A round 1807 on 8 August, eleven minutes before sunset, the Saratoga broke radio silence (as opposed to TBS short range traffic) to transmit a dispatch from Fletcher to Ghormley, information Turner, McCain, and Noyes. “Total fighter strength reduced from 99 to 78. In view of large number of enemy torpedo and bomber planes in area recommend immediate withdrawal of carriers. Request you send tankers immediately to rendezvous decided by you as fuel running low.” That message resulted from discussion in flag plot and brief consultation via TBS with Noyes, in temporary tactical command of the carrier group. Colonel Maas, the marine aviator on the TF-61 staff, participated in the deliberation and knew all the factors that influenced the decision. Two hours later he wrote a private analysis of the situation that declared (his emphasis): “Withdrawal of carriers at end of second day not ordered, but should have been. Fletcher proposed it on his initiative.” Maas believed Fletcher should “be commended for his judgment, courage, and tactical farsightedness.”1
To Fletcher’s legion of critics message 080707 of August 1942 defined his naval career. In 1943 Turner bitterly complained to Morison that Fletcher left him “bare-arse.” In 1945 he officially characterized the action as nothing less than “desertion,” undertaken for reasons “known only” to Fletcher. Vandegrift described Fletcher, “Running away twelve hours earlier than he had already threatened during our unpleasant meeting.” Official judgments were equally unsparing. On 23 August 1942 Nimitz condemned the withdrawal as “most unfortunate,” a comment he repeated in The Great Sea War. In December Admiral Pye, president of the Naval War College, called the pullout “certainly regrettable,” which risked “the whole operation.” In 1943 Admiral Hepburn’s final Savo report labeled Fletcher’s action a “contributory cause of the disproportionate damage” incurred in that battle. The Cominch Secret Information Bulletin No. 2 gave short shrift to the reasons attributed to Fletcher’s decision, including needless worry about “bombing and torpedo planes.” The 1950 Naval War College analysis of Savo expanded on Pye’s original points and concluded, “Such a precipitous departure” would “seriously jeopardize the success of the entire operation” and “prevent the inauguration of Task Two.”2
Historians accepted the official judgments without question. Morison wrote, “It must have seemed to [Turner] then, as it seems to us now, that Fletcher’s reasons for withdrawal were flimsy.” The carriers “could have remained in the area with no more severe consequences than sunburn.” Marine Brigadier General Griffith, historian as well as participant, completely concurred. He conceded, “Fletcher did have a point. We just couldn’t lose our carriers, but damn it, how the Marines suffered!” Vice Admiral Dyer did present Fletcher’s side, but his sympathies lay with Turner. The situation did not justify the carrier withdrawal. That was also the carefully considered judgment of Richard Frank, who concluded Fletcher, rightly or wrongly, placed the preservation of his carriers ahead of everything else. All other accounts derive from these key analyses.3
However, the supposedly airtight case against Fletcher is deeply flawed. Turner obscured the fact that he always intended to pull out the Amphibious Force incrementally. A key message from Turner boding some change in plan never reached Fletcher in time. The depiction of the fuel state of some ships is shockingly inaccurate and highly prejudicial to Fletcher. The fairest and most proper basis for analyzing Fletcher’s recommendation is to recreate, as best as can be done at this late date, the situation as he and others actually understood it at the time. There are three crucial factors: (1) necessity—what Fletcher knew of the situation at Guadalcanal and whether he believed it was vital to keep the carriers there after 8 August, (2) security—if he thought the carriers would be safe in light of what he deemed was their prime mission, and (3) supply—his estimation of the fuel in TG-61.1 and whether it was a valid reason to consider in withdrawing.
Turner predicated the invasion plans on most of the Amphibious Force leaving by sundown on D+1, 8 August, if the situation permitted. Considerable debate on 27 July centered on how long Fletcher should keep the carriers in hostile waters in support of the portion of the invasion force that must stay behind, possibly for several more days. Fletcher’s intent, whether clearly understood by others or not, was to remain until sundown 9 August (D+2). Of course, no partial withdrawal of the Amphibious Force ever took place. Excepting the hurt destroyer Jarvis, Turner held all his ships on station until the evening of 9 August. Thus the first important issue concerns Turner’s actual intentions on 8 August, as opposed to his later representations. Did Turner change his plans? Did he keep his superiors properly advised of his situation?
Maas’s personal situation estimate on the evening of 8 August alleged Turner “did not keep Fletcher informed.” A few days later he called communications “very bad” and wrote that “most of the time” Fletcher “got no information” or “very inadequate dope, and mostly very late,—too late to be useful.” Turner, though, stoutly maintained in a 16 August 1942 report on communications, “A reasonably complete history of the essential features of the operations was promptly transmitted.” That is highly debatable, especially in light of his claim that he took it upon himself to provide “operational bulletins to the superior command echelons,” particularly as Fletcher “could not break radio silence to report progress.” It is vital to note Turner suffered no similar restraint. As evidence, he gave Ghormley the text of the radio messages the McCawley sent out on 7–9 August to CTF-61 on the task unit commander’s frequency (2122 kilocycles). He asserted “all were received” and assumed Comsopac copied these messages as well.4 He erred on both counts. Ghormley’s endorsement disclosed that he could not receive messages on 2122 kilocycles and contended, correctly, “It is doubtful” Fletcher “could hear more than fragments of the blind transmissions on that frequency.” In truth Fletcher received very few of Turner’s messages on 2122. Ghormley blamed the failure of Turner’s “higher echelon communications” on the selected frequencies (drawn from standard Pacific Fleet procedure), “plus a lack of equipment in the transports.”5 However, Turner did not originate all his messages from the McCawley. His report mentioned, but did not identify, eight messages delivered on the morning of 8 August by boat to the Astoria for immediate transmission on a different frequency. With one exception noted below, all were highly routine. Ironically all of the Turner messages that the Astoria transmitted on 8 August went through without any difficulty and certainly contributed to the impression all fared well at Guadalcanal.6
Early on 8 August, Ghormley prodded Turner: “Report situation.” Not until mid-morning did he and Fletcher copy a long message from CTF-62 that related events as of 2100 on 7 August. Message 071030 of August 1942 (2130, 7 August local time) does not appear in Turner’s list of the above-mentioned “despatch reports” transmitted from the McCawley and thus was one of the Astoria messages received straightaway on the eighth. Prefaced “Situation summary Watchtower,” Turner’s only detailed update for 7 August is worth quoting in full:
Approach a complete surprise 18 enemy seaplanes destroyed on water one small anchorage. No enemy ships present. In both areas opened shore bombardment 1910 GCT 6th [0610, 7th, Z-11] with ships and planes light gun opposition soon silenced. Landings began Florida area 2040 [0740] initial opposition light Guadalcanal 2210 [0910] no opposition.
At 0220 GCT 7th [1320] about 25 enemy Type 97 bombers passed overhead dropped several bombs no damage proceeded toward carriers 2 shot down 2 damaged. About 10 enemy Type 99 dive bombers made 1 250 pound hit on Mugford 0400 [1500] considerable damage after deck house 2 guns 2 engines 15 missing 5 dead 9 seriously injured. 2 enemy planes shot down.
Situation at 0900 [2100, 7th] Guadalcanal all troops ashore estimated occupy west line Tenare [sic] River and on east security line about Longitude 160–06 in neither places in contact with enemy. On Florida side Halavo occupied no opposition. Tulagi occupied except east and fight continues Gavutu captured heavy casualties Tanambogo still in hands of enemy attack underway.
Tomorrow request maximum fighter cover 2 VSB squadrons continuously in area. Request scouting against approach enemy surface force from westward.
Early morning 8th expect to send out Santa Cruz Occupation Force less Pres Jackson, Wilson, plus McCawley, Fuller, Heywood, Trever, Mugford, some other APs later in day. Tonight 2 DMS searching for fighters shot down near Russell Island. One Wasp dive bomber shot down by Zero according to pilot wounded gunner lost.7
The salient point is that Vandegrift was safely ashore with the whole Lunga force, eleven thousand men, holding a three-mile front without opposition, but still a mile short of the great prize, the airfield. It would be reasonable to assume that by dark on 7 August, the X-Ray Squadron transports (as opposed to the cargo ships) were empty or nearly so. The marines still cleaned out the Tulagi area. Turner desired “heavy air support” and “maximum fighter cover” on the eighth. One task was accomplished easily, the second a disaster. The carriers twice searched north and west on the eighth. The message demonstrated Turner’s desire to hold to the plan of withdrawing a significant portion of the invasion force on 8 August. The units mentioned were heavy cruiser Quincy, three destroyers, a reduced Ndeni force of three transports and one cargo ship, two other X-Ray transports (including the McCawley), one Yoke transport, one high-speed minesweeper, and damaged Mugford. That left four X-Ray and three Yoke transports, most of which at least should also depart that day. Turner’s original plan called for the four APDs of Transdiv Twelve to re-embark the First Raider Battalion once it finished with Tulagi. The five cargo ships were to anchor close to the Lunga shore for as long as it took them to unload. Nothing in Turner’s message indicated he had changed any of these arrangements.
Other than flash reports of the noon attack and a short summary at 1430 of the damage, Fletcher received no additional information from Turner prior to his recommendation. That afternoon Fletcher summarized for TF-11 the situation based on the radio messages and observations by his aviators. Lunga, Gavutu, and Tulagi were occupied, leaving only Tanambogo. There was “no serious damage our forces from air attacks either yesterday or today.” So far the offensive cost the enemy nearly forty planes, but, “Our plane losses [are] light in comparison.” Subsequently Saratoga fighter pilots advised Fletcher that one transport was on fire a mile north of Red Beach, and nearby a destroyer, “apparently in trouble,” moved close to shore. Turner’s 1430 summary of the attack (which got through to Fletcher) identified the George F. Elliott and Jarvis as the stricken ships. The VF-5 pilots also reported that when they left, the X-Ray transports, deployed three abreast in column, maneuvered off the north coast of Guadalcanal, while to the north the Yoke transports could be seen “apparently engaged in developing into formation.” These were further hints of a possible imminent departure by Turner’s force. The last air support flight returned without attacking, another indication the fighting was nearly over. Maas’s situation estimate demonstrates Fletcher believed that afternoon the “transports [were] unloaded,—all troops were ashore,” and that only the cargo ships had to continue unloading. Fletcher heard the previous evening that the transports at Tulagi had not yet started unloading, but nothing from Turner or anyone else pointed to further difficulties there. Thus as far as Fletcher knew, most of the invasion force should have no reason to linger. Wright commented in 1952, “We in the Minneapolis were manning the same circuits that Fletcher did and about all we knew about happenings in the Guadalcanal-Tulagi area was that the shore operations had been completed successfully.” That was clearly how Nimitz judged the situation based largely on the same few messages. The War Plans Section commented, “Evidently the operation is proceeding satisfactorily.”8
Such sanguine appreciations of Watchtower’s progress were dead wrong. Although Turner had composed his 7 August situation report by 2130 the same evening, he did not actually have the Astoria transmit it until a dozen hours later. In the meantime circumstances changed radically. Late on the seventh after learning of heavy casualties sustained at Tulagi, Vandegrift asked Turner for one of the two battalions of the Second Marines held in reserve afloat. Already earmarked for Ndeni, the whole regiment was supposed to sail to Santa Cruz on 7 August after collecting its First Battalion. Turner had already postponed the Ndeni departure to the eighth, and by omitting the transport President Jackson, he conceded the First Battalion would not re-embark that day. Now at 0150 on 8 August, he pleasantly surprised Vandegrift by committing both reserve battalions to Tulagi, thereby putting all the Second Regiment’s riflemen ashore after sunrise. At 0217 the McCawley radioed Ghormley and Fletcher to note, ambiguously: “Owing to reinforcements Florida area [,] will not commence retirement as planned.” Typically, neither addressee received the message prior to a routine rebroadcast sometime late on 9 August. Questions remain as to Turner’s precise intentions, and what he meant by “will not commence retirement as planned.” Ostensibly that delay only applied to the four transports, one cargo ship, one heavy cruiser, and four destroyers of the Ndeni force that were to have gone out the evening of 7 August, and not to the rest of the transports that should soon be unloaded if they were not already. Although Turner abandoned the immediate execution of the Ndeni landing, his own pet project, he did not bother to advise his superiors. Possibly his unexpected generosity sought to provide enough riflemen to finish off Tulagi quickly to release the whole Second Marines for Ndeni.9
Turner left strong indications that later in the morning of 8 August he fully envisioned releasing the McCawley but aimed to remain himself. The Astoria found out, presumably when the McCawley’s boat brought the messages to be transmitted, that Turner and his staff would transfer on board later that morning. At breakfast the Astoria’s executive officer alerted correspondent Joe James Custer, who slept in the unoccupied flag sea cabin, that he must relinquish it because Turner would soon be on board. Later that day Custer heard that Turner postponed his arrival. As noted, Turner never explained what he intended to do with the Astoria and Vincennes after the landings, but it is likely he planned use them to support the marine raiders and Transdiv Twelve to “conduct raid operations as required.”10
The tardy unloading of all the transports and cargo ships disastrously affected Turner’s plans. The extended fighting in and around Tulagi tied up landing beaches and prevented the delivery of supplies. The transports there sat idle—hence Scott’s warning on the evening of the seventh. In contrast, the X-Ray Squadron at Lunga did not budge even during the air attack, and its troops certainly faced no opposition ashore. Yet by the evening of 7 August Turner already noticed “excessive delays occurring at Red Beach” in the unloading of supplies. He blamed the absence of marine pioneers to empty the boats. Even after they appeared, the “situation did not materially improve.” Heaps of crates were not being shifted inland fast enough. Finally at 0130 on 8 August, Turner suspended unloading because “the beaches were so congested.” Only after 0700 did the transfer of supplies resume at Red Beach. Word at 1037 of incoming planes caused Turner to get under way to meet the attack, which he did in sterling fashion. After the false alarm at 1355, things really fell apart. The ships required more than three hours to regain position and did not resume unloading until nearly sundown. Turner followed Vandegrift’s suggestion to relocate further unloading west to the beach situated east of the mouth of the Lunga River and north of the airfield, where the marines secured the shoreline. The new beach soon became clogged as well. The sailors and marines bickered over what should be done and who was at fault.11
By late afternoon on 8 August it must have galled Turner he was not going to get any of the invasion force out that day, despite all of his optimistic predictions at Pearl and at the 27 July conference. That applied especially to the X-Ray transports, all of which had disembarked their troops and nearly all the gear and supplies. Given this was the first Allied amphibious offensive in the Pacific, delays arising from inexperience could not fairly be held against anyone. But Turner did not candidly inform his superiors that something he directly supervised went even temporarily amiss. Perhaps he rationalized the laconic warning sent early on 8 August, that the “retirement would not commence as planned,” offered sufficient explanation. Yet if he had honestly reported the delays, Fletcher would have had powerful reason to risk staying. Turner did not imagine his one cautionary message did not reach his superiors in time, and his eagerness to shift blame led him to dissemble about many facts. Afterward he never admitted there was ever any reason to think most of the Amphibious Force might retire on the second day. Therefore he could depict Fletcher’s next act as outright “desertion.” Nor did Bates and Innis take note, despite being aware of Turner’s 071030 of August 1942. Instead they faulted Fletcher, when he made his recommendation, for having “failed to inform [Ghormley] fully of the delicate nature of the situation at Tulagi-Guadalcanal.” In truth, Turner failed to notify Fletcher of the “delicate nature of the situation.” Bates and Innis (and Morison) are effectively rebutted in the memoirs of Rear Adm. Geoffrey G. O. Gatacre, RAN, who was Crutchley’s staff officer for operations and intelligence in the Australia. In 1982 Gatacre wrote regarding Fletcher’s recommendation to withdraw the carriers, “The original plan called for a majority of the transports to leave the combat area by the night of 8th/9th August, and Admiral Fletcher may have been unaware that unloading was greatly behind schedule.” The evidence shows that Gatacre was entirely correct.12
Having charged without any basis in fact that Fletcher knew of Turner’s unloading woes but nevertheless abandoned him, critics postulated that his principal reason was unwarranted fear for the safety of his carriers. This insecurity arose from Fletcher having lost one carrier in each of his two recent battles and a disinclination to make it three for three, no matter how much that might compromise the overall mission. The critics ignore the possibility Fletcher might actually have learned something at Coral Sea and Midway, especially as he now found himself in the reverse role of his erstwhile opponents in supporting an amphibious thrust deep into enemy-controlled waters. Perhaps he really saw compelling reason to get the carriers clear of the invasion area and prepare for action. Just as he thought that with the marines ashore Turner’s immediate task neared its end, he understood the job of the carriers had only begun. In the short term they were the only shield against powerful naval and ground forces intent on destroying the marine lodgment. A large counter-landing would require strong carrier support to sweep away naval opposition before the actual landing force drew within range. Fletcher envisioned another grim carrier battle soon. He must be ready for enemy carriers at any time, despite rosy intelligence estimates from Pearl—the most recent on the afternoon of 8 August—placing the carriers in home waters. Intelligence could be wrong, and the resulting surprise quite deadly, as the Japanese at Coral Sea and Midway could attest, not to mention Kimmel at Pearl Harbor. Hindsight has obliterated the validity of Fletcher’s prudence.13
The tactical necessity of protecting a fixed point put heavy demands on the defensive capabilities of Fletcher’s carriers. They were accustomed to strike swiftly and draw clear of retaliation. Now they were exposed not only to the threat of opposing carriers, but also subs (messages reported several en route) and the more vigorous than anticipated land-based air. So far as Fletcher knew on 8 August, only eight B-17s struck the main Rabaul airfield the previous afternoon, while its bombers were busy attacking Turner. MacArthur announced no follow-up attack, except after dark on 8 August by a handful of RAAF Catalina flying boats. Moreover, encountering land-based Zero fighters so far from base proved a disheartening surprise. Only the Zero fighter’s incredible endurance permitted that mission and all the subsequent ones against Guadalcanal. Even worse, the awesome prowess of the land-based escort pilots only reinforced the negative comments voiced after Midway about the Grumman F4F-4 Wildcat fighter. Also distressing, despite coast-watcher warnings, the carriers failed to mount an adequate defense of the invasion force. The poor performance of radar, inexperience, bad luck, and bad decisions bungled the ever-difficult fighter direction. Turner’s ships were fortunate to escape as lightly as they did. As if all that was not enough, the unforeseen presence on 8 August of medium bombers wielding torpedoes so far from base boded ill for the carriers. As the Maas papers and other sources show, Fletcher considered all these factors as he mulled whether to propose the early withdrawal of the carriers.
Grumman Wildcat fighters, of course, constituted Fletcher’s principal defense against air attack. Just in two days, nine F4Fs were shot down, six ditched or crashed, and six others returned badly shot up and required lengthy repairs. Those twenty-one F4Fs comprised 20 percent of Fletcher’s fighter strength. The critics dismissed or minimized the significance of those losses, pointing out that the three carriers at Midway started with seventy-nine Wildcats, only one more than the seventy-eight F4Fs Fletcher currently had operational. What they ignored is that the whole Midway battle cost the carriers only nineteen F4Fs (six shot down and thirteen ditched or crashed), plus three that went down on the Yorktown. At Coral Sea a dozen F4Fs (eight shot down, four ditched) were destroyed, with nine more on the sunken Lexington. Thus by 1530 on 8 August, Fletcher had lost more F4Fs off Guadalcanal than in each of his previous battles. Yet he could only reckon the air fighting on D-Day and D+1 as preliminary skirmishes in a very long campaign, with no source short of Pearl for fighter replacements. The big brawls were yet to come. On 24 September Fletcher explained the need for carrier replacement air groups at convenient places like Suva, Nouméa, and Tongatabu. “In two days of supporting the landing against air attack, the carriers lost approximately 20% of their fighter strength. Yet they also had to be prepared to repel an expected attack by enemy carriers at some later date with no prospects of replacing their current losses.” Pederson stressed in his 1944 lecture, “The carriers could not afford to engage in a war of attrition over a period of time with land based planes” and still accomplish their mission of stopping a big counteroffensive against Guadalcanal. “If the carriers had remained to suffer further losses they would have been in no condition to do so.”14
The heavy fighter losses again elicited grave doubt over the combat effectiveness of the Grumman F4F-4 Wildcat. The 50 percent casualties suffered by the first F4Fs to engage land-based fighters carried significance beyond its stark number. It was a truism that sleek land-based aircraft outperformed heavier carrier planes, but such a difference between two nominal carrier fighters, the Zero and Wildcat, was deeply disturbing. In 1963 Fletcher remarked: “Nobody mentions the matter, for fear of bringing down the wrath of the aviators upon them, the Japanese Zero’s all wore Seven League Boots [and] our aviators gave them a lot of god damned respect.” Kinkaid also was “much interested in what returning pilots had to say about the situation in the landing area and, even more so, in their attitude regarding the Japanese air performance.” He recalled the “loss of fighter planes during [7 August] gave us pause,” and that “at first our pilots were greatly discouraged.” The truth of that is reflected in a routine message regarding machine guns and ammunition that Lt. Louis H. Bauer, commanding officer of VF-6 in the Enterprise, sent on 13 August to Pearl. Although pleased at the ruggedness and overall protection of the F4F-4, he dramatically concluded: “Pilots are anxiously awaiting faster and better fighters. Repeat pilots are anxiously awaiting faster and better fighters.” The Battle of the Eastern Solomons on 24 August and the brilliant success of the marine fighter pilots at Guadalcanal at least partially restored the fleet’s confidence in the Wildcat. Yet on 8 August, the evident failings of the F4F-4s when finally matched against Zeros flown from land bases weighed heavily on Fletcher’s mind.15
The badly flawed placement of fighters also severely compromised the air defense on 7–8 August. Analysis showed that inferior altitude, dispersal of assets, and remarkably effective fighter opposition devastated the screen combat air patrol on 7 August. The commanding officer of VF-5 likened the piecemeal commitment of his F4Fs to scraps tossed into a “Japanese meatgrinder.” Inadequate numbers of fighters over the Amphibious Force on 8 August, despite the early warning, stemmed primarily from the Saratoga’s inability to launch F4Fs on demand. Maas unfairly blamed Noyes, but Fletcher and Ramsey were at fault for furnishing the dawn Tanambogo strike that Turner requested. According to Ramsey, “In a large complement of mixed types such as was embarked in the Saratoga, flexibility in the use of fighters is seriously compromised.” Bruning, the FDO in the Chicago, bore the brunt of the criticism for fighter direction, largely unwarranted given that high terrain blocked his radar. He never detected the second strike on 7 August or the one on the eighth. The carriers fared no better, because they operated in the shadow of Guadalcanal’s high mountains. Kinkaid had recommended they move farther out to sea for unfettered radar coverage, but that would have curtailed fighter support for Tulagi and Lunga.16
If handling the combat air patrol was not problem enough, Fletcher learned on the afternoon of 8 August that the enemy cleverly employed a mixed assault force to divide the defense, very much as on 7 May against Crace. The returning combat air patrol pilots and Turner’s 1430 summary described a coordinated assault by eight horizontal bombers to draw off defending fighters (“Fooled us,” according to Maas), which opened the way for a low-level sneak approach by forty twin-engine torpedo planes. That assessment was wrong. There were no horizontal bombers, and the torpedo attack by twenty-three land-attack planes proved both inept and very costly, but Fletcher did not know that yet. The sudden appearance of numerous medium bombers toting torpedoes, land-based air’s most effective antiship weapon, provided another strong reason for concern, especially given the combat air patrol’s questionable performance. Fletcher had not expected torpedo planes, certainly not ones augmented by fierce Zero escorts, could even reach him off distant Guadalcanal. He queried Noyes whether the noon attackers “were actually carrying torpedoes,” and was assured they did. Maas wrote that night, “The use by Japs of long-range torpedo planes makes our present position untenable, dangerous, and foolhardy.” That was particularly true if the carriers could legitimately get clear without harming the overall mission.17
The basic role of the carriers in the Watchtower landings was, of course, to provide air support, in particular fighter cover. The week before the invasion Ghormley and McCain had, as noted, suggested Fletcher leave fighters behind at Lunga when the carriers withdrew. Fletcher considered detaching twenty-four F4Fs temporarily to Lunga, but that was not possible. First and foremost he had no place to send them. As far as he knew, the airstrip was still in enemy hands. Maas noted on the evening of 8 August, “Vandegrift failed to make Lunga Airfield first objective and prepare it for fighters. We could have left fighters and got out of area.” Kinkaid’s memoir confirmed that harsh assessment. Not until the evening of 10 August did Fletcher learn Vandegrift had overrun the Lunga airstrip on 8 August. Indeed, the Lunga field was never a feasible option, because it remained unfinished. But the Japanese thoughtfully left construction equipment so the marines could smooth it out by 11 August. However, even if conditions had allowed the carrier F4Fs to come into Lunga on 7 or 8 August, they had no aviation supplies and trained ground personnel on hand to service the planes. Regarding fighter support, it was Fletcher’s carriers or nothing.18
Fletcher strongly advised in his 24 September analysis of carrier operations that bases for land-based aircraft should be established, “Within the shortest possible distance of the objective prior to any invasion effort.” Land planes should stage into the objective as soon as the airfield was ready to take over air defense and release the carriers to withdraw. On 9 October Admiral Halsey went further. “However the danger exists that the more important and more fundamental fact may be lost sight of that the land plane bases and the operating units thereon should be available in supporting positions before [Halsey’s emphasis] the operation is undertaken at all. It is only by this provision in advance that the risking of carriers in restricted covering positions can be avoided.” Thus the question of protecting the carriers off Guadalcanal was much more complex than one might infer from Fletcher’s critics.19
Fletcher’s controversial message to Ghormley concluded with his request to send oilers “immediately to rendezvous decided by you as fuel running low.” Oil certainly was an important consideration in his mind. In 1947 he told Hanson Baldwin he had actually “recommended the withdrawal because of shortage of fuel.” Sixteen years later, he contradicted himself, explaining to Dyer that fuel was not the primary reason, that his “despatch didn’t say anything about needing to withdraw to fuel.” Instead if Ghormley approved, he “wanted to fuel as soon as tankers could reach me, as my staff had told me fuel was running low on some of the short-legged destroyers, and Task Force Sixty One had never been belly full since its formation.” Fletcher’s critics debated whether security or low fuel played the greater role in his perceived desire to bug out. They opted for the former but were prepared to denigrate him for both reasons. “It is idle to pretend that there was any urgent fuel shortage in this force,” Morison declared. Thus it is vital to determine precisely the fuel state of Fletcher’s ships on 8 August, its effect on the tactical situation, and whether he was justified in citing fuel as a factor influencing his operations, if not a critical reason to warrant withdrawal.20
Maas wrote on 8 August that the “task forces need fuel.” Fletcher recalled to Baldwin, “We were surprised to find the destroyers running short almost at once and some of the carriers and cruisers used much more fuel than we expected.” He blamed the excess consumption on the high-speed bursts necessary to gain enough wind velocity in the light breezes that prevailed for much of 7 August and the next day. The carrier task units certainly observed a sharp drop in fuel in their routine noon reports. Kinkaid’s war diary noted that as of noon on 8 August, “Fuel situation this force is becoming critical.” He “estimated the destroyers have fuel for about three days at 15 knots and the heavy ships have little more.” Noyes advised that his destroyers had only enough fuel for thirty-one hours at twenty-five knots. The destroyers of Fletcher’s own task unit were only a little better off, with oil for thirty-five hours at twenty-five knots.21
Rear Adm. Worrall Reed Carter’s official history of fleet logistics in the Pacific noted that Fletcher fueled his ships on 3–4 August, but a serious lack of foresight caused a shortage of oil at a critical juncture on the eighth. “Why Fletcher could not have refueled on 4 and 5 August and held on a day longer is not clear.” The reason is perfectly clear. Fletcher would have been delighted to top off his ships on 5–6 August, but because the Kaskaskia and Esso Little Rock had not come forward, no additional oil was available. Its absence proved crucial on 8–9 August.22
At least Admiral Carter conceded Fletcher might have been low on fuel, although wrongly assessed it to be his own fault. Other critics, however, assailed the notion TG-61.1 ever experienced a fuel shortage on 8 August, or at least one that could curb operations. Pye’s Savo analysis noted, “The Wasp and Saratoga in their reports indicated no concern about shortage of fuel.” Ship action reports, though, rarely mentioned fuel; that was the responsibility of the task force commander. Maj. John L. Zimmerman’s official Marine Corps monograph on Guadalcanal first cited the deck logs of Fletcher’s ships to prove fuel should not have been behind his haste to retire the carriers. Computing each ship’s “average daily consumption” from 1 to 8 August, Zimmerman gauged how long it could steam just with the fuel on hand on the eighth. All ships could have operated for “at least four days.” The carriers had enough oil for seventeen days, the battleship North Carolina eight days, the cruisers eleven days, and the destroyers “about 7 days,” except for the Benham (three days) and Grayson (two days). His methodology is grossly flawed. From 1 to 6 August the carrier task group did no high-speed steaming. Only on D-Day itself did the carriers consistently step up beyond twenty knots. Further, the figures listed for 8 August were the noon fuel reports only, before the additional large expenditure that afternoon.23
More formidable was the tag team of Naval War College analysts and Morison, who collaborated in their research. Griffith claimed their work “categorically refutes Fletcher’s statement that his ships were short of fuel.” Bates and Innis contended the Cimarron “topped off” a dozen of Fletcher’s destroyers on 3–4 August, and thereafter they “were refueled as necessary from the heavy ships of the various task groups.” A “check of the logs” revealed that on 8 August the destroyers of Fletcher’s own TU-61.1.1 (Saratoga) “averaged about seventy-five percent capacity,” whereas those of Kinkaid’s TU-61.1.2 (Enterprise) had only 42 percent and Noyes’s TU-61.1.3 (Wasp) 44 percent. Cruiser fuel in all three task units was 50 percent or more. Of the carriers, only the Enterprise “was running low and had fuel for three more days of operation.” Accordingly, “Although the fuel in the command was diminishing daily it was not at this time so critically low as to force retirement from the area.” Moreover, Fletcher’s intention to stay one more day, if Ghormley refused to give permission, demonstrated, “He was fully prepared to operate with the fuel on hand.” Morison concurred. “It can now be ascertained from ships’ logs that at noon 8 August the destroyers still had enough fuel for several days’ operations, and they could have been refueled from the cruisers and carriers, as well as from fleet oilers.” As in his discussion of the Wake relief, Morison provided a footnote that listed the fuel on 8 August for most of the ships in Kinkaid’s and Noyes’s task units (but just the Saratoga herself in Fletcher’s own task unit). Instead, regarding the Saratoga group, he wrote, “[The destroyer] Dale had just been topped off and was 97 percent full, the other five destroyers averaged three quarters full.” Morison noted, “The destroyers’ daily expenditures varied from 12,000 to 24,000 [gallons],” and fuel capacity ranged from 127,000 to 184,000 gallons. “The cruisers (capacity 618,000 to 839,000 gallons) were half full or better.” Thus according to Morison, the carrier group experienced no “urgent fuel shortage.”24
The only other historian to examine the oil situation of TG-61.1 in detail, Dyer listed fuel figures for 7–9 August for all of the ships he believed were present. He did not list capacities or assess actual fuel percentages. An author of a book on naval logistics, Dyer stressed the baleful influence of excess worry by naval commanders over what he called AFFAG (Ammunition, Fuel oil, Food, and Aviation Gasoline) to the detriment of completing the mission. On 8 August both Fletcher and Kinkaid suffered from the “virulent” form of AFFAG disease, and only the destroyer Grayson in Kinkaid’s task unit merited a “critical” grade for fuel shortage.25
In truth the Naval War College analysts, Morison and to a lesser extent Dyer, committed surprising errors in describing the fuel state of Fletcher’s ships on 8 August. Bates and Innis emphasized the destroyers were “topped off” on 3–4 August. Actually none were filled to full capacity, because there was not enough oil to go around. They also averred the destroyers of Fletcher’s TU-61.1.1 averaged about 75 percent of capacity on 8 August. Morison not only accepted their figures for five of the destroyers he thought were present (ignoring that the Hull and Dewey were detailed to TF-62), but he also commented the Dale had just fueled to 97 percent. Because no fleet oilers were with the force, that fuel would have come from a heavy ship. So, if the Dale could refuel, why not the other destroyers? In fact the war diaries and deck logs show that on 8 August only destroyers Phelps, Farragut, Worden, and Macdonough were actually with Fletcher’s own TU-61.1.1. Their logged fuel figures average just 46.8 percent of capacity, not “about seventy-five percent.” The absent Dale indeed fueled to 97 percent, but it was on 6–7 August from the merchantman Oliver Wendell Holmes while both ships lay at anchor at Espíritu Santo. Subsequently the Dale escorted two ships to Efate, arriving at noon on the eighth. Why Bates and Innis gave the average fuel state for the TU-61.1.1 destroyers as 75 percent remains a mystery. Yet their figures for the destroyers of Kinkaid’s TU-61.1.2 (42 percent) and Noyes’s TU-61.1.3 (44 percent) compare well to the 44.2 percent and 45 percent respectively derived capacities given in the authoritative 1945 U.S. Fleet Tactical Publication FTP-218, “War Service Fuel Consumption of U.S. Naval Surface Vessels.” In addition to placing the Dale in TU-61.1.1, Bates and Innis, Morison, and Dyer all inflated TU-61.1.3 by adding the Laffey (on her way to Pearl for repairs) and the Aaron Ward busy chaperoning the Cimarron. This error originated in the sloppy assumption that because these destroyers appeared in the TF-61 Op-Ord 1–42 (28 July 1942), they all had to be present on 8 August. As with Morison’s discussion of the Wake relief, the fuel consumption figure for destroyers (twelve to twenty-four thousand gallons per day) he gave with the 8 August fuel listing is misleading. Their vast expenditure in battle might easily consume half again as much oil as his larger amount.26
With an overall average of 45.3 percent of fuel capacity on 8 August, it can no longer be doubted all thirteen of Fletcher’s destroyers were low on fuel.27 The six cruisers averaged 52 percent fuel. It is clear when reckoning available fuel that Fletcher, Noyes, and Kinkaid used the standard practice of going by the lowest ship in each category. The basis for their estimates of steaming endurance was less reliable than those derived from empirical data unavailable in 1942. Even so, Fletcher and Noyes came close to the mark regarding destroyer fuel. Fletcher’s low destroyer, the Macdonough, had fuel for about thirty-six hours (1.5 days) at twenty-five knots, whereas he calculated thirty-five hours. Likewise Noyes reported his destroyers could steam thirty-one hours (1.3 days) at twenty-five knots, while the Sterett had enough oil for about thirty-six hours at twenty-five knots. By citing fifty hours (2.1 days) at twenty-five knots Fletcher erred with the cruisers. The actual figure was sixty-five hours (2.7 days) for the Minneapolis, ostensibly the low ship. Perhaps, though, he based his estimate on figures that Kinkaid provided for TU-61.1.2. Kinkaid’s fuel reports represented his ships having less fuel than later calculations could support. Dyer called his 7 August dispatch regarding destroyer fuel (two days at fifteen knots) “misleading,” if Fletcher quoted it properly in his 9 September 1942 report. Moreover, Kinkaid’s war diary for 8 August had complained, “Fuel situation [in] this force is becoming critical,” and that he “estimated the destroyers have fuel for about three days at 15 knots and the heavy ships have little more.” With the oil on hand the Grayson, by far his lowest destroyer, could only steam 3.6 days at fifteen knots, but the TF-16 war diary badly erred in stating the heavy ships had “little more.” So Kinkaid’s fuel reports to Fletcher may have been much too pessimistic, but Fletcher himself never had the luxury of personally counting every barrel in every ship. He had to take the word of his subordinates.28
Thus with the possible exception of the Grayson, the ships of TG-61.1 would not have run dry if they faced a third day of heavy steaming and combat off Guadalcanal. However, their fuel situation was certainly much more worrisome than generally perceived, and certain to have deteriorated rapidly. By noon of 9 August, after running mostly at fifteen knots after sundown on the eighth, average destroyer fuel decreased to 35.2 percent. That would have been considerably less had Fletcher stayed off Guadalcanal again conducting extensive air operations or dodging air attacks. Moreover, just parking the carriers within air support range for another day or two was not all there was to it, something all the critics ignored. Fletcher had to be ready to execute the second half of his mission, which entailed the prospect of fighting another carrier battle to prevent a landing force from retaking Tulagi and Lunga. One must therefore examine his options with regard to fuel to help assess its role in his desire to withdraw the carriers.
The Naval War College analysts and Morison assumed that Fletcher could simply refuel his destroyers from his heavy ships at will, such as Morison wrongly wrote of the Dale. Ordinarily such fueling was routine but required good weather and freedom from imminent air or sub attack. In September 1942, however, Fletcher informed Ghormley that fueling the destroyers from the cruisers “was not practicable.” They, too, were low on oil. He did tell Baldwin in 1947 he could “possibly” have “stayed one more day” if he had “fueled destroyers at night,” but that would have depended on whether he accepted the real risks of remaining on station. Such reduced fuel in the task group would have severely limited his ability to maneuver in a highly uncertain tactical situation. In January 1943 Rear Adm. George Murray independently confirmed Fletcher’s assessment. According to Murray, “The carrier task force problem, so far as refueling is concerned, hinges on destroyer consumption,” an “interesting sidelight [that] should be kept constantly in mind.” While “in an advanced area a task force cannot afford to approximate the low limit of fuel for the simple reason that it then is virtually immobilized for offensive operations involving 48 hours of high speed steaming.” That “low limit” pertained when the fuel in the heavy ships fell to “40/50% of their capacity,” without the offsetting advantage of having the destroyers at 90 to 100 percent of fuel. “With the heavy ships fueled to 40/50%, the destroyers to 90%, it is possible to carry out any offensive operations such as a raid involving a 2,000 mile operation.” By noon on 8 August, Fletcher’s TF-61 had already dropped below Murray’s minimum requirements with regard to the destroyers and the Enterprise (and was close with the gluttonous Saratoga), while cruiser fuel neared the lower limit. Also by the time Murray’s TF-17 reached Sopac at the end of August, the worst of the fuel supply problems had eased, and he could afford to be more sanguine with regard to steaming limitations. Fletcher faced a much more uncertain logistical situation in early August.29
What about the fleet oilers the critics thought just waited in the wings ready to be called forward? Turner told Baldwin in 1946 that Fletcher “did not want to risk bringing the tankers within range of Jap planes off Guadalcanal, so he planned to retire to the South in vicinity of the New Hebrides in order to refuel.” Thus Fletcher supposedly fretted his oilers could suffer a fate similar to the Neosho in May and deliberately kept them away. Turner managed to ignore his own real fuel worries during the approach and landing phase of Watchtower. In truth Fletcher never had the option on 8 August of summoning the oilers to Guadalcanal, because none were within a thousand miles. After fueling TF-61 on 3–4 August near Efate, the empty Cimarron and the destroyer Aaron Ward ventured south to Nouméa in the vain hope of finding the Esso Little Rock, which in the meantime turned up at Suva. The Cimarron group left Nouméa at dawn on the seventh bound for Suva to refill and come back out to Espíritu Santo. That whole trip of about fifteen hundred miles would require six days, counting time in harbor. In turn the Platte and Kaskaskia, both nearly full, departed Suva on the evening of 5 August, along with the Rainier and the destroyer Perkins, and were due at Nouméa on the afternoon of 8 August. Ghormley had thought to direct the destroyer Cummings (escorting a transport at Efate) to intercept the convoy on 6 August and release the Kaskaskia and Perkins to split off to Efate. The Cummings, however, could not join in time. The whole convoy continued to Nouméa, one thousand miles by sea from Fletcher’s current location close under southeast Guadalcanal. Thus even if one or both oilers cleared Nouméa by midnight on 8–9 August, Fletcher could not count on fuel reaching TG-61.1 before the evening of 11 August, unless he went south to get it. If he did stay off Guadalcanal, his fuel would soon be critically low, greatly restricting his mobility. The massive refueling necessary to fill the whole force would require at least two full days, immobilizing the carriers until 14 August.30
Should Fletcher have split the force and pulled out only part to fuel? Nimitz’s 23 August endorsement (the one in which he called the carrier withdrawal “most unfortunate”) speculated whether it was “practicable to fuel one carrier task force at a time, leaving two available for support of the operations.” The 1943 Cominch Secret Information Bulletin No. 2 similarly posed that possibility. In the meantime Fletcher addressed that issue. “Due to fact that all TFs were equally short of fuel,” his 9 September 1942 report to Ghormley explained, “it would not have been possible to retire one TF at a time.” Despite the assertions to the contrary, Fletcher was indeed correct. It is also doubtful, no matter how he rationalized later, that given his strong concern over the security of his carriers, he would have divided his ships and kept a weaker force off Guadalcanal. Nimitz, though, disregarded Fletcher’s reasoning. His The Great Sea War again proposed fueling the task groups one at a time.31
Fletcher also had to worry he might not even get all the fuel he needed no matter what he did. Not counting the old, slow, and small Kanawha busy fueling ships passing through Fiji, the Platte, Kaskaskia, and Cimarron (once she refilled) contained virtually all of the readily available oil in the entire Sopac area. If one should be sunk or incapacitated, her cargo could not be replaced until the arrival of the next batch of chartered tankers not expected at Nouméa before 12 August. Subs posed a real peril. On 26 July while escorting the Platte back to TF-61, the Worden picked up survivors from an army transport torpedoed seventy-five miles southeast of Nouméa. On 6 August Comsopac relayed a report from the Cimarron warning of three or four subs encountered twice within eighty miles of Nouméa. On both occasions the Aaron Ward dropped depth charges and claimed sinking two subs. The Cummings thought she bagged a sub near Efate. The presence of subs and too few escorting destroyers made any passage north by the oilers a risky prospect.32
In the absence of a perceived emergency that might have compelled Fletcher to keep the carriers on station despite the danger, it can be argued that inadequate fuel was another powerful reason, or at least a strong reinforcing reason, to withdraw TF-61.
At 1525 on 8 August Fletcher announced to Noyes, “In view of possibility of torpedo plane attack and reduction in our fighter strength I intend to recommend immediate withdrawal of carriers.” In the event Ghormley disapproved (the message just said “in case we continue present operation”), Fletcher thought they should return to the same waters as on the eighth, the strait between Guadalcanal and San Cristóbal. Noyes responded at 1615 “affirmative to both questions” and suggested if the carriers did stay, they should adopt the same operating plan as on 8 August. It is inconceivable Noyes did not discuss the situation with Capt. Forrest Sherman, his closest advisor, who evidently concurred with his answers to Fletcher. In 1949 Sherman wrote Morison: “On the afternoon of the eighth I felt that the Japanese response had been strong, but had no reason to be apprehensive over a temporary withdrawal. It was not until after the night action that the situation became critical.” That certainly implies that Sherman thought, as did Fletcher, that Crutchley’s Screening Group could take care of itself in the temporary absence of the carriers.33
Brig. Gen. Melvin J. Maas, U.S. Marine Corps Reserve, circa 1950.
Courtesy of Minnesota Historical Society
Written the evening of 8 August 1942, Maas’s invaluable explication of the reasons behind the recommendation to retire the carriers coincides closely with Fletcher’s recollections years later to Baldwin and Dyer. As quoted at the beginning of this chapter, Maas strongly supported Fletcher’s action by laying out all of the relevant factors of necessity, security, and supply: “To stay meant fighter protection for landing force and for finishing unloading cargo ships. Daily bombing attacks from Rabaul. To leave meant leaving Marines without air protection. To stay means endangering the carriers. We have too few to risk them at this time. The task forces need fuel. Attrition is rapidly reducing fighter strength of carriers. In addition to increasing danger from shore-based air attack, every hour increases danger from Jap subs, now gathering and heading for us.” Fletcher recalled to Dyer that he was fully aware he led most of the surviving U.S. carriers, and that there would be no replacements until mid 1943. Not only could Japan mass superior carrier strength in the Solomons, but it was also the most likely course of action. Fletcher’s greatest concern was risking the carriers in situations where they could not inflict greater or even comparable damage to the enemy. They must be preserved in the face of land-based air and sub opposition to deal with the enemy carriers certain to appear very soon. Nimitz had reimposed the restrictions that applied at Midway, and Fletcher “felt he had no choice but to obey his instructions.”34
Maas understood the “Marines cannot be dislodged by bombers. They will take a drubbing, and their losses will be greater without carriers.” However, the best chance Japan had of destroying the landing force was “by an expedition similar to ours.” He outlined Fletcher’s plan of action. “To be able to intercept [and] defeat such an effort, our carrier task forces must be fueled and away so as not to be trapped here.” Likewise Fletcher recalled to Baldwin in 1947 if he had stayed and fueled destroyers at night, “We would have to proceed quite slowly to our [fueling] rendezvous and would be caught in a very bad position if enemy carriers appeared on the ninth. I think we all expected they would show up shortly after the landing.” Maas explained on 8 August: “By withdrawing to, say, Noumea, or Tongatabu, we can be in a position to intercept and pull a second Midway on their carriers. If, however, we stay on here and then, getting very low on fuel, withdraw to meet our tankers, and if they should be torpedoed, our whole fleet would be caught helpless and would be cold meat for the Japs, with the resultant loss of our fleet, 2/3 of our carriers, and we would lose Tulagi as well, with all the Marines there and perhaps all the transports.” Maas then placed the withdrawal of the carriers in a wider context: “It is true, Marines will take a pounding until their own air gets established (about 10 days or so), but they can dig in, hole up, and wait. Extra losses are a localized operation. This is balanced against a potential National tragedy. Loss of our fleet or one or more of these carriers is a real, worldwide tragedy.” To put it bluntly, was the risk of having one or more carriers possibly crippled or sunk worth staying to protect five cargo ships that Fletcher thought would remain off Lunga? Maas thought no. A proud marine, he could hardly be accused of being unsympathetic to the Corps. Pederson in his 1944 lecture completely agreed with his colleague. “The correctness of the decision to withdraw the carriers is one that is still in dispute. When viewed from the viewpoint of the marines ashore it was questionable. But when looked at the broader strategic picture it has proved correct.”35