Following our development of a superconducting transition-edge-sensor (TES) microcalorimeter design that enables reproducible, high performance (routinely better than 3 eV FWHM energy resolution at 6 keV) and is compatible with high-fill-factor arrays, we have directed our efforts towards demonstrating arrays of identical pixels using the multiplexed read-out concept needed for instrumenting the Constellation-X X-ray Microcalorimeter Spectrometer (XMS) focal plane array. We have used a state-of-the-art, time-division SQUID multiplexer system to demonstrate 2×8 multiplexing (16 pixels read out with two signal channels) with an acceptably modest level of degradation in the energy resolution. The average resolution for the 16 multiplexed pixels was 2.9 eV, and the distribution of resolution values had a relative standard deviation of 5%. The performance of the array while multiplexed is well understood. The technical path to realizing multiplexing for the XMS instrument on the scale of 32 pixels per signal channel includes increasing the system bandwidth by a factor of four and reducing the non-multiplexed SQUID noise by a factor of two. In this paper we discuss the characteristics of a uniform 8×8 array and its performance when read out nonmultiplexed and with various degrees of multiplexing. We present data acquired through the readout chain from the multiplexer electronics, through the real-time demultiplexer software, to storage for later signal processing. We also report on a demonstration of real-time data processing. Finally, because the multiplexer provides unprecedented simultaneous access to the pixels of the array, we were able to measure the array-scale uniformity of TES calorimeter parameters such as the individual thermal conductances and superconducting transition temperatures of the pixels. Detector uniformity is essential for optimal operation of a multiplexed array, and we found that the distributions of thermal conductances, transition temperatures, and transition slopes were sufficiently tight to avoid significant compromises in the operation of any pixel.