Discovery and recovery
The U.S. Army compiled hundreds of pages of reports and documentation describing the recovery of the downed B-17-F in the mountains of Papua New Guinea in 1992. Below is an Army report summarizing their findings.
DEPARTMENT OF THF ARMY
United States Army Central Identification Laboratory
310 Worchester Avenue
Hickham AFB, Hawaii 96853-5503
| Background | Location and area | Field methods | Findings | Conclusions |
TAPC-PED-1-1 (600-8-1 im) 26 October 1994
MEMORANDUM FOR Commander, U.S. Army Central Identification Laboratory, Hawaii,
310 Worchestcr Avenue, Hickam Air Force Base, HI 96853-5530
SUBJECT: Search and Recovery Report 113/CIL/94, Archaeological Excavation of a B-17F Aircraft Crash Site (MACR #648), Hidden Valley), Upper Watut River, Morobe Province, Papua New Guinea, 4 August - 3 September 1993.
A. From a base in Port Moresby, Papua New Guinea (PNG), a B-17F (#41-24552) aircraft, in a formation with five other B-17's from the 65th Bomb Squadron, 43rd Bomb Group, attacked and bombed Lae (PNG) on September 15, 1943 between 10:20 and 10:55 A.M. Afterwards, the formation started on a return flight to Port Moresby. While approaching the city of Wau (PNG) and just prior to entering a thunderhead, B-17F #41-24552 was observed peeling off steeply from the formation and was never seen again. An extensive search the same day and the following day failed to locate any sign of the plane. All eleven crew members were eventually listed as Missing In Action/Body Not Recovered (MIAIBNR). In October of 1992, Mr. David Beebe located the probable crash site of B-17F #41-24552, and observed human remains across the area. He collected a bracelet with the name Howard (Eberly) inscribed on it which was eventually forwarded to Washington, D.C. Thereafter, the commander of the Central Identification Laboratory, Hawaii was notified and plans were made to survey and excavate the crash site.
B. From August 4 to September 3 1993 a joint United States Arm), and Papua New Guinea (PNG) team conducted a systematic archaeological excavation of a B- 17F crash site in the dense, highland montane forest (elevation 8300 feet) of Papua New Guinea. The great number of human remains and life-support gear recovered, plus evidence of unused escape equipment, suggested all eleven crewmembers were in the plane when it crashed. The site was excavated, photographed, and mapped. All human remains and personal effects and gear were returned to the Central Identification Laboratory, Hawaii, for final analysis. No further work is recommended at the site.
A. Hidden Valley, Upper Watut River drainage, Morobe Province, Papua New Guinea (Enclosure 1).
B. Grid Coordinate: 55M DG 5954/7648, Geographical Coordinates (long/lat) 07 degrees 27' 44"/146' 37' 82".
C. Map: Wau (PNG), Sheet: 8283 (Ed-1); Series: T 601; Scale: 1:100,000, printed in 1977, (Enclosure 2).
3. DESCRIPTION OF PROJECT AREA: (see images)
A. The crash site was located in "Hidden Valley" along the Upper Watut River drainage at 8300 feet above sea level. The wreckage was scattered alone the southern exposure of this high, narrow, steepsided valley, which ranged from 36' to 40' slope and represented a major work hazard throughout the project. Three wide, seasonal stream gullies run down the hillside and provide an undulating surface across the entire site area . The steep hill topography was broken only by small, flat terraces just above the Watut stream where the site camp vas constructed.
B. The local weather conditions were variable. Temperature ranged from 35 degrees Fahrenheit at night to 65 degrees at middav. The region was humid and damp with varying amounts of rainfall about every other day. Daily high winds made cloud cover and rain patterns very unpredictable. Their rapid movement into the valley often dictated helicopter travel to and from the site, and therefore daily/weekly work schedules. Several minor earthquake tremors added another unusual dimension to the local environmental conditions in the project area.
C. The region was covered with dense, plush temperate montane rain forest, dominated by a high canopy of straight, tall trees and a thick ground vegetation consisting of an upper blanket of lush, light-green moss (average: 10cm thick) overlying a thicker layer of soft, black humus or peat (average: 15cm thick). The humus consisted of thick, moist, decayed plant debris which represented a serious obstacle to team members when walking because of unexpected deep holes and obscured roots below ground. Dispersed throughout this understory were dense clusters of moist ferns, hazardous creeping vines and occasionally, low-lying thorn bushes.
D. The two most common trees in the area were Nothfagus spp. (montane beech) and Podocarpus sp. Most trees were mature, averaging 112cm in diameter, including many fallen logs. Though saplings (average: 56cm in diameter) were less common overall, more were observed in the project area. This was probably due to the destruction of larger trees during the crash. Numerous sapling trees growing in and around wreckage were ample evidence of forest growth since the crash (average diameter 38cm), Deep trunk scars exhibited on a few large trees probably resulted from the plane impact or embedded crash debris.
E. As noted, the upper layer at the site was a thick carpet of live moss over humus. Below this, was a fine, light brownish yellow clay (Munsell soil color chart #10YR 6/6). This nonporous, moist, highly organic soil was usually 10 to 30cm deep. However, on the western side of the site, it was much deeper in certain areas (1.6m) and was a slightly darker yellowish brown color (10YR 5/6). Many fragments of wreckage were embedded in the upper clay, but onlv a few large pieces were recovered from deeper levels. The lowest layer at the site was a highly weathered bedrock of Granodiorite characterized bv a whitish-yellow, very, loose, porous, coarse-grained, micaceous sand. This material ran beneath the entire project area, and its thickness varied front 60cm to several meters in exposed areas.
A. Initially, a helicopter landing zone and site camp were established some distance from the crash site, A walking survey was conducted to identify the distribution of wreckage and site boundaries. The site was extensively photographed prior to any archaeological activities.
B. Once the site boundaries were determined an excavation grid system of contiguous 8 x 8 meter units was established to facilitate the controlled, recovery of crash artifacts. A Sokkia DT6 Theodolite was used to plot the grid system. The initial grid measured 64 meters east to west and 48 meters north to south, but was amended several times over the course of the project to incorporate undiscovered wreckage or delete areas devoid of crash debris. The grid was subdivided in 2 x 2 meter units when concentrations of human remains were encountered, allowing more detailed recording of the data.
C. Work at the site was exceedingly hazardous due to the very steep slope, slippery ground vegetation, deep holes in the moss layer, unseen obstacles such as roots and vines, and sharp, pointed pieces of wreckage. To alleviate potential hazards, a series of safety measures were employed to protect the entire project team. Extensive wooden walkways were constructed along the hillside for safe walking across the site and to/from the lending zone. An efficient series of guide ropes were tied up the hillside to ensure safety and down travel. Another set of rope lines were erected to eliminate the dangerous transport of heavy buckets up/down the steep hill to and from the screens .
D. Initially, each grid unit was cleared of bulk vegetation prior to excavation, but crash artifacts were carefully left in place during this process. All units were excavated with trowels or hand rakes. Soil and plants were placed in metal buckets and transported by rope line and pulley to the screens. Though excavators and screeners often worked far apart, they tracked soil and artifacts from each unit, especially, when five or six units were excavated at once. To control the movement of soil from multiple units to the screens, bucket handlers coordinated this activity which expedited the excavation process. The site was excavated generally east to west - from areas of lesser artifact concentrations to greater concentrations.
E. As units were excavated, all crash artifacts were left in place to map their distribution across the site. Prior to the removal of artifacts, each completed unit was photographed. Screens consisted of rectangular wood box-frames covered on the bottom with 1/4 inch hardware cloth. They were tied to a four-legged wood frame creating a very efficient swinging screen. Soil and plant material were placed in the screen and carefully sorted for artifacts. Remains were placed in separate bags for later processing and analysis.
F. The site was excavated by natural stratigraphic layers The thick moss and humus layers were removed followed by the upper clay down to sterile soil.
G. A temporary lab was set-up at the base camp to sort, clean, analyze and package artifacts from the field. The work was conducted during the evening or on rain days all materials were stored in locker boxes and eventually returned to the Central Identification Laboratory, Hawaii, for final analysis.
A. The crash wreckage was distributed over a fairly compact area of the valley hillside measuring 76 meters east to west and 72 meters north to south. The total area excavated covered 424 square meters. The evidence indicated the plane disintegrated as it hit the high canopy, trees, and then the ground at high speed, and continued plowing through the dense montane forest. Though large trees accounted for most of the destruction, they simultaneously acted as a barrier to limit the scatter of plane wreckage over a wide area. Also, the steep terrain limited wreckage scatter, as large pieces (engines, propellers) were found embedded in the hillside. In addition, there was little evidence of a major impact explosion which would have spread debris over a much wider area.
B. Nearly all the wreckage was distributed across the steep hillside, except engine #4 (outboard right-wing) and the outer right wing. Discovered 150 meters northeast of the main crash site, the engine rested in the bed of the Upper Watut River. The wreckage distribution shows the plane was traveling approximately east to west when it hit the hillside. One large piece of vertical tail section was found which included the plane's identification number.
C. Apparently, after the initial survey of the site and prior to the excavations the wreckage was vandalized by unknown individuals. The PNG team officials indicated that a few local people scavenged the site looking for WWII memorabilia such as usable weapons, identification tags, patches, hats, and related items. Despite this, the original location of most wreckage was reconstructed across the site for fairlv accurate mapping. Unfortunately, some crash artifacts were removed from the project area. A number of human remains were displaced from their original locations, but were eventually collected and returned for laboratory analysis.
D. Most wreckage was recovered from the moss/humus layer and the first few centimeters of the clay soil. To ensure artifacts were not overlooked, each unit was excavated at least 10cm into sterile soil. As noted elsewhere, mostly, large heavy, pieces of wreckage such as engines and propellers, were embedded in the clay or sand as a result of initial impact. Exceptions were the embedded .50 cal. machine gun and human remains near the rear of the plane.
| D. Most wreckage was recovered from the moss/humus layer and the first few centimeters of the clay soil. To ensure artifacts were not overlooked, each unit was excavated at least 10cm into sterile soil. As noted elsewhere, mostly, large heavy, pieces of wreckage such as engines and propellers, were embedded in the clay or sand as a result of initial impact. Exceptions were the embedded .50 cal. machine gun and human remains near the rear of the plane. |  |
| .50 cal. machine gun |
Once exposed by excavation, all artifacts were left in place for mapping. Then significant items such as human remains, personal effects and gear, and life support equipment were removed for further analysis.
E. The bulk of the crash wreckage was metal fragments from the plane's super structure. Much of this material was ripped apart and scattered in small chunks (1.0-2.5m. sq. in size) across the site. The most recognizable pieces consisted of wing sections, engines, landing gear, propellers, tail wheel, and tail sections. Fuselage fragments were common but shredded in small chunks. Of great interest, were numerous fragments of fuselage "nose art," including the plane's name "Listen Here, Tojo."
 |
 |
| Prop at crash site |
Nose art of downed plane |
The distributional evidence suggests the plane was traveling east to west/northwest as it broke up while plowing through the dense forest. Scattered wing debris occurred everywhere, but was much less common on the western side, where most wreckage was from the rear of the plane. The front landing gear was uncovered in the far eastern (left wheel) and upper central (right wheel) areas. The left wheel, with shaft and axle, was well-preserved and included the entire rubber tire. It was isolated from other wreckage and was probably ripped off during the crash. The right gear was uncovered with engine 2 and exhibited the only evidence of a sustained fire which consumed the tire, fuel tanks and wreckage in a 4 x 4 meter area. Evidence of fire elsewhere was sporadic. The tail wheel was recorded in the western section and included a well-preserved tire.
 |
 |
| Tire at crash site | Debris in grid |
Wreckage from the nose, cockpit, turrets, waist and tail section were distributed east to northwest across the site. Scattered amongst this wreckage were thousands of small plane parts from the superstructure including gauges, panels, controls, instruments, engine parts, tubing, wires, window sections, turret pieces, and flooring. Scattered among the plane fragments were crew-related artifacts including human remains, personal effects and clothing, life-support/survival gear, field equipment and weapons.
F. The majority, of human remains were concentrated in three areas of the site. Area 1, associated with cockpit wreckage including throttle controls, yielded dense remains over a 16 meter square area. It was subdivided in 2 x 2m units for more controlled excavation. Here, elements were mostly disarticulated and broken, but represented portions of several skeletons and included fragments of skull, mandibles, vertebrae, shoulder bones, pelvis, limb bones and extremities (hands and feet). Parachute cord was entwined with many of these elements and other associated gear included life vest fragments, .45 cal. pistol, boots/shoes and other leather pieces. Interestingly enough, a set of officer flight wings, the co-pilot's (Townsend) I.D. bracelet and the pilot's (Eberly) wedding ring we re recovered from this unit (Enclosure 4, #3 and #4). The inscription inside the wedding ring reads, "From Billy Jean to Howard" (Eberly). The association of cockpit debris and pilot/co-pilot belongings leaves little doubt that they died in this vicinity. A few meters to the southeast, excavations uncovered the bombardier's ring, but it was not associated with many human remains (Enclosure 4-1 #I). Area 2 was centrally, located on the site and associated with wreckage from the navigator and upper turret sections (Enclosure 4). All remains were disarticulated and scattered around and beneath several large fallen trees. Common elements consisted of skull, limb bone, vertebrae and rib fragments and most of these, except skull pieces, exhibited gnaw marks. The remains covered a 6 meter square area with densities equivalent to those of Area 1. The assemblage was associated with leather goods (boots/shoes), life-support gear (life jacket pieces, first aid kit, parachute), field equipment (canteens, field blanket), pistol with holster, and personal effects. One pocket watch was recovered and an I.D. bracelet with the inscription "Louie." The third concentration of human remains (Area 3) was located on the western side of the site, associated with wreckage of several rear plane sections including the radio area, waist gunner positions, ball (belly) turret, tail, and tail gunner position. Most of the remains were scattered, disarticulated, and gnawed. Due to the concentration of skeletal remains and associated wreckage, the area was subdivided into 2 x 2 meter units for a more controlled recovery. Fragments of skulls were recorded in this area, plus limb, vertebrae, pelvis, rib and extremity elements. Several skull pieces were associated with human hair. The short, fine, straight, soft, brown hair was collected from several disturbed areas and bagged for further analysis. (See Identification for anthropological analysis. See DNA evidence for scientific confirmation.
G. A variety of personal effects were recovered from the crash site, including I.D. tags, rings, bracelets, coins, wallet, eyeglasses and watches. Most of these were associated with the three concentrations of human remains on the site. An unknown number of these artifacts were removed by vandals which was substantiated when several such items were returned to the team leader by the Wau (PNG) police. Included was the pilot's I.D. tag, a silver cross, and a .45 cal. pistol. As mentioned elsewhere, two name bracelets (Townsend and Louie), an I.D. Lag (Watson Hall), the pilot's wedding wing (Eberly), a bombardier's ring with the inscription "C.E.R." inside, and an empty photo wallet were recovered at the site. One metal flight wing insignia was discovered with the pilot's ring. Coins recovered included a 1941 dime, and an Australian pence. Several watches were recovered from the central and western sections.
 |
An unusual group of artifacts uncovered in the eastern part of the site (Feature 1), near skeletal Area I included the remains of an equipment bag. Outlined by leather seams, a zipper and handle, it contained a pair of metal-rimmed goggles, stopwatch, penlight, batteries, small lock and pencils. Several pairs of eyeglasses were excavated, but were not associated directly with human remains. All had metal rims in hard black cases without inscriptions. H. One of the most common artifacts at the site was leather goods, especially, footwear and jacket fragments.
 |
 |
 |
| .45 calibre pistol | Leather boot | Cross |
Generally, these items were well preserved due to the high altitude, but gnawing on footwear was common. Leather goods were scattered across the site and many were associated with skeletal remains but very few were recorded above the 24 meter line-- to east to west. Not Surprising, this is similar to the distribution of human remains. Leather clothing, consisting mostly of jacket fragments, was common, especially in the western half of the site. All specimens appeared to be ripped or torn and were scattered on the surface. Though many pieces were associated with the three skeletal areas, they were from disturbed contexts. Curiously, not a single insignia or identification patch was recovered. These disturbances are most likely linked to vandalism, since insignias and unit identification patches, etc., are highly prized WWII memorabilia.
1. Footwear distribution was similar to jacket fragments, except they were common in each of the three skeletal areas. Several types of footwear were encountered and included high and low top boots with thick rubber soles and heels, and shoes. Overall, nineteen boots/shoes were recorded, as well as a host of fragments, heels and soles. Many of these items exhibited gnaw marks, especially, around toe/ankle leather. Some specimens exhibited Australian manufacturing labels and were probably not standard issue footwear.
J. Life support gear was one of the largest and most varied artifact groups encountered at the site. It included fragments of parachutes, life vests, oxygen/gas masks, first-aid and survival kits, flair guns and life rafts. Parachute remains were very common across the site. Three nearly complete specimens were uncovered beneath nose and upper turret wreckage in the central area. The parachute cord fragments and release rings were scattered everywhere and occasionally, cords entwined human remains. The abundance of parachute remains represented more evidence that the crew crashed with the plane. Oxygen masks were encountered in several areas and were surprisingly, well-preserved, especially the rubber face shields and eye goggles. Only, two gas masks were recorded, both from the western tail section wreckage and both were well-preserved. Two other important items included first aid and survival kits. Found in several areas, the small, flat, rectangular metal first aid kits contained bandages, gauze and small vials of medicine. Only one was nearly complete, but two others were empty. Metal survival kits included fish hooks, a small pen light, matches and other unrecognizable materials. Fragments of these kits were uncovered in several areas of the site. Another item often found with human remains as well as elsewhere was the life vest. Though partly decomposed, yellow rubber vest fragments were easily recognized against the dark humus all across the site. A few were directly associated with human remains from Area 1. Also, fragments of two life rafts and several oars were excavated in the central and western areas. One was located next to the vertical tail stabilizer. Two badly damaged flair guns were removed from tile western area. Wau (PNG) police reported confiscating another one, in working condition, from a local villager who apparently, removed it from the site. Unfortunately, it was never returned. It was obvious from the sheer volume of life support gear recorded at the crash site that crew members perished with the plane. Much of this gear was in excellent condition because it was buried beneath layers of wreckage.
K. Identified weapons were the .45 cal. pistol with holster, .50 cal. machine gun and .30 cal. machine gun. These weapons were in surprisingly good condition considering the severity of the crash. Of the eleven standard issue .50 cal. guns aboard, 10 were recovered during the excavations.. Four .50's were identified in the central area associated with upper turret, nose and radio section wreckage. Six were uncovered near wreckage from the waist section, ball turret and tail gunner sections. The most common single artifact at the site was the .50 caliber round. Thousands were scattered across the site and many full .50 cal, ammunition cans were uncovered in several areas. In contrast, .30 cal. rounds were restricted to the central area where the only two .30 cal. machine guns were found. Two .45 cal. pistols and numerous rounds were recovered from the central part of the site. Neither was associated with human remains. One specimen was still in its holster and exhibited Custom made handles of bone or ivory. Two other .45's, apparently from the plane, were reported by the Wau (PN6) police but only one was returned (Photo 12).
6. CONCLUSIONS AND RECOMMENDATIONS:
A. Extensive archaeological excavations revealed the crash site of B-I7F #41-24552 (MACR Case #648) in the dense, highland, montane forest (8300' ASG) along the Upper Watut River drainage of Papua New Guinea. Traveling generally east to west at high speed, the plane crashed through the upper canopy, hit the ground and disintegrated as it continued plowing through the forest.
B. The shredded condition of the plane, the great quantity of scattered, disarticulated human remains and life-support gear, indicate the crew crashed with the plane. Recent vandalism prior to tile excavation caused irreparable damage to the site by unearthing human remains and associated personal effects and gear, displacing wreckage and removing personal effects and gear (I.D. tags, weapons, etc.). Evidence suggested some clothing (leather goods) and human remains had been associated and preserved since the crash but disturbed by vandals..
C. All areas adjacent to the site were surveyed. The site was completely excavated to sterile soil and all human remains, personal effects and gear were returned to the Central Identification Laboratory, Hawaii, for final analysis.
D. Considering the evidence presented above, no further archaeological work is recommended for this site.
8 Encls
1. General Location of Project Area
2. Topographic Location of Project Area
3. Site Map1/Wreckage Distribution | Map2
4. Distribution Map1 of Human Remains/Personal Effects | Map2
5. Photographs
Team Anthropologist
Copyright 2000, Peter V. Owens
Last revised, Feb. 22, 2000
