In previously proposed models of A beta soluble oligomers, the N-terminal domain A beta(1-16), which is missing in p3 peptides, protects the hydrophobic core of the oligomers from the solvent. Without this N-terminal part, oligomers of p3 peptides would likely expose hydrophobic residues to water and would consequently be less stable. We thus suggest, based on theoretical and experimental results, that p3 peptides would have a low propensity to assemble into stable oligomers, evolving then directly to fibrillar aggregates. These properties may explain why p3 would be devoid of any impact on synaptic function and moreover, strengthen the hypothesis that A beta oligomers are the principal synaptotoxic forms of A beta peptides in Alzheimer disease.