Abstract. We consider a modal language for affine planes, with two sorts of formulas (for points and lines) and three modal boxes. To evaluate formulas, we regard an affine plane as a Kripke frame with two sorts (points and lines) and three modal accessibility relations, namely the point-line and line-point incidence relations and the parallelism relation between lines. We show that the modal logic of affine planes in this language is not finitely axiomatisable. §1. Introduction. Recently the modal logics of space have began to draw considerable interest from logicians and computer scientists. See, e.g., [1]. Much of the interest seems to stem from the perceived use of modal logics for qualitative reasoning about spatial relations between objects, and the potential applications in computer science and knowledge representation.In this paper, we are concerned with the modal logics of projective and affine planes. In [2], geometries of points and lines were viewed as Kripke frames, the domain of each frame being the point-line incidence relation itself (i.e., the set of pairs (s, l ) where s is a point on a line l ). A completeness theorem for 'incidence geometries' was proved, using a non-orthodox 'irreflexivity' inference rule, and extensions to projective and affine geometries were considered.In [13], Venema viewed projective planes in a somewhat more straightforward way, as Kripke frames with two sorts (points and lines), and two modal accessibility relations (incidence between points and lines and between lines and points). He formulated a corresponding modal language with two sorts of formulas -point formulas (evaluated at points) and line formulas (at lines). He then presented a finite set of axioms, essentially expressing that the two accessibility relations are the converses of each other; every point lies on at least one line; any two points lie on at least one common line; and the duals of these two properties obtained by exchanging points and lines. The inference rules were orthodox: modus ponens, (well-sorted) substitution, and universal generalisation for each of the two sorts. Venema proved that the system is (strongly) sound and complete for projective planes. He also proved that the problem of determining whether a given formula is