And even if this particular crisis stabilizes, the decision-making approach that Trump used makes Kaiser Wilhelm look like a model of cool statesmanship, and its application in a crisis involving a real great power could be catastrophic. Nothing about these events supplies any confidence that his advisers can effectively direct him, or that the kind of advice available to him (see Rudy Giuliani, shadow secretary of state) won’t keep on getting worse. Nothing about the relative calm of years one through three guarantees that the test isn’t lurking in year four.

Or in year five or six or seven, should Trump be re-elected. Which points us to the central issue for Republican-leaning voters and Republican senators alike. Both groups have grown used to Trump, in part because human beings grow used to all things, but in part because the most alarmist predictions, mine included, did not accurately describe his first two years in office.

But those voters and legislators have to ask themselves, at the polls in 2020 or sooner in a Senate trial, what seems more likely to predict Trump’s governance going forward: The relatively restrained pattern of the McMaster-Mattis-Kelly period, or the unchecked impulses that just gave us death and betrayal and humiliation for no reason, none at all, save that our president is unfit for his job.