The Seven Site

Jump to: navigation, search
View from the south.
View from NNW to SSE.
View from the east.
The Seven Site lies on the crest and southwest slope of the surge ridge that forms the northeast side of the Nikumaroro atoll. It is about 1.1 km. from the southeast end of the island, and just west of the seven-shaped natural clearing from which it takes its name.


The site is densely vegetated in Scaevola frutescens (mao), with scattered (get list from Josh). A large Tournefortia (ren) tree stands on the ridge crest near the northeast edge of the site as we have defined it (Figure xx Contour map including elevation and lat/long; also work in KAP mosaic). This “big ren" – or its ancestor – can be identified in airphotos as early as 19xx. The Pisonia grandis (buka) forest begins just northwest of the site and continues sporadically to the north cape, interwoven with and bordered on the ocean and lagoon shores by dense thickets of mao. When we found the site in 1996, and cleared a portion of it in 2001, its most obvious surface features were a rectangular steel tank, 1 meter on a side, labeled “Tarawa Police” and containing six half-coconut shells, and a roughly circular excavated hole some 1.5 meters in diameter and 1 meter deep, with an evident though much eroded backdirt pile adjacent to the northwest.

Seven Site--KAP montage.


  • Dave Steadman report.
  • Get ’01 bones from SI and add to analysis.
  • Resolve cartridge problem.
  • Integrate revised Sharyn Jones report.
All units features corrugated plus e.jpg


We inspected the site in 1996, but undertook no further work there until discovery of the WPHC papers raised the possibility that it was where the 1940 bones discovery was made. We speculated that the hole might have resulted from burial and excavation of the cranium, and the tank might have been brought in as a rainwater catcher to support Gallagher’s detailed search. In 2001 we cleared a swath through the mao from the “Seven” to the hole, which we carefully excavated together with its backdirt. We mapped the vicinity and swept it with metal detectors. Results of the hole’s re-excavation were generally negative.

In the course of this work, however, we found two deposits of charcoal, ash, and burned animal bones on the ridge crest near the big ren. We interpreted these as the remains of cooking fires, and shifted the focus of investigation to their vicinity. We excavated six 1x1, 1x2, and 2x2 meter areas to explore these and other features, passing all excavated material through 6 mm.-mesh sieves. We also mapped all surface features and artifacts identified.

In 2007, with improved brush-cutting technology, we cleared a larger area, which we swept with metal detectors and mapped. We also scanned much of the site surface with a daylight ultraviolet light source in search of bones, which fluoresce in UV light. We further excavated the hole, together with two more fire features and a few locations where visual inspection or metal detecting suggested the presence of interesting material. Finally, we imaged the site and its vicinity using kite aerial photography (KAP).

Site Structure

The ridge on which the Seven Site lies is composed of coral rubble thrown up by wave action and stabilized by vegetation. There is little soil accumulation, though screening does produce a small amount of humic material. The surface is somewhat deflated, creating a loose “armor” or “pavement” of irregularly shaped coral pebbles and larger pieces of rubble. Immediately beneath this “armor” is a thin layer of somewhat smaller coral pieces; it is here that the humic soil and most of the cultural material is found. Below this level the chunks become larger again; at about 1 meter depth there is more or less consolidated coral with very large void spaces loosely filled with rubble.

All cultural material we found at the Seven Site was almost universally either on the surface or in the thin humic layer occupying roughly the top ten centimeters of the site.

Sequential Uses of the Site

Some historical and archaeological evidence, summarized briefly below, indicates that the Seven Site vicinity was used by personnel of the U.S. Coast Guard (USCG) during their sojourn on the island (1944-46). Other such evidence is clearly associated with use of the site by PISS colonists, before and/or after the USCG period. A third body of evidence suggests use of the site by the “castaway” whose bones were collected by Gallagher and his PISS colleagues in 1940. After briefly summarizing the relatively unambiguous evidence of USCG and PISS use, we will focus on the possible castaway associations.

U.S. Coast Guard-Related Material

Material obviously related to the 1944-46 U.S. Coast Guard loran station is ubiquitous on the site, though not in large quantities. The most common artifacts recovered were .30 caliber cartridges most plausibly associated with the M-1 carbines issued to the Coast Guardsmen. Veterans of Loran Unit 92 recall taking their carbines into the “bush” in the area for informal target practice. Fragments of ironware, some with the U.S. Coast Guard insignia, doubtless represent targets. Fragmentary bottles and radio parts found on the site are also probably associated with target use.

PISS-Related Material

The steel tank was almost certainly brought in by the PISS colonists, perhaps to support tree cutting or Gallagher’s search of the site for bones and artifacts in 1940, or perhaps as part of subsequent efforts to plant the area in coconuts. We recorded the remains of a corrugated steel roof adjacent to the tank, almost certainly representing a small structure whose roof was designed to drain rainwater into the tank. Although completely reduced to rust, this roofing can be identified as galvanized steel, and hence was probably Coast Guard material used by the colonists after 1946 (though the structure itself, and the tank, may be older).

On the surge ridge itself there are several more or less linear deposits of corrugated and sometimes flat iron sheets, reduced almost entirely to small flakes of rust. Unlike those near the tank, these are not steel, and were not galvanized. Toward the southeast edge of the site there are also a few somewhat heavier channel-shaped iron pieces. The purpose of this material on the site is somewhat mysterious, but we suspect that it was brought there in about 1940 to facilitate skidding large kanawa logs as part of the logging operation apparently carried out at that time (Note: Earlier discuss historical evidence of logging). A roll of asphalt construction siding, a portion of which was apparently spread on the ground at some point before the iron was laid down, is also probably of PISS origin.

Ferrous Metal Artifacts

Summary of rusty materials: Corrugated Iron, Corrugated Steel, Ferrous Artifacts In and Around the SL‐Fire Feature.

Possible Castaway Associations

Shellfish Features

We recorded three deposits of shellfish at the Seven Site – one made up of small bivalves representing genus Anadara, two of large bivalves of genus Tridacna. The Anadara feature comprises at least one hundred valves of the small shellfish, concentrated in an area about one meter across near the crest of the ridge (Figure xxx). It extends only about ten cm. into the rubbly ground, and is overlain by a thin layer of scattered fiberglass flecks derived from the gritty covering of the rolled asphalt siding. The asphalt siding, in turn, was apparently placed on the ground before the corrugated iron sheets were laid down; we found fragments of iron on top of the roll, but none under it. This suggests that the Anadara feature predates use of the site for whatever activities brought the siding and the iron there.

About two meters south of the Anadara feature we recorded a loose cluster of valves representing the “giant” clam genus Tridacna, most probably T. crocea. At least seventeen clams are represented by twenty-nine complete and fragmented valves. In most cases one valve of each pair is complete, while the other is often broken or even smashed into multiple fragments. Several examples show evidence of battering and/or prying around the byssal orifaces and on the siphon end. In one case (Fig. xxx), the tip of a small iron tool, apparently fabricated from the rim of a steel drum and found in metal detecting about three meters from the shell cluster, fits precisely into a chipped wound in the clam’s hinge.

The second Tridacna feature (Figure xxx) is a more linear concentration of shells, aligned roughly north-south. It lay about 10 meters from the first feature and just upslope (east) of the “SL” fire feature (see below). This cluster contains 32 valves representing 16 clams, plus six valves representing three clams scattered within a few meters downslope to the south and east. All are complete, with no sign of chipping, prying, or battering.


The average weight of a live Tridacna about the size of those on Nikumaroro is calculated to be 360 g, yielding about 39 g of edible meat . Each of the Tridacna clusters therefore, when fresh, should have weighed between six and seven kilograms – a modest load for one individual to carry from the lagoon southwest of the Seven Site or the reef northeast of it, and an easy burden for two. Tridacna can be observed living in both areas today. Each cluster would have produced 650-700 g of edible meat – a solid meal for one or two individuals (See Fun With Clams for basis of clam and meat weight calculations).

Archaeological and ethnographic data suggest that indigenous Pacific islanders did not and do not often collect Tridacna and bring them to an interior site to be processed, though they did do so in pre-contact times when Tridacna valves were used in the production of adzes and other implements. The usual practice, unless one has some need for a shell, is to cut the live clam’s adductor muscle and then extract the meat, leaving the valves at the clam bed. xxx Munro, for example, says

“Fishing methods for giant clams are exceedingly simple. In remote areas where the shells have no significant value and have strong byssal attachments to the reef, the flesh is simply excised from the shells by slipping a sharp knife along the inner surface of the shell to cut one end of the adductor muscle. This also applies to the larger species in which the shell is too heavy to be readily lifted from the water” .

While the Seven Site shells are not too heavy to be readily lifted from the water, they would have no particular value to the Nikumaroro colonists, and if they had been of any value one would expect them to have been taken to the village. The impression that the Seven Site Tridacna were harvested by someone not indigenous to the islands is heightened by the way the clams in the first cluster have been opened. It appears that someone first tried to pry them open on the hinge and/or siphon ends, and when this failed, simply smashed one valve of each animal with a rock. The linear arrangement of the second clam feature, and its proximity to the fire feature, suggest that the clams were placed around the edge of the fire, where the heat caused them to open, leaving no evidence of prying or bashing. This may indicate that the clam harvester(s) learned from experience with the first group of clams that a more patient technique was in order, but neither technique is characteristic of indigenous Pacific Islander Tridacna consumption.

If the Seven Site clams were not harvested by indigenous people (the PISS colonists), the remaining suspects are the Coast Guardsmen and the castaway(s). Coast Guard veterans of the loran station, while some recall using the Seven Site area as a place to shoot their guns, deny ever collecting and consuming shellfish there. This does not mean that one or more Coast Guardsmen did not at some point have a clambake on the site, but the condition of the clams in the first feature argues against the Coast Guardsmen as their tormenters. The Coast Guardsmen were equipped with heavy, serviceable knives, and would hardly have needed to fabricate a prying tool and chip away at the clam’s hinges. We think it most likely that the Tridacna features, and perhaps the Anadara feature, represent the subsistence activities of the castaway(s). It is notable that on a 1937 airphoto of the Seven Site vicinity (cite), faint lines can be made out that appear to be trails running between the site and a cove on the lagoon shore where today there is a bed of (mostly dead) Tridacna.

Fire Features

We excavated four features that we interpret as the remains of cooking fires were excavated: two in 2001, two in 2007 . One (Feature D) is on the ridge crest near the big ren tree; another (Feature M) is some fifteen meters to the west along the ridge crest. A third (Feature WR) is on the southwest slope of the ridge about ten meters south of the big ren, and the last (Feature SL) is about 24 meters southeast of the ren, on a southwest-tending spur of the ridge.

The features are difficult to see until the surface armor has been cleared away, so there may be more that we have not yet found. Each feature consists of burned coral, wood ash, some charcoal, and the burned and unburned bones of fish, birds, and sometimes sea turtles, together with burned and unburned artifacts of various kinds. None extends to a depth of more than 10-15 cm. Their horizontal boundaries are indistinct, but they all appear to range between one and two meters in diameter, and to be roughly circular to ovoid.

Feature D

Feature D lies on the ridge crest about five meters northeast of the big ren tree, shaded by the edge of its canopy. [ INSERT DISCUSSION OF FISH BONES, ARTIFACTS, BIRD BONES].

Feature M

Feature M, some fifteen meters from Feature D, is visually similar to the first feature and exhibited a paucity of artifacts. [ INSERT DISCUSSION OF BONES, ETC.]. A single radiocarbon sample taken from this feature yielded a “recent” age.

Feature SL

This fire feature adjoined the second of the Tridacna features, whose valves were relatively undamaged and which was organized in a linear fashion. It contained abundant bird bones and relatively fewer fish bones. [ INSERT DISCUSSION OF BONES].

Artifacts were numerous and puzzling in Feature SL, including:

  • Gidgies
  • Screw
  • Pin
  • Gutta-percha washers
  • Black tube
  • Rouge
  • Lubricant top
  • 40-cm. ferrous thing
  • Other ferrous
Feature WR

This feature, somewhat downslope to the west from the ridge crest, was much more complex than the first two features and contained more fish bone than any other.

Feature WR also contained a diverse collection of artifacts, including:

Cartridges Snap Bottles (Ric discuss) Wire Ferrous Rouge

Feature WR also produced a fragmentary cervical vertebra from a medium mammal identified as probably either sheep or goat. The bone was not burned, but showed cut marks suggesting that the animal from which it came had been butchered. We interpret this bone as most likely originating (locally) in a can of mutton stew.


The small number of fish represented by the remains in each fire feature suggests that each represents a single incident, with the likely exception of Feature WR, which may reflect several fish-cooking episodes. Feature SL appears to reflect an effort to cook birds.



Was the Seven Site the Site of the Bones Discovery?

Documentary evidence

From Navy PBYs, June 1941
Gallagher to Vaskess, 17 October 1940:
"Bones were found on South East corner of island about 100 feet above ordinary high water springs."
Gallagher to WPHC, 27 December 27 1940 transmittal letter:
"The larger of these packages is the coffin containing the remains of the unidentified individual found on the South Eastern shore of Gardner Island; the second package is the sextant box found in the immediate locality and contains all the other pieces of evidence which were found in the proximity of the body. ..."
"It is possible that something may come to hand during the course of the next few months when the area in question will be again thoroughly examined during the course of planting operations, which will involve a certain amount of digging in the vicinity"
We have photographic evidence of operations at the Seven Site about 100 feet above ordinary high water springs six months later in June 1941. The same photo shows no activity at the south east tip.
  • Reservation of land for Gallagher or a subsequent komitina (commissioner).

Photographic evidence

  • Evidence of logging
  • Evidence of clearing
  • Evidence of trails

Archaeolgical evidence

  • Fire features
  • Hole
  • Artifacts

Was the castaway Amelia Earhart?

Some materials collected at the Seven Site are consistent with this theory:

Why might the castaway have been attracted to the Seven Site?

GoogleEarth Niku Image.jpg

Arguments about why Earhart and/or Noonan would or should have moved from the best landing site near the Norwich City are essentially moot (arguable, undecidable, essentially irrelevant). IF that is where the castaway died, and if the castaway was, in fact, Earhart or Noonan, then we have to presume that he or she thought it was a good place to camp either temporarily or indefinitely. It is possible that the castaway arrived at the Seven Site after scouting what looks like, but is not, a freshwater pond near the southeast tip of the island.

  • Pleasant forest.
  • Breezes from prevailing easterly winds.
  • Easy access to both the sea shore and the lagoon.
  • Trees to climb to watch for ships.
  • Access to food: turtles, fish, clams, birds.
  • A ready supply of firewood in the buka forest?

The Kilts Account: Discussion and Speculation

By Tom King

Shortly before Peter MacQuarrie found Gerald Gallagher's correspondence about the bones discovery, we had a couple of highly qualified analysts of oral tradition look at the Kilts account, asking their advice about what in it had, to them, the ring of truth. Both said, in essence, that there was no particular reason to believe any of it.

With Peter's discovery in the Kiribati National Archives, however, we learned that at its core, the Kilts account was true. A skeleton was found, near one end of the island, by a party of islanders, with a cognac bottle and a woman's shoe, and though Gallagher wasn't an "Irish magistrate," didn't hop in the island's 4-oared boat to row the bones to Fiji, and didn't die en route, he was Irish, was in charge of the island, must have sent the bones in the 4-oared whaleboat Nei Manganibuka out to the schooner Nimanoa for the voyage to Fiji, and did die shortly after coming back from Fiji. It's not hard to imagine a conversation among Kilts, a colonist without much English, and an interpreter scrambling the facts of the bones story into the tale that Kilts wound up telling.

The one piece of the Kilts story that doesn't make sense as a loose interpretation of what we know actually happened is the business about the superstitious natives throwing the bones into the sea. Where does that come from? Here's my hypothesis.

I think there were two bones discoveries. I think that after Gallagher left for Fiji, the colonists found a second set of bones, probably on Nutiran. I think they held them, probably in a bag, for Gallagher's return. When he returned and promptly died, I think they decided that messing with bones was not a good idea, and disposed of them by taking them well away from the island and committing them to the deep.

This is not entirely a wild-assed guess -- just mostly one. Consider:

1. When Dirk Ballendorf interviewed residents of Nikumaroro Village in the Solomons (Provide link to Dirk's report), they told him that two skeletons had been found -- one at the SE end of the island, the other at the NW (Nutiran) end.
2. When Barb Norris, Kris Tague and I first interviewed Emily Sikuli, she said that the bones she remembered -- which she associated with the box her father had built -- had been found on Nutiran, "where the plane came down." When Ric and others interviewed her on video later, her story was rather garbled, suggesting to me that she'd been thinking about it and wasn't as sure of her facts as she'd been the first time around.

So there's some real indication of a second skeleton, found on Nutiran. There's no independent indication of its disposal at sea, however; that's found only in the Kilts account.

If there was a second skeleton, whose was it? If the skeleton at the SE end was Earhart's, then one on Nutiran, if it existed, might have been Noonan's. Noonan's death and burial on Nutiran would be consistent with our interpretation of the radio messages indicating that he was injured. However, the situation is complicated by the fact that we know that three bodies of Norwich City crewmen were buried on Nutiran, and others might have washed ashore and gotten buried by natural forces.

At this point, there's probably no way of finding out whether a second skeleton was found, and if it was, what happened to it. But it's one way of making sense of the only part of Floyd Kilts' account that doesn't otherwise bear any relationship to what's reliably documented as having happened.

Related Material