[The Reaper]

But mercenaries began to fade from the battlefield around the Age of Enlightenment. Partly this was because of breakthroughs in the science of warfare. Better weapons demanded less skill from the fighter. The experience of the mercenary was needed less. With a decently designed musket, a fresh soldier could be trained fairly swiftly and dispatched to the front. And then, too, the 18th and 19th centuries brought new ideas about the sanctity of the nation and the honor of the citizen in soldiering for it. ''Those who fought for profit, rather than patriotism,'' Singer writes, ''were completely delegitimated.'' Still, the British hired 30,000 German Hessians to help them battle the revolutionaries in the American War of Independence. Yet gradually the work of the mercenary grew more and more marginalized and disdained, and in the Geneva Conventions of 1949 it was essentially outlawed, at least in wars between nations.

Mercenaries carried on in the ignored and anarchic places of the world; through much of the second half of the 20th century, they played notorious roles in the insurrections of Africa. But then, in 1995, in the tiny West African country of Sierra Leone, private soldiering made a morality-twisting appearance. A rebel army was burning villagers alive and starting to develop its signature atrocity: hacking off the hands of civilians and letting them live as reminders of rebel power. Desperate, the country's ruler hired a South African firm, Executive Outcomes, that was run by a former apartheid-era military commander. It presented itself as something other than a violent, shadowy employment agency for apartheid-era veterans. It had glossy brochures outlining its military services. Its leader called himself a chairman. Its work wasn't that of ''mercenaries'' or ''dogs of war''; it would soon adopt the term ''private military company.''

In Sierra Leone, using a few aircraft and about 200 men, Executive Outcomes rapidly drove the rebel army of perhaps 10,000 back to the country's hinterlands. Brutality erupted again as soon as Executive Outcomes left, but the world had seen that a small, well-trained private force could accomplish immeasurable good.

Not long afterward, a London company led by a former British lieutenant colonel, Tim Spicer -- whose latest firm now has a nearly $300 million contract with the U.S. Department of Defense in Iraq -- tried again to rescue the West African country. Spicer failed but emerged as a kind of spokesman for the moral value of private military companies. ''The word 'mercenary,' '' he told The Daily Telegraph of London in 1999, ''conjures up a picture in people's minds of a rather ruthless, unaligned individual, who may have criminal, psychotic tendencies. We are not like that at all. All we really do is help friendly, reasonable governments solve military problems.'' (No matter that Spicer had once considered providing his help to Mobutu Sese Seko, the tyrannical dictator of Zaire, for a price.) Britain's foreign secretary, Jack Straw, and a former United Nations under secretary general, Brian Urquhart, were soon talking about the possible use of private military companies to aid the U.N. in stabilizing the world's conflict zones. The U.N. wasn't remotely ready to hire private armies to end civil wars, but a subtle shift in perception had started to take place.

In 2002, the U.S. government hired about 40 private gunmen, from the American company DynCorp, to keep President Hamid Karzai alive in Afghanistan. And in the spring of 2003, as Gen. Jay Garner, retired, established the Office of Reconstruction and Humanitarian Assistance, the short-lived precursor to the C.P.A., as the occupation's governing body in Iraq, the Pentagon put a small contingent of South Africans and Nepalese Gurkhas from the British firm Global Risk Strategies in charge of protecting him and his staff. ''That,'' Garner told me when we spoke last month, ''was the genesis'' of the rise of private security companies in Iraq.

The numbers, at the start of the occupation, were not large. Then, in the second half of 2003, as the C.P.A. expanded its presence across the country in its attempt to rule and rebuild, and as the insurgency mounted, the C.P.A. turned away from the coalition forces, which had been providing a measure of protection, and looked to the companies for safety. Andrew Bearpark, the C.P.A.'s director of operations during that period, explained to me that he was closely and strongly advised by the U.S. military in Iraq -- and financed by the Department of Defense -- to make this move.

Major contracts were put out for bidding. Triple Canopy was awarded its work in January 2004. Other companies received, or already had, their portions. (Meanwhile, the corporations actually doing the rebuilding, the hearts-and-minds element of the American occupation's campaign, were spending up to 25 percent of their U.S. government money on hired protection.) The deployment of private gunmen grew and grew into a profusion that may be explained partly by the subtle shift in perception that had removed some of the old mercenary stigma, and partly by the emphasis on outsourcing that had been gathering momentum in the U.S. military since the early 1990's (but that had been focused on logistical, unarmed support). Most immediately, though, the explosive growth may be explained by the strength of the insurgency in Iraq and by the apparent fact that there weren't enough troops on the ground to fight it.

''Sure, they are performing a military role,'' Garner said of the companies. Then, while noting that he wasn't criticizing the Department of Defense, he added, ''The gut problem is the force'' -- that is, the U.S. fighting force -''is too small.'' And Bearpark, who has lately become a consultant to a large security firm, maintained that private protection might sometimes be better than what a regular army could offer. The private teams are more streamlined and flexible, he argued; they are often better trained for the job; and they may be willing to take more risks, allowing officials to move more freely. But about the fundamental reason for the C.P.A.'s hiring of the companies, he said: ''The military just hadn't provided enough numbers. It was stretched to the limit.''

The Department of Defense is reluctant to discuss the role of security companies in Iraq and precisely how it got so big. Over several weeks I called the Pentagon repeatedly, asking whether the secretary of defense or one of his under secretaries had, at any point, deliberated about the presence of some 25,000 armed men or perhaps authorized it in one way or another, piecemeal or in its entirety. These questions -- which no one I spoke to was able to answer -- elicited from departmental press officers a series of unfulfilled promises to help me get an answer. In the end, they sent an officially approved written statement, which detoured fully around the questions but included the key line, ''P.S.C.'s are not being used to perform inherently military functions.''

The Pentagon's reticence on the issue may be due to uneasiness over the now-common accusation that it didn't adequately plan for battling an insurgency. (It may view questions about private gunmen as leading inevitably to questions about troop numbers.) But there is most likely an additional discomfort, a lingering problem with the companies' public image. For the shift in perception hasn't been complete; the hated word ''mercenaries'' still hovers near. With this problem, the firms are doing their best to help. Many of them have tried to rechristen themselves again, to further separate themselves from the past, from the old infamy of ruthless, insurrection-stirring white freelancers in Africa, to make their work palatable to all.

When I met with Lawrence Peter, a short man with a fiery voice, he raged that the press refused to accept the companies' newly chosen term: ''We are not private military companies! We are private security companies! Private security!'' He justified the distinction by saying: ''The work is defensive. We protect.'' Sometimes, though, the distinction seems secondary. No matter what you want to call Triple Canopy and its men, when the Mahdi Army -- a radical Shiite force loyal to the militant cleric Moktada al-Sadr -- attacked at Kut, the primary truth was that the company was fighting a war.

A current training adviser for Triple Canopy was, in early April 2004, in charge of defending the occupation's Kut headquarters. (For security reasons, it is Triple Canopy policy that employees now in Iraq or likely to return not be identified by their full names.) John, a tall spike of a man of 50, with a graying brush mustache, spent 26 years in the U.S. Army, much of that time with Delta. He was on the first invading helicopter into Grenada in 1983; his helicopter and the others behind him were riddled and ravaged by bullets, and three soldiers sitting near him were shot. ''I've never taken so much fire again till Kut,'' he told me in May. ''Kut was like stepping out into the air -- you know you might not exist any longer.''

Facing a river and enwrapped on three sides by the small, mostly Shiite city, the C.P.A. compound in Kut consisted of several one- and two-story concrete structures. The buildings had been Baathist offices and a hotel. There the coalition's regional ruler, known as the governorate coordinator, worked and lived with a crew of reconstruction officials and contractors, surrounded by 12-foot-high blast walls at the compound's perimeter -- except along the river, where, John told me, the governorate coordinator, a Brit, preferred that nothing obstruct the view. The city had been fairly peaceful. It was a ''sleepy hollow,'' John recalled his Triple Canopy boss telling him, joking that it was a suitable post for an old, graying man to guard.