We present an optical design for a spaceborne instrument, of about half m aperture, to perform a combined astrometic and photometric survey via a scan similar to that of Hipparcos. A CCD detector array with time delayed integration will permit an astrometic mission accuracy better than 50 microarcseconds for stars brighter than 10th magnitude. 1 1/2 orders better than Hipparcos. The passband is nominally 0.4 to 0.9 microns. For the instrument to have both high measurements rate and high accuracy, the optical system just satisfy several requirements. It should have aberration well under diffraction, for high precision in centroiding and as a means of keeping unmolded shifts of the image centroids small. The system should have a wide field of view so that there is a large overlap of successive scans, have a large field of view for scientific throughput, and have low image distortion so that the stellar images moved at constant rate along columns of detector pixels. The design presented meets these requirements using aspheric surfaces that are manufacturable. We have demonstrated that the instrument will determine the temperature of an observed star without requiring a dispersive element or color filters. The design is thus free of transmissive elements, and protected from the systematic errors that they might have induced, e.g., due to thermal variation variation and to chromatic effects. This study was inspired by our previous consideration of scientific throughput. Our study of data reduction from a scanning astrometic survey mission demonstrated that there is a substantial gain in mission accuracy if the spacecraft precesses without discontinuities such as those that result from gas jet firings. Our study of methods of processing the spacecraft showed that smooth rotation would be possible using solar radiation pressure, but only if the spacecraft rotation rate were increased. Maintaining the integration time for each object would require an optical design of shorter focal length. Meanwhile, our study of mission accuracy as a function of focal length showed that another increase of accuracy would result from shorter focal length, via the greater number of lower-accuracy measurements. Therefore we performed this optical study to find a design with shorter focal length, having a proportionate increase in infield of view. We conceived and investigated a family of short focal length, wide-field designs, and developed a methodology to facilitate selection from among them. The new baseline design achieves diffraction-limited images over a 2.2 degree FOV with a 1.1 degree square central blockage, and has a 7.5 m focal length.