What then, is the significance of correlation in answering the problem of causation? So far as the statistical method itself is concerned it shows only a mathematical functional relation between the conditions measured, not a physiological relationship. In other words a correlation between deficiency and delinquency might be explained by both conditions being related to some more fundamental factor which might be the causal factor involved. One cannot reason from correlation to direct causal connection. On the other hand, by correlation we may directly compare the relation between any one trait and various factors. We can find out, for example, whether the association of delinquency with deficiency is closer than the association of delinquency with other factors which it has been suggested are causes of delinquency. Goring's work allows us to compare the correlation of the tendency to be convicted of crime with deficiency and with many other constitutional and environmental factors which have been measured, and thus our attention may at once be directed to that factor which the present evidence indicates as most fundamental. Unless the measurement of the various factors is shown to be seriously faulty or incomplete the outcome should determine our point of view as to the main cause of delinquency, until new evidence is forthcoming. This is the problem of the next section. C. THE CAUSES OF DELINQUENCY. As we have noted above, the correlation of delinquency with various factors should give us a scientific point of view as to the main causal influence in criminality. Thanks to Dr. Goring this work has recently been carried far. His findings mark a new and higher scientific level in the study of criminology. No data are now available which modify his position in any important regard. I shall, therefore, attempt to give his evidence in the briefest possible manner, hoping that it may lead to a closer reading of his basal investigation. (a) CONSTITUTIONAL FACTORS. First comparing a dozen factors in the individual's own constitution which may be measured by the death rates, Goring found the tendency to be convicted of crime was correlated most closely with alcoholism, .39; sexual profligacy (syphilis and aneurism), .31; and epilepsy, .26; while it was found to correlate with intelligence, .66. The closeness of the relationship of defective physique to criminality was expressed by coefficients of .18 and .19. Among the inner factors investigated were many of Lombroso's characteristics of the so-called criminal physiognomy of which so much use is made by phrenologists, such as asymmetries, projection of the chin, complexion, form of the face and features, kind of hair, tattooing, left-handedness, temperament, etc.