We present chemical abundances of 57 metal-poor ( [ Fe/H ] < -1 ) stars that are likely constituents of the outer stellar halo in the Milky Way .
Almost all of the sample stars have an orbit reaching a maximum vertical distance ( Z _ { max } ) of > 5 kpc above and below the Galactic plane .
High-resolution ( R \sim 50000 - 55000 ) , high signal-to-noise ( S/N > 100 ) spectra for the sample stars obtained with Subaru/HDS are used to derive chemical abundances of Na , Mg , Ca , Ti , Cr , Mn , Fe , Ni , Zn , Y and Ba with an LTE abundance analysis code .
The resulting abundance data are combined with those presented in literature that mostly targeted at smaller Z _ { max } stars , and both data are used to investigate any systematic trends in detailed abundance patterns depending on their kinematics .
It is shown that , in the metallicity range of -2 < [ Fe/H ] < -1 , the [ Mg/Fe ] ratios for the stars with Z _ { max } > 5 kpc are systematically lower ( \sim 0.1 dex ) than those with smaller Z _ { max } .
For this metallicity range , a modest degree of depression in the [ Si/Fe ] and the [ Ca/Fe ] ratios are also observed .
This result of the lower [ \alpha /Fe ] for the assumed outer halo stars is consistent with previous studies that found a signature of lower [ \alpha /Fe ] ratios for stars with extreme kinematics .
A distribution of the [ Mg/Fe ] ratios for the outer halo stars partly overlaps with that for stars belonging to the Milky Way dwarf satellites in the metallicity interval of -2 < [ Fe/H ] < -1 and spans a range intermediate between the distributions for the inner halo stars and the stars belonging to the satellites .
Our results confirm inhomogeneous nature of chemical abundances within the Milky Way stellar halo depending on kinematic properties of constituent stars as suggested by earlier studies .
Possible implications for the formation of the Milky Way halo and its relevance to the suggested dual nature of the halo are discussed .