Returning to the question of how to measure such things operationally:
The easiest way is by surveying. Consider how much distance
corresponds to a degree of geodetic latitude. It's not a constant,
as should be obvious if you consider a highly eccentric ellipsoid.
A decent set of maps showing both lines of latitude and a scale
of miles suffices: compare high latitudes with low latitudes.
Even easier, use a software model. As previously mentioned, I've
had good experiences with the Geographiclib package. For the
earth it tells me:
south_pole to south_89: 111.694 km
south_44 to south_45: 111.122 km
equator to north_1: 110.574 km
Not a huge effect, but well outside the margin of error.