The possibility of employing gravitational-microlensing observations f
or high angular resolution studies of stars is explored. The decrease
in the amplification factor of a gravitational lens with increasing an
gular distance from its optical axis and the finite angular diameter o
f a stellar disk result in a divergence of the microlensing light curv
e from the curve corresponding to a point source. The effects of finit
e stellar angular sizes and limb darkening on the observed microlensin
g curve are treated. It is shown that the angular radius of a star, me
asured in fractions of the Einstein-cone radius, can be determined usi
ng a nonlinear least-squares fitting procedure. The currently availabl
e methods of solving this problem, which allow point estimates and con
fidence intervals to be obtained for the parameters, are discussed. Th
e wavelength dependence of limb darkening results in a chromatic effec
t during microlensing observations of stars, generally manifested itse
lf as reddening. Color variations in the standard broad-band UBV syste
m are estimated for various microlensing conditions. The effect is rat
her small, with color variations being no larger than 0(m).015. Howeve
r, it can be enhanced by a special choice of the detector passbands th
at ensure a maximum limb-darkening difference. An important feature of
the effect is its dependence only on the ratio of the angle between t
he lens and the true direction toward the star to the stellar angular
radius. Given the angular diameter of the star, the angular radius of
the Einstein cone of a gravitational lens can be estimated by comparin
g the observed and calculated color indices and flux amplification fac
tors.