I settled on the following simple device: I used sheets of dark gray cardboard. On each were pasted white paper dots of different form and in an irregular order. Each card had between ninety-two and a hundred and eight such white dots of different sizes. The task was to compare the number of spots on one card with the number of spots on another. Perhaps I held up a card with a hundred and four dots above, and below one with ninety-eight. Then the subjects of the experiment had to decide whether the upper card had more dots or fewer dots or an equal number compared with the lower one. I made the first set of experiments with eighteen Harvard students. I took more than the twelve men who form a jury in order to reënforce the possible effect, but did not wish to exceed the number greatly, so that the character of the discussion might be similar to that in a jury. A much larger number would have made the discussion too formal or too unruly. The eighteen men sat around a long table and were first allowed to look for half a minute at the two big cards, each forming his judgment independently. Then at a signal every one had to write down whether the number of dots on the upper card was larger, equal, or smaller. Immediately after that they had to indicate by a show of hands how many had voted for each of the three possibilities. After that a discussion began. Indeed, the two cards offered plenty of points for earnest and vivid discussion. During the exchange of opinion in which those who had voted larger tried to convince the party of the smaller, and vice versa, they were always able to look at the cards and to refer to them, pointing to the various parts. One showed how the distances on the one card appeared larger, and another pointed out how the spots were clustered in a certain region, a third how the dots were smaller in some parts, a fourth spoke about the optical illusions, a fifth about certain impressions resulting from the narrowness of the margin, and a sixth about the effect of certain irregularities in the distribution. In short, very different aspects were considered and very different factors emphasized. The discussion was sometimes quite excited, three or four men speaking at the same time. After exactly five minutes of talking the vote was repeated, again at first being written and then being taken by show of hands. A second five minutes' exchange of opinion followed with a new effort to convince the dissenters. After this period the third and last vote was taken. This experiment was carried out with a variety of cards with smaller or larger difference of numbers, but the difference always enough to allow an uncertainty of judgment. Here, indeed, we had repeated all the essential conditions of the jury vote and discussion, and the mental state was characteristically similar to that of the jurymen.