Late in February, as a result of continuing attrition and the destructive raid on Truk by U.S. carrier task forces,² the Japanese high command ordered all combat aircraft at Rabaul evacuated to that naval base. This, combined with certain other considerations, caused Eighth Area Army to revise its concept for the defense of New Britain.

With Cape Gloucester in their enemy's hands, all the rest of the island save for the northeastern tip ceased to have any value to the Japanese. Hence, to attempt defending the Talasea-Gasmata line, as stipulated in the original withdrawal orders, could constitute nothing more than a futile gesture. Accordingly, General Hitoshi Imamura issued new orders directing the further withdrawal of all elements on both coasts to the immediate Rabaul area³ where he still anticipated an early attack.

Chapter VIII left Colonel Sato with 65th Brigade headquarters at the inland village of Upmadung, carrying out his mission of transporting thither the stores accumulated at Iboki, including 60 tons of provisions Eighth Area Army had managed to send in by submarine.⁴ The 23 February departure of the doughty Matsuda,⁵ followed shortly by Katayama and Sumiya, left him once more low man on the totem pole when it came to covering the wholesale withdrawal.

The colonel got the provisions moved with time to spare. He left a small observation group at Iboki to report on enemy activities and an outpost of about platoon strength at Talawaga, near the mouth of the Aria River. The former withdrew before the Marines' approach on 24 February, and Sato

called in the latter a few days later when hostile patrols began probing in that direction. About the same time he received orders notifying him that the Komori Group would fall back on him from Arawe and directing him to take command of the combined force upon its arrival.

Immediately upon receipt of his withdrawal orders⁶ on 25 February, Major Komori dispatched his adjutant (by coincidence, a lieutenant named Sato) to Upmadung to make arrangements regarding provisions. Colonel Sato directed that five tons of supplies be concentrated as quickly as possible at a "three-fork road" south of a place called Augitni.⁷ He detailed troops to carry the material "with all possible speed," leaving their weapons and equipment under guard to facilitate the job.

Komori dispatched his force to the rendezvous point in successive echelons. He himself left Didmop with the rear guard on 28 February and reached Augitni via Vakan on 1 March. The following morning he started his hospital patients along the evacuation route and sent a detail of 250 men to Bulawatni to help transport provisions.⁸ To his delight, the first shipment to arrive contained a considerable quantity of saki, which Sato had picked up at Iboki and generously shared with his new associate.

This chanced to be the first time when all of the officers of the miscellaneous Merkus force had been together. The means being at hand, they decided to make it a festive occasion. Starting with toasts to the Emperor, the high command and big-hearted Colonel Sato, they went on improvising well into the evening. As Komori recorded in his diary: "It was fun." Evidently he acquired a gargantuan hangover, since he spent all next morning sleeping off the night before. He was still feeling queasy two days after the event, though attempting to explain his indisposition by writing: "Must have eaten too much last night."

But it was not in the cards that this period of leisure should last much longer. The Marines at Iboki became increasingly troublesome, though quite unaware that they faced anything more than starving stragglers. When one of their patrols entered Talawaga on 5 March, Colonel Sato felt the hot breath of pursuit on his neck and moved to the trail fork where he assumed command of the combined force. The next morning the Komori Group resumed its withdrawal in a northwesterly direction along the right prong of the trail fork, while Sato prepared to follow with his command on the 7th. At this point the ways of these two officers parted, and thenceforth each pursued his individual course toward the destiny that awaited him.

As noted in Chapter X, the 1st Battalion, 1st Marines had relieved the 5th Marines at Iboki for the Talasea operation. The patrol mentioned above consisted of Company A of that unit, commanded by Captain Harold F. Jennings, which had reached Talawaga in LCVP's by way of the Aria River, with the mission of proceeding overland to make contact with the Army force at Arawe. It soon found ample evidence that it had intercepted the main Japanese withdrawal route in the form of sick and dying stragglers. Yet the patrol killed several comparatively well fed soldiers with plenty of food and cigarettes in their packs, a discrepancy which rather puzzled Jennings. The patrol moved with caution, examining each abandoned bivouac area thoroughly and taking a number of prisoners. On 7 March Jennings passed through Bulawatni, within three to four miles of the trail along which Sato was proceeding in the opposite direction at the time, and he might have intercepted the enemy rear guard had he pressed

THE 1st BATTALION, 1st Marines, was based at Iboki for patrolling to the south and east.

on to the trail fork below Augitni. But the continuing southeasterly bearing of the trail he followed convinced him that it would take him far afield from his designated destination.⁹ So he halted the patrol a short distance below Bulawatni on the 8th and had his men cut a trail due west through the jungle to a north-south track that an Army patrol from Arawe had used recently. Then, too short on supplies to complete its mission, the patrol returned to Iboki, arriving on the 10th.¹⁰

On 6 March the Komori Group, with Major Tobuse's 1st Battalion, 141st Infantry in the lead, covered 16 kilometers on the trail running northeast from the fork toward the coast to reach what the Japanese called the "North Road." Along the way the commander was distressed to encounter many stragglers from the 53d Infantry, a sight which he described in his diary as "extremely pitiful." He could do little to help them, however, as it was problematical whether his own men could reach the next supply point on the amount of provisions they were able to carry, and which they refused to share with anyone else.¹¹ He still had an operational radio¹² at this time and learned from 17th Division of the Marine landing at Volupai.

After the first day on the trail, Komori's progress becomes difficult to follow, owing to his use of place names occurring in no other reports or maps. It was slow and difficult at best, because of rain and mud, and the necessity for wide detours to find fords through deep streams and safe passages through or around extensive swamps. His immediate force numbered about 200, and on 8 March he recorded the Tobuse Battalion a half-day's march ahead. He sighted occasional enemy reconnaissance planes, and on the 12th U.S. landing craft fired upon his men as they attempted to cross the Kuhu River, obliging them to take to the jungle and swamps. This finished the radio, and from then on he remained out of touch with practically everybody.

The broad Via River stopped his force on the 14th. When improvised rafts refused to float, the major and 15 others swam across. The rest, through weakness or timidity, declined to follow this excellent example, and it took two days and a wide detour inland to get the whole group to the right bank. Then, on the 17th, provisions ran out.¹³

The North Road led generally east-west at distances varying from two to three miles

Map 20
Japanese Withdrawal Routes
(Approximate)

inland from the coast.¹⁴ Thus the weakening men came by painful stages to Kometto (Eleanora Bay)¹⁵ and the welcome coconuts of Linga Linga Plantation on the 21st. But the wide Kapuluk River posed a formidable obstacle to further progress. "We are all weary," Komori recorded in his diary that night, "and Kou¹⁶ [the main supply dump] is still a long way off."

They tried first to swim the Kapuluk near its mouth, but 18 men of the 2d Company were swept away and 12 of them drowned, only the captain reaching the far bank safely. A patrol far upstream failed to find any fording point, so the troops spent the rest of the day building rafts. These managed to stay afloat, but broke apart so frequently that another two days were required for the passage. Then came the real heartbreaker on the 24th: they reached Kou only to find the provisioning post evacuated.

Again they staggered forward, living on coconuts and what food they could get from natives encountered. Since they possessed little to barter¹⁷ with, they usually ended by plundering the owners' scanty gardens--not the best sort of public relations for

COASTAL PATROL EXPLORES native village while pursuing Japanese eastward.

a force becoming increasingly in need of help. On 25 March an officer patrol encountered a U.S. patrol, and once more the group had to detour. On the 26th Komori hired a native to guide him to Numundo Plantation, at the eastern base of the Willaumez Peninsula. But evidently the man had a change of heart, for the major recorded next day that, lacking a guide, "we advanced using a compass."

They now entered a region of extensive swamps, intersected by another major obstacle: the two-pronged Kulu river. The column got across this stream on the 27th, but lost five more men in the swift current. The survivors emerged from the swamps on the 29th and followed the river southward.¹⁸ On the 31st Komori made his final diary entry, eloquent in its stark simplicity: "We are very tired and without food."

Apparently at this point malaria laid the major low.¹⁹ Ordering his men to continue toward Cape Hoskins, he took refuge in a native village, accompanied only by his executive officer, his orderly and a corporal attached to his headquarters. For lack of further diary entries, obscurity shrouds the last week of his life.

Contact at Kandoka

Colonel Sato resumed his march one day behind Komori (7 March), and his experiences generally parallelled the major's until he reached Linga Linga, on 23 March.

By rehabilitating stragglers during his stay at Upmadung, he had built up a conglomerate force of 400 to 600 men,²⁰ many of them in dubious condition, representing practically every unit that had fought in the Cape Gloucester area. He had learned of the Marines' Volupai landing from 17th Division, and two days out of Augitni he called all officers together to discuss the situation. Among these was 23-year-old Second Lieutenant Goremasamon Kawazu of the 30th Machine Cannon Company, who later described the event as follows:

He [Sato] pointed out quite honestly and frankly the seriousness of the situation. He drew a parallel between the predicament of Rabaul and the desperate situation in the Aleutians a year ago. He had no words of encouragement for his officers but expected them to carry out their duties in the traditional manner.²¹

On 19 March he learned from Komori that enemy patrols had infiltrated west of Willaumez Peninsula, and upon arriving at the Linga Linga coconut grove he was alarmed to discover evidences of recent bivouac by what appeared several hundred Americans. By this time sickness and hunger had whittled his own effective strength to about 250. As he was preparing to get these across the Kapuluk River the following morning (24 March) lookouts reported a number of U.S. landing craft entering Eleanora Bay (Kometto) and apparently heading toward the plantation.

Sato immediately prepared to repulse a landing and sent out patrols. The enemy craft sheered off, however, and later that day he sent what little equipment he had left across the river but remained with his men and weapons in defensive positions on the left bank.

He sighted more landing craft, together with reconnaissance planes on the 25th, but the Americans headed eastward without approaching closely. Encouraged, Sato put his force across the Kapuluk late that afternoon and bivouacked two kilometers beyond. Next morning he learned that the craft he had sighted had landed their troops near Kou, the former Japanese rationing point.

Probing cautiously in that direction, his advance party ran head-on into a Marine patrol.

This unit was the 3d Platoon of Company F, 1st Marines under Lieutenant W. C. Schleip. The Japanese had the good fortune to surprise their enemies in the act of crossing a stream; also to have with them one heavy and two light machine guns. With the fire of these and small arms they kept the Americans pinned down for approximately three hours, at which time approach of another Marine patrol caused them to break contact and withdraw.²²

To explain the presence of 1st Marines elements so far east requires a brief digression.

As noted previously, the 1st Battalion had been operating in the Iboki area at the time the Komori and Sato Groups pulled out. On 11 March this unit (less Company C) moved by water to Linga Linga and conducted vigorous patrolling for a week, leaving behind those traces of its recent presence which had alarmed Sato on the 23d. Companies E and F and a headquarters group of the 2d Battalion under Major C. H. Brush, Jr., then assumed responsibility for patrolling the Japanese withdrawal route, and were on their way to this mission when Sato's lookout sighted Company F's boats entering Eleanora Bay on the 24th.

By prearrangement Captain Petras, aide and pilot to the commanding general, reconnoitered ahead of the landing in one of the 1st Division's little cub planes--and spotted the Japanese at Linga Linga. This word he promptly signalled to the boats, then scouted farther until he found another suitable landing point. As an interested passenger, Petras had Brigadier General E. C. Long, once a pilot himself, who took over the controls while the captain drew a hasty

THE JAPANESE abandoned cases of ammunition as they retreated eastward.

sketch map of the area. This he dropped to one of the boats, then guided them in to the new beach, where the Marines came ashore some distance in Sato's rear.²³

In the meanwhile, Company E continued eastward and landed at the Kulu River mouth at the base of Willaumez Peninsula. There the Marines encountered only four enemy stragglers, killing three and taking one prisoner. Then they moved westward to Kandoka (Japanese Kou), where they were joined by two platoons of Company F, brought thither in the boats that Sato sighted on the 25th, and the combined force returned to the Linga Linga region. As a result of Lieutenant Schleip's fight the next day, however, Major Brush brought both reinforced companies back to Kandoka on the 27th. Failing to find any trace of the Japanese main body, he contented himself with setting up a trail block manned by the 3d Platoon and Weapons Platoon of Company E and returned to the command post, now set up on Yaluiai Island in Eleanora Bay.

Reason for the failure to make contact appears in the Sato Journal, which records the unit's position on that day as south of Kou and headed northeast. It describes the men as "physically and spiritually worn out" as a result of "yesterday's battle and the

MARINES ON PATROL take a break.

absence of rations." They gathered tree buds for food and "experienced terrible suffering." The presence of many wounded slowed progress further. At last, on the 29th, they reached the coast where they found some taro fields but had to work cautiously because of the known presence of Marines in the vicinity.

But time was running out on Colonel Jiro Sato, IJA. On 30 March a patrol from Kandoka spotted his rear echelon and reported counting 73 Japanese with machine guns and a radio.²⁴

This patrol consisted of eight men under Second Lieutenant Richard B. Watkins,²⁵ in command of the trail block. At 0900 that morning he had led his group inland from the village over some faint trails which he hoped the Japanese main body might be using, since diligent search of the main coastal route had failed to discover the force that Lieutenant Schleip had engaged. The Marines had proceeded about a mile and a half and were about to cross a stream in a sparsely wooded area when they sighted two Japanese standing with slung rifles, apparently resting.

Watkins had about decided to dispose of these when they moved off in a northeasterly direction, followed almost immediately by a large body of their compatriots. Lying where the Japanese could easily have spotted them, the Marines counted the 73 enemy soldiers and noted the equipment cited above. Conspicuous among them was a tall, burly officer being carried on a litter. The Japanese were cutting fresh trail through the jungle, and fortunately were too intent on their work to discover the patrol. They made excruciatingly slow progress, however, and Watkins did not dare to move until the last man had disappeared.

He then returned with all speed to Kandoka, sending a runner ahead with his report. At the village he met Major Brush who had come over from Yaluiai with one squad on what he had intended as merely a routine inspection. He promptly radioed his command post to send a rifle platoon and a 60mm mortar section to the scene and ordered Watkins to hold the trail block with one rifle squad, one machine-gun squad and two mortar squads while he himself set off with the rest of the troops available in an effort to overtake the Japanese from the rear.

Before his own departure that morning, Watkins had sent a six-man patrol under Sergeant Frank Chliek to an inland village some two miles south of Kandoka, right where the Japanese appeared to be heading. The lieutenant promptly dispatched a native messenger there with a warning; then, when firing broke out in that direction, he surmised, correctly, that Chliek had become engaged and took the remaining riflle squad and hurried to his assistance.

He arrived to find the sergeant and his men crouched on high ground by the east-west trail with dead Japanese all around them. The volume of fire was intense, the pattern not at all clear. As it turned out, Chliek had stumbled upon the column somewhere near its head and, taking advantage of his superior position, immediately opened

fire. Major Brush's force, which had been closing rapidly, was on the opposite side of the valley at this time; hence, at sound of the first shots, had hastened to the scene and struck the Japanese column's other flank. In order to avoid becoming involved in a fire fight with that group, Watkins ordered Chliek's patrol back to Kandoka and followed with his own squad at a slower pace. Brush caught up with him shortly, whereat a counting of heads showed that, miraculously, the Marines had not sustained a single casualty.

Here is the aftermath as Watkins recalls it:

On the following morning I took a 20 man patrol through the battle area. We counted 55 dead including 3 officers. It was quite easy to believe that perhaps 20 more died in scattered positions throughout the dense underbrush. The dead were all within a 100 yard stretch of trail. We encountered 2 more Japanese who had evidently just come upon the scene who were sitting side by side staring dazedly at the destruction and did not even turn their heads when we approached.

The burly officer previously observed on a litter proved to be Colonel Sato. Evidently Sergeant Chliek's patrol had struck the enemy column at precisely the point where he was being carried, for the colonel had only had time to leap to his feet and draw his sword to fight back before the Marines' fire riddled him.²⁶ Seeing their leader fall, those Japanese farther forward took off in mad flight, while those behind were cut off by the converging of the two patrols.

Thus perished one of the few Japanese to deserve much credit for the performance of the 65th Brigade on New Britain.

Although the rear echelon attacked by the two patrols comprised less than half of Sato's total force, the group, as such, ceased to exist with the death of the leader who alone had held it together. Units split into components, these into smaller parties, straggling eastward over a variety of trails, often hacking their own way by compass azimuth. Typical was the experience of young Lieutenant Kawazu. After escaping the Sato debacle, he rounded up 16 survivors of his unit and offered them the choice of following him or making their own way; only three chose to share his fate.²⁷

Now no Japanese capable of cohesive action remained west of Cape Hoskins--and precious few there.

Further Patrolling by 1st Marines

The 2d Battalion, 1st Marines remained in Linga Linga-Kandoka region ten days longer, patrolling vigorously to the east, west and south, but encountered only a diminishing trickle of stragglers. During the entire period of its operations here, 2/1 accounted for 183 Japanese killed and 65 taken prisoner, of whom 42 were turned in by natives.²⁸

How many Japanese reached the safety of Cape Hoskins remains problematical, but the number would appear to be far higher than the Marines suspected. During the early stages of the withdrawal, many were evacuated by boat from Karai-ai and Iboki, and, as noted in passing, Matsuda's headquarters and the remnants of 2/53 had escaped intact by water. The fact that the Marines failed to intercept any organized groups except Sato's rear guard, while numerous Japanese sources agree that such groups existed, would indicate that several of these got through before depletion of the coastal supply dumps brought about the most serious phase of attrition. In any event, a chart showing garrison units in the Rabaul area at the time of its surrender in

NATIVES ENTERTAIN with a "sing-sing."

1945 lists the 65th Brigade, with a strength of 2,700.³⁰

Early in March the 3d Battalion, 1st Marines had been earmarked for a special operation contemplated by division headquarters: seizure of Garove Island, location of the Japanese offshore barge staging base. At 0530 on the 7th of that month an amphibious patrol consisting of Lieutenant H. C. Duncan of Company K, Sub-Lieutenant W. G. Wiedeman, RANVR, and two natives landed from a PT boat under cover of darkness to reconnoiter beach conditions along the eastern shore. They found no Japanese installations in the immediate vicinity, but the beaches proved something less than ideal for an assault landing. Also they heard test firing from the southern part of the island that indicated the enemy had at least three artillery pieces and numerous 20mm machine cannon.³¹ When subsequent aerial reconnaissance drew unexpectedly heavy dual purpose 75mm fire,³² division concluded that the probable cost of seizure outweighed possible advantages to be gained and reluctantly cancelled the plan.

On 9 April the 3d Battalion relieved 2/1, which returned to Borgen Bay. But two days later, in compliance with unexpected orders 3/1 hurriedly rejoined the regiment,³³ for reasons that will soon appear. Thereafter, such wretched Japanese as remained

in the area between Iboki and Willaumez Peninsula were left to the mercy of nature and the natives.

Journey's End

At its pleasant outpost at San Remo Plantation the 2d Battalion, 5th Marines prepared to give passing attention to the brighter side of war. The date, 9 April: Easter Sunday (also pay day); the occasion: awarding of letters of commendation and meritorious promotions to personnel of the unit.

By now the harassing monsoon had given way to generally clear tropical weather. Although outposts and patrols still killed or rounded up a few Japanese stragglers, the duty was not too exacting, and everybody felt fine. In order to do the business up in style, Major Gayle had ordered a ceremonial parade and invited Colonel Smith and his regimental staff to do the honors. For several days the troops had been scrubbing their faded dungarees, brushing up on their parade ground manners, renewing their familiarity with close order drill. To top off the occasion, natives in nearby villages had agreed to entertain all hands with a "No. 1 sing sing" in the afternoon.³⁴

But there was still a campaign on, however unexciting it had become, and the battalion did not neglect routine security measures. Thus it happened that, at 0830 on the festive day, Company E's outpost came suddenly face to face with a group of four bedraggled Japanese. It was one of those abrupt encounters that allow little time for reflection or deliberation. The flurry of fire that followed killed three of the enemy and wounded the fourth, who promptly surrendered.

He gave his name as Corporal Isamu Kozuki, 1st Battalion, 81st Infantry, IJA.³⁵ And he identified the elder of two dead officers as his battalion and group commander.

Major Shinjiro Komori had reached his journey's end. And a long, rough road he had traveled, from his two Imperial commendations at Arawe to this dismal finish. Search of his pack revealed the diary from which his adventures have been reconstructed, and a rusty pistol--his only weapon.

Failure of his men to take Komori alive almost spoiled the day for Major Gayle. But the parade and "sing sing" came off as scheduled, and very successfully. Three days later Colonel Smith made brigadier general, and cheering word came from division headquarters regarding early relief.

For the 5th Marines, too, was nearing its journey's end, so far as concerned New Britain. With promotion of the regimental commander, Lieutenant Colonel Buse took over on an acting basis, and General Smith returned to Cape Gloucester to relieve General Shepherd as Assistant Division Commander. Units attached to the regiment followed, and representatives of the Army's 185th Infantry arrived to inspect positions preparatory to replacing the 5th, just as other 40th Infantry Division elements were doing in the Cape Gloucester area with a view toward relieving the entire 1st Marine Division.

Last clash of the campaign occurred on 22 April, when a 2/5 patrol intercepted a party of Japanese, killing 20, including two officers, and suffering the regiment's last fatality on New Britain. Three days later the whole 185th RCT moved into Talasea and San Remo, and at 1530 command of the area formally passed to the Army.

LCM's carrying the 1st and 3d Battalions back to Borgen Bay cleared Talasea at 1630, followed by those carrying 2/5 from San Remo at 2000. The movement was reported complete at 1930 on the 26th, and men of the 5th learned that the 1st Marines had departed this island of evil memory the previous day and that they would follow as ships became available.

Many Marines, cherishing fond memories, had hoped throughout the campaign that they would return to Australia when it ended. But it turned out they were headed for a very different sort of place.

Footnotes

1. Except as otherwise noted, detail concerning the movements of these two officers and their troops derives from the two diaries previously noted: ATIS 939, 1; and ATIS 999, 5.

2. 16Feb44 as a covering action for operations in the Marshalls.

3. Murayama gives 23Feb44 as date of issuance and summarizes orders in detail. Sato received 17th Div orders on 25 Feb. His own series of implementing orders appear in ATIS 918, 7, and ATIS 937, 3.

4. Murayama.

5. A map in Murayama shows brigade hq departing the Aria River and records its arrival at Cape Hoskins on 25Feb44.

6. ATIS 937, 1.

7. SA Det Op Order No 2, ATIS 937, 3. Augitni was sometimes transliterated "Agaribachite." The only certainty in regard to its location is that the U.S. map was widely in error. A miscellany of fragmentary evidence places it 4 to 6 miles south of village spotted as Bulawatni. Its location on accompanying maps is highly suppositional, based mainly on Japanese sketch map appearing in ATIS 733, 7.

8. Transport was handled by relays, each group of carriers dumping its load at a designated spot where another group picked it up. POW 917.

9. According to the most comprehensible captured Japanese map (ATIS 733, 7), the trail turned due south at Augitni and sharply southwest at Vakan, reaching Arawe via Didmop. On his withdrawal, Komori had covered Didmop-Augitni stretch in two days.

10. Material on this patrol derives from Capt Jennings' report of 14Mar44, contained in 1/1 Hist, 1721. On 11Mar the 3d Platoon of Co A left Iboki via the Army trail and reached Arawe on the 17th. Ibid.

11. POW 917.

12. The radio was broken down into four man loads for hand carry and had to be reassembled for use. POW 990.

13. ". . . thereafter lived on taro, natives' vegetables, birds and fish." POW 999.

14. POW 199.

15. It would appear that the Japanese used the name Kometto to apply both to the bay itself and some place located on its shore, probably Linga Linga.

16. Also appearing as Ko and Koho. Consensus of incidental evidence indicates that it was the place appearing on U.S. maps at Kandoka, a name occurring nowhere in Japanese records or interrogations.

17. "Stragglers were trading rifles and ammunition for taro from natives." POW 936.

18. This course was necessitated by establishment of a Marine road block at Garu where the main trail skirted to the north of the swamp area.

19. Exactly when Komori came down with malaria remains uncertain. POW 996 states that this occurred on 8 March, but nothing in the major's diary bears this out, so the logical assumption is that he became separated from his troops simultaneously with the diary's termination.

20. POW 1000; POW 936. Estimates vary widely, probably because the group moved in separate echelons.

21. POW 1000.

22. 2/1 Hist, 19. Journal of Sato Group states that their opponents "finally retreated toward the east." This action cost Schleip one killed and five wounded seriously enough to require evacuation to Borgen Bay. He estimated Japanese casualties as 20.

23. Petras.

24. 2/1 Hist, 20.

25. Details regarding these events derive from ltr Capt R. B. Watkins to CMC, undated, received in Apr52.

26. The manner of Col Sato's death is corroborated by POW 1002 who was among those toward the head of the column and thus escaped--for the time being.

27. POW 1000.

28. 2/1 Hist, 21. The natives became increasingly aggressive as enemy conditions worsened. Most prisoners they turned in were weakened stragglers who had taken refuge with them, but there were instances of natives attacking groups of comparatively strong Japanese with clubs and knives.

29. Murayama, who was at Cape Hoskins himself, states categorically that such was the case.

30. USSBS, 264. Gen Imamura set the number at 3,000. Ibid., 86.

31. Recon Patrols, G-2 Rpts. Duncan estmated 300 Japanese in occupation, but later prisoner interrogations indicated about 850.

32. Statement LtCol J. A. Pounds, 19Jan52.

33. 1st Mar Hist, 13.

34. 2/5 RofE.

35. POW 996.