The resistive anomaly, an increase of sample resistance above the normal st
ate value at the top of the superconducting transition, has been studied ex
perimentally in various aluminum nanostructures. The effect is not absolute
ly reproducible being dependent on the cooling history and on the particula
r arrangement of voltage and current probes for multiterminal samples. If t
he anomaly is clearly observed, its magnitude can be suppressed by a strong
bias current and/or magnetic field. It is shown that the main factor deter
mining the form of the R(T) transition in lift-off fabricated nanostructure
s is the inevitable film inhomogeneity, being a "fingerprint" of each parti
cular sample. The origin of the anomaly is attributed to a superposition of
two processes. First, the formation of deformed (not perpendicular to the
wire axis) N/S boundaries, and, second, the existence of a nonequilibrium r
egion inside the superconducting domain close to this N/S interface, charac
terized by a finite value of the electric field and the corresponding effec
tive resistance. Only a combination of the above effects can give a reasona
ble quantitative agreement with experimental data. [S0163-1829(99)02506-0].