I think I can offer an explanation for one part of the mystery that was bothering Eric. The many examples with epenthetic “o” are almost certainly from Japanese.
Japanese is very strict about the types of consonant clusters it allows. Impermissible clusters are generally broken up with /u/, which in Japanese is a high back vowel with very little lip rounding. The problem is that when the consonant before the epenthetic vowel is /t/, /u/ will be a very bad choice, because the combination /tu/ is actually pronounced [tsu], that is, the /t/ is affricated. The alternative of using /i/ as the epenthetic vowel is also out, because /t/ is palatalized before /i/. The result is that Japanese standardly uses /o/ as the epenthetic vowel after /t/, and /d/.
Indeed the examples tested by Eric are all words borrowed into Japanese, and they receive epenthetic /o/ in exactly that fashion. To top it off, most Japanese speakers seem to have a hard time “undoing” the epenthesis when trying to communicate in English, even in writing. That is they tend to think that the “o” really is a part of these words, and they commonly spell Sumatra “Sumatora”.
[Note: After writing this, I actually went and clicked through all the links in Eric’s table. I found most of the Google results to be very close to what Eric gave, except those for “Sumatora” (where Eric reported 2,300 and some hits, but I got over 4,000. I am assuming that this is the result of Google searches being biased depending on where they originate from.]