Well, let's see if we can project then, shall we? Look at Alabama and Georgia. The free black population in both states in 1850 were 2265 and 2931 for a 5196 total. Closs to Illinois' total of 5436. But 10 years later, the total in Illinois is 7628 and the total for Alabama and Georgia is 6190. I guess geometric growth curves work differently down south, huh?
So in other words you are quibbling over a 1,500 person difference in population growth? You might as well be whining that the 2000 census places two more neighborhoods in Congressional District 4 than Congressional District 5.
It's called statistical variance, BTW. Statistically there is no one perfectly fitting geometric curve. There is a fit line that approximates the data though, with individual states falling around it. Some will be a little higher producing a positive error term, some will be a little lower giving a negative error, some will land exactly on the line giving no error. But if the line fits fairly well it will show a geometric trend and predict geometric growth within a plus or minus variation. I don't expect you to comprehend mathematical techniques like that though, and even if you did you're too dishonest to admit it.