In accordance with the ANSI/IESNA American National Standard Practice for Roadway Lighting, RP-8-00, roadway lighting calculations were made to investigate Unit Power Density (UPD). Calculations were made for specific luminaires, roadway configurations and lighting system geometries, over both a range of mounting heights (Ht) and a range of overhang-setback distances (Oh). For each value of Oh, at each Ht, the luminaire cycle distance was increased until any one of the illuminance or luminance criteria are not met. The last distance that meets all the criteria (for both the illuminance and luminance methods) is the maximum for that run. As the mounting height was increased over successive runs, maximum distance consistently increased, before decreasing. Overhang-setback was initially set at the midpoint of its range and at least five adjacent values calculated, then additional variations as necessary to determine the maximum spacing. Optimum luminaire cycle distance typically occurred when the combination with height and overhang result in the uniformity and average criteria both reaching their limits. Both high pressure sodium (HPS) and metal halide (MH) sources were included in the calculations, including some in each of the four classifications of cutoff, and applied to roads of varied classifications (local, collector and major) and widths from one to six lanes. Results indicate that roadway lighting systems which meet both the illuminance and luminance criteria, have predictably lower UPD values with HPS lamps than with MH lamps. The best MH systems have UPD values 1.5 to 1.8 times the values for the best HPS systems applied to similar roadways. Systems that meet only the Luminance method criteria have significantly lower UPD values than systems which meet the criteria for the Illuminance method or both methods. In addition, as luminaire cutoff becomes more stringent, a system's UPD value tends to increase. Systems designed to the Luminance method and with luminaires with HPS lamps and semi-cutoff or non-cutoff distributions typically produce the lowest UPD values for a given roadway classification and width, but as roadways become wider the differences decrease. The increase in UPD when full cutoff is required is shown to be from 9 percent to more than 20 percent.