We have re-studied line-strength gradients of 80 elliptical galaxies .
Typical metallicity gradients of elliptical galaxies are \Delta { [ Fe / H ] } / \Delta \log r \simeq - 0.3 which is flatter than the gradients predicted by monolithic collapse simulations .
The metallicity gradients do not correlate with any physical properties of galaxies , including central and mean metallicities , central velocity dispersions \sigma _ { 0 } , absolute B-magnitudes M _ { B } , absolute effective radii R _ { e } , and dynamical masses of galaxies .
By using the metallicity gradients , we have calculated mean stellar metallicities for individual ellipticals .
Typical mean stellar metallicities are \langle { [ Fe / H ] } \rangle \simeq - 0.3 , and range from \langle { [ Fe / H ] } \rangle \simeq - 0.8 to +0.3 , which is contrary to what Gonzalez & Gorgas ( 1996 ) claimed ; the mean metallicities of ellipticals are not universal .
The mean metallicities correlate well with \sigma _ { 0 } and dynamical masses , though relations for M _ { B } and R _ { e } include significant scatters .
We find fundamental planes defined by surface brightnesses SB _ { e } , \langle { [ Fe / H ] } \rangle , and R _ { e } ( or M _ { B } ) , and the scatters of which are much smaller than those of the \langle { [ Fe / H ] } \rangle - R _ { e } ( or M _ { B } ) relations .
The \langle [ { Fe / H ] } \rangle - \log \sigma _ { 0 } relation is nearly in parallel to the [ Fe/H ] _ { 0 } - \log \sigma _ { 0 } relation but systematically lower by 0.3 dex ; thus the mean metallicities are about a half of the central values .
The metallicity-mass relation , or equivalently , the color-magnitude relation of ellipticals holds not only for the central part but also for the whole part of galaxies .
Assuming that Mg _ { 2 } and Fe _ { 1 } give [ Mg/H ] and [ Fe/H ] , respectively , we find \langle { [ Mg / Fe ] } \rangle \simeq + 0.2 in most of elliptical galaxies . \langle { [ Mg / Fe ] } \rangle shows no correlation with galaxy mass tracers such as \sigma _ { 0 } , in contrast to what was claimed for the central [ Mg/Fe ] .
This can be most naturally explained if the star formation had stopped in elliptical galaxies before the bulk of Type Ia supernovae began to explode .
Elliptical galaxies can have significantly different metallicity gradients and \langle [ Fe / H ] \rangle even if they have the same galaxy mass .
This may result from galaxy mergers , but no evidence is found from presently available data to support the same origin for metallicity gradients , the scatters around metallicity-mass relation , and dynamical disturbances .
This may suggest that the scatters have their origin at the formation epoch of galaxies .