REMARKS. The ball-and-field problem may be called a test of practical judgment. Unlike a majority of the other tests, it gives the subject a chance to show how well he can meet the demands of a real, rather than an imagined, situation. Tests like this, involving practical adjustments, are valuable in rounding out the scale, which, as left by Binet, placed rather excessive emphasis on abstract reasoning and the comprehension of language. The test requires little time and always arouses the child's interest. Our analysis of the responses of nearly 1500 subjects shows that improvement with increasing mental age is steady and fairly rapid. Occasionally, however, one meets a high-grade performance with children of 6 or 7 years, and a low-grade performance with adults of average intelligence. Like all the other tests of the scale, it is unreliable when used alone. VIII, 2. COUNTING BACKWARDS FROM 20 TO 1 PROCEDURE. Say to the child: "_You can count backwards, can you not? I want you to count backwards for me from 20 to 1. Go ahead._" In the great majority of cases this is sufficient; the child comprehends the task and begins. If he does not comprehend, and is silent, or starts in, perhaps, to count forwards from 1 or 20, say: "_No; I want you to count backwards from 20 to 1, like this: 20-19-18, and clear on down to 1. Now, go ahead._" Insist upon the child trying it even though he asserts he cannot do it. In many such cases an effort is crowned with success. Say nothing about hurrying, as this confuses some subjects. Prompting is not permissible. SCORING. The test is passed if the child counts from 20 to 1 _in not over forty seconds and with not more than a single error_ (one omission or one transposition). Errors which the child spontaneously corrects are not counted as errors. REMARKS. The statistics on this test agree remarkably well. It is plainly too easy for year IX, and no one has found it easy enough for year VII. The main lack of uniformity has been in the adherence to a time limit. Binet required that the task be completed in twenty seconds, and Goddard and most others adhere rather strictly to this rule. Kuhlmann, however, allows thirty seconds if there is no error and twenty seconds if one error is committed. We agree with Bobertag that owing to the nature of this test we should not be pedantic about the time. While a majority of children who are able to count backwards do the task in twenty seconds, there are some intelligent but deliberate subjects who require as much as thirty-five or forty seconds. If the counting is done with assurance and without stumbling, there is no reason why we should not allow even forty seconds. Beyond this, however, our generosity should not go, because of the chance it would give for the use of special devices such as counting forwards each time to the next number wanted.