Chapter 6
Securing the North: 23-30 September1


Major General William H. Rupertus, commanding the 1st Marine Division, initially envisioned the Peleliu operation as a tough but short campaign. The landings took place on 15 September under weather conditions that were ideal except for excessive heat. Despite fanatical Japanese resistance, the Marines secured a firm foothold on the island. The 5th Marines seized the prime objective of the operation, the vital Peleliu airfield, on the second day of the assault, while the 7th Marines accomplished the mission of driving the Japanese from the southern part of the island. The 1st Marines had already encountered the fringes of the vast enemy defense system in the central ridges of Peleliu, and early optimistic reports soon gave way to a more somber perspective of the situation. During the first week of operations the 1st Marines had borne the brunt of the assault. There was little likelihood that the campaign would become easier or less costly as it progressed.

Even the most pessimistic predictions of the difficulties that the Marines were to encounter proved conservative when the full extent of the Japanese defensive system was revealed to the assault troops. By evening of 20 September, having made only minor gains, the 1st Marines found progress blocked by ridges honeycombed with elaborate Japanese defenses consisting of layers of caves, dugouts, and cleverly concealed emplacements. Worse still, in five days of incessant fighting, the regiment had sustained nearly 1,700 casualties or more than half its original strength. The 5th and 7th Marines had suffered fewer casualties, but their advance was also stymied by the heavily fortified ridges. To most division personnel, there appeared the sobering realization that the attack had bogged down in the incredibly tough and skillfully defended terrain. The division paid an extremely heavy price of nearly 4,000 casualties for the ground that had been seized. The heavy losses in personnel resulted in a corresponding reduction in the combat efficiency of the division, which as early as 18 September had dropped from "excellent" to "very good." By evening of 19 September, continued casualties and fatigue further reduced the efficiency of the 1st Marine Division.

On 21 September, General Geiger, accompanied by members of his staff, visited


the CP of the 1st Marines to obtain a clearer picture of the situation.

It became rapidly apparent that the regimental commander was very tired, he was unable to give a very clear picture of what his situation was and when asked by the Corps Commander what he needed in the way of help he stated that he was doing all right with what he had.3

General Geiger and staff then proceeded to the division CP. After a look at the casualty reports, General Geiger told the division commander that, in his estimation, "the First Marines were finished."4 In the course of the ensuing discussion with General Rupertus, General Geiger expressed the view that the 1st Marines should be relieved and replaced by an Army regiment. The division commander attempted to forestall such action by asserting that the island could be secured in another day or two without the employment of Army troops, a patently impossible solution in view of the overall tactical situation. In the end, General Geiger ordered preparations made for the evacuation of the 1st Marines to the Russells and for the immediate attachment of a regiment of the 81st Infantry Division to the 1st Marine Division.

Reluctance on the part of the Commanding General, 1st Marine Division, to use Army units was not limited to this instance. It may have been rooted in earlier experiences, which did not always result in harmonious relations with Army commanders.5 On the other hand, General Rupertus may have distrusted a division that was new to combat and felt that the battle-hardened 1st Marine Division was capable of finishing the job it had undertaken without any help.

Prior to this latter incident [involving attachment of an infantry regiment to the 1st Marine Division], "the Corps Commander was disinclined to impose any particular line of action upon the division commander although more than once he had felt the urge to do so. Just what induced this reluctance on the part of General Rupertus was never understood by Corps. . . ." It is probable that he felt, like most Marines, that he and his troops could and would handle any task assigned to them without asking for outside help.6

At 1625, 21 September, IIIAC asked the Commanding General, 81st Infantry Division, if he could make a regimental combat team available for immediate movement to Peleliu to assist the 1st Marine Division in completing the seizure of the island. Within the hour General Mueller replied that the 321st Infantry Regiment was available as soon as it could complete its re-supply. Shortly thereafter, a group consisting of Rear Admiral George H. Fort, commander of the Western Attack Force, and Major Generals Julian C. Smith and Geiger arrived at division headquarters to confer with General Mueller. These four discussed the situation on Peleliu in detail and agreed on a general plan of movement for the 321st Infantry Regiment. Shortly before midnight, 21


September, the reinforced regiment received orders directing its loading and movement to Peleliu.

The Assistant Division Commander of the 81st Infantry Division, Brigadier General Marcus B. Bell, was designated liaison representative to IIIAC, to coordinate details of the movement from Angaur to Peleliu and attachment of RCT 321 to the 1st Marine Division. On 22 September an advance detachment of the regiment arrived at Headquarters, 1st Marine Division, near the airfield, to complete the necessary arrangements for the reception and disposition of troops. On the same date, the Commanding Officer, 321st Infantry Regiment, Colonel Robert F. Dark, reported to General Geiger on board the USS Mt. McKinley and then proceeded to the 1st Marine Division command post on Peleliu.

The main body of the 321st Infantry Regiment embarked early on 22 September. Embarkation of all troops and loading of equipment, except for amphibian tractors, continued throughout the day and was finally completed by 1630. The ships stood off Angaur during the night of 22-23 September. Shortly after daybreak the amphibian tractors were loaded into LSTs. The force departed from Angaur at 1000 and two hours later reached ORANGE Beach, Peleliu. Attached to the regiment were an engineer battalion less one company, two medical companies less two platoons, an amphibian tractor company, and Company A, 710th Tank Battalion, an 81mm Provisional Mortar Platoon, and several detachments of service troops. Upon landing at Peleliu, the first elements of RCT 321 to reach the island proceeded to the western shore, where 2/321 relieved remnants of 3/1 between 1400 and 1500.

No advance for the 1st Marines had been ordered for 23 September. Patrols pushed 1,000 yards northward along the west coast without encountering serious opposition. The areas directly behind the Marine lines were harassed by mortar and sniper fire from the high ridges to the east. In a last skirmish prior to its relief, Company L, 3d Battalion, was able to inflict heavy casualties on a large body of Japanese moving up a draw across the front towards Company I. The enemy force was wiped out by combined artillery and mortar fire.7

The relief of the remainder of the 1st Marines was effected quickly. All maps, overlays, and other information pertaining to the terrain and enemy positions were turned over to the Army unit, which had no maps of the area.8 The exhausted and depleted Marine regiment then moved out on foot or with such transportation as was available to a rest area on PURPLE Beach. There the units were reorganized and assigned defense sectors, which included the eastern coast of Peleliu and three islands offshore.

The men had hardly settled along PURPLE Beach before Colonel Puller informed them that they would go back into action after a three-day rest.9 This was not the intention of the division, however, and the regiment retained its defensive mission until departure from Peleliu on the last day of September. Since coming ashore, the 1st Marines


Men and supplies arrive across 500-yard reef off Peleliu
MEN AND SUPPLIES arrived across 500-yard reef off Peleliu. (USMC 95606)

81st Infantry Division troops join the battle of Peleliu
81ST INFANTRY DIVISION troops join the battle of Peleliu. (USMC A96738)


had suffered 1,672 casualties. In the 1st Battalion 71 percent of total strength had become casualties. Similarly, the casualty rate reached 56 percent in the 2d, 55 percent in the 3d, and 32 percent of authorized strength in regimental headquarters and weapons Company.10

During the remainder of its stay on Peleliu, the depleted 1st Marines in its rest area was able to account for a few snipers and a number of Japanese stragglers attempting to flee through the shallow water from the large peninsula across the bay. In this manner the regiment remained in action against the enemy until the very last day of its stay on the island.

In relieving the 1st Marines near Ngarekeukl on the West Road, the 3d Battalion, 321st Infantry Regiment, moved into positions directly south of the 2d Battalion, which had occupied the lines previously held by 3/1. The 1st Battalion occupied an assembly area to the rear, where it was kept in regimental reserve.


The arrival of an additional regiment on Peleliu ushered in a new phase in the bitter contest for the island. Even though, in the estimate of the 1st Marine Division, about two thirds of the original Japanese garrison had been put out of action, organized resistance was far from over. Driving the Japanese defenders from the Umurbrogol ridges, as well as the remainder of the central ridge system, promised to be a costly and time-consuming task. From their positions in and on the hills, the Japanese were able to interdict all movement along the East and West Roads, which generally paralleled the island coastlines north of the airfield. Just north of Peleliu, the island of Ngesebus and its fighter strip were still in Japanese hands. Enemy strength in the remaining Palau islands still exceeded 25,000. As a result, the possibility of an enemy countermanding was ever present. It had become abundantly clear that failure of the 1st Marine Division to score a breakthrough had resulted in a loss of momentum, which would have to be recovered before the Japanese hold on the island could be eliminated.

General Rupertus had for some time considered the possibility of launching a drive up the west coast of Peleliu to outflank the main Japanese defenses on the island. If successful, such an operation would open the dual possibility of attacking the Japanese center of resistance simultaneously from the north and south, and of crossing over a narrow body of water to Ngesebus to seize that island and its fighter strip.

In the minds of General Rupertus and his staff uncertainty prevailed as to Japanese intentions of sending reinforcements and materiel to Peleliu. Evidence on that score was conflicting. As early as 18 September two Japanese barges and one sampan had been observed unloading on the northwest coast of Peleliu. On the other hand, during the first week following the landing of the 1st Marine Division on Peleliu, repeated air searches of the islands to the north had shown no evidence of enemy movements to reinforce Peleliu.


There was no uncertainty on Babelthuap, for the Japanese were ready to reinforce their hard-pressed garrison on Peleliu. This mission fell to the 2d Battalion, 15th Infantry Regiment. From a military standpoint, General Inoue was not enthusiastic about sending this unit, because he believed that the Americans would attack Koror and Babelthuap once the operation on Peleliu had been concluded. His reason for reinforcing Peleliu was "for the sake of the garrison at Peleliu and to bolster the morale of the troops there."12

An advance detachment of the Japanese battalion left Babelthuap during the night of 22-23 September and stealthily headed for the northern tip of Peleliu. The Japanese felt that chances of detection were slight, for the first sustained rain since the beginning of the Peleliu operation fell throughout the night. The Marines on Peleliu likewise welcomed the rain:

During our first days on Peleliu rain proved an aid to fighting troops rather than a hindrance; the sandy ground absorbing the water readily and preventing any problem of mud. Rainfall increased the available water supply and provided a welcome break in the energy-sapping heat.13

The Japanese were not to reach their destination undetected. At 0535, the destroyer H.L. Edwards spotted seven enemy barges about 1,000 yards northeast of Akarakoro Point at the northern tip of Peleliu, obviously headed for the island. The destroyer opened fire, sinking one barge before the remainder reached the beach. A combination of bombing and strafing, naval gunfire, and artillery hit the barges on the beach, and by 0845 the cruiser Louisville reported all barges destroyed.

The enemy version of this incident agrees only in part with the American account, adding that "despite receiving severe enemy air and artillery fire at a point two kilometers off the coast, they made a successful landing at 0520 hours under the command of 1st Lieutenant Murahori".14

Aware of the fact that at least some Japanese reinforcements were reaching Peleliu and still uncertain as to the extent to which the Japanese might suPport their garrison on the island, General Rupertus on 23 September faced a difficult tactical decision. If the enemy had any weak spot on Peleliu, it was bound to be situated between the central ridges and the western shore. It was here that General Rupertus decided to commit the 321st Infantry Regiment in a drive up the western coast of the island. Such an advance would take the regiment northward from Phase Line O-3 to O-4 (See Map 8). The line of departure for the attack was located about 1,000 yards north of Ngarekeukl. The drive was to continue until a point north of the village of Garekoru was reached. During the advance, the left flank of the regiment was to remain anchored on the beach, while the right was to extend about 250 yards east of the West Road into the ridges of the Umurbrogol. The 1st and 2d Battalions, 7th Marines, were to support the Army drive by pressing against the Japanese


Map 8: Commitment of RCT 321, 24 Sep
Map 8: Commitment of RCT 31, 24 Sep


positions from the south and center, while 3/7 would advance to the right rear of RCT 321 over the high ground to cover the advance of the soldiers. If the situation called for bypassing the central Japanese defenses, the troops advancing northward were to remain on the lookout for any route that would permit the isolation of the Umurbrogol region from the north. The attack was to jump off at 0700, 24 September.15

The narrow coastal plain of western Peleliu, where RCT 321 was to operate during the drive to the north, varied in width between 50 and 750 yards. At the southern end of the West Road, where the soldiers relieved the 1st Marines, the coastal plain was about 500 yards wide. About halfway up the western coast of the island at a point where the coastal plain attained a width of 750 yards, was the village of Garekoru. Just south of the village an unimproved trail led eastward through the hills until it connected with the East Road. From its widest point near Garekoru the plain narrowed down to the north until its width near the northern tip of Peleliu was only about 50 yards.

Vegetation varied considerably on the west side of the island. To the south of Garekoru exposed coral was covered with dense tropical underbrush and trees. To the north of the village the soil and vegetation changed in character. The coral was gradually replaced by sand, and the tropical jungle gave way to coconut palms. From the point where the attack was to jump off on the morning of 24 September, slightly more than 6,000 yards separated the 321st Infantry Regiment from the northern tip of Peleliu. The broken, jagged coral cliffs dominating the western plain were honeycombed with Japanese defenses, which consisted of dugouts and caves extending from the very base to the tops of ridges, varying in elevation between 50 and 200 feet. Clearly, all the advantages of cover and concealment accrued to the defenders.

For the soldiers and men of the 7th Marines, the drive to northern Peleliu promised to be an arduous undertaking. Once again the stamina, aggressiveness, and raw courage of the men in the assault were to be pitched against the fanatical determination of a well provisioned foe who preferred death to surrender.


During the afternoon of 23 September, the 2d Battalion, 321st Infantry Regiment, occupied positions about 700 yards north of the airfield along the West Road. The 3d Battalion was echeloned closely behind it for immediate support. The 1st Battalion, designated as regimental reserve, was located in the vicinity of the regimental command post.

To obtain a more detailed picture of the tactical situation in his zone of attack, Colonel Dark dispatched patrols northward between the West Road and the seashore. Other patrols from 3/1 had previously reconnoitered this


ground, having advanced to Garekoru without meeting any heavy resistance. Nevertheless, the movements were observed by the Japanese occupying the ridges to the east of the road, whose fire on the patrols was without noticeable effect. At Garekoru the patrols encountered numerous land mines, aerial bombs emplaced as land mines, and a few defensive positions which were not being held in strength. At 1700, the patrols reported the entire area from Phase Line O-3 to Garekoru free of Japanese.

The receipt of such an optimistic report impelled General Rupertus to advance his timetable for the drive, scheduled to get under way on the morning of 24 September. RCT 321 received orders to send one battalion forward at once with the mission to advance to Phase Line O-4 and dig in for the night. The task was assigned to 2/321, which began to move north at 1730. One company, moving forward between the West Road and the coast, was able to advance unopposed under the cover of a low ridge. The Japanese on the central ridges promptly spotted the company advancing east of the road as it attempted to move over open ground. The unit drew such a heavy volume of rifle and machine gun fire that the advance promptly ground to a halt before it had made much headway. Even though it had not as yet drawn any enemy fire, the company on the left halted after an advance of 100 yards to maintain contact with the unit on its right. The fading hours of daylight found both companies withdrawing to their starting points, where they dug in for the night, after establishing a continuous line of defense.

Aside from the combat action along the western coast of Peleliu, several other developments had taken place on the island in the course of the day. On the beaches and at the airfield the combat engineers had continued the removal of bombs, mines, and duds. They cleared the main beach for about 1,000 yards, laid out perimeter roads around the airstrip, and constructed a road from the south end of the airfield to SCARLET Beach. Temporary repairs on the fighter strip were completed, as was a dispersal area for 9 night fighters and 24 day-fighter aircraft. Work had begun on the southwest half of the bomber strip and a temporary control tower was half finished. The urgency of work on the airfield was emphasized during the afternoon when a B-24, the first plane larger than a TBF to land on the Peleliu strip, made an emergency landing.

The relative quiet that settled over the island with the approach of darkness on 23 September was shattered by alerts that were sounded at 1823 and 1900. No enemy aircraft appeared, however, and no bombs were dropped. During an otherwise quiet night, an enemy mortar shell landed in the CP of 3/7, causing one casualty.

For the Japanese on Babelthuap the arrival of nightfall marked the time when the main body of the 2d Battalion, 15th Infantry Regiment could begin its embarkation for Peleliu. Versions of what transpired as the Japanese approached the northern tip of Peleliu differ. According to the Japanese:

The main body of our reinforcement force [2d Battalion, 15th Infantry Regiment] left the Palau Proper Islands on the


Corsair being readied for mapalm attack on enemy positions in the Umurbrogol
CORSAIR BEING READIED for napalm attack on enemy positions in the Umurbrogol. (USMC 100375)

Japanese ridge positions on Peleliu under air attack
JAPANESE RIDGE POSITIONS on Peleliu under air attack. USMC 98401)


night of the 23d. Nine of our landing barges arrived safely, but six of them were shelled and burned when taking the wrong landing route. However, most of the personnel in those six barges were able to land by walking in the shallows.17

Accordingto the American account of the action:

. . . during the night from 23-24 September, there was again considerable enemy barge activity to the north, with seven craft sighted in that area, attempting to reinforce the island. They were sighted at approximately 0330 on the morning of 24 September and were brought under our naval and artillery fire, LVT(A)s also were active in the northern waters, taking the barges under close-range fire. All the enemy craft were reported sunk.18

One of the units playing a major role in combatting the Japanese barges was 2/11, which at 0500 commenced firing on enemy barges approaching from the north and apparently heading for the northeastern tip of Peleliu. Destroyers also opened fire on the craft and kept them illuminated with star shells. An observer reported that the effect of the fire was excellent. He counted 8 or 10 barges sunk or damaged and saw many people swimming about the debris. Throughout the morning of 24 September, fire was placed on survivors who were attempting to wade ashore on Peleliu. The Japanese tendency to bunch up increased the effectiveness of the artillery fire. Final reports were that 11 barges had either been sunk or disabled.19

A prisoner of war, identified as a member of the engineer unit of the 2d Battalion, 15th Infantry Regiment, shed further light on the incident. According to him, the barges intercepted off Peleliu the previous morning had carried most of the supplies of the 2d Battalion, whereas the second echelon carried mostly troops. This prisoner estimated that 600 Japanese had gone ashore either on northeastern Peleliu or on small islands in the area.

Shortly before 0600, 24 September, an air, naval, and field artillery preparation was directed against the west side of the central ridges and suspected Japanese positions near Garekoru. At 0700, the 2d Battalion, 321st Infantry Regiment, jumped off in an attack designed to seize the area west of the central ridges to a line about 500 yards north of Garekoru. The 3d Battalion followed the attacking 2d in column. The beginning of the attack saw 3/7 echeloned to the right rear of the 2d Battalion, RCT 321, with the mission of screening the right flank of the advancing infantry.

This advance initially progressed on schedule as far as those elements moving through the coastal plain were concerned. Despite moderate small arms and mortar fire, the leading elements of 2/321 reached the junction of the West Road and a trail leading eastward by noon. (See Map 9). Fully aware of the importance of this juncture, the Japanese had established positions here that were more formidable than anything the infantrymen had encountered during the morning's advance. In a brief but violent action, the 2d Battalion overcame the Japanese at the road-trail


Map 9: Isolation of Umurbrogol, 26-27 Sep
Map 9: Isolation of Umurbrogol, 26-27 Sep


junction. Ejection of the defenders from this strongpoint resulted in the capture of an antitank gun, three machine guns, and a naval gun. The 2d Battalion continued its advance up the road to Garekoru, leaving rear elements to explore the trail, which later was to assume major significance. In its rapid northward advance, the battalion reached Phase Line O-4 to the north of Garekoru in midafternoon, when it halted, having reached the day's objective.

Elsewhere and particularly along the right flank, the advance did not proceed as envisioned. The rapid progress of 2/321 over the road had caused 3/7, moving over the ridges on the right, to fall behind. As a result, elements of 3/321 were sent up to the ridges to fill the gap. Soon, these elements ran into strong opposition along the low ridge that paralleled and dominated the road from a distance of 50 yards. In the face of strong opposition from this elevation, the infantry later pulled its right flank off the ridge and advanced up the coastal road. According to the Army version of the incident, withdrawal from the ridge occurred because the troops there wanted to maintain contact with the rest of the battalion.20

Regardless of the reasons leading to the withdrawal of troops from this vital terrain feature, it appeared likely that the Japanese might take advantage of the situation and reoccupy the ridge, a possibility that was glaringly apparent to the commanding officer of 3/7, who was anxiously watching this development.

It was perfectly obvious to anyone who stood on that ridge that its control by the Japs could have been disastrous to the whole effort. We had been ordered to maintain contact with them [RCT 321] by trailing their right flank in column along the crest of the ridge from the point where we had earlier tied in with 3/1. I myself was at this boundary when the movement by 3/321 commenced and I personally saw the whole thing. They moved forward along the ridge a few yards until they encountered the first enemy positions, then gave it all up as a bad idea, and bore sharply to their left front to the coastal road below. It was certainly not our mission to maintain the contact with them on the flat coastal road and turn the ridge over to the Japs. I therefore reported what I had seen immediately to CO, CT-7, as well as CO 3/321. The latter officer . . . promised that he would do everything within his power to get them back on the high ground. I watched several abortive efforts to do so before finally becoming convinced that if friendly troops were going to control the ridge that night it would have to be our I Company.21

At 1310, 3/7 informed regiment that Company K had been committed on the right of the battalion, while Company I had been committed on the left. In his report to the regiment Major Hurst emphasized that 3/321 had withdrawn from the hills towards the road, leaving an undefended gap on the left flank of 3/7, that the Japanese had occupied the hills, and that 3/7 was fighting to retake them.22

Five minutes later, while Company I was reoccupying the ridge, Captain Ferguson,


CO of the company, was killed.23 Shortly before 1500, 3/7 reestablished contact with the Army units on the West Road. As a result of their taking the initiative to recapture the vacated ridge, the Marines eliminated a serious threat to the flank of the advancing Army units at a cost of 17 casualties.

In the course of the afternoon, the leading elements of 3/321 probed the central ridges north of the Marines in an attempt to discover a route leading eastward. One company of infantry was finally able to gain a foothold on the first ridge line east of the road and about 600 yards south of Phase Line O-4. The main body of the battalion remained in the proximity of West Road during the night. At 1700, Company L, 3/7, relieved Company I, which went into bivouac as battalion reserve. At 1800, 3/7 dug in for the night on a 400- yard front with Company L on the left and Company K on the right.24

The 2d Battalion, RCT 321, reached Phase Line O-4 by midafternoon of 24 September. For the remainder of the day, the battalion pushed aggressive patrols up the West Road. A strong combat patrol, supported by Marine and Army tanks, moved northward against negligible opposition for nearly 2,000 yards. At this point, the patrol had nearly reached Phase Line O-5. From their vantage point the soldiers saw the extensive Japanese radio installations, which were situated about 600 yards north of Phase Line O-5, and numerous Japanese pillboxes, caves, and other defenses along West Road. Before dark the patrol withdrew to Garekoru without having encountered serious enemy resistance.

In the meantime, Company G, the leading unit of the 2d Battalion, began organizing for defense to protect the north flank of the regiment. Hardly had this defense been established when, at 1700, the enemy counterattacked and forced the 2d Battalion to withdraw about 200 yards. The infantrymen promptly retook this ground. A second enemy counterattack shortly after 1800 disintegrated before it ever got started when effective artillery concentrations were placed on the assembly area.25

Aside from the activity which took place along West Road, the most far-reaching action on 24 September was destined to occur to the east of the road, and just south of the village of Garekoru, where the advancing infantry had earlier in the day discovered a poorly defined trail leading eastward into the enemy-held ridges. For lack of a better name, and because elements of RCT 321 had been the first to encounter it, the route was designated the "321st Infantry Trail."26

Since this trail could represent the only lateral link between the West and East Roads on Peleliu, its tactical importance was immediately apparent. From the point where it branched off from the West Road, the trail led through a stretch of swampland before ascending into the enemy-held ridges. Patrols cautiously followed the trail as it meandered into the high ground, encountering only negligible resistance. In


orcler to take advantage of what could prove a weak spot in the Japanese defenses, Colonel Dark dispatched an infantry company of 2/321 along the trail.

As the company advanced into the ridges, Japanese resistance stiffened, particularly on Hill 100, which dominated the route of advance and blocked the progressof the companyto the east. Since this hill, which formed the northern bastion of the Japanese pocket in the Umurbrogol ridges, dominated not only the 321st Infantry Trail but also the East Road, its capture assumed major importance. In a brief but bitter engagement, the infantrymen seized thehill andthen grimly held on against all Japanese attempts to retake it. Fully aware of the decisive importance of holding the hill and, if possible, of gaining further ground to the south, Colonel Dark diverted most of the 3d Battalion to the east. Soon three additional infantry companies were attacking eastward along the trail into the area south of Hill 100. Before nightfall, the 3d Battalion had seized an escarpment south of Hill 100 and established contact with 3/7 farther to the southeast.

By evening of 24 September, it appeared that the lack of momentum which had forestalled practically all forward movement for several days, had ended. From his new division command post in the former Japanese administration building at the northern edge of the airstrip, General Rupertus pondered his next move.27 The advances made during the day had surpassed his expectations, and the tactical situation had radically changed for the better since morning. In addition, the sweep up the coast had brought a new high in the number of enemy prisoners, three being taken during the day.28

On the other hand, the successes achieved during the day were accompanied by difficult problems of evacuation and supply. Army engineers, already fully occupied in clearing Garekoru cf mines and boobytraps, had to improve both the narrow West Road and the 321st Infantry Trailj a job that continued after darkness during the night of 24-25 September. The men on the escarpment south of Hill 100 faced an even greater predicament, since they had to manhandle all supplies going to their exposed positions. During the return trip the supply party encountered staggering difficulties in evacuating the wounded over the inhospitable terrain. Yet the advantages gained by the day's advance far surpassed all resulting problems.

Elsewhere on Peleliu, extensive work on the airstrip had progressed to the point where the short southeast-northwest runway was fully restored. Even though heavy fighting still was in progress about two thousand yards to the north, a large number of aircraft were able to land, including two C-46S, one C-47, four PBYs, and the first echelon of VMF (N) -541, consisting of eight Hellcats. Henceforth the night fighters would be available for night operations.29

For the continued operations the following day, 3/7 shifted Company I from


battalion reserve to the left of Company L, a maneuver designed to extend the line held on the ridges an additional 250 yards to the north. The 1st Battalion was ordered to move up on the West Road, where it was to remain in Support of 3/321 until an opportunity arose to shift direction and advance southward into the central ridges.

There were no reports of night activity in any sector of the 1st Marine Division.30 Lest an erroneous impression be created that everything on Peleliu was entirely quiet, some action continued intermittently throughout the night. At 2200, Japanese threw hand grenades in front of the 3/7 sector.31 Also during the night, Japanese with demolitions fastened to their bodies, tried to infiltrate the 7th Marines lines. One Japanese, killed 10 feet in front of a halftrack, was found in the morning, a Molotov cocktail tied to one leg, explosives fastened to his back, and grenades stored in his pockets.32

The plan of operations for 25 September was to cut across the island near the 321st Infantry Trail. This maneuver would complete the isolation of Japanese forces in the Umurbrogol ridges. At 0700 elements of 2/321 jumped off from Hill 100 and in midmorning reached the edge of the East Road. Enemy resistance during this advance remained moderate and consisted primarily of rifle and machine gun fire from a key height across the East Road known as Hill B. The infantrymen paused in order to await the arrival of 3/321 for a combined drive across the road. When it attempted to expand the hold on the escarpment occupied the previous afternoon and to move onto East Road, the latter battalion ran into heavy fire from pillboxes and emplacements protected by steep walls and sheer cliffs guarding the northern approaches to the Umurbrogol defense system. There the drive of 3/321 bogged down and for the remainder of 25 September that battalion made no further gains. Any hope of a two-battalion assault against Hill B, which the Japanese were evidently holding in strength, had to be abandoned. For the remainder of 25 September, the Army troops near East Road remained in place.33

Despite the disappointing turn of events along the 321st Infantry Trail, progress appeared vastly more promising along the western coast of Peleliu. This applied primarily to developments along the West Road, where, at 0700, a strong Army combat patrol composed of infantry, tanks, and flamethrowers moved north from Garekoru village. The mission of this force was to destroy enemy installations that reconnaissance on the previous day had identified. This combat patrol advanced 1,200 yards, killing 30 Japanese and destroying four pillboxes and two large supply dumps before it reached Phase Line O-5. The patrol arrived at its objective without sustaining any casualties.

Such weak resistance along the West Road indicated to General Rupertus that the enemy had concentrated his main strength in the central ridges of Peleliu. A swift drive to the north over


West Road, coupled with a continuation of the RCT 321 drive to bisect the island, would result in the complete isolation of the Japanese in the central ridges while eliminating all enemy resistance in northern Peleliu. The possibilities that were now open to the division commander were greatly expanded, and General Rupertus decided that swift action was indicated. At 0945, division headquarters orally ordered the 1st Marines to take over the sector of the 5th in addition to its own. The 5th Marines was to pass through the lines held by 2/321 and attack towards the northern tip of Peleliu. Elements of the 321st Infantry Regiment engaged near East Road were to continue their attack eastward, but bypass the hard core of enemy resistance until movement up East Road became feasible. Then, in conjunction with the 5th Marines, they would launch a drive to the northern tip of the island.34

For the first time since the Marines had stormed ashore on Peleliu, all initiative had passed into the hands of the attacking force. The Japanese, holed up in their caves and dugouts, could still kill and maim; they were to inflict many more casualties before the fight for the island was over; but the road to the tip of the island was now open. Marines were moving up to eliminate any resistance that the enemy might still offer. Japanese hopes of receiving reinforcements were fading as more and more of the coastline in their hands had to be relinquished. The momentum of the American drive had been restored. At long last the beginning of the end of the costly and arduous operation was in sight.


Orders to move to the western portion of Peleliu found the 5th Marines in static positions on Eastern Peleliu, where the regimental mission had been to prevent enemy counterlandings. The 1st Battalion was deployed in the vicinity of the radio direction finding station near Ngardololok, the 2d Battalion was holding the islands north of the northeastern peninsula, while the 3d Battalion, less one company, occupied defensive positions along Purple Beach. The 1st Marines completed the relief of the 5th shortly after noon and assumed command of the 5th Marines sector at that time. In order to expedite movement of the regiment to the West Road, the battalions moved out on trucks at 1300, with 1/5 in the lead, followed by 3/5 and 2/5.36

By 1600, the 5th Marines had executed the passage of the lines, and the regiment passed through Phase Line O-4 near Garekoru. As 1/5 continued its advance up the West Road, it encountered erratic resistance from what appeared to be Japanese holdouts. The level terrain was devoid of the dense jungle growth abounding elsewhere on Peleliu and offered ideal conditions for the movement of tanks and LVT flamethrowers accompanying the advance battalion. Aside from occasional Japanese sniper and mortar fire, the advance


continued for about 500 yards without interruption to Road Junction 15, where teh West and East Roads met. This vital point was defended by a small Japanese force, which had installed itself on a ridge domiating the road forks.

Around 1700,m as the Marines approached this point, the enemmy opened fire, which served only t0o delay the Marines. When the firefight ended the Japanese had lost 20 killed. The battalion continued its advance for another 100 yards and seized the Japanese radio station, whose towers the forward elements of RCT 321 had spotted on the previous day. Hoaving secured this objective, the battalion established night defenses north of the radio station from the beach to the high ground east of the West Road (See Map 10). For the remainder of the evening and throughout the night, the forward elements at the radio station were continuously and heavily engaged with the enemy. The 1st Battalion, 5th Marines, was the recipient of direct fire form two 70mm guns firing out of caves less than 300 yards away. The Marines drew additional fire from enemy artillery and mortars located on the northern tip of Peleliu, as well as from two 37mm guns on Ngesebus Island. Small arms fire from three directions converged on the battalion.37

To reduce the enemy artillery and mortar fire that was proving so troublesom to 1/.5, the 11th Marines massed its batteries against suspected enemy artillery and mortar positions. Since direct observation was impossible, the regiment used a target list, firing on areaqs that looked promising on the basis of a chart and photographic inspection. Whenever such fire resulted in reduced enemy activity, continuous shelling at a slow, irregular rate was employed on the target throughout the night. This counterbattery fire proved very effective and helped the exposed battalion to maintain its forward positions.38

Nor was heavy mortar and artillery fire all that 1/t had to contend with during the night of 25-26 September. The Japanese launched three counterattacks ion the course of the evening against the hard-pressed Marines, but each of these attacks was repelled. At 0200, 26 September, a platoon of Company C took the initiative by launching a surprise counterattack, which resulted in the destruction of two particularly troublesom machine guns.39

Upon reaching Road Junction 15 in the wake of the advbance of 1/5, the 3d Battalioin pivoted to the southeast and headed down East Road, wehre it established night defenses on the road and along the western slopes of Hill 80. The latter hill, in itself an isolated terrain feature, owed its importance to its location, for it was the only link separating the Kamilianlul ridges to the south from the Amiangal ridges, which formed the northermost hill mass on Peleliu.

Unlike the effective resistance that 1/5 had encountered near the radio station, the enemy that 3/5 came upon was either disorganized or unawar5e of the Marines' presence on the ESt Road. According to an eyewitness:


Map 10: North Peleliu and Ngesebus, Northward Attack 5th Marines, D+10
Map: North Peleliu & Ngesebus
Northward Attack 5th Marines


. . . darkness had fallen and I was engaged in tying in I Company with the adjoining Second Battalion Unit under the command of Capt. Albert J. Doherty. Captain Doherty and I were standing on the road with a small group around us discussing our situation when the word was passed from my outpost that enemy troops were approaching. We immediately took cover, not knowing how large a force it was or whether it was cognizant of our presence. The enemy was obviously unaware of the 5th Marines advance because they approached making a great deal of noise. When they reached the approximate position where Captain Doherty and I had been my men opened fire killing all of the enemy and sustaining no casualties. The enemy group consisted of about 12 Japanese Marines.40

Once again, on the evening of the second day of continuous forward movement, General Rupertus had cause to be pleased with the progress that had been made during the day. Even though the effort of the Army troops to isolate the central Japanese defenses had been temporarily halted, this lack of progress had been more than compensated for by the rapid advance of the 5th Marines up the West Road. Except for a critical shortage of hydrogen for flamethrowers, the flow of supplies was moving smoothly.41

Behind the front lines, the engineers were keeping pace with the advance of the combat troops. By evening of 25 September, 500 men of a naval construction battalion were engaged in shore party work. Improvement of roads by the engineers was continuing. This work helped largely to offset the complaint voiced by RCT 321 "that the presence of the 5th Marine Regiment on the west coast further complicated the traffic control problem."42

Work on the airstrip was also making good progress. The Engineer Group, consisting of elements of the 1st Engineer Battalion and the 33d and 73d Naval Construction Battalions, had dug out and refilled most soft spots and graded, rolled, and watered the entire strip. Radar units and control equipment were in operation. The airstrip tower had been readied for around the clock operation.

After a night of incessant harassment by the Japanese, 1/5 spent the early hours of 26 September in consolidating its positions around the radio station and preparing for continuation of its attack later in the day to the northern tip of Peleliu. Shortly after 0600, 3/5 jumped off for an attack against Hill 80 with Company K on the left and Company I on the right. The attack carried the hill and by 0830 the assault force reached a swamp bordering the east coast of Peleliu. This advance of the 3d Battalion was of major importance because it effectively cut the island in two.

The 1st Battalion, 5th Marines, ran into stiff opposition from the Amiangal ridges dominating northern Peleliu. The northern portion of the L-shaped hill system consisted of ridges running generally from northeast to southwest for about 1,000 yards; the southern leg extended from northwest to southeast. The southern leg of the ridges was not


continuous but broken into four separate hills or knobs, designated from northwest to southeast as Hill 1, Hill 2, Hill 3, and Radar Hill, so named because it had at one time served as an enemy radar installation. These four knobs were to gain ill repute as Hill Row. The entire Amiangal ridge system was held in strength by the Japanese, particularly the portion paralleling the route of advance of 1/5. This part contained some of the most elaborate caves and tunnels on Peleliu.

The battalion had barely started out along the West Road when the enemy in and on Hill 1 opened up on the Marines with 37mm and 75mm guns as well as automatic weapons and mortars. This curtain of fire from the Amiangal ridges was reinforced by heavy fire from Ngesebus Island. All forward movement soon halted. Attacking eastward from the West Road, Company B assaulted the second knob, Hill 2, but also encountered opposition. Through sheer determination the company was able to gain a firm foothold on the hill by early afternoon. This accomplishment in effect served to outflank the Japanese on Hill 1, but Japanese resistance on the last mentioned hill continued throughout the day. An attempt by Company C to seize all of Hill 1 during the remaining hours of daylight was unsuccessful, and continuation of the assault had to await the following day.

During the bitter fighting in which 1/5 engaged during the night of 25-26 September and for most of the following day, 2/5 remained stationary on the southern flank of the regiment. At 1600 on 26 September, when it had become apparent that 1/5 could make no further progress, 2/5 was ordered to attack. The battalion advanced northward through the left wing of 1/5, carefully bypassing embattled Hill 1. In the course of its advance, the battalion drew heavy fire from Japanese emplacements in the plain, from the ridges on its right, as well as from Ngesebus. Enemy mortars proved especially troublesome. As a result of the heavy fire, Company F lost four of its supporting tanks before it had advanced very far beyond Hill 1.43 Attempts by the artillery to give all possible support to the battalion were largely ineffectual. "We fired frequent missions throughout the day on these mortars; the reported effect was that the mortars were neutralized while we fired but that they were not destroyed. The enemy apparently withdrew into the caves during the period of our fire."44

As evening approached, the 5th Marines occupied a jagged front line. The 1st Battalion, though out of contact with the 2d, was tied in with the 3d on the right. The mission of the 3d Battalion was to support either the 1st or 2d Battalion in the event of a major Japanese counterattack. Further progress of the 2d Battalion was impeded by a large antitank ditch, which blocked the approach to the remnants of the Peleliu phosphate plant. The Japanese had converted the reinforced concrete foundation of the otherwise demolished structure into a major defensive installation. Exposed to the enemy fire from Ngesebus Island and plunging fire from caves and defensive positions from the hills,


2/5 found itself in a very unenviable situation.

Further south on the island, at the juncture of the 321st Infantry Trail and the East Road, 26 September was to bring additional gains, though not without great difficulty. On the previous day, 2/321 had seized Hill 100 and advanced to the foot of Hill B, which dominated East Road and formed the last obstacle in the path of RCT 321's advance to the east coast. The mission of capturing this vital hill had been given to 3/321, which had been scheduled to launch the attack at 0700, with 2/321 in support. Even before 3/321 could get into its attack position along East Road it drew such furious fire from small arms, mortars, and automatic weapons in the northern strongpoints of the Umurbrogol Pocket that the movement bogged down.

When it became apparent around noon that the 3d Battalion would be unable to reach Hill B, the mission of seizing the hill was assigned to 2/321. To ensure the success of the attack, it was decided to launch a two-pronged assault against the hill. During the time that the 2d Battalion was organizing for the attack, a special task force composed of 7 medium tanks, 6 LVTs, 1 LVT flamethrower, and 45 riflemen, advanced northward over the West Road to its junction with East Road. There the column pivoted to the south, followed the East Road, and by 1500 had reached a point within 150 yards north of Hill B. Such a bold maneuver could not fail to attract enemy attention, and the column suddenly found itself under frenzied attack by 15 Japanese who, vastly outnumbered and outgunned, were promptly killed in the skirmish.

At 1600, when both attack forces had reached their jumpoff positions, the attack against Hill B commenced. White phosphorus mortar shells shrouded the hill in smoke, obscuring visibility of the defending force. The task force attacked from the north, while two companies of 2/321 attacked from the west and south. At the outset both attack forces encountered strong resistance, but in less than an hour the hill had changed hands and the entire Japanese force was wiped out. A number of Korean laborers, evidently less inspired than their taskmasters, preferred to surrender.

Despite the loss of Hill B, the Japanese took the initiative at least once during the day by attacking the command post of RCT 321 in force, after infiltrating the widely spaced Army and Marine lines. This surprise attack was launched by enemy troops armed with rifles and machine guns. The force created havoc momentarily but could not match the soldiers' fire. The Japanese then retreated, leaving 35 dead around the command post area.

Earlier that day, General Rupertus had felt that time for the capture of Ngesebus Island was ripe. "Improvement in our tactical situation led to the making of plans for an attack on Ngesebus to the north to be carried out the following day, but developments failed to warrant the pursuit of this action."45 Instead, an additional day was set aside for preparations for the assault on


Ngesebus as well as consolidation of the hold the 5th Marines had gained on the northern portion of Peleliu. Despite heavy resistance, elements of RCT 321 in the center of the island and the 5th Marines in the north had reached the eastern coast. The Japanese on Peleliu, though still capable of prolonged resistance, were now divided into two distinct pockets whose elimination was merely a matter of time.

Also on 26 September, "a new high in the number of prisoners for a single day was obtained when six POWs were brought in, doubling our total for the operation."46

The night of 26-27 September was to be typical. The Ilth Marines directed harassing fires against Japanese positions on the northern tip of Peleliu throughout the night.47 The Japanese, in turn, shelled the area occupied by 3/321. Fighting also flared up in the center of the island where several Japanese were killed as they attempted to infiltrate the positions of 2/321 in the vicinity of Hill 100 and Hill B. The 7th Marines spent a generally quiet night, marked only by an exchange of hand grenades in the sector of 2/7. Weapons Company killed three Japanese during the night, one of them equipped with a machine gun. One of the Japanese dead carried the following message.

American Brave Soldiers: We think you are much pity since landing on this island. In spite of your pitiful battle, we are sorry that we can present only fire, not even good water. We soon will attack strongly your army. You had done bravely your duty; now abandon your guns, and come in Japanese military with a white flag (or handkerchief) so we will be glad to see you and welcome you comfortably as we can well.48

The above Japanese offer found no takers among the soldiers and Marines on Peleliu. In fact, plans for the assault on Ngesebus were already completed. Naval demolition teams were busy offshore clearing the waters between PeleIiu and Ngesebus in preparation for the coming assault.

At 0800, 27 September, another milestone in the Peleliu operation was reached when, in a simple ceremony, the American flag was raised in front of the division command post. Even while this ceremony was in progress, the 5th Marines was once again embroiled in fighting of the utmost severity, in which a battalion command post was just as exposed to enemy fire as any other forward position. This applied to the CP of 2/5, which on the morning of 27 September was located in the radio station near the northern tip of Peleliu. This building was very well constructed, but had been thoroughly worked over by artillery and air bombardment.

The framework was still intact, however, and after clearing the dead Japs out, there were some rooms that were very suitable for a good CP. During the morning of the 27th several rounds of mortar fire were laid on the building with such accuracy that two of them went through the second story windows. Casualties were light but just knowing the Japs had the range on us wasn't so pleasant. Earlier that morning mortars were laid around


the CP, with devastating effect, and caused light casualties. Major Gayle, for example, had men killed all around him but he was lucky and never got a scratch, just sand blown in his eyes, and the helmet blown full of holes. Lucky he didn't have the helmet on.49

The 2d Battalion was to find progress extremely rough for the remainder of the day. First, the erstwhile phosphate plant, which the Japanese had turned into a blockhouse, had to be secured in an area that bristled with snipers. The problem confronting Colonel Harris was a formidable one. The blockhouse in front of 2/5 could not be taken until the antitank ditch blocking the approaches to it had been seized. This was a job that the infantry could not tackle without armored support. The regimental commander decided to utilize all arms available in reducing these obstacles. First of all, Colonel Harris called naval gunfire and artillery in on Ngesebus and any other targets suspected of harboring artillery or mortar positions.

A medium tank, equipped with a bulldozer blade, was pressed into service to level the antitank ditch, and filled it by 0830. A LVT flamethrower then was able to come within effective range of the Japanese fortification. Moments later, when the flame and smoke had cleared, all resistance from this stronghold had ceased and more than 60 dead Japanese remained in the rubble. While this action was in progress, patrols from Company E seized a small, weakly defended ridge abutting the road from the east. The 2d Battalion thereafter resumed its advance northward along the road as well as over the adjacent ridge. Company F, at the head of the column, soon found itself embroiled in some of the most bitter and frustrating action of the entire campaign. Aside from receiving heavy Japanese artillery and mortar fire, the company faced a series of pillboxes and field fortifications on level ground, and layer upon layer of caves in the hillsides.

Even though they were not aware of it at the time, the men of the 5th Marines had come upon the most skillfully constructed defenses on Peleliu. The Japanese Army had utilized all of the many natural caves possessing tactical value, adapting them for the emplacement of heavier weapons with great ingenuity. On the other hand, Japanese naval troops had preferred to construct their own caves with the help of the 214th Naval Construction Battalion, composed of men who had been professional miners and tunnel workers in civilian life. Since most of these Navy caves were located near the northern end of Peleliu, they proved a serious obstacle to the advance of the 5th Marines.

Sheer courage and heroism in themselves proved inadequate for the task. In the course of the morning of 27 September, Company F seized the two ridges forming the northwestern anchor of the Amiangal system and established observation posts on the crests. But this did not solve the problem of what to do about the Japanese occupying the caves about half way up the hill. Marine casualties mounted steadily, and evacuation of the wounded became more and more difficult. The assault on the ridges soon turned into a nightmare:


The hill we were taking was honeycombed with caves, and we used everything in the book in closing them. (Hand grenades, five gallon cans of gas, composition C wrapped around 81mm WP mortar shells, flame throwers, and finally a 155mm gun). . . . During the assault the Japs still fired out of the caves, even throwing our own grenades out at us. It was hard to believe that a 70 lb. compo charge wrapped around a 81mm mortar shell, which when set off fairly rocked the OP on top of the hill, did not kill the Nips in the caves.50

Since the cave openings completely dominated the road leading past the northernmost ridge of Peleliu, the advance of the entire 2d Battalion ground to a halt. The first tank attempting to squeeze through the narrow gap between the hill and the northwestern shore was hit, and the Marines occupying the crest of the hill were powerless to cope with the caves underneath.

This seemingly insurmountable problem was solved late in the afternoon when the troublesome bottleneck was eliminated by means of an expedient involving a combination of arms. The solution was complicated but proved effective. First, artillery placed a continuous barrage on Ngesebus, while naval guns shelled Kongauru, an island northeast of Ngesebus and connected to the latter by a causeway. Joining the shoot were nine tanks which placed smoke shells on the closest Ngesebus beach from positions on the West Road. Every fourth artillery projectile fired against Ngesebus was a smoke shell. Five LVT(A)s, equipped with 75mm guns, then pushed out into the channel and fired point-blank into the mouth of the cave. Under this covering fire, the tanks, supported by Company G, moved up the road and beyond the cave. An LVT flamethrower was then employed against the cave. The lines of the 2d Battalion consolidated this gain and dug in for the night.51

The 1st Battalion also made important gains on 27 September. Most of the action for the day centered around Hill 1, which Companies B and C attacked at 0930. once again the Marines were exposed to heavy fire from Japanese small arms and antitank guns emplaced in caves at the foot of the hill and on the slopes. After approaches for the supporting armor had been cleared, the rate of progress increased. In the course of the afternoon, 1/5 destroyed four 75mm and four 37mm guns, and a large number of automatic weapons. By 1700, having established a firm foothold on the hill, the Marines prepared to set up positions for the night. During the remaining hours of daylight, engineer demolition teams systematically eradicated every cave and each hollow even remotely resembling an enemy position.

In the course of 27 September, 84 Koreans and 7 Japanese were taken prisoner, a considerable increase over the previous total of 12 captured since the beginning of the operation.52

According to division intelligence estimates a total of 7,975 Japanese had been killed in 12 days of bitter fighting. Casualties for the 1st Marine Division were 768 killed, 3,693 wounded, and 273 missing in action, a total of 4,734.53 Similarly, casualties for RCT 321 during the first week of its commitment on


Peleliu were 46 men killed, 226 wounded, and 7 missing, a total of 279.54

Even though a major portion of Peleliu was now in American hands, tenacious resistance could be expected to continue from the remaining enemy pockets. Aggressive action was needed, not only for the elimination of these pockets, but also to remove from the adjacent islands those Japanese stiil capable of rendering effective support to their compatriots on Peleliu.


General Rupertus decided that Ngesebus Island was to be seized on 28 September, in a shore-to-shore operation executed with the assistance of all available supporting arms. The assault on the island was to be supported by a battleship, a cruiser, two destroyers, division and corps artillery, a tank company, a company of LVT(A)s, and a company of LVTs.

The mission of seizing Ngesebus was assigned to 3/5 about 1600, 27 September, when General Rupertus issued verbal orders to this effect to Lieutenant Colonel Lewis W. Walt, Executive Officer of the 5th Marines. Within the hour, members of 3/5 arrived at the 5th Marines command post to receive the attack order. Representatives of 1/7, held in reserve for the operation, also attended the meeting. The general plan of attack called for one hour of naval gunfire, air, and artillery preparation commencing at 0800. While the assault wave was covering the last 200 yards to the beach, Marine aircraft from the newly arrived VMF-114 would work the beach over. Sherman tanks were to form part of the first assault wave, flanked on either side by LVT(A)s and followed by LVTs loaded with the assault troops. The entire battalion was to embark in LVTs, and the waves were to land successively at two-minute intervals.56

H-Hour for the operation had been set to coincide with the lowest ebb of the tide to avoid water that would be too deep for the tanks, which could not be completely waterproofed.57

During the hours of daylight remaining on 27 September, 3/5, which had been held in reserve near Road Junction 15, relinquished its positions to 1/321 and assembled in preparation for the assault on Ngesebus. The Army battalion held a line extending from the junction in the north to Kamilianlul mountain, which was an extension of the Umurbrogol ridges north of the Wildcat Trail.

The curtain on the drama of Ngesebus opened on the morning of 28 September, when the massed fire of five artillery battalions from Peleliu, as well as heavy fire from warships and aircraft, blanketed the island. Near the northwestern shore of Peleliu, an impressive gathering of high-ranking officers had assembled to witness the operation. The group included such notables


as Admiral Fort, and Generals Julian C. Smith, Geiger, Rupertus, Mueller, Oliver P. Smith, and Bell. The weather was cool and cloudy and interspersed with frequent rain squalls.58

For the Corsair pilots of VMF-114, air support for the Ngesebus landings represented a very interesting and original assignment. The operation marked the beginning of combat work for the squadron, which had reached Peleliu only two days earlier. At 0630, the Corsairs hit the airstrip on Ngesebus with 500-pound bombs and strafed the entire island as well as Kongauru to the northeast. "Strafing runs were made just a few feet off the deck and a hail of lead laid all over the island."59 At 0840, 20 Corsairs preceded the landing craft and gave the island another heavy strafing. In the course of this attack, Japanese mortar positions were spotted and one specially prominent square blockhouse with an iron door was fired on and neutralized.

Whenever aircraft were not directly over the island, the artillery on Peleliu and naval guns offshore gave Ngesebus a heavy going over, starting at 0700 and concluding at 0905. Both quick and delay fuzes were used. Observers reported that the island was completely covered with fire.60 In the course of the preliminary bombardment, naval gunfire ships pounded the northern portion of Ngesebus and continued to fire on that part the island throughout the landing.

The 600-yard trip from Peleliu to Ngesebus proved generally uneventful, and the landing force proceeded as planned. Nevertheless, a few unforeseen contingencies arose that changed the sequence of events. Of the 16 tanks taking part in the operation, 3 failed to reach their destination after stalling and flooding in about three feet of water. The remaining tanks continued the crossing so cautiously that they could not keep up with the LVTs. In the end the infantry hit the beach long before the armor had completed the crossing. Another complication arose while the assault troops were en route to Ngesebus. Shortly after 0900, just as the assault waves departed Peleliu, spectators to the operation noticed that the naval gunfire had lifted and air attacks were slackening:

Upon investigation as to why the naval gunfire did not support the attack, it developed that a great deal of confusion existed as to what the location of the troops would be at H-hour. The troops had planned this attack in the manner of land warfare and to them H-hour was the time of crossing the line of departure--in this case, the shore line of Peleliu. The Navy, planning as for an amphibious assault, considered H-hour to be the time the troops hit the beach on Ngesebus. Consequently, when H-hour arrived, the fire support ships assumed the troops were hitting the Ngesebus beach and lifted their fires, when actually the troops were just moving off Peleliu.61

This temporary lapse in preparatory fire did not result in any serious disadvantage to the men of 3/5, particularly since Marine aircraft, quickly sizing up the situation, jumped into the breach and resumed the relentless strafing of the southern beaches on Ngesebus. The planes halted their assault runs


only when the assault force approached within 200 yards of the island.

The first wave hit the beach at 0911. Company K landed on the left, Company I on the right, and Company L followed in the rear as 3/5 reserve. The troops moved inland quickly and secured a beachhead against light resistance. Supporting LVT(A)s made short work of several pillboxes on the beaches. Companies I and K advanced to the northwest and linked up after crossing the airfield. To expedite the operation, the battalion commander committed Company L, which pivoted to the right and attacked towards the eastern shore of the island, encountering very little opposition in the process. The assault units suffered no casualties in the landing, while 50 of the enemy were killed or captured on or near the beaches.62

Within 12 minutes after the first assault wave had gone ashore on Ngesebus, the first tank lumbered across the beach. By 0930, all tanks and troop carriers except for three Shermans, had reached the island. Initial progress continued to be rapid, particularly in the zone of attack of Company L, which completed its mission of seizing the eastern part of Ngesebus within an hour and a half after going ashore. At 1300, a patrol of two tanks, three LVT(A)s, and one platoon of Company L landed on Kongauru Island off the northeastern tip of Ngesebus and secured the island against light resistance.

Companies I and K, attacking to the northwest, found the going considerably tougher in their zone of advance, particularly along the battalion left, where a series of ridges flanked the western shore of the island. There, the Japanese were offering determined resistance from caves and dugouts. Once again it became the unenviable task of the Marines to root the enemy out of these defensive positions, a job that was accomplished with the assistance of armor. In the midst of this operation, an enemy shell landed in the center of the battalion CP, wounding the battalion commander, Major John H. Gustafson, and a number of his men.63

By 1700, almost all of Ngesebus, except for a few hundred yards at the extreme northwestern tip of the island, was in American hands, though a few caves in the ridges still remained to be reduced. Because of the rapid progress made throughout the day, the presence of 1/7 was not required. The battalion reverted to division control as of 1500.64

The otherwise deadly serious business of seizing an enemy-held island was destined to feature at least one lighter episode. This was supplied by an aide to Major General James T. Moore, Commanding General of the 2d MAW. The aide, who also doubled as pilot, flew General Moore over Ngesebus Island in an L-5 observation airplane. In the course of this reconnaissance, the pilot observed a Japanese officer, equipped with sword and white gloves, directing the emplacement of a mortar. Upon completion of the flight and discharge of his passenger, the enterprising aide obtained several hand grenades and promptly headed back for a "bombing run" on the mortar position. The grenades were dropped with unobserved


Assault on Ngesebus Island as viewed from amphib8ous tank in third wave
ASSAULT ON NGESEBUS ISLAND as viewed from amphibious tank in third wave. (USMC 97006)

Marine assault troops advbance into Ngesebus as oil dump burns in background.
MARINE ASSAULT TROOPS advance into Ngesebus as oil dump burns in background. (USMC 102051)


results, but the Japanese opened up on the L-5 with a machine gun. The pilot received a bullet in the leg, which was to put him out of action for the remainder of the Peleliu campaign.65

The 3d Battalion spent a relatively quiet night on Ngesebus. On the morning of 29 September, Companies I and K resumed the attack. Progress was normal until the two companies had nearly reached the northern tip of Ngesebus, when a 75mm gun opened up at point blank range. The Marines quickly destroyed this weapon and went on to overcome the rest of the resistance on the island. At 1500, 29 September, Ngesebus was declared secure.

An hour later, 2/321 relieved the Marines and completed mopping up. Having accomplished the mission on the island, 3/5 returned to Peleliu. The battalion had secured the island at a cost of 15 killed and 33 wounded.66 In return, the Marines killed or captured 470 Japanese. Infantrymen of 2/321 were to account for another hundred of the enemy during the ensuing mop-up.67

After the capture of Ngesebus it became evident that the airstrip on the island was surfaced with sand so soft that the labor required to make the field operational was out of proportion to the benefits to be derived from it. As a result, no improvements were made to the airstrip, and the main advantage gained by the capture of the island was the final elimination of the bothersome fire from Ngesebus into the flanks and rear of the Marines advancing to the northern tip of Peleliu. At the same time, capture of the island deprived the Japanese of another possible staging area if they made another attempt to reinforce the Peleliu garrison.


Throughout 28 September, the spectacular operation of 3/5 against Ngesebus held the limelight. For the two battalions of the 5th Marines remaining on Peleliu, it was business as usual as the drive continued to dislodge the Japanese from their elaborate defenses on the northern tip of the island.

On the morning of 28 September, the Japanese in the northern part of Peleliu still held a pocket slightly longer than 2,000 yards on fairly level ground, except for Hill 3 and Radar Hill, and some of the tunnels in the northern leg of the Amiangal ridges, where they still resisted in strength despite the fact that the Marines were firmly established on the crests (See Map 11).

The attack of 1/5 against Hill 3 was preceded by a mortar preparation lasting over an hour. At 0845, Company C jumped off. Approaching the hill from the north and west, bazooka and demolition teams crept forward and eliminated dugouts and caves one after the other. A Sherman tank rendered valuable support by firing directly into the cave openings. In this manner the tank knocked out a particularly annoying machine gun, which had been firing on the advancing Marines from the southeastern slope of Hill 3. Before noon, 1/5 had secured a foothold on the crest


Map 11: Secuuring the North
Map 11: Securing the North


of the hill, though the Japanese still resisted in the caves below. Shortly before 1300, the enemy attempted to counterattack but was driven off by mortar fire. For the remainder of the afternoon Company C continued to consolidate its gains on the hill, capturing 15 Koreans in the process.69

At the very northern tip of Peleliu, 2/5 resumed its drive with the objective of seizing the flat ground to the north and east of the Amiangal ridges. Company G jumped off at 0700 and advanced through coconut groves near the eastern base of the ridges, where the Japanese had established elaborate fortifications. Despite heavy fire, the company continued its advance and by 1000 had killed 150 Japanese in the assault.70 For the first time on Peleliu it was observed that some of the Japanese remained completely apathetic in the face of the Marine attack and did nothing either to attack the Americans or to defend themselves. Company G continued its advance southward until it was able to bring small arms fire to bear against Radar Hill, the last ridge of the Amiangal system still entirely in Japanese hands. Here the company halted the advance.

Throughout the day Companies E and F of 2/5 worked with the demolition men trying to neutralize the caves, which honeycombed the hills. When the caves became untenable, some of the Japanese decided to make a break for it. A chase ensued when a group of about 70 Japanese suddenly poured out of the hillside and headed for the reef. Marines of Company F gave chase with three LVTs, overtook the fleeing enemy, and killed those refusing to surrender.71

Since on 28 September the Japanese retained only two small islands of resistance on Peleliu, naval air support was secured as of 1800 that date. During the night of 28-29 September, fighting flared up in the center of Peleliu. There the Japanese launched what the 1st Marine Division designated "the first in a series of desperation raids" out of the Umurbrogol pocket against the lines of the 7th Marines. The regiment suffered light casualties in repuls. ing these assaults which at times closed to within bayonet range.72 The Marines also drew mortar, machine gun, and rifle fire throughout the night.

The attack against the Japanese in the north of Peleliu reached its climax during 29 September, which also marked the end of large-scale operations in that part of the island. The 1st Battalion, 5th Marines, launched its assault against Radar Hill with flamethrowers, bazookas, and demolition charges. As anticipated, capture of the strongly de. fended hill could not be completed in one day. Even though the Marines reached the crest of Radar Hill on 29 September, a large cave underneath remained to be reduced.

East of the Amiangal ridges, Company G of 2/5 continued to mop up enemy remnants in this area. The company drew fire from caves in the ridges east of the phosphate plant. An approach


to these ridges from the east did not appear feasible because a swamp separated the Marines from these hills. Tanks stationed in the coconut grove fired point blank with their 75mm guns at the mouth of one of the most troublesome caves, temporarily sealing it. On the western side, the Japanese had blasted from the inside and cleared openings to two caves that the Marines had sealed. AS a result, the enemy was able to fire on both the east and west beaches.73 Atop the ridges, Company F sealed four caves only to have the Japanese blast them open from the inside. To add to the vexation and frustration of the Marines operating near the northern tip of Peleliu, the Japanese suddenly fired into the rear of Company G from caves that had been sealed on the previous day. Company E finally assaulted these positions and chased some of the Japanese onto the reef, where riflemen on LVTs annihilated them in short order. Nevertheless, in these caves an undetermined number of Japanese remained, capable of causing additional trouble when it suited them.

For all practical purposes, organized enemy resistance on northern Peleliu, except for the grimly defended pocket in the high ground north of the airfield, came to a virtual end on 29 September.74 The Japanese situation report for 29 September failed to indicate any great concern about events on northern Peleliu, mentioning only that ". . . on the front line, where our main forces are facing the enemy's main force, all is quiet; in fact, even some of our forces there are helping us out in the battle taking place in the north area of Peleliu."75 Twenty-four hours later, the tenor of the Japanese report changed, for in reference to northern Peleliu, Colonel Nakagawa reported that ". . . after a fierce battle, the enemy was finally successful in occupying the area. Our surviving forces are attempting to dash southward, cutting through the enemy in order to join the main force."76

On the morning of 30 September, the 1st and 2d Battalions, 5th Marines, continued to mop up the northern leg of the Amiangal ridges and Radar Hill. At 1000, both battalions were ordered to withdraw to the vicinity of the radio station prior to being relieved by elements of the 321st Infantry Regiment.

This relief took place during the afternoon of 30 September. While the Marines, somber and weary after their prolonged struggle for northern Peleliu, were en route to Ngardololok for a well-deserved rest, few imagined that more than a sprinkling of Japanese had remained in the Amiangal ridges. As it turned out, those Japanese remaining on northern Peleliu dug their way out of previously sealed caves, reoccupied new positions, and in a number of instances forced the bewildered Wildcats to retake positions that the 5th Marines had previosuly secured.TT At least two more days were required to end such resistance as remained in northern Peleliu, and even then isolated Japanese


continued to exist in dugouts whose entrances were partially blocked by debris.

During the struggle for northern Peleliu, the 5th Marines had killed and captured over 1,170 Japanese. Elements of RCT 321 accounted for another 175 following their relief of the Marines.78 The campaign for Peleliu during the last half of September had resulted in an estimated total of 9,076 enemy dead and 180 prisoners of war. During the same period of time, the 1st Marine Division had lost 843 killed, 3,845 wounded, and 356 missing, a total of 5,044 casualties.79 Except for the Umurbrogol pocket the entire island of Peleliu was in American hands. The last phase of the bitter struggle for Peleliu was about to begin.


Table of Contents ** Previous Chapter (III-5) * Next Chapter (III-7)


1. Unless otherwise noted, the material in this chapter is derived from: IIIAC Palaus Rpt; 1st MarDiv SAR; 1st MarDiv WarD, Sep44; 1st MarDiv D-2 Jnl; 1st MavDiv D-3 Periodic Rpt; Peleliu Comment File; Japanese CenPac Ops; Smith, Narrative; Hough, Assault on Peleliu; Morison, Leyte; Smith, Approach to the Philippines; McMillan, The Old Breed.

2. Additional sources used for this section are: IIIAC C-3 Rpts; 3/1 WarD; Vandegrift Letters.

3. Col William F. Coleman ltr to CMC, n.d., in Peleliu Comment File, hereafter Coleman ltr.

4. Ibid.

5. MajGen William H. Rupertus ltrs to LtGen Alexander A. Vandegrift (CMC), dtd 7Dec43, 4Feb44, 18Feb44, and 24Mar44, in Vandegrift Letters.

6. Wachtler ltr.

7. 1st MarHist, 23 Sep44.

8. 3/1 WarD, 23Sep44.

9. McMillan, The Old Breed, p. 319,

10. 1st MarHist, 26Aug-100ct44.

11. Additional sources used for this section are: IIIAC C-2 Rpts; 1st MarDiv FO 3-44.

12. LtGen Sadae Inoue, IJA, interview with LtCol Waite W, Worden, in Worden ltr.

13. 1st MarDiv SAR, Ph II, Anx A, p. 7.

14. Japanese CenPacOps, p. 109,

15. 1st MarDiv FO 3-44,

16. Additional sources used for this section are: IIIAC C-2 Rpts; 81st InfDiv OpRpts; 7thMar R-2 Jnl; VMF(N)-541 Hist; 2/11 SAR; 3/7 Rec of Events.

17. Japanese CenPacOps, p. 109.

18. 1st MarDiv SAR, Anx A, p. 7.

19. 2/11 AR, 24Sep44.

20. Smith, Approach to the Philippines, p. 537.

21. LtCol Edward H. Hurst ltr to CMC, dtd 15Mar50, in Peleliu Comment File, hereafter Hurst ltr II.

22. 7th Mar R-2 Jnl, 24Sep44.

23. 3/7 Rec of Events, 24Sep44.

24. Ibid.

25. 81st InfDiv OpRpt, 24Sep44.

26. Historical Committee, 81st Infantry Division, p. 139,

27. 1st MarDiv SAR, Anx A, p. 8.

28. Ibid.

29. VMF(N)-541 Hist, Feb44-Apr46.

30. 1st MarDiv SAR, Anx A, p. 8.

31. 7th Mar R-2 Jnl, 24Sep44.

32. IIIAC C-2 Rpt, 25 Sep44.

33. 81st InfDiv OpRpt, 24Sep44.

34. 81st InfDiv OpRpt, 25Sep44,

35. Additional sources used for this section are: 81st InfDiv OpRpt; 5th Mar URpts; 7th Mar War Jnl; 1/5 B-3 Jnl; 2/5 OpRpt; 2/11 SAR; 4/11 WarD, Aug-Nov44.

36. 5th Mar URpts, 26Sep44.

37. Smith, Narrative, p. 83.

38. Maj David R. Griffin ltr. to CMC, dtd 13Mar50, in Peleliu Comment File, hereafter Griffin ltr.

39. 1/5 B-3 Jnl, 25Sep44.

40. Maj John A. Crown ltr to CMC, dtd 13Feb50, in Peleliu Comment File, hereafter Crown ltr.

41. 1t MarDiv SAR, Anx A p. 10.

42. 81st InfDiv OpRpt, 25Sep44.

43. 2/5 OpRpt, 26Sep44.

44. 2/11 AR, 26Sep44.

45. 1st MarDiv WarD, Sep44.

46. Ibid.

47. 4/11 WarD, Sep44.

48. 7th Mar R-2 Jnl, 27Sep44, also 1st Mar Div SAR, Anx B, pp. 37-38.

49. 2/5 OpRpt, 27Sep44,

50. 2/5 Op Rpt, 27Sep44.

51. 5th Mar URpts, 28Sep44.

52. 1st MarDiv WarD, Sep44.

53. Ibid.

54. 321st RCT Unit Rpt 7, 30Sep44, Peleliu, as quoted in Smith, Approach to the Philippines, p. 549.

55. Additional sources used for this section are: 5th Mar URpts; VMF-114 WarD, Sep44; 2/11 SAR; 3/5 Rec of Events; Coleman ltr; Stuart ltr; Smith, Narrative.

56. LtCol Lewis W. Walt ltr to CMC, n.d., in Peleliu Comment File, hereafter Walt ltr.

57. LtCol Arthur J. Stuart ltr to CMC, dtd 25Apr50, in Peleliu Comment File, hereafter Stuart ltr.

58. 1st MarDiv SAR, Phase II, Anx A, p. 11.

59. VMF-114 WarD, Sep44.

60. 2/11 AR, 28 Sep44.

61. Coleman ltr.

62. 3/5 Rec of Events, 28 Sep44.

63. Ibid.

64. 5th Mar URpts, 29Sep44.

65. Smith, Narratitve, p. 90.

66. 3/5 Rec of Events, 29Sep44.

67. Smith, Approach to the Philippines, p. 548.

68. Additional sources used for this section are: 1st MarDiv WarD, Sep44; 1st MarDiv SAR; 5th Mar URpts; 1/5 B-3 Jnl; 2/5 OpRpt; Smith, Narrative.

69. 1/5 B-3 Jnl, 28 Sep44.

70. 2/5 OpRpt, 28 Sep44.

71. Ibid.

72. 1st MarDiv WarD, 28Sep44.

73. 2/5 OpRpt, 29 Sep44.

74. 1st MarDiv WarD, 29Sep44.

75. Japanese CenPacOps, p. 113.

76. Ibid. Smith, Approach to the Philippines, pp. 546-547.

78. Ibid., p. 548.

79. 1st MarDiv SAR, Phase II, Anx A, p. 13.

Transcribed and formatted for HTML by Patrick Clancey, HyperWar Foundation