This case is an addition to the growing body of case law1 developing around 'airport searches,' now a part of everyday air travel, resulting from the threat to the traveling public and to the airlines from hijacking. Here, in the course of a 'pat-down' by a United States Customs Security officer following activation of a magnetometer by appellant, a would-be passenger, counterfeit bills were found on appellant's person in a package wrapped in aluminum foil. Prior to trial on a one-count indictment charging possession of counterfeit currency with intent to defraud in violation of 18 U.S.C. 472, a suppression hearing was held on appellant's motion challenging the legality of the search and the admissibility of statements made by him subsequent to and as a result of the search. The district court denied all aspects of the motion and following conviction after a two-day jury trial appellant was sentenced to three years' imprisonment, to serve six months, with execution of the remainder suspended and probation imposed for two and one-half years.

Appellant, an older man who does not speak English but does speak Spanish, was approaching the boarding area of Pan American Flight 233 from Kennedy Airport to Santo Domingo at 7:25 a.m. on October 26 1972. Throughout the Pan Am terminal and at the boarding gate to Flight 233 there were signs posted in both Spanish and English stating that it is a federal crime to board an aircraft carrying a weapon and that passengers are subject to search. In addition, there was evidence that from time to time announcements are made at boarding time in both Spanish and English advising passengers that there would be a search conducted, although there was no specific evidence to the effect that these particular announcements were made as to Flight 233 or that appellant himself heard them.

The initial security procedure there in operation was basically that which has also been used for other international and domestic flights: the passenger places his carry-on luggage on a table where it is then searched, and the passenger then proceeds through a magnetometer, a device that is activated solely by metal on the person of the would-be passenger. There is no indication whatsoever that appellant up to the time of his passage through the magnetometer at the gate to Flight 233 had engaged in any conduct that could in the slightest way be considered suspicious. The much-heralded 'profile' procedure, whereby through observation a given individual is said to have an appearance or characteristics similar to those who have engaged in hijacking with the result that he is subject to subsequent further search or closer scrutiny, was not in use because it was not part of the screening procedure for international flights.

Appellant's boarding line consisted of approximately 30 or 40 individuals ahead of him. We are told that five or six of these were stopped and patted down after they set off the magnetometer. Again, there is no specific evidence that appellant saw these individuals being patted down. In any event, when his turn in line came, he did place his carry-on luggage on the table and it was presumably searched. He then passed through the magnetometer and activated it. The United States Customs Security officer, Ronald Bliss, then motioned for appellant to come over and open his coat. This he did, in the immediate vicinity of the magnetometer. Bliss then patted him down and felt a bulge in appellant's inside jacket pocket. At Bliss's request appellant removed the object, a package wrapped in aluminum foil with a frontal size approximately that of United States currency and a thickness of about one-quarter inch or a little bigger. Bliss indicated to appellant to open one end of it. Appellant complied, without protest, and Bliss observed what he thought was a counterfeit $20 bill. Appellant was then taken to another room, where the package was ascertained to consist of 51 bogus $20 bills, and after being shown a Spanish language 'rights card' by a Spanish-speaking Customs Security officer, appellant claimed that a stranger had approached him in front of his house and asked him to exchange the money in question. He explained the wrapping of the bogus bills (so that they would not get wet). He also had approximately $2,360 in genuine currency on his person that was not wrapped in foil. We do not know whether the magnetometer had been activated by the foil.

As we have said, the usual airport search consists of an examination of the individual passenger and the opening of carry-on luggage and a look at its contents.2 The search of the individual is made initially by a magnetometer; then, if the device is activated, further investigation is made.3 The search of the carry-on luggage may be made manually or by X-ray machines where available. Each aspect of the procedure is, however, a search within the meaning of the fourth amendment.

Even the unintrusive magnetometer walk-through is a search in that it searches for and discloses metal items within areas most intimate to the person where there is a normal expectation of privacy. United States v. Epperson, 454 F.2d 769, 770 (4th Cir.), cert. denied, 406 U.S. 947, 92 S. Ct. 2050, 32 L. Ed. 2d 334 (1972). While subject to the fourth amendment's requirement of reasonableness, nevertheless, such a search is carried out upon each and every passenger without a warrant or without probable cause.4

The airport search is a direct reaction to the wave of airplane hijackings which began in 1968, at which time popular feelings of fear and anger, and ultimately rage, called out for some program to safeguard air flights, and understandably so. Airplane hijacking is a particularly frightening crime. Many hijackers have been psychotic or political fanatics, for whom death holds no fear and little consequence, willing to bargain with the lives of defenseless passengers for money, transportation to a safe haven, the release of 'political' prisoners, or some other otherwise unattainable demand. Congress's penalty for hijacking reflects the public mood: death, if the jury so recommends. 49 U.S.C 1472(i)(1)(A). The present antihijacking system did not spring full blown into life. 'Profiles' and selective investigation came first; plainclothes sky marshals became regular forward seat passengers; anathema as it was to some to do so, diplomatic overtures were made to Cuba to eliminate it as a refuge for hijackers. Finally, technology with magnetometers, metal detection devices and X-rays was brought to bear. Today, the general methodology of the airport search has become more or less routine.5

Inevitably, the legality of these searches has been contested, and courts have responded differently.6 Our own circuit has indeed demonstrated in its confrontations with the problem7 that it has no unanimous point of view. This is not surprising perhaps when one considers the fact that neither component of the usual airport search of the person8 -- the use of the magnetometer or the frisk-- seems to fit readily within any of the traditional exceptions to the warrant requirement,9 and yet each seems reasonable in light of the overwhelming public acceptance of the search, and the necessity for it. It is helpful to consider these two components separately. Before doing this, however, we note our guideline for decision lies in the language-- through not the specific holding-- of Terry v. Ohio, 392 U.S. 1, 20, 88 S. Ct. 1868, 1879, 20 L. Ed. 2d 889 (1968), that 'the conduct involved in this case must be tested by the Fourth Amendment's general proscription against unreasonable searches and seizures,' since an airport search 'as a practical matter could not be subjected to the warrant procedure.' The ultimate standard of the fourth amendment on which we must base our opinion, therefore, is one of reasonableness. Cady v. Dombrowski, 413 U.S. 433, 93 S.Ct 2523, 37 L. Ed. 2d 706 (1973). And the reasonableness of a search depends upon the facts and circumstances and the total atmosphere of each case. Chimel v. California,395 U.S. 752, 765, 89 S. Ct. 2034, 23 L. Ed. 2d 685 (1969). Our inquiry here must be directed to the basic issue whether in the totality of circumstances such a search is reasonable.

In treating each component of the airport search, it is necessary to focus on the governmental interest which supposedly justifies intrusion on the privacy of the person, or simply to focus on the need to search. Terry v. Ohio, 392 U.S. at 20-21, 88 S. Ct. 1868. See also Camara v. Municipal Court, 387 U.S. 523, 536-537, 87 S. Ct. 1727, 18 L. Ed. 2d 930 (1967).

In 1969 the number of successful hijackings and hijacking attempts of United States carriers peaked at 33 and 40, respectively. FAA, Office of Air Transportation Security, Hijacking Attempts on U.S. Registered Aircraft (Oct. 5, 1973). As the airport searches became more widespread and other aspects of the anti-hijacking program took effect, the number of successful hijackings dramatically declined to 10 in 1972, and, surprisingly, there have been no successful hijackings of a commercial airliner in the United States in over a year. Id.10 These statistics would seem to indicate that the total anti-hijacking system has had some considerable success.

In addition to firearms alleged and real used in 117 incidents, other weapons used include knives, razors, bombs, a hatchet, an ice pick, and a tear gas pen. FAA, Office of Air Transportation Security, Hijacking Statistics (Aug. 1, 1973) 14. In this age of miniaturization a bomb may be fashioned out of plastic and placed inside a shaving cream can with a firing device with no more metal than a small fountain pen. In other words, no practical inspection will be able to insure absolutely the safety of airplane passengers. On the other hand, the overwhelming majority of weapons will respond to a magnetometer or be picked up in a carry-on baggage search. One of the prime purposes of the search, moreover, is deterrence, the knowledge that such searches are conducted acting to deter potential hijackers from even attempting to bring weapons on a plane. Depriving a hijacker of his weapon is critical, because by means of a weapon like a pistol or even a knife the hijacker may literally turn the plane itself into a weapon, threatening not only those within it, but those on the ground as well. In short, the plane may become a weapon of mass destruction that no ordinary person would have any way of obtaining except through a hijacking.

In determining the reasonableness of a search, one must also balance the need for the search against the invasion of privacy involved. There is a compelling need for a search to detect weapons before they are brought on an airplane. While this might justify some search of all prospective passengers at an airport, the question becomes whether it will justify in any given case the search as carried out and, for this case, the search of appellant here, first by use of the magnetometer.

I. USE OF THE MAGNETOMETER.

Initially it may be noted that the magnetometer search as we have it here could be viewed as 'inefficient' in that all passengers are searched, but only a fraction of one per cent have weapons. The magnetometer, although calibrated supposedly to be activated by a mass of metal approximating a .25 calibre pistol, often is activated by car keys, ladies' sewing scissors, briefcase hinges and latches, and the like. Indeed, one case says that there is evidence that up to 50 per cent of the passengers passing through a magnetometer may activate it.11 See United States v. Lopez, 328 F. Supp. 1077, 1086 (E.D.N.Y.1971). Thus the device is activated by numerous objects not useful as weapons, and if in each instance the person were subsequently to be frisked as a matter of course, a large number of unnecessary invasions of privacy would result.

This 'inefficiency' in the use of the magnetometer to search for weapons is troublesome for two reasons. First, whatever the extent of airplane hijackings in the future, one suspects that anti-hijacking measures and the accompanying airport search will remain a part of our way of life for some time to come. While the system was justified originally as a means to stop hijackings, now that it seems to be proven successful, its retention will be justified as necessary to stop their resurgence. In any case, the search may eventually remain without an actual, existing threat to justify it. Second, there is the possibility that the purpose of the airport search may degenerate from the original search for weapons to a general search for contraband. Again, the recent experience of courts with airport searches, including the present one, is not wholly reassuring from this point of view.12 And as potential hijackers are further deterred by the search, the only defendants before us as a result of airport searches will be those caught with contraband.13

To the extent that 'inefficiency' as we have referred to it is a necessary concomitant of a successful anti-hijacking system, the standard of reasonableness still prevails and the system must be viewed with due allowance therefor. To the extent, however, that the system were allowed to degenerate into bureaucratic shortcuts which served no legitimate purpose and still constituted an invasion of protected rights of privacy, then the system would be unreasonable per se. As was said by the Supreme Court in Terry, 'The scope of the search must be 'strictly tied to and justified by' the circumstances which rendered its initiation permissible.' 392 U.S. at 19, 88 S.Ct. at 1878, quoting Warden v. Hayden, 387 U.S. 294, 310, 87 S. Ct. 1642, 18 L. Ed. 2d 782 (1967) (Fortas, J., concurring). In other words, it is, and indeed for preservation of a free society must be, a constitutional requirement that to be reasonable the search must be as limited as possible commensurate with the performance of its functions.

It has been suggested that those who seek to travel on a common carrier have a lower 'expectation of privacy' regarding their person and the bags they carry than those who, say, merely walk on a public street. Such a suggestion has little analytical significance; if it were announced that all telephone lines would be tapped, it could be claimed that the public had no expectation of privacy on the telephone. What is clear is that the public does have the expectation, or at least under our Constitution the right to expect, that no matter the threat, the search to counter it will be as limited as possible, consistent with meeting the threat.

While the magnetometer may be 'inefficient' in that it searches every passenger, balanced against this is the absolutely minimal invasion of privacy involved. There is no detention at all; there is no 'probing into an individual's private life and thoughts . . ..' Davis v. Mississippi, 394 U.S. 721, 727, 89 S. Ct. 1394, 1398, 22 L. Ed. 2d 676 (1969). The passing through a magnetometer has none of the indignities involved in fingerprinting, paring of a person's fingernails (Cupp v. Murphy, 412 U.S. 291, 93 S. Ct. 2000, 36 L. Ed. 2d 900 (1973) (upheld where incident to a valid arrest)), or a frisk. The use of the device does not annoy, frighten or humiliate those who pass through it. Not even the activation of the alarm is cause for concern, because such a large number of persons may activate it in so many ways. No stigma or suspicion is cast on one merely through the possession of some small metallic object. Nor is a magnetometer search done surreptitiously, without the knowledge of the person searched. Signs warn passengers of it, and the machine is obvious to the eye.

Upon this balance we have no difficulty in reaffirming the conclusion of United States v. Bell, 464 F.2d 667 (2d Cir.), cert. denied, 409 U.S. 991, 93 S. Ct. 335, 34 L. Ed. 2d 258 (1972), that the use of a magnetometer is a reasonable search despite the small number of weapons detected in the course of a large number of searches. The absolutely minimal invasion in all respects of a passenger's privacy weighed against the great threat to hundreds of persons if a hijacker is able to proceed to the plane undetected is determinative of the reasonableness of the search.

In this regard we do not predicate decision upholding magnetometer searches in general on the basis of 'consent.' While a search which would otherwise be unlawful may through the consent of the person searched become lawful, Katz v. United States, 389 U.S. 347, 88 S. Ct. 507, 19 L. Ed. 2d 576 (1967); United States v. Callahan, 439 F.2d 852 (2d Cir.), cert. denied, 404 U.S. 826, 92 S. Ct. 56, 30 L. Ed. 2d 54 (1971), such consent entailing as it does the waiver of a constitutional right, must be freely and voluntarily given; it must not be directly or indirectly the result of coercion. Schneckloth v. Bustamonte, 412 U.S. 218, 228, 93 S. Ct. 2041, 36 L. Ed. 2d 854 (1973); United States v. Thompson, 356 F.2d 216 (2d Cir. 1965), cert. denied, 384 U.S. 964, 86 S. Ct. 1591, 16 L. Ed. 2d 675 (1966). To make one choose between flying to one's destination and exercising one's constitutional right appears to us, as to the Eighth Circuit, United States v. Kroll, 481 F.2d 884, 886 (8th Cir. 1973), in many situations a form of coercion, however subtle. Cf. Lefkowitz v. Turley, 414 U.S. 70, 79-82, 94 S. Ct. 316, 38 L. Ed. 2d 274 (1973). While it may be argued there are often other forms of transportation available, it would work a considerable hardship on many air travelers to be forced to utilize an alternate form of transportation, assuming one exists at all.14

II. THE FRISK.

We pass then to the question of the frisk after activation of the magnetometer. Normally a frisk is considered a gross invasion of one's privacy. In Terry the Court rejected the claim that it was only a 'petty indignity,' for, in a proper frisk, the

officer must feel with sensitive fingers every portion of the prisoner's body. A thorough search must be made of the prisoner's arms and armpits, waistline and back, the groin and area about the testicles, and entire surface of the legs down to the feet.

Priar & Martin, Searching and Disarming Criminals, 45 J.Crim.L.C. & P.S. 481 (1954), as quoted in Terry v. Ohio, 392 U.S. at 17 n. 13, 88 S. Ct. at 1877.

While the typical airport frisk may be more in the nature of a 'pat-down,' involving only the patting of external clothing in the vicinity of pockets, belts or shoulders where a weapon such as a gun might be secreted, in any given case the right to pat-down carries with it authorization for a full frisk since presumably, if we are authorizing anything, we are authorizing what is necessary to get the job done.

The fact that many persons undergo frisks or pat-downs at the airport boarding gate does not stop the frisk from being a gross invasion, but the frisk of many does mitigate the invasion to one to the extent that stigma attaching to one who endures it is lessened by being one of a crowd. This at least makes the airport frisk less noxious than a frisk on the street.

The intrusiveness of the airport frisk is also limited by foreknowledge of it. Signs warn of searches and one standing in line may see those in front being frisked. Thus, when one is himself frisked, it is not a surprise, and there is an element of voluntariness.15 Even after activating the magnetometer, the prospective passenger may refuse to submit to a frisk and instead forfeit his ability to travel by air, because this serves the purpose of the whole search procedure, which is not to catch criminals, but rather to keep armed hijackers from getting on airplanes. United States v. Davis, 482 F.2d 893, 908 & n. 41 (9th Cir. 1973). The signs and highly visible magnetometers and frisks serve as a deterrent, and indeed the hope is that the prospective hijacker will turn around and leave. Thus, while there is coercion to submit to the airport frisk, there is no compulsion, as there is in an ordinary street frisk, and the airport frisk is, as we say, less intrusive thereby.

It is unfortunate, but not necessarily fatal, that there is not an explicit warning that the person may avoid the frisk by not flying. Because, however, as we view it, the legality of the search does not rest on a 'consent' theory, but rather on the reasonableness of the total circumstances, the fact that a passenger may withdraw at any time plus the general knowledge that one is subject to a search if one flies is sufficient in light of the otherwise minimized invasion of the passengers' privacy.16

There is, of course, a compelling need for further investigation after an initial magnetometer reading showing metal. Without further investigation the magnetometer would not serve any valid purpose. It would be absurd to require the airlines either to deny passage to everyone activiating the magnetometer or to allow passage to any without discovering what the metal was. We refer back to our basic test: how may the search be limited commensurate with the performance of its functions? Put another way, what is the minimal invasion of privacy consistent with the need for the further investigation? Differences in equipment and facilities may change the actual procedure, but the rule is easy to state: exhaust the other efficient and available means, if any, by which to discover the location and identity of the metal activating the magnetometer before utilizing the frisk.

At the outset it should be noted that any further investigation after activation of the magnetometer is for the metal which did the activation; activating the magnetometer is not a general license to search for anything. Next, it is clear that compared with other methods the frisk is 'inefficient' in that a whole body may be frisked only to discover that the keys in the passenger's pocket set the machine off. It would be easier and more effective merely to ask the passenger to remove all his metal items and return through the magnetometer. This procedure is clearly preferred over the immediate frisk because, while still a search, it entails far less invasion of the privacy or dignity of a person than to have a stranger poke and pat his body in various places. Such a procedure is, as we have said, supra note 3, frequently used. It is proper for several reasons. It limits the search to the metal detected by the magnetometer. If the person passing through the magnetometer the second time does not activate the machine, the metallic items have all been made visible to the officer. If, on the other hand, the machine is activated again, the person has failed to divest himself of all metallic items on his own. While this in itself may not provide 'probable cause' or even a reasonable suspicion in Terry terms, it necessitates that further measures be taken. At this point the only further practical measure is a pat-down, or frisk if necessary. And it is reasonable at this point because no lesser readily available means presently exist for determining what unexplained metal object remains on the passenger's person. See United States v. Epperson, 454 F.2d at 772. This is the factual situaton-- a second walk-through followed by a second activation of the magnetometer-- that was involved in United States v. Mitchell, 352 F. Supp. 38 (E.D.N.Y.1972), aff'd without opinion, No. 73-1953, 486 F.2d 1397 (2d Cir., 1973).17

Of course, the second magnetometer walk-through is not the only alternative available to the authorities. Where airports are equipped with hand-held magnetometers they may be used to locate the metal on the passenger's body and he then may be required to remove it so that they may be assured that it is not a weapon. Another additionally proper possibility is to provide instruction to all passengers that metal items should be removed prior to passing through the magnetometer initially and to provide a table or whatever on which these items may be placed so that they may be viewed by an official. These examples are not to be considered exhaustive, but, again, the rule is easy to state: the frisk in the typical boarding situation we have been talking about is to be used only in the last instance.

It is clear that the search of appellant was not as limited in its obtrusiveness as it might have been. It would not have entailed any significant time to require him, for example, to remove any metal from his body and to pass through the magnetometer again. The fact that he spoke no English is no excuse, since a Spanish translator was available, and since this was a flight to Santo Domingo from a major airport there is no reason this possibility could not have been provided for.

In view of our determination that the frisk of appellant was unlawful, we need not reach the question whether the officer's opening of the aluminum foil would have been lawful. In light of the recurrent nature of such problems, however, a few words may be in order. It might be said that inasmuch as the officer is only looking for the item which activated the magnetometer, he could not open a container, which he knew upon feel could not contain the offending metal. Generally, this is so, because a passenger's activation of the magnetometer does not give the officer a license to search him generally, but only for the offending metal. This is not to say, however, that an officer may never investigate something which is not metallic. Hijackers as well as airport officers know of the existence of plastic explosives or even ordinary dynamite. If an officer comes lawfully upon a container which may conceal such items, he may require that they be opened to his inspection before the passenger is allowed to proceed on board with the container. United States v. Kroll, 481 F.2d at 886-887. In Kroll, the Eighth Circuit quoted with approval the statement of the district court that a search would be permissible "of that which may reasonably be deemed to conceal a weapon or explosives. Reasonableness, in this context, is a matter of probabilities." Id. at 886, quoting 351 F. Supp. 148, 153 (W.D.Mo.1972). Moreover, the probabilities may be affected by the unusualness of the container, the size or shape of the object, the failure or inability of the passenger to explain what is present therein, and other considerations.

To some it may seem that we are splitting hairs, and on its face there may seem to be little difference between what was done to appellant and the additional step or steps that we are requiring. Nevertheless, in this way we believe that the airport search is made effective and not frustrated, while at the same time an individual passenger's privacy is preserved to the greatest extent commensurate with safeguarding his fellow passengers from the possibility of hijacking. Only upon such minimizing of the invasion of privacy can the rigorous requirements of the fourth amendment be met.

Nothing we say herein is to be treated as prohibiting searches of persons where specific, articulable facts exist to support a reasonably prudent man's belief that his or others' safety might be in danger-- that is, where the requirements of Terry v. Ohio are met, as, e.g., in our own United States v. Bell, supra. Nor do we intend in any way to preclude questioning on the one hand or on the other hand such searches as may be proper where other information on the part of the authorities would make such action reasonable.

With this in mind, in the thought that the rule of law here enunciated has not been sufficiently clear to the airport authorities in the past, we announce it as prospective only in effect, applying it to the search in the case at bar and searches subsequent to the effective date of this opinion. Cf. Linkletter v. Walker, 381 U.S. 618, 622-629, 85 S. Ct. 1731, 14 L. Ed. 2d 601 (1965).

We reverse and remand with directions that the motion to suppress be granted.