The VVA Veteran® Online

July/August 2014


Gulf of Tonkin: Ambiguous Push to War


No dawn telephone call ever brought good news. At least Tom Hughes thought so. Thomas L. Hughes led the Bureau of Intelligence and Research (INR) at the State Department—the diplomats’ counterpart to the CIA. His job was to keep Secretary of State Dean Rusk up to speed. This Sunday morning proved no exception. A subordinate at the Department’s Operations Center told Hughes that a U.S. destroyer had just been attacked off the Vietnamese coast.

It was August 2, 1964, and the warship Maddox had indeed been in an intense battle with several torpedo boats. The incident took place in the afternoon in the Tonkin Gulf, before dawn in Washington. Hughes phoned Rusk. The secretary of state told Hughes to come over for a quick meeting at his home.

Hughes thought the gathering highly amusing in retrospect, but it was deadly serious at the time. The secretary of state and other top officials knelt over a map of Vietnam on the living room floor. A Navy briefer related the sketchy details that were known. A CIA man did not say much of anything.

Cyrus L. Vance, representing the secretary of defense, insisted that the attack had been unprovoked against a vessel innocently sailing in international waters. Rusk decided he would recommend a strong response. Most of the group went on to the White House, where they met with President Lyndon B. Johnson at 10:30 a.m.

In actuality the U.S.S. Maddox had been engaged in a DeSoto patrol, a secret mission in which the warship recorded emissions from coastal radars and other North Vietnamese radio transmissions for communications intelligence experts at the National Security Agency. South Vietnamese naval commandos simultaneously raided coastal targets in the North as part of a coordinated program of coercive measures known as Operation Plan 34-A.

The Maddox had nosed close to several North Vietnamese islands, although she had returned to international waters when the North Vietnamese torpedo boats arrived to contest the sea. The destroyer beat off the attack handily, sinking one torpedo boat and damaging the others.

Aside from the niceties of who was in what waters when, most of this was known when officials met with President Johnson. Washington was not on alert. LBJ asked for McGeorge Bundy, his national security adviser, and William P. Bundy, the assistant secretary of state for Far Eastern Affairs. They were relaxing on Martha’s Vineyard. Robert S. McNamara, the secretary of defense, was in Wisconsin climbing a mountain.

Johnson was told that someone had attacked a Navy warship. When LBJ asked who would do a thing like that, the answer was Hanoi. The president then asked if that was all that was known. The Navy answered yes. When LBJ pressed, the subject of 34-A operations came up immediately. Washington, therefore, knew from the beginning that a provocation had been involved.

Only a few minutes were necessary for LBJ to reach a conclusion. “It sounds to me like what happens in the dark at the movies in Texas,” the president told his advisers. “A fellow puts his hand on a girl’s knee. If nothing happens, he moves it up a few inches, and then a few more, until it gets slapped.”

Johnson then looked around the room. “Well boys, it looks like we just got slapped.”

The president left for a few moments. While he was out Tom Hughes scribbled a note and passed it to Secretary Rusk, who sat across from him. “Now that we know what happens in the movies in Texas,” Hughes wrote, “do you still want to go for a strong response?” Rusk desisted.

There was no protest when President Johnson returned and mandated restraint. A second destroyer, the C. Turner Joy, would reinforce Maddox, and both warships would head up the Gulf to assert U.S. presence. The Maddox should continue her intelligence collection for the NSA. The unit was instructed to avoid approaching the North Vietnamese coast when 34-A raids were under way. A dispatch to this effect went out from Gen. Earle Wheeler, chairman of the Joint Chiefs of Staff, at 12:25 p.m.

The Maddox’s DeSoto cruise was the latest in a series of offshore missions by U.S. destroyers. The first, in April 1962, had been off the Chinese port of Qingdao by the De Haven. The initial sally into the Gulf of Tonkin was made by the Agerholm that December. Then came the Edwards in April 1963 and the Craig eleven months later. By then DeSoto patrols had become routine.

A large empty van shell arrived at the U.S. Taiwan Defense Command in late 1963. It could be packed with interception equipment for a “direct support unit,” and the hardware could be changed depending on specific interests. Typhoon Gloria, which hit Taiwan in September 1963, left the van a stinking mess. It had to be cleaned up before the spooks could use it.

The Maddox sortie resulted from a request by Gen. William C. Westmoreland, the U.S. commander in Vietnam, for intelligence on North Vietnamese defenses being targeted by 34-A. These 34-A raids were more commonly referred to by the nickname “Kitkat.” For Maddox’s Tonkin Gulf cruise the van was equipped with four receiving posts to handle different frequencies and types of electronic transmissions. The Naval Security Group (NSG) detachment on the vessel consisted of a Taiwan-based officer, Lt. Gerrell D. Moore, plus fifteen sailors and Marines.

They, the Maddox’s skipper, and mission leader Capt. John J. Herrick were briefed on Taiwan. Cmdr. Herbert L. Ogier, Jr., captain of the Maddox, asked whether to expect Hanoi to react. Taiwan NSG chief Norman Klar replied that it was very unlikely. Mission commanders were told nothing of Kitkat. On the other hand, U.S. intelligence stations were informed that the DeSoto patrol’s purpose was to provoke a North Vietnamese response and record what that might be.

The destroyer sailed from Keelung on July 28, 1964. Quiet prevailed until the early morning of August 1, when electronic intelligence first observed North Vietnamese radar tracking the ship. Several radar units were involved and, with an interruption that night, the Vietnamese continued to report the destroyer’s position right through the August 2 attack.

During part of this time the radar station at My Duc sent its reports directly to the patrol boat T-146, which relayed messages to the North Vietnamese torpedo craft. While proceeding in the Gulf the Maddox actually saw the Kitkat boats. Based on electronic and communications intelligence from the NSG station at San Miguel, Lt. Moore warned Capt. Herrick of a probable attack some twelve hours before it took place.


Word of a sea fight in the Tonkin Gulf spread rapidly and was soon reported on commercial radio in the United States. Bill Gerhard, an NSA linguist with the B-Group section responsible for Vietnam, was with his family on the way to church that Sunday morning when he heard the news on the radio. Gerhard recalled saying, “Well, it looks like it will be a short church service today and I’ll [be getting] very little sleep for a while.”

That’s precisely what happened at Fort Meade. Cots were brought to B-205, the Vietnamese section, and a 24-hour watch set up. Gerhard, Air Force Lt. Col. Delmar C. Lang, the section chief, and division chief Milton Zaslow were the sparkplugs. They followed the Maddox dispatches as President Johnson decided on his response.

Then a remarkable thing happened: NSA reporting began to diverge. On August 3 the Vietnamese radar apparently stopped following the Maddox’s whereabouts. An agency summary compiled a month later termed radar activity “negligible” at that point. One Far Eastern station, observing the volume of Hanoi’s radio traffic on August 3, found it normal. But the facility at San Miguel reported North Vietnamese boats receiving instructions to shadow the Maddox, now joined by the C. Turner Joy. San Miguel recorded at least one position report. But there was nothing like the continuous stream of messages that had preceded the August 2 naval engagement.

Early in the morning of August 4, Washington time, a new NSA intercept arrived. The listening post at Phu Bai reported that the Vietnamese naval command at Haiphong had alerted a patrol boat for action and informed it that another boat also would participate. In addition, Haiphong was trying to fuel a torpedo boat. Based on this message, aboard the Maddox Lt. Moore told Capt. Herrick there would be a new attack. An hour later Herrick sent a dispatch saying his radar had detected ships and aircraft closing in. The seas were heavy and the air rife with atmospheric disturbances. The cloud ceiling was down to two thousand feet. What followed was a wild “fight.”

The story of that dark night in the Gulf, in which Turner Joy shot off 249 5-inch and 123 3-inch shells and Maddox launched 29 5-inch and 95 3-inch shells at phantom torpedo boats, has often been told. Capt. Herrick’s message at half-past midnight casting doubt on the reality of the events also is well known. The more important point is that all the key decisions in Washington were made between that initial warning message and Herrick’s dispatch disavowing his previous dispatches.

At 9:43 a.m. Washington time—the precise moment when the Maddox was firing star shells to try to see objects that had disappeared from its radar screen—President Johnson was on the telephone with Defense Secretary McNamara telling him he wanted to be more than reactive. An hour later McNamara, back on the line, said he was bringing “a list of alternative target systems” to lunch at the White House. McNamara, McGeorge Bundy, Rusk, and the Joint Chiefs discussed the target list at 11:40 a.m., fifty minutes before Capt. Herrick voiced his misgivings. The die was cast.

The Vietnamese message that the NSA interpreted as an attack order, it turns out, had been mistranslated. It had been an instruction to rescue sailors and recover a boat damaged in the action of August 2. Another message, taken as an after-action report, described the situation of August 2 and had been delayed in decryption. Coupled with the contradictory NSA reporting from the day before the phantom battle, the result was as confusing as the alleged attack itself.

Louis Sarris, the Vietnam desk chief for State’s Bureau of Intelligence and Research, was at the department’s Operations Center at the time. He recalls immediately recognizing the ambiguities. At the NSA, views were divided. “There was no question that the second attack a couple of days later was not an attack,” Delmar Lang said in 1988. Gerhard (whose identity is redacted from the source document) said: “We had nothing to indicate in SIGINT [signals intelligence] that an order had come out from anyone to go out and take on the vessels again.”

Gen. Gordon Blake, the NSA director, said later that “the SIGINT was very shaky” for the August 4 affair. But Milton Zaslow insisted “there have been many arguments since then that there was a spurious raid that never took place. That’s not what we felt we saw.”

The point was not lost on the White House. When the National Security Council staff gathered on the morning of August 5, the question of the NSA intercepts came up right away. The White House insisted on seeing them.

Several years later—after the 1968 Tet Offensive—NSA officials interviewed White House Situation Room chief Arthur McCafferty on the subject of how the president and his staff made use of SIGINT in decisions on Vietnam. McCafferty said that the White House used both “finished” intelligence—such as NSA summaries—and “raw” intercepts, and that the direct use of SIGINT began at 1600 Pennsylvania Avenue in August 1964.

A month later there was another DeSoto patrol, and an experience similar to the alleged August 4 incident befell two more American destroyers. President Johnson was heard to grouse, “They could be shooting at whales up there in the Tonkin Gulf,” and this time he rejected any idea of retaliation against North Vietnam.


But the episode raised the question of a postmortem on both affairs. At the time John McCone was the director of Central Intelligence and could have initiated such a study. He did not. Tom Hughes never heard McCone express any interest in a postmortem. In fact, he heard the opposite.

Adlai Stevenson, the U.S. ambassador at the United Nations, wanted to denounce Hanoi at the Security Council along lines similar to his stunning presentation condemning Russia during the Cuban Missile Crisis. For that he needed the intel that McCone refused to provide.

Gen. Blake at NSA did want a postmortem. Milt Zaslow reported at the end of August that the study was beginning, and Delmar Lang would be lead analyst. But the NSA postmortem was overtaken by events following the September non-incident, after which Lyndon Johnson ordered the chair of his President’s Foreign Intelligence Advisory Board (PFIAB), Clark Clifford, to do his own review. The report NSA furnished PFIAB in October proved quite different from the study Zaslow described. Signed by Lang, the NSA chronology relied on the delayed report and mistaken translation of August 4, minimized North Vietnam’s apparent lack of interest in the Maddox on the previous day, and emphasized only evidence that suggested a real battle.

The original of the mistaken translation disappeared from NSA files. Agency historian Robert Hanyok, revisiting this episode much later, discovered more than 120 relevant messages in the files and found that the Lang chronology had suppressed more than a hundred of them.

PFIAB’s conclusions remain secret to this day. But Clark Clifford, always a supremely political animal, had a vested interest in protecting Lyndon Johnson from charges that he acted recklessly in the Tonkin Gulf affair. How this affected the NSA report is nowhere explicit, but the agency seems to have bent to the prevailing winds.

In June 1972 Gen. Blake, interviewed for yet another retrospective on Tonkin, said of the Johnson White House: “It appears as if they latched onto the rather shaky SIGINT evidence and decided to retaliate possibly because they already wanted to [do] this [and the evidence] added the fuel that they needed.” Blake concluded: “It looks like we were stuck with the story.”

By then, McNamara and Rusk already had gone to Capitol Hill with the “true attack” version and obtained the Gulf of Tonkin Resolution. It took decades to unravel that story—and Tonkin facilitated the American war in Vietnam.

John Prados is the author of five books on Vietnam, including The Hidden History of the Vietnam War. He is a contributing writer to The VVA Veteran.

Chapter 32 Honors the Forgotten.Love a Parade: Harrisburg, Pennsylvania, Chapter 542 Asheville, North Carolina, Chapter 124A Monumental Tribute:
in downtown Amherst, Ohio
The VVA Veteran® is a publication of Vietnam Veterans of America. ©All rights reserved.
8719 Colesville Road, Suite 100, Silver Spring. MD 20910 | | contact us