That Ryan Doumit would return to the active roster as soon as cleared to come off the concussion list was never in doubt.
That it was Chris Colabello who lost his spot was a bit of a surprise.
Colabello's overall numbers in his 96 plate appearances with the Twins aren't that impressive: a slash line of .221/.302/.372. But he'd gotten rolling of late; since Aug. 6, in a span of eight games and 25 official at-bats, he had hit .360/.500.640. And, probably not coincidentally, he had drawn seven walks while fanning just six times, a drastic improvement (in his whole 96 PAs, he has nine walks and 27 strikeouts, an ugly ratio).
So why did he get bounced back to Triple A? Let's examine the options.
Doug Bernier is the only infielder on the bench. He wasn't seriously at risk.
Chris Herrmann -- like Doumit a catcher-outfielder -- seemed a likely candidate. But the Twins obviously like him as a defensive catcher, and particularly as the battery mate to Andrew Albers, who has yet to allow a run or to throw a pitch to anybody other than Herrmann.
Wilkin Ramirez -- like Colabello a right-handed hitter -- is out of options. Sending him down would require exposing him to waivers and clearing him off the 40-man roster, and with only a couple of weeks to go before the active roster is expanded, that's something the front office didn't want to do.
I think the Twins overrate Ramirez as a defensive outfielder, but there's no real doubt that he's a better defensive outfielder than Colabello or Doumit (but probably not better than Herrmann.) Colabello is almost certainly the better hitter.
Colabello will be back in September when the roster limit is raised. He would probably have stuck now except for the roster flexibility factor.
Still, this decision, minor as it is, defies my rule of the priorities: 2013 matters less than 2014 and beyond. I can envision Colabello playing a significant role (as a first baseman or designated hitter) in future years, and the Twins would do well to find out more about him. Taking two weeks out of that opportunity may not materially affect the evaluation, but it doesn't help.