Trindade: The Tracer hypothesis

Martin Shough

Back to Trindade index


May 1957 saw the first prototype flight of the Grumman WF2, a US carrier-borne, folding-wing AEW (Airborne Early Warning) plane with a huge ellipsoidal dorsal radome containing the new Hazeltine AN/APS-82 radar (Fig.1). The WF2, later known as the E1-B Tracer, was a direct development of the S2-F Tracker which was used in an ASW (Anti-Submarine Warfare) role. The AN/APS-82 was the first such AEW radar to go into service and had many advanced features including a stabilized antenna and Airborne Moving Target Indicator (AMTI).

The Tracer was operational officially from February 1958, but we suppose there was a pre-operational deployment in January 1958 from a carrier in the South Atlantic, possibly connected with the sale of Tracker ASW aircraft to Brazil at this time. The Tracer would be a strikingly unfamiliar sight to those on board the ship, and especially if the plane was slightly fogged up the appearance could be startling.


Fig 1. The Grumman WF2/E1-B 'Tracer'

The Tracer's was not the first dorsal radome. The first saucer-shaped dorsal installation was the rotating rotodome of the Lockheed WV-2E in 1956. But only a single prototype was flown and its dome was much smaller in relation to the plane. The large 4-engine Constellation airframe of the WV-2E was also land-based, whereas the Tracer could have flown from a carrier in the South Atlantic.


As early as 1959 Donald Menzel proposed that Barauna's photographs showed an aircraft rendered barely recognisable by being 'fogged up' with condensation. In 1999 Martin Powell pursued this idea and discerned what he believed to be blurred images of a twin piston-engined aircraft. He narrowed down a shortlist of aircraft capable of flying from the Brazilian mainland to Trindade and back (there was no airstrip on the island), and selected from this list a Beechcraft Bonanza as prime suspect. He offered an intriguing feature-for-feature comparison between subtle details in P1 and parts of the Bonanza airframe viewed from slightly below and to the right of the nose (see Fig.2).

However the motiveless presence of a twin Bonanza some 700 miles out over the Atlantic - at the very extreme limit of its range and therefore seriously vulnerable to fuel shortage in the event of any adverse weather or engine problems - was a mystery even to Powell. He therefore went on to suggest that Barauna later superimposed doctored images of a Bonanza, taken elsewhere, onto his Trindade photos in the darkroom, noting that he had plenty of time to fake a set of second-generation negatives from this montage before the Navy got hold of them for analysis.

The main disadvantage of Powell's scenario is that it ends up sacrificing what, at first sight, was its most attractive feature - the presence of a real "UFO" in the sky over Trindade. Compared to the technical and circumstantial difficulties incurred by theories of premeditated double-exposure or montage after the fact, or compared to the circumstantial difficulties even of an internal mask hoax in real time, this would have been a very great simplification. Can we rescue this advantage in another form?

Fig.2. Martin Powell's Bonanza theory (top left), compared with (bottom left) a Tracer leaving a carrier deck, wheels down. Above: the same image with wheels erased, diffused and with added noise. Below: with simulated motion blur

The sort of aircraft that might have comfortably made the trip from the Brazilian mainland to Trindade and back would be a specialised maritime reconnaissance aircraft with a redundancy of engines, or with the special gear needed to keep the engines lubricated during the long haul, like the P2 Neptune. However, even if we were able to defend the unlikely one-off presence of a land-based marine patrol aircraft on Jan 16 1958, this would still leave unexplained the fact that a Navy investigation into UFO sightings at Trindade had been ongoing since the report, a fortnight before, of a disc with dorsal and ventral bulges, seen by 6 witnesses including a Navy Medical Officer, other Navy personnel and civilians at different locations on the island. This was reported by radiogram to the Navy High Command on Dec 31 1957, and the Navy Intelligence report forwarded to the Minister on March 3 1958 specifically notes that

"the observers who sighted the UAOs know perfectly how to identify airplanes; all planes over the Island have been properly identified in all cases, with communications being reported to the Hydrography and Navigation Department"

If the object seen on Dec 31 was another ordinary aircraft somehow unknown to the Navy and thus not "properly identified", is it likely that witnesses would independently have described the same very distinctive illusory appearance later photographed from the deck of the Almirante Saldanha on Jan 16? The sort of fogging invoked by Menzel is rare and very sensitive to conditions. It is arguably unlikely even once in the low-level tropical air over Trindade, and extremely unlikely to have simulated a "Saturn-shaped" UFO on two separate occasions, once in conditions where the object was triangulated by the sightlines from different remote observers.

Hence the major advantages of the Tracer hypothesis are the inherently saucer-like configuration and the fact that the Tracer would be carrier-borne. If the carrier was stationed nearby but over the horizon on some extended exercise then the operating range for an aircraft ceases to be an issue, and repeated flypasts over a number of days could be explained.


The Tracer's twin turboprops would be noisy, but according to all reports there was no sound whatsoever. However the object passed NW of the ship and the surface wind was from the SE. We could suppose that engine noise might have been blown away by the wind, perhaps further masked by uproar on the deck and the surf noise that Barauna said could be heard from the island.

The Tracer could not achieve the high-subsonic speed estimated by the Navy reconstruction, but witnesses might have overestimated the angular rate of the 'UFO' drastically because of the excitement. This might be consistent with meteorological and photographic evidence that the photos probably span two minutes or so rather than the 30 seconds estimated.

This theory resurrects the possibility that a genuinely odd event occurred at Trindade and that Barauna took six pictures in good faith and in real time. In this case Barauna was as excited and mystified as anyone else, and everything happened just as reported. Capt. Bacellar and other crew members verified on board that the negatives showed the "UFO". At the time, no one deceived anybody.


We suppose that when Barauna printed up the negatives back home in his darkroom he was able to study enlargements under a glass and saw indications of what appeared to be a pair of engine pods. Suddenly the the wings 'popped out' and he realised that this was some sort of weird plane, partly obscured by the vapour which was reported by winesses on the spot. Disappointed, he decided this was too good an opportunity to pass up.

He made very large prints of all the negatives. He decided to keep his retouching to an absolute minumum and probably left three shots almost untouched, but on one giveaway shot taken when the plane was closest (P3) he painted or otherwise superimposed a Saturn-shaped profile over the shape of the WF2 Tracer. He then re-photographed all of the prints onto a new negative strip. He cut the six UFO shots off the end of the original neg strip and threw them away, leaving him with the need to invent an explanation of why the strip of negatives is no longer intact. He told the Navy that he did this because he tried reducing some of them individually in order to remedy a slight exposure error.


As Powell suggests, if Barauna fancied himself as a trick photographer he might have wanted to put one over on journalist Joao Martins after the Barra da Tijuca affair. Martins and a photographer had been involved in a notorious photo hoax a few years before, and Martins (at the instigation of researcher Dr. Olavo Fontes who had previously learned of the case through Navy contacts) happens to have been the first journalist to make contact with Barauna after the news of the case leaked from the President's office. Could Barauna somehow have engineered this 'coincidence'? It also happens that the Barra da Tijuca hoax had been republicised just three months previously, in Oct 1957. Barauna might have thought "anything you can do I can do better"; besides it would be a solid-gold story anyway because everyone on the ship believed they had seen a UFO. Barauna was the only one with photo evidence to the contrary.


Brazil had bought her first aircraft carrier from the UK some 14 months previously, and it was destined for an ASW role (Fig.3 below). But from June 1956 it was undergoing rebuild at Rotterdam, and was not recommissioned until 1960. Moreover since the Tracer was still in pre-production in Jan 1958 there could not have been an operational Tracer on a Brazilian deck anyway. Finally, of course, the Brazilian Navy would have explained the sighting - and rumbled Barauna's hoax - quite quickly if this had been the case.

One possibility is that there was a US carrier in the area with a Tracer undergoing final pre-operational trials - or perhaps being demonstrated to a potential purchaser.



Fig.3 The carrier Minas Gerais, built at Swan Hunter as the Royal Navy ship HMS Vengeance and sold to Brazil in December 1956 for $9 million as an attack/anti-submarine aircraft patrol platform. She was de-commissioned in 2001 and laid up at the Naval Dockyard in Rio.

Intriguingly on Feb 26 1958, just five days after the Trindade Island story broke in the Brazilian papers, Major-General Thomas Darcy, USAF, flew into Rio for a meeting with Brazilian defence officials. At the airport he was interviewed by the press. He was asked for an opinion on flying saucers and for details of the reason for his visit. After giving the official line on UFO reports he is quoted (in a letter of March 15 1958 from Olavo Fontes to Richard Hall) as saying

"There are many reasons [for my visit]. The first is related with the delivering of airplanes and instruments for anti-submarine defense. As to the others, I regret to inform that I can’t tell anything about them. . . . I am going to study the several problems which attract the interest of our countries, and I will visit, of course, the Air Base of Salvador."

Fontes points out that the AF Base at Salvador, capital of the State of Bahia, is the nearest point of departure on the mainland for a flight to Trindade Island. But of course the equally significant corollary is that Salvador would be the Air Base nearest to a carrier steaming northwest of Trindade Island, with ASW aircraft on deck either for demonstration or ready for delivery to Salvador Air Base pending completion of the refit of the Minas Gerais.

It ought to be possible to check if there were any US carrier operations in the area. Results so far are uncertain. I was able to discover one carrier that was definitely at sea somewhere in the Atlantic on 16 Jan 1958. The carrier was the USS Essex, which at the time was carrying S2-F Tracker ASW aircraft.

The S2-F or the "Stoof" was the the predecessor of the Tracer and the proven ASW aircraft operational at this time. It did not have a radome. The WF-2 Tracer - or the "Stoof with a roof" as it came to be known - was still a month from its first known operational deployment. But the Tracer had been flying for 7 months. Might the Essex have been carrying a demonstrator whilst delivering a consigment of Trackers, or perhaps whilst taking part in some pre-delivery training exercise with the Brazilian forces?

In January 1958 the Essex was fresh from participation in NATO exercise STRIKEBACK off the coast of Norway, where two of her S2-Fs were lost with all crew in a mid-air collision during a search for a downed F-4D from the carrier USS Saratoga. I can't nail down her position on January 16, but intriguingly she suffered another accident just two days later on Jan 18 when an AD-6 crashed on landing, causing huge damage and a major deck fire that resulted in her putting into Mayport, Fl. for repairs.

According to a contemporary press release this was said at the time to be the largest peacetime dollar loss in US Navy history - close to $5 million. A total of 12 planes were destroyed on deck or fell into the sea and 2 more were damaged. Amazingly they got the Essex underway again with a full compliment of AD-6s by Feb 2 and she sailed for the Med.

I found a crewman who served on the USS Essex from 1957-1961. He can't find out from his memorabilia exactly where they were sailing at the time of the crash - just somewhere "off the Atlantic coast" - but he was on the fire party that put out the deck fires after the Jan 18 crash. It took several hours and he remembers the event vividly. (As an interesting aside, he also remembers overpainting the USAF insignias on aircraft when the Essex delivered hundreds of mysterious 'civilians' to the Bay of Pigs in April 1961.) He says the Essex did carry the new Tracers operationally from a fairly early date, but his current best memory is that this was not until 1959. He has no recollection of a Tracer on board before that time, and no specific memory of any exercises off Brazil or demonstrations involving Brazilian forces. The earliest definite record that I have been able to find of a Tracer on board the Essex is dated 1967 (see Fig.4).

All in all, this history is not very consistent with the theory that the unlucky Essex was involved in the Trindade affair. Of course I can't rule it out. But so far I can find no definite evidence of the presence of any carrier in the South Atlantic at that time, and no evidence at all, therefore, of the presence of a WF2 Tracer.

Fig.4 An E1-BTracer leaving the deck of the carrier USS Essex in 1967. S2-F Trackers with wings folded at right (source unknown)


Knowing the size of the Tracer, a fairly large aircraft with a span of over 70', the angular sizes and displacements can be measured from the photographs to crudely estimate a ground track. The groundspeed implied can then be compared with the Tracer's maximum airspeed of 227 mph.

Making some simplifying assumptions the ground track given by connecting the positions P1 to P4 comes out at roughly 2.25 miles (3.6 km) for an object 70' across. Thus, a duration of 14 seconds (believed to have been estimated by the Navy as a minimum for six shots based on timing Barauna's operation of the camera) would imply a mean groundspeed of nearly 580 mph. More realistically, a duration of 30 secs (estimated by witnesses and given as a probable maximum in the Navy Intelligence report) yields a mean groundspeed of about 270 mph (we can ignore headwinds and tailwinds, which cancel over the whole track). Obviously this is still too fast for a Tracer, especially in view of the fact that 2.25 miles is not the entire track, because the southermost extent of the object's trajectory took it out of sight behind a mountain peak for "a few seconds" between P2 and P3. The course therefore involves what would have to be a roughly 180 banking turn whilst out of sight behind the central peaks of the island, a manoeuvre which could not be conducted at top speed. As a gross approximation let us round up the distance to 3 miles. Then in 60 secs or 120 secs, mean groundspeeds would be about 180 mph and 90 mph respectively. So a duration of, say, a minute or so might be consistent with a Tracer. This is somewhat longer than witness estimates but not necessarily inconsistent with results derived from cloud pattern and wave pattern displacements.

Fig.5. E1-B Tracer plan and elevations

Other aspects are harder to equate with a Tracer, however. The eyewitness evidence uniformly refers not only to the startling speed of the object, but also to its "abrupt" changes of speed and direction on an "undulating" flight path, another feature inconsistent with an aircraft of rather sedate performance. Also, as Brad Sparks has pointed out (personal communication) there must be a question as to whether a Tracer or comparable aircraft would have the tight radius of turn necessary to execute the reported course reversal over the island whilst briefly out of sight, and the aircraft aspect that might be consistent with the P1 and P2 images is inconsistent with the necessary direction of flight.

Ex hypothesi the attitude of the aircraft in relation to the camera at the time of P1 (see Fig.2) must have been almost head-on, heading only a few degrees south of the position of the ship or roughly SSW. In this aspect we have the illusion of a domed disc and, as Kentaro Mori has pointed out (personal communication), the black-painted nose of the Tracer 's radome (see Fig.1 & 2 and drawings in Fig.5) could correspond to the intriguing "dark spot" slightly offset from top-centre of the object in P1. But if this is a Tracer radome it is difficult to square with the presence of a similar spot at bottom-centre on P2. An image might be inverted in a hoax (see below), but a Tracer could certainly not fly upside down!

Another problem is that a Tracer approaching the ship at (say) 150 mph after P1, would close the distance in about 14 seconds, passing the stern at close range and heading out to sea. As Sparks pertinently points out, the UFO did not do this. Instead it moved transversely across the field of view with only small change in angular size to disappear behind a mountain.

All of the photographed positions of the aircraft would be to seaward of the highest elevations on the island and it would be impossible for it to have flown behind even the foremost peak which appears to the left of the object in P3 - if this is the unnamed 309m peak as deduced by Kentaro Mori. On the other hand if Tim Printy's identification of this peak as the 345m peak is correct then it could in principle have flown behind this - provided the pilot was motivated to risk flying low through the jagged terrain for unknown reasons. But more importantly, in order to disappear behind any island peak the approaching Tracer would have to turn very sharply onto a SSE or SE heading immediately after P1, maintaining approximately similar angular size and therefore similar range from the ship at the point of P2.

Not only is the suddenness of such a turn hard to equate with a Tracer, the orientation of the aircraft axis would have to rotate by some 45 - 70 in the direction of this new heading, placing it roughly broadside-on to the ship. Instead of the front elevation of the Tracer, as seen in Fig.2, photo P2 ought to show a side view of the fuselage. Yet P2 still shows what must be a near-head-on aspect almost identical to that in P1. And even if we wanted to try to argue that P2 does show a side view - a difficult case to make given the appearance - then because the length of the Tracer's fuselage is only 48' compared with the 72' wing span, the angular subtense would mean that the aircraft was much closer, at 700m or less, which aggravates another difficulty - the absence of engine noise.

I have found no direct data on the Tracer's engine noise spectrum. However the Tracer was similar to a DC-3 in terms of overall weight, size (neglecting the radome) and number of props, but with a much larger installed power of over 1500 HP (about 130% the DC-3) in its two heavy radial piston engines. This reflected the needs of carrier operation. The result was a much more noisy aircraft than a DC-3 (itself very noisy). I am given to understand, by an expert source, that an approaching DC-3 at low level can typically be heard a minute or two before it becomes visible to the eye, and would be audible at anything up to 4 - 5 miles (6.4 - 8 km), if winds were favourable.

Whilst the wind in this case was perhaps not favourable, insofar as the plane would be downwind of the ship, the surface breeze recorded in the ship's log was only 'light' and therefore not very unfavourable. It was suggested earlier that the reported sound of surf breaking onshore might have contributed, along with the excitement on deck, to masking engine noise; but of course the other side of this coin is that if surf could be clearly heard several hundred meters downwind in conditions of a light breeze then the breeze could not have been carrying sound away very effectively, and neither could the noise on deck have been significant. (Note: Plumes of surf can be seen on the photographs breaking on the E shore of the island. This is not inconsistent with the light SE wind since the local water transport is governed by the fetch of winds tens, hundreds and even thousands of miles away, moderated by geostrophic equilibrium and the shape of the ocean basins. In this case the momentum is dominated by the north-south flowing Brazil Current and wave refraction will bend the shoaling ocean swells towards the shoreline. See Ocean Current & Wave Motion near Trindade Island.)

The wind at a couple of thousand feet was very probably stronger than the surface breeze of course (we have no balloon soundings for the date and time in question, but there is good evidence that upper level winds - at tens of k/ft - must have been in the order of tens of knots). But even if we allow that the wind was sufficient to halve the perceived engine noise, it seems safe to infer that a DC-3 ought to have been audible from the ship at 3 - 4 kilometres in such conditions, and that therefore the low-frequency roar of the noisier Tracer should certainly have been very audible as it passed in front of the mountains at the much shorter ranges indicated by the angular subtense of the photographs.

In the most distant shot (P4) a Tracer would be only about 1.9 km from the ship. The angular width of the closest (P3) corresponds to a maximum of about 1 km. Plotting the course shows that the Tracer would have approached the ship from the northwest in P1, turning immediately to fly above the bay between the ship and the eastern shore, crossing inland over the Cabritos Beach and proceeding to the position of P2 side-on at a distance of only about 700m distance, with the mountainous spine of the island rising behind it to several hundred meters, then returning from P3 along a similar route. The natural rocky amphitheatre would be likely to confine and re-echo the engine noise on both inbound and outbound legs. It is hard to imagine that the noise would be anything less than loud.

Also, whilst the Tracer theory might remove the complications of a premeditated hoax it does not remove all problems associated with photographic fakery. Although the first, second and final photographs might all (with some strain admittedly) be interpreted as basically unretouched images of a blurred or fogged-up Tracer, photograph P3, in which the object appears at its largest and with best definition, clearly is not an unretouched image of a Tracer. Our hypothesis therefore has to involve an overpainting or montage to create the UFO on an enlarged print of P3, after which all four prints would then be rephotographed onto a new negative strip.

The reason that all four frames must have been similarly copied is that the shrinking of the density range, and consequent loss of shadow and highlight detail, incurred by a second generation copy would cause P3 to stand out as drastically different. But then this inevitable degeneration becomes a problem for all the photos, which appear to resolve excellent fine detail at both ends of a very wide density range. In this respect the Tracer theory is less satisfactory than the internal mask theory or some variant thereof, which in principle might produce a fake negative in real time without double-exposure or posterior montage (but which unfortunately would require premeditation and careful preparation before the trip and doesn't offer us a "real UFO" to explain the considerable circumstantial evidence that there was one.)

It is also interesting to notice that when a digital contour tracing is applied to a contrast-enhanced image of P3 the central flange resolves into a structure of dark patches or nodules. Although this is the sharpest image (on the original print) the pattern is not at all evident to the casual eye, but it can be compared with a related pattern of blurred striations that appears also in digital enhancements of P1 and P2 and which can be shown to extend from a similar equatorial nodular band in P4. The subtle relationship of these features does not seem to be consistent with the overpainting of P3. And in point of fact microscopic examination (of the negatives by the Brazilian Navy and its civilian consultants; of first-generation prints by later private researchers) has revealed no evidence such as sharp edges or grain pattern anomalies that might indicate retouching or montage.

We know that a deal between the US and Brazil to purchase ASW aircraft (probably S2F Trackers) was advertised during Major-General Darcy's visit as being underway at this time, explaining the possible presence of a US carrier in the area, and if there was a carrier there could have been a Tracer too. But surely this source would early on have been ruled out by the Brazilian Navy UFO investigation which had been underway for a fortnight by the date of the photo incident. Inquiries would have been made and aircraft movements accounted for, as described in the Navy Intelligence report. If the US carrier (ex hypothesi) was present for no more exotic reason than a trade arrangement openly discussed with the newspapers, we need to explain why information on a US overflight of Trindade would be secured away from Capt. Brandao and his intelligence staff.

Perhaps the Brazilian government simply had no knowledge of the existence of the Tracer, even though it may have been known that a US carrier was in the area. It is possible that the overflight of the island would have been considered a sensitive matter by both countries: By the US in respect of a new technology that was still technically secret, and by Brazil in respect of a territorial intrusion. If the overflight of Trindade had been ill-advised, the US might be keen to avoid potentially embarrassing official protests and questions from Brazil. Perhaps General Darcy's visit to Rio was in part a diplomatic smoothing of potentially troubled waters.

Another more Machiavellian explanation would relate the presence of a Tracer to those "other reasons" for Major-General Darcy's visit to Rio on February 26 1958, matters whereof the good General was disinclined to speak. Perhaps a Tracer was being used in connection with the Brazilian Navy's UFO investigation at the island, under what might have been a long-standing data-sharing and resources-sharing agreement with the US military. The Tracer's new state-of-the-art AEW radar would certainly be an ideal tool for coordinating a comprehensive air-sea surveillance. In this case the delivery to Brazil of "aircraft and equipment for anti-submarine defence", announced so publicly by General Darcy, would presumably be a convenient cover story designed to deflect inquiries away from the real reason for a US carrier presence in the area. But now we can hardly explain the absence of any of this from the confidential Navy Intelligence report as due to mere innocence. We would have to assume that this, too, was a play within a play.


The complete absence of engine noise is a strong argument against a Tracer or indeed any conventional aircraft. The argument above suggests that in order to be inaudible a Tracer would have to be several times as far from the ship as the 1 - 2 km distance consistent with the angular size. A larger aircraft several times as far away would have to be several times as fast and, to preserve consistency with these two requirements, have more engines. It isn't clear that these conditions would actually reduce the likely engine noise level, and moreover the type of large aircraft implied (obviously not carrier-based and therefore with no connection to ASW aircraft sales for which there is evidence) would certainly be harder to square with repeated overflights of a remote Brazilian island territory.

Considering the inconsistency between the airframe aspect and the track heading, one might try to adapt the Tracer hypothesis in the same way that Powell adapted his Bonanza hypothesis, i.e. by suggesting that Barauna used some photos of an approaching Tracer obtained elsewhere in order to make a posterior montage or double exposure. This could explain why the aircraft appears to be flying sideways, and why the black-painted nose of the radome (ex hypothesi, the "dark spot") appears to be dorsal in P1 and ventral in P2. But not only does this theory suffer from the same shortcoming as did Powell's (i.e., it fails to confront the circumstantial evidence that a sighting of some sort did take place on the ship and leaves unaddressed the technical questions connected with these photo-hoax methods) it has the additional difficulty that Barauna would be very unlikely to have been in a position to photograph a prototype Tracer.

In summary, the Tracer theory has some attractive features but overall is unsatisfactory in several crucial respects.


Powell, M., The Trindade Island UFO; A detailed Study of Photos 1 & 2

Back to Trindade index