Nearly all the people of the country seemed to have fled with Johnston's army; yet some few families remained, and from one of them I procured the copy of an order which Johnston had made at Adairsville, in which he recited that he had retreated as far as strategy required, and that his army must be prepared for battle at Cassville. The newspapers of the South, many of which we found, were also loud in denunciation of Johnston's falling back before us without a serious battle, simply resisting by his skirmish-lines and by his rear-guard. But his friends proclaimed that it was all strategic; that he was deliberately drawing us farther and farther into the meshes, farther and farther away from our base of supplies, and that in due season he would not only halt for battle, but assume the bold offensive. Of course it was to my interest to bring him to battle as soon as possible, when our numerical superiority was at the greatest; for he was picking up his detachments as he fell back, whereas I was compelled to make similar and stronger detachments to repair the railroads as we advanced, and to guard them. I found at Cassville many evidences of preparation for a grand battle, among them a long line of fresh intrenchments on the hill beyond the town, extending nearly three miles to the south, embracing the railroad-crossing. I was also convinced that the whole of Polk's corps had joined Johnston from Mississippi, and that he had in hand three full corps, viz., Hood's, Polk's, and Hardee's, numbering about sixty thousand men, and could not then imagine why he had declined battle, and did not learn the real reason till after the war was over, and then from General Johnston himself.

In the autumn of 1865, when in command of the Military Division of the Missouri, I went from St. Louis to Little Rock, Arkansas, and afterward to Memphis. Taking a steamer for Cairo, I found as fellow-passengers Generals Johnston and Frank Blair. We were, of course, on the most friendly terms, and on our way up we talked over our battles again, played cards, and questioned each other as to particular parts of our mutual conduct in the game of war. I told Johnston that I had seen his order of preparation, in the nature of an address to his army, announcing his purpose to retreat no more, but to accept battle at Cassville. He answered that such was his purpose; that he had left Hardee's corps in the open fields to check Thomas, and gain time for his formation on the ridge, just behind Cassville; and it was this corps which General Thomas had seen deployed, and whose handsome movement in retreat he had reported in such complimentary terms. Johnston described how he had placed Hood's corps on the right, Polk's in the centre, and Hardee's on the left. He said he had ridden over the ground, given to each corps commander his position, and orders to throw up parapets during the night; that he was with Hardee on his extreme left as the night closed in, and as Hardee's troops fell back to the position assigned them for the intended battle of the next day; and that, after giving Hardee some general instructions, he and his staff rode back to Cassville. As he entered the town, or village, he met Generals Hood and Polk. Hood inquired of him if he had had any thing to eat, and he said no, that he was both hungry and tired, when Hood invited him to go and share a supper which had been prepared for him at a house close by. At the supper they discussed the chances of the impending battle, when Hood spoke of the ground assigned him as being enfiladed by our (Union) artillery, which Johnston disputed, when General Polk chimed in with the remark that General Hood was right; that the cannon-shots fired by us at nightfall had enfiladed their general line of battle, and that for this reason he feared they could not hold their men. General Johnston was surprised at this, for he understood General Hood to be one of those who professed to criticise his strategy, contending that, instead of retreating, he should have risked a battle. General Johnston said he was provoked, accused them of having been in conference, with being beaten before battle, and added that he was unwilling to engage in a critical battle with an army so superior to his own in numbers, with two of his three corps commanders dissatisfied with the ground and positions assigned them. He then and there made up his mind to retreat still farther south, to put the Etowah River and the Allatoona range between us; and he at once gave orders to resume the retrograde movement.

This was my recollection of the substance of the conversation, of which I made no note at the time; but, at a meeting of the Society of the Army of the Cumberland some years after, at Cleveland, Ohio, about 1868, in a short after-dinner speech, I related this conversation, and it got into print. Subsequently, in the spring of 1870, when I was at New Orleans, on route for Texas, General Hood called to see me at the St. Charles Hotel, explained that he had seen my speech reprinted in the newspapers and gave me his version of the same event, describing the halt at Cassville, the general orders for battle on that ground, and the meeting at supper with Generals Johnston and Polk, when the chances of the battle to be fought the next day were freely and fully discussed; and he stated that he had argued against fighting the battle purely on the defensive, but had asked General Johnston to permit him with his own corps and part of Polk's to quit their lines, and to march rapidly to attack and overwhelm Schofield, who was known to be separated from Thomas by an interval of nearly five miles, claiming that he could have defeated Schofield, and got back to his position in time to meet General Thomas's attack in front. He also stated that he had then contended with Johnston for the "offensive-defensive" game, instead of the "pure defensive," as proposed by General Johnston; and he said that it was at this time that General Johnston had taken offense, and that it was for this reason he had ordered the retreat that night. As subsequent events estranged these two officers, it is very natural they should now differ on this point; but it was sufficient for us that the rebel army did retreat that night, leaving us masters of all the country above the Etowah River.

For the purposes of rest, to give time for the repair of the railroads, and to replenish supplies, we lay by some few days in that quarter—Schofield with Stoneman's cavalry holding the ground at Cassville Depot, Cartersville, and the Etowah Bridge; Thomas holding his ground near Cassville, and McPherson that near Kingston. The officer intrusted with the repair of the railroads was Colonel W. W. Wright, a railroad-engineer, who, with about two thousand men, was so industrious and skillful that the bridge at Resaca was rebuilt in three days, and cars loaded with stores came forward to Kingston on the 24th. The telegraph also brought us the news of the bloody and desperate battles of the Wilderness, in Virginia, and that General Grant was pushing his operations against Lee with terrific energy. I was therefore resolved to give my enemy no rest.

In early days (1844), when a lieutenant of the Third Artillery, I had been sent from Charleston, South Carolina, to Marietta, Georgia, to assist Inspector-General Churchill to take testimony concerning certain losses of horses and accoutrements by the Georgia Volunteers during the Florida War; and after completing the work at Marietta we transferred our party over to Bellefonte, Alabama. I had ridden the distance on horseback, and had noted well the topography of the country, especially that about Kenesaw, Allatoona, and the Etowah River. On that occasion I had stopped some days with a Colonel Tumlin, to see some remarkable Indian mounds on the Etowah River, usually called the "Hightower:" I therefore knew that the Allatoona Pass was very strong, would be hard to force, and resolved not even to attempt it, but to turn the position, by moving from Kingston to Marietta via. Dallas; accordingly I made orders on the 20th to get ready for the march to begin on the 23d. The Army of the Cumberland was ordered to march for Dallas, by Euharlee and Stilesboro; Davis's division, then in Rome, by Van Wert; the Army of the Ohio to keep on the left of Thomas, by a place called Burnt Hickory; and the Army of the Tennessee to march for a position a little to the south, so as to be on the right of the general army, when grouped about Dallas.

The movement contemplated leaving our railroad, and to depend for twenty days on the contents of our wagons; and as the country was very obscure, mostly in a state of nature, densely wooded, and with few roads, our movements were necessarily slow. We crossed the Etowah by several bridges and fords, and took as many roads as possible, keeping up communication by cross-roads, or by couriers through the woods. I personally joined General Thomas, who had the centre, and was consequently the main column, or "column of direction." The several columns followed generally the valley of the Euharlee, a tributary coming into the Etowah from the south, and gradually crossed over a ridge of mountains, parts of which had once been worked over for gold, and were consequently full of paths and unused wagon-roads or tracks. A cavalry picket of the enemy at Burnt Hickory was captured, and had on his person an order from General Johnston, dated at Allatoona, which showed that he had detected my purpose of turning his position, and it accordingly became necessary to use great caution, lest some of the minor columns should fall into ambush, but, luckily the enemy was not much more familiar with that part of the country than we were. On the other side of the Allatoona range, the Pumpkin-Vine Creek, also a tributary of the Etowah, flowed north and west; Dallas, the point aimed at, was a small town on the other or east side of this creek, and was the point of concentration of a great many roads that led in every direction. Its possession would be a threat to Marietta and Atlanta, but I could not then venture to attempt either, till I had regained the use of the railroad, at least as far down as its debouche from the Allatoona range of mountains. Therefore, the movement was chiefly designed to compel Johnston to give up Allatoona.

On the 25th all the columns were moving steadily on Dallas—McPherson and Davis away off to the right, near Van Wert; Thomas on the main road in the centre, with Hooker's Twentieth Corps ahead, toward Dallas; and Schofield to the left rear. For the convenience of march, Hooker had his three divisions on separate roads, all leading toward Dallas, when, in the afternoon, as he approached a bridge across Pumpkin-Vine Creek, he found it held by a cavalry force, which was driven off, but the bridge was on fire. This fire was extinguished, and Hooker's leading division (Geary's) followed the retreating cavalry on a road leading due east toward Marietta, instead of Dallas. This leading division, about four miles out from the bridge, struck a heavy infantry force, which was moving down from Allatoona toward Dallas, and a sharp battle ensued. I came up in person soon after, and as my map showed that we were near an important cross-road called "New Hope," from a Methodist meeting-house there of that name, I ordered General Hooker to secure it if possible that night. He asked for a short delay, till he could bring up his other two divisions, viz., of Butterfield and Williams, but before these divisions had got up and were deployed, the enemy had also gained corresponding strength. The woods were so dense, and the resistance so spirited, that Hooker could not carry the position, though the battle was noisy, and prolonged far into the night. This point, "New Hope," was the accidental intersection of the road leading from Allatoona to Dallas with that from Van Wert to Marietta, was four miles northeast of Dallas, and from the bloody fighting there for the next week was called by the soldiers "Hell-Hole."

The night was pitch-dark, it rained hard, and the convergence of our columns toward Dallas produced much confusion. I am sure similar confusion existed in the army opposed to us, for we were all mixed up. I slept on the ground, without cover, alongside of a log, got little sleep, resolved at daylight to renew the battle, and to make a lodgment on the Dallas and Allatoona road if possible, but the morning revealed a strong line of intrenchments facing us, with a heavy force of infantry and guns. The battle was renewed, and without success. McPherson reached Dallas that morning, viz., the 26th, and deployed his troops to the southeast and east of the town, placing Davis's division of the Fourteenth Corps, which had joined him on the road from Rome, on his left; but this still left a gap of at least three miles between Davis and Hooker. Meantime, also, General Schofield was closing up on Thomas's left.

Satisfied that Johnston in person was at New Hope with all his army, and that it was so much nearer my "objective;" the railroad, than Dallas, I concluded to draw McPherson from Dallas to Hooker's right, and gave orders accordingly; but McPherson also was confronted with a heavy force, and, as he began to withdraw according to his orders, on the morning of the 28th he was fiercely assailed on his right; a bloody battle ensued, in which he repulsed the attack, inflicting heavy loss on his assailants, and it was not until the 1st of June that he was enabled to withdraw from Dallas, and to effect a close junction with Hooker in front of New Hope. Meantime Thomas and Schofield were completing their deployments, gradually overlapping Johnston on his right, and thus extending our left nearer and nearer to the railroad, the nearest point of which was Acworth, about eight miles distant. All this time a continual battle was in progress by strong skirmish-lines, taking advantage of every species of cover, and both parties fortifying each night by rifle-trenches, with head-logs, many of which grew to be as formidable as first-class works of defense. Occasionally one party or the other would make a dash in the nature of a sally, but usually it sustained a repulse with great loss of life. I visited personally all parts of our lines nearly every day, was constantly within musket-range, and though the fire of musketry and cannon resounded day and night along the whole line, varying from six to ten miles, I rarely saw a dozen of the enemy at any one time; and these were always skirmishers dodging from tree to tree, or behind logs on the ground, or who occasionally showed their heads above the hastily-constructed but remarkably strong rifle-trenches. On the occasion of my visit to McPherson on the 30th of May, while standing with a group of officers, among whom were Generals McPherson, Logan, Barry, and Colonel Taylor, my former chief of artillery, a Minie-ball passed through Logan's coat-sleeve, scratching the skin, and struck Colonel Taylor square in the breast; luckily he had in his pocket a famous memorandum-book, in which he kept a sort of diary, about which we used to joke him a good deal; its thickness and size saved his life, breaking the force of the ball, so that after traversing the book it only penetrated the breast to the ribs, but it knocked him down and disabled him for the rest of the campaign. He was a most competent and worthy officer, and now lives in poverty in Chicago, sustained in part by his own labor, and in part by a pitiful pension recently granted.

On the 1st of June General McPherson closed in upon the right, and, without attempting further to carry the enemy's strong position at New Hope Church, I held our general right in close contact with it, gradually, carefully, and steadily working by the left, until our strong infantry-lines had reached and secured possession of all the wagon-roads between New Hope, Allatoona, and Acworth, when I dispatched Generals Garrard's and Stoneman's divisions of cavalry into Allatoona, the first around by the west end of the pass, and the latter by the direct road. Both reached their destination without opposition, and orders were at once given to repair the railroad forward from Kingston to Allatoona, embracing the bridge across the Etowah River. Thus the real object of my move on Dallas was accomplished, and on the 4th of June I was preparing to draw off from New Hope Church, and to take position on the railroad in front of Allatoona, when, General Johnston himself having evacuated his position, we effected the change without further battle, and moved to the railroad, occupying it from Allatoona and Acworth forward to Big Shanty, in sight of the famous Kenesaw Mountain.

Thus, substantially in the month of May, we had steadily driven our antagonist from the strong positions of Dalton, Resaea, Cassville, Allatoona, and Dallas; had advanced our lines in strong, compact order from Chattanooga to Big Shanty, nearly a hundred miles of as difficult country as was ever fought over by civilized armies; and thus stood prepared to go on, anxious to fight, and confident of success as soon as the railroad communications were complete to bring forward the necessary supplies. It is now impossible to state accurately our loss of life and men in any one separate battle; for the fighting was continuous, almost daily, among trees and bushes, on ground where one could rarely see a hundred yards ahead.

The aggregate loss in the several corps for the month of May is reported-as follows in the usual monthly returns sent to the Adjutant-General's office, which are, therefore, official:

Casualties during the Month of May, 1864
(Major-General SHERMAN commanding).

General Joseph E. Johnston, in his "Narrative of his Military Operations," just published (March 27, 1874), gives the effective strength of his army at and about Dalton on the 1st of May, 1864 (page 302), as follows:

During May, and prior to reaching Cassville, he was further reenforced.

And at New Hope Church, May 26th

His losses during the month of May are stated by him, as taken from the report of Surgeon Foard.

These figures include only the killed and wounded, whereas my statement of losses embraces the "missing," which are usually "prisoners," and of these we captured, during the whole campaign of four and a half months, exactly 12,983, whose names, rank, and regiments, were officially reported to the Commissary-General of Prisoners; and assuming a due proportion for the month of May, viz., one-fourth, makes 3,245 to be added to the killed and wounded given above, making an aggregate loss in Johnston's army, from Dalton to New Hope, inclusive, of 8,638, against ours of 9,299.

Therefore General Johnston is greatly in error, in his estimates on page 357, in stating our loss, as compared with his, at six or ten to one.

I always estimated my force at about double his, and could afford to lose two to one without disturbing our relative proportion; but I also reckoned that, in the natural strength of the country, in the abundance of mountains, streams, and forests, he had a fair offset to our numerical superiority, and therefore endeavored to act with reasonable caution while moving on the vigorous "offensive."

With the drawn battle of New Hope Church, and our occupation of the natural fortress of Allatoona, terminated the month of May, and the first stage of the campaign.

CHAPTER XVII.

On the 1st of June our three armies were well in hand, in the broken and densely-wooded country fronting the enemy intrenched at New Hope Church, about five miles north of Dallas. General Stoneman's division of cavalry had occupied Allatoona, on the railroad, and General Garrard's division was at the western end of the pass, about Stilesboro. Colonel W. W. Wright, of the Engineers, was busily employed in repairing the railroad and rebuilding the bridge across the Etowah (or Hightower) River, which had been destroyed by the enemy on his retreat; and the armies were engaged in a general and constant skirmish along a front of about six miles—McPherson the right, Thomas the centre, and Schofield on the left. By gradually covering our front with parapet, and extending to the left, we approached the railroad toward Acworth and overlapped the enemy's right. By the 4th of June we had made such progress that Johnston evacuated his lines in the night, leaving us masters of the situation, when I deliberately shifted McPherson's army to the extreme left, at and in front of Acworth, with Thomas's about two miles on his right, and Schofield's on his right all facing east. Heavy rains set in about the 1st of June, making the roads infamous; but our marches were short, as we needed time for the repair of the railroad, so as to bring supplies forward to Allatoona Station. On the 6th I rode back to Allatoona, seven miles, found it all that was expected, and gave orders for its fortification and preparation as a "secondary base."

General Blair arrived at Acworth on the 8th with his two divisions of the Seventeenth Corps—the same which had been on veteran furlough—had come up from Cairo by way of Clifton, on the Tennessee River, and had followed our general route to Allatoona, where he had left a garrison of about fifteen hundred men. His effective strength, as reported, was nine thousand. These, with new regiments and furloughed men who had joined early in the month of May, equaled our losses from battle, sickness, and by detachments; so that the three armies still aggregated about one hundred thousand effective men.

On the 10th of June the whole combined army moved forward six miles, to "Big Shanty," a station on the railroad, whence we had a good view of the enemy's position, which embraced three prominent hills known as Kenesaw, Pine Mountain, and Lost Mountain. On each of these hills the enemy had signal-stations and fresh lines of parapets. Heavy masses of infantry could be distinctly seen with the naked eye, and it was manifest that Johnston had chosen his ground well, and with deliberation had prepared for battle; but his line was at least ten miles in extent—too long, in my judgment, to be held successfully by his force, then estimated at sixty thousand. As his position, however, gave him a perfect view over our field, we had to proceed with due caution. McPherson had the left, following the railroad, which curved around the north base of Kenesaw; Thomas the centre, obliqued to the right, deploying below Kenesaw and facing Pine Hill; and Schofield, somewhat refused, was on the general right, looking south, toward Lost Mountain.

On the 11th the Etowah bridge was done; the railroad was repaired up to our very skirmish line, close to the base of Kenesaw, and a loaded train of cars came to Big Shanty. The locomotive, detached, was run forward to a water-tank within the range of the enemy's guns on Kenesaw, whence the enemy opened fire on the locomotive; but the engineer was not afraid, went on to the tank, got water, and returned safely to his train, answering the guns with the screams of his engine, heightened by the cheers and shouts of our men.

The rains continued to pour, and made our developments slow and dilatory, for there were no roads, and these had to be improvised by each division for its own supply train from the depot in Big Shanty to the camps. Meantime each army was deploying carefully before the enemy, intrenching every camp, ready as against a sally. The enemy's cavalry was also busy in our rear, compelling us to detach cavalry all the way back as far as Resaca, and to strengthen all the infantry posts as far as Nashville. Besides, there was great danger, always in my mind, that Forrest would collect a heavy cavalry command in Mississippi, cross the Tennessee River, and break up our railroad below Nashville. In anticipation of this very danger, I had sent General Sturgis to Memphis to take command of all the cavalry in that quarter, to go out toward Pontotoc, engage Forrest and defeat him; but on the 14th of June I learned that General Sturgis had himself been defeated on the 10th of June, and had been driven by Forrest back into Memphis in considerable confusion. I expected that this would soon be followed by a general raid on all our roads in Tennessee. General G. J. Smith, with the two divisions of the Sixteenth and Seventeenth Corps which had been with General Banks up Red River, had returned from that ill-fated expedition, and had been ordered to General Canby at New Orleans, who was making a diversion about Mobile; but, on hearing of General Sturgis's defeat, I ordered General Smith to go out from Memphis and renew the offensive, so as to keep Forrest off our roads. This he did finally, defeating Forrest at Tupelo, on the 13th, 14th, and 15th days of July; and he so stirred up matters in North Mississippi that Forrest could not leave for Tennessee. This, for a time, left me only the task of covering the roads against such minor detachments of cavalry as Johnston could spare from his immediate army, and I proposed to keep these too busy in their own defense to spare detachments. By the 14th the rain slackened, and we occupied a continuous line of ten miles, intrenched, conforming to the irregular position of the enemy, when I reconnoitred, with a view to make a break in their line between Kenesaw and Pine Mountain. When abreast of Pine Mountain I noticed a rebel battery on its crest, with a continuous line of fresh rifle-trench about half-way down the hill. Our skirmishers were at the time engaged in the woods about the base of this hill between the lines, and I estimated the distance to the battery on the crest at about eight hundred yards. Near it, in plain view, stood a group of the enemy, evidently observing us with glasses. General Howard, commanding the Fourth Corps, was near by, and I called his attention to this group, and ordered him to compel it to keep behind its cover. He replied that his orders from General Thomas were to spare artillery-ammunition. This was right, according to the general policy, but I explained to him that we must keep up the morale of a bold offensive, that he must use his artillery, force the enemy to remain on the timid defensive, and ordered him to cause a battery close by to fire three volleys. I continued to ride down our line, and soon heard, in quick succession, the three volleys. The next division in order was Geary's, and I gave him similar orders. General Polk, in my opinion, was killed by the second volley fired from the first battery referred to.

In a conversation with General Johnston, after the war, he explained that on that day he had ridden in person from Marietta to Pine Mountain, held by Bates's division, and was accompanied by Generals Hardee and Polk. When on Pine Mountain, reconnoitring, quite a group of soldiers, belonging to the battery close by, clustered about him. He noticed the preparations of our battery to fire, and cautioned these men to scatter. They did so, and he likewise hurried behind the parapet, from which he had an equally good view of our position but General Polk, who was dignified and corpulent, walked back slowly, not wishing to appear too hurried or cautious in the presence of the men, and was struck across the breast by an unexploded shell, which killed him instantly. This is my memory of the conversation, and it is confirmed by Johnston himself in his "Narrative," page 337, except that he calculated the distance of our battery at six hundred yards, and says that Polk was killed by the third shot; I know that our guns fired by volley, and believe that he was hit by a shot of the second volley. It has been asserted that I fired the gun which killed General Polk, and that I knew it was directed against that general. The fact is, at that distance we could not even tell that the group were officers at all; I was on horseback, a couple of hundred yards off, before my orders to fire were executed, had no idea that our shot had taken effect, and continued my ride down along the line to Schofield's extreme flank, returning late in the evening to my head-quarters at Big Shanty, where I occupied an abandoned house. In a cotton-field back of that house was our signal-station, on the roof of an old gin-house. The signal-officer reported that by studying the enemy's signals he had learned the key, and that he could read their signals. He explained to me that he had translated a signal about noon, from Pine Mountain to Marietta, "Send an ambulance for General Polk's body;" and later in the day another, "Why don't you send an ambulance for General Polk?" From this we inferred that General Polk had been killed, but how or where we knew not; and this inference was confirmed later in the same day by the report of some prisoners who had been captured.

On the 15th we advanced our general lines, intending to attack at any weak point discovered between Kenesaw and Pine Mountain; but Pine Mountain was found to be abandoned, and Johnston had contracted his front somewhat, on a direct line, connecting Kenesaw with Lost Mountain. Thomas and Schofield thereby gained about two miles of most difficult, country, and McPherson's left lapped well around the north end of Kenesaw. We captured a good many prisoners, among them a whole infantry regiment, the Fourteenth Alabama, three hundred and twenty strong.

On the 16th the general movement was continued, when Lost Mountain was abandoned by the enemy. Our right naturally swung round, so as to threaten the railroad below Marietta, but Johnston had still further contracted and strengthened his lines, covering Marietta and all the roads below.

On the 17th and 18th the rain again fell in torrents, making army movements impossible, but we devoted the time to strengthening our positions, more especially the left and centre, with a view gradually to draw from the left to add to the right; and we had to hold our lines on the left extremely strong, to guard against a sally from Kenesaw against our depot at Big Shanty. Garrard's division of cavalry was kept busy on our left, McPherson had gradually extended to his right, enabling Thomas to do the same still farther; but the enemy's position was so very strong, and everywhere it was covered by intrenchments, that we found it as dangerous to assault as a permanent fort. We in like manner covered our lines of battle by similar works, and even our skirmishers learned to cover their bodies by the simplest and best forms of defensive works, such as rails or logs, piled in the form of a simple lunette, covered on the outside with earth thrown up at night.

The enemy and ourselves used the same form of rifle-trench, varied according to the nature of the ground, viz.: the trees and bushes were cut away for a hundred yards or more in front, serving as an abatis or entanglement; the parapets varied from four to six feet high, the dirt taken from a ditch outside and from a covered way inside, and this parapet was surmounted by a "head-log," composed of the trunk of a tree from twelve to twenty inches at the butt, lying along the interior crest of the parapet and resting in notches cut in other trunks which extended back, forming an inclined plane, in case the head-log should be knocked inward by a cannon-shot. The men of both armies became extremely skillful in the construction of these works, because each man realized their value and importance to himself, so that it required no orders for their construction. As soon as a regiment or brigade gained a position within easy distance for a sally, it would set to work with a will, and would construct such a parapet in a single night; but I endeavored to spare the soldiers this hard labor by authorizing each division commander to organize out of the freedmen who escaped to us a pioneer corps of two hundred men, who were fed out of the regular army supplies, and I promised them ten dollars a month, under an existing act of Congress. These pioneer detachments became very useful to us during the rest of the war, for they could work at night while our men slept; they in turn were not expected to fight, and could therefore sleep by day. Our enemies used their slaves for a similar purpose, but usually kept them out of the range of fire by employing them to fortify and strengthen the position to their rear next to be occupied in their general retrograde. During this campaign hundreds if not thousands of miles of similar intrenchments were built by both armies, and, as a rule, whichever party attacked got the worst of it.

On the 19th of June the rebel army again fell back on its flanks, to such an extent that for a time I supposed it had retreated to the Chattahoochee River, fifteen miles distant; but as we pressed forward we were soon undeceived, for we found it still more concentrated, covering Marietta and the railroad. These successive contractions of the enemy's line encouraged us and discouraged him, but were doubtless justified by sound reasons. On the 20th Johnston's position was unusually strong. Kenesaw Mountain was his salient; his two flanks were refused and covered by parapets and by Noonday and Nose's Creeks. His left flank was his weak point, so long as he acted on the "defensive," whereas, had he designed to contract the extent of his line for the purpose of getting in reserve a force with which to strike "offensively" from his right, he would have done a wise act, and I was compelled to presume that such was his object: We were also so far from Nashville and Chattanooga that we were naturally sensitive for the safety of our railroad and depots, so that the left (McPherson) was held very strong.

About this time came reports that a large cavalry force of the enemy had passed around our left flank, evidently to strike this very railroad somewhere below Chattanooga. I therefore reenforced the cavalry stationed from Resaca to Casaville, and ordered forward from Huntsville, Alabama, the infantry division of General John E. Smith, to hold Kingston securely.

While we were thus engaged about Kenesaw, General Grant had his hands full with Lee, in Virginia. General Halleck was the chief of staff at Washington, and to him I communicated almost daily. I find from my letter-book that on the 21st of June I reported to him tersely and truly the condition of facts on that day: "This is the nineteenth day of rain, and the prospect of fair weather is as far off as ever. The roads are impassable; the fields and woods become quagmire's after a few wagons have crossed over. Yet we are at work all the time. The left flank is across Noonday Creek, and the right is across Nose's Creek. The enemy still holds Kenesaw, a conical mountain, with Marietta behind it, and has his flanks retired, to cover that town and the railroad behind. I am all ready to attack the moment the weather and roads will permit troops and artillery to move with any thing like life."

The weather has a wonderful effect on troops: in action and on the march, rain is favorable; but in the woods, where all is blind and uncertain, it seems almost impossible for an army covering ten miles of front to act in concert during wet and stormy weather. Still I pressed operations with the utmost earnestness, aiming always to keep our fortified lines in absolute contact with the enemy, while with the surplus force we felt forward, from one flank or the other, for his line of communication and retreat. On the 22d of June I rode the whole line, and ordered General Thomas in person to advance his extreme right corps (Hooker's); and instructed General Schofield, by letter, to keep his entire army, viz., the Twenty-third Corps, as a strong right flank in close support of Hooker's deployed line. During this day the sun came out, with some promise of clear weather, and I had got back to my bivouac about dark, when a signal message was received, dated—

KULP HOUSE, 5.30 P.M.

General SHERMAN: We have repulsed two heavy attacks, and feel confident, our only apprehension being from our extreme right flank. Three entire corps are in front of us.

Major-General HOOKER.

Hooker's corps (the Twentieth) belonged to Thomas's army; Thomas's headquarters were two miles nearer to Hooker than mine; and Hooker, being an old army officer, knew that he should have reported this fact to Thomas and not to me; I was, moreover, specially disturbed by the assertion in his report that he was uneasy about his right flank, when Schofield had been specially ordered to protect that. I first inquired of my adjutant, Dayton, if he were certain that General Schofield had received his orders, and he answered that the envelope in which he had sent them was receipted by General Schofield himself. I knew, therefore, that General Schofield must be near by, in close support of Hooker's right flank. General Thomas had before this occasion complained to me of General Hooker's disposition to "switch off," leaving wide gaps in his line, so as to be independent, and to make glory on his own account. I therefore resolved not to overlook this breach of discipline and propriety. The rebel army was only composed of three corps; I had that very day ridden six miles of their lines, found them everywhere strongly occupied, and therefore Hooker could not have encountered "three entire corps." Both McPherson and Schofield had also complained to me of this same tendency of Hooker to widen the gap between his own corps and his proper army (Thomas's), so as to come into closer contact with one or other of the wings, asserting that he was the senior by commission to both McPherson and Schofield, and that in the event of battle he should assume command over them, by virtue of his older commission.

They appealed to me to protect them. I had heard during that day some cannonading and heavy firing down toward the "Kulp House," which was about five miles southeast of where I was, but this was nothing unusual, for at the same moment there was firing along our lines full ten miles in extent. Early the next day (23d) I rode down to the "Kulp House," which was on a road leading from Powder Springs to Marietta, about three miles distant from the latter. On the way I passed through General Butterfield's division of Hooker's corps, which I learned had not been engaged at all in the battle of the day before; then I rode along Geary's and Williams's divisions, which occupied the field of battle, and the men were engaged in burying the dead. I found General Schofield's corps on the Powder Springs road, its head of column abreast of Hooker's right, therefore constituting "a strong right flank," and I met Generale Schofield and Hooker together. As rain was falling at the moment, we passed into a little church standing by the road-side, and I there showed General Schofield Hooker's signal-message of the day before. He was very angry, and pretty sharp words passed between them, Schofield saying that his head of column (Hascall's division) had been, at the time of the battle, actually in advance of Hooker's line; that the attack or sally of the enemy struck his troops before it did Hooker's; that General Hooker knew of it at the time; and he offered to go out and show me that the dead men of his advance division (Hascall's) were lying farther out than any of Hooker's. General Hooker pretended not to have known this fact. I then asked him why he had called on me for help, until he had used all of his own troops; asserting that I had just seen Butterfield's division, and had learned from him that he had not been engaged the day before at all; and I asserted that the enemy's sally must have been made by one corps (Hood's), in place of three, and that it had fallen on Geary's and Williams's divisions, which had repulsed the attack handsomely. As we rode away from that church General Hooker was by my side, and I told him that such a thing must not occur again; in other words, I reproved him more gently than the occasion demanded, and from that time he began to sulk. General Hooker had come from the East with great fame as a "fighter," and at Chattanooga he was glorified by his "battle above the clouds," which I fear turned his head. He seemed jealous of all the army commanders, because in years, former rank, and experience, he thought he was our superior.

On the 23d of June I telegraphed to General Halleck this summary, which I cannot again better state:

We continue to press forward on the principle of an advance against fortified positions. The whole country is one vast fort, and Johnston must have at least fifty miles of connected trenches, with abatis and finished batteries. We gain ground daily, fighting all the time. On the 21st General Stanley gained a position near the south end of Kenesaw, from which the enemy attempted in vain to drive him; and the same day General T. J. Wood's division took a hill, which the enemy assaulted three times at night without success, leaving more than a hundred dead on the ground. Yesterday the extreme right (Hooker and Schofield) advanced on the Powder Springs road to within three miles of Marietta. The enemy made a strong effort to drive them away, but failed signally, leaving more than two hundred dead on the field. Our lines are now in close contact, and the fighting is incessant, with a good deal of artillery-fire. As fast as we gain one position the enemy has another all ready, but I think he will soon have to let go Kenesaw, which is the key to the whole country. The weather is now better, and the roads are drying up fast. Our losses are light, and, not-withstanding the repeated breaks of the road to our rear, supplies are ample.

During the 24th and 25th of June General Schofield extended his right as far as prudent, so as to compel the enemy to thin out his lines correspondingly, with the intention to make two strong assaults at points where success would give us the greatest advantage. I had consulted Generals Thomas, McPherson, and Schofield, and we all agreed that we could not with prudence stretch out any more, and therefore there was no alternative but to attack "fortified lines," a thing carefully avoided up to that time. I reasoned, if we could make a breach anywhere near the rebel centre, and thrust in a strong head of column, that with the one moiety of our army we could hold in check the corresponding wing of the enemy, and with the other sweep in flank and overwhelm the other half. The 27th of June was fixed as the day for the attempt, and in order to oversee the whole, and to be in close communication with all parts of the army, I had a place cleared on the top of a hill to the rear of Thomas's centre, and had the telegraph-wires laid to it. The points of attack were chosen, and the troops were all prepared with as little demonstration as possible. About 9 A.M. Of the day appointed, the troops moved to the assault, and all along our lines for ten miles a furious fire of artillery and musketry was kept up. At all points the enemy met us with determined courage and in great force. McPherson's attacking column fought up the face of the lesser Kenesaw, but could not reach the summit. About a mile to the right (just below the Dallas road) Thomas's assaulting column reached the parapet, where Brigadier-General Barker was shot down mortally wounded, and Brigadier-General Daniel McCook (my old law-partner) was desperately wounded, from the effects of which he afterward died. By 11.30 the assault was in fact over, and had failed. We had not broken the rebel line at either point, but our assaulting columns held their ground within a few yards of the rebel trenches, and there covered themselves with parapet. McPherson lost about five hundred men and several valuable officers, and Thomas lost nearly two thousand men. This was the hardest fight of the campaign up to that date, and it is well described by Johnston in his "Narrative" (pages 342, 343), where he admits his loss in killed and wounded as

Total ............. 808

This, no doubt, is a true and fair statement; but, as usual, Johnston overestimates our loss, putting it at six thousand, whereas our entire loss was about twenty-five hundred, killed and wounded.

While the battle was in progress at the centre, Schofield crossed Olley's Creek on the right, and gained a position threatening Johnston's line of retreat; and, to increase the effect, I ordered Stoneman's cavalry to proceed rapidly still farther to the right, to Sweetwater. Satisfied of the bloody cost of attacking intrenched lines, I at once thought of moving the whole army to the railroad at a point (Fulton) about ten miles below Marietta, or to the Chattahoochee River itself, a movement similar to the one afterward so successfully practised at Atlanta. All the orders were issued to bring forward supplies enough to fill our wagons, intending to strip the railroad back to Allatoona, and leave that place as our depot, to be covered as well as possible by Garrard's cavalry. General Thomas, as usual, shook his head, deeming it risky to leave the railroad; but something had to be done, and I had resolved on this move, as reported in my dispatch to General Halleck on July 1st:

General Schofield is now south of Olley's Creek, and on the head of Nickajack. I have been hurrying down provisions and forage, and tomorrow night propose to move McPherson from the left to the extreme right, back of General Thomas. This will bring my right within three miles of the Chattahoochee River, and about five miles from the railroad. By this movement I think I can force Johnston to move his whole army down from Kenesaw to defend his railroad and the Chattahoochee, when I will (by the left flank) reach the railroad below Marietta; but in this I must cut loose from the railroad with ten days' supplies in wagons. Johnston may come out of his intrenchments to attack Thomas, which is exactly what I want, for General Thomas is well intrenched on a line parallel with the enemy south of Kenesaw. I think that Allatoona and the line of the Etowah are strong enough for me to venture on this move. The movement is substantially down the Sandtown road straight for Atlanta.

McPherson drew out of his lines during the night of July 2d, leaving Garrard's cavalry, dismounted, occupying his trenches, and moved to the rear of the Army of the Cumberland, stretching down the Nickajack; but Johnston detected the movement, and promptly abandoned Marietta and Kenesaw. I expected as much, for, by the earliest dawn of the 3d of July, I was up at a large spy-glass mounted on a tripod, which Colonel Poe, United States Engineers, had at his bivouac close by our camp. I directed the glass on Kenesaw, and saw some of our pickets crawling up the hill cautiously; soon they stood upon the very top, and I could plainly see their movements as they ran along the crest just abandoned by the enemy. In a minute I roused my staff, and started them off with orders in every direction for a pursuit by every possible road, hoping to catch Johnston in the confusion of retreat, especially at the crossing of the Chattahoochee River.

I must close this chapter here, so as to give the actual losses during June, which are compiled from the official returns by months. These losses, from June 1st to July 3d, were all substantially sustained about Kenesaw and Marietta, and it was really a continuous battle, lasting from the 10th day of June till the 3d of July, when the rebel army fell back from Marietta toward the Chattahoochee River. Our losses were: