Adam and Jamie set out to build a seaworthy craft that could hold them and several days' worth of supplies. They designed a canoe with an outrigger for stability, using a bamboo frame and a tape hull, and put together an impromptu worktable to keep dirt out of the tape as they worked. After six hours of construction, they launched the canoe and steered it into the oncoming breakers with little trouble. Seven hours later, Adam sighted land and they put ashore—on the island they had just left.

For a larger test, Jamie built a military-style flamethrower, while Adam made a new extinguisher that used pressurized water to drive the CO₂ out and improve the range. At a bomb range, Jamie's rig achieved a distance of 60 to 80 ft (18 to 24 m), and Adam stood behind a glass shield 60 ft (18 m) away (matching the video) to try and hold it off. A standard extinguisher proved ineffective; Adam's high-pressure rig did hold the flames back, but to a distance far below the full 30 ft (9 m). Since they were unable to replicate the video results, they declared the myth busted.

Almost none of the flamethrower construction details were shown on camera as possession of such a device is illegal where the show is filmed. Jamie noted that he and Adam were able to test this myth only under the supervision of several government agencies, referring to it as an "ordeal".

On a desert plain, they set up a new circle and placed Kari at its center to act as the film's kidnapping victim. Tory and three staffers drove around her to kick up dust clouds, after which one pulled up for her to get in and all four drove away in different directions. Grant tried to pick the right car, watching from a helicopter and using a similar camera setup as an actual drone. He succeeded in the first test due to the lack of dust clouds at the center of the circle; however, when the cars drove straight past Kari in a second test, he was unable to pick the one that took her. A repeat of this test, using a thermal imaging camera such as those used on drones, allowed him to find her easily.

Returning to the shop, the team looked into ways to foil the camera system and discovered that a CO₂ fire extinguisher could blind it effectively, due to the low temperature of the discharge. One last straight-run test in the desert, with the cars spraying CO₂ into the sky, kept Grant from being able to locate Kari. The team judged the myth busted, since the technique in the film did not work, but they noted that the underlying idea could be used successfully under different circumstances.

For small-scale testing of other wheel orientations, Adam and Jamie returned to the shop and built a cart to run on a treadmill. They found that if two wheels, diagonally opposite each other, were turned 45 degrees out of phase with the other two, the cart achieved a balance of smoothness and stability. A full-scale test at almost 20 mph (32 km/h) caused the tire treads and brake pads to come off the wheels. Despite the mechanical failures, they classified the myth as plausible, based on the sensor readouts and their own sensations.

At the bomb range, the Build Team attached a drum to a wheeled dolly and ignited the methanol remotely with a road flare. Their attempts only caused the kart to move a few feet, so they returned to the workshop for small-scale testing. Reducing the fuel/air ratio allowed a slight improvement in distance, but bubbling air into the methanol (for better vapor mixing) gave a much larger increase until the kart set itself and the test track on fire. When the team attached a funnel-shaped exhaust nozzle and injected the fuel through an atomizer (creating a pulse jet engine), they could safely propel the kart to a distance of 15 ft (4.6 m).

For the second full-scale test, they attached the nozzle to a full-sized drum, fitted a scuba air tank to operate the atomizer, and mounted bicycle wheels for reduced friction. This design only moved in short bursts of acceleration and achieved a top speed of 5 mph (8.0 km/h). Declaring the myth busted, they brought in pulse jet expert Robert Maddox to upgrade the engine; his design propelled the kart to 50 mph (80 km/h).

They then built a mockup of a typical hotel room and gave themselves a 6-hour time limit to secure all the furniture to the ceiling, with sheets, cushions, and other items glued in their proper places. Although a few items gave them trouble, including a television set with a waffle-pattern base and a plant in a wicker basket, they were able to finish the job in time.

In a second test, they dressed Buster in a rubberized wetsuit and glued him into a fiberglass racing seat in the car, with the intent of improving the materials' mechanical strength and adhesion. He still tore out of his seat on impact, leading the team to declare the myth busted at this point. For a final test, they attached one steel plate to Buster's back and a second to a specially built seat frame, to act as "ideal" surfaces for adhesion. When they glued the plates together, he stayed in his seat through impact.

For a comparison, they fired a Civil War-era cannon into a line of pallets loaded with concrete blocks, and found that it could penetrate one pallet and part of a second. A brush fire close to the range forced a brief delay in the test, after which a faulty heating element forced them to shut down for repairs. In their second test, they were able to destroy one pallet and found the ball stopped at the front edge of the second. Although their cannon did not have as much destructive potential as a real one and was not suited for practical use, they judged the myth confirmed.

They built a full-sized vehicle and launch ramp and set up at a rock quarry, placing the ramp on a ridge overlooking a lake. With 500 fireworks on the back and Buster strapped in, the vehicle achieved a straight flight that carried it 700 ft (213 m) to a touchdown on the other side of the lake. The team reversed their original "busted" decision and declared the myth confirmed. However, they also noted that a person trying it would have a very poor chance of surviving the impact on landing.

They then set up a trailer with a high-speed camera and measuring scale and towed it behind a pickup truck at 45 mph (72 km/h) as Jamie drove. When Adam fired the pistol, the speed and trajectory of the ricochet matched those from the stationary test. However, two 5-shot test runs with a target car in place of the trailer gave no penetration, due to bullets being deflected by other metal components in the underside. With the myth busted, Adam drove a car lined with bulletproof vests at 45 mph (72 km/h), triggering a pistol to shoot out one of his own tires. Although he did not immediately lose control, the tire had gone completely flat after 0.5 mi (805 m) and he was forced to stop. He and Jamie observed that shooting a pursuer's tires was much more effective than trying to bounce a shot up through the floor.

For full-scale testing, the Build Team set Buster up 50 ft (15 m) from an explosion site and fitted a trigger to set off the blast 0.5 s after he had jumped. A control run with no explosive gave a range of 27 ft (8 m); this result did not change when a 10 lb (4.5 kg) charge of C-4 was used. Switching to the lower-velocity explosive ANFO in 10, 25, and 100 lb (4.5, 11.3, and 45.4 kg) quantities did not improve his range. Finally, Tory built a steel frame and attached a blanket, sprayed with truck bedliner resin, to serve as a shock wave catcher and strapped Buster to it. Both a control run and a test with 100 lb (45 kg) ANFO gave a range of 23 ft (7.0 m), leading the team to judge the myth busted.

They decided to do small-scale tests in order to determine the amount and/or type of wrap needed to make the landing safe (defined as 10 g or less), dropping a human analog onto the wrap instead of covering it. Using a fall height of 6 ft (1.8 m) and a plastic pipe filled with cannonballs, they found that 16 in (41 cm) of wrap gave between 22 and 25 g. Further tests on Buster from 15 ft (4.6 m) allowed them to design an alternating layer/coil wrapping design that reduced the impact to 15 g.

Under Jamie's supervision, the crew fitted Adam with safety equipment, covered him in three layers of this design, and dropped him from 15 ft (4.6 m). The resulting force on landing was 9 g; however, Adam found the total weight of the padding (150 lb (68 kg)) very uncomfortable and hit his head on the support board underneath him. When Buster was wrapped up and dropped from the full 35 ft (11 m), he experienced forces of 29 g (torso) and 48 g (head), indicating that he would survive with some degree of injury. Since the circumstances shown in the video would have proven fatal, Adam and Jamie declared the myth busted.

After resetting the seat and turning the car upside down, the team fired the rockets but the car did not move at all. Grant noted that the rockets' thrust was too small compared to the car's weight (approximately 4000 lbs). A second test, with enough rockets added to generate this much thrust, still failed to flip the car and also set the interior on fire. Declaring the myth busted, they explained that the effect in the film had been produced using a high-pressure nitrogen cannon rather than a rocket. They then built one themselves which successfully flipped the car.

Next they investigated the scenario of a charging knife wielder. Adam and Jamie determined that Jamie could run 24 ft (7.3 m) before Adam could draw, cock, and fire his pistol. When Jamie ran at Adam with a foam knife from this distance, Adam was able to shoot him; a second test at 20 ft (6.1 m) gave the same result. However, when Jamie started at distances of 16 ft (4.9 m) or less, he successfully stabbed Adam without being shot. They decided that a knife wielder at close quarters could have an advantage over a gunfighter, depending on the circumstances, and declared the myth busted.

Building the Fire Dragon to an overall length of 6 ft (1.8 m), the Build Team took it into the desert to fine-tune the ignition timing. The goal was to launch the arrows at the peak of the flight, using a fuse cut to an appropriate length; although the stabilizing fins fell off in midair, the arrows did launch at once.

After a Fire Dragon warhead achieved a range of 800 yd (732 m), as opposed to a single rocket arrow's 650 yd (594 m), the team set up 10 of the weapons and aimed them at a group of dummies and a castle wall 800 yd (732 m) downrange. Seven of the arrow salvos fired, but only a few arrows landed in the target zone and no targets were hit. The team classified the existence of the Fire Dragon as plausible, but noted that the weapon was highly unreliable.

Finally, Adam and Jamie posed as Jack and Rose, respectively, and tried to climb onto a board modified to accommodate their superior weight. Although Jamie managed to fit on just fine, when Adam tried to climb on with him, the board was unable to support them. Then they got the idea to remove Jamie's life vest and strap it to the underside of the board, which increased its buoyancy enough to support the two, leading them to declare the myth plausible. But James Cameron said that while that meant the prop designers had gotten the size wrong, the script from the movie said that Jack had to die.

Next, they dug three trenches with the same shapes as the tanks, and lined the walls with color-coded plywood to both distinguish the trenches from each other and compensate the shapes as much as possible. Finally, they began testing with explosives, first with an aboveground control, in which sensors at 20, 30, 40 and 50 ft (6, 9, 12 and 15 m) registered pressure differentials of 39, 12, 7 and 5 psi (269, 83, 48 and 34 kPa), respectively. In the straight trench test, the same sensors registered 397, 65, 38 and 21 psi (2,737, 448, 262 and 145 kPa). The sharp-angled trench test showed pressure differentials of 60, 19, 12 and 7 psi (414, 131, 83 and 48 kPa). Finally, the round-angled trench test yielded pressure differentials of 76, 21, 13 and 8 psi (524, 145, 90 and 55 kPa). Because the sharp-angled trench showed the lowest pressure readings out of the three trench tests, Adam and Jamie dubbed the myth plausible.

To investigate the suitabilities of different balloon configurations to cushion the clown passengers, Grant built a rig to drop 50 lb (23 kg) of weight on the balloons inside a large polycarbonate tube. In a control test with no balloons, the weight pulled 100 Gs at impact. They then tested with the same balloon setup as in their previous car crash test, which only reduced the impact to 98 Gs, popping all of the balloons in the process. However, subsequent tests with separate configurations of balloon "animals", small balloons and large balloons yielded respective impacts of 6, 2.8 and 2 Gs. But in a car crash test with the best-performing balloons (the large ones), the passenger suffered 115 Gs on the head and 130 Gs on the chest, both of which would still have been fatal impacts.

Despite calling the myth busted at that point, the team continued testing with various balloon configurations of their own designs. They tested a large ball constructed from balloons made to resemble smoked sausages, built by Kari; a giant makeshift balloon bubble wrap, built by Grant; and giant mylar balloons stuffed with smaller latex balloons, built by Tory. When tested in the polycarbonate tube, the three designs showed respective impact results of 3.8, 1.7 and 2.8 Gs. But when Grant's design, the best performing of the three, was tested in a car crash, the passenger pulled 220 Gs on the head and 230 Gs on the chest, a larger impact than the large balloons.

They built a full-size, 80 ft (24 m) bridge from planks and steel cables and strung it up across a dry dock at Mare Island Naval Shipyard. Jamie stood near one end, ran toward it as the other end blew out, and tried to jump for the end bar after taking two more steps. He fell as soon as the bridge collapsed, and a second attempt (with Adam standing closer to the end and trying to jump for the bar) also failed. They noted that as soon as the cables were cut, the bridge lost the tension holding it up and the jumper's support was gone.

They declared the myth busted at this point, but Jamie wondered if a person could simply hold on to the handrails. Once the rope rails were wrapped with tape, he ran toward the end as the bridge fell and was able to hold on. He tried to climb up, but tired and slipped down when he was just short of the edge. However, Jamie joked that he probably would have succeeded if he genuinely thought his life was on the line.

At the Port of San Francisco, the team put a boat in the water and began dumping ice into it. The boat only began to sink after more than 4,000 lb (1,814 kg) of ice had been added, more than the world record hail-fall of 4 ft (1.2 m) in a single afternoon. However, it then bobbed up to the surface due to buoyancy forces that allowed the ice to float free of the boat once it was underwater.

The team decided to explore a third scenario, having a hailstone land in the boat while it was in the water. Grant set his cannon up to fire vertically downward into a floating boat, using the fiberglass and wood hulls. The fiberglass did not break at 300 mph (483 km/h), while the wood did puncture and take on water. A second test at 80 mph (129 km/h) did not puncture the wood, but a final shot at 150 mph (241 km/h) did. They classified the myth as plausible, noting the specific set of circumstances needed to sink a boat.

Disguised as an old groundskeeper, Adam met with 10 volunteers and asked them to go into the cabins, saying that something bad had happened in one of them. One at a time, the volunteers spent 3 minutes sitting in each cabin, then made their guess as to which one Adam was referring to. After only 2 of them guessed correctly, Adam and Jamie judged the myth busted, but noted that having all 10 volunteers visit the cabins in the same order may have affected the results.

The team brought in 12 volunteers and had each one smell 6 pairs of samples, guessing which (if any) samples in each pair were fear sweat. None of the volunteers got more than 2 out of 6 right, but odor scientist Pamela Dalton correctly guessed 5 out of 6 when she took the test.

To obtain a fresh sample of fear sweat, Tory rappelled to the bottom of a 165 ft (50 m) vertical shaft in Moaning Cavern, fighting acrophobia as he descended. That sample was shuffled with 19 non-fear ones, and Pamela attempted to pick out the fear sweat sample from the group of 20. Even though she did not succeed, the team classified the myth as plausible, noting the need for years of experience in the field of odor detection.

For a control run, each man ran the course using a wad of steel chain with the same weight as the cadaver. Adam posted a time of 3:06; Jamie struggled at the window and finished in 4:33. Both of them began to tire noticeably toward the end of the run. With the cadaver, though, they made much better time due to its weight distribution. Adam finished in 1:20, Jamie in 1:55, leading them to call the myth plausible. Adam and Jamie noted that unlike the steel weight, cadavers are actually much easier to carry since they can be slung over the shoulder to distribute the weight more evenly, and unlike the wad of chain, the cadaver could be moved more incrementally (lifting an arm or a leg at a time, rather than the whole thing all at once).

Eleven of the 12 chose correctly, prompting Adam and Jamie to investigate the ease of identifying other body parts—namely, palms and teeth. Using the same methodology, 7 of 12 were correct on the palm test, 10 of 12 on the teeth.

They debated the effectiveness of the "potty dance": distracting attention from the full bladder versus putting more pressure on it. The second day, they repeated the trial but allowed themselves to do the dance. This time, Tory lasted 1:30, Grant 1:52, Kari 2:46. On the third day, they tried various methods to relax as much as possible. Tory lasted 2:05, Grant 2:46, Kari 2:43.

The team judged the myth plausible, but noted that results of different methods can vary widely from person to person.

When the team switched to 0.0625 in (1.6 mm) glass, Matt was still unable to penetrate it but did manage to shatter one needle; with 0.03125 in (0.8 mm) antique glass, he smashed it on his first throw. The team declared the myth busted for human throwers at this point, but Tory modified an air rifle to fire the needles at high speed. Using 0.125 in (3.2 mm) glass, they got no result at 140 mph (225 km/h), chipped the glass at 160 mph (257 km/h), and finally punched a clean hole at 180 mph (290 km/h).

They worked separately on modifications to improve traction on the uphill run. Adam added 50 lb (23 kg) of weight to his bicycle's frame and filled its tires with corn syrup to flatten them out, while Jamie mounted a 120 lb (54 kg) set of lead training wheels and attached another 20 lb (9 kg) weight in front. They laid out a race course at the bottom of the pool, but neither man was able to ride his modified bicycle on its uphill portion without using his feet to push ahead. They classified the myth as busted, since they were able to ride on flat and downhill runs.

They repeated the test with two simulated cadavers and brought in an orthopedic surgeon to assess the extent of injuries with a portable X-ray machine. The drop with no mattress resulted in a paralyzing spinal cord injury and shattered pelvis, while the drop onto the mattress led to even worse pelvic and spinal injuries and a dislocated knee and elbow.

Declaring the myth busted at this point, Adam and Jamie decided to find out if there was a way to make the jump and survive. Jamie trained with a professional stuntman to learn jumping/landing techniques, working his way up to greater heights. When he developed severe back pain overnight, though, he left the final attempt up to the stuntman, who jumped from the full 35 ft (10.7 m) and came safely to a stop in just over 4.5 ft (1.4 m) of water.

The Build Team fired a Civil War-era cannon, loaded with an iron ball and 1 lb (454 g) black powder, and measured a speed of 1,200 ft/s (366 m/s). Their first attempt to match that speed with their own small-scale cannon led to the accident, and the testing was immediately called off to assess damage and start an investigation.

Several months later, they resumed their work at a remote rock quarry and secured permission to fire stone balls through the full-sized cannon, using a cardboard canister to protect the barrel. On the advice of a geologist, they chose sandstone, limestone, and granite for further testing and fashioned one ball from each material. They scaled down the powder charge based on each ball's weight to match the speed of the iron one, then set up pallets of bricks to stand in for a castle wall.

A control test with an iron ball penetrated 2 layers of the wall and destroyed 4 bricks. The sandstone penetrated 1 layer and destroyed 2 bricks; the limestone, 1 layer and 3 bricks; the granite, 2 layers and 4 bricks. In addition, every stone ball broke apart on impact. Having found a stone that could meet both of the original criteria, the team declared the myth plausible.

While Alton prepared the ingredients, Adam and Jamie built or modified containers to hold the dishes and Jamie added insulation under the hood where needed. They loaded the containers into the chosen spots - which included near the exhaust manifold and the transmission - and drove for 4 hours, again escorted by the CHP, stopping occasionally to check the food temperatures and re-position dishes in order to prevent overcooking. At their destination, they tasted the dishes, agreeing that if the turkey was not satisfactory, then the myth was busted. However, the cooking with the engine compartment was successful, and they ended up with a dinner that included the following dishes:

• Turkey breast loaves
• Turkey legs confit-style with olive oil and thyme
• Turkey thighs on herbs
• Dressing
• Cream of corn with rosemary
• Sweet potatoes
• Mini pumpkin pies (In an added twist, Alton had never tested the recipe in a real oven before.)
• Gravy soup (The only dish that did not cook as originally planned - it did not thicken, but was deemed a great soup.)

Assisted by a nutrition professor, Kari designed three holiday meals for the other two, to be eaten on consecutive days. After the first meal, a large one with turkey, Tory scored 47, Grant 52. The second meal had the same number of calories, but replaced turkey with protein powder mixed in the mashed potatoes, and led to scores of 48 and 42. After the third meal, identical to the first but with smaller portions, they scored 63 and 67. The Build Team judged the myth as busted, since the amount of food eaten had much more of an effect on reaction time than the turkey did.

After Tory noted that the texture of the meat was a major clue as to its identity, Kari ground 20 samples (12 exotic, 8 chicken) and cooked them as patties with no additional flavoring. This time, Tory and Grant scored 18 and 19, respectively, leading the Build Team to call the myth busted. However, there was one exotic meat (boar) that both Tory and Grant mistook for chicken.