Thorndike and Rock 1 have offered evidence that "a satisfying after-effect could strengthen the connection which it followed and to which it belonged in cases where the learner did not know what the connection was." The value of this evidence, however, rests upon an assumption which seems open to question, namely, that lack of awareness of what was being learned was indicated by continued slow 'learning' without a sudden increase in the percentage of successes. This point can best be made clear by a description of the procedure of the experiment in question.The instructions to the subjects were as follows: 2"'I shall say 'Ready' and then say a word. You will say the first word or words that you think of. I will record the time in fifths of a second. Then I will say another word. You will say the first word or words that you think of. I will record the time as before. We will continue this. At the end of the experiment, you will receive a money bonus depending upon how quickly you respond on the average. Also I shall say 'Right' if the word that you say is one of ten or more that we have arbitrarily decided to call right as responses to the word in question. I shall say 'Wrong' if the word that you say is one of ten or more that we have arbitrarily decided to call wrong as responses to the word in question. Sometimes I shall say nothing because the word that you say is in neither our list of Rights nor our list of Wrongs for the word in question. At the end of the experiment, you will receive a money bonus depending on a system of credits for Rights and penalties for Wrongs. Speed is the main thing. The number of Rights versus Wrongs is a minor thing.'" A number of words were then read to him, one at a time. "In each case, if his response is clearly due to the sequential connections used in speaking and writing • • • he is rewarded by the announcement 'Right.' If his response is