Jeffrey LewisAn "Isolated Incident"

I’ve acquired a copy of the one-page Bullet Background Paper On Minot-Barksdale Weapons Transfer Incident distributed by the Air Force.

The paper lists “five procedural errors that led to the incident:”

  • First, Airmen assigned to the Minot weapons storage area failed to examine each one of five pylons (a pylon is a self-contained package of six cruise missiles) located in the storage area.
  • Second, the driver of the trailer that moves the pylons to the aircraft began loading the pylons before the required pylon inspection was complete.
  • Third, the driver failed to inspect the missiles before hooking up the trailer for transport.
  • Fourth, when the driver radioed the munitions control center to verify the status of the pylons, the control center failed to access a database (as required) that would have alerted them that one of the pylons was not properly prepared for transfer.
  • Finally, the B-52 navigator neglected to check all missiles loaded for transport, as required.

Wow, that’s pretty amazing. My only complaint, really, is that in the paper and the press briefing, the Air Force describes the mistake as an “isolated incident.”

Such incidents may be exceedingly rare, but seldom isolated — all organizations experience accidents.

Anyway, gotta go marry off Czerwinski …


  1. Geoffrey Forden (History)

    I seem to have had the misperception that nuclear weapons would be kept separate, would have much more security than conventional munitions, and would have been very, very hard to confuse with conventional weapons. I remember being told that tactical nukes were kept in special bunkers, inside special locked boxes with barbed wire stretched over them so that they couldnt be stolen (at the time, that sounded like insufficient protection…little did I know). Why isnt that the case? Is it possible that Russian nukes, after all the protection we have sold to them, are better protected than ours? Whats going on here?

  2. Tim H (History)

    I highly recommend reading the insights in this article, soon to be given further exposition in The American Conservative magazine.

  3. Joseph Logan (History)

    Any of these might pass the smell test as an “isolated incident” except #4. “Failed to access a database” could reasonably be expected to set off alarms in the chain of events mentioned above were it an isolated incident. That it didn’t should arouse some fairly straightforward questions about how we can be certain the database was accessed on most other occasions. Unlike a visual inspection, checking a database leaves a record, one that can be audited and reported. Were I on the Armed Services Committee, I would immediately say “prove it”.

  4. bob massman (History)

    Those are Administrative controls. Where are the Engineered Controls? It’s still amazing this mishap could occurr and indicates they need a serious review and upgrade to their safeguards systems.

  5. Muskrat

    Are arms controllers empowered to perform marriages like the captains of ships? I’ve often heard treaties compared to marriages… unhappy ones, usually.

  6. Andy (History)


    I have not read anything that suggests that these nuclear weapons were misidentified as conventional weapons. After all, the AGM-129 has no conventional version – so the missiles themselves have three possible configurations: A live nuclear warhead, a training warheads, or warhead.

    What apparently happened has nothing to do with confusion between conventional munitions and unconventional, but the failure to download the nukes in the first place followed by the failure of anyone down the line checking to ensure they had, in fact, been downloaded as procedure required.

  7. Lao Tao Ren (History)

    How do you get at the truth of these incidents when the institutions clam up and Congress have to rely on whistleblowers and leakers?

    Maybe this is a useful lesson for the arms control community…. never unilaterally disarm before leaking!

    Oct. 19, 2007 – 8:11 p.m.
    The Nuclear Bombshell That Never Went Off
    By Jeff Stein, CQ National Security Editor

    Say you’re a member of Congress, and a Pentagon expert tells you that top officials are secretly letting Taiwan go nuclear, to contain China’s emerging threat.

    Do you: (1) start an investigation, with an eye toward hearings to grill officials on the facts, or (2) drop it and stand aside as officials run your whistleblower out of town?

    In the real-life case of Pakistan and nuclear weapons, the answer from Congress has been (2).

  8. Jelco Cathlon (History)

    So could it be that some neo-con plot was uncovered by mistake and that those cruise missiles where getting readyed to be delivered somewhere in the middle east to be launched on Iranian targets; didn,t a senior administrator commit or was suicided in the aftermath?

  9. Ron

    The elephant in the room is the computer handshake that occurs when the pylons hit the hardmount.

    I have been told that the aircraft commander would have known precisely what he was carrying.

    There is much more to this story than has been reported.

  10. Tim (History)

    I am putting together a stats midterm, so I have to look at this in terms of a basic probability problem.

    Q: If we assume that the probability of missing each of these checks is equal and independent, what must be the probability of missing a check to have this happen one time in (a) 10,000? In (b) 100,000?

    (a) (1/10000)^1/5 = 0.1585 = 16%
    (b) (1/100000)^1/5 = 0.1 = 10%

    Now the probabilities are unlikely to be equal and might not be independent, but any way you slice it, missing a check more than 10% of the time is astoundingly lax discipline.

  11. Tim (History)

    Actually, it is much worse than that now that I think about it. All the checks failing only results in transport if there are warheads (improperly) there in the first place.

    If that has happened 100 times, then it would take a failure rate of 40% for each check to expect one failure. Warheads have probably not been mistakenly queued up for transfer anything like 100 times.

    Bottom line is that nobody was checking.