Skip to main content

The Navy Department Library

Related Content
  • Operations
Document Type
  • Monograph-Research Report
  • Primary Source Document
Wars & Conflicts
  • World War II 1939-1945
File Formats
  • Image (gif, jpg, tiff)
Location of Archival Materials
  • NHHC-Library

Solomon Islands Campaign: XI

Kolombangara and Vella Lavella, 6 August - 7 October 1943 

Office of Naval Intelligence. United States Navy



PDF Version [53.9MB]



Solomon Islands Campaign: XI



Vella Lavella


6 August – 7 October 1943 



Office of Naval Intelligence, United States Navy



Office of Naval Intelligence

Washington, D. C.


1 October 1943.


    Combat Narratives are confidential publications issued under a directive of the Commander in Chief, U. S. Fleet, and Chief of Naval Operations, for the information of commissioned officers of the U. S. Navy only.

    Information printed herein should be guarded (a) in circulation and by custody measures for confidential publications as set forth in Articles 75 ½ and 76 of Navy Regulations and (b) in avoiding discussion of this material within the hearing of any but commissioned officers. Combat Narratives are not to be removed from the ship or station for which they are provided. Reproduction of this material in any form is not authorized except by specific approval of the Director of Naval Intelligence.

    Officers who have participated in the operations recounted herein are invited to forward to the Director of Naval Intelligence, via their commanding officers, accounts of personal experiences and observations which they esteem to have value for historical and instructional purposes. It is hoped that such contributions will increase the value of and render ever more authoritative such new editions of these publications as may be promulgated to the service in the future.

    When the copies provided have served their purpose, they may be destroyed by burning. However, reports acknowledging receipt or destruction of these publications need not be made.





Director of Naval Intelligence





8 January 1943. 

    Combat Narratives have been prepared by the Publication Branch of the Office of Naval Intelligence for the information of the officers of the United States Navy.

    The data on which these studies are based are those official documents which are suitable for a confidential publication. This material has been collated and presented in chronological order.

    In perusing these narratives, the reader should bear in mind that while they recount in considerable detail the engagements in which our forces participated, certain underlying aspects of these operations must be kept in a secret category until after the end of the war.

    It should be remembered also that the observations of men in battle are sometimes at variance. As a result, the reports of commanding officers may differ although they participated in the same action and shared a common purpose. In general, Combat Narratives represent a reasoned interpretation of these discrepancies. In those instances where views cannot be reconciled, extracts from the conflicting evidence are reprinted.

    Thus, an effort has been made to provide accurate and, within the above-mentioned limitations, complete narratives with charts covering raids, combats, joint operations, and battles in which our Fleets have engaged in the current war. It is hoped that these narratives will afford a clear view of what has occurred, and form a basis for a broader understanding which will result in ever more successful operations.






Commander in Chief, U. S. Fleet, and Chief of Naval Operations.



Charts and Illustrations

  Facing page                
Chart:  New Georgia Group.                                                                  
Illustration:   Vila Airfield 
Diagrams illustrating Battle of Vella Gulf:   
     DesDiv FIFTEEN opens fire (2346)  on page 
     DesDiv FiFTEEN opens fire (2356)  on page
Chart:  Battle of Vella Guf.   
Illustration:  B-17 over Gizo Island  12 
Chart:  Night Action 17-18 August.   
Illustration:  Barakoma Airfield  13 
Illustrations  between pages 26 and 27 
     Scenes from action 17-18 August.   
     Bow of Patterson after collison   
     Bow of McCalla.   
     O'Bannon after Battle of Vella Lavella.   
     Selfridge and O'Bannon.   
     Selfridge after Battle of Vella Lavella.   
     Starboard view of Selfridge  
     Selfridge returning from Battle of Vella Lavella.   
Chart:  Battle of Vella Lavella, 6-7 October.   
Illustration:  Wrecked Zero on Vila Airfield  60 
Chart:  Solomon Islands Area.   
Illstration:  Kolombangara after our landing  61 

Diagram:  Selfridge under torpedo attack
on page 63 



Battle of Vella Gulf, 6-7 August 1943 
     Preliminary Arrangements 
     Conclusions  11 
     Destruction of Barges, 9-10 August  12 
Landing on Vella Lavella, 15 August 1943  14 
     Strategic Considerations  14 
     Task Force Organization  15 
     Preliminary Operations  17 
     Approach and Landing  18 
     Antiaircraft Action  19 
     Operations Ashore  22 
Night Action off Vella Lavella, 17-18 August 1943  22 
     Preliminary Maneuvers  23 
     Engagements with Destroyers  24 
     Destruction of Barges  27 
     Conclusions  28 
Supplying Vella Lavella, 17 August-3 September 1943  28 
     Second Echelon  29 
     Third Echelon  31 
     Fourth and Fifth Echelons  32 
     Passage of Command  33 
Blockading Operations, 1 September-5 October 1943  34 
     Minelaying West of Kolombangara  35 
     PT Boat versus Barge  36 
     Air Combat and Patrol  38 
     Evacuation of Kolombangara  40 
Battle of Vella Lavella Iland, 6-7 October 1943  47 
     Military Situation  47 
     Preliminary Movements  48 
     Contact and Approach  50 
     Action with Group I  52 

Battle of Vella Lavella Iland, 6-7 October 1943--Continued.                                            
     Selfridge Engages Group II  55 
     Arrival of Supporting Group  57 
     Enemy Losses  59 
     End of a Campaign  60 
Notes on the Japanese Torpedo Attack in the Battle of Vella Lavella  61 
     I.  Chevalier  61 
     II.  Selfridge  62 
Appendix A:  Supply Echeleons to Vella Lavella  66 
Appendix B:  Symbols of U. S. Navy Ships  67 
Appendix C:  Designations of U. S. Naval Aircraft  69 


New Georgia Group

Vila Airfield and Blackett Strait

Kolombangara and Vella Lavella

6 August – 7 October 1943 


    The capture of the airfield at Munda Point on 5 August 19431 ended the first phase of our northward march through the New Georgia Group. Despite the fact that a total of 1,671 Japanese dead was counted and that heavy additional casualties were known to have been inflicted by Allied naval, artillery and air bombardments, some Japanese were able to withdraw to the north and effect a junction with other troops holding out at Bairoko Harbor, the last major center of Japanese resistance on New Georgia. A few others, probably high-ranking officers for the most part, were evacuated by barges to the enemy base at Vila-Stanmore on Kolombangara. This road of escape, however, was effectively denied to the majority of the survivors by our light surface vessels, which reported intercepting and sinking several troop-carrying barges in Kula Gulf during the final stages of the fighting at Munda.

    Meanwhile, as two columns of our ground forces pushed through the jungle in pursuit of the fleeing enemy, Army engineers and Navy Seabees began reconstruction of Munda airfield, which was found to be in reasonably good condition despite the intensive bombardments which preceded its capture. Allied use of the airstrip, it was felt, would effectively neutralize the field at Vila-Stanmore, besides bringing our fighters and bombers within much closer range of the enemy’s last three remaining air bases in the Solomons – Kahili, Ballale and Buka.

    Even though partially neutralized, the base at Vila-Stanmore remained a stumbling block in the path of our northward drive. Indications were that the Japanese had no intention of withdrawing from the area, despite the fact that its potential usefulness had greatly diminished. It was believed, on the contrary, that the enemy intended to augment his garrison there, and to this end was preparing to move in troops and equipment from the north in barges and destroyers under cover of darkness.


1 See Combat Narrative, “Operations in New Georgia Area, 29 June – 5 August 1943”.


Preliminary Arrangements

    The work of preventing these reinforcement operations fell to our destroyers and PT boats. For some weeks they had been systematically searching out enemy supply concentrations throughout the New Georgia Group. As our offensive against Munda gained momentum, they had contributed in no small measure to the success of the operation by breaking up enemy reinforcement attempts. Now the task would have to be continued in the waters surrounding Kolombangara.

    On 4 August, the day before Munda finally fell, Comdr. Frederick Moosbrugger, commanding a Task Group of six destroyers, reported to the headquarters of Rear Admiral Theodore S. Wilkinson on Guadalcanal and was informed that the Admiral wished to make a sweep of Vella Gulf with two destroyers and a number of motor torpedo boats, with the object of intercepting and disrupting enemy barge traffic. Conversations on this subject with MTB officers led to the calling of a conference to meet the following day aboard the destroyer Dunlap (Lt. Comdr. Clifton Iverson) at Purvis Bay. Attending this conference were Comdr. Moosbrugger as Commander of Task Group MIKE;2 Comdr. Rodger W. Simpson, ComDesDiv FIFTEEN in TG MIKE; Comdr. R. W. Calvert, ComMTB Flotilla ONE; Lt. Comdr. Henry Farrow, and other MTB officers. At this conference the MTB representatives presented “much valuable information” regarding barge traffic and destroyer sightings in the Vella Gulf area, and tentative plans for an operation were worked out. The operation did not materialize because of the need for the MTB’s elsewhere, but the information obtained at the conference proved invaluable in the action that was to follow.

    Comdr. Moosbrugger, as CTG MIKE, had the following vessels at his disposal:

    DesDiv TWELVE (less Gridley), Comdr. Moosbrugger.

              Dunlap (F), Lt. Comdr. Iverson.

             Craven, Lt. Comdr. Francis T. Williamson.

            Maury, Comdr. Gelzer L. Sims.


2 Numbers identifying task forces have been omitted from all Combat Narratives for reason of security. In place of these numbers will be found the Navy flag name for the first letter of the surname of the commanding officer of a task force.

  DesDiv FIFTEEN (less Wilson), Comdr. Simpson.

               Lang (F), Comdr. John L. Wilfong.

              Sterett, Lt. Comdr. Frank G. Gould.

             Stack, Lt. Comdr. Roy A. Newton.

    Late in the afternoon of 5 August, he received a dispatch from Admiral Wilkinson directing him, with his two divisions, to sortie from Tulagi at 1230 on 6 August, proceed to Vella Gulf by a route south of the Russells and Rendova so as to arrive at Gizo Strait at 2200, and make sweeps of the Gulf, avoiding minefields. If he made no enemy contacts by 0200 on 7 August, he was to return at maximum speed to Port Purvis, passing north of Kolombangara. Admiral Wilkinson later informed Comdr. Moosbrugger that he believed the Japanese intended to reinforce the Vila-Stanmore area during the night of 6 August, using destroyers and possibly a cruiser.

    The following morning Comdr. Moosbrugger invited Comdr. Simpson to a breakfast conference. The battle plan adopted was one previously conceived for a similar situation by Comdr. Anleigh A. Burke, who had been relieved by Comdr. Moosbrugger just prior to this operation. The following assumptions were made: (a) there was a remote possibility that enemy submarines might be encountered in Gizo Strait; (b) Gizo Strait was not minded; (c) enemy MTB’s might be operating in the Vella Gulf area; (d) enemy snoopers would be active; (e) enemy troop-carrying barges equipped with the equivalent of 40 mm guns might be encountered in the Gizo Strait area, close to the fringing reefs north and west of Gizo Island, in the Blackett Strait area, and near the western shore of Kolobangara Island; (f) enemy destroyers would approach either from the north through Vella Gulf, or through Wilson Strait and Gizo Strait (the latter seemed improbable, as our PT boats had thoroughly combed that area); (g) enemy cruisers might be present; (h) the enemy force might consist of two groups, well separated; (i) the enemy surface forces would be at a disadvantage, their decks loaded with troops; (j) the element of surprise was in our favor, and must be exploited; (k) in a night surface engagement under favorable conditions, the destroyer’s primary and most devastating weapon was the torpedo; and (l) that American gunfire was superior to that of the Japanese.

    Detailed plans for the entire operation were developed and transmitted to the commanding officers. The vessels involved were to pass through

Gizo Strait in column of division columns at 15 knots,3 entering Vella Gulf at moonset. On passing abeam of Liapari Island, DesDiv FIFTEEN was to form on bearing 1500 T. about 4,000 yards from DesDiv TWELVE and sweep at 15 knots on course 124 degrees T. within a mile or two of Gizo Reefs. Thence they were to head north, close under the west shore of Kolombangara, searching for barges. DesDiv FIFTEEN, equipped with 40 mm guns, was selected as the inshore division, and DesDiv TWELVE, with 44 torpedoes, was designated as the offshore division to engage any destroyers or heavier ships which might arrive earlier than expected. It was planned to destroy barges detected on this sweep only on condition that reports from the spotting Black Cats gave definite information that there were no destroyers or heavier ships in the area. If such information were not forthcoming, contacts on barges were to be passed up until the second trip around.

    If destroyers or larger vessels were encountered, Division TWELVE would close to fire torpedoes, retire to about 10,000 yards until the torpedoes hit, then open gunfire. Division FIFTEEN was to cover Division TWELVE while the latter approached the enemy for torpedo attack. Unless the enemy discovered and opened fire on Division TWELVE during its approach, Division FIFTEEN was to wait until the torpedoes hit before opening gunfire. It would also make a secondary torpedo attack if a favorable opportunity was presented.


    Task Group MIKE departed Purvis Bay, Florida Island, 6 August at about 1130. The departure one hour prior to the scheduled time was considered necessary because of the condition of the Maury’s engineering plant, which limited her speed to 27 knots. In the vicinity of Savo Island the task group assumed a special semi-circular antiaircraft formation with the Dunlap at the center and proceeded south of the Russells and Rendova Islands to Gizo Strait.

    At 1730 a relay contact report was received from Plane One of Flight 15 saying that a Japanese force had been sighted at latitude 040 50’ S., longitude 1540 40’ E., course 1900 T., speed 15 knots. It was estimated that this force would reach Vella Gulf about midnight if it proceeded by direct route at a speed of 21 knots.


3 The 15-knot speed was chosen because it made for efficient sound searches, produced a wake difficult for snoopers to detect, and permitted maintenance of the predetermined timetable.

    Six Black Cats were assigned to CTG MIKE for coverage and search, operating in two groups of three each. Bad weather and radar trouble were encountered, however, and communication was never established with either group.

    In accordance with the battle plan, the Task Group arrived at Point Option, latitude 080 03’ S., longitude 1560 41’ E. at 2200, slowed to 15 knots and began a careful search for enemy craft. Upon traversing Gizo Strait DesDiv FIFTEEN formed on bearing 1500 T., distant 4,000 yards from DesDiv TWELVE. Our forces were disposed in a line of division columns with ships at intervals of 500 yards. At 2228 course was changed by division column movement to 1240 T. to sweep the reefs fringing Gizo Island and the approach to Blackett Strait. That done, course was changed to 0000 T. at 2250 for a sweep up the west coast of Kolombangara Island.

    After the moon set at 2226 the night was extremely dark. The sky was completely overcast, with a ceiling at about 4,000 feet. Surface visibility varied from about 3,000 to 4,000 yards, depending upon the rain squalls that descended at frequent intervals. The wind was from the southeast, force 2. The sea was smooth, the Gulf being nearly landlocked. Our force was apparently not sighted by enemy planes throughout the approach.

The Engagement

2333           Dunlap makes radar contact.

2341-42      DesDiv TWELVE fires 24 torpedoes.

2344           CTG orders DesDiv TWELVE to execute “Turn 9.”

2346           Explosions seen on targets. DesDiv FIFTEEN changes course to 2300 T. and                                    opens fire.

2351           Enemy destroyer turns over and sinks.

2352          DesDiv TWELVE changes course to 1800 T. to join in gunfire. Whole target area flame.   Many explosions.

2355          CTG orders Dunlap to open fire on smaller ship. Other ships of division join in firing. DesDiv FIFTEEN changes course to 090 degrees T. and joins in firing.

0000         Target disappears.

0017         Enemy destroyer observed against flames of a large destroyer and sunk by gunfire from DesDiv FIFTEEN.

0027         Large enemy destroyer presumed sunk by torpedoes from DesDiv FIFTEEN.

At 2318 the flagship Dunlap reported what appeared to be a good radar contact bearing 0900 T., range 4,500 yards, and requested verification by

the other ships. Upon receiving no verification the contact was abandoned as false. At 2323 course was changed to 0300 T., speed to 25 knots to continue along the northwest coast of Kolombangara.

    Shortly after discarding the first contact as a phantom, the Dunlap at 2333 made a second radar contact, bearing 3590 T., range 23,900 yards, and requested verification from other vessels. Craven confirmed the contact reporting, “I have three targets; looks mightly nice to me”; to which the Dunlap replied over TBS, “We have four.” Reports from the Combat Intelligence Center indicated that the enemy ships were in column, apparently using a shallow zigzag on course 1650 T. and 180 0 T., at speed between 25 and 30 knots.

    At 2340, when the Dunlap’s Torpedo Officer announced a track angle of 2900 T., and the distance between the opposing forces was being closed at a rate of about 50 knots, CTG MIKE gave orders over TBS for DesDiv TWELVE to take course 3350 and prepare to fire torpedoes. DesDiv FIFTEEN turned by column movement to follow DesDiv TWELVE, then changed course to 2700 T. and then to 1900 T. The two division of TG MIKE now drew apart rapidly with Division TWELVE on a course that, if continued, would sweep the port flank of the enemy column while Division FIFTEEN was preparing to cut back on a southwesterly course to cross the bow of the enemy in excellent position to open gunfire.

    Between 2341 and 2342 DesDiv TWELVE fired 24 torpedoes at a range of 4,300 to 4,820 yards. Since surface visibility at this time was less than 4,000 yards, none of the enemy ships had been sighted at the time of torpedo attack. At 2345 Division TWELVE made a simultaneous turn to the right to maneuver clear of possible enemy torpedo fire and to take station for further action, leaving Division FIFTEEN to engage the enemy with gunfire during the maneuver.

    While results of the attack were awaited, the chief torpedoman’s mate of the flagship Dunlap reported to the bridge of his vessel that all torpedoes “appeared to run hot, straight and true.” At 2346 the first explosion was observed among the targets. There followed a series of violent explosions, variously reported as to number. The general consensus of opinion seems to have been that there were four hits on three vessels, occurring from left to right; these were followed by another series of violent explosions totaling between seven and ten.

    The torpedo attack created such havoc among the enemy forces that from the time its results were observed it was evident that the battle

Division 15 opens fire as torpedoes of Divison 12 hit targets

was won. Two ships were rocked by continuous explosions, and a third was enveloped in a mass of flame with successive explosions. The latter target was at first identified as a cruiser, though later information indicated that she was probably a large destroyer. Red flames and heavy black smoke proved unmistakably that she had been hit in her oil tanks. A large fire developed in the center of the explosions and spread over a considerable area.

    It seems probable that the enemy did not know of the presence of our force until he was hit, since there were no indications of any evasive action taken by his ships. It was suggested that the position of our ships, close to the shoreline of Kolombangara may have prevented detection by Japanese radar. The enemy force seemed to be thrown into confusion from the start of the engagement. Some enemy gunfire was reported, but it was of short duration and entirely ineffective.

    Immediately after the first torpedo explosion, Division FIFTEEN, now broad on the port bow of the enemy column, swung right on course 2300 T. to cut across its course, and opened fire with all guns. The Lang picked a target to the left of the flames by FD radar. An early hit illuminated the target, and the Sterett and Stack joined without signal in concentrating gunfire upon the same ship. About the third salvo, a yellow fire broke out amidships and spread rapidly. The target, now seen to be a destroyer, returned an ineffective fire, but under the combined pounding of three ships she rolled over and sank at 2352.

    In the meantime the Stack discovered what was believed to be a good target in the area of the large destroyer that was aflame, and the Commanding Officer, Lt. Comdr. Roy A. Newton, ordered the starboard torpedo battery fired. It was believed that at least one hit was obtained, but results were obscured by the fire.

    At 2352 Division TWELVE, having completed its evasive maneuver, changed course to 1800 T. by turn movement to join in the gunfire. At 2355 Comdr. Moosbrugger ordered his flagship to open fire upon a target southeast of the burning area. The battery of the Dunlap was immediately joined by those of the other two ships of Division TWELVE. At 2356, just as Division FIFTEEN was changing course to 090 degrees, gunflashes were observed coming from an enemy vessel, which was evidently firing on our division as it turned. This was the target upon which Division TWELVE had just opened fire from the opposite side. The Japanese destroyer was

now smothered by fire from both divisions and “literally torn to pieces.” She disappeared at 0000.

    At 0003 the commander of TG MIKE ordered Division TWELVE to change course to 310 T. and pass a few thousands yards to the north of the flaming area. His purpose was to take a position near the northwest entrance to Vella Gulf so as to be prepared to intercept any other enemy force approaching.

    Radar search indicated only one target remaining in the area, the bearing of which revealed it to be the large burning destroyer, which could be clearly seen at this time. Ships of both divisions took turn in keeping up and spreading the fires that raged topside. At 0010 there was a terrific explosion on this ship that mounted 600 to 700 feet in the air. Division FIFTEEN, which had been assigned “mopping up” duties, took course 0500 preparatory to finishing the burning ship with torpedoes. At 0017, when our ships were about ready to fire torpedoes, the procedure was interrupted by a rare and unexpected opportunity.

Division 12 opens fire adfter completing evasive maneuver 15 reverses course to join in firing

    An apparently undamaged destroyer, not indicated on the radar scope, moved slowly into silhouette against the flames from the burning ship. Division FIFTEEN immediately opened fire at a range of less than 5,000 yards. All three ships began hitting at once, their salvos rending the topside and starting fires. An early salvo from the Sterett, striking just aft of amidships, evidently hit the enemy vessel’s magazine. After a violent explosion the bow of the destroyer rose to an angle of sixty degrees, and she sank stern first in a few seconds.

    At 0020 CTG MIKE, being informed that Comdr. Simpson was going to sink the remaining ship with torpedoes, ordered Division TWELVE to change course to 0900 T. in order to clear range for Division FIFTEEN. Just at this point the Dunlap reported a torpedo to starboard and took an evasive turn. The same ship reported another torpedo at 0035 and a third at 0059. At none of these times, however, was a torpedo sighted from her bridge, and it was concluded that the reports were erroneous.

    Comdr. Rodger W. Simpson in the meantime turned to the task of finishing the remaining Japanese ship, directing each ship of his division to fire two torpedoes at the burning vessel. Three violent explosions resulted. The target was said to have “looked like a bed of red hot coals thrown a thousand feet in the air.” At 0027, after more explosions, she disappeared from sight as well as from all radar screens in the force, and was presumed to have sunk. Oil and debris continued to burn in the area for an hour and a half.

    Radar search revealing no further targets in the surrounding area, both divisions, after joining up and adjusting their formations, maneuvered to search the area of burning oil. Much burning wreckage cluttered the area, filling the air with a mingled smell of burning fuel, diesel oil and wood. Observing many survivors in the water, Comdr. Moosbrugger directed Division TWELVE to attempt to pick up some of them. However, at this time the Maury reported that engineering difficulties would probably prevent her from maintaining her speed, a maximum of 22 at that time, for more than an hour. At 0118, therefore, Comdr. Moosbrugger retired down the “Slot”4 with Division TWELVE after directing Comdr. Simpson of Division FIFTEEN to attempt to pick up survivors for intelligence purposes.


 4“The Slot” was previously designated on Solomon Island charts as New Georgia Sound. On later charts this bod of water is given the name of “The Slot”, in conformity with common usage.


    Steaming through the survivor area at 25 knots, Division FIFTEEN required about four minutes to traverse waters that were filled with men clinging to rafts and wreckage. To the Commanding Officer of the Lang it seemed that “the sea was literally covered with Japs”—so thick that their bodies were seen to be thrown up in the phosphorescent wake of the vessel. From all sides the survivors lifted a cry that sounded like “Know-we, Know-we”, chanted in unison with considerable volume.5 “It was a weird unearthly sound punctuated at times by shrieks of mortal terror.” When speed was reduced and efforts were made to pick up survivors someone in the water blew a whistle, the chanting stopped and the men all swam away from the ship.

It was concluded from the number of men seen in the water, a number clearly in excess of that accounted for by the ships’ crews, that all four vessels were probably carrying troops and supplies to Kolombangara or reinforcements for Vila. Enemy casualties were thought to have been quite heavy, since there was no opportunity for orderly abandonment in the case of any one of them. Moreover, the area surrounding the large destroyer had been a flaming sea of oil, and the waters surrounding the other vessels had been smothered with gunfire, much of it 5-inch AA common.

Unable to get any Japanese survivors aboard, or to discover any other targets, Division FIFTEEN retired from Vella Gulf about 0200 and proceeded to Tulagi. The reassembled task group discovered no material damage to any ship and no personnel casualties whatever as a result of its highly successful engagement.


As is usual in these night actions, evidence as to the exact extent of the enemy’s losses is not conclusive. There seems to be no doubt that there were four Japanese vessels involved in the engagement. All the observers among our forces whose reports are on record believed that all four were destroyed. On the other hand, several Japanese survivors who were picked up later insisted that one of the destroyers escaped. Our own observers consistently refer to the “cruiser” present among the Japanese vessels. The gunnery officer of the Dunlap, who was in a good position for observation, identified it as belonging to the Kuma or the Natori class. Contrary to this, however, is the testimony of the prisoners, who


5 The word was probably Kowai, a Japanese ejaculation of fear.


agreed (even to the names of the ships) that all fur of the ships were destroyers. At the present time, therefore, all that can be claimed conclusively are three destroyers sunk and one heavily damaged.

    As a result of this battle heavy enemy reinforcements were prevented from reaching Kolombangara. This probably decided the Japanese to evacuate the Island instead of holding it. “Had they elected to defend it,” observed CINCPAC, “it would have cost us a diversion of effort, either in ground forces to take it by assault, or in air and surface forces to keep it neutralized as we by-passed it and moved further to the northwest.”

Destruction of Barges in Vella Gulf, 9-10 August

    On 9 August Admiral Wilkinson received word from COMSOPAC that another Tokio Express composed of cruisers and destroyers appeared to be preparing another attempt to reinforce the garrison at Vila and that troop-carrying barges might be present also. Admiral Wilkinson ordered Comdr. Moosbrugger’s Task Group to proceed to Vella Gulf, arrive at Gizo Strait at 2200, and make a sweep of the Gulf to destroy enemy shipping.

    Task Group MIKE was composed of the same ships that had fought the Battle of Vella Gulf on 6-7 August, except for the substitution of the Gridley, Lt. Comdr. Jesse H. Motes, Jr., for the Maury in DesDiv TWELVE and the substitution of the Wilson, Lt. Comdr. Walter H. Price, for the Stack in DesDiv FIFTEEN.

    The two Division Commanders, Comdr. Moosbrugger and Comdr. Simpson, held a short conference before the departure of the task group. It was decided that the battle plan used on the night of 6-7 August would be followed and the same assumptions would be made. Since contact with the enemy had not been made on that occasion until our task group was nearly out of the Gulf, it seemed probable that the Japanese assumed that our ships had come up the Slot instead of through Gizo Strait. It was considered unnecessary, therefore, to vary the route on the 9th.

    Task Group MIKE formed in the vicinity of Savo Island and took a course south of the Russells and Rendova Island. Three Black Cats were assigned to the group for scouting, but again no contact was established with any of the Cats during the night.

    Entering Vella Gulf by way of Gizo Strait about 2258, the task group swept partly around Gizo Island, then at 2344 changed course to 0000 T.


B-17 over Gizo Island

Battle of Vella

Night Action North of Vella Lavella

Brakoma Airfield on Vella Lavella

to make a sweep of the western shore of Kolombangara. The sea was smooth, the wind from the southeast, force 4; the sky was overcast and visibility was between 700 and 3,000 yards.

    Two radar contacts were made on small unidentified targets soon after the group entered the Gulf. The first contact, established at 2313, was probably a barge near the shore of Liapari Island. The second contact, probably another barge, was made at 2347 on the opposite side of the Gulf, off Sandfly Harbor, Kolombangara. The Task Group Commander ordered that neither target be taken under fire. His primary mission was to destroy the Tokio Express, and since radar search of the Gulf had not been completed he did not wish to reveal the presence of the task group.

    Comdr. Moosbrugger changed course to 0300 T. at 0003, and increased speed to 25 knots. At 0014 a third radar contact was made which after a few minutes appeared to be three enemy barges. Radar search had been carried far enough by this time to indicate that there were no enemy destroyers in the Gulf. Believing that the Express might be approaching from the north, Comdr. Moosbrugger directed Division TWELVE to hold fire and continue northward, while Division FIFTEEN took the targets under fire.

    At 0020 Division FIFTEEN sighted three barges at 1,000 yards and fired on these targets intermittently until 0032. The barges returned fire with 25 caliber machine guns, but caused no casualties. Two of these barges were possibly sunk and one damaged.

    Division TWELVE completed a northern sweep of the Gulf, and Division FIFTEEN rejoined on the southward sweep of the coast of Vella Lavella. At 0018 the Gridley made contact on a target near Liapari Island that was thought to have been the barge located earlier near the same place. Five ships opened fire on this target which disappeared from all screens about 0137, and was believed to have sunk.

    The Task Group turned northward again at 0138 and picked up another contact bearing 0640 T., distant 7,340 yards, which may have been one of the three barges engaged at 0020. The Group closed range and Division TWELVE and the Lang opened fire at 0151, continuing to fire for five minutes. The contact disappeared from the screen and was presumed to have sunk.

    After operating in Vella Gulf for nearly four hours, Comdr. Moosbrugger retired through Gizo Strait and turned southward at 0241 to return to his base by the same route used in his approach.


    In all seven radar contacts were made, two of which may have been twice contacted. Only brief glimpses of the targets could be gained. From this incomplete evidence it was estimated that they were barges between 80 and 100 feet in length. They were tracked at speeds up to seven knots. Their small caliber gunfire was ineffective, causing no casualties to our forces. 


Strategic Considerations

    One week after the Battle of Vella Gulf another major step was taken in the Solomons Campaign. The capture of Munda and the consolidation of control over New Georgia had left Arundel Island, Kolombangara, and Vella Lavella, as well as several smaller islands, to be occupied. Of these, Kolombangara with a garrison of more than 5,000 Japanese troops, strong fortifications, and an airfield was easily the most formidable. Up until 12 July, Kolombangara had been scheduled as our next objective. On that date, however, when the fall of Munda seemed to be impending and it seemed likely that this airfield could be used to extend the range of our fighter cover, Admiral Halsey decided to cancel the earlier plan in favor of an assault upon Vella Lavella.

    This decision marked an important departure from the strategy hitherto followed in the South Pacific. Our advance so far had been from one enemy-occupied and defended island to the next. The primary objective in both Guadalcanal and New Georgia had been an enemy airstrip. The new strategy was to by-pass the enemy defenses on Kolombangara and his airfield at Vila and advance our island perimeter many miles beyond to an island with negligible defenses and no airfield. Although there were significant differences, this strategy had been in part anticipated in the Aleutians Campaign when we by-passed Kiska and landed on Attu 11 May 1943.

    Lying northwest of Kolombangara, athwart the path of the supply routes of enemy bases on that island, Vella Lavella would furnish bases for more effective patrol of both Vella Gulf and Blackett Strait. The latter provided the favorite rote for Japanese barges running supplies to the garrison at Vila. It was intended to establish a minor naval base and an airstrip on the island from which Japanese shipping and air bases of southern Bougainville might be attacked.

    It was known that Vella Lavella had received little attention from the Japanese during their occupation of the Solomons in the summer of 1942.


Current intelligence reports estimated enemy strength in the island to be perhaps 250 men. These men, equipped with light antiaircraft guns, were concentrated principally on the northwest coast, from which they operated barge supply points.

    The coastal area of Vella Lavella is a narrow strip varying in width from 100 yards to one mile. Beyond this the ground rises abruptly toward a central ridge from 2,000 to 3,000 feet in height. Apart from the coconut plantations along the coast, usually located opposite channels giving entrance through the main reefs, the country is overgrown with heavy jungle through which it is impossible to see more than a few yards.

    On the night of 21-22 July, a party of six Army, Navy and Marine officers was landed on Vella Lavella by PT boat for the purpose of making a comprehensive reconnaissance of the southern part of the island. The party investigated Liapari Island, Biloa Anchorage, landing beaches, and a site for an airstrip. Taken off six days later, along with the rescued crew of a Catalina patrol plane, the reconnaissance mission reported suitable beaches and a site for a PT base at either Liapari or Nyanga. Because of good drainage, safe approaches, suitable beaches and bivouac areas, Barakoma was selected as the site for the landing and for the construction of an airstrip. Furthermore, the locality was not occupied by the Japanese and no opposition in force was expected.

Task Force Organization

    Acting upon Rear Admiral Theodore S. Wilkinson’s recommendations, Admiral Halsey designated 15 August as D-day,6 and authorized a force composed of Naval, Army and Marine Units for the operation.

The composition of the main body of Task Force WILLIAM under command of Admiral Wilkinson was as follows:

    Advanced Transport Group, Capt. Thomas J. Ryan, Jr.

        TransDiv TWELVE, Comdr. John D. Sweeney

            Stringham (F), Lt. Comdr. Ralph H. Moreau

            Waters, Lt. Comdr. Charles J. McWhinnie

            Dent, Lt. Comdr. Ralph A. Wilhelm

            Talbot, Lt. Comdr. Charles C. Morgan

    TransDiv TWENTY-TWO, Lt. Comdr. Robert H. Wilkinson

        Kilty (F), Lieut. John W. Coolidge.


6 It is an interesting coincidence that 15 August also became D-day for the landing on Kiska, which had been by-passed by the taking of Attu.


        Ward, Lt. Comdr. Frederick W. Lemly

        McKean, Lt. Comdr. Ralph L. Ramey

    DesDiv FORTY-ONE, Capt. Thomas J. Ryan, Jr.

        Nicholas (F), Lt. Comdr. Andrew J. Hill, Jr.

        O’Bannon, Lt. Comdr. Donald J. MacDonald

        Taylor, Lt. Comdr. Benjamin Katz

        Chevalier, Lt. Comdr. George R. Wilson.

        Cony (FF), Comdr. Harry D. Johnston

        Pringle, Comdr. Harold O. Larson 

Single Transport Group, Capt. William R. Cooke, Jr.

    LCI Unit, Comdr. James M. Smith

        LCI’s 61, 23, 67, 68, 332, 334, 222 (F), 330, 331, 333, 21, 22

    DesDiv FORTY-THREE, Capt. William R. Cooke, Jr.

        Waller (F), Comdr. Laurence H. Frost

        Saufley, Comdr. Bert F. Brown

        Philip, Lt. Comdr. William H. Groverman, Jr.

        Renshaw, Lt. Comdr. Jacob A. Lark 

Third Transport Group, Capt. Grayson B. Carter

    LST’s 354, Lieut. Bertram W. Robb

           395, Lieut. Alexander C. Ford

           399, Lt. (jg) Joseph M. Fabre

    DesDiv FORTY-FOUR (less Pringle, Cony), Comdr. James R. Pahl

        Conway (F), Comdr. Nathaniel S. Prime 

        Eaton, Comdr. Edward L. Beck

        SC’s 760, 761 



    MTB Flotilla CAST, Comdr. Allen P. Calvert 

        New Georgia MTB’s

            18 MTB’s 

        Kolombangara MTB’s

            8 MTB’s 

Northern Landing Force, Brig. Gen. Robert B. McClure, U. S. A.


        Headquarters Detachment

        4th Marine Defense Battalion (less 155 mm Gun Group; Btry. D [90 mm];

        1/2 Btry. G[S/L] and Tank Platoon).

        35th Regimental Team (less certain detachments)


7 After performing convoy duty to Barakoma beach these destroyers were to form a screen for landing operations.


Naval Base Force, Capt. George C. Kriner

        58th Construction Battalion (less rear echelon)

        Naval Base Units including Boat Pool No. 9 

Preliminary Operations 

    During the week preceding the operation considerable air strength was reported in Kahili and Rabaul, and destroyers were observed in the Buin-Shortlands area. Photographs of the Rabaul area taken on the 13th revealed—besides 19 cargo vessels—two cruisers, five destroyers, six submarines, and smaller warships. Airdromes in the area held an estimated total of 271 aircraft. In the harbor were 25 floatplanes and four flying boats. All these hostile forces had to be taken into account in the plan of operations.

    Surface forces of the Third Fleet were held in reserve to provide protection against any major enemy forces that might put in an appearance. Motor torpedo boats based on Rendova and Lever Harbor were to screen our forces by picket lines to the south, the west, and the northeast of Vella Lavella during the night of 14-15 August, returning to base at daylight on the 15th.

    Shore-based aircraft from Munda were to support the operation. An attack upon Rabaul by aircraft from the Southwest Pacific Area was requested in view of the concentration at that point, but General MacArthur reported this to be impracticable because of other commitments.

    In spite of adverse weather through most of the week preceding the attack on Vella Lavella, Allied planes struck at enemy bases and depots in the Vila, Buin and Rekata Bay areas. TBF’s and SBD’s dropped 36 tons of bombs on gun positions at Disappointment Cove and Kape Harbor, Kolombangara, on the 13th. On the 12th and 13th Liberators dropped a total of 49 tons of explosives on the airfield at Kahili, destroying many planes and shooting down 11 enemy fighters. The seaplane base in Rekata Bay was attacked seven times and the Vila airdrome was raided almost nightly during the week.

    As a final preliminary to the action, an advance party was landed at Barakoma on the night of 12-13 August to mark the channels and beaches to be used by landing craft and to select bivouac and dispersal areas and defense positions. Another duty assigned this advance party, which consisted of only 25 men, was to take custody of a large number of Japanese prisoners reported to be held by native sentries. Upon arrival, however, it was discovered that no prisoners were in hand, but that several


hundred refugees from Kolombangara and survivors from the enemy ships sunk during the night of 6-7 August were at large. These Japanese were reported to be armed with hand grenades, clubs, and a few firearms. Reinforcement was immediately requested and after some delay the troops were moved in from Rendova by motor torpedo boats after dawn on the 14th.

Approach and Landing at Barakoma

    In the meantime, loading of equipment and supplies on the LST’s at Guadalcanal had begun on 12 August and was completed the following day. On the 13th troops embarked, conducted a debarkation drill, and reembarked. At 2130 that evening at Kokumbona a single enemy plane came in low for an attack on the Eaton. The destroyer opened fire and observed its tracers hitting the approaching plane, which quickly ignited, whirled off to the west and crashed inland.

    With the exception of one reinforced battalion of the 35th Combat Team which embarked in the Russells, all units of the Main Body of the Task Force embarked at Guadalcanal. The three Transport Groups, each with its screen of destroyers, departed independently on 14 August on a schedule arranged to time their arrival at Barakoma on the 15th as follows:

  Depart  Arrive 
  Guadalcanal  Barakoma 
  14 August             15 August 
Advance Transport Group                                                            1600                            0610                    
Second Transport Group  0800  0710 
Third Transport Group           0300  0800 

It was hoped thus to avoid undue exposure to air attack by giving each group full use of the beaches and cutting to a minimum the time other groups would be kept awaiting opportunity to unload.

    Admiral Wilkinson chose as his flagship the Cony, on which General McClure also embarked. The Pringle was designated primary fighter director ship with the Conway as relief. The O’Bannon and the Taylor, as bombardment ships, were to open fire on any shore battery which threatened our landing craft. Screening destroyers were to maneuver to seaward of the transport area.

    The weather during the approach to Barakoma was excellent, the sea calm, the sky nearly cloudless, and at night there was a brilliant moon.


All Task Groups took the same course, lying south of the Russells and Rendova, thence north through Gizo Strait. Northwest of Rendova the three groups "leap-frogged.” Shortly before dawn the Third Group with LST’s was slowly overhauled and passed by the Second Group with the LCI’s, while almost simultaneously the Advance Group with the swift APD’s overtook and passed both the slower Groups. The Units finally entered Gizo Strait in correct order. So far, but for the approach of an enemy snooper plane which apparently passed without spotting our force, the passage had been uneventful.

    The Advance Group arrived off Barakoma at the break of dawn and began unloading troops and equipment at 0615. Debarkation proceeded with dispatch, and was completed by 0715. The APD’s departed on the return trip to Guadalcanal at 0730 with a screen of four destroyers. It was with the beaching of the LCI’s of the Second Group at 0715 that the schedule was broken. It was discovered that only 8 of the 12 landing craft could be accommodated at one time by the three beaches. This condition plus an error in communications from the beach party delayed the completion of unloading the last four LCI’s until about 0900. In the meantime the LST’s had arrived at 0800 according to schedule and were awaiting their turn at the beach.

Antiaircraft Action

    Friendly aircraft from Munda were overhead by 0605. Since the landing was carried out within 90 miles of Kahili, the largest Japanese air base in the Solomons, it was no surprise when the radar of the fighter director destroyer picked up bogies at 0747. The original raiding party was considerably larger than the number of planes that succeeded in breaking through to deliver the attack. Our fighters, which successfully intercepted the attack, were credited with shooting down many planes.

    At 0759 between 15 and 20 enemy fighters and dive bombers commenced an attack on the destroyers of the screen, ignoring for the moment the more vulnerable targets presented by the beached LCI’s and the slow LST’s. The flagship Cony sustained three near hits, two within 50 yards of the ship, but there was no damage to the vessel and only one man was slightly wounded. Planes also dove on other destroyers, straddling the Philip’s bow narrowly and bracketing LST 395 with two near hits, but without damage to either. Antiaircraft fire from the attacked vessels was effective in helping drive off the attack. The Nicholas, O’Ban-


non, Taylor and Chevalier, still in the area during the attack, also joined in the firing.

    The LCI’s completed unloading by 0900 and retired with all the remaining destroyers except the Conway and the Eaton. The Advance Group and the Second Group returned separately to Guadalcanal.

    Between 0900 and 0915 the three LST’s beached and began unloading. LST 395 struck a ledge of broken coral short of dry land, necessitating the construction of a ramp. This was quickly accomplished with the aid of a bulldozer that had been placed in the bow of the ship for such a purpose. Difficulties in clearing jungle growth for space to deposit supplies, and in keeping at work the personnel assigned unloading duty threw the LST’s off schedules. Instead of departing at 1600 the Group did not retire before 1800.

    Unloading proceeded laboriously under a burning sun and a perfectly clear sky. Visibility was so excellent that Corsairs and Kitty Hawks circling at altitudes up to 24,000 feet were clearly visible to the unaided eye.

    After the attack of 0800 the ships remained unmolested until 1227 when the heaviest attack of the day began, continuing about 15 minutes. The Japanese planes, dive bombers of the Aichi 99 type plus Mitsubishi Zeros, were intercepted, but two groups consisting of eight to twelve bombers broke through and attacked. One of these groups attacked the beached LST’s; the other two struck at the destroyers. Bombers were seen circling at 14,000 feet and starting their glides, some of which were steep, others shallow.

    The first bomber to dive on the Conway was shot down, crashing within 50 feet of that ship. Three of its bombs burst close off the ship’s starboard bow. The salvo from the second plane fell even closer on the starboard bow, one bomb within 20 yards. The third plane was shot down and crashed 500 yards from the Eaton. The next three bombers sustained no apparent damage, but their bombs missed badly.

    Eight bombers came in high and fast out of the sun to attack the LST’s. Near hits were scored on two of the ships. No material damage was done, though one man was seriously wounded. One of the landing ships claimed to have downed two bombers with antiaircraft gunfire and another at least one.

    Scarcely had the bomber attack been broken up when the Conway spotted seven fighters, Zeros and Vals, hedge hopping over a saddle in the ridge of the island, closing in for a strafing attack on the landing area.


The gun control officer of the Conway opened effective fire before the planes were sighted by the landing ships, and brought down one. Amply warned by the destroyer’s fire, the crews of the LST’s put up a tremendous fire from small and medium weapons which completely frustrated the attack. One of the landing ships claimed three planes, another two, and the third one plane shot down. Capt. Grayson B. Carter, commander of the Third Transport Group, considered the repulse of this attack “the highlight of the day.” In all, our forces reported they shot down ten planes in the noon attack.

    Unloading was resumed and proceeded uneventfully until warning of a fourth attack came at 1724 from the Combat Air Patrol. This attack group, a large one coming in from the northwest, was intercepted and all but broken up by the Air Patrol. Only about eight single planes succeeded in getting through. They dived wildly, released bombs indiscriminately, and fled low over the water without inflicting any damage.

    Shortly after this attack, the Third Transport Group completed unloading except for about 130 tons of cargo aboard two LST’s and retired at 1800 to avoid exposure to night attacks while without fighter cover.

    Defense against air attacks during the day had been highly successful so far as the ships and their personnel were concerned. Ashore, however, twelve of the landing force had been killed and forty wounded, though damage to material was slight.

    On the return to Guadalcanal under a brilliant full moon and a clear sky, two of the three Transport Groups were subjected to repeated air attacks. From five to eight Mitsubishi 97’s attacked the Second Group, consisting of 12 LCI’s and four destroyers, when it was south of Rendova Island. The planes attacked singly at intervals from 2050 to 2140, dropping, it was believed, six torpedoes in all. No hits were made, however.

    As in the daylight attacks, it was the Third Transport Group that took the most prolonged punishment. Between 2034 and 2330 the Group underwent six horizontal bombing attacks, and for four hours the convoy was surrounded or partially surrounded by flares and float lights dropped by the enemy. The bombs were usually dropped in patterns of eight, and were thought to be not over 100 pounds in weight. So often were the awkward landing ships bracketed and straddled that one of their commanding officers was sure that LST meant “large slow target.”


 All of these ships were sprayed with shrapnel. One reported some 200 holes punctured in bulkheads. Another reported 14 near hits from bombs within 100 feet of the vessel. The destroyers made frequent use of smoke screens. That our ships survived the attacks with only minor damage and insignificant casualties was attributed, among other things, to “a perfectly phenomenal supply of good luck.” Several officers commented on the excellence of Japanese night search and attack.

Operations Ashore

    In the course of D-day the three Transport Groups had landed at Barakoma a total of 4,600 troops, including 700 Naval personnel; 2,300 tons of equipment and supplies, including eight 90-mm antiaircraft guns; 15 days’ supplies of all classes, and three units of fire, except for the 90-mm guns, for which one Marine Unit (300 rounds per gun) was landed. The supplies discharged were considered sufficient to sustain the landing party well beyond the scheduled arrival of the next echelon.

    The 4th Defense Battalion of the Fleet Marine Force immediately upon landing began setting up antiaircraft guns in temporary shore positions. By 1530, sixteen 50 caliber machine guns, eight 20-mm, and eight 40-mm guns were in position. Marine gunners were credited with shooting down five enemy dive bombers during the day. By 1800 two searchlights were in readiness. During the night of 15-16 August shore installations were attacked 12 times by enemy planes, but only minor damage was inflicted.

    Troops of the 35th Regimental Combat Team had landed without opposition and proceeded with their task of establishing a temporary defense perimeter. This was accomplished by noon without any Japanese resistance having been encountered, and field artillery was emplaced in temporary positions by 1700. In the meantime naval base units, including the 58th Construction Battalion, commenced work upon docks, ramps, roads, airstrips and dispersal areas. 


Preliminary Maneuvers

    Early in the afternoon of 17 August a plane contact with enemy forces from Bougainville was reported at Admiral Wilkinson’s headquarters. The report gave the location, disposition, course and speed of four destroyers making up a Tokio Express probably destined either for the relief of Kolombangara or an attack on the new base at Barakoma.


    Admiral Wilkinson dispatched Capt. Thomas J. Ryan, Jr., commanding officer of Destroyer Squadron TWENTY-ONE, with four destroyers of Division FORTY-ONE to intercept the enemy force. Task Unit ROGER was organized as follows:

    Nicholas (F), Lt. Comdr. Andrew J. Hill, Jr.

    O’Bannon, Lt. Comdr. Donald J. MacDonald.

    Taylor, Comdr. Benjamin Katz.

    Chevalier, Lt. Comdr. George R. Wilson.

    Forming column open order, distance 500 yards, in the order named above, the Task Unit departed Purvis Bay, Tulagi Island, at 1527, 17 August, and proceeded “up the Slot” at 32 knots, course 3050 T. This route took the force east of New Georgia Island to arrive at a position 30 miles bearing 3050 T. from Visu Visu Point, on New Georgia, at 2300, and thence north of Kolombangara.

    According to the original plan, Capt. Ryan’s force was to be augmented by two destroyers from DesDiv FORTY-THREE, which was then screening the Second Echelon on its return from Barakoma. The Philip and the Saufley, under the command of Capt. William R. Cooke, Jr., were to join Task Group ROGER en route to intercept the enemy, and destroy barges expected to be under escort by the Japanese destroyers. However, upon hearing by TBS that the Waller, the third destroyer of DesDiv FORTY-THREE, was in difficulty, and knowing that Capt. Cooke was escorting LST’s through Gizo Strait, Capt. Ryan ordered the Philip and the Saufley to return to Capt. Cooke.

    A minor casualty aboard the Chevalier threatened to diminish the strength of the Task Unit still further. At 1630 her number two main feed pump became steam-bound and necessitated the vessel’s dropping out of formation to make necessary repairs. Since the casualty had been foreseen, repairs were quickly made and the Chevalier rejoined at about 1655.

    Between 2035 and 2253 numbers of bogies were tracked, closing and opening range. Some of these were probably friendly planes. At any rate no attack was made on the force. All hands were called to general quarters at 2210.

    Toward midnight gun flashes were seen from the enemy force then under attack by our TBF’s. At 0025, one of our torpedo planes that had participated in the attack reported four enemy destroyers midway between


Vella Lavella and Choiseul Island on course 1200 T., speed 12 knots. Repeated efforts to obtain further information from planes, including two Black Cats, were unavailing because of poor radio reception at both ends.

    Shortly before this report, the weather had cleared. A brilliant moon, almost full, which had been obscured by frequent rain squalls until 0010, now provided visibility for more than eleven miles. The sea was calm, the wind from the southeast between force 1 and 2.

Engagement with Destroyers

    0029         O’Bannon reports surface contact.

    0039         Enemy aircraft drops flares, scores near bomb hit on Nicholas.

    0056         Enemy DD’s open fire.

    0058         Nicholas opens fire followed by O’Bannon and Taylor.

    0100         Chevalier fires five torpedoes.

    0110-13    Explosions noted on two enemy destroyers, fire on third.

    0121         Engagement with enemy destroyers broken off. 

    At 0029, when our forces were off the northeastern tip of Vella Lavella Island, in latitude 070 26’ 25” S., longitude 1560 45’ 05” E., the O’Bannon reported a surface contact bearing 3130 T., distance 23,000 yards. The Chevalier made contact at almost the same time, and the enemy was visible within two minutes. A number of barges were located between our formation and the enemy destroyers. Capt. Ryan ordered a series of changes in course in order to close the destroyers.

    It was believed that enemy planes, which radar plot indicated had been circling Task Unit ROGER since 0014, had sighted and reported its presence to the enemy ships. With the assistance of spotting planes, excellent visibility, and a full moon, the enemy was in all probability aware of the presence of our ships when the O’Bannon first made radar contact. At any rate the Japanese destroyers had formed up and changed course as if to avoid engagement.

At 0039 enemy planes dropped flares that illuminated our formation and also a stick of bombs that landed about 100 yards off the starboard bow of the Nicholas at the head of our column.8 Eleven minutes later the planes dropped two more flares, just as our formation was passing near the group of enemy barges. These now appeared to consist of numerous small barges towed by two tugs, and two large ones. For the moment our ships passed them up, with the intention of returning later to destroy them.


8The Commander of the O’Bannon believed these bombs fell 100 yards off the starboard quarter of the Nicholas.


At 0056, when the range was about 14,000 yards, the Japanese destroyers opened fire—about two minutes before our ships commenced firing. Initial enemy salvos were over our formation some 600 yards, though the pattern of gunfire was reported to be good, and the third or fourth salvo fell short, almost finding the range. The Chevalier received a near hit about 50 feet off the starboard beam, which drenched personnel on the bridge with its splash. For approximately two minutes after the enemy opened fire, the Nicholas, O’Bannon, Taylor and Chevalier were in column on course 2700, and three of the Japanese destroyers, also in column, were on a course of about 1800. For this prief period, while almost crossing our bow, the enemy gunners enjoyed a tactical advantage.9  Additional advantages of the enemy were a bright moon, against which our force was silhouetted, and the flares previously dropped by planes. After the first few salvos, however, enemy gunfire became gradually less accurate until at the close of the engagement shells were landing between two and three thousand yards short.

At about the time the Japanese opened fire, one of their destroyers swung out of formation to close range rapidly, apparently with the intention of firing torpedoes. Accepting the torpedo menace, Capt. Ryan continued to close range until 0058 when the formation executed a column right to open gunfire, coming to course 3500 T., all ships following the Nicholas without signal. At 0058 the Nicholas, and at 0100 O’Bannon and Taylor commenced fire upon the enemy column, at ranges varying from approximately 8,500 yards to 11, 350 yards.

At 0100 the Chevalier, rear ship in the column, fired five torpedoes spread 20, low speed, with full radar control at the enemy destroyer that had advanced out of formation bearing 3000 T., distant 6,800 yards. Immediately after firing her torpedoes she joined in the gunfire. The concussion from Gun No. 3 fired an additional torpedo. The targets at that time were on course 2000 T., speed 30 knots.

Soon after our ships opened fire the Japanese destroyers executed a turn to the northwest, deserted the barges, and fled at high speed. For a few moments after the enemy turned, our ships enjoyed a tactical advantage similar to that the Japanese had enjoyed when they first opened gunfire. Hits were reported on all four targets, some of which were believed badly damaged.


9See track chart.


For ten minutes after we opened fire, radical maneuvers to avoid enemy torpedoes were necessary, during which firing had to be suspended intermittently. Japanese torpedo fire was quite accurate. Although none of our vessels was hit, one torpedo wake was reported to have crossed about 25 to 30 yards ahead of the Chevalier, and another to have passed close astern of the Taylor. The Nicholas also reported torpedo wakes.

    The Task Unit had orders to fire torpedoes when ready. None of the vessels except the Chevalier, however, believed that the range was suitable for a torpedo attack, since torpedoes had been set on intermediate speed for the action. The Chevalier planned to fire a half salvo from mount two, but this jammed, and mount one had to be used. Torpedoes in the former were on intermediate speed; those in mount one on low speed. At 0110 observers on the Chevalier saw a heavy explosion on the bow of the rear destroyer in the enemy line and the Nicholas heard a distinct underwater explosion. The destroyer on which the explosion was observed turned left, decreased speed, and seemed to smoke heavily.

    At 0111 the column reformed after evasive maneuvers, and a turn was made to the left on course 300 degrees T., which was roughly parallel to the course of the retiring enemy. Firing was continued at a range of 15,500 to 17,600. The Taylor and O’Bannon concentrated their gunfire on the third destroyer in the enemy line, obtaining a number of hits. An explosion was observed on this target, which also began to smoke.

    In this phase of the action the Nicholas suffered serious casualties in her five-inch batteries. In guns Nos. 1 and 2, after approximately 200 rounds had been fired by each, the rammers failed to retract and jammed in the forward position. A cartridge also jammed in gun No. 4 and proved exceptionally hard to remove. The O’Bannon also suffered casualties in her No. 3 gun as the result of a dented cartridge case and rammer trouble, which caused the gun to miss about ten salvos. These casualties were remedied before the final phase of the action.

    Believing that he could not afford to leave the enemy barges unmolested any longer, Capt. Ryan ordered the engagement with the destroyers broken off at 0121. At that time the column was closing with the enemy and hitting again. Three of the Japanese destroyers were still in sight, one of them well ahead of the other two. The third appeared to be dropping slowly astern. 10


10Capt. Ryan believed the four destroyers to be of the Fubuki-Amagiri class. O’Bannon and Taylor believed one was of the Terutsuki class. A Black Cat pilot was certain that one of the vessels was much larger than the other three and reported it to be a cruiser. 


Scenes from night action of 17-18 August

Bow of Patterson after collisons with McCalla

Bow of McCalla after the collison

Bow of O'Bannon after Battle of Vella Lavella

Selfridge and O' Bannon

Bow of Selfridge after Battle of Vella Lavella

Starboard view of Selfridge 

Selfridge returning from Battle of Vella Lavella

    The extent of our success in this engagement is not definitely known. At 0128 observers on the Chevalier saw a heavy explosion on the enemy destroyer that had dropped astern, and about 0129 her silhouette disappeared from the horizon simultaneously with the disappearance of her pip from the SG radar screen. This was not the vessel which the Chevalier claimed to have hit with a torpedo. Lt. Comdr.     MacDonald of the O’Bannon reported that no enemy vessels were observed to sink, and that his CIC11 officer reported that all four destroyers went off the SG screen out of range. A fighter pilot later reported sighting a burning destroyer off Warambari Bay at the northern tip of Vella Lavella, at 0620, 18 August. This may have been one of the destroyers engaged. All that can be positively claimed, however, seems to be two destroyers damaged. There was no damage to our ships or personnel.

Destruction of Barges

    Upon breaking off action with the enemy destroyers, Task Unit ROGER reversed course to search for the barges passed earlier. At 0148 the O’Bannon’s radar showed the barges bearing 213 degrees T., distance 13,000 yards, in a position only slightly to the north of where they were originally sighted. Course was then changed to close and the order given to fire when ready. The radar screen revealed four or five contacts. Range was closed by simultaneous turns until visual contact was established. Some of the barges, not self-propelled, were towed by tugs, one of which 75 feet long with a flush deck and a cabin amidships, was probably of wood. The number of smaller barges under tow was indeterminate. When firing was opened about 0155 the barges scattered to the west at eight to ten knots. One sank and others disintegrated. Machine gun fire appeared to be as effective as the 5-inch battery for these targets. By 0227, after about 20 minutes of firing, only one barge of the original four or five remained on the radar screen, and it was left burning furiously.

    At 0235 the O’Bannon reported a new surface contact bearing 183 degrees T., distance 8,000 yards. Capt. Ryan in the Nicholas turned south to investigate this contact, and ordered the Taylor, O’Bannon and Chevalier to continue their search for barges in the area of original contact. The Nicholas discovered one large barge similar to our LCT, set it on fire, and watched it burn intensely and sink at a range of less than 1,000 yards.


11Combat Intelligence Center.


Another barge was set afire in the same area. In all, at least four large barges plus an indeterminate number of small barges were thought to have been destroyed. There was no evidence of gunfire or resistance of any kind from the barges.

    At 0306 unidentified planes began an attack on the Task Unit. A stick of bombs fell across the O’Bannon, two landing close to the port quarter and two more ahead on the starboard bow; another stick dropped about 300 yards on the starboard beam of the Nicholas, and two bombs fell within 50 to 100 feet of the Taylor. The O’Bannon had already opened fire on these planes when IFF identified them as friendly. This identification, however, proved to be coming from our own ships, which had their IFF turned on at the time, and it was concluded that the planes were not friendly.

    Proceeding south close to the reefs six miles off Vella Lavella, the Taylor, Chevalier and O’Bannon fired star shells at 0328 to illuminate the northeast coast of the island. No further targets were discovered. At 0400 the Nicholas, with Capt. Ryan, rejoined the formation, and a sweep of Vella Gulf was made. Aside from firing upon an unidentified plane, and another bombing attack which landed a stick of bombs close to the O’Bannon, the retirement to Tulagi was without incident.


    The reason for the enemy’s precipitous desertion of his barges is a matter for speculation. The fact that no personnel was observed aboard the barges and that when set on fire they burned furiously with periodic explosions leads to the supposition that they were carrying supplies only. It seems probable that the combatant vessels were carrying troops. If so this might explain their reluctance to engage and their willingness to leave the barges unprotected.

    The bright moonlight, which extended visibility to more than eleven miles, reduced the tactical advantage our forces enjoyed in their radar. Had maximum visibility been 6,000 yards or less, it seems probable that success against the destroyers would have been more nearly equivalent to that achieved in several other similar engagements in this area.


    The amphibious phase of the Vella Lavella campaign continued for twenty days after the initial landing and was not complete until four


more echelons landed supplies, equipment, and personnel from Kokumbona and Kukum on the coral beaches of Barakoma. Although the enemy failed to oppose these operations with surface vessels, his air force did not neglect the opportunity our supply ships presented while moving slowly through the narrow waters of Gizo Strait and Blanche Channel or lying beached at Barakoma.

    The 35th Regimental Combat Team, concerned with establishing and maintaining a defense perimeter, had accomplished its mission without meeting any resistance. Patrols made a few contacts with small unorganized groups of Japanese who offered no resistance and kept moving northward. The military situation had therefore remained virtually unchanged since midday of 15 August.

The Second Echelon

    The Second Echelon, under the command of Capt. William R. Cooke, Jr., was scheduled to arrive at Barakoma two days after the initial landing. This consisted of LST’s 339, 396, and 460 screened by Destroyer Division FORTY-THREE (less Renshaw), including the Saufley, Philip and Waller (F), and SC 1266.12  Besides equipment and supplies, these ships were also transporting detachments of the 35th Combat Team, the 4th Defense Battalion, and the 58th Construction Battalion.

    Departing from the Guadalcanal area on 16 August the echelon arrived at Barakoma at 1625 the following day. In planning the operation it had been thought that all echelons subsequent to the initial landing group could be more safely unloaded at night. Shortly after our fighter cover withdrew, two waves of enemy planes, one at 1850 and a second at 1910, attacked the beached vessels with bombs and strafing. Expecting the attacks to be continued through the night, Brig. Gen. McClure ordered the LST’s, though still not unloaded, to withdraw from the beach and return in the direction of Rendova pending further orders. With destroyers in screening stations the convoy proceeded down Gizo Strait. There the third enemy plane attack began at 2110, and continued for two hours and seventeen minutes.

    From the first attack at 1850 on 17 August, the convoy was under almost continuous attack by enemy planes until about 0200, 19 August. With only brief respite, men were at general quarters stations for a period of about 36 hours. During the greater part of this period no air cover was


12For organization of the 2nd Echelon see Appendix A.


present because of darkness or unfavorable weather conditions. Although the fighter director control ship reported repeatedly that enemy planes were in the vicinity, the attacks occurred as a surprise in all but the technical sense, in that they were made under conditions of reduced visibility, such as fog, low clouds, or through smoke screens. Enemy planes were invariably in attack positions before they could be sighted.

    That the convoy suffered no damage as a result of these bombing and torpedo attacks was probably due in large part to the effectiveness of smoke screens laid by the destroyers, and also to the fact that the enemy for the most part chose the screening vessels, rather than the LST’s as targets. With radical maneuvers and intense gunfire, these ships escaped damage.

    At 2237, while engaging bombing and torpedo planes, the Waller and Philip collided in the narrow waters of Gizo Strait. At the moment of collision both ships were laying smoke screens and making radical changes of course at approximately 30 knots. The Waller was engaging one plane while the radar screen showed two more approaching. Under the circumstances the accident was considered unavoidable. In a sideswipe collision the Waller struck the port side of the Philip. The damage was not serious enough to prevent the ships from continuing their screening duties through the night, though the following morning the Waller’s damage was found more extensive than had been thought originally, and it was considered necessary to send her to Tulagi.

    The convoy continued toward Rendova until 0143, when the course was reversed to return in the direction of Barakoma. At 0220 flames suddenly burst from the ventilators of LST 396. It was quickly apparent that the fire was beyond control and that the ship would have to be abandoned. The burning ship launched its only boat and threw life rafts overboard. Capt. Cooke ordered the Saufley and SC 1266 to pick up survivors, which was accomplished without the loss of a single life. The rescue was effected under hazardous exposure to exploding ammunition and falling debris.

    At 0232 a heavy explosion blasted out the after section of the main deck, and subsequent explosions tore the vessel apart, sending flames hundreds of feet in the air. She sank at 0320. The cause of the fire was not determined. No bomb hit was observed and there had been no underwater explosion. Although several theories were advanced to account for the first explosion, all the evidence indicated that a gasoline


vapor explosion occurred in the port shaft alley, and that this was the origin of the fire that rapidly destroyed the ship.13

    After completing their rescues the Saufley and SC 1266 had to beat off yet another air attack before arriving at Barakoma. There they found that the LST’s had already beached at 0700. Before our ships left, enemy planes made two more bombing attacks. A near hit lifted LST 339 so high on the beach that she had difficulty in retracting. Damage inflicted upon enemy planes was impossible to assess with accuracy, but it was thought possible that four planes were shot down during daylight attacks and two during the night attacks. The convoy’s return to Guadalcanal after the unloading of the LST’s was completed without further attack, though the presence of bogies kept the men at general quarters until 0200, 19 August.

Third Echelon

    With minor changes the First Echelon to Vella Lavella also acted as Third Echelon, which consisted of LST’s 354, 398 and 398 escorted by Destroyer Division FORTY-FOUR (less Cony) plus Pringle and SC 505. Departing Guadalcanal on 20 August these ships arrived at Barakoma at 0700, 21 August. En route up Gizo Strait, the formation was subjected at 0540 to torpedo, bombing and strafing attacks. Smoke screens and antiaircraft fire were used against the attack. The Pringle sustained slight damage as a result of strafing that killed two men and wounded several seriously.

    After the LST’s were beached the destroyers retired to Rendova, and SC 505 took up off-shore patrol. At 1015 Aichi 99 bombers accompanied by Zeros appeared over the beach at high altitude, having penetrated our fighter air cover. The enemy planes, after maneuvering uncertainly, apparently in search of the destroyers that had already departed, attacked the SC but retired without pressing home their attack on the LST’s.

    A third air attack, developing about 1515, was pressed home with determination against the LST’s by a formation of more than 20 planes, consisting of Aichi 99 dive bombers, Mitsubishi 97 torpedo bombers, and Mitsubishi Zeros. A few planes swerved off without dropping their bombs, and a few dropped their loads prematurely, but some ten planes dived steeply through intense antiaircraft fire directly upon the beached


13See Bureau of Ships War Damage Report No. 46 for a complete description of the loss of LST 396.


vessels. Although there were no direct hits, bombs landed thickly in the area between and around the LST’s. Near hits, one within 25 feet, lifted the bow of one ship clear of the beach, knocked off a hatch cover, jammed the elevator, and knocked personnel on the main and tank decks off their feet. Bombs exploding on the beach covered the decks with coral rocks, some of them over 100 pounds in weight, knocked down gunners, and put about 50 percent of the LST batteries out of commission. A fire broke out on LST 398 but was quickly extinguished. In all, personnel casualties amounted to four killed and eleven seriously wounded.

    The 58th Construction Battalion was highly commended for its zeal, not only in unloading but in fighting. Ignoring danger, the men of this Battalion carried out the Task Force Commander’s order to “go on and stay on till the job is done.” They continued at their task under the hottest attack, and manned batteries when their gunners were knocked out. Because of their excellent work, the LST’s were unloaded, except for a few tons remaining on two of the ships, and were able to withdraw at 1600. The return trip under escort was without incident.

Fourth and Fifth Echelons

    The next two echelons were considerably more fortunate than previous ones in their experience with enemy aircraft. Improving weather, for one thing, enabled air cover from Munda to give more effective protection, and lessons learned from previous experiences could be profitably applied.

    LST’s 399, 339 and 397, loaded with twenty 2 1/2 –ton trucks each, in addition to rations and ammunitions, were escorted to Barakoma by the Saufley, Renshaw, Cony and SC 733, the echelon under command of Capt. Cooke. Aside from mistakenly firing upon one friendly plane at 0350 and driving off a formation of enemy planes with the assistance of friendly fighter planes at 0615, the convoy made the approach without incident. All LST’s were beached by 0855, 26 August, and unloading proceeded with speed and efficiency, facilitated by use of the trucks aboard each vessel. This was the first echelon to escape air attack during unloading operations. The return voyage was similarly uneventful.

    Capt. Cooke also commanded the Fifth Echelon, using the same screening destroyers in addition to the two destroyer transports, Dent and McKean, each of which carried about 175 men. LST’s 398, 341 and


353 departed with the convoy from Kokumbona at 0400, 30 August, and five LCI’s with the 1st Battalion of the 145th Infantry joined the force off Munda the following morning at 0145. The night’s passage was uneventful except that unidentified planes were fired upon twice.

    Beaching and unloading were staggered, the Dent and McKean proceeding independently at 0600. These vessels disembarked their troops and commenced retirement independently by 0800, when the remainder of the convoy arrived. The LCI’s completed disembarkation within the next hour and started to form for the return trip under escort by the three destroyers.

    About this time, at 0900, enemy planes that had been under radar observation for several minutes closed for an attack on the destroyers. Nine dive bombers, apparently Aichi 99’s, divided their attention among the three vessels, four attacking the Saufley, two the Cony, and three the Renshaw. All ships suffered near hits, and direct hits were avoided only by radical maneuvers at high speed. All three sustained superficial damage topside, and two men on the Cony were injured by fragments. Damage inflicted on the enemy was indeterminate, though two planes were possibly shot down.

    After escorting the LCI’s to a point within 15 miles of Rendova, the destroyers returned to Barakoma for the unloaded LST’s at 1400. The return passage was accomplished without mishap.

Passage of Command

    The amphibious phase of the Vella Lavella operation was considered complete at midnight, 3 September 1943, when the island passed to the command of Gen. McClure, New Georgia Occupation Force.

    The unloading of the Fifth Echelon left the forces on Vella Lavella with approximately 40 day’s rations, fuel and other supplies for 30 days, and five units of fire.14 During the 20 days from the initial landing at Barakoma to the passage of command the following troops and supplies were landed at Vella Lavella by ships and landing craft of the Amphibious Force:


14Supply activity subsequent to the 5th Echelon was as follows:

    7 September                6th Echelon                        3 DD’s, 2 LST’s

    18 September              7th Echelon                        6 DD’s, 6 APD’s, 5 LST’s, 6 LCI’s.

    25 September              8th Echelon                        7 DD’s 8 APD’s 6 LST’s.

    1 October                     9th Echelon                        6 DD’s, 6 APD’s, 5 LST’s.


Rations  1,097 tons 
Drummed petroleum products  843 tons 
Ammuniion  2,247 tons 
Vehicles  2,528 tons 
     263 cargo type   
     190 construction type   
     134 special type      
     587 total vehicles   
Other frieight  1,911 tons 
Total cargo  8,626 tons 

    In the period between 15 August and 3 September the 35th Regimental Combat Team enlarged the defense perimeter, extending it from a point north of the Barakoma River on the east to Varisi on the west coast. By the end of August strong outposts were established at Malasova and Supato to the west and Marivari to the east. The whole southern end of the island was mopped up and patrols pushed up the east coast to Orete Cove.

    Behind these lines the unhindered development of the Barakoma area went forward according to plan. The advanced naval base was responsible for the immediate construction of beach points, roads and anchorages. As a means of supplying units with which road connections were not available, the No. 9 Boat Pool, with its 16 LCVP’s and LCM’s, became indispensable. Two or three coxswains were wounded in the course of their duty when the boats were taken ashore under fire. By 3 September the 58th Construction Battalion had succeeded in clearing ground for the airfield strip and had begun the work of grading. The establishment of an advanced base was moving forward along lines that were to establish the pattern of subsequent island operations. 


    On 25 August, Bairoko Harbor, the last enemy stronghold on New Georgia, fell to Allied troops. Thus eight weeks after the Allies made their first landing on the island, all organized Japanese resistance ceased on New Georgia. Considerable equipment and food and a few stragglers were found at Bairoko by the occupation forces.

    Two days after the occupation of Bairoko Harbor, Allied ground


forces crossed the narrow waters of Hathorn Sound and landed on Arundel Island. Although the landing was unopposed, resistance soon developed in the northern part of the island. At about the same time our troops completed mopping up operations on Baanga Island, southeast of Arundel, and occupied the minute islands between Baanga and Arundel almost without opposition.

    The last week of August and the entire month of September were spent in consolidating and developing positions won earlier in the campaign, and in exploiting strategic advantages gained through by-passing enemy positions south and east of Vella Lavella. Besides the strong Japanese garrison on Kolombangara, determined enemy forces clung to positions on the northeast coast of Arundel Island, and bases of lesser importance Ganongga, southwest of Vella Lavella, and to the east on Santa Isabel.

    The fact that these positions were by-passed and flanked did not mean that they could be written out of the campaign. When the enemy could no longer depend upon the nightly Tokio Express for supplies, he resorted to a large fleet of barges that hugged the reefs and coast by night and faded into the overhanging jungle by day. When barges no longer sufficed, he ran in submarines and float planes, and finally planes that landed supplies by parachute.

    In answer to Japanese tenacity, we maintained constant patrol of the air and sea approaches of the besieged enemy and laid additional mine fields. The object was to bomb and starve the Japanese out of their positions without the expenditure of men and material involved in a direct assault.

Minelaying West of Kolombangara

    Blackett Strait, opening off Vella Gulf west and south of Kolombangara, was the most sheltered and convenient route for Japanese shipping to the Vila-Stanmore area. A minefield had been laid in the Strait on the night of 6-7 May 1943, with gratifying results.15 It was decided that a new minefield should be laid in the same area in the early morning of 25 August.

    This operation was undertaken by a task group composed of the Pringle, Comdr. Harold O. Larson (CTG LOVE), with the light minelayers Preble, Lt. Comdr. Frederic Steinke, Montgomery, Lt. Comdr. Dwight L. Moody and Breese, Lt. Comdr. Alexander B. Coxe, Jr. These


15 See Combat Narrative Solomons Campaign: IX, Bombardments of Munda and Vila-Stanmore, p. 57.


ships were covered by DesDiv FORTY-ONE, the Nicholas (F), O’Bannon, Chevalier and Taylor.

    Arriving north of Kolombangara at 2348, Task Group Love proceeded to the mining area off Wilson Cove by hugging the western shore of the island. Beginning at 0042 the group laid a three line standard minefield from a point bearing 2690 T., distance 1,100 yards from Wilson Cove entrance for three miles on course 2000 T. In each of the three rows 61 mines were laid, or a total of 183 mines, with no prematures. It is possible, however, that the operation was detected by the enemy.

    While the mining was in progress the covering force, DesDiv FORTY-ONE, was attacked by many Japanese planes near the mouth of Vella Gulf. By occupying the enemy’s attention fully for hours, the covering force enabled the mining group to proceed without interruption.

    The original plan of retirement from the mining operation, a wide turn in succession from the left, had to be abandoned because of inadequate rehearsal and because the center minelayer drew ahead of her position during the mining. CTG LOVE announced over TBS that he would bring the minelayers out by turn movement. All ships acknowledged, and the signal was given to execute the turn.

    A few seconds later the Preble and the Montgomery collided. Both ships were heavily damaged, and only by means of efficient damage control were their commanding officers able to extricate them from enemy waters.

    During the retirement around the northern coast of Kolombangara the formation was not spotted until it was rounding Tuki Point. At this time the Montgomery, which was pushing up a considerable bow wave on account of her damage, was attacked by an enemy plane that dropped a stick of bombs close to her. About the same time a shore battery opened fire on the Montgomery from a nearby position, but did not find her range. The formation escaped further attack by running into a rain storm that prevented a dawn attack from the air.

PT Boat versus Barge

    Motor Torpedo Boats were found to be well adapted to the task of blockade in the Solomons area, which is one of comparatively smooth water and offers many temporary operating positions close to targets. Main bases of operation for the MTB’s in the Central Solomons were Lever Harbor and Rendova Harbor.


    Except when they were kept in harbor because of destroyer operations up the Slot, the torpedo boats were on patrol every night. Their primary task was the interception of enemy barges supplying enemy forces in the Kolombangara-Vella Lavella Area, but many additional tasks were assigned them: transportation of personnel between bases, rescue of the personnel of downed planes, removal of battle casualties, and coastal reconnaissance. The destructive power of these small craft against barges was augmented by the installation of a 37 mm gun in the bow.

    Operating chiefly at night in enemy waters, the MTB’s were subject to more than the usual navigational hazards. The waters around Vella Lavella and Kolombangara, studded with coral reefs, are frequently charted incorrectly if at all. Moving at high speed in strafing runs against barges a number of these craft were run aground and lost. On the night of 6-7 September, PT’s 118 and 172 piled up on a reef on the western side of Tambama Bay while making a run on three enemy barges. PT 172 put 40 to 50 feet of reef astern before it stopped, and PT 118 went on the reef up to the stern. Neither vessel could be salvaged and both had to be destroyed on the reef, though their personnel was rescued. Two more MTB’s were damaged by grounding on the night of 16-17 September.

    Frequently their operations placed the PT’s under fire of enemy shore batteries. Their natural enemy, however, was the float plane, which constantly harassed them and succeeded in damaging four of our boats. One enemy float plane was shot down by PT 156 on the night of 19-20 September, while bombing the boat formation. One tragic mistake was made while four PT’s were strafing enemy barges along the northern coast of Kolombangara at dawn on 30 September. One of three approaching Corsairs failed to receive recognition signals from the boats and strafed PT 126, wounding one officer and killing two men. The PT returned the fire and shot down the Corsair.

    Although action occurred chiefly at night, the PT’s were also useful on day operations. As an example, there was the patrol mission assigned five boats from Rendova on 14 September. Three boats in Section T searched the coves from Doveli Cove to Takisukuru on the northern and western coast of Vella Lavella, while the two boats of Section R covered the northeastern coast of the island from Masiri Bay to Lambu Lambu. Proceeding inside unchartered reefs, and rounding one palm-fringed point after another to push into bays occupied by enemy forces, the PT’s shot up houses,


wharves, and observation stations, and sought out and fired camouflaged barges and cargoes. In all, five unmanned barges were destroyed, three of which were loaded.

    Results of night encounters with barges were usually indefinite. Damaged barges often slid past reefs or into coves, where it was possible to elude detection. Daylight patrols and occupation forces, however, discovered that many of the attacks were more successful than at first believed. During September MTB’s destroyed at least 12 barges, many of them loaded, and damaged eight or ten others.

    Toward the last of the month a new PT base was established at Lambu Lambu Cove on the northeast coast of Vella Lavella.

Air Combat and Patrol

    While the new airstrip at Barakoma was being completed, the airfields of Guadalcanal and Munda sent a series of attacks against enemy shipping and bases. In the meantime, the Japanese airforce displayed increasing unwillingness to risk daylight attacks on our positions. Not a single daylight attack was ventured during the week 5 September through 11 September, during which Air Combat Intelligence reports stated that the enemy air force was “reduced to virtual impotency.” This can probably be explained by the problems facing the Japanese in other areas and by the heavy losses they suffered in June, July and August.16

    The inability of the enemy to cope with our aggressive air tactics for the time being might be illustrated by his failure to provide protection for his barge and supply lines. On 9 September in one day’s strafing missions, Corsairs attacked and destroyed nine Japanese barges. Of these barges, five were destroyed by a flight of four Corsairs in and near Warambari Bay. This is within 80 miles of the enemy air base at Kahili, yet no effort was made to protect the barges.

    Aircraft took a heavier toll of enemy barges than did the PT boats throughout September. A prowling Liberator sank at least two off Choiseul the night of 2 September; two Corsairs damaged three on the 7th, and on the following day Corsairs destroyed two more barges around Kolombangara. Mitchells and SBD’s claimed several barges damaged and sunk incidental to bombing attacks upon bases, but the strafing tactics of the F4F’s were more deadly in this type of work.


16 According to ACI reports, Japanese combat losses in those three months in the South Pacific area aggregated 781 planes, an average of 260 a month.


    Japanese supply dumps and positions on southern Kolombangara, from Surumuni Cove on the east to Ringi Cove on the west, were subjected to almost daily air attack. The airstrip at Vila, the most frequent target, was kept in a very nearly unusable condition throughout September by bombing squadrons. On the afternoon of 4 September, 91 dive bombers, torpedo bombers and medium bombers struck the Vila area and Dulo Cove. Another attack by 9 Liberators followed a few minutes later. On the following day heavy bombers and dive bombers hit Vila again. Just before noon on 7 September, 23 B-25’s bombed the large depot and supply areas east of Ringi Cove, starting fires, and Corsairs followed up at dusk with a strafing attack on Ringi and Webster Coves. On the 8th, 12 Mitchells struck the Vila airstrip, and on the 9th 30 SBD’s, 23 TBF’s and 12 B-25’s blasted Vila gun positions under cover of fighters from Munda. At noon the following day 16 B-25’s attacked the barge and bivouac area west of Dulo Cove, exploding a possible ammunition dump; their attack was followed up a half hour later by 20 SBD’s that strafed the same area. At 1210 on 11 September, 24 SBD’s and 15 TBF’s bombed and strafed enemy positions at Bambari, Parapatu and Surumuni. This was followed fifteen minutes later by a strike at gun positions northeast of Vila by 18 B-25’s.

    During the second half of September, while the Kolombangara positions were not neglected, our air force turned its attention chiefly to Japanese air bases in the Bougainville area, particularly Kahili and Ballale. From 14 to 30 September, 15 air attacks were made on these two bases alone. Much of the bombing was carried out by heavy bombers, of which Army Liberators made up a considerable part. Between 75 and 100 enemy planes and much material were thought to have been destroyed in these raids.

    Results of the strategy of strangulation and blockade became apparent as one enemy position after another was abandoned or mopped up. By the first of September patrols reported that no Japanese appeared to remain on the mainland of Santa Isabel Island, and that the enemy had left behind a large quantity of foodstuffs and ammunition. Indications were that the Japanese were also preparing to evacuate Rekata Bay, once an advance seaplane base for attacks on Guadalcanal. Patrols on Ganongga, south of Vella Lavella, reported on 13 September the destruction of a large enemy bivouac area on the northeast coast of the Island. Living quarters plus a quantity of food, clothing, weapons and other stores had been abandoned.


    At 1445, 21 September, organized enemy resistance ceased on Arundel Island. For three weeks enemy forces, reinforced from Vila across Blackett Strait, had put up bitter opposition to our ground forces. Probably some of the Japanese troops were evacuated to Vila, but 345 enemy dead were counted, indicating that from 500 to 600 were killed in the operation. Our forces now controlled the south side of Blackett Strait, thus virtually closing one more supply route to Vila.

Evacuation of Kolombangara

    On 27 September the airfield at Barakoma was placed in operation. Caught between the jaws of the Allied vise—Munda on the southeast and Barakoma on the northwest—and receiving no strong air or surface support, the Japanese forces on Kolombangara found themselves in a rapidly deteriorating position. Although the extent of enemy forces present on the Island was not known at the time, a prisoner captured later indicated that they numbered about 10,000.

    It was expected that the Japanese would seize the opportunity afforded by the moonless nights of 1 and 2 October for evacuating these forces, and that they would employ barges of various sizes in large numbers as well as combatant vessels before and during those nights. Barge traffic as far as Choiseul was possible under cover of darkness, and high speed surface support from Bougainville would be able to come down the Slot and return between sunset and sunrise.

    In anticipation of such an attempt, our forces, consisting mainly of destroyers, made nightly sweeps of the area. Japanese float planes, however, kept the Tokio Express thoroughly informed of the strength and location of our forces. Apparently under orders to give direct support to the evacuation only when this could be done without serious opposition, the Express presented our destroyer commanders with a cleverly calculated dilemma. The Japanese destroyers would approach cautiously, then retire at high speed to the Shortlands, possibly to decoy our forces away from the vicinity of the barge traffic. Our destroyer commanders had to choose almost nightly between attempting to surprise the Express at the cost of letting the barges escape, or opening fire on the barges at the expense of warning the Express.

    The decision was usually made by the appearance of Japanese snooper planes, indicating that the presence of our force was known; whereupon fire was opened upon barges. Snoopers were almost constantly on the radar screens, and their tactics proved very effective. They would circle


the vicinity of a task force for several hours before dropping flares or attacking. The commander of a task force would be loath to open fire before he was sure that he was observed, lest the success of its mission be prejudiced.

    Since major units of enemy forces might be used to relieve Kolombangara, Task Group MIKE, under Rear Admiral Aaron S. Merrill, whose primary mission was to prevent such reinforcement, included cruisers. It was organized as follows:

Cruisers, Rear Admiral Merrill

    Montpelier (F), Capt. Robert G. Tobin

    Denver, Capt. Robert P. Briscoe

DesRon TWENTY-TWO, Capt. William R. Cooke, Jr.

    Waller (F), Lt. Comdr. William T. Dutton

    Renshaw, Lt. Comdr. Jacob A. Lark

    Cony, Comdr. Harry D. Johnston

    Eaton, Comdr. Edward L. Beck

    This Task Group, under operational control of Rear Admiral Wilkinson, carried out nightly searches north and northwest of Vella Lavella from 22 September to 25 September without making any important enemy contact. On the night of 25-26 September, an additional Task Group joined in the operations north of Vella Lavella. Task Group BAKER, commanded by Capt. Frank E. Beatty, was organized as follows:

Cruisers, Capt. Beatty

    Columbia, Capt. Beatty.

    Cleveland, Capt. Andrew G. Shepard

DesRon TWENTY-THREE, Capt. Martin J. Gillan, Jr.

    Charles Ausburne (F), Comdr. Luther K. Reynolds

    Claxton, Comdr. Harold F. Stout

    Dyson, Comdr. Roy A. Gano

    Spence, Comdr. Henry J. Armstrong 

    Enemy light forces were estimated to be ready to enter the Slot and Vella Gulf and did so on the night of 25 September. We hoped to trap these forces between our two Task Groups. The enemy, instead of falling into this trap, planted a submarine patrol where it could attack the cruisers of Task Group BAKER as they entered the Slot. The Columbia and the Cleveland of Task Group BAKER both reported torpedo wakes in their vicinity. A contact was obtained upon a possible surfaced submarine which was taken under fire, with unknown results. Cooperation between enemy plane and submarine was indicated by the presence of an enemy aircraft which was taken under fire just before the submarine attack. No


damage resulted from the attack, but thereafter cruisers were not risked on such missions. Instead, they were held in readiness as a covering force.

    On 27 September a task unit under the command of Capt. Gillan, composed of the destroyers of Task Group BAKER, Ausburne (F), Claxton, Dyson and Spence plus the Foote was sent on an offensive patrol against submarines and barges. The usual antiaircraft action was necessitated by approaching snooper planes. At 0045 on 28 September contact was made with enemy barges which were taken under sustained and rapid fire with full radar control. At least four separate fires were observed, and it was estimated that three or four of the barges were destroyed.

    On the night of 29 September a task group under the command of Capt. Frank R. Walker on the Patterson, Lt. Comdr. Albert F. White; with the Foote, Comdr. Alston Ramsey; Ralph Talbot, Comdr. Richard D. Shepard; and McCalla, Lt. Comdr. Halford A. Knoertzer, made a sweep of the north coast of Kolombangara Island and into Vella Gulf, crossed the Slot to the south coast of Choiseul Island and returned to the northwest coast of Kolombangara. There was no moon and the night was dark, with a 5/10 sky and frequent rain squalls. Visibility on surface objects was about 3,000 feet.

    At 2230 the Ralph Talbot and McCalla made radar contact bearing 2250 T., distance 4,500 yards, just as the task group was changing course in column to 1400 T. about 18 miles north of Kolombangara. At 2237 the McCalla was detached to develop the contact and left the formation which continued on a southward course. She sighted two enemy barges bearing 2900 T., distance about 3,000 yards, at 2243 and with permission from Capt. Walker opened fire on them. Firing was continued on this and subsequent groups of barges that appeared on the FD Radar. The barges resembled our own LST’s but were about two-thirds their size. Many of them were taken under fire and it was believed that seven to eight were damaged or sunk.

After six minutes of firing it was discovered that the gyro compass repeater system was operating erractically and was temporarily useless, though it appeared to return to agreement with the master gyro after repairs. During the engagement with barges several salvos of three to five-inch with the barges several salvos of three to five-inch projectiles landed fifty yards abeam the McCalla on the engaged side. Upon learning of the possible presence of larger enemy vessels, Capt. Walker at 2258 directed the Task Group to close the McCalla, and ordered that vessel to rejoin the formation. Salvos of large projectiles


continued to land in the McCalla’s wake, and as she was changing course to rejoin, two “extremely heavy explosions”, apparently from aircraft bombs were heard and felt astern.

    At 2304 the McCalla made radar contact on enemy barges located beyond and to the right of the other three destroyers of the Task Group, bearing 608 degrees T., distance about 7,000 yards. The Commanding Officer decided to engage the enemy prior to rejoining his task group as ordered, and communicated his decision to the Task Group Commander. The Group in the meantime was proceeding on a northerly course leaving the enemy on its starboard quarter. At 2309 these ships opened fire on targets to northwestward instead of taking McCalla’s contacts under fire.

    At 2315 the McCalla changed course to 320 degrees to continue closing the Task Group. At this time the targets she had selected, now on her starboard beam, distance 6,000 yards, were seen to open fire in the direction of the Task Group. Tracers from one of the targets indicated weapons of heavier type tan automatic guns, and the Task Group Commander was informed. The McCalla slowed to 20 knots to take her targets under fire. The Task Group then distant between 6,000 and 7,000 yards, changed course to 1200 T. and began closing the McCalla from a head-on bearing of 3100 T. The latter ship changed course to 3000 T. and at 2318 began firing on enemy contacts to the starboard. Several hits were observed on a group of seven or eight barges, and five pips disappeared from the radar screen during the firing.

    At 2320 the Task Group was reported to be approaching within 4,000 yards of the McCalla, bearing slightly on her starboard bow. The McCalla ceased firing at 2321 and turned left to clear her own ships and get in position to fall in astern of the column. Shortly thereafter her Executive Officer reported that the PPI17 scope showed the formation to be dead ahead. The Commanding Officer ordered the rudder put full left and went to the starboard wing of the bridge.

When the smoke of firing cleared, and night vision was regained, he sighted the Patterson 1,500 yards dead ahead and closing rapidly. At this moment he was informed that the bridge had lost steering control. 18 


17 Position Plot Indicator.

18 The Commanding Officer of the McCalla believed that this casualty and the earlier failure of the gyro compass repeater system may have been caused by a combination of the aircraft bombs exploding astern, which caused “considerable shock damage” and the shock of the ship’s own gunfire. He believed the latter caused most of the damage. “It appears to have been unusually intense,” he wrote, “probably due to low angle of fire due to short range and small size of targets.”


 The rudder angle indicator showed full left rudder. The casualty was immediately reported over TBS to all ships, which were instructed to keep clear. Control was shifted to the after steering room which reversed rudder to the 30 degrees right position apparently in an effort to steady on the heading at the time of the steering casualty. Collision with the Patterson now appeared inevitable, and at 2325 the commanding officer rang up full astern.

    At 2326 the McCalla struck the port bow of the Patterson at a point about even with the No. 1 gun. The McCalla lost way rapidly and the Patterson surged ahead, tearing the bow of the McCalla at frame 24. When the latter vessel backed clear and turned, her entire bow forward of frame 24 dropped clear of the ship and sank immediately. The Patterson, when collision became unavoidable, had backed her engines full and had come to full right rudder. The impact of the collision completely demolished and severed the Patterson’s bow at frame 23, flooding the ship forward to bulkhead 36 and flooding the ship below the first platform forward of bulkhead 48. Three men stationed in No. 1 magazine were trapped and killed, and 16 other men suffered minor injuries.

    Prompt and effective damage control measures enabled the two vessels to continue under their own power. The Ralph Talbot and the Foote circled the damaged ship, engaging and driving off enemy planes. Although all four ships remained in the area until daylight, no further barges were encountered. It was believed that the attempted evacuation was broken up for the night.

    On the following night Japanese attempts at evacuation were again broken up by our destroyers. The intercepting force on the night of 30 September was under command of Comdr. Alvin D. Chandler, ComDesDiv FORTY-TWO, in the Radford, Comdr. William K. Romoser, with the Saufley, Comdr. Bert F. Brown, and Grayson, Lt. Comdr. Henry O. Hansen. Initial contact with enemy barges was made at 2107 at a position 13 miles bearing 3440 from Tuki Point, Kolombangara. Much difficulty was experienced in maintaining contact by radar, and visibility was so low that the Radford was directed to illuminate with star shells. The barges, approximately 12 in number, immediately dispersed. The destroyers picked individual targets and proceeded to sink them. Of the 12 targets, six were destroyed, three were unaccounted for, and three fast craft escaped in the direction of Kolombangara.


    As anticipated, the attempt at complete evacuation came on the moonless nights of 1 and 2 October. On the first of these nights our striking force was composed of destroyers under command of Capt. Cooke, ComDesRon TWENTY-TWO, in the Waller, with Renshaw, Eaton, and Cony, and Comdr. Chandler in the Saufley with the Radford, Grayson and LaVallette (Lt. Comdr. Robert L. Taylor). This force was supported and covered by Task Group MIKE (Admiral Merrill in the Montpelier with the Denver and the four destroyers previously assigned to Task Group BAKER—the Ausburne, Claxton, Dyson, and Spence). At 2204 the cruisers and ships of the covering force were ordered to retire to the east and Capt. Cooke’s destroyers proceeded under orders to rendezvous with Comdr. Chandler’s group north of Kolombangara.

    It was soon apparent that there would be no lack of targets this night. The two groups made contact with enemy barges independently about the same time, before the rendezvous took place, and began operations independently. During the night the striking force encountered five groups totaling an estimated 35 barges. They were of various types ranging in size from 200 to 75 feet in length. The silhouette of the larger ones was that of the LST cut down to a very low freeboard.

    Firing with radar control proved unsatisfactory because of the size of targets and confusion created by splashes of projectiles. Salvo fire, with star shell illumination, however, proved quite effective. With several ships operating together, it was impossible to evaluate results of gunfire accurately. CINCPAC arrived at the estimate of 20 barges sunk during the night, besides an unknown number damaged. The water was thick with bodies. Comdr. Chandler was convinced, after examining the water in which barges were sunk, that his group alone killed almost 1,000 Japanese.

    At 2304 a Black Cat reported three enemy destroyers off the northwest coast of Vella Lavella on a southeasterly course. They showed no disposition to close and protect their barges. Our destroyer groups broke off engagement with the barges, however, joined up and proceeded north. Many barges contacts were passed by during the maneuver and undoubtedly a large number escaped. The search for the enemy destroyers proved unsuccessful.

    Our destroyers were constantly harassed by enemy snooper planes, which dropped the usual flares and float lights as well as a few bombs.


At 2339 three bombs landed 20 to 25 feet off the Saufley’s port side. Some slight material damage resulted and 18 of her personnel were injured, two of whom later died. Although this striking group remained in the area until 0430 no further casualties were suffered.

    The intercepting force on the following night was composed of Capt. Cooke, ComDesRon TWENTY-TWO, in the Waller with the Eaton and Cony and Comdr. Harold O. Larson, ComDesDiv EIGHT, in the Ralph Talbot with the Taylor, and Terry (Comdr. George R. Phelan). There was no covering force of cruisers on the night of 2-3 October, as there had been on the previous night.

    Again our destroyer commanders were confronted with the choice of pursuing enemy combat units or confining themselves to barges. One group had already opened fire on barges when radar contact was established on a group of five to seven targets at first believed to have been destroyers.19 These targets were in two groups in latitude 070 46’ S., longitude 1560 59’ E., proceeding at 25 to 28 knots and maneuvering rapidly.

    The Talbot, Taylor and Terry opened the attack at 2325 by firing high speed torpedoes at 3,500 yards, target angle 0900, speed 28 knots. Some observers noted geysers of water among the targets shortly after, and three underwater explosions were heard, but no torpedo hits were claimed. After retiring to evade enemy torpedoes, our destroyers reformed, closed to 8,000 yards and opened gunfire. The Terry scored repeated hits on one target which slowed to almost zero speed, separated from the formation, and disappeared from all radar screens. It was believed sunk. The enemy returned several salvos, then retired at high speed in a northwesterly direction.

    Capt. Cooke’s destroyers then took up the chase at a speed of 32 knots, which was maintained until 0020, when it became apparent that the enemy could not be closed for effective gun or torpedo fire. It was also suspected that our destroyers were being drawn off to permit the passage of barge traffic, a suspicion that seemed verified by later events. Pursuit was abandoned and the destroyers returned in search of barges.

    Our destroyers soon engaged large numbers of barge groups and fired upon them intermittently from 0028 until 0224. They were of several sizes, some with canoes and rowboats in tow. An estimated 20 of the


19 These were variously identified as “motor-torpedo craft,” “gunboats” and “old destroyers” of the Wakatake class (820 tons). Reports agree that they were from 600 to 1,000 tons displacement.


large barges were sunk, as well as many damaged and an unknown number of small craft destroyed. Few were thought to have escaped.

    Undoubtedly the mass evacuation of Kolombangara had been attempted on the nights of 1-2 and 2-3 October, although some traffic continued on the two nights following. Comdr. Chandler in the Nicholas, with the Chevalier and O’Bannon, sank two more barges six miles northwest of Tuki Point, Kolombangara on a night sortie 3 October. A few hours later, at 0057, 4 October, the same destroyers sank a Japanese gunboat of about 65 tons 15 miles south of Luti, Choiseul. The following day, four Corsairs and four Airacobras strafed and destroyed 16 enemy barges between West Cape, northwest Choiseul, and Pearl Island, off central Choiseul.

    Officer patrols returning from Kolombangara on 4 October reported that the enemy had evacuated Voluna and Hunda, on the southwestern coast, on 26 and 27 September. Many Japanese helmets and rifles were found abandoned along the Vila-Hamberi road. The area between Gatere and Sambira was reported clear of the enemy. On 5 October coast watcher patrols found the Vila and Stanmore area completely evacuated, and on the 7th Bambari Harbor on the east coast was reported clear. On the 8th patrols moved up the east coast to Surumuni Cove and up the west coast beyond Hunda Cove. The Vila patrol found a large amount of abandoned equipment, including artillery, AA guns, ammunition, and eight to ten demolished planes.

    A Japanese prisoner captured on the 8th indicated that 10,000 Japanese had been evacuated between 28 September and 2 October—3,000 to Choiseul by barges and the remainder to Rabaul by destroyers. During the height of evacuation our forces destroyed some 60 barges besides damaging many more. 


Military Situation

    While the evacuation of Kolombangara was in progress, the campaign to clear Vella Lavella of the enemy was approaching a climax. The final phase of military operations differed in several respects from earlier phases. Whereas the object had originally been to secure a perimeter, it now became the purpose to exterminate the enemy as quickly as possible. The complete security of the island had become necessary for the rapid development of offensives further north.


    On 18 September elements of the 3rd New Zealand Division arrived to relieve the 35th Regimental Combat Team. Maj. Gen. H. E. Barrowclough, Divisional Commander, assumed command of the Northern Landing Force, and Brigadier L. S. Potter, commanding the 14th New Zealand Brigade Group, was placed in charge of the operation. Brigadier Potter decided upon a plan to pocket the Japanese in the northwest corner of the island by moving two of his Battalion Combat Teams simultaneously up the east and west coast. The operation was to be semi-amphibious in nature, with the use of LCVP’s and LCM’s to transport personnel and supplies from bay to bay in a series of bounds.

    The plan was put into operation on 21 September. Retiring slowly to the west, the 500 to 700 Japanese were reported to be tired and poorly armed. Captured documents indicated a low state of morale and a general hope for evacuation. In view of these circumstances the stubborn resistance which the New Zealanders met came as a surprise. Harassed by small parties of machine gunners and snipers, they were for a period reduced to progress of 300 to 600 yards a day.

    By 5 October, however, the 35th Battalion Combat Team had reached Marquana Bay on the west and the 37th Battalion Combat Team had arrived at Warambari Bay on the east. The Japanese were hemmed in upon the narrow strip of land between, with fire of two field batteries, mortars, and machine guns registering in this area from both sides. A prisoner captured during the afternoon said that the 500 remaining Japanese were well organized but realized that they were trapped. They were short of food, tired of fighting, and many wished to surrender, but were prevented by their officers. General Potter decided that the time had come to close the pincers.

Preliminary Movements

    This was the situation when on 5 October Admiral Wilkinson decided to send a force to intercept the Tokio Express which might be expected either to complete the evacuation of Kolombangara or to evacuate the forces cornered in northwest Vella Lavella.

    Of the nine destroyers available to Admiral Wilkinson at the time, six were already committed to escort a convoy of LST’s and APD’s leaving Guadalcanal the morning of the 6th. The three remaining destroyers, commanded by Capt. Frank R. Walker, were sent up the Slot to break up the evacuation. This force was composed of the following ships:


        Selfridge (F), Comdr. George E. Peckham

        Chevalier, Comdr. George R. Wilson

        O’Bannon, Comdr. Donald J. MacDonald. 

No major contacts were made that night. The following morning the same force was ordered to continue the patrol along the southern coast of Choiseul. 

    About 1800, 6 October, more definite word regarding the Japanese evacuation forces reached Admiral Wilkinson from Admiral Halsey. This force was reported to consist of an unknown number of destroyers, three PT’s and six SC’s which could be off northwest Vella Lavella by 2230. 

    Three destroyers were immediately detached from the convoy then underway south of New Georgia and ordered to rendezvous by 2300, ten miles west of Sauka Point, Vella Lavella, with Capt. Walker’s destroyers. This group, under command of Comdr. Harold O. Larson, consisted of the following destroyers: 

        Ralph Talbot (F), Lt. Comdr. Richard D. Shepard

        Taylor, Lt. Comdr. Benjamin Katz

        LaVallette, Lt. Comdr. Robert L. Taylor 

    If this rendezvous could be effected in time, of which there was some doubt, it was believed our forces would be sufficient to defeat any probable enemy strength. 

    Capt. Walker, after receiving instructions regarding the rendezvous, was later informed by Admiral Wilkinson that there were possibly nine destroyers in the Express, which was estimated to have passed through Bougainville Strait at 2000. Further information was received that Comdr. Larson’s group would not arrive at the rendezvous until 2340. It was understood that if the rendezvous was delayed, Capt. Walker’s group was to intercept the enemy by itself. 

    About dusk two squadrons of B-25’s and 16 P-38’s attacked Kahili. From them came a report that four destroyers or possibly light cruisers had been sighted at 1830 five to ten miles north of Torau Bay headed south. 

    At 1900, when Capt. Walker received orders concerning the rendezvous, his force was ten miles south of Kakasa, Choiseul, on course 3500. Not long thereafter snoopers were picked up by radar, and from that time until after the ensuing action from one to three bogies were on the screen continuously. It was evident that they were tracking our force persistently, though they remained at ranges of from 8,000 to 10,000 yards,



occasionally closing to 5,000 yards. Frequent maneuvers by column and turn, as well as changes of speed and seeking cover under rain squalls, were unavailing in the effort to shake off the snoopers. At 2121 they began dropping float lights, indicating by this means the heading of our destroyers. 

    A bright half moon was overhead and visibility was excellent, ranging from 13,000 to 20,000 yards. There were a few scattered squalls. The sea was calm and smooth, with a long swell running. The wind was westerly, force one. 

    At 2226 Capt. Walker of the group made the first of several efforts to call Comdr. Larson’s destroyers by TBS. He got no answer. 

Contact and Approach 

    At 2230, when Capt. Walker’s group was northwest of Vella Lavella, in approximately latitude 070 32’ S., longitude 1560 24’ E., speed 25 knots, course was changed to 2100 for the rendezvous, then about seven miles distant. Immediately after the change of course, at 2231, a surface radar contact was made almost simultaneously by all three ships. 

    The radar indicated two enemy groups. Group I was to the left of the other, bearing 2770, distance 19,500 yards. Group II was to the right and about a mile to the north of Group I, distance also about 19,500 yards, bearing 2810. Tracking was begun, and an urgent contact report sent out to Admiral Wilkinson, but no acknowledgment was received. Both groups of enemy ships appeared to be changing simultaneously from an easterly course through north to west. Having apparently located our force, the enemy was maneuvering accordingly. 

    At 2234 the Selfridge sighted the enemy ships on the horizon off her starboard bow. As visual and radar indications became clearer enemy Group I was seen to consist of five ships in column, four of which were destroyers, and one somewhat larger, possibly a cruiser, in the middle position. Group II was composed of four destroyers. Although Comdr. Larson’s destroyers could not be expected in the area for another hour, and further efforts to call the group by TBS were unsuccessful, Capt. Walker immediately closed with his three destroyers for an attack upon the nine enemy ships. The order of his ships in column was Selfridge, Chevalier, and O’Bannon, keeping station at a distance of 500 yards. 

    The two enemy groups took a southwesterly course at about 2240, Group II on course 2300, speed 20 knots, and Group I some 2,000 yards nearer and to the left on course 200 degrees, speed 26 knots. It quickly became apparent


that the larger and faster Group I was interposing itself between our force and their Group II—presumably their evacuation group—which kept farther away, though without definitely retiring. Group I continued in a wide swing around the bow of our advancing column, crossing from starboard to port. At 2245 this group was distant 14,500 yards, heading 1730, speed 27 knots. 

    Capt. Walker, changing course several times to intercept Group I, turned to the left at 2245 on course 2400 T., speed 30 knots, thus bringing the enemy again on his starboard bow. The head-on approach course was continued for the next ten minutes. 

    In the meantime at 2243 the Chevalier reported radar contact on small enemy craft bearing 3360 T. at 13,000 yards, and two minutes later another contact on two craft bearing 0060 T. at 12,500 yards. These craft were thought to be motor torpedo boats, and Capt. Walker ordered that they be watched and fired upon if they approached. At the same time he reported a small “pip” bearing 0060, distant 4,060 yards retiring north at over 30 knots. This was tracked as far as 8,000 yards by which time attention was concentrated on the approaching destroyers. At 2243 the Selfridge also reported a radar contact bearing 0080, distant 10,700 yards, possibly a PT boat. The O’Bannon sighted a PT boat several thousand yards away on her starboard beam at 2246, and personnel on deck believed they saw two PT boats. All these reports and sightings came in the tense and crowded minutes preceding engagement with the main enemy force. Until 2255 our destroyers were at a disadvantage, in that the enemy was in an excellent position for firing torpedoes, while our own tubes could not be brought to bear. All attention was thus concentrated upon enemy Group I. The PT contacts were not taken under fire, and the CIC was absorbed in plotting the course of the enemy destroyers. 

    At 2252, when almost dead ahead of our column, distant about 10,000 yards, enemy Group I turned left on course 1150 to sweep down our port flank. Capt. Walker was faced with the alternative of turning right to engage on a reverse course. The latter choice would shorten the engagement with the five ships of Group I, but would possess the advantage of taking our group to a position where it could engage Group II. Also, Comdr. Larson’s destroyers, expected in the area at 2340, might possibly arrive in time to engage Group I when it was left behind.


Action with Group I 

    2255          Task Group fires torpedoes.

    2256 ½      Task Group opens gunfire.

    2259          Chevalier contacts two enemy PT’s on starboard beam.

    2301          Chevalier is hit by torpedo.

    2303          O’Bannon rams the Chevalier.

    2305          O’Bannon’s target blows up and sinks.

    2310          Selfridge’s target dead in water and burning, sinks at 2326. 

    Capt. Walker at first announced over TBS that he was turning left to fire torpedoes, but immediately countermanded this saying he was turning right to fire. At 2255 he came right to course 2700 and executed the signal to fire a half salvo of torpedoes. The three destroyers immediately fired, the O’Bannon inadvertently firing six instead of five torpedoes. Enemy Group I was plainly visible, range about 6,000 yards from the Selfridge. The base torpedo course of the Selfridge was 1750 T., the average torpedo run between 5,500 and 6,000 yards. A good solution was thought to have been secured upon Group I, which continued to advance upon a steady course. Evidence regarding the effectiveness of this fire is inconclusive, although several explosions were felt and observed by the three vessels about the time for the torpedoes to have reached their targets. 

    The three destroyers continued on the same general course with shallow zigzag maneuvers after firing. The signal to open gunfire was executed about 1½ minutes after that for torpedoes. Since they were on a course opposite to that of the enemy, the Selfridge opened fire on the rear enemy ships, the Chevalier on the fourth, and the O’Bannon on the third. Hits were scored immediately and we continued to hit during the next six minutes of furious engagement with Group I. 

    The Selfridge believed her target, the fifth ship, was hit by a torpedo. The enemy vessel turned right and slowed, and a few minutes later was afire. Approximately 100 rounds were fired at this target and many hits were observed. The first and third enemy ship returned to lay a smoke screen around the burning target. At 2310 the damaged ship was dead in the water and burning furiously. The circling destroyers abandoned her in a few minutes. She continued to burn until about 2326, when she blew up with a terrific explosion. 

    The fourth enemy ship, fired on and hit by the Selfridge and the Chevalier, was reported to have disappeared from sight and from the


radar screen about the time our torpedoes arrived. Although no fires were seen, there was a dense column of smoke over this vessel before it disappeared from sight. She was thought to have been hit by one or more torpedoes. 

    The O’Bannon opened fire on the third and largest ship in the enemy column. The Commanding Officer described his target as “long and low in the water with an inverted Y stack, spilling aft,” and believed her to be a Yubari class light cruiser, though CINCPAC later identified her as the destroyer Yugamo. This vessel turned left between the two opposing columns about 2251; approaching to a much closer range than the four other ships. At 2256 she reversed course sharply to the right. This caused the O’Bannon’s fire to cross that of the other destroyers. Hit with many salvos, the O’Bannon’s target burst into flame four minutes after gunfire was opened. Five minutes later she was dead in the water. At 2305 she blew up with an explosion so violent that it was seen by ships of Comdr. Larson’s group, still many miles distant from the area. This ship was listed as “definitely sunk.” The two remaining ships of the enemy column were possibly hit, but not badly damaged. 

    Only the O’Bannon reported any return gunfire from the target in this phase of the action, and that was ineffective, the salvos landing slightly astern and 600 yards over. 

    The Chevalier had just been directed at 2259 to prepare to illuminate the targets of the Selfridge in Group II, when the CIC informed her commander that two small high speed targets were approaching on the starboard beam. Believing that he was not justified in taking his main batteries off the destroyer targets, Comdr. Wilson swung left to bring his 1.1-inch machine gun to bear on the small craft. At 2300 he observed the Selfridge turning right to engage Group II and ordered “right full rudder.” 

    At 2301 or 2302 while on course 3000 T. the Chevalier was hit by a torpedo.20 Two concussions followed, the explosion of the torpedo, and almost simultaneously, what the Commanding Officer believed was the explosion of No. 2 Gun Magazine. All personnel on the bridge were stunned. Upon regaining consciousness, Comdr. Wilson learned that the entire bow forward of the bridge had been blown off and the forward motion of the after section was threatening to submerge the ship “like a


20 For discussion of direction and origin of the torpedo attack see “Notes on the Japanese Torpedo Attack in the Battle of Vella Lavella,” p. 61.


submarine.” All communications were out and it was too late to warn the O’Bannon by blinker tube. 

    The ships had been ordered to stay closed up, and Comdr. MacDonald of the O’Bannon noted shortly before 2300 that he was less than 500 yards from the Chevalier, next in column. When he saw her explode, he immediately ordered hard right rudder and checked fire, intending to clear the torpedoed vessel on his port hand. Completely out of control, however, the after section of the Chevalier also continued to swing right, closing rapidly. Comdr. MacDonald ordered emergency full speed astern and passed the word to stand by for a collision. 

    At approximately 2303 the O’Bannon rammed the Chevalier in the after engine room on the starboard side. Slowed by backing engines and stopped by the Chevalier’s starboard shaft, the O’Bannon penetrated only 10 to 12 feet. The collision stopped the Chevalier’s starboard shaft and flooded her after engine room, thus stopping her port shaft. This halted the diving motion and, according to her commanding officer, probably prevented her from going under immediately. 

    The bow of the Chevalier forward of frame 55 had broken off with the explosion and submerged immediately except for a small part of the bow that remained above water. After the collision all light and power were lost. The forward fireroom bulkhead was damaged and the fireroom was secured after the water reached the burner level. The fuel oil suction lines aft had been cut and the after engine room was flooded above the upper gratings and rendered untenable. The bridge was a shambles. An appreciable list to the starboard left only six inches of freeboard on that side, with the main deck frequently awash. 

    While making preparations to abandon ship, Comdr. Wilson took time to order the torpedoes in tube set on low speed, the gyro angle set at zero, and the tube trained on the blazing enemy vessel slightly abaft his port beam. He directed that one torpedo be fired every 30 seconds at that target as it lay dead in the water circled by two destroyers. The burning ship blew up shortly after the Chevalier fired her last torpedo, but whether as the result of the torpedo attack is not known. 

    The principal damage to the O’Bannon was the sharp deflection to the starboard of her bow below the first platform and forward of frame 10. A secondary damage was the deflection of the entire bow forward of frame 30 slightly to the starboard shell from main deck to keel. All equipment seemed to be operating, however, and the ship after backing clear of the


Chevalier went ahead at one-third speed to screen the damaged ships, which by then included the Selfridge. 

Selfridge Engages Group II

    2259                  Selfridge checks fire, turns right on course 300 degrees. Enemy Group II turns right on course 330 degrees and opens fire.

    2300                  Selfridge opens fire on enemy destroyers.

    2304-2306         Torpedoes sighted coming from three directions.

    2306½              Selfridge hit by torpedoes.

    2310                  Communications established between Capt. Walker’s Group and Capt. Larson’s supporting group.

    2314                  Two remaining enemy destroyers of Group I retire to northwest. 

    It will be recalled that shortly before the Chevalier was torpedoed, the Selfridge ceased firing on Group I and turned right to engage Group II. At about the same moment (2259) the ships of enemy Group II made a radical turn to the right, course 3300, and assumed an echelon formation. Only three ships were sighted in this formation. The fourth was hidden in a rain squall, but was revealed by radar to be about two miles north of the other three ships, making slower speed. 

    The two rear enemy destroyers opened fire on the Selfridge as they turned. All of their salvos missed badly, falling 2,000 to 4,000 yards short. It was concluded that these ships did not have radar. 

    The Selfridge, now on course 3300 T., speed 30 knots, trained her batteries on the right hand destroyer on her port bow. At 2300 she commenced firing on this target, a destroyer with two raked stacks, bearing 2580 T., distance 10,600 yards. The Selfridge fired about 180 rounds of 5-inch common at this target, full radar control being employed initially, then partial radar control, salvo fire. Observers reported seeing shells hit and heavy smoke billow from the Japanese vessel. Range was approximately 11,000 yards when fire was checked. It is believed this ship was damaged, perhaps seriously.21   

    The Selfridge continued to engage and pursue the superior enemy force alone even after 2303, when the damage to the Chevalier and the O’Bannon was reported. 

    The Selfridge shifted fire to a second enemy destroyer, also a two stacker, about 500 yards to the rear and left of the previous target. Partial


21 A search plane, at 0607 the following morning, reported floating debris at latitude 070 39’ S., longitude 1560 04’ E., which might have come from this ship. Upon this evidence Capt. Walker believed she was “possibly sunk.”


radar control, salvo fire was again used, and some hits were observed. The Selfridge expended approximately 90 rounds of 5-inch common upon this target, which was probably damaged. 

    The final target was the ship that had fallen behind the other three, and at this stage of the action was said to be about 2,500 yards from the previous target and far enough to the north to be slightly on the starboard side of the Selfridge. Reports of her position and range, however, are confused. This was probably due to the fact that all observers were fully occupied in spotting torpedoes before and during the time this target was under fire. 

    At 2304, when the Selfridge was on course 3200 one torpedo was seen porpoising about 2,000 yards on the port beam on an approximately parallel course. This torpedo was thought to have been fired from Group I. At about 2305 two torpedoes were seen approaching about 300 on the port bow. The captain ordered left full rudder, but immediately countermanded this order because of a report of a torpedo close aboard on the starboard bow, and ordered right full rudder. Speed was changed to 25 knots. He then rushed to the starboard side, but could see nothing because of other personnel crowding the bridge windows. Hurrying back to the port forward bridge windows, he saw a torpedo wake crossing the bow from the starboard and going down the port side about 25 yards away. Another torpedo wake was seen about 400 yards farther away at angle showing it was fired from a position on the starboard bow. The ship had swung around about 300 by this time and was heading about 0100. No other torpedoes being reported on the starboard, left full rudder was given and an effort was made to locate the torpedoes previously reported approaching on the port bow. The Selfridge had swung left to about 3350 T. at 2306½, when she was hit on the starboard bow at about frame 40. At about the same instant she suffered “another very probable torpedo hit” on the port side.22 Her bow was sheered off and everything forward of the bridge was wrecked. All power was lost temporarily, and only the TBS receiver was still in operation. 

    This ended the action. Only 11½ minutes had elapsed from the moment our group fired torpedoes to the torepedoing of the Selfridge. At the end


22 For a discussion of the probable origin of the torpedo attack on the Selfridge see “Notes on the Japanese Torpedo Attack in the Battle of Vella Lavella,” p. 62. For the extent of damage, see photographs following page 26.


of the action the Chevalier was torpedoed, rammed and sinking; the O’Bannon was limping with a badly damaged bow, and the Selfridge had her bow sheered off. Yet this crippled force of three destroyers was left in command of the field at the end of the battle. Except for two destroyers of Group I that continued to circle a burning enemy ship, those of the nine enemy ships that had not been destroyed or disabled had been put to flight. Apparently warned by snoopers of the approach of our supporting force, the two remaining enemy destroyers retired at 2314 and were tracked fleeing to the northwest. Group II was also withdrawing at best speed. 

Arrival of Supporting Group 

    A serious disadvantage throughout the engagement was the failure of communications between Capt. Walker’s group and the supporting group under Comdr. Larson. Repeated efforts to establish communication between the two groups over TBS failed until 2310, when the O’Bannon achieved contact with the Taylor

    At 2300, while the supporting group was still nine miles south of the rendezvous, gunflashes of a full scale night action could be seen to the north. In the next 10 minutes of approach, however, while the battle was at its climax, Comdr. Larson could not distinguish friend from foe. Within a few minutes he sighted the two destroyers to the north that were circling the burning ship, but he could not be sure at that time that they were enemy vessels. 

    About the same time, at 2312, Comdr. MacDonald of the O’Bannon, to the north of these targets, ordered his torpedo battery to stand by to fire on the circling destroyers. A good solution was obtained, but fire was held because Comdr. Larson’s group, directly south and beyond the targets, might have been in the line of fire. While both groups hesitated, the enemy destroyers opened range and retired before firing could be commenced. 

    At 2326 the Chevalier reported that she was unable to remain afloat. After attempting for about 20 minutes to go alongside, the O’Bannon gave up the attempt because of the sharp rudder effect of her damaged bow.23 While the maneuvering was in progress a Japanese plane dropped a bomb close to the Chevalier and about 1,500 yards ahead of the O’Bannon. Earlier, just after the collision of the two ships, a plane had dropped a float


23 See photograph following p. 26.


light close aboard which was now of great assistance in locating survivors. When the ships were within 50 yards of each other, boats were lowered to bring off the wounded. Remaining survivors, floating on life rafts, were towed to the ship. The O’Bannon picked up 16 officers, and 234 living and seven dead enlisted men. Total casualties of the Chevalier as reported by Capt. Walker were: 1 officer missing, 2 officers (Comdr. Wilson and his Executive Officer) wounded, and 53 enlisted men dead or missing. Most of the injured had fractured legs and arms and a few suffered from fuel oil poisoning. 

    After leaving two motor whaleboats in the water for the use of survivors who might have escaped observation, the O’Bannon retired alone at 0130. She proceeded west of Simbo Island and through Blanche Channel, making an engine speed of 13 knots, and arrived at Tulagi without further incident. 

    In the meantime Comdr. Larson reported to Capt. Walker for duty, and was directed to lead his destroyers toward Marquana Bay in search of enemy barges. Our vessels were subjected to an attack from the air that caused no damage. No barges were found. 

    Returning to the area where the Selfridge and Chevalier were hit, the LaVallette from Comdr. Larson’s group made an unsuccessful search for more survivors. Since there appeared to be no chance of saving the Chevalier, the LaVallette was ordered to remain behind after the other vessels had retired, and sink her. After a careful search of the wreck by a boarding party, the LaVallette drew off to 2,100 yards and fired one torpedo. The Chevalier blew up with an explosion that sent a cloud of black smoke towering over 500 feet. Her severed bow was located about a mile to the west and sunk with depth charges. After a further search for survivors and possible sections of the bow, the LaVallette headed south before dawn. 

    The Selfridge in the meanwhile discovered the full extent of the damage she suffered from the torpedo attack. Flooded aft to frame 67, the ship was considered a total loss forward of about frame 55. In addition considerable blast damage occurred aft, varying with the distance from the center of explosion. Her engines appeared to be undamaged except for vibration, however, and her guns were kept manned. After torpedoes and other weights were jettisoned and the forward fireroom bulkhead shored up, it appeared that she would float and could make slow speed. At 2340 a small craft, thought to be enemy, was seen to approach about 3,500 yards and then retire, but no further enemy contact was made.


Between 2358 and 0040 the Selfridge backed slowly toward the southeast on a retiring course. 

    At 0041 the Taylor came alongside and all wounded, in addition to 14 officers and about 161 men, were transferred to her. The casualties of the Selfridge included 13 enlisted men dead, 36 men missing, and 11 including one officer, wounded. 

    The Selfridge then proceeded at five knots, slowly working up the RPM for 16 knots as she proved capable of better speeds. She retired to the southeast, screened by the Ralph Talbot and Taylor and later joined by the LaVallette. Fighter plane cover had arrived overhead by 0515. At 0715 radar contact was made with many enemy planes coming in for an attack. Skilful direction by the Taylor’s fighter director unit prevented any planes from getting within five miles of the ships. The group reached Tulagi in safety at 1300 on 8 October. 

Enemy Losses 

    The evidence regarding enemy vessels destroyed or damaged in this action is still inconclusive. The only sinking positively established is that of the destroyer Yugumo. The day following the action PT boats picked up 78 Japanese survivors, all of whom said they were from this ship. According to them, she was sunk by a torpedo after severe damage by gunfire. 

    Interrogation of the survivors indicated that the Kazegumo and the Akigumo were present with the Yugumo. All these ships were described as “special class DD’s” and were said to be acting as escorts for six other DD’s which left Rabaul for Vella Lavella. Other enemy ships present were identified as the Samidare, Shigure, Fumikaze, and Isokaze. The last three named were reported to be acting as transports and carrying barges for evacuation purposes. It was reported that eight enemy vessels were seen retiring toward Rabaul, although it is possible that other ships may have joined those surviving the action. 

    All reports by eye witnesses concur that several other ships were severely damaged and some possibly sunk. Comdr. Larson, although arriving too late to take part in the action, saw three ships from enemy Group I burn, explode, and sink. An officer from the Staff of the Commander Destroyers Pacific Fleet, aboard the the Selfridge during the engagement, confirms this impression of enemy losses.


End of a Campaign 

    New Zealanders of the 35th and 37th Combat Teams, waiting through the night of 6-7 October to shut the jaws of their pincers upon the Japanese the following morning, heard plainly the guns of the naval battle to the north. When dawn came, they moved forward for the final mop-up around the Bay. Patrols from both Teams returned reporting no contact whatever. 

    It was under cover of the naval action that the evacuation of the last Japanese troops on Vella Lavella was accomplished. Some 400 men were thought to have been taken off the island during the night. A more thorough investigation of the area by patrols of both combat teams was made on the 8th, but only dead Japanese and abandoned equipment was found. At 0900 on 9 October the leading companies of each combat team met, and the commanding officer signaled headquarters that resistance had ceased and his task was completed. 

    Considering the important strategic advantages gained by the campaign thus ended, the cost was quite small. An important offensive base had been gained at a cost of fewer than 150 killed. The success of the Vella Lavella campaign can be assessed from the fact that the airstrip was in use six weeks after the original landing. Within two months it could accommodate almost 100 planes, and was in full operation against Bougainville. Thus the next campaign was opened almost simultaneously with the closing of the previous one. 

    Perhaps of equal or even greater importance than the strategic advantages gained by the campaign of Vella Lavella and Kolombangara were the lessons learned. The principle of seizing unoccupied territory for the development of an airfield, soon to be repeated in the Bougainville operation, was worked out in this campaign. The success of the operation clearly demonstrated the soundness of the strategy of by-passing enemy strongholds, then blockading and starving them out. Although it might appear superficially that this plan was developed in the Aleutians, actually the factors controlling decisions in that campaign were of a different nature. This operational pattern, later repeated so often in the Central and Southwest Pacific, was rehearsed successfully for the first time at Kolombangara and Vella Lavella.


Wrecked Zero on Vila Airfield

Battle of Vella Lavella

Solomon Islands area

Kolombangara after our landing



    The circumstances under which the Chevalier and the Selfridge were torpedoed raise a number of difficult problems. “Nothing was seen, visually or by radar,” writes CINCPAC, “which we can definitely point to as having hit the Selfridge, and this holds also for the Chevalier.” CINCPAC adds to the multiplicity of theories advanced by the action reports by suggesting the possibility that the ships may even have been hit by a torpedo plane, since several enemy planes were overhead at that time.

    Comdr. Wilson was strongly of the opinion that the Chevalier was hit on the starboard bow and that enemy torpedo boats were responsible for the hit. “My decision not to take the fast closing small targets under fire was wrong,” he wrote. “I should have immediately fired on them, since I was not being hit by enemy shell fire. We could probably have sunk them and then returned to the kill unharmed. The PT boats definitely got us since we were hit on the starboard bow, I am sure, as is almost everyone else.”

    Comdr. Wilson offered another suggestion, however, which would indicate a different explanation of the torpedo attack. “We should have turned away at least 90 degrees after we fired torpedoes,” he wrote. Had that been done, “we could have inflicted as much damage as we did and ourselves remained uninjured.”

    Capt. Walker in his final report on the battle did “not entirely agree” with the theory that the attack came from the torpedo boats. In his opinion, “visibility was sufficiently good to permit visual observation of motor torpedo boats traveling at high speeds if they had closed for an attack.” The two pips picked up by the O’Bannon’s SG operator shortly before the Chevalier was torpedoed were at first evaluated as low flying aircraft. The pips were on the screen only a short time. Deck personnel did report sighting torpedo boats, however, and the Executive Officer concluded that these were the pips spotted. The CIC of the Chevalier seems to have had no doubt concerning the approach of two small craft on the starboard beam about two minutes before being torpedoes, but these craft were not reported to have been sighted.


    It was the opinion of both Capt. Walker and Comdr. MacDonald that the Chevalier was torpedoed on the port side, not on the starboard as Comdr. Wilson believed. This is also the conclusion of studies made by the Bureau of Ships.24 No PT boats were indicated by radar or sighted on the port side. No torpedo wake was reported to have been seen on either side.

    The most reasonable assumption seems to be that the Chevalier was struck by a torpedo from one of the five enemy destroyers in Group I. The opposing columns passed abeam of each other on reverse courses at between 5,000 and 6,000 yards. A salvo from five Japanese destroyers would have covered a wide area. Considering the record of accuracy the Japanese have established in the use of the torpedo, and the limited evasive action taken by our group, it was only reasonable that some hits might be expected. In the opinion of CINCPAC the destroyers of Group I rather than the torpedo boats hit the Chevalier, “because it is almost incredible that the former should not have hit somebody.”


    Comdr. Peckham of the Selfridge was of the opinion that torpedo boats were the source of the successful attack on his ship. According to him, “the failure of the CIC on all three destroyers to keep track of the possible torpedo boat groups entailed the torpedoing of this vessel.” No word of the movements of the small craft reached him after he was told of the initial contact on two groups of possible torpedo boats made by the Chevalier and the Selfridge at 2243. He believed that these two groups could have reached a firing position on both sides of Selfridge in time to fire the torpedoes that hit her. On the other hand, he was convinced that his ship “was not torpedoed by either group of Japanese destroyers.”

    Any discussion of the source of the torpedo attack should be prefaced by the reminder that instantaneous estimates of the direction and position of torpedo wakes sighted and reported at night under conditions of excitement are subject to some degree of error. Six torpedoes in several different directions were reported in rapid succession while the ship was maneuvering. Opportunities for error were plentiful.

    The only torpedo of those reported that could have originated from enemy destroyer Group I was the one seen porpoising 2,000 yards on the 


24 War Damage Report No. 46.


Selfride under torpedo attack, 6-7 October 1943

port beam on an approximately parallel course while the Selfridge was on course 3200. The next two torpedoes, reported about a minute later, were seen approaching about 300 on the port, bow in other words from almost the opposite direction, and from an angle that scarcely permits the assumption that they were fired by Group II. Almost immediately a torpedo was reported “close aboard on starboard bow,” apparently from a third direction, although the direction is not given. Left rudder was ordered, then changed at once to right rudder. The ship was coming to 0100 when the Captain saw a fifth torpedo wake crossing his bow from the starboard and going down the port side, then a sixth torpedo about 400 yards farther away also passing down the port side at an angle indicating it was fired from a position on the starboard bow. “Left full rudder” was at once ordered, and an effort was made to relocate the two torpedoes previously reported approaching on the port bow.

    In less than three minutes some six torpedoes had been sighted coming from at least three different directions. None of these torpedoes could have come from the ships that the Selfridge was firing upon during that interval. Only one could have come from the enemy destroyers previously engaged and left astern.

    The Selfridge was struck on the starboard bow at frame 40 by a torpedo coming at an angle of between 400 and 500 to the bow. At about the same instant, according to the commanding officer, there was “another very probable torpedo hit” on the port bow at frame 30, striking the ship at an angle of between 200 and 300 to the bow. Thus torpedoes fired from positions separated by as much as 800 struck simultaneously. The wake of neither torpedo was sighted, although it was the belief of Comdr. Peckham that one of the torpedoes previously sighted approaching on the port bow (though not resighted) scored the hit on the port bow.

The theory attributing these simultaneous hits to torpedo boats that had maneuvered into position unobserved cannot be dismissed as impossible, although it does seem improbable. One other possible source of the attack would seem to be the Selfridge’s fourth target in Group II. The location of this ship, hidden in a rain squall, is a matter of doubt. Capt. Walker reported, however, that this ship was far enough north at the end of the action to be on the starboard side of the Selfridge.25  Finally, the possibility that at least some of the torpedoes sighted may have been dropped by enemy planes should be taken into account.


25 The Gunnery Officer of the Selfridge reported that ship was “sharp on the port bow.”


    “The really significant fact,” concludes CINCPAC, “is that U. S. forces again suffered successful torpedo attack, with a strong possibility of its having been made by secondary craft or a detached group, which, although known to be present, were either lost track of or given too little attention, while our main interest was being centered on the engagement with nearer and more imposing enemy forces . . . It has been the unexpected torpedoes that have done the damage—torpedoes in the salvo that could not possibly reach us, from the enemy group that had been overwhelmed by our gunfire, or from the isolated and unidentified pip which somehow got dangerously close to us before we knew it.”




Second Echelon, Northern Force, Capt. William R. Cooke, Jr.

    Transport Unit, Capt. Cooke

        LST’s 339, 396, 460, Lt. Comdr. Roy W. Lajeunesse

        DesDiv FORTY-THREE (less Renshaw), Capt. Cooke

            Waller (F), Comdr. Laurence H. Frost

            Saufley, Comdr. Bert F. Brown

            Philip, Lt. Comdr. William H. Groverman, Jr.

    SC 1266

Landing Unit, Maj. Frank G. Umstead, USMC

    Detachment 35th CT.

    155mm Gun Group, 4th Defense Battalion

    Battery D (90 mm), 4th Defense Battalion

    ½ Battery G (S/L), 4th Defense Battalion

    Detachment 58th Construction Battalion.

Third Echelon, Northern Force, Capt. Grayson B. Carter.

    Transport Unit, Capt. Carter

        LST’s 354, 395, 399.

        DesDiv FORTY-FOUR (less Cony)

            Conway (F), Comdr. Nathaniel S. Prime

            Eaton, Comdr. Edward L. Beck.

            Pringle, Lt. Comdr. George DeMetropolis

    SC 505

Landing Unit, Capt. Burtis W. Anderson, USMC.

    Detachment 35 CT

    Tank Platoon, 4th Defense Battalion

    Detachment 58th Construction Battalion.




AB              Crane Ship                                                       AVC             Catapult Lighter                           
ACM  Auxiliary Mine Layer  AVD  Seaplane Tender (converted DD) 
AD  Destroyer Tender  AVP  Seaplane Tender (small) 
AE  Ammunition Ship  AW  Water Distilling and Storage Ship      
AF  Provision Store Ship  AY  Auxiliary Tender, Small 
AG  Miscellaneous Auxiliary  BB  Battleship 
AGC  Amphibious Force Flagship  CA  Heavy Cruiser 
AGP  Motor Torpedo Boat Tender  CAZ  Auxiliary, Unallocated as to Type (Conversion) 
AGS  Surveying Ship  CB  Large Cruiser 
AH  Hospital Ship  CL  Light Cruiser 
AK  Cargo  Vessel  CM  Mine Layer 
AKA  Cargo Vessel, Attack  CMc  Coastal Mine Layer 
AKN  Net Cargo Ship  CT  Destroyer (Building for Brazil) 
AKS  General Stores Issue Ship  CV  Aircraft Carrier 
AKV  Aircraft Supply Ship  CVB  Aircraft Carrier, Large 
AM  Large Mine Sweeper  CVL  Aircraft Carrier, Small 
AMb  Base Mine Sweeper  CVE  Aircraft Carrier, Escort 
AMc  Coastal Mine Sweeper  DD  Destroyer 
AN  Net Layer  DE  Destroyer Escort Vessel 
AO  Oiler  DM  Light Mine Layer (high speed) 
AOG  Gasoline Tanker  DMS  Mine Sweeper (high speed) 
AP  Transport  IX  Unclassified Auxiliary 
APA  Transport, Attack  LCC(1)  Landing Craft, Control, Mk.I 
APc  Coastal Transport  LCC(2)  Landing'Craft, Control, Mk. II 
APD  Troop Transport (high speed)  LCI(L)  Landing Craft, Infantry (large) 
APH  Transport for Wounded  LCM(2)  45' Landing Craft, Mechanized, Mk. II 
APL  Hotel Barge (Barracks Ship)  LCM(3)  50' Landing Craft, Mechanized, Mk. III 
APM  Mechanized Artillery Transport  LCM(6)  56' Landing Craft, Mechanized, Mk. VI 
APS  Auxiliary Cargo Submarine  LCP(L)  36' Landing Craft, Personnel (large) 
APV  Aircraft Transport  LCP(R)  36' Landing Craft, Personnel (with ramp) 
AR  Repair Ship  LCP(N)  Landing Craft, Personnel (nested) 
ARB  Repair Ship, Battle Damage  LCR(L)  Landing Craft, Rubber (large) 
ARG  Internal Combustion Engine Tender  LCR(S)  Landing Craft, Rubber (small) 
ARH  Heavy Hull Repair Ship  LCS(L)(3)  Landing Craft, Support (large), Mk. III 
ARL  Repair Ship, Landing Craft  LCS(L)(3)  Landing Craft, Support (small), Mk. III 
ARS  Salvage Vessel  LCS(S)(1)  Landing Craft, Support (small), Mk. I 
ARV  Aircraft Engine Overhaul and Structural Repair Ship  LCS(S)(2)  Landing Craft, Support (small), Mk. II 
AS  Submarine Tender  LCT(5)  Landing Crat, Tank, Mk. V 
ASR  Submarine Rescue Vessel     
ATF  Ocean Tug, Fleet     
ATA  Ocean Tug, Auxiliary     
ATO  Ocean Tug, Old     
ATR  Ocean Tug, Rescue     
AV  Seaplane Tender (large)     


Landing Craft, Tank, Mk. VI                     YCK                 Open Cargo Lighter                       
LCV  Landing Craft, Vehicle  YCV  Aircraft Transportation Lighter 
LCVP  Landing Craft, Vehicle and Personnel  YDG  Degaussing Vessel 
LSD  Landing Ship, Dock  YDT  Diving Tender 
LSM  Landing Ship, Medium  YF  Covered Lighter; Range Tender; Provision Store Lighter 
LST  Landing Ship, Tank  YFB  Ferry Boat 
LSV  Landing Ship, Vehicle  YFT  Torpedo Transportation Lighter 
LVT(1)  Landing Vehicle, Tracked (unarmored)  YG  Garbage Lighter 
LVT(A1)  Landing Vehicle, Tracked (armored)  YHB  Houseboat 
LVT(2)  Landing Vehicle, Tracked (unarmored)  YHT  Heating Scow 
LVT(A2)  Landing Vehicle, Tracked (armored)  YMS  Motor Mine Sweeper 
LVT(3)  Landing Vehicle, tracked (unarmored)  YMT  Motor Tug 
LVT(A3)  Landing Vehicle, Tracked (armored)  YN  Net Tender 
LVT(4)  Landing Vehicle, Tracked (unarmored)  YNg  Gate Vessel 
PC  173' Submarine Chaser  YNT  Net Tender (tug small) 
PCE  180' Patrol Crafft Escort  YO  Fuel Oil Barge 
PCE(R)  180' Patrol Craft Escort, Rescue  YOG  Gasoline Barge 
PCS  136' Submarine Chaser  YOS  Oil Storage Barge 
PE  Eagle Boat  YP  District Patrol Vessel 
PF  Frigate  YPK  Pontoon Storage Barge 
PG  Gun Boat  YR  Floating Workshop 
PGM  Motor Gun Boat  YRD(H)  Floating Workshop, Dry Dock (Hull) 
PT  Motor Torpedo  YRD(M)  Floating Workshop, Dry Dock (Machinery) 
PY  Yacht  YS  Stevedore Barge 
PYc  Coastal Yacht  YSD  Seplane Wrecking Derrick 
SC  110' Submarine Chaser  YSP  Salvage Pontoon 
SS  Submaine  YSR  Sludge Removal Barge 
YA  Ash Lighter  YTB  Harbor Tug, Big 
YAG  District Auxiliary, Miscellaneous  YTM  Harbor Tug, Medium 
YC  Open Lighter  YTL  Harbor Tug, Little 
YCF  Car Float  YTT  Torpedo Testing Barge 
    YW  Water Barge 




Class of airplane            
Model designation               Manufcturer                           U.S. name             
  Navy  Army     
Fighter, 2-eng  FO 
Lockheed "A"  Lightning 
Northrop, Haw  Black Widow 
  F7F    Grumman   
  XF5U    Chance-Vought   
Fighter, 1-eng  F2A1    Brewster  Buffalo 
  F3A    Brewster  Corsair 
  F4U    Chance-Vought  Corsair 
  FM    Eastern, Linden  Wildcat 
  XF2M    Eastern, Linden   
  FG    Goodyear  Corsair 
  F2G    Goodyear  Corsair 
  F4F    Grumman  Wildcat 
  F6F    Grumman  Hellcat 
  FR    Ryan   
  XF8B    Boeing, Seattle   
  XF15C    Curtiss, Buffalo   
  XFD    McDonald   
  XF8F    Grumman   
Scout Bomber, 1-eng (Dive)  SB2A1 
Brewster  Bermuda 
Canadian Cat  Helldiver 
Curtiss, Col  Helldiver 
Curtiss, Col  Helldiver 
  BTC    Curtiss, Col   
Douglas, E.S.  Dauntless 
  BTD    Douglas, E.S.   
  XBT2D    Douglas, E.S.   
Fairchild, Mont  Helldiver 
  BTM    Martin, Balto  Mercury 
  SBN1    N . A. F.   
  SB2U1    Vought-Sikorsky  Vindicator 
  XBK    Fleetwings   
Torpedo Bomber, 1-eng  TBY    Consolidated-Allen-town  Seawolf 
  TBD1    Douglas, E.S.  Devastator 
  TB2D1    Douglas, E.S.   
  TBM    Eastern, Trent  Avenger 
  TBF    Grumman  Avenger 
Patrol Bomber, boat, 4-eng  PB2Y    Consol, S.D.  Coronado 
2-eng  PB2B 
Boeing, Vane  Catalina 
  PBY-5, 5A 
Consol, S.D., N.O.  Catalina 
  P4Y    Consol, N.O.   
  PBM    Martin, Balto  Mariner 
N. A. F.  Catalina 
  XPBB    Boeing, Seattle  Sea Ranger 

See footnotes at end of table.



Class of airplane                Model designation           Manufacturer                       U.S. name                     
Bomber, Land: 4-eng  PB4Y-12 
Consol, S.D.  Liberator 
Consol, S.D.  Liberator 
2-eng.  PBO1 2 



Lockheed  Hudson 
N. Amer., K.C.  Mitchell 
  PV-1    Lockheed "A"  Ventura 
  PV-2    Lockheed "B"  Harpoon 
  XP2V    Lockheed "A"   
Observation Scout, 1-eng.  OY-12 
Consol, Wayne  Sentinel 
  SOC1    Curtiss, Col.  Seagull 
  S03C1    Curiss, Col.  Seamew 
  SC    Curtiss, Col.  Seahawk 
  S2E-1    Edo, L. I. C.   
  XOSE    Edo, L. I. C.   
  OS2N    N. A. F.  Kingfisher 
  OS2U    Vought-Sikorsky  Kingfisher 
Utility, 2-eng  JRB2 
Beech  Expeditor 
Cessna  Bobcat 
Douglas, S.M.  Havoc 
Grumman  Goose 
Grumman  Wildgeon 
Martin, Omaha  Maruder 
Utility, 1-eng.  J2F 
Columbia, V.S.  Duck 
Beech  Traveler 
  GK2 3 
Fairchild, Hager  Forwarder 
  GH    Howard  Nightingale 
  AE (HE) L-4  Piper  Grasshopper 
Transport, 4-eng., Land  RY-1-22 
Consol, F.W.  Liberator 
Consol, S.D.  Liberator 
Douglas, Chicago  Skymaster 
  XR60    Lockheed "A"  
  R2Y    Consol, S.D.   
Transport, 2-eng, Land  RB 
Budd, Phila  Conestoga 
Curtiss, L.B.  Commando 
  R3D2    Douglas, L.B.   
  R4D-1, 52 
C-47, 47A
Douglas, L.B.  Skytrain 
  R4D-2, 42 
C-49, 49A
Douglas, S.M.  Skytrooper 
Douglas, S.M.  Skytrooper 
C-56, 60
Lockheed  Lodestar 
Transport, 4-eng, Sea.  PB2Y3R    Consol, S.D.  Coronado 
  JRM (XPB2M-1R)  Martin, Balto  Mars 
  JR2S-2    Vought-Sikorsky  Excalibur 
Transport, 2-eng., Sea  PBM-3R    Martin, Balto  Mariner 
Training, 1-eng., Primary  N3N    N. A. F.   
Piper  Grasshopper 
  NR1 2 
Ryan  Recruit 
  NP1    Spartan   


17, 18, 27 

Boeing, Wichita (Stearman)  Caydet 
  N2T    Timm  Tutor 
  XN5N    N. A. F.   
  XNL2    Langley Aviation   

See footnotes at end of table.



Class of airplane          
Model designation            Manufacturer                       U.S. name                      
Training, 2-eng., Adv.  SNB-1 
Beech  Kansan 
Beech  Navigator 
Training, 1-eng., Adv.  SNV2 
Consol, Downey  Valiant 
  SNC1    Curtiss, St. Louis  Falcon 
  NH    Howard   
N. Amer. Dal  Texas 
Special Purpose, 2-eng  TDR    Am. Aviation   
  TD3R    Am. Aviation   
{Brunswick Balke Collender   
  TD2R    Interstate Aviation   
N. A. F.   
Special Purpose, 1-eng.  TDN    Bruns. Balke Coll   
  TDR    Interstate Avia   
Radio Plane Co   
Helicopter, 1-eng.  HNS-12 
  HO2S-12  R-5  Sikorsky   

All models obsolete.

These airplanes are Army-procured for Navy.  All others are under Navy cognizance.



Published: Tue Apr 14 11:03:58 EDT 2020