The question before us is whether Border Patrol agents had reasonable suspicion to stop German Espinoza Montero-Camargo and Lorenzo Sanchez-Guillen. The defendants, who were driving separate automobiles in tandem, made U-turns on a highway at the only place where the view of the agents manning a permanent stationary checkpoint was obstructed. Following the turns, the two cars, both bearing Mexicali license plates, stopped briefly in an area that is often used as a drop-off and pick-up point for undocumented aliens and contraband. The U-turns occurred shortly after the cars passed a sign stating that the previously closed Border Patrol facility was now open. Based on these and other factors, the district court concluded that the stop, which occurred some fifty miles north of the Mexican border, was justified, as did the majority of the three judge panel that considered the question. We took the case en banc to reconsider the reasonable suspicion question. Although we affirm the result reached by both the district court and the panel majority, we reject some of the factors on which they relied.

On the afternoon of October 15, 1996, a passing driver told border patrol agents at the Highway 86 permanent stationary checkpoint in El Centro, California, that two cars heading north, with Mexicali license plates,1 had just made U-turns on the highway shortly before the checkpoint. Upon receiving the tip, two Border Patrol Agents, Brian Johnson2 and Carl Fisher,3 got into separate marked patrol cars and headed south to investigate. Approximately one minute later (and about one mile from the checkpoint), the two agents saw a blue Chevrolet Blazer and a red Nissan sedan, both with Mexicali plates, pull off the shoulder and re-enter the highway heading south.4

According to the agents, the area where they first observed the cars is used by lawbreakers to drop off and pick up undocumented aliens and illegal drugs, while evading inspection. Its use for such purposes is due in part to the fact that the view of that part of the highway area from the Border Patrol checkpoint is blocked. The location, according to Agent Johnson, is the only place where it is feasible to turn around both safely and with impunity. After that point, the road narrows and is in plain view of the checkpoint. The highway itself runs through the open desert and there is a fence on either side.

Both agents testified that almost all of the stops made by the Border Patrol at the turnaround site resulted in the discovery of "a violation of some sort . . . " involving either illegal aliens or narcotics.5 In contrast, Agent Johnson said that similar stops made in connection with turnarounds near other checkpoints did not result in arrests nearly as frequently. He attributed the difference to the fact that travelers routinely miss their turnoffs to camping sites near those other checkpoints. Before the northbound Highway 86 checkpoint, however, there are no exits, driveways, or roads nearby that a driver might accidentally pass by. In fact, the only exit off of Highway 86 in that area is a private driveway to the Elmore Ranch, some two miles from the turnaround point.6

The place where the agents saw that the vehicles had stopped following the U-turn was a deserted area on the side of the southbound highway located opposite the large sign on the northbound side advising drivers that the checkpoint was open. As Agent Johnson testified, the sign was the first indication to northbound drivers that the Border Patrol's facility was operational. The checkpoint in question had been closed for some time and had reopened only a day or two earlier.

At the suppression hearing, Agent Johnson testified that the majority of people going through the El Centro checkpoint are Hispanic. This demographic makeup is typical of the larger region of which the city El Centro is a part. In Imperial County, where El Centro is located, Hispanics make up roughly 73% of the population. See U.S. Census Bureau, "Population Estimates for Counties by Race and Hispanic Origin: July 1, 1999." Agent Johnson also testified that as he pulled behind the Blazer, he noted that both the driver and the passenger appeared to be Hispanic. Johnson stated that when the driver and passenger noticed him behind them, the passenger picked up a newspaper and began reading. This, according to Agent Johnson, further aroused his suspicions. Johnson then stopped the Blazer, identified himself as a Border Patrol agent, and asked about the citizenship of the two occupants. In response to Johnson's inquiries, the driver, Lorenzo Sanchez-Guillen, and his passenger, Sylvia Renteria-Wolff, showed Agent Johnson I-586 cards, which allow Mexican citizens to travel up to 25 miles inside the United States for no longer than 72 hours at a time. As the Blazer had been stopped approximately 50 miles from the border, Johnson then brought the two occupants to the checkpoint for processing.

In the meantime, Agent Fisher continued to follow the second car, a red Nissan sedan. According to Fisher, when he and Agent Johnson first drew near the two cars, the Nissan began to accelerate. As Fisher caught up with the vehicle, he could see that the second driver also appeared to be Hispanic. Fisher ultimately pulled theNissan over after following it for approximately four miles. Appellant German Espinoza Montero-Camargo was the driver. After stopping the car, Agent Fisher, with the aid of Agent Johnson, who had returned to help him, searched the trunk and found two large bags of marijuana. A subsequent search of the Blazer back at the checkpoint turned up a loaded .32 caliber pistol in the glove compartment and an ammunition clip that fit the pistol in the passenger's purse.

Montero-Camargo, Sanchez-Guillen, and Renteria-Wolff were charged with conspiracy to possess marijuana with intent to distribute in violation of 21 U.S.C. SS 846 and 841(a)(1), as well as possession of marijuana with intent to distribute in violation of 21 U.S.C. S 841(a)(1). SanchezGuillen was also charged with being an illegal alien in possession of ammunition in violation of 18 U.S.C. S 922(g)(5) and S 924(a)(2) and aiding and abetting the carrying of a firearm during the commission of a drug trafficking crime in violation of 18 U.S.C. S 924(c)(1) and (2). The three defendants filed a pre-trial motion to suppress on the ground that the vehicle stop was not based on reasonable suspicion. When the district court denied the motion, Montero-Camargo entered a conditional guilty plea to conspiracy to possess and possession of marijuana with the intent to distribute; he reserved the right to challenge on appeal two of the district court's determinations, including the denial of the motion to suppress.7 Sanchez-Guillen went to trial, and a jury convicted him of conspiracy to possess and possession of marijuana with the intent to distribute, as well as being an illegal alien in possession of ammunition. He raises a number of issues on appeal.8

In denying the motion to suppress, the district court conceded that the government's case "was somewhat weak," but concluded that, upon considering "all the factors that the officers had in their possession at the time that each of them made the stops, . . . there was a sufficient founded suspicion to make an investigatory stop." Those factors, as the district court categorized them, included: 1) the tip about a U-turn made in the middle of the highway just before the checkpoint by two cars with Mexican license plates; 2) the alleged driving in tandem and the Mexicali license plates which supported the inference drawn by the officers that these were the two cars identified by the tipster; 3) the area in question, which, based on the officers' experience with previous stops, is "a notorious spot where smugglers turn around to avoid inspection" just before the first sign indicating that the checkpoint was in fact open; 4) the fact that the occupants of both cars appeared to be of Hispanic descent; and 5) the fact that the passenger in the Blazer picked up a newspaper as the Border Patrol car approached. The district judge concluded that when these factors were considered in light of the officers' experience, they supported a finding of reasonable suspicion.

On appeal, Montero-Camargo and Sanchez-Guillen argued that the district court erred in denying the motion to suppress.

The panel majority agreed, however, with the district court's conclusion. It did so by listing, without further explication, a number of factors,9 including: apparent avoidance of a checkpoint, tandem driving, Mexicali license plates, the Hispanic appearance of the vehicles' occupants, the behavior of Renteria-Wolff, the agent's priorexperience during stops after similar turnarounds, and the pattern of criminal activity at the remote spot where the two cars stopped.10 Although we reach the same result as both the district judge and the panel majority, we do so on the basis of a more selective set of factors.

The Fourth Amendment "applies to all seizures of the person, including seizures that involve only a brief detention short of traditional arrest." United States v. Brignoni-Ponce, 422 U.S. 873, 878 (1975). Accordingly, the Fourth Amendment requires that such seizures be, at a minimum,"reasonable." Id. In order to satisfy the Fourth Amendment's strictures, an investigatory stop by the police may be made only if the officer in question has "a reasonable suspicion supported by articulable facts that criminal activity may be afoot . . . ." United States v. Sokolow, 490 U.S. 1, 7 (1989) (internal quotation omitted) (citing Terry v. Ohio, 392 U.S. 1, 30 (1968)).

Like probable cause determinations, the reasonable suspicion analysis is "not `readily, or even usefully, reduced to a neat set of legal rules' " and, also like probable cause, takes into account the totality of the circumstances. Sokolow, 490 U.S. at 7-8 (quoting Illinois v. Gates, 462 U.S. 213, 232 (1983)). Although the level of suspicion required for a brief investigatory stop is less demanding than that for probable cause, the Fourth Amendment nevertheless requires an objective justification for such a stop. See id. at 7. As a result, the officer in question "must be able to articulate more than an `inchoate and unparticularized suspicion or `hunch' of criminal activity." Illinois v. Wardlow, 120 S. Ct. 673, 676 (2000). Rather, reasonable suspicion exists when an officer is aware of specific, articulable facts which, when considered with objective and reasonable inferences, form a basis for particularized suspicion. See United States v. Cortez, 449 U.S. 411, 418 (1981); United States v. Salinas, 940 F.2d 392, 394 (9th Cir. 1991).

The requirement of particularized suspicion encompasses two elements. See Cortez, 449 U.S. at 418. First, the assessment must be based upon the totality of the circumstances.11 Id. Second, that assessment must arouse a reasonable suspicion that the particular person being stopped has committed or is about to commit a crime. See id. ; see Terry v. Ohio, 392 U.S. at 21 n.18 ("[t]his demand for specificity in the information upon which police action is predicated is the central teaching of this Court's Fourth Amendment jurisprudence"). Accordingly, we have rejected profiles that are "likely to sweep many ordinary citizens into a generality of suspicious appearance . . . ." United States v. Rodriguez, 976 F.2d 592, 595-96 (9th Cir. 1992) (concluding that the factors cited in the case--namely, a Hispanic man carefully driving an old Ford with a worn suspension who looked in his rear view mirror while being followed by agents in a marked car-described "too many individuals to create a reasonable suspicion that this particular defendant was engaged in criminal activity"); see also United States v. Rodriguez-Sanchez, 23 F.3d 1488, 1492(9th Cir. 1994) (holding that reasonable suspicion cannot be based "on broad profiles which cast suspicion on entire categories of people without any individualized suspicion of the particular person to be stopped").

In Brignoni-Ponce, the Court listed factors which officers might permissibly take into account in deciding whether reasonable suspicion exists to stop a car. Those factors include: (1) the characteristics of the area in which they encounter a vehicle; (2) the vehicle's proximity to the border; (3) patterns of traffic on the particular road and information about previous illegal border crossings in the area; (4) whether a certain kind of car is frequently used to transport contraband or concealed aliens; (5) the driver's "erratic behavior or obvious attempts to evade officers;" and (6) a heavily loaded car or an unusual number of passengers.12 422 U.S. at 884-885. With time, however, "[s]ubsequent interpretations of these factors have created a highly inconsistent body of law," and we have given them varying weight in varying contexts. United States v. Hernandez-Alvarado, 891 F.2d 1414, 1416 (9th Cir. 1989).

As the list of factors set out in Brignoni-Ponce suggests, sometimes conduct that may be entirely innocuous when viewed in isolation may properly be considered in arriving at a determination that reasonable suspicion exists. In United States v. Sokolow, the Supreme Court held that:" `[i]n making a determination of probable cause the relevant inquiry is not whether particular conduct is `innocent' or `guilty,' but the degree of suspicion that attaches to particular types of noncriminal acts.' That principle applies equally well to the reasonable suspicion inquiry." 490 U.S. 1, 10 (1989) (citations and footnotes omitted) (quoting Illinois v. Gates, 462 U.S. 213, 243-44, n. 13 (1983)); see also United States v. FrancoMunoz, 952 F.2d 1055, 1057 (9th Cir. 1991). In short, conduct that is not necessarily indicative of criminal activity may, in certain circumstances, be relevant to the reasonable suspicion calculus. See Wardlow, 120 S. Ct. at 677. At the same time, however, innocuous conduct does not justify an investigatory stop unless there is other information or surrounding circumstances of which the police are aware, which, when considered along with the otherwise innocuous conduct, tend to indicate criminal activity has occurred or is about to take place. See People of the Territory of Guam v. Ichiyasu, 838 F.2d 353, 355 (9th Cir. 1988).

In all circumstances, "the officer is entitled to assess the facts in light of his experience in detecting illegal entry and smuggling." Brignoni-Ponce, 422 U.S. at 885. Nevertheless, "[w]hile an officer may evaluate the facts supporting reasonable suspicion in light of his experience, experience may not be used to give the officers unbridled discretion in making a stop." Nicacio v. INS, 797 F.2d 700, 705 (9th Cir. 1986), overruled in part on other grounds in Hodgers-Durgin v. De La Vina, 199 F.3d 1037, 1045 (9th Cir. 1999); see also United States v. Jimenez-Medina, 173 F.3d 752, 754 (9th Cir. 1999). In other words, an officer's experience may furnish the background against which the relevant facts are to be assessed, see Cortez, 449 U.S. at 418, as long as the inferences he draws are objectively reasonable; but "experience" does not in itself serve as an independent factor in the reasonable suspicion analysis.

2. The Factors Considered by the District Court

As noted above, the district court based its determination that reasonable suspicion existed on a series of factors: 1) the U-turn made before the checkpoint by the two cars; 2) the driving in tandem and the Mexicali license plates; 3) the area at which the U-turn occurred included a well-known drop-off point for smugglers; 4) the Hispanic appearance of the three defendants; and 5) Renteria-Wolff's picking up the newspaper after glancing back at the patrol cars.13 Although we agree with the district court that reasonable suspicion did exist to justify an investigatory stop, we conclude that some of the factors on which the district court relied are not relevant or appropriate to the reasonable suspicion analysis. We begin by considering the factors in that category, before turning to address those which the district court properly considered.

In concluding that reasonable suspicion existed, both the district court and the panel majority relied in part upon the Hispanic appearance of the three defendants. We hold that they erred in doing so. We first note that Agent Johnston testified at the suppression hearing that the majority of people who pass through the El Centro checkpoints are Hispanic, and thus, presumably have a Hispanic appearance.

As we stressed earlier, reasonable suspicion requires particularized suspicion. See Wardlow, 120 S. Ct. at 676; see also Hernandez-Alvarado, 891 F.2d at 1417; Jimenez-Medina, 173 F.3d at 754. Where, as here, the majority (or any substantial number) of people share a specific characteristic, that characteristic is of little or no probative value in such a particularized and context-specific analysis.14 See Rodriguez-Sanchez, 23 F.3d at 1492 (holding that reasonable suspicion cannot be based "on broad profiles which cast suspicion on entire categories of people without any individualized suspicion of the particular person tobe stopped"). As we put it in Rodriguez, "[w]e are not prepared to approve the wholesale seizure of miscellaneous persons . . . in the absence of well-founded suspicion based on particular, individualized, and objectively observable factors which indicate that the person is engaged in criminal activity." 976 F.2d at 596 (holding that a stop cannot be upheld where the factors tendered as justification are "calculated to draw into the law enforcement net a generality of persons unmarked by any really articulable basis for reasonable suspicion . . . .") (emphasis added).

The likelihood that in an area in which the majority--or even a substantial part--of the population is Hispanic, any given person of Hispanic ancestry is in fact an alien, let alone an illegal alien, is not high enough to make Hispanic appearance a relevant factor in the reasonable suspicion calculus.15 As we have previously held, factors that have such a low probative value that no reasonable officer would have relied on them to make an investigative stop must be disregarded as a matter of law. See Gonzales-Rivera, 22 F.3d at 1446. Moreover, as we explain below, Hispanic appearance is not, in general, an appropriate factor.

In reaching our conclusion, we are mindful of BrignoniPonce, in which, a quarter-century ago, the Supreme Court affirmed this court's decision reversing the denial of Brignoni Ponce's motion to suppress and held that a stop could not be justified by ethnic appearance alone. In that case, the Court held that "[e]ven if [Border Patrol officers] saw enough to think that the occupants were of Mexican descent, this factor alone would justify neither a reasonable belief that they were aliens, nor a reasonable belief that the car concealed other aliens who were illegally in the country." Brignoni-Ponce, 422 U.S. at 886. In a brief dictum consisting of only half a sentence, the Court went on to state, however, that ethnic appearance could be a factor in a reasonable suspicion calculus.16

In arriving at the dictum suggesting that ethnic appearance could be relevant, the Court relied heavily on now-outdated demographic information.17 In a footnote, the Court noted that:The 1970 census and the INS figures for alien registration in 1970 provide the following information about the Mexican-American population in the border States. There were 1,619,064 persons of Mexican origin in Texas, and 200,004 (or 12.4%) of them registered as aliens from Mexico. In New Mexico there were 119,049 persons of Mexican origin, and 10,171 (or 8.5%) registered as aliens. In Arizona there were 239,811 persons of Mexican origin, and 34,075 (or 14.2%) registered as aliens. In California there were 1,857,267 persons of Mexican origin, and 379,951 (or 20.4%) registered as aliens.

Brignoni-Ponce, 422 U.S. at 886 n.12. Brignoni-Ponce was handed down in 1975, some twenty-five years ago. Current demographic data demonstrate that the statistical premises on which its dictum relies are no longer applicable. The Hispanic population of this nation, and of the Southwest and Far West in particular, has grown enormously--at least five-fold in the four states referred to in the Supreme Court's decision. According to the U.S. Census Bureau, as of January 1, 2000, that population group stands at nearly 34 million. Furthermore, Hispanics are heavily concentrated in certain states in which minorities are becoming if not the majority, then at least the single largest group, either in the state as a whole or in a significant number of counties.18 According to the same data, California has the largest Hispanic population of any state--estimated at 10,112,986 in 1998, while Texas has approximately 6 million. As of this year, minorities-Hispanics, Asians, blacks and Native Americans--comprise half of California's residents; by 2021, Hispanics are expected to be the Golden State's largest group, making up about 40% of the state's population. Today, in Los Angeles County, which is by far the state's biggest population center, Hispanics already constitute the largest single group.19

One area where Hispanics are heavily in the majority is El Centro, the site of the vehicle stop. As Agent Johnson acknowledged, the majority of the people who pass through the El Centro checkpoint are Hispanic. His testimony is in turn corroborated by more general demographic data from that area. The population of Imperial County, in which El Centro is located, is 73% Hispanic. In Imperial County, as of 1998, Hispanics accounted for 105,355 of the total population of 144,051. More broadly, according to census data, five Southern California counties are home to more than a fifth of the nation's Hispanic population.20 See Dick Kirschten, The Emerging Minority, Nat'l J., Aug. 14, 1999. During the current decade, Hispanics will become the single largest population group in Southern California, see A Lesson in How to Count, The Economist, Nov. 13, 1999, and by 2040, will make up 59% of Southern California's population. Accordingly, Hispanic appearance is of little or no use in determining which particular individuals among the vast Hispanic populace should be stopped by law enforcement officials on the lookout for illegal aliens.21 Reasonable suspicion requires particularized suspicion, and in an area in which a large number of people share a specific characteristic, that characteristic casts too wide a net to play any part in a particularized reasonable suspicion determination.22

Moreover, the demographic changes we describe have been accompanied by significant changes in the law restricting the use of race as a criterion in government decision-making. The use of race and ethnicity for such purposes has been severely limited. See Adarand Constructors v. Pena, 515 U.S. 200 (1995); J.A. Croson v. City of Richmond, 488 U.S. 469 (1989). Relying on the principle that " `[o]ur Constitution is color-blind, and neither knows nor tolerates classes among citizens,' " Croson, 488 U.S. at 521 (Scalia, J., concurring) (quoting Plessy v. Ferguson, 163 U.S. 537, 559 (1896) (Harlan, J., dissenting)), the Supreme Court has repeatedly held that reliance "on racial or ethnic criteria must necessarily receive a most searching examination to make sure that it does not conflict with constitutional guarantees. " Wygant v. Jackson Bd. of Ed., 476 U.S. 267, 273 (1986) (quoting Fullilove v. Klutznick, 448 U.S. 448, 491 (1980)). In invalidating the use of racial classifications used to remedy past discrimination in Croson, the Court applied strict scrutiny, stating that its rigorousness would ensure that:

the means chosen `fit' this compelling goal so closely that there is little or no possibility that the motive for the classification was illegitimate racial prejudice or stereotype. Classifications based on race carry a danger of stigmatic harm. Unless they are strictly reserved for remedial settings, they may in fact promote notions of racial inferiority and lead to a politics of racial hostility.

Croson, 488 U.S. at 493. The danger of stigmatic harm of the type that the Court feared overbroad affirmative action programs would pose is far more pronounced in the context of police stops in which race or ethnic appearance is a factor. So, too, are the consequences of "notions of racial inferiority" and the "politics of racial hostility" that the Court pointed to. Stops based on race or ethnic appearance send the underlying message to all our citizens that those who are not white are judged by the color of their skin alone. Such stops also send a clear message that those who are not white enjoy a lesser degree of constitutional protection--that they are in effect assumed to be potential criminals first and individuals second.23

It would be an anomalous result to hold that race may be considered when it harms people, but not when it helps them.24

We decide no broad constitutional questions here. Rather, we are confronted with the narrow question of how to square the Fourth Amendment's requirement of individualized reasonable suspicion with the fact that the majority of the people who pass through the checkpoint in question are Hispanic. In order to answer that question, we conclude that, at this point in our nation's history, and given the continuing changes in our ethnic and racial composition, Hispanic appearance is, in general, of such little probative value that it may not be considered as a relevant factor where particularized or individualized suspicion is required. Moreover, we conclude, for the reasons we have indicated, that it is also not an appropriate factor.25

We now turn to another factor on which the United States relies, namely Renteria-Wolff's behavior. Both the district court judge as well as the panel majority concluded that Renteria-Wolff's behavior--more specifically, her picking up a newspaper after glancing at the patrol car in the rear-view mirror--was a relevant factor in the reasonable suspicion analysis.We disagree. In general, although eye contact, or the lack thereof, may be considered as a factor establishing reasonable suspicion, we have noted that whether the contact is suspicious or not "is highly subjective and must be evaluated in light of the circumstances of each case." United States v. Robert L., 874 F.2d 701, 703 (9th Cir. 1989); see also United States v. Magana, 797 F.2d 777, 781 (9th Cir. 1986); United States v. Pulido-Santoyo, 580 F.2d 352, 354 (9th Cir. 1978). The skepticism with which this factor is treated is in large part due to the fact that reliance upon "suspicious " looks can so easily devolve into a case of damned if you do, equally damned if you don't. See Gonzales-Rivera v. INS , 22 F.3d 1441, 1446-47 (9th Cir. 1994); Nicacio, 797 F.2d at 704; see also United States v. Mallides, 473 F.2d 859, 861 n.4 (9th Cir. 1973) (collecting cases). Accordingly, we have noted that that factor is "of questionable value . . . generally."26 United States v. Munoz, 604 F.2d 1160, 1160 (9th Cir. 1979) (per Kennedy, J.).

In this case, Agent Johnson testified that, as he approached the Blazer from behind, he observed that Renteria-Wolff appeared to glance quickly in the rear view mirror before picking up a newspaper and reading it. It is unclear from the record whether Johnson could in fact have seen such a glance as he drove up behind the Blazer. In any event, it is a common, if not universal, practice for drivers and passengers alike to take note of a law enforcement vehicle coming up behind them. In fact, the most law-abiding of citizens frequently adjust their driving accordingly.

Further, we give no weight to the fact that Sylvia Renteria-Wolff picked up a newspaper after glancing at the patrol car. Agent Johnson did not suggest that by this action she sought to conceal her face so that he would not recognize her. Had Renteria-Wolff continued to keep her eyes on the patrol car behind them after her initial glance, Agent Johnson might well have found it equally suspicious--because she paid too much, rather than too little attention to him. It is, in fact, difficult to imagine what Renteria-Wolff could have done at that point that might not have appeared suspicious to a Border Patrol agent.27 It is for this very reason that we reached the conclusion we did in then-Judge Kennedy's opinion in Munoz.

We recognize that in its recent decision in Wardlow, the Supreme Court noted that evasive behavior may be a "pertinent factor in determining reasonable suspicion." Wardlow, 120 S. Ct. at 674. However, nothing in Wardlow --or the three Supreme Court cases it cites to illustrate that proposition-runs contrary to our conclusion that Renteria-Wolff's conduct provides no basis for reasonable suspicion. The three earlier cases all involved obvious, unambiguous attempts to evade contact with law enforcement officials--conduct very different from what was observed by the Border Patrol agent as he followed the car in which Renteria-Wolff was riding. In the first case, namely Brignoni-Ponce, the Supreme Court categorized evasive behavior as "obvious attempts to evade officers" or to hide. See Brignoni-Ponce, 422 U.S. at 885 (emphasis added). In the second case, Sokolow, the Court held that evidence that the suspect took an evasive or erratic path through an airport in an apparent attempt to avoid police might also be relevant to the reasonable suspicion determination. See Sokolow, 490 U.S. at 8. In the third, Florida v. Rodriguez, 469 U.S. 1, 6 (1984) (per curiam), the Supreme Court held that articulable suspicion existed where threemen spoke furtively among themselves after seeing officers approaching them, where one was twice overheard during that conversation urging the others to "get out of here," and where one of the three, Rodriguez, in fact turned around and attempted to flee. See id. at 6. As noted above, all three cases described actual--and obvious--attempts to evade or to hide from law enforcement officers. Moreover, Wardlow itself, of course, involved headlong flight, which the Court termed "the consummate act of evasion . . . ." Wardlow, 120 S. Ct. at 674. We do not mean to suggest that these cases establish the outer parameters for what is evasive behavior. Rather, we conclude only that glancing in a rear view mirror and then picking up a newspaper to read is not. Such actions are simply not the sort of evasive conduct that the Supreme Court has held is properly part of the reasonable suspicion calculus, nor did the officers suggest that it was the type of behavior they had observed in the past when wrongdoing was afoot. Accordingly, we conclude that, like Hispanic appearance, Renteria-Wolff's behavior was not a relevant or appropriate factor to consider in determining reasonable suspicion.

The question then is what factors are both relevant and appropriate to the reasonable suspicion analysis in this case. Those factors are, in a certain sense, interwoven, and they draw their significance, in part, from one another. The first of these factors to consider is the U-turn or turnaround.28 In United States v. Ogilvie, 527 F.2d 330, 332 (9th Cir. 1975), this Court held that "turning off the highway and turning around [are] not in themselves suspicious . . . ." Accordingly, "the proximity of the turn to the checkpoint, regardless of the legality of the checkpoint, [is] not a sufficient foundation on which to rest reasonable suspicion." Id. (emphasis added).

In the panel decision, the majority and the dissenting judge disagreed as to whether Ogilvie prohibited reliance on the U-turn in this case. We side with the majority and conclude that it does not. Ogilvie simply holds that a turnaround alone is not enough in and of itself to create reasonable suspicion,29 see Ogilvie, 527 F.2d at 332. Indeed, in subsequent decisions, this Court has made it clear that a turnaround combined with other factors may be considered as part of a reasonable suspicion analysis. See United States v. GarciaBarron, 116 F.3d 1305, 1307 (9thCir. 1997) ("[a]pparent efforts to avoid checkpoints combined with other factors have generally been found to constitute `reasonable suspicion' "); United States v. Medina-Gasca, 739 F.2d 1451 (9th Cir. 1984). Even more so, a U-turn.

In concluding that the U-turn in this case constitutes a significant factor, we note a number of circumstances that combine to make it so, some of which also constitute independent factors in the reasonable suspicion analysis. First, a U-turn on a highway is very different from reversing direction by using a designated highway exit. The use of a highway exit is both frequent and legal; in contrast, a U-turn on a highway is unusual and often illegal. While it is not clear whether the U-turn here was legal, the other surrounding circumstances render the reversal-in-direction one that may properly be given significant weight in our reasonable suspicion analysis.30 One of those circumstances is the fact that the two cars made their U-turn and immediately stopped at the side of the highway in an isolated, desert area frequently used to drop off or pick up undocumented aliens or contraband. Another is that the U-turn occurred just after a sign indicating that an upcoming checkpoint had been re-opened. Finally, it is highly unlikely that the reason for the U-turn was that the cars had accidentally passed their exit point. There is only one turn-off anywhere in the area before the checkpoint, and that turn-off leads to a private road rather than one that members of the general public might use.

We also rely on the characteristics of the area in which the cars stopped after the reversal-in-direction as an independent factor. We note initially that an individual's presence in a high crime area is not enough to support reasonable, particularized suspicion that the individual in question has committed or is about to commit a crime. See Brown v. Texas, 443 U.S. 47, 52 (1979) (holding that an investigatory stop was not justified when police officers detained two men walking away from each other in an alley in an area with a high rate of drug trafficking because "the appellant's activity was no different from the activity of other pedestrians in that neighborhood"). Still, "officers are not required to ignore the relevant characteristics of a location in determining whether the circumstances are sufficiently suspicious to warrant further investigation." Wardlow, 120 S. Ct. at 676 (quoting Adams v. Williams, 407 U.S. 143, 144, 147-48 (1972)).

The citing of an area as "high-crime" requires careful examination by the court, because such a description, unless properly limited and factually based, can easily serve as a proxy for race or ethnicity. District courts must carefully examine the testimony of police officers in cases such as this, and make a fair and forthright evaluation of the evidence they offer, regardless of the consequences. We must be particularly careful to ensure that a "high crime" area factor is not used with respect to entire neighborhoods or communities in which members of minority groups regularly go about their daily business, but is limited to specific, circumscribed locations where particular crimes occur with unusual regularity.31 In this case, the "high crime" area is in an isolated and unpopulated spot in the middle of the desert. Thus, the likelihood of an innocent explanation for the defendants' presence and actions is far less than if the stop took place in a residential or business area.32

Finally, we consider the tandem driving as well as the Mexicali license plates. The panel majority treated the two as independent factors giving rise to reasonable suspicion. In contrast, the district court took a different approach, relying on the "tandem driving" and Mexicali license plates solely for the purposes of linking the cars described by the tipster to the ones observed by the Border Patrol officers. We conclude that, under the circumstances present here, both occurrences may be given some direct weight in the reasonable suspicion analysis. They do not, however, constitute substantial factors, either singly or collectively.

With respect to tandem driving, we have held that two or more cars traveling together, although not sufficient in itself to establish reasonable suspicion, may nonetheless "be indicative of illegal smuggling activity." See United States v. Robert L., 874 F.2d 701, 704 (9th Cir. 1989); see also United States v. Medina-Gasca, 739 F.2d 1451, 1453 (9th Cir. 1984); United States v. Saenz, 578 F.2d 643, 646-47 (9th Cir. 1978); United States v. Larios-Montes, 500 F.2d 941, 943-44 (9th Cir. 1974). However, the circumstances of the tandem driving will in the end determine whether that factor is relevant. While two cars driving together is intrinsically innocuous, here the fact that the two cars not only turned around in tandem in the middle of a highway, but then pulled off the shoulder together and stopped where criminal activity often took place, makes the tandem driving relevant. The fact that the cars had Mexicali license plates may also provide some additional weight, given all the other circumstances. While having Mexican plates is ordinarily of no significance, where the criminal act suspected involves border-crossing, the presence of foreign license plates may be afforded some weight in determining whether a stop is reasonable.

In this case, the two cars driven in tandem by MonteroCamargo and Sanchez-Guillen made U-turns on a highway, at a place where the view of the border officials was obstructed, and stopped briefly at a locale historically used for illegal activities, before proceeding back in the direction from which they had come. The U-turn occurred at a place at a location where it was unlikely that the cars would have reversed directions because they had missed an exit. Moreover, the vehicles in question bore Mexicali license plates and the U-turn occurred just after a sign indicating that a Border Patrol checkpoint that had been closed for some time was now open. We conclude that these factors, although not overwhelming, are sufficient to constitute reasonable suspicion for the stop. In reaching that result, however, we firmly reject any reliance upon the Hispanic appearance or ethnicity of the defendants. We also do not consider Renteria-Wolff'sbehavior in glancing at the Border Patrol car in the rear view mirror and then picking up and reading a newspaper.

In affirming the district court's ruling, we note that the agents' initial decision to investigate the tip and to pursue the two vehicles was made without any knowledge on their part of the defendants' ethnicity or Hispanic appearance. Agents Johnson and Fisher observed that appearance only when the officers subsequently caught up with the defendants' cars. Moreover, the agents had enough information to justify the stop before they became aware of the defendants' likely ethnicity. Under these circumstances, there is no need to remand the matter to the district court for reconsideration of its decision. Instead, we AFFIRM the district court's denial of the motion to suppress.

KOZINSKI, Circuit Judge, with whom Judges T.G. NELSON, KLEINFELD and SILVERMAN join, concurring:

What happened in this case is perfectly clear. It is revealed in a direct answer to a simple question from the district court:

THE COURT: Right. But why did you stop [the defendants] at that time?

THE WITNESS [AGENT JOHNSON]: Well, we stopped them --there's only--the only reason that we--especially on that side that we have the people stop and turn around at that particular point, is because they're in violation of some immigration or some criminal code that make[s] them not want to be inspected by our checkpoint. So they try to turn around at that point and head back up.1

That's the whole story: The border patrol agents stopped the two cars in which defendants were travelling because they had turned around just short of the Highway 86 checkpoint, which raised the entirely plausible inference that they were up to no good. Everything else the agents said had contributed to the stop--the license plates, the newspaper, the acceleration, the tandem driving, the Hispanic appearance, the furtive glance--is window dressing, designed to get around our opinion in United States v. Ogilvie, 527 F.2d 330 (9th Cir. 1975), where we held that avoiding a checkpoint by reversing direction is not sufficient to establish reasonable suspicion. See id. at 332.

Ogilvie is just plain wrong and we should overrule it. Turning in one's tracks just before reaching a law enforcement checkpoint is precisely the kind of behavior that properly gives rise to reasonable suspicion. It is possible that a motorist will do so for entirely legitimate reasons, but "commonsense judgments and inferences about human behavior" suggest that the maneuver was designed to avoid the checkpoint. Illinois v. Wardlow, 120 S. Ct. 673, 676 (2000). Ogilvie seems to require that the motorist have done something more to arouse suspicion, like disobeying the traffic laws or driving erratically, 527 F.2d at 332, but the opinion doesn't explain why any of those things would make it more likely that the motorist was trying to evade the checkpoint. After all, a motorist wishing to avoid police scrutiny will slink away as unobtrusively as possible, not peel rubber and disappear in a cloud of dust.

The majority retains Ogilvie and tries to distinguish it, but its efforts are sadly unconvincing. The first ground on which the majority tries to distinguish Ogilvie is the so-called U-turn factor: Ogilvie reversed directions by means of an exit ramp and overpass, whereas the defendants here used that reckless and exotic maneuver--the U-turn. See Maj. Op. at 4017. In fact, the majority is so taken with the idea that defendants made a U-turn, it uses the term no fewer than 15 times.

But nowhere in the record does any witness say that defendants made a U-turn. Agent Johnson testified that "all [the passing motorist] said was there was acouple of vehicles turning around south of the [check]point." RT at 17 (Dec. 23, 1996).2 A U-turn, in the technical sense the majority uses that term, see Maj. Op. at 4017 & n.30, starts and ends on the highway, with the vehicle winding up in a lane moving in the opposite direction. By contrast, a turnaround can be accomplished by pulling off the road, changing direction in the offroad area, and then driving back onto the highway heading in the opposite direction.3 This would be consistent with the testimony of the agents that when they first came upon the two vehicles, they were off the highway, just making their way back onto it. See RT at 14 (Dec. 23, 1996); id. at 7 (Jan. 6, 1997).4

Why does it matter? Because once we get rid of the "unusual and often illegal" U-turn, Maj. Op. at 4017, the difference between this case and Ogilvie evaporates. Ogilvie found herself on a freeway, facing a police barricade:

Four marked government vehicles were parked on the median strip separating northbound and southbound lanes of I-19; cone shaped red markers were placed to cause northbound traffic to merge to a single lane; two stop signs with flashing red lights were placed beside the single lane. All cars going north were being stopped.

527 F.2d at 331. She used the only path open to her to avoid the checkpoint--she took an exit, traversed the overpass and merged into the southbound lane.

What the defendants did here is directly analogous, given the nature of the road they were on. They were travelling on a highway, not a freeway, so there were no exits or overpasses. When they spotted the sign indicating that the checkpoint was open, they took advantage of "the only place where it's really feasible to turn around safely" before the checkpoint. RT at 16 (Dec. 23, 1996). This is just what Ogilvie did. For her it meant using the exit; for them it meant pulling off the road, turning around, then getting back onto the highway. The majority's attempt to salvage Ogilvie by painting the turnaround here as somehow dangerous or illegal falls of its own weight.

Even less convincing is the majority's attempt to invoke "characteristics of the area" as a distinguishing factor. Maj. Op. at 4018. To read the majority opinion, one might get the impression that the area just south of the checkpoint is a combat zone rivaling Prohibition-era Chicago. The majority does not pause to query why a barren stretch of highway in the middle of the desert would become such an active center of criminal activity, nor why criminals would be so stupid as "to drop off or pick up undocumented aliens or contraband," id., within a mile of a border checkpoint, when they have countless miles of road where they could safely make such exchanges.

The answer to this mystery is that the description of the area as "high crime" isn't supported by the record either. What the two arresting agents said is this: On several prior occasions, they had stopped vehicles that had done precisely what these defendants had done--reversed direction at or near the spot where theyfirst became aware that the checkpoint was operational--and, in almost every case, the vehicles were carrying contraband or aliens. The agents did not testify that they had ever apprehended anyone actually using this spot as a drop off point.5 They watched the spot routinely because, "That's where we expect the turnaround to be, if they do turn around, because that's the only place where it's really feasible to turn around safely, at that point." RT at 16 (Dec. 23, 1996).

What this evidence does show--the only thing it shows--is that Ogilvie is bunkum: People who turn around right before a checkpoint generally do have something to hide. Far from distinguishing Ogilvie, the majority's emphasis on the "high crime area" illustrates just how silly Ogilvie is. The only thing about this "area" that leads to a high incidence of arrests is the presence of the checkpoint, which prompts criminals to reveal themselves by turning around--just as happened in Ogilvie. Had the Ogilvie checkpoint stayed in place long enough to catch a few more people, that mild-mannered stretch of I-19 might have turned into a "high crime area" as well.

The majority's contorted efforts to preserve an ancient and ill-conceived precedent would be amusing, were this not such serious business. What factors law enforcement officers may consider in deciding to stop and question citizens minding their own business should, if possible, be carefully circumscribed and clearly articulated. When courts invoke multifactor tests, balancing of interests or fact-specific weighing of circumstances, this introduces a troubling degree of uncertainty and unpredictability into the process; no one can be sure whether a particular combination of factors will justify a stop until a court has ruled on it. It also creates an incentive for officers to exaggerate or invent factors, just to make sure that the judges who review the case will approve their balancing act.

I understand that it's not always possible to eliminate uncertainty, and that weighing and balancing is the stuff of many legal doctrines. But what excuse is there for resorting to a totality-of-the-circumstances approach when a single factor--the turnaround right before the checkpoint--alone justifies the search? And what excuse is there for language like this, which calls for an advanced degree in philology to comprehend:

We conclude that, under the circumstances present here, both occurrences [tandem driving and Mexicali license plates] may be given some direct weight in the reasonable suspicion analysis. They do not, however, constitute substantial factors, either singly or collectively.

Maj. Op. at 1139. What on earth does this mean? First, consider all the hedges and qualifiers ("under the circumstances present here . . . some direct weight . . ."). What purpose do they serve? What guidance do they give? There are so many slippery surfaces, the human mind can find no purchase in wrapping itself around it.

But then ponder the meaning of the entire passage: These factors may be given some weight, but they are not substantial factors. So we not only have a multifactor test, not only do we ask district courts and police in the field to weigh and balance all the factors, we now have different classes of factors--regular and jumbo. How many regular factors add up to make a substantial factor? And how many substantial factors amount to reasonable suspicion? I have no clue, which makes me think that cops on their beats all over this circuit will have some trouble figuring it out as well.6

But there is a darker side to the majority's verbal Macarena. The opinion recognizes the danger in allowing the police to characterize an area as "high-crime " to establish a basis for reasonable suspicion, but then proceeds to do just that, based on nothing more than the personal experiences of two arresting agents. As I discuss above, the agents didn't even claim this was a high crime area, but let's say they had. What in this record would support their conclusion? Both agents testified only that they had detected criminal violations after stopping people in the area. How often? One agent said he'd been involved in 15-20 stops over eight and a half years, and "[could]n't recall any . . . where we didn't have a violation of some sort." RT at 14-15 (Dec. 23, 1996). The other agent testified to "about a dozen" stops in the same period, all but one of which led to an arrest. Id. at 3, 30 (Jan. 6, 1997).

Without hesitation, the majority treats this as a crime wave, but is it really? Does an arrest every four months or so make for a high crime area? Compare United States v. Thornton, 197 F.3d 241, 248 (7th Cir. 1999) ("In less than one year there had been some 2,500 drug arrests in the five-block-byfive-block area where the incident occurred."); United States v. Morales, 191 F.3d 602, 604 (5th Cir. 1999) ("In the past year alone, the Agent had detained approximately 600 illegal aliens on this stretch of the highway."). Can we rely on the vague and undocumented recollections of the officers here? Do the two officers' figures of "15-20" and "about a dozen" reflect separate pools of incidents, or do they include some where, as here, both officers were involved? Are such estimates sufficiently precise to tell us anything useful about the area? I wouldn't have thought so, although I could be persuaded otherwise. But my colleagues don't even pause to ask the questions. To them, it's a high crime area, because the officers say it's a high crime area.

Just as a man with a hammer sees every problem as a nail, so a man with a badge may see every corner of his beat as a high crime area. See, e.g., Price v. Kramer, 200 F.3d 1237, 1247 (9th Cir. 2000) (police officers sought to justify stop on the ground that Crenshaw Boulevard in Torrance was a "high crime area known for `gang activity' "). Police are trained to detect criminal activity and they look at the world with suspicious eyes. This is a good thing, because we rely on this suspicion to keep us safe from those who would harm us. But to rely on every cop's repertoire of war stories to determine what is a "high crime area"--and on that basis to treat otherwise innocuous behavior as grounds for reasonable suspicion-strikes me as an invitation to trouble. If the testimony of two officers that they made, at most, 32 arrests during the course of a decade is sufficient to turn the road here into a high crime area, then what area under police surveillance wouldn't qualify as one? There are street corners in our inner cities that see as much crime within a month--even a week. I would be most reluctant to give police the power to turn any area into a high crime area based on their unadorned personal experiences. I certainly would not reach out to decide the issue.

The majority purports to draw support for its methodology from the Supreme Court's Wardlow opinion, where the Court held that "relevant characteristics of a location " may be considered in the reasonable suspicion calculus. Maj. Op. at 4018 (quoting Wardlow, 120 S. Ct. at 676). This misses the point entirely. The question is not whether the characteristics of the area may be taken into account, but how these characteristics are established. In our first opinion to interpret this language from Wardlow, the majority adopts a methodology for establishing the characteristics of the area that is about as rigorous as the recipe for Leftovers Casserole.

Not to worry, the opinion says, because we take the police at their word only when it's an "isolated" area where people don't "typically carry on legitimate activities. " Maj. Op. at 4019 n.32. But we're talking here about a highway--one that sees enough traffic to make a checkpoint worth the government's trouble. How can an area traversed by hundreds, perhaps thousands, of people every day be considered "isolated," and why is it that people who travel on rural highways do not "carry on legitimate activities"? If we're willing to ignore all the cars and people, I suppose the Santa Monica Freeway is an isolated area too. Then again, perhaps the majority considers the shoulder of the highway as "isolated" where the highway itself is not. A more contrived distinction is difficult to imagine.

Perhaps the majority imagines that by retaining Ogilvie, it avoids eroding Fourth Amendment liberties. Instead it accelerates the process. The patently suspicious behavior in Ogilvie is the kind of thing we want police to act on: It's a specific event, easy to observe and verify. It consists of a deliberate act by a given individual, and does not require any speculation, impressions, hunches or broad characterizations about an area or a class of people. It gives rise to a reasonable inference that the individual taking the evasive action has cause to fear scrutiny. It's not a perfect inference; they never are. But it's vastly more objective and less open to manipulation than the amalgamation of ambiguous factors from which the majority constructs founded suspicion today. Such foundations are not of concrete, but quicksand.