An organism that seems to be identical to Orskov's 'Sarcina mirabilis' [Orskov, J. (1930) Acta Pathol Microbiol Scand Suppl III, 519-541] has been rediscovered in specimens from the upper respiratory tract of humans. Six strains were studied, and the results, which conformed to Orskov's description of S. mirabilis, were as follows. Rough to smooth colonies grow on many plated media and show extremely polymorphic cell morphology with round cells with diameters from 1 to > 10 microns. The smallest cells were often motile with circular movements. Strains were Gram-negative, facultatively anaerobic, oxidase and urease positive, and weakly catalase positive. Nitrate and nitrite were reduced, and glucose, fructose, sucrose and mannitol were fermented. Polysaccharide was produced on sucrose agar. Electron microscopy showed coccoid cells with a bundle of three to nine flagella, a Gram-negative cell-wall morphology, and aggregates of irregular cells held together by a common surface layer. The mean mol% (G+C) of the organisms was 65.0. 16S-ribosomal RNA sequencing revealed that the organism belongs to the beta subgroup of Proteobacteria, separate from all other described genera, but most closely related to Burkholderia. The name Lautropia mirabilis is proposed for this organism.