>> then there is the issue of spherical geometry <<
And don’t forget about compass errors. I’ve seen colonial-era land surveys that were off as much as thirty degrees from True North — that is, much more than the “normal” magnetic deviation.
(Maybe the errors were sometimes due not only to pure carelessness, but also due to nearby iron-ore deposits?)
Absolutely.
There are so many variables to consider:
Magnetic variation that changes with location.
Individual compass deviation.
Local magnetic anomalies.
User error.
There is a long, narrow triangular piece of land near where I grew up that falls between two surveyed properties. The adjacent parcels are supposed to share a common boundary but don't.
The triangular piece of property between them doesn't exist on the records and no one owns it but it is there.
I read somewhere that this is not uncommon on land that was surveyed many years ago before modern tools and techniques were available.