This paper examines and refutes a conjecture to the effect that the solution set for a general (real) static errors-in-variables problem is a finite union of sets that are described by a finite number of linear inequalities. The conjecture is disproved by detailed examination of particular errors-in-variables problems with four variables. The solution set in this case is described by five surfaces, all intersecting in straight lines, but in general one of these surfaces is not flat.