compare the relative size observed from the front and back of the tower is what he did and then factor in the distance between the two observation points to come up with the data. I mean, the error will obviously be substantial but there is other evidence to further support his following assumptions