Schofield in person had followed the advance of Judah's division, and reconnoitred along the ridge parallel to Rocky Face on the east. It was plain that there was little chance of getting near Buzzard Roost by following Harker's path along the knife-like summit, and he was disposed to let Judah try the effect of a night attack upon the fortified outpost at the enemy's signal station in front of Harker. [Footnote: Id., pt. iv. p. 83.] Sherman realized that he could not hope to carry the Dalton lines from the west and north, and that Johnston was too well satisfied with his defensive position to leave it unless some part of our army was compromised by making a false move. McPherson, however, was entering Snake Creek Gap with so little opposition as to show that the importance of that pass was not understood by Johnston, if indeed he knew of its existence. Sherman therefore determined to keep up active demonstrations with watchful observation of the enemy for another day, whilst the decisive part of McPherson's movement should go on, and was already planning to transfer Hooker's Corps to McPherson's column as soon as the latter should hold the outlet of his gap. He wrote to Schofield, "We must not let Johnston amuse us here by a small force whilst he turns on McPherson." He sometimes suspected this was being done, and had been uneasy during the day at the absence of cannonade from Johnston's lines. [Footnote: Official Records, vol. xxxviii. pt. iv. pp. 83, 84.] The orders for the 9th were that Thomas should continue to push along the crest of Rocky Face from the north and make demonstrations on other parts of his line, whilst Schofield cautiously swung his left flank out toward the east at right angles to the principal ridge and made a strong reconnoissance of the enemy's lines in the immediate front of the town. At midnight Sherman learned that Hooker had made an effort to carry Rocky Face at Dug Gap, two or three miles south of Buzzard Roost, and had failed with considerable loss to Geary's division, which was engaged.
At daybreak on the 9th, my own camp was astir. The division advanced beyond the left flank of the position of Hovey's, then swung the left forward and moved southward astride of the ridge parallel to Rocky Face on the east. Judah's division connected our movement with the left flank of the Fourth Corps across the intervening valley. Hovey's division marched in rear of my left flank as a reserve. McCook's division of cavalry covered the extreme flank at Varnell's Station, under orders to demonstrate on the direct road to Dalton as our infantry advanced. [Footnote: Id., pp. 98-100.] The enemy resisted with strong outposts and skirmish lines posted in several strong barricades of timber and stones. We drove him from these and continued the movement till we confronted the main line of intrenchments. Schofield intended to attack these as soon as Newton's division of the Fourth Corps (which was our pivot) should be able to force the position in its immediate front on the crest of Rocky Face, but Newton was obliged to report that Harker's brigade had failed in its effort, [Footnote: Official Records, vol. xxxviii. pt. iv. p. 102.] and Schofield ordered us to stand fast where we were.
McCook had found a superior force of Confederate cavalry under Wheeler on the Dalton road; his advanced brigade under Colonel La Grange had been roughly handled, and that officer was captured. [Footnote: Id., p. 96.] General Stoneman was, however, advancing from Charleston with the cavalry of the Army of the Ohio, and the affair was of no great significance, though the Confederates claimed a considerable victory for their horse. [Footnote: Id., p. 683.]
Our movement had been an interesting one. As we went forward on the ridge, we could see Judah's line keeping pace with us in the valley and on the lower slopes of Rocky Face, whilst Newton's men continued the line to the summit, where Harker was having a sharp combat in which both artillery and small arms were brought into play. Off beyond our left was a separate rounded height, Potato Hill, on which the enemy had artillery which annoyed us, and to which our own guns answered. The space between was filled with skirmishers, horse and foot, and a rattling fusillade accompanied our march. It was evident that the lines before us were very formidable and held in force, and that the reconnoissance had been pushed as far as possible; to go further would commit us to a desperate attack upon intrenchments. [Footnote: When Johnston's retreat gave us possession of Dalton, we found the works of a very strong character, putting that front quite beyond a coup-de-main. I examined them myself later in the campaign.] But Sherman did not desire to do this. He wished to keep the enemy employed so that he could not send a great force against McPherson, and thus to give the latter a chance to make a success of the movement against Resaca. [Footnote: Official Records, vol. xxxviii. pt. iv. p. 98.] Toward evening he directed Schofield to fall back to a strong defensive position again, as from the news he got from McPherson he was sure Johnston must either attack us or retreat on the next day, and he wished to be ready for a prompt transfer of his army to Snake Creek Gap. But Schofield thought a night movement too uncertain in that broken and tangled country, especially as he had not been pleased with the handling of Hovey's division during the day, and obtained permission to bivouac for the night where we were, sending a couple of infantry regiments to support McCook's cavalry and cover our flank. [Footnote: Id., pp. 99, 119.]
During the night Sherman learned that McPherson had not succeeded in taking Resaca or breaking the railroad, and had retired to the mouth of Snake Creek Gap. Johnston was, of course, now aware of the turning movement, and before morning we had evidence that he was changing the positions of his army to meet the new situation. Sherman immediately turned his whole energy to transferring his army to McPherson's position. Hooker's Corps leading off was followed by Palmer's, and this by ours. Howard's was ordered to remain in position covering the Chattanooga railway, and to follow Johnston directly through Dalton when he left his intrenchments. The movement could not be begun till the 11th, as Stoneman with the cavalry of the Army of the Ohio was marching from Cleveland, and another day was needed to enable him to get upon our left flank, the place assigned him in the combined advance. [Footnote: Id., pp. 112, 113, 126.] Three days' rations in haversacks and seven more in wagons gave provisions for a short separation from our base, and orders to send back all baggage to Ringgold were strictly enforced.
At daybreak of the 10th I advanced my skirmishers to reconnoitre the enemy's lines, which were found to be still held in force, and his troops on the alert. We then proceeded to wheel the whole of the corps backward in line of battle, ready to halt at any moment, and engage the enemy if he should come out and attack us. My division being on the flank, it was to regulate the movement, Judah's conforming to mine on the right, and Hovey's in reserve immediately in rear of mine. [Footnote: Official Records, vol. xxxviii. pt. iv. pp. 123, 131.] We were under a warm skirmish fire of infantry in front, and the enemy's cavalry on our left flank also followed up the movement sharply. Reinforcing the skirmish line till the enemy was driven back, a good position in rear was selected for my second line and it was made to lie down. My first line was then marched slowly to the rear over the other, to another position, where it halted and lay down in turn, whilst the other rose and marched to the rear in a similar manner. Making the troops lie down avoided the danger, incident to such a manoeuvre under fire, that the men in second line would be confused by the passing of the first line through their ranks and break their organization. [Footnote: Officers experienced in war know that manoeuvres which are easy and of fine effect on parade are difficult and even dangerous under fire, and that it is wise to simplify the tactics as much as possible. Marshal Saint-Cyr, whose reputation for tactical skill was second to none in the wars of the French Republic and Empire, thus speaks of the matter in his comments on the battle of Novi, apropos to the break of the French division Watrin, which was in two brigade lines: "La première, attaquée avec vigueur par le général Lusignan appuyé par Laudon, ne soutint qu'un moment le choc, et se rabattit sur la seconde; elle espérait se reformer en arrière de celle-ci, en faisant ce qu'on appelle une passage de ligne; mais il fut démontré une fois de plus, que cette manoeuvre, qui fait un assez bel effet à la parade, ne peut réussir à la guerre lorsqu'on est suivi par un ennemi actif. La premiere entraîna la seconde dans un mouvement rétrograde; de plus elle y apporta assez de confusion pour que ces deux lignes réunies crussent n'avoir d'autre parti à prendre que celui de la fuite," etc. Mémoires, vol. i. p. 257. There can be no question as to the general soundness of this criticism, and we should not have continued the movement described if we had been attacked in force. We should then have fought where we stood, bringing the reserves to support the front line. It justifies, however, the precaution of selecting carefully the alternate positions and making the rear line lie down.] When we came opposite the positions assigned us in the extension of the Fourth Corps line, the division changed front to rear on right battalion and so swung into its place. [Footnote: Official Records, vol. xxxviii. pt. ii. p. 675.] Sherman had sent Captain Poe, his chief engineer, to observe our movement from the crest of Rocky Face held by Newton's troops, and congratulated Schofield upon it, saying it "was described to me by Captain Poe, as seen from the mountain, as very handsome." [Footnote: Id., pt. iv. p. 121.] In his full report made at the close of the campaign, General Schofield referred to it as "a delicate and difficult one, owing to the character of the ground, the position and strength of the enemy, and our comparative isolation from the main army." He adds: "I regarded it as a complete test of the quality of my troops, which I had not before had opportunity of seeing manoeuvre in presence of the enemy." [Footnote: Id., pt. ii. p. 510.]
Schofield had been so dissatisfied with General Hovey that on the same day he asked to have him removed from the command of the division, notwithstanding his high personal esteem for him and his confidence in his personal gallantry. The trouble seemed to be in the comprehension of orders and in the grasp of the surrounding circumstances. Sherman did not feel at liberty to act on the request, as Hovey had been assigned to the new division, before it took the field, in fulfilment of a promise of General Grant under whom Hovey had served in the Vicksburg campaign, and had been recommended for promotion as a recognition of good conduct at the affair of Champion Hill. [Footnote:Id., pt. iv. p. 122. Brigadier-General Alvin P. Hovey had been a Judge of the Supreme Court of Indiana, and a "War Democrat" in politics. His subsequent withdrawal from the army and his connection with Sherman's famous protest against promotions given under stress of personal and political influences at Washington would not be entirely clear without mention of the incident here told.]
Johnston seems to have heard rumors of Sherman's original plan to send McPherson's column against Rome, much further in rear, and he remained under the impression that this was the meaning of the movements he now heard of, until McPherson was in possession of Snake Creek Gap. [Footnote: Official Records, vol. xxxviii. pt. iv. pp. 674, 675.] On the 7th he had urged Polk to hasten his concentration at Rome, and ordered Martin's division of cavalry to Calhoun to cover the communications with Polk, and protect the railroad south of the Oostanaula. Brigadier-General Cantey was at Resaca with at least four thousand men, his own and Reynolds's brigades with fourteen pieces of artillery. [Footnote:Id., pp. 679, 682.] Movements toward his rear were reported to Johnston, and all his subordinates were on the alert to find out what it meant; the cavalry was ordered to watch all gaps south of Dug Gap, but no mention is made of Snake Creek Gap till McPherson had passed through it. [Footnote: Id., pp. 681, 683, 686, 687.] Then Cantey was told to hold Resaca firmly, and call on Martin for assistance if he needed it. Cars were sent to bring a brigade from Rome, intrenchments were made to cover the south end of the Resaca bridge; Major Presstman, chief engineer, was sent to mark out more extensive works about Resaca, and Hood was ordered there with considerable reinforcements. As soon, however, as it was known that McPherson had retired to Sugar Valley, Hood was called back to Dalton, and Johnston requested Polk to hasten in person to Resaca and take command, hurrying forward his corps as fast as possible. [Footnote: Id., pp. 687, 689.] This was the situation on the evening of the 10th of May.
When we took our position on the ridge of Rocky Face as the left of the line, the division was somewhat exposed to a flank attack, and I ordered the fallen timber on the hillside to be thrown together to make obstruction to any hostile advance, besides the usual tactical precautions of outposts and reserves. This, like the slashing made at Red Clay a few days before, was suggested by the difficulty of knowing what was going on around us in a country covered by dense forests with only small cultivated openings here and there. In this instance it was only the gathering of logs and tree-tops already lying on the ground, and utilizing them as a means of delaying an enemy till our lines could be formed. From such beginnings grew up our more and more elaborate system of intrenched camps; a natural evolution of campaigning in a country only partially cleared, with no roads worthy of the name.
To pass such a defile as Snake Creek Gap with an army was no small undertaking. Hooker was ordered to clear a second track, so that two lines could march by the flank at once, but this could only be imperfectly done in the time at command. [Footnote: Official Records, vol. xxxviii. pt. iv. pp. 126, 135, 145.] Careful orders in detail were made, fixing the time for each corps and division to move, keeping the roads filled night and day. Wagons were sent by the rear to Villanow, and the regular subsistence trains were stopped at Ringgold and Tunnel Hill till the Confederate army should be dislodged. For night marching men were stationed with torches at the forking of paths, and boards were nailed to trees as finger-posts.