Don't forget that most of these people were illiterate and may have had no idea about actual birth dates. It's absolutely not unusual to find ages that are different in each census for an individual. I find that the best thing to do is to usually look for a person within a 10-year radius of the one I am researching, and even then they kind of fool you. For instance, one person showed his birth in 1860 as 35, and in 1870 and 56, then in 1880 as 45, yet I absolutely know it was the same person because of whom they were living with. By 1900, I guess he gave up and said he was 90 at that time, so you see what I mean. The best thing to do is usually to take the first age you find him at since you figure that would be closest to his birth and they must have had some idea how hold he was then. So I would say, an age with only a couple of years discrepancy is gold, and 5 is a steal, go with the first age.

Deloris