Chapter 6
Saito's Last Battle

The Japanese were in a trap, the jaws of which were ever tightening. But, like cornered rats, they were capable of frenzied reaction.

Every indication supported the belief that the enemy remaining in the island's upper end were ill-equipped, ill-supplied, and, in many cases, ill. Communications had suffered almost a complete breakdown, hospital facilities were practically non-existent, and the shortage of food and water was telling on their morale. Passionate phrases of patriotism, which Japanese commanders erupted on the slightest provocation, were hard to swallow on an empty stomach.

And yet, to say that the Japanese were demoralized would be stating an untruth. After all, who had imposed this sorry condition but the hated American devils? And what better way is there to vent one's wrath and relieve one's gnawing hunger than to kill an American? Though these questions we surmise, the Japanese reaction in the last days of the campaign points convincingly to their accuracy.

As noted previously, Holland Smith's order for 5 July had designated 1200 as the attack hour. In the hours of daylight prior to noon, however, the two assault divisions were to conduct preliminary moves necessary to initiate a coordinated attack.

Tanapag Plain

For the 27th Division, assigned the left sector (see Map 21), these preliminary actions involved moves of approximately 1,600 yards for units along the beach and only minor adjustments for those on the high ground inland.

Terrain was varied in the division zone: the left half included an 800-yard expanse of coastal flatland, while the right faced the turbulent maze of hills and ravines that is inner Saipan; near the center, an abrupt escarpment divided the extremes. The coastal flat and the hills which immediately dominate it were assigned the 105th Infantry, while the rugged interior fell to the 165th. The 106th Infantry had been pinched out of the attack on the previous day and reverted to division reserve.

In addition to mopping up the Flores Point area, the 105th Infantry was to push its front abreast of the 165th Infantry, occupying the high ground inland from Tanapag. Events of the day, as will be seen, prevented the regiment from achieving the desired gains.

The left assault battalion (the 2d) inched along the beach, investigating extensive defensive installations which the Japanese had prepared to counter U. S. landings. Although virtually no resistance was encountered, each position had to be carefully explored, lest dangerous thorns be left behind. The units moved steadily along the beach, through the scattered remains of Tanapag Village, and toward Road Junction 2. As the battalion neared that junction, the Japanese opened suddenly, fiercely, with a fusillade of machine-gun fire. The advance stalled.

As was so often the case, the enemy could not be located. And the Japanese discouraged

curiosity with frequent, well-placed bursts of fire. Finally it was believed that the fire was issuing from a battered Japanese landing barge beached off to the left front. Two medium tanks were immediately dispatched along the coast road to destroy this supposed source of the trouble. After blasting the barge and apparently silencing the Japanese fire, the tanks swung off the road for the return to U. S. lines. This proved a bad move: the vehicles had driven into a Japanese mine field and one was completely wrecked before the truth was known. As if by signal the Japanese opened fire again, this time more heavily than before.

The rest of the day was spent in attempting to locate the enemy positions, all to no avail. The 2d Battalion dug in for the night just short of Road Junction 2. (See Map 21.)

The 105th Infantry's right assault battalion (the 3d) faced a zone that included a portion of the plain and also the rugged hills that closely fringed it. To negotiate this sector, it was necessary to advance one company (L) along the top of the bluffs while another (K) moved below on the flat. To the north of the battalion's lines, two narrow gorges could be seen, each cutting deep into the escarpment. Along the high ground between the two ran the bumpy, winding cross-island road. The 3d Battalion's right and the 165th Infantry's left would bisect the necks of the draws while the open mouths yawned toward units on the coastal flat. The near draw caused no trouble, but the second (later named "Harakiri Gulch") developed into a stubborn core of enemy activity.

The 3d Battalion moved well at first, but as the left company (K) approached a palm grove about 800 yards east of Tanapag Village a hail of enemy fire immobilized it. The fire emanated from the grove itself as well as the choppy ground to the east. A thrust by machines of the 762d Tank Battalion into the coconut grove failed to diminish the Japanese resistance and the company held up for the night. Higher on the bluffs the right company (L) had traversed the first draw without difficulty, but in moving across the road toward Harakiri Gulch, it was greeted by fire from cliff positions on the far side. Several battalions of 27th Division artillery were registered there, along with point-blank bursts from a 3d Battalion antitank gun, but the enemy was not dislodged and became less hospitable than ever. In this unenviable position the battalion dug in for the night; Company L atop the spur facing Harakiri Gulch, Company K below and just southeast of the palm grove.

During the day the 1st Battalion had remained in regimental reserve, occupying an assembly area south of Tanapag Village.¹

The 165th Infantry was not required to conduct as extensive preliminary operations as the 105th. The change in direction and reassignment of zones had placed it well ahead. At 1300, however, the regiment launched its attack,² 2d Battalion on the right, 3d on the left. The former, in contact with the 4th Marine Division, moved without difficulty, but the latter, battling the upper reaches of the Harakiri Gulch cross-compartment, was stopped after a short 100 yard gain. Like elements of the 105th Infantry to its left, the 3d Battalion encountered heavy machine-gun fire from the draw's northern wall as well as mortar fire, which often made upright movement suicidal.

With the right extended well ahead in contact with the 4th Marine Division and the left checked on the high ground overlooking Harakiri Gulch, the regiment established its night defense. The 1st Battalion had remained in an assembly area just northeast of Radar Hill in reserve.³

The day's advance in the 27th Division zone had been disappointing, but General Griner hoped that the next day would see the unit moving rapidly again. In a message issued at 1800, 5 July, the general instructed the 105th Infantry and the left battalion of the 165th to move out at 0700, 6 July, and advance their lines to Makunsha, an elongation of the line held by the right battalion of the 165th Infantry. This was to be accomplished by 0900, so that the division could attack with units abreast. Stressing the

DOMINATING POSITION, captured by 165th Infantry, flanked Tanapag-Makunsha Plain. For once, the Japanese were caught in the open, vulnerable to fire such as that being delivered by this heavy machine gun.

importance of the move, Griner's message stated: "It is imperative that the left of the Division line be advanced to be abreast of the other units for the coordinated attack at 0900. The early advances will be pushed aggressively."⁴

From O-7Z to O-8A

General Schmidt's order to his 4th Marine Division on 5 July specified that the 23d Marines conduct preliminary operations for the seizure of O-7Z (the regiment had patrolled just short of this objective on the previous day). This would provide a favorable line of departure for the day's attack. The move was to be completed by noon so that the other two regiments could pass through and continue to division objective O-8A, some 1,400 yards forward of O-7Z. Included as the principal terrain feature of intermediate objective O-8A was the high ground on the north side of Karaberra Pass, a deep gouge in Saipan's upper middle. From the pass, the objective line ran in a southeasterly direction to a cove on the east coast, labelled Inai Fahan (see Map 21).

The scheme was executed with only minor hitches.

The first of these occurred when the 23d Marines took somewhat longer than expected to occupy O-7Z. Moving with three battalions abreast, the regiment made rapid progress on both flanks, but the center unit (1st Battalion) was delayed by the nightmarish terrain through which it was forced to move. To remedy this situation the two flank units, already on the objective, stretched their lines to cover the vacant middle. Colonel Jones reported the objective occupied at 1315,⁵ and the passage began. The task of mopping-up its difficult zone of action took the 1st Battalion until 1500, at

which time it joined the rest of the regiment in division reserve.⁶

Assuming the left half of the division front, the 24th Marines made good progress. Initially the regiment had been slowed by the failure of the 25th Marines to appear on the right flank; but, subsequently, when Colonel Hart ordered his battalions to move out anyway, the sweep was rapid. By 1600 the unit was in possession of its portion of O-8A, having seized the high ground on both sides of Karaberra Pass and, to prevent Japanese infiltration, extended its lines through the pass itself.

The 1st Battalion, moving in contact with the 27th Division, experienced little difficulty. The speed of its movement was regulated chiefly by units on its flanks. For the right assault unit, the 2d Battalion, the going was tougher: wooded draws and ravines lying within its zone contained many troublesome caves. The enemy in these required systematic extermination. As has been so often indicated, cave cleaning was a slow process. Karaberra Pass brought more of the same: here grenades and antipersonnel mines were rolled down upon the Marines as they trudged the valley floor. This particular area was extremely defensible and, had the Japanese chosen to make it so, could have been much worse than it was. By late afternoon the 24th Marines occupied the high ground on both sides of the pass as well as the floor itself.

The 3d Battalion, 24th Marines, had remained in reserve throughout the day. Two of its companies (K and L) had spent the previous night in the Army division's lines, but by 1100, 5 July, both had been relieved. With the coming of darkness the battalion was ordered to establish a secondary line, one company (K) being attached to the 1st Battalion to bridge a break on the left flank.

Since the 24th Marines had moved out in the attack prior to the arrival of the 25th Marines on the right flank, no contact existed between the two units until after 1500, when a patrol from the 25th Marines began operating through the area. When the advance stopped for the night a 400-yard separation still existed. To prevent the Japanese from using this as a funnel through which to pour substantial forces (individual infiltrators were next to impossible to stop), the 25th Marines detailed a strong combat patrol with orders to rove the interval during the night. In addition, heavy machine guns were emplaced and mortars and artillery registered to cover the area should the need arise. The effectiveness of this security was not tested by the Japanese, however.⁷

Colonel Batchelder's 25th Marines had a difficult time schedule to meet: it was to pass through the right half of the 23d Marines, a maneuver involving a lateral move of about 2,500 yards. This move, moreover, could not commence until the 1st and 3d Battalions had been relieved by Army units. By about 1000 the relief took place, and the lengthy hike began. Delayed in its start, the regiment was late in effecting the passage, arriving at about 1330. Batchelder ordered his 1st and 2d Battalions to attack abreast, 1st on the right. The 3d Battalion would follow the assault battalions at 500 yards, using a route generally in the center of the regiment's zone.

The 1st Battalion started the day in an exciting way. In the early morning hours, prior to relief by Army units, the Japanese had set up a machine gun in Company A's thinly-spread front lines and began spattering bursts through the battalion command post. The proximity of friendly troops precluded the use of mortars against this target and dictated infantry assault. Designated to expel these intruders were the men of the attached 1st Platoon, Company A, 20th Marines, whose normal engineer functions were temporarily shelved. At 0540 the engineers attacked, killed or routed the Japanese, and restored the position. This incident cost a total of 12 Marine casualties.

After passage of the 23d Marines' lines, the 1st Battalion drove forward without encountering resistance. The heat and humidity were terrific, however, and the long hike resulted in

many heat exhaustion casualties.⁸ Also proving a hindrance to swift movement were several aerial bombs (250-pound) buried along the coastal road. These bombs, installed with noses barely protruding from the ground, were intended to destroy U. S. vehicles moving along the road. Fortunately, these explosives were spotted before damage was done. The battalion commander, Lieutenant Colonel Mustain, ordered the attached engineers to begin detonating the inverted bombs; and, since the task appeared too great for a single platoon to accomplish, he requested services of the engineers attached to the 3d Battalion. A bulldozer was provided to fill the holes left by the explosions, and by 1700 the road was open for traffic.

The itinerant 2d Battalion, meanwhile, moved on the left half of the 25th Marines' front. This well-travelled unit had performed directly under the control of not only its parent regiment, but 4th Division, Northern Troops and Landing Force, 2d Division and 8th Marines as well. Its "Cook's tour" had left little time for rest. Back with familiar faces once again, the battalion advanced steadily to its assigned objective.

By 1700 the 25th Marines dug in on O-8A. With the evening rations came the first fresh socks, and the Marines gratefully replaced their limp and soggy ones, which were unceremoniously interred.⁹

In addition to the 14th Marines, the 4th Marine Division had the benefit of reinforcing fires from the XXIV Corps Artillery during the afternoon. One small, troublesome plateau forward of the 2d Battalion, 24th Marines, had been effectively pounded by the massed fires of eight battalions of artillery.¹⁰

For the 2d Marine Division, in NTLF reserve, the day was quiet and uneventful.¹¹

Night of 5-6 July

About 2030 the 24th Regiment reported 250 Japanese to its front, apparently organizing for an attack. Artillery, mortar, 37mm and machine-gun fire converged upon the area and the Japanese dispersed. Later, at about 2300, there were minor attempts at infiltration in the regiment's zone, but no counterattack materialized.

Patrols from the 27th Division observed approximately two companies of Japanese northwest of Hill 767 as well as scattered smaller groups in the vicinity of Makunsha and northeast of that village. Here again, artillery fires kept the enemy stirred up and disorganized, and nothing more than desultory rifle fire developed.

Japanese aircraft made a feeble effort during the night. Just after dark three to five planes came over Saipan but failed to press anything resembling an aggressive attack. Two bombs, dropped near Aslito Airfield, achieved the rather unspectacular feat of slightly damaging one U. S. plane and cracking a windshield on another. While skittering nervously about the area, one of the raiders was downed by a U. S. night fighter. From the half-hearted, hesitant manner in which the raid was executed, the impression was derived that the planes were attempting air evacuation of key personnel. Just after midnight a single Japanese plane appeared over the island and was shot down by a night fighter.¹²

D-plus 21--6 July

The Ditch and the Gulch

NTLF plans for 6 July had prescribed a push to O-8 (see Map 21) with the remaining terrain on the island divided equally between the two assault divisions. To allow the left elements of the 27th Division an opportunity to come abreast, the 4th Marine Division and the right battalion of the 27th Division would delay their attack until 0900.

If this scheme had worked out, it is likely that the two divisions would have moved side by side to the end of the island. But, as will be seen, the early morning preliminary attack of the Army division's left units met stiff resistance, and the plan was thwarted before it could even begin. When it was apparent to Holland Smith that the going would be slow and difficult on the left and rapid on the right, he decided to change his plan.

The modification, issued shortly after 0900 on 6 July, altered the division boundaries to the extent that the 4th Marine Division would expand to the northwest, pinch out the 27th Division north of Makunsha, and take over the entire frontage for the sweep to Marpi Point. This move would not only enable the 27th Division to mop up properly the Tanapag-Makunsha-Harakiri Gulch area, but might also cut off enemy retirement to the north.

The 105th Infantry, which had launched its preliminary attack at 0700, was deeply embroiled in trouble by the time that news of the change arrived.

The scheme of maneuver for the 2d Battalion took into account the minefield which had been discovered on the previous day. To avoid this area, Major Edward A. McCarthy planned to swing his companies in column along the beach until past the minefield, at which time a two-company front could be assumed.

The battalion moved out under a protective rolling barrage fired by the 249th Field Artillery Battalion. The first 100 yards' advance was rapid, carrying the unit almost past the minefield; but, as the artillery lifted to targets farther to the northeast, the soldiers suddenly met a hail of grazing fire spitting across the coastal flat. Despite efforts of the battalion commander to regain momentum, the attack stalled.

Then began the difficult task of locating the source of the enemy fire. Though no one could realize it until much later in the day, the Japanese were well established in a shallow ditch only 150 yards to the front. This natural trenchwork, a drainage rut running from the base of the escarpment to the sea, was relatively insignificant when contrasted with the cliffs and gulches on the right flank and had not been credited with being as strongly occupied as it was. As long as artillery fire showered on the area the Japanese had kept their heads down; but, as soon as it had lifted to the northeast, they had risen from their shelter and stopped the 2d Battalion's advance.

Tanks were immediately requested, but a communication failure delayed their arrival. While waiting, the 2d Battalion improved its position, broadened its front, and inched about 50 yards closer to the ditch. Work on the minefield was also progressing; one platoon from Company A, 102d Engineer Battalion, had undertaken the hazardous task shortly after 1000. Necessarily working in the prone position, the platoon made slow progress. The "mines" were the same type that had worried the 25th Marines on 5 July: buried 250-pound aerial bombs with noses fused to explode under the weight of vehicles.

At about 1000 five medium tanks from the 762d Tank Battalion arrived at Major McCarthy's command post. How and where should the tank thrust be made? (As yet no one had identified the ditch as the trouble spot.) The road was a possibility, but this could be conveniently interdicted by Japanese antitank weapons, and, also, the spindly bridge across the ditch would likely be mined or wired for demolition. The route selected was the railroad track running through the center of the battalion's zone.

Moving single file along the thin road bed, the tanks were soon stymied when the lead tank snarled its tracks on the narrow steel rails. While work was in progress to clear a path for the second tank to move around the first, Japanese antitank weapons opened up and achieved hits on both. Although shells pierced the vehicles, neither was put out of action. The turn of events had clearly indicated that this route was a poor one. By use of a cable both tanks were hauled clear of the area.

While the tanks were executing their abortive gesture, a rifleman in the left company (E) spotted some movement in the ditch ahead. The target seemed ideal for the 60mm mortars, but these, unfortunately, were out of ammunition. An effort by an SPM (Self Propelled Mount)

VICTIMS OF DOREY'S ATTACK, these Japanese blocked the 6 July advance of the 2d Battalion, 105th Infantry, along the Tanapag Plain until Dorey's tanks arrived on the scene.

was also disappointing; after moving up the coast for a short distance the driver, discouraged by near misses from the same antitank weapon that had hit the two tanks, abandoned the undertaking.

The next attempt, though boldly conceived, was also unsuccessful. One rifle squad rushed the position, but was repulsed after a wild skirmish.

At last the source of the trouble had been definitely located. But now what was to be done about it?

The situation was generally unchanged until about 1530 (except for the commitment and ultimate detachment of the battalion reserve, Company G, which will be described later). At this time three Army light tanks, previously operating with the 3d Battalion around Harakiri Gulch, appeared in the 2d Battalion's zone in search of a mission. This they were not denied. Upon learning about the situation, Lieutenant Willis K. Dorey, USA, decided to roam out and examine the ditch.

Dorey moved to the ditch unhampered by antitank fire and found a tanker's dream target. Japanese soldiers were jammed along the ditch, almost shoulder to shoulder. One of his tanks was put out of action when an enemy soldier placed a magnetic mine against its side; but, after leaving his third tank to protect the cripple, Dorey ranged on alone. Using machine guns and cannister on the tightly-packed foe, Dorey slaughtered from 100 to 150 as he blasted the ditch from end to end.

At 1700, following this massacre, men of the 2d Battalion advanced with no difficulty. Such few Japanese as remained alive were incapable of resistance. The long, hard struggle for the ditch was over.

The 3d Battalion, 105th Infantry, had found the going no easier. This unit, as noted before, was straddling the cliff line with one company on the flat and another atop the bluffs. The morning attack would depart from the line seized on the previous day: the left just short of the palm grove and the right facing precipitous Harakiri Gulch. In accordance with General Griner's order, the battalion moved out at 0700 to bring its lines abreast of units farther to the right.

Selecting a defiladed route, the left company (K) moved to the edge of the grove without receiving fire. Once the unit emerged from its cover, however, two machine guns contested further approach. Savior in this situation was Lieutenant Dorey,¹³ whose tanks ploughed into the grove and destroyed the two machine guns. This relieved the pressure and the soldiers quickly swept to the northern edge of the trees.

Fire from the upper lip of Harakiri Gulch's mouth, meanwhile, had made all movement across the flat costly. Any further advance would place the left company directly in the fire-lanes. Capture or effective neutralization of the gulch area was therefore imperative if forward movement with a reasonable degree of safety were to be made.

With hopes of neutralizing the fire from this direction, or at least minimizing it, Lieutenant Dorey's tanks commenced working over the area. This measure appeared effective; the volume of enemy fire diminished and, finally, stopped. When, at about 1000, Dorey was forced to leave the area to replenish ammunition, the small arms fire of the 3d Battalion maintained the spray of rounds against the cliffs. The enemy remained dormant.

At this point in the proceedings word was received of the change in division boundary and the effect this would have on the scheme of maneuver. The weight would shift from left to right as the division pivoted toward the sea. Strong pressure would be exerted by the 165th Infantry to swing down from the hills and to the beach north of Makunsha.

With the shift in balance, the 2d Battalion, 105th Infantry, would take over the entire frontage on the coastal plain, allowing the 3d Battalion to utilize its full strength on the bluffs above. This involved relief of Company K by Company G. The latter had spent the early morning hours mopping up in the vicinity of Tanapag Village. Relief took place at about noon following an unusual train of events.

Upon receiving notification of the change in attack plans, Company K's commander decided to undertake seizure of a small knoll north of the palm grove. Since Dorey's tanks had returned after supplying themselves with ammunition, these would be used to prevent Japanese interference from Harakiri Gulch and the cliffs around it.

Almost as soon as Company K moved out to seize the knoll, however, the enemy opened with devastating effectiveness from the cliffs to the right. It appeared that they had been biding their time awaiting just such a move. The right of the company was stopped. As the left lunged forward in a determined effort to carry the objective, the Japanese also started moving: from the cliffs towards the ditch immediately in rear of the knoll. What threatened to be a savage hand to hand struggle never materialized, for at this precise moment a blasting, crashing explosion rent the air, hurling Japanese and pieces of Japanese in all direction. The terrific concussion knocked most of the men of Company K from their feet, but the Japanese suffered more. What had been a thriving surge was suddenly--nothing. As a result of this blast, all enemy fire ceased in this area for nearly an hour.

What had caused this terrific explosion? The most plausible explanation seems to be that the leading Japanese of the counterattack group had blundered upon a mine, detonation of which set off a number of others. Disorganized and slightly dazed by the concussion, Company K did not immediately move to the objective. By the time the unit had

regained its composure and was ready to move up, Company G arrived and effected the relief.

The tanks had not escaped excitement either. While cruising near the mouth of Harakiri Gulch, the machines were met by two Japanese soldiers who ran out and attached a magnetic mine to one and threw a Molotov cocktail at the other. Both tanks were put out of action. Lieutenant Dorey ordered the surviving vehicles to depart from the area. The crews of the disabled tanks made good their escape on foot.

Higher on the bluff, meanwhile, the right company (L) of the 3d Battalion made unsuccessful ventures into Harakiri Gulch during the morning. Of the belief that it was imperative to build up suporting fires before attempting to cross the draw, the company commander attempted to move base of fire units into positions from which to cover the advance of the rest of the company. As part of the same scheme, tanks moved up from the gulch's longer axis into covering positions. Though the efforts continued throughout the morning, all failed. In the process, one tank was lost when a Japanese rifleman placed a magnetic mine against it.¹⁴

While Company L exercised its unsuccessful stratagems against Harakiri Gulch during the morning, the change in attack plans had been made known. For the afternoon attack the 27th Division would still have four battalions in the assault, but the emphasis changed from left to right. The new scheme is well described by an Army historian:

The basic plan of movement called for the 3d Battalion, 105th Infantry to attack across Harakiri Gulch and up onto the high, almost level plateau that covered the area between there and Paradise Valley. The companies of these two battalions would then wheel left, go down the face of the cliffs from above, and sweep out across the Tanapag Plain to the sea, each arriving on the plain at a point progressively farther east toward Makunsha. The 2d Battalion, 165th would proceed north down Paradise Valley, cleaning out this strongpoint, and reaching the beach just above Makunsha. While this operation was going on in the hills, the 2d Battalion, 105th Infantry, on the division's left wing, was to make a limited attack northeast along the beach.¹⁵

This simple, direct plan was destined to failure, however.

The 3d Battalion, 105th Infantry, failed to gain during the afternoon. The commander of the left assault company (L), decided to undertake no further moves until he had located and destroyed the enemy positions which had stalled his morning efforts. The right company (I), after moving about 75 yards down into the gulch, met rifle fire and was stopped. Later in the afternoon the company withdrew to its starting point on top of the hill.

A determined afternoon thrust against Harakiri Gulch was made by the 1st Battalion, 165th Infantry,¹⁶ but here again results were not encouraging. The brunt of the 1st Battalion's fight fell to Company A.

This company faced the upper part of the gulch at a point where the 3d Battalion, 165th, had been stalled on the previous day. Rather than attempt another cross compartment move, the change of direction would allow the company to advance down the draw's longer axis. After descent to the gulch floor, the company cautiously crept through a series of man-made ditches that had carved the area to resemble a huge washboard. All eyes were straining toward the rocky slopes for some sign of the Japanese among the large umbrella-like trees and little straw shacks that dotted the valley. At first, all was quiet. Then, unexpectedly, the tiny shacks rocked with the force of explosion after explosion. The chain of blasts continued for about 15 minutes, during which time the soldiers kept under cover in the ditches. When silence returned and investigation was possible, the startling truth was discovered: about 60 Japanese, scattered through the small abodes, had placed grenades against their abdomens and committed suicide. It was this incident, that gave the gulch its name.

Hopes that these suicides would mark the end

of Japanese resistance in the draw were soon dispelled; accurate grazing fire stopped the unit as it resumed the push. In addition, the shacks were now ablaze. The explosion of the grenades had turned the combustible huts into fiery patches, and the heat forced soldiers nearby to pull back. To expose one's body for even a moment was enough to attract a well-aimed burst, and casualties mounted. Handling their fallen comrades proved a difficult task for the men of the company. The number of men killed or wounded while attempting to evacuate other wounded was alarming; loss to would-be rescuers was altogether out of proportion to the number of personnel being saved. It was a vicious circle.

When it became apparent that Company A's attack had stalled, the company commander ordered the unit to withdraw. Extrication was not easy, since the soldiers were under heavy fire; and it was after 1500 when they were finally reassembled at the gulch's upper entrance.

Probably the greatest contributory factor in this failure was the fact that the company had had no assistance on either flank. The 3d Battalion, 105th Infantry, as mentioned before, had not advanced during the afternoon. Had this unit moved, it might have relieved some of the pressure on Company A's left. On the right, terrain prevented all but one platoon of Company C, 165th Infantry, from attempting to help the gulch attackers, and even this platoon was slow in starting. By the time it finally got underway, Company A had recognized the attack as a lost cause and was withdrawing, so that the platoon's lonely venture onto the slopes of Harakiri Gulch was violently repulsed.

Farther to the right, the 2d Battalion was unable to proceed down Paradise Valley in its assigned zone and had received permission to maneuver through the 1st Battalion's sector. This move, of course, was stymied by the 1st Battalion's failure to make progress.

In an effort to press the assault down to the Makunsha Plain area, the 3d Battalion, 106th Infantry, was attached to the 165th with orders to attack between the two flank units. By the time this order (issued at 1500) was translated into action, however, darkness was approaching and the battalion failed to move into the lines. The unit was ordered, therefore, to remain in an assembly area southwest of Hill 767.

In midafternoon, with forces at Harakiri Gulch stalemated and the battalion along the beach immobilized (prior to Lieutenant Dorey's eminently successful tank foray against the ditch), General Griner ordered Colonel Bishop to commit his reserve (1st Battalion), to swing past the gulch strong point and continue to the beach 1,200 yards beyond Road Junction 2. To throttle the gulch while the 1st Battalion moved past, Company G would swing into position facing up the draw. For this mission the company was attached to the 3d Battalion.

The attack proceeded generally according to plan. Moving out at 1715, the 1st Battalion advanced rapidly. Enemy fire against the unit ranged from moderate on the left to heavy on the right, in direct proportion to the proximity of the high ground where numerous caves overlooked the coastal flat. With the assistance of SPM's, men of the 1st Battalion neutralized the cliff dwellers and continued their surge toward Makunsha. By dark between 800 and 900 yards had been gained, the unit having carried to about 150 yards from the beach.

At this point, however, the battalion commander, Lieutenant Colonel O'Brien, discovered that his front had progressed beyond the 2d Battalion along the beach and that his troops had swung partially across the 2d Battalion's front. To improve this situation, O'Brien ordered his forward companies to pull back and tie in with the 2d Battalion. By about dark most of O'Brien's unit had accomplished this, but the reserve (Company C), which had concentrated its attention and effort on the caves on the right flank, did not completely assemble in the defense area until 2100. Because of darkness at the time of its arrival, Company C did not find or occupy its assigned position in the two-battalion defense area, but, instead, overlapped a portion of the 2d Battalion. Also, the late adjustment in the lines masked the fires of several heavy machine guns and one antitank gun positioned in the 2d Battalion area. O'Brien commented on this unsatisfactory state of affairs, but it was considered unwise to make any shifts in the dark. Later in the night when trouble

developed, these masked weapons opened fire, pinning down one company (A) of the 1st Battalion and causing some confusion.

Since units facing Harakiri Gulch had not moved, the right flank of the perimeter was open, Company G, 300-odd yards to the southeast, being the closest friendly unit on that flank. The latter company, unsuccessful in sealing the mouth of the gulch, had disposed itself on dominating terrain from which it could cover most of the area by fire. (See Map 22 for positions of 27th Division units on the night of 6-7 July.)

There is some indication that the exact dispositions of the 105th Infantry were unknown to both the regimental and division commanders. According to the 27th Division G-3 Journal, the only message relating to the gap was received from the 105th Infantry on the late afternoon (1845) of 6 July: ". . . G Co. filling gap between 1st and 3d Battalions."¹⁷ At 2000, having received no further information on this, General Griner issued a field message which merely directed a continuation of the attack the following day.¹⁸ Had the division commander known that the gap existed, he still would have had time to order a reserve unit to establish a connection. The 3d Battalion, 106th Infantry, was in reserve southwest of Hill 767, less than 2,000 yards from the gap.

Despite the discouraging events around Harakiri Gulch, the late afternoon advance along the coastal flat was encouraging indeed. But on the other hand, everyone from the corps commander to the front line riflemen was well aware of the enemy capability to strike at any point along the lines and that any position might find itself in the vortex of an all-out banzai attack. Since the 15 June landings the western coastal plain had seemed the most likely area for such a thrust. An enemy attack there would be no surprise, but no one foresaw the consequences of the battle that later developed.¹⁹

Expansion of the Front

After jumping off at 0900 on 6 July, the 4th Marine Division's attack moved well for about 20 minutes. Then a temporary halt was demanded by a change in the NTLF scheme of maneuver. With the 4th Marine Division taking over the entire island sector for the final push, it was necessary to insert another regiment into the lines. Accordingly, General Schmidt ordered the division reserve (23d Marines) to move in on the division left flank and expand the frontage to the west coast at a point just beyond Makunsha. In effect, the division would attack both northeast and, temporary at least, northwest.

This deployment, so easy on paper, took considerable time. To get into its newly-assigned zone, the 23d Marines hiked 4,300 yards, arriving at its line of departure at about 1415.

Here the regiment faced a steep, 300-foot cliff which had to be descended to reach the coastal plain and, eventually, the beach. Jumping off shortly after 1415, the Marines discovered that the cliff was cloaked with heavy, scrubby underbrush and pitted with scores of natural caves. Working troops down the declivity was a challenge in itself, there being only two paths. As soon as Marines began filtering down these, Japanese riflemen and machine gunners countered with well-aimed fire. Thus denied, the regiment busied itself in a search for other routes of descent to the plain.

Since this portion of the long cliff line dominated the Black Beaches (northeast of Makunsha), positions encountered were probably part of the prepared beach defenses. From commanding terrain atop the cliff, however, the Marines could observe large numbers of Japanese fleeing to the north along the plain. Many of these, no doubt, had been flushed by the pressure of the 27th Division. The Marines cascaded the fire of rifles, machine guns, mortars, and half-tracks upon these popular targets.

By dark, after a series of unsuccessful efforts to work small detachments down the cliff on ropes, the 23d Marines established a defense along the lip of the precipice for the night.

NORTHERN SAIPAN. Mt. Petosukara was seized by the 25th Marines on 6 July, and Mt. Marpi by the 2d Marines on 9 July. Rocky coastline offered last ditch positions to stubborn Japanese.

Contact was established with the 27th Division along the cliff top, but none was attained with the 24th Marines on the right.²⁰

In the division center the 24th Marines started the day's activities with the 165th Infantry to its left and the 25th Marines to its right; but, with the alteration in division boundaries, the 23d Marines replaced the Army regiment. The juggling of units took some time and the 24th had to wait; an independent move would greatly endanger the unit's left flank.

About 1400, however, the 24th Regiment was permitted to push forward again, and the momentum was maintained until commanding ground about 400 yards short of Objective O-8 was reached. During this surge only sporadic resistance was encountered. The large gain (1,400 to 1,800 yards) far outstripped the 23d Marines on the left and the aforementioned gap developed. To protect this flank the 24th Marines' reserve (3d Battalion) was committed, but a physical tie-in was not accomplished.

During the day the left assault battalion (1st) had encountered an ingeniously-designed dummy position, containing sham antiaircraft guns, gunners and radar. The 2d Battalion, on the other hand, found a very real foe in an irregular cliff formation to its front. As the Marines approached, about 200 Japanese began a hasty retreat to the north, speeded by fire from 75mm half-tracks, 37mm guns, and rockets. The infantry then picked up the pressure, pursuing the Japanese as long as daylight permitted.

Though the regiment's left had not been tied in, the right was in firm contact with the 25th

Marines, which likewise had enjoyed rapid movement.²¹

The 25th Marines' zone on Saipan's east coast was similar to that of the 105th Infantry on Tanapag Plain. Here, however, the coastal flat ended in steep drops to the sea rather than a gradual merging of beach and water. On the inland side of the flat there were also cliffs, but these rose precipitously above the plain to a second, smaller plateau, in turn fringed by another slope which rose to a third plateau. Along this escalading terrace the 25th Marines would move until the levels fused and culminated into Mt. Petosukara, the dominating terrain short of O-8. (See Map 21.)

To negotiate the hodgepodge, the 1st Battalion would move along the coastal flat, while the 2d Battalion straddled the cliffs and plateaus inland. The reserve would move along the second plateau in rear of the assault units.

Off at 0900, the regiment advanced rapidly. In the 1st Battalion's zone, well suited to vehicular movement, 13 tanks from Company C, 4th Tank Battalion, preceded the sweep. One of these was used as a mobile observation post to spot naval gunfire. Missions and corrections were relayed from the tank to Shore Fire Control Party 251 and from there to the firing ship. This device worked very well and a number of missions were directed in this manner. Principal naval gunfire targets were enemy positions in the cliff lines, both the one at the water's edge and the one which fringed the coastal plain.

A glance at the map and a knowledge of the Japanese system of defense was enough to convince the 25th Marines that trouble would come from the cliffs. It did. Men of the 1st Battalion, particularly, were exposed to plunging fires as they trudged through the hot cane fields. But, employing coordinated tank-infantry, the Marines systematically destroyed the troublemakers and pushed on.

More difficult was the problem of the cliff which dipped from the plain to the water's edge. Here Japanese soldiers could remain concealed in the caves until Marines attempted to approach; then, suddenly, they would open with devastating bursts. Usually, the cliffs were too steep to permit descent from the top, and an approach along the narrow strip of beach at the bottom was a sure invitation to disaster. Naval gunfire was helpful in sealing some of the cave entrances, but the caves were many and the job was a slow one.

Among the worst features of the situation was the fact that most of the caves housed more noncombatants than soldiers. Interpreters were unceasing in their efforts to get cave occupants to surrender, and many of these efforts proved fruitful. In all too many instances, however, civilians were dissuaded by Japanese soldiers who threatened or took the lives of any who would weaken to the Americans' invitations. After several Marines were killed in sincere efforts to extricate civilians, patience became somewhat worn. After that, the occupants were invited to come out; and, if they hesitated over-long, the caves were blown.²²

Realizing that seizure of the coast line would demand considerable time, Lieutenant Colonel Mustain ordered the assault companies of his 1st Battalion to push on, leaving his reserve to work the caves. With assault units relieved of this detailed, time-consuming activity, the 1st Battalion swept rapidly forward. The infantrymen were glad to have the tanks along, particularly when two enemy vehicles hove into view shortly after noon. Under the fire power of the U. S. armor the enemy withdrew. The tanks were glad to have the infantry along, too: while firing into the fringing cliff on the left flank, they were rushed by several Japanese, who were cut down by surrounding Marines before they could attach the magnetic mines they carried. Incidents like these nurtured a growing mutual respect.

The 2d Battalion, meanwhile, kept pace through its nightmarish zone. Pushing a coordinated attack over such terrain was a strain in itself, even had there been no crafty

Map 21
Progress Lines for
5 and 6 July 1944

defenders present. By late afternoon, however, the battalion had moved to Mt. Petosukara.

Establishment of the defensive position for the night was very difficult, inasmuch as a vertical gap of about 500 yards existed between battalions. In the reshuffling process, which was still underway after dark, the 1st Battalion flushed a covey of Japanese directly into the 2d Battalion's lines. In an effort to prevent more Japanese from entering its lines, the latter called down an emergency artillery barrage which fell among the approaching left units of the 1st Battalion. Fortunately, the fire was lifted before any Marine casualties resulted. After a brief flurry, the enemy was repulsed.

The situation regarding the separation between the two battalions was reconciled shortly and they dug in for the night, 2d Battalion generally on the west and 1st Battalion on the east slopes of Mt. Petosukara. The gap was never filled, although a platoon from Company C was placed on a dominating point between them.

Just at dusk, 700 to 800 civilians came through the 1st Battalion's lines to surrender. This produced a delicate and dangerous situation, since Japanese soldiers could have followed the civilians with a strong attack. Though considerable apprehension was caused by this possibility, no attack developed.

The 3d Battalion, followed in reserve, committed only one company during the day. This company (K) became embroiled in a sizeable scrap when one of its platoons searched out some caves in a by-passed hill and found a virtual beehive. After a lively skirmish, the hill was cleared of the enemy and 61 Japanese bodies were counted.²³

Night of 6-7 July

The night of 6-7 July was to be long-remembered. It produced the most devastating banzai attack of the entire war. Compared with the stroke that was launched against the 105th Infantry, all other Japanese efforts during the night take on an extremely pale color. And yet, carrying the comparison further, these "pale" attacks were more violent than any that had gone before, and for those individuals who found themselves in the midst of one of these, even the smallest conflict may assume elephantine stature. Ordinarily, the fury of a battle is judged on its over-all effect, numbers of troops involved, and casualties but fails to consider the effect, on any one individual.²⁴ The smaller thrusts of this momentous night will be covered first.

Individuals of the 2d Battalion, 25th Marines, spent an unpleasant night on Mt. Petosukara. After a series of smaller probing missions the Japanese struck the Marines' right with about one rifle company. Coming in the midst of a heavy downpour, the attack, for a time, appeared to be making headway. But 60mm mortar shells, dropped within 50 yards of the front lines, together with small-arms fires and artillery, broke the enemy. But many Japanese, who seeped through the 2d Battalion's lines, set about a vigorous program of harassment. There was no sleep for anyone atop Mt. Petosukara for the remainder of the night. Early in the morning the battalion aid station was attacked by a small group, but corpsmen and doctors, plus Marines in the area, rallied to the threat and repulsed the enemy. Daylight revealed that between 60 and 70 Japanese had been killed in the abortive punches at the 2d Battalion, 25th Marines.²⁵

Elsewhere along the 4th Division front there were numerous instances of minor infiltration attempts, with all units reporting activity. For the 3d Battalion, 24th Marines, stationed to protect the regiment's exposed left flank, the action was particularly vicious. Just before dawn the enemy rushed forward, armed with grenades and "idiot sticks," encouraging themselves with loud shouts and screams. Though this thrust failed to penetrate the Marines' lines, some of the enemy fell within

five yards of the foremost foxholes. The action report of the 3d Battalion, 24th Marines, estimated the number of enemy killed in this attack at "more than 200."²⁶

The Banzai Attack

Since the early stages of the Saipan planning General Holland Smith had been convinced that, if and when an all-out banzai attack were launched at Saipan, it would come along the western coastal plain. It was this belief (shared by all officers on the Corps staff) that prompted Holland Smith to caution Major General Watson to keep the left of his 2d Marine Division strong as it advanced up the coast. Inasmuch as Watson and his staff had arrived at a similar conclusion, the division, in its long surge from the Charan Kanoa beaches to Tanapag seaplane base, had been ever alert to Japanese attack along the coastal plain.

On 2 July, with the Japanese becoming more and more compressed in the island's northern end, the Corps commander issued an order warning all units to take special precautions against a mass attack and to tie in--by physical contact--all front lines at night.

On 6 July, Holland Smith, accompanied by Major General Watson, visited the 27th Division command post and while there warned General Griner that "a banzai attack probably would come down Tanapag Plain late that night or early the next morning." Further, Holland Smith cautioned Griner to "make sure that his battalions were physically tied in." To this, according to General Smith, General Griner gave assurance that "his division was prepared for just such a Japanese move."²⁷

Holland Smith's warning was timely.

In regard to this conference, General Griner later wrote:

He [Holland Smith] makes no mention of the change in orders which was the purpose of his visit. The first evidence any of us had of the impending counterattack was the report of a Jap prisoner, captured in the early evening of July 6 by the 105th Infantry, who was sent to the Corps G-2 for further interrogation.²⁸

The attack which hit the 1st and 2d Battalions, 105th Infantry, on the early morning of 1 July was the culmination of a long series of reverses for Saipan's defenders. General Saito had virtually given up hope on 25 June when he wrote: "There is no hope for victory in places where we do not have control of the air. . . ." And since then nothing had occurred to warrant optimism. By 6 July, according to the NTLF G-2 Report, the Japanese had two courses of action remaining: "to fight a withdrawing action ending in complete annihilation on the northern tip of the island, or to attempt to muster their disorganized and crumbling forces into one all out 'banzai' charge." Saito chose the latter.

Huddled in Saito's cave command post at about 0600, 6 July, his subordinates heard the general's order:

I am addressing the officers and men of the Imperial Army on Saipan.

For more than twenty days since the American Devils attacked, the officers, men, and civilian employees of the Imperial Army and Navy on this island have fought well and bravely. Everywhere they have demonstrated the honor and glory of the Imperial Forces. I expected that every man would do his duty.

Heaven has not given us an opportunity. We have not been able to utilize fully the terrain. We have fought in unison up to the present time but now we have no materials with which to fight and our artillery for attack has been completely destroyed. Our comrades have fallen one after another. Despite the bitterness of defeat, we pledge "Seven lives to repay our country."

The barbarous attack of the enemy is being continued. Even though the enemy has occupied only a corner of Saipan, we are dying without avail under the violent shelling and bombing. Whether we attack or whether we stay where we are, there is only death. However, in death there is life. We must utilize this opportunity to exhalt true Japanese manhood. I will advance with those who remain to deliver still another blow to the American Devils, and leave my bones on Saipan as a bulwark of the Pacific.

As it says in the "SENJINKUN" [Battle Ethics], "I will never suffer the disgrace of being taken alive,"

Map 22
Front Lines in Tanapag Plain Area
on Night of 6-7 July

Map 23
Progress Lines for
7, 8 and 9 July 1944

and "I will offer up the courage of my soul and calmly rejoice in living by the eternal principle."

Here I pray with you for the eternal life of the Emperor and the welfare of the country and I advance to seek out the enemy.

Follow me.

If the subordinates, had, indeed, followed the example of General Saito, there would have been no attack. The tired general, feeling that he was too aged and infirm to be of use in the counterattack, held a farewell feast of saki and canned crab meat and then committed harakiri.²⁹

Several days prior to Saito's final order the assembly of remaining Japanese forces had begun. Some Japanese were out of reach behind U. S. lines, others were hidden in the deep shadows of caves, but all available were mustered. Marpi Point, Paradise Valley and Harakiri Gulch were sites for mobilization and reorganization. Many provisional units were formed in an effort to achieve some degree of tactical unity among the assorted groups and individuals.

Weapons and equipment were wanting and some Japanese were armed only with grenades or crudely-fashioned spears. One Japanese staff officer (Major Kiyoshi Yoshida) who participated in the battle estimated the total participants at 1,500, many with no weapons. In the light of subsequent events, however, his estimate appears too low. Assuming that he was misinformed on this point, it is not at all surprising; even at the assembly points, U. S. artillery and mortar fire hampered Japanese efforts to organize their troops and an accurate count was impossible.

As the conglomerate force moved toward Makunsha, which would serve as a point of departure, it was subjected to persistent, heavy concentrations of artillery fire. This resulted in a limping approach with leaders calling upon every conceivable device to maintain the unity of their commands. Greater silence by the sprawling force would have helped; early in the evening patrols from the 27th Division Reconnaissance Troop had detected unusual activity. This had resulted in the large number of unobserved fire concentrations thrown into the Makunsha vicinity.

Despite what has been said about the unsatisfactory state of weapons and equipment, the willingness of the individual Japanese to die honorably and to take seven American lives to repay his country somewhat compensated for the shortage. Here was a determination which was seldom--if ever--matched by the fighting men of any other country.

Preceding the main force as it approached U. S. lines were a number of strong patrols. In their reconnaissance-in-force missions the patrols poked and prodded the lines, seeking vulnerable points and goading the soldiers to open up and reveal positions. These explorations, no doubt, divulged the 300 yard gap between the 1st and 3d Battalions, 105th Infantry. The blow fell at 0445. First and hardest struck were the isolated positions held by the 1st and 2d Battalions, 105th Infantry. The attack on these units hit from front, flank and, after moving through the gap, the rear. Almost as soon as the attack was launched, communications to the rear were cut. It was then simply a matter of two isolated battalions of soldiers fighting for their lives. This they did and did well. Some of the soldiers stacked so many dead Japanese forward of their positions that it was necessary to move to get fields of fire.

Major McCarthy, commanding the 2d Battalion, described the onslaught:

It reminded me of one of those old cattle stampede scenes of the movies. The camera is in a hole in the ground and you see the herd coming and then they leap up and over you and are gone. Only the Japs just kept coming and coming. I didn't think they'd ever stop.³⁰

The sheer weight of this attack, its ponderous momentum, carried it through the soldiers' lines. In addition, hundreds of Japanese moved past the isolated defense area, using the gap on the right.

The fanatical surge then carried to the 3d Battalion, 105th Infantry, located on the high ground overlooking Harakiri Gulch. Here the soldiers were on better defensive terrain; the Japanese had to climb up to them. The lines held, and the Japanese suffered staggering

LIEUTENANT COLONEL WILLIAM J. O'BRIEN, USA, 1st Battalion, 105th Infantry, posthumously awarded the Medal of Honor for refusing evacuation after he was seriously wounded during the Tanapag Plain banzai attack and instead manning a jeep-mounted .50-caliber machine gun until his death.

losses. This portion of the battle did not cease with the coming of daylight on 7 July but continued until midafternoon.

Next to feel the impetus of the determined thrust was the 3d Battalion, 10th Marines, in firing positions about 500 yards southwest of Tanapag Village.³¹ The Japanese mass that struck this unit was apparently that portion of the banzai force which had moved through the gap between the 1st and 3d Battalions, 105th Infantry. Hardest hit were Batteries H and Headquarters and Service, in position on the left of the railroad track (see Map 22). Batteries I and G, set up on the right of the track also got into the fight, although they were not as closely engaged as the other two. The only battery that actually fired its howitzers during the melee was H, the others being forced to remain silent by the presence of other artillerymen to their front.

First Lieutenant Arnold C. Hofstetter, an officer of Battery H, tells the story of the vicious action:

Small arms and machine gun fire was heard to the front and right front at considerable distance at about 0300, July 7, 1944. No information as to source could be obtained. Later, the fire appeared to come closer and, since it appeared that the position might be attacked, the gunners were told to cut time fuzes to 4/10 second in preparation for close in fire.

About 0515, just as it was getting light, a group of men seen advancing on the battery position from the right front at about 600 yards. It was thought that Army troops were somewhere to the front, so fire on this group was held until they were definitely identified as Japs at about 400 yards. We knew that our men manning the listening post were somewhere to the front, so the firing battery was ordered to open fire with time and ricochet fire on the group to the right. Firing was also heard from the machine guns on the left.

After the howitzers started firing, it sounded to me like numbers 3 and 4 were not firing enough, so I went to these pieces to get them firing more. I got them squared away and stayed with number 4 until Japs broke through wooded ravine to the left, and I heard that word had been passed to withdraw. The firing battery fired time fuze and percussion fuze so as to get a close ricochet. Some smoke shell was fired. Cannoneers were shot from their posts by machine guns and small arms, . . . which interrupted the howitzer fire and finally made it impossible to service the piece.

The remainder of the firing battery fell back about 150 yards from the howitzers, across a road, and set up a perimeter defense in a Japanese machinery dump. This was about 0700. We held out there with carbines, one BAR, one pistol, and eight captured Jap rifles, Japs got behind us and around us in considerable strength. They set up a strong point in a point of woods to our rear. . . . About 1500, an Army tank came in from the right and got the strong point and Army troops relieved us.

I estimate that 400-500 Japs attacked the position. They used machine guns, rifles, grenades, and tanks.³²

Some mortar fire fell in the area, but it may have been from the . . . rear.

During the early phases of the battle, before Battery H was forced to abandon its firing positions, one of its howitzers was turned completely around to engage a Japanese medium tank which had ploughed through to the rear. At a range of less than 50 yards the artillerymen sent 105mm shells crashing into, and destroying, their armored target. Eventually, however, the pressure from Japanese infantrymen was too great; Battery H's outnumbered survivors pulled back. In the frantic confusion and haste of the retreat, they failed to remove the breech blocks or firing locks from the howitzers. The Japanese, oddly enough, neither used nor destroyed the weapons while they had control of the area.

Battery I had been attacked at 0455, with Japanese advancing astride the railroad track. The first thrust was repulsed and the Japanese fell back to reorganize. A series of actions, followed this same pattern, continued until 0650, when all of the battery's small-arms ammunition was expended. At this time the battery commander (Captain John M. Allen) ordered the firing locks removed from the howitzers and the unit to fall back to the positions of Battery G. Here the two batteries continued the fight until relieved by elements of the 106th Infantry later in the day.

Located directly behind Battery H, headquarters and service personnel of the battalion were forced to retire after heavy, close-in fighting. Among the 136 casualties in the battalion was the commanding officer, Major William L. Crouch, who was killed. Japanese killed in the encounter numbered 322.³³

Even as the Japanese surged directly against

PRIVATE FIRST CLASS HAROLD C. AGERHOLM, 4th Battalion, 10th Marines, awarded the Medal of Honor for making repeated trips through a fire-swept area in a jeep-ambulance to load and evacuate singlehandedly approximately 45 wounded men of the 3d Battalion, 10th Marines, before he was mortally wounded by an enemy rifleman.

the 3d Battalion, brother artillerymen of the 4th Battalion, 10th Marines, in position to the southeast, received pressure from the fringe of the main enemy tide. In defense of its firing positions, the battalion killed about 85 Japanese. Despite the turmoil of the banzai attack, the 4th Battalion responded on the morning of 7 July to an urgent request for support originated by the 23d Marines, fighting down the cliffs above Makunsha.³⁴

By 0530 the lunging stab had pierced another 400 to 600 yards to the command post of the 105th Infantry. Rallying to the defense of the installation, headquarters personnel met the Japanese in what was to be the end of the savage surge. The soldiers fought from deserted Japanese positions and exchanged blow for blow, bullet for bullet, with their attackers. Their defense perimeter was never penetrated; and, after several hours of close-in fighting,

the battle was over. The enemy survivors withdrew.³⁵

While most of the credit for the repulse goes to the personnel fighting the Japanese at close hand, the fires of the 27th Division artillery also contributed prominently to the outcome. From 0515 to 0615, 7 July, the three 105mm howitzer battalions fired a total of 2,666 rounds--an average, of more than 44 rounds per minute for a period of an hour.

Meanwhile, as the main prong of the attack spent its fury in the manner just described, two secondary efforts struck up Paradise Valley and Harakiri Gulch. Emerging head-on into the prepared positions of the 2d Battalion, 165th Infantry, the Paradise Valley Japanese were massacred as they spurted from the narrow funnel of the ravine. The other effort was just as funereal. Striving to break out of Harakiri Gulch and seize the high ground to the south, the enemy were slaughtered by the 1st Battalion, 165th, and the 3d Battalion, 105th, in position along the rim of the gulch. At dawn, the 3d Battalion, 106th, was inserted in the center of this position.

The same cross-compartment move which had been attempted by the U. S. soldiers during the day was now being tried--from the opposite direction--by the Japanese. The shoe was on the other foot, and the Japanese met murderous fire as they exposed themselves. Neither of these secondary efforts achieved anything positive for the Japanese, and in the process virtually exterminated the entire force. Not content that their efforts had been repulsed, the Japanese continued to expend themselves in futile charges throughout most of 7 July.

After the Japanese banzai attack had passed over them, men of the 1st and 2d Battalions, 105th, were cut off from friendly troops. Of those present when the attack started, only 25 per cent now remained, the rest being killed, wounded or missing. In addition, they had no communications with other units and there was little ammunition left. Carrying or dragging the wounded, the decimated force started trudging to the south to reach friendly lines.

Japanese units, repulsed by units farther south, trickled in small, disorganized bands toward Makunsha, their starting point, and more often than not met the battered remains of the 1st and 2d Battalions on the way. This resulted in a series of uncoordinated skirmishes and finally forced the soldiers to form a perimeter defense at the village of Tanapag, both flanks tied into the water. In this position they fought off a succession of thrusts which continued for about 15 hours.

As if fate had not been cruel enough, about noon of 7 July, Army artillery observers spotted the soldiers' perimeter at Tanapag and, presuming it to be an enemy outpost, directed fire down upon it. A great number of the survivors, driven into the water by the deluge, were picked up by U. S. destroyers at the reef's edge. The remainder was evacuated by means of amphibian tractors and DUKW's. Casualties (killed, wounded and missing) in the two battalions were heavy: 1st Battalion--349; 2d Battalion--319.³⁶

Thus ended one of the most devastating single battles of the war. The mop-up would take another two days, but the fight was gone from the Japanese survivors. Just how many Japanese were committed to this struggle is unknown. Major Yoshida's estimate of 1,500 was doubled by a Korean civilian, Mamoru Iwaya, who also participated in the attack. Both agreed that there was a shortage of weapons (Iwaya said about one-third of the personnel carried rifles and the rest had spears). Estimates of the number of machine guns: Yoshida said one heavy and "a few" light machine guns, Iwaya said "approximately 10" light machine guns. The latter also thought that the attack was personally led by General Saito, who, he thought, was killed in the fight. In a large jumbled force such as this, no one

BANZAI AFTERMATH. Japanese dead litter Tanapag Plain in the zone defended by the 2d Battalion, 105th Infantry.

could be sure just how many participants there were.³⁷

The following narrative account by Major General Griner, the 27th Division Commander, gives a concise summary of the over-all picture:

At 0530, 7 July, the Commanding Officer, 105th Infantry, telephoned that tremendous masses of the enemy were attacking his 1st and 2d Battalions, that terrific mortar fire was falling on them, and that the telephone lines had gone out. He estimated the enemy attackers to number fifteen hundred (1500). I alerted the 106th Infantry (less one (1) battalion)³⁸ for movement and telephoned General Erskine, Chief of Staff, NTLF, giving him the estimate of the enemy strength. Further investigation of the situation at the front led me to believe that the enemy strength was underestimated, and at 1100, 7 July, I again telephoned General Erskine, stating that this was the main enemy counterattack and consisted of no less than three thousand (3000) enemy soldiers. I stated that I was employing the Division Reserve and requested that a battalion of the 2d Marine Division be turned over to me as soon as practicable at a specified location. He attached to the Division the 3d Battalion, 6th Marines. Our tank losses from enemy action were somewhat large. Twice we requested that an additional platoon of tanks, which represented only a small fraction of those available to NT&LF, be made available to the Division. These requests were denied, obviously because that headquarters did not accept my version of the importance

SERGEANT THOMAS A. BAKER, USA, Company A, 105th Infantry, posthumously awarded the Medal of Honor for refusing, even though wounded, to leave the line. When his unit was forced to withdraw during the savage banzai attack, Baker, unable to move by himself, requested that he be propped against a tree from which he could carry on the fight alone. Later, when the sergeant's body was found, his weapon was empty, and eight Japanese lay dead before him.

of the action then in progress. At this time, the 4th Marine Division was experiencing no opposition in the advance to the northeast. I issued orders about noon to the Commanding Officer, 106th Infantry, to counterattack the enemy to relieve the 1st and 2d Battalions, 105th Infantry, and to recapture a Marine Artillery Battery which, unknown to this Division,³⁹ had taken positions in the Division Zone of Action six hundred (600) yards in rear of the forward line of the preceding night. It later appeared that this artillery was in support of the Marine regiment, 2d Marine Division, which was on the left of the 4th Marine Division.⁴⁰ The 106th Infantry advanced five hundred (500) yards in front of their line of Departure but failed by three hundred (300) yards to relieve the battered elements of the two battalions of the 105th Infantry. This incident was witnessed by Major Fuller,⁴¹ Liaison Officer from NT&LF, who reported to his headquarters that the advance could easily have been continued and subsequently gave me the same report. Colonel Stebbins, the Regimental Commander, upon being interrogated later, stated that the decision to advance no farther was his own and made despite the protest of one of his battalion commanders, that although the enemy fire being received was not heavy, he realized that masses of Japs were being passed over who could fire in his rear. He stated that the mopping up of the area already gained consumed the remaining period of daylight. When the line stabilized, I inserted the 3d Battalion, 6th Marines, between the right of the 106th Infantry and the left of the battalion of the 105th Infantry which was still intact⁴² with instructions that the 3d Battalion, 6th Marines would advance only to conform to the movement of the right of the 106th Infantry. I also made arrangements for the evacuation by DUKWs of the beleaguered elements of the two battalions, 105th Infantry. I then called the Chief of Staff, NT&LF, as was my custom as the daylight hours drew to a close, to report the existing situation. I advised that the line was firmly established, without break or gap of any kind, and stated that another battalion of the 2d Marine Division would provide sufficient reserve if made subject to my call. The Chief of Staff then advised that the 2d Marine Division had been moved forward to a line about fifteen hundred (1500) yards in rear of the then existing front and that General Smith was contemplating passing them through the 27th Division at daylight the next morning. General Erskine then asked if the 27th Division could complete the job and wipe out the enemy. I replied that we could. He stated that he would call back. Within thirty minutes he telephoned that the 2d Marine Division would pass through the 27th Division at daylight and that the Division (less 165th Infantry) would pass to NT&LF reserve. The 165th Infantry would be attached to the 2d Marine Division.⁴³

As indicated, the 1st and 2d Battalions, 106th Infantry, were ordered at about noon of 7 July to counterattack and relieve the battered battalions of the 105th. Though this move regained the positions of the 3d Battalion, 10th Marines, and pushed farther to the north, the advance was halted 200 to 300 yards short of the Tanapag Village perimeter. Since physical relief had not been accomplished, it remained for the two battalions to be evacuated by amphibian

tractors. The LVT's of the 773d Amphibian Tractor Battalion supplied the transportation and at 2200, 7 July, the last survivor left the perimeter.

With the 106th Infantry's advance through the battlefield, the true scope of the enemy attack was revealed: Japanese bodies littered the plain area. An accurate count would have to await complete mop-up, however.

When the 106th Infantry stopped, an interval existed between it and other 27th Division units along Harakiri Gulch. To fill this, the 3d Battalion, 6th Marines, was committed. As night fell the 27th Division's front was solid, extending from the beach, across the plain, through Harakiri Gulch's mouth, along the top of the cliffs as far as Paradise Valley, then inland to a connection with the 4th Marine Division.

Against the coastal portion of this line the Japanese executed a vigorous thrust during the night of 7-8 July. Though the strength, organization and planning of this second attack were weak by comparison with the previous effort, the vigor and determination with which it was executed did not vary. Shortly after midnight a report came from the observation post atop Hill 767; Japanese force moving south. Enemy "feeler" patrols probing for weak spots were engaged as early as 0300, but the information that they took back to their commanders must not have been encouraging. No holes existed.

The attack, launched about 0430, struck the 1st Battalion, 106th Infantry, along the beach. Grazing defensive fires swept across the flat beach like a scythe, cutting all in its path. The Japanese never seriously threatened penetration of the lines and those that attempted to swim around the flank were spotted and killed. Shortly after 0700 the Japanese attack petered out and died. The 106th Infantry estimated that it had slaughtered about 1,000 of the enemy, while at the same time its own losses were practically negligible.

As actions of the tanks on 7 July have not been mentioned thus far, it will be well to summarize their activities. Available to the 105th Infantry were two platoons of light tanks and one platoon of mediums. Early in the morning of 7 July before the regimental commander was fully aware of the tragic circumstances of his 1st and 2d Battalions, he ordered one platoon of lights to proceed to the front lines. After moving along the beach road as far as Road Junction 5 (in the vicinity of Battery H's firing positions), two of the tanks were lost due to enemy actions and a third slipped off the road and "bellied up." The personnel of the latter were rescued by the platoon of mediums which also had ventured up the road as far as the road junction. The presence of enemy soldiers along the route made it apparent that a tank thrust, unprotected by infantry, would be a costly operation, and the tanks withdrew to the command post of the 105th Infantry. There they assisted the headquarters personnel in the defense of the installation.

The next tank excursion was not made until shortly after noon. This followed the shelling, by friendly artillery, of the beleaguered band of soldiers in the vicinity of Tanapag. It will be recalled that the intensity of this fire had driven many soldiers out onto the reef. Once there, they were subjected to small-arms fire from Japanese located along the beach. In an attempt to eliminate this source of trouble and relieve the situation, three light tanks sortied. Bends in the shore line masked their fire, however, and the three vehicles ventured upon the reef in search of direct firing positions. This proved a bad maneuver. All three fell into a hole and had their motors drowned out.

Meanwhile, two medium tanks advanced along the beach road and reached the isolated perimeter. There Major McCarthy, the 2d Battalion commander (and one of two officer-survivors), went on board one machine and used its radio to describe his dire situation to the regimental commander. On the tanks' return trip, Major McCarthy followed on foot with a small detachment. After the latter's arrival at the 105th Infantry command post, the entire situation was reviewed, providing the first complete summary of what had taken place. This resulted in Griner's orders to the 106th Infantry to push out and relieve the men at the battered perimeter.⁴⁴

Meanwhile, the 2d Marine Division had been alerted and moved to a position which would

2D DIVISION MOP-UP following banzai attack. Marines had to be alert for wounded Japanese who could still fire one more shot or hurl one more grenade.

afford depth to the west coast defenses. A line from Tanapag Seaplane Ramp extending about 3,000 yards inland was established by dark of 7 July. The NTLF Operation Order for 8 July instructed the 2d Marine Division, less detachments,⁴⁵ to "advance in the present 27th Infantry Division zone of action, pass through elements of [that] division on its front line, attack, mop up and destroy enemy elements. . . . Upon passing through the 27th Infantry Division assume operational control 165th Infantry and 3d Battalion, 6th Marines, . . . assume tactical control within the assigned zone of action at 0630 [8 July]." Upon passage of its lines the 27th Division would revert to NTLF reserve.⁴⁶

The 2d Division's Mop-Up of 8 and 9 July

No effective resistance was offered the 2d Marine Division after it had passed through the 27th Division on the morning of 8 July, although large numbers of poorly armed and totally disorganized enemy were encountered. Some Japanese used the familiar device of hiding in caves and firing from the deep shadows, but the bulk seemed eager to die and made headlong rushes at the Marines. This foolish expenditure simplified the mop-up.

The 6th Marines, advancing on the division left, uncovered one strong pocket of resistance in a wooded swampy area just east of Tanapag Village. Here an estimated force of 100 holed-up Japanese set about a lively defense of their positions. Flame-throwers, tanks, and 75mm half-tracks blasted at the strong point, but the fight continued. Since this sector promised to be extremely time-consuming and since it was

desired that the mop-up continue, one company (F) stayed behind to contain and destroy the enemy force while the remainder of the regiment swept on. By dark the 6th Marines had reached the water's edge and stopped for the night.

Meanwhile, many Japanese had waded out upon the reef to escape the onslaught. This was but a temporary reprieve, however inasmuch as amphibian tractors, embarking combat patrols, were sent to encourage the enemy to surrender and, if they refused, to eliminate them. The 6th Marines' action report characterized the day's activities as an "attack . . . against a disorganized, swarming mass of Japs . . . waiting for the final death blow. . . ." Adding the hundreds found strewn through its zone to the number that it disposed of, the 6th Marines made a "conservative estimate" for 8 July of 1,500 dead Japanese in its area.⁴⁷

A correspondent, who surveyed much of the battle area on 7 and 8 July, was moved to write "The whole area seemed to be a mass of stinking bodies, spilled guts and brains."⁴⁸

On more rugged terrain inland the 8th Marines encountered much the same type of resistance. Here, with more caves to investigate, the sweep was somewhat slower. One wooded pocket encountered, a deep ravine, caused a minor delay. The tactic that previously had worked so well was again employed: one company stayed behind to contain and mop-up while the rest continued the push. The 8th Marines' zone, being inland of the main path of the banzai attack, contained fewer enemy bodies; but, even so, there was conclusive evidence that a terrific battle had taken place.⁴⁹

As noted before, the 165th Infantry had been attached to the 2d Marine Division for the mop-up of 8 July. Actually, however, this regiment had begun mopping up on 7 July almost as soon as the Japanese attack dissipated. Fighting through deeply-gouged Harakiri Gulch was much easier on 7 July than on previous days. The banzai attack relieved the soldiers of the

SERGEANT GRANT F. TIMMERMAN, 2d Tank Battalion, posthumously awarded the Medal of Honor for his action during the 8 July mop-up. Timmerman blocked with his body a Japanese grenade which would have fallen through his tank's open turret hatch upon the other crewmen inside.

worst of their tasks. By dark the 1st Battalion had moved through Harakiri Gulch and had reached the plateau overlooking the coastal plain. The regiment was thus placed in a favorable position for the descent.

On the 8th, however, the soldiers delayed their sweep, awaiting arrival of the Marines on their left flank. By about 1130 the Marines were abreast and the push was resumed. Moving down the cliff proved difficult. Wooded fissures and eroded cuts provided excellent hide-outs for die-hard stragglers from the banzai attack and eradication of these was a Herculean task. Independent smaller actions--a job for squads and platoons rather than a coordinated battalion attack--were necessitated by the great number of indentations in the cliff face.

While the 1st Battalion, 165th Infantry, stalked the stragglers in the cliff, the 2d Battalion by-passed Paradise Valley, leaving the

FOLLOWING BANZAI ATTACK, 2d Division passed through the 27th Division and mopped up the area. Here Marines, deployed as skirmishers with the welcome support of a medium tank, move along Tanapag Plain in search of enemy stragglers.

mop-up to the 3d Battalion. This resulted in a rapid surge to Makunsha Village for the former. The valley was still a cesspool into which many stragglers had drained, and the task would take time. Darkness cut short the 3d Battalion's work on 8 July, and the unit withdrew to positions at the valley's upper entrance for the night.

On 9 July, while the 2d Battalion waited in Makunsha, the 1st Battalion finished its cliff task and moved to the coast. The blasting this cliff received had altered its appearance; much of the underbrush was gone and many of the caves and crevasses had been filled. Paradise Valley was also cleared of Japanese during the day. The 3d Battalion systematically combed the valley's jagged slopes and by about noon emerged on the coastal plain.⁵⁰

Activity in the sectors of the 6th and 8th Marines followed the pattern established on 8 July. The 8th Marines met little difficulty and devoted a share of its fire power (including attached tanks) to support the 165th Infantry.

Action against the enemy that had swum out to the reef continued. Efforts to induce these to surrender were to little avail: only 14 prisoners were taken by the LVT's patrolling there, while about 100 resisted and either were killed or took their own lives. One Japanese officer was observed to behead four of a

group of his soldiers before he himself was killed by the Marines.

One episode, described by First Lieutenant Kenneth J. Hensley of Company G, 6th Marines, indicates the difficulties experienced in convincing the Japanese on the reef that they should surrender. Hensley was in command of four amphibian tractors and 24 riflemen on 9 July when the incident occurred.

. . . We closed to about 100 yards from the 50-60 enemy on the reef and motioned them to come to the boat.⁵¹ They motioned us away. One officer armed with a rifle and standing on a rock, apparently controlling the 50-60 enemy soldiers, aimed his rifle at the boat several times. We covered this officer with machine guns and closed with the LVT(4)'s to about 50 yards. We again motioned for them to come out to the boats. The Jap officer then fired at the leading boats, and a machine gun, apparently about .50 caliber, opened fire on the boats putting two holes completely through one LVT(4). Many of the enemy threw hand grenades from the reef and fired rifles at the leading boats. Immediately all boats opened fire with all small arms and annihilated these 50-60 enemy on this section of the reef.⁵²

In addition to its reef-sweeping details, the 6th Marines also cleared the pocket east of Tanapag Village during the day. One hundred Japanese bodies were counted there.

Following the precedent of their fallen comrades, Japanese emerged during the night from their hiding places and attempted to pierce the strong U. S. defense. On the night of 8-9 July the 165th Infantry killed about 75 in front of its positions. Similarly, the next night 150 more fell from such abortive incursions. The 6th Marines reported killing "50 or more" on each night as the enemy attempted to sneak through the lines. Just where they were going, or what they were trying to accomplish, is not clear. These efforts amounted only to a gesture, or a final gasp, from a dying Japanese force.⁵³

And so, the Tanapag-Makunsha plain area was again in U. S. hands. As indicated throughout the narrative of this phase of the Saipan fight, the Japanese had suffered an almost unbelievable massacre. And yet not unbelievable, for densely sprawled through the area for all to see was the hideous evidence. By the 27th Division's count, 4,311 enemy dead lay in the battle area.

To determine accurately just which Japanese were killed by the various participating units in this battle is impossible. There are certain clues, however. In the sectors of the two battalions of the 105th Infantry which bore the brunt of the initial attack, 2,295 Japanese bodies were counted. Another 2,016 lay to their rear. How many of these fell prior to the climactic banzai attack as a result of artillery, naval and air bombardment which had been almost continuous since 13 June? And how many, on the other hand, were killed during the savage attack and the mop-up that followed?

The questions started very soon after the battle. On 12 July the G-1 of the 27th Division submitted a letter to NTLF on the subject of burial of Japanese dead. General Griner added an endorsement to the letter which said in part:

In view of press releases and official dispatches indicating the numerical strength of the enemy counterattack on 7 July as between 300 and 500, basic report is forwarded indicating that by actual count, 4,311 enemy soldiers were buried in the area between the farthest advance of the 1st and 2d Battalions, 105th Infantry, and the farthest advance of the enemy. In the same area, 406 bodies of American soldiers were found.

A further breakdown of enemy dead indicates that 2,295 were killed in the combat area of the 1st and 2d Battalions, 105th Infantry, and 2,016 in the combat area of the 1st and 2d Battalions, 106th Infantry, the 3d Battalion, 6th Marines, and the 3d Battalion, 105th Infantry. These dead represented the number of enemy that pushed beyond the defensive line of the 1st and 2d Battalions, 105th Infantry.

Basic report is forwarded in the hope that by proper distribution of the facts, accuracy of historical records and press releases will be preserved.⁵⁴

General Schmidt, who had recently relieved Holland Smith as Commanding General, NTLF,⁵⁵ answered General Griner's letter on 15 July. Schmidt did not question the number of

dead Japanese found and buried but seriously doubted that these had all fallen as a result of the banzai attack. Also, Schmidt's letter called attention to the interrogation of Major Yoshida, who had estimated that only 1,500 Japanese had participated in the attack.

The areas over which the enemy made its counterattack [continued General Schmidt's letter] against the 105th Infantry on 7 July have, from time to time, been subjected to naval gunfire, bombing, strafing attacks and artillery fires from 13 June to 7 July, and it is evident that these fires produced some casualties. It is probable that some of the dead remained on the ground as a result of the initial occupation by the 105th Infantry, and subsequent advance of the 106th Infantry and elements of the 2d Marine Division in the same zone of action.

Staff officers and observers, who visited the scene of attack soon after the ground was recaptured by the 2d Marine Division, have reported very large numbers of enemy dead within the zone and have also stated that many of the dead, due to the advanced state of decomposition, had apparently been in the area for some time. Some photographs turned in to this headquarters bear out these reports.

More than 300 freshly killed Japanese were counted on 8 July, in front of and in the positions of the 3d Battalion, 10th Marines (Artillery) which were overrun. It is probable that these Japanese constituted a large part of the enemy force which penetrated the positions of the 105th Infantry.

General Griner was not long in answering. His letter of 16 July said in part:

It was not suggested that each enemy soldier was killed by infantry elements of this division, or by its infantry and artillery, although the 27th Division Artillery was the only artillery in direct support of the action. Our Naval gunfire, aerial bombing, and Corps Artillery did good work everywhere on the island and at all times. Any estimate of the casualties inflicted by them on the enemy is acceptable, provided it is applied uniformly in all Division Zones of Action.

On the morning of 8 July, over 24 hours after the NTLF liaison officer with the 27th Division, Major Regan Fuller, had estimated the enemy counterattack force at 1,500, NTLF had sent a dispatch to Admiral Turner which stated that the 1st and 2d Battalions, 105th Infantry, had been "overrun by a considerable force estimated to be at least 300 to 400 Japanese, supported by two tanks."⁵⁶ It was to this inaccurate dispatch that Griner next referred and demanded correction, saying in part:

The great mass of the enemy dead found in the area between the extreme advance of the 105th Infantry and that of the 106th Infantry (2295) obviously were killed by the 105th Infantry since the enemy counterattack had subsided completely prior to the advance of the 2d Marine Division at noon, 8 July.

In view of the fact that 2,016 enemy dead lay in the area in rear of the positions of the 1st and 2d Battalions, 106th Infantry, the statement . . . that it is probable that the 300 Japanese dead found in front of the 3d Battalion, 10th Marines (Artillery) "constituted a large part of the enemy force which penetrated the positions of the 105th Infantry" is not a reasonable conclusion. No mention is made of the fact that this position was relieved by the 2d Battalion, 106th Infantry, which reasonably could have been expected to have inflicted casualties in doing so.

As the only General Officer who was present on this front at any time prior to noon, 8 July; as the Commander of all troops operating in that zone, except a Battalion of Marine Field Artillery which had entered the area without the knowledge of this headquarters and had emplaced certain field guns 520 yards in rear of the most advanced line, not however for the purpose of supporting the troops in this area; and as one who followed in detail each event as it occurred, I submit that my personal observations should receive greater credence than is indicated by the record. I viewed perhaps a thousand enemy dead and did not see more than two or three in an advanced state of decomposition. Paragraph 127 of the Army Medical Field Manual, "Field Sanitation" (FM 8-40), states that at a temperature of 85ƒ F maggots will be formed in about eight hours. The staff officers and observers did not arrive on the scene until at least thirty-one (0500, 7 July, to 1200, 8 July) hours had elapsed after the launching of the enemy counterattack.⁵⁷

Realizing that a controversy was developing at Saipan, Admiral Nimitz, Commander in Chief, Pacific, sent a dispatch to Admiral Spruance, Commander Fifth Fleet, on 9 July ordering an investigation of the banzai attack. For this task, Spruance appointed two members of his staff, Colonel James E. Jones, USMC, and Lieutenant Colonel George S. Eckhardt, USA. These officers developed the facts of their report by conversations with General Holland M. Smith, General Griner, members of the NTLF and 27th Division staffs, and "such representatives of the troop units involved as were available for explanation of the action on the ground."⁵⁸

On 19 July Spruance submitted a report to Admiral Nimitz containing his conclusions relative to the counterattack. Spruance's report failed to establish the exact number of enemy that had participated in the attack, saying: "Estimates of the enemy strength in the assault run from 1500 to 3000." Though this figure is not specific, it serves to correct the first report from NTLF of "300 to 400." The last paragraph of the report contains comments on the conduct of troops and commanders involved in the fight:

1. The gap between the 1st and 3d Battalions 105th Infantry could have been readily closed by swinging the left of the 3d Battalion forward. In any case, this gap should have been covered by local reserves and/or planned machine gun fires. There is no evidence of such measures being taken.

2. The 3d Battalion 105th Infantry although not as heavily engaged as the other two, [made] no effort to go to the relief of the 1st and 2d Battalions. The reserve company of the 3d Battalion was not committed except to protect the battalion command post which was under attack.

3. Although the position of the 27th Division was penetrated to a depth of 1000 to 1500 yards, this penetration was always firmly contained.

4. There is no question that our troops fought courageously in this action. The 1st and 2d Battalions and Headquarters Company, 105th Infantry and the 3d Battalion, 10th Marines deserve particular mention.

5. All evidence points to the fact that the front line units of the 27th Division were alert to the imminence of a Japanese attack at dawn on 7 July 1944.⁵⁹

Whatever conclusion the reader may draw from the facts presented, the statement in subparagraph (d) of Admiral Spruance's report should emerge in bold relief above the clouds and fog of controversy: "There is no question that our troops fought courageously in this action."

D-plus 22, 23 and 24 (7, 8 and 9 July)

To the Shore at Marpi Point

It will be recalled that on 6 July the 4th Marine Division had been ordered to take over the entire front at a point just beyond Makunsha. From there it would drive to Saipan's northeastern tip. Assumption of its broad zone demanded commitment of all three infantry regiments of the division, and on 7 July NTLF attached an additional regiment, the 2d Marines, to assist in delivering the last punches. Artillery support had been bolstered also; the 3d and 4th Battalions, 10th Marines, attached to the 4th Division, moved to advanced positions from which they could deliver fires forward of the 23d Marines. As already indicated, however, these two battalions suffered heavy personnel losses in the banzai attack and were in no condition to provide effective support. The 2d 155mm Howitzer Battalion, V Amphibious Corps, executed general support missions from its newly occupied positions about 500 yards inland from Garapan's south edge.

The Japanese banzai attack had an important bearing on the final drive of the 4th Division. Many of the enemy who would have contested the last few miles of Saipan had already expended themselves in the Tanapag onslaught. The several thousand who accelerated their appointment with death could have provided a considerable challenge in the rocky approaches to Mt. Marpi. This result was not apparent to the men of the 4th Division, since they were not in the attack's path but was true nonetheless. But this is not to say that the last days at Saipan would go unopposed.

On 6 July the center and right regiments (24th and 25th respectively) had made rapid progress, reaching the dominating ground just short of O-8. The 23d Marines, however, met rugged terrain in its expedition to the west coast. A precipitous cliff, carved with enemy positions, would have to be descended before the unit could begin its final, half-mile drive to the west coast. The newly-attached 2d Regiment was assigned a zone between the 23d and 24th. On its four-regiment front the division attacked at 0930.

The two right regiments moved at a fast clip on 7 July, encountering only minor trouble in the form of sporadic small-arms fire from small bands or patrols of the enemy. In addition, the 1st Battalion, 24th Marines, reported fire from a large-caliber gun positioned in a cave. This weapon was pushed out for one or two rounds then hastily retracted before the

TAKING COVER behind a medium tank, these Marines try to locate the source of the automatic fire cracking about them. Man in foreground carries an SCR 300 radio with short antenna. Oak planking on side of tank is to prevent Japanese magnetic mines from sticking there.

Marines could bring down counter-fires. Eventually, however, the enemy gunner selected the wrong moment to peek his gun from the hiding place, doing so just as a shell landed in the entrance. The cave gun and its elusive gunner were destroyed.

One difficulty during the day was presented by the great number of civilians pouring through the lines. Until these had approached close enough for recognition, it was impossible to distinguish them from enemy soldiers, and it was necessary to proceed with a great deal of caution.

By early afternoon of 7 July the two right regiments had pushed so far ahead of units on the left that General Schmidt ordered them to hold up and wait.⁶⁰

The going had been tougher on the left. Here the problem was to find suitable routes from the cliff top to the flat below. In a maneuver designed to exploit the 24th Marines' success, Colonel Stuart's 2d Regiment moved to its right through the the 24th Marines' zone and reoriented its direction of attack straight toward the west coast. By dark of 7 July, Stuart's Marines were disposed along the high ground overlooking the coastal plain (see Map 23).

THROUGH A BATTERED CANEFIELD on the west coast, men of the 2d Marines, accompanied by a medium tank from Company A, 2d Tank Battalion, push toward the northern end of the island. Hundreds of the enemy were killed during this phase of the operation.

For the 23d Marines the activities of 7 July were a continuation of those begun the previous day. This regiment had to cross the coastal plain and reach the beach just northeast of Makunsha, and no amount of detouring would allow it to execute this until the cliff line flanking Karaberra Pass was cleared of enemy forces. Some imaginative Marine named the high ground overlooking the pass "Prudential Hill,"⁶¹ and the title stuck. To attempt to traverse the plain without first clearing the cliffs which dominated it would have been an invitation for the Japanese to deliver fire into the regiment's rear.

This was the problem then: the Marines faced the west coast and the Japanese did also. The Marines needed the base of the cliff for a line of departure but this was dominated by the Japanese in the cliff. In addition, the base was mined and covered with plunging fire. The struggle continued throughout 7 July and most of the caves and positions were destroyed. The coming of darkness, however, indicated the desirability of returning to the high ground for the night with plans to get a fresh start on 8 July. From positions overlooking the coastal flats, the 23d Marines observed large straggler groups moving toward Marpi Point and brought long-range machine-gun and 75mm half-track fires to bear on them. According to the 23d Marines' action report, "over 500 were killed."⁶²

Fighting in and around Karaberra Pass was close-in, personal and, frequently, individual Marine against individual Japanese. Because of this, men of the 3d Battalion, 23d Marines, were especially alert for anyone with oriental facial characteristics as they cleaned the area.

ANTICIPATING A BLAST, Marines hug cover near island's northern end. At times like these, one may only hope and pray.

So it was that one dark-complexioned person, dressed in a Marine uniform but showing a poor command of the English language, narrowly escaped death before he was finally permitted to surrender. After being marched to the 3d Battalion command post, he explained to Major Treitel, commanding the unit, that he was a Marine, of Puerto Rican ancestry, who had the misfortune of resembling the Japanese. It was the third time he had been taken prisoner by fellow Marines at Saipan. Major Treitel described him as "one frightened hombre," and well he might have been.⁶³

As the division stopped for the night of 7-8 July the lines took on a peculiar appearance. The left had not progressed, while the right had moved from 700 to 1500 yards. Connecting the two extremes was the 2d Marines, facing the west coast. The night passed quietly.

For the attack on 8 July, General Schmidt decided to hold the right of his line stationary, while the left moved to the sea. This, in effect, was a reorientation of the direction of attack, since units would be moving to the northwest rather than the northeast. If successful, however, this maneuver would greatly reduce the division frontage, which on the morning of 8 July was 6,300 yards. The scheme affected units as follows: the 25th Marines and the right battalion of the 24th (2d) would hold their positions facing to the northeast, while the remainder of the division swept to the northwest (see Map 23).

The plan was a good one, having as its most unique feature the support provided by rocket craft--LCI(G)'s--lying off Makunsha and firing into the cliffs directly below the 23d Marines. Though delayed somewhat by the terrain, occasional road blocks and mines, and desultory fire from caves, the advance to the coast was rapid. After placing the finishing, blasting alterations on Prudential Hill's western landscape, the 23d Marines dashed across

the coastal flat and reached the beach at 1410.⁶⁴

The 2d and 24th Marines, meanwhile, kept generally abreast and secured their assigned zones by 1530. A great number of Japanese were killed during this move and in the subsequent mop-up along the coast. Many of those encountered were no doubt survivors of the banzai attack which had struck the 27th Division. In three operations (Guadalcanal, Tarawa, Saipan), men of the 2d Marines could not recall a single day in which they had killed as many Japanese. It is probable that the 1st Battalion, 2d Marines, with its score of about 800 enemy killed on 8 July, accounted for more Japanese on that day than it had throughout the rest of the operation.⁶⁵ One company commander observed that "hunting was exceptionally good."⁶⁶

An unusual employment of 37mm guns during the day is worthy of note. In order that the 1st Battalion, 24th Marines, might use all available men in the push across the coastal plain, Major Roger G. B. Broome (Weapons Company, 24th Marines) volunteered to assume, with two 37mm guns and a few riflemen, a position from which to protect the right flank as the unit swept to the coast. After the

BELOW MT. MARPI'S AUSTERE HEIGHT (833 feet), 4th Division's tanks and infantry push across flat toward the final objective on 9 July. Many positions on Marpi's sheer northern face were by-passed by the speedy moves, and Marines had to return later for the mop-up.

UNCOMPLETED MARPI AIRSTRIP was overrun by 4th Division's 9 July dash to Saipan's northern coast. Marines appear alert after 25 days of combat.

infantry had departed, Broome's isolated position was rushed by a numerically superior group of Japanese. During the skirmish, the 37mm crews fired their pieces at ranges of 10 to 20 yards, taking up the brief slack between rounds by throwing grenades and firing small arms. For a time the issue was in doubt, but the Marines held. This exceptional employment of a weapons unit was necessary and effective in this situation.⁶⁷

The 2d and 24th Marines, after reaching their objectives, spent the remainder of the day mopping up the jagged coast line.

Earlier in the afternoon, when it was apparent that the attack to the northwest would be successful, General Schmidt approved the 25th Regiment's request to advance its lines to the dominating terrain 600 yards to its front. Patrols had located no enemy in that area, and Colonel Batchelder deemed it advisable to capitalize on the situation. The advance had the additional advantage of providing observation over the Marpi Point area.

SAIPAN WON, the landing force command and unit representatives attend official flag raising at Charan Kanoa.

During the day the division's left had moved 1,500 yards in seizing the western coastal area and had advanced the northern portion of its line approximately 600 yards closer to the island's tip. At one point in the proceedings (shortly after noon), General Schmidt had entertained hopes of reaching Mt. Marpi during the day and issued an operation order to that effect. Time consumed mopping up beach areas, however, caused the general to defer the plan until the next day.

Night activity was confined to the coastal areas, where groups of civilians and some military personnel came from the caves to surrender.⁶⁸ Surrender was not on the minds of the screaming Japanese that attacked the 1st Battalion, 2d Marines, however. This battalion, set up on the lip of the coastal cliff, received a series of miniature banzai attacks that lasted most of the night. The point of attack was a logical one: at the right center of the battalion's front, where a 30-40 yard break existed. This break tilted sharply to the water's edge, affording the Japanese a corridor for the climb to the top of the cliff. Whatever the cause of this unusual formation, it had been apparent to the men of the 1st Battalion that this area had dangerous potentialities. In view of this, machine guns had been emplaced on either shoulder with sectors of fire interlocking over the mouth of the chute.

Into these prearranged fires the Japanese hurled themselves in attack after attack and, despite the repulse of all attempts, succeeded

ELECTING TO LIVE, these Saipan civilians cast their lot with the Americans. This puzzled, tired group eyes apprehensively one of the United States invaders to whom they surrendered.

--by the sheer momentum of their thrust--in knocking out both machine guns. These were replaced, however, and the slaughter continued. Later, one of the replacements was also destroyed. Vast quantities of incendiary grenades were used during the night to light the funnel and expose the enemy, but mortars or artillery could not be used because of the proximity of the Marines to their attackers. Machine guns, rifles and grenades were enough, however; the enemy soldiers were stopped before they could really get started. As had been the case in other such attempts, the Japanese seemed to feel an obligation to get killed in an attack, as if some higher honor would thus be accorded them. The coming of daylight revealed enemy dead thickly stacked in the narrow confines of the corridor. No accurate count of the bodies was completed, but the battalion commander estimated about 100.⁶⁹

Objective O-9 (extreme northeastern tip of Saipan) had seemed very distant on 15 June 1944 when the 2d and 4th Marine Divisions hit the Charan Kanoa beaches. It seemed to get no closer as the days grew to weeks, and the number of casualties climbed from one digit to five,

and the young matured to grim old age. As long as a single wooded hummock separated the front lines from the farthest tip of land, objective O-9 had been an ambition--little more.

Ambition became reality on 9 July when the 4th Division spurted to the shores of Marpi Point. The final thrust was made with three regiments abreast, from right to left, 25th, 24th, and 2d. Men of the 23d Marines, meanwhile, spent the day mopping up along the northwest coast, assisted by the LVT(A)'s of the 2d Armored Amphibian Battalion which fired on cave positions from seaward.

The attack of 9 July produced nothing startling in the way of tactical situations or enemy reactions. Firm pressure was applied all along the front and a fast pace maintained. Indeed, many units, anxious to be the first to reach Marpi Point, moved faster than reasonable caution would dictate. Many enemy groups were by-passed with the result that much time and effort were spent extricating holed-up Japanese after 9 July.

Along the west coast in the 2d Marines' zone, isolated enemy units somewhat retarded the push, but generally the Japanese realized the futility of coming out to do battle and instead chose to remain in the myriad caves along the coast. Extermination of all of these was not to be accomplished on 9 July--or 9 August, or 9 September, for that matter.

By 1615 all three regiments had reported that they were in possession of O-9.⁷⁰ And as the Marines looked to the north, all that they could see was water. There was no objective O-10.

Good news travels fast. In the same minute that the 4th Marine Division reported seizure of O-9, Admiral Turner, the Expeditionary Force Commander, declared the island secured. The time was 1615, 9 July 1944.

This was merely a formal declaration, however. It meant that all organized resistance had ceased and that the entire island--technically at least--was in our hands. Or, as one Marine put it: "It means that if you get shot now, you were hit in your own rear areas." For the men of the 2d Battalion, 24th Marines, word that the island was "secured" came as something of a surprise. In the words of the battalion's action report:

At 1615 a message was heard from an unknown station by tank liaison officer over tank TCS radio, that the island was secured. Part of F Company was pinned down on the beach for four hours by intense enemy machine gun and rifle fire from the western cliff face which resulted in one dead and ten wounded. This company was not able to pull completely back on to the high ground until darkness set in. It took five hours to evacuate the wounded up the treacherous cliff.⁷¹

For the Japanese, hundreds of whom still lurked in the many caves and recesses, "secured" had no meaning. Mopping up of these die hards continued for months thereafter. More than a year later, even after the final surrender of Japan, there were still Japanese soldiers at large on Saipan.

Naval Activity

During the closing days of the operation, the Navy bent a good share of its efforts toward the improvement of facilities at Saipan. Tanapag Harbor would figure prominently in the post-battle value of Saipan, and the task of clearing mines, wrecked barges and ships was already underway. By 6 July the Minesweeper and Hydrographic Survey Group reported that a 150-foot channel had been swept from Tanapag Harbor to the northwest docks. This set the signal for initiation of work on the piers and seaplane ramps in the Tanapag-Garapan area.

There was no diminution in the pressure exerted against other islands in the Marianas. Tinian, Guam, and Rota were peppered daily with bombs in preparation for contemplated landings. On 7 July Admiral Spruance announced the target dates for two of them Guam--21 July, Tinian--24 July.

Japanese air activity in the final days of the campaign was weak and ineffective. During the night of 6-7 July, while Saito's loyal subordinates prepared for the final banzai, 12 to 15 planes made two raids against Saipan and ships near the island. Although the Condition Red lasted from just after dark until after midnight, the raiders failed to achieve any hits. In the process, U. S. night fighters

from the fast carrier groups downed two of the intruders and the destroyer Hudson accounted for a third. Again, some Japanese planes attempted to land at Marpi Point Airfield and at North Field on Tinian but, apparently, were frustrated by ships' gunfire.

The final air attack executed against Saipan during the capture and occupation phase of the operation came at 1945 on 7 July and lasted until after midnight. In all, the Japanese launched nine separate raids during the evening but, aside from the dust, noise and excitement kicked up by the 25 bombs dropped on Aslito Airfield, achieved nothing. Night fighters rising from Saipan and from the fast carriers divided the honors, each shooting down one enemy plane.

Much more devastating attacks were launched against the airfields at Saipan in the months to follow, but the feeble display of 7 July marked the last one prior to "securing" the island on 9 July.⁷²

As noted at various points throughout this narrative, the Japanese ground troops were not happy with the poor showing of their aircraft. Another indication of this appears in the diary of an unidentified noncommissioned officer stationed on Tinian: "6 July-Did Vice Admiral Kakuda (Commander in Chief 1st Air Fleet) when he heard that the enemy had entered our area go to sleep with joy?"⁷³

ELECTING TO DIE, hundreds of Saipan civilians refused surrender offers, flung children and selves upon the jagged coastal rocks.

TANAPAG HARBOR'S MANIAGASSA ISLAND, a nuisance and threat to American shipping, was seized on 13 July by the 3d Battalion, 6th Marines. Ships in this pre-D-Day picture are Japanese. Mutcho Point is in right background.

The Crowning Horror

Following the declaration that the island was secured, efforts to induce cave-occupants to surrender were intensified. Interpreters, using public address systems, pleaded with people in caves to come out. The device was not only attempted from land but from sea as well. LCI gunboats moved close inshore and broadcast promises of good treatment, for which they were answered with fire from Japanese soldiers in the caves. Even some of Saipan's leading citizens, who had surrendered and received good treatment, talked to those in the caves, urging them to yield. But, for the effort expended, the results were not encouraging. The primary reason for this failure was that the people had been saturated with Japanese propaganda to the effect that the Americans intended to torture and kill them. This had been repeated so often that the people came to believe it.⁷⁴

At this time the very zenith of horror occurred. Hundred of civilians, believing that the end had come, embarked on a ghastly exhibition of self-destruction. Casting their children ahead of them, or embracing them in death, parents flung themselves from the cliffs onto the jagged rocks below. Some waded into the surf to drown or employed other gruesome means of destroying themselves. How many civilians died in this orgy of mass hysteria is not known. A commander of a patrol craft (YP) said that progress of his boat around Marpi Point at this time was slow and tedious because of the hundreds of corpses floating in the water.

Maniagassa Island

One small objective in the Battle for Saipan remained. This was tiny (250 yards wide, 300 yards long) Maniagassa Island in Tanapag Harbor. Lying like a possessive apostrophe some 2,500 yards northwest of the seaplane base, the isle was known to be occupied. It was conceivable that Japanese located there could harass activities within the harbor as well as provide Japanese on other islands with information relative to U. S. ship and plane movements. Even a small radio set could relay messages to other islands in the Marianas for further transmission to Japan.

The special assignment of capturing Maniagassa went to the 3d Battalion, 6th Marines, commanded once again by Lieutenant Colonel John W. Easley.⁷⁵ Preparations for this miniature operation were as complete as those for a full-scale campaign. Air observers from the 2d Marine Division conducted the reconnaissance, took the necessary aerial photos, and directed supporting artillery fires. Naval gunfire support was furnished by one LCI(G) firing its 40mm guns. The 5th Amphibian Tractor Battalion provided 25 LVT's to transport the battalion, while five LVT(A)'s from the 2d Armored Amphibian Battalion rendered close-in support.

With detailed preparations completed, the landing was executed at 1100, 13 July 1944. In the 15 minutes preceding this, the 10th Marines rained 900 105mm howitzer and 720 75mm shells upon the island. Nearly the entire objective was within the bursting radius of each shell, and the target was well saturated. Marines of Companies I and K hit the beach unopposed and hastily set about completion of their mission. Within an hour Maniagassa was overrun. Of the 29 Japanese defenders, 15 were captured and the remainder killed. The Marine casualty report was singular: one man wounded in action. One rifle platoon and a 60mm mortar section were left to garrison the new possession, and the remainder of the attacking force returned to Saipan.⁷⁶

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

Footnotes

1. War Department Historical Division Booklet, "Small Unit Actions, The Fight on Tanapag Plain," 66-73, hereinafter cited as The Fight on Tanapag Plain. 105th Inf Report, 9.

2. Originally scheduled for noon, the attack was delayed until 1300 by NTLF order.

3. 165th Inf Report, 8.

4. 27th Div Field Message 6, 1800, 5 July.

5. Contradictions are found in the 4th Marine Division report which gives the time as 1155 and in the 25th Marines report which complained that it was forced to fight 400 yards before reaching the designated line of departure.

6. 23d Mar Report, 47. L. R. Jones.

7. 24th Mar Report, 23. 1st Bn, 24th Mar Report, 11-12. 2d Bn, 24th Mar Report, 10-11. 3d Bn, 24th Mar Report, 10.

8. These were allowed to rest for a day or two and returned to their units.

9. 25th Mar Report, 8. 1st Bn, 25th Mar Report, 19-20. 3d Bn, 25th Mar Report, 12.

10. 4th Mar Div Report, Section VI, 33. This unusual density of fire was available for two reasons: first, only two divisions were now in the assault on a narrowed front; second, few counterbattery missions remained at this late stage in the operation.

11. NTLF G-3 Journal, 1800 4 July to 1800 5 July.

12. Although never proved, it was believed that this plane had succeeded in landing on the field (Banederu) on Marpi Point, had loaded whoever or whatever it had come for, and had taken off again. It was at this stage of the proceedings that the U. S. night fighter intercepted the plane and sealed the mystery.

13. The same officer who later in the day furnished such vital assistance at the ditch.

14. The normal practice of sending infantrymen along with the tanks as protection had not proved feasible in the gulch. Once inside the draw, riflemen made easy targets for the well-hidden Japanese.

15. The Fight On Tanapag Plain, 89.

16. This unit had relieved the 3d Battalion, 165th, early on the morning of 6 July, the latter reverting to regimental reserve.

17. 27th Inf Div G-3 Journal, 1845, 6 July.

18. 27th Inf Div Field Message 7, 2000, 6 July.

19. The Fight on Tanapag Plain, 86-113. 106th Inf Report, 9-10. 165th Inf Report, 8.

20. 4th Mar Div Report, Section VI, 33-34. 23d Mar Report, 47. 2d Bn, 23d Mar Report, 6. L. R. Jones.

21. 24th Mar Report, 24. 1st Bn, 24th Mar Report, 12. 2d Bn, 24th Mar Report, 11. 3d Bn, 24th Mar Report, 10.

22. Although cited in this one specific instance of the 25th Marines' move along the coast, this same procedure prevailed in other units before and after this particular example.

23. 25th Mar Report, 9. 1st Bn, 25th Mar Report, 20-23. 3d Bn, 25th Mar Report, 13.

24. Major General Merritt A. Edson was once asked which was the toughest operation in the Pacific. His classic answer is apropos: "The one in which you got hit!"

25. Barringer. Joslin.

26. 3d Bn, 24th Mar Report, 11.

27. Quotations from Coral and Brass, 194. Another reference to this same conference may be found in On To Westward, Robert Sherrod, 131.

28. Ltr from MajGen G. W. Griner, Jr., to MajGen O. Ward, 17Dec49. The "Jap prisoner" to whom Griner refers was a leading seaman from the 55th Keibitai who stated that his unit had been ordered to make an all-out attack at 2000. NTLF G-2 Report, 57.

29. NTLF G-2 Report, 57-58. For a detailed Japanese account of the days leading up to this climactic episode, and the suicide itself, see Appendix IX, page 284.

30. The 27th's Battle For Saipan, 15.

31. The 3d and 4th Battalions, 10th Marines, now attached to the 4th Marine Division, had occupied this area during 6 July to provide supporting fires for the 23d Marines in their sweep to the northern tip of the island. These two artillery units had been formed into a groupment under the commanding officer of the 4th Battalion (Lieutenant Colonel Kenneth A. Jorgensen). After arrival in position areas the later officer had contacted the commanding officer of the 105th Infantry, informed him of the artillery positions and, at the same time, secured information as to the 105th Infantry's front lines. Telephone and radio communications were established between the Marine artillery and the 105th's command post, and a Marine liaison officer remained with the Army commander. Interview with Maj W. P. Oliver, Jr., 16Jan50.

32. The presence of Japanese tanks in this attack is of interest mainly because of the curiosity of these remaining operative after over three weeks of steady reverses. The vehicles were three in number; one an amphibious tank, one a medium and one a light.

33. The foregoing account of the activities of the 3d Battalion, 10th Marines, is derived from the following sources: Memorandum to the Commanding General, Expeditionary Troops from Brigadier General Graves B. Erskine, 19Jul44; Report of enemy counterattack into position of 3-10 on 7 July 1944, ltr from CO, 10th Marines to CG, 2d Mar Div with Enclosures (A), (B), (C), (D), and (E).

34. Interview with Maj W. P. Oliver, 16Jan50.

35. Major Regan Fuller, NTLF Liaison Officer with the 27th Division, reported that "there was nothing going on but a little sniping by 1100 in vicinity of the C. P." Ltr from Maj R. Fuller to CMC, 12Jan50.

36. Among the dead was the 1st Battalion commander, Lieutenant Colonel William J. O'Brien, who posthumously was awarded the Medal of Honor for his actions during the banzai attack.

37. 27th Div Interrogation of Mamoru Iwaya (Korean) 2d class worker (civilian), 7Jul44. NTLF Interrogation of Major Kiyoshi Yoshida, intelligence officer of 43d Division Headquarters, 11Jul44. The 27th's Battle For Saipan, 15-16. 165th Inf Report, 8-9. 105th Inf Report, 10.

38. The one battalion referred to is the 3d battalion which was attached to the 165th Infantry, in position on the rim of Harakiri Gulch.

39. Though the 27th Division C. P. was not informed of the presence of Marine artillery, the 105th Infantry C. P. was. See footnote 31 on page 224.

40. This was the 2d Marines, attached to the 4th Division at 0920, 6 July.

41. Major Regan Fuller, USMC.

42. The battalion here referred to is the 3d Battalion which was still on the high ground overlooking the mouth of Harakiri Gulch.

43. Source of this quote is Major General George W. Griner's Certificate of 12 July 1944, para. 1.

44. Love, 495-501.

45. Detachments included: 2d Marines and Company A, 2d Tank Battalion, which were attached to the 4th Marine Division and the 1st Battalion, 29th Marines, which was attached to the Saipan Garrison Force.

46. NTLF Operation Order 25-44.

47. 6th Mar Report, 15-16.

48. On to Westward, Robert Sherrod, 140.

49. 8th Mar Report, 9. 2d Mar Div Report, Section VI, 25.

50. 165th Inf Report, 8-9. 2d Mar Div Report, Section VI, 26. Love, 475, 498, 514-518.

51. "Boat" is here used to refer to amphibian tractors.

52. Statement of 1stLt K. J. Hensley, Appendix K to NTLF G-2 Report.

53. 2d Mar Div Report, Section VI, 26. 8th Mar Report, 13. 165 Inf Report, 9.

54. CG, 27th Inf Div, 1st Endorsement to ACofS, G-1 27th Div, ltr of 12Jul44.

55. Schmidt turned over the 4th Marine Division to Major General Clifton B. Cates (later nineteenth Commandant of the Marine Corps) on 12 July 1944.

56. NTLF G-3 Journal, 1800, 6 July to 1800, 8 July.

57. Ltr from MajGen Griner to CG, NTLF, 16Jul44.

58. This investigation has often been erroneously referred to as the "Spruance Board."

59. "Report on Japanese Counterattack at Saipan on Morning of 7 July 1944" from Commander, Fifth Fleet to Commander in Chief, U. S. Pacific Fleet and Pacific Ocean Areas, 19Jul44.

60. 4th Div Report, Section VI, 34-35. 24 Mar Report, 24. 25th Mar Report, 9.

61. Though this name was derived from the Prudential Life Insurance Company's trademark, Webster's definition, "Prudential--That which relates to or demands the exercise of discretion," is also apt.

62. 23d Mar Report, 47-48.

63. Treitel.

64. Dillon.

65. Interview with Capt C. Schultz, 16Jan50.

66. Brooks.

67. Ltr from Col 0. Lessing to CMC, 11Jan50. Major Broome died from wounds received during this skirmish.

68. 4th Mar Div Report, Section VI, 35-36. 23d Mar Report, 48. 24th Mar Report, 24. 25th Mar Report, 9.

69. A count, initiated by Gunnery Sergeant Claude E. Moore, was abruptly terminated when the latter received four holes in his buttocks from a single bullet. While stooping to count the bodies, the sergeant presented a profile view of his posterior to a Japanese survivor in a nearby cave, who inflicted the unusual wound. The bullet hit no bones and the sergeant was not too uncomfortable to see the humor of the situation. Kyle.

70. 4th Mar Div Report, Section VI, 36-37.

71. 2d Bn, 24th Mar Report, 13. Rothwell.

72. TF 51 Report, Encl. A, 13-14, and Annex 1 to Encl A.

73. CINCPAC-CINCPOA Item #11,405.

74. Another reason often mentioned is that Japanese soldiers would not permit the civilians to surrender and killed those who weakened. This was not borne out, however, by an atrocity investigation which was conducted by five officers from the NTLF G-2 Section. Approximately 150 Japanese and Chamorro civilians were interrogated on this point, and none testified that they were threatened or used as shields by Japanese soldiers. There is a strong possibility, however, that those who would have testified differently are dead.

75. This officer had commanded the 3d Battalion, 6th Marines, during the D-Day landings, but had been wounded and evacuated by D-plus 1. He returned to duty in time for the closing stages of the operation.

76. 6th Mar Report, 17.