It is always interesting to note the type of errors made. The most common error is to omit one or more of the digits, usually in the first part of the series. If the child's ability is decidedly below the test he may give only the last two or three out of the five or six heard. Substitutions are also quite frequent, and if so many substitutions are made as to give a series quite unlike that which the child has heard, it is an unfavorable sign, indicating weakness of the critical sense which is so often found with low-level intelligence. In case of extreme weakness of the power of auto-criticism, the child in response to the series 9-8-1-7-6-, may say 1-2-3-4-5-6, or perhaps merely a couple of digits like 8-6, and still express complete satisfaction with his absurd response. After each series, therefore, the examiner should say, "_Was it right?_"[54] Very young subjects, however, have a tendency to answer "yes" to any question of this type, and it is therefore best not to call for criticism of a performance below the age of 6 or 7 years. [54] "_Was it wrong?_" is not an equivalent question and should not be used. Digit series of a given length are not always of equal difficulty, and for this reason it is never wise to use series improvised at the moment of the experiment. We must avoid especially series of regularly ascending or descending value, the repetition at regular intervals of a particular digit, and all other peculiarities of arrangement which would favor the grouping of the digits for easier retention. It remains to mention two or three further cautions in regard to procedure. It is best to begin with a series about one digit below the child's expected ability. If the child has a probable intelligence of about 6 or 7 years, we should begin with four digits; in case of probable 10-year intelligence we begin with five digits, etc. On the other hand, we should avoid beginning too far down, because then the result is too much complicated by the effects of practice and fatigue. It is not necessary, and often it is not expedient, to give the digits tests of all the different years in succession; that is, without other tests intervening. While this may be permissible with older children, in young children the power of sustained attention is so weak that no single kind of test should occupy more than two or three minutes. Children below 6 or 7 years should ordinarily be given the tests in the order in which they are listed in the record booklet. In his 1911 revision of the scale Binet unfortunately shifted this test from year VII to year VIII. Goddard follows his example, but Kuhlmann retains it in year VII. The data from more than a dozen leading investigations in America, England, and Germany agree in showing that the test should remain in year VII. VII, 4. TYING A BOW-KNOT