When I search for the page using the site:domain, it's not there at all.
I'm assuming you're also including unique quoted text strings in the search.
The confusing part here is that you see the pages indexed before they are dropped, which makes me think it's some kind of filter. If it were a dupe content filter, the dropped page, unless penalized, in my experience would generally come back in a week or so... but it would always also show up on a search for an exact text string in quotes from the page.
There was a period, just as Panda was beginning, when Google was
not returning searches for quoted titles. As my theory goes, Google felt that the quoted string... in this case, the page title... was not predictive of a useful result, and thus the phrase-based index dropped it from that "layer" of the algo. Note that "layer of the algo" here is my way of describing my hypothetical construct of what the algo is doing. I would make sure that the text strings that you include in your search do
not include the page title, or sections of the page title. I would try some searches with a variety of short five or six word unique phrases.
Beyond that, I'd follow up the Blog subdirectory clue for what it's worth... perhaps try adding a somewhat similar test page to a different subdomain on your site and see what happens.
I'd want a much larger sample size, though... more pages over time... before calling this a symptom of a problem, but it's wise to keep on top of this. I would keep careful records of the dates, etc, in case others start pointing to similar problems happening around the same time.
If you've got a date reference now... say when each page was indexed and when it disappeared... that might be helpful to add to this thread.
