"Commander-in-chief." No other phrase captures so concisely the power of the president. In no other arena is the president's power so unconstrained as it is in national defense. With the War Powers Act a dead letter and Congress unusually compliant, he can commit the world's most potent military virtually at will, staking American lives, treasure, and prestige with consequences - intended and otherwise - that ripple down the decades.

And yet, at the same time, in no other arena is the power of the president so severely restricted. It takes years to recruit, organize, and train new units of troops. High-tech weapons require one or two decades to develop, and even policy pronouncements must grind laboriously through the bureaucracy before they can emerge as executable plans.

The commander-in-chief may wield the sword unchecked - but it takes years of labor and myriads of people to reforge that sword.

Few presidents have pushed up against this paradox as hard as George W. Bush has done. He started advocating his views before he became president. As a candidate in 2000, he could have stuck with the safe Republican criticisms of Bill Clinton as a draft-dodger who overcommitted and underfunded the military.

Instead, Bush went further, and a bit out on a limb, by calling for a radical "transformation" of Industrial Age services that relied on mass - big forces, slow buildups, massive firepower - into an Information Age military that relies on agility - smaller forces, rapid deployment, precision strikes.

To be fair, Clinton officials and congressional Democrats, including Al Gore's running mate, Sen. Joe Lieberman, D-Conn., had spent years advancing their own forms of "transformation," but Bush decried them for being too cautious. He even pledged to "skip a generation of technology" by canceling current programs (never specified) and investing the savings in cutting-edge research.

The idea unnerved military officials, defense contractors, and their congressional patrons. And Bush's choice as Defense secretary, the hard-driving and distinctly undiplomatic Donald Rumsfeld, only aggravated this "Iron Triangle" even more. By September 10, 2001, it seemed as if the Bush-Rumsfeld transformation had run aground on the limits of the commander-in-chief's power to command.

Then nearly 3,000 Americans died on a sunny Tuesday morning. President Bush did not choose war. But he did choose how to wage it and where, with an intensity that is pushing the U.S. military to its limits and is forcing both innovation and exhaustion. It is hard to imagine any plausible president - e.g., Al Gore - not invading Afghanistan. But Iraq was Bush's war.

Admittedly, the Clinton administration had spent years bombing Iraq and (under Republican pressure) had made Saddam Hussein's overthrow a long-term goal. Arguably, as George Friedman of Strategic Forecasting (known as Stratfor) points out, "Even more than Bush, [a President] Gore would have come under tremendous political pressure" to act aggressively in the Middle East.

But a Gore administration invasion of Iraq? "Probably not," said Michele Flournoy, a former Clinton official now with the Center for Strategic and International Studies. "And if we did do it, we certainly would not have done it this way."

The Bush administration, overriding both diplomatic and military traditionalists, attacked Iraq not only without an international consensus, but with a relatively small force that trusted in speed, smart weapons, and long-range sensors - in short, in transformation - to make up for mass. The gamble paid off brilliantly in the initial phase: The invasion force took Baghdad with half the troops and less than half the casualties that it had taken to free Kuwait in 1991. But the high-tech force lacked the old-fashioned manpower to pacify what it had conquered.

The result, not entirely intended, is that the kind of armed nation building that Bush rejected as a candidate has become his consuming priority as a president. Bush's unilateralist crusading is the opposite of Clinton's multilateralist "meals-on-wheels" peacekeeping, said Kenneth Allard, a retired Army colonel-turned-analyst. "But what has happened is much the same thing: We now have American troops all over hell's half-acre. Regardless of the motivation, the result is the same."

So as George W. Bush seeks re-election as a war president, he is running up against the limits of the president's war power. Between the administration's pushing of reform from above and junior officers' adaptation under fire from below, the military has made real strides in how it operates at the cutting edge: more cooperation among the Army, Air Force, Navy, and Marines; more daring in substituting speed and smart weapons for massed firepower; more experiments in new organizations for Army brigades and Navy battle groups; more innovation all around.

Yet the bulk of the equipment that the military operates now and plans to buy in the future - most prominently, short-range fighters and big-hulled ships - remains the same as it was under Clinton, or even under Bush's father. Looking at today's procurement budget, said analyst Andrew Krepinevich of the Center for Strategic and Budgetary Assessments, "you can call this 'Clinton-heavy,' and you can call Clinton 'Bush I-lite.' "

And for all the increased spending on defense in recent years, the extraordinary commitment in Iraq has pushed the Army in particular toward its breaking point. As president, Bush has certainly changed the U.S. military. The crucial question is whether the consequences he intended will outweigh those he did not.

Transformation

As a candidate, George W. Bush pledged to "skip a generation of technology" to invest in radical leaps forward in weaponry. But he never specified what he would skip, or what he would skip to, with the notable exception of missile defense. Unlike a lot of the vaguer parts of the "transformation" vision, missile defense has definitely materialized: Radar sites and interceptor missile silos are already under construction, thanks to the administration's decisions to revoke the Antiballistic Missile Treaty, to accelerate testing, and to spend more than $9 billion in fiscal 2004 alone - to the dismay of critics who consider the program unreliable and rushed.

The Bush administration has invested in other innovations as well. It is spending $1.4 billion in fiscal 2004 for unmanned systems like the Predator drone. And it is spending $207 million - more than three times the fiscal 2000 figure - for the Joint Forces Command's research, development, and experimentation program.

But overall, these dollars are dwarfed by those for more-traditional programs, with almost $11 billion going for manned fighters alone. In fact, Bush has continued all three of the Clinton administration's tactical aviation programs: $2.9 billion in 2004 for the upgraded SuperHornet version of the Navy F-18, originally a 1970s design; $3.6 billion for the high-performance, high-cost Air Force F-22, first conceived in the 1980s; and, $4.3 billion for the F-35 Joint Strike Fighter (still in development), started in the early 1990s as a relatively affordable workhorse to be shared by the Air Force, Navy, and Marines.

Even under Clinton, budgeteers warned of a fiscal "train wreck" as the three programs ramped up concurrently. Strategists, meanwhile, questioned the fixation on light-payload, short-range fighters over heavy-hauling, long-range bombers, which the military had no plans to buy. Pre-9/11 rumor had Rumsfeld toying with the idea of restarting production of the "stealth" bomber, the B-2, but the idea went nowhere, so fighters still rule the budget and the skies.

The pre-9/11 Pentagon was also abuzz with the Rumsfeld team's consideration of cutting Navy aircraft carriers and two of the Army's 10 combat divisions. Global war wiped out those plans. But even after the invasions of Afghanistan and Iraq, Rumsfeld continued to push the Navy and, especially, the Army in a way he never did the Air Force. There was no "skipped generation," but the administration made real cuts, and real additions.

The Navy, for example, has long favored relatively large destroyers like the USS Cole and the planned next-generation DD(X), while it has resisted investing in smaller vessels optimized for shallow waters like the Persian Gulf. "But the Bush administration gave the Navy the ultimatum, 'Do something small, fast, and agile, or you don't get DD(X),' " said Anteon Corp. analyst Scott Truver. "It was that stark." The result is an experimental program for a "Littoral Combat Ship" one-fifth or less the size of a destroyer.

The Bush administration has also increased the Navy's shipbuilding budget, which still goes mainly for traditional vessels. This mix, however, is under serious review: The Littoral Combat Ship threatens funding for destroyers; the Navy is studying cuts to its expensive submarine fleet, with some former nuclear-missile subs being converted to smart-weapon platforms; and some planned amphibious landing ships for the Marines could be replaced by "seabasing," a system of giant cargo ships that are large enough to land an airplane on and that could supply invasion forces from far offshore without needing a beachhead or a seaport.

More subtle, but more radical, than any new construction are proposed changes in how the Navy operates its ships. The traditional round of deployments - six months deployed overseas, 18 months refitting and training, repeat - has already yielded to the demands of supporting the Afghanistan and Iraq invasions. Now the Navy is making such "surges" standard procedure. Under a plan called "Sea Swap," ships do not sail back after six months but instead have their current crew flown home and replacements flown in. This plan reduces the number of ships required to maintain forward patrols.

In one sense, there's nothing new here. All of these ideas originated within the Navy itself, back in the 1990s. The first Sea Swap exercise was under way before Bush's inauguration; seabasing supersizes existing support ships to enable a long-held Marine desire to strike deep inland; and the Littoral Combat Ship was conceived by Vice Adm. Arthur Cebrowski in 1998.

But it took the Bush administration to shove Cebrowski's idea into the budget, and to elevate Cebrowski himself from a lone maverick to director of Rumsfeld's special office for transformation. Similarly, Adm. Vernon Clark, the chief of naval operations, is pushing all these changes, and was first appointed under Clinton. But it was the Bush administration that backed Clark's program against Navy traditionalists and gave him an extra two years as chief to implement it - an extension in office unequalled by any chief since the legendary Arleigh Burke, who oversaw the introduction of missiles and nuclear power to the Navy from 1955 to 1961. "What has been the Bush impact on the Navy?" asked one congressional source. "It has been to make Clark probably the most powerful chief of naval operations we've had in a long time."

Indeed, the most important impact of the Bush administration may not be in programs but in people. Although the administration has hardly equaled the mass firings of old officers by Gen. George Marshall at the start of World War II, "what Rumsfeld has sought to do is nothing less than a complete recasting of military culture ... by ruthlessly replacing representatives of 'old think' with a cadre of senior officers sharing his views," said Loren Thompson, a well-connected consultant to defense contractors and an analyst at the Lexington Institute. In contrast to Clinton-era officials, Thompson said, "Rumsfeld has not been afraid to criticize and second-guess and punish the most senior of officers."

Nowhere has this harsh mode of management been more apparent than in the Army. When selecting a new chief of staff last year, Rumsfeld passed over every active-duty general officer to name a retired commando, Gen. Peter Schoomaker. The previous chief, Gen. Eric Shinseki, had pushed what was then considered a radical transformation program under Clinton, but he found himself hammered by Rumsfeld as too conservative. The administration canceled the Army's high-tech Crusader cannon as too heavy to transport to distant battlefields; conversely, it criticized the Army's air-deployable Stryker armored vehicle as too light to survive on its arrival in battle. (The Stryker survived the skepticism and is now used in Iraq.) Schoomaker continued the cuts by canceling the Comanche scout helicopter, the Army's prize aviation program.

But these cuts do not mean that the Army is "skipping ahead" to any radical new weapons. Quite the contrary. The Comanche funds are going to repair and upgrade the current helicopter force, flown hard and shot up in Iraq and Afghanistan. And the centerpiece of Shinseki's program to re-equip the entire Army with high-tech vehicles, the Future Combat System, is in budgetary trouble and is likely to be reconceived as a test bed for technologies that can be spun off into the current force. Supplemental requests have focused on funding near-term needs, such as up-armored Humvees, missile-jammers for helicopters, and body armor.

Yet this is not a simple retrenchment. Even more so than in the Navy, the transformation in the Army is not about new superweapons but about new ways of using the hardware already on hand. The Army has sped up the fielding of computer communications networks for retrofitting onto existing armored vehicles; it has begun reorganizing its brigades for the high-speed, widely dispersed - closely coordinated with air support from the other services - that those networks enable; and it has started overhauling its personnel assignment system to produce stable teams of troops ready to deploy on short notice.

Like the innovations in the Navy, the entire Army program, from networks to self-sufficient brigades to team-building, originated inside the service in the '90s. An iconoclastic colonel named Douglas MacGregor laid out the entire package in his book Breaking the Phalanx back in 1997. And MacGregor, who retired recently, is deeply skeptical that the Army will go far enough, pointing, for example, to the "tremendous resistance" to streamlining the traditional hierarchy and giving the reorganized brigades the self-sufficient headquarters that he called for in his book. "Rumsfeld has all the right intentions," said MacGregor, "but at the end of the day, I see no evidence that it will make any difference."

Nevertheless, the Army is at least trying to implement in existing units ideas once reserved for hypothetical war games and far-future plans. And even before appointing Schoomaker, the Bush administration was imposing a new way of warfare on the Army and the military as a whole. Witness in Afghanistan the reliance on Special Forces backed by long-range aircraft armed with smart bombs, as opposed to the use of conventional Army units. Contrast the massive buildup, prolonged air bombardment, and deliberate ground advance of the 1991 Persian Gulf War with the headlong dash to Baghdad in 2003, when a much smaller Army-Marine ground force - hastily retrofitted with a bare-bones computer network to coordinate operations - relied on air support and sheer speed to compensate for its exposed flanks.

Officers who fought in this war speak of extraordinary adaptation. "Problems were thrown at us that we had no way to predict," said Navy Capt. Charles Wright, whose carrier-based air wing participated in both Afghanistan and Iraq. The network technologies introduced during the 1990s were "awesome," Wright said, but human operators improvised new ways to use them on the fly. This meant compressing traditional procedures to call in air support and using network chat rooms to revise targeting plans.

But now U.S. troops are under a new kind of pressure to adapt - and their technology may not help them. "The Rumsfeld crowd thought that technology uber alles was the thing that they could count on," said retired Col. Allard. "But they clearly miscalculated the number of ground troops to pacify Iraq. That is a labor-intensive operation." The Bush administration has forced the military into a higher gear than most thought possible. The gears, however, are starting to strip.

Overstretched

War imposes change. It also stifles change. The same immediate demands that impel troops to adapt or die also devour resources that could be invested in lasting transformation. Although post-9/11 defense spending is comparable, in constant dollars, to the height of spending during the Cold War in the mid-'80s - when much of the current equipment was built - "in the Reagan buildup, virtually all the money was going for modernization. Now a huge chunk is going for consumables," such as fuel, ammunition, and spare parts, said Steven Kosiak of the Center for Strategic and Budgetary Assessments. "It's not going to have the same kind of long-term impact on U.S. defense capabilities."

The stress of war has hit all the services, but none harder than the Army. The crucial shortfall is not in money or machines, but in manpower. Of the Army's 500,000 soldiers, almost 140,000 are in Iraq, and they are supported by another 30,000-plus in Kuwait - on top of continuing Cold War commitments to Germany and South Korea, and to Clinton-era peacekeeping in the Balkans. The Web site maintained by GlobalSecurity.org called "Where are the Legions?" counts 15 of the Army's 34 active-duty combat units (brigades and regiments) as currently deployed. That means that nearly as many units are abroad as at home, when historical experience shows that a long-term commitment, as with the British in Northern Ireland, requires three or four units recuperating and training for each one deployed.

Over administration objections, both chambers of Congress have voted to increase the Army by 20,000 to 30,000 soldiers (House and Senate figures differ). But "the training base can only put so many people through so fast," Krepinevich said. "You might be able to stand up three brigades a year" - about 15,000 troops - "but you need them now."

In the meantime, Army planners are even considering cannibalizing that training base. For 20 years, the National Training Center at Fort Irwin, Calif., has been the crucible where Army units learned hard lessons in war games against the elite 11th Armored Cavalry Regiment. Now the 11th Cavalry may be sent to Iraq - and its years of specialized expertise in role-playing a tough, realistic "opposing force" may be replaced by whatever local National Guard unit is available. Many officers fear that this threat to the National Training Center is just part of a general unraveling of the professional Army, which they painstakingly rebuilt after the ordeal of Vietnam and the end of the draft left behind a "hollow force."

So far, in contrast to the '70s, or the mid-'90s for that matter, Army recruiting numbers have held steady. So while the prospect of being shot at in Iraq has scared away some 18-year-olds, post-9/11 patriotism and a relatively weak economy have encouraged at least as many youths to enlist. "If the job market continues to improve," said one congressional staffer, "we're going to emerge this fall into an era of relatively weak recruiting."

The active-duty force is not enough, however. During the Vietnam War, President Johnson relied on the draft to fill the ranks and left the Reserves and National Guard at home. Today, reviving the draft is a political impossibility, and President Bush has called up the Reserves and Guard in numbers unequalled since the Second World War: Since 9/11, he has ordered more than 385,000 citizen-soldiers to active duty, either for homeland security or duty abroad. (That figure double-counts some weary souls who have been called up twice.)

Forty percent of the force now in Iraq is from the Reserve or Guard. Derided as "weekend warriors" during Vietnam, reservists were already spending months in Bosnia or the "no-fly" zones over Iraq during the Clinton years. Now, like active-duty troops, they face yearlong tours of duty in Iraq - which can be and have been extended. Compared with the 1990s, "everybody's deployments seem to be longer," said Joyce Raezer of the National Military Family Association. "And the predictability is gone. When you think the end is in sight, you get the word, 'Nope, a little longer.' "

Beyond the extensions of an entire unit's tour, tens of thousands of individual troops in units slated for Iraq, or in short-staffed specialties, are subject to "stop-loss" orders, which forbid planned departures from the service, even to retire. So far, the overall rate at which Army soldiers voluntarily re-enlist as their terms of service end has stayed stable and on target since 9/11.

But recent reporting at bases to which units are now returning from Iraq has showed some alarming drops locally in re-enlistment, particularly among the midcareer sergeants who form the backbone of the Army. "That's the canary in the mine," said retired Gen. Robert Scales, former commandant of the Army War College. "The thing that broke the Army in Vietnam was the disappearance of the professional noncommissioned officer corps."

Because Vietnam consumed so many troops for so many years, "you could come out of Vietnam and count on spending 18 months in some other place - Korea, the United States, Europe - and then you'd get orders back to Vietnam," recalled retired Gen. Donn A. Starry. "They'd go back the second time, but the third or fourth time, the wife would say, 'Don't expect to find me here when you come back.' " Thousands of veteran sergeants, and their irreplaceable expertise, left the Army.

Will history repeat itself? "It is very depressing for a lot of these folks to see that their unit is on the schedule for a second rotation into Iraq," Raezer said. On the upside, she added, and unlike during Vietnam, "the level of support for the soldiers and their families from civilian organizations, employers [of Guard and Reserve troops], just ordinary Americans, and Congress inspires awe - and it is continuing."

Another crucial difference is that since Vietnam and the end of the draft, the military has restructured its pay and benefits system specifically to retain the long-serving professionals at the core of the force. The upside is that troops are kept in the ranks by what personnel experts call "the golden handcuffs." The downside is that personnel costs eat an ever-larger share of the budget. By fiscal 2005, the Congressional Research Service estimates, the cost per service member will be 30 percent higher than it was in 1999.

The upswing dates to the late 1990s. With the federal budget in surplus but recruitment and retention rates worrisomely low, service members' advocacy groups and the Joint Chiefs of Staff lobbied Congress to fund major new benefits. Troops got bigger housing allowances and automatic annual pay raises fixed at half a percentage point above inflation. Some of the most generous additions were to the retirement plan, including a 25 percent increase in pensions for current troops who serve a full 20 years and an expansion of health benefits for both current and future retirees. Coupled with the same kinds of cost growth as in civilian health care, the new medical benefits caused the Pentagon's unified medical budget to leap by 25 percent in 2002 alone and to increase at 11 percent annually since, to almost $30 billion in 2004.

Bush inherited, not invented, these exploding benefits. Early in his term, he voluntarily upped military pay even above the automatic increase required by law. But in the fiscal 2004 budget, the White House Office of Management and Budget reined in increases somewhat, and the automatic annual raises are about to expire. Nevertheless, costs continue to rise, especially for health care, where spending will double by 2022, according to the Congressional Budget Office.

Meanwhile, Bush has also inherited, and retained mostly unchanged, the Clinton-era plan for procuring new weapons, a plan that also counts on major long-term spending increases. Overall, by the same CBO figures, the Defense budget is on track to grow from $383 billion in fiscal 2004 to $439 billion in fiscal 2009 - above the peak of either Vietnam or the Reagan buildup, in constant dollars. And CBO warns that its estimates do not include likely weapons overruns or the supplemental bills for funding the global war in Iraq, Afghanistan, and elsewhere.

Thanks to Congress's generous supplementals, "the Department of Defense has not really been asked to absorb any of the costs" of Iraq, said CSBA's Kosiak. But historically, defense spending rises and falls in cycles, Kosiak warned, and "if we're not going to see budgets of $450-plus billion a year for the next decade, we're not going to be able to afford everything in the administration's plans."

Arguably, all of this is just the cost of doing business. Generous benefits help to retain high-quality troops despite exhausting tours of duty in Iraq; high-performance fighters and big ships preserve the U.S. dominance of the air and sea, which the new "transformational" approach to warfare takes for granted. But the cost of doing business is getting high. Decisions made in the 1990s put the Defense budget on upward autopilot; Bush, so far, has only accelerated the climb. At some point, some president may have to make hard choices. This president has not done that yet.