The problem of the lateral instability of deep beams under the action of bending loads applied in tiieir planes of greatest flexural rigidity has been a subject of investigation by engineers since the introduction of the first rolled sections, over a century ago. The elementary theories for the determination of the bending stresses sufficient to cause failure of a beam by buckling have been evolved by a number of mathematicians, but there has been little experimental verification of these results by tests under prop@rly controlled conditions. Although the buckling of beams has been compared with, the failure of columns under end loads, the possibility of such beam failure in practice has been of far less importance than that of stanchion buckling--due,, primarily, to the fact that in most cases considerable restraints are applied to a beam under working conditions which invariably increase the loads necessary to cause instability. Also the materials in general use- have low specific strengths with limiting design stresses usually lower than the critical buckling stresses for beams of normal proportions. With the introduction as structural materials of high strength steels, and also of aluminium alloys having low elastic modulii, the problem becomes more crucial from the design aspect, and recently it has been found desirable to include design data, based on the fundamental theory, in new codes of Structural Engineering practice. The recent draft Code of Practice for Structural use of Steel in Buildings l contains clauses relating to the lateral stability of steel beams in which design formulae, deduced from the theoretical results for beams under idealised conditions, are suggested as alternatives to the existing empirical formulae in terms of the slenderness ratio L. This improved design data will give estimates of the permissible stresses in a beam which will generally be in error on the safe side and in many cases of beams used as floor joists or frame members it can be shown that failure due to lateral buckling cannot occur at flange stresses lower than the yield stress of the material, snd even then only as a secondary form of failure. The qualitative analysis of the influence of the various factors influencing the stability of beams under practical conditions may prove of great assistance to designers who at present sacrifice economy in their materials due to reduction of the allowable stresses as a precaution against buckling, -which may prove unnecessary when the actual conditions of loading and support are considered.