1. After a critical analysis of Temperament theories, we were led to the conclusion that a satisfactory solution of the problem must in the first place account for all those behaviour qualities which have commonly been considered Temperamental. These qualities and behaviour patterns should be reducible to a few fundamental variables, functionally distinct, and capable of a physiological explanation. The resulting typology, a convenient classificatory scheme, should be dynamic so that each type is intelligible in the light of the interrelation of these variables. Only quantitative variation should be possible within each type, the temperamental make -up of an individual depending entirely on the ratio in which the variables are present in him. As the function of each in terms of behaviour, at any given strength, both singly and in combination, is known, his temperamental behaviour pattern can be accurately described.
The diagnosis of individual differences, either by assessment of criterion qualities or direct measurement of variables by means of clinical tests, should be possible. This would satisfy the requirements of both General and Individual Psychology.
2. Our analysis led to our acceptance of the three variables of the Heymans typology: Secondary - Primary Function, Activity - Inactivity, Emotionality - Non-Emotionality, as a basis for further research. We have found that they satisfy all the requirements stipulated: They are real behaviour Units each functionally distinct, though certain complex behaviour patterns can only be understood in terms of their interrelations; they can also have a modifying influence on each other's function. It is therefore necessary to distinguish between their true causative function (i.e. the quality concerned would not exist with a minimum value of the variable) and a modifying or facilitating function (i.e. the quality concerned would be somewhat quantitatively different, more or less marked or intense, with a minimum value of the variable.) The resulting Typology is dynamic; we found the attributes of each type entirely consistent with its variable ratio.
3. There is no evidence that distribution of types is different in different age-groups.
4. On our data, girls are more often Secondary Functioning than boys; contrary to the finding of other investigators, we found no difference in respect of Emotionality between the Sexes.
5. The theoretic function in terms of behaviour of the variable ratios can be modified by environmental influences. The precise nature of the environmental influence is not known. It can profoundly modify the operation of the Function mecha- nism, enhancing, facilitating, braking, the most pronounced qualities of each according to the degree to which habit for- mation has taken place. Environment favours the development of delinquency in the Primary Functioning, of extreme inertia in the Secondary Functioning. It may actually determine the strength of the Activity and Emotionality Variables.
6. A useful relationship has been discovered between two of the variables and Intelligence. Given a certain level of Intelligence, the Active are more likely to be scholastically eminent and effective than the Inactive, and within each group, the secondary more than the Primary. It is also highly pro- bable that the Active have more often a high I.Q. - have more "g" - than the Inactive.
7. A study of the various researches of the Heymans school and of our own, reveals a universality in the distribution of types in the community. In an unselected sample, the Active are more often Secondary Functioning and more often Emotional, than the Inactive; further that there is a tendency for the Primary to be more Emotional than the Secondary. Hey - mans attempted to explain this relationship on evolutionary grounds. Its true nature will only be clearly understood, when the organic basis of each of the variables is accurately known.
8. A study of the technique of Perseveration testing re- veals the inadequacy of nearly all tests in giving a measure of either the neuro- muscular or the subjective aspects of the concept. Yet there is strong evidence that such a factor exists, and tentative suggestions are made for its measurement.
A statistically significant relationship is established between the Function variable and that neural quality which determines the Flicker Threshold, which previous investigators have taken to be a manifestation of Perseveration. There is evidence that this relationship is explicable in terms of neural conduction and that the initial assumption as regards the neural basis of Secondary and Primary Function is warranted.
9. The Activity variable refers to willed Action and is distinct from mere physical Activity. It is probably identical with the conative element in Webb' s "w" factor.
10. The Emotionality variable cannot be measured by such tests as the Pressey and Woodworth Tests. Our evidence does not enable us to decide whether a unitary organic process, a central emotional core, underlies this behaviour variable.
It is probably determined, however, by some function or functions of the Endocrine glands and the Autonomic nervous system.
11. The Kretschmer Temperament groups, Cyclothymes and Schizothymes, are based on differences in Emotional discharge. Within each important qualitative Temperament differences occur and the scheme fails to satisfy our postulated requirements. Constitution does not single out subjectively unitary Temperament divisions. The Cyclothymes are largely made up of Alpine and Pyknic types on the organic side, Primary and Secondary Active types on tile subjective side. The schizothymes, at any rate in Kretschmer's group, largely of Asthenics and Asthenic Alpine types on the organic side, of Primary and Secondary Emotional Inactives on the subjective side. The Non -Emotional Inactives are not dealt with by this typology at all.
12. The Heymans Scheme thus forms the starting point for an adequate Temperament theory. The salient elements in other noted typologies can be accounted for on the basis of its variables in conjunction with environmental factors. By means of these variables, it has succeeded, more than any other theory, in reducing Subjective Temperament phenomena to a practical, useful and fundamental typology without sacrificing the dynamic subjective interrelation and the direct connection between this interrelation and organic processes.