In situ neutron diffraction experiments during tension-compression deformation were performed on five Fe-Cr-Ni alloys with the volume fraction of ferrite (a) ranging from 0.0 to 100 %. Tensile deformation was applied in a step by step manner up to a strain of 1.3-2.0 % followed by compressive deformation, and neutron diffraction spectra were recorded during temporary stops of a deformation machine with fixed crosshead. (111) reflection of austenite and (110) of ferrite, respectively, were measured simultaneously by using a position sensitive detector. Elastic lattice strains in both constituent phases were evaluated from measured diffraction spectra as a function of external load. Based on these experimental results, heterogeneous deformation behavior in the a-g dual phase alloys is discussed considering the Bauschinger effect. It is concluded that large compressive residual lattice strains detected in the g phase after tensile pre-straining, causes the large Bauschinger effect in a-g dual phase alloys.KEY WORDS: in situ strain measurement; neutron diffraction; a-g dual phase alloy; tension-compression deformation; Bauschinger effect.fractometer TKSN400 installed at the medium-power reactor LWR-15 in Nuclear Physics Institute (NPI). A twocrystal sandwich monochromator has permitted us to use neutrons with two different wavelengths simultaneously; l 1 ϭ0.147 nm (Ge(311)) and l 2 ϭ0.27 nm (Si(111)). The diffractometer was also equipped with a deformation device enabling both tensile and compressive tests up to a maximum loading of Ϯ20 kN.The geometric arrangement of the in situ strain measuring system is illustrated in Fig. 1, where width of two Cdmasks was of 2.0 mm. Shape and dimensions of tensioncompression specimens were the same as those used in Ref. 7). Overall specimen strain was measured by means of a strain gauge glued on specimen surface. From here on, it is called macroscopic strain. A specimen was strained in tension in a step-by-step manner up to 0.013-0.020 followed by unloading, and subsequently deformed in compression similarly down to reversed strains of 0.010-0.0013. Diffraction spectra were collected during temporary stops upon a crosshead-fixed regime. An exposure time of 4 ks was necessary to obtain sufficient statistics to perform a reliable analysis of diffraction profiles. After the crosshead was fixed, the applied stress decreased rapidly within 0.4 ks and afterwards slowly by less than 10 MPa as reported in Ref. 7). Such relaxation behavior was observed in all examined cases. Diffraction profiles were measured for 4.0 ks and the uncertainty of the applied stress during the diffraction experiments is then considered to be approximately of 10 MPa. If we employ the testing method with constant load control, the resulting data will suffer from similar ambiguity in strains because of creep deformation. Therefore, such a kind of experimental error cannot be avoided in this case. In the present analysis, the applied stresses and the macroscopic strains recorded after holding time of 4 ks were us...