Snake Creek


An Episode in the Retreat of the Confederate Army

of Tennessee from Dalton, Georgia, May, 1864

By Wilbur G. Kurtz

In that considerable body of literature dealing with the Atlanta Campaign of 1864, one item has remained a challenging hiatus – – the unguarded Snake Creek Gap. This aroused some curiosity at the time among certain Confederate officers and in after years there was some critical comment in that quarter, but the matter generally has received a silent treatment that apparently stems from a concerted effort either to ignore it or consign it to Limbo.

This is all the more arresting since no account of the retreat from Dalton makes sense without relating that the Federal forces under Sherman maneuvered Johnston from his embattled position there by merely marching through an unguarded mountain pass in his rear. Even the veriest fireside historian pauses to wonder how this could have happened, but if he continues his reading with the hope that all will be cleared up, he’s doomed to disappointment.

Nothing could have pleased the Federal high command better if General Johnston had, in sortie after sortie from the mountain fastness of Dalton, exposed his forces to the superior fire-power of the Cumberland, Tennessee and Ohio armies. But Johnson had no intention of doing that and by the same token, Sherman wasn’t minded to storm the gaps north and west of the town with the fell purpose of breaking through by sheer weight of numbers. The Federal problem then was to maneuver Johnston into the open and try conclusions with whatever was presented.

Dalton, shire town of Whitfield, is situated in the rolling valley of the Connasauga River. On the west is a longish knife-like ridge called Rocky Face — its western scarp largely of perpendicular cliffs, hence its name. Two gaps, one natural, one artificial give access to the immediate environs from the West. The upper pass, three miles north-west of the town is called Buzzard Roost or Mill Creek Gap. It marks the place where the raging waters of Mill Creek finally forced a passage through Rocky Face and the stream, now that its struggle, has long since abated, flows quiescently eastward toward the Connasauga. In addition to the stream the gap is traversed by the ancient highway now known as U.S. 41, and the Western & Atlantic R.R., now known as the Nashville, Chattanooga and St. Louis R.R.

Three and one-half miles southwest of Dalton is the lower pass. In the 1840’s or ‘50’s, certain enterprising gentry cut a road to a notch in the profile of Rocky Face from both sides, deepened the notch by excavation and thereby established a road of sorts from Dalton to Villanow and LaFayette; by common usage it was called Dug Gap.

Rocky Face tapers off seven miles northward of Dalton and the broken area permits passage from the village of Tunnel Hill, via Harris Gap, to the head of Crow Valley which is eastward of and parallel to Rocky Face as far southward as Mill Creek. Low, parallel ridges lie eastward of Crow Valley and these, together with the valley, pointing as they do toward Dalton from the north, constitute portentous but defensible routes thereto.

Rocky Face feathers out some seven miles below Dalton but a corrugated terrain prolongs it in a parallel but staggered series of fragmentary ridges to the southwest. Between two of these fragments is a long, narrow valley with a trickle in it that flows southward to the Oostenaula River. This rivulet, “spanned by a heron’s wing, crossed by a stride,” is dignified by the name of Snake Creek, which would have no significance on the map but for the fact that one can enter its valley from the western side of Rocky Face and emerge on the east side of it and the succeeding ridges with little or no handicap from the rocky road that is parallel to, and sometimes identical with, the stream. It is a true mountain pass, but Snake Creek Gap is its name.

This gap is some four miles in length. A few miles from its upper or western end is the crossroads hamlet of Villanow — a name given it by some pioneer settler who had read Jane Porter’s Thaddeus of Warsaw. At Villanow, State Highway No. 2 runs northward to join U.S. highway 41 a little above Mill Creek Gap. The road westward from Villanow, en route to LaFayette, crosses Taylor’s Ridge at Shipp’s Gap some four and one-half miles distant. Between Shipp’s Gap and Ringgold Gap, Taylor’s Ridge is broken by two other gaps: Gordon Springs Gap and Nickajack Gap.

The lower or eastern and of Snake Creek Gap is twelve miles, as the crow flies, from Dalton railroad station. Fifteen miles south of Dalton on the W. & A. R.R. is Resaca, named for a Mexican War battle. It is on the north bank of the Oostenaula River, which is formed by the junction of the Connasauga and the Coosawattee rivers. Resaca was an insignificant hamlet said to have had a population of 191 in 1860 – its establishment due to the location of a railroad and bridge construction camp in the 1840’s. In 1864 Resaca was connected by two wagon roads with Dalton; one the present U.S. 41 highway and the other, a much more traveled road then, was close to and parallel with the eastern side of Rocky Face. This Mountain Road, as General Howard named it, branched to Dug Gap about four miles south of Dalton; branched lower down to Tilton and at the county line forked, the two branches crossing the one from Snake Creek Gap to Resaca. The west fork went via Sugar Valley P.O. and Lay’s Ferry to Rome; the east fork to Calhoun of McGinnis’ Ferry, then to Calhoun where it joined the present U.S. highway 41.

Above the upper end of Snake Creek Gap were a number of neighborhood roads west of Rocky Face that connected Ringgold and Tunnel Hill with the gap, so that one could go from these places to Resaca, Calhoun and Rome without passing through Mill Creek Gap and Dalton.

It was only by accident that Bragg’s army stopped at Dalton on the retreat from the disaster at Missionary Ridge. Pursuit by the Federals had been halted at Ringgold Gap by General Patrick R. Cleburne’s heroic defense of that pass, November 27, 1863, and the Army of Tennessee settled down in winter quarters at Dalton for want of a better place to go. General Johnston superseded General Bragg as army commander, December 27, and headquarters were set up at the Cook house, better known as the Mrs. William C. Huff place on Selvidge Street. In March 1864, Johnston transferred headquarters to the Tibbs home on N. Hamilton Street, where he remained until the night of May 12, when Dalton was evacuated by the Confederate army.

Johnston had no high opinion of Dalton as a defensive position. He states that “It had neither intrinsic strength nor strategic advantages. It neither fully covered its own communications, nor threatened those of the enemy”1. Rocky Face on the West, while a formidable barrier against aggression, favored the Federals by masking their movements from the West, and its limited extent permitted access to the Eastward through Harris’ Gap north of Tunnel Hill and through Snake Creek Gap near Resaca. There were no natural barriers of any consequence east and southeast. But that four mile stretch of the narrow valley of Snake Creek was a perfect corridor to the railroad bridge over the Oostenaula at Resaca; but for that railroad, stretching some ninety miles to Atlanta, Johnston could have remained as long as he did at Dalton.

Only those military movements bracketed by the dates May 1 and 12 are pertinent to this account, In brief, they are as follows: Sherman moved the 4th and 23d Corps from Cleveland, Tennessee, via Varnell’s Station and Catoosa Springs, to the northward of Dalton — Tunnel Hill being the first objective. The 14th Corps advanced from Ringgold toward Mill Creek Gap on what is now U.S. highway 41. The 20th Corps crossed Taylor’s Ridge at Nickajack and Gordon Springs gaps, having marched thither from Chattanooga by way of the Chickamauga battlefield. McPherson’s Army of the Tennessee (the 15th and part of the 16th Corps) moved from Chattanooga via Lee and Gordon’s Mill and at Rock Springs Church turned off the LaFayette road and moved toward Taylor’s Ridge – – the 16th Corps seizing Shipp’s Gap the night of the 7th, and marching toward Villanow on the 8th, followed by the 15th Corps which had crossed the ridge at Gordon Springs Gap. While Geary’s division of the 20th Corps made a demonstration in force at Dug Gap on the 8th, the 16th Corps, followed by the 15th, occupied Snake Creek Gap, unopposed. On the 9th, McPherson issued from the eastern end of Snake Creek Gap and assailed the Confederates at Resaca, commanded by Brig. Gen. James Cantey. Being strongly entrenched, Cantey’s forces prevented McPherson’s two corps from reaching the railroad and the bridge and, fearing that his left and rear were exposed because of the Mountain Road (the one a just east of Rocky Face). McPherson withdrew to the mouth of the gap. Here he remained until the 13th, while Sherman was moving the rest of the army through the gap – – excepting the 4th Corps which moved directly by way of Dalton to Join him in the bitter three days of battle on the Resaca hills. This sketchy outline remains to be filled in by the circumstances which relate to the unguarded gap.

Sherman’s strategy as of the 8th and 9th, which is still a matter of controversy, seems to have been one of demonstrations at the upper end of Rocky Face and in Crow Valley, with like demonstrations on the West at Mill Creek and Dug gaps, while McPherson’s two corps plunged into the narrow Snake Creek pass with instructions to press eastward, seize and break the railroad at Resaca.

To checkmate these moves, Johnston’s forces were placed on Rocky Face at each side of Mill Creek Gap at such points where natural formations were more or less assailable, and also in a line of field works across Crow Valley which ran from Rocky Face to Potato Hill near the Cleveland railroad. Dug Gap was occupied by the 1st and 2nd Arkansas mounted rifles (dismounted) of Reynolds’ brigade. These Arkansans, commanded by Colonel James A. Williamson, 2 had taken over at Dug Gap on the 7th, relieving an unidentified contingent.

We are here concerned with such precautions as were taken to appraise the Confederate high command of the approach of the Federals toward Dalton. Johnston knew that the bulk of Sherman’s forces were to the northward of the town, and there was sufficient evidence that a force under McPherson was poised at Guntersville in northern Alabama, for a move to the Confederate left by way of Rome and Kingston, Georgia — a move that would sever the life-line of the Dalton defenders, the Western & Atlantic RR. But Johnston did not discover, until too late, that Sherman had changed his plan and diverted McPherson from Rome to objectives much nearer Dalton, and pursuant to this change, McPherson had been hastily shifted from Guntersville to Chattanooga from which point a rapid march would carry him by way of Shipp’s Gap to the rear of Dalton at or near Resaca 3. Sherman could not have known, of course, that this sudden shift of McPherson’s Army of the Tennessee would leave the Confederate high command in the firm belief that the original plan would be carried out – a belief that persisted until McPherson was actually in possession of Snake Creek Gap. It should be remarked that the Federal commands, which had observed the Dalton defenders all the previous winter, had full knowledge of the existence of Snake Greek Gap and its relationship to the railroad in the immediate rear of Dalton, and this is the more remarkable in the light of subsequent events which seem to indicate that the Confederate staff had little or no knowledge of its existence, to say nothing of its strategic implications. No other hypothesis adequately explains what happened.

Sherman’s diversion of McPherson from the road to Rome to an objective near Dalton was necessary because that general had not been able to assemble four divisions counted on when the original plans were made, hence a more restricted operation was deemed advisable.

To meet this alleged menace from the Roman front, Johnston and the Richmond authorities laid down a barrage of dispatches to Lieutenant General Leonidas Polk at Demopolis, Alabama, directing him to mobilize all his forces at Rome. As for the approaches to Dalton, they were patrolled by Wheeler’ a cavalry. Scouts were maintained along the Tennessee River west of Chattanooga, also in the Chickamauga valley and to the northeast of Dalton.

In his report, Wheeler stated that the left of his picket line in the Chickamauga valley was at Shipp’s Gap 4 in Taylor’s Ridge and, inferentially the two gaps northward — Gordon Springs and Nickajack — were patrolled. Ringgold Gap, in the same ridge, had been in Federal hands since Cleburne’s retreat from there December 27, 1863.

By May 1st, it was no secret that Sherman’s forces were in motion. On that day, General Mackall, Johnston’s Chief of Staff, sent instructions to Major General Joseph Wheeler, whose cavalry was mainly posted at Tunnel Hill, to order his scouts then in observation along the Tennessee River from Bridgeport to Gunter’s Landing (Guntersville), to report all Federal moves to the commanding officer at Rome, also to check on rumors of Federal supplies en route from Chattanooga to Ringgold 5. This order was designed to take care of any and all Federal advances southward of the Tennessee River between Guntersville and Chattanooga.

On the 4th, headquarters again called Wheeler’s attention to the Federal concentration at Gunter’s Landing,6 and both Johnston at Dalton and Inspector General Cooper at Richmond wired Polk at Demopolis to hasten all available troops and go in person to Rome to meet this threat toward Johnston’s left and rear 7.

These dispatches, together with one sent to Bragg (same date) indicate that Johnston was sufficiently aware of the transfer of Federal troops from northern Alabama to Georgia, but while a concentration on the immediate Dalton front was part of the Federal strategy, there was nothing to indicate an abandonment of the move on Rome. Certain other dispatches of the 4th and 5th, foreshadowed formidable Federal concentration from Chattanooga to Ringgold and from Cleveland to Varnell’s Station 9. These tended to indicate that scouting should be balanced by patrols immediately westward. Perhaps this inspired Mackall’s dispatches of the 5th; having learned of Federal forces south of Ringgold on the main Dalton road, it was not impossible that some of those troops might by-pass Mill Creek Gap and penetrate to the southward, hence an augmented patrol of the Chickamauga valley westward was necessary. Wheeler was directed “to acquire the most authentic information of the present position of the enemy this side and to the east of Taylor’s Ridge” 11. “East” was obviously intended for “West,” but in the light of subsequent directives the emphasis seems to have been on the area between Rocky Face and Taylor’s rather than beyond the latter, for the McPherson phantom was more of an obstruction to clear vision than Taylor’s Ridge. Whatever scouting was done in that quarter and whatever reports were sent in, headquarters either failed to get the reports or scanned such as were received with a purblind conception of their import.

Confirmation of this is found In the following: on the 6th Johnston intercepted three batteries of Robertson’s artillery belonging to Wheeler’s command en route from Oxford Mississippi, to Carterville Georgia, and diverted them to Rome12, and on the 7th, Johnston read a dispatch that confirmed his every belief in the McPherson phantom. Some scout reported he had seen a division of McPherson’s troops between Lee and Gordon’s Mill and LaFayette, and the rest of the corps was closing up. This was all true enough, but Johnston, in a dispatch to Polk, was positive that this meant McPherson was threatening “Our left and rear by the Oostenaula from LaFayette”13, therefore the urgency to concentrate at Rome was greater than ever. Guntersville didn’t fit this picture but McPherson did; indeed any report on or glimpse of, Federal troops on any road west of Taylor’s Ridge was interpreted as a far-flung move upon the left rear of Dalton – – an excellent example of twisting facts to conform to a fixed idea. As a matter of fact, McPherson was on the LaFayette road, as reported, but he turned left or southeastward at Rock Spring Church and marched his two corps to Shipp’s and Gordon Springs gaps of Taylor’s Ridge.

Ironically, on the same day, the 7th, Mackall reiterated his order of the 5th to Wheeler, this time specifying a patrol of the valley west of Rocky Face and south of Mill Creek Gap. These scouts were to report to general headquarters by way of “the gap five miles below Dalton on the Villanow road”14. This was to be done if Wheeler’s men were driven, as was anticipated, from Tunnel Hill into Mill Creek Gap. As it happened, ‘Wheeler’s cavalry was driven from Tunnel Hill by the 4th and 14th corps by 11:00 a.m.15 and pursuant to Mackall’s order, cavalry patrols should have been in the valley that afternoon and evening, west of Rocky Face and Dug Gap — the one “five miles below Dalton.”

But if a cavalry patrol was in the valley that afternoon and evening, it could not have discovered any Federals unless the patrol had scouted to the westward as far as the gaps of Taylor’s Ridge and continued the vigil throughout the night. It would seem there were no Scouts either in the valley or at the gaps of the ridge that afternoon and evening, or if so, no word of the movements of McPherson and Hooker reached headquarters. Just what engaged Wheeler’s attention that afternoon after his retreat from Tunnel Hill to Mill Creek Gap, is clear only in one respect; he sent Grigsby’s cavalry, during the late afternoon or evening, not to the valley west of Rocky Face, but to Dug Gap by a route eastward of it. If this was not contrary to Mackall’s orders, what was it?

Colonel J. Warren Grigsby commanded a brigade of Kentuckians in Hume’s division of Wheeler’s cavalry corps. Perhaps the best-known officer in the brigade was the commander of the 9th Kentucky, Colonel W. C. P. Breckinridge. Some twenty years later, Colonel Breckinridge wrote what is perhaps the only account of Grigsby’s brigade at Dug Gap16. We infer from his narrative that the brigade reached Dug Gap from Mill Creek Gap by way of the road east of Rocky Face and on arrival, found the two small Arkansas regiments commanded by Colonel J. A. Williamson, previously mentioned, who had been sent there by Mackall earlier in the day. These troops, the 1st and 2nd Arkansas (dismounted rifles) belonged to Reynold’s brigade which, together with Cantey’s brigade, had been transferred from the Department of the Gulf to reinforce Johnston’s command at Dalton, but by May 7th, only Cantey’s brigade and the two Arkansas regiments had reached Resaca where Cantey was halted and the Arkansans were sent forward to Dug Gap17.

Arrived at the eastern approach of the gap, Dortch’s battalion of Grigsby’ s brigade ascended to the summit and joined Williamson’s regiments in camp, presumably leaving the rest of the brigade to bivouac at the base of the ridge. Here a belated order from Wheeler was received by Grigsby to place a regiment in the valley south and west of Mill Creek Gap. This was in reiteration of the order Mackall had issued to Wheeler at 7:00 a.m. and if we are curious why Wheeler hadn’t complied with the Mackall directive immediately after his withdrawal from Tunnel Hill, we share that curiosity, not to say annoyance, with Mackall who wouldn’t have heard about it if Williamson hadn’t reported the arrival of Dortch’s battalion at the top of the gap. Wherefore, in the reiteration of the 7:00 a.m. order, Mackall waxed peremptory, even to quoting Johnston “that it is absolutely necessary that a good regiment of cavalry under an officer on whom you can fully rely, be sent into the valley south and west of Mill Creek Gap. Please give positive orders on this subject; it will not do to leave that valley open tonight”18. Pursuant to this order, Breckenridge’s 9th Kentucky was sent down the western scarp to patrol the valley as directed.

It required three orders from headquarters to got cavalry scouts into the valley between Rocky Face and Taylor’s Ridge; one of the 5th and two on the 7th. And the irony of it was that these orders envisaged nothing more than a precaution against superstitious Federal advances southward from Ringgold Gap. Not once are the gaps of Taylor’s Ridge south of Ringgold mentioned in the three orders.

And as if this was not enough of irony, an overweening reliance a upon Wheeler’s alertness was betrayed in a casual remark by Mackall in a dispatch of May 5th: “The general understands that your picket lines hold Taylor’s Ridge”19. In his report Wheeler stated that the left of his picket line on the west of Dalton was at Shipp’s Gap. But the irony in this case was not that Wheeler had no pickets on the ridge, but rather in the inevitable misunderstanding that awaited any report of Federal movements in that quarter, as we have seen, added to which was the failure of the pickets, if any, to apprise headquarters the night of the 7th that far from marching on Rome, McPherson was in Shipp’s Gap. It remained for Breckinridge’s troopers to discover that McPherson’s Army of the Tennessee was swarming on practically every road between Rocky Face and Taylor’s Ridge.

The 9th Kentucky, after descending the western scarp at Dug Gap, made this discovery in the small hours of the 8th. For once, headquarters was favored with definite news from one of Wheeler’s officers west of Rocky Face. Breckinridge dispatched Mackall from: “One mile and a half from Dug Gap on Villanow, Dug Gap and Dalton Road, May 8, 1864, 10:15 a.m. The Kentucky colonel reported that Federal cavalry had advanced eastward from the Presbyterian Church, mentioned in a previous dispatch (Where is that dispatch?) and another body of horsemen was on the Dalton-Villanow road (State Highway No. 2) only three miles from his pickets, headed southward from the direction of Huff’s house20. From the same point, Breckinridge sent another dispatch to Mackall at 1:30 p.m. He had ascertained that Kilpatrick’s cavalry, supported by a body of infantry, was moving toward Villanow; Dr. Anderson’s house was cited as being on Kilpatrick’s line of march21.

These two dispatches, checked with Federal reports, are singularly accurate, but it is doubtful if Breckinridge realized at the time that the troops seen by his scouts had crossed Taylor’s Ridge and had not arrived at Anderson’s and Huff’s by merely marching southward from Ringgold. Breckenridge’s scouts saw the advance of the right wing (Geary’s division) of Hooker’s 20th Corps, which had reached Huff’s on the night of the 7th after crossing Taylor’s Ridge at Gordon Springs Gap — the rest of the corps having reached Anderson’s and Trickum via Nackajack Gap, and bivouacked in Dogwood Valley. Geary’s division had marched by the springs and Huff’s to one of the Thornton farms where two of his brigades camped, while a third, Ireland’s, was detached to accompany Kilpatrick’s horsemen to Villanow to meet and support Dodge’s advance from Shipp’s Gap to Snake Creek Gap, May 8th.

It is certain that Breckinridge knew nothing at this time of McPherson’s objective – Snake Creek Gap and Resaca. The Colonel was soon preoccupied with the advance of Buschbeck’s and Candy’s brigades of Geary’s division, which drove the Kentucky troopers back to Dug Gap.

These brigades moved into Babb settlement about 3:00 p.m. Geary planted his rifled Rodmans in Joel Babb’s backyard, and after a few shells the struggle for the gap began — memorialized in bronze on the Brown monument on the Atlanta Capitol lawn. Geary stated that this battle was for the purpose of covering the march of McPherson’s Army of the Tennessee through Villanow to Snake Creek Gap. In that it was eminently successful. Another angle, however, was the hope that Dug Gap could be seized so that the Federals would have ready access to Johnston’s retreating columns when and if Dalton was abandoned, but Geary’s climbing columns were repulsed by the horsemen in the sky.

Numerous reports of heavy firing at Dug Gap prompted Mackall to order General Cleburne to support Grigsby’s brigade. At this time Cleburne was posted east of Dalton on the middle Spring Place road, and two of his brigades, Lowrey’s and Granbury’s, made a rapid march in spite of the oppressive weather, reaching the ascent about 7:00 p.m. General Hardee accompanied General Cleburne22. By dark Cleburne’s two brigades had relieved Grigsby’s cavalry and Geary had withdrawn his baffled forces.

Though Grigsby’s men had been on duty some fifteen or twenty hours, the brigade, on being relieved by Cleburne’s division, was ordered to the eastern mouth of Snake Creek Gap. Had Breckenridge not written his narrative, it is probable we would never have had the details Of this move23. He points out that this was the only provision made to hold the gap and adds that “if a cavalry force had been started to Snake Creek Gap at the moment Grigsby was ordered to Dug Gap, it would have reached there before McPherson, and held it during the night of the 7th, during which time infantry support could have reached there”24. That this provision, when made, was at the lower, not the upper end of Snake Creek Gap, signally betrays the fatal neglect which had unaccountably persisted. It is, of course, idle to conjecture about what would have followed a forehanded occupation of the gap by the Confederates, but however forlorn the hope, a brave defense of this key to the fields would have made better reading in after years than a record of stupid neglect.

Grigsby began his night ride at 10:00 o’clock, taking the road just east of Rocky Face. The eastern outlet of the gap was reached at dawn. Grigsby had been informed that a company of Georgia troops was picketing the approaches to the gap, but the only pickets he encountered were those of the 9th Illinois (mounted) and the 39th Iowa, just arriving at the eastern outlet about the time Grigsby’s men got there. The 9th Illinois was soon joined by the 66th Illinois and perhaps by the rest of Burke’s 16th Corps brigade, and Grigsby’s horsemen were driven toward Resaca. Colonel Breckenridge salvaged, ironically, some credit from this affair by declaring that “prompt information of his (McPherson’s) movements (was sent) to our army headquarters.”

Indicative of confusion at the Tibbs house are three telegrams from Mackall to Cantey, on the 8th. Being in ignorance that forenoon of McPherson’s move toward Villanow and Snake Creek Gap, Mackall wired Cantey at Resaca to come up to Dalton with “one thousand men of your brigade” and leave word for Reynolds with his brigade to join him at Dalton later. Mackall followed this with a directive to Reynolds to remain and take command at Resaca, his troops to relieve Cantey’s. At 12:10 p.m. Cantey’s orders were countermanded but he was alerted for a later move and enjoined to watch toward Villanow and LaFayette. At 11:50 p.m. Cantey was instructed to hold Reynold’s brigade when it arrived25. These contradictory telegrams were the result of discovering that Dalton, beset on two fronts, was in need of reinforcements. Schofield, Howard, Palmer, and Hooker were swinging into position in Crow Valley; at the north end of Rocky Face; at Mill Creek and Dug gaps. The reference to Villanow in the dispatch of 12:10 p.m. was doubtless inspired by the Breckinridge dispatches of the forenoon which indicated the presence, in force, of Federals at and northward of Villanow; the 11:50 p.m. dispatch directing Reynolds to remain at Resaca was a direct result of Geary’s push at Dug Gap that afternoon, but none of these dispatches directed a specific coverage of the Snake Creek Gap approach to Resaca.

To say that McPherson’s advance columns, having issued from the eastern mouth of Snake Creek Gap on the 9th, brushed aside Grigsby’s horsemen, would be a mere figure of speech. Grigsby’s disposition of his men sufficed to delay McPherson several hours and it was not until late afternoon that he was able to deploy his force and assail Cantey in his trenches near Camp Creek on the Resaca front. Reynolds’s brigade had not yet arrived when Grigsby’s men joined Cantey in the works26. McPherson’s instructions called for the breaking of the railroad — not a general engagement with entrenched Confederates, so finding the forces in his front too strongly emplaced and his left flank imperiled by the forks of the Mountain Roads27, McPherson exercised the discretion given him and withdrew to the mouth of the Gap. Sherman’s comments on this performance indicated extreme annoyance; he pointed out to McPherson that such an opportunity came only once in a lifetime, overlooking the while both the paucity of the force he had sent on so critical a mission and the solid achievement of seizing the gap.

However, a detachment of eighteen men from the 9th Illinois (mounted) managed to reach the railroad two miles south of Tilton, but the Confederate cavalry prevented the destruction of the railroad, though the wires were cut and a wood station was burned28. This forehanded cavalry was the brigade of the ubiquitous Grigsby29. Learning that a Federal force was threatening his right, Grigsby galloped toward Tilton and reached the vicinity of Green’s Wood Station30 in time to prevent the destruction of the railroad by Captain Hughes of the 9th Illinois Infantry detachment.

But Grigsby found other matters to interest him near Tilton. His morning dispatch announcing the actual presence of McPherson’s troops in Snake Crack Gap had brought General Johnston southward, in person31. If Hardee thought his own presence was needed at Dug Gap, it is not surprising that Johnston went to Tilton or below, to look over the situation; any Federal force in Snake Creek Gap was certainly more alarming than the attempt of two brigades to scale the abrupt scarp of Dug Gap.

And southward, likewise, marched Hood with three divisions: Hindman’s of his own corps, Cleburne’s and Walker’s of Hardee’s32. But the night of May 9th saw McPherson back in Snake Creek Gap; the threat had abated for the moment and Hardee, up at Ault’s Mill, anxious because of the persisting threat to his right, was assured that his divisions would return — but they never did. On the 13th, the battle of Resaca began.

On May 11th, Mackall addressed a pertinent inquiry to Wheeler. He must have recollected the orders he had issued to him on the 5th and 7th. Here, and the 11th, reports were coming in that Crow Valley was being abandoned — that the Federals at upper Rocky Face and at Mill Creek Gap seemed to be vanishing, but there was Snake Creek Gap in full control of McPherson’s forces and who was to gainsay that most, if not all, of Sherman’s forces would soon be in that corridor to Resaca.

Had Wheeler replied to, or complied with, the directives of May 1st, 5th and 7th? As for replies or reports, if any, they are no part of the published record; indeed the absence of dispatches signed by Wheeler between May 1st and 13th, is conspicuous to anyone sufficiently curious to examine the record. Had there been replies and reports as numberless as the leaves at Vallombrosa — and all destroyed — there survives a sufficiency of evidence unmistakably showing that Johnston was badly served by his cavalry. As we have seen, three orders were required to get a patrol into the valley and, because of the delay, it failed to discover that two of the three Federal corps there were engaged, not in a remote flanking movement to the left rear, but in a direct thrust thither by way of the Snake Creek corridor.

Mackall’s inquiry betrays that, as of date, headquarters had no accurate knowledge of the whereabouts of most of the Federals. McPherson’s two corps had been in possession of Snake Creek Gap more than two days; Grigsby had held them at bay for over six hours and Cantey had fought then off on the Resaca front during the late afternoon of the 9th; the Dalton garrison had been robbed of three divisions to stave of the menace to the southward and Johnston, in person, had ridden down to and below Tilton where he could have been a spectator to Grigsby’s repulse of the Illinois lads who, though eighteen in number, had had the temerity to attempt the destruction of the railroad — yet Mackall gravely informed Wheeler that “a corps is supposed to be held in the mouth of Snake Creek Gap threatening Resaca.” Here is his inquiry:

Dalton, May 11, 1864 7:30 a.m.

General Wheeler, Cleveland and Dalton Road:

Did the system of scouts established by you just before the advance of the enemy include the valley between Taylor’s Ridge and Rocky Face? Is it still in operation? It is very important now that the force and movements of the enemy between those two ridges from Ringgold to Snake Creek Gap should be accurately known, and, as cavalry cannot be kept in observation in the valley now, General Johnston wishes you to try sending scouts in from your position to ascertain. Grigsby and Allen will receive orders to attempt the same from the south. A corps is supposed to be held in the mouth of Snake Creek Gap threatening Resaca.

Respectfully, your obedient servant,

W. W. Mackall

Chief of Staff.33

Just what reaction was there at headquarters when Grigsby’s dispatches announced that McPherson was at the eastern mouth of the gap? Dalton was now untenable; the piled-up effort to withstand the Federals at Rocky Face and Crow Valley, excellent while it lasted, was no longer useful, but nothing of record indicates surprise, mortification or chagrin. Even when writing of the event ten years later, Johnston expressed none of these emotions and there is no hint from him that anyone had been derelict in duty. One may speculate in vain on what Wheeler thought; his biographers34 confine their accounts as of May 9th, to his preoccupation with McCook’s cavalry near Varnell’s Station — where Colonel Oscar LaGrange was captured.

General Mackall began the above dispatch with a question that sounded as if he were leveling his sights upon a culprit guilty of gross neglect, but the rest of it belies such intention. One is confirmed in the belief that the directives of the 5th and 8th had no reference to Federal approach from the lower gaps of Taylor’s Ridge, but to mere flanking movements by troops that passed the ridge at Ringgold. Presumably, Mackall had nearly forgotten these orders and now with recollection of them returning, he thinks that a few scouts on the Federal flanks might clear up the mystery of the Vanishing Federals. Twenty minutes after he wrote this dispatch, he ordered Wheeler to go up to the northward of Rocky Face and learn if Sherman was shifting his forces toward the Oostenaula.

General Johnston’s comments avoid the obvious crux of the situation: the unguarded gap. He pontifically proclaims that he had full knowledge of all the strategic points of Dalton’s defense and to prove it, he calls attention to the “impractical nature of the country” between the mouth of the gap and the fortified hills of Resaca, threatened as it was by two approaches from the north (the forks of the Mountain Road) which permitted a move upon the Federal Left which, as we have seen, was a contributing factor to McPherson’s failure on the 9th. “This obstacle,” wrote Johnston, “to a rapid march by the United States Army was not unknown to the Confederates. We had examined the country very minutely and learned of its character thoroughly. We could calculate with sufficient accuracy, therefore, the time that would be required for the march of so great an army from Tunnel Hill to Resaca through the long defile of Snake Creek Clap, and by the single road beyond that pass. We knew also in how many hours our comparatively small force, moving without baggage trains and in three columns on roads made good by us, would reach the same point from Dalton.”35

If this sort of rationalization sounds naïve, it should be understood as deliberately so. Admitting previous knowledge of the gap and the full implications thereof, but meeting the situation only by measuring distances, counting roads and estimating time, was a calculated risk that might be assumed because of necessity, but General Johnston alleged no necessity for this course. It is hardly necessary to point out the curious use made, in his reasoning, of the estimated time required for the Federals to move from Tunnel Hill to Resaca, when three Federal corps were well south of Tunnel Hill — two of them actually in Snake Creek Gap — before he knew where they were. These alleged calculations were invoked after ten years, to bolster up a specious citation of strategy that has assured many of his readers that everything was under control, by directing their attention to the generalship displayed in ascertaining when to let go at Dalton, but no accounting was made of failure to know how the enemy got to the place that forced him to let go. This is like unto the good man of the house who, knowing that thieves are abroad, bars the front door and all the windows, but leaves the back door open, reassured by the knowledge that he is only a dozen paces from the open portal.

But General Johnston was not that naïve. The ten years that elapsed between the Dalton affair and the publication of his narrative, found many of those concerned still alive, and since a few of his enemies were writing books that elaborated on his shortcomings, he confined his remarks to defense and counter-attack only in domains traversed by his antagonists. As it happened, not one of them mentioned the affair. To have proclaimed a culpable neglect in securing Snake Creek Gap would have been a gratuitous contribution to the arsenal of his detractors, and furthermore, he was not given to pointing out shortcomings of his friends when, after all, his own inadvertent participation was inexorably bound up therein. In other words, Johnston was a victim of circumstances which he was supposed to control, but was, in the nature of events, unable to do so, because of an overweening reliance upon subordinates who were derelict in duty. Hence, the paragraph about timing the march from Dalton to Resaca is a mere poker-face assertion that everything was as it should have been, and posterity could take it or leave it.

If Johnston was reticent about the facts, why was everybody else? His detractors hesitated not to blazon his shortcomings, real or alleged, from Dalton to Atlanta — from President Davis on down the list: Bragg, Hood, Hardee, and a regiment of newspaper editors safely ensconced in bomb-proof sanctums. The answer is: few persons knew of what had happened. Bragg was not present; Hood and Hardee were preoccupied with the Rocky Face and Crow Valley defenses.

There were instances of contemporary criticism, however, voiced by a few who saw the situation first hand and close up, but their observations did not see print until years later. Colonel Breckenridge wrote his account twenty years afterward. He was still sensitive about the march of McPherson to Snake Creek Gap only a few miles behind that smoke-screen thrown up at Dug Gap by Geary’s rifled Rodmans. The Colonel avers that Johnston should have sent a cavalry force, even at the eleventh hour, to Snake Creek Gap at the time Grigsby was ordered to Dug Gap; this force could have held the pass, he asserts, until reinforcements made the position impregnable. He disclaims any offer of criticism and he leans backward in assuring his readers that he does not know who was responsible for the neglect, and he will venture no guess. So far as records go, this is perhaps the only statement by an officer of Wheeler’s cavalry, regarding the neglect in fortifying the gap.

Major General Patrick R. Cleburne, who knew something about holding mountain passes with a small division against superior forces, made his comment the following August 16th, some one hundred days later: “How this gap,” he wrote, “which opened upon our rear and line of communications, from which it was distant at Resaca only five miles, was neglected I cannot imagine. General Mackall, Johnston’s Chief-of-Staff, told me it was the result of a flagrant disobedience of orders, by whom he did not say. Certainly the commanding general never could have failed to appreciate its importance. Its loss exposed us at the outset of the campaign to a terrible danger and on the left forced us to retreat from a position where, if he adhered to his attack, we might have detained the enemy for months, destroying vast numbers of men, perhaps prolonged the campaign.”36

It must occur to the reader that the nearest Confederate force to the lower end of Snake Creek Gap on the 7th and 8th, was Brigadier-General Cantey’s command at Resaca.37 Johnston’s retention of Cantey’s brigade there on the 7th was in line with his belief in McPherson’s march toward Rome, and the placing of Williamson’s Arkansans at Dug Gap indicates that if Johnston knew of the Snake Creek Gap corridor, he deliberately ignored it as of no consequence in the situation as he then understood it. Cantey cannot be held responsible for any dereliction; he was a newcomer to the ground and doubtless never heard of Snake Creek Gap until the 9th, there being no orders directing his attention thereto.

Indeed, the conviction persists that it was not until after Grigsby’s dispatch from Snake Creek Gap reached headquarters, that Johnston included the gap in his reckonings, for during the forenoon of the 9th, and prior to the receipt of news of Federals in the gap, he sent a dispatch to Cantey, warning him that the Federals might cross the river and attack the railroad bridge from the south, and to meet this possible contingency a new line of earthworks was advised and a brigade was sent down from Dalton.38

Johnston’s allegations, in his narrative, about the gap, make curious reading in the light of this and other dispatches to Cantey. Despite his assertion of full knowledge of the gap, we find him taking defense measures at every point but the one that was his undoing. Could this have stemmed from total ignorance, or at best, a mere superficial knowledge; if the latter, was it obscured by a collective blind-spot when the defense situation was scanned at headquarters?

The Federal General Cox stated that the Confederate staff, due to the lack of adequate maps, had little or no accurate knowledge of the topography of the country, and in the opinion of Federal officials, this led to a belief by Johnston that his position could only be turned by a much longer detour and one involving many more contingencies for the Federals.39

The question remains, then: on what specific information, if any, was Grigsby sent to Snake Creek Gap? In other words, who became aware at the eleventh hour of its potential menace, not to say, existence? Johnston was probably indicting his May 9th dispatch to Cantey at the moment when Grigsby’s horsemen were contending with Dodge’s troops at the lower end of the gap. Breckenridge does not say who issued the order to Grigsby, but he does say that Grigsby was erroneously informed that there were no Federal troops there and curiously enough, despite the mandate to Wheeler to report on all cavalry engagements since May 6th, he is silent about Grigsby’s mission to Snake Greek Cap though he gives some account of the Dug Gap affair. Indeed his report is remarkable for the conspicuous absence of references to Snake Creek Gap; the nearest he came to one is the phrase: “All but two divisions of the enemy had turned our left flank, moving toward Resaca.”40

General Wheeler’s report is dated June 1, 1864, just twenty-four days after Snake Creek Gap was occupied by the Federals. He began his report by stating that for a few days prior to May 6, his three brigades picketed the front and flank of the Confederate army at Dalton “from Shipp’s Gap on our left to the Connecauga River on our right.” Follows then the statement that during this period the Federals drove in his pickets toward Tunnel Hill — all the commands being on or eastward of the road to Chattanooga. He made no mention of pickets driven in from the three gaps in Taylor’s Ridge where McPherson and Hooker crossed, despite his citation of Shipp’s Gap as the left of his picket line. With his mention of Grigsby at Dug Gap, he is silent on the significance of that little battle in relation to the Federal occupation of Snake Creek Gap and, as cited above, the transfer of Grigsby’s men to that gap and the all-day conflict with McPherson on the Resaca front is a conspicuous hiatus in his report. Practically, all of his reported operations were north and northeast of Dalton where his force Was preoccupied with the left of the Federal line; indeed, it would seem that but for Grigsby’s mission to Dug Gap, he was occupied nowhere else within the purview of Mackall’s order to him of May 1st and 5th.

We have no choice but to assume that no official inquiry was made — no charges of dereliction preferred against any one, and that by the consent of all involved, the matter was consigned to the Limbo where a myriad of other dubious affairs have been reposing since the beginning of time.

If General Mackall had been a little more specific in what he said to General Cleburne, at least one mystery would have been cleared up — one less of the group that continues to intrigue readers of the annals of the Army of Tennessee – – among them the perplexing complications of why, Johnston did not fight at Cassville, and the winter’s tale of that night march of Schofield’s army past the bivouac fires of the somnolent Confederates on the Columbia Pike at Spring Hill.



1 Johnston, Narrative of Military Operations, p. 277

2 O.R., XXXVII : 4, 673, Mackall to Williamson, May 7. Mackall’s order to Williamson directed him to proceed to Villanow Gap. That Dug Gap was meant is indicated by Johnston’s account of the defense of the place against General Geary’s assault on the and Williamson’s continued presence there on the 9th and 10th. Ibid., 683, 686. The order to Williamson, obviously, was inspired by the Federal approach from the North, and not by anticipation of a move from Taylor’s Ridge on the West.

3 Sherman., Personal Memoirs., II, 25.

4 O.R., XXXVIII : 3, 943, Wheeler’s report.

5 O.R., XXXVIII : 4, 656, Mackall to Wheeler, May 1.

6 Ibid., 660, Johnston to Wheeler, May 4.

7 Ibid., 661, Cooper to Polk; Johnston to Polk, May 4.

8 Ibid., 659, Johnston to Bragg, May 4.

9 Ibid., 660, Breckinridge to Grigsby, May 4, 4:00 P. m.; 665, Kelly to Buford, may 5.

10 Ibid., 664 Mackall to Wheeler, May 5. The reference is to Palmer’s 14th Corps in the vicinity of Tunnel Hill.

11 Ibid., 664, Mackall to Wheeler May 5, 12:00 noon.

12 Ibid., 668, Mackall to Martin, May 6, 12:00 noon.

13 Ibid., 675, Johnston to Polk, May 7, 2:45 p.m.

14 Ibid., 672, Mackall to Wheeler, May 7, 7:00 a.m.

15 O.R., XXXVIII : I, 188, Major General O.O. Howard’s report.

16 Battles and Leaders, IV, 277. Breckinridge’s narrative and the brief statement made by Wheeler concerning Grigsby’s brigade at Dug Gap are interesting examples of how word manipulation may mislead the casual reader who does not have the Mackall dispatches before him. Breckinridge states:

“On May 7th, our cavalry was driven through Mill Creek Gap. On that night, after we had gone into camp, Colonel Grigsby, who commanded the Kentucky cavalry brigade, was ordered to send a regiment to the front of Dug Gap to guard the approaches to it. In obedience to that order the 9th Kentucky Cavalry passed over Rocky Face and near midnight bivouacked on Mill Creek, about a mile from, and in front of, Dug Gap.”

General Wheeler states, as of May 7th –- O.R., XXXVII : 3, 944:

“At about 11:00 o’clock we were forced to abandon the town (Tunnel Hill), and by 3:00 o’clock were drawn back to our fortifications, where our infantry line of battle was formed. At dark I sent a regiment of Grigsby’s brigade to reinforce the picket at Dug Gap. May 8th, the remainder of Grigsby’s brigade was sent to Dug Gap, and with the rest of my command I moved to the Cleveland road . . . .”

As a factual sequence, the Breckinridge statement square with the events but if the elements of time and place are added, the affair assumes another color. Where did Grigsby go into camp? Mackall did not specify what cavalry of Wheeler’s should be sent to the valley west of Rocky Face, and he certainly did not order Wheeler to send any brigade to Dug Gap, but Wheeler stated that at dark he sent a regiment of Grigsby’s brigade to reinforce the picket there, and sent the rest of the brigade thither on the 8th. Was this pursuant to an order from headquarters that rescinded the one of May 7th, 7:00AM? If so, why did Williamson’s report of Dortch’s presence at Dug Gap inspire the peremptory order from Johnston to move a regiment into the valley at once? Clearly, Wheeler had not obeyed the order of 7:00 A.M., if he ever got it; patently, he had taken it upon himself to send a brigade to Dug Gap instead of the valley west of it. If the 9th Kentucky “passed over Rocky Face Ridge” after going into camp with the brigade, then Grigsby camped at the eastern foot of Rocky Face on the Dug Gap road. Wheeler’s statement that on the 8th, “the remainder of Grigsby’s brigade was Sent to Dug Gap” is but another way of saying that the remainder of the brigade ascended to Dug Gap from their camp east of it, and joined Williamson.

17 O.R., XXXVIII :4, 673, Mackall to Williamson, May 7. Johnston, Narrative of Military Operations, P. 305.

18 O.R, XXXVIII : 4, 673, Mackall to Wheeler, May 7.

19 Ibid., 664, Mackall to Wheeler, May 5

20 Ibid., 677, Breckinridge to Mackall, May 8, 10:15 a.m. The Presbyterian Church was Chattanooga Church, its site now occupied by Mt. Vernon Presbyterian Church, Whitfield County, on State Highway No. 2, two miles southwest of Rocky Face P.O., which is on U. S. Highway 41. The William H. Huff house is still standing; it is on the road between Gordon Springs and State Highway No. 2.

21 The Dr. Anderson house is still standing; it is immediately northeastward of Dunagan’s Church on the road to Mt. Vernon Church and Rocky Face P. 0. General Hooker had his headquarters at Dr. Anderson’s, May 7, 8, 1864,– O.R., XXXVIII : 4, 58, 77, 92, 93. Matt Deck resided in the Anderson house in 1950.

22 O.R., XXXVIII : 3, 720, Cleburne’s report.

23 Ibid., 721. Cleburne makes a mere mention of it.

24 Battles and Leaders, Vol. IV, p. 279

25 O.R., XXXVIII :4 679, Mackall to Cantey, May 8.

26 Ibid., 682, Mackall to Wheeler, May 9, 4:00 p.m.

27 Ibid., 106, McPherson to Sherman, May 9, 10:30 p.m.

28 O.R., XXXVIII : 3 376, Dodge’s report.

29 O.R., XXXVIII : 4, 683, Mackall to Cleburne, May 9.

30 John F. Green (1796-1873) was a prosperous planter prior to the war, whose plantation bordered the Connasauga River some two miles north of Resaca, most of the acreage being the scene of the fighting in the battle of Resaca, May 14th and 15th. (O.R., XXXVIII : 3, 817). He resided in a massive log house on a knoll overlooking the Connasauga – – a structure that still survives, in part, and one of the two houses of the battlefield area that dates from the 1860 decade. Adjacent to and north of the house, Colonel Green operated a water-tank and a wood station (one-half mile apart) on the Western and Atlantic Railroad, and east of the house, he maintained a river ferry. Green’s water-tank was supplied from a small river tributary that flows from Nance’s Spring near the war-time site of Union Church. Several hundred yards west of the tank stood the bullet-ridden Scales house (demolished in the 1940’s) which was the storm-center of the contest between Brigadier General Alpheus Williams’s 20th Corps division and Major General Alexander P. Stewart’s division of Hood’s Corps, May 15. (O.R., XXXVIII : 2, 28; 3, 817)

Green’s wood station, north of the tank, was one of the stops made by the Andrews Raiders to replenish fuel for the locomotive General which had been seized at Big Shanty in Cobb County, April 12, 1862 (Pittenger, The Great Locomotive Chase, 137). Since there was no other wood station between Resaca and Tilton, this was clearly the one referred to in General Dodge’s report of the burning of a wood station south of Tilton by the 9th Illinois Mounted Infantry. Colonel Green donated the land for the cemetery which members of his family organized after the war as a burial place for the Confederate dead of the battle of Resaca. (History of Gordon County, Georgia, Miss Lulie Pitts, p.302)

31 O.R. XXXVIII : 4. 663, Mackall to Cleburne at Dug Gap,, May 9.

32 Johnston, Narrative of Military Operations, P. 307.

33 O. R., XXXVIII : 4, 692, Mackall to Wheeler, May 11, 7:30 a.m.

34 General Joseph Wheeler and the Army of Tennessee, J. W. DuBose, N.Y., 1912. Fighting Joe Wheeler, John P. Dyer, L. S. U. Press, 1941.

35 Johnston, Narrative of Military Operations, p.315.

36 O.R., XXXVIII : 3, 721, Cleburne’s report.

37 At this time, Brigadier General James Cantey commanded a brigade which, together with Reynolds’s brigade, had been transferred from the Department of the Gulf to reinforce Johnston at Dalton where both brigades subsequently became a part of Polk’s command and, together with Preston’s artillery, formed a division commanded by Cantey. O.R., XXXII : 2, 582; 3, 861; XXXVIII : 3, 646. Toward the last of June, Major General E. C. Walthall replaced Cantey in command of the division. Ibid., 653.

38 O.R., XXXVIII : 4, 684, Mackall to Cantey, May 9.

39 Cox, Atlanta, P.31

40 O.R., XXXVIII : 3, 945, Wheeler’s report.

Source: Wilbur G. Kurtz papers, Atlanta Historical Society, Inc.

Mss 130 Box 36/Folder 11. All rights reserved by the Atlanta Historical Society, Inc.

Update to Reference:

Kenan Research Center

Atlanta History Center

ahc.MSS 130 Box 49/Folder14. All rights reserved by the Atlanta Historical Center.