In Bact. prodigiosum the induction of mutations to different types of colony colour (from red „r“ to white „w“ or rose „h“), dwarf colonies, and sulfathiazol resistance by UV of 248, 254, 265, 280 and 303 mμ was investigated. The cells were irradiated in thin layers on small agarblocks placed in the spectral line of a quartz-monochromator, or in stirred suspensions under a germicidal lamp (prevailing wave length 254 mμ). The induced colour mutations appear as colonies with a number of mutated sectors or mosaic spots, not as wholy changed colonies. Postcultures of sector-(S)-colonies show an average of 30% mutants (w, h) and about 4,5% S-colonies besides normals (r). This percentage of mutants varies with the dimension of the sectors in the mother-S-colony. Second postcultures of these different types correspond to the type of their mother colony except for a few spontaneous mutants normally arising in the strain. The statistics of sector numbers in all colonies (spontaneonsly and irradiated) represents the superposition of two Poisson-distributions with means m = 0,1 and m = 5 sectors/colony considering sector dimensions up to about 1/64 of the colony surface. Thus two genotypes are present, the normal („r“, m = 0,1) and the polysector type („S“, m = 5), of which the latter is increased by irradiation. Considering only sectors until about 1/16 of colonysurface the mean sector number m of the S-type is about 2. A mathematical analysis examines 3 possible hypotheses of the sector production in S-colonies: 1. in the mother cell a number (z) of free duplicating particles (e.g. plasmagenes) are present which segregate randomly in cellfissions and are partially inactivated by UV-„mutation”. Cells with particles are normal (r) without mutated (w, h). 2. A chromosom is split into a number of „monids“ of which only a part is mutated and which are separated in the cellfissions. 3. An unstable state of genetic material (mutable allele?) is induced by UV-irradiation which turns over to mutated and normal genotype during colony growth. This 3rd „labilisation“-hypothesis covers the facts well, but the 1st and 2nd (partial mutation) are impossible. The action spectra of all 3 mutation types and the cell-„killing” have a maximum at 265 mμ. The shape of the action spectrum fits the nucleic acid absorption. The „nucleoid“ of the bacterial cell is most probably the spot uptaking the UV-energy and transferring it to the place of mutation (gene). The dose (254 mμ)-curves of the 3 mutations (S, Z, Sth) and cell killing (T) show single hit type. The possibility that only selection is the reason of ascent of mutationrate is excluded by 2 facts: 1. The increase of the absolute number of mutations per irradiated cell and 2. the parallelism rather than divergence of dose-mutation curves with different spontaneous rates. The S-mutation rate rise until 15 to 20% falling slightly at higher doses but other mutations and killing increase further. This difference between the mutation groups is interpreted as production of a UV-resistance in the irradiated cells independently of the mutations by a multihit process which inhibits at higher doses only the S-mutation but not the rarer ones or the killing. Regarding the facts known until now the mutation is a process in which the primary energy absorption (microphysical „hit“) is followed by a reaction chain. From this chain 4 steps are to be seen now: 1. an oxydation or peroxydation, 2. a photo-reversible reaction, 3. a photostable and 4. a genetically unstable state (mutable allele?) causing the sector production.