As Bush considers a new war strategy, advocates of boosting US troop levels argue that step is the surest solution, despite the higher costs and potential risks. They say it is possible, affordable and perhaps the last chance to rescue a war effort after more than four years and 3,000 American deaths.

Yet to be explained is what could be achieved by adding several thousand US troops to the roughly 140,000 in Iraq and how the move would affect a gruelling guerrilla war that the Bush administration now acknowledges military strength alone cannot win. The idea would be to use the extra forces to help create a measure of security in Baghdad, where sectarian and insurgent attacks occur daily. That means more raids with Iraqi forces to clear and hold certain neighbourhoods.

If successful, this approach would be a first step toward creating conditions that would allow more effective use of economic reconstruction money and other job-creating efforts. That, in turn, would strengthen prospects for political stability.

One option the US military has proposed to Bush is a modest troop increase - perhaps 9,000 in the coming months, possibly to be followed by a similar addition in the summer. Most would go to Baghdad, the critical battleground; some could go to western Anbar province, the focal point of the Sunni insurgency.

No one questions US military superiority, but some wonder about the effectiveness of a burst of new troops.

Democratic senator Jack Reed, who has visited Iraq numerous times, said in a telephone interview he is sceptical of sending more troops because he doubts the administration will make it part of a new, broader approach.

"Unless there is a comprehensive approach involving political decisions by the Iraqis, economic development and security, then it's not going to work," said Reed, a member of the Senate Armed Services Committee.

Anthony Cordesman, an Iraq-watcher at the Centre for Strategic and International Studies, wrote in an analysis this week that the argument over whether to send more troops misses a larger, more important point: finding a way to heal Iraq's political wounds to reverse the drift toward civil war.

"There will be little point in surging US military forces or in trying to build more effective Iraqi forces unless the US and the Iraqi government can find a way to halt this drift" toward ethnic and sectarian division, Cordesman wrote.

Adding troops would go against the grain of US public opinion, which favours ending the war rather than getting deeper into it.

Until now, the administration has believed in the need to achieve political conciliation before real security could be established. The US military has been buying time for the Iraqis to solve their internal conflicts, trying to contain the insurgency while training Iraqi forces.

During 2006, even as US commanders believed they soon could reduce the number of American troops in Iraq, sectarian fighting worsened, casualties escalated and Bush decided to reassess his strategy. Among the advice he has received in recent weeks is a bold troop-buildup plan, written in part by a retired general who was the Army vice chief of staff when the Iraq war began in March 2003.

That general, Jack Keane, argues for sending an extra seven Army brigades and Marine regiments, about 32,000 troops, in two phases beginning in March and April.

The first infusion would be 25,000, followed by an additional 7,000 several months later. In his view, the current strategy of passing security responsibility to the Iraqis to establish the peace is failing.

Frederick W Kagan, an American Enterprise Institute scholar who wrote the plan with Keane and others, asserts that such a troop increase must be sustained for at least 18 months.

That is a tall order for a military stretched thin by its current commitment in Iraq.

General John Abizaid, top commander of US forces in the Middle East, told Congress in November that 20,000 more troops could be deployed, but the Army and Marine Corps are too taxed to sustain the increase for long.

Kagan dismisses the idea, championed by some in Congress, that a troop increase should last only four months to eight months.

"That would be a disaster," Kagan said in a telephone interview.

Those who argue for a longer-term increase worry that any short boost would give insurgents and others an opportunity to hunker down and wait until the extra troops went home.

Robert Burns

Washington 

Robert Burns has covered the military for the Associated Press since 1990.