The mechanisms of magnetization reversal in small magnetic particles have been much discussed in the last decades and prompted intense research activities, motivated in particular by applications in magnetic recording technology [1]. However, experiments were performed, in general, on large assemblies of particles, and the dispersion of morphologies, compositions, orientations, and separations of the magnetic entities limited the interpretation of the results. Furthermore, interactions between particles were difficult to take into account. Single particle studies were possible only in few cases [2]. Recently, insights into the magnetic properties of individual and isolated particles were obtained with the help of near field magnetic force microscopy [3], electron Lorentz microscopy or holography [4], and micro-SQUID (superconducting quantum interference device) magnetometry [5]. It is now possible to make a clear link between experiments performed on nanometer-sized single objects (particles, wires, etc.) and the numerical calculations based on the Brown micromagnetic equations [6]. We report the first study of isolated nanoscale wires with diameters smaller than 100 nm, for which singledomain states could be expected. The cylindrical geometry, with its large shape anisotropy, is well suited for comparison with theory. We obtained unique insight into the process of magnetization reversal by measuring histograms of the switching field as a function of the orientation of the wires in the applied field, their diameter, and the temperature. Furthermore, we measured the probability of switching as a function of the applied field and the temperature. Our studies reveal that the magnetization reversal proceeds by a distortion of the magnetization followed by a nucleation and a propagation process. The observed behavior illustrates the fundamental importance of the study of single, isolated magnetic particles in comparing models and experiments. We developed planar microbridge dc SQUID [7], made of Nb (thickness 20 nm), which were shown to be able to detect 10 4 mB [8]. The SQUID is made of a thin (20 nm) Nb layer in order to prevent flux trapping. The experimental setup allows measurements of hysteresis loops in magnetic fields of up to 0.5 T and temperatures below 6 K, with a time resolution of 100 ms. Ni wires were produced by electrochemically filling the pores of commercially available nanoporous tracketched polycarbonate membranes of thicknesses of 6 to 10 mm [9]. The pore size was chosen in the range of 30 to 100 nm [10]. In order to place one wire on the SQUID detector, we dissolved the membrane in chloroform and put a drop on a chip of some hundreds of SQUID’s. Magnetization measurements were performed on SQUID’s with a single isolated wire. Scanning electron microscopy (SEM) (Fig. 1) was used to determine the position and morphology of the wire. The surface roughness was around 5 nm, corresponding to our SEM resolution.