Absolute gf-values for 118 Fe I lines in the visible region of the spectrum have been measured by shock-tube emission spectroscopy. Special attention was given to choosing relatively faint lines in a region of reasonably well-defined solar photospheric continuum and to eliminating systematic errors. For the latter purpose, the gas kinetic temperature was measured accurately by an ultrasonic technique for every shock, and a special study was made in which the various experimental parameters were varied over extreme ranges and in which the validity of the assumptions of local thermodynamic equilibrium (LTE) and of optical thinness were verified. Our results are in fairly good agreement with Corliss and Tech's free-burning-arc measurements for low-excitation lines but show large disagreements for lines having high-excitation potentials. The magnitude of the discrepancies is largely dependent upon upper excitation potential and to a lesser extent upon wavelength. Our measurements for low-excitation and high-excitation lines agree to within a factor of about 2 with values obtained by other independent investigations, which include shock- tube, wall-stabilized-arc, atomic-beam, and beam-foil measurements. These results suggest that the solar photospheric abundance of iron may need to be revised upward in the light of the various recent measurements of gf-values for high-excitation lines, which are consistently smaller than reported free-burning-arc values by nearly 1 order of magnitude. These free-burning- arc f-values and the furnace-absorption f-values upon which they are based have repeatedly led to the paradox of solar photospheric iron abundance about an order of magnitude lower than coronal. According to a line-profile fitting analysis by Ross (1970), our Fe I gf-values correspond to an increased photospheric abundance by a factor of about 5.