dissenting:
The bottom line of the seizure and subsequent search conducted in this case is the so-called “drug courier profile” routinely employed by drug enforcement officials. Use of the profile is a prime example of organized enforcement efforts that are rapidly eroding our protection against unwar*1219ranted, arbitrary intrusions by public officials. This erosion is made easier since, in the isolation of the judiciary, we tend to see its impact only on wretched peddlers of misery like the one before us. The periodic intrusions on the rest of us — when, from time to time, our conduct fits the objective behavior relied on by these officers — tend to be lost or ignored when we assess the reasonableness of the conduct in cases like this one.1 Even worse, the device ratified here not only readily lends itself to unreviewable racial bias but in all probability already incorporates and routinely employs it. See United States v. Vasquez, 612 F.2d 1338, 1353 n. 10 (2d Cir.1979) (Oakes, J., dissenting) (agent’s testimony that race plays role in formulation of suspicion).
We know little or nothing about the characteristics that make up the profile. Nor do we know about the standards or criteria that guide an agent’s application of it to particular individuals.2 We have only broad, self-serving police assurances that reliance on the profile and the agent’s judgment is well-founded.
In this case, we are told that suspicion was first aroused when the defendant, while getting flight information, looked repeatedly at the agent. I find that remarkable only in that it describes so many perfectly innocent people. We are further told that the defendant’s flight was from Fort Myers, Florida via Tampa to Atlanta for a connecting flight to Tulsa, Oklahoma. That Fort Myers is a drug source city is equally unremarkable. If we are to believe the representations that parade before us, every city in the United States large enough to have a substantial public airport is a “drug source city.” See United States v. Pulvano, 629 F.2d 1151, 1155 n. 1 (5th Cir. 1980). To this we add that the defendant was flying on a one-way ticket — we are not told what percentage of innocent travelers fly one way — and that the agent saw the defendant looking out a window at the plane.
On this information, the agent initiated an interview in which he learned that this suspicious and devious man traveled under his own name and listed his home phone. He also learned that the defendant was going to Tulsa to visit his ex-wife and children. The defendant’s nervousness is supposed to add some justification for the police conduct that followed. I strongly doubt, however, that even a federal judge would not appear nervous under similar circumstances, no matter how innocent his behavior. The agent then pressed the defendant about drugs and was refused permission to search his luggage. I know many people who resent unjustified police interference and who, like the defendant, lawfully would refuse permission to search their bags even though they contain nothing illegal.
After that interview, the agent notified counterparts in Tulsa of his “suspicions.” During the defendant’s flight to Tulsa, the agent learned of the defendant’s previous arrests and of a “rap sheet” entry regarding a “mental disability.” Record, vol. 4, at 199-200. In my opinion, the cheek of the “rap sheet” adds nothing in this case unless people who have previously been arrested are excepted from fourth amendment protection.3 In addition, I am at a loss and alarmed that some justification for the intrusion seems to come from evidence that the defendant had a “mental disability.”
*1220While the contact in Atlanta, where the defendant readily identified himself, delivered his ticket, and answered questions, undoubtedly fits the euphemism of “inoffensive contact between a member of the public and the police,” United States v. Mendenhall, 446 U.S. 544, 555, 100 S.Ct. 1870, 1877, 64 L.Ed.2d 497 (1980) (Stewart, J., concurring), what happened in Tulsa cannot remotely be described as either a nonseizure or a reasonable intrusion.
As the defendant was about to leave the Tulsa airport, a DEA agent and a Tulsa police officer stopped him and took him back into the airport where they and three other uniformed officers strategically placed themselves to detain the defendant while they sent for a search dog. Even a casual observer of news events is aware of police shootings of “suspects.” One in this defendant’s place would be foolhardy to attempt to leave. Surely when we decide whether “a reasonable person would have believed that he was not free to leave,” id. at 554,100 S.Ct. at 1877,' we do not wish by the standards we employ to encourage people to get shot or otherwise have violent encounters with the police. While the defendant was held for fewer than fifteen minutes before the dog “T.J.” arrived, the detention clearly amounted to a seizure. It was intended to and did hold him until the dog arrived. Cf. Dunaway v. New York, 442 U.S. 200, 207 & n. 6, 99 S.Ct. 2248, 2253-54 & n. 6, 60 L.Ed.2d 824 (1979) (detention for custodial interrogation constitutes seizure). If any doubt as to whether the defendant was seized could exist, the lead agent’s reading the defendant his Miranda rights lays it to rest. According to the testimony — and it would be embarrassing to suggest otherwise — the defendant was not then free to go. It is at that point that the basis for the police conduct must be measured. What the dog found is therefore irrelevant although certainly not without its own problems.4
Use of the “drug courier profile” seems to me to be a particularly invidious practice, especially when it is left to the shifting subjective vagaries of the individual officer. See Vasquez, 612 F.2d at 1352-53 & n. 10. We should not begin to ratify it until both its standards and its potential for abuse have been thoroughly scrutinized. My general concerns are more elaborately and articulately spelled out in the dissent of Judge Oakes in Vasquez, 612 F.2d at 1348-53. They seem to me equally supported by the Supreme Court’s per curiam opinion in Reid v. Georgia, 448 U.S. 438, 100 S.Ct. 2752, 65 L.Ed.2d 890 (1980).
Since there was insufficient cause to justify the seizure, I would reverse the conviction.
. Although the DEA apparently does not keep statistics, one agent has estimated that “an experienced agent can have a 60-70%” successful observation rate. Comment, Reformulating Seizures — Airport Drug Stops and the Fourth Amendment, 69 Calif.L.Rev. 1486, 1488 n. 11 (1981). Even without discounting for inexperience and the tendency to overstate one’s successes, this estimate demonstrates that intrusions on innocent people are widespread.
. What little we do know from this or other records is consistent with innocent behavior. E.g., United States v. Pulvano, 629 F.2d 1151, 1155 n. 1 (5th Cir. 1980); United States v. McCaleb, 552 F.2d 717, 719-20 (6th Cir.1977).
. The “rap sheet” check did reveal an earlier drug-related offense; however, there was only one such offense for which the defendant was arrested in 1969, twelve years before the airport confrontation in this case. Record, vol. 4, at 182.
. T.J. was trained to sniff out bombs. We are of course told of his success record in ferreting out explosives even though we are carefully not told the success record of his counterpart who initiated contact in Atlanta. However, this was a search for narcotics — not explosives. It would strain credulity or worse to suggest that the officer brought the defendant back into the airport to detain him around other passengers for a period of time if they remotely suspected a bomb. Thus, it is at best questionable whether the dog’s signaling could reasonably be relied on to justify any further detention.