Meanwhile, Vandegrift had to fortify his shoreline against an expected enemy counterinvasion. An intelligence alert had warned that a Japanese invasion force was gathering at Rabaul and was expected to arrive within four days. Preparation of defenses required manpower, diverting men from the unloading work. Turner refused Vandegrift’s request to be reinforced with the 1,400 officers and men of the 2nd Marines, which the admiral continued to insist on holding in reserve for possible deployment to Ndeni.
Throughout the morning, ships filed out of Sealark Channel. Hundreds of marines watched glumly as the battle-mauled Chicago, still smoking and with most of her bow blasted away, vanished beyond Taivu Point. By sundown, none remained. A supply officer estimated that fewer than half of the crates packed into the transports in New Zealand had been landed. The marines, now alone, were precariously short of such vital necessities as food, ammunition, heavy artillery, antiaircraft guns, communications equipment, barbed wire, fuel, and spare parts. And still they wondered: where was the Japanese army, and when and where would it reveal itself?
Chapter Three
ADMIRAL KING, SOUND ASLEEP ON HIS DOCKED FLAGSHIP DAUNTLESS at the Washington Navy Yard, was shaken awake in the early morning hours of August 12. “Admiral, you’ve got to see this,” said his duty officer, who had never before interrupted the boss’s sleep. “It isn’t good.”1
With disbelieving eyes, King read Turner’s dispatch reporting the loss of four Allied cruisers with heavy loss of life, and the hurried withdrawal of the transports and cargo ships from Ironbottom Sound. He asked that the dispatch be decoded again, in the vain hope that it was somehow mistaken. It was not.
The news kicked King in the teeth. WATCHTOWER was his invention, his hobbyhorse, and his responsibility. He had insisted on the risky expedition with full knowledge that Allied shipping resources and airpower were strained to the snapping point. He had trodden over the well-reasoned joint objections of the region’s two theater commanders. “That, as far as I am concerned, was the blackest day of the war,” he later said. “The whole future became unpredictable.”2
The next morning, in his office on the second “deck” of Main Navy (the headquarters building on Constitution Avenue), he studied the track charts forwarded by Turner and tried to envision how the Japanese fleet could have stolen into Ironbottom Sound undetected. “I just can’t understand it,” he admitted to Admiral Harry W. Hill, who had dropped in to see him.3 The analogy to Pearl Harbor was impossible to ignore. The earlier surprise attack had inflicted greater casualties and material damage, but the beating at Savo had been meted out in wartime, against ships operating in enemy-dominated seas, when their commanders and crews ought to have been hypervigilant to every likely threat. The navy’s honor, reputation, and self-respect were on the block.
FDR received the news from his naval aide, Commander John L. McCrea, who drove the dispatch from Washington to “Shangri-La,” the president’s rural presidential retreat in Maryland (later renamed Camp David). The president, McCrea recalled, “was heartsick about it. There wasn’t anything he could do about it.”4
King ordered that news of the defeat be concealed from the press. He dispatched two trusted officers, Admiral Arthur J. Hepburn and Captain DeWitt C. Ramsey, to fly to Noumea to investigate its causes. Not for a moment did he consider scaling back the commitment to WATCHTOWER, however—indeed, he moved at once to reinforce Ghormley. In a memorandum to the president on August 13, King outlined his plan to send the battleships South Dakota and Washington, accompanied by a cruiser and six destroyers, through the Panama Canal and on to Noumea. Two more cruisers would be transferred from Britain to the east coast and kept in readiness for possible transfer to the Pacific.5 On the same day, King asked General Marshall to provide more army air units to the region, “regardless of commitments elsewhere. . . . In my opinion, the Army Air Forces in Hawaii and the South Pacific Area must be reinforced immediately to a much greater extent than appears now to be in prospect.”6 Marshall acquiesced without dissent, even though the USAAF chief Henry “Hap” Arnold was adamantly opposed to diverting any of his strength from Europe or the pending invasion of North Africa. Marshall was perfectly aware that King had FDR’s ear, and that the commander in chief wanted it done.
As Turner’s ships limped back toward Noumea, survivors of the sunken cruisers began the grim work of reconstructing the fatal events. The senior surviving officer of each ship was required to submit an action report—but the loss of records and the death or disablement of many key witnesses and participants required them to work largely from memory. All knew the action would receive intense scrutiny. Admiral Crutchley, to his credit, did not mince words. On the morning of August 10, he wrote Turner: “The fact must be faced that we had an adequate force placed with the very purpose of repelling surface attack and when that surface attack was made, it destroyed our force.”7 How had it happened?
Hepburn and Ramsey’s investigation shone a harsh light on the failure of air and submarine reconnaissance to discover Mikawa’s force and alert the task force to its approach. The division of the region into two theaters, one controlled by Ghormley and the other by MacArthur, had posed a communications hitch right at the vital boundary that Mikawa crossed on the night of August 8–9. Several Allied planes had spotted the Japanese ships farther up the Slot, but some of the sighting reports failed to get through to Turner in time, and others were inaccurate or incomplete. A report of “seaplane tenders” among the Japanese ships threw Turner off the scent—he assumed they must be headed for Rekata Bay on Santa Isabel and would launch a seaplane torpedo attack on August 9. Admiral McCain’s PBYs had been foiled by bad weather, but McCain had failed to inform Turner that the patrol flights had not occurred. Turner might have used the cruiser planes in his task force to conduct his own air searches, but he did not.
A fateful convergence of errors and bad luck was behind the debacle, and no senior naval commander in the WATCHTOWER expedition could count himself entirely blameless. Hepburn’s report concluded that the defeat could not be attributed to a single root cause. Captain Bode, of the Chicago, was the only officer formally censured (for failing to send an alert when his ship was attacked). In his comments on the report, submitted to Secretary of the Navy Frank Knox five weeks after the battle, King stressed that “this was the first battle experience for most of the ships participating in the operation and for most of the flag officers involved, and that consequently it was the first time that most of them had been in the position of ‘kill or be killed.’ . . . They simply had not learned how and when to stay on the alert.”8 The officers and men had been on Condition One Alert, with virtually the entire crew on watch, for more than forty-eight hours. The men were human; they could not function indefinitely without rest. Collective physical and mental exhaustion had overcome the task forces, rendering them vulnerable to surprise attack.
A more contentious question was Fletcher’s abrupt decision to withdraw the aircraft carriers on the afternoon of August 8, which the Hepburn report called “a contributory cause” of the disaster. Had he stayed until the morning of the ninth, as previously planned, Fletcher could have done nothing to prevent the catastrophe off Savo Island—but it is possible that his air groups could have delivered retribution from the air on Mikawa’s fleeing column. Fletcher’s early withdrawal has remained one of the livelier controversies of the Pacific War. It drew pungent criticism from Turner and Vandegrift, both of whom seemed to have regarded it as a personal betrayal. In his memoir, the normally even-tempered Vandegrift used the incendiary term “running away,” with its blunt intimation of cowardice, to describe Fletcher’s departure.9 Fletcher’s decision was pilloried by the influential Samuel Eliot Morison in his quasi-official history of U.S. naval operations in the Second World War.
The subject has been parsed, scrutinized, and debated by generations of historians. Little more can be usefully said, except to bring out some of the most salient points. Fletcher, in almost perfect contrast
to Turner, seemed content to accept history’s judgment of his conduct. He did not, like Turner, review and provide detailed notes on Morison’s draft manuscript covering the events of WATCHTOWER. He published no memoir. Retiring in 1947 to a farm in rural Maryland, he largely removed himself from the cut and thrust of historical debate. He either forgot or falsely denied that he and Turner had clashed, during the July 26 planning conference on the Saratoga, over the question of how long the carriers would stay. “Turner and his staff were very pleased” with the arrangements made at the conference, Fletcher told a New York Times reporter in 1947, and added: “At no time was there any friction between Turner and myself.”10 That was plainly inaccurate, as several other witnesses have attested.
In his dispatch to Ghormley on the evening of August 8, Fletcher had offered two reasons for his proposed withdrawal: the low-fuel state of his task force, and heavy losses of F4F fighters on August 7 and 8. In preparing his volume on the Guadalcanal landing (volume 5, The Struggle for Guadalcanal) Morison obtained the navy’s records of the actual fuel state of each of the ships in Fletcher’s task force, and showed that Fletcher’s fuel situation was far from critical. All the cruisers were at least half full. The destroyers’ fuel condition varied, but none had less than 40,000 gallons of fuel, and their daily fuel expenditure ranged from 12,000 to 24,000 gallons. Morison’s biting conclusion: “Thus it is idle to pretend that there was any urgent fuel shortage in this force. . . . Fletcher’s reasons for withdrawal were flimsy. . . . [H]is force could have remained in the area with no more severe consequences than sunburn.”11 John B. Lundstrom provides evidence that Fletcher may have received incomplete or inaccurate information about the fuel state of his screening vessels.12 The admiral could act only on the basis of what he knew. On the other hand, it is a task force commander’s responsibility to obtain proper reports from ships under his command.
The second issue, heavy fighter losses in air combat on August 7 and 8, cannot be idly dismissed. On Dog-Day, half of the American fighters that engaged the enemy in air combat were sent down in flames. Overall fighter losses in the two days had come to twenty-one, leaving reserves of seventy-eight Wildcats on the three carriers. Fighters, by a wide margin, were the most valuable weapon in Fletcher’s arsenal. They were the only aircraft that could properly defend Turner’s ships against air raids, but they were equally needed to fly cover over the carrier task forces throughout the daylight hours. Fletcher’s air operations during the first two days of WATCHTOWER had been the busiest in the history of carrier warfare. Wear and tear to equipment, and the simple exhaustion of the aircrews, were considerations that no responsible task force commander could afford to ignore.
The heart of the controversy was the value of the carriers themselves. Their best protection was constant movement and finding concealment in thick weather whenever possible. Operating for several days “chained to a post,” in a fixed location south of Guadalcanal, invited devastating counterattack by air or submarine. Japanese twin-engine medium bombers, armed with torpedoes, had the range to reach Fletcher’s task force from Rabaul. The submarine menace grew inexorably the longer his ships remained corralled in a finite geographic zone. Three months earlier, in these very same waters, Fletcher had lost the Lexington to air attack at the Battle of the Coral Sea. The Yorktown, his former flagship, had been shot out from under him at Midway. The Saratoga, his current flagship, had been torpedoed by a Japanese submarine in January and knocked out of action for four months. Though Fletcher did not yet know it, the Saratoga would catch another torpedo on August 31, and the Wasp, the newest carrier to arrive in the theater, would be destroyed by submarine attack in mid-September. The Hornet would succumb to air attack in October (at the Battle of the Santa Cruz Islands), leaving just one American flattop (the Enterprise) in the Pacific.
This much is incontrovertible: the risk that one or more American carriers would be lost during the WATCHTOWER expedition was not negligible. But how should that risk be balanced against Turner’s need for continued air protection? The carriers had provided most of the navy’s offensive striking power since escaping destruction on December 7, 1941. For the time being (until the new Essex-class carriers could be commissioned and brought into the fleet) they were scarce and valuable assets. Fletcher correctly assumed that the Japanese aircraft carriers would sooner or later come down into the lower Solomons. In that case it would be his overriding duty to fight another carrier slugging match like those of Coral Sea or Midway. It was in that context that Fletcher’s withdrawal must be judged.
The proper use of the carriers during WATCHTOWER was a first-order strategic question, and it should have been resolved in advance. That it was not is unsurprising, given how hastily WATCHTOWER had been planned and executed. Turner had hoped for a full five days to unload the division’s cargo, but plans for the operation had specified that the transport fleet would withdraw in three days, and neither Turner nor the marines had ever been promised longer than “two or three” days of carrier air cover. Neither King, Nimitz, nor Ghormley had provided clear instructions to resolve the discrepancy. The decision therefore fell to Fletcher, whom Ghormley designated the overall commander of the expedition. Many officers who took part in the July 26 command conference on the Saratoga were taken aback by the evident rancor between Fletcher and Turner and came away with the impression that Fletcher lacked confidence in the operation. Ghormley’s presence had been urgently needed at that conference—had he been at the table, he could have adjudicated the question and resolved any doubt. A reasonable share of culpability must be attributed to WATCHTOWER’S cumbersome and ambiguous command arrangements, and thus to King himself.
SUNRISE ON AUGUST 10 found Task Force 62 zigzagging generally east, at about the midpoint between Guadalcanal and Noumea, with the damaged Chicago and Ralph Talbot struggling to remain in company. Turner’s ships were loaded with hundreds of wounded marines and sailors, 831 hospitalization cases in all. Four submarine contacts were reported between dawn and dusk, and a torpedo wake was once observed to pass by the bow of the McCawley. The destroyers hunted the surrounding waters with depth-charge barrages, and the task force executed a series of radical turns. The stricken Chicago, falling behind, was ordered into Efate with a destroyer screen. No ship suffered a torpedo hit, and on August 15, the bulk of Task Force 62 arrived safely at Noumea Harbor. The wounded were removed to hospital ships by stretcher. All would be sent to Australia or the States.13
The 1st Division’s hopes now depended on constant and generous supply by sea. On August 15, four old flush-deck “four-piper” transports dropped anchor at Lunga Roads and began unloading supplies—aviation gasoline (400 drums), aerial bombs (300), .50-caliber ammunition, lubricating oil—and ground personnel for Marine Aircraft Group 21, which would fly in as soon as the airstrip was in a condition to receive planes. On the following afternoon, the Fomalhaut, loaded with heavy construction equipment for Guadalcanal and Tulagi, stood out of Noumea accompanied by three destroyers. On that same date, Turner asked McCain to “load all APDs [high-speed transports] to capacity with food and send one division to Guadalcanal and the other to Tulagi.”14
With 16,000 mouths to feed, Vandegrift grew concerned about his food reserves. According to Turner’s records, entered into his war diary on August 9, “Sufficient food was landed at the Tulagi area for about eleven days and at the Guadalcanal area for about thirty-six days.”15 The marines did not agree with those figures, however—Vandegrift radioed this on August 15: “From rations on hand and consumed to date estimate about twelve days’ rations landed Guadalcanal. Further loss due to weather and handling reduced this to ten days. No opportunity should be lost to forward rations to this command.”16 In the 1st Division’s final report, submitted in 1943, the figures recorded for August 15 were seventeen days of regular field rations, three days of Type C Rations, and another ten days’ supply of captured Japanese food. The discrepancies are likely explained by the spoilage of rations packed in cardboa
rd boxes, which tended to disintegrate in the rain, and by the uncertain amount of captured enemy food. At any rate, there is no doubt that the marines went hungry during those early weeks in the Solomons. Vandegrift ordered reduced rations on August 12, and most men subsisted on two meager meals per day. Private William Rogal, manning a foxhole on Tulagi, was issued one C ration per day “and sometimes not even that.” He and his fellow marines scavenged for Japanese provisions in warehouses along the Tulagi waterfront and found a few sacks of barley. They made a kind of barley soup, an awful mush, but ate it avidly. On the transports, Rogal recalled, sex had been the habitual topic of conversation among bored marines. Now it was food, because “the single overriding emotion during the weeks we existed in that jungle retreat was hunger!”17
Turner continued to pressure Fletcher to shield his cargo ships, but Fletcher was keen to keep his task force together and on the move. He anticipated a major Japanese counterstrike, and with good reason. The intelligence picture remained clouded, but air reconnaissance, coast-watcher reports, and bits and pieces of “Ultra” (decrypted enemy radio intercepts) seemed to portend a major Japanese fleet movement into the lower Solomons. Overflights of northern Bougainville confirmed that the Japanese were building new landing strips south of Buka Airfield.18 MacArthur’s B-17s flew over Rabaul and Kavieng almost every day and snapped aerial photographs. Between August 12 and August 16, these photos revealed a significant buildup of naval force, and the airfields at Rabaul appeared to have been reinforced with fighters and bombers. The Japanese were apparently constructing another airfield at Buin on southern Bougainville, about 400 miles northwest of Guadalcanal, which would give them a much closer springboard for air attacks. Considerable enemy air activity was reported at Lae, on the northern coast of New Guinea. On August 15, Japanese aircraft dropped six loads of ammunition and food to Japanese troops scattered on the western end of Guadalcanal (four fell within or near the marine lines and were captured). The Americans still had not pinpointed the location of Japanese carrier forces, but they could be at sea and on their way to the Solomons. On August 21, Fletcher told Ghormley that he considered it “inadvisable to send cruisers and destroyers into CACTUS [Guadalcanal] nightly,” because of their exposure to submarine attack.19 Better to keep them at sea, on the move, and prepared to repel the expected enemy naval offensive.