OPINION
¶ 1 Daniel Magner (“defendant”) appeals from his convictions and sentences for possessing and transporting marijuana for sale. He asserts his convictions are founded upon evidence seized as a result of an illegal detention of his person. We agree and therefore reverse.
BACKGROUND
¶ 2 On Tuesday, November 28, 1995, Department of Public Safety Officer Jeff Brownlee (“Brownlee”) stopped defendant on Interstate 40 east of Flagstaff for driving 71 m.p.h. in a 65 m.p.h. zone. During the stop, Brownlee made several observations of defendant and his vehicle. Based upon his training and the inferences he drew from those observations, Brownlee became suspicious that defendant was involved in transporting drugs. After issuing defendant a written warning for speeding and returning his driver’s license and registration, Brown-lee requested defendant’s permission to search the vehicle. When defendant refused, Brownlee informed defendant that he was not free to leave, ordered him from the car, and called for a drug detection dog to come to the scene.
¶3 Because the Flagstaff dog was unavailable, a dog from Williams was called but did not arrive for approximately forty-five minutes. The dog was put to work and alerted on defendant’s car, at which point Brownlee placed defendant under arrest. Defendant surrendered his keys, and the officer opened the car trunk and discovered marijuana inside.
¶4 A grand jury indicted defendant for possession of more than four pounds of marijuana for sale and transportation of more than two pounds of marijuana for sale, each a class 2 felony. Defendant moved to suppress the marijuana, arguing, inter alia, that he was detained without reasonable suspicion of criminal activity, and that he was detained for an unreasonable amount of time. After *396 an evidentiary hearing, the trial court denied the motion.
¶ 5 Defendant and the state agreed to a bench trial based upon a stipulated record. The trial court found defendant guilty on both counts and sentenced him to concurrent, minimum terms of four years imprisonment. Defendant timely appealed.
ISSUES
¶ 6 At the outset, we note that defendant concedes that Brownlee’s traffic stop was proper and defendant also does not dispute the propriety of his arrest following the dog’s alert. The first question defendant raises is whether reasonable suspicion supported the officer’s detention of defendant in order to secure the drug dog. The second question is whether the period of his detention while awaiting the arrival of the drug dog was excessive.
ANALYSIS
I. The Detention of Defendant
A. The Evidence
57 7 To justify detaining a person for investigative reasons, an officer must be able to articulate specific facts which, together with rational inferences from those facts, reasonably warrant the suspicion that the defendant had committed, or was about to commit, a crime.
State v. Killean,
¶ 8 At the evidentiary hearing in this matter, Brownlee testified that he had more than eight years of experience in drug interdiction and had attended numerous drug interdiction courses conducted by law enforcement agencies. He then articulated several specific observations he made of defendant, together with the inferences he drew from those observations.
¶ 9 Brownlee observed that defendant generally refused eye contact, that defendant’s eye twitched the one time he did make contact, and that defendant was unusually upset at being stopped. Brownlee concluded from this behavior that defendant was nervous. He observed that defendant had his car registration on the passenger seat, which caused Brownlee to wonder whether defendant had a gun in the glove box. He observed that defendant’s wearing of a tie was inconsistent with the jeans and sneakers he was also wearing, and that a tie was unusual for a cross-country trip. He concluded that defendant was attempting to present as a businessman to any passing patrolman, thus lessening his chances of being stopped. Through questioning defendant, Brownlee learned that defendant was traveling to Arkansas from Tucson, a city he knew to be a source for illegal drugs, and defendant’s account of his stay in Tucson seemed confusing to Brownlee. He observed that defendant’s car was dirty, a factor which the officer knew from his training was consistent with “people who are involved with criminal activity” who “travel from Point A to Point B as fast as they can” without cleaning their cars. Finally, the officer noted a black overnight bag in the back seat and inferred that since most people like to keep their luggage in a safe place, defendant risked placing this bag on the seat to keep the contents of the trunk hidden. To Brownlee, the cumulative effect of these observations and inferences raised a suspicion that defendant was transporting illegal drugs.
¶ 10 By its denial of defendant’s motion, the trial court found as true Brownlee’s observations and also found rational the inferences Brownlee drew from those observations. Based on the totality of these factors, the trial court concluded that Brownlee had reasonable suspicion to detain defendant.
B. The Standard of Review
¶ 11 A trial court’s ruling on the propriety of an investigative detention presents a reviewing court with a mixed question of law and fact and, consequently, with a “mixed” standard of review. We review
de novo
the ultimate question whether the total
*397
ity of the circumstances amounted to the requisite reasonable suspicion.
State v. Gonzalez-Gutierrez,
¶ 12 In this matter, Brownlee was the only witness and the trial court was not required to make any conflicting credibility determinations. We therefore find no abuse of discretion in the trial court’s accepting as accurate the observations that Brownlee made and also accepting that he drew the inferences to which he testified. However, since it is only rational inferences that can contribute to an ultimate finding of reasonable suspicion,
Terry v. Ohio,
C. The Individual Factors
Defendant’s Nervousness
¶ 13 Brownlee explained that he thought defendant was nervous because defendant generally did not make eye contact with him, and defendant’s eye “really started twitching” during the one instance when there was eye contact. Brownlee further explained that defendant was unusually upset at being stopped. According to the officer, motorists who are stopped for a traffic violation normally are upset, but not to the extent of refusing eye contact or attempting some degree of congeniality.
¶ 14 We agree that the officer’s observations supported a conclusion of some nervousness on defendant’s part. However, courts must be wary of granting much weight to a law enforcement officer’s subjective observation that a defendant was nervous.
United States v. Fernandez,
¶ 15 However, “dramatic” indications of nervousness may contribute substantially to a suspicion of criminal activity.
United States v. Green,
¶ 16 We do not believe that defendant’s actions fall into the “dramatic” category. “[AJvoidance of eye contact is consistent with innocent as well as criminal behavior.”
State v. Robinson,
The Vehicle Registration
¶ 17 When Brownlee observed the registration on the passenger seat, he concluded that defendant kept it there in order to avoid opening his glove box, raising the suspicion in Brownlee’s mind that defendant had a gun there. The difficulty with Brown-lee’s conclusion, however, is that it cannot be arrived at without making assumptions not supported by any evidence. First, Brownlee apparently assumed that defendant’s registration form was always kept on the seat, rather than removed for the stop, and never questioned defendant on this point. For all we can tell from this record, defendant could have been keeping it in a door pocket or any other compartment in the car within quick and easy reach, and retrieved it as he was being pulled over. Brownlee’s second assumption is that registration forms are always kept in the glove box, or, as he also observed, in a person’s wallet. However, common sense and experience do not support this assumption, and again, Brownlee did not ask defendant where he normally kept the form. Thus, while Brownlee’s assumptions could have been correct, his inference based on those assumptions is considerably weakened by the lack of any evidentiary support for them.
Defendant’s Attire
¶ 18 Brownlee inferred that defendant sought to deflect suspicion by wearing a tie to project an image as a businessman. The officer felt that the tie was not typical for a long drive from Tucson to Arkansas. He based this inference, not on his drug detection training, but on his experience as an officer patrolling “the road.”
¶ 19 We do not find anything suspicious solely from the combination of a tie with jeans and sneakers. Such attire is de rigueur in some current fashion circles and would not be unusual if found on a person stopped on a city street. Nevertheless, we do agree that, in the context of a cross-country trip, wearing a tie is unusual. However, the officer did not inquire of defendant why he was so attired, losing the opportunity to either dissipate or strengthen his suspicion. Moreover, the officer did not testify from his drug interdiction training that attire such as defendant’s has been shown to be typical of drug couriers.
Tucson as a “Source City”
¶ 20 Brownlee did not explain how he knew that Tucson was a “source city” for drugs. However, it is not unreasonable to infer that based on his experience, this was factually accurate and was known to him. As a component of reasonable suspicion, traveling to or from a “source city” for drugs is sometimes recognized as a legitimate factor when assessing reasonable suspicion.
See Green,
¶ 21 However, this factor has no weight standing alone since obviously very many innocent persons travel to and from Tucson and other source cities.
See Reid v. Georgia,
*399
2754,
¶22 Here, the officer did ask further questions of defendant regarding his reasons for traveling from a source city. Defendant explained to Brownlee that he had been in Tucson to visit family for Thanksgiving. Brownlee, however, thought defendant’s answer was “confusing.” Defendant stated he had been in Tucson for “a couple days” whereas Brownlee observed that the time of the stop was some five days after Thanksgiving. In Brownlee’s opinion, this raised further suspicion regarding defendant’s travel from Tucson.
¶ 28 We are unable to discern why the officer was confused. As reported by the officer, defendant’s answer was responsive to the officer’s question. When asked where he had been, defendant responded he had visited his family in Tucson for “a couple days” for Thanksgiving. He was not asked nor did he say what he meant by a “couple” of days; he was not asked nor did he say what the dates of his particular holiday period were and whether that period encompassed Thanksgiving Day; he was not asked nor did he say whether he had made any other stops after leaving Tucson to, for example, sightsee or visit friends. In short, the officer’s confusion was his own, not caused by defendant, and could easily have been either dispelled or confirmed by a few further questions of defendant. As far as this record is concerned, defendant’s explanation for being in Tucson was plausible.
The Dirty Car
¶24 Brownlee explained that his “training and experience” taught him that people involved in criminal activity “travel from Point A to Point B as fast as they can” without taking time to clean their ears. Although Brownlee described defendant’s car as dirty and unkempt, the only specific indicator to which he testified was an overflowing ashtray. Nevertheless, he concluded that this was sufficient “dirt” to warrant the inference that defendant was in the category of people involved in criminal activity.
¶ 25 While the officer’s general inference about dirty cars rationally can be drawn, it is not any more rational than a conclusion that the dirty, unkempt appearance of a car suggests that the occupant is lazy or someone who chooses to wait to clean the vehicle until the trip is concluded, knowing that continued travel will simply clutter the vehicle again.
See State v. Lovegren,
¶ 26 Moreover, the weakness of the “dirty car” observation is illustrated by the fact that an officer’s suspicion can be aroused by a vehicle that looks “too clean.”
See United States v. Baron,
The Overnight Bag
¶ 27 Brownlee explained that the overnight luggage bag in the back seat was indicative of criminal activity because “most people who travel like to secure their items in a safe place.” From this, Brownlee in *400 ferred that defendant was hiding something in his trunk.
¶28 Again, the inference drawn by the officer is one that can be drawn. However, there are other plausible alternative explanations for leaving an overnight bag in a back seat. For example, a traveler may do so if he wanted easy access to items such as a shaving kit or toothbrush. Alternatively, a motorist’s trunk might be loaded to capacity with other legitimate travel items. Indeed, Brownlee admitted that it was “not uncommon for people who are traveling across country to have ... personal items in the car with them.” Brownlee asked no questions of defendant regarding the bag, and we therefore cannot know whether the explanation would have been plausible, indicating innocence, or implausible, indicating criminal activity.
D. Totality of the Circumstances
¶ 29 In conducting our
de novo
review evaluating the totality of the circumstances, we do not have a mathematical formula into which we can insert these circumstances and calculate a proper result.
See State v. Stricklin,
¶ 30 As we have already indicated, we do not find an abuse of discretion in the trial court’s finding that Brownlee made the observations to which he testified. Moreover, we do not find an abuse of discretion in the trial court’s finding that the inferences the officer drew from his observations were “rational,” in the sense that they were inferences that could be drawn and thus were not completely contrary to reason.
Chapple,
¶ 31 We believe our analysis demonstrates that there were other equally strong or stronger rational inferences of innocent behavior that could have been drawn from each of the observations made by the officer. Moreover, these other inferences were not diminished or eliminated by inquiries that could easily have been made. Thus, the continuing viability of these alternative inferences renders de minimis the weight which otherwise might have been given to Brown-lee’s inferences.
¶ 32 Implicit in our analysis is the conclusion that Brownlee would have been authorized to continue defendant’s detention for a brief period to ask further questions about the circumstances Brownlee deemed suspicious.
See United States v. BrignoniPonce,
II. The Dissent
¶ 33 The dissent criticizes us for addressing individually the circumstances asserted *401 by the state as constituting reasonable suspicion. To the extent this criticism asserts that we have failed to evaluate these circumstances as a “totality,” it is incorrect. As our analysis clearly shows, we have performed a de novo review of the same circumstances as were before the trial court, and reviewed them as a whole. We simply disagree with the dissent that these circumstances, taken together, comprise reasonable suspicion.
¶34 To the extent the dissent’s criticism can be taken as disapproval of our subjecting the individual circumstances to some critical analysis, we are at a loss to know how we are to characterize the strength of suspicion attaching to the “totality” of a set of circumstances without also determining the strength each individual circumstance contributes toward the overall assessment. If a police officer’s inference from a particular circumstance is, in our opinion, fallacious, weak, or contradictory to other evidence, we cannot simply ignore that shortcoming and blindly accept that inference. And in order to determine whether a particular inference may be defective, it is obviously necessary that it be critically analyzed. We therefore conclude that such analysis is both proper and necessary in conducting our de novo review.
¶ 35 The dissent also alludes to the necessity that the reviewing court take into account the special expertise with which a police officer observes and interprets apparently innocent conduct. We agree that if an officer connects his observations and inferences to some specialized training he has received, this could enhance the strength attributable to his conclusions. The difficulty here, however, is that officer Brownlee attributed his inference to specialized knowledge with regard to only one factor, that dealing with the “dirty” car. This conclusion, however, was substantially undermined both by the weakness of its premise, which consisted only of an overflowing ashtray, and by the contradictory use to which such an observation is subject. Brownlee did not testify that any of his specialized drug interdiction training taught him that exhibiting nervousness of some particular degree, keeping a registration form on the seat, wearing a particular form of attire, or placing an overnight bag on the back seat were, collectively or individually, signs of a drug courier. Consequently, we reject the notion that we have disregarded inferences “that might well elude an untrained person.”
Cortez,
III. The Length of the Detention
¶ 36 Defendant argues that the one-hour wait between the stop and the arrest was so “excessive” as to be unreasonable under the Fourth Amendment. Because we have reversed the trial court on the issue of reasonable suspicion, we need not address this issue.
CONCLUSION
¶ 37 The trial court committed error when it found that Brownlee had reasonable suspicion for the detention of defendant. Accordingly, we reverse defendant’s convictions and remand for proceedings consistent with this opinion.
¶38 I agree with the majority’s conclusion that our review is
de novo.
However, this court should give deference to the trial court’s factual findings, including findings regarding the officer’s credibility and the reasonableness of inferences that he drew.
State v. Gonzalez-Gutierrez,
¶ 39 The majority opinion demonstrates the difficulty of addressing and second guessing “individually” each factor giving rise to the officer’s reasonable suspicion, rather than, as the Supreme Court dictates, taking these factors into account considering the “totality of the circumstances-the whole picture.”
United States v. Cortez,
¶ 40 In United States v. Cortez, the Supreme Court recognized that trained law enforcement officers often draw inferences and make deductions “that might well elude an untrained person.”
[W]hen used by trained law enforcement officers, objective facts, meaningless to the untrained, can be combined with permissible deductions from such facts to form a legitimate basis for suspicion of a particular person and for action on that suspicion.
¶ 41 At the time of defendant’s stop, Officer Brownlee had been with DPS for eight and a half years. Brownlee’s service with DPS involved principally drug interdiction, and, to further his education and training, he had attended numerous schools and seminars that taught what to look for in drug trafficking. After stopping this defendant, Brown-lee made several objective observations and drew inferences therefrom that, in his opinion, gave him reasonable suspicion 1 to take the further non-intrusive investigative step of calling for a drug dog to be brought to the scene. If these factors are not sufficient to have given Brownlee reasonable suspicion for further investigation, law enforcement officers in the state will no longer have a clear standard as to what factors will justify such a decision. Those persons transporting narcotics do not advertise that fact the way a Good Humor truck advertises its contents.
¶42 I do not believe the trial court abused its discretion in finding that Brown-lee’s suspicion was reasonable. I simply cannot agree with the majority that neither Brownlee nor the court reached “a practical, common-sense conclusion about human nature.” Op. at 527.
¶43 With regard to the length of the detention issue, which is not addressed by the majority because of its decision on the reasonable suspicion issue, I would affirm. The 40-50 minute detention in order for the nearest drug dog to travel 40 miles from Williams was not unreasonable. I would find no abuse of discretion.
Notes
. For every case cited by the majority for the proposition that a factor relied on by Brownlee was
not
reasonable, there is a case that would support the proposition that it
was
reasonable.
See, e.g., United States v. Delaney,
