Signalements : le nombre credible de "soucoupes volantes"

Most readers of this report will perhaps be convinced that alien spaceships or some other unknown phenomena can be involved in only a very small percentage of all UFO reports or perhaps in none. Yet there is a curious tendency on the part of many students of the problem to imply that the sheer number of reports somehow proves that there must be some physical reality involved. For example, J. B. McDonald (1968) argues before the House Committee on Science and Astronautics, in a one-paragraph statement on witness credibility: "... It seems tedious to enlarge here on those obvious matters. One can be fooled of course; but it would be rash indeed to suggest that the thousands of UFO reports now on record are simply a testimony to confabulation, as will be better argued by some [selected cases]." Jones, who argues against the probability of any substantial number of flying saucers says: "There have been so many flying saucers seen by now, if we were to believe the accounts, that surely one of them must have broken down or left some trace of its visits. It is true that one can explain the absence of relics by supposing ... fantastic reliability ...

It would seem to me that if one begins by studying both witness reliability and selected cases, and if one thereby realizes that it is quite conceivable and probable for the great bulk of reports to be simple mistakes and fabrications, then arguments invoking the enormous number of reports become irrelevant. We are concerned by only a small "residual" of puzzling reports.

This raises another approach to the UFO "residual" reports. We could attempt to answer the question: what is the maximum frequency of spaceships that could actually have penetrated our airspace and still leave us with such meager evidence as we have for their existence? Obviously if a 30-ft. metal disk hovered over the Capitol for some hours, we would have a multitude of photos, video tapes, and other hard evidence from different observers in different positions.

Une mesure de la réaction du public face à des phénomènes spectaculaires et céleste non familiers peut être obtenue de l'étude des signalements de boules de feu. 6 boules de feu spectaculaires furent étudiées jusqu'ici en utilisant les analyses de C. P. Olivier de la Société Météoritique Américaine (1962, 1963, 1967) et les signalements dans Ciel et Télescope. Parmi ceux-ci, la durée la plus longue fut de seulement 31 s pour l'objet du 25 Avril 1966 ; encore une fois même pour un objet d'une telle courte durée, un certain nombre de photographies furent prises. Dans d'autres cas, des traînées de poussière d'une durée allant jusqu'à 17 mn furent photographiées le largement signalées. Les observations Zond IV s'appliquent également. Ces données permettent des estimations de la fréquence des signalements visuels comme photographiques.

Les boules de feu étaient plus brillantes que la pleine Lune dans la plupart des cas. Sous elles n'apparaissent pas comme des sources ponctuelles, mais comme un disque ayant environ la moitié de la taille de la Lune. Certaines d'entre elles étaient suffisamment brillantes pour attirer l'attention des personnes à l'intérieur ; some of them were accompanied by thunder-like explosions. All attracted national publicity. In short, they are remarkable enough to have attracted attention and photographs, and are thus considered comparable to hypothetical, well-observed "flying saucers" in public response.

The analysis must take into account the number of inhabitants in the area of visibility as well as the duration of visibility. We may call the product of the number of inhabitants times the duration, the "exposure" of the phenomenon. We can ask how the total number of actual witnesses is related to the exposure.

Pour les durées de période courte (quelques mn) il est raisonnable de s'attendre à ce que le nombre de témoins (une fraction du nombre d'habitants) soit proportionnel à l'exposition. Ceci peut également est supposé quant au nombre de signalements détaillés récupérés par les enquêteurs qui les solicitent, et quant aux nombre de photographies. Dans les cas de boule de feu et de Zond IV il existe des données donnant le nombre de témoins, le nombre de signalements récupérés, ou le nombre de photographies. Ainsi, si N est le nombre total d'habitants, et t est la durée de l'événement (en s), nous avons une théorie du 1er ordre de la forme :

Nombre de témoins = N W = C_WNt

Nombre de signalements récupérés = N_r = C_rNt

Nombre de photographies récupérées = N_P = C_PNt

Il est possible d'identifier la constante de proportionalité, C d'après les signalements mentioned ci-dessus. Les valeurs dérivées sont listées en Tableau 4. La constante 1/C has dimensions man-sec/witness (or /report, /photographer). For example, the Air Force files on Zond IV yield 78 reports for a two-minute phenomenon visible from a region inhabited by an estimated 23 000 000 persons, giving 3.5 x 10⁷ man-sec to generate one report. It is clear that the number of photographs generated will depend on the duration of the phenomenon in a more complex way than indicated in our simple equation, since with durations longer than some limit, more witnesses will have time to obtain a camera. In this approximate and first-order treatment, this complication is neglected.

Application of Table 4 can be illustrated by the fireball reports. The original data suggest about 500 reports in five years for these very bright objects. We assume that the average fireball is visible roughly 10 sec. These figures allow us to solve the equation (cited above) for the number of inhabitants through whose skies pass fireballs in five years. If it takes 6 x 10⁶ man-sec. to generate one report (Table 4), then the fireballs must have been exposed to about 300 000 000 people. This figure is expected to be accurate to something better than an order of magnitude. That is, every citizen of the United States evidently has such a fireball in his sky about once every few years (whether or not he is outside and sees it). This is in good accord with known data - Vedder's (1966) estimate of the flux of meteors of magnitude -15 is one every three to four years over an area of the size of the United States.

The question before us is how many of the UFO reports could correspond to real objects in view of the available data. Is a "residual" of even 2% of the cases reasonable? We have three relevant statistics:

1. National opinion surveys indicate that roughly 5 x 10⁶ persons of the total U. S. population believe they have seen UFOs in 20-year interval since they were first reported. If 2% of these represent really strange unknowns, we should have 1 x l0⁵ witnesses.
2. There have been roughly 15 000 recovered cases, representing perhaps 45 000 individuals' reports. A 2% residual would give 900 reports of unknowns.
3. The project study suggests that the "residual" photographs of unidentifieds number of the order of 20.

Combining these three statistics with the three constants from Table 4 we derive three independent estimates of the total number of citizens exposed to the "high-strangeness residual UFOs" in the last 20 years; viz., 2x10⁷; 1x10⁸; and 2x10⁹. It can be seen that the accuracy is no better than an order of magnitude. However, taking 200 000 000 persons as a representative value, the implications are clear. The results suggest that merely to generate the 2% residual, every person in the country has had an UFO visible above his horizon once in the last 20 years.

Of course, since most man-hours in this country are spent indoors, or asleep, or paying no attention to the sky, it is not surprising that very few people have reported seeing such craft. But taking into account the array of automatic surveillance equipment operating in this country, it does border on the incredible that the "hard" evidence should be so scanty. The statistic is similar to the five-year statistic for bright fireballs, and although the "evidence gathered over an arbitrary five-year time span for the existence of bright fireballs" is similar to that gathered over 20 years for "flying saucers" the "fireball evidence" is perhaps more convincing : it includes detection by automatic survey cameras, large numbers of witnesses per incident, and more reliable witnesses. To accept as many as 2 % residual cases as examples of extraordinary aircraft, then, is to accept that an UFO could fly around the country in such a way as to be potentially visible to, or in the sky of, every citizen for 40 s without being positively recorded or conclusively reported.