Editing Gulf of Tonkin incident (section)

===Second alleged attack===
[[File:Sonar Console C Mk 1 Mod 0, Westinghouse Electric Corporation - National Electronics Museum - DSC00484.JPG|left|thumb|upright=.8|A sonar console]]

On 4 August, another DESOTO patrol off the North Vietnamese coast was launched by ''Maddox'' and ''Turner Joy'', to "show the flag" after the first incident. This time their orders indicated that the ships were to close to no less than {{convert|11|mi|km}} from the coast of North Vietnam.<ref name="autogenerated1" /> During an evening of rough weather and heavy seas, the destroyers received radar, sonar, and radio signals that they believed signaled another attack by the North Vietnamese navy. For some two hours (from about 21:40 to about 23:35, local time) the ships fired on radar targets and maneuvered vigorously amid electronic and visual reports of enemies. Despite the Navy's claim that two attacking torpedo boats had been sunk, there was no wreckage, bodies of dead North Vietnamese sailors, or other physical evidence present at the scene of the alleged engagement.<ref name="Prados">{{cite web |url=http://www.gwu.edu/~nsarchiv/NSAEBB/NSAEBB132/essay.htm |access-date=October 29, 2009 |archive-date=September 25, 2009 |date=August 4, 2004 |title=Electronic Briefing Book No. 132 - Essay: 40th Anniversary of the Gulf of Tonkin Incident |website=National Security Archive at [[George Washington University]] |last=Prados |first=John |archive-url=https://web.archive.org/web/20090925161716/http://www.gwu.edu/~nsarchiv/NSAEBB/NSAEBB132/essay.htm |url-status=live }}</ref><ref name=Ha01/>

At 01:27 local time (13:27 [[Washington, D.C.|Washington]] time), Herrick sent a cable in which he acknowledged that the second attack may not have happened and that there may actually have been no Vietnamese craft in the area: "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."<ref name="Ellsberg9-10">{{Harvnb|Ellsberg|2002|pp=9–10}}.</ref><ref name=Ha01/>

{{Quote box
| quoted = true
| bgcolor = #ADD8E6
| width = 40em
| align = right
| quote = ...{{nbsp}}And ultimately it was concluded that almost certainly the [4 August] attack had occurred.  But even at the time there was some recognition of a margin of error, so we thought it highly probable but not entirely certain.  And because it was highly probable—and because even if it hadn't occurred, there was strong feeling we should have responded to the first attack, which we were positive had occurred—President Johnson decided to respond to the second [attack]. I think it is now clear [the second attack] did not occur{{nbsp}}...
| source = –Defense Secretary [[Robert McNamara]], 1996<ref name=":2"/>
}}

One hour later, Herrick sent another cable, stating, "Entire action leaves many doubts except for apparent ambush at beginning. Suggest thorough reconnaissance in daylight by aircraft."<ref name="Ellsberg10">{{Harvnb|Ellsberg|2002|p=10}}.</ref> In response to requests for confirmation, at around 16:00 Washington time, Herrick cabled, "Details of action present a confusing picture although certain that the original ambush was bona fide."<ref name=Ellsberg10/> It is likely that McNamara did not inform either the president or Admiral U. S. Grant Sharp Jr. about Herrick's misgivings or Herrick's recommendation for further investigation.<ref>Gareth Porter, ''Perils of Dominance: Imbalance of Power and the Road to War in Vietnam'' (University of California Press: 2005), pp. 194–95.</ref> At 18:00 Washington time (05:00 in the Gulf of Tonkin), Herrick cabled yet again, this time stating, "the first boat to close the ''Maddox'' probably launched a torpedo at the ''Maddox'' which was heard but not seen. All subsequent ''Maddox'' torpedo reports are doubtful in that it is suspected that sonarman was hearing the ship's own propeller beat"{{sic}}.<ref name="Ellsberg10" />

In the face of growing uncertainties over the course of the day regarding whether the attack had occurred, the Johnson administration ended up basing its conclusion mostly on communications intercepts erroneously assessed to be North Vietnamese preparations to carry out an attack and a North Vietnamese after action report.  Later analysis showed those communications to have concerned the recovery of torpedo boats damaged in the 2 August attack and North Vietnamese observations of (but not participation in) the 4 August U.S. actions.<ref name=Ha01/>