Data on physical properties of steels have been collected from the open lit
erature and put into a database. The influence of composition on the densit
y of steels has been analyzed. An overview over former studies is given. Th
e steels have been investigated by regression analysis in two groups, i) fe
rritic and low alloy steels, and ii) austenitic steels. For ferritic steels
two models are provided. The first model is based on the assumption that a
ll C is bound in cementite and other solutes are insoluble in cementite, Th
e second model employs the result of a thermodynamic analysis where the amo
unt of cementite and the solubilities in ferrite and cementite were determi
ned with computational thermodynamics. The non-linear effect of Cr and Mn i
n cementite was computed and regression analysis of the effect of solutes o
n the density of ferrite was performed. For Ti-stabilized austenitic steels
, the amount of TiC and the solubilities were assessed in a thermodynamic a
nalysis. The effect of solutes on the density of austenite was studied by r
egression analysis. For estimations of the density of steels containing com
ponents that are not covered by the regression analysis, the regression coe
fficients can be supplemented with literature data or theoretically determi
ned values. The results obtained by the present regression analysis are: Cu
and Mo increase the density of ferritic steels, and C, Cr, Mn, S, Si, and
V decrease it. TiC. C, Cr, Mn, N, Si, and Ti reduce the density of austenit
ic steels and Cu, Co. Mo, and Ni increase it.