We will never know the full civilian toll of the Iraq War. Two years ago a Johns Hopkins University public health study concluded that something like 100,000 civilians had been killed in the first 18 months of the Iraq War, more than half of them women and children and most as a result of coalition action.

		Barbara Bodine, a visiting scholar at the MIT Center for International Studies, was U.S. ambassador to Yemen and twice served as a U.S. diplomat in Iraq, including first coordinator for reconstruction in Baghdad and the central provinces in 2003.

The science was valid but arcane; our presidential campaign was in full swing. The numbers contradicted our notion of how we conduct war and contravened the post-Vietnam aversion to "body counts." The report was dismissed. Commentators now cite the 100,000 figure as accepted fact.

That number is tragically out of date. Johns Hopkins researchers returned to Iraq in the summer of 2005 to update their report with the help of al-Mustansiriya University in Baghdad and funding from the Massachusetts Institute of Technology. Their conclusion: more than 600,000 Iraqis, 2.5 percent of the population, have died violently since 2003.

This is a staggering number. How can this possibly be? Unquestionably violence in Iraq has climbed exponentially, despite the best efforts of coalition and Iraqi forces, while the source has shifted from primarily insurgency to sectarian civil war. Despite what seems to be unceasing media reports about coalition military operations, jihadi car bombings, sectarian executions and assassinations, increasingly the violence in Iraq is out of our sight.

Much of Iraq is off-limits to Westerners and much of Iraq is too insecure for Iraqis. The Iraq Body Count and the U.S. Defense Department's numbers extrapolate from press reports, morgue statistics, anecdotal evidence and other second-hand information. But reports of deaths in the Arabic and English press focus on the spectacular or high level. And inaccessible government facilities and the Muslim dictate to bury the dead by sunset, if possible, distort the validity of hospital and morgue numbers.

The slow attrition of professionals, bureaucrats, technocrats, even barbers and bakers -- by one source there is one assassination per hour in Basra -- goes underreported.

Passive surveillance tallies, such as those conducted by The Iraq Body Count and Defense Department, do monitor trends. The proportion of violent deaths attributed to coalition forces may have peaked in 2004, although the actual number continues to climb. Gunfire remains the most common cause of death, although those attributable to car bombings have increased steadily.

The Johns Hopkins figures reflect the same trends. But unlike numbers dependent on passive surveillance, the Hopkins survey again was undertaken, at great personal risk to its teams, throughout Iraq by professionally trained Iraqis who visited a broad range of households. The numbers and their details as a result are significantly higher.

It doesn't take a scientist to understand the numbers, the methodology or why direct surveys would produce higher estimates than passive surveillance. The method is the same accepted worldwide to measure public health crises such as epidemics -- 47 clusters (almost twice that of the 2004 survey) among 16 of 18 governorates were examined. The teams were medical doctors with previous survey experience. The details of this survey can be fairly debated, but the science is good; the analysis conservative; the conclusions objective.

What should be openly debated this time around are the implications of these numbers. Whatever solace we may try to find in the decline in the proportion of violent deaths attributable to coalition forces, irrespective of the rise in absolute numbers, is more than offset by the rise in sectarian and jihadi car bombings, mass executions and assassinations.

With each spin in the cycle of violence the legitimacy, the viability, of an Iraqi government unable to protect its people diminishes. Perhaps there can be no legitimacy without security, but the reverse holds true as well. Efforts to train a national army and a police force work only if there is a state that commands the loyalty of the army and the police and also is recognized and accepted by the people as legitimate, not just legal. Without that, you are not training an army or a police force, just more effective militia.

What the Johns Hopkins numbers suggest is that the human infrastructure of Iraq is being systematically destroyed. This is a population decimated by eight years of grotesque war with Iran, a brief but bloody run-in with the world during Desert Storm, and an assault on its own Kurdish and Shiite populations. It is a nation that was left staggered under 10 years of sanctions while being bled by a kleptocracy. Generations of young men have been repeatedly purged, which continues. Add the flight of the middle class and the technocrats, and the question becomes -- who will be left to rebuild Iraq?