On the morning of December 7, 1941, two Army privates at a mobile radar set on the north shore of Oahu saw the largest return they had ever seen. They checked their equipment, tracked the contact and telephoned it in, roughly fifty minutes before the first bombs fell. The report reached a lieutenant who read it as a flight of American bombers he knew to be due that morning. A separate warning from Washington was still moving through commercial telegraph channels when the attack began. This piece walks through the record of both failures and what they show about the receiving end of any warning system.

Opana Point, 7:02 a.m.: what the radar showed and what the operators did

In late 1941 the Army’s Aircraft Warning Service on Oahu was a small network of mobile SCR-270 radar sets placed around the island. Each set reported by telephone to an Information Center at Fort Shafter, near Honolulu. The Opana station sat on high ground near Kahuku at the northern tip of the island. That morning it was manned by Privates Joseph Lockard and George Elliott of the Signal Corps. The Hawaiian Department had scheduled the stations to operate from 4:00 to 7:00 in the morning. When the Opana crew’s watch ended at 7:00, the truck that was supposed to take them to breakfast had not yet arrived, so they left the set running and Elliott, the less experienced of the two, got more practice on it.

The SCR-270 was a long-range early warning set working in the VHF range. It showed a return as a spike on an oscilloscope trace, which gave the operator a range, and the operator read the bearing from the position of the rotating antenna. It could not measure altitude, and it had no electronic means of telling friendly aircraft from hostile ones. Estimating the number of aircraft from the height and character of the spike was a matter of operator judgment. I raise this because the set did exactly what it was built to do, which was to tell somebody that something large was in the air at a given range and bearing. Identifying what that something was belonged to the people downstream.

At about 7:02, a return appeared to the north at a range usually given as roughly 130 miles, based on the station’s plot and the operators’ later testimony. Lockard’s first suspicion was that the set was malfunctioning. He checked it and concluded the return was real. The contact was larger than anything either man had seen on that equipment. They plotted it as it closed on the island, and after some discussion between them Elliott called the Information Center. The call reached the switchboard operator, Private Joseph McDonald, who located the only officer present, Lieutenant Kermit Tyler, and Lockard then spoke with Tyler directly. The two privates kept tracking the flight after that conversation until they lost it in the ground return near the island, about twenty miles out and around 7:40 by the commonly cited timeline.

The Information Center at Fort Shafter: a system still in training

The Information Center was designed to work much like the filter and operations rooms the British had built. Station reports would be plotted on a central table, a controller would decide which tracks were hostile, and pursuit aircraft would be sent to intercept them. The site’s article on the Dowding system covers how Britain made that process work, with filter officers and electronic identification of friendly aircraft. In December 1941 the Oahu center was not at that stage. The inquiries described an organization still in a training status. Its staffing was incomplete, and the relationship between the Army’s warning service and the Navy, whose aircraft also flew around the island, had not been fully worked out.

The plotters who had worked the board during the 4:00 to 7:00 window had left when it closed. Tyler was a pursuit pilot with the 78th Pursuit Squadron, assigned to the center that morning as pursuit officer. By most accounts it was only his second tour there. He had not been trained as a controller, and the job he was filling had no written description of what to do with an unidentified track. He was the only officer in the room. A report from a radar station therefore reached a person who held the position without the training, staff or instructions to act on it.

The center also had no reliable way to know what friendly aircraft were airborne at any moment. Nobody routinely fed it schedules of Army bomber movements or Navy carrier flights. With no electronic identification on the radar, a large formation on the scope was simply a large formation, and identifying it depended entirely on whether the person at the table happened to know about a matching friendly flight. The design of the 1941 system made this unavoidable. The question was what that person would conclude, and nothing in the procedure helped him test his conclusion before he gave it.

The B-17s and the radio station: an explanation that fit what Tyler knew

Tyler knew that a flight of B-17 bombers, usually given as twelve aircraft, was due in from the mainland that morning. He had also heard Honolulu radio station KGMB playing music through the night on his drive to Fort Shafter. A bomber pilot he knew had told him the station stayed on the air overnight when bombers were being ferried in, so that the crews could use the broadcast to navigate. The B-17s were coming from California, which lay to the northeast. A large formation approaching from roughly the north fit that expectation well enough for a man with no training and no better information. Tyler also later said he had considered that the contact might be Navy aircraft.

Tyler told Lockard, in substance, not to worry about it. That is how the exchange is consistently reported from the testimony, and I have paraphrased it rather than quoting a particular wording. The accounts given to the later inquiries differ on one point that matters a great deal: how much of the contact’s size actually reached Tyler. Lockard’s account stresses that the return was unusually large. Tyler’s account is that he did not understand it to be much larger than the expected bomber flight. Both versions are in the record, and I do not think a reader should resolve the conflict in either direction. The fact that the size of the contact could be disputed afterward is itself part of the failure.

The B-17s did arrive that morning, in the middle of the attack. They had been configured for a long ferry flight rather than for combat, and they landed wherever they could while Japanese aircraft were over the island. Several were damaged. Tyler’s explanation was therefore based on a real flight that really was approaching Oahu at about that time. It was a plausible explanation that happened to be wrong. A trained controller would be expected to treat that kind of conclusion as a hypothesis and check it, and Tyler had neither the tools nor the procedure to do so.

The receiving end: where a delivered report stopped being a warning

The Opana report passed every test a radio technician would apply. The equipment worked, the telephone circuit worked, and the call reached the designated destination in a few minutes. The failure came afterward, at the point where a person had to turn a report into a decision. Communications people tend to treat that step as somebody else’s problem. Nothing in the exchange required the size of the contact to be stated in a fixed form and read back. Nothing gave the radar crew a channel around the receiving officer if they disagreed with him. Nothing required the receiving officer to escalate an unexplained track to anyone with authority to act. The whole warning function rested on one man’s interpretation, and that man had been placed in the chair without preparation.

In my years of reviewing recorded radio and telephone traffic for quality assurance, the version of this failure I saw most often was quieter. A field unit or a caller reports something that does not fit the picture the dispatcher already holds, and the dispatcher acknowledges it in a way that sounds complete but files it under the existing picture. When the recording is played back, the sender usually thought the message had been understood, and the receiver usually thought it had been answered. The only fix I have seen work is procedural: require specific content, such as quantity, location and the sender’s confidence, to be stated and repeated back, and give a named person the job of deciding what happens to anything that remains unexplained.

A question that comes up whenever this history is taught is what would have happened if the Opana report had reached the commanders at 7:15. That is a counterfactual, and the record cannot settle it. The inquiries discussed it, and historians still argue about how much warning could have been put to use in the time available, given the alert status of the aircraft and guns on the island that morning. I will not pretend to answer it. The point that stands on the record is narrower: a detection was made and reported in time to reach a decision-maker, and the system in place had no mechanism to make sure it did.

The Ward’s report: the same pattern on the Navy side

The Army radar contact was the second unexplained warning that morning. At about 6:45, more than an hour before the air attack, the destroyer USS Ward, patrolling off the entrance to Pearl Harbor, fired on a small submarine and dropped depth charges on it. Its commanding officer, Lieutenant William Outerbridge, reported the attack by radio to the Commandant of the Fourteenth Naval District. The report stated that the Ward had fired on and depth-charged a submarine operating in the defensive sea area. I am paraphrasing the substance here, and the exact text is in the inquiry record for anyone who wants it.

The report moved through the Navy’s duty officers slowly. Part of the reason given in the record is that unconfirmed submarine contacts had been reported in the area before and had not held up, and the staff wanted confirmation before treating this one as real. They were still verifying it when the aircraft arrived. Here a warning reached the duty level and was discounted because the receiving organization had learned, from earlier false reports, to treat this category of message as probably mistaken. Fire and EMS communications centers know that pattern as alarm fatigue: the commercial building that has sent three false fire alarms this month gets a slower and more skeptical response on the fourth, regardless of what the dispatch card says.

The Ward report and the Opana report were never combined. They travelled up separate Army and Navy chains, and no one sat at a point where a submarine engagement at the harbor mouth and an unidentified formation approaching from the north could have been seen together. Each report on its own could be explained away. In a single operations picture, the two together would at least have raised the question of whether they were connected. The Fort Shafter center had no Navy feed, so nothing it received could have prompted that question. Nobody in either chain had the job of comparing the two services’ unexplained reports.

November 27: a warning read as a sabotage alert

The interpretation problem appeared ten days before the attack as well. On November 27, 1941, the War Department sent Lieutenant General Walter Short, commanding the Hawaiian Department, a message saying that negotiations with Japan appeared to have ended for practical purposes and that hostile Japanese action was possible at any moment. The message directed him to undertake reconnaissance and other measures he judged necessary. It also cautioned him against alarming the civilian population or revealing his intentions. On the same day the Navy sent Admiral Husband Kimmel, commanding the Pacific Fleet, a message that described itself in its opening words as a war warning. The exact texts are in the inquiry record, and I have paraphrased them here.

Short placed his command on the lowest of the three alert levels in the Hawaiian Department’s plan. That level was oriented toward sabotage by the island’s large population, which the command was concerned about, rather than toward external attack. He replied to Washington that his department had been alerted to prevent sabotage and that he was maintaining liaison with the Navy. One practical consequence, documented in the inquiries, was that Army aircraft were parked close together in the open, where they were easier to guard against saboteurs on the ground and far more exposed to an attack from the air.

A communications person should read that exchange closely, because the loop was technically closed. The warning went out, the recipient acknowledged it, and his acknowledgment stated plainly how he had interpreted it. The War Department received that reply. The later inquiries pointed out that nobody in Washington compared it against what the original message had intended and sent a correction. A readback is useful only if the person who sent the message actually checks it against what was meant. If the sender files the acknowledgment without reading it, the readback gives the appearance of verification without any of its value.

December 7 in Washington: the warning that went by commercial telegraph

On the morning of December 7, Washington had decrypted a long Japanese diplomatic message along with an instruction that it be delivered to the Secretary of State at 1:00 p.m. Washington time. Hawaii then kept a time zone ten and a half hours behind Greenwich, so 1:00 p.m. in Washington was about 7:30 a.m. in Honolulu. General George Marshall, the Army Chief of Staff, read the material in late morning and drafted a warning to the Pacific commands in longhand. The warning said the Japanese were presenting what amounted to an ultimatum at 1:00 p.m. Eastern time and had been ordered to destroy their code machines. It said Washington did not know what significance the hour had but that the commands should be on alert accordingly. Marshall later explained to the inquiries that he did not use the transpacific telephone because he was concerned it could be intercepted.

The Army signal center found that its own radio circuit to Hawaii was not getting through that morning because of atmospheric conditions. The officer handling the message sent it by commercial service instead, by Western Union to San Francisco and then by RCA radio to Honolulu, believing that route would be fastest. The message was filed around noon Washington time. It did not carry a priority marking that would have moved it ahead of routine traffic, and the record indicates Marshall was not told it had gone by the commercial route. It reached the RCA office in Honolulu shortly before the attack began and then had to be carried by messenger to Fort Shafter, through a city under attack. It was decoded there and reached Short in the afternoon, hours after the attack was over. Accounts differ by some minutes on each step, and the congressional record is the reference for the timeline.

Even if it had arrived on time, the message would not have told Hawaii it was about to be attacked. It warned that something might happen at a particular hour, at a place Washington could not name. The communications failure is still clear enough to state plainly. The primary circuit was down. The alternate route was chosen on an assumption about its speed rather than on a tested figure. The precedence marking that existed to move urgent traffic ahead of routine traffic was not used. And the originator did not learn which path his message had taken until long after it mattered.

What the inquiries established, and the myths that outlived them

The attack produced a long series of official investigations. The Roberts Commission, chaired by Supreme Court Justice Owen Roberts and appointed in December 1941, reported in January 1942 and concluded that Kimmel and Short had been derelict in their duties. In 1944 the Army Pearl Harbor Board and the Navy Court of Inquiry spread responsibility more widely, including to officials in Washington. Further investigations followed, and the Joint Congressional Committee on the Investigation of the Pearl Harbor Attack issued its report in 1946, with majority and minority views that did not agree on how to divide the blame. The hearings and report were published by the Government Printing Office and are available through federal depository libraries and digitized government collections. Anyone teaching this material should read those documents themselves rather than rely on a summary, including this one.

Several stories have outlived the record. The claim that the radar failed is a myth, as covered above. The claim that Kimmel and Short received the Opana warning and ignored it is also a myth: the report stopped at the Information Center and did not reach either commander before the attack. Tyler is sometimes described as having been court-martialed or formally blamed, and that is also untrue. He was not court-martialed. The inquiries generally treated him as an inexperienced officer placed in a role he had not been prepared for, and he continued to serve as a pilot during the war. The theory that Washington knew in advance of an attack on Pearl Harbor specifically and withheld the warning has circulated since the 1940s. The majority of the Joint Committee did not find support for it, and I treat it as a claim rather than as history.

What the record does support is less dramatic and more useful to people who run communications systems. The detection hardware worked. Every warning that morning, from the Ward, from Opana and from Washington, either reached a receiving point that could not interpret it or travelled by a route that was too slow. The arguments about Kimmel’s and Short’s reputations have continued for decades, including in Congress, and they are beyond the scope of this article. The communications findings, by contrast, are not seriously disputed.

What to do at your agency

• The communications center supervisor pulls the recordings of three recent calls in which a field unit or caller reported something the telecommunicator did not expect, and writes one paragraph on each stating whether quantity, location and the reporter’s confidence were asked for, logged and read back.
• The radio system administrator or communications manager identifies the backup route by which state warning point or National Weather Service messages reach the dispatch floor when the primary circuit is down, sends a real test message over it, and records the measured delivery time in the continuity plan.
• The communications manager adds one item to the next supervisors’ meeting already on the calendar: the name and phone number of the person an overnight telecommunicator escalates an unexplained report to, and what that person is authorized to do on receiving it.
• The training officer writes into the existing warning-receipt procedure a requirement that any warning received from the state or the National Weather Service be acknowledged with a readback of the hazard and protective action, and that the sender’s confirmation of that readback be logged.
• The emergency management director confirms whether the list of expected activity, including planned events, scheduled drills, military or air medical training flights and controlled burns, actually reaches the dispatch floor, who updates it, and whether each entry carries an end time so that old entries do not become explanations for new reports.
• The fire or EMS chief asks the CAD administrator for the current list of precedence or priority settings used on outbound alerts and pages, and confirms in writing that life-safety warnings are sent at the highest setting by default rather than at the operator’s discretion.

Takeaways

• Privates Joseph Lockard and George Elliott at the Opana radar station detected the incoming formation at about 7:02 a.m. on December 7, 1941, and reported it correctly, so any account that blames the radar or the operators is wrong.
• The report reached Lieutenant Kermit Tyler at the Fort Shafter Information Center, a facility still in training status, where he was the only officer present and had not been trained as a controller.
• Tyler matched the contact to a real flight of B-17 bombers due from California that morning, and the testimony given to the inquiries differs on how much of the contact’s size reached him.
• The destroyer Ward reported attacking a submarine more than an hour before the air attack, and the report was slowed in part by caution after earlier submarine reports had not been confirmed.
• Short’s November 27 reply to Washington stated that his command was alerted against sabotage, and the inquiries noted that nobody in Washington compared that readback against the warning’s intent or corrected it.
• Marshall’s December 7 warning went by commercial telegraph without a priority marking after the Army radio circuit to Hawaii failed, and it reached Short hours after the attack ended.
• Whether earlier warning would have changed the outcome is a counterfactual that the record does not settle, and it should not be taught as though it does.
• A warning system is only as good as its receiving end, which needs fixed message content, readback that the sender actually checks, a named person to escalate to, and alternate routes whose delivery times have been measured.