On April 22, 2004, about 1632 Alaska daylight time, a wheel-equipped Arctic Aircraft Corporation S-1B2 airplane (Arctic Tern), N55AD, sustained substantial damage following a loss of engine power and subsequent in-flight collision with trees during the emergency descent/landing, about 9 miles south-southwest of Willow, Alaska. The 14 CFR Part 91 instructional flight was operated by the student pilot/airplane owner, and operated in visual meteorological conditions. No flight plan was filed for the local flight. The student pilot reported minor injuries, and the certificated flight instructor received serious injuries. The flight departed the Willow airstrip, Willow, Alaska, about 1600. The NTSB investigator-in-charge had a telephone conversation with the student pilot and flight instructor on April 22, about 1710. They both related essentially the same information. The purpose of the flight was to provide instruction to the pre-solo student pilot in preparation for a private pilot certificate. It was the instructor's first flight in the student's airplane, and in an Arctic Tern. The instructor took the rear seat in the two-place, tandem seat cabin. After a routine preflight inspection, engine run-up and departure climb, some general air work was performed. During the air work, about 2,500 msl, the instructor said he simulated a loss of engine power by reducing the throttle to idle, and instructed the student to select an emergency landing site and initiate a landing approach to the site. The student turned the airplane toward an abandoned airstrip near Deshka Landing. When the airplane had descended to approximately 1,200-1,500 msl, the instructor told the student to "recover" and apply engine power. The student replied that the engine would not respond. The instructor told the student to continue the landing approach, but the student indicated he wanted the instructor to make the emergency landing. The instructor said the airplane was too high to land on the small, partially obstructed airstrip, and he elected to fly the airplane slowly into trees beyond the runway. The airplane received structural damage to the wings, fuselage, and empennage. In follow-up conversations with the student, and in his written report to the NTSB, the student stated that approximately 25-28 gallons of fuel was aboard the airplane at the time of the accident. He said during the initial simulated loss of engine power, the carburetor heat control was never pulled on, and the instructor never told him to pull it on. He also said that he was not told to try restarting the engine by using the starter motor. The student believes the loss of engine power was precipitated by not using carburetor heat, and the subsequent carburetor icing. He said the carburetor heat was checked during the initial pretakeoff engine run-up, and it worked fine. He also noted that there is no carburetor heat control in the rear seat, where the instructor was seated, but there is a throttle and flight controls. The instructor wrote that he simulated an engine out procedure by reducing the throttle to 1000 rpm, and " I stepped him through the standard engine-out procedure- fuel (switch on fullest, mixture rich), carburetor heat, magnetos, while turning toward Deshka Landing." He noted that he could see the top row of gauges from the back seat "by straining" but could not easily see the position of the engine controls. He indicated that when he exited the airplane after the accident, he saw a "pulled knob on the left side panel." Photographs of the accident site and the airplane disclosed that the fuselage had compressed several inches, and that the engine had been pushed rearward and downward. The carburetor heat control knob on the left side of the panel appeared to be out approximately 1/2 of its full travel. On May 19, 2004, under the direction of an NTSB air safety investigator, and with two Anchorage FAA Flight Standards District Office aviation safety inspectors present, the accident engine was mounted on another airframe at a repair facility in Big Lake, Alaska. The engine did not have its carburetor air box attached due to damage received in the accident. The engine was subsequently started and run through various power ranges without any noted mechanical anomalies. The closest weather reporting station to the accident site was at Willow, about 9 miles north-northeast of the accident site. The 1451 Alaska standard time METAR reported the sky condition as 3,000 feet scattered, 5,000 feet broken. Temperature was 46 degrees F, and Dew Point 21F. A carburetor icing probability chart is included in the docket of this report. It depicts the potential for carburetor icing at various temperature and dew point combinations. Using the temperature and dew point of 46 F and 21F, respectively, the chart indicates there is a potential for carburetor icing at cruise and glide power settings. FAA Advisory circular 61-23C, Chapter 2, Carburetor Icing, states, in part: "...if the temperature is between -7C (20F) and 21C (70F), with visible moisture or high humidity, the pilot should be constantly on the alert for carburetor ice. During low or closed throttle setting, an engine is particularly susceptible to carburetor icing."