A little observation will convince one that mentally retarded subjects, even when they possess a reasonable degree of fluency in recognizing printed words, do not sense shades of meaning. Their reading is by small units. Words and phrases do not fuse into one mental content, but remain relatively unconnected. The expression is monotonous and the voice has more of the unnatural "schoolroom" pitch. They read more slowly, more often misplace the emphasis, and miscall more words. In short, one who has psychological insight and is acquainted with reading standards can easily detect the symptoms of intellectual inferiority by hearing a dull subject read a brief selection. The giving of memories is also significant. Feeble-minded adults who have been well schooled are sometimes able to read the words of the text fairly fluently, but are usually unable to give more than a scanty report of what has been read. The scope of attention has been exhausted in the mere recognition and pronouncing of words. In general, the greater the mechanical difficulties which a subject encounters, the less adequate is his report of memories. The test has, however, one real fault. School children have a certain advantage in it over older persons _of the same mental age_ whose school experience is less recent. Adult subjects tend to give their report in less literal form. It is necessary, therefore, to give credit for the reproduction of the ideas of the passage rather than for strictly literal "memories." The selection we have used is, with minor changes, the same as Binet's. His selection was divided into nineteen memories. The one here given has twenty-one memories. Binet used the test both in year VIII and year IX, requiring two memories at year VIII and six memories at year IX. When we require eight memories, as we have done, the test becomes difficult enough for non-selected school children of 10 years. Location in year X seems preferable, because it insures that the child will almost certainly have had the schooling requisite for learning to read a selection of this difficulty, even if he has started to school at a later age than is customary. Naturally, placing the test higher in the scale makes it more a test of report and less a test of ability to recognize and pronounce printed words. X, 5. COMPREHENSION, FOURTH DEGREE The questions for this year are:-- (a) "_What ought you to say when some one asks your opinion about a person you don't know very well?_" (b) "_What ought you to do before undertaking (beginning) something very important?_" (c) "_Why should we judge a person more by his actions than by his words?_" The PROCEDURE is the same as for the previous comprehension tests. Each question may be repeated, but its form must not be changed. It is not permissible to make any explanation whatever as to the meaning of the question, except to substitute _beginning_ for _undertaking_ when (b) seems not to be comprehended.