Twenty half-dumbbell shaped rods were cast in a silverless gold-palladium alloy that contained a minimum of 42% fresh alloy and the rest once-melted alloy. Each pair of half-dumbbells was assembled and either presoldered or postsoldered using a vacuum-oven and a torch-soldering technique. There was no significant difference in the ultimate tensile strength values between the presoldered and postsoldered joints within the same technique group, but the torch-soldered joints for the same solder were significantly stronger (p less than .01). It was believed that slow heating and cooling has deleterious effects on the ultimate tensile strength joints (UTS) possibly because of the microstructural changes and the highly oxidizable content of the postsolder used. Another 15 samples cast using once, twice, and three-times remelted parent alloy were presoldered and postsoldered with the torch only. In each of these groups, specimens joined by presolder showed significantly higher strength than those joined by postsolder (p less than .02). Remelting the parent alloy affected the strength differences between the presoldered and postsoldered joints. It was assumed that on remelting, the nature of the parent alloy is changed because some original elements are volatilized and newly formed oxides dissolved into it. Because it contains copper, the postsolder is more affected by this contamination than presolder. The UTS of the parent alloy, as well as that of the presolder and postsolder, were also determined. As expected, the parent alloy possessed the highest strength and the postsolder, the lowest.