Miami-Dade retired narcotics detector dog Franky watches during a demonstration in Miami in this Dec. 6, 2011, photo. Franky's super-sensitive nose is at the heart of a question being put to the U.S. Supreme Court. (Alan Diaz/AP)

Aldo the German shepherd and Franky the chocolate Lab are drug-detecting dogs who have been retired to opposite ends of the ultimate retiree state.

But their work is still being evaluated, and on Wednesday it will be before the Supreme Court. The justices must decide whether man’s best friend is an honest broker as blind to prejudice as Lady Justice, or as prone as the rest of us to a bad day at the office or the ma­nipu­la­tion of our partners.

The Supreme Court in the past has tended to agree with the first view. Justice John Paul Stevens, now retired, wrote for the court in a 2005 case that a drug-sniffing dog reveals “no information other than the location of a substance that no individual has any right to possess.”

But the two cases on the docket present an aggressive challenge to the notion that a dog’s “alert” to the presence of drugs is enough to legally justify a search of someone’s home or vehicle.

Florida v. Jardines asks whether it was constitutional for Miami-Dade County police, acting on a tip, to bring Franky to Joelis Jardines’s front door. Franky alerted to the smell of marijuana, the police used that to obtain a warrant, and Jardines was arrested on suspicion of turning his home into a “grow house.”

Miami-Dade detective Douglas Bartelt and narcotics detector canine Franky give a demonstration in Miami in this Dec. 6, 2011, photo. (Alan Diaz/AP)

Florida v. Harris asks a more basic question of whether judges should be skeptical of Fido’s qualifications. It builds on research that shows a high rate of false alerts and cases of ma­nipu­la­tion by a dog’s handler.

Justice David H. Souter, also now retired, sounded the alarm about the reliability of police canines in his dissent in the 2005 case, writing that the “infallible dog . . . is a creation of legal fiction.”

The Florida Supreme Court went further last year in the Harris case when it threw out the evidence in a 2006 traffic stop in the Florida Panhandle that featured Aldo.

“Courts often accept the mythic dog with an almost superstitious faith,” Justice Barbara J. Pariente wrote. “The myth so completely has dominated the judicial psyche in those cases that the courts either assume the reliability of the sniff or address the question cursorily; the dog is the clear and consistent winner.”

The Florida court said that judges should look at the “totality of circumstances,” including a dog’s training and certification records, field performance, and evidence of the handler’s training and experience.

Bill Heiser, founder of Southern Coast K9, is helping to raise the next generation of Aldos and Frankys on 12 acres of sandy soil in central Florida. He agrees that training and handling are key.

It takes about four months — training four or five times a day — to produce a dog for law enforcement. The dogs at Southern Coast work for the attention of their handlers — “Mom” or “Dad,” Heiser calls them — and receive it and their toy only when they have properly alerted to the scent of the drugs they are taught to detect.

An analysis of the 2011-2012 Supreme Court session, including justice voting patterns and key cases.

He demonstrates how a fox-red Lab named Rowdy races through a room filled with distractions and sticks his nose in identical boxes containing a tennis ball, his toy of choice. But he stops and sits only at the box that also contains dope — Heiser uses real drugs, obtained from law enforcement — and waits for his trainer’s praise.

If a dog is properly trained, Heiser said, a bag of McDonald’s fare won’t distract him, a female’s scent won’t delay him, a suspect’s looks won’t interest him.

“He doesn’t know he’s finding cocaine or heroin or marijuana or meth or crack cocaine,” Heiser said. “All he knows is ‘That odor means my toy’s there. And if I do what I’m supposed to do, what I’m trained to do, Mom gets involved or Dad gets involved.’ That’s what he works for, that love and attention.”

The state of Florida, backed by the Obama administration, said such training and certification should be enough for courts.

The kind of “novel, sniff-by-sniff records” the Florida Supreme Court would require would overburden law enforcement and “upend settled law across the nation,” the state said in a brief written by Gregory G. Garre, who served as solicitor general in the George W. Bush administration.

But the law professors, civil libertarians and criminal defense lawyers who support Clayton Harris, the man stopped by Aldo’s handler, Liberty County officer William Wheetley, say that Florida’s proposed standard is not enough.

For one thing, no national standard for certification exists. “There’s no such thing as a well-trained narcotics detection dog,” said Jeffrey Weiner, a criminal defense lawyer from Miami. “It means whatever a trial judge or appellate judge or Supreme Court justice wants it to mean.”

Glen Gifford, Harris’s attorney, told the justices in his brief that Aldo’s certification was out of date and that there was no real record of how he and Wheetley worked together. Aldo alerted to the scent of drugs on the handle of Harris’s vehicle, and a search found the ingredients for methamphetamine inside.

But in the tiny town of Bristol, that was not the only time Wheetley had pulled Harris over. And Aldo alerted to drugs that time, too, but a search came up empty.

That could simply mean that drugs had recently been in the truck but were now removed, or that they were expertly hidden.

But Gifford points to other research that says dogs can alert to other scents or simply have bad days.

And a study last year at the University of California at Davis — disputed by some in the dog-handling industry — was mentioned in several briefs. It indicated that handlers had much to do with when a dog alerted.

It brought together 18 K-9 teams and ran them through a test facility at which the handlers had been told that some targets had been marked and some had not.

Together, the teams racked up 225 false alerts. Only one team was perfect. It was the one that did not alert at all, because there were no drugs in the facility.