On August 17, 2002, about 1330 eastern daylight time, a Bellanca 8GCBC, N86714, was substantially damaged during a forced landing in Guilford, Connecticut. The certificated commercial pilot was not injured, and visual meteorological conditions prevailed. No flight plan had been filed for the local flight which departed Skylark Airport (7B6), East Windsor, Connecticut, about 1115. The banner tow flight was conducted under 14 CFR Part 91. According to the pilot, the flight had been "going smoothly," and he was heading toward a new display area at 1,300 feet, when the engine "started making noise accompanied by severe vibration and loss of power." The pilot located a field where he could safely drop the banner, and descended to 500 feet to make the drop. He had hoped that the loss of the banner's drag would have enabled the airplane to climb. However, the airplane would not climb, and the vibrations became more severe, so the pilot decided to land in the field while he still had control of the airplane. During the approach, the pilot had to clear trees, and "came in a little high and fast." After touchdown, he "tapped the toe brakes," and the airplane nosed over. A post-flight examination of the Lycoming O-360-C2 engine revealed that the number 3 cylinder was cracked and its exhaust valve was seized. The cylinder was forwarded to the Safety Board Materials Laboratory for further examination. According to the engineer's factual report, the cylinder consisted of two major components: an aluminum head and a steel alloy barrel, which had been screwed together while the head was at an elevated temperature of 600 degrees F. Identification numbers stamped in the head revealed that it had been produced by Engineering Components, Inc. (ECI), and barrel identification numbers revealed that it was an original Lycoming part that had been rebuilt by ECI. The examination also revealed that there was a continuous fracture around the head, in the vicinity of the fourth and fifth cooling fins from the bottom. In addition, the head was missing paint in the vicinity of the two spark plugs and exhaust port, and the interior of the exhaust port exhibited a white deposit. The upper portion of the cylinder head was mechanically removed by a gear puller, with the combustion chamber padded to prevent surface damage, and a piece of wood placed in the cylinder bore. The upper portion of the head slid slowly and easily off the barrel, "indicating that there was no material attaching the upper portion of the cylinder to the lower portion." The fracture face on the lower part of the head had an area that was "flat, perpendicular to the barrel wall, and devoid of any deformation edges, features indicative of fatigue cracking." Within the fracture zone, there were multiple "ratchet marks" adjacent to the inside surface, "indicating a large number of fatigue origin areas." The remainder of the head was removed by sawing through it and the barrel. A measurement of head basic thread dimensions revealed that they conformed to the requirements of the manufacturer's engineering drawing. An examination of threads 1, 2, and 3 (from the bottom) revealed indentations and displaced material on the flanks below the thread roots, and on the flanks above the roots of threads 2 and 3. Hardness tests on two specimens revealed that both were within limits. In addition, optical emission spectroscopy was performed on a section of the lower portion of the head which revealed that the chemical composition was within limits. An initial examination of the barrel revealed "no obvious indications of damage." A measurement of basic thread dimensions revealed that the thread profile, with the exception of the first three threads, conformed to the engineering drawing. The first three threads displayed a flat crest per the taper requirement of the drawing; however, there were "sharp corners between the flat crowns and the adjacent thread flank instead of the specified radius." The sharp corners were continuous around the threaded barrel, but "appeared to decrease from thread 1 to thread 3." A higher magnification examination of a section of barrel threads 1, 2 and 3 revealed protruding material. Thread 1 displayed protruding material at the upper and lower corners. Flow lines in the microstructure were "consistent with material from the crown of the thread flowing toward the flanks, resulting in the protruding material at the intersection of the crown and flanks." There was also a continuation of the flank surface under the flowed (protruding) material, "a feature...consistent with lap," defined as "a surface imperfection, appearing as a seam, caused by the folding over of metal, fins or sharp corners and then rolling or forging them onto the surface, but not fusing them." A protruding feature on the thread correlated to a similar-shaped indentation on the thread flank of the cylinder head, below thread root 1. Threads 2 and 3 also had protruding material which appeared to match corresponding indentation shapes on the cylinder threads. According to the engine logbook, all four cylinder assemblies were replaced by "new" ECI cylinder assemblies on March 1, 1999, and had accumulated approximately 635 hours of service at the time of the accident.