At the time Columbus sailed, the earth being round was common knowledge (it's clear to anyone who's seen a ship disappear over the horizon, hull before masts). The difference between Columbus and his contemporaries is that he was going on maps/atlases that had *half* the accepted circumference. Had there not been a continent about as far away as he thought 'India' would be, they would have perished.

BTW, you show Eratosthenes' calculated circumference about 33% high. I think it was a lot closer than that.

Eratosthenes figured the angular difference between two different locations (Alexandria and Syene), then multiplied by the distance between the two to get the circumference.

Using an air-distance between Alexandria and Syene (now Aswan) of 525.23 mi with this same method, I get 26,261.5 miles for circumference - as compared to the actual equatorial circumference of 24,901.4 mi, or about .05% off. A bit more if you use an average of equatorial and circumpolar.

To be fair, nobody's really sure exactly what his measurement was, as it was in 'stadia'; there were several lengths associated with that. On top of that, there's no way Eratosthenes was working with an accurate distance between the two points.