Written by AgentSwarm on 18 October 2026 from three public documents: the Rogers Commission report (Volume I), House Report 99-1016, and Richard Feynman's Appendix F. A briefer planned it, one writer per section wrote it from its assigned parts of the sources, and every statement cites the part (and, for the House report, the page) it comes from. Not yet reviewed. About AgentSwarm.
The Space Shuttle Challenger was lost on 28 January 1986 during mission 51-L, with structural breakup beginning at approximately 73 seconds after liftoff [rogers-commission-report-volum-04]. The Rogers Commission identified the failure of the joint between the two lower segments of the right Solid Rocket Motor as the cause, with the destruction of the seals that prevent hot gases from leaking through the joint [rogers-commission-report-volum-04] [rogers-commission-report-volum-01]. The House Committee on Science and Technology concurred that the failure was in the aft field joint of the right-hand Solid Rocket Motor, and it judged that the joint design was faulty [house-report-99-1016-investiga-p001]. The thesis of this report is that a known, recurring joint anomaly was normalized within the organization under schedule and policy pressure, and that a cold-weather launch exceeded the experience base [house-report-99-1016-investiga-p019] [feynman-personal-observations-01]. The sections that follow cover the technical cause, the design history, the launch decision, the organizational conditions, where the three bodies of evidence agree and differ, and what changed afterward.
The Rogers Commission concluded that the loss of Challenger was caused by a failure in the joint between the two lower segments of the right Solid Rocket Motor, specifically the destruction of the seals meant to stop hot gases from leaking through the joint during propellant burn [rogers-commission-report-volum-04]. The Solid Rocket Booster (SRB) is a Marshall-managed element whose Solid Rocket Motor (SRM) case is built from steel segments joined by tang-and-clevis joints held by 177 steel pins per joint [rogers-commission-report-volum-02] [house-report-99-1016-investiga-p199]. In a tang-and-clevis joint, the tang is the male end of one segment and the clevis is the female end of the adjoining segment, and the two O-rings sit in grooves in the clevis [house-report-99-1016-investiga-p199]. Joints made between the four casting segments at the launch site are called field joints [rogers-commission-report-volum-02]. The primary O-ring is the first seal against propellant gas; the secondary O-ring is a backup [house-report-99-1016-investiga-p199]. Putty, a zinc chromate material, sits in the joint to insulate the O-rings from hot gas [house-report-99-1016-investiga-p069]. The insulation lining the case protects the steel from the propellant's heat [house-report-99-1016-investiga-p069]. When the motor ignites, the case bulges and the joint rotates, which opens the gap between the tang and the O-ring face [house-report-99-1016-investiga-p199]. The Commission's tests found that the gap opening increase at the aft field joints was approximately 0.029 inches for the primary O-ring and 0.017 inches for the secondary O-ring [rogers-commission-report-volum-07]. The O-rings must move into their sealing positions within a few hundred milliseconds of ignition, because the gap opens quickly and the seal is only effective once the pressure acts behind it [rogers-commission-report-volum-07] [house-report-99-1016-investiga-p199]. Low temperature degrades this process: O-ring resiliency is severely reduced when cold, so the ring recovers slowly and can remain compressed in its groove [rogers-commission-report-volum-07]. The Commission's subscale tests showed that sealing at a 0.004-inch initial gap was not achieved at 25 degrees Fahrenheit and was marginal at 40 and 50 degrees [rogers-commission-report-volum-07]. Joint temperatures on the 51-L flight were calculated at 28 degrees Fahrenheit, plus or minus 5 degrees, at the lowest-temperature joint, with the right aft field joint possibly the coldest [rogers-commission-report-volum-07]. Putty can also delay pressure reaching the O-rings: in one test with conditioned putty at 75 degrees, pressure was delayed 530 milliseconds at the primary O-ring, and at 20 degrees the delay was 1.9 seconds [rogers-commission-report-volum-07]. The leak check, which pressurizes the volume between the O-rings, was raised from 50 psi to 200 psi, and the Commission noted that higher pressure makes blow holes in the putty more likely [rogers-commission-report-volum-07] [rogers-commission-report-volum-17]. Blow holes in putty allow hot gas to reach the primary O-ring, which erodes it; the Commission reported that after the 200 psi check was adopted, over half the Shuttle missions experienced field joint O-ring blow-by or erosion [rogers-commission-report-volum-17]. Water that entered the joints during pad exposure of about seven inches of rain could freeze and unseat the secondary O-ring in tests, though the House Committee notes this as a likely rather than proven condition [rogers-commission-report-volum-07] [house-report-99-1016-investiga-p069]. The House Committee's most probable sequence places smoke at ignition near a strut connecting the booster to the External Tank, where the joint was colder and the gap too small, and where a putty hole probably existed [house-report-99-1016-investiga-p069]. The Committee judged that the primary O-ring was unseated or eroded there and the secondary O-ring unseated, and that the breach clogged with burned char and aluminum oxide within about three seconds, stopping the smoke [house-report-99-1016-investiga-p069]. Turbulence and the throttle-up to 104 percent then probably broke the clog free and reopened the joint, after which a flame was seen near the strut and burned through the External Tank [house-report-99-1016-investiga-p069]. Photographs show the flame and the plume's attachment to the External Tank ring frame, and telemetry records the first evidence of flame at 58.8 seconds [rogers-commission-report-volum-04] [rogers-commission-report-volum-02]. The Commission's sequence of events places the first evidence of flame on the right SRM at 58.788 seconds after ignition and the start of divergent booster behavior near 72 seconds [rogers-commission-report-volum-04]. The House Committee found that the design of the field joint was unsatisfactory and could not reliably contain the burning propellant gases across the operating range, and that the O-ring materials and putty were unsatisfactory, particularly in winter [house-report-99-1016-investiga-p069]. The Committee also noted that the fluorocarbon O-ring material was the wrong choice at low temperatures, and that NASA and Thiokol could not control the quality of these materials because the formulas were proprietary [house-report-99-1016-investiga-p069] [house-report-99-1016-investiga-p078]. A central dispute concerned the "static seal" assumption: the O-rings were treated as static-seal gaskets, but the Committee argued the Shuttle joint was not static because the gap varied during ignition, and that the leak check did not prove the primary O-ring seated [house-report-99-1016-investiga-p199]. What the evidence shows is that the joint failed at the right aft field joint and that cold, putty behavior, and gap opening all reduced sealing margin; what it does not prove is the exact path of hot gas, because the Committee itself describes its sequence as the most probable one, not a certainty [house-report-99-1016-investiga-p069]. The temperature link is supported by the cold-joint analysis and the resiliency tests, but the failure mechanism at the joint itself was reconstructed from wreckage and data rather than observed directly [rogers-commission-report-volum-07] [house-report-99-1016-investiga-p069].
The Solid Rocket Booster (SRB) joint that failed on 28 January 1986 was not a new design choice made in haste; its basic concept was inherited from an earlier motor, and its problems were visible for years before the accident. [rogers-commission-report-volum-15] The Shuttle motor design was based primarily on the Air Force's Titan III solid rocket, and Thiokol hoped to reduce new design problems and cut costs by borrowing from the Titan design. [rogers-commission-report-volum-15] The Titan joints used one O-ring intended to be able but not required to take the brunt of combustion pressure, whereas the Thiokol O-rings were designed to take that brunt, with only an insulating putty as an additional gas barrier. [rogers-commission-report-volum-15] In November 1973 the SEB Design, Development and Verification Team Report rejected proposals to test motor performance at 40 and 90 degrees Fahrenheit, stating that the data could be obtained from normal variation in ambient conditions. [house-report-99-1016-investiga-p060] In December 1973 NASA selected Morton Thiokol for the solid rocket motors even though Thiokol ranked fourth of four bidders in the design category, because NASA placed greater importance on cost reduction and Thiokol had an attractive cost proposal. [house-report-99-1016-investiga-p060] Tests in 1977 showed that joint rotation opened the joint gap rather than closing it as the original analysis predicted, reducing rather than increasing pressure on the O-ring in the milliseconds after ignition. [rogers-commission-report-volum-15] In January 1978 John Miller of Marshall wrote that tang distortion could cause O-ring/tang separation and recommended redesign of the clevis joints on all on-coming hardware. [house-report-99-1016-investiga-p060] In January 1979 Marshall's Miller memo objected that the primary O-ring was being forced to function in a way that violates industry and government O-ring application practices, and that the secondary O-ring had been verified by tests to be unsatisfactory. [rogers-commission-report-volum-15] The Commission found that the Parker Seal experts' first thought was that the O-ring was being asked to perform beyond its intended design, and that a different type of seal should be considered. [house-report-99-1016-investiga-p297] No records showed Thiokol was informed of the 1979 supplier visits, and the O-ring design was not changed. [rogers-commission-report-volum-15] The first in-flight erosion of a primary O-ring occurred on STS-2 in November 1981, with 0.053 inch erosion in the right SRB aft field joint, and it was not reported in the Level I Flight Readiness Review for STS-3. [rogers-commission-report-volum-15] The joint was classified Criticality 1R from November 1980, and in December 1982 Marshall management accepted that the secondary O-ring was no longer functional once the joints rotated at 40 percent of maximum expected operating pressure, so the classification was changed to Criticality 1. [rogers-commission-report-volum-15] The Criticality 1 waiver approved by Level I on 28 March 1983 was described by the Commission as having been signed so that STS-6 could fly, a connection that Weeks denied. [rogers-commission-report-volum-16] Engineers and managers differed on whether the secondary seal was redundant, with Hardy and Mulloy holding that it was redundant except in worst-case tolerance stackups, a caveat absent from the Critical Items List. [rogers-commission-report-volum-16] In 1983 Thiokol's analysis of secondary O-ring redundancy considered joint or hardware geometry only, without considering the resiliency of the ring as affected by temperature. [rogers-commission-report-volum-15] After erosion was found on STS 41-B in February 1984, the O-ring erosion was first discussed as a technical issue at a Flight Readiness Review in March 1984, and the 0.090 inch maximum-erosion value was accepted based on a limited computer program with a 0.095 inch sealing test. [house-report-99-1016-investiga-p060] Following the erosion on STS 51-C, a Thiokol briefing called the field joint blow-by an "acceptable risk" and called the secondary O-ring a "redundant seal" even though the field joint had been Criticality 1 for two years. [house-report-99-1016-investiga-p060] The Commission found that the leak check pressure was raised to 200 psi, and that the new putty and the higher stabilization pressure were associated with more frequent O-ring damage. [house-report-99-1016-investiga-p060] Thiokol's 19 August 1985 briefing to NASA Headquarters listed frequency of O-ring damage increasing since Randolph putty, higher leak test stabilization pressures, and high performance motors were adopted. [house-report-99-1016-investiga-p060] On 31 July 1985 Boisjoly wrote that the rationale for flying the joint design was now suspect and that a field joint scenario would be "a catastrophe of the highest order-loss of human life." [rogers-commission-report-volum-18] The August 1985 briefing did not include temperature data on resiliency, and McDonald later said temperature may have been lost because a long range of temperatures had not been run. [house-report-99-1016-investiga-p166] The Committee noted that the briefing recommended an accelerated pace but concluded it was safe to continue flying, and that no participant recommended grounding the Shuttle. [house-report-99-1016-investiga-p166] In December 1985 Thiokol requested closure of the O-ring erosion problems, and the problem assessment system erroneously recorded a closure, so the problem did not appear as an open launch constraint at the 51-L Flight Readiness Review. [rogers-commission-report-volum-18] Acceptance of anomalies was framed as "within experience base" in the reviews after STS 51-1, and this wording recurred in the reviews for flights 61-A through 61-C. [house-report-99-1016-investiga-p060]
The evening of 27 January 1986 is where a known joint problem was, in the Rogers Commission's words, accepted as an acceptable flight risk rather than resolved [rogers-commission-report-volum-19]. The Commission found that Thiokol and NASA management came to accept erosion and blow-by as unavoidable and an acceptable flight risk [rogers-commission-report-volum-19]. The Commission's chronology sets out the sequence of calls among Kennedy, Marshall, and Thiokol personnel [rogers-commission-report-volum-13]. The first phase of the teleconference began at 5:45 p.m. EST, when Thiokol presented concerns about O-ring performance at low temperature and an opinion that launch should be delayed [rogers-commission-report-volum-10]. Stanley Reinartz, Manager of the Shuttle Projects Office at Marshall, testified that he made the decision not to escalate the concern to Level II [rogers-commission-report-volum-10]. The second phase began at about 8:45 p.m. EST, after Thiokol's charts reached Kennedy by telefax [rogers-commission-report-volum-10]. The temperature forecast for launch was 29 F at 9:00 a.m. and 32 F at 2:00 p.m. on 28 January [house-report-99-1016-investiga-p207]. The coldest previous launch was 51-C at 53 F, and Thiokol's recommendation was that O-ring temperature be at least 53 F at launch [house-report-99-1016-investiga-p207]. Thiokol's secondary O-ring resiliency chart showed recovery in 600 seconds at 50 F against 2.4 seconds at 75 F [house-report-99-1016-investiga-p207]. Blow-by had been seen on SRM-15 (51-C) at an O-ring temperature of 53 F and on SRM-22 (61-A) at 75 F, which weakened the temperature argument [house-report-99-1016-investiga-p207]. Boisjoly testified that he could not quantify his concern because he had no data to do so [rogers-commission-report-volum-10]. The Commission's later analysis found that all four flights with O-ring temperatures at 63 F or below showed thermal distress, against three of twenty flights at 66 F or above [rogers-commission-report-volum-19]. The managers' comparison used only flights with observed distress, not the full flight history [rogers-commission-report-volum-19]. Robert Lund presented the conclusion that the launch should not proceed outside the 53 F data base, and Thiokol management then requested a caucus off the net [rogers-commission-report-volum-10]. Jerald Mason asked Lund to put on his management hat, and the final launch rationale was read from a handwritten note by Joe Kilminster [rogers-commission-report-volum-10]. Boisjoly testified that no engineer or non-management participant spoke in favor of launch [rogers-commission-report-volum-10]. The House Committee concluded that the engineers' opinion, absent contrary evidence, should have been accepted until better information became available [house-report-99-1016-investiga-p078]. In effect, the question shifted from whether the joint had been shown to be safe to whether it had been shown unsafe, a reversal the Committee's account of the teleconference supports [house-report-99-1016-investiga-p236]. The Commission stated that the O-ring concerns never reached Jesse Moore or Arnold Aldrich, the Level I and II program officials [house-report-99-1016-investiga-p236]. The House Committee could not determine whether Moore or Aldrich would have stopped the launch had they been informed [house-report-99-1016-investiga-p236]. Ice on the launch pad was a separate concern: the ice team leader recommended a scrub, but that objection was apparently not conveyed to the STS Program Manager [house-report-99-1016-investiga-p078].
Programmatic pressure, communication failures, and weak safety oversight formed the organizational setting in which the 27-28 January 1986 launch decision was made. The Rogers Commission found that the Shuttle was declared operational and that the system was not prepared for an "operational" schedule [rogers-commission-report-volum-22]. The Commission stated that the flight rate "did not appear to be based on assessment of available resources and capabilities and was not reduced to accommodate the capacity of the work force" [rogers-commission-report-volum-22]. Spare parts were deferred under budget reductions, and 45 of approximately 300 required parts were cannibalized for Challenger before mission 51-L [rogers-commission-report-volum-22]. Engineers at Johnson said that their concerns about late manifest changes "do not carry much weight" against political advantages, a point the Commission recorded in the words of Leonard Nicholson [rogers-commission-report-volum-22]. The Commission's view was that the "operational" label should not imply any less commitment to quality or safety [rogers-commission-report-volum-22]. Arnold Aldrich described five communication or organization failures that affected the launch decision; four related directly to faults within the safety program: lack of problem reporting requirements, inadequate trend analysis, misrepresentation of criticality, and lack of involvement in critical discussions [rogers-commission-report-volum-20]. In 1983, a change to the reporting requirements removed the reporting of flight safety problems, flight schedule problems, and problem trends, so that Level II "lost all insight" into these issues [rogers-commission-report-volum-20]. None of the Level III monthly problem reports issued by Marshall during 1984 and 1985 was distributed to Level II management [rogers-commission-report-volum-20]. The Commission found no trend analysis of the field joint O-ring history, which turned abrupt in January 1984, when more than half of the missions from the tenth flight onward had blow-by or erosion [rogers-commission-report-volum-20]. The field joint was classified Criticality 1 since 1982, but most problem reports listed it as Criticality 1R, which "perhaps" led managers to believe redundancy existed [rogers-commission-report-volum-20]. Marshall's launch constraint on the nozzle joint after mission 51-B was not reported to Level II or Level I, although the governing document required it [rogers-commission-report-volum-20]. The Commission concluded that the Marshall Safety, Reliability and Quality Assurance organization reported to the director of Science and Engineering, which lacked independence from the producer of the hardware [rogers-commission-report-volum-20]. The House Committee found that the SR&QA organization at Marshall had apparently been declining in number since the Apollo program, and that it showed no input into Problem Assessment System reports on joint erosion [house-report-99-1016-investiga-p226]. The House Committee also found no evidence that SR&QA had evaluated the joint erosion problem or presented its evaluation at Level III Flight Readiness Reviews [house-report-99-1016-investiga-p226]. The Committee noted that the Thiokol recommendation was given to NASA management, and that Thiokol's written positive recommendation for launch contrasted with the Rockwell managers' refusal to give an unqualified go for ice concerns [house-report-99-1016-investiga-p157]. The House Committee also found that the present system let contractors "express concern" without having to say "stop the flight, it is unsafe" [house-report-99-1016-investiga-p157]. Waiver practice shows the same pattern: NASA personnel did not necessarily apply "a sufficiently rigorous engineering analysis" to waivers of launch commit criteria, as with the ullage pressure transducers qualified to "10 degrees" [house-report-99-1016-investiga-p257]. The House Committee added that the waivers did not directly contribute to the accident, but that the low temperatures during the night of the 27th and morning of the 28th most probably did [house-report-99-1016-investiga-p257]. Feynman reported that certification criteria in Flight Readiness Reviews "often develop a gradually decreasing strictness," so that prior flights without failure were accepted as an argument for safety [feynman-personal-observations-01]. Feynman set the estimates of loss-of-vehicle risk, which ranged from about 1 in 100 for working engineers to about 1 in 100,000 for management, and argued that the management figure reflected "management's fantastic faith in the machinery" [feynman-personal-observations-01]. In his account, erosion of the O-rings was accepted through a "safety factor of three" that he argued was a misuse of the term, because erosion was "a clue that something was wrong" [feynman-personal-observations-01]. Feynman concluded that NASA management "exaggerates the reliability of its product, to the point of fantasy" [feynman-personal-observations-01]. This conclusion is Feynman's interpretation of the estimates he collected, not a finding of the Rogers Commission [feynman-personal-observations-01].
**Common ground.** The House Committee found no credible alternative to failure of the right-hand Solid Rocket Booster (SRB) aft casing joint, and explicitly adopted the Rogers Commission’s finding that the External Tank did not cause or contribute to the accident [house-report-99-1016-investiga-p019]. The Committee concluded that sufficient warnings existed to prevent launch; Feynman likewise treated previous seal erosion and blow-by as warnings, not evidence of safety [house-report-99-1016-investiga-p078] [feynman-personal-observations-01]. Agreement on organizational weakness is reflected in the House’s endorsement of Rogers’s recommendation for a Safety, Reliability and Quality Assurance (SR&QA) office reporting directly to NASA’s Administrator [house-report-99-1016-investiga-p030].
**Different diagnoses of the decision process.** The House did not simply characterize the failure as inadequate communication: it found ample opportunities to surface problems, no evidence that managers suppressed information they considered significant, and a failure of technical decision-making rather than technical communication [house-report-99-1016-investiga-p030]. It judged that Flight Readiness Review procedures operated as designed but could not compensate for faulty engineering judgment [house-report-99-1016-investiga-p078]. Marshall should have reported the eve-of-launch teleconference to senior program managers, the Committee concluded, but it found no evidence that doing so would have changed the outcome [house-report-99-1016-investiga-p030].
Rogers’s analysis also extended beyond the immediate launch decision, describing flight-rate demands unsupported by resources and an inadequately prepared transition to operations [rogers-commission-report-volum-22]. The House explicitly extended responsibility to Congress and the Executive Branch for Shuttle policy whose safety consequences they largely failed to appreciate, and identified continuing SR&QA staffing, oversight, and role-definition weaknesses [house-report-99-1016-investiga-p030]. Its additional findings included insufficient understanding of Space Shuttle Main Engine (SSME) blade fractures, violations of component certification limits, and unresolved mission-critical hardware anomalies—not alternative causes of Challenger’s loss [house-report-99-1016-investiga-p019]. On ice, it separately concluded that launch should not have proceeded until the pad escape platform was cleared [house-report-99-1016-investiga-p019].
**Feynman’s reliability argument.** Feynman reported loss-probability estimates ranging roughly from engineers’ 1 in 100 to management’s 1 in 100,000, criticizing management’s reliability claims as exaggerated rather than presenting those figures as equally substantiated estimates [feynman-personal-observations-01]. He found deterioration in SSME certification standards analogous to acceptance of booster seal damage, despite generally high-quality engine engineering [feynman-personal-observations-01]. In contrast, he praised avionics software’s bottom-up testing and independent adversarial verification, while distinguishing persistent sensor and actuator problems from that strong software practice [feynman-personal-observations-02]. His central objection was that repeated survival of poorly understood anomalies did not establish safety, and that empirical models required explicit attention to uncertainty [feynman-personal-observations-01].
NASA's redesign of the solid rocket motor (SRM) field joint was conducted and supervised by personnel from several NASA centers, including Marshall, the Astronaut Office, and individuals from outside NASA [house-report-99-1016-investiga-p118, page 118]. The Rogers Commission recommended an independent SRM design oversight committee from the National Research Council to review certification requirements and oversee the design and test program [rogers-commission-report-volum-26, page 198]. It also required that joint integrity be insensitive to dimensional tolerances, assembly, environmental effects, and internal pressure, with certification tests over the full operating range, including temperature [rogers-commission-report-volum-26, page 198]. NASA scheduled its review of the Critical Items List, hazard analyses, and FMEAs for completion by March 1987, with outside auditing [house-report-99-1016-investiga-p118, page 121]. The House Committee questioned whether independent reviewers had enough propulsion expertise, since rocket propulsion has been called a "black art" with few experts [house-report-99-1016-investiga-p118, page 121]. On risk management, the Committee found that NASA did not explicitly use a centralized program coordinating all the factors of an adequate risk management program [house-report-99-1016-investiga-p146, page 146]. It found that launch constraints were often waived after a rationale for accepting the problem was developed, rather than after correcting the problem [house-report-99-1016-investiga-p146, page 152]. Rogers recommended a policy governing the imposition and removal of launch constraints, and that Flight Readiness Reviews and Mission Management Team meetings be recorded [rogers-commission-report-volum-26, page 200]. The Committee found no basis for calling the Flight Readiness Review procedure flawed, but noted that FRRs cannot prevent a flight over a design flaw that management had already judged an acceptable risk [house-report-99-1016-investiga-p146, page 156]. Rogers recommended an Office of Safety, Reliability and Quality Assurance reporting directly to the Administrator and independent of other NASA functions [rogers-commission-report-volum-26, page 199]. The Committee found that SR&QA functions had lost effectiveness through headquarters staff reductions and diffusion into field centers [house-report-99-1016-investiga-p187, page 187]. The Committee recommended that the Shuttle not fly again until NASA can assure that safety criteria have been met, and that the early 1988 date be treated as flexible [house-report-99-1016-investiga-p118, page 118]. It recommended that the new safety office have authority to enforce scheduling that leads to safe overtime rates [house-report-99-1016-investiga-p118, page 123]. Rogers recommended a flight rate consistent with NASA's resources and rigorous controls on cargo manifest changes [rogers-commission-report-volum-26, page 201]. The Committee also urged that the new SR&QA office join procurement and award-fee processes, noting Thiokol's eligibility for an incentive fee of probably about $75 million at the time of the accident [house-report-99-1016-investiga-p187, page 190]. Open questions the Committee raised include what function the new Safety Office would perform in redesigning the SRB field joint and in NASA's recertification plan [house-report-99-1016-investiga-p118, page 118]. It also noted that NASA's lack of statistical data on certain components limits the usefulness of sound engineering judgment [house-report-99-1016-investiga-p146, page 146]. The following lessons are my interpretation, not findings of the investigations:
[rogers-commission-report-volum-26, page 198].[rogers-commission-report-volum-26, page 198].[rogers-commission-report-volum-21, page 161].[rogers-commission-report-volum-21, page 161].[house-report-99-1016-investiga-p146, page 152].