Introduction

Abstract

The concept of least squares estimation is inextricably linked with the name of Karl Friedrick Gauss. Although Legendre was responsible for the first published account of the theory in 1805 and, indeed, first coined the term ‘least squares’, it was Gauss who developed the method into a statistical tool, embedding it within a statistical framework involving a probabilistic treatment of observational errors. Gauss’s first published exposition on least squares appeared in his famous Theoria Motus Corpo- rum Coelestum which appeared in 1809 when he was 31 years of age. But, as Sprott (1978) has pointed out in his excellent review of Gauss’s contributions to statistics, the basic ideas were most probably formulated while he was still in his twenties.