A first student project to put pedals on a wheelchair for exercise and propulsion was unsuccessful. The need remained and in June of 2005 the “Eureka” event occurred. Seeing a five-year-old on her training-wheel-equipped bicycle suggested that a fifth wheel could be added in the center between the wheelchair's two large rear wheels, and a mast supported by the fifth wheel's axle could extend forward to support a front axle and pedal set. A chain drive completed the propulsion system. There are no pedal-powered wheelchairs currently on the market. Around 2001 a product (EZChair) without retractable pedals was on the market but withdrawn. A team at the University of Buffalo invented and patented a pedal-powered wheelchair in 1993 (US Patent 5,242,179), but it was not commercialized. Also, a Japanese company designed and built a series of fifth-wheel wheelchair designs. Between 2006 and late 2008 we built many prototypes incorporating geometries that permitted retracting the pedal. For compactness a “Pedalong” with three telescoping tubes was built but it proved impossible to secure tightly. In the next design twin telescoping tubes passing above and to the rear of the rear axle provided the desired extension. A clamp at the front of the outer tube provided tightness of the assembly. In the Northwestern research program (see below), there was some success, but awkwardness in operation prevented commercialization. In October 2008 a major design change from a fifth wheel in the center to a powering of the two standard rear wheels was begun. This required a new chain path geometry and addition of a differential to the drive train. With the new design user control, arm-powering and braking through the rear wheels is retained, and chair stability is improved. Twelve individuals with chronic post-stroke hemiplegia ($>6$ months post-stroke event) participated in a study to examine the metabolic energy expended when participants performed a 6-minute walk test, a 6 minute leg-propelled wheelchair trial (using the Pedalong), and a 6 minute arm-propelled wheelchair trial. VO2, VCO2, and distance traveled were measured using a portable metablic cart system and wheel-based distance measurement system. The Pedalong and walking trials showed equivalent oxygen consumption levels, but manual pushing was, on average, significantly less. All three modes (walking, leg-propelled and arm-propelled) resulting in similar distances traveled within the 6 minute period. The leg-propelled trials generated the greatest amount of VCO2 during expiration compared with the other modes. This means that more of the available oxygen is being utilized (metabolized) during the leg-propelled mode and so, a greater number of calories were being burned during this 6-minute test.