It is not always easy to judge whether or not a given reaction should be classed as non-specific. A study of our material made with special reference to this type of reactions has enabled us to select the following list of words, any of which, occurring in response to any stimulus word, is classed as a non-specific reaction: article, articles bad beautiful, beauty fine good, goodness great happiness, happy large man necessary, necessity nice object (noun) people person pleasant, pleasantness, pleasing, pleasure pretty small thinking, thought, thoughts unnecessary unpleasant use, used, useful, usefulness, useless, uselessness, uses, using woman work It should be mentioned that some of these words occur as reactions to one or several stimulus words with such frequency (_citizen--man_, value 27.8 per cent; _health--good_, value 9.4 per cent) as to acquire in such instances a value as high as that of strictly specific reactions. *Doubtful Reactions* have already been defined (p.40): any reaction word which is not found in the table in its identical form, but which is a grammatical variant or derivative of a word found there, is placed in this group. § 4. INDIVIDUAL REACTIONS; EXPLANATION OF GROUPS AND METHODS OF APPLICATION. *Normal Reactions.*--Inasmuch as the frequency tables do not exhaust all normal possibilities of reaction, a certain number of reactions which are essentially normal are to be found among the individual reactions. In order to separate these from the pathological reactions, we have compiled an appendix to the frequency tables, consisting mainly of specific definitions of groups of words to be included under each stimulus word in our list. This appendix will be found at the end of this paper. A word of explanation is perhaps due as to the manner in which the appendix has been compiled. It was developed in a purely empirical way, the basis being such individual reactions, given by both normal and insane subjects, as seemed in our judgment to be obviously normal. It must be acknowledged that the appendix falls short of all that might be desired. In the first place, its use involves to some slight extent the play of personal equation, and it therefore constitutes a source of error; in the second place, it is in some respects too inclusive while in other respects it is not sufficiently so. However, the error due to personal equation is slight; the inclusion of certain "far-fetched" or even frankly pathological reactions may be discounted by bearing in mind that the general value of this group is not equal to that of the group of common reactions; and the number of strictly normal reactions which are not included is after all small. Our experience has shown us that the appendix constitutes an important aid in the analysis of individual reactions.