E. coli strain RH 2859 carrying a mutation in rplX gene (Cabezon, 1977) does not assemble 50S subunits when grown at non-permissive temperature, while the assembly of functional 30S subunits proceeds normally. Abortive 30-32S particles, containing 23S RNA or its precursor form and a certain number of 50S proteins, accumulate. Protein analysis of those particles indicates that several 50S proteins are either missing or present in reduced amount (L7/12, L20, L24, L26, L27, L28, L32 and L33). Also newly synthesized 50 S proteins appear to exchange with preformed 50S subunits. The rate of synthesis of all r-proteins has been determined in the mutant grown at non-permissive temperature. The rate of synthesis of seven 30S proteins decreases between 1.5 to 2.5 times whereas that of most 50S proteins is 2 to 5 times lower than in wild type cells. As shown by studies on r-protein stability at 43°, for some proteins part of the decrease is due to degradation of newly synthesized proteins. However all 30S proteins are stable in the period of time tested, in contrast possibly because they are overproduced with respect to present need several 50S proteins are unstable (L2, L5, L6, L11, L13, L14, L15, L16, L19, L22, L24, L27, L28, L30, L32). Finally an estimate of the pool size of free r-proteins in the sap of mutant cells cultured at the non-permissive temperature was determined: Four 30S proteins and one 50S protein (S20, S11, L31, S14 and S2) are present in the pool at a level much higher than others. Our results suggest that the amount of r-protein in mutant cells growing at non-permissive temperature is controlled on the one hand by degradation of the overproduced protein and/or on the other hand at the level of synthesis, possibly by a feed back mechanism exerted by free proteins present in the soluble pool. © 1979 Springer-Verlag.