PLANESWe conclude a series of papers on 4-dimensional stable planes admitting a non-solvable Lie group A of automorphisms. Here we consider the case where A contains a subgroup qb ~ Spin 3. For dim A >~ 6, the possible planes are completely determined (Theorem 5). Apart from classical examples, we get a single new plane, which is analogous to a 2-dimensional plane constructed by Strambach [21]. For dim A = 5, we can determine the action of the group and roughly describe the possible planes. However, there remains a seemingly difficult problem of deciding which planes of this possibly very large class actually exist, or at least to give some examples.Stable planes are a generalization of topological projective planes designed to include geometries similar to hyperbolic planes. There are also examples which are not obtained as subplanes of projective planes, such as Strambach's plane mentioned above and its 4-dimensional analogue; cf. also [5, § 5]. A stable plane ~ = (M,50) consists of two things. One, a point space M, throughout assumed to be locally compact and Hausdorff and of finite covering dimension d > 0. Two, a system 50 of subsets of M, called lines. Different points x,y~M are joined (belong to) a unique line L = x v y. The operation v and its dual/x (intersection of different lines) are continuous with respect to a suitable topology on 5 °, and the domain of definition of ^ is open (stability of intersection). If ^ exists everywhere, then we get a topological projective plane. The following facts can be found in [3]: The smallest possible values old are 2 and 4 (cf. also [10]), and then the pencil 50x of all lines through x~M is homeomorphic to a sphere of dimension d/2. Each line is closed in M and homeomorphic to an open subset of 50~. The automorphism group F of g is a locally compact topological transformation group of M and of 50. It is a Lie group if d = 2 or if d = 4 and dim F ~> 5[4].This group is the main tool for the investigation of planes; in particular, it serves to single out classes of planes which can be completely understood. In most cases, one starts from a condition of the kind 'dim F >~f' and hopes for a complete classification if f is somewhere in the vicinity of 2d, which is about half the maximal value for projective planes. For stable planes with lines homeomorphic to ~2 this works with f= 10 [3], and in fact, f=9 (unpublished).The present paper concludes a series whose results corflbine to show that Geometriae Dedicata 21 (1986), 1-12.