Microbubbles are actively being investigated as contrast agents for ultrasound imaging and as ultrasound-mediated drug delivery vehicles. Gas-filled microcapsules stabilized by multilayers of oppositely charged polyelectrolytes [poly(allylamine hydrochloride) and poly(styrene sulfonate)] have recently been reported. These rigid microbubbles are promising owing to their high stability, tunable surface charge and easy functionalization of the polyelectrolyte surface. Bubbles stabilized by polystyrene beads have also been studied. However, rigid shells, which are made either from polymers or crystallized lipids, increase the resonance frequency of the microbubbles dramatically. As maximum echogenicity is obtained for an ultrasound frequency close to the bubble resonance frequency (which varies as the inverse of bubble size), high-intensity ultrasound is required to obtain successful echo imaging of hard-shell bubbles. High-intensity ultrasound pulses can compromise shell integrity and may limit the use of hard-shell microbubbles with small diameters (1–5 mm) for drug delivery. Because they are supple, monolayers of self-assembled phospholipids in the fluid state allow production of small bubbles that resonate at much lower frequency. Their stability, however, remains a critical issue. Under the action of Laplace pressure, which increases as the inverse of their radii, bubbles dissolve in aqueous media all the more rapidly when they are smaller. Naked gas bubbles smaller than 10 mm persist only for a few seconds in an aqueous environment. Enclosing the bubbles within a phospholipid shell allows an increase in bubble persistence by a factor of 10 or more, provided the aqueous phase is saturated with the gas. Micron-size bubbles with half-lives sufficiently long for practical diagnostic ultrasound examination have been obtained using phospholipids and a perfluorocarbon filling gas. 12] Perfluorocarbons retard bubble dissolution very effectively, due to very low solubility in water. The half-life of the commercially available soft-shell bubbles is limited, however, to a few minutes. We found that, against expectations, and contrary to all published experimental evidence, small (~1.55 0.20 mm in radius) bubbles covered with a shell of fluid phospholipid and dispersed in an aqueous phase can last longer, by an order of magnitude, than larger bubbles (~5.40 0.50 mm) made of the same components. This was made possible because we were able to prepare monodisperse populations of bubbles and determine the size and stability characteristics of the bubbles accurately using a new acoustic method (Supporting Information). Initial bubble size distributions and stability over time were determined by measuring the attenuation coefficient of ultrasound waves, that is, the reduction in amplitude of a wave package that propagates through the bubble dispersion. The technique is based on the fact that, as a bubble is a resonator, the product of the bubble’s radius, r, and its resonance frequency, f0, is about 3. [9] The attenuation spectra of ultrasound by the microbubbles were monitored using a multi-frequency analysis set-up (Supporting information). The microbubbles investigated contained N2 saturated with perfluorohexane (C6F14, PFH) as filling gas. [12] The bubble wall consisted of 1,2-dimyristoyl-sn-glycero-3-phosphatidylcholine (DMPC). Two DMPC concentrations, 10 or 50 mm, were investigated. Bubbles were obtained by sonicating DMPC (10 or 50 mm) in an isotonic solution (Isoton II), the volume above the dispersion being filled with PFH-saturated N2. The bubble dispersion was diluted with 14 mL of Isoton and the bubbles were allowed to float. 300 mL (for the 10 mm-concentrated dispersion) or 1 mL (for the 50 mm-concentrated dispersion) were pipeted after 3 min and 20 min, respectively, and transferred to the ultrasonic cell. The ratio of the partial pressure of oxygen in the aqueous phase to that at saturation was approximately 1. We checked that the ultrasound field did not modify the size characteristics or stability of the bubbles (Experimental Section). The low DMPC concentrations (10 mm), after 3 min of flotation of the dispersions at 25 8C, lead consistently to a monomodal population of bubbles with an average radius of 5.40 0.50 mm (Figure 1). The bubble radius and radius distribution are confirmed both by static light scattering (5.90 0.50 mm) and by optical microscopy (6.3 1.2 mm) [Figure 1].