Operation Watchtower: Why Guadalcanal, and the Landings of 7–8 August 1942

Higgins boats carrying men of the 1st Marine Division run in toward the beach near Lunga Point, Guadalcanal, in the first days of Operation Watchtower, August 1942. The landing east of the Lunga met no opposition at the waterline. U.S. Navy photograph 80-G-30521 (National Archives); public domain, via Wikimedia Commons.

Guadalcanal · Chapter 1 of 9

Operation Watchtower: Why Guadalcanal, and the Landings of 7–8 August 1942

How an unfinished Japanese airstrip on a British island pulled the United States into its first offensive of the Pacific war, and what happened on the first two days.

2 July – 8 August 1942Guadalcanal · Tulagi · Gavutu–Tanambogo~32 min read

At 0909 on Friday, 7 August 1942, local time, the first landing craft carrying assault troops of the 5th Marines grounded on a stretch of sand east of Lunga Point on the north coast of Guadalcanal. Nobody shot at them. Across the channel, twenty miles to the north, the 1st Marine Raider Battalion had been wading ashore on Tulagi for an hour through coral shallows that stopped its boats short of the beach, and by midday a parachute battalion fighting as ordinary infantry was under fire on the tiny island of Gavutu. Within thirty-six hours the Marines held the nearly finished Japanese airfield on Guadalcanal and, at a cost of 144 dead, the harbor islands across the sound. It was the first American amphibious offensive of the Second World War, and the start of a six-month campaign on land, at sea and in the air.

The operation was code-named Watchtower. The men who carried it out had a less dignified name for it. It was organized in about five weeks, after a June 1942 victory at Midway that nobody had planned for in advance, against an objective that most of the planners had not heard of until the first week of July. The 1st Marine Division had been told it would have six months of training in the South Pacific before it saw action. It got roughly six weeks, much of it spent unloading and reloading its own ships in the rain at Wellington, New Zealand.

A note on clocks and sources. Times are local Guadalcanal time (GMT+11) unless marked; Japanese records generally used Tokyo time (GMT+9), two hours behind. The narrative rests mainly on the three American official histories: John Miller Jr.’s Army volume Guadalcanal: The First Offensive (1949); Major John L. Zimmerman’s Marine Corps monograph The Guadalcanal Campaign (1949); and Pearl Harbor to Guadalcanal (1958) by Frank O. Hough, Verle E. Ludwig and Henry I. Shaw Jr. Where they disagree, as they often do on troop counts, the competing figures are given. The night of 8–9 August, off Savo Island, is the subject of Chapter 2.

Before the war: the Protectorate and its people

Guadalcanal is the largest island of the southern Solomons, a double chain of volcanic and coral islands running roughly northwest to southeast for some 900 miles east of New Guinea. Its spine is a range of mountains that rises steeply above a rain-forested southern coast; on the north side the mountains fall away to foothills, grassland ridges and, along the Lunga River, a coastal plain of coconut plantations and kunai grass. That plain was the only extensive flat ground in the southern Solomons that could take a large airfield with little effort, a fact that both sides came to appreciate in 1942. Miller’s Army history places the center of the north coast 565 nautical miles southeast of Rabaul, the Japanese stronghold on New Britain.

The islands were the British Solomon Islands Protectorate, declared in the 1890s and administered from the small island of Tulagi, tucked against the much larger Florida (Nggela) Island across the channel from Guadalcanal. Tulagi’s harbor was the best protected anchorage in the southern Solomons, and the island held the Resident Commissioner’s offices, the police barracks, the hospital, a club, a cricket ground, Chinese-owned stores and the handful of wharves that served the copra trade. Nearby Gavutu, linked by a causeway to the even smaller Tanambogo, carried wharves and trading-company buildings and, from 1939, a seaplane ramp built by the British administration and the Royal Australian Air Force. Most of the Protectorate’s population, however, lived nowhere near these little European outposts. Solomon Islanders, Melanesian and Polynesian communities speaking dozens of languages, lived in coastal and hill villages across the group, and many of the men had worked on the plantations or served in the Armed Constabulary.

The Pacific war reached the Protectorate before the Japanese did. According to the Solomon Islands Historical Encyclopaedia, Europeans began leaving Tulagi in late December 1941, women and children first, and the last European civilians sailed on the SS Morinda on 8 February 1942. Most of the Chinese community was left behind; some later helped the coastwatchers, some found passage to Sydney or Nouméa, and the rest went into hiding in villages. Japanese aircraft bombed Gavutu on 22 January 1942, and on 11 February the Resident Commissioner, William Marchant, moved his headquarters to Auki on Malaita and then into the hills behind it, taking or sending away the government’s land records. A Solomon Islands Defence Force was raised from the prewar constabulary and volunteers and eventually numbered about 800 men. The Anglican bishop of Melanesia, Walter Baddeley, also withdrew to Malaita and stayed in the islands.

A few officials and planters stayed on Guadalcanal itself, hidden in the hills with teleradios supplied by the Royal Australian Navy. They were part of the coastwatching organization built up before the war by Lieutenant Commander Eric Feldt under the Navy’s Directorate of Naval Intelligence, which Miller describes as a branch of the Australian Coastwatching Service. Among them was the district officer for Guadalcanal, a 27-year-old Scot named Martin Clemens. Their reports, and the islanders who carried food, messages and warnings for them, are a thread that runs through every chapter of this collection; the Coastwatchers of the Solomon Islands page tells their story in detail.

The Japanese arrive: a seaplane base at Tulagi and an airfield at Lunga

Japan’s southern advance in early 1942 had moved faster than its own planners expected. Rabaul fell on 23 January, Buka in the northern Solomons on 13 March, and fighter strips went in on Bougainville. The next step was a two-pronged operation to take Port Moresby on the south coast of Papua and to establish a seaplane base at Tulagi, from which flying boats could patrol the Coral Sea and the approaches to New Caledonia and Fiji. After heavy air raids on 1 and 2 May, a detachment of the Kure 3rd Special Naval Landing Force came ashore at Tulagi on 3 May from the minelayer Okinoshima without opposition; the small Australian and British party had withdrawn after coastwatchers reported ships approaching. Zimmerman’s monograph names the detachment commanders as Lieutenant Juntaro Maruyama on Tulagi and Lieutenant (junior grade) Kakichi Yoshimoto on Gavutu.

The next morning, 4 May, aircraft from USS Yorktown struck the shipping in Tulagi harbor and sank the destroyer Kikuzuki, the opening blow of the Battle of the Coral Sea. That battle, fought over the following four days, cost the Japanese the light carrier Shōhō and turned back the Port Moresby invasion force. It did not dislodge them from Tulagi. Through May and June the Yokohama Air Group moored its big four-engined Kawanishi H6K flying boats in lines off Gavutu and Tanambogo, and in early July floatplane fighters arrived to operate from Halavo Bay on Florida. Pacific Wrecks, using Japanese sources, counts 536 naval personnel at the Tulagi base, besides laborers; Zimmerman put the figure “about 500 on Tulagi” alone.

Across the channel on Guadalcanal, the Japanese at first did nothing about airfields. Zimmerman notes that survey parties went ashore near the Lunga about a month after Tulagi was taken, that the survey was finished late in June, and that construction began “in earnest” early in July. Two naval construction units, the 11th and 13th, came ashore with their equipment on 6 July; with guard troops and specialists they made up the force that Miller’s footnotes put at about 2,230 men in the Lunga area and that Shaw, drawing on later research, puts at 2,571. Many of the workers were laborers rather than soldiers; Shaw’s 1992 account describes the men who fled the landing as mostly Korean laborers. Miller writes that the airfield was meant to base sixty naval aircraft and was to be finished by 15 August. Local men from Guadalcanal villages were also put to work, a detail recorded by Pacific Wrecks.

The airfield was what changed American plans. The coastwatchers on Guadalcanal reported the clearing and grading near the Lunga, and American aerial reconnaissance confirmed it. Miller records that word of the Japanese move reached the Joint Chiefs of Staff in Washington on 6 July. On 17 July, two officers of the 1st Marine Division staff, Lieutenant Colonel Merrill B. Twining and Major William McKean, rode along on a B-17 reconnaissance flight to look at the north coast for themselves; Shaw writes that what they saw and photographed showed no extensive beach defenses. Zimmerman adds that Japanese fighters chased a B-17 away before it could photograph Lunga Point, a reminder that the reconnaissance picture was patchy.

King’s offensive: the 2 July directive and Tasks One, Two and Three

The idea of a counteroffensive in the South Pacific belonged above all to Admiral Ernest J. King, Commander in Chief of the U.S. Fleet and Chief of Naval Operations. King had argued since early 1942 that the line to Australia had to be held, and that the best way to hold it was to move forward up the Solomons toward Rabaul. Midway, on 4–6 June, sank four Japanese fleet carriers and gave him the opportunity. (The codebreaking that made Midway possible is told on the Station Hypo page.) Miller records that on 25 June King proposed a task force under Admiral Chester W. Nimitz to strike about 1 August, with immediate objectives in the Solomons and the Santa Cruz Islands and the New Guinea–New Britain area as the final goal.

There was an immediate argument about who would command. The Solomons lay inside General Douglas MacArthur’s Southwest Pacific Area, and the Army’s planners, backed by General George C. Marshall, favored a direct attack on Rabaul under MacArthur, who had proposed exactly that in June. The Navy would not risk its few remaining carriers in the confined waters off New Britain under an Army commander. Miller presents the dispute as primarily between King and Marshall, with MacArthur’s objections running alongside it, and records that the two chiefs settled it between 29 June and 2 July.

The result was the “Joint Directive for Offensive Operations in the Southwest Pacific Area Agreed on by the United States Chiefs of Staff,” issued on 2 July 1942. Its ultimate objective was the New Britain–New Ireland–New Guinea area. It divided the work into three tasks:

To put Task One under Nimitz without moving it out of MacArthur’s area by fiat, the directive moved the boundary between the Southwest Pacific Area and the South Pacific Area west to longitude 159° East, effective 1 August. The line ran just west of Guadalcanal and Tulagi, placing the lower Solomons inside the South Pacific Area, whose commander, Vice Admiral Robert L. Ghormley, had formally assumed command at Auckland, New Zealand, on 19 June. Guadalcanal was not named in the directive. The decision to add it came days later, as the reports of the airfield arrived; within a week, the planners were treating the airfield as the central objective and Tulagi as the secondary one.

Neither of the two commanders in the theater thought it could be done on time. Ghormley flew to Melbourne on 7 July to confer with MacArthur, and on 8 July they sent a joint dispatch recommending that Task One be postponed. Their reasons, as Miller summarizes them, were that only one amphibious division was available and heavy losses would leave it unable to support later tasks; that the transports would be exposed for about two days off Guadalcanal and Tulagi beyond the reach of land-based fighter cover; and that Allied air strength in the theater was too small to protect the invasion. They wanted to wait until there were forces enough to carry out all three tasks “in one continuous movement.” The Joint Chiefs refused on 10 July. The Japanese might finish their airstrip. The only concession was time: on 16 July Ghormley told Nimitz that the second echelon of the Marine division had reached Wellington late and that bad weather had slowed the loading, and set D-Day for 7 August.

“Operation Shoestring”: a division in New Zealand

The force available for Task One was the 1st Marine Division, and it was not ready. Its commander, Major General Alexander A. Vandegrift, had brought it from the United States expecting a long period of training in New Zealand; Shaw writes that the division had been assured six months in the South Pacific before it saw action. Division headquarters and the 5th Marines reached Wellington on 14 June. The second echelon, including the 1st Marines, was still at sea. The 7th Marines had been detached to Samoa, so the 2nd Marines of the 2nd Marine Division, coming from San Diego, were attached in its place, together with the 1st Marine Raider Battalion, the 1st Parachute Battalion and the 3rd Defense Battalion, an antiaircraft and coast-defense unit. Ghormley called Vandegrift to Auckland on 26 June and told him what was coming.

The division’s ships had been loaded in the United States for an administrative move, with cargo stowed to use every cubic foot rather than in the order it would be needed on a hostile beach. For an assault it had to be “combat loaded,” with each transport carrying a complete battalion landing team and its weapons, ammunition, rations and vehicles stowed so that the first items needed came out first. That meant unloading everything at Wellington and loading it again. The work was done at Aotea Quay, where, Zimmerman notes, there were “labor difficulties” with the unionized waterfront workers. In the end the Marines did nearly all of the dockside work themselves, with civilians operating only the cranes and loading machinery. Miller records that embarkation of the first combat group began on 2 July and was complete by 13 July, and that each ship had a working party of 300 men in three eight-hour reliefs around the clock.

The second echelon arrived on 11 July and had to unload and reload between then and 22 July. It was the southern winter. Miller describes cold, driving rain and low morale, and cardboard ration cartons that turned soggy and burst on the quay, spilling their contents into the mud. Shaw calls it “a never-to-be-forgotten scene on Wellington’s docks.” Not everything would fit, and some of the division’s vehicles and heavier equipment stayed behind in New Zealand, a decision that the shortages of August and September would make painful.

The haste was apparent to everyone involved. In a 1992 Naval History article, Commander Daniel J. Franken describes the operation as so hastily planned that it was called “Operation Shoestring.” The historian Robert Citino, writing for the National WWII Museum, describes it as “a colossal improvisation, concocted on the fly,” and notes that the Marines, “with their sardonic sense of humor,” had a better name for it than Watchtower. Intelligence was thin. Lieutenant Colonel Frank B. Goettge, the division intelligence officer, flew to Australia in early July to collect maps and interview former residents of the islands; Zimmerman records that the maps he arranged never reached the division, and that the coastal sketch the Marines used “contained no usable indications of ground-forms or elevations.”

Estimates of the Japanese garrison ran high. Ghormley’s intelligence staff estimated about 3,100 men in the Tulagi–Guadalcanal area, Turner’s plan used 7,125, and the division’s own estimate of 20 July reached 8,400. Turner’s figure, as Zimmerman gives it, assumed 1,850 on Tulagi and Gavutu–Tanambogo and 5,275 on Guadalcanal. The lowest of these, Ghormley’s, turned out to be closest to the truth.

The commanders and the plan

Ghormley’s Operation Plan No. 1-42 of 16 July laid out the command chain. Ghormley himself, who King had assumed would command Task One in person, stayed at his headquarters, first in Auckland and then aboard a ship at Nouméa, New Caledonia. Tactical command went to Vice Admiral Frank Jack Fletcher as commander of the Expeditionary Force, Task Force 61, which Miller counts at eighty-two ships. Under Fletcher were two main forces. The Air Support Force under Rear Admiral Leigh Noyes, twenty-six warships and five tankers in three groups, was built around the carriers Saratoga, Enterprise and Wasp, with the new battleship USS North Carolina in the Enterprise group. The Amphibious Force under Rear Admiral Richmond Kelly Turner, Task Force 62, had twenty-three transports and cargo ships and twenty-eight warships.

Within Turner’s force, the British officer Rear Admiral Victor A. C. Crutchley, Royal Navy, commanded the screening group of Australian and American cruisers and destroyers, and the cruiser-destroyer fire support groups were assigned to bombard the two landing areas. The landing force was Vandegrift’s 1st Marine Division. Outside Fletcher’s command but essential to it was Rear Admiral John S. McCain’s Task Force 63, which controlled all Allied land-based aircraft in the South Pacific, from Army B-17s to Navy Catalina patrol planes and New Zealand Hudsons. Miller counts 291 aircraft in McCain’s command by the end of July, scattered from New Caledonia to Fiji, Efate and the new strip at Espiritu Santo. McCain’s searches were supposed to watch the approaches from Rabaul; MacArthur’s air forces in Australia and New Guinea were to search and strike from the other side of the boundary.

Franken’s article identifies a weakness in this arrangement: Fletcher was both the overall tactical commander and, in practice, the carrier commander, and the carriers were the asset he was most anxious to preserve. Vandegrift was subordinate to Turner until the landing force was established ashore.

The plan itself was simple in outline. Turner’s transports were divided into Transport Group X-Ray, four transport divisions for Guadalcanal, and Transport Group Yoke, two divisions for Tulagi and the other harbor islands. On Guadalcanal, after a naval and air bombardment, Combat Group A, the 5th Marines less one battalion, would land on a 1,600-yard beach, designated Red Beach, about 6,000 yards east of Lunga Point, and push west along the coast. Combat Group B, the 1st Marines, would land behind it, pass through and drive southwest to seize a feature called the “Grassy Knoll” that commanded the plain. Across the sound, the Raiders and the 2nd Battalion, 5th Marines, would take Tulagi, and the 1st Parachute Battalion would take Gavutu four hours after the Tulagi landing; a company of the 2nd Marines would land first on Florida Island to cover the flank. The rest of the 2nd Marines was the division reserve. Miller’s strength figures give a total landing force of over 19,000; Zimmerman thought 18,722 probably close to the mark and noted division records of 19,105 and 19,546.

Koro: a rehearsal that failed and a time limit that held

The Amphibious Force sailed from Wellington at 0900 on 22 July. Zimmerman records that the various forces met on 26 July at a rendezvous 367 miles south of Fiji, and the combined Expeditionary Force reached the island of Koro, in the Fiji group, the next day. Rehearsals followed from 28 to 31 July (Miller; Zimmerman gives 28 to 30). They went badly. Coral reefs made the chosen beaches impractical for actual landings, so the boats approached and turned back. Landing craft had to be conserved, which made realistic practice impossible. Radio silence ruled out practicing communication between the troops and the supporting aircraft. Vandegrift later called the rehearsal “A complete bust,” and Miller quotes him weighing its “dubious” benefits against the loss of “priceless time.” It did give the boat crews practice in debarking troops and timing waves, and it brought the senior commanders face to face for the first time.

That meeting, held aboard Fletcher’s flagship Saratoga on 26 July, was the most consequential conference of the operation. Fletcher, Turner, Vandegrift and McCain attended; Ghormley did not, and neither did Crutchley. According to Shaw, Fletcher told the others that he could not “keep the carriers in the area for more than 48 hours after the landing.” Franken describes Turner and Vandegrift as “stunned.” Turner had planned on several days of carrier cover while his transports unloaded; Vandegrift argued that he needed four days to get his supplies ashore. Fletcher’s reasoning, as later historians have reconstructed it, was that his three carriers were half of the U.S. Navy’s operational carrier strength in the Pacific after the loss of Lexington at Coral Sea and Yorktown at Midway, and that keeping them in a fixed area within range of Rabaul’s bombers and Japanese submarines invited a repetition. Shaw writes that Fletcher reluctantly agreed to stay a little longer than he first proposed; Miller states that he intended to withdraw the carriers before D+3, short of the time needed to unload completely, and that Ghormley knew this.

Miller records that it was also agreed at Koro to land first at Tulagi and later the same morning at Guadalcanal, and that McCain said it would take ten days to fit the Navy’s F4F Wildcats with the extra fuel tanks needed to fly up from the New Hebrides to the captured airfield. The point was clear to everyone at the table. The Marines would have the carriers for two days, perhaps three. After that, until aircraft were flying from Lunga, the transports and the men ashore would depend on what McCain’s long-range planes could do from 560 miles away.

The approach under cloud

The Expeditionary Force left the Koro area on 31 July on a southwesterly course, as if bound for Australia, and then turned north. On 4 August the transports Betelgeuse and Zeilin joined from Nouméa, and Miller counts the whole force from that point at eighty-two ships carrying over 19,000 landing troops. On 5 and 6 August the weather closed in: overcast, haze and rain squalls cut visibility to about four miles, and the ships kept radio silence. Zimmerman calls the conditions “extremely favorable” and notes, from Japanese sources, that bad weather grounded the Japanese patrol planes at Rabaul on both days. No Japanese search plane found the force.

The force passed south of Guadalcanal and rounded its western tip, Cape Esperance, in the dark. Miller records that the ships raised Savo Island, the round volcanic cone at the western mouth of the sound, at 0200 on 7 August. At 0240 the two transport groups separated: Group Yoke, with Crutchley’s cruisers covering, turned north of Savo toward Tulagi, while Group X-Ray steamed south of Savo toward the Guadalcanal shore. The carriers stayed south and west of Guadalcanal, out of sight of land. On Tulagi and at Lunga, the Japanese were asleep.

7 August: Red Beach

At 0614 the cruisers and destroyers of the Guadalcanal fire support group opened fire on the coast; the Tulagi group followed two minutes later. Carrier aircraft came over at first light. Miller counts 44 carrier planes over Guadalcanal and 41 over Tulagi in the first strikes, and records that eighteen Japanese seaplanes at the Tulagi anchorage were destroyed. The 25th Air Flotilla war diary at Rabaul, quoted by Zimmerman, logged a report from Tulagi that an enemy force of twenty ships had entered and was bombarding the shore while preparing to land. Miller sums up the Japanese surprise in one line: until the afternoon “not one shot had been fired at the Amphibious Force.”

The X-Ray transports reached their positions about 9,000 yards off Red Beach around 0650 and began lowering their boats. Vandegrift confirmed Zero Hour for Guadalcanal as 0910. The first wave, the 1st and 3rd Battalions of the 5th Marines, 1/5 on the right and 3/5 on the left, crossed the line of departure on time. Shaw puts the first touchdown at 0909; Hough’s official history at 0910. The beach was empty. There were no Japanese positions, no wire and no fire. The assault battalions moved inland a few hundred yards and set up a perimeter while the regimental headquarters came ashore at 0938 and the 1st Marines began landing in a column of battalions from about 0930, the 2nd Battalion leading.

The Japanese on the plain had fled. The bombardment and the sight of the transports had sent the construction workers and their guards west toward the Matanikau River and the villages beyond it, leaving their camps, their stores and their unfinished airstrip behind. The Marine official history describes an enemy who had retreated “in evident haste and confusion,” leaving equipment and stores in place. There was no plan for a defense of Lunga. The construction units had not been organized or armed to fight a division, and the nearest Japanese army troops were hundreds of miles away.

The advance inland was slower than the planners had imagined. The 1st Marines moved off southwest at 1115 toward the Grassy Knoll and soon found that the objective did not exist where the plan said it did. Hough’s history explains that the Grassy Knoll, “visible from the ships, could not be seen from the beach,” and that it lay much farther inland than reported; it was in fact Mount Austen, several miles from the coast and far beyond a day’s march. Colonel Cates’s regiment struggled through grass, jungle and stream beds in heavy heat, carrying loads its men had not been conditioned for in the weeks aboard ship, and by late afternoon had covered about a mile. It dug in for the night. On the coast, 1/5 crossed the mouth of the stream the Marines’ maps called the Tenaru at 1330 and moved toward the Ilu. (The Marine maps transposed several river names along this coast, a confusion that runs through the American accounts; Chapter 4 explains it.) The objective was abandoned. Vandegrift ordered the division to turn west toward the airfield instead.

The real problem on 7 August was the beach. The plan had assigned only about 490 men of the 1st Pioneer Battalion to unload the boats at Red Beach, and Miller notes that many of the boats lacked bow ramps, so every crate had to be lifted over the gunwales. Amphibian tractors, used in combat for the first time, proved useful but were too few. While the rifle companies sat idle inland, supplies piled up at the waterline faster than the shore party could move them. By nightfall, Miller writes, about 100 landing craft were waiting to unload and another 50 were standing offshore; at 2330 the shore party commander reported the situation “entirely out of hand.” The Marine Corps history records that the division asked the ships to stop sending cargo, and that its officers later concluded they had needed “at least 100 men for each vessel” to handle it.

Tulagi: Raiders against the Kure landing force

Tulagi is a narrow, wooded ridge of an island, about two miles long, with a spine Zimmerman puts at about 350 feet high running through its northwestern two-thirds and a cluster of hills, ravines and the former government buildings at its southeastern end. The Japanese expected any landing to come at the southeastern end, where the wharves and beaches were. The Marines therefore landed on the western side, at a place designated Beach Blue, where coral reefs extended well offshore. Zimmerman describes the boats grounding on the coral 30 to more than 100 yards from the beach, and the Raiders of Companies B and D wading in through water from waist to armpit deep. There was no fire from the shore; Zimmerman records one Raider killed early in the day by an accidental discharge.

Ten minutes after the Raiders landed, the garrison radioed Rabaul:

Enemy troop strength is overwhelming. We will defend to the last man.

— Japanese radio message from Tulagi, about 0810, 7 August 1942, as translated from the 25th Air Flotilla war diary and quoted in Zimmerman, The Guadalcanal Campaign (1949), ch. II

The Raiders turned southeast and swept down the island with Companies B and D on the high ground and along the north shore, Company A and Company C following. They met little opposition at first and reached Phase Line A, across the island, at 1120; Major Justice M. Chambers, commanding Company D, was wounded there. Beyond that line resistance hardened. Company C fought for nearly an hour to clear Hill 208, and the advance slowed as it reached the Japanese defensive area around the old government station, the cricket ground and a steep-sided ravine. The defenders were men of the Kure 3rd Special Naval Landing Force and the naval air personnel of the seaplane base, fighting from caves and dugouts in the coral. The 2nd Battalion, 5th Marines, under Lieutenant Colonel Harold E. Rosecrans, had landed behind the Raiders at 0916, cleared the northwest end of the island and then moved to support them. By late afternoon Edson reported that about 500 Japanese had withdrawn into the southeastern ridge and its ravine, around Hill 281, and the Raiders dug in facing them.

That night the Japanese attacked. Hough counts four separate attacks against the Raider lines during the darkness, followed between 0030 and 0530 by quiet infiltration toward Edson’s command post. One attack struck the refused flank of Company A, where, according to the official history, the Marines killed 26 Japanese within 20 yards of their line. Small groups got through to the rear areas, and the fighting that night was close and confused. None of the attacks broke the line. The official accounts describe the Japanese as fighting in small groups, with little coordination, and almost none surrendering.

On 8 August Companies E and F of 2/5 passed through the Raider lines, and the Marines closed in on the ravine and Hill 281 from both ends. The caves were cleared one at a time with grenades, explosives and rifle fire. After an attack at 1500 the last pocket in the ravine was overrun, and by nightfall of 8 August, Hough writes, “Tulagi was labeled secure.” Individual Japanese and small groups were hunted for several days afterward, and a few escaped by swimming to Florida Island. Miller’s Army history records that only three of the Tulagi garrison surrendered. The 1st Raider Battalion and its commander, Merritt Edson, would become central to the defense of Guadalcanal itself a month later; see the Edson’s Raiders page and Chapter 5.

Gavutu and Tanambogo: the hardest fight of the landings

Gavutu and Tanambogo are two small knobs of coral and rock, each crowned by a hill, joined by a causeway a few hundred yards long. Gavutu’s Hill 148 overlooked the seaplane ramp and the old prewar buildings; Tanambogo’s hill overlooked Gavutu. The islands had been the main base of the Yokohama Air Group’s flying boats, and their defenders, naval aviation ground crews, landing-force troops and laborers, had had three months to dig in. Unlike Tulagi, there was no undefended beach to land on. Vandegrift’s order set the assault for H plus four hours, so that the boats could be reused after the Tulagi landing, and the naval bombardment that preceded it had wrecked the seaplane ramp the Marines were to land on.

The 1st Parachute Battalion landed near noon, in Shaw’s words “in three waves, 395 men in all.” The parachutists had not jumped; there were no transport aircraft to carry them, and they went in as light infantry from Higgins boats. The first wave came ashore near the ramp under fire from both islands. Company A got ashore without casualties, but Companies B and C, following, were caught in crossfire from Gavutu’s Hill 148 and from Tanambogo. Vandegrift later estimated losses of about 10 percent among the troops in that first landing area. Major Williams was wounded early, and Major Charles A. Miller took command. Fighting from cave to cave up the slope, the battalion held most of Gavutu by 1430, though the north face of Hill 148 remained exposed to fire from Tanambogo. Brigadier General Rupertus, commanding the Tulagi area, initially reported casualties of 50 to 60 percent among the parachutists; Zimmerman later attributed that figure to “inadequate communications.”

Tanambogo still had to be taken. Rupertus sent Captain Crane’s Company B of the 2nd Marines, which had already landed and re-embarked at Haleta on Florida that morning. It reached Gavutu at about 1800 and, after a short naval bombardment, attempted a landing on Tanambogo after dark. A shell set a fuel dump on fire, lighting up the beach and the boats, and the Japanese opened fire on everything in the water. One boat crew was almost entirely killed or wounded. Crane withdrew with his wounded; about a dozen men left ashore found their way back during the night, two of them by rowboat, the rest reaching Miller’s command post on Gavutu around midnight. On Gavutu itself, a strong group came out of a cave under the southern slope of Hill 148 in the night and counterattacked without success.

On 8 August the division reserve was committed. The 3rd Battalion, 2nd Marines, under Lieutenant Colonel Robert G. Hunt, landed on Gavutu in four waves beginning at 1000. Hough records that a misdirected bomb from an American dive bomber fell on the positions of Company K, killing three Marines and wounding nine. Light tanks were landed at 1315 to support an attack across the causeway and from the sea. On Tanambogo one tank, separated from its infantry, was stopped and set afire by Japanese who swarmed it at close quarters; the official history records 42 Japanese bodies found around it afterward. Lieutenant Sweeney was killed in the fighting on Tanambogo. Company I landed on Tanambogo at 1620 and, with Company K pushing across the causeway, had taken the southeastern two-thirds of the island by 2100. Organized resistance ended the following day, 9 August. Zimmerman describes the defenders as fighting from “dugouts and caves in the coral” until they were killed.

How many Japanese died on the three islands will not be known exactly; Zimmerman and the Marine official history both say so in nearly identical words. The official history estimates 750 to 800 Japanese in the Tulagi–Gavutu–Tanambogo area, of whom 23 were taken prisoner and about 70 escaped to Florida Island, according to a later intelligence summary. Shaw’s figures imply 886 in the area; Zimmerman’s monograph gives about 1,500 including some 600 laborers, a figure later research reduced. Pacific Wrecks, drawing on Japanese records, notes that most of the prisoners taken on Gavutu–Tanambogo were Korean laborers. On any count, the garrison was almost entirely destroyed. The pattern the Marines met here, of small groups fighting from caves and almost none surrendering, would recur across the Pacific for three years.

Rabaul strikes back: the air attacks and the coastwatchers

Rabaul reacted within hours. The Japanese naval air units there, the 25th Air Flotilla with the Tainan Air Group’s Zero fighters and twin-engined Mitsubishi bombers of the 4th Air Group, were sent south against the landing area. For the bombers and their escorts it was a round trip of more than 1,100 miles; Richard Frank’s history describes it as the longest fighter mission of the war to that date. Accounts of the size of the first strike differ. Hough’s history speaks of about 18 bombers; Miller of 25; later Japanese-based research summarized in recent accounts gives 27 Type 1 bombers escorted by 17 or 18 Zeros. A separate formation of nine Aichi D3A dive bombers flew from Rabaul at the very limit of their range.

The first warning came from southern Bougainville, where Paul Mason, a planter turned coastwatcher, was hiding in the hills above Buin with a teleradio. The Australian Dictionary of Biography records his signal as “Twenty-four bombers headed yours.” Hough writes that the warning passed through the watchers’ network and was relayed to the fleet, by way of the Allied stations in Australia, within about 25 minutes. That gave the carrier fighter director time to put Wildcats in the path of the strike. Shaw writes that Fletcher’s pilots met the Japanese about 20 miles northwest of the landing area. The transports got underway and maneuvered, which stopped unloading, and the bombers, attacking from high altitude at about 1320, hit nothing of importance. A second wave, the dive bombers, came in about 1500 and hit the destroyer Mugford with a 250-pound bomb, causing about 20 casualties. Miller estimates the two raids cost about three hours of unloading time.

The fighting in the air on 7 August was costly for both sides. The Naval History and Heritage Command’s H-Gram 009 estimates that about half of the U.S. Navy fighters that met the strikes, some twenty, were shot down over the two days, while the Japanese bombers “suffered severe losses and inflicted minimal damage.” Japanese claims were wildly inflated: the H-Gram records that the Japanese air commander claimed two heavy cruisers, a cruiser, two destroyers and nine transports sunk. Shaw counts 21 Wildcats lost by the carrier squadrons on 7 and 8 August. Among the Japanese pilots that day was Petty Officer First Class Saburō Sakai of the Tainan Air Group, already one of Japan’s leading fighter pilots. After shooting down a Wildcat, he attacked a formation of American carrier bombers he took for fighters (accounts differ on whether they were Dauntlesses or Avengers) and was hit in the head by their return fire. Partly blinded, he flew his damaged Zero back to Rabaul, a flight of more than 550 miles that took several hours, and survived; his wounds kept him out of combat for nearly two years.

On 8 August the Japanese came back with torpedo bombers. Again the warning came from Bougainville. Shaw credits it to Jack Read, the assistant district officer at Buka Passage; the Marine official history names the watcher as Pilot Officer Cecil John Mason, RAAF. Turner stopped unloading again and got his transports underway in formation. About noon some forty twin-engined bombers came in low over Florida Island and across the sound to drop torpedoes, into heavy antiaircraft fire and attacks by carrier fighters. The Marine official history credits ships and fighters with shooting down 12 and shore antiaircraft with two more. The bombers lost heavily, but they did damage.

The transport USS George F. Elliott, which had carried the 2nd Battalion, 1st Marines, from Wellington, was struck by a bomber that, after repeated hits from antiaircraft fire, swerved and, in the words of her Navy ship history, “crashed into the ship amidships.” Fires spread out of control and ruptured the water mains. Salvage proved impossible, and the gutted ship was sunk the same day, with much of her cargo still aboard. The destroyer USS Jarvis was hit by a torpedo that tore a large hole in her side. She was ordered to retire to the south for repairs, sailed alone that night past Savo Island, and the next day was found by Japanese bombers and lost with all hands; the H-Gram puts her loss at 233 men. Her story belongs to Chapter 2.

On 8 August, too, the first Japanese attempt to reinforce the islands ended at sea. The Naval History and Heritage Command records that the old submarine S-38 sank the transport Meiyo Maru on 8 August, thwarting the enemy’s “first attempt to reinforce their ground forces on Guadalcanal.” Neither side knew yet that this was the start of a long contest over who could get men and supplies to the island, a contest the Japanese would come to fight with destroyers running down “the Slot” at night.

8 August: the airfield and what the Japanese left behind

On Guadalcanal the advance resumed on the morning of 8 August with a new objective, the airfield. At 0930 the 1st Battalion, 5th Marines, with Company A of the 1st Tank Battalion, crossed the Ilu and moved west along the beach toward the Lunga, while the 1st Marines moved out from their night positions inland. Opposition remained negligible. Lieutenant Colonel Lenard B. Cresswell’s 1st Battalion, 1st Marines, reached the airfield, by Miller’s account by about 1600, and overran it. The Japanese had made no attempt to destroy it. At 1430 the 5th Marines crossed the Lunga north of the airfield and pushed on toward Kukum, the Japanese camp and wharf area west of Lunga Point, which Miller says was taken by 1500 after a small rearguard fight. A little after 1500, Hough records, Company D reached the main Japanese encampment.

What they found surprised everyone. Miller’s list, compiled from the division’s records, includes a nearly completed runway; two radio stations; ice plants; two electric power plants; an air compressor plant; rifles, machine guns and 70-mm and 75-mm guns with ammunition; gasoline; two radar sets; and about 35 serviceable trucks. There were warehouses of rice, tea, hardtack, dried kelp, noodles, canned goods, beer and sake. There were also diaries and documents, which became valuable sources of intelligence. Hough’s history describes “undamaged food, ammunition, engineering material, electrical gear,” and radio equipment left behind by an enemy who had retreated “in evident haste and confusion.” The Marines adopted the ice plant at once; Shaw records that they would soon be eating Japanese canned fish and rice, smoking Japanese cigarettes and filling Japanese rice bags as sandbags.

The captured equipment mattered more than the food. When the transports left the next day, they took with them most of the division’s heavy engineering gear, which had not yet been unloaded. Shaw notes that captured Japanese gear helped complete the runway; the trucks and hand tools left at Lunga did work the division’s own missing equipment should have done. On 12 August a Navy PBY Catalina landed on it, and on 20 August the first Marine fighters and dive bombers flew in. By then it had been named Henderson Field, and the code name for the island, Cactus, had given its aircraft their name. That story is the subject of Chapter 3.

The view from Rabaul

For the Japanese, the landings were a surprise in every sense. Their planners did not expect a major American offensive in the South Pacific so soon after Midway, and the reports from Tulagi on the morning of 7 August at first suggested a raid rather than an invasion. The H-Gram records that the Japanese army at Rabaul, Lieutenant General Harukichi Hyakutake’s 17th Army, dismissed the landings as a reconnaissance in force; its attention was fixed on the overland advance on Port Moresby across the Owen Stanley Range in New Guinea, which had begun in July. The airfield at Lunga was a navy project; the army had no troops on Guadalcanal and, at first, little reason to think the island its problem.

The navy reacted more quickly. Vice Admiral Gunichi Mikawa had arrived at Rabaul in late July to command the newly formed 8th Fleet, responsible for the outer South Seas, including the Solomons and New Guinea. On 7 August, while the air flotilla was sending its bombers south, Mikawa assembled every heavy ship within reach: five heavy cruisers, two light cruisers and a destroyer. The H-Gram records that he had them underway by mid-afternoon on 7 August, and that the Navy General Staff in Tokyo, under Admiral Osami Nagano, at first rejected his plan as too reckless before accepting it on the recommendation of Admiral Isoroku Yamamoto. Mikawa intended to run down the channel between the islands, later called the Slot, and attack the transport anchorage at night.

The Japanese on the islands experienced the landings very differently from their commanders. On Tulagi and Gavutu–Tanambogo, landing-force troops and air-unit ground crews who had expected a quiet rear-area posting fought until they were killed, many of them in caves. Their final messages to Rabaul, preserved in the 25th Air Flotilla’s war diary and translated after the war, are some of the few contemporary Japanese records of the fighting. On Guadalcanal, the construction workers, many of them Korean laborers and civilians, fled west into the jungle with what they could carry. They had lost their camp, their food and their tools in a single morning. Within weeks the survivors, joined by soldiers landed from destroyers, would be living on short rations in the jungle west of the Matanikau, the beginning of the hunger that led Japanese soldiers to call the island Gatō (餓島), “Starvation Island.” Chapter 8 follows that story.

The cost of two days, and Fletcher’s request

By the evening of 8 August the Marines held everything they had been sent to take except the last corner of Tanambogo, which fell the next day. On Guadalcanal the cost had been slight; the official histories record no organized resistance at Red Beach or on the way to the airfield and give no separate casualty figure for it. In the harbor islands it had been heavy: 144 Marines dead and 194 wounded. The Navy had lost George F. Elliott, had Jarvis torpedoed and Mugford damaged, and had lost about a fifth of its carrier fighters. Against that, the harbor garrison had been all but destroyed and the airfield was in American hands.

Unloading had fallen far behind. The air attacks had twice stopped work for hours, the beach was choked, and on the night of 7–8 August the division had asked the ships to stop sending cargo ashore. Red Beach was extended west toward the Lunga on 8 August, but most transports and cargo ships still held much of their loads. Miller writes that less than half of the planned sixty days of supplies was eventually landed; he puts food on hand at about thirty days. Shaw, counting differently, writes that the forces ashore had 17 days’ rations after counting captured Japanese food, and only four days’ supply of ammunition.

Then came two pieces of news. During the morning of 8 August an Australian Hudson patrol plane had sighted Mikawa’s ships east of Bougainville, but its report reached Turner only hours later and did not make clear that a cruiser force was coming. Turner received word of the approaching Japanese force at about 1800. At 1807, Hough and Miller record, Fletcher sent a message to Ghormley citing a fuel shortage and plane losses that had reduced his fighters “from 99 to 78,” and again requested permission to withdraw his carriers until sufficient land-based aircraft and fuel were available to support the shipping. Ghormley consented. The carriers retired southward the next morning.

Without the carriers Turner did not believe he could keep his transports in the sound under daylight air attack. That evening he called Vandegrift and Crutchley to a conference aboard his flagship McCawley off Lunga Point, and told them that the Amphibious Force would leave at 0600 the next morning. Hough records that Vandegrift considered the retirement left him in a “most alarming” position. To attend, Crutchley took his flagship, HMAS Australia, out of the screening force guarding the western approach south of Savo Island. While the three men talked in McCawley’s cabin, Mikawa’s cruisers were already coming down the Slot. The Battle of Savo Island began about 0140 on 9 August.

Fletcher’s withdrawal remains one of the most argued decisions of the Pacific war. The case against him, made by Morison and many Marine veterans, is that his fuel was not critically short, that 78 fighters was still a strong force, and that by leaving early he exposed the transports and the cruisers at Savo and left the Marines without air cover or much of their supply. The case for him, made most fully by Lundstrom, is that the carriers could not be replaced in the near term, that Japanese carriers were known to be somewhere to the north, and that he had told everyone at Koro what he would do. The Navy would lose Wasp and Hornet in the next three months. Either way, the Marines would spend the next twelve days with no friendly aircraft overhead, holding a perimeter around an airfield they could not yet use.

Two days that began in silence on Red Beach ended with the airfield taken, the harbor islands won at heavy cost, the transports preparing to leave and a Japanese cruiser force bearing down on the anchorage. The improvisation that the Marines called Operation Shoestring had achieved its first object almost exactly as planned. What it could not yet do was keep what it had taken. That would require ships the Navy was about to lose, aircraft that had not yet arrived, and six months of fighting.