You are mistaken in order to filter out these visits from your Google Analytics - they are real species, you are simply not sure of their origin.
I think the message you read about people bookmarking your site is this: http://www.google.com/support/forum/p/Google%20Analytics/thread?tid=61e3d404f3c9c2bb&hl=en
A bookmark is a red herring - this is a more likely reason:
If you look at the line returned from the full analysis of the user agent, it clearly says: Mozilla / 5.0 (iPhone, U, CPU iPhone OS 4_0_1, like Mac OS X, fr-fr) AppleWebKit / 532.9 (KHTML, like Gecko) Mobile / 8A306
This is AppleWebKit.
This means that all these users come from iPhone / iPad applications using the built-in browsing features, invoking AppleWebKit capabilities inside their code.
As stated above, most iPhone apps contain a browser that leverages the capabilities of AppleWebKit - this is the activity that you see on your website.
Here are some possible scenarios:
- People subscribe to their blog’s RSS feed using an application such as Reeder. When they view a message, it records a pageview. When they click on the iOS app, it records the rebound.
- Someone is sharing a link to your site via Twitter. Anyone who clicks this link using their own iOS app on Twitter. 1 visit, 1 page view, 1 rebound.
- Flipboard, Facebook, Google mobile app, Bing app, etc. etc.
Perhaps you can use campaign tracking codes to find out which medium (and ultimately which application) controls traffic?
UPDATE: here's an article from Land Search Land on the topic of lost referrer data from mobile apps.
Some parts of interest:
Smartphone applications are not able to transfer the "link to page" information to the analytics of your site, for the same reason that Word documents or PDF files are not written as referrers of sites: they are separate from the applications. Even for mobile applications in which browsers are embedded (for example, Facebook, Google, Twitter, Groupon), there is simply no “link page” data for click-through.
If you expect to see that this traffic is nicely classified in some kind of analytical report, you will wait some time ....
The end result is that your “direct” traffic and sales count are overstated and that a significant percentage of your actual mobile search and mobile social traffic is underestimated. This means that you are probably not investing enough in mobile devices and are providing conversion credit to other download channels.
In the Google Search app:
But for mobile organisms, you're out of luck. Google Analytics not only doesn’t register Google as a referrer, and mobile keywords that drive traffic also become incognito.
Finally:
You cannot control what you cannot measure. Until you can measure the impact of mobile applications on mobile ROI, it’s easy for you to imagine leaving money on the table.
Again, this is not your statistics, which is corrupted, this is your measurement. Filtering your mobile visitors is a bad idea.