When a body dies, it can lose weight in a lot of ways. Various sphincter muscles loosen, which can trigger the sudden release of feces and urine and other fun things. Moisture can evaporate after it's expelled through pores or via the "final breath" exiting the lungs. I guess if you're decapitated your body would lose something like 5 kilograms instantly. But as far as weight loss that doesn't have an obvious biological origin... we don't really know. It seems unlikely.

There's a fairly widespread belief that the instant a human dies, it loses a measurable and unexplainable amount of weight, which can only be attributed to the weight of the "soul" escaping. Oddly enough, often running in tandem with this belief is the idea that the soul — everyone's soul — weighs exactly 21 grams. This notion was no doubt popularized by the 2003 movie 21 Grams, but the movie didn't invent the idea — it's been around since the early 1900s, and can be specifically traced to experiments done by one Dr. Duncan MacDougall of Haverhill, Mass., beginning in 1901.

Dr. MacDougall wanted to prove that the soul existed as a material thing; that it took up space and had weight. Reasoning that dead people did not have souls, he decided to compare, as precisely as possible, the weight of a body immediately before and immediately after the moment of death. A sudden decrease in weight would prove the existence of a material soul.

You can probably see the ethical and logistical difficulties here — even in 1901, you couldn't just throw a person on a scale and ask him to stand there patiently while you killed him. MacDougall's bright idea was to set up shop at the Dorchester, Mass., Consumptives' Home, where he studied a series of six terminally ill patients dying from tuberculosis and other diseases. (Presumably they consented to be part of the experiment, as they had no real hope of recovery.)

MacDougall rigged up a contraption that placed a patient's bed on one side of a balance and weights on the other side — supposedly sensitive to two-tenths of an ounce. When it became apparent that a patient was about to expire, MacDougall and his assistants would carefully eye the balance to detect any sudden change in weight at the moment of death.

For patient No. 1, MacDougall wrote:

"The patient's comfort was looked after in every way, although he was practically moribund when placed upon the bed. He lost weight slowly at the rate of one ounce per hour due to evaporation of moisture in respiration and evaporation of sweat.

"During all three hours and forty minutes I kept the beam end slightly above balance near the upper limiting bar in order to make the test more decisive if it should come.

"At the end of three hours and forty minutes he expired and suddenly coincident with death the beam end dropped with an audible stroke hitting against the lower limiting bar and remaining there with no rebound. The loss was ascertained to be three-fourths of an ounce."

Eureka! There it was! Three-fourths of an ounce, or approximately 21 grams. Was this the weight of a soul?

MacDougall continued:

"This loss of weight could not be due to evaporation of respiratory moisture and sweat, because that had already been determined to go on, in his case, at the rate of one sixtieth of an ounce per minute, whereas this loss was sudden and large, three-fourths of an ounce in a few seconds.

"The bowels did not move; if they had moved the weight would still have remained upon the bed except for a slow loss by the evaporation of moisture depending, of course, upon the fluidity of the feces. The bladder evacuated one or two drams of urine. This remained upon the bed and could only have influenced the weight by slow gradual evaporation and therefore in no way could account for the sudden loss.

"There remained but one more channel of loss to explore, the expiration of all but the residual air in the lungs. Getting upon the bed myself, my colleague put the beam at actual balance. Inspiration and expiration of air as forcibly as possible by me had no effect upon the beam. My colleague got upon the bed and I placed the beam at balance. Forcible inspiration and expiration of air on his part had no effect. In this case we certainly have an inexplicable loss of weight of three-fourths of an ounce. Is it the soul substance? How other shall we explain it?"

After drawing conclusions from his six human subjects, MacDougall tried the experiment again with 15 dogs. (These were not terminally ill dogs — he couldn't find a reliable supply of them — but dogs who were given lethal injections that had the added benefit of rendering them immobile, and easier to weigh.) MacDougall found, or at least reported, that none of the dogs lost any weight — consistent with the belief that dogs don't have souls and humans do.

So what did these experiments prove? Well, nothing, actually. They proved that MacDougall, despite having the best of intentions, had his failings as a researcher.

Somewhat surprisingly, from all reports, Dr. MacDougall was neither mad scientist nor quack. He was a respected physician with an M.D. from Boston University (where he was class president), and he made great efforts to conduct his study "scientifically," controlling for all external factors, as shown in the second excerpt above. His findings were published in esteemed journals like American Medicine and the Journal of the American Society for Psychical Research, and the New York Times reported on his research. So you could say he had strong credentials.

But MacDougall's research was hampered by the limitations of 1900s technology and — especially — the small sample size of his study. He did his utmost to keep the patients in a "closed system" to get accurate readings, but it's not as if they were held in sealed containers on a digital scale accurate to 1/100th of a gram. Can we really believe a difference of three-quarters of an ounce was significant?

Plus, his study was limited to six patients, and only four of them were able to be satisfactorily weighed due to problems with the scale. Of those, only the first patient displayed the now famous 3/4 ounce (21 gram) instant drop in weight. The second didn't lose any weight until 15 minutes after his breathing stopped. The third lost 1/2 ounce at death, then another full ounce several minutes later. The fifth lost 3/8 ounce, but then the scale wouldn't rebalance when more weight was added.

Certainly one (and only one) vivid example of a body losing three-quarters of an ounce at death is not enough to conclude that the same happens to all humans.

MacDougall hoped that others would carry on his work, but few have bothered. The medical world and the law don't look favourably on the use of dying humans as guinea pigs for frivolous experiments. Does the soul have weight? Maybe. But probably not. And it would take more than experiments like MacDougall's to prove it anyway.

Sources: Lost Magazine: A Soul's Weight (this is a very interesting article!), Snopes.com, MacDougall's report in the Journal of the ASPR