It was late for a presidential address, almost midnight on Aug. 4, 1964, but the situation was dire. Since midmorning, the Pentagon and the White House had been following reports of a battle in which North Vietnamese gunboats attacked two US destroyers, USS Maddox and USS Turner Joy, in the Tonkin Gulf off the coast of North Vietnam.
President Lyndon B. Johnson had let a previous attack on Maddox two days before in the Tonkin Gulf go by with a stern warning. He would not refrain again. In his midnight address, Johnson called the second attack “open aggression on the high seas” and said that air strikes were on the way to retaliate. Sixty-four sorties in several waves from the carriers Ticonderoga and Constellation struck gunboat bases and supporting facilities in North Vietnam, where it was already the middle of the day on Aug. 5.
On Aug. 7, Congress passed the Tonkin Gulf resolution, authorizing Johnson to “take all necessary measures” to repel attacks and prevent further aggression. Within days, the US Air Force moved into Southeast Asia in force. The conflict escalated, and by spring the United States was engaged in a full-scale theater war. The Tonkin Gulf resolution was the only congressional authority given, then or later, for the commitment of force in Vietnam.
USS Maddox under way during the early 1960s. North Vietnamese PT boats attacked Maddox and USS Turner Joy on Aug. 4, 1964. The attack, and an alleged one two days later, shaped the authority under which the war would be fought.(US Navy photo)
The public’s understanding of the Aug. 4 encounter was summed up vividly by Time magazine. “Through the darkness, from the west and south, the intruders boldly sped,” the Time article said. “There were at least six of them, Russian-designed ‘Swatow’ gunboats armed with 37 mm and 28 mm guns, and P-4s [PT boats]. At 9:52 they opened fire on the destroyers with automatic weapons, this time from as close as 2,000 yds. The night glowed eerily with the nightmarish glare of air-dropped flares and boats’ searchlights. For three-and-a-half hours, the small boats attacked in pass after pass. Ten enemy torpedoes sizzled through the water.”
Except for the air-dropped flares, the Time story was wrong in every detail. There were no gunboats, no PT boats, no torpedoes, no three-and-a-half-hour battle. The Aug. 4 attack never happened. It was not until years later that accurate accounts emerged of events that night in the Tonkin Gulf.
The First Incident
In the election campaign of 1964, Johnson faced the hawkish Barry M. Goldwater, who was nominated at the Republican National Convention in July. Challenged by Goldwater to take a stronger stand on Vietnam, Johnson sought to look moderate but firm. He was waiting for an opportune moment to ask Congress to adopt a resolution, drafted by his aides, expressing support for his handling of the war.
US armed forces had been in Vietnam for three years, mostly in an advisory and support capacity, although air commandos were secretly flying combat missions. Deeper in the shadows was the “Studies and Observation Group,” its bland-sounding name a cover for clandestine activities under Operations Plan 34-A, approved by Johnson in January 1964. One SOG enterprise sent fast patrol boats, based at Da Nang and manned by South Vietnamese crews, to raid up and down the North Vietnamese coastline. Like all 34-A operations, the raids had to be authorized in advance by the Secretary of Defense, the Secretary of State, or the President.
Returning from a raid the morning of July 31, the fast patrol boats passed within sight of Maddox, which was just entering the Tonkin Gulf for a signals intelligence patrol. The mission of such patrols, code-named DeSoto, was “to update our overall intelligence picture in case we had to operate against North Vietnam,” said Adm. U. S. G. Sharp, head of US Pacific Command. Lashed to the decks of the ships were signals intercept vans, operated by naval security personnel. The DeSoto patrols were under Navy control and separate from the 34-A raids, but the North Vietnamese did not know that.
Cmdr. Herbert L. Ogier was captain of Maddox but also aboard was the mission commander, Capt. John J. Herrick, commander of Seventh Fleet Destroyer Division 192. On the afternoon of Aug. 2, Maddox intercepted an order for North Vietnamese P-4 torpedo boats—called PT boats by the Americans—to attack the destroyer. Maddox turned toward the open sea and was about 15 miles from shore when the PT boats caught up.
The pursuers rapidly became the pursued. Maddox opened fire with its five-inch guns and summoned air support from Ticonderoga in the South China Sea. The PT boats fled, ineffectively scattering torpedoes as they went. The action lasted 37 minutes. The North Vietnamese believed they had shot down one airplane and hit another; in fact, one F-8E Crusader was damaged slightly by the deck guns but landed safely. Maddox sustained a single bullet hole and no casualties. The Americans believed they had sunk all three PT boats; actually, they had damaged one of them.
Johnson decided against a military response. “We concluded that an overeager North Vietnamese boat commander might have been at fault or that a shore station had miscalculated,” he said. He issued a stern warning that the North Vietnamese should “be under no misapprehension as to the grave consequences” that would ensue if they attacked again.
Maddox resumed its patrol, joined by Turner Joy. They were instructed to stay at least 11 miles offshore, one mile less than the territorial limit claimed by North Vietnam, which the US did not recognize. Johnson gave the two destroyers specific orders to destroy any vessels that attacked them.
The fast patrol boats from Da Nang conducted another raid the night of Aug. 3-4. That day, Maddox entered the Tonkin Gulf with Turner Joy 1,000 yards astern. They divided the radar surveillance duties. The Maddox radar was set for long-range detection and Turner Joy’s radar was tuned for shorter range, including the tracking of surface targets.
Skunks and Bogies
Secretary of Defense Robert McNamara points out communist supply routes during a 1965 press conference. In his 1995 memoirs, McNamara insisted that evidence of the first attack was “indisputable” and that the second attack was “probable.”
About 6:15 p.m., the Marine signals intelligence unit at Phu Bai in South Vietnam notified Maddox of a probable North Vietnamese operation that night against the destroyers. Herrick promptly forwarded the warning up the line. It got instant attention in Washington where it was still morning, 11 hours earlier than in the Gulf of Tonkin. Nerves were on edge in anticipation of another attack.
Not until later would it be understood that the warning was Phu Bai’s interpretation of an intercepted message segment, and that the actual communication dealt mainly with the towing and refueling of boats. One line of the intercept directed unspecified “military operations” with no reference to the US ships.
At 7:46 p.m., the Maddox radar picked up the first three “skunks”—potentially hostile surface contacts—about 40 miles to the east, opposite the direction from which an enemy approach might have been expected. Several bogies, or aerial blips, also popped up. A steady stream of reports flowed from the Tonkin Gulf through channels back to Washington.
Over the next two hours, Maddox tracked at least eight skunks and Turner Joy reported four more at shorter range. At 9:37 p.m., the Maddox sonar detected a noise spike, which was interpreted as an incoming torpedo. The two ships maneuvered evasively as the Maddox sonar reported an astounding total of 26 torpedoes. Turner Joy’s sonar did not detect any torpedoes. The destroyers opened fire at 9:39 p.m., eventually expending more than 300 rounds, plus depth charges and star shells.
The night was exceptionally dark and visibility was hampered by intermittent drizzle. However, in later interrogation, deck and bridge crews on both destroyers said they saw torpedo wakes, a searchlight, smoke, and the silhouette of attacking boats.
The situation looked completely different to the carrier pilots flying overhead. They could see no ships other than Maddox and Turner Joy and the only wakes they saw were from the destroyers themselves. The aircraft launched rockets and strafed the general area of reported radar contacts.
Commander James B. Stockdale, leading the fighters, said the Maddox air controller “kept giving rapid-fire, blow-by-blow descriptions of ongoing sea battles that for the life of me I couldn’t find on my horizons.” In his memoirs, published in 1984 after his years as a POW in North Vietnam and receiving a Medal of Honor, he recalled that night in the Tonkin Gulf.
“I had the best seat in the house from which to detect boats—if there were any,” Stockdale said. “I didn’t have to look through surface haze and spray like the destroyers did, and yet I could see the destroyers’ every move vividly.” There were “no boats, no boat wakes, no ricochets off boats, no boat gunfire, no torpedo wakes—nothing but black seas and American firepower.”
The destroyers ceased firing at 11:44 p.m., with no casualties or damage to either ship. It was just after noon, 12:44 p.m. EDT in Washington.
The Smoking Gun
Throughout the morning, Herrick had reported the destroyers under attack and returning fire. He soon became less certain. North Vietnam had only 12 torpedo boats, each carrying two torpedoes, so the entire fleet could not have launched the 26 torpedoes identified by the Maddox sonar.
Shortly after midnight, at 1:27 p.m. Washington time, Herrick sent a message saying, “Review of action makes many reported contacts and torpedoes fired appear doubtful. Freak weather effects on radar and overeager sonarmen may have accounted for many reports. No actual visual sightings by Maddox. Suggest complete evaluation before any further action taken.”
By then, the President had already reached his conclusions, and according to Assistant Secretary of State William P. Bundy, “It became very clear that he was in no mood for discussion.”
In a telephone call to the Pentagon at 1:37 p.m., Sharp said many of the reports were probably the result of “freak effects on radar and overeager sonarmen,” who “get keyed up with a thing like this and everything they see on the sonar is a torpedo.” Even so, Sharp said there was “no doubt” that an attack occurred.
Herrick was bombarded with inquiries from Washington and other levels, jumping channels and the chain of command. Unwilling to set aside the sightings and inputs from his crews, Herrick certified at 2:48 p.m. Washington time that the attacks were “bona fide.”
Any remaining uncertainty was swept aside by signals intelligence intercepts that Secretary of Defense Robert S. McNamara described as “unimpeachable.” The smoking gun was a North Vietnamese “after-action report” furnished to McNamara that afternoon by the National Security Agency. It read, “Shot down two enemy planes in the battle area, and one other plan[e] was damaged. We sacrificed two ships and all the rest are okay. The combat spirit is very high and we are starting out on the hunt and [are waiting to] receive assignment. … The enemy ship could have been damaged.”
This was not the usual format for an after-action report, and inexplicably, the “two ships” in the text NSA provided had been “two comrades” in the original translation. Much later, Louis Tordella, deputy director of NSA, told the Senate Foreign Relations Committee that the message almost certainly referred to the Aug. 2 skirmish, not to Aug. 4. The “two comrades” were most likely casualties on one of the boats from the earlier incident. The Vietnamese language intercept on which the translations were based was missing from the NSA files.
During the Aug. 2 attack, there had been numerous intercepts of boat-to-boat communications and instructions to the attacking boats. No such communications were heard on Aug. 4.
Johnson was champing at the bit. For political reasons, he dared not look weak or indecisive. He wanted the retaliatory air strikes to be delivered at 7 p.m. to coincide with a televised address to the nation, but to his dismay, the reprisals slipped forward. Ticonderoga was short of available aircraft. A second carrier, Constellation, was pressing toward Yankee Station, east of Da Nang, but was not yet in range of the targets. The first sorties were launched about an hour before Johnson began his address at 11:37 p.m.
Two aircraft were lost during the strikes, one of them an A-4C Skyhawk flown by Lt. j.g. Everett Alvarez Jr., who was captured and became the first American POW of the Vietnam War.
Future President Nixon and Republican presidential candidate Goldwater issued public statements of support and on Aug. 7, Congress adopted the Tonkin Gulf Resolution, giving Johnson the backing and authority he wanted.
The Search for Proof
Capt. John Herrick (l), DeSoto commander, and Cmdr. Herbert Ogier, Maddox commander, on the destroyer in August 1964. Herrick would come to believe that the “torpedoes” heard were in fact echoes from Maddox’s outgoing sonar beams.(US Navy photo)
The Pentagon, obviously uneasy with the evidence in hand, called for supporting information in a pointed message sent to US Pacific Command Aug. 6, with copies to the destroyers and the carriers.
“An urgent requirement exists for proof and evidence of second attack by DRV [Democratic Republic of Vietnam] naval units against TG 72.1 [the destroyers] on night 4 Aug. as well as DRV plans and preparation for the attack, for previous attacks, and for any subsequent operations,” the message said. “Material must be of type which will convince United Nations Organization that the attack did in fact occur.”
The carrier pilots were “re-debriefed” Aug. 7 and Aug. 11. The Department of Defense sent two high-ranking officials for the second re-debriefings: Both of them were lawyers.
The JCS message drew additional confirmation that the attack had happened, and the re-debriefings found an A-1 crew that had seen bursts of light (possibly gunfire from destroyers) and tracers (possibly from other US aircraft). The report sent forward amplified the possibility that this was a sighting of the enemy. In 1968, McNamara cited the crew statement in testimony and construed it as a significant element of proof of attack.
A Harris poll showed a huge jump in public approval for Johnson’s handling of the war. Vietnam was neutralized as a campaign issue, and in November, Johnson won the election in a landslide.
Johnson had doubts of his own but he did not agonize over the possibility of error. He told Undersecretary of State George W. Ball, “Hell, those dumb, stupid sailors were just shooting at flying fish!” Later, when many in Congress criticized the war, Johnson said, “Congress gave us this authority, in August 1964, to do whatever may be necessary.”
Questions and Challenges
Leaks and rumors about the Tonkin Gulf continued to circulate. The Senate Foreign Relations Committee, to which a number of the leaks had been directed, managed to obtain the official logbooks for Maddox and Turner Joy and opened hearings in February 1968. The White House attempted to block the inquiry, but was unable to do so.
The key witness was McNamara, who repeated his assurance that “while sonar and radar readings may be subject to interpretation and argument because of sea and atmospheric conditions, we had intelligence reports of a highly classified and unimpeachable nature which established, without question, that the attacks took place on both Aug. 2 and Aug. 4.”
Sen. J. William Fulbright, the committee chairman, accused McNamara of “selective declassification of security material,” but the committee released its report without stating a conclusion. The DeSoto commander, Herrick, told United Press International that he still believed some of the sonar contacts had been torpedoes.
Congress repealed the Tonkin Gulf Resolution in 1970, but did so in the context of the war powers of the President and made no particular statement about the facts of the Tonkin Gulf incidents.
Nevertheless, the credibility of the government’s story was hanging by a thread. “I think it was a complete phony,” Goldwater said in 1980. “I think that Johnson plain lied to the Congress and got the resolution.”
Further revelations appeared in books and articles. After reviewing the facts, Herrick told US News & World Report in 1984 that most likely, there had been no torpedoes. “It was the echo of our outgoing sonar beam hitting the rudders, which were then full over, and reflected back into the receiver,” he said. “Most of the Maddox’s, if not all of the Maddox’s, reports were probably false.” Herrick stated that conclusion again in 1985 and 1986.
On the other hand, in 1986 an official history, The United States Navy and the Vietnam Conflict, said that the evidence of a second attack in the Gulf of Tonkin was conclusive. McNamara moderated his position in his memoirs in 1995 but did not concede altogether. “The evidence of the first attack is indisputable,” he said. “The second attack appears probable but not certain.”
The Missing Pieces
Eventually, two historians finally pieced together the full story. Edwin E. Moise’s Tonkin Gulf and the Escalation of the Vietnam War, published in 1996, used official documents and correspondence to reconstruct the night of Aug. 4 in great detail, plotting on maps the courses of the ships and each of the radar skunks.
Further information about signals intelligence, previously secret, appeared in, “Skunks, Bogies, Silent Hounds, and the Flying Fish,” by Robert J. Hanyok in NSA’s classified journal Cryptologic Quarterly in 2001. The article included the texts of the key intercepts and other details. It was declassified for public release in November 2005 with some sections blacked out.
After the Aug. 2 attack, NSA set up a team to collect and analyze the field intercepts. Responsibility was left to midlevel management, which probably felt the pressure to produce “proof” of the Aug. 4 attack.
According to Hanyok, the NSA summaries were “deliberately skewed to support the notion that there had been an attack.” In preparing its reports, NSA used only 15 of the 122 available intercepts, selecting those that fit the official scenario. “US Sigint never intercepted anything associated with an [Aug. 4] attack,” Hanyok said.
Meanwhile, McNamara had recanted about the Aug. 4 attack. Interviewed for “The Fog of War,” a documentary film in 2003, he said, “Events afterward showed that our judgment that we’d been attacked that day was wrong. Didn’t happen.”
Edward J. Marolda, lead author of the 1986 Navy history, changed his mind as well. In a summary done in 2005 for the Naval History and Heritage Command, he said, “More recent analysis of that data and additional information gathered on the 4 Aug. episode now makes it clear that North Vietnamese naval forces did not attack Maddox and Turner Joy that night in the summer of 1964.”
Some Navy veterans insist that Moise, Hanyok, and all the others are wrong and that the Aug. 4 attack did happen. Their arguments essentially boil down to the accusation that eyewitness sightings by deck and bridge crews have been disregarded or discounted. Their points are overwhelmed by the weight of other evidence.
The misjudgments of Aug. 4-5 were not deliberate. For the most part, a chain reaction of mutually reinforcing mistakes built up to conclusions decision-makers were predisposed to believe.
These events did not start the war. It had begun much earlier—instigated, directed, and sustained by the North Vietnamese. Both North Vietnam and the United States were deeply committed. Escalation was probably inevitable, but it was the Tonkin Gulf imbroglio that determined how and when the buildup of US forces took place and shaped the authority under which the war was fought.
Whatever else, a cloud hangs permanently over the record of that amazing night in the Tonkin Gulf and the actions related to it.