Historical Note on the Origin of the Normal Curve of Errors