Treatment by polymyxin B sulfate and ethylenediaminetetraacetate separated a 40 kilodalton (kDa) protein from the nonpigmented Serratia marcescens and even from the nonpigmented bacteria of the pigmented strains, whereas the same treatment separated the 100 kDa protein associated with the pigment formation from the pigmented bacteria. Lysozyme treatment separated the 100 kDa and/or 40 kDa proteins correlated with the pigmented level. The 40 kDa protein was not an outer membrane protein but a flagellin. These results suggest that the flagella formation was inversely related with the pigment formation.