While northern Peleliu and the offshore islands were being secured, pressure had been maintained against the principal pocket of Japanese resistance on Umurbrogol Mountain, as that section of the central ridges lying south of 321st Infantry Trail was designated on the maps supplied to the Allied forces fighting in the Palaus. After the 321st Infantry had divided the island, the Umurbrogol Pocket was about 1,900 yards long north to south on its eastern side; approximately 1,200 yards long on the west; and, on the average, 550 yards wide east to west. No offensive operations were undertaken against the pocket from 22 September--the date of 7th Marines' last strong efforts--until the 27th of the month, when elements of the 321st Infantry began attacks on the north side. On the latter date the pocket was being contained on the south, southwest, and west by Marine units; on the northwest, north, and northeast by Army troops. There were no troops on the east side, where the central ridges fell steeply to East Road, itself bounded on the east by a dense swamp which lay along the shores of the shallow bay separating Peleliu's eastern and western arms.

Compressing the Umurbrogol Pocket

The Terrain

Umurbrogol Mountain was no mountain in the usual sense of the word.¹ (Map 21) It was, rather, a chaotic jumble of steep coral ridges, the tallest peak among which was about 300 feet high. There were innumerable caves throughout the pocket, ranging from large caverns with small entrances to open, shallow shelters big enough for only one man. Some caves extended through ridges; these and many others had two or more entrances, permitting the Japanese to move from one position to another without appearing above ground.

As left by nature, the Umurbrogol Pocket was much like the Ibdi Pocket area on Biak Island, but larger and rougher. Like the Ibdi Pocket, the Umurbrogol originally had a thick cover of tropical trees and dense jungle undergrowth which, as the result of continued air, naval, and artillery bombardment (including extensive employment

Map 21
Terrain of Umurbrogol Pocket

of aerial napalm strikes), was gradually knocked down or burned away. Again, as on Biak, the Japanese had improved upon nature. There were many artificial or semi-artificial caves which had been constructed to protect approaches to the inner sections of the pocket, and the Japanese had improved almost every natural cave. Where no caves were available or could be constructed, the enemy employed rock faults and crevices for defensive positions. Digging new entrances to existing caves or even cutting new levels within some caves, the Japanese were well prepared to execute a long and bloody holding action along the many ridges. These ridges, with the exception of the Five Sisters group at the southern side of the pocket, were generally parallel and oriented north-northeast to south-southwest. Steep-sided and fissured, many of them had razor-back summits upon which no cover could be found. The ridges were separated by deep draws, gullies, and wider valleys, the floors of which were strewn with coral boulders or coral outcroppings similar to stalagmites. Steep as they were, the sides of some ridges also were covered with such chunks and outcroppings.

At the southeast corner of the pocket, paralleling East Road, lay Walt Ridge; to the north, across a 70-yard-wide draw, was Boyd Ridge,² beyond which lay an unnamed, broken ridge line extending north to 321st Infantry Trail. Lying west of Boyd and Walt Ridges was Mortimer Valley,³ the western wall of which was named Five Brothers Ridge. West of the Five Brothers was Wildcat Bowl,⁴ bounded on the west by China Wall, on the southwest by the Five Sisters, and on the southeast by a 225-foot-high peak designated Old Baldy.⁵ West of Five Sisters and China Wall lay a narrow defile known as Death Valley, beyond which was a coral plateau comprising broken ridgelets, sink holes, and jumbled coral formations overlooking West Road.

At the northern end of the pocket area the ridges were closer together, often less well defined, narrower, and sharper, and generally more broken. Above (north) of China Wall lay Wattie and Baldy Ridges, on or near which elements of the 3d Battalion, 321st Infantry, had already fought. Wattie Ridge abutted the cliffs above West Road, Baldy Ridge lay east of Wattie, and, east of Baldy, across a narrow gorge from the north end of Boyd Ridge, was Ridge 120.⁶ North from the line Wattie-Baldy-Ridge 120 lay 700 yards of unnamed ridgelets, peaks, ravines, and deep draws extending to 321st Infantry Trail. From the trail, the 321st Infantry started moving south against the Umurbrogol Pocket on 27 September.

The 321st Infantry at the Umurbrogol, 27-29 September

After its important tactical success in dividing Peleliu at 321st Infantry Trail, the 321st Infantry was ordered to send one battalion to northern Peleliu and to employ the rest of the regiment in a drive south from the trail into the northern end of the

SOUTHEAST CORNER OF UMURBROGOL POCKET. East road and Walt Ridge in right foreground. Five Brothers Ridge in center, between Mortimer Valley and Wildcat Bowl. China Wall, abutting Five Sisters, forms the western rim of Wildcat Bowl.

Umurbrogol Pocket.⁷ On the evening of 26 September the 1st Battalion, 321st Infantry, was designated for the push north. At the same time the 2d Battalion (less Company G, but with Company A, 710th Tank Battalion, and Company K, 321st Infantry, attached) was ordered to move south from 321st Infantry Trail to secure the ground to the X-ray phase line. (See Map IX.) This new line, lying roughly midway between the third and fourth phase lines, was drawn northwest to southeast across the central ridges south of 321st Infantry Trail and marked what was believed to be the northern edge of the inner core of Japanese defenses. While the 2d Battalion attacked south toward the X-ray phase line, the 3d Battalion (less Company K but plus Company G) was to hold at Hill 100 and at the area around the north end of Wattie and Baldy Ridges.

The 2d Battalion, 321st Infantry, started south at 0700 on 27 September, with Company K advancing on the right along the ridges above East Road. Company F, with the still-organized Neal Task Force of tanks, LVT's, and infantry in support, pushed south from Hill B along East Road. Company E remained on Hill B, while Company H provided fire support from positions near the same hill. In Company K's sector, extremely

broken terrain was the chief obstacle to progress during the morning (although plenty of small arms fire was received) and by noon the unit had moved only 200 yards south from 321st Infantry Trail. Company F, keeping pace, was harassed by Japanese small arms fire from positions to its right front beyond Company K.

During the afternoon, Japanese fire on Company K increased and the unit suffered numerous casualties, including the company commander. Harassing fire from Japanese on the east slopes of the central ridges continued against Company F. Direct fire from tanks, which used many white phosphorus shells, and fire from a flame thrower mounted on an LVT of the Neal Task Force, helped to reduce many of the enemy positions, but others were inaccessible to such fire or to infantry assault from the low ground along East Road. By late afternoon the intensity of fire on both companies had not appreciably diminished.

Company K suffered so many casualties during the afternoon, and the forward positions of that unit and Company F seemed so dangerously exposed, that a withdrawal for the night was ordered. Company K managed to push west and north over the central ridges to Company I's positions on the ridge lines at the southwest base of Hill 100. Company F and the Neal Task Force moved back to the base of Hill B and set up defenses along East Road.

Early the next morning, 28 September, Company K was withdrawn from the forward combat area because of losses and fatigue. The company then moved to Hill B, where it replaced Company E. To continue the attack south a composite unit, under the commander of the 3d Battalion, was organized around Companies E, F, and I⁸ of the 321st Infantry, and Company A, 710th Tank Battalion. The objective for the day was to secure and organize for defense all the ground between 321st Infantry Trail and the X-ray phase line.

Company I pushed south along the ridges and Company E, with tanks in support, advanced down East Road. Both units quickly resecured the ground which had been taken the previous day, and by noon advance elements of the two companies were within 200 yards of the X-ray phase line. Company L, on the west side of the ridges and north of Baldy and Wattie Ridges, joined the advance in midafternoon, swinging south in contact with Company I's right. Systematically attacking and cleaning out each Japanese cave or crevice position found, the three companies reached the X-ray phase line late in the afternoon. This time there was no withdrawal, and the units set up night defenses on the ridges or along East Road.

On the morning of the 29th, the 2d and 3d Battalions, 321st Infantry, were relieved of further responsibility for operations against the Umurbrogol Pocket and moved north to replace the 5th Marines at northern Peleliu and on the offshore islands. The 7th Marines took over along the X-ray phase line.

Marines Versus the Umurbrogol Pocket 29 September-15 October

When the 321st Infantry's elements left the pocket, it was estimated that the Japanese had been compressed into an area averaging

400 yards in width and 900 yards in length.⁹ This estimate was somewhat misleading and, in regard to the width, actually in error. On the east the pocket was some 1,000 yards long from the X-ray phase line south to the East Road causeway north of Asias, where the Weapons Company of the 7th Marines was holding containing lines. On the west, along the ridges between the western ends of the X-ray and third phase lines, elements of the 7th Marines and other troops from Marine artillery, amphibian tractor, engineer, and pioneer units were holding a line about 750 yards long. The distance between the Weapons Company north-northwest across the ridges to the third phase line was over 900 yards, a distance along which the 2d Battalion, 7th Marines, held containing lines.¹⁰ On the northern side, across the ridges at the X-ray phase line, the pocket was only 400 yards wide. A few hundred yards farther south a line drawn with the same northwest to southeast orientation as the X-ray would have been about 650 yards wide.

General Rupertus planned to have the 1st and 3d Battalions, 7th Marines, attack south from the X-ray phase line, while other Marine units on the western and southern sides of the pocket held their positions. Company A, 710th Tank Battalion, was made available to the 7th Marines for this new push because the Marines' own 1st Tank Battalion had already left the island. The new attack began at 0800 on 30 September.(Map 22)

From 30 September through 2 October, Japanese opposition, heavy rain, fog, and sickness, prevented the 1st and 3d Battalions, 7th Marines, from making much progress south. The 1st Battalion, down to 90 riflemen fit for duty, had to be relieved after securing an area 150 yards wide and 300 long on East Road and flanking ridges below the X-ray phase line. On 3 October the 2d Battalion, 7th Marines, began new attacks from the south, securing a foothold along the eastern side and top of Walt Ridge, with the help of mediums of the 710th Tank Battalion. Other tanks of the same unit, cooperating with men of the Weapons Company, 7th Marines, probed into the southern end of Mortimer Valley, but had to retire before dark. On the same day the 3d Battalion, 7th Marines, brought partially up to strength by men pulled out of Marine engineer and amphibian tractor units, seized the eastern slope of Boyd Ridge, north of Walt Ridge.

The next day, 4 October, elements of the 3d Battalion, 7th Marines, secured a temporary foothold on the northern end of Ridge 120, but were forced to withdraw through a draw lying between Boyd Ridge and the unnamed ridges to the north. At the end of the day Companies I and L, 7th Marines, were down to a combined strength of about 80 men from an authorized number of 470. The 3d Battalion, 7th Marines, was through; the 1st Battalion could muster few more than 100 men fit for duty; and the 2d Battalion reported in at 30 percent of efficiency. Like the 1st Marines before it, the 7th Marines was finished as a regiment, and on 5 October the 5th Marines began moving into positions to relieve the 7th.

There was little action on the 5th or 6th other than some attacks into ground which

Map 22
Marines at Umurbrogol Pocket
30 September-15 October 1944

could not be held at Baldy Ridge and Ridge 120. On the 7th,¹¹ six tanks of Company A, 710th Tank Battalion, together with most of the infantrymen of the 3d Battalion, 5th Marines, moved into Mortimer Valley behind an hour-long artillery and mortar barrage. At 0700, four of the six tanks pushed up into the west side of the valley's mouth. Two were hit but not severely damaged by Japanese antitank fire from a position impossible to find. The four then moved out of the valley, south of a water-filled sink hole at the mouth, and, with the remaining two, reapproached the mouth from the east, behind the cover of the south end of Walt Ridge. Moving back into the mouth about 1030, the lead tank struck a mine and lost a track. The next three passed around the damaged vehicle, firing 75-mm. ammunition at all likely enemy positions. About 100 yards into the valley, another tank was hit by a Japanese antitank gun located at the western base of Walt Ridge and was forced to retire to repair its jammed 75-mm. gun. A third tank moved forward and received three hits, none of which caused much damage. It was then decided to pull the tanks back out so that artillery could fire on Five Brothers and the western base of Walt Ridge. In withdrawing, four tanks received minor damage.

About 1400, six tanks with two companies of the 3d Battalion, 5th Marines, attempted a second push. Company L moved northwest into Wildcat Bowl and Company I started into Mortimer Valley. No penetration could be made at Wildcat Bowl, although Company I managed to get about 200 yards into Mortimer Valley. This time, when the tanks were stopped by the troublesome antitank gun at the base of Walt Ridge, the gun was discovered and destroyed. As the tanks began running out of ammunition during the late afternoon, the infantry positions in Mortimer Valley became untenable, and under cover of smoke which the tanks' guns provided, withdrawal was made to the morning line of departure. It was ten days before American troops again attempted to operate in the valley.¹²

No other offensive action was undertaken on 7 or 8 October, but pressure was maintained against the Japanese pocket by continued artillery fire and aerial bombardment, which now began to destroy so much foliage and undergrowth that visibility along the various ridges was greatly increased. On the 9th the attack was resumed on the northern side of Wattie and Baldy Ridges, where no firm foothold could be secured. Patrols of the 2d Battalion, 5th Marines, probed into a deep gulch between Wattie and Baldy Ridges, finding there the bodies of twelve 321st Infantry soldiers, grim reminders of the previous efforts of the 3d Battalion, 321st Infantry, in the same area during the period of 25-27 September.¹³

On 10 October elements of the 2d Battalion, 5th Marines, secured Wattie, Baldy, and 120 Ridges and pushed south the next day to gain the top of tactically important

Hill 140, lying immediately north of the northern end of Five Brothers Ridge. On the 12th a Marine 75-mm. pack howitzer was dragged up Hill 140 to fire on Five Brothers. Sandbag fortifications were erected for the protection of the weapon and its crew.

Exactly who should be credited with carrying the first sandbag into the Peleliu ridges had never been clearly established, but he started something which was to become increasingly important . . . . The lack of cover and impossibility of digging-in had repeatedly obliged attacking troops to relinquish hard-won gains as untenable. Obviously, the sandbag provided an answer of a sort, and many were in use at this time . . . so long as the operation remained essentially one of movement, the problem of sandbagging successive positions . . . presented formidable difficulties. Nowhere was there any sand inland . . . the heavy bags had to be carried into position already filled, no small undertaking in that crazily upended country. It remained for the 81st Division, following relief of the Marines, to develop a technique with such ingenious refinements as to make the sandbag into something closely resembling an offensive weapon, in which capacity it played a crucial part in the final reduction of the Pocket.¹⁴

On 13 and 14 October, elements of the 5th and 7th Marine Regiments secured much additional ground on the western side of the central pocket between the third and X-ray phase lines, terrain so rough and broken that it was deemed useless as a route of attack into the Umurbrogol. Two days later these elements were pulled out, having cleared an area some 700 yards long north to south and as much as 200 yards east beyond the earlier containing lines above West Road. Men from Marine artillery, engineer, and amphibian tractor units moved up from the old lines to occupy the additional yardage.

Entr'acte: The Relief of the 1st Marine Division

From 12 through 15 October, when elements of the 321st Infantry began moving up to relieve Marine units, there was little action at the Umurbrogol Pocket beyond the clearing of the ridges overlooking West Road. To the 15th, the 321st Infantry had spent its time mopping up at northern Peleliu and on the offshore islands already seized, and had taken a few more islets off Ngesebus and Kongauru. At northern Peleliu, debris was cleared, defenses were constructed, roads and trails improved, and enemy dead buried. Mopping up was a continuing process. During the period 4-8 October, for instance, the 321st Infantry killed 171 Japanese in the north or on the offshore islands, losing 8 men killed, 30 wounded, and 1 missing. Many of the Japanese killed during this period appeared to be stragglers who were attempting to escape from the Umurbrogol Pocket.¹⁵

About mid-October, a number of command changes occurred in the Palaus area. On the 12th, for instance, the 1st Marine Division was relieved of all responsibilities other than continuing the fight in the Umurbrogol Pocket. The 321st Infantry took over the defenses along the eastern arm, while the Island Garrison Force assumed responsibility for the area south of the pocket. On the

morning of 12 October, the command post of the III Amphibious Corps moved ashore and General Geiger, the corps commander, declared that the assault and occupation phase of operations on Peleliu was ended.¹⁶

The exact meaning of this announcement is not clear, especially in relation to Admiral Fort's somewhat similar declaration of 30 September, stating that Angaur, Peleliu, Ngesebus, and Kongauru had been captured and occupied. However, General Geiger's announcement seems to have been made in preparation for the 1st Marine Division's imminent departure from the Palaus. The statement also bears relation to the passing of control of operations in the Palaus from the U.S. Third Fleet and Admiral Halsey (as then represented by Admiral Fort's Western Attack Force headquarters) to the Headquarters, Forward Area Central Pacific (Task Force 57), under Admiral Hoover. On the 13th General Geiger issued orders alerting the 321st Regimental Combat Team to relieve the 1st Marine Division elements still at the Umurbrogol Pocket. The next day, control over all operations in the Palaus passed from Admiral Fort to Admiral Hoover's command.¹⁷

On 12 October, in accordance with General Geiger's orders, the 321st Regimental Combat Team began occupying new defenses on Peleliu's eastern arm. These movements were hardly well started when, on the 14th, the combat team had to begin execution of the corps commander's new orders to relieve Marine elements still at the Umurbrogol Pocket. During the morning of the 14th, the 2d Battalion, 321st Infantry, withdrew from Ngesebus and other offshore islands to move to an assembly area near Hill 100 on 321st Infantry Trail. Simultaneously the 1st Battalion, 323d Infantry, fresh from unopposed operations at Ulithi Atoll, began preparations to relieve Marine units at the southwest corner of the pocket.¹⁸

Actual relief of Marine units began on 15 October, when the 2d Battalion, 321st Infantry, took over the lines held by the 5th Marines across the northern side of the pocket and on the salient at Hill 140. The next day the 1st Battalion, 323d Infantry, took over the containing lines on the western and southwestern sides of the pocket while the 3d Battalion, 321st Infantry, relieved the Marine units on Walt and Boyd Ridges, along the eastern edge. The relief of Marine elements in central Peleliu was almost complete, and control of operations at the Umurbrogol passed to Colonel Dark, the commander of the 321st Regimental Combat Team.¹⁹

With the relief of marines at the Umurbrogol, the remainder of the battered 1st Marine Division began preparations for leaving the Palaus. The 7th Marines left for rest and rehabilitation in the Solomons on 22 October. The 5th Marines, attached

operationally to the 81st Infantry Division, remained in defensive positions in northern Peleliu and offshore islands until 27 October, when it began loading. The regiment left on the 30th.²⁰ By 18 October, when the last marines had been relieved from their Umurbrogol Pocket positions, the 1st Marine Division and attached units had suffered 6,526 casualties--1,252 killed (including perhaps 115 missing, presumed dead) and 5,274 wounded.²¹

Overcoming the Final Resistance

Of the 1st Marine Division's total casualties, about 1,000 were incurred as the 5th and 7th Marine Regiments compressed the Umurbrogol Pocket during the period 29 September through 16 October. The action of the two regiments succeeded in reducing the pocket to a rough oval averaging about 400 yards east to west and about 850 yards north to south.²² Some 2,000 Japanese had been defending the pocket when the marines began their final offensive, and estimates of enemy strength remaining in the pocket after the marines left varied from 300 to 2,000. From Japanese sources, it appears that the marines killed some 850-1,000 Japanese during the period 29 September-16 October, leaving approximately 1,000 Japanese alive in the pocket as the 321st Infantry moved back in.²³

The 321st Infantry to 20 October

Offensive operations against the Japanese remaining in the Umurbrogol Pocket were resumed about 1500 on 16 October, when elements of the 2d Battalion, 321st Infantry, made an attempt to move from their new positions on Hill 140 across a deep draw to Brother No. 1, the northernmost of the five peaks on Five Brothers Ridge.²⁴(Map 23) The intent was to secure a position for supporting weapons and observation in preparation for a general attack south by the 1st and 2d Battalions, 321st Infantry, on the morning of 17 October. The plan for the latter day was to have the 1st Battalion push south along the broken ridges west of Hill 140 while the 2d Battalion seized Five Brothers. The rest of the reinforced regiment was to hold its positions and provide fire support whenever possible.

At midafternoon on the 16th, one platoon of Company G, 321st Infantry, went down into a deep ravine between Hill 140 and Brother No. 1. The platoon expected to receive

Map 23
321st Infantry at Umurbrogol Pocket
16-25 October 1944

fire support for its attack up Brother No. 1 from Company F, which was located on a ridge immediately west of Hill 140, but once in the ravine the platoon was met by strong enemy machine gun and rifle fire originating at positions in defilade from Company F's supporting weapons. The Company G platoon soon suffered casualties which included the company commander, who had gone forward to direct the attack. The platoon had to be withdrawn, leaving behind some dead and wounded who could not be brought out until after dark and then only by individual acts of heroism. During the night Japanese troops, either foraging for food or trying to harass the American lines, were active all along the Umurbrogol front. In retaliation and in preparation for the next morning's attack, American artillery and 4.2-inch mortars²⁵ placed intermittent fire on known or suspected enemy positions and also employed area fire.

The 1st Battalion, 321st Infantry, started south from its previous positions behind the 2d Battalion at 0700 on the 17th. Resistance during the morning was limited to scattered rifle fire, but the terrain proved almost impassable and by noon the leading elements had gained little more than 100 yards. During the early afternoon, after moving about 50 yards farther south, Company A was stopped by fire from a pillbox and near-by cave positions located on the second of seven more or less parallel but extremely broken ridges which lay between Hill 140 and West Road. Company B was now committed on the first ridge (to A's left) in an attempt to provide fire support for A's attack. But as it moved forward, Company B's rear and flanks were exposed to fire from enemy rifles and heavier weapons on Five Brothers Ridge. Further progress in the 1st Battalion zone proved impossible, and late in the afternoon both companies withdrew about fifty yards to covered positions. The day's efforts had gained about 125 yards in a southerly direction. On the other hand, about 150 yards of new ground had been taken east to west in the broken terrain west of Hill 140.

The 2d Battalion's attack against Five Brothers was led by Company E, but again no advance across the ravine from Hill 140 to Brother No. 1 could be made. When this assault stalled, other elements of the 2d Battalion, supported by 710th Tank Battalion armor, entered Mortimer Valley through a draw between Walt and Boyd Ridges in an effort to reduce Brothers Nos. 1 and 2 by attack from the east. Delays incident to bulldozing a tank road through the draw prevented significant progress from the new direction, but tanks and flame throwers, late in the afternoon, neutralized some caves along the western base of Walt Ridge and on the east side of Brother No. 1. It was estimated that these efforts accounted for about forty Japanese.

Action against Five Brothers was resumed at 0900 the next morning, 18 October. Tanks and LVT-mounted flame throwers again pushed through the gap between Walt and Boyd Ridges to reduce caves along the east side of the Brothers and the west side of Walt Ridge. Meanwhile, a heavy concentration of 4.2-inch and 81-mm. mortar fire was placed along the top of Five Brothers Ridge by a platoon of Company D, 88th Chemical Weapons Battalion, and the Provisional Mortar Platoon, 710th Tank Battalion. Company E followed the mortar fire southward and at 1000

LVT FLAME THROWER IN ACTION at Umurbrogol Pocket.

reached the top of Brother No. 1. About forty-five minutes later, part of the company scrambled up the summit of Brother No. 2, seventy-five yards to the south.

This success, achieved with unexpected ease, made it possible to send Company F forward to pass through Company E and seize Brother No. 3. By 1315 Company F had gained the northern slopes of the third peak, but was almost immediately pinned down by enemy fire which broke out from positions on Brothers Nos. 4 and 5, to the south and southwest, and from the southern base of Walt Ridge, to the southeast. At approximately 1500 hours about two platoons of Japanese infantry began a well-organized counterattack from Brothers Nos. 4 and 5. Company F had no cover on the razorback summit of Brother No. 3 and was subjected to increasingly heavy fire from enemy supporting weapons. Company E, on Brothers Nos. 1 and 2, was in like straits. Neither company had yet had time to bring forward sandbags, without the protection of which it was impossible to hold ground gained on the now bare slopes of the Umurbrogol Pocket's broken ridges and gorges. Constant artillery and mortar bombardment,

as well as high explosives and napalm strikes by planes of Marine Air Group II, based on the Peleliu airdrome, had knocked down, burned away, or otherwise destroyed almost the last vestiges of vegetation in the pocket area. There was no cover or concealment other than that which could be provided by sandbags.

Reinforcements sent forward from the 1st Battalion could not reach Companies E and F in time to be of help, for by 1600 the Japanese counterattack was fully developed. Within an hour Companies E and F were forced off the three peaks, unable to hold their exposed positions against the combination of infantry assault and fire from Japanese supporting weapons. Most of the men withdrew northward to set up defenses at the base of Brother No. 1, but some, cut off, slid down the eastern slopes of the peaks and made their way across Mortimer Valley and out through the draw between Walt and Boyd Ridges. There, the men were provided with covering fire from tanks and elements of the 3d Battalion, 321st Infantry.

Operations had not been much more successful in the 1st Battalion's zone, where strong enemy resistance was again encountered by Companies A and C along the rough ground west of Hill 140. Both units discovered many occupied caves and enemy fortifications of varied types, including one large cave boasting steel doors. The latter position, once an enemy communications center, was neutralized and overrun during the afternoon, but no other appreciable gains could be made. Most of the day was spent constructing sandbag fortifications and emplacing 75-mm. pack howitzers for direct fire on the many enemy defensive installations in the battalion area.²⁶

The Japanese, even before their counterattack against Five Brothers, had sent infiltraters into action on the southwest side of the pocket, chiefly by means of tunnels and connecting caves, to points some 300 yards outside the recently established containing lines of the 1st Battalion, 323d Infantry.²⁷ Late on the afternoon of the 17th, such infiltraters had forced mortar units in position behind the lines of the 1st Battalion, 323d Infantry, to withdraw. Company I, 7th Marines, which had not yet left the island,²⁸ was sent into the area to mop up, meeting stronger resistance than had been anticipated. The next morning Company L, 7th Marines, supported by one tank of the 710th Tank Battalion, resumed mopping up. The tank hit an aerial bomb buried as a mine and was destroyed by ensuing fires and explosions. The Marine company commander, who had been riding in the tank, was killed, as were three members of the tank crew. Two men escaped. Company L, 7th Marines, withdrew at dusk.

The next morning, 19 October, a company of the Army's 154th Engineer Battalion, supported by artillery of the 4th Battalion, 11th Marines, moved back into the area to blow up or seal remaining caves. Opposition was still strong and on the morning

of the 20th, Company C, 323d Infantry, had to be withdrawn from its containing lines near Five Sisters to support the engineer demolition teams. Tanks were again moved into the area, now known as the Southern Pocket, but mopping up was still no easy task. During the night of 22-23 October, the remaining Japanese even forced one platoon of Company C, 323d Infantry, back about 100 yards. Company E, 323d Infantry, which had recently arrived on Peleliu, was thereupon sent up from defenses along West Road to reinforce Company C. The next morning the two rifle companies pushed most of the remaining Japanese north toward Death Valley and China Wall, but final mopping up at the Southern Pocket was not completed until 3 November, after the line of attack had been changed from north to south. Altogether, over 100 Japanese were killed in the Southern Pocket, at the cost of 22 Army and Marine Corps troops killed and 51 wounded. While the action had no significant effect upon operations farther north, it is indicative of the Japanese ability to infiltrate over (and literally through) the Umurbrogol's broken terrain and to hold reoccupied positions against great odds.

While this minor operation had been going on, pressure had been maintained against the Umurbrogol from the north. The 19th, the day after the 2d Battalion, 321st Infantry, had been driven off Five Brothers Ridge, was marked principally by patrolling in preparation for new attacks, while Marine air units undertook napalm strikes against Five Brothers and Wildcat Bowl, to the west. The strike was principally effective in driving into the open a few Japanese who, running out of overheated caves, were summarily shot down by alert 321st Infantry riflemen. The 1st Battalion, 321st Infantry, gained 25-50 yards along the ridges west of Hill 140 with the aid of 75-mm. pack howitzer fire and the now omnipresent sandbags. The battalion's men, prone on the ground, inched their sandbags forward with rifle butts or sticks, laboriously expanding a hold and almost realizing the infantrymen's dream of portable foxholes.

Thus far, operations of the 321st Infantry and attached units had been conducted under General Rupertus' direction, but on the morning of 20 October responsibility for all ground operations in the Palaus passed to General Mueller of the 81st Division. The next higher echelon was the Western Carolines Sub-Area, commanded by Rear Adm. John W. Reeves, the local representative of Admiral Hoover's Forward Area Central Pacific.

The first day of operations under the control of the 81st Infantry Division was spent principally in reconnaissance, and on the morning of 21 October, after another napalm strike, the 1st Battalion, 321st Infantry, resumed the attack. Company A, on the right, gained almost 100 yards along its front during the morning, progressing farther toward the north end of China Wall and the northwest corner of Wildcat Bowl. No progress was made by Company C on the left (east), but the unit maintained contact with 2d Battalion elements at Hill 140. During the afternoon all elements of the 1st Battalion were held up by intense enemy fire from Brother No. 3 and ridges at the north end of China Wall. In the 2d Battalion area a volunteer patrol fought its way back up to the summit of Brother No. 1, practically by infiltration, late in the afternoon. Company E rushed forward to consolidate this hold and a human chain was formed to pass sandbags up the hill. By dark, sandbag fortifications were completed and preparations were

under way to attack south again on the next day.

The plan of attack for 22 October was ambitious. The 1st Battalion (less Company B, which had replaced Company C, 323d Infantry, on containing lines at the southwest corner of the pocket) was to continue forward to seize additional high ground northwest of Wildcat Bowl. The 2d Battalion was to seize all of Five Brothers Ridge. The 3d Battalion was to take all remaining enemy positions on the west side of Walt Ridge and was to occupy Mortimer Valley. The 1st Battalion, 323d Infantry (less Company C, but with Company B, 321st Infantry, attached), was to push from the south and west against Five Sisters and Death Valley. In preparation for this attack, all available mortars plastered Japanese-held areas from 0515 to 0530, and from 0545 to 0615 Marine air units delivered another napalm strike against Wildcat Bowl and the southern end of Five Brothers. At 0645 ground action began.

By 0845 the 2d Battalion, 321st Infantry, had secured all of Brothers Nos. 1 and 2; fifteen minutes later one platoon was atop No. 3. During the rest of the day, sandbag fortifications were built on this new ground, but the expectation of securing the entire ridge was not realized. To the west, the 1st Battalion could make little progress in a series of bazooka and rifle grenade attacks against numerous Japanese cave and crevice positions. Company A, 323d Infantry, moved north over broken terrain west of Death Valley in an attempt to outflank or take from the rear some of the enemy positions holding up the 321st Infantry's 1st Battalion. Although the two 1st Battalions managed to establish contact in the rough ground, their combined efforts failed to gain much ground west of Death Valley. During the night a Japanese counterattack from the north end of China Wall forced the left (north) platoon of Company A, 323d Infantry, back to West Road, creating a gap between the two 1st Battalions' lines. Fortunately, the Japanese made no attempt to exploit this success and at dawn on 23 October Company A was able to reoccupy its abandoned positions without difficulty.

More spectacular advances had been made by the 3d Battalion, 321st Infantry. Company I, supported by two platoons of tanks, three M-10 tank destroyers, and two LVT's mounting flame throwers, pushed into Mortimer Valley via the gap between Walt and Boyd Ridges. Delayed but a little while by Japanese who fought from holes and fissures around Grinlinton Pond, at the base of Brother No. 4, the infantry-armor force completed a sweep down the valley by 1100. Caves and crevices along the bases of Five Brothers and Walt Ridges were blasted, and then attention was devoted to caves on the northeast slopes of Five Sisters. At least thirty-four of the enemy were known to have been killed by these operations, and it was believed that many more were sealed in caves. At dusk, with the aid of sandbags, Company I formed a defense line along the western base of Walt Ridge, while Company L, previously holding the eastern slopes, moved over the crest to the western side. Company K remained on Boyd Ridge. On the south, no permanent gains could be made by elements of the 1st Battalion, 323d Infantry, operating in the Five Sisters-Death Valley area.

The Japanese, still strong enough to react vigorously to the 321st Infantry's advances, attempted to infiltrate through the lines Company I had set up on the floor of Mortimer Valley, probably in an effort to reoccupy caves on Walt Ridge's western

MOVING INTO MORTIMER VALLEY. Grinlinton Pond, right foreground; Walt Ridge, right background; Five Brothers, left background.

base. These efforts were unsuccessful, as was an attack against Company F, on Brother No. 2. The latter attack cost the Japanese at least 20 men killed.

On 23 October the only significant advance was made by Company E, 321st Infantry, which proceeded south along Five Brothers Ridge to secure Brother No. 4 and fortify it with sandbag emplacements. In the 3d Battalion's zone, redispositions .were effected to close the southern entrance to Mortimer Valley, but in the 1st Battalion's area progress was negligible and resistance apparently undiminished. On the 24th, the 1st Battalion gained 50-75 yards, laboriously inching forward its sandbags; the 3d Battalion continued sealing caves along both sides of Mortimer Valley; and the 2d Battalion consolidated its gains on Five Brothers Ridge. The next day, the 323d Infantry, all of which had now arrived from Ulithi, began relieving the 1st and 2d Battalions, 321st Infantry. On 26 October control of further operations against the Umurbrogol Pocket passed from Colonel Dark to Col. Arthur P. Watson, the commander

of the 323d Regimental Combat Team. To that date the 321st Infantry had lost 146 men killed and 469 wounded on Peleliu.²⁹

The 323d Infantry Finishes the Job

Company I, 323d Infantry, replaced those elements of the 2d Battalion, 321st Infantry, that were holding along Five Brothers Ridge, while Companies K and L took over the sandbagged positions of the 1st Battalion, 321st Infantry, at the northwest corner of the Umurbrogol Pocket.³⁰(Map 24) The 2d Battalion, 323d Infantry, went into lines south of Five Sisters and the 1st Battalion held its positions at the southwestern and western sides of the pocket. The 3d Battalion, 321st Infantry, attached operationally to the 323d, remained in its sandbagged emplacements along Walt Ridge and the floor of Mortimer Valley.

As the 323d Infantry took over, Japanese strength in the Umurbrogol Pocket was estimated at "anywhere between 300 and 1200,"³¹ but according to Japanese sources, there were only 700 effective troops still alive, including those lightly wounded. The 321st Infantry had probably killed about 400 Japanese in the pocket during the previous ten days. The pocket was now reduced to an average north-south length of approximately 600 yards. On the north it was still about 475 yards wide, but deep salients had been driven south into Mortimer Valley and along Five Brothers Ridge. On the east Japanese resistance was limited to a few caves along the eastern base of the Five Brothers. At the south, across the mouth of Mortimer Valley to the west side of the entrance to Death Valley, the width was less than 350 yards.³² The gains made by the 321st Infantry and the attached 1st Battalion, 323d Infantry, had cost those units 392 casualties--74 men killed, 315 wounded, and 3 missing.³³ There was an unknown number of battle fatigue and sickness cases.

During the period 26 October through 1 November heavy rains, fog, and accompanying poor visibility severely limited the 323d Infantry's operations at the Umurbrogol Pocket. The time was employed principally to improve defenses, extend sandbag fortifications, and to enclose with sandbags the 75-mm. pack howitzers for support of later offensive action. Throughout the period mortar barrages and napalm strikes continued, but little artillery could be brought to bear because of the proximity of the American front lines to the remaining Japanese positions.

The Japanese made almost nightly counterattacks during the six days, aiming most of their efforts against Five Brothers. One such attack, on the night of 26-27 October, forced Company I, 323d Infantry, off the forward slopes of Brother No. 4, but the company retook the positions at dawn the next day and the effort cost the Japanese

Map 24
323d Infantry at Umurbrogol Pocket
26 October-27 November 1944

34 men killed. During the next night an estimated 35 Japanese were killed by Company K on the ridges near the northwest corner of Wildcat Bowl. The enemy was obviously short of water, for each night a few more men were killed as they attempted to get water from Grinlinton Pond, around which first Company I and later Company L, both of the 321st Infantry, held a sandbag perimeter.

More active operations were resumed on 2 November when the 2d Battalion, 323d Infantry, attacked Five Sisters from the southwest; elements of the 1st Battalion moved north along the ridges west of Death Valley, pursuing the last Japanese from South Pocket; and part of the 3d Battalion started south along ridges and broken terrain at the north end of China Wall. The latter's attack gained only 25-50 yards as Company L sandbagged its way forward literally inches at a time. The 2d Battalion met with more success and established troops atop Old Baldy and the Five Sisters peaks. The 1st Battalion's Company C got nowhere in the Death Valley region. From 3 to 12 November no significant gains were made by any unit, as slow patrolling and sandbagging continued. Heavy rains began to fall on the 4th, culminating in a typhoon which did not blow itself out until the 8th. During this period the 81st Cavalry Reconnaissance Troop and elements of the 321st Infantry secured additional offshore islets north of Peleliu.

Attacks at the Umurbrogol started in earnest again on 13 November, with the 1st Battalion, 323d Infantry, pushing eastward from the ridges west of Death Valley and the 2d Battalion moving north into that gorge and Wildcat Bowl. The 1st Battalion's attack gained little ground, and that only by sandbagging almost every foot of the terrain. In the 2d Battalion's zone Company G moved about 75 yards north along the more easterly of the two razor-back ridges which formed the China Wall, but no other advances were made. No gains of any significance were made on the 14th or 15th: but on the latter day there was discovered in the 1st Battalion's area what appeared to be an important center of resistance in rough, broken ground west of the center of Death Valley. On the 16th, oil was sent forward through a hose from fuel tanks set up in covered positions 300 yards distant and poured into a large cave which seemed to be the center of the new-found defenses. Ignited by white phosphorus hand grenades lobbed into the cave, the flaming oil produced such promising results in driving Japanese into the open or killing them, that the same method of conquest continued to be used by the 1st Battalion and in other parts of the pocket.

During the period 16-21 November patrolling continued in all areas, with tanks and LVT-flame throwers operating throughout Wildcat Bowl and Death Valley, attacking enemy caves at the bases of China Wall and Five Brothers. Engineer flame throwers and demolition teams, sometimes accompanied by armored bulldozers, followed the tanks and LVT's closely, destroying or sealing all enemy caves which could be reached. The tank-LVT sorties prompted many of the Japanese remaining in the pocket to make desperate efforts to escape and during the night of 17-18 November, for instance, about 33 were killed trying to flee. On the next day, according to enemy sources, there were only 150 Japanese still capable of fighting left in the Umurbrogol Pocket.

By noon on 21 November infantry patrols were able to operate without encountering

much resistance throughout Wildcat Bowl and the southern half of Death Valley; the last Japanese on Five Brothers Ridge were killed on the 23d. The remaining Japanese held final defenses on China Wall and on the broken ridges west of the northern half of Death Valley. By dusk on 23 November further resistance was confined to the China Wall, still a nearly impregnable position. Steep sides, both east and west, made it impossible for wheeled or tracked vehicles to gain access to China Wall from Death Valley or Wildcat Bowl, and the terrain at the north end of China Wall was so broken as to be impassable for armor. A scooped-out area between the two sides of China Wall gave the remaining Japanese protection from direct fire, and the proximity of the 323d Infantry's forward elements to the enemy positions now made it too dangerous to use mortar fire against the hollow.

On 22 November one company had ascended the north end of China Wall; another had closed in an additional 75 yards from the west-northwest; and a third had gained 25-50 yards at the southern end of the Wall. By the end of the day the remaining Japanese were compressed into an area 125 yards wide east to west and 285 yards long north to south. There were no gains on the 23d, but the next day a few more yards were made on all fronts. In order that tanks and LVT-flame throwers could reach the central hollow of China Wall, engineers began constructing a ramp up the east wall at the north end of Wildcat Bowl. This ramp was completed on the afternoon of the 25th, and the 323d Infantry was ready to make a final attack.

Unknown to the 81st Division, the 57 Japanese still remaining alive in the pocket had given up the fight on the 24th and had burned the 2d Infantry's colors, traditional sign of final defeat. The 57 survivors were split into 17 small teams which were instructed to hide out during the day and raid American lines at night. A last message was sent to General Inoue, the Palau Sector Group commander, still on Koror Island, advising him of the end. The next day, 25 November, a prisoner reported that Colonel Nakagawa and General Murai had committed suicide the previous night. Subsequently, the 81st Division established positive identification of the bodies.

The 323d Infantry had noted on the afternoon of the 24th that resistance had almost completely vanished, and rapid gains were obviously possible on the next morning. But the troops advanced cautiously, carefully searching every foot of ground, for they expected perhaps a trap or at least a last, desperate, banzai charge by the remaining Japanese. On the 26th, tanks and LVT-flame throwers moved up the completed ramp to begin blasting away at caves and other defenses along China Wall's hollow center. Again gains were rapid, but the 323d's men, realizing the fight was over and not wishing to take unnecessary risks, were prevented by scattered rifle fire from clearing the last tiny pocket of resistance before dark. The next morning the attack was continued by eight rifle companies which closed in from all directions. There was no resistance--not a shot was fired on either side. At 1100 Colonel Watson declared that hostilities were over. The operation which General Rupertus, commanding the 1st Marine Division, had expected to last only four days,³⁴ had actually continued for almost two and one-half months. In the words of the 81st Infantry Division, "The enemy had fulfilled his determination to fight unto death."³⁵ That fight

had cost the 323d Infantry 118 men killed and 420 wounded.³⁶

Results of Operations in the Palaus

In the Palaus--on Angaur, Peleliu, and the smaller islands off Peleliu--approximately 13,600 Japanese were killed, over 11,000 of them on Peleliu alone. In addition, some 400 enemy prisoners were taken, chiefly Korean and Okinawan troops. During its operations the 81st Infantry Division and attached units suffered over 3,275 battle casualties--542 killed and 2,736 wounded or injured in action. The reinforced 1st Marine Division's battle casualties were approximately 1,250 killed (including about 115 missing) and 5,275 wounded, a total of about 6,525. Thus, Army and Marine Corps troops killed or captured approximately 14,000 Japanese during the Palaus operation while incurring around 9,800 battle casualties themselves. Many more American troops were rendered temporarily or perhaps permanently hors de combat through nonbattle injuries, battle fatigue, heat exhaustion, sickness, or disease. The 81st Division, for instance, had about 2,500 men hospitalized for these reasons.³⁷

The Palaus operation was extremely costly--it was one of the bloodiest battles of the war. It was so costly, in fact, that one wonders if the results were worth the effort. Doubts are easily raised in the light of the fact that eleventh-hour changes in plans for subsequent operations--notably the invasion of the Philippines--made it impossible to fit the Palaus into the operational role originally planned for them.³⁸

Air support for the Philippine landings as originally planned was to have been provided from a heavy bomber field on Angaur Island. Shorter-range planes based on Peleliu were to fly missions against the Palaus themselves and also against other Japanese-held islands in the western Carolines in order to keep Japanese air and naval bases in that area neutralized. The Angaur field was to have been completed on 15 October, on which date Army B-24's, of the Seventh Air Force's 494th Bomb Group (H), were to have begun strikes against targets in the Philippines. But as a result of terrain difficulties and inadequate gasoline storage facilities, the Angaur field was not ready for use by bombers until 21 October, the day after American forces had landed on Leyte, in the central Philippines. Then it was found that the 494th Bomb Group needed more training (which it acquired by flying missions over the Palaus and Yap) before it could carry out combat missions over the Philippines. In the end, it was not until 17 November that the first bomber mission was flown against a Philippine target from a field in the Palaus.³⁹

The Angaur field did become a useful heavy bomber base and aircraft staging point for later phases of the Philippine campaign, especially during operations on Luzon.⁴⁰

PELELIU, showing base development. At the time photograph was taken, jungle growth had again covered the ridges of Umurbrogol Pocket. (Aerial view, October 1946.)

Employment of the Palaus as a fleet base was negligible, although Kossol Passage at the northern end of the group was successfully used as an emergency fleet anchorage. Ulithi Atoll, seized by the 323d Regimental Combat Team of the 81st Infantry Division in an unopposed operation during 22-24 September,⁴¹ proved an extremely valuable base for the U.S. Pacific Fleet during various phases of operations in the Philippines. Later in the war Ulithi also served as a staging area for fleet and amphibious forces moving against Okinawa.⁴² Thus, operations in the western Carolines, including the Palaus, did secure for the Allies valuable air and naval bases. The Palaus were not as valuable as had been anticipated nor were the islands used as extensively or for quite the same purposes as originally planned.

There remains the possibility of conjecture as to whether the Philippines could have been invaded without the Palaus in Allied hands. Among the highest-ranking Allied commanders in the Pacific, only Admiral Halsey, commander of the U.S. Third Fleet, believed it safe to bypass the Palaus, and he expressed such an opinion as early as mid-June.⁴³ From the vantage point of hindsight, it would appear that Admiral Halsey was right and that it may not have been absolutely necessary to take the Palaus. Nevertheless, with the information available to them in the summer and early fall of 1944, Admiral Nimitz, General MacArthur, and the Joint Chiefs of Staff all believed that only by securing the Palaus could the Allies dominate Japanese bases in the western Pacific and insure the safety of forces moving toward the Philippines.⁴⁴

Footnotes

1. This subsection is based on: Incl 1, Terrain and Int Sum, pp. 1-2, to 323d Inf Opns Rpt Peleliu; Sketch maps and sketches included in 81st Inf Div Opns Rpt Peleliu; George McMillan, The Old Breed. A History of The First Marine Division in World War II (Washington, 1949), sketch map p. 34; Hough, The Assault on Peleliu, pp. 136-40, 149-52, 159-61, 194-97.

The terrain configuration shown on Map 21 gives only a general impression of the extremely twisted and broken terrain of the area. The relief was taken from the best map available, one prepared in 1945 from aerial photographs. Some of the terrain features that appear on hasty maps and sketches drawn by participating units cannot be readily identified on this map. Nor do the units' maps and sketches themselves agree on the terrain configuration of the area. Compare with Maps 22, 23, and 24.

2. Names varied on Peleliu according to the unit operation. The names in this chapter generally follow the usage employed by the 81st Division.

3. So called by the 321st Infantry. The 1st Marine Division knew it as the Horseshoe or Horseshoe Valley; Marine air units called it Big Slot; the 323d Infantry called it East Valley.

4. Known to the 323d Infantry as Main Valley and to Marine air as Little Slot.

5. Known to the Marines as Hill 300 and considered by them a part of Five Sisters.

6. Also known as Hill 120, a terrain feature which formed the bulk of the ridge.

7. This subsection is based principally on: 321st RCT Jnls and Jnl files, Peleliu; 81st Div Opns Rpt Peleliu, pp. 25-33.

8. The 321st Infantry seems to have had a penchant for splitting up its infantry battalions, a practice which it had previously undertaken extensively on Angaur. Apparently the Neal Task Force was disbanded on the evening of the 27th or on the 28th.

9. Narrative, p. 13, of Phase II, Opns Annex, to 1st Mar Div Opns Rpt Palau. The remainder of this section is based principally on: Hough, The Assault on Peleliu, pp. 143-67; 710th Tank Bn Opns Rpt Peleliu, pp. 11-12.

10. The battalion had been holding these lines since 21 September, limiting its activities to patrolling. When it came out of these lines on 2 October, it had suffered about 150 casualties, including 33 men killed, during that period of comparative inactivity. Hough, The Assault on Peleliu, p. 144.

11. The action described here as taking place on 7 October is placed on 8 October by the 710th Tank Battalion's report, but the regimental and battalion records of the 5th Marines set it on the 7th. The tank unit's report was written in December (none of its journals can be found), whereas the Marine records, which were checked, were maintained during the action.

12. According to the Marine Corps' account, the sortie into Mortimer Valley had not been made to establish a permanent foothold but merely to reduce fire from Japanese guns which were harassing the airdrome and the lines of the 3d Battalion, 5th Marines. The report of the 710th Tank Battalion states that after the day's action, the marines decided that Mortimer Valley could not be taken from the south, and that no further effort was to be made from that direction.

13. See above, Ch. XXIII.

14. Hough, The Assault on Peleliu, pp. 164-65. The 321st Infantry installed an overhead cable system powered by a 2½-ton truck. This system, later expanded and improved by the 323d Infantry, was employed to move sandbags, ammunition, and supplies up ridges and across ravines to front-line units, and was also used to move down to West Road (by means of Stokes litters which the 81st Division borrowed from naval forces) many of the more severely wounded men, whose lives were probably saved by this relatively rapid means of evacuation.

15. 81st Inf Div Opns Rpt Peleliu, pp. 39, 41-45.

16. Ibid., p. 47; III Amph Corps Opn Plan 12-44, 10 Oct 44, in Encl L, Opns Orders, to III Amph Corps Opns Rpt Palaus; Encl B, Opns, p. 11, to III Amph Corps Opns Rpt Palaus.

17. III Amph Corps Opns Order 13-44, 13 Oct 44, in Encl L, Opns Orders, to III Amph Corps Opns Rpt Palaus; CTF 32 Opns Rpt Palaus, pp. 95-96; Ltr, Adm Fort to Gen Ward, 15 Nov 50, no sub, in OCMH files. On 29 September command in the western Carolines had passed from Admiral Wilkinson's Joint Expeditionary Force (Task Force 31) to Admiral Fort's Western Attack Force (Task Force 32), and Admiral Wilkinson was assigned a new command for the invasion of the Philippines. On 20 October responsibility for ground operations in the Palaus passed from the III Amphibious Corps to the 81st Division, and General Geiger and his staff moved off, ultimately to take part in the invasion of Okinawa.

18. 81st Inf Div Opns Rpt Peleliu, p. 47: 321st RCT Unit Rpts 21 and 22, 14 and 15 Oct 44, respectively, in 321st RCT Jnl file, Peleliu.

19. 81st Inf Div Opns Rpt Peleliu, pp. 49-51: Hough, The Assault on Peleliu, p. 167.

20. Narrative, pp. 19-21, of Phase II, Opn Annex, to 1st Mar Div Opns Rpt Palaus; Hough, The Assault on Peleliu, pp. 168-70; 81st Inf Div Opns Rpt Peleliu, pp. 49-57, 69-71.

21. Hough, op. cit., p. 183; Narrative, pp. 19-20, of Phase II, Opns Annex, to 1st Mar Div Opns Rpt Palaus. The 1st Marines lost 1,749 men, the 5th Marines 1,378, and the 7th Marines 1,497, a total of 4,624 casualties in the three infantry regiments. These figures are from the 1st Marine Division's report and are cited in Hough, op. cit., page 183.

22. 321st RCT 8-2 and 8-3 Overlays, in 321st RCT Jnl file Peleliu. Hough, The Assault on Peleliu, p. 166, n. 73, has it that the pocket was reduced to an area about 400 by 500 yards, an estimate which appears small on the basis of the evidence in the 321st RCT's records. The 81st Infantry Wildcat Division in World War II, p. 161, states that the pocket was about 750 by 500 yards in size on 15 October, but this estimate has been ripped out of context from an estimate made by the S-2 of the 323d Infantry on 26 October, as contained in Incl 1, Terrain and Int Summary, p. 1, to 323d Inf Opns Rpt Peleliu.

23. The Palau Operations, pp. 142, 261. This Japanese source states that there were about 1,150 Japanese left on Peleliu as of 13 October and that as of the 23d there were 700 effectives left, including slightly wounded men.

24. This subsection is based on: 321st RCT Jnls and Jnl files, Peleliu; 81st Inf Div Opns Rpt Peleliu, pp. 49-69; The 81st Infantry Wildcat Division in World War II, pp. 162-79.

25. The mortars were from Company D, 88th Chemical Weapons Battalion, a platoon of which had arrived from Angaur during the afternoon and had been emplaced on the southwest side of the pocket near the lines of the 1st Battalion, 323d Infantry.

26. The 75-mm. pack howitzers had been turned over to Army units by the 1st Marine Division. Many of the weapons had been left on the island more by accident than design. The men of the 11th Marines' two pack howitzer battalions had left Peleliu, but heavy seas had made it impossible to load their weapons. Hough, The Assault on Peleliu, p. 161, n. 62. Other Marine 75's had been taken over in place by Army troops at the request of the 81st Division.

27. The following story of action at the Southern Pocket, as the area was designated, was reconstructed from: Incl II, 1st Bn Rpt, 90 Oct-3 Nov 44, n. p., to 393d Inf Opns Rpt Peleliu; Hough, The Assault on Peleliu. p. 169.

28. Operations from 17-20 October were still under the control of Headquarters, 1st Marine Division, and the 3d Battalion, 7th Marines, was division reserve.

29. Casualty figures from Ltr, Gen Mueller to Gen Ward, 29 May 51.

30. This subsection is based on: Sec. II, Narrative, pp. 3-37, to 323d Inf Opns Rpt Peleliu; Incl 9, Unit Jnl, to 323d Inf Opns Rpt Peleliu; Incl 10, Periodic Rpts, to 323d Inf Opns Rpt Peleliu; 81st Inf Div Opns Rpt Peleliu, pp. 71-97; The 81st Infantry Wildcat Division in World War II, pp. 177-200; The Palau Operations, pp. 161, 198-200.

31. Incl 1, Terrain and Int Sum, p. 1, to 323d Inf Opns Rpt Peleliu.

32. These figures, based on information provided by the overlays of the 321st and 323d Infantry Regiments, do not agree with the dimensions given in the 323d Infantry's Terrain and Intelligence Summary.

33. 321st RCT Unit Rpt 23, 16 Oct 44, as compared with 321st RCT Unit Rpt 33, 26 Oct 44, both in 321st RCT Jnl file Peleliu.

34. Hough, The Assault on Peleliu, p. 35.

35. 81st Inf Div Opns Rpt Peleliu, p. 97.

36. Casualty figures from Ltr, Mueller to Ward, 29 May 51.

37. These statistics are reconstructed from numerous mutually irreconcilable and contradictory sources including, among others, the G-1 and G-3 Jnls and Jnl files of the 81st Inf Div; the Rpts and Jnls and Jnl files of the 321st, 322d, and 323d Inf Regts and RCT's; the 81st Inf Div's Opns Rpt; the 1st Mar Div's Opns Rpt; the III Amph Corps Opns Rpt; The 81st Infantry Wildcat Division in World War II; Hough, The Assault on Peleliu; Ltr, Mueller to Ward, 29 May 51.

38. See above, Chs. XIX and XX.

39. MS, The Pacific--MATTERHORN to Nagasaki, Ch. IX, pp. 30, 48: Ch. I, p. 74.

40. Ibid., Ch. IX, p. 48.

41. 81st Inf Div Opns Rpt, Ulithi Atoll and Other Western Caroline Islands, 21 Sep 44-4 Jan 45, pp. 11-23.

42. Roy E. Applemen, and others, Okinawa: The Last Battle, in U.S. ARMY IN WORLD WAR II (Washington, 1949), p. 43.

43. Fleet Admiral William F. Halsey and Lt Comdr J, Bryan, III, Admiral Halsey's Story (New York, 1947), pp. 194-95. This is not hindsight on Admiral Halsey's part, since he cites his War Diary for 18 June 1944 to prove his point. Admiral Halsey based his objections to the Palaus operation at least partially on a belief that the invasion would result in prohibitive casualties. Other commanders, notably Maj. Gen. John R. Hodge, commanding the XXIV Corps, apparently early expressed the belief that it would be possible to bypass Yap, at least, well before the Yap operation was canceled during the eleventh-hour changes in the concept for the final steps of the approach to the Philippines. The opinion concerning Yap was possibly shared by one of the Joint Chiefs of Staff's subordinate committees, the Joint Staff Planners, well before September 1944.

44. The 81st Infantry Division, with the exception of the 321st Regimental Combat Team, left the Palaus for New Caledonia in December 1944 and January 1945. The 321st RCT was finally relieved by the 111th Infantry (a separate regiment not part of any division) and left for New Caledonia early in February 1945. In addition to operations in the Palaus and Ulithi, elements of the 81st Division seized Ngulu Atoll, between Yap and the Palaus, on 16-17 October 1944; Pulo Anna Island in the Sonsoral Group, between the Palaus and Morotai, on 20 November; Kayangel Atoll, north of Kossol Passage in the Palaus, on 28 November-1 December; and Fais Island, southeast of Ulithi, on 1-4 January 1945. On all these islands radar or loran stations were set up. See 81st Inf Div Opns Rpt, Ulithi and Other Caroline Islands, 21 Sep 44-4 Jan 45, pp. vi, 33-78.