I have been watching a WebmasterWorld thread named Loads of Supplemental and the titles are all corrupt. It is true, there are many examples of pages in Google's supplemental index that have inaccurate titles. It appears that Google is listing the title plus some of the first piece of content at the top of the page.
I have an example, but the site owner asked I keep the URL and any revealing information hidden. Take a look at this search result;
The part after the blank area in the clickable title is actually from the content at the top of the page that shows the day and date. This result is also obviously in the supplemental index, as you can see in the green font on the right side of the image.
This is a very interesting observation. Google isn't always doing this, it may depend on the query at hand. Is this a sign that Google may be producing dynamic titles based on queries? I know Google sometimes opts for DMOZ titles, but this is way different than that.
But by taking the top text, which is normally standard header text, in our case date information, but in many cases text such as "home," "about us," etc. this can make for a poor search experience.
Forum discussion at WebmasterWorld.