Large sites or very detailed server administrators, sometimes go through the process of blocking rogue spiders and bots. To make sure you do not block legit spiders, like GoogleBot, they may use a verification process to ensure GoogleBot is really GoogleBot.
To do that they use reverse DNS verification to properly ensure GoogleBot is really GoogleBot.
A WebmasterWorld thread is reporting that now a specific IP range of GoogleBot is indeed GoogleBot but the DNS verification process is failing.
webcentric from WebmasterWorld said when he does the verification the forward DNS throws a socket exception: "no such host is known." This issue seems to only occur on the 66.249.70.0 /24 IP range.
Others have confirmed the issue, Google has not yet responded to the issue.
So if you are using these techniques, you may be blocking GoogleBot by accident because GoogleBot is not letting you properly verify.
Forum discussion at WebmasterWorld.
Update: Google's John Mueller responded to me on Google+ about this saying:
Oops, we'll get that set up before we continue using those IP ranges!
