On May 15, 1999, approximately 1230 Pacific daylight time, a Lake LA-4, N152WB, impacted the metal segmented circle at the base of the wind sock during a landing attempt on runway 19 at Coeur D'Alene, Idaho. The certified flight instructor (CFI) and his student, who holds a private pilot license, were not injured, but the aircraft, which was owned by the CFI, sustained substantial damage. The 14 CFR Part 91 instructional flight, which originated at Coeur D'Alene about three hours earlier, was being operated in visual meteorological conditions. No flight plan had been filed, and there was no report of an ELT activation. According to the instructor pilot, the student allowed the aircraft to get high on approach while attempting to land in variable, quickly shifting winds. The instructor told the student to steepen the approach and to center the nose of the aircraft on the runway centerline, which he monetarily did. Just prior to the flare, the instructor noticed that the wind sock was "blowing from side to side," and he again "...instructed rudder correction." Then according to the instructor, during the flare, the nose of the aircraft "...veered sharply to the right and the nose pitched up." The student was unable to bring the aircraft back into correct alignment with the runway during the touchdown, and he allowed the aircraft to land hard while "...sideways on the runway." The instructor then took over just after the aircraft touched down, but it "...jerked sharply to the right and off the runway." The instructor was unable to get the aircraft realigned with the runway using full left rudder and left brake, so he added a burst of power because he thought the increased airflow over the rudder would cause it to be more effective. The addition of power did not help in the instructor's attempt to turn the aircraft back to the left, and instead, the rate of turn to the right increased as the aircraft departed the runway and impacted the segmented circle. While the aircraft was being disassembled after the accident, the left side engine pylon mount upper strap was found to have failed through a weld that appeared to be part of a previous repair. A portion of the strap, to include the weld line, was sent to the NTSB materials laboratory for analysis. During that analysis, it was determined that there were crack arrest position marks along portions of the weld that were consistent with propagation of a fatigue crack. It was further determined that when the repair weld was made, it had inadequate penetration, minimal fusion, and numerous voids. Also, the strap, which was made of low-alloy steel, was welded using a stainless steel electrode. A review of the log books did not reveal when the repair had been made, nor whether Service Letter #43, which calls for inspecting the straps for cracks every 100 hours, had been complied with. It was also noted that with the strap fractured, the engine mounting pylon would twist when a cylinder was grasped by hand and forward pressure was applied.