But since our table shows this association to have an empirical value of 7.6 per cent, it becomes immaterial which of its logical relations is to be considered the strongest. It is mainly important, from our point of view, to separate reactions possessing an empirical value from those whose value is zero. § 6. AN EMPIRICAL PRINCIPLE OF NORMAL ASSOCIATION. On a general survey of the whole mass of material which forms the basis of the first part of this study, we are led to observe that _the one tendency which appears to be almost universal among normal persons is the tendency to give in response to any stimulus word one or another of a small group of common reactions_. It appears from the pathological material now on hand that this tendency is greatly weakened in some cases of mental disease. Many patients have given more than 50 per cent of individual reactions. It should be mentioned that occasionally a presumably normal subject has given a record very similar to those obtained from patients, in respect to both the number and the nature of the individual reactions. A few subjects who gave peculiar reactions were known to possess significant eccentricities, and for this reason we excluded their records from the thousand records which furnished the basis for the frequency tables; we excluded also a few peculiar records obtained from subjects of whom nothing was known, on the ground that such records would serve only to make the tables more cumbersome, without adding anything to their practical value. The total number of records thus excluded was seventeen. It will be apparent to anyone who examines the frequency tables that the reactions obtained from one thousand persons fall short of exhausting the normal associational possibilities of these stimulus words. The tables, however, have been found to be sufficiently inclusive for the practical purpose which they were intended to serve. Common reactions, whether given by a sane or an insane subject, may, in the vast majority of instances, safely be regarded as normal. As to individual reactions, they cannot all be regarded as abnormal, but they include nearly all those reactions which are worthy of special analysis in view of their possible pathological significance. What can be said further of individual reactions, whether normal or abnormal, will appear in the second part of this contribution. PART II. ASSOCIATION IN INSANE SUBJECTS. § 1. GENERAL SURVEY OF PATHOLOGICAL MATERIAL. The pathological material which forms the basis of the present part of our study consists mainly of two hundred and forty-seven test records obtained for the most part from patients at the Kings Park State Hospital. The different groups from which the cases were selected, together with the number from each group, are shown in Table I. TABLE I.