The issue on this appeal is whether in January 2000, the police in Andover, Massachusetts, had probable cause to arrest Barbara Valente for planting anonymous bomb threats in her place of work.

In August 1997, Valente began work as an on-site manager for ADECCO Employment Services ("ADECCO"), then known as TAD, at a Hewlett-Packard ("HP") plant in Andover, Massachusetts. The HP plant was a large one, comprising seven buildings and employing over 2,500 people. Between August 1997 and January 1999, seven anonymous bomb threats and fourteen anonymous notes expressing spite or workplace dissatisfaction appeared at the plant. By way of example, one bomb threat read:

There is a bomb planted in building 2. To get you and to get Sheila and to get anyone who keeps us here on a dangerous day. It will go off when I leave at two! Goodbye to you.

Several of the notes were found by Valente — one purportedly signed with her first name — and twenty-one of the twenty-four notes were in the building in which she worked. After the first bomb threat in November 1997, the Andover police began an investigation. The detective then heading the case identified a suspect (not Valente) by handwriting, and HP retained a handwriting analysis firm with adequate credentials: McCann and Associates. McCann deemed the samples inconclusive and the investigation lapsed.

In September 1998, Valente began three months of maternity leave to care for her newly adopted child. During this time, two more notes were discovered and in January 1999, the investigation resumed under the charge of Detective William Wallace. More handwriting samples were obtained, this time from nine employees including Valente. McCann determined that for the fourteen notes as to which it could draw conclusions, only Valente could not be ruled out as a suspect; but McCann also said that she could not be "conclusively" identified based on the samples available.

More samples of Valente's handwriting were secured and in November 1999, McCann concluded that it was "more probable than not" that Valente was the author of three bomb notes, including two found when she was on leave, and eleven of the other notes. Wallace met with McCann representatives who, with slides or similar means, displayed the similarities on which they relied and repeated their conclusion. Wallace also determined that Valente could have visited the HP building in question during her leave, although he had no proof that she had done so.

On January 5, 2000, at Wallace's request Valente came to the police department. In the discussion that followed, Wallace deemed Valente's rather terse disclaimers not what he expected from an innocent person — she simply said she knew nothing about the notes; he also found her to be very nervous, her skin reddening and breaking out in hives. At the close of the interview, Wallace and his supervisor agreed that Valente should be arrested and a warrantless arrest was effected on three counts of creating a bomb scare, each of which carried a maximum punishment of 20 years' imprisonment under Massachusetts law. Mass. Gen. Laws. ch. 269, § 14(b)(1), (c) (2002).

Valente was never prosecuted. A criminal complaint was dismissed when the prosecutor failed to meet a discovery deadline. During the period after her arrest and when Valente was not working at the plant, yet another note was found and suspicion fell on another employee. In July 2001, Valente brought the present civil rights action against Wallace, his supervisor and the town, 42 U.S.C. § 1983 (2000), charging that she had been arrested without probable cause in violation of the Fourth Amendment.1

After discovery, the district court granted summary judgment on the merits in favor of the police, ruling from the bench that they had probable cause for the arrest, a ground that also disposed of any derivative claim against the town. In the alternative, the district court found that the two individual officers were entitled to qualified immunity. Valente now appeals. Our review on summary judgment is de novo. See Roche v. John Hancock Mut. Life Ins. Co., 81 F.3d 249, 253 (1st Cir.1996).

For a warrantless arrest, the Fourth Amendment is taken to require "probable cause," Wong Sun v. United States, 371 U.S. 471, 479, 83 S. Ct. 407, 9 L. Ed. 2d 441 (1963), and the broad outlines of the concept are familiar. See generally, 2 LaFave, Search and Seizure ch. 3 (3d ed.1996). But the case law on probable cause harbors one central ambiguity and a host of smaller issues. The ambiguity exists because the Supreme Court has told us that probable cause means more than "bare suspicion" but less that what would be needed to "justify ... conviction." Brinegar v. United States, 338 U.S. 160, 175, 69 S. Ct. 1302, 93 L. Ed. 1879 (1949). A good deal of territory lies in between.

Within this territory, the Supreme Court has said that the question is whether the evidence would "warrant a man of reasonable caution" in believing that a crime has been committed and committed by the person to be arrested. Beck v. Ohio, 379 U.S. 89, 96, 85 S. Ct. 223, 13 L. Ed. 2d 142 (1964). See United States v. Reyes, 225 F.3d 71, 75 (1st Cir.2000). The emphasis is on calculating likelihoods. E.g., Brinegar, 338 U.S. at 175, 69 S. Ct. 1302. Whether this excludes all other factors and whether the likelihood must be "more likely than not" are questions arguably unsettled;2 but, centrally, the mercurial phrase "probable cause" means a reasonable likelihood. Illinois v. Gates, 462 U.S. 213, 235, 103 S. Ct. 2317, 76 L. Ed. 2d 527 (1983).

In our case, the bomb threats constituted a crime so the only question is whether the police had probable cause to believe that Valente was the culprit. The test is objective and turns on what a reasonable police officer would conclude based on the evidence actually available at the time (and not on unknown facts or subsequent events). Roche, 81 F.3d at 254. On the facts known to Wallace at the time, we hold that a reasonable police officer would be warranted in the belief that Valente had sent some of the notes.

The main reason for this conclusion is that a seemingly qualified expert handwriting examiner had said that the notes were more likely than not written by Valente. This is not conclusive as to probable cause (for reasons to which we will return) but it is a powerful start. Cf. Roche, 81 F.3d at 255 (voice identification). And it is worth noting that Wallace did not just take the expert's summary conclusion; he met with the McCann representatives and watched them explain graphically why they were persuaded.

If Valente's fingerprints had been found inside a threat letter first opened by the police, it could hardly be doubted that probable cause would exist. Handwriting analysis is a less rigorous means of identification, e.g., Mnookin, Scripting Expertise, 87 Va. L.Rev. 1723, 1726-27 (2001), depending inter alia on how distinctive the handwriting might be, the number and type of samples, and the competence of the examiner. See United States v. Mooney, 315 F.3d 54, 62-63 (1st Cir.2002). Normally, the examiner can do no more than speak of probabilities. But here the examiner did say, "more likely than not," which is the most that the Fourth Amendment requires. See note 2 above.

This is not the end of the story. Plenty of other evidence might be available to the police to reinforce — or to rebut — an expert's view that the suspect more likely than not wrote a threat note. Suppose the note contained information almost certainly known only to the suspect or, conversely, information that the suspect would be most unlikely to know. Or there might be information about motive and access, two staples of criminal investigation. The initial expert judgment might have to be adjusted; conceivably, "probable cause" could vanish based on new data.

Here, the notes had begun not long after Valente joined the company, and most were found in the building where she worked. See United States v. Brown, 457 F.2d 731, 733 (1st Cir.1972). All of this is mildly helpful to Wallace, although how far depends in part on the size of the work-force, turnover, and similar data; some of the facts, and what the police knew on these points, are obscure. The police do appear to have checked to see whether Valente could have had access during her three-month leave and found that this could not be ruled out.

There is also Valente's behavior during the interview at the police station. According to Wallace, she was both unduly taciturn (he said that the usual innocent person protests much more strongly when wrongly accused) and extremely nervous. Frankly, to a layman, neither of these symptoms seems very revealing. However, the case law does give some weight both to demeanor evidence, e.g., United States v. One Lot of U.S. Currency, 103 F.3d 1048, 1055 (1st Cir.1997), and to the experience of the police. E.g., United States v. Ortiz, 422 U.S. 891, 897, 95 S. Ct. 2585, 45 L. Ed. 2d 623 (1975). In fairness to Wallace, he said that the interview was primarily to see if Valente could counter the inference he had already drawn.

Thus, at the time of the arrest, the police had expert evidence that Valente was more probably than not the note writer; she had apparent access to the site throughout the period before her arrest; most of the notes were discovered near or not far from where she worked; there was no specific motive attributable to her but disgruntled employees are not unknown; and there was apparently no other then-current suspect against whom a strong case existed. This, in our view, is probable cause for an arrest. See Roche, 81 F.3d at 254-55.

What has been said above largely disposes of Valente's arguments on this appeal. Her counsel claims that we should look only at the McCann report because Wallace, in a deposition taken two and a half years after the arrest, said that this was the only evidence of guilt and her nervousness was not mentioned in the arrest report. But Valente's access, proximity to the notes and nervousness at the interview were contemporaneous events known to the police and doubtless considered by them even if not formally designated evidence of guilt.

Finally, Valente says that a psychological profile commissioned by HP allegedly suggested that the culprit had traits that differed from Valente's. However, while handwriting is an inexact science, psychological profiling appears to be even more inexact; handwriting experts have been routinely used in courts for a century now, Mnookin, supra, at 1726, while psychological profiling remains primarily a law enforcement device for narrowing the field of suspects and is rarely admissible in court. 1 Giannelli & Imwinkelried, Scientific Evidence § 9.7 at 479 (3d ed.1999). If the profile counts at all in the probable cause calculus, it does not alter the result in this case.

Affirmed.