The effects of different poly(ethylene glycol) (PEG) attachment strategies upon the adhesion of a Gram-negative bacteria (Pseudomonas sp.) was tested. PEG was covalently immobilized, at the lower critical solution temperature of PEG, to a layer of branched poly(ethylenimine) (PEI). PEI was both physically adsorbed to a stainless-steel (SS) substrate and covalently immobilized to a carboxylated poly(ethylene terephthalate) (PET-COOH) surface. On both substrates, the PEI and PEG grafting conditions were optimized so that the levels of surface coverage after each step were maximized and were the same on both substrates, as judged by X-ray photoelectron spectroscopy and time-of-flight secondary ion mass spectrometry (ToF-SIMS). Also, ToF-SIMS imaging showed that both substrates were chemically uniform after each surface modification step. Thus, the two surfaces differ only in the mode of attachment of PEI to the substrate. In bacterial adhesion experiments, the optimal SS-PEG surface was not capable of reducing the number of adherent Pseudomonas sp. when compared to the controls. However, the PET-PEG surface reduced the level of adhesion by between 2 and 4 orders of magnitude for up to 5 h. ToF-SIMS analysis showed that both PEG surfaces adsorbed low but comparable levels of proteinaceous growth medium components (tryptic soy broth), as indicated by the addition of unique amino acid fragment ions in the spectra, most likely small peptides. Thus, bacterial adhesion was strongly dependent on the PEG immobilization strategy and not on the extent of peptide/protein adsorption. However, for the best PEG surfaces the residual bacterial adhesion is most likely from recognition of the small amount of adsorbed peptides. This highlights the necessity for preventing the adsorption of small biological species that can even penetrate PEG layers of high graft density, in the quest for the ultimate "nonfouling" surface.