Except for personnel and machines of the 710th Tank Battalion acting in support of the Marines in the Umurbrogol, RCT 321 had taken little active part in the Peleliu fighting since completion of its mission in the north on 2 October. Yet the men kept busy at useful tasks: building defense installations against possible counterlanding, burying enemy dead, clearing up the inevitable debris of battle. And, lest their part appear wholly passive, in one four-day period (4-8 October) they managed to kill 171 Japanese.1
On 7 October, the 2d Battalion, stationed on Ngesebus and Kongauru, began a series of minor amphibious operations along Peleliu's northern approaches. A number of small islands lay in this region, leading like stepping stones into the central Palaus, with the enemy quite possibly in occupancy. With the idea of finding out, Company F, reinforced, boarded LVT's on the early morning of 9 October and crossed the few thousand yards of reef to Garakayo, the nearest and largest. The troops encountered some machine-gun fire from a ridge overlooking the southern beach and from the neighboring island of Cordoray, but succeeded in securing it by 1500. Five Japanese were killed in the process and five more during the following night.
On the 10th Company G replaced F and set about platoon reconnaissance of the other nearby islands named with such characteristic Palaus tongue-twisters an Ngemelis, Arimasuku and Garyo. Whenever they were so fortunate as to find an island not already named, the soldiers happily christened it after one of themselves, with the result that their special map shows Murphy, Galligan and Turner among all the Ngargersiuls and Gorokottans. They found no more Japanese, however.
As preparations progressed for evacuation of the 7th Marines, the 321st occupied that regiment's defense sector as well as its own, and on 13 October the assault battalions were ordered to relieve the 5th Marines in the Umurbrogol on the 15th and 16th. All these dispositions threatened to spread the regiment rather thin, but this was somewhat compensated for by arrival and attachment of the 1st Battalion, RCT 323, from Ulithi, with the other two battalions scheduled to follow shortly.
"KANSOKUYAMA," the Japanese called this height. They had an observation post on the summit, reached by ladders leading upward through the interior cave system. Marines knew it as "Hill 300," the Wildcats as "Old Baldy."
With completion of the relief and assumption of command, the assault units of the 321st were deployed around the Pocket as follows: 3/321 held the eastern perimeter with positions along the crests of Walt and Boyd Ridges and extended below the mouth of the Horseshoe; 2/321 was atop Hill 140 and extending west from there, with the 1st Battalion on its right to the southward; 1/323 held the south-southwestern zone, facing the Five Sisters and Death Valley.
For two days (16 and 17 October) elements of the 2d Battalion tried vainly to get from Hill 140 to the top of No. 1 of the Five Brothers. The ravine between was too narrow, the ascent too steep, cover too scarce in view of the heavy and accurate fire from the enemy's commanding positions. On the battalion's right, 1/321 made an advance of 200 yards on the 17th but was stopped cold on the following day and commenced to sandbag its position and bring up pack howitzers.
The tactical importance of gaining a foothold on the summit of the Five Brothers was self-evident to all hands; and RCT 321 was especially anxious to have such an achievement to its credit by the time the 81st Division assumed command. Authority was therefore obtained from the 1st Marine
TANK-INFANTRY ACTION inside the Horseshoe.
Division (which, as has been noted, retained over-all command until 20 October) to stage an all-out effort on the 18th.2
To support this assault, tanks and LVT flame-throwers entered the Horseshoe over a track which had been bull-dozed through the draw north of Walt Ridge where they could bear directly on many of the Japanese positions capable of angling fire against the narrow, exposed crest of the Brothers: from the lower slopes of Walt Ridge, Hill 300 and the Five Sisters. Then heavy concentrations of 81mm and 4.2 inch mortar fire were placed on the Brothers themselves.
Under cover of this Company E succeeded in scaling Brother No. 1 at 1100, and 45 minutes later took No. 2. In an effort to exploit these gains while the enemy was still off balance, Company F passed through Company E and was established on the summit of No. 3 by 1315.
At this point, however, success came to an end. On that narrow crest the men were frightfully exposed, with Japanese on three sides of them and no time for the laborious process of bringing up sandbags. Before Company F could consolidate its position, the men found themselves pinned down by a vicious cross-fire from enemy emplacements which the tanks in the Horseshoe
were unable to get at. Under cover of this, the Japanese staged a strong counterattack from the two southernmost Brothers and by 1700 succeeded in driving the men of the 2d Battalion from all three of the seized crests.3
With establishment of the 81st Division command post on 20 October, the top Marine echelons departed at once by air: General Geiger and the corps staff that morning, General Rupertus and the division staff at 2300. In parting, CG, 1st Marine Division, paid tribute to the Army unit which had fought with him so long and so well in the following terms:
"The performance of duty of the officers and men of Regimental Combat Team 321, throughout the assault phase on Peleliu, Ngesebus and the northern outpost islands warrants the highest praise. It was a pleasure to have this unit serve as part of my command during this extremely difficult operation, and I express the sentiments of every officer and man of the 1st Marine Division in wishing them good luck in future operations against the enemy. In the eyes of the entire 1st Marine Division they have earned a 'Well Done.'"
With his assumption, of command, Major General Paul J. Mueller announced that he proposed undertaking a major seige operation, and that was precisely what he did. During the initial phases of an amphibious operation, time is so all-important that often heavy losses must be sustained in order to capitalize it. But this period was long since past on Peleliu. From now on there would be no more ambitious but premature assaults such as that on the Five Brothers. Reduction became a step-by-step business; grim, methodical, inexorable; characterized more by hard work than hard fighting. Not a life would be lost that could possibly be spared.
The number of Japanese remaining in the Pocket was unknown, of course, and estimates varied widely. Asked by the 81st Division for a tentative figure, the 1st Division staff guessed roughly 500, raised by other guessers to a maximum of 1200, "a figure which was subsequently ascertained to be underestimated."4 At the end of the operation the 81st reported: "Over 1500 Japanese were known to have been killed and fifty-eight captured",5 an estimate which would appear to disregard the possibility that a few of these might have been disposed of by the Marines during the month they had fought over this terrain. But three days after the Wildcats took over, Colonel Nakagawa, whose reports were seldom exactly pessimistic, informed his superiors on Koror: "Our total garrison units number about 700 soldiers, including the slightly wounded;" and six days later: "Garrison units number about 500 still able to fight."6
Whatever their numbers, fight they did, and in the best Japanese tradition.
Although many Marines7 remained on the island following termination of Marine Corps command, their part was minor. After 20 October the story of Peleliu is the Army's story. Army agencies have told it elsewhere and told it well. It will be only summarized here in order to round out the picture.
The 81st Division employed only two of its regimental combat teams on Peleliu. The 322d Infantry remained mopping up and in garrison on Angaur, though as matters straightened out on that island most of its reinforcing elements were detached and transferred to the scene of the main fighting. All elements of RCT 323, less a small detachment left to garrison Ulithi, had arrived by 26 October, on which date they
relieved the weary 321st, less 3d Battalion which remained on the now comparatively inactive eastern perimeter.
Thereafter these four battalions carried on the Pocket battle to the bitter end a month later, though 81st Cavalry Reconnaissance Troop participated in two more island seizures to the north: Gorokottan on 11 November and Ngeregong on the 15th. The latter island had been seized previously by personnel from the LCI gunboats operating in the region, but these had been chased out by an enemy counterlanding. The Reconnaissance Troop went in prepared for any contingency, with artillery and air preparation and LVT(A)'s and LCI's for direct support. But only three Japanese were found, all of them dead.
Back in the Pocket the methodical attrition continued. No one was in any hurry, and the Wildcats' caution was exceeded only by the thoroughness with which they did the job. Mortar concentrations and napalm air strikes prepared the way for each major attack, and extraordinary labor was undertaken to bulldoze routes for tanks and LVT flame-throwers to inaccessible enemy strong points. At times a satisfactory day's advance measured ten to fifty yards, and there were cases of men crawling ahead on their bellies pushing sandbags in front of them with poles. If any one implement could be said to have done more than any other to seal the fate of the Japanese in the Umurbrogol, it would be the humble sandbag.
The Marines' idea of emplacing heavy weapons in strategic high ground for direct fire support was elaborated and ramified by the 81st. There seemed no limit to the Wildcats' ingenuity, especially that of the engineers. When no amount of labor would suffice to get LVT's into one particularly troublesome area, they rigged up what must stand as one of the most unique flamethrowers in anybody's war: a 300-yard pipe line leading to the target from a fuel truck parked on the West Road, complete with booster pumps to insure pressure and equipped with a nozzle which enabled the
TWIN VALLEYS THAT WERE DEATH TRAPS: Horseshoe (left) and Wildcat Bowl, walled in by Walt Ridge, Five Brothers and China Wall. Japanese command post near northern end of the latter.
operator to play flame on the Japanese positions like water from a hose.
A battery of flood lights was mounted so as to focus on the pond within the Horseshoe which served as the enemy's only stable water source. And near the end the 306th Engineer Battalion built a coral ramp that enabled tanks to get from the floor of Wildcat Bowl to the summit of the rugged China Wall. Some of the conveyor systems developed to get supplies up into the ridges and to evacuate the wounded looked like pure Rube Goldberg.
The seige developed according to a traceable design but seldom exactly according to plan. The first heavy pressure was applied from the north and northwest. Elements of 2/321 got back to the top of Brother No. 1 three days after they had been thrown off on 21 October. This time they had seen to it that ample sandbags were readily available; these were passed to the crest hand-to-hand via human chain, and the position was
secured. The following day No. 2 was taken and sandbagged in turn, and on 23 October Nos. 3 and 4 got the same treatment. No. 5, lower in elevation, was dominated by strong positions still enemy held and hence let alone for the time being.
On 22 October elements of 3/321 came down from Walt Ridge and sandbagged a line of defense on the floor of the Horseshoe from which they were never driven, thereby shortening the perimeter considerably on the east. But with the attack stalled in the west and northwest by a combination of terrible terrain and strong resistance, interest next shifted to the south.
Rain and poor visibility stopped offensive operations for the first few days after RCT 323 took over, but on 2 November that regiment's 2d Battalion, attacking in the wake of napalm and mortar preparation, seized the Five Sisters and Hill 3008 with comparative ease, sandbagged them and hauled up pack howitzers. But an attempt to move a tank-infantry patrol up Death Valley was repulsed the following day.
Attrition continued to gnaw away, hampered by the weather9 as well as the Japanese. Armored bulldozers concentrated on the approaches to the two main declivities, and on 13 November tanks and an LVT flame-thrower, supported by infantry, made the first successful sortie into Wildcat Bowl. This was part of an all-out push from every direction and succeeded so well as a diversion that Company G was able to get from the Sisters onto the southern end of the China Wall against light opposition. This curious formation was really a double ridge, the serried pinnacles along either flank separated by a narrow, jumbled gulch which required something approximating alpine tactics merely to negotiate. By the time the men had toiled and scrambled 75 yards along this, they came under heavy fire from positions farther northward, and the invaluable sandbags were brought up.
A similar gain was made by Company B, attacking from the west, but the advance of the 3d Battalion from the north could be measured in feet, when it could be measured at all. The terrain here was very bad, but certainly no worse than atop the China Wall, and the opposition determined. The reason was surmised at the time and later verified: nearby lay the hard core of all Japanese resistance, the CP of the doughty Nakagawa, who was not a man to neglect defense of his approaches.
The rest of the story is one of unceasing pressure from the north, south and west, with the resulting gradual attrition. Units in assault were relieved at short intervals, and the sandbags continued creeping forward. The action was grinding and monotonous, with few highlights.
By 17 November, tanks had worked onto the southern portion of Death Valley with the aid of armored bulldozers, and by dint of great exertion commenced slugging their way northward from there. Tank-infantry patrols operated virtually at will within Wildcat Bowl, with results that must have been highly discouraging to the enemy. By 21 November it became apparent that the Japanese, moving under cover of darkness, had withdrawn altogether from their remaining caves in the western flanks of the Brothers to concentrate for a last stand in the seemingly impenetrable China Wall.
By taking quick advantage of this development, the Wildcats achieved their most significant
gain of the past two weeks. With advance of the 3d Battalion from the north still held up by impossible terrain and desperate resistance, Company F, supported by tanks and LVT flame-throwers, commenced a sortie up the length of the Bowl at 0700 on 22 November. Forty-five minutes later the foremost elements reached the northern extremity and commenced the precarious ascent of the China Wall itself. By the time the Japanese discovered what was up and were able to react, the company was consolidating a tactically important position on the high ground immediately south of the 3d Battalion lines, trapping those enemy remaining in the area between the two units and imperiling those holding up the drive from the west: above and beyond Death Valley.
Simultaneously with this action, Company H attacked through Company G in the 2d Battalion's position atop the China Wall to the south and sandbagged a 75 yard gain from this direction. And then, as though to prove the validity of the new development, Company I secured the last of the Five Brothers without encountering any opposition. This permitted the containing line to close across the saddle between the Brothers and Sisters, greatly reducing its length.
At last the pattern of Victory was taking shape for the final kill. With a foothold secured at the northern end of the China Wall, Company A, 306th Engineer Battalion, commenced construction of that remarkable coral ramp from the floor of the Bowl to the crest of the double-pinnacled ridge as a means for bringing flame and armor against the enemy's last and heretofore inaccessible strong points.
During that night (24-25 November) 45 Japanese,10 including two officers, were killed in what proved to be the dying convulsion of organized resistance on Peleliu. The story was obtained from a prisoner captured at about 1100 the following morning. According to his account, Colonel Nakagawa and Major General Kenjiro Murai11 had ordered this last attack in an effort to break through the beseiging cordon and conduct guerilla warfare against the rear areas. They had then notified headquarters to the effect that the end had come, burned the regimental colors,12 and ceremoniously shot themselves.
Although no other witnesses remained to verify the suicides, the rest of this characteristically Japanese scene is substantiated by the opening phrases of the last message received from the Peleliu garrison: "Our defense units were on the verge of being completely annihilated; therefore, the unit destroyed the 2d Infantry Regiment flag which they had in their possession. . . . All documents were burned."13
Throughout the day of the 25th the circling lines continued their inexorable closing, moving slowly, searching the jumbled ground with infinite care to insure that not a single enemy be overlooked. The 1st Battalion, advancing from the west, flushed 30 Japanese and captured the prisoner referred to above. Twenty-one more were killed during the night in what appeared to be individual attempts to escape rather than concerted action.
The 306th Engineers' ramp was completed late that afternoon, and first thing the
NAPALM STRIKES, such as this against the Five Sisters, were spectacular, but, as the Japanese commander reported, "Since perfecting our defenses against fire, we suffered no losses."
following morning tanks and an LVT flamethrower began the tricky negotiation of this route to the ridge's summit. Otherwise action on 26 November was a repetition of the previous day's, with the strongest resistance still encountered in the area at the head of Death Valley.
No fireworks, literal or metaphorical, signalized the end of one of the most stubbornly contested campaigns in the history of this or any other war. The units in assault jumped off in all zones at 0700 on 27 November. As the 81st Division Operation Report puts it succinctly: "Resistance to this multisided attack seemed to disintegrate completely." At 1030 elements of the 2d Battalion, moving northward along the China Wall, came face to face with their comrades of the 3d Battalion, working southward. Across the few yards intervening, they could see men of the 1st Battalion perched on the rim of Death Valley. For long moments the weary Wildcats looked at one another in an uncanny silence, trying to realize that this was all there was; there wasn't any more.
At 1100 Colonel Arthur P. Watson, commanding RCT 323, reported officially to General Mueller that the Peleliu operation was over. "The enemy had fulfilled his determination to fight to the death."14
Six Weeks Attrition by
81st Infantry Division
1. 81st OpnRpt, 11, 45. Many of these had been lurking in caves which had been overrun, but others, discovered on the beaches evidently trying to escape from the island, were believed to have infiltrated from the pocket. Such would-be deserters turned up in increasing numbers as the long grind wore on, their occurrence probably symptomatic of the deterioration of Japanese morale under the remorseless pressure.
2. O. P. Smith PerNar.
3. The narrowness of these razorback crests made it impossible either to fight effectively or to withdraw in an orderly manner. BrigGen O. P. Smith witnessed this action through glasses from below and records: "All of the men did not withdraw north up the ridge, but many of them slid down the steep slopes of the Brothers into Horseshoe Valley below." Smith PerNar.
4. 81st OpnRpt, II, 6.
5. 81st UnitHist, 200.
6. Tada Record, 23Oct, 26Oct44.
7. Notably the aviation units operating from the airfield, service troops attached to the Island Command, two battalions of artillery (4/11 and 3d 155mm Howitzer Bn.) and crews of the LVT flamethrowers.
8. Here that bugaboo of variant place names arises again. The Marines thus called this elevation and considered it a part of the Five Sisters formation though it was separated from the others by a fairly wide saddle. The Wildcats called it "Old Baldy", and it so appears on all their maps, which gives rise to confusion with the Marines' "Baldy Hill" nearly 1,000 yards to the north. Other names applied by the 81st were Wildcat Bowl, China Wall and Grinlinton Pond, none of which had been named previously. They also renamed the Horeshoe Mortimer Valley in honor of Captain Joseph F. Mortimer of the 323d Infantry who was killed in action there. But where the Marines had operated before the new names were applied, this account will continue to adhere to their terminology.
9. Rain accompanied by the edge of a typhoon halted offensive action altogether 4 to 9 November, and repeated spells of foul weather constituted a minor nuisance until the end of the campaign.
10. There were doubtless a few isolated individuals still holding out in the region, but the garrison commander reported only 56 men under a Capt Nemoto as of 1700 on 24Nov. Tada Record.
11. This was the only intelligence obtained during the entire operation which placed General Murai on Peleliu. The Japanese CP occupied a deep, two-chambered cave 40 feet down a vertical shaft near the northern end of the China Wall, but this was not fully explored until the bodies it contained had decomposed beyond positive identification. For further elaboration on this minor mystery, see Appendix F.
12. In Japanese military tradition, the burning of any unit's colors signified that that unit had ceased to exist and should be stricken from the Army lists.
13. Tada Record, 24Nov44. Text of the message as eventually relayed to Tokyo read: "The splitting of the men into 17 teams was completed at 1700 hours. Following the commander's wishes, we will attack the enemy everywhere. This will be the last message we will be able to send or receive."
14. 81st OpnRpt, 97.
Transcribed and formatted for HTML by Jerry Holden for the HyperWar Foundation