The U.S. Census Bureau has released some data from the 2000 census, including statistics on surname frequency. The survey includes responses from almost 270 million names-- check out the pdf file at the Census Bureau site for a very interesting discussion of the numbers of surnames in the country, the number of people with the seven most common surnames, etc.
In the table below, the first column is the name varient; the second is the popularity of the surname per 100,000 people; the third is a cumulative proportion of the surname plus all surnames occurring more frequently; and the last the rank. There was a total of over 6 million surnames total (although many "singetons" are believed to be mistakes by people who didn't follow the directions).
The way to think about the cumulative proportion is if you put those 100,000 sample people in a line based on their surname frequency, you would have to go through nearly two-thirds of the people-- 64,750 people-- before you met the first Slavin. (And yes, their "fractional people" is a little freaky, but that's the world of statistics for you.)
Note: Sleaven, Sleavin, and Slavy were not found in the sample. More information, including a file with all surnames appearing at least 100 times in the census (it's a mighty big file!) can be found at the Census Bureau's website.
Copyright © 2005 Larry Slavens. All rights reserved.