It was whilst this movement was progressing, on the 20th of July, and was near its completion, that Hood made the attack already planned by Johnston, upon Thomas's columns, crossing Peachtree Creek by several roads converging at Atlanta. It involved the right of Howard's corps, the whole of Hooker's, and the left of Palmer's. It was a fierce and bloody combat, in which the Confederates lost about 6000 men in killed and wounded, whilst the casualty lists of Thomas's divisions amounted to 2000. Again, on the 22d, the second part of Johnston's plan was tried, and Hardee's corps, moving by night through Atlanta and far out to the southward of Decatur, advanced upon the flank of McPherson's army, whilst Cheatham at the head of Hood's own corps advanced from the Atlanta lines and continued the attack upon the centre and left of McPherson and upon the right of Schofield. A great battle raged along five miles of front and rear, but at evening the worsted Confederates retired within the fortifications of the city, a terrible list of 10,000 casualties showing the cost of the aggressive tactics. The losses on the National side were 3500, heavy enough, in truth, but with very different results on the relative strength of the armies and their morale. But the end was not yet. On the 28th McPherson's army, now under the command of Howard, was marching from the left wing to the right, to extend our lines southward on the west side of Atlanta, when once more Hood struck fiercely at the moving flank at Ezra Church, but again found that breastworks grew as if by magic as soon as Howard's men were deployed in position, and again the gray columns were beaten back with a list of 5000 added to the killed and disabled. Howard had less than 600 casualties in the action. It was only a week since Johnston had been relieved, and matters had come to such a pass in his army that the men stolidly refused to continue the assaults. From our skirmish line their officers were seen to advance to the front with waving swords calling upon the troops to follow them, but the men remained motionless and silent, refusing to budge. [Footnote: For details of these engagements, see "Atlanta," chaps, xii.-xiv.]
During the first half of August Sherman extended his lines southward, until my own division, which was the right flank of the infantry lines, was advanced nearly a mile southeast of the crossing of the Campbelltown and East Point roads on high ground covering the headwaters of the Utoy and Camp creeks. We were here somewhat detached and encamped accordingly in a boldly curved line ready for action on the flanks as well as front. It was now the 18th of August and Sherman devoted the next week to the accumulation of supplies, the removal of sick and wounded to the rear, getting rid of impedimenta, and general preparation for a fortnight's separation from his base. My position had been selected with reference to this plan, as a pivot upon which the whole of the army except the Twentieth Corps should swing across the railways south of Atlanta.
The movement began on the 25th, and we stood fast till the 28th, when we began our flank movement on the inner curve of the march of the army, taking very short steps, however, as we must keep between the army trains and the enemy. On the 30th Schofield moved our corps from Red Oak Station, on the West Point Railroad, a mile and a half directly toward East Point, so as to cover roads going eastward toward Rough-and-Ready Station on the Macon road. We were hardly in position before our skirmishers were briskly engaged with an advancing force of the enemy's cavalry, and we felt sure that it was the precursor of an attack by Hood in force. It proved to be nothing but a reconnoissance, and showed that Hood was strangely misconceiving the situation. Its chief interest to me at the moment was in the experiment it enabled me to make of the speed with which my men could cover themselves in open ground in an emergency. The division was astride the East Point road, the centre in open fields where no timber could be got for revetment, and only fence rails to give some support to the loose earth. Giving the order to make the light trench of the rifle-pit class, where the earth is thrown outward and the men stand in the ditch they dig, in fifteen minutes by the watch the work was such that I reckoned it sufficient cover to repel an infantry attack, if it came. It would be an extraordinary occasion when we did not have more warning of an impending attack; and the incident will illustrate the confidence we had that in forcing the enemy to assume aggressive tactics, the campaign was practically decided.
On the 31st, as Sherman's left wing, we held the Macon Railway at Rough-and-Ready Station, Howard, as right wing, was across Flint River, closing in on Jonesboro, whilst the centre under Thomas filled the interval. Hood had sent Hardee with his own and Lee's (late Hood's) corps to defeat what was supposed to be a detachment of two corps of Sherman's army, and a sharp affair had occurred at the Flint River crossing, where Howard succeeded in maintaining his position on the east side. On hearing of our occupation of Rough-and-Ready, Hood jumped to the conclusion that it was preliminary to an attack on Atlanta from the south, and ordered Lee's corps to march in the night and rejoin him at once. Getting a better idea of the situation before morning, he stopped Lee and prepared to evacuate Atlanta. On September 1st Sherman closed in on Jonesboro, his latest information indicating that two corps of the enemy were assembled there. Late in the day he learned of the disappearance of Lee's corps, but assumed that Hood was assembling somewhere near. He tried hard to concentrate his forces to prevent Hardee's escape, but his scattered army could not be united till nightfall.
In the night Hood blew up the ordnance stores at Atlanta, and hastening to join Lee by roads east of Sherman's positions, he marched on Lovejoy Station. Hardee evacuated Jonesboro also, and before morning the Confederate army was assembled again upon the railroad, five miles nearer to Macon. Atlanta was occupied by the Twentieth Corps on the 2d, and Sherman ordered his army to return to the vicinity of that city for a period of rest. Hood's conduct for the past three days had been the result of complete misapprehension of the facts; but its very eccentricity had been so incomprehensible that no rule of military probabilities could be applied to it, and before Sherman could learn what he was doing, the time had passed when full advantage could be taken of his errors.
The condition of Hood's army at the close of the campaign was anything but satisfactory to him. His theory was that his offensive tactics would keep up the spirit and energy of his men and constantly improve their morale. When he found that they were, on the contrary, discouraged and despondent, and could not be induced to repeat the assaults upon our positions which had followed each other so rapidly in the last days of July, he querulously laid the blame at the door of his subordinates. He called the attack upon Howard's advance at Flint River "a disgraceful effort" because only 1485 were wounded, and asked to have Hardee relieved and sent elsewhere. [Footnote: Official Records, vol. xxxviii. pt. v. pp. 1021, 1023, 1030. Hardee had before asked to be relieved. (Id., pp. 987, 988.) For Hood's final, urgent request and the result, see vol. xxxix. pt. ii. pp. 832, 880, 881.] True, he had telegraphed Hardee that the necessity was imperative that the National troops should be driven into and across the river, and that the men must go at them with bayonets fixed; but it was his own old corps, now under Lieutenant-General S. D. Lee, that made the principal attack and was repulsed. Lee was not one of the officers who might be presumed to be discontented with Johnston's removal, but had been brought from the Department of Mississippi, at Hood's suggestion, to take the corps when the latter was promoted, and had won Davis's admiration by his zeal. [Footnote: Id., p. 892, and vol. lii. pt. ii. p. 713.] It would be hard to find better proof that the trouble lay in the consciousness of the men in the line that they were asked to lay down their lives without a reasonable hope of benefit to their cause. The discouragement pervaded the whole army, and is seen in Hood's own dispatches hardly less than in others. [Footnote: Hood to Davis, September 3, two dispatches, Id., vol. xxxviii. pt. v. p. 1016. In another, p. 1017, he repeated an earlier suggestion to remove the prisoners from Andersonville. When Johnston had done this, it was made one of the charges against him. See Davis to Lee, Id., vol. lii. pt. ii. p. 692. For Hardee's opinion of the situation, see Id., vol. xxxviii. pt. v. p. 1018.] In a labored letter to Bragg on September 4th, he unconsciously shows how his own total misunderstanding of Sherman's movements was the prime cause of his disaster, whilst the shame at the result leads him to charge it upon others. As to the spirit of the army, nobody has given more telling testimony, for he says, "I am officially informed that there is a tacit if not expressed determination among the men of this army, extending to officers as high in some instances as colonel, that they will not attack breastworks." [Footnote: Official Records, vol. lii. pt. ii. p. 730. This letter seems to have come to light since the first publication of the records of the campaign, and is found in the supplemental volume.]
In the correspondence between Johnston and the Confederate government regarding the numerical force of his army, he naturally emphasized his inferiority to Sherman in numbers as an explanation of his cautious defensive tactics and his retreating movements. The introduction into the Southern returns of a column of "effectives" as distinguished from the number of officers and men "present for duty," [Footnote: Ante, vol. i. p. 482.] led to a habitual underestimate by their commanding officers. On several occasions Johnston defended his conduct of the campaign by asserting that his army was less than half the size of Sherman's, [Footnote: Official Records, vol. xxxviii. pt. iv. p. 795.] and this necessarily led to an examination of his returns. These regular numerical reports are of course the ultimate authority in all disputes, and we find the Richmond government doing just what the historian has to do,--comparing the estimates of the general with his official returns. Officers of all grades and of the highest character fall into the error of memory which modifies facts according to one's wish and feeling. Thus at the beginning of this campaign we find General Bragg, speaking for the President, saying that General Polk's "estimates and his official returns vary materially." [Footnote:Id., vol. lii. pt. ii. p. 659.] Nobody could be freer from intentional misstatement than the good bishop-general. We find the same discrepancies at the East as well as the West. Lee, Jackson, Longstreet, and their subordinates fall into the same error. It is therefore the canon of all criticism on this subject, that nothing but the statistical returns in the adjutant-general's office shall be received as proofs of numbers, though, of course, the returns must be read intelligently.
Conscious of straining every nerve to reinforce the great armies in the field, Mr. Davis naturally asked what it meant when the army in Georgia was said to be so weak. General Bragg assisted him with an analysis of Johnston's last returns. Writing on June 29th, he refers to the last regular return, that of June 10th, which is the same now published in the Official Records. In using it, therefore, we agree with the Confederate government at the time in making it conclusive. It shows that Johnston's army had present for duty 6538 officers and 63,408 enlisted men, or, in round numbers, was 70,000 strong. [Footnote: Official Records, vol. xxxviii. pt. iv. p. 805; Id., pt. iii. p. 677.] The "effectives" are given as 60,564; but this, as we know, is the result of subtracting the number of the officers and non-commissioned staff from the aggregate present for duty. But in addition to the troops named, Bragg very properly adds that Johnston "has at Atlanta a supporting force of reserves and militia, estimated at from 7000 to 10,000 effective men, half of whom were actually with Johnston near Marietta." We thus have from Confederate authorities the proof that the army was nearly 80,000 strong on June 10th, after the first month of the campaign had closed, including the engagements at Dalton, Resaca, New Hope Church, Dallas, and Pickett's Mill.
To complete the examination of the same return, it is necessary to notice that the "aggregate present" is given at 82,413, or 12,500 more than the "present for duty." This includes "extra-duty men," such as clerks at headquarters of the organizations from Johnston's own down to brigades and regiments, men permanently detailed for any special service, men in arrest, etc. [Footnote: Hood's dispatch of September 5, Id., pt. v. p. 1021; and his Order No. 19, vol. xxxix. pt. ii. p. 835.] It is here that good administration in an army seeks to reduce the number of those who are withdrawn from the fighting ranks, and to make the "aggregate present" agree as closely as possible with the "present for duty." I shall presently note the result of such an effort.