Ah. Yep; good catch. It is possible--with commercial, direct-to-consumer DNA testing marking its 18th year--that people born early in the 20th century may have had a cheek swabbed (no saliva tests until significantly later than 2000) and a sample stored at FTDNA until affordable autosomal DNA came along. (The first relatively inexpensive autosomal DNA test--and it was for medical purposes, not genealogy--is only 10 years old, offered by 23andMe. The next major player to enter the fray, and expressly for genealogy, was FTDNA in 2010, which had formed in 2000 and had already been testing yDNA and mtDNA. AncestryDNA, though not yet with labs that they owned, began selling tests in 2012.)
So if you factor in an elderly relative who tested, and the possibility of unusually long generations in the line (my grandfather, for example, was born in 1866...152 years ago!), we can certainly see those DNA connection suggestions propagate to much earlier dates than we might otherwise expect. That's why it's so much more accurate to view autosomal DNA in terms of the recombination and sharing at each birth event rather than calendar years.
I actually wish WikiTree could cap the direct line at the 4g-grandparents but still go to eight birth events laterally to 3rd cousins. There are far too many people who think those suggestions are some sort of evidence of DNA matching, and going back to 6g-grandparents is misleading...being able to arrive at solid autosomal DNA evidence back beyond 4g-grandparents is not only extremely difficult, but highly unlikely. There are several reasons why, but any proposed autosomal evidence linking more than six generations back from the test-taker is very likely in error and should be considered suspect.