USS Kilty (DD-137), Empress Augusta Bay, 1943

USS Kilty (DD-137), Empress Augusta Bay, 1943

U.S. Destroyers: An Illustrated Design History, Norman Friedmann .The standard history of the development of American destroyers, from the earliest torpedo boat destroyers to the post-war fleet, and covering the massive classes of destroyers built for both World Wars. Gives the reader a good understanding of the debates that surrounded each class of destroyer and led to their individual features.


Náutica é o conjunto das tecnologias de navegação no mar, a palavra "nau" significa barco.

O termo náutica é hoje empregado não somente para se referenciar a arte de navegar, mas a toda atividade ligada a embarcações aquáticas ou praticada em água do mar, rio ou lagoa. Sejam estas atividades de lazer, reparos, indústria, comércio, esportes, dentre muitas outras.

Avaliação de artigos
A ser desenvolvido e implementado.


During World War II, a variety of new and experimental units were organized by Marine Corps to enhance the capabilities of the Corps. For the first time under one cover, this historical reference pamphlet tells of the development, deployment, and eventual demise of the five types of special units: raiders, parachutists, glider forces, barrage balloon squadrons, and base defense battalions. Official records of the Marine Corps and appropriate historical works were utilized in compiling this chronicle.

Final editing of the manuscript was accomplished by Henry I. Shaw, Jr., Chief Historian and Head, Histories Section, Historical Branch. Miss Kay P. Sue typed the final draft. Maps were prepared by Sergeant Jerry L. Jakes. All illustrations are official Department of Defense photographs from the files of the Combat Pictorial Branch, G-3 Division of this Headquarters.

USS Kilty (DD�) was a Wickes-class destroyer in the United States Navy. She was the first ship named for Admiral Augustus Kilty.

USS Sigourney (DD-643) was a Fletcher-class destroyer, the second ship of the United States Navy to be named for James Sigourney, an officer during the War of 1812.

USS Charles Ausburne (DD-570), a Fletcher-class destroyer, was the second ship of the United States Navy to be named for Charles L. Ausburne, a sailor in World War I who was posthumously awarded the Navy Cross.

USS Chevalier (DD-451), a Fletcher-class destroyer, was the first ship of the United States Navy to be named for Lieutenant Commander Godfrey Chevalier.

USS Saufley (DD/DDE/EDDE-465), a Fletcher-class destroyer, was a warship of the United States Navy named for pioneering Naval Aviator, Lieutenant Richard Saufley, USN.

USS Waller (DD/DDE-466), a Fletcher-class destroyer, was a ship of the United States Navy named for Major General Littleton Waller, USMC (1856�).

USS Taylor (DD/DDE-468) was a Fletcher-class destroyer of the United States Navy, named for Rear Admiral William Rogers Taylor (1811�). She was laid down on 28 August 1941 at Bath, Maine, by the Bath Iron Works Corp. launched on 7 June 1942, sponsored by Mrs. H. A. Baldridge and commissioned on 28 August 1942 at the Charlestown Navy Yard near Boston, Mass., Lieutenant Commander Benjamin Katz in command.

USS Fullam (DD-474) was a Fletcher-class destroyer in the United States Navy during World War II. Fullam was named for Rear Admiral William Fullam (1855-1926).

USS Spence (DD-512), a Fletcher class destroyer , was laid down on 18 May 1942 by the Bath Iron Works, Bath, Maine launched on 27 October 1942 sponsored by Mrs. Eben Learned and commissioned on 8 January 1943, Lieutenant Commander H. J. Armstrong in command. The ship was named for Robert T. Spence, superintendent of the construction of USS Ontario (1813), and captain of USS Cyane (1815).

USS Guest (DD-472), a Fletcher-class destroyer, was a ship of the United States Navy named for Commodore John Guest (1822�).

USS Pringle (DD-477), a Fletcher-class destroyer, was a ship of the United States Navy named for Vice Admiral Joel R. P. Pringle (1873�).

USS Sigsbee (DD-502), a Fletcher-class destroyer, was a ship of the United States Navy named for Rear Admiral Charles D. Sigsbee (1845�).

USS Conway (DD/DDE-507), a Fletcher-class destroyer, was the second ship of the United States Navy to be named for William Conway, who distinguished himself during the Civil War.

USS Cony (DD/DDE-508), a Fletcher-class destroyer, was a ship of the United States Navy named for Joseph S. Cony (1834�), a naval officer during the Civil War.

USS Foote (DD-511), a Fletcher-class destroyer, was the third ship of the United States Navy to be named for Rear Admiral Andrew Hull Foote (1806�), who served during the Civil War.

USS Robinson (DD-562), a Fletcher-class destroyer, was the second ship of the United States Navy to be named for Captain Isaiah Robinson, who served in the Continental Navy.

USS Patterson (DD-392), a Bagley-class destroyer, was the second ship of the United States Navy to be named for Daniel Todd Patterson, an officer of the US Navy who served in the Quasi-War with France, First Barbary War, and the War of 1812.

USS Lardner (DD-487), a Gleaves-class destroyer, was the second United States Navy ship to be named for Rear Admiral James L. Lardner, a Naval officer during the American Civil War. Lardner received 10 battle stars for World War II service.

USS Harrison (DD-573) was a Fletcher-class destroyer of the United States Navy. She was second Navy ship of that name, and the first named in honor of Captain Napoleon Harrison (1823�).

USS Farenholt (DD-491) was a Benson-class destroyer in the United States Navy during World War II. She was the second ship named for Admiral Oscar Farenholt.

Subaru EJ253 Engine

Please note that this article considers the EJ253 engine as it was supplied in Australian-delivered vehicles specifications for other markets may differ.

EJ253 block and crankcase

The Subaru EJ253 engine had an aluminium alloy block with 99.5 mm bores – with cast iron dry-type cylinder liners – and a 79.0 mm stroke for a capacity of 2457 cc. The cylinder block for the EJ253 engine had an open-deck design whereby the cylinder walls were supported at the three and nine o’clock positions.

The crankcase for the EJ253 engine had five main bearings and the flywheel housing was cast with the crankcase for increased rigidity. Like other EJ Phase II engines, the crankshaft thrust bearing was positioned at the rear of the crankshaft to reduce the transfer of natural engine frequencies to the transmission and driveline.


Cylinder head

The EJ253 engine had an aluminium alloy cylinder head with cross-flow cooling. The EJ253 engine had a hollow-type single overhead camshaft (SOHC) for each cylinder bank. Due to the cylinder head offset, the left camshaft was longer than the right camshaft to align the cam belt sprockets. Both camshafts were driven by a single belt which had round profile teeth for quiet operation and was constructed of wear-resistant double canvas and heat resistant rubber materials with a wire core.

For the EJ253 engine, the four valves per cylinder were actuated by shim-less type buckets (i.e. one-piece, solid valve lifters).

Intelligent Active Valve Lift System (i-ALVS)

Subaru’s Intelligent Active Valve Lift System (i-AVLS) was first introduced on the post-August BL/BP Liberty and BP Outback. With i-AVLS, one intake valve for each cylinder which could utilise a low lift camshaft lobe profile or a high lift camshaft lobe profile. At low engine speeds, the operation of the low/mid lift camshaft profile on one intake valve increased the speed of the air entering the combustion chambers from that port and created an imbalance in pressure as air entered the cylinder. This pressure imbalance created a swirling pattern and better air/fuel mixture formation, thereby increasing torque output.

At high engine speeds, the rocker arms of each cylinder’s two intake valves were locked together such that high-profile camshaft lobe acted on both of them. With the higher lift, intake resistance to air as reduced to enhance top-end power. Based on engine load, driving requirements and atmospheric conditions, the ECM would determine which camshaft lift profile to engage.


Injection and ignition

The EJ253 engine had multi-point sequential fuel injection and centrally located spark plugs. Initially, the EJ253 engine had two ignition coils (one for each pair of cylinders, i.e. 1-2 and 3-4) which fired the spark plugs directly twice per cycle. The ignition knock control system had ‘fuzzy logic’ that enabled the maximum ignition advanced angle to be used without detonation since the programme continually adapted to changes in environmental conditions and fuel quality.

It is understood that the EJ253 engine had a MAF (mass airflow) sensor rather than a MAP (manifold absolute pressure) sensor for more accurate measurement of intake air volume. Furthermore, the EJ253 engine had a compression ratio of 10.1:1 the injection and firing order was 1-3-2-4.

Subaru EJ20G Engine

For the EJ20G engine, the crankshaft was supported by five bearings that were made from aluminium alloy, while the no. 3 thrust bearing had a metal flange to receive thrust force. Furthermore, the corners of the crankshaft journals and webs, and the crank pins and webs, had a fillet-roll finish to increase stiffness.

  • The top piston ring had an inner bevel design
  • The second piston ring had an interrupt design to reduce oil consumption and,
  • The oil control ring had a slit design.

Cylinder head and camshafts

The cross-flow cylinder head for the EJ20G was made from die-cast aluminium and had double overhead camshafts (DOHC) per cylinder bank. A single timing belt was used to drive the four camshafts, while the back of the belt also drove the water pump. The timing belt consisted of a strong and inflexible core wire, wear-resistant canvas and heat-resistant rubber material. For quiet operation, the teeth on the timing belt had a round profile. For the GC/GM Impreza WRX, a hydraulic belt-tensioner maintained timing belt tension. The timing belt cover was a made from a synthetic resin moulding and used rubber at the mating surface of the cylinder block to absorb noise and vibrations.

Each camshaft was supported by three journals with three camshaft caps, while each camshaft flange was supported by a groove in the cylinder head to receive thrust force. During their manufacture, the camshaft ‘nose’ was subjected to a chill treatment to increase wear resistance and anti-scuffing properties.

The EJ20G engine had four valves per cylinder that were actuated by hydraulic lifters.

TD05 Turbocharger

For the GC/GM Impreza WRX, the EJ20G engine had a water-cooled, Mitsubishi TD05 turbocharger the rotational speed of the turbine ranged from approximately 20,000 rpm to 150,000 rpm and peak boost is understood to be around 11 to 12 psi.

To prevent excessive boost pressure, which could cause knocking and heavier thermal loads on the pistons, the EJ20G engine had a wastegate valve. Once boost pressure reached its maximum, the wastegate valve would open so that part of the exhaust gas would bypass the turbine and flow into the exhaust pipe.

The turbocharger was lubricated by the engine oil and used full-floating type bearings to form lubrication films. Furthermore, engine coolant from the coolant drain hose (under the engine cylinder head) flowed to a coolant passage in the turbocharger bearing housing. After cooling the bearing housing, the coolant flowed into the coolant filler tank via a pipe.

The EJ20G engine had an air bypass valve to prevent the suction noise that can otherwise occur when the throttle valve is suddenly closed and causes a sudden rise in air pressure between the turbocharger and the throttle body. The air bypass valve was actuated by the vacuum created by the closure of the throttle valve and allowed the suction air to bypass the turbocharger and flow upstream, thereby lowering the pressure in the air passage.


Injection and ignition

The EJ20G engine had pentroof combustion chambers which featured a wide ‘squish’ area. The EJ20G engine had multi-point fuel injection via gallery-type (or side-feed type) fuel injectors. For the GC/GM Impreza WRX, the EJ20G engine used a hot-film type mass air flow sensor to calculate intake air volume. The injection and firing order for the EJ20G engine was 1-3-2-4.

The EJ20G engine had centrally mounted spark plugs and a compression ratio of 8.0:1. Furthermore, a piezo-electric type knock sensor installed on the cylinder block which converted knocking vibrations into electric signals.

USS Kilty (DD-137), Empress Augusta Bay, 1943 - History

If the initial plans for the direct assault on the Buin area or the Shortlands had been carried out, the two small islands of the Treasury Group would have been bypassed and left in the backwash of the campaign. Instead, with the change in plans to strike directly at Empress Augusta Bay, the islands of Mono and Stirling became important as long-range radar sites and torpedo boat anchorages. Moreover, in an attempt to deceive the enemy as to the direction of the attack on Bougainville and convince him that the ultimate Allied aim might be the Buin area or the Shortlands, the seizure of the Treasurys was given added emphasis by being set as a preliminary to the Torokina landings. To help this deception succeed, reconnaissance patrols to the Shortlands and diversionary operations on the island of Choiseul&mdashplus low-flying photo missions over the Shortlands&mdashwere scheduled by IMAC to increase the enemy&rsquos conviction that the follow-up objective was the Shortlands.

This could have been a natural assumption by the enemy. The Treasurys are about 60 miles northwest of Vella Lavella and only 18 miles south of the Shortlands. While the size of the Treasurys limited consideration as a major target, Mono and Stirling were close enough to Shortland Island to cause the Japanese some concern that they might be used as handy stepping stones by SoPac forces. But then again, the Treasurys are only 75 miles from Cape Torokina&mdasha fact which the Allies hoped might be lost on Bougainville&rsquos defenders.

The Treasury Islands are typical of other small islands jutting out of the sea in the Solomons chain. Mono is a thickly forested prominence of volcanic origin, with abrupt peaks and hill masses more than 1,000 feet high in the southern part. These heights slope gradually in an everwidening fan to the west, north, and east coasts. The shores are firm, with few swamps, and rain waters drain rapidly through deep gorges. The island is small, about four miles north to south and less than seven miles lengthwise.

Stirling Island to the south is smaller, more misshapen. Fairly level, this island is about four miles long and varies from 300 yards to nearly a mile in width. There

are several small, brackish lakes inland, but the island is easily traversed and, once cleared of its covering forest, would be an excellent site for an airfield. Between these two islands is a mile or more of deep, sheltered water&mdashone of the many anchorages in the Solomon Islands to bear the name Blanche Harbor. The combination of these features&mdashairfield site, radar points, good anchorage&mdashwas the factor which resulted in the seizure of the Treasurys as part of the Bougainville operation.

Early information about the islands was obtained by an IMAC patrol which spent six days in the Treasurys in August, scouting the area, observing the movement of the Japanese defenders, and interrogating the natives. In this latter instance, the loyal and friendly people of the Treasurys were a remarkable contrast to the suspicious and hostile Bougainville inhabitants. Additional details were received from rescued aviators who found Mono Island a safe hiding place after their planes had been forced down by damage incurred in raids over Buin and the Shortlands. This first-hand intelligence was augmented by aerial photographs. The reports and photos indicated that the best landing beaches were inside Blanche Harbor, on opposite shores of Mono and Stirling. The only beaches suitable for LSTs, however, were on Mono between the Saveke River and a small promontory, Falamai Point.

As limited as this information was, the amount of intelligence on the enemy dispositions on the two islands was even more meager. The Japanese strength was estimated at 135 men, lightly armed. These were bivouacked near Falamai but maintained a radio station and observation posts in various areas. Natives reported that much of the time the Japanese moved about Mono armed only with swords or hand guns. Stirling Island was apparently undefended.

The 8th New Zealand Brigade Group, attached to I Marine Amphibious Corps for the seizure and occupation of these islands, arrived at Guadalcanal from New Caledonia in mid-September. Although the New Zealanders would form the bulk of the assault troops, the GOODTIME operation was IMAC-directed and IMAC-supported. The landing force comprised about 7,700 officers and men, of whom about 1,900 were from I Marine Amphibious Corps support troops&mdashantiaircraft artillery, construction battalions, signal, and boat pool personnel. Marines attached to the brigade task organization included a detachment from the IMAC Signal Battalion and an air-ground liaison team from General Harris&rsquo ComAirNorSols headquarters.

On 28 September, Brigadier Row, the landing force commander, was informed of the general nature of the GOODTIME operation, and planning in conjunction with Admiral Fort began immediately, although there was only enough information available to the commanders of the task group and the landing force to formulate a plan in broad outline. The task was far from easy, for the Southern Force was confronted with the same logistical and transportation problems that faced the Empress Augusta Bay operation.

Fort and Row decided that the main assault would be made in the area of Falamai, where beaches were suitable for LSTs. Stirling Island would be taken concurrently for artillery positions. No other landings were planned but after Row was informed that the long-range radar would have to be positioned on the northern side of Mono to be of benefit to

Map 12: Treasury Islands Landings, I Marine Amphibious Corps, 27 October 1943

the Bougainville operation, another landing at Soanotalu on the north coast was written into the plans.

Final shipping allocation to Fort&rsquos Southern Force included 31 ships of six different types&mdash8 APDs, 8 LCIs, 2 LSTs, and 3 LCTs for landing troops and supplies, 8 LCMs and 2 APCs for heavy equipment and cargo. The limited troop and cargo capacity of this collection of ships and landing craft restricted the Southern Force&rsquos ability to put more than a minimum of troops and supplies ashore initially, but this problem was solved by reducing the strength of the brigade&rsquos battalions and limiting the number of artillery weapons, motor transport, and engineering equipment in the first echelon. The brigade&rsquos assault units included 3,795 troops with 1,785 tons of supplies and equipment. Succeeding echelons were scheduled to sail forward at intervals of five days.

The final plans, issued by Row&rsquos headquarters on 21 October, directed the 29th and 36th Battalions to land nearly abreast near Falami Point, with the 34th Battalion landing on Stirling Island. Simultaneously, a reinforced infantry company accompanied by radar personnel and

Seabees would go ashore at Soanotalu in the north. The two battalions on Mono would then drive across the island to link up with the Soanotalu landing force while naval base construction got underway at Stirling.

The initial landings in Blanche Harbor were to be covered by a naval gunfire support group of two destroyers, the Pringle and Philip. Liaison officers of IMAC planned the gunfire support, as the New Zealand officers had no experience in this phase of operations. While the brigade group expected to have no trouble in seizing the islands, the naval support was scheduled to cover any unforeseen difficulties. The gunfire plan called for the two destroyers to fire preparation salvos from the entrance to Blanche Harbor before moving in toward the beaches with the landing waves to take targets under direct fire. The IIIPhibFor, however, took a dim view of risking destroyers in such restricted waters. The desired close-in support mission was then assigned to the newly devised LCI(G)&mdashgunboats armed with three 40-mm, two 20-mm, and five .50 caliber machine guns&mdashwhich were making their first appearance in combat. Two of these deadly landing craft were to accompany the assault waves to the beaches.

After one final practice landing on Florida Island, the brigade group began loading supplies and embarking troops for the run to the target area. Admiral Fort&rsquos Southern Force was divided into five transport groups under separate commanders, and these groups departed independently when loaded. The slower LSTs and LCMs left first, on the 23rd and 24th of October, and were followed the next day by the LCIs. The APDs sailed on 26 October.

The Southern Force departed with a message which delighted the New Zealanders as typical of the remarks to which Americans at war seemed addicted: &ldquoShoot calmly, shoot fast, and shoot straight.&rdquo 2

At 0540 on the 27th, the seven APDs of the first transport group lay to just outside the entrance to Blanche Harbor and began putting troops over the side into landing craft. Heavy rain and overcast weather obscured the beaches, but the preassault bombardment by the Pringle and Philip began on schedule. The USS Eaton moved to the harbor&rsquos mouth and took up station as fighter-director ship as the destroyers registered on Mono Island. The firing was accomplished without assistance of an air spotter, who later reported radio failure at the critical moment. This probably accounts for the disappointing results of the preparatory bombardments, which proved to be of little value except to boost the morale of the assault troops. The Pringle&rsquos fire was later declared to be too far back of the beach area to be helpful, and the bombardment by the Philip left a great deal to be desired in accuracy, timing, and quantity.

A fighter cover of 32 planes arrived promptly on station over the Treasurys at 0600, and, under this protective screen, the assault waves formed into two columns for the dash through Blanche Harbor to the beaches. Unexpectedly, enemy machine gun fire from Falamai and Stirling greeted the assault boats as they ploughed through the channel. At 0623, just three minutes before the landing craft nosed into the beaches on opposite sides of the harbor, the preassault cannonading ceased and the two LCI gunboats&mdashone on each

flank of the assault wave&mdashtook over the task of close support for the landing forces. At least one 40-mm twin-mount gun, several machine guns, and several enemy bunkers were knocked out by the accurate fire of these two ships. Promptly at 0626, the announced H-Hour, New Zealand troops went ashore on Mono and Stirling.

At Falamai, the 29th and 36th Battalions moved inland quickly against light rifle and machine gun fire, mostly from the high ground near the Saveke River. Casualties in the first wave were light&mdashone New Zealand officer and five sailors wounded&mdashand the second wave had no casualties.

The New Zealanders began to widen the perimeter as more troops were unloaded. At 0735, enemy mortar and medium artillery fire registered on the beach area, causing a number of casualties and disrupting unloading operations. Both LSTs were hit, with one ship reporting 2 dead and 18 wounded among the sailors and soldiers aboard. The other ship reported 12 wounded. Source of the enemy fire could not be determined. The Eaton, with Admiral Fort on board, ignored a previous decision not to enter Blanche Harbor and resolutely steamed between the two islands. This venture ended, however, when enemy planes were reported on the way, and the Eaton reversed course to head for more maneuvering room outside the harbor. Assured that the air raid was a false alarm, the destroyer returned to Blanche Harbor and added its salvos to those of the LCI gunboats. This fire, directed at likely targets, abruptly ended the Japanese exchange.

By 1800, the two battalions had established a perimeter on Mono Island and were dug in, trying to find some comfort in a dismal rain which had begun again after a clear afternoon. Evacuation of casualties began with the departure of the LSTs. With the exception of one LST, which still had 34 tons of supplies aboard when it retracted, all ships and landing craft had been unloaded and were on their way back to Guadalcanal by the end of D-Day. The casualties were 21 New Zealanders killed and 70 wounded, 9 Americans killed and another 15 wounded.

The landings at Stirling and Soanotalu were uneventful and without opposition. There were no casualties at either beachhead. At Stirling, the 34th Battalion immediately began active patrolling as soon as the command was established ashore. The Soanotalu landings proceeded in a similarly unhindered manner. A perimeter was established quickly, and bulldozers immediately went to work constructing a position for the radar equipment which was to arrive the next day.

The fighter cover throughout the day had shielded the troops ashore from enemy air attacks. The escorting destroyers, however, were hit by an enemy force of 25 medium and dive bombers at about 1530, and the USS Cony took two hits. Eight crewmen were killed and 10 wounded. The fire from the destroyer screen and the fighter cover downed 12 of the enemy planes. That night the bombers returned to pound the Mono Island side of Blanche Harbor and, in two raids, killed two New Zealanders and wounded nine.

Action along the Falamai perimeter the night of 27 October was concentrated mainly on the left flank near the Saveke River, the former site of the Japanese headquarters, and several attacks were beaten back. The following day, patrols moved forward of the perimeter seeking

the enemy, and one reinforced company set out cross-country to occupy the village of Malsi on the northeast coast. There was little contact. Japanese ground activity on the night of the 28th was light, and enemy air activity was limited to one low-level strafing attack and several quick bombing raids&mdashall without damage to the brigade group.

By 31 October, the entire situation was stable. The perimeter at Falamai was secure, Malsi had been occupied without opposition, and radar equipment at Soanotalu had been installed and was in operation. With the arrival of the second echelon on 1 November, the New Zealanders began an extensive sweep of the island to search out all remaining enemy troops on the island. The going was rough in the high, rugged mountain areas, but, by 5 November, enemy stragglers in groups of 10 to 12 had been tracked down and killed. The New Zealanders lost one killed and four wounded in these mop-up operations.

Undisturbed for some time, the perimeter at Soanotalu was later subjected to a number of sharp attacks, each one growing in intensity. The Soanotalu force was struck first on 29 October by small groups of Japanese who were trying to reach the beach after traveling across the island from Falamai. These attacks continued throughout the afternoon until a final charge by about 20 Japanese was hurled back. Construction of the radar station continued throughout the fighting. Enemy contact on the next two days was light, and the first radar station was completed and a second one begun.

On the night of 1 November, a strong force of about 80 to 90 Japanese suddenly struck the perimeter in an organized attack, apparently determined to break through the New Zealand defense to seize a landing craft and escape the island. The fight, punctuated by grenade bursts and mortar fire, raged for nearly five hours in the darkness. One small group of enemy penetrated the defenses as far as the beach before being destroyed by a command group. About 40 Japanese were killed in the attack. The Soanotalu defenders lost one killed and nine wounded. The following night, 2 November, another attempt by a smaller Japanese force was made and this attack was also beaten back. This was the last organized assault on the Soanotalu force, and the remainder of the Japanese on the island were searched out and killed by the New Zealand patrols striking overland.

By 12 November, the New Zealanders had occupied the island. Japanese dead counted in the various actions totaled 205 the New Zealanders took 8 prisoners. It is doubtful that any Japanese escaped the island by native canoe or swimming. In addition, all enemy weapons, equipment, and rations on the island were captured. The Allied casualties in this preliminary to the Bougainville operation were 40 New Zealanders killed and 145 wounded. Twelve Americans were killed and 29 wounded.

During the period of fighting on Mono Island, activity on Stirling was directed toward the establishment of supply dumps, the building of roads, and the construction of advance naval base and boat pool facilities. Although several minor enemy air raids damaged installations in the early phases of the operation, the landing at Empress Augusta Bay diverted the attention of the enemy to that area and ended all Japanese attempts to destroy the force in the Treasurys.

If the Japanese had opportunity to speculate on the significance of the Treasurys invasion, the problem may have been complicated a few hours later by a landing of an Allied force on the northwest coast of Choiseul Island, just 45 miles from the southeastern coast of Bougainville. The landing was another ruse to draw Japanese attention from the Treasurys, point away from the Allies&rsquo general line of attack, and divert the enemy&rsquos interest&mdashif not effort&mdashtoward the defense of another area. More specifically, the Choiseul diversion was calculated to convince the Japanese that the southern end of Bougainville was in imminent danger of attack from another direction. The salient facts which the Allies hoped to conceal were that the real objective was Empress Augusta Bay, and that the Choiseul landing force consisted only of a reinforced battalion of Marine parachute troops.

Actually, the raid on Choiseul was a small-scale enactment of landing plans which had been discarded earlier. Choiseul was considered as a possible objective for the main Northern Solomons attack but when the decision was made that the Allied attack would strike directly amidships on the western coast of Bougainville, the Choiseul idea was dropped. Then, when the suggestion was advanced by Major James C. Murray, IMAC Staff Secretary, that, because of the size and location of Choiseul, a. feint toward that island might further deceive the Japanese as to the Allies&rsquo intentions, the diversionary raid was added to the Northern Solomons operation.

Choiseul is one of the islands forming the eastern barrier to The Slot and as one of the Solomon Islands, it shares the high rainfall total, the uniform high heat and humidity common to other islands of the chain. About 80 miles long and 20 miles wide at the widest point, Choiseul is joined by reefs at the southern end of two small islands (Rob Roy and Wagina) which seems to extend Choiseul&rsquos length another 20 miles. The big island is not as rugged as Bougainville and the mountain peaks are not as high, but Choiseul is fully as overgrown and choked with rank, impenetrable jungle and rain forest. The mountain ranges in the center of the island extend long spurs and ridges toward the coasts, thus effectively dividing the island into a series of large compartments. The beaches, where existent, vary from wide, sandy areas to narrow, rocky shores with heavy foliage growing almost to the water&rsquos edge. Other compartments end in high, broken cliffs, pounded by the sea.

The island was populated by nearly 5,000 natives, most of whom (before the war) were under the teachings of missionaries of various faiths. With the exception of a small minority, these natives remained militantly loyal to the Australian government and its representatives. As a result,

coastwatching activities on Choiseul were given valuable assistance and protection.

Combat intelligence about the island was obtained by patrols which scouted various areas. One group, landed from a PT boat on the southwest coast of Choiseul, moved northward along The Slot side of the island toward the Japanese base at Kakasa before turning inland. After crossing the island to the coastwatcher station at Kanaga, the patrol was evacuated by a Navy patrol bomber on 12 September after six days on the island.

Two other patrols, comprising Marines, naval officers, and New Zealanders, scouted the northern end of the island and Choiseul Bay for eight days (22-30 September) before being withdrawn. Their reports indicated that the main enemy strength was at Kakasa where nearly 1,000 Japanese were stationed and Choiseul Bay where another 300 troops maintained a barge anchorage. Several fair airfield sites were observed near Choiseul Harbor, and a number of good beaches suitable for landing purposes were marked. Japanese activity, the patrols noted, was generally restricted to Kakasa and Choiseul Bay. 4

During the enemy evacuation of the Central Solomons, Choiseul bridged the gap between the New Georgia Group and Bougainville. The retreating Japanese, deposited by barges on the southern end of Choiseul, moved overland along the coast to Choiseul Bay where the second half of the barge relay to Bougainville was completed. This traffic was checked and reported upon by two active coastwatchers, Charles J. Waddell and Sub-Lieutenant C. W. Seton, Royal Australian Navy, who maintained radio contact with Guadalcanal.

Seton, on 13 October, reported the southern end of Choiseul free of Japanese, but added that at least 3,000 to 4,000 enemy had passed Bambatana Mission about 35 miles south of Choiseul Bay. On 19 October, the coastwatcher reported that the enemy camps in the vicinity of Choiseul Bay and Sangigai (about 10 miles north of Bambatana Mission) held about 3,000 Japanese who were apparently waiting for barge transportation to Bougainville. Seton indicated that the Japanese were disorganized, living in dispersed huts, and were short of rations. They had looted native gardens and searched the jungle for food. Further, the Japanese were edgy. All trails had been blocked, security had been tightened, and sentries fired into the jungle at random sounds. 5

After this information was received at IMAC headquarters, Vandegrift and Wilkinson decided that a diversionary raid on Choiseul would be staged. On 20 October, Lieutenant Colonel Robert H. Williams, commanding the 1st Marine Parachute Regiment, and the commanding officer of his 2nd Battalion, Lieutenant Colonel Victor H. Krulak, were summoned from Vella Lavella to Guadalcanal. At IMAC headquarters, Williams and Krulak conferred with Vandegrift and his staff. The orders to Krulak were simple: Get ashore on Choiseul and make as big a demonstration as possible to convince the

Map 13: Choiseul Diversion, 2nd Parachute Battalion, 28 October - 3 November 1943

Japanese that a major landing was in progress. In addition, reconnaissance would be conducted to determine possible sites for a torpedo boat patrol base.

The IMAC operation order, giving the code name BLISSFUL to the Choiseul diversion, was issued on 22 October. Based on information and recommendations from Seton, the Marines&rsquo landing was set for the beaches in the vicinity of Voza village, about midway between Choiseul Bay and Bambatana Mission. There the beaches were good, friendly natives would help the invading forces, and there reportedly were no enemy troops. Moreover, it was firmly astride the main route of evacuation of the Japanese stragglers from Kolombangara and points south. After receiving the order, Krulak went to the airstrip on Guadalcanal, and, while waiting for a plane to take him back to his command, wrote out the operation order for his battalion&rsquos landing.

This was to be the first combat operation of the 2nd Battalion as well as its first amphibious venture. Although equipped and trained for special assignments behind enemy lines, these Marines&mdashknown as Paramarines to their comrades&mdashnever chuted into action because suitable objectives were usually beyond the range of airborne troops and the necessary transport planes were in chronically short supply. The 1st Parachute Battalion, however, had taken Gavutu and Tanambogo Islands before going to Guadalcanal to take part in the defense of the airfield there in 1942. This battalion had then formed the nucleus for the present 1st Parachute Regiment, now consisting of three battalions in IMAC reserve at Vella Lavella. Each battalion, of three rifle companies each, was armed with a preponderance of light automatic and semi-automatic weapons. The nine-man squads in Lieutenant Colonel Krulak&rsquos rifle platoons carried three Johnson light machine guns 6 and six Johnson semi-automatic rifles each company had, in addition, six 60-mm mortars.

Krulak&rsquos return to his command set off a flurry of near-frenzied activity, since the battalion had a minimum of time for preparation. For the next four days, officers and men worked almost around the clock to assemble equipment, make final plans, and brief themselves on the task ahead. On the 24th, Coastwatcher Seton and two of his native guides arrived at Vella Lavella to meet Krulak&rsquos officers and men and give them last-minute information. After being briefed by Seton, Krulak requested and was given authority by IMAC to operate in any direction on Choiseul, if consistent with his mission.

Equipment and supplies for the operation were presorted into four stacks and late on the afternoon of the 27th of October the parachute battalion and its gear was embarked on board eight LCMs borrowed from the Vella Lavella boat pool. Krulak&rsquos three companies were reinforced by a communications platoon, a regimental weapons company with mortars and light machine guns, and a detachment from an experimental rocket platoon (bazookas and rockets) from IMAC. Total battalion strength was 30 officers and 626 men. In addition, one naval officer accompanied the battalion for reconnaissance purposes related to the possible establishment of a torpedo boat base.

At dusk, when four APDs which had just completed the Treasury landings arrived

off Vella Lavella, the troops and equipment were transferred from the LCMs to the McKean, Crosby, Kilty, and Ward in a quick operation that was completed in less than 45 minutes. The destroyer division, with the USS Conway acting as escort, sailed from Vella Lavella at 1921. The Conway&rsquos radar would locate the landing point in the dark.

Moving in column through the night, the convoy was sighted shortly after 2300 by an enemy snooper plane which dropped one bomb, scoring a near miss on the last APD in line. Shortly before midnight, at a point some 2,000 yards off the northwest coast of Choiseul, the convoy stopped, and a reconnaissance party in a rubber boat headed toward shore to scout the landing area. A signal light was to be shown if no enemy defenders were encountered. While waiting for the signal, Krulak ordered Companies F and G into the landing boats.

After waiting until 0019 (28 October), Company F headed toward the beach with Company G close behind. The operation order had directed Company G to make the initial assault, but the APDs had drifted apart and the Kilty with Company F embarked was closer to shore. Since no light on shore was yet discernible, the Marines expected opposition. The landing, however, was uneventful, and the patrol was waiting on shore. Observers on ship reported later that the light was visible at 0023, just four minutes after the parachute companies began the run for the beach. After setting the troops ashore, the landing craft immediately returned to the transports to bring in a load of supplies. 7

A lone enemy plane detected the Conway standing patrol duty seaward, and dropped two bombs near the ship. The Conway, reluctant to draw attention to the landing, did not return the fire, and the enemy plane droned away. An Allied escort plane, assigned to protect the convoy against such attacks, drew considerable criticism, however, for not remaining low enough to spot such bombing runs.

Two hours after arrival in the area, the convoy reversed course and steamed back to Vella Lavella, leaving behind four landing craft (LCP(R)) with their crews for the battalion&rsquos use. These craft were dispersed under the cover of overhanging mangroves near the offshore island of Zinoa, and the Marines turned to moving supplies off the beach. Seton, who landed on Choiseul with the battalion, disappeared into the bush and returned almost immediately with a group of native bearers. With their help, the Marines moved into the jungle. The transfer was none too soon enemy reconnaissance planes appeared at dawn to bomb the area but without success.

Early on the morning of the 28th, a base of operations was established on a high jungle plateau about a mile to the northwest of Voza and outposts were set up on the beach north and south of the village. Security was established and wire communications installed. The plateau, behind natural barriers of rivers and high cliffs, was an ideal defensive spot and a necessary base camp for the heavy radio gear with which IMAC had equipped the parachute battalion.

During the day of 28 October, while the Marines established their camp, another enemy flight appeared and raked the beachhead with a strafing and bombing attack. The effect was wasted. The Marines had dispersed their equipment had been moved and good camouflage discipline had been observed. Too, the natives had obliterated every sign of a landing at Voza and established a dummy beachhead several miles to the north for the special benefit of Japanese planes seeking a target.

Informed by Seton&rsquos guides that the Marine battalion, was bivouacked between a barge staging-replenishment base at Sangigai about eight miles to the south and an enemy outpost at the Warrior River about 17 miles to the north, Krulak on the morning of the 29th sent out reconnaissance patrols to the north and south. These groups were to locate trails, scout any enemy dispositions, and become familiar with the area.

Krulak personally led one patrol toward Sangigai, going overland toward the Vagara River which was about halfway between the Marine camp and the enemy base. While part of the patrol headed inland toward the high ground to the rear of Sangigai, to sketch the approaches to the village, the Marine comander led the rest of the patrol to the river. There the hidden Marines silently watched a group of about 10 Japanese unloading a barge and since this appeared to be an excellent opportunity to announce the aggressive intentions of the Krulak force, the Americans opened fire. Seven of the Japanese were killed, and the barge sunk. Krulak&rsquos section then returned to the base, followed shortly by the other half of the patrol. After the attack order was issued, a squad was sent back over the trail to the Vagara to hold a landing point for Krulak&rsquos boats and to block the Japanese who might be following the Marines&rsquo track. The patrol ran into a platoon of the enemy about three-quarters of a mile from the original Marine landing point and drove the Japanese off. 8

At 0400 the following morning, 30 October, Krulak led Companies E and F, plus the rocket detachment, toward Voza for an attack on Sangigai. The barge base had been marked as a target since 22 October. To help him in his assault, and give the impression of a larger attacking force, Krulak requested a preparatory air strike on reported Japanese positions northwest of the base. Estimated enemy strength was about 150 defenders, although Seton warned that Sangigai could have been reinforced easily from the southwest since the Marines&rsquo landing.

Krulak&rsquos attack plans were changed at Voza, however, since one of the four boats had been damaged a few minutes earlier in an attack by Allied planes. The strafing ended when the fighter pilots discovered their error and apologized to the boat crews with a final pass and a clearly visible &ldquothumbs-up&rdquo signal. The requested air strike at Sangigai hit at 0610 with better results. While 26 fighters flew escort, 12 TBFs dropped a total of more than two tons of bombs on enemy dispositions.

Unable to use the boats for passage to the Vagara, Krulak ordered his troops to begin a route march overland from Voza. Seton and his native guides led the way, followed by Company F (Captain Spencer H. Pratt) with a section of machine guns

and the rocket detachment. Company E (Captain Robert E. Manchester) and attached units followed. At about 1100, Japanese outposts on the Vagara opened fire on the Marine column. Brisk return fire from the parachutists forced the enemy pickets to withdraw towards Sangigai.

Following the envelopment plan he had formulated on the 29th, Krulak sent Company E along the coastline to launch an attack on Sangigai from that direction while the remainder of the force, under his command, moved inland to attack from the high ground to the rear and east of Sangigai. The assault was set for 1400, but as that hour drew near, the group in the interior found that it was still a considerable distance from the village. The mountainous terrain, tangled closely by jungle creepers and cut by rushing streams, slowed Krulak&rsquos force, and, by H-Hour, the column was still not in position to make its attack effort. When the sound of firing came from the direction of Sangigai village, the second force was still moving towards its designated jump-off point. Seton&rsquos natives, however, indicated that the enemy were just ahead.

Company E, moving along the beach, reached its attack position without trouble. Although the assault was delayed a few minutes, the company opened with an effective rocket spread and mortar fire. As the Marines moved forward, the Japanese defenders hastily withdrew, abandoning the base and the village to flee to the high ground inland. The Marine company entered the village without opposition. 9

The enemy&rsquos withdrawal to prepared positions inland fitted perfectly into Krulak&rsquos scheme of maneuver. The Japanese moved from the village straight into the fire of Company F, and a pitched battle that lasted for nearly an hour ensued. An enveloping movement by the Marines behind the effective fire of light machine guns forced the Japanese into several uncoordinated banzai charges which resulted in further enemy casualties. As the Marines moved once more to turn the enemy&rsquos right flank, the Japanese disengaged and about 40 survivors escaped into the jungle. A final count showed 72 enemy bodies in the area. Krulak&rsquos force lost four killed. Twelve others, including Krulak and Pratt, were wounded.

Company E, possessors of Sangigai, had been busy in the interim. Manchester&rsquos company, using demolitions, destroyed the village, the Japanese base and all enemy supplies, scuttled a new barge, and captured a number of documents, including a chart of enemy mine fields off southern Bougainville. The Marines then withdrew to the Vagara to board the four landing craft (the disabled boat had been repaired) for the return to Voza. Krulak&rsquos force, after burying its dead, retraced its path to the Vagara and spent the night in a tight defensive perimeter. 10 Early the

next morning, 31 October, the landing craft returned to carry the parachutists to Voza and the base camp.

With the battalion reassembled once more, the Marines prepared ambushes to forestall any Japanese retaliatory attacks, and aggressive patrols were pushed out along the coast to determine if the Japanese were threatening and to keep the enemy off balance and uncertain about Marine strength. A Navy PBY landed near Voza the following day to evacuate the wounded Marines and the captured documents and, on the same day, in answer to an urgent request by Krulak, 1,000 pounds of rice for the natives and 250 hand grenades and 500 pounds of TNT were air dropped near Voza. Several brisk engagements between opposing patrols were reported on this day, 1 November, but the base camp was not attacked. Seton&rsquos natives, however, reported that Sangigai had again been occupied by the Japanese.

After Krulak returned to the base camp on 31 October, his executive officer, Major Warner T. Bigger, led a patrol to Nukiki Village, about 10 miles to the north. No opposition was encountered. On the following day, 1 November, Bigger led 87 Marines from Company G (Captain William H. Day) toward Nukiki again to investigate prior reports of a large enemy installation on the Warrior River. Bigger&rsquos instructions were to move from Nukiki across the Warrior River, destroying any enemy or bases encountered, and then move as far north as possible to bring the main Japanese base at Choiseul Bay under 60-mm mortar fire. Enemy installations on Guppy Island in Choiseul Bay were an alternate target.

The patrol moved past Nukiki without opposition, although the landing craft carrying the Marines beached continually in the shallow mouth of the Warrior River. Since the sound of the coxswains gunning the boats&rsquo motors to clear obstructions was undoubtedly heard by any enemy in the area, Bigger ordered the Marines to disembark. The boats were then sent downriver to be hidden in a cove near Nukiki. Bigger&rsquos force, meanwhile, left four men and a radio on the east bank of the river, and all excess gear including demolitions was cached. Mortar ammunition was distributed among all the Marines. The patrol then headed upriver along the east bank, and the Warrior was crossed later at a point considerably inland from the coast.

By midafternoon, the natives leading the patrol confessed to Bigger that they were lost. Although in the midst of a swamp, the Marine commander decided to bivouac in that spot while a smaller patrol retraced the route back to the Warrior River to report to Krulak by radio and to order the boats at Nukiki to return to Voza. In response to Bigger&rsquos message, Krulak asked Seton if he had any natives more familiar with the country north of the Warrior River the only man who had visited the region was sent to guide the lost Marines.

The smaller patrol bivouacked at the radio site on the night of 1-2 November and awoke the next morning to the realization that a Japanese force of about 30 men had slipped between the two Marine groups and that their small camp was virtually surrounded. Stealthily slipping past enemy outposts, the patrol members moved to Nukiki, boarded the boats, and returned to Voza. After hearing the patrol&rsquos report, Krulak then radioed IMAC for fighter cover and PT boat support to

withdraw the group from the Choiseul Bay area.

Bigger was unaware of the activity behind him. Intent upon his mission, he decided to continue toward Choiseul Bay. After determining his position, Bigger ordered another small patrol to make its way to the river base camp and radio a request that the boats pick up his force that afternoon, 2 November. This second patrol soon discovered the presence of an enemy force to Bigger&rsquos rear, and was forced to fight its way towards Nukiki. This patrol was waiting there when the landing craft returned to Nukiki.

The main force, meanwhile, followed the new guide to the coast and then turned north along the beach toward Choiseul Bay. Opposite Redman Island, a small offshore islet, a four-man Japanese outpost suddenly opened fire. The Marines quickly knocked out this opposition, but one Japanese escaped&mdashundoubtedly to give the alarm.

Because any element of surprise was lost and thinning jungle towards Choiseul Bay provided less protection and cover for an attacking force, Bigger decided to execute his alternate mission of shelling Guppy Island. Jungle vegetation growing down to the edge of the water masked the fire of the 60-mm mortars, so Bigger ordered the weapons moved offshore. The shelling of Guppy was then accomplished with the mortars emplaced on the beach with part of the baseplates under water. The enemy supply center and fuel base was hit with 143 rounds of high explosives. Two large fires were observed, one of them obviously a fuel dump. Bigger&rsquos force, under return enemy fire, turned around and headed back toward the Warrior River.

The Japanese, attempting to cut off Bigger&rsquos retirement, landed troops from barges along the coastline and the Marine force was under attack four separate times before it successfully reached the Warrior River. There the patrol set up a perimeter on the west bank and waited for the expected boats.

Several men were in the river washing the slime and muck of the jungle march from their clothing when a fusillade of shots from the opposite bank hit the Marine force. The patrol at first thought it was being fired upon by its own base camp, but when display of a small American flag drew increased fire, the Marines dove for cover. Heavy return fire from the Marine side of the river forced the enemy to withdraw. Seizing this opportunity, Bigger directed three Marines to swim across the Warrior to contact the expected boats and warn the rescuers of the ambush. Before the trio could reach the opposite shore, though, the Japanese returned to the fight, and only one survivor managed to return to the Marine perimeter.

Even as the fierce exchange continued, the Marines sighted the four boats making for the Warrior River from the sea. An approaching storm, kicking up a heavy surf, added to the difficulty of rescue. Under cover of the Marines&rsquo fire, the landing craft finally beached on the west shore, and the Bigger patrol clambered aboard.

One boat, its motor swamped by surf, drifted toward the enemy shore but was stopped by a coral head. The rescue was completed, though, by the timely arrival of two PT boats&mdashwhich came on the scene with all guns blazing. 11 While the

patrol boats raked the jungle opposite with 20-mm and .50 caliber fire, the Marines transferred from the stalled craft to the rescue ships and all craft then withdrew. A timely rain squall helped shield the retirement. Aircraft from Munda and PT boats provided cover for the return to Voza.

The time for withdrawal of the battalion from Choiseul was near, however, despite the fact that Krulak&rsquos force had planned to stay 8-10 days on the island. On 1 November, another strong patrol, one of a series sent out from the Voza camp to keep the enemy from closing in, returned to the Vagara to drive a strong Japanese force back towards Sangigai. From all indications, the Japanese defenders now had a good idea of the size of the Krulak force, and aggressive enemy patrols were slowly closing in on the Marines. Seton&rsquos natives on 3 November reported that 800 to 1,000 Japanese were at Sangigai and that another strong force was at Moli Point north of Voza.

After the recovery of the Bigger patrol from Nukiki, IMAC asked Krulak to make a frank suggestion as to whether the original plan should be completed or whether the Marine battalion should be removed. The Cape Torokina operation was well underway by this time, and IMAC added in its message to Krulak that Vandegrift&rsquos headquarters considered that the mission of the parachute battalion had been accomplished. Krulak, on 3 November, radioed that the Japanese aggressiveness was forced by their urgent need of the coastal route for evacuation, and that large forces on either side of the battalion indicated that the Japanese were aware of the size of his force and that a strong attack, probably within 48 hours, was likely. The Marine commander stated that he had food for seven days, adequate ammunition, and a strong position but that if IMAC considered his mission accomplished, he recommended withdrawal.

Commenting later on his situation at this time, Krulak remarked:&ndash

As a matter of fact, I felt we&rsquod not possibly be withdrawn before the Japs cut the beach route. However, we were so much better off than the Japs that it was not too worrisome (I say now!) The natives were on our side&mdashwe could move across the island far faster than the Japs could follow, and I felt if we were not picked up on the Voza side, we could make it on the other side. Seton agreed, and we had already planned such a move. Besides that we felt confident that our position was strong enough to hold in place if necessary. 12

On the night of 3 November, three LCIs appeared offshore at a designated spot north of Voza to embark the withdrawing Marines. In order to delay an expected enemy attack, the Marines rigged mine fields and booby traps. During the embarkation, the sounds of exploding mines were clearly audible. Much to the parachutists&rsquo amusement, the LCI crews nervously tried to hurry embarkation, expecting enemy fire momentarily. Krulak&rsquos battalion, however, loaded all supplies and equipment except rations, which were given to the coastwatchers and the natives. Embarkation was completed in less than 15 minutes, and, shortly after dawn on the 4th of October, the Marine parachute battalion was back on Vella Lavella.

Krulak&rsquos estimate of the Japanese intentions was correct. Within hours of the Marines&rsquo departure, strong Japanese forces closed in on the area where the parachute battalion had been camped. The enemy had been surprised by the landing

and undoubtedly had been duped regarding the size of the landing force by the swift activity of the battalion over a 25-mile front. Then, after the operation at Empress Augusta Bay got underway, the Choiseul ruse became apparent to the Japanese, who began prompt and aggressive action to wipe out the Marine force. The continued presence of the Allied group on Choiseul complicated the evacuation program of the Japanese, and, once aware of the size of the Krulak force, the enemy lost no time in moving to erase that complication.

Before the battalion withdrew, though, it had killed at least 143 Japanese in the engagement at Sangigai and the Warrior River, sunk two barges, destroyed more than 180 tons of stores and equipment, and demolished the base at Sangigai. Unknown amounts of supplies and fuel had been blasted and burned at Guppy Island. Mine field coordinates shown on the charts captured at Sangigai were radioed to the task force en route to Cape Torokina, vastly easing the thoughts of naval commanders who had learned of the existence of the mines but not their location. Later, the charts were used to mine channels in southern Bougainville waters that the Japanese believed to be free of danger.

The destruction of enemy troops and equipment on Choiseul was accomplished at the loss of 9 Marines killed, 15 wounded, and 2 missing in action. The latter two Marines were declared killed in action at the end of the war. 13

The effect of the diversionary attack upon the success of the Cape Torokina operation was slight. The Japanese expected an attack on Choiseul the raid merely confirmed their confidence in their ability to outguess the Allies. In this respect, the Japanese were guilty of basing their planning on their opponents&rsquo intentions, not the capabilities. There is little indication that enemy forces in Bougainville were drawn off balance by the Choiseul episode, and enemy records of that period attach little significance to the Choiseul attack.

This may be explained by the fact that the main landing at Cape Torokina took place close on the heels of Krulak&rsquos venture and the ruse toward Choiseul became apparent before the Japanese reacted sufficiently to prepare a counterstroke to it. Certainly, the size and scope of the landing operations at Empress Augusta Bay were evidence enough that Choiseul was only a diversionary effort.

Enemy reaction to the Allied moves was a bit slow. The Japanese knew that an offensive against them was brewing what they could not decide was where or when. The Seventeenth Army was cautioned again to keep a watchful eye on Kieta and

Buka, and General Hyakutake in turn directed the 6th Division to maintain a firm hold on Choiseul as well as strong positions in the Shortland Islands. Then, the Japanese defenders on Bougainville waited for the next developments.

After the Allied landings in the Treasurys, the Japanese thinking crystallized: Munda was operational Vella Lavella was not. Therefore, the only targets within range of New Georgia were the Shortlands or Choiseul. And based upon this reasoning, the Allies scarcely would attempt a landing on Bougainville before staging bases on Mono or Choiseul were completed. Reassured by this assumption, the Japanese relaxed, confident that the next Allied move would come during the dark quarters of the moon&mdashprobably late in November.

With the Allied move toward Choiseul, the Japanese were more convinced that the Allied pattern was predictable. With a firm foothold on Mono and Choiseul, the Allies would now move to cut Japanese lines and then land on the southern part of Bougainville in an attempt to seize the island&rsquos airfields. Basing their estimates on the increased number of Allied air strikes on Buka and the Shortlands, the Bougainville defenders decided that these were the threatened areas. All signs pointed to a big offensive soon&mdashprobably, the Japanese agreed&mdashon 8 December, the second anniversary of the declaration of war.

The enemy had no hint that such an unlikely area as Empress Augusta Bay would be attacked. The defense installations were concessions to orders directing that the western coast be defended, and the troops at Mosigetta&mdashthe only force capable of immediate reinforcement to the Cape Torokina area&mdashwere alerted only to the possibility that they might be diverted on short notice to the southern area to defend against an assault there.

Japanese sea and air strength was likewise out of position to defend against the Bougainville thrust. Admiral Mineichi Koga, commander in chief of the Combined Fleet at Truk, had decided earlier to reinforce Vice Admiral Jinichi Kusaka&rsquos Southeast Area Fleet and the land-based planes of the Eleventh Air Fleet at Rabaul so that a new air campaign could be aimed at the Allies in the South Pacific. This operation, Ro, to start in mid-October, was to short-circuit Allied intentions by cutting supply lines and crushing any preparations for an offensive. To Kusaka&rsquos dwindling array of fighters, bombers, and attack planes, Koga added the planes from the carriers Zuikaku, Shokaku, and Zuiho&mdash82 fighters, 45 dive bombers, 40 torpedo bombers, and 6 reconnaissance planes.

Koga&rsquos campaign, though, was delayed. Allied radio traffic indicated that either Wake or the Marshall Islands would be hit next, and to counter this threat in the Central Pacific, Koga sent his fleet and carrier groups toward Eniwetok to set an ambush for the Pacific Fleet. After a week of fruitless steaming back and forth, the Japanese force returned to Truk, and the carrier groups moved on to Rabaul. The Japanese admiral had at first decided to deliver his main attack against New Guinea, but the Treasurys landings

caused him to swerve towards the Solomons. Then, when Allied activities between 27 October and 1 November dwindled, the fleet again turned toward New Guinea to take up the long-delayed Ro operation. The elements of the Japanese fleet reached the area north of the Bismarcks on 1 November, just in time to head back towards the Solomons to try to interrupt the landings at Cape Torokina.

Transcribed by Chris Gage


1. Unless otherwise noted, the material in this section is derived from: ComSoPac Oct-Nov 43 WarDs ThirdPlt NarrRept IMAC AR-I IMAC C-2 Repts, 27 Oct-13 Dec 43 (Bougainville Area OpFile, HistBr, HQMC), hereafter IMAC C-2 Repts IMAC C-2 Jnl, 27 Oct-27 Nov 43 (Bougainville AreaOpFile, HistBr, HQMC), hereafter IMAC C-2 Jnl IMAC D-2 MiscRepts (Treasury Is), 27 Oct 43, dtd 10 Nov 43 (Treasury Is AreaOpFile, HistBr, HQMC) 8(NZ) BrigGru Rept on Ops, Treasury Is (Op GOODTIME), dtd 30 Nov 43 (Treasury Is AreaOpFile, HistBr, HQMC) 8(NZ) BrigGru OpO No. 1, Op GOODTIME, dtd 21 Oct 43 (Treasury Is AreaOpFile, HistBr, HQMC) ONI, Combat Narrative XII: Henderson, &ldquoNaval Gunfire Support&rdquo Rentz, Bougainville and the Northern Solomons Miller, Reduction of Rabaul Morison, Breaking the Bismarcks Barrier.

2. Quoted in Gillespie, New Zealand History, p. 148.

3. Unless otherwise noted, the material in this section is derived from: ComSoPac Oct-Nov 43 WarD SoPacFor CIC. Study of Choiseul Island, dtd 19 Sep 43 (Choiseul AreaOpFile, HistBr, HQMC) IIIPhibFor AR IMAC AR-I, Anx Q, BLISSFUL IMAC C-2 Jnl IMAC OpO No. 2, dtd 22 Oct 43 (Choiseul AreaOpFile, HistBr, HQMC) CO, 2nd ParaBn, PrelimRept, Op BLISSFUL, dtd 5 Nov 43 (Choiseul AreaOpFile, HistBr, HQMC) 2nd ParaBn UnitJnl, 27 Oct-4 Nov 43 (Choiseul AreaOpFile, HistBr, HQMC) 2nd ParaBn OpO No. 1, dtd 23 Oct 43 (Choiseul AreaOpFile, HistBr, HQMC) MajGen Victor A. Krulak ltr to ACofS, G-3, HQMC, dtd 17 Oct 60, hereafter Krulak ltr Rentz, Bougainville and the Northern Solomons Isely and Crowl, Marines and Amphibious War.

4. IMAC Patrol Rept. on Choiseul Bay, 22-30 Sep 43, dtd 4 Oct 43: 3rd MarDiv Rept of Patrol to Kakasa, 6-13 Sep 43, dtd 16 Sep 43 (Bougainville AreaOpFile, HistBr, HQMC).

5. The Japanese uneasiness had an excellent basis. Sub-Lieutenant Seton &ldquohad organized 25 natives into a quasi-military force, armed them (Japanese weapons) and, on 2 October, ambushed an [enemy] group in a landing craft, killing seven.&rdquo Krulak ltr.

6. &ldquoThe Johnson light machine gun was more an auto-rifle than a machine gun: more a machine gun than the BAR.&rdquo Ibid.

7. DesDiv 44 AR for night of 27-28 Oct, Initial Landing of Marine Paratroopers on Choiseul Island, dtd Nov 43 USS Conway AR, 27-28 Oct 43, dtd 25 Nov 43 (Choiseul AreaOpFile, HistBr, HQMC).

8. This encounter left Krulak &ldquoin no doubt that we needed to go at them quickly, because they were obviously aggressive.&rdquo Krulak ltr.

9. The Marines with Krulak saw the first enemy position &ldquowithin a few minutes of E Company&rsquos opening fire. In this sense, the timing was extremely lucky. Had the enveloping column been 30 minutes slower, the Japs would have gotten away from E Co into the bush. As it was, the sentence in the operation order &lsquoPrevent enemy withdrawal into the mountains&rsquo (War Diary-1600 29 Dec) worked out well.&rdquo Ibid.

10. The original plan was for the boats to make two trips on the 30th, but by the time Company E got back to base it was getting dark. The battalion executive officer cancelled the return trip in view of the dangers of running the boats along the reef shelf at night. Krulak&rsquos radio had broken down and so he had no way of learning of this decision, although he guessed that this was the case. Still, it was an anxious night.

11. One of these boats was commanded by Lieutenant John F. Kennedy, USNR, later the 35th President of the United States. Krulak ltr.

13. This is the casualty figure given by Rentz, Bougainville and the Northern Solomons, p. 114. Few accounts of the Choiseul attack are in accord on Marine casualties. Muster rolls of the battalion indicate 9 KIA, 12 WIA, and 5 MIA. Of those missing, four were later declared dead and one believed a prisoner of war. IMAC C-2 Jnl, 4 Nov 43, and the report of the diversion attack, Operation BLISSFUL, p. 4, indicate that 9 KIA and 16 WIA is correct. IIIPhibFor AR, pp. 3-4, states that 8 KIA, 14 WIA, 1 MIA, and 1 captured is correct. IMAC AR-I, p. 11, gives the casualties as 8 KIA, 14 WIA, and 1 MIA. ONI, Combat Narrative XII, p. 24, gives the losses as 9 KIA, 15 WIA, and 2 MIA. The figure given by Rentz undoubtedly takes into cognizance a 13 Dec 43 message from Coastwatcher Seton to the effect that the bodies of two Marines, one of them bound as a prisoner, had been found near the Warrior River. ComSoPac Dec 43 WarD.

14. Unless otherwise noted, the material in this section is derived from: SE Area NavOps-III Seventeenth ArmyOps-II Rentz, Bougainville and the Northern Solomons Miller, Reduction of Rabaul Morison, Breaking the Bismarcks Barrier.

Subaru EZ30D Engine

The EZ30D engine had a die-cast aluminium cylinder block with 89.2 mm bores and an 80.0 mm stroke for a capacity of 2999 cc. Within the bores, the EZ30D engine had monoblock cast iron cylinder liners. Compared to the 2.5-litre EJ251 four-cylinder engine, the EZ30D was 20 mm longer and 40 kg heavier.

For the EZ30D engine, the water jackets were independent for the right and left halves of the block. The block halves were bolted together with 19 bolts (all located on the right bank of the engine), while the cylinder block and transmission cased were attached with 11 bolts.

The crankshaft for the EZ30D engine was supported by seven main bearings.


  • At heavy loads below 3600 rpm, the valve remained closed to create separate intake runners, increasing intake air speed and creating a ram effect for greater power and,
  • At heavy loads and high engine speeds (i.e. above 3600 rpm), the valve would open to reduce airflow resistance – this created a ‘scavenging’ effect that allowed each cylinder to receive high-pressure waves of intake air as they rebounded off the opposite side of the manifold.

Cylinder head and camshafts

The EZ30D engine had an aluminium cylinder head with chain-driven double overhead camshafts (DOHC) per cylinder bank. The camshafts were made from carbon steel pipes and had sintered metal lobes. During construction, the lobes were affixed to the steel pipes using a sintered metal paste the camshafts were then baked until the paste hardened. For the EZ30D engine, the lobes of the camshafts were offset by 1 mm to rotate the camshaft bucket and shim to reduce wear.

The EZ30D engine had two timing chains – the left timing chain had 148 links and the right timing chain had 134 links. The timing chains were sprayed with oil from a jet located on the oil pump relief valve housing and did not require maintenance since oil pressure and spring tension maintained chain tension.

Valve timing

Injection and ignition

The EZ30D engine had sequential, multipoint fuel injection via ‘air assist’ fuel injectors the injection and firing order were 1-6-3-2-5-4. Dual air/fuel sensors were used to monitor fuel mixture after combustion and continual adjustments were made to reduce exhaust gas emissions.

The EZ30D had a direct ignition system with an individual ignition coil for each cylinder (i.e. ‘coil-on-plug’), eliminating the need for a distributor and spark plug wires. Dual knock sensors enabled the ECU to adjust ignition timing in response to combustion noise feedback.

The EZ30D engine had a compression ratio of 10.7:1 95 RON premium unleaded petrol was specified so that the maximum possible ignition angle could be utilised.


The EZ30D engine had one exhaust port per cylinder head and three catalytic converters: one in each of the left and right exhaust manifolds and another in the mixing chamber. The E-OBD system monitored the operation of the front catalytic converters.


En la incursión participaron entre 656 y 725 hombres del 2º Batallón de Paracaidistas de los Estados Unidos , dirigido por el teniente coronel Victor Krulak . El 27 de octubre, en Vella Levalla, la fuerza atacante se embarcó en ocho lanchas de desembarco LCM antes de ser transferida a cuatro transportes de alta velocidad (APD): McKean , Crosby , Kilty y Ward , que acababan de ser liberados de apoyar el desembarco de Nueva Zelanda. tropas en las Islas del Tesoro. El número exacto de tropas japonesas en la isla ascendía a entre 3.000 y 7.000, y estaban bajo el mando del mayor general Minoru Sasaki . La mayoría de estas tropas se basaron alrededor de Kakasa y alrededor de la bahía de Choiseul, donde se mantuvo un pequeño puerto para varias barcazas. Escoltado por el destructor USS Conway , el convoy que transportaba al grupo de asalto estadounidense avanzó durante la noche hacia Choiseul en ruta fue atacado por un solo avión japonés, y uno de los APD estuvo a punto de fallar. Justo antes de la medianoche, el convoy llegó al área de reunión a unas 2,000 yardas (1,800 m) de la costa del lugar de aterrizaje elegido en Voza, en el lado noroeste de la isla. Un grupo de reconocimiento avanzado fue enviado a tierra en botes de goma y luego las Compañías F y G llevaron a los marines a tierra y, a pesar de esperar resistencia, aterrizaron sin oposición a principios del 28 de octubre. Durante el desembarco, el destructor de escolta, Conway , mientras estaba parado frente a la playa de desembarco, fue atacado. por un avión japonés con varias bombas. Estas bombas fallaron y el destructor mantuvo el fuego para que no llamara la atención sobre las operaciones de aterrizaje.

Después de que se estableció la cabeza de playa, la lancha de desembarco comenzó a transportar suministros y equipo a tierra, y con la ayuda de alrededor de 80 porteadores locales organizados por Seton que habían desembarcado con los marines, el batallón de Krulak estableció una base sólida en una meseta de 1,000 yardas (910 m). al oeste de Voza, con comunicaciones por radio y puestos avanzados de alerta temprana establecidos al norte y al sur. Desde allí, Krulak tenía la intención de llevar a cabo incursiones en un frente amplio para simular una fuerza mayor, con acciones que eventualmente se llevarían a cabo alrededor de Sangigai y en la parte occidental de la isla hasta Nukiki y el río Warrior en un área que abarca aproximadamente 25 millas (40 millas). km). Poco después de la llegada, como parte del plan de engaño, Krulak envió un mensaje de radio no codificado al IMAC de que había desembarcado una fuerza de 20.000 en la isla que estaba a punto de comenzar las operaciones. En dos horas se habían descargado las provisiones y los APD y la escolta del destructor habían partido del área para el viaje de regreso a Vella Levalla. Cuatro lanchas de desembarco LCVP y sus tripulaciones navales se destacaron en este momento, detalladas para ayudar a los marines a moverse hacia arriba y hacia abajo por la costa. Los aviones japoneses atacaron la playa de desembarco poco después de que los marines llegaran a tierra, y luego nuevamente durante el día y la noche, incluido un ataque a la isla de Zinoa, donde los marines habían escondido sus lanchas de desembarco. Estos ataques resultaron ineficaces ya que los marines habían dispersado rápidamente su equipo. También se hicieron esfuerzos para disfrazar la cabeza de playa y se construyó un sitio ficticio más al norte para desviar la atención de los japoneses.

Durante la tarde del 28 de octubre, Krulak envió una patrulla para explorar una base de barcos PT a lo largo de la costa occidental. Mientras tanto, la red de inteligencia local de Seton también envió patrullas para identificar las posiciones defensivas japonesas más cercanas, determinando que la concentración principal estaba alrededor de Sangigai, con otro grupo más pequeño a 30 millas (48 km) al noroeste. El 29 de octubre, se enviaron patrullas de reconocimiento al norte y al sur para recopilar información sobre las disposiciones, defensas y avenidas de aproximación japonesas y realizar bocetos de campo de lugares importantes otras patrullas, con especialistas adjuntos del Ejército y la Armada, reconocieron Moli Point y varios otros lugares para determinar su idoneidad como bases de botes PT o como ubicaciones para estaciones de radar. Una de las patrullas atacó y mató a siete japoneses y hundió una barcaza durante una patrulla hacia el río Vagara cerca de Sangigai. Más tarde, se envió una fuerza de bloqueo a Vagara, donde se enfrentó con un pelotón de japoneses. En respuesta, Krulak determinó que lanzaría un ataque contra la principal posición japonesa en Sangigai al día siguiente para interrumpir cualquier intento de atacar su base alrededor de Voza.

El 30 de octubre, dos compañías (E y F), lideradas por Krulak y apoyadas por tropas de apoyo equipadas con cohetes y morteros y una sección de ametralladoras, atacaron la base de barcazas en Sangigai. Partiendo a las 04:00 horas, se dirigieron hacia las lanchas de desembarco alrededor de la isla Zinoa, con la intención de utilizarlas para avanzar por la costa. Se había organizado un ataque aéreo preliminar, que consistía en 12 torpederos Grumman TBF Avenger , escoltados por 26 cazas. En cambio, estos aviones ametrallaron por error la nave de desembarco estadounidense, dañando una de ellas. Como resultado, las dos compañías de asalto de Krulak se vieron obligadas a abrirse camino hacia su objetivo, a pie, lideradas por el australiano Seton y sus guías locales. Mientras tanto, el ataque aéreo golpeó a Sangigai a las 06:10 horas. Un enérgico tiroteo tuvo lugar alrededor de Vagara alrededor de las 11:00 horas cuando los centinelas japoneses dispararon contra los marines que avanzaban, pero finalmente se vieron obligados a retirarse a la posición principal japonesa alrededor de Sangigai. Las dos compañías marinas se dividieron para lanzar el ataque desde dos direcciones. La espesa jungla y el terreno accidentado frenaron el avance de la Compañía F, bajo el mando de Krulak, que se estaba moviendo hacia el interior para adoptar una posición de flanqueo, y aún se estaba moviendo hacia su posición cuando estalló el fuego al norte del pueblo alrededor de las 14:00 horas. En este momento, la Compañía E, bajo el mando del capitán Robert E. Manchester, había avanzado a lo largo de la costa, se retrasó solo unos minutos y puso en acción sus morteros y cohetes, bombardeando una cresta mantenida por los japoneses a unas 500 yardas (460 m) al norte de la pueblo, y luego lanzó su ataque. Aseguraron la aldea unos minutos después. Los defensores japoneses se habían retirado de Sangigai, moviéndose hacia el interior donde avanzaron hacia el fuego de la Compañía F ubicada en el terreno elevado al este de la aldea cerca del río Kolombanara. A partir de las 14:30 horas, se produjeron nuevos combates durante los cuales los marines intentaron envolver y luego flanquear a los japoneses que recurrieron a varios ataques directos. Finalmente, después de aproximadamente una hora, los 40 supervivientes japoneses finalmente se dispersaron y se retiraron a la jungla. La batalla resultó en la muerte de 72 japoneses, mientras que los marines perdieron cuatro muertos y 12 heridos Krulak estaba entre los heridos. Posteriormente, los marines registraron la aldea en busca de inteligencia y destruyeron el equipo y las instalaciones japonesas restantes. Luego se retiraron a Vagara donde se estableció una posición defensiva nocturna. Por la mañana, el grupo de asalto se embarcó en varias lanchas de desembarco y regresó a Voza.

Durante el 31 de octubre y el 1 de noviembre, los infantes de marina emprendieron más acciones de patrullaje alrededor de su base en previsión de un contraataque japonés en este momento, ocurrieron varios enfrentamientos menores entre patrullas. Los japoneses también volvieron a ocupar Sangigai, aunque no atacaron la base de los marines alrededor de Voza. Los infantes de marina heridos fueron evacuados por un hidroavión consolidado PBY Catalina y se enviaron más suministros, incluida la comida para los guías locales de Seton. El 1 de noviembre, una gran patrulla de 87 infantes de marina de la Compañía G, bajo el mando del oficial ejecutivo del batallón, el mayor Warner T. Bigger, fue enviada al norte en lanchas de desembarco hacia Nukiki. Aterrizando cerca del río Warrior, la patrulla desembarcó con la intención de atacar a los japoneses alrededor de la bahía de Choiseul con sus morteros de 60 mm. Los botes estaban escondidos en una ensenada cerca de Nukiki y un destacamento de cuatro hombres quedó en un campamento base con una radio en la orilla oriental del río, junto con el equipo pesado y los explosivos. La patrulla partió luego hacia la Bahía Choiseul, pero luego de cruzar el río los guías asignados a la patrulla se perdieron y los marines se vieron obligados a establecer un vivac y permanecer durante la noche mientras un grupo más pequeño se destacaba para restablecer el contacto con Voza y solicitar que se hiciera otro guía. enviado.

Durante toda la noche, un grupo de japoneses cortó a los dos grupos. Sin embargo, el 2 de noviembre, el grupo más pequeño pudo volver a embarcar en Nukiki y regresar a Voza sin ser detectado, mientras que el grupo más grande bajo el mando de Bigger siguió adelante con su misión después de separar a varios hombres para regresar al río para solicitar la extracción en el tarde. Al ser atacado cerca de la isla Redman desde un pequeño puesto de avanzada que fue rápidamente destruido, Bigger decidió atacar a su objetivo alternativo, el suministro japonés y el depósito de combustible en la isla Guppy. Se dispararon ciento cuarenta y tres ráfagas de mortero desde la playa, incendiando la base. Cuando los marines comenzaron a retirarse, los japoneses desembarcaron una fuerza en la costa detrás de ellos para intentar interrumpir su retirada. Luchando contra cuatro ataques separados, la fuerza de Bigger estableció una posición en la orilla occidental del río Warrior, para esperar la llegada de los barcos para retirar la fuerza. En este punto, la incursión se detuvo momentáneamente cuando una emboscada japonesa atrapó entre 40 y 50 marines. Tres infantes de marina resultaron gravemente heridos, uno de ellos de muerte. Diez de los infantes de marina fueron recogidos y rescatados por el barco torpedero a motor PT-59 , bajo el mando del teniente John F. Kennedy , aunque el fuego del PT-59 le dio tiempo a la Marina para rescatar a muchos otros infantes de marina supervivientes a bordo de otro barco PT. PT-236 , cubierto por apoyo aéreo desde aviones basados ​​en Munda.

Después de la extracción de la fuerza de Bigger, el comando estadounidense comenzó a considerar retirar a los marines de Krulak de Choiseul. Los japoneses, al darse cuenta del pequeño tamaño de la fuerza que se enfrentaba a ellos, habían adoptado patrullas cada vez más agresivas hacia Vagara, donde se habían producido más enfrentamientos de patrullas el 1 de noviembre. La información recopilada por Seton de los habitantes locales indicaba que los japoneses habían reunido entre 800 y 1.000 hombres en Sangigai. , mientras que otras tropas se concentraban al norte de la base marítima de Voza. En consulta con Seton, Krulak hizo los preparativos para moverse a través de la isla si su base estaba aislada de la playa, pero finalmente los comandantes estadounidenses decidieron retirar a los marines antes de que esto sucediera. Luego, los infantes de marina se retiraron de la isla a bordo de tres lanchas de desembarco LCI en la mañana del 4 de noviembre, luego del exitoso alojamiento de las tropas aliadas en Cabo Torokina. Para cubrir su retirada, se colocó un campo de minas y se colocaron otras trampas explosivas Mientras los marines cargaban sus suministros y equipo, se escucharon varias explosiones mientras las patrullas japonesas entraban al campo minado. Después del embarque, los marines regresaron a Vella Lavella, escoltados por cinco barcos PT bajo el mando del teniente Arthur H. Berndston.

Câble à fibres optiques blindé

Fondé en 2010, Dongguan Qingying Industrial Co., Ltd (QY) est une société sophistiquée de fibre située dans deux villes, Dongguan et Chongqing, en Chine. Avec plus que le secteur de l'usine 10,000m2, QY est maintenant une société avec la capacité intégrée de la R&D, de la production, du commerce et des ventes internationales.

QY est un ISO9001 : société certifiée par 2015 avec des centaines de clients en Europe, l'Amérique, l'Asie du Sud-Est et Australie. Nos produits principaux sont connecteur installable sur place optique de fibre (connecteur rapide), corde de correction, corde de correction blindée, tresse, diviseur de PLC, atténuateur, et beaucoup d'autres produits de FTTH.

Avec une expérience de dix ans d'industrie optique de fibre, QY croit qu'apprenant, coopérant, et partageante avec des clients est la meilleure manière de se développer avec des clients. Pendant la prochaine décennie, QY sera agressivement de développer des produits nouveaux par demande de clients, et améliore la qualité des produits existants pour des clients. Les employés de tout le QY attendent avec intérêt la portion et Co-travaillent avec des clients pour des affaires de fibre à l'avenir.

Watch the video: The IPV-1 System Patrol Craft. an imperial vessel that sort of looked like an egg