Sbeitla, Gafsa, Fériana and Thelepte: The Long Retreat

In the four days after the destruction of Combat Command A at Sidi Bou Zid, the southern half of the American front in Tunisia was given up. Gafsa was evacuated in the night of 14–15 February 1943; Sbeitla was held under attack until the evening of 17 February and then abandoned; Fériana and the Thelepte airfields, the best forward air base the Allies had in Tunisia, were left the same day; and by the evening of 18 February the Americans had fallen back through Kasserine to the mountains of the Western Dorsal. Behind them, Rommel and von Arnim argued over what to do with their victory. This chapter follows the retreat place by place and asks what the record shows about the “panic” at Sbeitla, the burning of Thelepte and Rommel’s real intentions.

Why it mattered

The battle of Kasserine Pass is usually told as a three-day fight in a mountain gap on 19–22 February 1943. But the Americans did not arrive in the pass by choice. They were pushed there by four days of retreat across the high plain of west-central Tunisia, a retreat that began the night after the Germans struck at Sidi Bou Zid on 14 February, the story told in Chapter 3. In those four days the U.S. II Corps gave up an area the size of a large American county, with three towns, a railway, depots of food, fuel and ammunition, and the airfields at Thelepte from which its fighters had flown since early January. The fighting along the way was small by the standards of the war. The decisions were not.

On the Allied side, the retreat was the first time an American army in the Second World War gave up large amounts of ground to the Germans, and it was carried out under a tangled chain of command: an American corps commander, Maj. Gen. Lloyd R. Fredendall, taking orders from a British army commander, Lt. Gen. Kenneth Anderson, with Eisenhower, the Allied commander in chief, approving the major decisions from Algiers and Constantine. How well it was done is still argued over. The official U.S. Army historian, George F. Howe, called the withdrawal from Sbeitla “orderly and effective”; the night before it, rear-echelon units in the town had stampeded. Both are true, and the gap between them is part of the story.

On the Axis side, the same four days produced the quarrel that shaped everything that followed. Erwin Rommel, whose army had retreated nearly 1,500 miles from El Alamein, saw a chance to break into the Allied rear at Tébessa. Hans-Jürgen von Arnim, the commander of Fifth Panzer Army in northern Tunisia, did not. The order that settled the argument, issued by the Comando Supremo in Rome shortly before midnight on 18 February, sent Rommel north toward Le Kef by way of the Kasserine Pass and the Sbiba gap. Whether that was the right decision, and whether Rommel’s own plan would have worked, are among the most debated questions of the campaign.

A basin of Roman towns and forward airfields

Central Tunisia is shaped by two mountain chains. The Eastern Dorsal (Dorsale Orientale) runs south from near Pont-du-Fahs past Fondouk and Faïd toward Maknassy; the Western, or Grand, Dorsal (Grande Dorsale) runs southwest from the Tunisian Ridge past Sbiba and Kasserine toward the Algerian border. Between them lies a high, dry basin, mostly open plain and olive groves, crossed by a few roads and a narrow-gauge railway. In February 1943 the Americans held the passes of the Eastern Dorsal at its southern end and lived in the basin behind them; the main U.S. supply base and II Corps headquarters lay to the west, in Algeria, around Tébessa.

The places in this chapter are old. Gafsa, an oasis town in the south with springs, palm groves and a phosphate railway, is the ancient Capsa, sacked by the Roman general Marius during the war against Jugurtha in 107 BC. Fériana (modern Feriana), about 40 miles to the north on the road to Kasserine, sits beside the ruins of Roman and Byzantine Thelepte, and the American airfields of 1943 took the old name. Kasserine (Al Qasrayn), a small town on the railway at the eastern mouth of the pass that now bears its name, stands on the site of Roman Cillium. Sbeitla (Sbeïtla), about 17 miles northeast of Kasserine in a straight line, grew up next to the ruins of Sufetula, whose three temples on its forum are among the best-preserved Roman buildings in North Africa; it was near Sufetula, in 647, that an Arab army defeated and killed the Byzantine exarch Gregory, opening North Africa to the Arab conquest. In 1943 Sbeitla was a road junction of some importance, where the road from Faïd met the roads north to Sbiba and west to Kasserine (Howe).

Thelepte mattered most to the airmen. The airfields lay on hard, dry ground a few miles from Fériana. Craven and Cate’s official air force history describes it as “a good dry-weather field” in semidesert country, a rare thing in a Tunisian winter when the fields farther north turned to mud (Craven and Cate, vol. 2). From early January 1943 the U.S. XII Air Support Command based fighters there: the 33rd Fighter Group with P-40 Warhawks from 7 January, the 81st Fighter Group with P-39 Airacobras from late January, and the 31st Fighter Group with Spitfires from 7 February, with A-20 light bombers of the 47th Bombardment Group nearby (Maurer, Air Force Combat Units; Craven and Cate). The fields had no radar warning. Craven and Cate record that the French gendarmes who telephoned in sightings reported every aircraft as hostile.

Behind the basin, the Western Dorsal was the obvious place to make a stand. Anderson had wanted, since January, to hold the Eastern Dorsal lightly with mobile forces and to keep his main strength on the more defensible Western Dorsal; the Americans’ scattered dispositions on the eastern ridge in mid-February were the product of compromises between that idea, French political needs and American wishes, described in Chapter 2. Through the Western Dorsal ran the roads that mattered: from Kasserine through the pass to Thala and Tébessa, from Sbeitla north through Sbiba to Le Kef, and from Fériana northwest to Tébessa past the passes at Bou Chebka and El Ma el Abiod.

Giving up Gafsa, 14–15 February

Gafsa was the southern anchor of the American front and its most exposed point. It had changed hands before: a small Axis detachment had held it briefly in November 1942, and American paratroopers under Col. Edson Raff had then moved in. In February 1943 the town was the seat of the Advance Headquarters of II Corps, in the Hotel de France, under Col. Frederic B. Butler, who directed a mixed American and French command reaching southeast to the village of El Guettar together with Colonel Morlière of the French Constantine Division. Two squadrons of the British 1st Derbyshire Yeomanry, an armored car regiment, patrolled to the east and south (Howe, ch. XXI).

The Allied commanders had long doubted that the town could be held. Howe writes that the high command decided, “after some irresolution,” that Gafsa could not be reinforced enough to withstand any probable enemy attack, and that the plan in case of a strong attack was for the garrison to fall back on Fériana, where a counterattack could be mounted. On 13 February Eisenhower met Anderson at Fredendall’s headquarters near Tébessa, then drove forward through the night to Sbeitla and Sidi Bou Zid, returning to Tébessa shortly before daylight on 14 February. He left later that morning for Constantine, stopping on the way to look at the Roman ruins at Timgad, and the news of the German attack at Sidi Bou Zid caught up with him on the road. Over the next two days he took part in the decisions to concentrate the Allied strength in central Tunisia and to give up Gafsa (Howe, ch. XXI).

The order to evacuate came from First Army on 14 February, as reports from Sidi Bou Zid made the scale of the German attack clear and Allied intelligence still expected a second blow from the south. First Army’s plan was to move the French out on the first night and the Americans on the second. II Corps objected that the secret could not be kept for two nights and that the enemy would interfere with the second move, and in the event the whole garrison went out in the night of 14–15 February (Howe, ch. XXI). The Americans and French drove north up the road to Fériana. The 51st Medical Battalion and the 48th Surgical Hospital moved farther back. The Derbyshire Yeomanry squadrons screened the eastern flank of the column, and the 1st Ranger Battalion, which had raided the Italian outpost at Sened Station only three nights before, was the last unit to leave the town (Howe; Darby and Baumer, Darby’s Rangers).

Howe calls it “a night of rather confused and excited activity,” especially among the civilians, who remembered the Axis occupation of November and did not want another. The plan called for removing all supplies and transport and booby-trapping and mining the town. Not everything went as planned. The railway bridge north of Gafsa was blown too early, before all the rolling stock had been moved; the trains were run west instead to Metlaoui, on the phosphate line, and hidden in a tunnel. In the old kasbah, the walled citadel in the center of the town, six tons of French ammunition were blown up. The blast damaged nearby buildings and injured some of the people living in them (Howe, ch. XXI, citing the II Corps records). The American air force history adds that the withdrawal from Gafsa was the signal for the air command’s evacuation plan, Plan A, which had been issued on 10 February; the advanced airfield at Sbeitla was cleared the same night (Craven and Cate).

The force that came out of Gafsa was a scratch command of the kind II Corps had been forming all winter. Howe lists it as the 3rd Battalion of the 26th Infantry (1st Infantry Division), the 1st Ranger Battalion, the 175th Field Artillery Battalion, squadrons of the Derbyshire Yeomanry and French units of the Constantine Division; from 8 February the 1st Battalion of the 168th Infantry had been held at Fériana as corps reserve. It was commanded by Col. Alexander N. Stark Jr., commander of the 26th Infantry. Two days later the same colonel would be sent to command the defense of the Kasserine Pass, the story of Chapter 5.

Morgenluft: Rommel’s battle group comes north

The German operation against Gafsa had been planned for nearly two weeks. On 4 February Rommel had revived, in a memorandum to the Comando Supremo, an idea he had first raised in November: while part of his German-Italian Panzer Army held the old French fortifications of the Mareth Line against the British Eighth Army, its mobile units would strike north at Gafsa from the southeast, and von Arnim’s mobile units would attack from the northeast. Rommel set three conditions: the attack had to come within days, it had to be made by concentrated and superior forces, and it needed one commander across the boundary between the two Axis armies. The Comando Supremo first ordered an attack on the Gafsa area to destroy the Allied forces there, with Rommel in command of the mobile elements of both armies (Howe, ch. XXI).

That arrangement did not survive. At a meeting with Kesselring and von Arnim on 9 February, reconnaissance reports that American units were leaving Gafsa led to a new plan. Fifth Panzer Army would attack first, at Sidi Bou Zid, in an operation named Frühlingswind (Spring Wind). Later, Rommel would strike Gafsa with an assault group drawn from the Afrika Korps, joined by elements of the 21st Panzer Division. Rommel’s part was named Unternehmen Morgenluft (Operation Morning Air). At a further meeting on 13 February at La Fauconnerie, Rommel was pessimistic about what his group would meet at Gafsa; von Arnim was sure it would be lightly held; and Generalmajor Heinz Ziegler, who commanded Frühlingswind in the field, promised that the 21st Panzer Division would be sent south to reinforce the Gafsa attack at the first opportunity (Howe, ch. XXI). At 15:30 on 14 February, with the Sidi Bou Zid battle going well, Ziegler ordered the 21st Panzer to assemble for the move on Gafsa, “probably starting at noon next day.” The division never went. Von Arnim kept it for the drive on Sbeitla.

Rommel’s assault group, the Kampfgruppe DAK, was a German-Italian force of about division strength under Oberst Kurt Freiherr von Liebenstein, the former commander of the 164th Light Africa Division. Howe lists its German core as the headquarters of the Afrika Korps, Panzer Grenadier Regiment Afrika, two weak battalions of the 1st Luftwaffe Jäger Brigade, one tank battalion of the 15th Panzer Division with 26 tanks, the 33rd Reconnaissance Battalion, two artillery battalions, a rocket-projector battery, an engineer company, an antitank company and the headquarters of the 135th Flak Regiment with five heavy and two light antiaircraft batteries. The Italian Centauro Armored Division added two infantry battalions, two artillery battalions, a motorized Bersaglieri battalion, a tank battalion with 23 tanks and a battery of assault guns (Howe, ch. XXII, n. 6). In all, about 49 tanks, a small fraction of the armor von Arnim had committed at Sidi Bou Zid.

The group moved north from the Gabès area cautiously. Rommel was worried about his lack of reserves, feared a reverse, and overestimated the Allied strength around Gafsa (Howe, ch. XXII). Axis troops found the town empty on 15 February and occupied it that evening; the Allied air history speaks of “a small enemy column” entering in the night (Craven and Cate). Reconnaissance went southwest toward Metlaoui and the oasis of Tozeur and northwest up the road toward Fériana. Kampfgruppe DAK had taken its objective without a fight, a day after the Americans had left. Rommel’s original aim, the destruction of the Allied forces in the Gafsa basin, had already slipped away.

The line swings back, 15–16 February

While Gafsa was being emptied, the 1st Armored Division was losing its second battle at Sidi Bou Zid. On 15 February Combat Command C counterattacked across the plain to relieve the infantry cut off on the djebels Lessouda and Ksaira and was itself shattered, as told in Chapter 3. That evening Anderson accepted what the losses meant. In a message to Eisenhower on 15 February he described a “very exposed southern flank” and proposed a voluntary withdrawal to the Western Dorsal, where he would fight hard; because his reserves were now committed to covering the move, he could not make the diversionary attack in the north that Eisenhower had suggested. Eisenhower approved, and at 17:00 First Army warned II Corps that an early withdrawal had been decided (Howe, ch. XXII, citing Msg O-420, Anderson to Eisenhower, 15 February 1943).

The same day Anderson released Combat Command B of the 1st Armored Division to II Corps. CCB, under Brig. Gen. Paul McDonald Robinett, had been held in the north, near Maktar, as a reserve for the French XIX Corps sector, and its release was the first time since January that the 1st Armored Division had most of its strength under its own commander. It moved south by two routes in the night of 15–16 February and passed to division control at Sbeitla (Howe, ch. XXII). The division’s commander, Maj. Gen. Orlando Ward, now had three combat commands around the town: the remnant of Combat Command A, which had lost most of its tanks and two infantry battalions on 14 February; the remnant of Combat Command C; and the fresh CCB. After midnight on 14–15 February the corps boundary had also been shifted south, so that Thala, Sbiba and the Fondouk gap passed to General Louis-Marie Koeltz’s French XIX Corps (Howe, ch. XXI).

On 16 February Anderson forbade further counterattacks. II Corps was to concentrate on holding Sbeitla, Kasserine and Fériana while the withdrawal was prepared (Howe, ch. XXII). The Germans gave the Americans a little time. Howe headed a section of his chapter “The Axis Forces Squander a Day”: after the victory of 15 February, von Arnim’s divisions spent most of the 16th collecting their forces, and only in the afternoon did the 21st Panzer Division move west toward Sbeitla. Von Arnim’s own plan was still limited. He ordered Ziegler to strike at Sbeitla and then turn north toward Fondouk; for 18 February he intended to use the 21st and 10th Panzer Divisions around Sbeitla and Fondouk-Pichon and to advance only to the crests of the Eastern Dorsal (Howe, ch. XXII). At 22:50 on 16 February the Comando Supremo directed him to exploit his success with the strongest forces available. Kesselring, who was visiting Hitler’s headquarters in East Prussia, ordered the attack on Sbeitla to go ahead.

Robinett reached the field in the afternoon of 16 February and placed CCB south of the Sbeitla-Faïd road, southeast of the town, in time to reconnoiter and organize what Howe calls an active defense. Robinett had recommended from the start a pullback to the Western Dorsal and a defense on ground of the Americans’ choosing, and he had a reputation, as one recent account puts it, for annoying his superiors with his outspoken views (Knighton, drawing on Orr Kelly, Meeting the Fox). Now he had the chance to fight the kind of battle he had argued for.

The night of 16–17 February at Sbeitla

The Germans came up to Sbeitla at dusk on 16 February. A battle group under Gerhardt probed the forward American positions; another, under Pfeiffer, from the 21st Panzer Division, was to follow Lt. Col. Crosby’s withdrawing 3rd Battalion, 13th Armored Regiment, closely and attack the bridge about five miles east-southeast of the town (Howe, ch. XXII, citing the German combat report on Faïd). The history of the 443rd Antiaircraft Artillery Battalion records that Tiger tanks reached the road junction east of the town that the Americans called Kern’s Crossroads by evening, that the Germans attacked Sbeitla shortly after dark, and that they drew back after three of their tanks were destroyed (443rd AAA Battalion history, Texas Military Forces Museum).

What happened next was the worst moment of the retreat. Fire from Pfeiffer’s detachment fell into the olive groves near the command post of Combat Command A, among vehicles refueling and supply dumps. An improvised mine-laying unit was scattered. Vehicles began to move west, others followed, and by Howe’s account a “dense mass of churning traffic” poured through Sbeitla in the dark. The troops were exhausted after two days of defeat, and they had the sense, as Howe puts it, that those in command were “playing by ear” (Howe, ch. XXII, citing the history of the 3rd Battalion, 13th Armored Regiment). Trucks, half-tracks, ambulances and supply vehicles jammed the road toward Kasserine. Some men simply left. The combat units in the line east and south of the town held their positions, but to those who saw the roads that night it looked like a rout.

At headquarters the picture was no better. Around midnight Ward concluded that Combat Command A was facing a very large armored force and reported to Fredendall that the situation was “extremely grave,” estimating about nine Mark VI (Tiger) and eighty Mark IV tanks (Howe, ch. XXII). It was an estimate made at night from fragmentary reports. At about 01:30 on 17 February Anderson, through his liaison officer with II Corps, Brigadier C. G. McNabb, authorized Fredendall to withdraw Ward’s command from Sbeitla and Stark’s force from Fériana. Then came a confusion of orders. A division order issued at about 03:00 said that Sbeitla “must be held at all costs until 1100.” About an hour later it was replaced by an order to hold a line south and east of the town until ordered to withdraw. The change reflected a disagreement between Fredendall, who wanted to be out of Sbeitla by late morning, and Anderson, who wanted it held all day (Howe, ch. XXII, n. 27).

During the night American troops destroyed the ammunition dump and the railway bridge at Sbeitla (443rd AAA Battalion history). When the Germans did not attack at dawn, the defenders had their reprieve. Their advanced elements watched the American truck columns leave the town while the main German force waited for units coming up from Sidi Bou Zid, to make a combined attack at noon (Howe, ch. XXII).

Robinett’s stand, 17 February

Robinett had spent the 16th arranging his defense on the ground in front of the town. His medium tanks of the 2nd Battalion, 13th Armored Regiment, under Lt. Col. Henry E. Gardiner, were in defilade in folds and wadis about four miles southeast of Sbeitla, with the self-propelled howitzers of the 27th Armored Field Artillery Battalion behind them. The 601st Tank Destroyer Battalion held the outpost line in front. Two companies of the 2nd Battalion, 6th Armored Infantry, under Lt. Col. Elton W. Ringsak, held the hills of Djebel el Koumine and Djebel Sabel Dilou on the flank (Howe, ch. XXII, citing the CCB after-action report and the 601st’s report).

Shortly after noon on 17 February the 21st Panzer Division attacked with an armored group under Stenkhoff. Its tanks first struck CCB’s northern flank, which the 601st Tank Destroyer Battalion was screening. At about 13:15 the Germans made a frontal attack straight at Gardiner’s battalion. The hidden tanks held their fire and then fired together, and the volley knocked out five or more German tanks. Under fire from the 27th Armored Field Artillery, Stenkhoff’s group pulled back (Howe, ch. XXII, citing the German combat report and the 21st Panzer Division war diary). Another account, drawing on Orr Kelly, describes a second German attack at about 14:15 and Robinett’s receipt of the order to withdraw soon after (Knighton).

The rest of the division was already going. Combat Command A began to withdraw at 11:30, moving north under dive-bomber attack through Sbiba and then swinging west in a wide loop to Tébessa; British engineers closed a gap in the Sbiba minefield behind it. Combat Command C made its first bound west toward Kasserine at 14:30 (Howe, ch. XXII). CCB began to shift west at about 14:30, in stages, each element covering the next. Gardiner’s battalion was the last to disengage, between 17:30 and about 19:00, when it was dark. It lost nine tanks, Gardiner’s own among them (Howe, ch. XXII, citing the CCB report, Robinett’s memoir and Gardiner’s own account). It refueled at a division gasoline dump beside the highway and went on through the Kasserine Pass to a bivouac between the pass and Thala. The 601st Tank Destroyer Battalion, cut off on the southern flank, was scattered during the night and came back in pieces.

German troops entered Sbeitla after 17:00 (Howe, ch. XXII). They found that most of the supplies had been moved or destroyed, and that the Americans had blown bridges, the water mains and the aqueduct that carried water from the springs near Sbeitla to the port of Sfax. At midnight the remaining units of Combat Commands A and C, with rear elements of CCB, began the last march to Kasserine. The 21st Panzer Division reported that it had wrecked the 1st Armored Division and the 168th Infantry, and claimed 2,546 prisoners and 103 tanks destroyed over the four days of fighting; the Germans also counted 27 American tanks and tank destroyers disabled or destroyed on the field southeast of Sbeitla (Howe, ch. XXII, nn. 29, 31). The claim about the 1st Armored was overstated. Howe’s judgment is that Combat Command B came out of Sbeitla in shape to fight another battle, and it did, at Djebel el Hamra four days later.

Fériana and Thelepte, 17 February

Nearly 40 miles to the southwest, the same night’s decisions emptied Fériana and Thelepte. After Anderson’s authorization at 01:30 on 17 February, Fredendall ordered Stark to leave a small covering detachment of infantry, light tanks and 155 mm guns at Fériana until 18:00 and to place the rest of his command, with elements of the French Constantine Division, on the heights north of Thelepte. He warned Col. Paul L. Williams of the XII Air Support Command at Thelepte that his aircraft had to be flown off at daybreak (Howe, ch. XXII, citing the II Corps G-3 journal).

The air command had been preparing for this since 10 February, when its Plan A was issued. Thelepte had been partly cleared on 16 February; the A-20 light bombers of the 47th Bombardment Group had already gone to Youks-les-Bains, the steel-matted field near Tébessa, by midnight on the 16th. On the 17th the two fighter groups still at Thelepte, the 31st with Spitfires and the 81st with P-39s, flew out with what Craven and Cate count as “altogether 124 operational aircraft.” The 31st went to Tébessa and the 81st to Le Kouif, an airfield north of Tébessa; the 52nd Fighter Group, with more Spitfires, was already at Youks. The last mission took off from Thelepte at 10:30, and a security detachment inspected the fields between 11:00 and 12:00 (Craven and Cate, vol. 2, citing the XII Air Support Command report and the group histories).

What could not be moved was destroyed. In Craven and Cate’s account, “60,000 gallons of aviation gas were poured out,” the rations were blown up, and “eighteen aircraft, of which five were non-reparable,” were burned; about 3,000 troops and most of the units’ equipment were got out. “Nothing was left for the enemy,” they concluded. Howe gives a different count. Citing a report by Capt. John B. Upton preserved in the history of the 19th Engineer Combat Regiment, he writes that 34 planes that could not be flown away were demolished. He also cites a German radio message of 19 February reporting that the Afrika Korps had salvaged 20 tons of aviation gasoline and 30 tons of lubricants at the fuel depot, along with French ammunition and fuzes (Howe, ch. XXII, n. 35). Howe’s verdict is that “Allied demolitions left the Thélepte airbase with little of military value.” Both statements can be true at once: the airfield was useless to the Germans as an air base, but not every drum of fuel was destroyed.

Kampfgruppe DAK came up the road from Gafsa the same morning. Rommel had ordered Liebenstein to take Fériana by surprise without getting tied down in a costly fight. The German records show that the Allied rearguard pulled out of Fériana at about noon after a short engagement, earlier than the 18:00 Fredendall had set, and the Afrika Korps group then pushed on to Thelepte (Howe, ch. XXII). Somewhere on that road Liebenstein’s vehicle hit a mine. He was wounded, and Gen. Karl Bülowius, Rommel’s chief engineer, took over the group. By the evening of 17 February the airfields from which the Americans had flown for six weeks were in German hands.

By the numbers

Thelepte, 17 February 1943: aircraft flown out and destroyed

Aircraft of the U.S. XII Air Support Command at the Thelepte fields on the day they were abandoned, as counted by the two U.S. official histories. Bars are scaled to the largest value.

Operational aircraft flown out (Craven and Cate)124

Aircraft that could not fly, demolished (Howe, citing Upton)34

Aircraft burned (Craven and Cate)18

… of which non-repairable (Craven and Cate)5

Kasserine village and the new line, 17–18 February

The roads from Sbeitla and Fériana met at Kasserine, a small town on the railway at the eastern mouth of the pass, and every unit coming out of the basin from the east and south had to pass through it or near it. Fredendall gave the job of keeping the door open to Col. Anderson T. W. Moore, whose force, the 19th Engineer Combat Regiment and the 1st Battalion of the 26th Infantry, was to hold a line east of Kasserine village until Ward’s troops had gone through and then to organize the Kasserine Pass itself for defense (Howe, ch. XXII). In the night of 17–18 February the columns of the 1st Armored Division came down the road from Sbeitla in the dark, past Moore’s positions and into the pass. Combat Command A, which had gone north through Sbiba, reached Tébessa by its long loop to the west.

By 18:00 on 18 February Combat Command B had passed through the Kasserine Pass and Tébessa to the wooded mountains near El Ma el Abiod, southeast of Tébessa, where it could block the roads from Fériana and Thelepte; the rest of the 1st Armored Division took positions southeast of Tébessa to protect it from the east and south (Howe, ch. XXII). The Kasserine village defenses had been thinned out as the columns passed, and on 18 February German reconnaissance from Kampfgruppe DAK entered the undefended village. They captured about sixty French soldiers who were arriving from the east and met there a reconnaissance unit of the 21st Panzer Division coming from Sbeitla: the two Axis offensives had joined hands. Another German detachment drove Allied outposts back into the passes near El Ma el Abiod (Howe, ch. XXII).

To the north, Anderson moved fast to close the road from Sbeitla to Le Kef. He sent to the Sbiba gap first the British 1st Guards Brigade, with a combat team of the U.S. 1st Infantry Division attached, then the U.S. 34th Infantry Division, and then the British 26th Armoured Brigade under Brigadier Charles Dunphie. The headquarters of the British 6th Armoured Division under Maj. Gen. Charles Keightley went to the French XIX Corps to command the defense of the Sbiba gap, and the corps boundary was moved to run from four miles south of Sbiba to Hadjeb el Aïoun. Three battalions of the 34th Division, with French troops attached, took positions northeast of Sbiba (Howe, ch. XXII). Thala, behind the Kasserine Pass, would become the British 6th Armoured Division’s concern as well, the story of Chapter 6.

The new Allied line, then, ran along the Western Dorsal: the Sbiba gap in the north; the Kasserine Pass in the center, held by Moore’s engineers and infantry and soon by Stark; and the passes south and east of Tébessa at Dernaia, Bou Chebka and El Ma el Abiod, held by Stark’s former Gafsa force, the French and CCB. II Corps was back roughly where Anderson had wanted it a month before, but it had arrived there tired, disorganized and short of equipment. The Signal Corps photographers at Tébessa on 18 February recorded French troops loading their equipment onto railway cars and Moroccan cavalry in the streets; the 443rd Antiaircraft Battalion’s history recorded that the reorganization of its batteries had to make good the loss of 13 gun carriers and trailers between Faïd and Kasserine, and that morale had dropped among the veteran units, which had hoped for rest (443rd AAA Battalion history).

Rommel against von Arnim, 17–19 February

The Axis victory had been won by two armies with two commanders who disliked and distrusted each other. Rommel, a field marshal since June 1942 and a national hero in Germany, commanded the German-Italian Panzer Army, which was falling back into southern Tunisia ahead of Montgomery, and he had been told he would soon be relieved for reasons of health. Von Arnim, a Generaloberst who had come from the Eastern Front in December 1942, commanded Fifth Panzer Army in the north. Neither was subordinate to the other. Above them stood Kesselring, the German commander in the Mediterranean, at Frascati near Rome, and the Italian Comando Supremo under Gen. Vittorio Ambrosio, which held formal command of the Axis forces in North Africa.

On 17 February, as his group took Fériana and Thelepte, Rommel floated the idea of a thrust on Tébessa, but judged his own forces too weak to attempt it unless they were reinforced by von Arnim’s mobile divisions and supported by a holding attack elsewhere. Von Arnim’s reply set the idea aside. By the next day the air reconnaissance showed Allied columns streaming west from Kasserine and from Bou Chebka toward Tébessa, which suggested that the passes of the Western Dorsal were lightly held. After an exchange of messages and a telephone call with von Arnim, who rejected his proposals outright, Rommel at 14:20 on 18 February went over his head. He sent a message to the Comando Supremo and Kesselring proposing “an immediate enveloping thrust” on Tébessa and the country north of it, provided Fifth Panzer Army’s supplies were adequate, and asking for the 10th and 21st Panzer Divisions to be put under his command and assembled at Thelepte and Fériana (Howe, ch. XXIII).

Kesselring, just back at Frascati from Hitler’s headquarters, backed him. He radioed both army commanders that it was “essential to continue the attack toward Tébessa” and to the north, and said he would discuss the question that day with Mussolini and Ambrosio. Late on the 18th Rommel asked the Comando Supremo to rush the 21st Panzer to Thelepte and the 10th Panzer to Kasserine so that the offensive could start the next evening; his objective was still Tébessa (Howe, ch. XXIII).

The answer reached Rommel at his advanced headquarters shortly before midnight on 18 February. The Comando Supremo spoke of “a unique opportunity” to force a decisive success in Tunisia and ordered a deep thrust north to threaten the rear of the British V Corps and, if possible, to cut it off. Rommel was given the 10th and 21st Panzer Divisions and told to attack toward Maktar and Tadjerouine, with Le Kef as his first objective, concentrating his forces on a line from Sbeitla to Tébessa. Fifth Panzer Army was to prepare a holding attack in the north (Howe, ch. XXIII). The order did not name the routes, but the roads to Le Kef ran through the Western Dorsal at the Sbiba gap and, by way of Thala, through the Kasserine Pass. Rommel decided to try both and to put his main effort into whichever proved more weakly held: the 21st Panzer Division would attack toward Sbiba, the Afrika Korps group toward the Kasserine Pass, and the 10th Panzer Division would gather at Sbeitla to exploit whichever succeeded.

Rommel was bitterly disappointed. In Howe’s summary, he regarded the order as a rejection of his wide envelopment through Tébessa and thought it appallingly shortsighted, because it aimed the main attack straight into the Allied reserves gathering around Thala and Le Kef rather than around their flank; his own account in The Rommel Papers makes the same complaint. On 19 February Kesselring flew to Tunisia. He found that von Arnim had read the order as permitting a thrust toward Tébessa; von Arnim himself now proposed a concentric attack on Le Kef and down the Medjerda valley toward Béja, with the 21st Panzer attacking from Sbeitla and the 10th from Pichon. Kesselring rejected that, and told von Arnim that his own intention had included both Tébessa and Le Kef, a point the order had failed to make clear to Rommel (Howe, ch. XXIII). Von Arnim also kept back part of the 10th Panzer Division, including its heavy tank battalion with the Tigers, in his own sector. In the night of 19–20 February the Comando Supremo placed the Italian First Army, the 10th and 21st Panzer Divisions and the Afrika Korps under “Group Rommel,” effective at 06:00 on 20 February. By then the attacks on the Kasserine Pass and at Sbiba had already begun.

The GI in retreat

For the American soldiers in the basin, the four days after Sidi Bou Zid were a blur of night marches, air attacks, rumors and fatigue. Most of them had been in action for the first time only days or weeks before. The 1st Armored Division had lost two of its battalions of tanks and much of its infantry; men from the 168th Infantry who had escaped from the djebels came in by twos and threes, and units like the 443rd Antiaircraft Battalion were still recording stragglers, men listed as missing, reaching their lines at Sbeitla on 16 February (443rd AAA Battalion history). Few of them knew why they were going back or where the retreat would stop. Howe’s phrase for the feeling in the ranks at Sbeitla, the sense that those in command were “playing by ear,” came from the history of one of the tank battalions.

The roads carried more than soldiers. Gafsa’s civilians had spent the night of 14–15 February in what Howe calls confused and excited activity, and the 443rd’s history records that one of its officers, searching toward Fériana on 17 February for a missing convoy, found only refugees on the roads toward Fériana and Gafsa, both already in German hands. That night, at 22:00, Ward ordered the battalion to unload its trucks and use them to carry ammunition and gasoline out of Kasserine before midnight; whatever could not be moved was destroyed (443rd AAA Battalion history). At Tébessa on 18 February the Signal Corps photographed Moroccan cavalry in the streets and French troops loading their equipment onto trains.

Americans at home followed the retreat in the papers. Wartime censorship and the distance from the front softened the news, and the first dispatches described a planned withdrawal. The Associated Press correspondent Hal Boyle called the move out of Sbeitla and Fériana a “strategic evacuation,” with the troops destroying what they left behind; his dispatch was carried by American newspapers including The Charlotte News in North Carolina on 19 February 1943 (as summarized in the Charlotte News archive at wjcash.org). The phrase was not false. It was not the whole truth either.

The most famous American reporter in North Africa, Ernie Pyle of the Scripps-Howard newspapers, was with the troops in southern Tunisia and wrote about the withdrawal from the Sbeitla valley. His dispatch, datelined “The Tunisian front” and marked as delayed, was written before the Allies retook the lost ground and appeared in American newspapers in early March 1943; Pyle later collected his North African columns in Here Is Your War (1943). He described a withdrawal that started before dawn one morning and went on without a break for twenty-four hours, and that looked little like the panicky retreat his readers might have imagined. Pyle wrote under censorship and for a home-front audience, and he described what he saw; the stampede of the night of 16–17 February is not in it.

In their words

The withdrawal of our American forces from the vast Sbeitla Valley, back through Kasserine Pass, was a majestic thing in a way.

Rick Atkinson’s narrative in An Army at Dawn (2002), built on soldiers’ letters, diaries and interviews as well as the official records, shows both sides of the retreat: the confusion of men who did not know where it would end, and the steadiness of the units that did their jobs. Gardiner’s own diaries, published only in 2021, add the view of a battalion commander who had been a Chicago lawyer three years before. For many of the men who came through the Kasserine Pass on the night of 17–18 February, the hardest thing was not the danger but the sense of a defeat they did not understand. The next four days would test whether they could recover from it.

Timeline

The long retreat, 4–20 February 1943

1. 4 Feb 1943Rommel proposes to the Comando Supremo a strike on Gafsa by the mobile forces of both Axis armies.
2. 9 FebKesselring, Rommel and von Arnim revise the plan: Frühlingswind (von Arnim, Sidi Bou Zid) first, then Morgenluft (Rommel, Gafsa).
3. 10 FebThe air support command issues Plan A for the evacuation of its forward airfields.
4. 14 FebFrühlingswind destroys most of Combat Command A at Sidi Bou Zid; First Army orders Gafsa evacuated.
5. Night 14–15 FebThe Gafsa garrison withdraws to Fériana, the 1st Ranger Battalion last out; the Sbeitla airfield is cleared.
6. 15 FebCombat Command C’s counterattack fails; Anderson proposes withdrawal to the Western Dorsal; at 17:00 First Army warns II Corps; CCB is released; in the evening Kampfgruppe DAK occupies Gafsa.
7. 16 FebAnderson bans counterattacks; CCB takes position southeast of Sbeitla; at 22:50 the Comando Supremo orders von Arnim to exploit; the Germans reach Sbeitla at dusk.
8. Night 16–17 FebGerman fire causes a stampede of vehicles through Sbeitla; at about 01:30 Anderson authorizes withdrawal from Sbeitla and Fériana.
9. 17 Feb, 10:30Last air mission from Thelepte; 124 aircraft flown out; fuel and unflyable aircraft destroyed.
10. 17 Feb, about 12:00–19:00Kampfgruppe DAK takes Fériana and Thelepte; CCB repulses the 21st Panzer Division at Sbeitla, then withdraws; Germans enter Sbeitla after 17:00.
11. Night 17–18 FebThe 1st Armored Division marches through Kasserine; Moore’s force organizes the Kasserine Pass.
12. 18 FebDAK reconnaissance enters Kasserine village; at 14:20 Rommel proposes a thrust on Tébessa; CCB reaches El Ma el Abiod by 18:00.
13. 18 Feb, before midnightThe Comando Supremo orders Rommel to attack toward Le Kef with the 10th and 21st Panzer Divisions.
14. 19 FebKesselring flies to Tunisia; the attacks on the Kasserine Pass and at Sbiba begin; flying is stopped by a sirocco.
15. 20 Feb, 06:00“Group Rommel” takes command of the Axis attack forces.

Judging the retreat

The decision to withdraw was sound. After the losses of 14 and 15 February, II Corps could not hold the basin against two panzer divisions and Rommel’s group, and the Western Dorsal was the line Anderson had wanted from the start. The withdrawal was proposed by a British army commander and approved by an American commander in chief, and most historians since Howe have accepted that it was the right call. What went wrong was the way the Allied front had been set up before the attack, scattered in small packets along the Eastern Dorsal, and the way the battle at Sidi Bou Zid had been fought. Those are the subjects of Chapter 2 and Chapter 3.

The execution was mixed. The evacuation of Gafsa was done in a single night, earlier than First Army had planned, and with some damage to the town and its people. The air command’s evacuation of Thelepte, planned a week ahead, was the best-run part of the retreat: every aircraft that could fly got out, the ground echelons got out, and the base was wrecked. At Sbeitla the conflicting orders between Anderson and Fredendall, issued in the small hours, left the division unsure how long to stay; a night stampede of rear-echelon traffic showed how close to breaking the most battered units had come; and yet the combat units held well past the time they had been ordered to, and Robinett’s command fought a model delaying action before slipping away in the dark. The losses of the four days were heavy in equipment, much of it destroyed by the Americans themselves, and lighter in men than at Sidi Bou Zid.

On the Axis side, the record is one of opportunities taken slowly. Von Arnim’s divisions spent much of 16 February reorganizing; Rommel’s assault group approached an empty Gafsa with great caution; and the two commanders spent 17 and 18 February arguing about objectives while the Americans settled into the passes of the Western Dorsal. By the time the Comando Supremo’s order arrived, the Allies had had three days to bring up reserves. That delay is one reason the battles of 19–22 February in the passes turned out as they did.

Questions readers ask

Was Gafsa lost in battle?

No. The Allied commanders had already decided that Gafsa could not be held against a serious attack, and once the Germans struck at Sidi Bou Zid on 14 February, First Army ordered it evacuated. The American, French and British troops left in the night of 14–15 February, and Rommel’s assault group entered the empty town on the evening of 15 February. American troops under Patton retook Gafsa on 17 March 1943.

How many American aircraft were destroyed at Thelepte?

The sources differ. The U.S. Army Air Forces’ official history says eighteen aircraft that could not fly were burned, five of them beyond repair, and that 124 operational aircraft were flown out. The U.S. Army’s official history, citing an engineer officer’s report, says 34 aircraft were demolished. Both agree that the airfield was left useless to the enemy. About 60,000 gallons of aviation gasoline were poured out, though the Germans reported salvaging 20 tons of it.

Did Rommel want to take Tébessa?

By 18 February, yes. That afternoon he proposed an enveloping thrust through Tébessa into the Allied rear, if he were given both of von Arnim’s panzer divisions. Kesselring supported him and von Arnim opposed him. The Comando Supremo’s order that night aimed the attack at Le Kef instead, and Rommel attacked through the Kasserine Pass and at Sbiba on 19 February. A thrust by part of his force toward Tébessa was stopped at Djebel el Hamra on 21–22 February.

How it connects

The retreat followed directly from the destruction of Combat Command A and the failed counterattack at Sidi Bou Zid on 14–15 February, told in Chapter 3, and from the thin dispositions along the Eastern Dorsal that are the subject of Chapter 2. The men it carried back to the Western Dorsal fought the next battles: Moore’s 19th Engineers and Stark in the Kasserine Pass on 19–20 February (Chapter 5), and Robinett’s CCB at Djebel el Hamra and the British at Thala on 21–22 February (Chapter 6). The XII Air Support Command’s war, from Thelepte to Youks-les-Bains and back, belongs to Chapter 7.

The ground lost in these four days was regained within weeks. The Axis forces fell back from the passes on 22–23 February, American troops walked back through Kasserine toward Fériana on 26 February, the Thelepte airfields were reoccupied on 1 March, and Patton’s II Corps retook Gafsa on 17 March and went on to El Guettar, the story of Chapter 8. The cost, and the long argument over what Kasserine taught the U.S. Army, are in Chapter 9.

Walking the ground today

The route of the retreat can be driven in a day on the modern roads of the Kasserine and Gafsa governorates. Gafsa, now a regional capital and the center of Tunisia’s phosphate industry, still has its kasbah, the old citadel where the French ammunition was blown up in 1943, and the Roman pools that gave ancient Capsa its water. The road north to Feriana runs across open steppe; the ruins of ancient Thelepte lie beside the town, and the wartime airfield nearby still has an active runway, but it is fenced and its use is unclear, so it should be viewed only from public roads (Wikipedia, “Thelepte Airfield”). Kasserine is a much larger town than the village of 1943, with Djebel Chambi, Tunisia’s highest mountain, rising to the west above the pass.

Sbeitla is the most rewarding stop. The archaeological site of Sufetula, at the edge of the modern town, with its forum, three temples, churches and Byzantine forts, is open to visitors, and from the high ground southeast of the town, along the road toward Sidi Bouzid, it is possible to see the kind of open, rolling country, broken by wadis and olive groves, in which Robinett hid his tanks on 17 February. Conditions in the border areas of west-central Tunisia, especially around Djebel Chambi, have at times been unsettled; travelers should check current U.S. State Department advice and use local guides. Americans who died in these battles and were not repatriated lie in the North Africa American Cemetery at Carthage, near Tunis, described in Chapter 9.

Sources

• George F. Howe, Northwest Africa: Seizing the Initiative in the West (U.S. Army in World War II; Washington, D.C.: Office of the Chief of Military History, 1957), especially ch. XXI, “The Enemy Strikes at U.S. II Corps”, ch. XXII, “The Enemy Drives Back the Allied Southern Flank”, and ch. XXIII, “Rommel’s Thrust Through Kasserine Pass”, with their notes on the II Corps, 1st Armored Division and Combat Command B records and the German war diaries.
• Wesley Frank Craven and James Lea Cate, eds., The Army Air Forces in World War II, vol. 2, Europe: Torch to Pointblank (University of Chicago Press, 1949), ch. 4 and ch. 5, for the XII Air Support Command, Thelepte and the evacuation of 17 February.
• Maurer Maurer, Air Force Combat Units of World War II (Office of Air Force History, 1983), and Air Force Historical Research Agency material, as summarized in Wikipedia, “Thelepte Airfield”.
• 443rd AAA Automatic Weapons Battalion (SP) Association, history of the battalion in the Tunisian campaign (1998), Texas Military Forces Museum.
• Ernie Pyle, Scripps-Howard column on the withdrawal from the Sbeitla valley, The Pittsburgh Press, 5 March 1943, as transcribed in the TimeGhost “Roving Reporter” archive; Ernie Pyle, Here Is Your War (New York: Henry Holt, 1943).
• The Charlotte News, 19 February 1943, as summarized in the W. J. Cash / Charlotte News archive (wjcash.org), for Hal Boyle’s AP dispatch.
• On the heroes: Wikipedia, “Paul McDonald Robinett”; Andrew Knighton, “Paul Robinett, Legendary American Tank Commander of World War Two”, War History Online, 2018 (drawing on Orr Kelly); University of North Georgia Press, 1,271 Days a Soldier: The Diaries and Letters of Colonel H. E. Gardiner, ed. Dominic J. Caraccilo (2021), and its press materials; Home of Heroes and TracesOfWar on Gardiner’s Distinguished Service Cross.
• Wikipedia: “William W. Momyer”, “Karl Bülowius”, “Kurt Freiherr von Liebenstein”, “Battle of Sidi Bou Zid”; J. Rickard, “Operation Morgenluft, 16–18 February 1943”, historyofwar.org (2017).
• Books: Martin Blumenson, Kasserine Pass (Houghton Mifflin, 1967; reissued as Rommel’s Last Victory) and “Kasserine Pass, 30 January–22 February 1943,” in Charles E. Heller and William A. Stofft, eds., America’s First Battles, 1776–1965 (University Press of Kansas, 1986); Rick Atkinson, An Army at Dawn: The War in North Africa, 1942–1943 (Henry Holt, 2002); Orr Kelly, Meeting the Fox: The Allied Invasion of Africa, from Operation Torch to Kasserine Pass to Victory in Tunisia (Wiley, 2002); Paul McD. Robinett, Armor Command (McGregor & Werner, 1958); George F. Howe, The Battle History of the 1st Armored Division (Combat Forces Press, 1954); I. S. O. Playfair et al., The Mediterranean and Middle East, vol. IV (HMSO, 1966); B. H. Liddell Hart, ed., The Rommel Papers (Harcourt, Brace, 1953); William O. Darby and William H. Baumer, Darby’s Rangers: We Led the Way (Presidio, 1980); Charles Whiting, Kasserine: First Blood (Stein and Day, 1984).

Photographs: U.S. Army Signal Corps prints in the U.S. National Archives (digitized by the Signal Corps Archive project), U.S. Air Force historical material, the Archivio Centrale dello Stato (Italy) and the German Bundesarchiv, via Wikimedia Commons; modern photographs by Carole Raddato and TOUMOU under Creative Commons licenses. The German and Italian wartime photographs were taken by official propaganda photographers, and several Signal Corps captions do not record the exact place, as noted in the captions. Insignia and aircraft art are original illustrations, not official emblems.