The Real Story
In 1907, Duncan MacDougall of Haverhill, Massachusetts, in the USA, did in fact claim to have weighed the human soul... and controversy continues to this day regarding both the nature of his experiments and the interpretation of the results.
The controversy started in April 1907, when the New York Times published their short article stating that MacDougall had 'weighed the soul,' the details of which are essentially as the legend above describes them. MacDougall had not intended to publish his results so soon, but the clearly incorrect representation of his experiments that was publically published in the New York Times prompted him to pull his notes together and put down a more correct accounting of his strange proceedings. MacDougall's account of his experiments -- entitled "Hypothesis concerning soul substance together with experimental evidence of the existence of such substance" -- was simultaneously published in two journals, American Medicine and the Journal of the American Society for Psychical Research, in May 1907. In this article, MacDougall first explained why he believed the soul could be measurably weighed, and then briefly detailed the results of twenty-one experiments upon the matter; and in the Journal of the American Society for Psychical Research, further details and proof was provided in the form of correspondence that MacDougall had shared with a respected physician of the time, Dr. Richard Hodgson.
MacDougall's arguments are a mix of both then-current scientific beliefs about matter in the universe and philosophical beliefs regarding the nature of personality and spirit. Essentially, he argued that the soul must be a measurable, space-occupying substance within a human body largely because it required the ability to maintain an individual personality... and the only matter in the universe believed to not have weight was 'ether', which was also believed to be a single universally large connected layer of stuff with no variance, and therefore no ability to preserve a personality cast into it. So, in short, because MacDougall believed humans have souls that retain their personalities after death, he also believed these could not be part of the universe that was uniform (with no individual differences)... which meant this soul had to be made of some other matter, which scientific beliefs then dictated had to have a mass and weight of some sort. Having stated this belief, MacDougall arranged a series of macabre experiments to test it.
The inventive doctor had a specialized bed set up that was built upon a scale that could register a change in weight of as little as two-tenths of an ounce... and then he arranged for six people who were dying of different illnesses to do the dying part on his special bed as he monitored their weight (I imagine MacDougall was a very persuasive person!). During the experiments, all subjects were cared for and made comfortable, and any loss of weight caused by sweating and evaporation of water from the breath was carefully measured and compensated for. Upon cessation of life, any waste voided (urine or feces) remained on the bed/table, and so didn't effect the weight being monitored. His human subjects were all chosen for the fact that they were dying from conditions and diseases that did not lead to great physical movement, thus eliminating the possiblity of the scales being tipped simply by a violent action on the subject's part.
His first experiment, executed on April 10, 1901, was with a man dying of tuberculosis. After three hours and forty minutes of observation, and seemingly conincident with the man's death, the scales tipped suddenly and indicated a loss of three-fourths of an ounce from the man's frame.
Since this was the first experiment, MacDougall wondered if he had missed an important point in framing the test; namely, was he just measuring the air escaping from the man's lungs upon death? After his first experiment, MacDougall climbed onto the bed himself and had Dr. Sproull, who was assisting him, put the scales into balance. He then sucked and blew air with as much force as he could to see if the scales would register a change; they did not. He repeated this test with Dr. Sproull on the bed, and MacDougall himself measuring; again, there was no noticible movement in the scales. With this noted, MacDougall was soon testing his next subject.
Dr. Sproull also assited MacDougall during his second experiment, which took place sometime before November 10, 1901. The next subject was a man dying from consumption; he was on the bed for four hours before he apparently stopped breathing, but twitches of the eyelids and/or lips continued to be seen for another fifteen minutes after this point. After the last facial movements had been noted, the scales slowly tipped down over the next fifteen minutes, eventually noting a loss of one-half of an ounce. Dr. Sproull listened for a heartbeat, and could find none. MacDougall then re-checked his measurements... and now found that it took an ounce and a half, plus 50 grains more, to bring the scales into balance... apparently, an additional ounce and 50 grains of weight had gone missing within the three minutes that the man's heartbeat was being checked! However, due to this man's gradual demise, MacDougall was unable to be sure of what the exact moment of death for the patient was, which brought into question his ability to prove the weight loss was associated with it. The scales were re-adjusted to watch for further weight loss, but nothing else vanished for the remaining forty-five minutes of observation.
MacDougall's next four human experiments were performed sometime before May 22nd, 1902. His third subject, another man dying of tuberculosis, initially showed a loss of half an ounce in weight coincident with death, but an additional ounce was found to be missing after he was checked for a heartbeat... however, MacDougall admitted, the scales were jarred during the attempt to check for a heartbeat, and this likely caused the weight on the bar to shift and create the illusion that more weight had gone missing. Though MacDougall didn't say it, later researchers have been quick to point out that the same sort of mistake could easily explain the additional ounce and 50 grains thought lost after the heartbeat was checked in the doctor's second experiment.
More unfortunately, MacDougall's fourth subject -- a woman dying in a diabetic coma -- expired before he even set the scales to balance. In addition to this problem, the doctor also suffered from "interference by people opposed to our work;" thus he regarded this experiment of no value... but he still stated how much weight he estimated was lost. MacDougall's fifth subject, another man dying of tuberculosis, showed a three-eighths of an ounce weight loss upon his death. Strangely though, when weights were used to bring the scales back up again, the scales stayed up when the weights were removed; it was fifteen minutes before they fell back to proper alignment. This detail was basically reported, but otherwise ignored, by MacDougall, who then stated that "our scales in the case were very senitively balanced." MacDougall's sixth subject died too fast for the scales to be set properly, and so was also deemed of no value by him... but yet again, MacDougall was sure to mention what his estimate of the weight loss was, even though he couldn't justify it.
So, in total, MacDougall ran his experiment sucessfully on just four human subjects... and the results were different in each case. Coincident with his last four human experiments, MacDougall also ran twelve experiments using dogs as his subjects. The initial weight of the dogs varied from twenty to sixty-five pounds, and his scales were set to a sensitivity of a sixteenth of an ounce. In all cases, following the same precautions he observed with his human subjects, he had absolutely no change in the weight of the dogs as they died on his scales. A further three tests on canines had the same results. It must be noted at this point, however, that none of the dogs died of natural causes, as had the human subjects... instead, the dogs were subdued with two different drugs to prevent them from moving during the experiment. It is further implied by MacDougall's notes on the matter that the same drugs used to subdue the animals were, in fact, the likely agent that also killed them.
MacDougall concluded that the missing weight seen in his human experiments was in fact caused by the soul leaving their bodies. He doesn't explicitly say it, but it's also clearly implied in his notes that he feels the lack of weight loss in the dog experiments was an indication that they did not have souls... and he believed this would be true of animals in general. He also readily admitted that more experiments would have to be performed to verify his results, and actively asked Dr. Hodgson to encourage others to also perform his experiment so they could either confirm or deny the results he himself had acheived.