In the minds of the men who fought there, the exotic-sounding name "Umurbrogol Mountain" became associated with some of the most unpleasantly exotic terrain on the face of creation. Marines had been fighting in this outlandish high ground since D-plus 2, and some attempt has been made to describe it. But words are inadequate, photographs not much better. One has to see it fully to believe it.

The generally accepted geological explanation is that Peleliu was once a part of the ocean floor which, thickly encased in coral growth, had been forced above the surface by subterranean volcanic action. Many areas of the island still retain a semblance of the natural submerged coral formations to be seen on any tropical-reef. But where the pressure had been strongest, the ground had buckled and cracked to form a maze of ridges and defiles, the whole littered with jagged boulders and rubble which had been torn adrift by the violent action. This pressure accounts for the elevations attaining no great height before splitting apart. It also accounts for the broken nature of the terrain and the extreme steepness of most of the slopes, many of them sheer cliffs. The thin topsoil above this bedrock had been capable of sustaining scrub jungle just dense enough to screen the contours from aerial photography, thereby rendering the preliminary maps of this particular region so egregiously inaccurate in detail.

The same pressure which caused these weird surface contortions had created innumerable underground faults in the coral limestone; cracks and fissures in the solid rock, many of which had been further enlarged by erosion into natural caves, cluttered with the stalactities and stalagmites characteristic of caverns in all varieties of limestone. Since rock faults seldom follow any traceable patterns, these natural caves occurred in all shapes and sizes, and at every conceivable level from the floors of valleys, up the faces of sheer cliffs to the tips of craggy pinnacles. It was the Japanese exploitation of these natural features which caused the fighting in the Umurbrogol Pocket to develop as it did.

The terrain within the Pocket, as defined at the end of the third operational phase (1 October), differed only in degree from that in the Kamilianlul and Amiangal systems to the northward: the ridges were higher, longer, and packed more closely together in greater depth. The highest elevation achieved an altitude of only about 300 feet, but that was almost straight up. No caves here were as large and elaborate as some of the engineering feats encountered

Map 14
Umurbrogol Pocket

STALACTITE FORMATION in natural cave "improved" by the Japanese and "revised" by the Americans.

in the northernmost ridges--but there were a lot more of them.¹

Natural caves, of course, did not always occur in positions of tactical importance, and a number of them were taken without great difficulty or abandoned by the enemy. A case in point was Colonel Nakagawa's original command post, located in a "balcony"² cave well to the south of the final Pocket. This boasted most of the comforts and conveniences of permanent quarters: wooden decks, electric lighting, communication facilities, a well equipped galley, partitioned quarters with built-in bunks. Yet because he rightly judged its approaches to be indefensible, the colonel evacuated it on D-plus 2 and retired without haste to that inaccessible position which was destined to be the last hard core of organized resistance on Peleliu.

However, where a natural cave did possess tactical importance, especially if it were adaptable to the emplacing of heavy weapons,

the Japanese went to great lengths to improve it. Blast walls, or revetments, of logs, coral-filled oil drums and reinforced concrete were placed to protect the entrance, or entrances, from direct fire or the force of nearby explosions. A few were even equipped with doors of steel armor plate which could be swung open in order to run out artillery pieces, mortars or automatic cannon, then closed again to provide cover and concealment after a few rounds had been fired. Elaborate communication equipment was subsequently discovered in several of these.

All approaches were protected with great shrewdness by the construction of artificial, or semi-artificial, caves so placed as to bring mutually supporting and interlocking fires to bear against the flanks or rear of attackers. In this respect there were limiting factors in the amount of work required and shortage of labor personnel.³ As a result, many such were so small and simple as to prove death traps once the Marines were able to get to them.

But many others proved quite the opposite. They assumed a number of shapes and sizes, depending on local conditions. Wherever possible, bays and ells were blasted or chiseled in which the occupants could take refuge from direct fire through the cave's mouth. The ultimate desideratum was another entrance, or entrances, on the opposite side of the ridge or out of view of the attackers around a shoulder, through which the defenders could duck out for a safe cigarette or two during such time as the Marines were earnestly blasting and burning their empty refuge. In that region of jumbled razorback ridges, construction of multiple entrances proved practicable in a surprising number of cases. Thus, during the earlier stages of the operation, there were repeated instances of caves being declared secure in all good faith, whereas in actuality they were reoccupied by their unharmed defenders within minutes of the time the attack ceased, and the whole business would have to be done all over again.

But if the Marines learned the hard way, at least they did learn. As the fighting developed in the Umurbrogol, it became standard procedure to search out and seal every opening bearing the remotest resemblance to a cave mouth, regardless of whether there was reason to believe it occupied. No doubt much energy and high explosive were wasted in this manner, but by the time a systematic job had been completed, attacks proceeded much more smoothly. There was simply no easy, inexpensive way.

The impossibility of judging the nature of a cave or the number of its defenders from the outside doubtless led to some exaggerated estimates of enemy casualties. But sometimes this had the opposite effect. There is one authenticated case of eight Japanese being reported killed defending the mouth of what appeared a smallish cave, which on subsequent investigation proved to be an enormous cavern containing the bodies of 80.

Greatly enhancing mutual support of enemy positions was the generally parallel orientation of the ridges. Following the peninsula's northeast-southwest axis, they faced one another, often at very close range, across a succession of gullies, draws, box canyons and miscellaneous declivities; steep-sided, fissured and strewn with those ubiquitous boulders and coral outcrop. Thus, Marines on one narrow, open ridge crest were vulnerable to fire from concealed Japanese in the faces of the flanking ridges, even though there might be friendly troops atop those ridges, too, prevented by the steepness of the slope from getting at the enemy below them. There were numerous instances of such troops lowering explosives over the cliffs on ropes in an effort to explode them in the cave mouths below--and of the Japanese reaching out and cutting the ropes.

From the infantryman's point of view, perhaps the greatest curse of this evil ground was the limitations it imposed on the effectiveness of supporting arms. Flat trajectory naval gunfire had proved valueless against invisible targets impervious to anything short of a direct hit, and most of them in defilade, anyhow. Massed artillery fires were useful in keeping the enemy holed up when that suited the attackers' purpose, but any permanent effect it had was wholly a matter of luck.⁴ And with constriction of the Pocket, massed fires had to be abandoned to avoid danger to friendly troops.

Direct fire by individual heavy weapons could be extremely effective,⁵ but all too often no amount of labor or ingenuity sufficed to get such weapons into the desired positions. Tanks were invaluable wherever they were able to operate; but in such up-ended terrain, this excluded many of the most critical areas. The same applied to the LVT-mounted flame-throwers, and prodigious time and labor were expended bulldozing routes for both types of machine into some highly improbable territory. But in the end the business of reduction usually resolved itself into the infantryman struggling in across all but impassable ground with such weapons as he could carry on his back.

The air arm helped, or tried to. Five hundred and 1000-pound bombs made an impressive display and doubtless rattled the teeth of many underground Japanese, but caused no visible slackening in enemy resistance. Napalm fire bombs were tried. Later, in order to minimize danger to friendly troops in the constricted Pocket, unfused belly tanks filled with napalm were dropped to saturate a chosen target area which was then ignited by lobbing in white phosphorus mortar shells. There is at least one substantiated report of the intense heat driving a small group of Japanese into the open where they could be destroyed. But perhaps the most useful purpose served by napalm was burning away what vegetation remained, thereby depriving the enemy of concealment and occasionally revealing positions heretofore unsuspected.⁶

The Umurbrogol Pocket, as defined at the time responsibility for its reduction passed to the 7th Marines (29 September), has been described as roughly 900 yards long by 400 yards wide.⁷ Certain of its southern features have been touched upon in passing in Chapter IV, but may be reviewed briefly in connection with the accompanying maps and photographs.

The valley known as the Horseshoe⁸ offered what appeared the most practical avenue of ingress into the heart of the Pocket from the southeast. This was bounded outboard by the elevation which came to be called Walt Ridge, which formed the Pocket's eastern wall in this area; inboard by a generally parallel formation: the Five Brothers. Ringing it to the north, three other major elevations frowned down into the valley at various angles: Hill 140, Ridge Three and Boyd Ridge, the latter also part of the eastern wall and separated from Walt

UMURBROGOL POCKET from the southeast, Walt Ridge center foreground, Boyd Ridge to its right north of draw. Smoke rises from western face of Five Brothers, north of which lie Hill 140 and Ridge 120. Five Sisters at left, abutting China Wall in background. Swamp-surrounded inlet in foreground prevented approach from eastern peninsula.

Ridge by a steep-sided draw. The East Road, approaching from the south parallel to the extension of Hill 200 (see Chapter IV), angled sharply eastward near the base of towering Hill 300 and crossed the swamp at the mouth of the Horseshoe on the causeway previously described, just beyond which a secondary road branched off to a dead end inside the valley.

West of the Horseshoe and separated from it by the Five Brothers lay a generally similar valley at a somewhat higher elevation subsequently christened "Wildcat Bowl."⁹ Bounding this on the west, in turn, was the towering, butteressed China Wall, a double-crested ridge extending northward from the Five Sisters. The latter formation, together with Hill 300, partially closed the Bowl to the south, while an unnamed transverse ridge closed it completely on the north. The narrow, rough defile known as Death Valley skirted the western base of the China Wall, separating it from the serried high ground of jumbled coral formations which reached almost to, and dominated, the West Road.

North of the Horseshoe the ridges were piled closer to one another, and the defiles were narrower: Wattie Ridge, Baldy, Baldy Ridge, Ridge 120, the Three Knobs (see Map 14). Here was concentrated some of the heaviest and ultimately most successful Marine action.¹⁰

When Colonel Nakagawa chose to fight a defense to the death in prepared positions in the Umurbrogol, he automatically sacrificed anything resembling mass mobility for his force.¹¹ Thereafter the Japanese potential for jeopardizing island security on any considerable scale ceased to exist. The sole object became to make the conquest as costly and time-consuming as possible: to kill for the mere sake of killing, and to pin down U.S. troops against their employment elsewhere.

THE HORSESHOE as it appeared during the later stages. Five Brothers (left), Walt Ridge (right), Hill 140 (center background).

To this end, the defenders followed the general policy of obliging the Marines and soldiers to come to them, often exercising an extraordinary degree of fire discipline to avoid giving away their positions. One could stand at the Horseshoe's mouth in broad daylight and study at leisure the precipitious slopes and sheer cliffs that wall it in, almost physically aware of the weightless impact of scores of hostile eyes. Yet there would be no sign of the enemy: no movement, no shots; only a lonely silence. Occasional small patrols operated there, and in similar spots elsewhere, with impunity. But let an important position be approached, or a sufficient force be committed to furnish a profitable target, and a fury of fire of all varieties would be angled in from every direction and every altitude the hills provided.

Nights were something else again. Infiltrating Japanese, as in all Pacific operations, constituted a continuous nuisance and minor menace. Mass mobility might be no longer possible, but small-scale activity and limited counterattacks were the rule throughout the Pocket. Lurking in concealment by day, with the coming of darkness the little men swarmed out of their holes like so many gophers, bent on raising hell. No doubt, some of their actions were dictated by the whims of individuals, but there was ample

evidence that many parties were acting under orders in carrying out predetermined missions.

As always, the infiltrators' object was to cause all possible damage and confusion in the Americans' rear areas. They went about this in two principal ways.

Individuals and small groups sought to reoccupy positions previously taken, as was noted previously in connection with earlier phases of the operation. Owing to their complete familiarity with the region, they usually knew exactly where they wanted to go, the best route for getting there, and the nature of the particular installation they wished to use. Moving under cover of darkness, through weirdly jumbled terrain of the ridges, they often reached their destinations despite all efforts to intercept them. Once ensconced in their chosen hideouts, they would snipe at targets of opportunity until their inevitable destruction, often managing the business so shrewdly that they remained undetected for days.¹² Conversely, of course, the more cautiously they operated, the less damage they managed to do. It is an ironic commentary on the effectiveness of infiltration as a tactic that most of their few victims were noncombatant souvenir hunters who were turning out in such hordes as to constitute nearly as great a nuisance as the Japanese infiltrators.

The presence of so many of these gentry was largely the result of the inability of the military police company to maintain an effective straggler line through having been weakened by evacuation of one of its platoons with the 1st Marines. While a few 1st Division Marines no doubt were guilty of this practice, mostly it was confined to aviation ground personnel, sailors ashore from the transports, and miscellaneous service troops with nothing better to do.¹³ Disgusted front line Marines did not much care whether these people got themselves hurt or not. The trouble was that every time one was wounded, good men had to expose themselves to the same peril in order to rescue him.

In justice, however, it should be recorded that not all of the men, who had no business being forward were motivated solely by souvenir hunger or morbid curiosity. As the struggle dragged on and casualties mounted, a feeling of unrest and insufficiency developed among troops in the rear areas. Many such who could get away from their assigned duties came forward with no other object than to lend a hand to their embattled comrades: bringing up supplies, serving as stretcher bearers, helping man the containing lines. Writes the battalion commander of 2/7: "I believe that . . . the unknown officers and men from the Marine air units who slipped away from the field and plagued me daily to let them participate 'in any way' should be mentioned."¹⁴

The second objective of Japanese infiltration was potentially more serious. Groups laden with demolition charges repeatedly attempted to penetrate the lines, evidently with the intention of getting through to the airfield and destroying planes and installations there; or, failing that, to blow up any artillery, tanks or other vehicles encountered along the way. Although they did some trifling damage, none ever reached the

airfield.¹⁵ Most such parties were stopped short of the lines, many being blown to kingdom come by their own explosives. While this kept U.S. nerves on edge, it was certainly an easier method of destruction than mining them out of the ridges and canyons.

Reduction of the Pocket evoked a high degree of the proverbial Yankee ingenuity and brought into play, at one time or another, every weapon at the Marines' disposal, plus a few improvised for the occasion. And when the job devolved on the 81st Division, the Wildcats not only improved on some of their predecessors' techniques but dreamed up a few new ones of their own. Many of the lessons learned here would be found useful in subsequent operations; but if Peleliu, Iwo Jima and Okinawa proved anything at all, it was that there is no single over-all tactical solution to Japanese cave fighting: no easy way.

The story of the Umurbrogol comprises the sum of the stories of the units which operated there, and will be developed here in connection with each of these in turn. To understand the nature of the fighting, it will be well to note at the outset that it was conducted for the most part by small units: "With the exception of a very few cases, most of the objectives were taken by platoons or squads of men, for as a whole the fronts were small and the terrain too irregular for company operation and maneuver."¹⁶ Note, too, that by now the toll of casualties had reduced the Marine units far below T/O strength: the terms squad, platoon, company, battalion did not designate the same number of men they had two weeks earlier.

HILL 200: Marines on top were unable to get at holed-up Japanese in the precipitous slopes below. Airfield in background.

Operations of the 7th Marines

As recorded in Chapter IV, the 1st Battalion, 7th Marines, relieved those elements of RCT 321 facing the northern perimeter of the Pocket on 29 September, at which time responsibility for operations throughout the Umurbrogol passed to 7th Marines command.

This date found the depleted regiment deployed thinly to encircle the Pocket on three sides, the dense swamp containing it on the east as effectively as any number of troops could have done. The western ridges were lower and less well defined than those surrounding the Horseshoe and Wildcat Bowl. In many places they could scarcely be called true ridges, being rather coral high ground, pitted, creased, broken into a score of crazy cross-compartments, box canyons, sump holes and low but sharp precipices. It was recognized from the first as the most

impractical terrain in the entire region over which to attack, and the line established here was pushed inland just far enough to safeguard the West Road against enemy fire. Thereafter until almost the end of Marine participation action here was purely defensive, and the line remained largely static.

It may be noted in passing that this particular line was by no means so sharply defined as the maps might indicate. Some sectors were manned in strength only at night. At one time or another, for various reasons, portions of it came down out of the hills altogether and were positioned along the West Road. Mapped locations are intended to be only approximate, or indicative.

On 29 September the greater part of this line was held by the 3d Battalion where it had extended the front up the peninsula behind the advance of the 321st. But, as will be seen, within a few days this area was turned over to detachments from various supporting arms in order to make all possible assault infantry available for offensive operations.¹⁷

The mission of containing the Pocket to the south had been assigned to the Regimental Weapons Company, which also acted in support of the several offensive drives from this direction. This unit took up a position facing the mouth of the Horseshoe across the swamp, its left extending in front of the bases of Hill 300, and the Five Sisters, at a respectful distance. This area was the scene of the enemy's strongest and most persistent infiltration attempts, both because its terrain was adapted to such tactics and because it provided the most direct route to the airfield, the infiltrators' main objective. To cope with this, personnel from Regimental Headquarters and Service Company, detachments from other units, and individual volunteers were brought up to reinforce the line at night.¹⁸

The 2d Battalion was atop the ridge on the Weapons Company's left, a northerly extension of Hill 200, in which the Japanese were still holed up in some strength and expended much time and energy attempting to infiltrate. The line followed this crest to a point where Colonel Berger's command post faced the Five Sisters and China Wall at a range of 25 to 40 yards across the mouth of Death Valley. Beyond here it angled off across the coral badlands toward the West Road to tie in with the right of the 3d Battalion. The 2d Battalion had occupied this position since 21 September, playing what was intended as a relatively inactive role; yet some indication of the peril prevailing in this area is conveyed by the fact that in the nine days preceding its relief, 2/7 sustained 149 battle casualties.¹⁹

The initial offensive plan of the 7th Marines for reduction of the Pocket called for the 1st Battalion to drive southward along the East Road and the ridges immediately adjacent to it, with elements of the 3d Battalion cooperating on the right. To make this last possible, it was necessary to stretch the containing line still thinner (the battalion front covered 1150 yards at this time) and for a short period to divide 3/7 into two separate task forces: one to defend, facing east; the other to attack southward.

Map 15
Umurbrogol Pocket
20 Sept-1 Oct

In the 1st Battalion Zone, Company B jumped off on schedule at 0800 on the morning of 30 September, and by 1025 had accomplished its mission: seizure of a high ridge west of the road. Here it was in position to support an attack down the road by Company A, which was to leapfrog at this point in order to secure the next ridge to the south.

However, heavy rain and low-lying fog made visibility so poor that the second phase of the attack was delayed until 1245. The intervening time was utilized by an engineer demolition team in sealing all caves in the overrun territory. As soon as visibility permitted, a mortar barrage was laid on the second objective, and Company A, supported by a tank and an LVT flame-thrower, attacked down the East Road to such good effect that an advance of 300 yards to the south and 150 yards inland was achieved. Company C moved in to help hold the captured territory, and a platoon of engineers was sent forward to aid in preparing it for defense.

A side light on this action offers ground for interesting conjecture regarding the enemy's plans and motives.

During this advance [of Company A] enemy troops were observed running down the southernmost valley . . . in small units of four to eight men. . . . Subsequent knowledge of the general situation as of D+15 indicates that our forces had reached the northern outposts of the enemy final defensive area. The attempted movement of enemy troops north during A-1-7's attack was probably from a central reserve, portions of which were committed to counter our encroachment anywhere on their perimeter.²⁰

The attack group of the 3d Battalion had been extending eastward throughout the morning in order to narrow the front of the 1st Battalion for the latter's attack. But the same atrocious weather conditions which delayed 1/7 had the effect of causing the two units to lose physical contact. Company L sent one patrol groping through the rainy fog in an effort to tie in once more. Another, moving forward farther to the west, came face to face with a formidable obstacle, christened descriptively "Baldy Hill": a rocky eminence rising like a high wart on the northern nose of a long, narrow ridge. As they explored possible approaches for an attack, members of the patrol sighted a number of Japanese on the hill and shortly thereafter were brought under heavy mortar fire. With failing visibility and the lateness of the hour dictating an end to the day's operations, both patrols were withdrawn for the night.

There was no notable advance the following day (1 October). The 3d Battalion group, with Company L in assault, moved out bright and early (0720) with the aim of rectifying the lines prior to an attack on Baldy. But the forward movement had progressed no more than 75 yards before the men came under such a heavy volume of rifle and machine gun fire from that formidable hill as to indicate the inadvisability of frontal assault. A supporting 155mm barrage had to be called off because of fragmentation falling in the lines.

Nor did it prove possible to establish physical contact between the battalions. As one observer noted: "The terrain in the gap was very rocky with numerous sheer cliffs and difficult draws."²¹ Visual contact was reported at 1034, and that had to suffice.

On its own part, the 1st Battalion was in no shape to be of much help. A check-up showed only 90 riflemen remaining fit for duty.²² These held a front extending into the swamp on the left, across the road and approximately 150 yards inland over the southern nose of the unnamed ridge immediately north of Boyd Ridge and separated from it by a deep draw. Clearly incapable of further offensive action, the battalion was overdue for relief, and steps were taken to effect this the next day.

The fact is that the entire 7th Regiment was in deplorable condition, with casualties

approaching the ruinous percentage which had necessitated evacuating the 1st Marines from the island. Even Major Hurst's 3d Battalion, which had suffered least to date, reported combat efficiency below 50 percent for the first time during the campaign. Lieutenant Colonel Berger reported his 2d Battalion's combat efficiency at 30 percent, adding gloomily, "The men are very tired."²³ Yet these weary, battered remainders gathered themselves for one last all-out effort which provided the 7th Marines' greatest contribution to the pinching out of the Umurbrogol Pocket.

In order to make "fresh"--if you could call them that--assault troops available, some extensive regrouping was necessary. Detachments of artillerymen--"Infantillery", they liked to call themselves--took over those sectors of the containing lines held by 3/7. This latter unit was reinforced for the occasion by a platoon of the Weapons Company, Company C, 1st Engineer Battalion, and 52 men of the 1st Amphibian Tractor Battalion, and relieved 1/7 on the latter's newly acquired front. The 3d Battalion, 5th Marines, now available after completing the mop-up of Ngesebus three days before, was attached operationally to the 7th Marines and relieved 2/7, which moved to a bivouac area south of the Weapons Company sector for a night's rest preparatory to the coming attack.

The plan for 3 October called for a series of coordinated efforts aimed at one main objective: seizure of the remaining enemy-controlled stretch of the East Road and the ridges which dominated it, in order to gain staging points from which flank attacks could be launched against the strongholds which had checked the drives from the north and south.

Furthermore, the fighting on the northern perimeter had given the East Road a new importance. So long as the enemy controlled this lower stretch, all supplies brought into that sector and all wounded evacuated from it had to travel north on the West Road as far as Road Junction 15, then turn southward along the upper portion of the East Road where it angled across the peninsula: a trip of several miles to cover a distance which could be measured in hundreds of yards.

The main effort called for a converging attack to seize the remaining ridges which dominated the road. The 2d Battalion would attack from the south with Walt Ridge as its objective, the 3d Battalion from the north to capture Boyd Ridge.²⁴ Once the crest lines had been occupied and the units had made contact, they would face westward along the newly established front against the heart of the pocket. To avoid the danger of colliding head-on since the two forces would be moving straight toward each other, the attacks were to be successive rather than simultaneous: 3/7 would not jump off until 2/7 had achieved maximum penetration and could mark its flank with smoke.

In coordination with these movements, 3/5 was to attack the Five Sisters and extend eastward in order to strengthen the line on the south, and the Weapons Company, in conjunction with Army tanks, would support the attacks by moving both into the Horseshoe and up the East Road. To make this unit available for these missions, its static defensive line was taken over by the weary 1st Battalion.

The attack from the south jumped off at 0703 in the wake of an intensive barrage by 155's and the concentrated 81mm mortars of five battalions, the latter throwing

SEVENTH MARINES HALF-TRACKS advancing toward Horseshoe. Japanese in Hill 300 (right background) were not ousted until 2 November.

smoke during the closing stages in order to screen the advance. From his position so long held atop the flanking ridge, the commanding officer of 2/7 had been a distressed witness to the bloody repulses suffered by 2/1 and 1/7 before this same objective. Now that his own battalion's turn had come, he had made his preparations with great care, using both air and ground reconnaissance. The approach route selected followed the trail through the swamp²⁵ over which Captain Pope had withdrawn nearly two weeks before. Resistance was light, and so rapid was progress that the leading platoon of Company G, in assault, reached the top of Walt Ridge at 0730, the rearward elements following closely. Up to this point the battalion had suffered no casualties,²⁶ but all hands knew from bitter experience that this happy condition would not obtain for long. Sure enough, as the company carried out its mission by expanding its gains, the men came under a savage cross fire from the Five Brothers, facing them across the Horseshoe, and from the heights to the northward. And the regimental command

post began to receive increasingly urgent calls to send up stretchers.²⁷

Tanks and halftracks from the Weapons Company moved into the Horseshoe to cope with enemy installations which were harassing the sweep of Company G along the axis of the ridge's military crest. Also, mortar fire was called down on the fourth of the Five Brothers which was proving especially objectionable.

In the meanwhile, elements of Company G, on top of the ridge, were obliged to face in two directions: west, to cope with the flanking fire pouring in from that direction, while the drive endeavored to move northward along the crest. Company E was ordered up with the mission of passing through Company G's right and continuing the attack to the north. At 0900, however, progress was halted by "a slight saddle at this point which was covered by a murderous cross fire [from the Five Brothers]. Two out of every four men attempting to get across were hit."²⁸ It was necessary to wait until ropes and ladders could be brought up, and until the engineers could blast a covered approach along the east face of the ridge at the top of a 90-foot vertical cliff.

While Company E waited for ropes and the tanks for ammunition, Company F moved up the road to assault the ridge still farther along. The leading elements had progressed part way up the slope when they were ordered to halt, then to withdraw in order to be available for a subsequent mission. With no way of telling when the advance could be resumed, 2d Battalion elements along the ridge crest were directed to consolidate their positions and mark the northernmost flank with purple smoke as a signal for the 3d Battalion to begin its attack southward.

The 3d Battalion sighted Company G's smoke and jumped off at 1020. All went well until the advance elements reached the draw separating the unnamed ridge which they occupied from the objective ridge (Boyd). The foremost squad of Company K²⁹ got past this, perhaps because the enemy were not yet aware of attackers' intention. But the remainder of the platoon was promptly pinned down by fire funneled through the draw from the high ground farther inland, thereby holding up the rest of the company. Tanks which had been supporting the stalled attack of the 2d Battalion³⁰ moved forward to help beat down the volume of fire. In this they were not wholly successful, and a new means was sought for placing Company K on its objective.

Company K got the remainder of its personnel to the summit³¹ of Boyd Ridge by the expedient of detouring the draw through the swamp which lay just east of the road. This provided excellent concealment and some defilade, but the dense growth and wet underfooting made for slow progress. It was not until nearly 1530 that the mission was accomplished. However, so practical had this resort proved that steps were taken

at once to pass Company I over the same route, and around the next draw in an effort to tie in with the 2d Battalion to the south.

During this time 2/7 had continued its efforts to expand its hold on Walt Ridge. At 1550, Company E finally got over the covered trail blasted by the engineers. Now Company F, previously withdrawn, was committed to close the gap between the two battalions by making contact with Company I. Patrols of these two units, moving toward each other, met in the swamp facing the draw between the two ridges shortly after 1600.

This patrol contact, of course, was merely preliminary to closing a solid front. In order to achieve this final step, Company F pushed its left up the slope of Walt Ridge to tie in with the right of Company E, which move was reported accomplished at 1750.³² Simultaneously Company F extended its own right northward while tanks attempted to neutralize the draw with covering fire. Company I acted in a similar manner in the 3d Battalion zone, and was reported tied-in with F at 1730.³³ However, subsequently it proved necessary to break this physical contact in order to remain tied-in with Company K on Boyd Ridge, and Company I refused its left flank, maintaining only visual contact with F.³⁴

In the meanwhile, the weary men of Company G had been relieved on top of Walt Ridge by Company B and passed into battalion reserve. Thus, the regimental position as defined that evening shows Companies B, E, and F in line (left to right) along the crest of Walt Ridge, with the latter echeloned down the slope where it tied-in, to all practical purposes, with the 3d Battalion left. On its part, the 3d Battalion had Companies I and K in line (l to r), curving back from where the former, facing generally south, had visual contact with Company F, up Boyd Ridge where K faced west along the ridge's military crest. Company L, uncommitted that day, remained in 3/7 reserve.

Colonel Nakagawa evidently got out his rose-colored glasses to view this particular action. According to the version received at Koror: "At 0730, 110 tanks³⁵ and about two infantry battalions began to attack the central hills from north and south. Our garrison units repelled them and withdrew. . . . In this district about 100 enemy troops infiltrated our front line secretly but were exterminated during the evening."³⁶ An interesting sidelight on these unequivocal statements occurs in an entry dated eight days later: "Higashiyama [Walt Ridge] is still in enemy hands."

While these events were taking place on the eastern perimeter, the 3d Battalion, 5th Marines, attacked the Pocket strongly from the south. This thrust was mainly supporting in nature, designed to distract the enemy's attention from the operations of the 7th Marines, deter any attempts to reinforce the eastern perimeter, and if possible gain positions from which the main attack could be supported by direct fire. The battalion's objective was the Five Sisters, that gaunt palisade which had checked all northward advance since the 1st Marines had come up against it two weeks earlier.

The approach was exceedingly difficult, paralleling the ridge on the left in which

many Japanese were still holed up.³⁷ With Company L mopping-up behind it, Company K pushed forward and reached the base of the objective about noon. Here Company L moved up abreast, passing over the mopping-up job to Company I, and the two companies in assault ascended Sisters Number 1, 3, 4 and 5 (numbering from west to east). On the extreme left, one platoon of Company K, supported by a tank, moved into the ravine known as Death Valley, in an effort to get at Sister Number 2, which lay slightly to the north of the others.

A combination of difficult terrain and enemy resistance prevented the latter unit from making any important penetration, but the main assault carried to the jagged summits. No doubt, this success could be attributed in part to the steady attrition of the past two weeks which had knocked out supporting positions and installations in the Sisters themselves one by one. However, later in the afternoon the men of 3/5 began to come under an increasing volume of effective small arms fire from positions which they were unable to locate. Actually, this could have, and probably did, come from every direction save their immediate rear. The narrow crest was wholly devoid of cover save for occasional accidents of terrain, and the adamant coral, as everywhere inland of the beaches, defied efforts to dig in. Discreetly the battalion withdrew and set up a defense line for the night only a hundred yards or so in advance of the line from which it had jumped off that morning.³⁸

While many enemy installations had doubtless been liquidated, the Japanese in the area were still capable of reacting with such determination that the bodies of 21 would-be infiltrators were found, mostly within the forward lines, at dawn next day.

The Weapons Company, 7th Marines, having accomplished its support mission, withdrew from the Horseshoe according to plan. Thus, night of 3 October found the containing lines in just about the same position they had been that morning, save for the significant gains accomplished by the 7th Marines atop Boyd and Walt Ridges along the eastern perimeter.

On this same day (3 October) the division was saddened by the death of Colonel Joseph F. Hankins, one of its best liked officers and the highest ranking casualty sustained on Peleliu: an event especially tragic in that some surviving accounts made it appear unnecessary.

Hankins, a battalion commander in the 1st Marines at Cape Gloucester, had been appointed provost marshal and commander of the big, unwieldy Division Headquarters Battalion upon his promotion to full colonel. In the former capacity he was responsible, among many other things, for the security of the West Road. During the day or two preceding, this vital artery had been brought under particularly troublesome fire from a Japanese sniper, or snipers, posted in the high ground dominating a stretch of evil repute known as "Dead Man's Curve." The Military Police Company, in its attenuated form, was unable to cope with the situation; so, at about 1600 on the afternoon in question, Colonel Hankins, a member of several famous Marine Corps rifle teams, armed himself with an M-1 and a pair of binoculars and announced that he was going to do a little countersniping.³⁹

Whatever the colonel's plans on setting out, the situation discovered on the scene quickly changed them. The following account by an officer of the 1st Motor Transport Battalion represents the correlated versions of several eyewitnesses:

Map 16
Seizure of Eastern Ridges

Colonel Hankins appeared at the curve in the road where the Military Police were regulating the one-way traffic. An LVT had become immobilized across the road directly in the open and two or three trucks were jammed up in the near proximity of this LVT. The men, under the heavy fire of small arms from the nearby cliff had deserted their vehicles and taken refuge on the reverse slope of the road. Colonel Hankins proceeded to the middle of the road in order to restore traffic to normal condition and had actually gotten the crews back on the vehicles when he was struck by a sniper's bullet and killed instantly.⁴⁰

To avenge Colonel Hankins and eliminate the danger once and for all, the commanding general ordered up a company from 2/5, then in division reserve. After a study of the situation the battalion commander sent Company E into the general area of the high ground dominating Dead Man's Curve. Though it could not be known at the time, this grotesque jumble of broken coral formations constituted one of the approaches to Colonel Nakagawa's command post and in consequence was held in some strength. In the fighting that ensued, the company pushed about 75 yards inland to secure the military crest, and 150 to 200 yards northward to occupy all of the area from which it had been determined that the fire on the road had originated. Holding detachments from the 11th Marines followed the assault troops to organize the new positions, and Company E was relieved upon completion of its mission.⁴¹ From then onward the West Road was free of sniper fire during the remaining time the 1st Marine Division operated on Peleliu.⁴²

The day following Colonel Hankins' death (4 October), saw the last offensive efforts of the 7th Marines, as a regiment. For the 3d Battalion, it brought unmitigated tragedy.

The two tactically important ridges which bounded the Pocket on the east were now securely in the Marines' grip, and the day was spent in expanding and consolidating the positions there. However, the East Road remained a perilous passageway so long as the Japanese continued able to interdict it with fire funnelled through three major gaps: The Horseshoe, the draw separating Walt and Boyd Ridges, and the narrower draw between Boyd and the unnamed ridge to the north of it.⁴³ Clean-up of the two latter was the chore which fell to the 3d battalion, and a bloody business it turned out to be.

Company I undertook mop-up of the southern draw, that between Walt and Boyd Ridges, with collaboration of Company F, on the right of the 2d Battalion. Here tanks could be employed with some effect, but the job proved tedious, nerve-racking and costly. However, it was done to such effect that the two battalions were able to tie in physically that night, though some fire continued to be received from the high ground inland.⁴⁴ Two officers were killed in the process, and Company I emerged from the action with only one officer and 31 men fit for duty.⁴⁵

Although what befell Company L rates only as a small unit action, it will be related here in some detail as exemplifying a number of characteristic aspects of the fighting in the Umurbrogol.

As the accompanying sketch shows, the high ground from which the Japanese could fire most directly down the northern draw included three semi-isolated hills or knobs. These were steep and rugged, varying in height from 60 to 90 feet, and rose just

Map 17
ACTION OF COMPANY L, 7th MARINES, on 4 October 1944.

north and northeast of Ridge 120, a steepsided razorback running parallel to Boyd Ridge from which it was separated by a gorgelike valley. On the west of Ridge 120 and dominating it lay the height known as Baldy, which had frustrated attacks from the north for the past several days. So closely packed-in were all tactical features of the Peleliu landscape that the nearest knob was a scant hundred yards from the road, while the ridge was less than 150 yards in.

It should be borne in mind, too, that, while Company K was firmly ensconced on Boyd Ridge and had mopped-up its eastern slope in order to control the road, no effort had been made by infantry to enter the valley beyond, though tanks operating in the vicinity of the knobs had pounded Japanese installations in the ridge's western face which Marines on the summit were unable to reach.

Company L was ordered into the draw at 1430 with the mission of seizing the three knobs. This was accomplished by 1515 without loss and in the face of resistance so light as to astonish all hands. There, according to the original plan, the day's work would have ended. But Major Hunter Hurst had reason to feel that he had gained a great advantage. Just ahead lay Ridge 120, an ideal jumping off place for an attack against the flank and rear of that

troublesome Baldy.⁴⁶ Moving in Company C, 1st Engineer Battalion, to hold the captured knobs, he ordered the attack continued.

Again resistance was surprisingly light. By 1605 an entire platoon had mounted the northern nose of Ridge 120. Here several enemy positions were destroyed, and the men began a slow, systematic movement south along the ridge's axis, parallel to the front of Company K, facing them less than a hundred yards away across the narrow gorge.

At 1645⁴⁷ the platoon on the ridge began receiving fire from Baldy. There were several casualties, and the men moved quickly into defilade on the eastern military crest. This brought them into a clear line of fire of the enemy still holed up, with automatic weapons, in the lower slopes of Boyd Ridge, who proceeded to cut loose with everything they had.

Only then did the whole brutal truth become evident. Maintaining their excellent fire discipline, the Japanese had refrained from showing their hand until the maximum of U.S. troops had been irrevocably committed. Now, for the men caught in this savage cross fire, that coverless ridge crest became a death trap.

And there were other Japanese emplaced in the lower slopes of Ridge 120 and the captured knobs, covering the only possible line of retreat: down the cliffs and out through the draw.

What followed had the aspects of a protracted nightmare.⁴⁸ The enemy were now using everything that the constricted space permitted, and from every angle: small arms, automatic weapons, 20mm machine cannon, mortars. The senior non-commissioned officer on the scene, Gunnery Sergeant Ralph Phillips, was one of the first hit, shot dead by a machine gun burst. Two others were killed in a matter of minutes, and there were wounded all over the place. The three Navy corpsmen who had accompanied the platoon performed prodigies in caring for these, but two of them were killed before it was over.

One of the worst features was that the steepness of the slope made evacuation of helpless men nearly impossible. The platoon leader, Second Lieutenant James E. Dunn, was hit while lowering himself over the cliffside in an effort to lead his men to comparative safety and fell to his death on the jagged coral of the ravine floor.

The men on the ridge fought desperately. But so well concealed and so strong were the Japanese positions that the Marines were seldom able to see their enemies or to inflict serious damage when they did. On the floor of the draw Captain James V. Shanley, commander of Company L, viewed their predicament with increasing dismay. He ordered a tank up the narrow defile. The jumbled terrain checked it before it could reach a position where its guns would bear effectively, but at least it was far enough forward to provide cover of a sort for the command post and a rallying point behind which the wounded could be carried, once they could be got out. Then he called for smoke, and the battalion commander ordered Company K to blanket the area accordingly to screen the evacuation.

Frustrated heretofore in their efforts to furnish substantial aid to their comrades on the opposite ridge, the men of Company K responded energetically, hurling white phosphorus grenades down into the gulch. Fortunately, the wind was just right, and eddying smoke commenced to blot out the scene. Under its cover the few remaining able-bodied men began the backbreaking, heartbreaking job of getting the wounded down the precipitous slopes and to the cover of the tank. Although the Japanese were

now firing blindly most of the time, men continued to be hit.

Nearly all of those still alive had made it, when Captain Shanley, peering out through the thinning smoke, saw two wounded Marines staggering toward him, clinging to each other for mutual support while bullets kicked up spurts of powdered coral about them. Some 30 yards away their strength gave out, and both fell. Springing forward, crouched low, the captain ran to them. He carried the first man to safety in the lee of the tank, and returned for the other. A mortar shell burst immediately behind him, and he went down mortally wounded.⁴⁹ The acting company executive officer, Second Lieutenant⁵⁰ Harold J. Collis, charged into the smoke to help him, only to fall at his side, killed instantly by a Japanese antitank gun.

It was over by 1820. All of the wounded who had survived to get as far as the tank had been safely evacuated. But of the 48 Marines who attempted to seize Ridge 120, only five emerged from the draw unscathed. With Company L reduced to the strength of a single platoon and Company I down to 31 men and one officer, the 3d Battalion was now about on a par with the other two as regards combat efficiency.

In the other sectors 4 October was uneventful by comparison. The 3d Battalion, 5th Marines, repeated its moves of the previous day against the Five Sisters and Death Valley, with almost identical results: some notable advances made which it proved impracticable to hold. The 2d Battalion, 7th Marines, attempted to move men and tanks up the East Road across the mouth of the Horseshoe, only to have two tanks knocked out and the men pinned down by a rejuvenated volume of enemy fire pouring from that dangerous defile; so devoted most of the day to consolidating positions along the crest of Walt's Ridge. Elements of the 3d Armored Amphibian Tractor Battalion reported as reinforcements to 1/7, and some elements of that unit retired to a bivouac area behind the lines.

The 7th Marines was now through as an assault unit on the regimental level, and its relief by the 5th Marines began the following day (5 October). The 1st Battalion and Regimental Weapons Company went out first. The battered 3d Battalion continued its efforts to mop-up the draws to the north and south of the Boyd Ridge position. During the afternoon it was thought advisable to withdraw depleted Company I, so Company F took over this sector, extending the line of 2/7 from Walt up onto Boyd Ridge, which was held through the night.

The final relief was completed on 6 October, with both of these battalions departing to join the other elements of the 7th Marines in bivouac in the Ngardololok area. Aside from a few limited combat missions (to be discussed hereinafter), they were destined to remain in general reserve, saddled with no more onerous duties than static beach defense.

The regiment had richly earned this rest.

Final Drive of the 5th Marines

Bearing in mind that the 5th Marines had been fighting almost continuously for nearly a month under some of the worst combat conditions encountered anywhere in the Pacific, an apt foreword to any account of the last major effort of the 1st Marine Division on Peleliu is contained in a letter from an officer who was present:

I don't think that there can be any true picture of that final drive without some description of the great weariness of the Marines who participated in it. The Division had optimistically said that the 5th would be one of the first outfits to leave Peleliu, yet after securing the northern end of the island everyone knew that we would be committed again. Now the 1st and 7th Regiments were for the most part gone, but the 5th was back at it again. The men and officers were superb during this last phase

"MEN AND OFFICERS were superb during this last phase but very, very tired."

but very, very tired. . . . I have never forgotten Major Gayle's answer, when questioned as to how the fighting men were doing, as it typified the character and ability of all those who remained of the 5th: "Every Marine fighting in those hills is an expert. If he wasn't, he wouldn't be alive."⁵¹

As related in the previous section of this chapter, the 3d Battalion, 5th Marines, operationally attached to the 7th Marines, had been engaged in the Umurbrogol since 2 October and had already made two assaults on the Five Sisters. This unit reverted to parent control on the morning of 6 October when the 5th Marines, as a regiment, completed relief of the 7th in the drive to reduce the Pocket.

The tactical concept with which this regiment approached its new assignment is best explained in the words of its commander, Colonel Harris: "The 5th Marines, after careful air and ground reconnaissance, reversed the direction of all prior attacks and made the main drive down from the north.⁵² The methodical reduction of enemy positions was possible in driving southward due to the compartmentation of the terrain. It was this slow but steady eating away of the Jap defenses that gave the real payoff. Too

much credit cannot be given for the splendid support furnished by our artillery, tanks, engineer company, and LVT flame-throwers."⁵³

According to orders, the 1st Battalion relieved 2/7 and the 2d Battalion took over the positions held by 3/7. The commanding officers of these two units had reconnoitered the ground the previous afternoon. However, the regimental overlay for the first day shows a situation somewhat different from that existing on disastrous 4 October. Following the debacle of Company L, Major Hurst had found it expedient to evacuate the not readily tenable summits of the three captured knobs and further to shorten his line by withdrawal of badly battered Company I. Thus, the front taken over by 1/5 included both Walt and Boyd Ridges, and the already depleted battalion was spread thinly over some 1200 yards.

On the north, 2/5 faced Baldy Hill and its supporting network of ridges and knolls from a respectful distance. On the south, 3/5 had been withdrawn to a bivouac area following its second attack on the Five Sisters, preparatory to new offensive efforts from this direction.

The containing line on the west remained essentially unchanged, manned by supporting troops. In its northern sector Major

POCKET viewed from northeast. Boyd Ridge at jog in East Road (foreground). Paralleling it (right to left) are Ridge 120, Ridge Three and Hill 140. Immediately to right of 120 lies Baldy, with Wattie Ridge beyond it.

Harold T. A. Richmond, executive officer of 1st Battalion, 5th Marines, commanded a group made up mostly of detachments from the 1st Pioneer Battalion. On his right lay Lieutenant Colonel Richard A. Evans' group of "Infantillery": detachments from the corps artillery and 11th Marines, stiffened by volunteers from miscellaneous units. Closing the cordon sanitaire at the south were men of the amphibian tractor battalions. As these sectors were destined to continue static for several days longer, immediate interest focuses elsewhere.

The 2d Battalion immediately encountered the same contact problem that had plagued the 7th Marines when operating in the same area the previous week. On the left, Company E, occupying the ridge paralleling the East Road north of Boyd Ridge (now actually occupied by Lieutenant Colonel Boyd's command), found itself separated by a long, steep canyon from Company G on the high ground to the westward, facing Baldy. For the first few days Company F in battalion reserve, was used to patrol this gap, setting up a defensive line at night to prevent infiltration. Otherwise the action in this zone was wholly aggressive.

Company E attacked within half an hour of arriving in position: at 0900 on 6 October. The direction of the push was westward, into the badlands where Captain Shanley's company had met disaster two days earlier. Two of the previously abandoned knobs were recaptured, but again the troops came under heavy and accurate fire from many of the same positions which had frustrated their predecessors. Not only was further advance impossible, but the men risked having their heads blown off if they so much as raised them above the knobs' crests. But their rear was secure for the moment, so there they stayed. Company E was destined to spend four days making this area sanitary, while bulldozers cut a track for tanks and flame-throwers toward hitherto inaccessible points still farther in, in preparation for the crucial drive southward.

Meanwhile Company G staged a frontal assault on Baldy. The line of advance lay through the same terrain where 3/7 had been turned back on 30 September and 1 October, and the attack was ordered only under pressure and with considerable reluctance.⁵⁴ However, even though lacking strong fire support, a unit of approximately platoon strength reached the summit. This could not be reinforced for the simple reason that the occupied position would accommodate no more; nor was it possible to bring up there any weapon larger than a BAR. Baldy was considered untenable, and with the approach of night the assault platoon withdrew to the original jumping-off place.⁵⁵

The offensive effort of 7 October was delivered from the southeast by the 3d Battaloin, supported by tanks--or vice versa. Following two and a half hours of intensive artillery preparation, six army tanks advanced to the mouth of the Horseshoe at 0900 and commenced firing into all enemy positions which could be located on either flank: the lower slopes of Walt Ridge on the right, and the Five Brothers on the left. Those elements of the 1st Battalion in position atop the former moved over the crest in order to bring their firepower to bear in support. Several of the tanks were hit by Japanese heavy weapons, but none was seriously damaged. Much encouraged, they withdrew at 1045 to replenish their fuel and ammunition, determined to return and give the Horseshoe the full treatment.

The second attack jumped off at 1215 with a fire team of Marines protecting each tank, supported by two LVT flame-throwers and a platoon of the 1st Engineer Battalion for demolitions work and mine removal. Again heavy fire was received from the dominating terrain. Company L, on the left, attempted to drive into the declivity, later known as Wildcat Bowl,⁵⁶ which angles westward a short distance within the mouth of the Horseshoe through a narrow pass between the Brothers and the Sisters, bounded on the west by the as yet untested buttressed ridge later christened the China Wall. This combination was simply too much, and the attackers were stopped before being able to make any significant penetration.

Company I, attacking up the main valley, did somewhat better, advancing as far as the terrain permitted tanks to travel at this time: about 200 yards. Men and machines attacked numerous caves with every means available and killed many Japanese. Altogether, this was the largest single dose of destruction the Horseshoe defenders had received to date. But, again, once the tanks had run out of ammunition, the position became untenable for the infantry.⁵⁷ And so naturally powerful were the enemy positions that it was another ten days before U.S. troops and tanks attempted again to operate in those confines; nearly the end of October before the Horseshoe could be called anything approximating secured.

The 2d Battalion spent this day (7 October) paving the way for events to come: locating the enemy's strongest positions and taking the preliminary steps toward destroying them. The success of Company G's patrol the previous day indicated that Baldy Hill served the Japanese mainly as as observation post. The same did not apply, however, to the ridge extending southward from it. The crest and flanks of this were still wooded, and Marines attempting to advance anywhere within range invariably drew heavy fire from hidden positions somewhere along its axis.

Much the same could be said of other spurs of the ridge system. It was determined, therefore, to destroy this cover systematically in order to reveal the enemy positions before launching an infantry attack. Once LVT flame-throwers could be bulldozed within range, the vegetation could be burned off with comparative ease. While engineers labored toward this end, mortars were set up wherever they could be brought to bear. A single 60mm mortar on the ridge north of Boyd, for instance, fired 3,000 rounds during this phase.

Patrols from Company E came down from the knobs and operated with bazookas against a number of caves which had been located in that area: in the lower slopes of the ridges and of the knobs themselves. In the Company G sector, small patrols moved to the front, probing for a soft spot to the west which would provide access to the high ground and make possible an attack on Baldy Ridge from the rear. This was the same basic objective, it will be recalled, that had motivated Captain Shanley's ill-fated effort from the east three days earlier.

These activities continued for the next two days, as a trail was worked farther and farther up the gorge in order to bring tank guns to bear on both Ridge 120 and the western face of Boyd Ridge. There was no hurry now. The tactical situation⁵⁸ called

for thoroughness rather than haste, and Colonel Harris resisted all pressure⁵⁹ to speed things up at the cost of unnecessary losses.

There was no activity in the other sectors so far as the infantry was concerned. The 3d Battalion remained in reserve in its bivouac area to the south, and along the eastern ridges the 1st Battalion contented itself with countersniper fire. Pressure against the Pocket was maintained by artillery fire from both north and south, and on 8 October Marine Corsairs delivered two air strikes against the Five Sisters, Five Brothers, and the ridges beyond, using 1,000-pound bombs and napalm.

Near the 2d Battalion command post, then located hard by the West Road about on a line with the Japanese northern perimeter, heavy weapons were being positioned to support this unit's impending attack. Fired point-blank into the sheer cliff that barred approach from this angle, the powerful shells pulverized the coral until the face of what had been a precipice began to assume the aspect of a steeply inclined ramp, greatly simplifying the problem of gaining access to the high ground.⁶⁰

Attacking on the morning of 9 October, a platoon of Company G, commanded by Second Lieutenant Robert T. Wattie, scaled

TANK-INFANTRY SORTIE attacks Japanese positions in the Five Brothers.

this new approach and succeeded in reaching the top of the narrow ridge which forms the western spur of Baldy. This was a position of tactical importance for two reasons: (1) It commanded a clear field of fire to the West Road and constituted a potential danger so long as it remained in enemy hands; (2) About midway of its length it connected with Baldy Ridge, thereby providing direct access to that primary objective. It was promptly named in honor of the platoon leader who had captured it and appears as "Wattie Ridge" on subsequent maps and reports.

Lieutenant Wattie led his men southward along the crest for about 100 yards, knocking out several Japanese positions. But at this point the platoon came under heavy fire from the ridge to the east and from a large cave situated at the head of the dead-end gulch that separated the two, and he withdrew his men to more tenable ground to the northward to give supporting weapons a chance to work over the enemy positions.

By this time the LVT flame-throwers had been bulldozed within range and had ample time to burn off the concealing vegetation from the crest of Baldy Ridge and the jumble of razorbacks to the east of it. An attempt was made to subdue that troublesome cave in the gulch by air attack, but it proved so located as to make a bombing strike dangerous to friendly troops nearby. The direct fires of the supporting artillery were then concentrated on the area. A lucky shot, or shots, precipitated a most satisfactory landslide, with such good results that wary infantrymen approaching to finish the job next morning found the cave entrance sealed as effectively as they could have done it with demolition charges applied by hand.⁶¹

A battery of 105mm howitzers, an Army M-10 mounting a 76mm, and a 37mm were now available to support the 2d battalion's attack. On the morning of 10 October (D-plus 25) the southward push jumped off in earnest, with Companies G and E both in assault.

In the former unit's zone Lieutenant Wattie led his men back over the ground they had relinquished the previous day, proceeded beyond, and commenced to move over to Baldy Ridge. Here they ran head-on into some Japanese who had different ideas. A sharp skirmish ensued, with both sides using small arms and hand grenades in considerable quantities. Then the Marines gathered themselves together and carried the enemy position with a rush. From this point they swept northward along the crest against only scattered opposition, burning out or blasting every installation encountered. By noon they had secured the entire ridge and the formidable heights of Baldy itself from the rear, in accordance with the original plan.

Seizure of this dominating terrain secured Company E's right flank against the deadly cross fire that had previously trapped Captain Shanley's company. This unit's own operations of the past three days, plus the advent of tanks and LVT flame-throwers in the draw, had already secured its left. Thus, Company E, launching its assault at 1215, was able to secure the full length of Ridge 120 with comparative ease.

In conjunction, other elements of Company G secured Ridge 3, a semi-detached razorback south and slightly to the east of Baldy Ridge. When the 2d Battalion halted to consolidate the new positions, the situation overlay showed the line pushed well forward on the left over some of the most difficult and stubbornly contested terrain in the entire Umurbrogol Pocket. The two companies were tied-in only by fire, however, owing to a steep gulch between them, and one platoon of Company F was moved forward and attached to Company G as a special reserve to cope with this situation.

Next morning (11 October) the 2d Battalion resumed its drive on what had been selected as its final objective: Hill 140, a position of the utmost tactical importance, situated just north of the Five Brothers, from which it was separated by a sheer, narrow declivity. A heavy weapon emplaced here would be capable of placing direct fire not only into the nearest of the Brothers and part of the Horseshoe, but down the draw between Walt and Boyd Ridges, dominating this and quite possibly converting what had been a potent peril to the East Road into a practical route of entry into the very heart of the Pocket.

The battalion commander was convinced that Hill 140 would provide a fine position for at least one such weapon for the support of future operations. Major George E. Bowdoin, executive officer of the 4th Battalion, 11th Marines, had been considering the possibility of getting artillery pieces into the hills to provide direct fire. He believed that the 75mm pack howitzer would prove wholly practicable for this purpose and

Map 18
Drive From the North

already had taken steps to see that several such were made available.⁶²

Company G continued southward along Baldy Ridge, and progressed satisfactorily until reaching the ravine which separated the southern end of that elevation from Hill 140. Company E attacked on a parallel line on the left, sweeping along the eastern slope of Ridge 3 until stopped short of the objective by a heavy volume of small arms fire from the front and right; i.e., the northern face of Hill 140 and the slope of Baldy Ridge. Elements of Company G moved down from their vantage point on the high ground to help eliminate fire from the latter position, while forward elements of both companies endeavored to neutralize the enemy to the front with their own fire. At this point Company F, the remainder of which had been brought forward in close reserve at the beginning of the attack, passed through Company E, moved up the ravine, and carried the objective positions by assault from farther to the west.⁶³

Hill 140 was announced secure at 1500. The remainder of the afternoon was spent in consolidating the new position and mopping-up enemy caves which had been overrun in the advance or which lay within reach in other directions. In the course of this work an LVT flame-thrower which had managed to work its way up the draw within range detonated a 12-inch naval shell in one of the occupied caves, with a resultant blast that blew away a large bit of the hillside. Fortunately, this unlooked-for event caused only one Marine casualty. Altogether, so thorough had been the preparation and so well executed the attack, capture of one of the most important single positions in the whole Umurbrogol had cost the 2d Battalion only two killed and ten wounded.

Apparently the significance of this fighting was not immediately apparent to the Japanese. On 10 October Nakagawa reported: "There were no large battles during the day and no changes in the front lines." And again on the 11th: "All through the day there were no heavy engagements," adding, "Areas which are secured by our garrison are Oyama, Suifuzan, Kansokuyama, the main districts in the central hills," though conceding at last the loss Higashiyama;⁶⁴ a definition of position which squared quite accurately with that shown on Marine situation overlays of the period.

On the 12th, however, a note of concern crept in. Evidently dealing with events of the previous day and night, the colonel informed headquarters: "The enemy seems to have made Suifuzan their main objective. They attacked Suifuzan and the west hill of Suifuzan [translation not entirely clear] with flame-throwers. The enemy penetrated our front lines but were repelled by night attack."⁶⁵

The Japanese did indeed stage a counterattack on Hill 140 that night. But the position was admirably suited to defense, and 2/5, unaware that it was in the process of being "repulsed," beat off the enemy without undue difficulty. None penetrated the line, and their final dispersal marked the

TACTICAL CONFERENCE. Strain of the protracted campaign stamped on their faces, three Marine commanders study situation map. Left to right: General Geiger, Colonel Harris, General Rupertus.

end of the 2d Battalion's active participation in the Peleliu campaign. At 0930 on 12 October, relief of the weary men was begun by the only slightly less weary 3d Battalion, and they came down out of the hills for a well-earned rest.

There was little activity in the other sectors during the 2d Battalion's drive. The 1st Battalion's mission had been simply to hold its line along the eastern ridge, countering random enemy fire and sealing up such undestroyed caves as could be found in its area. Colonel Boyd improved this time by tightening the containing cordon on the south: moving the line facing the Horseshoe's mouth from the far side of the swamp to the lee of the causeway, tying in on the right at the base of Walt Ridge. On the morning of 10 October 1/5 had been relieved by the 2d Battalion, 7th Marines, and moved into a bivouac area to the south. As the latter unit had previously occupied the same line, officers and men were entirely familiar with the positions, and the relief was completed without casualties or notable incidents. Thus, 2/7 had been on the left of 2/5 throughout the two final days of the latter's attack, but had been no more active there than had its predecessors.⁶⁶

Following its sortie into the Horseshoe on 7 October, the 3d Battalion had been in regimental reserve near the command post. On 10 October Company K had been sent up into the hills to eliminate some enemy snipers who had become unexpectedly active in caves located in an area far behind the present front lines which the 1st Marines had overrun and believed secured as early as D-plus 3.

Whether these Japanese had lurked undiscovered in these supposedly cleaned out caves all this time, or had infiltrated from the Pocket to reoccupy them, was anybody's guess. But they were there, firing down a draw at troops passing along the main road and generally making a nuisance of themselves. And so craftily did they play their little game that the men of Company K never did find them, though they received some rifle fire and five mortar shells in the process of trying. They blasted shut every cave opening they could find and returned the next day to finish the job. They were successful in that somewhere along their methodical course they put an end to the sniping, but in all of their two days work there they glimpsed not a single live Japanese.

Any resemblance between the uneventful relief of the 1st Battalion and that of the 2d, two days later, was purely coincidental. Instead of moving into a comparatively quiet sector already familiar to the relieving unit, the 3d Battalion was taking over a front which had been seized only the previous afternoon and with which even the troops being relieved had had small chance to become acquainted. And it was anything but "quiet", comparatively or otherwise. Before the day's doings were over, 22 Marines

of the two battalions had become casualties.

The line⁶⁷ held by the 2d Battalion on the morning of 12 October was in the form of a deep salient, nowhere achieving any great width. At the apex Company F, facing southward, had no friendly troops on its immediate left, where the side of Hill 140 fell away in a sheer cliff into the northern end of the Horseshoe. Company E was echeloned rearward along the eastern slopes of Ridge 3 and Baldy Ridge, in visual contact with elements of 2/7 on Boyd Ridge across the narrow canyon. On the right, Company G was bent back almost as sharply along the opposite slope of Baldy and over onto Wattie Ridge.

According to the 3d Battalion plan, Company L relieved Company F on Hill 140. Since the narrow canyon in Company E's sector could be controlled by patrols operating from Boyd Ridge, there was no need to continue this front in any force, so the remaining two companies of 3/5 were placed in position on the right preparatory to straightening out the front by a push southward pivoting on Hill 140: Company K tying in with Company L, and Company I extending the flank westward across Wattie Ridge.

There was trouble in all sectors. The commanding officer of Company K found the troops he was to relieve so thoroughly pinned down by enemy fire that the machine gunners were sighting along the under side of their weapon's barrel. A stranger in some of the strangest territory in the world, he raised his head in an attempt to orient himself--and was killed instantly by a Japanese bullet.

Over to the westward things were just as bad. Late the previous day, in an effort to rectify the line, a reserve platoon of Company F had pushed forward into the high ground above the West Road. Heretofore this region had been kept under control, more or less, by concentrations of interdicting artillery and mortar fire. With friendly troops in actual occupancy, this fire had to be called off. The Japanese, reacting promptly, reoccupied the area in some force. Exercising their excellent fire discipline, they kept their presence hidden until a platoon of Company I moved in; then, taking advantage of the inevitable confusion attendant on performing front line relief, they cut loose with heavy rifle and machine gun fire from positions enfilading both platoons. This was a situation calling for discretion, and the troops were drawn back under smoke grenade concealment to allow new concentrations of mortar fire to work over the area again.

This day, 12 October (D-plus 27), saw two other significant developments.

First, was issuance of a new map which at last undertook to show the terrain of the Pocket as it actually was rather than as it had appeared in preliminary aerial surveys when overgrown with vegetation. This was a hurried sketch job prepared by the Intelligence Section of the 5th Marines: inaccurate as regards relative elevations and with contour lines deceptive. However, it was reasonably accurate on the horizontal plane, and certainly a vast improvement over anything existing heretofore. With its advent, some uniformity in nomenclature began to appear in unit reports.⁶⁸

The second event of importance was the emplacing of artillery in the high ground.

PACK-HOWITZERS were hauled to high ground to fire point-blank across canyons.

It has been noted that the 2d Battalion especially wanted Hill 140 for this purpose, and even though this unit was in the process of being relieved, no time was lost in carrying through the plan. Tackle was rigged to swing a disassembled 75mm pack howitzer to the summit of Wattie Ridge, from which point it was manhandled to the forward position, reassembled, and set up behind sandbags to fire into the Horseshoe and the western base of Walt Ridge. A second one was similarly installed in the sector of the containing line held by the artillery-infantrymen, now commanded by Lieutenant Colonel Edson A. Lyman, 11th Marines, who had relieved Lieutenant Colonel Richard A. Evans, IIIAC Artillery, on 9 October.

This latter gun was destined to have a brief but colorful career. Colonel Lyman emplaced it on the western rim of Death Valley to bear on the Five Sisters and China Wall. The point directly across from it was believed to be Colonel Nakagawa's observation post from the fact that what were apparently high ranking Japanese officers, wearing white gloves, had been seen there on occasion studying the terrain through binoculars; a supposition which the violence of enemy reaction seemed to bear out. The first rounds "routed out a covey of Nips," and shortly thereafter the crew was brought under heavy small arms fire at the deadly range of about 75 yards. By the time 40 rounds had been fired, one man was hit and "it was deemed expedient to secure." Two more were killed at daybreak by a hornet's nest of hidden enemy, and the position was considered too perilous for further artillery operations.⁶⁹

Exactly who should be credited with carrying the first sandbag⁷⁰ into the Peleliu

ridges has never been clearly established, but he started something which was to become increasingly important as the grim campaign of attrition rolled on its inexorable course. The lack of cover and impossibility of digging-in had repeatedly obliged attacking troops to relinquish hard-won gains as untenable. Obviously, the sandbag provided an answer of a sort, and many were in use at this time in more or less permanent positions, such as the gun emplacement on Hill 140 was intended to be.

But so long as the operation remained essentially one of movement, the problem of sandbagging successive positions on any large scale presented formidable difficulties. Nowhere was there any sand inland of the beaches, which meant that the heavy bags had to be carried into position already filled, no small undertaking in that crazily upended country. It remained for the 81st Division, following relief of the Marines, to develop a technique with such ingenious refinements as to make the sandbag into something closely resembling an offensive weapon, in which capacity it played a crucial part in the final reduction of the Pocket.

The 1st Battalion, 5th Marines, remained inactive in its bivouac area throughout the day of 12 October, but at 1900 one element moved out in a novel attempt at counter-infiltration by placing Marines in concealed positions on the Five Sisters. The force was a small combat patrol from Company C, led by First Lieutenant Roy O. Larsen, Assistant R-2, and its approach was covered by a diversionary shelling within the Horseshoe. As the Regimental War Diary records: "The plan was to get on top of these hills and dig in during the remainder of the night." The men managed to climb to the saddle between Sisters number 1 and 2, where they encountered a group of about 35 Japanese who dropped a grenade in their midst, scattering them so that "full control was never regained." The War Diary account adds, somewhat picturesquely "There were a great many more Japs in these hills as they could be heard talking and sliding around in the coral." Larsen discreetly withdrew.

Morning of 13 October found the 3d Battalion the only unit of the 5th Marines still on the lines, and the only unit of any regiment with immediate orders for resuming the offensive. Hereafter its main efforts would be concentrated to the westward, with the dual purpose of straightening out the salient formed by Hill 140, and providing greater security for the West Road by constricting the Pocket further from this direction.

It will be recalled that the terrain here was considered impractical for attack by either party, and by a sort of tacit mutual consent had been a quiet sector ever since the Pocket had been isolated. The containing lines against sporadic Japanese infiltration, fully manned only at night, followed the most defensible contours, in some places being withdrawn altogether from the high ground during the day to the low ridge just seaward of the West Road. Now 3/5 began probing southward into the flank of the ground in front of the containing lines.

Under cover of artillery and mortar barrages, a patrol from Company K pushed forward 75 yards into the jumbled terrain without encountering any resistance, destroying some Japanese rifles and ammunition discovered during the process. On its right, a similar patrol from Company I penetrated 150 yards, also without resistance, and preparations were made for a serious advance the following day.

At 0600 a napalm strike by aircraft against the region into which it was proposed to attack ushered in the morning of 14 October. Two hours later Company I jumped off in the wake of a heavy mortar preparation. This time resistance was encountered, described in the Regimental War Diary as "severe enemy sniper fire." The methodical advance continued nevertheless, and when the men halted at 1630 to prepare

positions for the night, they had made a gain of approximately 250 yards.⁷¹

Simultaneous with this advance by Company I, what was left of the 1st Battalion, 7th Marines, which had been brought up from a rest area for the purpose, attacked northward from the southern perimeter in a converging movement, gaining 125 yards before being halted by enemy fire. A further advance of 150 yards during the next two days completed the job.⁷² Since both of these operations had been moving across the front of the old containing lines, these had only to displace forward in order to consolidate the territory thus gained. When this had been done, the new lines were as much as 200 yards deeper into the high ground and something like 400 yards shorter from north to south than they had been nine days earlier.⁷³ The West Road was not subjected to sniper fire again during the time the Marines remained in occupancy.

By now, even Japanese optimism was wearing a bit thin. Describing an attack on "the western districts of Suifuzan yesterday," Nakagawa for once refrained from making claims of victory: "A powerful unit of our garrison force in a daring night attack attacked the enemy in this area. At present a fierce combat and severe artillery fire is taking place, with results unknown." And on the next day: "A unit of the enemy penetrated into Suifuzan and western part of Higashiyama.⁷⁴ Our units holding the southern part of Suifuzan together with another strong raiding unit are repulsing the enemy. . . . The enemy intercepted our movements in the western part of Higashiyama with a mortar attack which lasted all through the night. They also attacked with flame-throwers attached to tanks."⁷⁵ And with that last somewhat puzzling statement ended Japanese cognizance of Marine operations on Peleliu.

The other two companies of 3/5 confined their 14 October activities to improving their positions, sealing caves, patrolling. Company L brought up additional sandbags to Hill 140 and laid concertina wire around the position beyond grenade range which was instrumental in thwarting a Japanese infiltration attempt later that very night. And that was the last gasp of the Peleliu campaign so far as these units were concerned. By 1100 the following morning (15 October), their relief by elements of the 321st Infantry had been completed and they were on their way to a rest in the northern defense zone. The positions which they turned over, and from which the Army regiment commenced its attritional operations against the Pocket, are indicated (approximately) on Map 19.

As early as 27 September, the U.S. flag had been raised at the 1st Marine Division command post to symbolize that Peleliu was secured--as indeed it was to all practical purposes, though it could not be so

Map 19
Final Marine Action

termed officially. On 12 October,⁷⁶ as 3/5 was beginning its final push, the "Assault Phase" of the operation was declared at an end.

This perhaps unfortunate phrase was used at the time and occurs in the official reports of IIIAC (p. 11), 1st Marines Division (II, p. 18), and 81st Infantry Division (II, p. 47). It gave rise to some pungent comment among the Marines still very much in assault in the Umurbrogol ridges, and the 81st Division Wildcats, for whom six weeks of hard fighting remained.

Actually the term had technical significance in that it signalized a transfer of command functions from the assault forces to the Central Pacific administrative echelons Forward Area (Vice Admiral J. H. Hoover), Western Carolines Sub Area (Rear Admiral J. W. Reeves, Jr.), and Island Command (Brigadier General H. D. Campbell).⁷⁷ Admiral Fort in his report used the term "Attack and Occupation Phase"⁷⁸ and explains further:

I was directed by Admiral Halsey to turn over command of the Palaus to Vice Admiral Hoover, who was in command of the Forward Area, when the Attack and Occupation Phase terminated. This was when the islands had been occupied and when Base Development had been initiated and could proceed without enemy interference, and when the Assault Troops could be relieved by the Garrison Forces. This phase was completed on 13 October 1944, and I transferred command to CTF 57, Vice Admiral Hoover, at 1200 on 14 October, 1944.⁷⁹

So at long last a turning point had arrived, and it was possible to initiate steps for evacuating the rest of the 1st Marine Division to the doubtful charms of its "rehabilitation" base. But even Pavuvu would look good now.

Relief and Departure of 1st Marine Division

The operation which the commanding general had predicted would be over within four days had now lasted a month--and was still going on. It has been seen how the 1st Marines had been evacuated after that regiment's over-all strength had been cut by 58 per cent in a week of what was possibly the most savage fighting the campaign produced. Succeeding weeks had whittled away the 7th and 5th Marines in turn until their casualties began to approximate that ruinous figure. Furthermore, these two regiments had had a far more protracted dose of war in Peleliu's debilitating heat, soggy rains and backbreaking terrain, and their assault potentials had been as seriously impaired. The troops stood in great need of rest and rehabilitation, and now, at last, their turn was coming.

The permanent relief of the 1st Marine Division by the 81st Infantry Division began on the morning of 15 October when the 2d Battalion, 321st Infantry, took over the area held by 3/5 across the northern end of the Pocket: the Hill 140 salient and the line extending westward from it. The following day the infantry regiment's 1st Battalion relieved 1/5⁸⁰ atop Walt and Boyd Ridges and facing the mouth of the Horseshoe. At about the same time elements of the 323d Infantry, newly arrived from the successful seizure of Ulithi, completed taking over the southern and western containing lines, and at 1245 on 16 October command of operations in the Umurbrogol passed officially to the commanding officer, 321st Infantry.⁸¹

BATTLE WEARY MARINES could be relieved at last.

This completed relief of the 5th Marines, the 2d Battalion having moved out four days earlier. But the regiment was destined to spend two more weeks at Peleliu in the capacity of general reserve, occupying positions designed to defend against possible Japanese counterlanding: the 2d Battalion on Ngesebus, Kongauru, and Garekayo;⁸² the 1st Battalion along the northern extremities of the long peninsula; and the 3d Battalion deployed along the East Road, facing seaward. This defensive conception was a bit euphemistic, as everyone was pretty well convinced that the danger of counterlandings was long since past;⁸³ so, although they went through a few perfunctory military motions, the troops managed to rest.

The 7th Marines remained committed longer but got away from the island more quickly. Movement Order No. 4-44, dated 16 October, directed embarkation of the regiment for return to base in the Russells as soon as this could be managed, and elements commenced moving to PURPLE Beach to commence loading operations.

At the time that command passed to the 321st Infantry, 1/7 was still engaged in its northward drive to make possible advancement of the containing lines from the west, and its immediate relief was not thought

expedient. Upon completion of its mission later that afternoon, however, the unit was relieved in position by Company B, 323d Infantry, and returned to its bivouac area where it reverted to parent control at 0800 on 17 October. Shortly thereafter the battalion moved in turn to PURPLE Beach and was embarked in the transport Sea Sturgeon.

The 3d Battalion had been ordered into division reserve on 14 October and remained in that capacity for five days without actually being employed in the lines. Company K was attached to the 1st Battalion on the 15th and supported that unit's push, but reverted to parent control before sundown. Then, on the 17th and 18th, the battalion attained the dubious distinction of being the last 1st Division unit to see action on Peleliu.

This was directed against a group of Japanese infiltrators who had reoccupied caves in the area of Company E, 1st Medical Battalion, a short distance south of the Pocket, where they indulged in some lively sniping. The action began at 1840 on 17 October. Company I moved up and engaged the enemy in a short, sharp fire fight but was unable to reduce the positions before dark. The following morning Company L relieved Company I and resumed the attack, only to discover the Japanese present in considerably greater strength than previously estimated. A tank came forward in support but had the misfortune to strike a large land mine and was very thoroughly blown up. The company commander⁸⁴ had been directing the tank's fire and was killed, together with three crew members, two others being wounded. Resistance still existed in the area at nightfall, but 81st Division elements relieved Company L the next morning, and the weary battalion reverted to regimental control.

PRIVATE FIRST CLASS RICHARD E. KRAUS earned the Medal of Honor when he intentionally smothered an exploding grenade with his own body thus saving the lives of three men.

Company K had been sent to the beach the previous day to lend a hand at loading ship, and on the morning of the 20th the rest of the battalion arrived at the same destination, where by 1730 all hands had joined the 1st Battalion aboard Sea Sturgeon. Loading was completed the following day, and the ship shoved off on 22 October, arriving at Pavuvu on the 30th.

The 2d Battalion, which had been engaged in some patrolling in the northeastern islands, did not commence loading until 26 October. The men thereupon ran into unexpected complications of a somewhat bizarre nature, which the battalion commander describes as follows:

The Sloterdyke was a Dutch merchantman (under a Dutch captain with a polyglot crew) leased by the Army with an Army TQM and a Navy armed guard aboard and was carrying a skeleton Marine

infantry battalion and a Marine artillery battalion (4/11). By dint of the Marines manning the winches and booms, we were able to load and depart on the 30th.⁸⁵

The good ship Sloterdyke (or Sloterdykk, or Sloterdyk--sources vary as to the spelling) with her colorful complement arrived "home" at Pavuvu on 7 November in the convoy bearing the 5th Marines and its reinforcing elements.

This combat team had been organized as a task force under Brigadier General O. P. Smith, assistant division commander, attached operationally to the 81st Division. The force was not employed in the active fighting, however, the several battalion units remaining in the defensive positions previously described, except for a few minor changes, until 26 October. On that date 81st Division Field Order No. 23 directed their relief by Army elements, which was accomplished by 1200. On the 27th trucks became available and began shuttling the troops from northern Peleliu to PURPLE Beach.

Lack of suitable shipping delayed embarkation. No regular personnel transports were available, and most of the freighters which were used as resupply ships lacked both accommodations for troops and the loading devices necessary to handle some of the heavy equipment. Even after an adequate ship had been found (transport Sea Runner),⁸⁶ loading out was seriously complicated by rough seas. In the end it proved necessary to leave some of the equipment behind with a detail of men to take care of it.⁸⁷ Not until 30 October did the weary Marines see the last of that island for which they had fought so long and so bloodily. Some sour punster dubbed it "Nothing Atoll".

Footnotes

1. See Appendix D for further description of caves and their utilization.

2. A type of natural hillside cave with an extremely wide horizontal mouth, partially obstructed by stalactites and stalagmites, which it was all but impossible to seal. On the other hand, such an opening provides far less cover than a small one. Many caves of this type are extremely large, this one extending, with some improvement on nature, entirely through the ridge and out the other side on a different level. Japanese Military Caves on Peleliu. CinCPac-CinCPOA Bulletin 173-45, 27-33.

3. Japanese Navy elements on Peleliu used artificial caves almost exclusively and had what amounted to a monopoly on the skilled labor available. See Appendix D. Army people did not find their sister service's ideas of cooperation especially amusing. See Appendix F.

4. "From the light artillery standpoint, the most emphatic lesson of Peleliu was that 75's fired futilely at Jap caves. 1/11 fired 20,000 rounds at one two hundred yard area in front of the 1st Marines without disturbing appreciably anything below the foliage and topsoil." Ltr Maj J. R. Chaisson to CMC, 8Mar50.

5. Employment of a 155mm gun against the northern ridges was described in Chapter VI. This gun was later emplaced to angle direct fire in through the mouth of the Horseshoe. There was one somewhat bizarre incident, well authenticated, of its gunner sighting in on an individual Japanese and picking him off at 300 yards like a clay pipe in a shooting gallery.

6. "The enemy plan seems to be to burn down the central hills post to ashes by dropping gasoline from airplanes." Tada Record, 7Oct44.

7. Exact distances in this region cannot be gauged accurately as the three different maps used at one time or another are not in scale with each other. Insofar as it is possible to orient terrain features on the standard 1:20,000 map, length of the Pocket at this time would appear to be closer to 1,200 yards.

8. Appearing in reports of the 81st Division as "Mortimer Valley." Place names used herein, unless otherwise noted, are the ones applying at the close of Marine action at Peleliu.

9. Both of these names were applied by the 81st ("Wildcat") Division and constitute exceptions to the rule cited in previous footnote. They are used herein in the interests of easier orientation, inasmuch as Marine reports have no standardized terms for them. A few units referred to the Bowl as "Peleliu Pocket." Aviators, to whom terrain appears much different than it does to ground troops, called it the "Little Slot," as differentiated from the "Big Slot," their name for the Horseshoe.

10. This would appear to be the area lumped together in Japanese reports under the inclusive designation "Suifuyama" or "Suifuzan" (suffixes yama and zan are interchangeable, both meaning "mountain").

11. In a report just prior to the period under discussion, Nakagawa estimated Japanese strength in the Umurbrogol as "about 2½ battalions put together," with "the main body of 2d Bn, 2d Infantry and part of 2d Bn, 15th Infantry, not yet under our control." Tada Record, 28Sept44.

12. This writer had the experience of coming under fire from a cave in the nose of a ridge which had been secured and thoroughly explored more than a week previously. The incident occurred just at dusk, at a range of approximately 300 yards, making it all but impossible to determine the exact source of the shots. Although the sniper had betrayed his general location, most of the next day was required to find and destroy him. Had he chosen a nearer target or an hour of better visibility, he would have been located with little difficulty, but might have done more damage.

13. "These souvenir hunters and the morbidly curious were mostly service troops and aviation personnel on their first operation and had apparently been given a thorough 'snow job' in some rear area to the effect that all Japs died with a jeweled sword in hand. . . . An effective straggler line would have saved many lives." Scantling.

14. Berger.

15. At least, none of which U.S. troops were ever aware. However, by dint of what must be construed as wishful thinking, Colonel Nakagawa was able to inform his superiors: "One part of our unit infiltrated into the air base on the night of the 5th and threw the enemy into confusion and at the same time they set fire to the air base between 0030 and 0250 on the 6th." Tada Record, 6Oct44.

16. Ltr Maj A. J. Doherty to CMC, 15Feb50.

17. Characteristic of the Corps tradition which considers every Marine basically an infantryman, secondarily a specialist, and trains him accordingly. LVT personnel became increasingly available as landing facilities were developed, artillerymen as constriction of the Pocket precluded massed fires, pioneers on relief from shore party duties. Engineers subsequently assumed some of this duty, but operated mostly with the combat teams to which they were attached. Before the end, details were drawn from every Marine ground unit on the island, including Division Hq Bn and 16th Field Depot.

18. Major Joseph E. Buckley, CO of the Weapons Company, had his own method for obtaining "volunteers." As previously noted, souvenir hunters had become a wholesale nuisance. Any men found in Buckley's area without good reason for being there were promptly seized, handed weapons and placed in the line, where they were held by force if necessary. If they behaved themselves, he notified their unit commanders of their whereabouts and employment; otherwise he did not bother, and they were carried as AWOL for as long as Buckley chose to hold them.

19. 33 killed, 116 wounded. Berger.

20. Ltr Maj D. P. Wyckoff to CMC, 13Mar50.

21. 1/7 HistRpt.

22. "At this time diarrhea was prevalent; many men could barely walk as a result of this malady." Gormley.

23. "By juggling men from Headquarters Company and from E Company, some balance had been achieved and rifle company strengths read as follows: E Company 67, F Company 60, G Company 82. . . . These figures . . . (were) much lower than Regiment had reported to Division . . . and I am afraid the result did not exactly endear me to Regiment." Berger.

24. It should be remembered that these particular place names had not been applied at this time. In the 2/7 Journal (but nowhere else) "Walt" is referred to as "Hill X", and definitely located by listing of the target squares in which it lay. "Boyd" is unnamed in any reports of this particular action and designated only by its target squares.

25. Berger.

26. The bodies of many dead Marines were found scattered about the ridgetop, however, sad reminders of Captain Pope's gallant stand there throughout the night of D-plus 4 (see Chapter IV)--Oral statement of Capt. F. T. Farrell, 20Apr49.

27. 2/7 Journal.

28. 2/7 WD.

29. This squad consisted of nine men commanded by Lieutenant Charles R. Hickox, Jr., one of the company's platoon leaders. The presence of this small force on the north face of the objective proved far more valuable than an otherwise discouraging situation would indicate. Lt. Hickox was in continuous radio contact with the battalion commander and was able to place his men in position to be of great assistance to the subsequent attack. Hurst.

30. Two tanks and an LVT flame-thrower had been detailed to support 3/7 that morning, and had to travel up and down three-quarters of the peninsula to reach the battalion. This subsequent reinforcement from the south stands out as the first successful attempt to use the direct line of communication along the East Road. 3/7 WD.

31. The terms "summit" and "top" as they appear in the reports are apt to be deceptive. Boyd Ridge is a narrow razorback, its topographical crest bare and untenable by troops. The rifle units of Company K carried it in one rush, without artillery preparation but covered by machine-gun fire. With seizure of the objective, the company was deployed across the eastern slope with leading elements on the military crest where they remained "for two days and two nights, swapping hand grenades with the Japs on the western slope." Hurst.

32. 2/7 Journal.

33. 3/7 Journal.

34. 3/7 WD. Company F also refused a flank to the edge of the swamp. Since the draw was not much more than 50 yards wide, actual physical contact was not considered essential. Hurst.

35. The Japanese were capable of some amusing deviations from the hard facts of life, but this extraordinary figure would appear the result of inaccurate translation.

36. Tada Record, 3Oct44. This report concludes with an interesting example of Japanese intelligence work: "Judging from the enemy's organization, equipment, tactics and action, units . . . were believed to be Marines with one part of the Australian Army. Their strength was estimated at about five infantry battalions."

37. This ridge was the northerly extension of Hill 200, on top of which 2/7 had been posted for so long without being able effectively to get at the Japanese below. Flanking fire from these positions had played an important part in the repulse of all previous attacks moving in this direction. The thorough job of cave sealing performed by 3/5 on this day nullified fire from this particular ridge as a major tactical factor henceforth. Berger.

38. "The enemy unit that attacked Kansokuyama, our main post in the southeast central hills, was its best picked company. However, more than half of them were killed." Tada Record, 3Oct44.

39. O. P. Smith PerNar. "The Old Breed", 337.

40. DeBell.

41. 5th Mar WD, 4Oct44.

42. O. P. Smith op.cit.

43. By now the lower East Road was being used successfully for supply and evacuation. Fire coming through the draw was partially thwarted by interposing moving tanks as shields for the traffic: supply and ammunition vehicles coming up, and stretcher parties evacuating the wounded.

44. Later, after the 81st Infantry Division had taken over the final mopping-up of the Pocket, engineers bulldozed a tank route through this draw to provide ingress to the Horseshoe from the north. This proved of great value, as the Japanese were discovered subsequently to have both a 75mm field piece and a 47mm antitank gun bearing on the mouth of the valley from excellent concealed positions in caves.

45. 7th Mar R-2 Journal.

46. As will be seen, this was essentially the tactic used subsequently for the successful capture of Baldy and the ridge leading south from it, of which Ridge 120 was actually a spur.

47. This is the hour recorded by 3/7 War Diary, 15.

48. Many details of this action, subsequently verified, derive from an account by T/Sgt. Jeremiah A. O'Leary, a Marine Corps Combat Correspondent, written shortly after the event and published in True Magazine, October, 1945.

49. Captain Shanley had been awarded the Navy Cross during the Cape Gloucester operation. For his action on this day he received, posthumously, a gold star in lieu of a second Navy Cross. He died that evening.

50. Collis's promotion to First Lieutenant had been received at regimental headquarters that morning, but he never learned about it.

51. Peppard.

52. It may be noted that both RCT 321 and the 7th Marines had attacked from the north. However, in both cases the main efforts were concentrated against the enemy's eastern perimeter with the aim of gaining control of the East Road.

53. Harris comments on preliminary script.

54. Harris Interview. Officers from all infantry units who have been consulted speak of the constant pressure from above to speed up the action and get the operation over with.

55. Authority for last statement is 5th Mar War Diary. The battalion commander (interviewed 28 June 1949) states succinctly of this position: "It didn't have any future." It was his contention that lines drawn on a situation overlay are meaningless, per se: that the only position worth risking men's lives to hold is one capable of emplacing weapons with which to support a further advance. Since this one obviously was not, he preferred to withdraw and probe elsewhere in hopes of finding a better one.

56. This declivity could be classed as possibly the worst death trap on the island. Not until more than six weeks later, after the Sisters, the Brothers, and the Horseshoe had all been secured, were 81st Division units, which had relieved the Marines, able to operate effectively within the Bowl. Colonel Nakagawa's CP was located in a deep dual cave in the China Wall near the northwestern end of the gulch. (See Chapter VIII.)

57. This sortie was made for the sole purpose of knocking out caves suspected of containing the heavy weapons which were firing on the airfield and southern perimeter lines. In this it was entirely successful, as no further fire of this nature was received. There was no intention of retaining a foothold within the Horseshoe at this time. Harris Interview.

58. The containing lines had rendered the Japanese so ineffectual outside of the Pocket that there was pseudo-serious talk of simply setting up barbed wire around the whole area, classifying it as a prisoner of war enclosure and ceasing all further assault operations.

59. "Sometimes, if a pressure is not exerted, a battle may be allowed to deteriorate into a stalemate simply because of the peculiarities of mass inertia. On the other hand, unreasonable goading . . . may breed a resentment which . . . may result in an effect just the opposite of the one desired. . . . This sensitiveness on the part of the commander is the mark of a great commander as distinguished from one who is merely good. In this particular instance . . . I am certain that no pressure was received by Corps from Commander Expeditionary Troops and Landing Force and that none was upon the Division by the Corps Commander." Wachtler.

60. Subsequently the engineers improved this and constructed two additional ramps of a similar nature which gave ingress to LVT flame-throwers into the western ridges. Harris Interview.

61. Patrols operating in this gulch came across the decomposing bodies of 12 U.S. soldiers, grisly reminders of the 321st Infantry's grim efforts to find a practicable route of attack across the peninsula in order to isolate the Umurbrogol Pocket, more than two weeks before.

62. The two pack howitzer battalions of the 11th Marines had been evacuated several days before, but fortunately the heavy seas raised by the near-typhoon had prevented loading of their guns aboard ship.

63. Verbal statement of Lieutenant Colonel Gayle interviewed June 1949. 5th Mar War Diary states that Company G's attack carried to the objective and "secured Hill 140". No hour is given, but farther on the same account states: "By 1500 F Co had secured Hill 140."

64. Tada Record. Exact orientation of these Japanese place names is difficult, as the sketch map showing them is completely out of scale with any U.S. maps. "Oyama" is the area of Nakagawa's CP and would appear to include both the China Wall and Five Brothers, together with immediate approaches from north and west. "Suifuzan" is an inclusive term embracing all of the Umurbrogol system north of Oyama, most of which was in U.S. hands by this time. "Kansokuyama" is Hill 300, possibly including the Five Sisters. "Higashiyama" is Walt Ridge, perhaps Boyd Ridge as well. See Map 14.

65. Ibid.

66. The depleted battalion suffered 31 additional battle casualties during this period, however, and its total strength was down to 277 on 14 October.

67. It should be understood that this was not a "line" in any strict military sense; rather, a series of positions, held in no great strength and often in only visual contact with one another.

68. As has been noted earlier, place names were improvised hurriedly and without due consideration as to whom should be so honored. Thus, when the mapping party found LtCol R. W. Boyd occupying the elevation seized a week earlier by Maj Hunter Hurst's 3d Bn., 7th Marines, it was tagged on the map "Boyd Ridge". Similarly, when LtCol L. W. Walt, executive officer, 5th Marines, was discovered inspecting positions on the ridge immediately to the south, where Capt E. T. Pope had won the Medal of Honor on D-plus 4, and which had been secured on 3 October by LtCol S. S. Berger's 2/7, it became "Walt Ridge" henceforth. A number of people were most unhappy about this--and still are.

69. Lyman.

70. Sandbags were ordered used by the 5th Marines in the attack on the northern end of Peleliu. First use was made in emplacing the 155mm gun in that area and subsequently on the 155mm and 105mm positions placed to angle fire into the mouth of the Horseshoe. At the time of the events here related, the positions atop Walt and Boyd Ridges were fortified with both sandbags and armor plate taken from disabled LVT(A)'s. The latter, however, proved too heavy and too awkwardly shaped to be adaptable on any wide scale. Harris Interview.

71. The inadequacy of such maps as are available of the Pocket area makes exact orientation of positions exceedingly difficult. Those indicated on the accompanying maps are intended to be no more than approximate. According to 3/5 Record of Events, Company I's advance carried to Target Square 140-T. This would appear to place it abreast of the Five Brothers and 150-200 yards west of the China Wall. Although the Regimental War Diary refers to an overlay, none concerning this movement can be found in the records. It would be inconclusive in any event since the China Wall, as such, is not definitely located on the map it was designed to overlie.

72. The 1st Battalion, 7th Marines, was attached operationally to the 5th Marines for this action. The route of its first day's attack took it past the western flank of the Five Sisters. On the second day the battalion attempted to penetrate Death Valley but was able to gain only about 50 yards. The mission was completed on the third day by advancing up the next draw to the westward. 1/7 Historical Report.

73. It was estimated that these several operations reduced the area of the Pocket by 30% to 40%. It now measured roughly 400 yards x 500 yards. O. P. Smith, Personal Narrative. Here again distance is impossible to scale accurately owing to map variations.

74. These allusions to Higashiyama are not clear, as the Japanese had conceded its loss several days previously and made no claims to having retaken it.

75. Tada Record, 14, 15Oct44.

76. Coincidentally, IIIAC command post was set up on shore at 0800 on this date. IIIAC OpnRpt, Enc B, 11.

77. Mission of the Island Command as defined in Western Carolines Sub Area Order #2-44, 2Nov44: "Maintenance of an air base for offensive and defensive use; defense of the Command." This group assumed responsibility for unloading and construction on 28Sept, prisoners of war on 21Oct, and air defense on 10Dec. It did not assume responsibility for ground defense until 13Jan45.

78. CTS 32 OpnRpt, 95.

79. Fort.

80. This battalion had relieved 2/7 in these positions on 14 October but had remained on the defensive during these two days.

81. To which the 1st Bn., 323d, was attached operationally at this stage. It should be noted, however, that over-all command of the attack remained in the hands of the 1st Marine Division for another four days, pending the arrival and establishment of the 81st Division CP on 20 Oct.

82. A smaller island lying northeast of Ngesebus. For its seizure on 9 October see Chap. VIII.

83. Further evidence of the futility of relying on reason when fighting the Japanese was furnished during the dark morning hours of 18Jan45, when the enemy actually did stage a counterlanding on Peleliu. By that time most of the 81st Division, in turn, had left the island. With the coming of dawn those elements remaining combined with units of the Island Command to hunt out the intruders with no great difficulty, killing 71 and capturing two. 81st Unit Hist, 200.

84. Captain H. W. Jones, formerly CO of the Battalion Headquarters Company, had taken over this command only a short time before, following the death of Captain Shanley. The 3d Battalion was especially hard on rifle company commanders: of four who served in that capacity during the Peleliu operation, three were killed in action and one wounded. Hurst.

85. Berger.

86. Most 5th Marines personnel were embarked in Sea Runner, but at least two other vessels took part in this movement. Benedict.

87. A detail of 13 men with 15 vehicles of 1st MT Bn. They left on or about 13Nov44. Ltr 1stLt J. B. Darnell to CMC, 13Mar50.