We have used grazing incidence wide-angle x-ray scattering (GIWAXS) on oriented lipid multilayers to measure chain order and to examine liquid-liquid coexistence in the system DOPC/DPPC/cholesterol, a model for the outer leaflet of the cell plasma membrane. Coexistence of liquid-disordered (Ld) and liquid-ordered (Lo) domains is thought to be related to "rafts" in the cell membrane, cholesterol-rich lipid heterogeneities which provide platforms for protein sorting. Many of the methods used for measuring liquid-liquid coexistence in model membranes require a potentially perturbing probe, while x-ray scattering is probe-free. In unoriented (powder) x-ray data, scattering from the Ld and Lo phases looks very similar, whereas in GIWAXS patterns from oriented samples, these phases are easily distinguishable because of the differences in their chain orientational order. By using a simple analytical model to relate the GIWAXS data to the chain orientational distribution, we fit our data to obtain the average chain orientational order parameter, Smol. While this type of analysis has been well-used for liquid crystals, it is not commonly applied to model membrane systems. For DOPC/cholesterol and DPPC/cholesterol mixtures, composition and temperature dependent trends in Smol determined by GIWAXS are consistent with earlier NMR data. Addition of 40% cholesterol to liquid-phase DPPC or DOPC more than doubles Smol. In addition to measuring chain orientational order parameters for binary mixtures of DOPC/cholesterol and DPPC/cholesterol, we have measured GIWAXS for ternary mixtures where fluorescence microscopy and NMR indicate the coexistence of Ld and Lo phases below the miscibility transition temperature, Tmix. In order to fit to the GIWAXS data for these mixtures at low temperature, we required two values of Smol, which we interpret as evidence of coexisting Ld and Lo phases. Our Tmix values based on x-ray work agree reasonably (to within the 5-10 C temperature steps used) with the Tmix values based on the NMR and microscopy work of Veatch et al. (Veatch and Keller, 2003b; Veatch et al., 2004; Veatch et al., 2007b). This approach provides a new method for examining phase coexistence in model membranes without the need to add a potentially perturbing probe.