On April 23, 2018, about 1150 eastern daylight time, a Bellanca 7GCBC airplane, N8737V, was substantially damaged when it was involved in an accident near Andover, New Jersey. The pilot was seriously injured. The airplane was operated as a Title 14 Code of Federal Regulations Part 91 personal flight. The pilot reported that, before departure and during the preflight inspection, he sumped about 32 ounces of fuel from each wing, the gascolator, and the lower sump. During taxi to the fuel pumps to refuel, he smelled smoke, which he thought might have been due to a bird's nest, so he and the airport manager removed the top and bottom cowls and thoroughly inspected the area around all the cylinders and the bottom cowl area and found no nests. He subsequently fueled the airplane with 13 gallons of fuel and conducted and engine run-up, and all indications were normal and remained normal throughout the takeoff roll. However, after taking off and as the airplane was about 500 ft above ground level (agl), the pilot noticed that the rpm was decreasing. He applied carburetor heat, but the rpm continued decreasing. When he repositioned the carburetor heat to cold, he saw that the rpm was still decreasing, so he decided to turn left toward the airport and subsequently landed uneventfully. He then completed a walk-around inspection with no anomalies noted. An airframe and powerplant mechanic, reported that he saw the airplane take off and that the engine sounded normal when the airplane passed him at midfield. He added that, shortly thereafter, he heard the power "tapering off." He added that the engine sounded normal during the landing roll. The airport manager reported that, after the pilot refueled the airplane, he saw him remove the engine cowl to look for a bird's nest, and that he did not find any. Shortly after the airplane took off from runway 03, he saw the pilot turn the airplane toward runway 21 and then land uneventfully. After the landing, the pilot asked the mechanic to help him look for anything that might have caused the loss of rpm. The mechanic reported that he found the intake cuff moderately deteriorated, installed incorrectly, and clamped to the airbox instead of the cowl. The mechanic provided the pilot with a new clamp but did not watch him install it. The mechanic noticed that the No. 3 exhaust manifold had “vibrated the nuts loose and it was hanging;” the remaining exhaust nuts were tight. Although he stated that they then replaced two gaskets, and the mechanic installed new nuts, washers, and lock washers, the mechanic reported that he just supplied the pilot with the parts and did not watch him install them. The mechanic examined the airbox and venturi, and found no obstructions, and he did not see anything that would have prevented fuel going to the carburetor. The mechanic observed a kink in the main fuel supply line, which he replaced. The mechanic reported the line was made out of an unreinforced automotive fuel or coolant hose. Subsequently, the pilot conducted a full-throttle engine run-up and a walk-around inspection with no anomalies noted. He checked the fuel level, and it showed over half a tank. The mechanic stated that, after the engine run-up, the pilot asked him if the items they found earlier could have contributed to other power loss-related events the engine had experienced in the past year and that he agreed that they could have been contributors. He added that, before further trouble shooting of the engine problem was performed, the pilot took off for the next flight. The pilot reported that he taxied the airplane to the end of the taxiway and conducted another full-throttle engine run-up for about 45 seconds with no anomalies noted. After a normal takeoff and just after clearing the end of the runway and the beginning of a lake and when the airplane was about 300 ft agl, the pilot noticed the rpm decreasing “quickly.” He immediately made a slight right turn toward a field as a possible emergency landing site, but due to the rapid decrease in rpm, he had to lower the airplane’s nose to keep it flying, which eliminated the field as a landing choice because the airplane would not be able to clear a tree line in its path. Subsequently, the pilot turned left toward the departure airport, but about a third of the way through the turn, the engine “stopped producing any usable power.” The pilot added that he “pulled the nose up as hard as…[he] possibly could right before impact" with the water. The pilot egressed the airplane and was subsequently rescued by first responders. The airplane sank in 44 ft of water and was recovered about 24 hours later. The mechanic reported that, although he did not see the accident sequence, he did hear the same power reduction about the same time after takeoff as it had occurred after the first takeoff. The airport manager said he saw the airplane take off the second time and then veer slightly right, at which point, the engine seemed to “sputter and stall” when the airplane was about 200 ft. The airplane then “broke left and descended at a sharp angle,” but the pilot leveled the airplane just before it impacted the lake. Review of the airplane’s maintenance records revealed that the engine’s most recent field overhaul occurred on May 1, 2015, at a tachometer time of about 3,081 hours. The most recent annual inspection occurred on May 4, 2017, at a tachometer time of about 3,419 hours. The pilot did not provide engine or airframe times from the last annual inspection to the time of the accident. Examination of the airframe and engine revealed no evidence of any preaccident mechanical malfunction or failures that would have precluded normal operation.