By Brandon L. Garrett, Harvard University Press, Cambridge, Mass. 367 pages, $39.95

‘Every prison that men build,” according to Oscar Wilde, “is built with bricks of shame.” In “Convicting the Innocent,” Professor Brandon L. Garrett of the University of Virginia School of Law will persuade many readers to take seriously the poet’s evaluation. The book calls, loudly and clearly, for the attention of the legal profession.

The author has assembled and studied the first 250 cases of persons, all but four of them men, convicted of crimes but later exonerated by DNA tests. He began the project in collaboration with Barry Scheck and Peter Neufeld of the Innocence Project of the Benjamin N. Cardozo School of Law. His subjects, whom he calls “exonerees,” spent an average of 13 years in prison before their exculpation.

The bulk of the book deals with the question why and how, under our system of justice with its constitutional and other safeguards, so many individuals could be found guilty and imprisoned for crimes they did not commit. The author asks, what went wrong?

Garrett rejects the “few bad apples” rationalization of these miscarriages of justice. He blames “systemic failures.” Serious flaws, he contends, exist in each of the principal techniques by which criminal convictions are accomplished.

As a preliminary observation, the author shows how the criminal justice system concentrates its fire on minorities. Of his 250 subjects, 70 percent were black, Latino or Asian. Later in the book, he notes that “[W]ealthy, employed people with high education and clear records are far less likely to be suspected of crimes, more likely to be able to hire a skilled lawyer, and much less likely to be wrongly convicted.” Something to think about.

To begin his review of the ways in which his subjects were convicted, Garrett takes up confessions, “incredibly powerful evidence at trial.” Of the 250 individuals whose cases he has studied, 40 were convicted on the basis of such admissions. He calls the prevalence of false confessions a “puzzle.” How, he asks, can someone “voluntarily” admit to a crime he did not commit? To complicate the matter, the credibility of most of the confessions involved in his study was fortified by the inclusion of details that a person innocent of the crime could not know.

Many readers will remember the Central Park jogger case. In 1989, a young woman was beaten, raped and left for dead in the park. Five teens from Harlem were arrested. Having signed confessions that included exhaustive particulars of the crime, all were convicted.

Fortunately, New York made DNA testing available to convicted persons (Crim. Proc. Law §440.30[1-a]). In 2002, after DNA analysis, Supreme Court exonerated the five and helped identify the true culprit, ordering the teens’ release. They had spent between seven and 13 years behind bars.

Supported by persuasive documentation, the author shows how the devices employed in police interrogation of suspects help solve the confessions puzzle.

Details of a crime often are fed to a suspect by leading questions and then incorporated in the written confession signed for submission in evidence. Other aspects of an interrogation, for example, its duration, the physical set-up in which it takes place, and the subtle insinuation that confession will somehow make it easier for the suspect, can also play their parts.

The author proceeds to examine in like detail and with like documentation the other standard methods of securing a conviction. He takes up eyewitness identification, forensic science and “jailhouse informants,” each of which was involved in the convictions of a substantial number of his subjects.

All of these techniques, he explains, include features that load the dice against a defendant. He emphasizes, however, that police and prosecutors, in employing these methods, are not necessarily acting in bad faith.

The most disturbing of the author’s contentions is his “tip of the iceberg” argument. Far from all crimes leave biological evidence usable in DNA analysis. Homicide by firearm and armed robbery, as examples, normally produce no such evidence. It follows, he contends, that DNA saves only a small fraction of the erroneously convicted. (Most of his subjects, 171 of 250, had been found guilty of rape.)

“[L]egal change can be much slower than technological change,” maintains the author. Thus, resort to DNA testing is often resisted by prosecutors and sometimes denied by judges. The spurious argument that “overwhelming” evidence of guilt makes such testing pointless still works now and then.

The book deals not solely in negative terms. Time and again Garrett shows that DNA can inculpate the guilty as well as free the innocent.

He also offers suggestions how each of the flaws he has pointed out can be remedied. In eyewitness identifications, for example, he maintains that the traditional line-up should be double blind, that is, with both the officer in charge and the witness unaware who in the line-up is the suspect. The witness should be told, moreover, that the line-up may not include the perpetrator at all.

“How men their brothers maim,” exclaimed Wilde in his condemnation of prisons. How many of our brothers (and sisters) have we in this country maimed by taking part in or condoning wrongful criminal convictions? This troubling, highly readable book makes that question one that cannot be, must not be, ignored.

Walter Barthold is retired from the practice of law in New York City.