The other curve (II) in the figure, that secured by the method of Identification, seemed to the investigators to be the most accurate. It is not subject to the errors due to expression and to contrast, and it has the advantage of allowing the subject the right to recognise the square. It is shown to him again, with no information that it is the same, and he decides whether from his remembrance of the earlier one, it is the same or not. The only objection to this method is that it requires a great many experiments in order to get an average result. To be reliable, an average must be secured, seeing that, for one or two or a few trials, the student may guess right without remembering the original square at all. By taking a large number of persons, such as the three hundred students, this objection may be overcome. Comparing the averages, for example, of the results given by the men and women respectively, we found practically no difference between them. This last point may serve to introduce a distinction which is important in all work in experimental psychology, and one which is recognised also in many other sciences--the distinction between results obtained respectively from one individual and from many. Very often the only way to learn truth about a single individual is to investigate a number together. In all large classes of things, especially living things, there are great individual differences, and in any particular case this personal variation may be so large that it obscures the real nature of the normal. For example, three large sons may be born to two small parents; and from this case alone it might be inferred that all small parents have large sons. Or three girls might have better memories than three boys in the same family or school, and from this it might be argued that girls are better endowed in this direction than boys. In all such cases the proper thing to do is to get a large number of cases and combine them; then the preponderance which the first cases examined may have shown, in one direction or the other, is corrected. This gives rise to what is called the statistical method; it is used in many practical matters, such as life insurance, but its application to the facts of life, mind, variation, evolution, etc., is only begun. Its neglect in psychology is one of the crying defects of much recent work. Its use in complicated problems involves a mathematical training which people generally do not possess; and its misuse through lack of exactness of observation or ignorance of the requirements is worse than its neglect.