he was still pushed back, steadily and continuously. Trench after trench and strong point after strong point were wrested from him. The great majority of his frequent counterattacks failed completely, with heavy loss; while the few that achieved temporary local success purchased it dearly, and were soon thrown back from the ground they had for the moment regained.

The enemy had, it is true, delayed our advance considerably, but the effort had cost him dear; and the comparative collapse of his resistance during the last few days of the struggle justified the belief that in the long run decisive victory would lie with our troops, who had displayed such fine fighting qualities and such indomitable endurance and resolution.

[Sidenote: Mouquet Farm in hands of British.]

Practically the whole of the forward crest of the main ridge on a front of some 9,000 yards, from Delville Wood to the road above Mouquet Farm, was now in our hands, and with it the advantage of observation over the slopes beyond. East of Delville Wood, for a further 3,000 yards to Leuze Wood, we were firmly established on the main ridge, while further east, across the Combles Valley, the French were advancing victoriously on our right. But though the centre of our line was well placed, on our flanks there was still difficult ground to be won.

[Sidenote: High ground from Ginchy to Morval.]

From Ginchy the crest of the high ground runs northward for 2,000 yards, and then eastward, in a long spur, for nearly 4,000 yards. Near the eastern extremity of this spur stands the village of Morval commanding a wide field of view and fire in every direction. At Leuze Wood my right was still 2,000 yards from its objective at this village, and between lay a broad and deep branch of the main Combles Valley, completely commanded by the Morval spur, and flanked, not only from its head northeast of Ginchy, but also from the high ground east of the Combles Valley, which looks directly into it.

[Sidenote: The French near Combles.]

Up this high ground beyond the Combles Valley the French were working their way toward their objective at Sailly-Saillisel, situated due east of Morval, and standing at the same level. Between these two villages the ground falls away to the head of the Combles Valley, which runs thence in a southwesterly direction. In the bottom of this valley lies the small town of Combles, then well fortified and strongly held, though dominated by my right at Leuze Wood and by the French left on the opposite heights. It had been agreed between the French and myself that an assault on Combles would not be necessary, as the place could be rendered untenable by pressing forward along the ridges above it on either side.

[Sidenote: Difficulties in way of French advance.]

The capture of Morval from the south presented a very difficult problem, while the capture of Sailly-Saillisel, at that time some 3,000 yards to the north of the French left, was in some respects even more difficult. The line of the French advance was narrowed almost to a defile by the extensive and strongly fortified Wood of St. Pierre Vaast on the one side, and on the other by the Combles Valley, which, with the branches running out from it and the slopes each side, is completely commanded, as has been pointed out, by the heights bounding the valley on the east and west.

[Sidenote: Close cooperation necessary on right.]

On my right flank, therefore, the progress of the French and British forces was still interdependent, and the closest cooperation continued to be necessary in order to gain the further ground required to enable my centre to advance on a sufficiently wide front. To cope with such a situation unity of command is usually essential, but in this case the cordial good feeling between the allied armies, and the earnest desire of each to assist the other, proved equally effective, and removed all difficulties.

[Sidenote: Enemy defense on main ridge over Thiepval.]

On my left flank the front of General Gough's army bent back from the main ridge near Mouquet Farm down a spur descending southwestward, and then crossed a broad valley to the Wonderwork, a strong point situated in the enemy's front-line system near the southern end of the spur on the higher slopes of which Thiepval stands. Opposite this part of our line we had still to carry the enemy's original defenses on the main ridge above Thiepval, and in the village itself, defenses which may fairly be described as being as nearly impregnable as nature, art, and the unstinted labor of nearly two years could make them.

[Sidenote: British advance on Thiepval defenses.]

[Sidenote: Positions might be rushed.]

Our advance on Thiepval and on the defenses above it had been carried out up to this date, in accordance with my instructions given on July 3, 1916, by a slow and methodical progression, in which great skill and much patience and endurance had been displayed with entirely satisfactory results. General Gough's army had, in fact, acted most successfully in the required manner as a pivot to the remainder of the attack. The Thiepval defenses were known to be exceptionally strong, and as immediate possession of them was not necessary to the development of my plans after July 1, 1916, there had been no need to incur the heavy casualties to be expected in an attempt to rush them. The time was now approaching, although it had not yet arrived, when their capture would become necessary; but from the positions we had now reached and those which we expected shortly to obtain, I had no doubt that they could be rushed when required without undue loss. An important part of the remaining positions required for my assault on them was now won by a highly successful enterprise carried out on the evening of September 14, 1916, by which the Wonderwork was stormed.

[Sidenote: Plan of combined attack.]

[Sidenote: Main effort against Rancourt and Fregicourt.]

The general plan of the combined allied attack which was opened on September 15 was to pivot on the high ground south of the Ancre and north of the Albert-Bapaume road, while the Fourth Army devoted its whole effort to the rearmost of the enemy's original systems of defense between Morval and Le Sars. Should our success in this direction warrant it I made arrangements to enable me to extend the left of the attack to embrace the villages of Martinpuich and Courcelette. As soon as our advance on this front had reached the Morval line, the time would have arrived to bring forward my left across the Thiepval Ridge. Meanwhile on my right our allies arranged to continue the line of advance in close co-operation with me from the Somme to the slopes above Combles, but directing their main effort northward against the villages of Rancourt and Fregicourt, so as to complete the isolation of Combles and open the way for their attack upon Sailly-Saillisel.

A methodical bombardment was commenced at 6 a.m. on September 12, 1916, and was continued steadily and uninterruptedly till the moment of attack.

[Sidenote: Bombardment and infantry assault.]

At 6.20 a.m. on September 15, 1916 the infantry assault commenced, and at the same moment the bombardment became intense. Our new heavily armored cars, known as "tanks," now brought into action for the first time, successfully co-operated with the infantry, and, coming as a surprise to the enemy rank and file, gave valuable help in breaking down their resistance.

[Sidenote: Tanks enter Flers.]

[Sidenote: High Wood carried.]

[Sidenote: Capture of the Quadrilateral.]

The advance met with immediate success on almost the whole of the front attacked. At 8.40 a.m. "tanks" were seen to be entering Flers, followed by large numbers of troops. Fighting continued in Flers for some time, but by 10 a.m. our troops had reached the north side of the village, and by midday had occupied the enemy's trenches for some distance beyond. On our right our line was advanced to within assaulting distance of the strong line of defense running before Morval, Les Boeufs, and Gueudecourt, and on our left High Wood was at last carried after many hours of very severe fighting, reflecting great credit on the attacking battalions. Our success made it possible to carry out during the afternoon that part of the plan which provided for the capture of Martinpuich and Courcelette, and by the end of the day both these villages were in our hands. On September 18, 1916, the work of this day was completed by the capture of the Quadrilateral, an enemy stronghold which had hitherto blocked the progress of our right toward Morval. Further progress was also made between Flers and Martinpuich.

[Sidenote: Results of four days' fighting.]

The result of the fighting of September 15, 1916, and following days was a gain more considerable than any which had attended our arms in the course of a single operation since the commencement of the offensive. In the course of one day's fighting we had broken through two of the enemy's main defensive systems and had advanced on a front of over six miles to an average depth of a mile. In the course of this advance we had taken three large villages, each powerfully organized for prolonged resistance. Two of these villages had been carried by assault with short preparation in the course of a few hours' fighting. All this had been accomplished with a small number of casualties in comparison with the troops employed, and in spite of the fact that, as was afterward discovered, the attack did not come as a complete surprise to the enemy.

[Sidenote: Prisoners taken.]

The total number of prisoners taken by us in these operations since their commencement on the evening of September 14, 1916, amounted at this date to over 4,000, including 127 officers.

[Sidenote: General attack launched.]

[Sidenote: Objectives taken.]

Preparations for our further advance were again hindered by bad weather, but at 12.35 p.m. on September 25, 1916, after a bombardment commenced early in the morning of the 24th, a general attack by the Allies was launched on the whole front between the Somme and Martinpuich. The objectives on the British front included the villages of Morval, Les Boeufs, and Gueudecourt, and a belt of country about 1,000 yards deep curving round the north of Flers to a point midway between that village and Martinpuich. By nightfall the whole of these objectives were in our hands, with the exception of the village of Gueudecourt, before which our troops met with very serious resistance from a party of the enemy in a section of his fourth main system of defense.

[Sidenote: French take Rancourt.]

[Sidenote: Combles occupied.]

On our right our allies carried the village of Rancourt, and advanced their line to the outskirts of Fregicourt, capturing that village also during the night and early morning. Combles was therefore nearly surrounded by the allied forces, and in the early morning of September 26, 1916, the village was occupied simultaneously by the allied forces, the British to the north and the French to the south of the railway. The capture of Combles in this inexpensive fashion represented a not inconsiderable tactical success. Though lying in a hollow, the village was very strongly fortified, and possessed, in addition to the works which the enemy had constructed, exceptionally large cellars and galleries, at a great depth under ground, sufficient to give effectual shelter to troops and material under the heaviest bombardment. Great quantities of stores and ammunition of all sorts were found in these cellars when the village was taken.

[Sidenote: Gueudecourt carried.]

[Sidenote: Few casualties.]

On the same day Gueudecourt was carried, after the protecting trench to the west had been captured in a somewhat interesting fashion. In the early morning a "tank" started down the portion of the trench held by the enemy from the northwest, firing its machine guns and followed by bombers. The enemy could not escape, as we held the trench at the southern end. At the same time an aeroplane flew down the length of the trench, also firing a machine gun at the enemy holding it. These then waved white handkerchiefs in token of surrender, and when this was reported by the aeroplane the infantry accepted the surrender of the garrison. By 8.30 a.m. the whole trench had been cleared, great numbers of the enemy had been killed, and 8 officers and 362 of the ranks made prisoners. Our total casualties amounted to five.

[Sidenote: Tactical value of the main ridge.]

The success of the Fourth Army had now brought our advance to the stage at which I judged it advisable that Thiepval should be taken, in order to bring our left flank into line and establish it on the main ridge above that village, the possession of which would be of considerable tactical value in future operations.

[Sidenote: New attack on Thiepval.]

Accordingly at 12.25 p.m. on September 26, 1916, before the enemy had been given time to recover from the blow struck by the Fourth Army, a general attack was launched against Thiepval and the Thiepval Ridge. The objective consisted of the whole of the high ground still remaining in enemy hands extending over a front of some 3,000 yards north and east of Thiepval, and including, in addition to that fortress, the Zollern Redoubt, the Stuff Redoubt, and the Schwaben Redoubt, with the connecting lines of trenches.

[Sidenote: Strong enemy resistance.]

The attack was a brilliant success. On the right our troops reached the system of enemy trenches which formed their objectives without great difficulty. In Thiepval and the strong works to the north of it the enemy's resistance was more desperate. Three waves of our attacking troops carried the outer defenses of Mouquet Farm, and, pushing on, entered Zollern Redoubt, which they stormed and consolidated. In the strong point formed by the buildings of the farm itself, the enemy garrison, securely posted in deep cellars, held out until 6 p.m., when their last defenses were forced by a working party of a pioneer battalion acting on its own initiative.

[Sidenote: Thiepval taken.]

On the left of the attack fierce fighting, in which "tanks" again gave valuable assistance to our troops, continued in Thiepval during that day and the following night, but by 8.30 a.m. on September 27, 1916 the whole of the village of Thiepval was in our hands.

[Sidenote: Prisoners.]

Some 2,300 prisoners were taken in the course of the fighting on the Thiepval Ridge on these and the subsequent days, bringing the total number of prisoners taken in the battle area in the operations of September 14-30, 1916, to nearly 10,000. In the same period we had captured 27 guns, over 200 machine guns, and some 40 trench mortars.

[Sidenote: Stuff and Schwaben Redoubts.]

On the same date the south and west sides of Stuff Redoubt were carried by our troops, together with the length of trench connecting that strong point with Schwaben Redoubt to the west and also the greater part of the enemy's defensive line eastward along the northern slopes of the ridge. Schwaben Redoubt was assaulted during the afternoon, and in spite of counterattacks, delivered by strong enemy reenforcements, we captured the whole of the southern face of the redoubt and pushed out patrols to the northern face and toward St. Pierre Divion.

Our line was also advanced north of Courcelette, while on the Fourth Army front a further portion of the enemy's fourth-system of defense northwest of Gueudecourt was carried on a front of a mile. Between these two points the enemy fell back upon his defenses running in front of Eaucourt l'Abbaye and Le Sars, and on the afternoon and evening of September 27, 1916, our troops were able to make a very considerable advance in this area without encountering serious opposition until within a few hundred yards of this line. The ground thus occupied extended to a depth of from 500 to 600 yards on a front of nearly two miles between the Bazentin-le-Petit, Lingy, Thilloy, and Albert-Bapaume roads.

[Sidenote: Destremont Farm carried.]

Destremont Farm, southwest of Le Sars, was carried by a single company on September 29, 1916, and on the afternoon of October 1, 1916, a successful attack was launched against Eaucourt l'Abbaye and the enemy defenses to the east and west of it, comprising a total front of about 3,000 yards. Our artillery barrage was extremely accurate, and contributed greatly to the success of the attack. Bomb fighting continued among the buildings during the next two days, but by the evening of October 3 the whole of Eaucourt l'Abbaye was in our hands.

[Sidenote: Fourth Army attacks.]

At the end of September, 1916, I had handed over Morval to the French, in order to facilitate their attacks on Sailly-Saillisel, and on October 7, 1916, after a postponement rendered necessary by three days' continuous rain, our allies made a considerable advance in the direction of the latter village. On the same day the Fourth Army attacked along the whole front from Les Boeufs to Destremont Farm in support of the operations of our allies.

[Sidenote: Enemy's trenches east of Gueudecourt taken.]

The village of Le Sars was captured, together with the quarry to the northwest, while considerable progress was made at other points along the front attacked. In particular, to the east of Gueudecourt, the enemy's trenches were carried on a breadth of some 2,000 yards, and a footing gained on the crest of the long spur which screens the defenses of Le Transloy from the southwest. Nearly 1,000 prisoners were secured by the Fourth Army in the course of these operations.

With the exception of his positions in the neighborhood of Sailly-Saillisel, and his scanty foothold on the northern crest of the high ground above Thiepval, the enemy had now been driven from the whole of the ridge lying between the Tortille and the Ancre.

[Sidenote: Germans make repeated counterattacks.]

[Sidenote: British situation satisfactory.]

Possession of the northwestern portion of the ridge north of the latter village carried with it observation over the valley of the Ancre between Miraumont and Hamel and the spurs and valleys held by the enemy on the right bank of the river. The Germans, therefore, made desperate efforts to cling to their last remaining trenches in this area, and in the course of the three weeks following our advance made repeated counterattacks at heavy cost in the vain hope of recovering the ground they had lost. During this period our gains in the neighborhood of Stuff and Schwaben Redoubts were gradually increased and secured in readiness for future operations; and I was quite confident of the ability of our troops, not only to repulse the enemy's attacks, but to clear him entirely from his last positions on the ridge whenever it should suit my plans to do so. I was, therefore, well content with the situation on this flank.

Along the centre of our line from Gueudecourt to the west of Le Sars similar considerations applied. As we were already well down the forward slopes of the ridge on his front, it was for the time being inadvisable to make any serious advance. Pending developments elsewhere all that was necessary or indeed desirable was to carry on local operations to improve our positions and to keep the enemy fully employed.

[Sidenote: Strong enemy positions in eastern flank.]

[Sidenote: Enemy resistance weakens.]

On our eastern flank, on the other hand, it was important to gain ground. Here the enemy still possessed a strong system of trenches covering the villages of Le Transloy and Beaulencourt and the town of Bapaume; but, although he was digging with feverish haste, he had not yet been able to create any very formidable defenses behind this line. In this direction, in fact, we had at last reached a stage at which a successful attack might reasonably be expected to yield much greater results than anything we had yet attained. The resistance of the troops opposed to us had seriously weakened in the course of our recent operations, and there was no reason to suppose that the effort required would not be within our powers.

[Sidenote: Necessity to gain spur and heights.]

The last completed system of defense, before Le Transloy, was flanked to the south by the enemy's positions at Sailly-Saillisel, and screened to the west by the spur lying between Le Transloy and Les Boeufs. A necessary preliminary, therefore, to an assault upon it was to secure the spur and the Sailly-Saillisel heights. Possession of the high ground at this latter village would at once give a far better command over the ground to the north and northwest, secure the flank of our operations toward Le Transloy, and deprive the enemy of observation over the allied communications in the Combles Valley. In view of the enemy's efforts to construct new systems of defense behind the Le Transloy spur, was extended and secured time in dealing with the situation.

[Sidenote: Rain and fog a hindrance.]

Unfortunately, at this juncture, very unfavorable weather set in and continued with scarcely a break during the remainder of October and the early part of November. Poor visibility seriously interfered with the work of our artillery, and constant rain turned the mass of hastily dug trenches for which we were fighting into channels of deep mud. The country roads, broken by countless shell craters, that cross the deep stretch of ground we had lately won, rapidly became almost impassable, making the supply of food, stores, and ammunition a serious problem. These conditions multiplied the difficulties of attack to such an extent that it was found impossible to exploit the situation with the rapidity necessary to enable us to reap the full benefits of the advantages we had gained.

[Sidenote: Enemy has time to reorganize.]

None the less, my right flank continued to assist the operations of our allies against Saillisel, and attacks were made to this end, whenever a slight improvement in the weather made the co-operation of artillery and infantry at all possible. The delay in our advance, however, though unavoidable, had given the enemy time to reorganize and rally his troops. His resistance again became stubborn and he seized every favorable opportunity for counterattacks. Trenches changed hands with great frequency, the conditions of ground making it difficult to renew exhausted supplies of bombs and ammunition, or to consolidate the ground won, and so rendering it an easier matter to take a battered trench than to hold it.