. They demonstrated, by means of variation in colony morphology, that this substance could transform their rough type 2 Streptococcus pneumoniae strain R36A into a smooth type 3 strain. It has become accepted as fact, from modern textbook accounts of these experiments, that smooth pneumococci make capsule, while rough strains do not. We found that rough-to-smooth morphology conversion did not occur in rough strains R36A and R6 when the ability to synthesize native type 2 capsule was restored. The continued rough morphology of these encapsulated strains was attributed to a second, since-forgotten, morphology-affecting mutation that was sustained by R36A during strain development. We used a new genome-PCR-based approach to identify spxB, the gene encoding pyruvate oxidase, as the mutated locus in R36A and R6 that, with unencapsulation, gives rise to rough colony morphology, as we know it. The variant spxB allele of R36A and R6 is associated with increased cellular pyruvate oxidase activity relative to the ancestral strain D39. Increased pyruvate oxidase activity alters colony shape by mediating cell death. R36A requires a wild-type spxB allele for the expression of smooth type 2 morphology but not for the expression of smooth type 3 morphology, the phenotype monitored by Avery et al. Thus, the mutated spxB allele did not impact their use of smooth morphology to identify the transforming principle.The bacterium Streptococcus pneumoniae owns the important distinction of having introduced the world to the concept that DNA is the hereditary material. In their landmark studies, Avery, MacLeod, and McCarty identified DNA as the substance that could convert or transform one type of S. pneumoniae into another (1). The indicator of the transformation event was the ability of an unencapsulated strain of S. pneumoniae with rough colony morphology (R) to acquire a new encapsulated smooth (S) appearance (1).The R and S nomenclature has long been used to describe the colony morphology of S. pneumoniae: R colonies appear rough and irregular on solid media, while S colonies appear smooth and shiny. In the celebrated work of Griffith, these terms were originally used to correlate colony morphology with virulence properties (4). Later, R and S came to describe the encapsulation state of the bacterium because it was found that, in contrast to S colonies, R colonies do not produce typespecific polysaccharide capsule. The R designation is actually an umbrella term that has been used to describe many morphologically distinct unencapsulated S. pneumoniae mutants. Although all strains producing R colonies have lost their ability to make capsule, not all R colonies look the same, because the strains in question may have accumulated additional mutations that further alter colony morphology.One R variant that has sustained multiple colony morphology-altering mutations is S. pneumoniae R36A, the strain used by Avery et al. to demonstrate transformation (1). The history of R36A indicates that the R phenotype is the result of at least two inde...