[17] Freud, Selected Papers on Hysteria and Other Psychoneuroses, 3d edition, translated by A.A. Brill, N.Y. Nerv. and Ment. Dis. Pub. Co. Nervous and Mental Disease Monograph, Series No. 4. [18] Havelock Ellis, in an appendix to his study on the Sexual Impulse, 1903, gives a number of autobiographic reports of normal persons treating their first sexual feelings in childhood and the causes of the same. These reports naturally show the deficiency due to infantile amnesia; they do not cover the prehistoric time in the sexual life and therefore must be supplemented by psychoanalysis of individuals who became neurotic. Notwithstanding this these reports are valuable in more than one respect, and information of a similar nature has urged me to modify my etiological assumption as mentioned in the text. [19] The above-mentioned assertions concerning the infantile sexuality were justified in 1905, in the main through the results of psychoanalytic investigations in adults. Direct observation of the child could not at the time be utilized to its full extent and resulted only in individual indications and valuable confirmations. Since then it has become possible through the analysis of some cases of nervous disease in the delicate age of childhood to gain a direct understanding of the infantile psychosexuality (Jahrbuch für psychoanalytische und psychopathologische Forschungen, Bd. 1, 2, 1909). I can point with satisfaction to the fact that direct observation has fully confirmed the conclusion drawn from psychoanalysis, and thus furnishes good evidence for the reliability of the latter method of investigation. Moreover, the "Analysis of a Phobia in a Five-year-old Boy" (Jahrbuch, Bd. 1) has taught us something new for which psychoanalysis had not prepared us, to wit, that sexual symbolism, the representation of the sexual by non-sexual objects and relations--reaches back into the years when the child is first learning to master the language. My attention has also been directed to a deficiency in the above-cited statement which for the sake of clearness described any conceivable separation between the two phases of autoerotism and object love as a temporal separation. From the cited analysis (as well as from the above-mentioned work of Bell) we learn that children from three to five are capable of evincing a very strong object-selection which is accompanied by strong affects. [20] Some persons can recall that the contact of the moving air in swinging caused them direct sexual pleasure in the genitals. [21] "Those who love each other tease each other."