Effect of systematic gf-value errors on stellar curves of growth

From an analysis of 101 Fe I lines in the spectrum of the star 0 UMa, a value has been derived for the magnitude of the systematic increase, noted by several laboratory studies, in the Corliss and Tech (1968) gf-values with increasing energy of the upper level. This result is in reasonable agreement with the findings of the laboratory measurements and shows that the best estimate of the linear increase is 0.2 dex eV-1 for lines with upper excitation potentials above 2.5 eV. The effect of this systematic error on the curve of growth is to spread the observational data along the direction of the abscissa. Such a dispersing of the data could be misinterpreted possibly as due to a large microturbulence velocity and thus lead to an abundance which is correspondingly too small. Key words: Fe I gf-values - stellar curve of growth - microturbulence