Chapter X: The Withdrawal Begins

The success of the Japanese landings at Lingayen Gulf and Lamon Bay ended all hopes for an American victory in the Philippines. Only one day after the landing to the north, on 23 December, General MacArthur decided that he would have to fall back to Bataan and fight a delaying action there until help could arrive. This decision, made only under the greatest necessity, was the basic strategic decision of the campaign in the Philippines.

"WPO-3 Is In Effect"

Before the war, General MacArthur had determined that he would meet a Japanese attack by offensive action, not by what he considered to be the passive defense provided for in WPO-3. Accordingly, he had ordered his force commanders to meet the Japanese at the beaches and to drive them back into the sea. There was to be "no withdrawal from beach positions." The first Japanese landings between 8 and 10 December had caused no change in the strategy.

Once the Japanese had landed, General MacArthur had to consider seriously the prospect of an eventual withdrawal to Bataan and the evacuation of Manila. To prepare President Quezon for the worst, he sent word to him on the morning of the 12th to be ready to move to Corregidor on four hours' notice.

Shocked and wholly unprepared for this "startling message," Quezon arranged a conference with MacArthur that night at the Manila Hotel. At the meeting, MacArthur explained that there was no immediate cause for concern, and that he was only "preparing for the worst in case the Japanese should land in great force at different places." In such an event, it would be unwise, he told Quezon, to have his forces scattered. he intended to concentrate his army on Bataan, and to move his headquarters, the High Commissioner's office, and the Commonwealth Government to Corregidor and declare Manila an open city. "Do you mean, General," asked Quezon, "that tomorrow you will declare Manila an open city and that some time during the day we shall have to go to Corregidor?" MacArthur's answer was an emphatic "No." He did not seem to be certain that the move would even be necessary, and was evidently only preparing the President for such a possibility. The meeting closed with Quezon's promise to consider the matter further. Later he consented, with reluctance, to move to Corregidor if necessary.[1]

The possibility of a withdrawal seems to have been in the minds of other officers in MacArthur's headquarters before the main Japanese landings. During an inspection of the 21st Field Artillery sector along Lingayen Gulf, Col. Constant L. Irwin, MacArthur's G-3, showed little interest in the


tactical placement of the guns. He seemed concerned, instead, with the location of the ammunition and supply routes, selected to conform with the mission of holding at the beaches. "He took a look at our ammunition disposition and the dangerous supply routes," wrote Colonel Mallonée, instructor of the 21st Field Artillery, "and very violently announced that it would be impossible to withdraw the ammunition in time to save it. . . ."[2] This was the first time, remarked Mallonée, that he heard the word "withdraw." He explained to Colonel Irwin that his orders were to hold at all costs, and repeated Wainwright's order: "We must die in our tracks, falling not backward but forward toward the enemy." The answer of the G-3 officer, was, "Don't believe everything you hear."[3]

Colonel Mallonée, as well as the chief of staff and senior instructor of the 21st Division, was now thoroughly confused about the mission and after a conference decided to request clarification from General Wainwright's headquarters. They were told that the mission was still to hold at all costs, but, added Colonel Mallonée, "by the manner in which it was issued it was evident that there is considerable doubt in the minds of the North Luzon Force command as to whether the mission is actually as given."[4]

As early as 12 December, then, General MacArthur was preparing the ground for measures that would have to be taken if he decided that it was necessary to withdraw to Bataan. When General Homma landed his 14th Army at Lingayen Gulf ten days later, on 22 December, MacArthur still made no change in his plan. But his message to General Marshall on that date shows that he now believed he might have to withdraw quickly. He estimated that the Japanese disembarking from the seventy to eighty transports in Lingayen Gulf had a strength of 80,000 to 100,000 men, and reported that he had on Luzon only about 40,000 men "in units partially equipped." He anticipated that "this enormous tactical discrepancy" would force him "to operate in delaying action on successive lines through the Central Luzon plain to final defensive position on Bataan."[5] When forced to do so, he told General Marshall, he would declare Manila an open city to save the civilian population and move his headquarters, together with the Philippine Commonwealth Government and the High Commissioner's office, to Corregidor, which, he said, "I intend


to hold."[6] General Marshall immediately replied that his proposed line of action was approved and that he was doing his utmost to send aid.[7]

The fighting in North Luzon on 22 and 23 December and the rapid advance by the Japanese to Rosario apparently convinced MacArthur that the time had come to put the scheme for withdrawal into effect. General Wainwright's request on the afternoon of the 23d for permission to withdraw behind the Agno River must have confirmed this decision. To these military considerations must be added General MacArthur's desire to save the city of Manila from destruction.

But the chief reason for the withdrawal order was the failure of the troops to hold the enemy. Up to this time General MacArthur seems to have had the greatest confidence in the fighting qualities of the Philippine Army reservists and in the ability of his forces to hold the central Luzon plain. The events of the 22d and 23d forced a revision of this view. "General MacArthur, viewing the broken, fleeing North Luzon Force," wrote Colonel Collier, a sympathetic observer, "realized that his cherished plan of defeating an enemy attempt to advance toward Manila from the north was not now possible. . . ."[8]

MacArthur's position on 23 December 1941 was somewhat akin to the position in which General Yamashita found himself three years later, when the victorious Americans were preparing to invade Luzon. Realizing that his opponent's air and naval forces were far superior to his own, that American ground forces were free to land on any beaches they chose, and that their superior mobility and fire power were too great for him, he concluded that the Japanese would be unable "to conduct warfare on flat land." Yamashita, therefore, decided to withdraw from Manila and the central Luzon plain, and to fight a delaying action to "divert American forces in Luzon so as to keep them from attacking Japan as long as possible." Unlike General MacArthur, Yamashita hoped to accomplish his objective by withdrawing into the mountains of northern Luzon. He might have been more successful if he had retired to Bataan, as the Americans had four years earlier. From there he could have maintained his forces intact and have denied the Americans, for a time at least, the use of Manila Bay.[9]

The decision having been made to withdraw to Bataan, USAFFE notified all force commanders that "WPO-3 is in effect."[10] Nothing more was required. WPO-3 was an old plan, well known to all U.S. Army officers who had been in the Philippines six months or more. Under it, the Philippine Department headquarters after the experience of numerous maneuvers, had selected


certain delaying positions along the central Luzon plain. These positions had been reconnoitered and were considered fairly strong defensive lines along the route of withdrawal to Bataan. It only remained to issue written orders to supplement the announcement that WPO-3 was in effect.

The next morning, 24 December, at 1200, the USAFFE staff was called to a conference. General Sutherland announced the decision and stated that the headquarters was to be moved to Corregidor that evening. Each man was to take with him only field equipment and one suitcase or bedroll. By special order all officers in the headquarters, except those of high rank had been promoted one grade. To the War Department General MacArthur sent news of his decision, as well as the further information that the Japanese had landed at Atimonan and Mauban that morning.[11] "Tonight I plan to disengage my forces under cover of darkness," he wrote. "For the present, I am remaining in Manila, establishing an advanced headquarters on Corregidor." After evacuating the High Commissioner and the Commonwealth Government, he told the Chief of Staff, he would declare Manila an open city.[12]

On the afternoon of the 24th, President Quezon and High Commissioner Sayre, with their personal and official families, sailed to Corregidor aboard the interisland steamer Mayan. Many Philippine officials simply packed a few belongings and left the city, despite the order that all Commonwealth officials would remain at their posts.[13]

The headquarters began to move out on the Don Esteban after 1900 that day. "It was a beautiful moonlit night," wrote Colonel Collier, "and the cheerful, peaceful murmuring of the rippling waves from the cutting prow of the ship belied the havoc of war."[14] It was Christmas Eve, and the men sat around on deck talking in hushed tones and watching the flames rising from the Navy's fuel dump where over 1,000,000 gallons of oil had been fired earlier in the day. The Don Esteban docked at Corregidor at 2130, and the next morning Headquarters, USAFFE, opened on the island. That day, MacArthur reported to the War Department that his headquarters had moved.[15] A rear echelon, headed by Brig. Gen. Richard J. Marshall, Deputy Chief of Staff, remained behind in Manila to close out the headquarters and supervise the shipment of supplies and the evacuation of the remaining troops.[16]

There was much to do in the days that followed to prepare Bataan for the troops destined to make their last stand there. On the morning of the 24th, Col. Lewis C. Beebe, G-4, USAFFE, and Brig. Gen. Charles C. Drake, Quartermaster, were called to General Marshall's office and there told of the decision to withdraw all troops on Luzon to Bataan and to evacuate Manila. General Drake was instructed to move his base of operations to Bataan immediately and to check on the reserves at Corregidor


to be sure that there was enough to supply 10,000 men for six months. Small barges and boats required to move the supplies from Manila to Corregidor and Bataan were quickly gathered, and within twenty-four hours Corregidor was completely stocked with the supplies for a six months' campaign. At the same time, all supplies were immediately started on their way to Bataan by every available means--water, truck, and rail. Ammunition had already been stored in the peninsula, together with certain defense reserves including 300,000 gallons of gasoline, lubricating, oil, and greases, and about 3,000 tons of canned meats and fish.[17]

In Manila, the rear echelon worked valiantly to get all the supplies out of the city before the Japanese moved in. Those small craft not transferred to Corregidor and Bataan were destroyed; demolitions were carried out with efficiency and dispatch. By the time General Marshall and his men moved out on New Year's Eve, most of the supplies that might possibly be of value to the enemy had been destroyed.[18]

At the same time that a revised supply plan was put into effect, a revised plan of operations was quickly worked out. The object of these plans was to gain time to prepare defenses on Bataan and to permit an orderly withdrawal into the peninsula. Wainwright's North Luzon Force was to hold the Japanese north of the key city of San Fernando, Pampanga--where Route 7, the main highway leading into the Bataan peninsula, began--until 8 January, then withdraw into Bataan.[19] This would provide time for the South Luzon Force to move up past Manila and into Bataan and give those troops already on Bataan an opportunity to establish a line. The withdrawal was to be in five phases, or along five lines. On each line Wainwright's men were to hold only long enough to force the enemy to prepare for an organized attack. The object was to delay, not defeat, the enemy and to reach Bataan intact.[20]

General Parker's South Luzon Force was to withdraw west and north along successive defense lines through and around Manila, across the Pampanga River, spanned by the two bridges known collectively as the Calumpit Bridge, to San Fernando, and then to Bataan. All of the South Luzon Force was to clear the bridge before 8 January. The Calumpit Bridge therefore became a critical point in the plan for withdrawal. It had to be held until all the troops in the South Luzon Force passed over.[21]

To prepare defensive positions on Bataan, the Bataan Defense Force was organized on the 24th. General Parker was placed in command and given two Philippine Army divisions, the 31st and 41st (less 42d Infantry), in addition to the troops already in Bataan to do the job. Command of the South Luzon Force, which consisted during


the withdrawal of the 51st Division (PA), one regiment of the 1st Division, the 42d Infantry, plus supporting tanks and SPMs, passed to General Jones.[22]

The only troops in Bataan when Parker reached there at 1700 of the 24th were the Philippine Division (less 57th Combat Team and one battalion of the 45th Infantry) and a provisional air corps regiment. The 14th Engineers (PS) marked out the defensive positions and the Philippine Army troops, when they arrived on the peninsula, moved into these positions and began to dig foxholes and put up wire. Brig. Gen. Clifford Bluemel's 31st Division (PA), stationed along the Zambales coast was the first into Bataan. Its movement was completed by 26 December. Two days later the 41st Division (PA), less elements, took up its position along the skeleton line.[23]

The plan for the withdrawal of the forces in north and south Luzon called for a difficult maneuver requiring accurate timing and the closest co-ordination. Should the forces in north and south Luzon fail to pull back to Bataan, or should the Japanese seize the road net leading into the peninsula, then the strategic objective of the withdrawal, the denial of Manila Bay to the enemy, would be jeopardized.

The North Luzon Force Plan

The North Luzon Force plan of withdrawal was based on the five delaying positions or lines selected and reconnoitered during peacetime. Separated by the estimated distance which could be covered in one night's march, these lines utilized the terrain features advantageous in defense--rivers, high ground, and swamps. Each was anchored on high ground and took full advantage of natural barriers. They lay across the face of the central Luzon plain and covered the main approaches to Manila, Routes 3 and 5. (Map 6)

The first defensive line, known as D-1, extended in an easterly direction from Aguilar, south of Lingayen Gulf on Route 13, through San Carlos to Urdaneta on Route 3. As Col. William F. Maher, Wainwright's chief of staff, has observed, the D-1 line "was simply a line on which we hoped to be able to reorganize the badly disorganized forces north of the Agno River."[24]

The second position, the D-2 line, extended in general along the arc of the Agno River, one of the formidable natural barriers in the central plain. After holding for one day on this line, the troops were to retire next to the D-3 line, stretching from Santa Ignacia on the west through Gerona and Guimba to San Jose on the east. The D-4 line was approximately twenty-five miles long and extended from Tarlac on the left (west) to Cabanatuan on the right. Small rivers and streams intersected this line, which, at Cabanatuan, was anchored on the Pampanga River.

The final and most southerly position, called the D-5 line, stretched from Bamban in front of Mt. Arayat, across Route 5 to Sibul Springs. Southeast of Mt. Arayat, between the Pampanga River and Route 5, was the Candaba Swamp, which broke the central plain into two narrow corridors leading toward Manila. Of the five lines, only the last, the D-5 line, was to be organized for a protracted defense. Plans called for a stand here until the South Luzon


MAP 6: Withdrawal in the North


Force could slip behind the North Luzon Force, up Route 3, into San Fernando.[25]

During its withdrawal to Bataan, the North Luzon Force was to be supported by General Weaver's Provisional Tank Group, whose job it would be to cover the withdrawal, sweep enemy avenues of approach, and halt hostile mechanized movement. The tanks were deployed on alternate sides of the road, at curves and bends, to achieve maximum sweep of their weapons with a minimum of exposure. Always they were to take care that they left themselves a route of escape. When required to withdraw, the tanks were to move back one at a time, under cover of the forward tank. The tankers were to select their positions after a careful reconnaissance, and with an eye to fields of fire, alternate positions, avenues of approach, and emergency escape routes.[26]

The success of the withdrawal would depend to a large degree on the engineers. Their task was twofold: to maintain roads and bridges ahead of the retreating columns, and to destroy the bridges and block the roads already passed to halt the enemy advance. Demolitions and the construction of obstacles before the D-1 line were to be accomplished by the front-line units; North Luzon Force engineers, consisting principally of the engineer battalion of the 91st Division (PA), were made responsible for all work south of that line. The destruction of railroad bridges was left to a special detachment of demolition experts from MacArthur's headquarters, attached to North Luzon Force. Demolitions were to be executed by the engineers when ordered by the division or covering force commander and when the tanks and vehicles of the last elements of the rear guard had cleared the bridge.[27]

The term line, applied to the five delaying positions, is misleading. Actually the front was too wide to be held continuously by the forces available to General Wainwright. Unit commanders were given considerable leeway in occupying their positions and usually could do little more than place their troops so as to cover the most likely routes of approach. Each line was to be occupied before dawn, held during the day, and evacuated at night, the troops withdrawing to the next line. Their withdrawal would be covered by a shell, a small part of the retiring force, which was to remain in position until just before dawn when it was to pull back hastily to rejoin its parent unit on the line below. This shell, in theory, would consist of an infantry-artillery team, but in practice often included only one of these arms.

By occupying these positions successively and holding them with a shell while the bulk of the force retired to the safety of a prepared position to the rear, MacArthur hoped to force the enemy to halt and deploy for an attack before each position. By the time he was ready to attack, the line would be evacuated. In this way, the Japanese advance southward would be considerably delayed, and time would be gained to prepare defenses on Bataan and to permit the South Luzon Force to pass into the peninsula


behind the North Luzon Force. The danger of the scheme lay in the Japanese control of the air, which made it possible for them to play havoc with the retiring road-bound tanks and artillery. The risk was a calculated one, but the danger was minimized by limiting important movements to the hours of darkness.[28]

The supply of the troops during the withdrawal would be difficult. The problems ordinarily encountered in supplying large bodies of mobile troops during a retrograde movement would be complicated by the shortage of supplies and trained supply officers, the necessity of moving a large amount of equipment to Bataan, and the destruction of those supplies which could not be saved. The fact that most of the men were inadequately trained, poorly equipped, and often undisciplined would add considerably to the difficulties.[29]

To the Tarlac-Cabanatuan Line

On Christmas Even the North Luzon Force stood generally along a line extending from Tayug on the east through Urdaneta and San Carlos to Aguilar on the west. (Map 4) All units were under orders to hold for twenty-four hours before falling back to the Agno.

On the right (east) was the 26th Cavalry (PS). That afternoon the Scouts had been forced to retreat from Binalonan across the Agno River to Tayug, thus actually anchoring the North Luzon Force at the start of the withdrawal on the D-2 line. At Tayug, the cavalrymen had relieved the 71st Engineer Battalion (PA) covering the river crossing and had joined the 91st Division (PA) and the remnants of the 71st.[30]

West of Tayug, holding the center of the North Luzon Force line from Urdaneta to San Carlos, was General Brougher's 11th Division (PA). Also in the center was the 192d Tank Battalion, at this time the only armor in support of the North Luzon Force. On the afternoon of the 24th it was moving south toward the Agno, under orders to deploy along the south bank. Already on its way toward the river was the 194th, which had left Manila that morning with orders to assemble in the vicinity of Carmen.[31]

Extending the North Luzon Force line west from San Carlos to the Zambales Mountains, straddling the Agno, was General Capinpin's 21st Division (PA). Stationed initially along the southern shore of Lingayen Gulf, this division had not yet come in contact with the enemy. Its orders were to withdraw at 1900 on the 24th in two columns along the two roads, one on each side of the river.

Withdrawal to the Agno

At the appointed hour, 1900 of 24 December, the 21st Division began to withdraw.[32] Wire communication between the


division command post and front-line units was discontinued and signal troops began reclaiming the wire for later use. The first units to move out were the 22d Infantry and the 1st Battalion, 21st Field Artillery. Blowing the large bridges to the rear, they retired down the road toward San Carlos.

West of the Agno, the 21st Infantry began to withdraw from its beach positions at about 1900. By 2130 of Christmas Eve, division headquarters had reached its new command post on Route 13, eleven miles south of San Carlos. So quiet had the night been that Col. Ray M. O'Day, division instructor, turning on his radio to hear the midnight mass, "looked up at Heaven and could hardly believe it was a war-torn world."[33]

The withdrawal continued all through the night. By about 0400 of 25 December the bulk of the 21st Infantry had reached Aguilar and, when the sun rose, its 3d Battalion moved across the Agno in Bancas to take up positions along the east-west road to San Carlos. It was not until late afternoon that the last covering units reached the D-1 line. They had been held up by delays in the destruction of many small bridges, and in one case, premature demolition of a bridge had forced the abandonment of precious vehicles. There had been no hostile contact during the withdrawal.

The rest of the North Luzon Force spent a less peaceful Christmas. The enemy, prevented from reaching the Agno on 24 December by the stiff defense of the 26th Cavalry, continued his efforts the next day. With Binalonan in his possession, General Tsuchibashi, the 48th Division commander, could now split his force into two columns. One he sent south on Route 3 to Urdaneta, where the 11th Division was posted; the other went east toward Tayug. (Map 6) The column along Route 3 would consist of the 1st and 2d Formosa Infantry with the 4th Tank Regiment. The remainder of the 48th Division (less 1st Battalion, 47th Infantry at Damortis), concentrated in the Pozorrubio-Binalonan area during the night of 24-25 December.[34]

At 0200 of Christmas morning, the 1st and 2d Formosa and the 4th Tank Regiment moved out against Urdaneta, which was defended by elements of the 11th Division's 13th Infantry (PA). The fight lasted all morning but the Japanese proved too strong for the Filipinos and by noon had control of the town. The 11th Division then began falling back toward the Agno.[35]

Meanwhile, on the right flank of the North Luzon Force there had been a shuffling of units. The 71st Division, ordered to San Fernando, Pampanga, for reorganization, was moving out of the line. The 91st Division, with the 26th Cavalry attached, was under orders to pull back to the next line at 2100, leaving a shell on the river until dawn of the 26th. The cavalry was to hold the river line at Tayug to cover General Stevens' withdrawal and to protect the force right flank. A shell from the 91st Division, the 92d Combat Team, was to take up a


position to Pierce's left, along the Agno as far south as Carmen.[36]

VILLASIS-CARMEN BRIDGE over the Agno River on Route 3. (Photograph taken in 1935.)

By evening of 25 December, the 11th Division, in the center, stood on the Agno River and was in its D-2 positions. Defense of Carmen and its important bridge, rebuilt by the 91st Engineer Battalion, was assigned to the 1st Battalion, 21st Infantry, in force reserve since the second week of hostilities. To its left (west) along a 2,000-yard front west of Carmen, was the 13th Infantry. The rest of the 11th Division sector, extending to Bautista, was held by the 11th Infantry. The 21st Division was on the left, and 92d Combat Team and the 26th Cavalry on the right of the D-2 line. Spread thin along the Agno River between Carmen and Route 13, a distance of twenty-five miles, was the 194th Tank Battalion (less Company C) which had reached the river at 1900 the previous night. Tank support on the right side of the line was provided by the 192d Tank Battalion, which covered the sixteen miles from Carmen to Tayug.[37]


AGNO RIVER between Bayambang and Carmen. Bayambang, foreground, stretches along the southwestern bank of the river.

--172 & 173--

When the move was completed and all of the North Luzon Force had reached D-2, General MacArthur reported to Washington, "Our position now along the Agno River."[38] Thus far, the withdrawal had proceeded satisfactorily. The Japanese had attacked at only one point and had achieved their objective, but had not disrupted the American scheme of withdrawal. Already the important bridges across the Agno, at Bayambang and Villasis, were ready for destruction.

Fight on the Agno

The D-2 line, from three to twelve miles behind the D-1 line, depended primarily on the curving Agno River for its strength. Both flanks were guarded by high ground. The two critical points on the line were Tayug and Carmen, both important road junctions. A break-through at Tayug would open the right of the North Luzon Force to a hostile flanking movement; a Japanese penetration at Carmen would split the defenses in the center. Failing to hold either of these vital points, the North Luzon Force would have to abandon its position and perhaps its plan of withdrawal.

While Wainwright was pulling back to the Agno, the Japanese had not been idle. Shortly after noon on 25 December, and advance element of Lt. Col. Kuro Kitamura's 48th Reconnaissance Regiment, moving east from Binalonan, met patrols of the 26th Cavalry at Asingan, across the river from Tayug. By 1900 Kitamura's troops had driven the Scouts back to the river where the 2d Squadron was already in position on the opposite shore. Only the soft mud of the riverbank had prevented the Japanese tanks from crossing immediately. The struggle continued into the night and at 0200 the next morning, when the Japanese finally reached the opposite shore, the Scouts broke off the action. By 0400 Tayug was in enemy hands. Since further opposition was futile, Colonel Pierce withdrew to the 91st Division line at Umingan, ten miles to the southeast. Blowing eight bridges between Tayug and San Quintin as it retired, the decimated 26th Cavalry passed through General Stevens' line at 0545. Later in the day, under North Luzon Force orders, it continued south toward Bataan as force reserve.[39] The Scouts had fought with great effect in the five days since the Japanese landings and had contributed in a large degree to the enemy delay. Their discipline and courageous stands at Damortis, Rosario, and Binalonan had shown that the Philippine soldier, properly trained, equipped, and led, was the equal of any.

While the 48th Reconnaissance Regiment was attacking the 26th Cavalry at Tayug, the second of General Tsuchibashi's columns--consisting of the 2d Formosa, a battalion of the 1st Formosa, and the 4th Tank Regiment--was moving due south against Carmen. During the evening of 25 December, this force entered unoccupied Villasis on Route 3, only a mile north of Carmen and the Agno River. After a preliminary air strike behind the lines by twelve planes of the 8th and 16th Light Bombardment Regiments, the Japanese opened the assault against Carmen, crossing the Agno near Villasis after sunset of the 26th. The 2d Formosa and the 4th Tank Regiment, with artillery in support, met opposition


TARLAC RAILROAD STATION after Japanese bombing, above;
salvaging a carload of .30-caliber ammunition, below.


from the 37-mm. guns of the 194th Tank Battalion, which, having only armor-piercing shells, was unable to hold up the Japanese advance.[40]

Late in the afternoon of the 26th, when news of the withdrawal of the 26th Cavalry on the right reached Wainwright, he ordered the 11th Division to fall back through Carmen to Route 3, then south to the D-3 line. Before the move could get under way, the Japanese shattered the 1st Battalion, 21st Infantry, at Carmen, inflicting two hundred casualties and capturing Maj. Robert Besson, the battalion commander. By 1930 Carmen was in enemy hands. The Japanese pushed on vigorously, a battalion of the 1st Formosa striking the 92d Combat Team on the right of the 11th Division line. Two hours later the enemy was in Rosales, three miles to the east of Carmen.

With Route 3 in Japanese hands, the 11th Division was forced to fall back via the Manila Railroad, which extended along the western (left) edge of its sector. There was no other route of retreat in this area. Behind the division front was a large, roadless area covered with rice fields. The only routes leading to the rear were on the division flanks--Route 3 on the east and the Manila Railroad on the west. Swift action o n the part of General Brougher in commandeering and dispatching a locomotive and several freight cars from Tarlac that night made possible the escape of the troops.[41]

The Provisional Tank Group encountered greater difficulty in withdrawing than had the infantry. Col. Ernest B. Miller, the 194th Tank Battalion commander, had told General Weaver at 1830 of the 26th that the enemy might soon cross the Agno and that there remained "nothing but the tanks to stop it."[42] Actually, the Japanese were already across the river. Weaver ordered Miller to hold at the D-2 line until 0500 the following day. The 192d Tank Battalion to the east was also ordered to hold, but Colonel Miller as the senior tank officer was authorized to withdraw both battalions sooner if Japanese action threatened to cut their line of retreat.[43]

In the 192 Tank Battalion area the tactical situation made compliance with General Weaver's order impossible. Around dusk on 26 December, Col. John H. Rodman, commanding the 92d Combat Team, informed Col. Theodore Wickord, the 192d commander, that the infantry was pulling back on the right to form a line from Carmen to Umingan. When the 92d pulled back at about 2100, Wickord's battalion also moved out. It moved east past Carmen, then south, before the Japanese could blockade


the route of escape, and reached the D-3 line without difficulty.[44]

Meanwhile, the 194th Tank Battalion made its own way south as best it could. The tanks of Company A fought their way through a Japanese roadblock at the edge of Carmen and retreated down Route 3. Above San Manuel, about six miles south, Colonel Miller, the battalion commander, organized a roadblock with three tanks; all the other he sent to the rear. Shortly after, a single half-track with a 75-mm. gun (SPM), commanded by Capt. Gordon H. Peck, came down the road after having cut its way through the cane fields. Placing himself under Colonel Miller's orders, Peck took his place at the roadblock. At about 2300, General Brougher, the 11th Division commander, arrived at San Manuel. He explained that his division was moving back by rail and asked that the tanks cover the railroad until the Filipino troops could pass through to safety. It was finally agreed that the block would be held as long as possible before the tanks and the SPM fell back five miles to Moncada, where the railroad crossed Route 3. The troop trains carrying the 11th Division were expected to pass through that town at 0400 on 27 December.[45]

All was quiet at the roadblock until a few hours before dawn. At about 0245, after the last stragglers had cleared the block, a Japanese armored column, apparently advance elements of the 4th Tank Regiment, reached the spot. Fire from the American tanks and SPMs swept the highway and adjoining ditches. The Japanese were taken completely by surprise and after fifteen minutes pulled back. Fearing encirclement by Japanese infantry, Miller and Peck then struck out for Moncada.[46]

The tanks and the SPM that had formed the roadblock moved slowly down Route 3 in the dark hours before dawn. They reached the rail crossing in Moncada only a scant ten minutes before the 11th Division troop trains entered the town. Once the trains had cleared the danger point, Colonel Miller continued his march south and reached the D-3 line at Gerona at about 0830 on the morning of 27 December. Here he was joined by the survivors of the battalion's Company D. Cut off from retreat, the company had come south along an old carabao cart trail, the Manila Railroad tracks, and Route 3. It found the bridge just below Moncada destroyed and was forced to leave its fifteen tanks north of the stream. This decision had been made in the hope that some of the men could return later with guides and bring the tanks south. This expectation could not be fulfilled and the tanks were lost for the rest of the campaign.[47]


The D-3 Line

Approximately forty miles in length, the D-3 line stretched across the Luzon central plain midway between Lingayen Gulf and San Fernando, Pampanga, from a point just west of Santa Ignacia on Route 13 to San Jose in Nueva Ecija Province, at the junction of Routes 5 and 8, Deployed along this line were the 91st, 11th, and 21st Divisions (PA), supported by the Provisional Tank Group and the 75-mm. guns (SPM).

The right (east) flank, resting on the foothills of the Sierra Madre, was held by the 91st Division which had taken up positions across Route 5 and on the south bank of a small river in the vicinity of San Jose.[48] Between Route 5 and Gerona on Route 3 were the 11th Division and the bulk of the Tank Group--the 194th at Gerona and the 192 to its right. The 21st Division, whose two columns had reunited at Camiling, was in position between Gerona and Santa Ignacia at the edge of the Zambales Mountains.[49]

Despite occasional alarms there was no action on the D-3 line on 27 December. That night the North Luzon Force made ready to fall back to the D-4 line. The 91st Division began pulling out at about 1730 and by 0430 had reached the south bank of the Pampanga a few thousand yards below Cabanatuan. Two hours later the entire unit was ordered into the line between Cabanatuan and Carmen, Nueva Ecija, a barrio on the road ten miles west of Cabanatuan and not to be confused with the village of the same name on the Agno. At Carmen the 91st Division tied in with units of the 11th Infantry that had withdrawn from the D-3 line during the night and were deployed from Carmen west to La Paz. The 21st Division stood on the left of the 11th Division, extending the line to Tarlac, where Route 13 joined Route 3 and the main track of the Manila Railroad. The tanks were in general support.

On the Agno River the Japanese halted to consolidate their position and bring up more troops. During the 27th, artillery, armor, and service troops moved forward to join the 48th Division. The 47th Infantry and a battalion of artillery, in reserve near Pozorrubio since 24 December, together with the 7th Tank Regiment, were dispatched to Tayug. Infantry and artillery units occupied San Quintin to the south and patrols pushed forward into undefended Umingan. On the 48th Division right (west), the 1st Formosa consolidated its hold on Rosales. One battalion of the regiment remained at Urdaneta, and another went on to Carmen to relieve Colonel Tanaka's troops who then moved back across the Agno to Villasis for rest.[50]

By 28 December the North Luzon Force was on the D-4 line. In the face of a well-trained and better equipped enemy, it had fulfilled its mission--to hold the Agno line until the night of 26-27 December and to withdraw to the Tarlac-Cabanatuan line. Now from positions along this line, the


troops in North Luzon awaited the next attack.


As the front-line unites moved back, the troops to the rear began to carry out the supply plan. On 24 December General MacArthur's headquarters had ordered the evacuation of Fort Stotsenburg and the destruction of its 300,000 gallons of gasoline and large amounts of high octane fuel. Lt. Col. Wallace E. Durst, Post Quartermaster, was able to save about 50,000 gallons of gas by shipping some of it to the rear and issuing the rest to vehicles in the immediate area. "No material amount of gasoline," reported Durst's assistant, Lt. Col. Irvin Alexander, "was abandoned to the enemy."[51] In addition to gasoline, Stotsenburg stocks included 8,000 pounds of fresh beef, about 100,000 components of dry rations, large supplies of clothing, and air corps ammunition and equipment. When the post was finally abandoned, almost nothing of value was left, according to Colonel Alexander. All supplies, he said, had been shipped to Bataan or issued to troops in the Stotsenburg area.[52]

The evacuation of Fort Stotsenburg long before the approach of enemy forces, aroused much criticism from officers who disagreed sharply with Colonel Alexander's optimistic statements on the amount of supplies saved. Colonel Collier exaggeratedly described the evacuation of Stotsenburg as a "frenzied departure" in which "warehouses filled with food, clothing, and other military supplies were left intact." Also left behind, he reported, were 250,000 gallons of gasoline and several obsolete but serviceable planes.[53] General Drake, MacArthur's quartermaster, reported that only a portion of the reserve supplies stocked at Stotsenburg had been removed before its evacuation.[54]

On the afternoon of 25 December, as North Luzon Force fell back to me D-2 line, Lt. Col. Charles S. Lawrence, commanding the Tarlac Depot, had informed Drake that evacuation of the depot would be necessary very soon. In the absence of orders to the contrary, he said, he would issue all his supplies, including five days' subsistence for the North Luzon Force, at one time and head for Bataan with his men.[55] That night he learned from Lt. Col. Gyles Merrill, Wainwright's supply officer, that the line through Tarlac would be occupied on the night of 27 December. Merrill suggested that Lawrence place his remaining rations in dumps at Tarlac, to be picked up by the troops as they withdrew. With Wainwright's approval Lawrence placed the supplies in separate dumps, one for each division or separate unit. Troops of the 21st Division Headquarters Company were posted as guards.


This done, Lawrence and his men left for Bataan.[56]

The evacuation of Stotsenburg and Tarlac was typical of the hurried movement of supplies once the plan of withdrawal had gone into effect. "The troops withdrew so fast," reported General Drake, "that we could not put into operation any of our withdrawal plans to cover this movement."[57] There was scarcely time to remove "a few defense reserve supplies" from McKinley and Stotsenburg and no time to evacuate the depots established before the war at Tarlac and Los Banos. Fortunately, many of the supplies left behind were picked up by the units as they withdrew, and much of the remainder was destroyed.

Closely related to the difficulty of supply and evacuation was the scarcity of motor vehicles on Luzon. Even the addition of civilian vehicles did not solve this problem. "The fact is," wrote Colonel Lawrence, "that there was not sufficient motor equipment in the Philippines to begin to meet fully all the requirements."[58] This shortage was made more serious by the failure of commanders to return the vehicles which brought their supplies. Even more reprehensible was the hijacking and commandeering of vehicles along the highways, often by commanders who feared that they would not have the transportation to move their troops and equipment in an emergency. These practices "resulted in confusion and caused a complete interruption in motor transport service during the period of evacuation of supplies to Bataan."[59]

The Tarlac-Cabanatuan Line

The original plan of withdrawal called for only a brief halt at the D-4 line, just long enough to force the enemy to stop and prepare for a co-ordinated attack. A determined stand would be made on the D-5 line. On 27 December General Wainwright changed this plan. Fearing that a quick withdrawal from D-4 would leave too little margin for error between his last defensive line and the vital bridges across the Pampanga River at Calumpit, over which the South Luzon Force would have to pass, he decided to hold at Tarlac and Cabanatuan, the D-4 line. Late that night he issued new orders to his North Luzon Force abandoning D-5 as the final line of defense. "D-4 will be held at all costs until ordered withdraw," he announced. "Maximum delay will be effected on each position. Withdrawal plan later."[60]

The final plan for holding the D-4 line and for the withdrawal to follow utilized the existing deployment of units already on the line. The 91st Division was assigned the eastern edge of the central plain, the zone between the Pampanga River, which paralleled Route 5, and the mountains to the east. The critical point in this sector was Cabanatuan, where the roads from the north converged into Route 5 which led south toward Manila. When ordered to withdraw, the division would move down Route 5 to Plaridel, a distance of forty-five miles, thence west to Calumpit where Route 3 crossed the Pampanga River.


The 11th Division was on the left of the 91st, in the area between Carmen and Route 3. It was to retire along the secondary roads in its sector. The 21st Division was on the western edge of the central plain, covering Tarlac and Route 3. Its line of retreat was along Route 3 to Angeles, thence to Bataan by Route 74.[61] As a further protection to the Calumpit bridges and the South Luzon Force route of withdrawal, the 194th Tank Battalion, reduced to twenty tanks, was pulled out of the D-4 line by MacArthur's headquarters on the 29th and ordered back to Apalit, three miles northwest of Calumpit, to a position of readiness. The day before, Company A of the 192s had been shifted from the 91st Division sector to the area west of the Pampanga and now, with a platoon of the 194th, formed the only tank support between the Pampanga and Route 3. The rest of Colonel Wickord's battalion remained in position east of the Pampanga, in support of the 91st Division.[62]

When all units were on the line, General MacArthur reported to the War Department that he was "endeavoring to temporarily hold hard in the north" until the North and South Luzon Forces could join at San Fernando after which he would "pivot on my left into Bataan." American and Filipino troops were "tired but well in hand." In this report, MacArthur mistakenly estimated that his North Luzon Force alone was facing three Japanese divisions. These enemy troops, he pointed out, were excellent, and their equipment "modern and extensive." Although the Japanese were not then exerting heavy pressure against his line, MacArthur believed that this inactivity would soon end. The enemy, he warned, was "undoubtedly setting up a powerful attack both north and south simultaneously designed to pin me down in place and crush me."[63]

General MacArthur's estimate of the enemy's intentions was correct. The arrival of the 48th Division at the Agno River had completed the landing phase of the operation. General Homma was now ready to drive on through Cabanatuan and Tarlac to Manila.

As of noon, 27 December, the North Luzon Force position seemed to the Japanese to favor a rapid advance. American air power had been knocked out and the Philippine garrison was effectively cut off from reinforcement. Three of the divisions which had opposed their landings, the 11th, 71st, and 91st, as well as armor and cavalry, the Japanese believed, had suffered decisive defeats. The Japanese were also aware of General MacArthur's move to Corregidor and of the transfer of at least one division--the 31st--to Bataan. On the basis of his intelligence estimate General Homma reasoned correctly that MacArthur planned a delaying action "in one corner of Bataan" and on Corregidor.[64]

Despite this correct evaluation of American intentions, the consensus in the 14th Army staff was for a continuation of the drive on Manila. The mission assigned by Imperial General Headquarters was to take Manila, and it is doubtful that Army had the authority to divert any of its forces from that mission. As Lt. Col. Yoshio Nakajima, 14th Army intelligence officer, wrote: "Since the mission of the 14th Army was to occupy Manila, the main force proceeded


to that city."[65] Some even felt that, since Manila was the main objective, the withdrawal to Bataan "expedited the completion of our mission."[66]

The plan finally adopted for the advance from the Agno River utilized one division, reinforced, supported by armor and aircraft. The main effort was to be made on the east, along Route 5, and the immediate objective was Cabanatuan. The 48th Division would jump off from the Agno River on the 28th and advance toward that town. Simultaneously, the Kamijima Detachment, consisting of elements of the 9th Infantry and supporting artillery, would move from its positions along the Lingayen coast to Carmen to protect the right flank of the 48th Division. From there it would presumably advance down Route 3 toward Tarlac. The only concession made to the obvious American withdrawal to Bataan was to order General Tsuchibashi to send an infantry regiment with heavy artillery support to Tarlac to assist the 9th Infantry in its effort to move speedily down the central plain and seize the road net leading into the peninsula. Supporting the 48th Division advance were the 4th and 7th Tanks Regiments, 14th Army artillery, and the 5th Air Group.[67]

Threat on the East

The key to the right flank of the D-4 line was Cabanatuan. Situated on the banks of the Pampanga River, the town is an important road junction on Route 5. The river, about 100 yards wide at this point, and unfordable by motor vehicles, flows swiftly in its twisting and irregular course. Approaching Cabanatuan from the mountains to the northeast, the Pampanga passes the town about 3,000 yards to the north then turns sharply south to flow west of the town and continue its errant way in a southwesterly direction toward Manila Bay. At Cabanatuan two bridges span the swiftly flowing river: one to the north and another to the west. It was in the general vicinity of these bridges that the Japanese first attacked the D-4 line.

The 14th Army advance from the Agno began on schedule on the morning of 28 December, at the same time that General Homma moved his command post to Binalonan. In the lead were the 4th and 7th Tank Regiments, a battalion of the 2d Formosa, and a battalion of the 48th Mountain Artillery which advanced through San Quintin to San Jose. From there, they struck southeast, crossed the Pampanga at Rizal, and by 29 December had reached Bongabon, in position to threaten the right flank of the D-4 line.

The 48th Division followed in two columns. The west column, consisting of the 1st Formosa supported by a battalion of artillery, left Rosales before dawn of the 29th and marched southeast through Guimba, then east to Baloc on Route 5, north of Cabanatuan. The east column, consisting of the 2d Formosa, 47th Infantry, 48th Reconnaissance, and artillery and engineer units, followed behind the tank regiments to San Jose, where Route 5 intersected Route 8, and then followed the former toward Cabanatuan.[68]

At Cabanatuan, the main strength of the 91st Division, the 92d Combat Team,


waited for the attack. In and around the town were the 2d and 3d Battalions, and to the left extending to the Pampanga, was the 1st Battalion. Both bridges had been blown and were considered impassable for wheeled traffic, but not for foot troops. Moreover, the river was fordable north of Cabanatuan.[69] On the morning of 29 December, the left (east) column of the 48th Division reached the Pampanga northwest of Cabanatuan, but it was the tanks, driving down from Bongabon, that reached the town first. As the tankers approached, the 47th Infantry, under cover of an artillery bombardment, began crossing the river. It was now late in the afternoon, and the 92d Combat Team, outflanked and faced by a superior enemy, fell back. That night the Japanese entered Cabanatuan.[70]

The Japanese did not stop at Cabanatuan. Led by Maj. Gen. Koichi Abe, 48th Division infantry group commander, they continued south along Route 5 on 30 December. Followed by two battalions of the 48th Mountain Artillery and a battalion of 150-mm. howitzers of the 1st Field Heavy Artillery Regiment, the 47th Infantry pursued the withdrawing 91st Division toward Gapan, about fourteen miles below Cabanatuan. Just north of that village the defenders crossed the Penaranda River, destroying the steel highway bridge over that stream. Urged on by Lt. Louis I. Bentz, Jr., about sixty-five Filipinos of the 92d Infantry formed a line along the south bank of the river, while the remainder of the regiment, bolstered by three hundred high school ROTC boys who had arrived that morning from Manila, occupied a mile-long line from the village west to the Pampanga. The 47th Infantry hit this line late in the afternoon and broke through with little difficulty. By nightfall the enemy had entered the town. The remnants of the 91st Division withdrew toward Baliuag, twenty-five miles south on Route 5, where they planned to reorganize.

The rapid advance of the Japanese along Route 5 jeopardized the American right and resulted in a shortening of the D-5 line. The North Luzon Force right flank would now have to be anchored on Mt. Arayat, west of Route 5, instead of Sibul Springs to the east. Route 5 lay open and the enemy was well on his way toward the Calumpit area. Unless he was held, the withdrawal of the South Luzon Force would be threatened.

Advance in the Center

The center of the D-4 line, from the Pampanga to Tarlac, was held by the 11th Division. Paralleling the front was an east-west road. The critical points in the line were Zaragoza and La Paz, held by the 11th Infantry. The 2d Battalion was in front of La Paz, the 3d Battalion to the east above Zaragoza, and the 1st Battalion in reserve about 5,000 yards to the south. Company A, 192d Tank Battalion, was in general support near Zaragoza. The only route of withdrawal was down a secondary road from La Paz to Concepcion, about thirteen miles, then west to Route 3.

In the initial deployment of the 11th Infantry no provision had been made for


guarding the eastern entrance to the critical east-west road which ran behind the line. The 92d Infantry on the right was supposed to protect that flank, but Maj. Russel W. Volckmann, acting 11th Infantry commander, was uneasy about this arrangement. Recognizing the importance of the road and the vulnerability of his position he shifted his line so that troops of his 3d Battalion were in position to guard the road. A roadblock was established on the west side of the bridge across the Dalagot River, leading into Zaragoza, and a platoon of tanks placed in position there. The bridge was prepared for demolition, but the river was easily fordable by foot troops. The organization of the roadblock was a wise precaution, for the Tarlac-Cabanatuan road had already been exposed on the east by the withdrawal of the 91st Division.[71]

The assault against the 11th Division was made by the Kanno Detachment, consisting of the 3d Battalion, 2d Formosa, supported by a battalion of the 48th Mountain Artillery, substantially the same force which had landed at Vigan on 10 December. This force was the one which General Tsuchibashi had assigned to assist the Kamijima Detachment in its drive toward San Fernando. Its mission was to move south along Route 5 to Cabanatuan, then push west to outflank Tarlac, which Colonel Kamijima was approaching from the north. This maneuver would cover the right flank of the 48th Division and, if executed speedily and successfully, would turn the North Luzon line and cut off the retreat of the American troops in the center.[72]

The Kanno Detachment jumped off from Talevera, north of Cabanatuan, at 0100 on 30 December. preceded by bicycle-mounted infantry, the unit cleared Cabanatuan, already in Japanese hands, shortly after and pushed on along the Cabanatuan-Tarlac road, disregarding security measures. At 0315 and alert tanker of the 192d Tank Battalion observed a large number of cyclists in column approaching Zaragoza. When the Japanese reached the American position they were greeted by point-blank fire from the alerted tanks. At the mercy of the tanks, the cyclists lost an estimated eighty-two men before they could make their escape.

It was still dark when the action ended. The tank commander, fearing infiltration by enemy infantry, withdrew his platoon across the Zaragoza bridge, then insisted that the bridge be blown though the 11th Infantry troops were still on the other side. The commander of the engineer detachment had no choice but to comply and lit the time fuses. So surprised was the troop commander when the bridge was blown that he ordered an investigation immediately and incorrectly concluded

. . . that the engineer lieutenant had left the destruction of the bridge to his platoon sergeant and departed for the rear. The platoon sergeant detailed a private and departed with the rest of the men. The private, not to be outdone, had found a civilian, instructed him how to light the dynamite, paid him one peso and then left to join his platoon. The civilian, after hearing the shooting, became excited and blew the bridge.[73]



The premature destruction of the bridge took the tanks out of the action and left the infantry, still on the far side of the shallow river, without the support of the armor.

When daylight came the Kanno Detachment struck the roadblock with heavy rifle and mortar fire. Part of the detachment had swung around to the north and now began to exert pressure from that direction. Fearing that his battalion might be outflanked, the commander pulled his men back across the river. By noon, they were established in positions along the west bank. Despite heavy casualties and the presence of a strong hostile patrol above La Paz, the battalion commander felt he could hold the enemy at the river line.

Shortly after noon the Japanese artillery opened fire against the 3d Battalion, preparatory to an infantry attack. After a twenty-minute barrage by 75-mm. guns of the 48th Mountain Artillery, the Kanno Detachment began to cross the river. Unable to halt the enemy, the 3d Battalion moved west along the Zaragoza-La Paz road. Colonel Kanno brought his men safely across, then halted the advance until he could get his heavier weapons across the river. The 3d Battalion, about 500 yards to the west and supported by tanks, awaited the attack. At 1415 a Japanese antitank gun moved into the Japanese line and directed its fire against the Americans. It was finally knocked out, but only after it had destroyed the lead American tank.

With the lead tank gone and their location known to the enemy, the tanks began to pull back. Since they were not under 11th


Infantry control, there was no way to keep them in position. The Japanese immediately unleashed a heavy barrage, threatening the American positions. Major Volckmann, who was on the scene, organized a counterattack with the battalion reserve. The counterattack opened at 1500 and, although no ground was gained, it evidently surprised the Japanese and led them to believe the defenders were stronger than they actually were. When the Japanese fire slackened, the 3d Battalion withdrew again, this time about 1,500 yards to the west along the La Paz road. By 1630 the men were in their new positions.

No sooner had the 3d Battalion taken up its new position that it received orders to pull back. These orders originated in Wainwright's headquarters, where it had become apparent during the day that the entire line was threatened by the 48th Division's breakthrough at Cabanatuan. Division commanders were ordered to pull back to the D-5 line. General Brougher, accordingly, directed his men holding the center of the line to withdraw through La Paz to Concepcion. The 11th Infantry immediately began to assemble at La Paz. By 1730 the 3d Battalion had fallen back across the bridge just east of that point, the remainder of the regiment retiring before it. When all the troops were across, the bridge was destroyed. At this moment the Kanno Detachment appeared along the Zaragoza road and was met with machine gun fire. With its rear momentarily secure, the battalion retired toward the D-5 line.

Of the 550 men of the 3d Battalion only 156 remained. Many of these were wounded. But the Japanese had been stopped effectively. By delaying Kanno for twenty-four hours, the 3d Battalion had prevented him from reaching Tarlac on 30 December in time to join in the attack on that town. It had thus frustrated a maneuver which might well have turned the left anchor of the North Luzon Force line.

Fight on the West

At the western end of the D-4 line stood the ruined city of Tarlac, its streets a shambles from the repeated strikes of enemy bombers. just south of the city, the 21st Division, as yet untried in battle, awaited the advance of the Japanese. On the gently sloping ground to the west was the 21st Infantry guarding the bridge where Route 13 crossed the Tarlac River. The 22d Infantry, on its right, straddled Route 3. In reserve was the 23d Infantry, eight miles south of Tarlac at Santa Rosa. The terrain, except for the area in which the 21st was deployed, was low and level, consisting largely of rice fields and offering little opportunity for cover. The infantry derived what protection it could from dry cornstalks, bamboo trees, and swamps. The only consolation the rifleman could draw from his position was that he had a clear field of fire.[74]

The Kamijima Detachment, which was assigned the mission of assaulting Tarlac, had shown a curious reluctance to advance below the Agno River. Heavy casualties during the landings had made Colonel Kamijima, in the words of 14th Army Chief of Staff Maeda, "very cautious."[75] Such reluctance might well expose the right


(west) flank of 48th Division, and General Maeda, whose interest in Bataan had led him to emphasize the importance of the advance on Tarlac, took steps to correct the situation. He reprimanded Kamijima for his excessive caution and ordered him to move across the Agno.[76]

By 29 December the Kamijima Detachment had apparently progressed to a point just north of Tarlac. On that day the 3d Battalion, 21st Infantry, reported that it had been fired on by Japanese patrols. The 23d Infantry was ordered to reconnoiter and organize a position along the high ground between Santa Rosa and San Miguel, east of Route 3. At the same time the rear of the 22d Infantry was strengthened. That night the men of the 22d found occasion to open fire against Japanese patrols. Their fire was not returned, and it is possible that the imagination of the men in combat for the first time was responsible for the many Japanese patrols reported south of Tarlac.[77]

Shortly after noon of the 30th, advance elements of the 9th Infantry led by Colonel Kamijima himself entered Tarlac. With only two companies of infantry Kamijima refused to push on. At about 1500 the remainder of the 9th Infantry (less the 3d Battalion) and the two batteries of the 22d Field Artillery reached the area. Thus reinforced, Colonel Kamijima felt strong enough to attack and sent his men against the 22d Infantry positions along Route 3. The defenders held firm, inflicting severe losses on the 9th Infantry and killing Colonel Kamijima himself.[78]

During the course of the action, the 22d Infantry noted a number of men advancing down the road from Tarlac. These men were first thought to be 13th Infantry troops retiring from positions east of the city, but just before they reached the stream in front of the American line they were identified as enemy troops and fired upon. A few minutes later, five American tanks and two SPMs broke out of Tarlac and fought their way down toward the stream. Their retreat had been cut off by Colonel Kanno's advance along the Cabanatuan-Tarlac road, and after much difficulty they had pushed their way through enemy-held Tarlac. The 21st Division troops recognized the tanks and half-track and furnished them with artillery support in their flight to the stream. But here they met an insuperable obstacle and the men had to abandon their vehicles. With the exception of one crew whose tank was hit, all the men reached the 21st Division lines safely. Attempts to rescue the vehicles were unsuccessful and the artillery was ordered to destroy them.[79]

Late in the afternoon the 21st Division received orders to withdraw under cover of darkness to the D-5 line. That evening units began moving out of their D-4 positions. Pressure on the 22d Infantry had died down, but now the 21st Infantry came under heavy attack. As the division pulled back, this regiment supported by the 3d Battalion, 21st Field Artillery, covered the withdrawal alone. During the fight the 21st Infantry received many casualties and was badly battered. Finally, still intact but greatly weakened, the regiment began to pull back. The artillery battalion remained


in position to cover the infantry's withdrawal. Long after its scheduled hour of retirement, the artillerymen, led by their American instructor, 1st Lt. Carl J. Savoie, continued to fire.

To the rear the division covering force waited impatiently and anxiously for the 3d Battalion to pass through its line. When the trucks and guns of the battalion finally came down the road, Colonel Mallonée noted that the men "were tired, worn, hungry--but cocky, proud, aggressive."[80] They had good reason to feel cocky. The battalion, unaided, had held up the Japanese advance and made possible the successful withdrawal of the 21st Infantry.

". . . every man of the 21st Infantry who came out of Tarlac . . . alive should get down on his knees and thank God for that red-headed son of a bitch [Savoie]. He was everywhere he was needed at the right time. . . . He kept the guns in almost three hours after he could have withdrawn to give us a chance to break off. We were all out and the enemy back into Tarlac before he pulled up a gun."[81]

By dawn, 31 December, the 21st Division was on the D-5 line. The 21st Infantry at Bamban, fifteen miles south of Tarlac, was here joined by its 1st Battalion. This battalion had been detached and placed in North Luzon Force reserve earlier and had seen action on the Agno line in the fighting around Carmen. The Japanese 9th Infantry was also reinforced when its 3d Battalion caught up with the rest of the regiment. The enemy force at Tarlac was further strengthened on the 31st by the arrival of the Kanno Detachment and by Lt. Col. Katsumi Takahashi's 8th Field Heavy Artillery Regiment. This greatly increased Japanese force spent the day preparing to push south along Route 3.[82]

In the brief period of seven days, from Christmas Eve to the year's end, there had been a radical change in the situation in northern Luzon. The Japanese, who on 24 December had just secured their beachhead, now threatened Manila and the road net into Bataan. The enemy had broken out of his initial lodgment and was now moving rapidly in two columns down the broad central plain of Luzon.

The North Luzon Force had withdrawn approximately fifty miles from its first defense line to its D-5 positions at Bamban and Arayat. The left and center had retired with moderate success, but the right flank was in grave danger. On that flank, General Homma had placed the main strength of the 48th Division supported by two regiments of tanks and increasing amounts of artillery and other supporting arms. Should the right flank give way, the withdrawal of the South Luzon Force to Bataan might well be imperiled.

The first part of the withdrawal had been completed. Although it had been successful, there had been difficult moments. Communications had broken down at times, supply had proved difficult, and some of the bridges had been blown too soon. The defense lines had sometimes been hastily and inadequately manned, or not occupied at all. "Not a single position," wrote the assistant G-3 of USAFFE, "was really occupied and organized for defense. Troops were barely stopped and assigned defensive sectors


before they stampeded into farther withdrawal, in many instances without firing a shot."[83] This view portrays the withdrawal at its worst. Not all troops stampeded, and there were numerous instances of heroism under fire and determined stands. For the most part, the withdrawal was conducted as well as it could be with the untrained and ill-equipped Philippine Army troops.


Table of Contents ** Previous Chapter (9) * Next Chapter (11)


[1] Quezon, The Good Fight, pp. 194-98. Present at the meeting also were Col. Manuel Nieto, the President's aide, and Lt. Col. Sidney L. Huff, MacArthur's aide.

[2] Mallonée, Bataan Diary, I, 56; see also ltr, Brig Gen Constant L. Irwin, ret., to Ward, 13 Jun 51, OCMH. The author discussed the question of the withdrawal with Generals Sutherland and R.J. Marshall and with Colonel Collier.

[3] Mallonée, Bataan Diary, I, 56. The conversation between Irwin and Mallonée took place in the presence of the senior American instructor of the 21st Division (PA), Col. Ray M. O'Day, the division chief of staff, and several other officers. Colonel O'Day did not get the idea of a withdrawal from Irwin, but did say that such a remark was made in reference to the ammunition. Ltr, O'Day to author, 16 Nov 49, OCMH.

[4] Mallonée, Bataan Diary, I, 57.

[5] Rad, MacArthur to Marshall, No. 3, 22 Dec 41, AG 381 (11-27-41 Gen) Far East. Although MacArthur stated that he had only 40,000 men on Luzon, an analysis of the units present indicates that the number of troops was actually much higher. The strength of the American garrison, even without the air force, could not have been less than 20,000, including the 12,000 Philippine Scouts. To this figure must be added the strength of seven Philippine Army reserve divisions and one regular division, as well as the Constabulary, inducted into the service of the United States in December. Many of the units were undoubtedly at two-thirds strength, but even at half strength, the total number of troops on Luzon at this time could not have been less than 75,000-80,000. The number of Japanese troops who landed at Lingayen between 22 and 28 December was about 43,000.

[6] Ibid.

[7] Rad, Marshall to MacArthur, 22 Dec 41, AG 381 (11-27-41 Gen) Far East. This message implied approval of MacArthur's plans by the President, for it contained the statement that the President had seen all of MacArthur's messages.

[8] Collier, Notebooks, II, 38.

[9] A. Frank Reel, The Case of General Yamashita (Chicago, 1949), pp. 21-22. Most of the Japanese officers who read this volume in manuscript form did not agree with the author that a withdrawal to Bataan by Yamashita would have resulted in a more successful defense. Comments of Former Japanese Officers Regarding the Fall of the Philippines, p. 41, OCMH.

[10] Wainwright, General Wainwright's Story, p. 36. Wainwright received the order on the night of the 23d, but General Parker did not get the order until about 1000 of the 24th, when he was ordered to Bataan to organize the Bataan Defense Force. Ltr, Parker to War, 16 Jan 52, OCMH.

[11] Rad, CG USAFFE to AGWAR, 24 Dec 41, AG 381 (11-27-41 Gen) Far East. MacArthur mistakenly reported that the Japanese were standing off Nasugbu. No landing was ever made there.

[12] Rad, MacArthur to Marshall, 24 Dec 41, AG 381 (11-27-41 Gen) Far East.

[13] Collier, Notebooks, I, 80-81; II, 40.

[14] Ibid., II, 42.

[15] Ibid., 44; rad, MacArthur to AGWAR, 25 Dec 41, AG 381 (11-27-41 Gen) Far East.

[16] USAFFE-USFIP Rpt of Opns, pp. 33, 400; interv, author with R.J. Marshall, 7 Apr 48.

[17] QM Rpt of Opns, pp. 20-21.

[18] Interv, author with R.J. Marshall, 7 Apr 48; Carlos P. Romulo, I Saw the Fall of the Philippines (New York: Doubleday & Company, Inc., 1942), pp. 68-90.

The first specification in the charge against General Homma when he was tried as a war criminal in Manila in 1946 was the violation of an open city. Since Manila was used as a base of supplies, and since a U.S. Army headquarters was based in the city and troops passed through it after 26 December, it is difficult to see how Manila could be considered an open city between 26 and 31 December 1941. Nevertheless, the charge against General Homman stood. USA vs. Homman, specification of charges.

[19] The date was later changed when it was found that North Luzon Force could not hold until the 8th.

San Fernando, Pampanga, should not be confused with San Fernando, La Union.

[20] USAFFE-USFIP Rpt of Opns, p. 33; Collier, Notebooks, II, 47.

[21] Ibid.; ltr, Sutherland to CG 51st Div (PA), 24 Dec 41, sub: Opns Orders, AG 371 Phil Rcds.

[22] USAFFE GO 54, 27 Dec 41; SLF and II Corps Rpt of Opns, pp. 16, 19.

[23] SLF and II Corps Rpt of Opns, p. 19; USAFFE-USFIP Rpt of Opns, pp. 33-35.

[24] Ltr, Maher, formerly NLF CofS, to author, 11 Nov 49, OCMH.

[25] Collier, Notebooks, II, 82; NLF and I Corps Rpt of Opns, p. 11.

[26] Prov Tank Gp Rpt of Opns, pp. 10-11; ltr, Col Ernest B. Miller to Ward, 31 Dec 51, copy in OCMH.

[27] Memo, NLF Engineer for CofS NLF, 1 Aug 42, sub: Orgn and Opns of NLF Engineers 8 Dec41-6 Jan 42, pp. 1-10, copy made for author by Col Harry A. Skerry, the NLF Engineer, and on file in OCMH. This memo will hereafter be cited as Skerry, NLF Engineer Rpt of Opns.

[28] Ltr, Maher to author, 11 Nov 49, OCMH; Collier, Notebooks, II, 82; Mallonée, Bataan Diary, I, 67; NLF and I Corps Rpt of Opns, p. 14.

[29] QM Rpt of Opns, pp. 4-5.

[30] Lt. Col. William E. Chandler, "26th Cavalry (PS) Battles to Glory," Part 2, Armored Cavalry Journal (May-June 1947), p. 11; Skerry, NLF Engineer Rpt of Opns, pp. 3, 5, 9.

[31] Miller, Bataan Uncensored, pp. 89-91; Prov Tank Gp Rpt of Opns, pp. 10-11; 11th Inf (PA), Beach Defense and Delaying Actions, 8 Dec 41-5 Jan42, p. 19, OCMH. The latter is Part 2 of an unsigned and undated account entitled The 11th Infantry Regiment (PA). Part I is entitled Mobilization and Training.

[32] The plan of withdrawal of the 21st Division (PA) can be reconstructed in some detail from personal papers loaned to the author and on file in OCMH. Mallonée, Bataan Diary, I, 52, 67-70; O'Day, 21st Div (PA), II, 10-11; Brief History of 22d Infantry (PA), p. 3; Rpt on Opns of 21st Inf (PA), 7 Dec 41-9 Apr 42, p. 2; Capt Grover C. Richards, Outline of Steps to a POW Camp, pp. 3-4; ltr, [CO, 21st Inf] to TAG PA, 30 Dec 45, sub: Opns of 21st Inf (PA), 7 Dec 41-9 Apr 42, p. 2.

[33] O'Day, 21st Div (PA), II, 11.

[34] 14th Army Opns, I, 52-53; Answers to Questionnaire on Phil Campaign, 5 Aug 49, ATIS Doc 49692; Statement of Col Moriji Kawagoe, CofS 48th Div, in Statements of Japanese Officials on World War II, II, 126-27.

[35] 14th Army Opns, I, 53; Col Glen R. Townsend, CO, 11th Inf (PA), Defense of the Philippines, p. 12, OCMH; O'Day, 21st Div (PA), II, 11.

[36] Maj William E. Chandler, An Outline History of the 26th Cavalry (PS), p. 4, OCMH; see also the same author's article, "26th Cavalry (PS) Battles to Glory," Part 2, Armored Cavalry Journal (May-June 1947), pp. 11-12; ltr, Selleck to Board of Officers, 1 Feb 46, sub: Reinstatement of Rank, p. 9, OCMH; unsigned account of 92d Inf (PA), p. 1, OCMH.

[37] Prov Tank Gp Rpt of Opns, p. 11; Miller, Bataan Uncensored, pp. 89-91; Engineers of the Southwest Pacific 1941-1945, Vol. I, Engineers in Theater Operations (Washington, 1947), p. 6 illustration, and p. 15 n. 24; NLF and I Corps Rpt of Opns, p. 8; 11th Inf (PA) Beach Defense and Delaying Action, p. 19, OCMH; ltr [CO, 21st Inf] to TAG PA, Opns 21st Inf (PA), p. 2.

[38] Rad, CG USAFFE to TAG, 26 Dec 41, AG 381 (11-27-41 Gen) Far East.

[39] 14th Army Opns, I, 53; Chandler, "26th Cavalry (PS) Battles to Glory," p. 12; ltr, Maher to author, 11 Nov 49, OCMH; rads, Maj Lumontod to USAFFE, Nos. 29 and 31, 26 Dec 41, AG 370.2 (19 Dec 41) Phil Rcds.

[40] 14th Army Opns, I, 53; 5th Air Gp Opns, p. 39; Statement of Gen Kawagoe, ATIS Doc 62707, Statements of Japanese Officials on World War II, II, 128; Answers to Questionnaire on Phil Campaign, 5 Aug 49, ATIS Doc 49692; Miller, Bataan Uncensored, p. 95.

[41] 14th Army Opns, I, 53; 11th Inf (PA) Beach Defense and Delaying Action, pp. 19-20, OCMH; 2d Lt Louis I. Bentz, Jr., 82d Inf (PA), p. 1, Chunn Notebooks; O'Day, 21st Div (PA), I, 4; Brougher, Notes on Withdrawal of the 11th Infantry from Agno River Position, pp. 1-2, copy in OCMH.

[42] Miller, Bataan Uncensored, p. 97.

[43] Prov Tank Gp Rpt of Opns, p. 11; Miller, Bataan Uncensored, pp. 97-98; ltr, Miller to War, 31 Dec 51, OCMH; Weaver, Comments on Draft MS, Comment 13, OCMH. Miller claims he was given no authroity to withdraw earlier if necessary. He also claims that no holding orders were issued to he 192d. Weaver stated that his orders were given to both battalion commanders. The 192d has prepared no reports and efforts by the author to secure additional material on tis question from the battalion commander have been unsuccessful. Colonel Miller intimates that General Weaver was keeping the tanks in a "rat trap." There is no direct evidence other than Colonel Miller's published statements to substantiate such a view.

[44] Prov Tank Gp Rpt of Opns, p. 12; ltr, Weaver to Wainwright, 20 Nov 45, OCMH; Miller, Bataan Uncensored, p. 98.,

[45] Miller, Bataan Uncensored, pp. 97-102; GO 10, Hq Prov Tank Gp in the Field, 14 Feb 42, copy given the author by General Weaver and on file in OCMH. Prov Tank Gp Rpt of Opns, p. 12; ltr, Miller to Ward, 31 Dec 51, OCMH; Weaver, Comments on Draft MS, Comment 14, OCMH.

[46] Miller, Bataan Uncensored, pp. 102-03; GO's 5 and 10, Hq Prov Tank Gp in the FIeld, 13 Jan and 14 Feb 42.

[47] Miller, Bataan Uncensored, pp. 103-104, 108-09; Prov Tank Gp Rpt of Opns, p. 12; Capt Jack C. Altman, 194th Tank Bn, p. 3, Chunn Notebooks; Weaver, Comments on Draft MS, Comment 15, OCMH. During an interrogation at the end of the war General Homa stated that the 9th Infantry had reported the capture of twenty-three tanks stranded north of a river above Paniqui after the bridge had been blown. Interv, Col Walter E. Buchly with Homma, Manila, Mar 46, copy in OCMH.

There was investigation of the blowing of the bridge and the loss of the tanks in prison camps in Formosa and Manchuria in 1944 adn 1945. It was then shown that Company D had reached the Moncada bridge fifteen hours after the last tanks of its battalion headquarters and fourteen hours after the last infantry elements had crossed. No tank guides from either Company A or Battalion Headquarters, 194th Tank Battalion, had been left behind to direct the withdrawal of Company D. Comments of Col Skerry on Draft MS, Comment 4, OCMH.

[48] Bentz, 92d Inf (PA), pp. 1-2, Chunn Notebooks.

[49] O'Day, 21st Div (PA), II, 12; Mallonée, Bataan Diary, I, 82; Miller, Bataan Uncensored, p. 107; Weaver, Comments on Draft MS, Comment 16, OCMH.

[50] 14th Army Opns, I, 50, 53-54.

[51] Col Alexander, Narrative Rpt of QM Activities at Fort Stotsenburg, pp. 1-2, App. A to QM Rpt of Opns.

[52] Ibid., pp. 2-3; Col Alexander, Personal Recollections of Bataan and Later, pp. 48-50, copy in OCMH.

[53] Collier, Notebooks, II, 84; II, 2. See also ltr, Lt Col John E. Olson to author, 10 Jan 52, OCMH. In his comments on this manuscript, Colonel Alexander insisted that the supplies had been evacuated and that nothing was left behind. Ltr, Alexander to Ward, 25 Dec 51, OCMH.

[54] QM Rpt of Opns, p. 22; see also Mallonée, Bataan Diary, I, 108; O'Day, 21st Div, I, 12-13; Drake, Comments on Draft MS, Comment 8, OCMH.

[55] Col Lawrence, Tarlac Advance QM Depot Rpt of Opns, p. 6, App. A to QM Rpt of Opns.

[56] Ibid.

[57] QM Rpt of Opns, p. 22.

[58] Lawrence, Tarlac Advance QM Depot Rpt of Opns, pp. 7-8, App. A to QM Rpt of Opns.

[59] QM Rpt of Opns.

[60] This order is quoted in Mallonée, Bataan Diary, I, 90. See also O'Day, 21st Div (PA), II, 12. The author has been unable to find the original of this order. It is entirely possible that the change in mssion of North Luzon Force was initiated by MacArthur's headquarters.

[61] NLF and II Corps Rpt of Opns, p. 12.

[62] Prov Tank Gp Rpt of Opns, p. 13; Miller, Bataan Uncensored, p. 111.

[63] Rad, CG USAFFE to TAG, 28 Dec 41, AG 381 (11-27-41 Gen) Far East.

[64] 14th Army Opns, I, 55-56.

[65] Interrog of Col Nakajima, 25 Apr 47, Interrogations of Former Japanese Officers, I, Mil Hist Div, GHQ FEC.

[66] Interrog of Col Motoo Nakayama, Apr 47, ibid.

[67] Ibid.; 14th Army Opns, I, 56-57, 61-64.

[68] 14th Army Opns, I, 60-61, 83; Statement of Gen Kawagoe, ATIS Doc 62707, Statements of Japanese Officials on World War II, pp. 128-30.

[69] Bentz, 92d Inf (PA), p. 2, Chunn Notebooks; Skerry, Comments on Draft MS, Comment 6, OCMH. The 192d Tank Battalion was in the area, but there is no information on its position or employment at this time.

[70] The American sources for this action are sketchy and vague as to times and places. Where doubts exist or where the records are irreconcilable, the Japanese account of the action has been used. 14th Army Opns, I, 63, 83. Ltr, Luther R. Stevens to Capt Edwin B. Kerr, 30 Dec 52; ltr, Col John H. Rodman to Maj Gen Orlando War, 1 Feb 52, both in OCMH.

[71] Interv, author with Col Volckmann, May 48; Maj W.J. Lage, Opns of 3d Bn, 11th Inf (PA) at Zaragoza, 28-29 Dec 41 (paper prepared for Advanced Infantry Officers Course, 1947-48, The Infantry School); Weaver, Comments on Draft MS, Comments 18 and 19, OCMH. Major Lage's account is one day off, and the author had made necessary corrections.

[72] 14th Army Opns, I, 83.

[73] Lage, Opns of 3d Bn, 11th Inf (PA), p. 16. Another explanation of these events is to be found in the report of the 11th Engineer Battalion, a portion of which is quoted in Skerry, Comments on Draft MS, Comment 10, OCMH.

[74] O'Day, 21st Div (PA), II, 12-13; ltr, [CO, 21st Inf] to TAG PA, Opns of 21st Inf (PA), p. 2; Brief Hist of 22d Inf (PA), p. 3; Hist of 21st Div (PA), pp. 18-19. This last document was obtained from General Capinpin and like the others is on file in OCMH.

[75] Interrog of Gen Maeda, 10 May 48, Interrogations of Former Japanese Officers, Mil Hist Div, GHQ FEC, I.

[76] Ibid.; 14th Army Opns, I, 60-61.

[77] O'Day, 21st Div (PA), II, 13; Richards, Steps to a POW Camp, p. 6.

[78] 14th Army Opns, I, 63; Interrog of Gen Maeda, 10 May 47, Interrogations of Former Japanese Officers, Mil Hist Div, GHQ FEC, I; USA vs. Homma, p. 3055; Mallonée, Diary, I, 100; Brief Hist of 22d Inf (PA), p. 3.

[79] Mallonée, Bataan Diary, I, 101-02; O'Day, 21st Div (PA), II, 13-14; separate, unsigned one-page history, entitled 194th Tank Battalion, OCMH; Miller, Bataan Uncensored,, p. 111. The identity of the tanks is not clear in the sources.

[80] Mallonée, Bataan Diary, I, 105; O'Day, 21st Div (PA), II, 14.

[81] Mallonée, Bataan Diary, I, 105, quoting Lt Col. William A. Wappenstein, CO, 21st Infantry.

[82] 14th Army Opns, I, 64.

[83] Collier, Notebooks, II, 83.

Transcribed and formatted for HTML by Patrick Clancey, HyperWar Foundation