An Engineer
Looks at the


In early July 1947, Mac
Brazel, foreman of the Foster sheep ranch northwest of Roswell, New Mexico, discovered a
large quantity of extremely unusual, widely scattered, and highly fragmented lightweight
debris on a pasture. The Army initially attributed it to the misidentified remnants of a
downed weather balloon and an attached radar target. Fortyseven years later, the Air
Force explained the debris as similar wreckage from a topsecret Project Mogul balloon
train. More than three years have passed since I first examined the debrisfield issue quantitatively in "The Roswell Debris: A Quantitative Evaluation of the Project Mogul Hypothesis," IUR, March/April 1995. This article updates my initial research. The debris fieldMost of the debris consisted of lightweight, thingage, flat or slightly curved, sheetlike fragments exhibiting physical properties atypical of late1940s technology. There appears to have been two distinct types at the site: (1) foillike pieces easily crumpled by hand which completely recovered their original shape and showed no signs of wrinkling when released; and (2) pieces which could not be deformed or damaged by any means, even when whacked with a 16pound sledgehammer. I refer to both kinds of debris, which would not burn, as thinshell material. (Common thinshell products—which do not possess such extraordinary deformation characteristics—include Saran wrap, balloon film, aluminum foil, and motorvehicle and aircraft body panels.) Major Jesse Marcel, intelligence officer of the 509th Bomb Group based at Roswell Army Air Field (RAAF), inspected the site shortly after Brazel reported the debris to the Chaves County sheriff in Roswell. Marcel described a big field: debris ". . . about as far as you could see—three quarters [of a] mile long and two hundred to three hundred feet wide." It was "scattered all over—just like you’d explode something above the ground and [it would] just fall to the ground." The shortest pieces were "four or five inches. It was [as if it were from] something of some greater area that had been together." A rectangular field shape is implicit—but by no means certain—in Marcel’s description. For that configuration, the debris would have littered an area equal to 250 ft (average width) x 4,000 ft (about ¾ of a mile) long = 1,000,000 ft^{2} (nearly 23 acres). Methodology for quantitatively evaluating the Mogul hypothesis Evaluating the Mogul hypothesis quantitatively constitutes a threestep process.
To carry out the first step in the evaluation process, I had to formulate a mathematical model of the debris field. Its development is detailed in my 1995 IUR article, but a brief description is provided below. Debrisfield model The debrisfield model uses surface area to estimate the quantity of unconventional thinshell material on the Foster ranch. Surface areas of other debris forms that covered significantly less of the field are neglected. To simplify the analysis, the field is represented mathematically as a flat plain with flat shell fragments distributed over it in a variable, twodimensional pattern. One very small region near one (the presumed front) edge of the field is assumed to have a moderately dense debris concentration—20% ground coverage. (The actual site must have featured at least one, fairly large region of closely packed debris because the sheep refused to cross it; they had to be driven around it.) The density drops off rapidly away from this location, exhibiting an extremely spotty coverage over the rest of the simulated pasture. The unknown debris distribution is approximated by equations which attempt to simulate the major elements contained in the most reliable firststatement, firsthand, and secondhand witness testimony. My objective was to generate a conservative approximation of that pattern—one that would estimate the smallest possible amount of onground thinshell material—to minimize the risk of unfairly biasing the evaluation outcome against the Mogul hypothesis. Consistent with that approach and for additional reasons discussed in the 1995 article, I selected a parabola instead of a rectangle as the most plausible debrisfield shape. The model’s parabolic field has the same overall dimensions described by Major Marcel: 250 feet wide (at its far end) by 4,000 feet long. But it encompasses only 667,000 ft^{2} (about 15 acres) of pasture, 33% less ground area than that contained in the rectangular configuration. Initial evaluation: Mogul Flight 9 In the fall of 1994 the Air Force posited that Mogul Flight 4 was the source of the debris on the Foster ranch. That explanation seemed so improbable I applied my newly developed model to a different balloon train, Flight 9. Researcher Karl Pflock, in his 1994 monograph, Roswell in Perspective, speculated that the debris was merely polyethylene (plastic) balloon fragments from downed balloons comprising that aborted experiment. My analysis showed that if all the balloons in Flight 9 had landed and shredded at the site, nearly four such trains would be needed to generate enough debris to litter 15 acres of ranch land. Then, in his 1997 book UFO Crash at Roswell: The Genesis of a Modern Myth, coauthor and former Mogul project engineer Charles B. Moore revealed that Flight 9 actually used neoprene (synthetic rubber) balloons. Since Flight 9 was thus eliminated as a candidate debrisfield causative agent, I took another look at the Flight 4 hypothesis. Updated evaluation: Mogul Flight 4 According to Charles Moore, Flight 4 consisted of 28 neoprene, meteorologicalsounding (i.e., weather) balloons attached to a 600footlong master line of braided nylon cord, three ML307B rawin radar targets, possibly one or more silkcanopy parachutes, and a variety of test equipment such as a sonobuoy microphone, radio transmitter, dry cells, and plastic containers holding solid and liquid ballast. All components and systems were ordinary offtheshelf items; only the Mogul program objective was classified.
(personal correspondence dated March 31, 1995), "These targets consisted of nine righttriangular segments with 24inchlong bases and heights. Each segment consisted of a panel of aluminum foil laminated to some fairly tough, heavyduty paper and deployed on balsa wood struts." (Moore is also on record saying that the struts were coated with glue to stiffen and render them noncombustible.) Moore recently told Roswell researcher Robert J. Durant that each target was constructed of only seven triangular segments having the same dimensions given above. I elected to use nine triangles—yielding a singletarget surface area of 18 ft^{2}—in order to maximize the total amount of Mogulsupplied thinshell material. This gives the benefit of the doubt to—and helps—the Mogul hypothesis.
A quantitative assessment. The quantitative evaluation methodology outlined earlier will now be applied to Mogul Flight 4. The objective: Ascertain if there was enough thinshell material in that balloon train to account for the reportedly unusual shelllike stuff on the Foster ranch site. As noted previously, the radartarget laminate—aluminum foil bonded to paper—was the only possible source for that type of material. Let Area 1 denote the modelestimated surface area of onground, thinshell debris. This parameter, calculated in my initial study (and still applicable), is equal to 6,880 ft^{2}. To give the Mogul hypothesis a fighting chance, assume that all three targets landed on the pasture and were broken apart and shredded into pieces of varying size. The total laminate surface area (Area 2) is equal to 3 targets ´ 18 ft^{2 }/ target = 54 ft^{2}. Compare the two surface areas by dividing Area 1 by Area 2. We obtain 6,880 ft^{2} ¸ 54 ft^{2} = 127.4. This number tells us that it would take more than 127 Flight 4 balloon trains — more than 381 radar targets — to litter the field! Mercy. Model accuracy. Obviously we can’t validate the accuracy of the above finding. But a crude, indirect measure of the model’s simulation capability can be defined: C—the overall debris coverage ratio. This parameter is equal to the total modelgenerated surface area of onground, thinshell material fragments divided by the total field ground area. A reasonably accurate model will have a C value roughly similar to that which could (theoretically) be computed for the actual site. We can safely assume that the Foster ranch debris field consisted of mostly open range, with individual pieces or clusters of debris scattered over it. Since there had to be sufficient thinshell material available to create at east one densely packed region as well as to define the perimeter of the vast area described by Marcel, a value of C between 0.01 and 0.10—i.e., between one to ten percent coverage—appears reasonable. The model’s debris coverage density is 6,880 ft^{2} ¸ 667,000 ft^{2} = 0.0103 (1.03% overall coverage)—at the lower end of the arbitrary acceptable range. Nearly 99% of the simulated field is debrisfree or uncovered, confirming the desired conservative nature of the model. Summing up. My mathematical analysis indicated better than a two orderofmagnitude disparity between the amount of thinshell material on the ground (by all accounts—a lot) and what Mogul Flight 4 could have supplied (very little). This enormous difference provides a comfortable margin for error. For example, even if the model’s surfacearea estimate is off (too high) by a factor of 10, it would still require nearly 13 balloon trains (via 688 ft^{2} ¸ 54 ft^{2}) or 39 targets to generate the shredded shell fragments found at the site. Still think that a single balloon train caused the Roswell debris? Gigo? Roswell detractors might say that my analysis is flawed because decadesold eyewitness testimony is factored into the debrisfield model. Because Major Marcel’s large fieldsize estimate constitutes a vital input to the model, it is a likely target for critics. They might claim that he grossly overestimated the field size, resulting in a classic case of GIGO—garbage input (a grievously poor fieldarea estimate) generating garbage output (the 127plusballoon train result). The merits of this charge can be checked out by doing a simple whatif analysis.
Notes: (1) N is the number of Flight 4 balloon trains needed to litter the field with radar target laminate. N = M/R, where R = 54 ft 2 is the total laminate surface area comprising three rawin radar targets. (2) P = ( R/M )100% is the percentage of onground, thinshell debris that a single Flight 4 (three radar targets) can account for. Whatif analysis for the Mogul Flight 4 hypothesis Assume that Major Marcel overestimated the length of the debris field. The effects of such an error on the model’s output and associated hypothesis evaluation outcome are examined by varying the length of its parabolashaped field. Calculations and commentary. Table 1 presents the results of the analysis for several field lengths ranging from Marcel’s 4,000foot estimate to 250 feet. The width of the field is held constant (at 250 feet) for the sake of simplicity. Figure 3 illustrates the relative sizes of the four field lengths considered.
Figure 3. Relative sizes of the parabolic debris fields examined in the whatif analysis. Each parabola is 250 feet wide at its far end. The smallest length listed in Table 1—250 feet —assumes that Marcel completely butchered his estimate, imagining the debris to be scattered over an area 16 times longer than it actually was. Even for this mindnumbing scenario—a pasture with less area than a football field—it would still take 10 Mogul Flight 4 launches to supply the thinshell material found on the ground. Alternatively, one Mogul train would account for only 10% of the debris. One final note. The Air Force claimed in its 1994 report that the debris was strewn over a circular area 200 yards (600 feet) in diameter. (A July 9, 1947, story in the Roswell Daily Record, attributed to an interview with Mac Brazel in the newspaper’s office, is the basis for that contention.) My model cannot simulate a circular debris field, so we can’t evaluate the Mogul hypothesis for this pattern using a modelgenerated M value in the methodology employed above. However, we can make a ballparktype estimate using information provided in Table 1. The table shows that a 2,000footlong parabola has an area of 333,000 ft^{2}—just 18% more than the 283,000 ft^{2} in a 200yarddiameter circle—and that it would take nearly 65 Flight 4 balloon trains to deposit debris over that narrow strip of land. The number of Mogul balloon trains required to deposit enough thinshell material to outline the perimeter of a 200yarddiameter area can be crudely approximated by scaling that number using the ground area ratio of the two field shapes. We get (283,000 ft^{2 }/ 333,000 ft^{2}) ´ 65 = 55 balloon trains—admittedly a very rough estimate. However, its doubledigit magnitude—and attendant large allowable margin for error—indicates that even the Air Force’s own explanation doesn’t fit the Mogul hypothesis. Reducedsize field coverage ratios. Using the terms defined in Table 1, the previously discussed field coverage ratio can be expressed as C = M / A. C corresponding to field lengths L = 2,000, 1,000, and 250 feet is equal to 0.0105, 0.0108, and 0.0129, respectively, indicating that the smaller fields are also about 99% debris free. The values of N and P listed in Table 1 are thus—at the very least—reasonable firstapproximation type estimates. Summary and conclusion A mathematical model idealized the debris field as a variablelength, parabolashaped region sparsely covered with fragments of an extraordinary thinshell material. It was assumed that Mogul Flight 4 created the debris field, leaving behind metalizedpaper, rawinradartarget remnants having a known total surface area. Modelpredicted and Mogul Flight 4supplied thinshell material surface areas were compared. One Mogul balloon train could account for only an extremely small fraction of the reported debris, even if Major Jesse Marcel had badly overestimated the field size. Clearly, Project Mogul Flight 4 could not have been responsible for the debris found on the Foster ranch. Indeed, the analysis illustrates in a most compelling fashion just how absurd the Air Force’s Mogul hypothesis really is. 
Editor's Note: Robert A. Galganski has a master’s degree in
civil engineering and is employed as a ground vehicle crash safety systems research and
development specialist. He received the Ufologist of the Year award at the 1997 National
UFO Conference in Springfield, Ohio, for his contributions to Roswell Incident research. Article
from the International UFO Reporter, Summer 1998, Volume 23, Number 2.
Copyright Ó
19972004
CUFOS. All rights reserved 