B H ^=^fl Mbi 
 
 \U\ 
 
 I 
 
 lliii;^^ 
 
 «iu 
 
 II 
 
 i 
 
 ■ 1 1 t "^ 
 
 h ''.'<'. 
 
 
 

 -S 
 
A Study of Memory 
 
 FOR 
 
 Connected Trains of Thought 
 
 BY 
 
 E. N. HENDERSON 
 
 Submitted in Partial Fulfillment of the REqyiREMENTs for the 
 
 Degree of Doctor of Philosophy in the Faculty of 
 
 Philosophy, Columbia University 
 
LT3/0&3 
 
 pRies OP 
 
 IKE NEW ERA PRINTINO COMPANY, 
 LANCASTER, PA. 
 
CONTENTS. 
 
 Page. 
 
 Introductory. Education and Experimental Psychol- 
 ogy I 
 
 Part I. Education and the Experimental Study of 
 
 Memory 6 
 
 Educational Bearing of Certain Experimental Work on 
 
 Memory 7 
 
 Data Gathered from Experiments with Classes of School 
 
 Children 8 
 
 Results of Laboratory Experimentation ii 
 
 Summary of Results 19 
 
 Part II. The Special Research 26 
 
 The Method 26 
 
 The Tests 28 
 
 The Classes of Subjects tested 31 
 
 The Results 32 
 
 A. Amounts of Loss ,'. 3^ 
 
 Effect of Age and Training on Power to Learn and 
 
 to Remember 38 
 
 Correlation of Scores According to Ability 40 
 
 Comparison of Results in Tests 3, 4, and 5 49 
 
 Effect of Method of Learning on Results 50 
 
 Comparison of Standing in Tests with Teacher's 
 
 Marks 51 
 
 Summary of Conclusions on Amountsof Loss 52 
 
 B. Character of the Changes in the Later Reproductions 53 
 
 1. Recognized Words 54 
 
 2. Accounts of Introspection of Graduate Students. 58 
 
 3. Comparative Strength of Memory for Details 
 
 and for Larger Topics 62 
 
 4. History of the Ideas regarding the Contents of 
 
 the Passages 64 
 
 a. Types of Changes in the Reproductions 65 
 
 b. Modifications in the Reproduction of ' The 
 
 Dutch Homestead.' 70 
 
 iii 
 
 O '- •- <-\ 
 
 fl 
 
IV CONTENTS. 
 
 c. Comparison of the Generalizing Processes 
 
 in Tests 3, 4 and 5 75 
 
 d. Comparison of tlie Generalizing Processes 
 
 in Adults and Children 80 
 
 e. Extent of Loss of Details by Generalizing 
 
 Processes 82 
 
 f. Summary 83 
 
 Educational Significance of the Results of theSpecial Research 85 
 
 Appendix 88 
 
■Ti. 
 
 INTRODUCTORY. EDUCATION AND EXPERIMEN- 
 TAL PSYCHOLOGY. 
 
 Any one who, working from the point of view of a student 
 of education, ventures upon the territory of the experimental 
 psychologist is forced to come to some conclusion, for better or 
 for worse, as to whether experimental psychology has anything 
 to give to the intelligent appreciation of education. A decision 
 on this matter will rest largely upon point of view. One's gen- 
 eral philosophy will go far toward determining whether one says 
 that the teacher has no real use for the experimentalist or that 
 education must be revolutionized by this comparatively new sci- 
 ence. From the historical point of view, it is interesting to 
 note that consistency should keep experimental psychology the 
 willing bond-servant of the science of education. For, if we 
 regard the attempt of Herbart to deal quantitatively with con- 
 sciousness as furnishing the initial impulse toward the develop- 
 ment of experimental study of mental phenomena, then it is 
 plain that in its beginnings this science was created to further 
 an analysis, the great aim of which was a theory of teaching. 
 Herbart was primarily interested in education, and his psycho- 
 logical theories were probably influenced largely by the bias of 
 his educational principles. The experimental psychology that 
 gives us what it declares to be a purely academic analysis of 
 consciousness — an analysis professedly useless to the teacher 
 — such a science is, indeed, an ungrateful child. As a matter 
 of fact, is it not likely that a science of education, if ever there 
 is such a subject, must be based on a psychology that does not 
 fall short of a study of acts of will in such a way as to further 
 practical efficiency in dealing with children and men? May 
 not our generation be justified in saying not only that knowl- 
 edge is power, but that all significant knowledge must be a 
 source of practical strength? And if we have so far not pro- 
 foundly modified educational practice by reason of our psycho- 
 
2 E. N. HENDERSON. 
 
 logical investigations, should we lose hope? The Baconian 
 method realized signally feeble results in the hands of Bacon 
 himself. After all, will not any verdict on the relation of ex- 
 perimental psychology to education have to wait a few more 
 years on practical results? It is in the hope that we may be 
 working toward such results, possibly building better than we 
 know, or than those about us know, that the following study is 
 submitted, claiming to be on the one hand in the interest of ed- 
 ucation, and on the other a venture in the field of experimental 
 psychology. 
 
 As a matter of fact, it is only by addressing itself to peculiarly 
 educational problems that experimental psychology is apt to 
 gain any results of substantial educational importance. The 
 teacher, reading over the dreary tables of data that record ex- 
 perimental results, finds, it is true, a little that is significant for 
 his purposes. Possibly it would be unjust to say that this is so 
 little that it does not repay the effort expended in its search. 
 Certainly, where the investigations were conceived in the in- 
 terest of education, we may expect and get something. The 
 so-called * child-study ' movement has aimed to place teaching 
 on a ' scientific ' basis. As a rule, however, it has been scarcely 
 scientific enough itself, and many of its results are insufficiently 
 criticised. But it would be hard to measure the effect of the 
 spirit and attitude of child study upon ways of thinking and 
 methods of work of teachers, and upon the curriculum gener- 
 ally. However unscientific from the point of view of a labora- 
 tory science, it was, nevertheless, as a rule more accurate and 
 conclusive than mere common sense, and though it may have 
 worshipped idols, they probably replaced the fetishes of a more 
 debased superstition. The teacher who endeavors to apply the 
 theory of culture epochs will very likely work more intelligently 
 than one who simply follows in blind routine the grind of 
 centuries. 
 
 Child study may be said to have stood for the thesis that the 
 school programme should conform to the needs of the child 
 that is, rather than those of the adult that is to be. In fact, it is 
 claimed that only in being true to the nature of the child can we 
 properly prepare for adult life. But the emphasis is laid, not 
 
A STUDY OF MEMORY. 3 
 
 SO much on education as preparation for the future, as on edu- 
 cation as interesting and satisfying for the present. Thus 
 inspired, the new education has introduced a muUitude of de- 
 vices to enrich the work of the school-room, to provide motives 
 for it, and to render it interesting. It has done this, confident 
 that the essential preparation for life may be gained far more 
 easily and effectively in this way than by the unenlivened dis- 
 cipline of the old-fashioned masters. It has assumed that the 
 natural way of doing things is the effective way, and has busied 
 itself in hunting the natural way — which was to be revealed 
 by the study of child nature, of the genesis of mind. But the 
 assumption remains as yet unverified. No one has proved that 
 the new methods of education are better than the old in prepar- 
 ing for life or even in teaching the rudiments of learning. The 
 new education has, indeed, striven to be practical, and it may 
 with justice be granted that a persistent attempt to be a thing 
 ought ordinarily to approximate nearer to it than haphazard 
 work. But it can scarcely be affirmed that the old education 
 neglected the practical. In that it aimed to store up in child- 
 hood treasures that were to be used much later, it was apt to be 
 miserly and to collect riches that were never used. Worse ! 
 The thief of forgetfulness all too often stole away most of its 
 accumulations. On the other hand, while the new education may 
 not build palaces never to be occupied, it runs the risk of not 
 getting beyond mere play-houses. May we not question whether 
 the palace crumbled into ruins will not be as useful to the man 
 as the toy structures of the child? Does it not remain to be 
 proved that the games, the imitations, the natural occupations 
 of the child furnish the most effective means of preparing him 
 for the life that is to come? Is that education which is practical 
 in preparing for immediate needs more practical in the long 
 run than the one that thinks almost solely of the remote future? 
 I do not think that we can doubt that our modern schools are 
 more interesting than their forerunners. The old-fashioned 
 schoolmasters were not concerned very much in pleasing the 
 child. But as yet we have no means of determining adequately 
 the effectiveness of results in education. That our educational 
 methods succeed, too often means that they gratify the preju- 
 
4 E. N. HENDERSON. 
 
 dices of parents or excite the interest of pupils, rather than that 
 they prepare adequately for the future. The judgment of time 
 may often be counted on to eliminate the unfit in this business 
 of training for life. But a school ought to be able to test its 
 methods before they land it in bankruptcy. Nor should its 
 success, like that of a dramatic company, depend on the ability 
 to gratify the fancy. We need usable criteria for determining 
 the success of school work. These criteria should be based on 
 the needs of life rather than on the work of the school. Other- 
 wise they merely test the child's success in certain work with- 
 out reference to the real value of the work itself, and the real 
 standing of the child in the race of life. Such tests would be 
 able to anticipate the careers of our graduates more confidently 
 than those now in vogue. 
 
 In this period of educational experimentation, of storm and 
 stress in school circles, the need of scientifically determined 
 methods of examination by which we can get something like an 
 adequate test of the effect of our various school methods is 
 especially desirable. From whom are these methods of testing 
 to come if not from the educational psychologist? Certainly no 
 one is in a better position to view various subjects and methods 
 impartially. The specialist in any department is apt to have 
 his eyes turned toward his course of study. He will examine 
 to find whether his students have learned what was taught them. 
 Doubtless this is a very important kind of test. But who is to 
 determine what value this learning has for life? Some will say 
 that there is no way of settling this question except by waiting 
 for the practical test of living. Unquestionably this is in many 
 cases true ; possibly in all cases. Yet if it were uniformly so, 
 who is to collect the facts by which the verdict of life is brought 
 home to the teacher? For this is not so simple a matter as it 
 may appear. Suppose we could show that most of our ablest 
 men knew Greek. Would this show that their ability was due 
 to their study of Greek, or that, as Professor Thorndike sug- 
 gests, most people with considerable ability happen to have 
 been taught Greek. To disentangle apparent from real evi- 
 dence in these matters, to determine in how far a college exami- 
 nation in history is comparable to a life examination in the same 
 
A STUDY OF MEMORY. 5 
 
 subject, or in the citizenship to which it is supposed to con- 
 tribute ; such tasks certainly require special training, ability to 
 gather a special kind of evidence, and the entire time of a 
 special body of workers. A hopeless task, you may think, but 
 certainly the matter is of such importance as to make any gain 
 that may result from substituting scientific methods for guess- 
 work a gain worth while. To settle by adequate tests the 
 relative efficiency of different methods of learning to read, to 
 write, to compute, would mean immense advance. Indeed, the 
 imagination can picture an age in which the random work of 
 our schools will be as much surpassed as is the scientific method 
 of antiquity by modern experimentation. In such an age no 
 one will mistake even experienced judgment for proof, and the 
 number of questions in regard to which men will plunge into 
 controversy instead of resorting to investigation will have been 
 materially reduced. 
 
 It would seem then a not insignificant division to say that 
 educational psychology works at two problems. The first is 
 that of the nature of mental growth, and the second concerns 
 the effect of mental training. These problems are mutually 
 interdependent. Yet from the educational point of view each 
 is worthy of distinct treatment. For the one is concerned with 
 processes, the other with results ; the former starts on foot new 
 methods, the latter investigates the effect of these methods. 
 That has its eyes turned toward the child, this directs its gaze 
 toward the man. Both can profitably employ, indeed both 
 must, to be effective, employ some experimentation. Statis- 
 tical investigation may or may not be experimental. Methods 
 of testing are themselves to be determined and justified only by 
 experiment, and they are the means of determining the results 
 of still more extensive educational experiments. We are not 
 without examples of this psychology of the effects of training. 
 The study by Professors Thorndike and Wood worth, ' The 
 Influence of Improvement of One Mental Function upon the 
 Efficiency of Other Functions'' is an illustration. Such re- 
 searches offer evidence that experimental psychology is not 
 utterly disconnected with education, and encourage further 
 study along similar lines. 
 
 1 Psych. Rev., 1901, Vol. VIII., Nos. 3, 4, 6. 
 
E. N. HENDERSON. 
 
 PART I. EDUCATION AND THE EXPERIMENTAL 
 
 STUDY OF MEMORY. 
 
 Any test of the result of school work is at bottom a test of 
 memory. We may raise on our educational banners the rally- 
 ing cry 'not information but culture, not knowledge but power, 
 but the culture and the power are none the less results of the 
 activity of a function that may best be given the general name, 
 memory. A habit, we must admit, is only a memory ground in 
 so well that we can act effectively without conscious adjustment. 
 
 Wherein the information-giving type of education may fall 
 short is, therefore, not in its attempt to give knowledge, but in 
 its failure to select the kind of knowledge that can most effec- 
 tively be stored in memory, and to see to it that this knowledge 
 becomes part of the life of the individual. The reforming 
 maxim that we should ' learn nothing by heart ' owes what- 
 ever truth it contains to its implied criticism of the knowledge 
 that is committed to memory in this way. The great problem 
 of education is, after all, how most usefully to store the memory, 
 how to create the habits of thought and action that will be 
 most valuable in future life. Hence the examination, or test 
 of what is committed to memory, is an indispensable part of 
 school practice. No system can get along without it. Exam- 
 inations mean standards : definite, positive work. To have no 
 examinations means loose work and uncertainty about results. 
 The character of the examination is the test of the aim and 
 quality of the course of study. A purely academic examina- 
 tion is indicative of a purely academic course. The most practi- 
 cal examination is that which requires an application of the 
 material committed to memory. This sort of examination is in 
 many lines difficult, yet teachers are coming to feel that it is 
 the only true method. It is not quite so easy to grade pupils 
 by it as by answers involving information in the specific form 
 in which that information was given, but, what is far more 
 
A STUDY OF MEMORY. 7 
 
 important, the examination of this sort correlates far more 
 closely with that given by life itself. As a matter of fact, it is 
 not so easy to determine the grading of life as it is to range test 
 papers according to percentages. 
 
 If it be admitted that the storing of memory is the object of 
 education, and that the examination should exist to reveal the 
 results of the training, any investigations of the manner in 
 which memory operates and how best to test it are certainly 
 important. Teachers have been working on these problems so 
 long that a well-criticised and fairly effective school practice 
 may be said to exist. Yet with all the reading of examination 
 papers that goes on, there has been, so far as I know, no care- 
 ful study of them as revelations of a psychological process. 
 For one thing, conditions are so different with different pupils 
 that the examination tests rather these various conditions than 
 the fate in the normal mind of the material presented in a cer- 
 tain way. Teachers do indeed recognize in the examination 
 a means not only of grading the pupils, but also of estimating 
 their own work. But this estimate is made in a very uncritical 
 way. The average of the marks of the class may, it is true, be 
 taken as the measure of whether the teacher has gained her 
 aim, but it indicates no definite relationship between conditions 
 of presentation and retention. The experienced examiner knows 
 in a general way what questions she may expect to have an- 
 swered by the average student. But we may hope, I take it, 
 far more from investigations conducted under controlled and 
 known conditions than from mere general experience. 
 
 Educational Bearing of Certain Experimental Work 
 
 ON Memory. 
 
 It has already been suggested that here and there among 
 the reports of psychological experimentation may be found 
 material of interest to the teacher. This is possibly even more 
 true of work on memory than of that in any other field. Some 
 account of it may not be amiss in a paper of this sort. The 
 material may be said to consist of two main kinds : (i) Data 
 gathered from experiments with classes of school children ; (2) 
 results of laboratory experimentation that have an educational 
 value. 
 
8 E. N. HENDERSON. 
 
 I. Data Gathered f 7' om Exferhnents -with Classes of 
 
 School Children.^ 
 
 This material we may naturally suppose to be of especial 
 interest to the teacher. The method of most of the experiments 
 was that of reading or exhibiting at a uniform rate a series of 
 numbers, letters, syllables, words or objects, and asking the 
 children to write out the series in the original order. The num- 
 ber of units in the series was by some observers varied, and 
 the results noted. Others contented themselves with a fixed 
 number of units, but the nature of the units in the series was 
 varied. All observers tested children of various ages, and 
 some contrasted the abilities of girls and boys. The results 
 may be divided into two classes : (i) Those indicating the 
 growth of power to reproduce the series with increasing age ; 
 (2) Those indicating the kind of material easiest reproduced, and 
 the methods of presenting the series that were most efficacious. 
 
 I. Comparing such of the data of different observers as are 
 sufficiently similar to admit of this, we discover that they agree 
 in finding that the older children reproduce the series better.^ 
 
 1 The researches the results of which have been considered in this review 
 are : (i) 'The Growth of Memory in School Children,' T. L. Bolton, Am. Jour- 
 nal of Psych., Vol. TW .,-p^. 362-380. (2) ' Memory, an Experimental Study,' 
 E. A. Kirkpatrick, Psych. Rev., 1S94, pp. 602-609. (3) ' Influence de I'age sur 
 la memoire immediate,' B. Bourdon, Revue Phil., No. 38, pp. 148-167, {4) ' La 
 m^moire des mots,' A. Binet et V. Henri, VAnnee Psych., Vol. I., pp. 1-23. 
 Also 'La memoire des phrases,' pp. 24-59. (s) 'Experiments on Memory 
 Types,' C.J. Hawkins, Psych. Rev., Vol. IV., pp. 289-294. (6) ' Ex. Unter- 
 such. ii. d. Gedachtnissentwickelung b. Schulkindern, A. NetschajefF, Ztsch. 
 filr Psych., Vol. XXIV., pp. 321-351. (7) Research with same title, MarxLob- 
 sien, Ztsch. fur Psych., Vol. XXVII., p. 34. 
 
 2 The most evident comparison lies between the results of Bolton and Bour- 
 don, both of whom read series of digits at a uniform rate to classes, and asked 
 for written reproductions. Bolton's scores indicated amounts of forgetting, but 
 they can be converted into scores of amount remembered, and in this form they 
 appear in the following table. The scores indicate the average percentage of 
 numbers in a series retained by children of a certain age. 
 
 Age of Subjects. 
 
 Experi- 
 
 Nos. of 
 
 8 
 
 9 
 
 10 
 
 II 
 
 12 
 
 13 
 
 14 
 
 15 
 
 menter. 
 
 Digits. 
 
 Years. 
 
 Years. 
 
 Years. 
 
 Years. 
 
 Years. 
 
 Years. 
 
 Years. 
 
 Years. 
 
 Bolton, 
 
 6 
 
 43-7 
 
 54-5 
 
 57.6 
 
 67.6 
 
 68. 7 
 
 72.3 
 
 76.5 
 
 74-7 
 
 Bourdon, 
 
 6 
 
 22 
 
 59 
 
 56 
 
 61 
 
 74 
 
 43 
 
 90 
 
 47 
 
 Bolton, 
 
 7 
 
 22 
 
 35-4 
 
 33.8 
 
 37-5 
 
 48.5 
 
 50.8 
 
 639 
 
 59-3 
 
 Bourdon, 
 
 7 
 
 
 
 13 
 
 28 
 
 5 
 
 37 
 
 33 
 
 43 
 
 40 
 
 Bolton, 
 
 8 
 
 
 
 
 
 
 
 16 
 
 26.3 
 
 29-5 
 
 34-5 
 
 33-8 
 
 Bourdon, 
 
 8 
 
 
 
 
 
 
 
 
 
 19 
 
 19 
 
 33 
 
 20 
 
A STUDY OF MEMORY. 9 
 
 Bolton's subjects were evidently as a rule brighter than those 
 of Bourdon. The latter's scores also show greater variability, 
 and indeed a very ragged sort of progress. This is doubtless 
 due to the fewness of his subjects. Bolton had 100 children of 
 each age. But it is by no means sure that the increase of ability 
 with age is due to growth in power to retain. As Professor 
 Thorndike indicates/ it is probably the result of a complicated 
 group of factors, including better attentiveness, greater power of 
 comprehending instructions, etc. The rate of improvement from 
 year to year cannot be determined with sufficient definiteness to 
 be of any significance. Different observers report nothing that 
 indicates clearly anything but variability in this rate. Bolton's 
 experiments with series of 6 digits show improvement of about 
 75 per cent, from 8 to 14 years of age. Bourdon's show 400 per 
 cent, improvement in the same time. Kirkpatrick, using series 
 of 10 monosyllabic words, finds an improvement of about 34 
 per cent, for girls and 60 per cent, for boys from the primary 
 grades to the high school. Lobsien, using series of 9 words of 
 1,2 or 3 syllables each, finds between the ages of 9 and 14 an 
 almost equal improvement for boys and nearly 80 per cent, for 
 girls. Kirkpatrick, Bourdon and Netschajeff unite in finding 
 the greatest improvement somewhere between the ages of 8 and 
 14, but Binet, using series of 5 to 7 monosyllables, finds almost 
 no improvement between 7 and 13. This is all the more strik- 
 ing, inasmuch as he used series of the same length as those of 
 Bourdon. With many of his series Bourdon reports the best 
 reproductions from children of 14, the older subjects proving no 
 better, if not actually worse. Netschajeff seems to corroborate 
 this notion. It is possible that these older pupils were lackadais- 
 ical in the experiment. Kirkpatrick finds improvement between 
 the high school and college, but the college students tested were 
 doubtless on the average of higher ability than the younger sub- 
 jects. In general I think we may say that the experiments 
 show the most rapid improvement in power of reproduction of 
 series before the age of 14. After that, if there is any advance, 
 it is slight. 
 
 As regards the relation between girls and boys, we may 
 
 1 ' Notes on Child Study,' p. 78. 
 
lO E. N. HENDERSON. 
 
 conclude that younger girls do better than boys of the same age. 
 Whether this advantage is overcome after adolescence, we cannot 
 tell. The testimony of Lobsien and Netschajeff shows no such 
 change in the superiority of the girls, while that of Kirkpatrick 
 indicates a reversal of the earlier relationship. 
 
 2. Passing to the question of the relative ease of reproducing 
 different classes of material, we find similar contradictions and 
 variations. Netschajeff finds number series the hardest of eight 
 types of series to reproduce. Lobsien finds them easier than any 
 other of the same types except those made up of objects seen.^ 
 Kirkpatrick finds series of shown words in general easier 
 to reproduce than series of spoken words. Hawkins reverses 
 this relation.^ Certain points of agreement come out promi- 
 nently, however. A series of objects seen is invariably reported 
 as remembered better than a series of words. In Kirkpatrick's 
 experiment series of objects were best retained ; words pre- 
 sented to the eye were next in the order of ease of reproduction. 
 Three days after his original experiment, he asked for a second 
 reproduction. The number of objects remembered was three 
 times as great as that of the words in the visually presented 
 series. It is not strange that we remember the objects that we 
 see about us better than what we read. Yet Kirkpatrick's ex- 
 periment must be granted to be striking, and not without its 
 lesson. The difficulty in remembering abstract words, that 
 both Netschajeff and Lobsien found, is to be expected, and 
 when one see Lobsien's lists, one wonders whether the children 
 were philosophers enough to understand many of them. 
 
 ^The nature of the series used by these experimenters is indicated by the 
 following list. The order is that of ease in reproduction. 
 
 Netschajeff. Lobsien. 
 
 1. Objects. I. Objects. 
 
 2. Words indicating things seen. 2. Numbers. 
 
 3. Sounds. 3. Words indicating things seen. 
 
 4. Words indicating sounds. 4. Words indicating tastes. 
 
 5. Words indicating tastes. 5. Sounds. 
 
 6. Abstract words. 6. Words indicating sounds. 
 
 7. Words indicating feelings. 7. Words indicating feelings. 
 
 8. Numbers. 8. Abstract words. 
 
 '^ Kirkpatrick is probably nearer the general fact here, as his results are cor- 
 roborated for adults (see p. 17). 
 
A STUDY OF MEMORY. II 
 
 It is also noticeable that younger children find series of more 
 than eight quite difficult to manage, indicating either a feebler 
 immediate memory, or, what is more likely, less power of in- 
 tense concentration of attention than they come to possess later. 
 Hawkins made the interesting observation that with younger 
 pupils a series of fifteen words could better be reproduced if the 
 words were shown successively, one every 2 seconds, than if 
 they were all given at once, and the subject allowed 30 seconds 
 to look them over. With older pupils this difference disap- 
 peared. One is tempted to think, however, that the explana- 
 tion offered by Hawkins, /. e., that the span of immediate mem- 
 ory grows greater, may not be the entire account of the matter. 
 Probably improved methods of committing to memory causes 
 the older students to waste less time in learning the series shown 
 simultaneously. 
 
 Netschajeff found that in general those who from indepen- 
 dent evidence were found to be of the visual type had better 
 memories than those of the motor or acoustic types. He thinks, 
 however, that the motor element entered into the efforts of most 
 subjects to commit to memory words shown, so that those with 
 good visual memories really enjoyed the advantage of motor 
 memories reinforced by those of sight. 
 
 Several studies have been made on the power of children to 
 reproduce sentences or stories.^ The only conclusions of impor- 
 tance refer to the growth with age of ability to recount the 
 matter dictated. Experiments made by Lay ^ indicate that with 
 college students a passage containing abstract material is much 
 more difficult to retain than a concrete one. 
 
 2. Results 0/ Laboratory Experimentation. 
 
 When we pass to the material offered by researches in the 
 experimental laboratories to the study of educational problems, 
 we find that the contributions may be classified under the fol- 
 lowing heads : (i) Conditions favorable or unfavorable to com- 
 mitting to memory ; (2) the nature of the memory image as 
 compared with the original experience. 
 
 1 Binet (see reference, p. 8) ; Shaw, ' A Test of Memory in School Children,' 
 Ped. Sent., 1S96-97, pp. 60-78. 
 
 2L,ay, 'Mental Imagery,' Psych. Rbv., Monograph Supplement, No. 17. 
 
12 E. N. HENDERSON. 
 
 I. The ordinary way of fixing a thing in mind is by 
 repetition. Practice creates habits. But laboratory researches 
 have established some interesting things about the effect of 
 repetition and the method thereof. To the teacher it might be 
 worth while to know these. Ebbinghaus/ probably the earliest 
 systematic experimenter in these fields, committed to memory 
 series of nonsense syllables by repeating them. The difficulty 
 of learning was estimated by the number of repetitions required. 
 The rate of forgetting was also tested. He found that after a 
 series was learned further repetitions fixed it better in memory. 
 There would seem, however, to be a limit to this effect. Bair, 
 who investigated carefully the results of practice,^ concludes 
 that after a certain time a physiological limit is reached, beyond 
 which further practice does not increase either speed or 
 accuracy — although it keeps the power at its climax. It is 
 interesting to note that the mere ability to go over a series accu- 
 rately does not mark the attainment of the highest excellence 
 of which memory is capable. Such a limit may be conceived 
 to be either the greatest speed in repetition or the longest reten- 
 tion in memory. Moreover, it is likely that practice after the 
 highest speed is attained will not be utterly useless in causing 
 longer retention of the series. Again, we ought to notice that 
 memorizing a series of nonsense syllables is very likely largely 
 a matter of motor associations, and so approximates closely 
 to learning a motor habit. May not the same thing be said of 
 nearly all the exercises that come under the head of learning 
 by heart? They are as mechanical as learning to walk. 
 Spelling and multiplication are gymnastic exercises. It is of 
 great importance to know when repetition has done its work 
 and drill may stop. Further, Bryan and Harter,' in their study 
 on learning telegraphy, found that after the limit of improve- 
 ment seemed to have been reached, and progress had for a 
 long time ceased, certain operators enter upon a second stage 
 of growth, and struggle toward a new ' physiological limit.' 
 Here we have an illustration of a common phenomenon in the 
 
 1 'Ueber das Gediichtniss,' Leipzig, 18S5. 
 
 2 'The Practice Curve,' Psych. Rkv., Monograph Supplement, No. 19. 
 
 'Psych. Rev., Vol. IV., p. 27, and Vol. VI., p. 345. 
 
A STUDY OF MEMORY. 1 3 
 
 class room. Certain pupils seem to stop for awhile, develop- 
 ment is ' arrested ' as it were. But the experienced teacher 
 knows that this is often not the end, and awaits patiently the 
 renascence of capacity for growth. How best to fill in such an 
 interval, or to initiate the new era? These are problems for the 
 educational psychologist. 
 
 The effectiveness of learning depends not alone on mere 
 number of repetitions, but also on the way in which these are 
 distributed. "^ It is a common opinion that little or no gain comes 
 from practice after fatigue sets in. ' Whatever can be said of 
 this,^ both Ebbinghaus and Jost^ agree that it is better to have 
 the repetitions scattered over considerable time, rather than con- 
 centrated within a short period. Jost found that a series of 
 twelve syllables was better learned if repeated ten times a day 
 for three days than if repeated thirty times in one day. Ebbing- 
 haus indicated that there are limits to the extent of scattering that 
 is desirable. Here we have the problem involved in the division 
 of time in the school program. Teachers know that they get 
 better results by having frequent short drills than by lengthen- 
 ing the periods and increasing the intervals between them. 
 They must be glad to find their general idea corroborated by 
 the psychologist, but they have yet to learn by experimental 
 evidence the distribution of time that will be most suitable in 
 various cases. 
 
 In regard to the length of the series which can best be 
 grappled with as a unit of work, the experimenters come a little 
 nearer giving practical information. With nonsense syllables a 
 single repetition was found by Ebbinghaus to suffice for the 
 learning of seven syllables. As the number of syllables was 
 increased, the number of repetitions required to learn the series 
 increased, at first very rapidly, then more slowly. These 
 longer series committed by a greater number of repetitions were, 
 I conjecture, retained better than the shorter series learned more 
 easily. Most school work where mechanical memory alone is 
 involved does not require the serial learning of more new units 
 than can be grappled in a single act of thought. Hence the 
 
 , iSee 'Mental Fatigue,' g. L. Thorndike, Psych. Rev., Nov., 1900. 
 ^Ztsch. f. Psych. ^ 1897, pp. 436-472. 
 
H E. N. HENDERSON. 
 
 repetitions are to fix in memory rather than to enable a bare 
 reproduction. Longer series enter in where the units are con- 
 nected because they belong to a common topic, as in the case of 
 sentences, of passages, etc. Miss Steffens ^ contributes some 
 information in regard to the length of the unit that can be most 
 effectively used in learning such material. She found that if, 
 instead of learning certain stanzas of ' Childe Harold ' verse by 
 verse, as her subjects were inclined to do, each stanza was read 
 repeatedly as a whole, fewer repetitions were necessary, and 
 the time of memorizing was reduced. She also found the same 
 to be true of longer series of nonsense syllables. Doubtless, 
 the reason why we prefer to take the shorter section is because 
 our progress is then more evident. On the other hand, most of 
 us have suddenly realized that we knew passages of consider- 
 able length when we have made no effort to commit them piece- 
 meal, but have read or heard them a number of times. We 
 may well suppose that we waste much time in learning by short 
 sections. But where a thing is to be learned by heart, doubtless 
 the sections might also be too long. The psychologists here 
 give us a valuable suggestion, but not a definite rule. 
 
 When we contrast the effect of repetition upon memory with 
 that of vividness, we are at once confronted with the difficulty 
 of generalizing in regard to the latter. The lack of any stand- 
 ard to which we can refer various kinds of vivid stimuli make 
 such experiments as those of Miss Calkins^ inadequate to the 
 formulation of any general law. By associating numerals in 
 pairs she ascertained that a number would recall one occurring 
 with it several times better than one paired with it only once but 
 printed in a striking color. Moreover, the number was less apt 
 to call up a recent associate with which it had been paired but 
 once, than a vivid or a frequent one. The teacher may con- 
 clude that undesirable but strong associations may be overcome 
 by constantly repeating others. This, however, is one of the 
 commonest methods in education, and Miss Calkins' experi- 
 ments do not indicate any definite way of determining degrees 
 of vividness in such a way that we can establish for any case 
 
 ^ Ztsch. f. Psych. ^ 1900, pp. 321-380. 
 
 * Psych. Rev. Monograph Supplement, No. 2. 
 
A STUDY OF MEMORY. 15 
 
 how much repetition will overcome a given vivid association. 
 It may be that such a law is unattainable. 
 
 The effect of rhythm on the ease with which we can commit 
 series of nonsense syllables was elaborately investigated by 
 Miiller and Schumann.^ That rhythm makes memorizing 
 easier is a commonplace truth, but the specific value of differ- 
 ent classes of rhythm for this purpose had never been investi- 
 gated. I do not see, however, that Miiller and Schumann have 
 given teachers any news that it is important for them to learn. 
 The study of the relation between the motor adjustment and the 
 memory of series is more significant educationally. Strieker's 
 well-known experiment, of trying to think the sound ' bubble ' 
 with the mouth open, brings out clearly the extent to which 
 motor elements enter into our acoustic memories. Cohn^ in- 
 vestigated the effect of pronunciation on memorizing series. 
 Certain series were shown to the learner, and also pronounced 
 by him. Others were shown, but all articulation by the learner 
 was inhibited. In still other cases the subject pronounced some 
 other sound or counted as the letters were being read. Those 
 who seemed to be of the visualizing type were not bothered 
 much by not articulating the letters nor by counting. On the 
 other hand, those who were of the acoustic-motor type were 
 seriously hampered by the lack of acoustic-motor associations 
 or by the presence of interfering ones. Smith, ^ whose experi- 
 ments antedate those of Cohn, found that with all his subjects 
 counting increased the number of errors in committing to mem- 
 ory a shown series. He also found that the characters of the 
 deaf and dumb alphabet could be learned easier if they were 
 made by the hands as they were being observed. 
 
 The importance of the motor element in committing to mem- 
 ory is emphasized especially by Bair,^ who experimented on the 
 effect of breaking the sensory-motor associations involved in a 
 habit, as compared with breaking the order of the sensory im- 
 pressions. His subjects learned to write with rapidity a series 
 
 ^ Ztsch.f. Psych., Vol. VI., pp. 81-190 and pp. 257-339. 
 
 "^Ztsch./. Psych., Vol. XV., pp. 161-183. 
 
 ^ Am. Journal of Psych., 1896, pp. 453-490. 
 
 * Psych. Rev., Monograph Supplement, No. 19. 
 
l6 E. N. HENDERSON. 
 
 of letters on a typewriter. Then the keys were capped so as 
 to change the arrangement of the letters thereon. This was 
 found to retard more the writing of a learned series than the 
 change of the order of the letters in the series without modify- 
 ing the lettering of the keys. It follows, Bair concludes, that 
 the association between a sense impression and the motor re- 
 sponse to it is in the case of such habits the hardest association 
 to break. This is probably due to the fact that a new associa- 
 tion of the sort is hard to form. I have already noted that all 
 learning by heart at school is nothing more than a gymnastic 
 drill. As James indicates,^ the association in such cases is from 
 sensory stimulus to motor response, the feeling of which consti- 
 tutes the stimulus for the next response. Frequently we find it 
 hard to anticipate the next sensory element without going 
 through the motions that precede it. In such cases the associa- 
 tion between sensory elements is comparatively weak. There 
 is, however, a sensory element here that such analysis neglects. 
 In spelling village for example, the thought of the word as a 
 whole, possibly also of the object, enters in to determine the 
 series of letters reproduced. The letters v-i-l-l-a might as 
 easily lead on to i-n as to g-e. What prevents this is the total 
 idea in the mind. But while the process of spelling is going 
 on the motor feelings constitute the cues by which the succes- 
 sion of letters is controlled. The general thought ' spelling 
 village ' plus the feeling that we have pronounced a suffices 
 to stimulate the pronunciation of g before we have thought of 
 it. The ordinary stock association is between thinking a and 
 then pronouncing it. In this case the feeling of having pro- 
 nounced a leads to the pronunciation of g. That this should 
 happen is due to the general thought in the mind. At this 
 point it may be worth while to suggest that the change in the 
 order of the letters in Bair's experiment may have destroyed not 
 so much the sensory associations between the letters as the sen- 
 sory-motor association between the feeling of having struck one 
 key and the motor impulse to strike the next. If this is the 
 case, he simply compared the strength of two sensory-motor 
 associations. At any rate, we may be fairly sure that sensory- 
 1 'Psych.,' Vol. I., p. ii6. 
 
A STUDY OF MEMORY. 1 7 
 
 motor associations play the most important part in all thor- 
 oughly learned series that we are taught in the schools, and we 
 are also comparatively sure that it would be harder for us to 
 pronounce the symbol a as we now 3o za than to learn to spell 
 sate/ after we had spent some time in putting these letters into 
 TV as. 
 
 In leaving the subject of the importance of the motor asso- 
 ciations in committing to memory, I wish to emphasize the 
 thought that in the learning of serie§ they are especially effect- 
 ive. If one tries to recall the episodes of a story, he will find, 
 I think, that the points come out much more clearly and effect- 
 ively if he iel/s the7n instead of merely imagining them. Where 
 thought works out into movement we find it easiCT to travel 
 straight on toward our goal, whereas unexpressed imaginations 
 do not lead on so decisively, but waver to and fro. There can 
 be no doubt that expression is indispensable to learning most 
 things in the school room, and that memories thus reinforced 
 are far more effective than those of experiences in which the 
 pupils have merely assumed passive attitudes. But while ex- 
 pression is probably from the educational point of view the most 
 valuable aid to memory, it is well to note that material furnished 
 to both eye and ear is better retained than that presented to 
 either alone. This fact was established by Miinsterberg and 
 Bigham.^ They also found that in general visually presented 
 series were better remembered than aurally presented ones, a 
 conclusion that was verified by Miss Calkins. 
 
 There remains to chronicle the experiments that have been 
 performed testing the effect on learning and remembering of 
 various sorts of distraction and the extent to which certain habits 
 interfere with others. W. G. Smith ^ performed experiments 
 similar to those of Cohn and T. L. Smith already reported, ex- 
 cept that his purpose was not so much to segregate the effect of 
 the muscular responses in memorizing, as to estimate the effects 
 of various sorts of distraction. His results can scarcely be 
 generalized to advantage, but it is interesting to note that he 
 found the learning of twelve syllables to be interfered with less 
 
 ' Psych. Rev., 1894, pp. 34-38. 
 
 ^Mind (New Series), Vol. IV., pp. 47-73. 
 
1 8 E. N. HENDERSON. 
 
 by carrying on addition at the same time than by repeating the 
 sound la to the beat of a metronome. The addition, though 
 complicated, was so well drilled that it required less thought 
 than the much simpler exercise. However, it must be noted 
 that to keep time required sensory attention, and thus interfered 
 with the sensory attention to the syllables. We can think about 
 one thing and do another more easily than we can think about 
 two things at once. This information will probably not surprise 
 the experienced teacher. 
 
 In regard to the effect on our powers of recall of distracting 
 conditions in the interval between memorizing and recall. Big- 
 ham^ has^ade some interesting contributions. He found that 
 the memWy was in general affected more by acoustical distur- 
 bances than by optical ones ; also that if what was to be retained 
 consisted of visual images, optical disturbances affected it more, 
 but acoustical disturbances were found to be especially destruc- 
 tive to auditory memories. Teachers generally have commented 
 on the disturbing effect of noises. Possibly a m(>re valuable 
 reflection that the study of distraction suggests concerns the ex- 
 tent to which different functions act independently. The con- 
 ception of the mind as consisting of a great number of com- 
 paratively independent aptitudes rather than of a few general 
 powers is a still further amplification of this thought, and ex- 
 periments on the effect of a habit upon the power to learn one 
 that is apparently antagonistic tend to justify the notion. Berg- 
 strom^ found that if a pack of cards were sorted with piles 
 placed in certain positions, this interfered with the natural ease 
 of sorting them into piles differently placed. Miinsterberg,^ 
 however, maintains that apparently contradictory habits do not 
 destroy each other, but that each can subsist beside the other, 
 and be called forth on occasion. At first it is hard to create the 
 new habit, but once formed it is easy to shift back and forth 
 from the one to the other. Bair corroborates this view of 
 Miinsterberg, both from experiments on the typewriter, and by 
 repeating Bergstrom's experiments with a view toward finding 
 
 1 Psych. Rev., 1894, pp. 453-461. 
 
 "^ Am. Journal of Psych., Vol. VI., p. 438. 
 
 3 'Gedachtnisstudien. Beitrage zur experimentellen Psychologic,' Heft 4. 
 
A STUDY OF MEMORY. 19 
 
 whether the initial interference due to the readjustment neces- 
 sary in forming a new habit would persist after both habits 
 might be said to be formed. He found that practice constantly- 
 reduced the interference of shifting. 
 
 2. On the subject of the quality of the memory image as 
 related to the original experience, what seems to me the most 
 suggestive contribution that laboratory psychology has to give 
 is that our memories represent, as it were, a sort of generalized 
 experience. They are modified in the direction of a standard 
 value. Leuba ^ suggests such a transformation. Philippe^ 
 noted the loss of particular factors in the memory images of 
 certain objects, as a statue of Venus, a bangle, a cigarette, etc. 
 Bentley,^ experimenting with colors, discovered that the 
 memory image was as a rule darker, but when the subject was 
 in the light it seemed lighter. Warren and Shaw* noticed that 
 in selecting by memory a square of a certain size from a 
 number of squares of different sizes a tendency to choose 
 from near the center of the list was observable. Xilliez^ ob- 
 served that in endeavoring to repeat a series of numbers 
 just heard there was a tendency to make the intervals be- 
 tween their values smaller. These are typical of that regress 
 toward a general washed-out image which plays so important a 
 part in discussions of the general idea. The nature of the 
 images we retain is, of course, a matter of first importance to 
 the teacher. Experimental psychology, however, has not con- 
 tributed anything more definite to increase the knowledge of 
 teachers regarding the fate of the ideas they so sedulously im- 
 plant in the minds of their pupils. 
 
 Summary of Results. 
 
 What then are the results of experimental researches on 
 memory that may be applied to education? We have learned 
 that children grow in power to reproduce series, and that this 
 growth is most rapid before fourteen. This increased power is 
 
 '^ Ant. Journal of Psych., Vol. V., pp. 370-384. 
 
 ^ Revue Philos., Vol. XVI., pp. 508-524 ; Vol. XVIIL, pp. 481-493. 
 
 ^ Am. Journal of Psych., 1899, pp. 1-4S. 
 
 * Psych. Rev., Vol. II., pp. 239-244. 
 
 ^ L' Annie Psych., Vol. II., pp. 193-200. 
 
20 E. N. HENDERSON. 
 
 due to greater ability to memorize, which in turn is due either 
 to increased power of concentration, to greater knowledge, or 
 to greater natural retentiveness. We have learned that boys 
 before adolescence, and probably for several years after, do 
 worse at such reproductions than girls of the same age ; and 
 that later on they probably catch up. We have been told that 
 children remember series of objects far better than series of 
 words, and that words indicating abstractions are very poorly 
 retained. We know that after a thing is learned by heart 
 further effort may strengthen our grip on it, though we do not 
 know any way to determine practically the point at which such 
 effort ceases to avail. We know that a certain effort expended 
 at intervcHs over considerable time is likely to produce better 
 results than if it were continuous, but we do not know the limits 
 of effective scattering, nor whether the principle applies to 
 anything save to learning by heart. We have found that 
 larger units can frequently be memorized as wholes with 
 greater economy of time than if we learned them in sections, 
 but we do not know the size of the unit most favorable, nor 
 have we any rule that can be applied to different classes of 
 material. Repetition we have found to fix associations better 
 than a certain kind of vividness. The principle is, however, 
 so indefinite as to have little practical value, save that it may 
 encourage us to repeated efforts to eliminate by repetition un- 
 desirable associations. We can be sure that rhythm helps to 
 memorize, that expression is of enormous assistance, that two 
 senses working together fix a series of ideas better than one, 
 that apparently contradictory habits can exist side by side, and 
 that disturbing sounds are especially destructive to memories 
 of what has been learned, although sight memories are particu- 
 larly affected by ocular distractions. Doubtless this is not all 
 that experimental research on memory has to yield to educa- 
 tion. But it represents about all. What can be said of it? 
 
 In the first place most teachers will say that they knew all 
 that before. So they did — in a general way. Nevertheless, 
 it may be agreed that the facts gain clearness and emphasis by 
 being shown by experimental methods, not to speak of the ad- 
 vantage of having opinions converted into certainties. Again, 
 
A STUDY OF MEMORY. 21 
 
 the criticism will be made that the facts thus derived are proved 
 for special cases, and that no general rule such as can be ap- 
 plied in teaching is to be drawn from them. This is doubtless 
 a serious criticism. Moreover, even if general principles could 
 be drawn from the experiments, they would have to be modified 
 to apply to special educational questions. The art of teaching, 
 it may be urged, is one of continuous adjustment to special sit- 
 uations. Not so much general psychological principles as spe- 
 cial applications of these, such as experience alone can make, is 
 what is desirable. Nor is this criticism to be dismissed as the 
 protest of a mere rule-of-thumb artisan. What we need is ex- 
 periment directed toward express educational problems. Such 
 work, doubtless, will derive general principles rather than spe- 
 cial rules, but they will be principles directly applicable to edu- 
 cation, because derived from a consideration of the school en- 
 vironment. They doubtless will not attain such accuracy as 
 that afforded by laboratory research. However, numbers and 
 statistical methods will compensate for this. On the other 
 hand, the teacher, fully conscious that each subject, each les- 
 son, each pupil, constitutes an individual problem, ought not to 
 forget the value of a broad generalization summing up experi- 
 ence in the large. Principles prevent experience from crystal- 
 lizing into poverty-stricken methods that appeal to their users, 
 not so much from their educational efficiency as because these 
 people are accustomed to them. Therefore, let us have prin- 
 ciples, not to supplement experience, but to enable us to 
 master it. * 
 
 Herein we discern the peculiar value that a study of what 
 experimental psychology has done may have to school men. It 
 has given us principles to guide further research. While it may 
 not contribute in the most direct and manifest manner to educa- 
 tional practice, its methods and its results furnish the basis on 
 which we must proceed. The experimentalist is at present en- 
 gaged in clearing our farm. Some small patches he has culti- 
 vated, and reaped therefrom a scanty harvest, but for the most 
 part his task has been to remove forests, underbrush and rocks, 
 and to make possible the productive tasks. We must know his 
 work in order not to duplicate both his failures and successes. 
 
22 E. N. HENDERSON. 
 
 when we experiment on educational problems. His work has 
 been as a rule minutely critical, but it has concerned very 
 simple and abstract matters. We need researches that will 
 reach into the concrete work of the school and test its results. 
 
 One striking contrast is noticeable between the material 
 used in most of the experimental work on memory that has 
 been reviewed and the kind of memory work done in the 
 school room. The former is largely concerned with power to 
 retain unconnected, often meaningless series, whereas the school 
 very rarely asks for work depending on this power alone. The 
 experiments, therefore, bear on the most mechanical tasks of 
 the pupil. But a more important kind of experiment would 
 concern itself with memory for connected thought. Even 
 spelling is a rational process. The letters make a whole by 
 the character of which they are to a great extent determined. 
 What interests the teacher is not so much the absolute strength 
 of the most wooden sort of memory, but the methods of associ- 
 ation by which the memory is stocked most richly and effect- 
 ively. To ascertain these, and also what becomes of the asso- 
 ciation groups with the lapse of time, may well furnish abundant 
 opportunity for research. 
 
 As a matter of fact the physiological and psychological proc- 
 esses involved in the remembering of a connected train of 
 thought are probably far different from what we have in the case 
 of a mere mechanical habit. The link of the common theme 
 greatly increases the number of units that can be retained. 
 Binet and V. Henri ^ found that by the use of a mn^onic de- 
 vice the number of digits that can be recalled on a single read- 
 ing may be raised from eight to thirty-six. They also noted ^ 
 a far smaller amount of forgetting in phrases than in series 
 containing the same number of disconnected words. I per- 
 formed a rough experiment as follows. I took from a popular 
 novel sentences of twenty words each; e. g.^ " A sprinkle of 
 blood, occasionally quite a dash of it, reddened the leaves and 
 tufts of grass along his pathway." From similar sentences I 
 selected phrases having four words each, and putting five of 
 
 ^ Revue Philos., No. 37, pp. 11 4-1 19. 
 ^L^ Annie Psych., No. I., pp. 1-59. 
 
A STUDY OF MEMORY. 23 
 
 these together made series of twenty words each ; e. g., " Into 
 their selfish dreams — like an eagle dancing — something im- 
 portant for him — forethought, zeal, and preparation — his fid- 
 dle and bow." Again, taking several of the sentences I ar- 
 ranged the words in as many series of twenty words each, the 
 combinations making no sense either as a whole or in phrases. 
 I read these various series at a uniform rate and in a monotone, 
 and then asked the listening students to write down the words 
 they remembered in as near the original order as possible. 
 Several readings of each series were given. One student in- 
 variably got the sentences perfectly in two readings, sometimes in 
 one. The series of phrases took him always two and sometimes 
 three readings, while the series of disconnected words took 
 from four to six readings, to learn. The difference was equally 
 manifest when we compare the reproductions after a single read- 
 ing. The words in the sentences were recalled with from no 
 omissions to at most five ; those in the series of phrases with 
 from five to ten omissions ; while there was an average of eight 
 omissions in the case of the disconnected words. A second 
 student required from two to four readings in order to get the sen- 
 tences, from three to five to get the phrases, and from six to 
 eleven to get the series of disconnected words. A third student 
 made on a single reading one slight error in reproducing the 
 sentence, omitted seven words from the phrase series, and eight 
 from the series of disconnected words. The latter series re- 
 quired of him three readings, while the other two were per- 
 fected on the second trial. 
 
 The results of this experiment are such as we might cer- 
 tainly expect, but the consideration of the difference in the 
 mental conditions involved reveals the difference between school 
 memorizing and that which we have reviewed. In recalling 
 the sentence the idea of the meaning is dominant. In the one 
 quoted the general picture of a bleeding man traversing a path- 
 way suggests the cue, ' a sprinkle of blood,' and each succeed- 
 ing thought, — /. e., 'quite a dash of it,' 'leaves and tufts of 
 grass,' etc., — helps to fill in the details of the same picture — 
 a picture that persists while the sentence is being spoken. One 
 subordinate thought does not replace another. All are in the 
 
24 E. N. HENDERSON. 
 
 mind at the same time. All that their serial order means is 
 that they are emphasized in succession. To keep the original 
 order of the thought and the special arrangement of words 
 involves the mere mechanism of association by serial contiguity. 
 But it is much more involved in remembering the series of 
 phrases, where the thought of one phrase does not fuse with 
 that of the others. It would, however, be an error to suppose 
 even here that there is no common mental state to which all 
 the separate thoughts contribute. We feel in a vague unan- 
 alyzed way that we must recall the five phrases of four words 
 each that we have just heard. This feeling guides and stimu- 
 lates our recall, much as the picture of the bloody trail leads to 
 repeating the sentence. The mental state is difficult to analyze. 
 It involves a recall of the total condition under which the phrases 
 were learned with sufficient accuracy to range them in order 
 and to exclude phrases belonging to other series. This vague 
 guiding consciousness becomes quite prominent when some time 
 later we endeavor to recall such a series. But it is doubtless 
 present and just as important in the guidance of an immediate 
 reproduction. It may be characterized as the attention absorb- 
 ing itself in the situation in which the series was learned. Such 
 an attitude exists as preliminary to the recall of the sentences, 
 but it is quickly replaced by a sense of the meaning they con- 
 vey — a far more potent ' open sesame ' to the details of the 
 series. 
 
 When we pass to the series of twenty disconnected words, 
 the attitude of attention is reinforced still less by inherent com- 
 bining forces. The main force of serial association by contig- 
 uity is required to do most of the work. Each of the twenty 
 meanings conveyed by the twenty words is gone with the utter- 
 ing of the word that expresses it. It gets little attention either 
 in learning or recall. It is quite otherwise with the meanings 
 of the words in the sentence. Each shares all the emphasis 
 that every other gets, because it is part of that other. Ebbing- 
 haus has given the average memory span for nonsense syllables 
 as seven. We may suppose that for disconnected words it is 
 from eight to ten. But with connected words it ranges with 
 average minds from twenty to forty, depending upon the diffi- 
 
A STUDY OF MEMORY. 25 
 
 culty of the form of expression and the degree of unity in the 
 thought. 
 
 School work is concerned largely with meanings. Words 
 sink into mere means of expression. The kind of experiment 
 on memory that most directly concerns the teacher, therefore, 
 will be that which deals with different sorts of associated 
 material. To find out how well comparatively different classes 
 of such material are retained, to investigate the efficiency of 
 different ways of learning it, to compare the way in which the 
 memories vary with the lapse of time, this comes nearer the 
 living problems with which education has to deal. We want 
 an experimental psychology of rhetoric and methods of teach- 
 ing, as well as of the relation of school training to practical life. 
 
 The following research deals with connect^ material. It 
 does not profess to have attained any results of strikingly prac- 
 tical value. It may claim, however, to have worked in material 
 such as constitutes an important part of school-room work, and 
 to have done something to indicate the possibility of far more 
 extensive and valuable researches regarding the products of 
 school training. 
 
26 E. N. HENDERSON. 
 
 PART II. THE SPECIAL RESEARCH. 
 
 The experiments, the results of which it is the special object 
 of this paper to discuss, consist in having certain passages of 
 connected matter learned in a given time, and reproduced after 
 definite intervals. A careful comparison of the successive 
 reproductions with each other and with the original is then 
 made, with the object of finding out both the amount and the 
 character of the material retained on each occasion. The 
 investigations bear on the relation between power of learning 
 readily and of retaining what is learned, the relative amount of 
 forgetting after different intervals, and the relation between 
 memory for ideas and memory for words. Moreover, since 
 classes of different ages were tested, the results indicate some- 
 thing about the effect of age and training on the power to learn 
 and to remember. In two cases adults were given the same 
 tests as children, thus furnishing sharply contrasted conditions 
 of age and training. Inasmuch as one class of students took 
 part in several tests, a comparison of their relative rank in each 
 gives some notion of the influence of the kind of material upon 
 the relative ability to learn and remember. Finally, a careful 
 study of what ideas are reproduced on the first occasion and 
 their fate as evidenced in the later reproductions sheds some 
 light on the qualitative modifications of our ideas regarding 
 certain things that occur with the lapse of time. The tests, it 
 will be seen, consist of material like that which students of dif- 
 ferent subjects might be expected to learn and to reproduce in 
 some fashion in the various exercises of the school room. They 
 have therefore a direct bearing on the significance of the results 
 of examinations, and on the fate of the ideas implanted by the 
 teacher in the minds of the pupils. 
 
 r"^ The Method. 
 
 / The members of the class to be tested were informed that 
 I they were about to be asked to commit to memory as much of 
 
A STUDY OF MEMORY. 27 
 
 a certain passage as they could in three minutes. They were 
 told that this was ample time in which to go over the passage a 
 number of times and to memorize a large part of it. The object 
 of this statement was to allay the excitement that the thought of 
 so short an interval might arouse, while insuring the putting 
 forth of all the energies in as concentrated an efifort as the indW 
 vidual could evoke. I do not think that many of the persons 
 who took part felt their efficiency materially interfered with by 
 the confusion of nervous excitement. As soon as these general 
 instructions were given, a typewritten copy of the passage to be 
 studied was handed to each member of the class, to be kept face 
 downward until a signal was given. They were told that they 
 should read the passage over at least twice, and then memorize 
 it in any way they wished. This instruction aimed at insuring 
 some attention to the ideas of the latter part of the passage. 
 The class was also informed that a written reproduction of the 
 passage would be called for, in which the words should be 
 given where remembered. In case they were forgotten, the 
 ideas that were retained were to be expressed as the subject 
 chose. At a given signal the papers were turned, and the work 
 began. It was stopped by having the subjects turn over the 
 papers at the end of the three minutes. 1 
 
 Two__days later I returned and asked for a second reproduc- 
 tion of the passage, giving instructions as before. In many 
 "cases I asked that the words which they were confident had oc-/ 
 curred in the original passage be underscored. Four weeks later 
 I returned again, and asked for a third reproductrdn, giving sim-;^- 
 ilar instructions. With the graduate students in Columbia Uni- 
 versity, who memorized on separate occasions three different 
 passages, it was necessary to announce in the beginning that 
 other reproductions would be expected. They found no diffi- 
 culty, however, in carrying out my request to banish the thought 
 of the passage from their minds during the interval. A similar 
 warning was given to the mature students in the summer session 
 of 1902 at Columbia University who took part in the experi- 
 ment. Thus a comparison of notes was avoided. To the other 
 classes nothing was said, and my second and third appearances 
 were surprises. I assumed that the later reproductions would 
 
28 E. N. HENDERSON. 
 
 be influenced far less by whatever discussion might appear 
 spontaneously than by the effect of suggestion in keeping the 
 passage before their thoughts, in case I warned them not to talk 
 about it, assigning the reason that I expected them to write it 
 out again. Moreover, with the younger classes a request that 
 they do not communicate with each other would doubtless have 
 stimulated rather than repressed such discussion. From a com- 
 parison of papers T am certain that there was very little gained 
 by intervening conversations. I took the testimony of the sub- 
 jects regarding the matter. Most declared that they had not 
 talked about the passage at all. A few had discussed it, but 
 an examination of their later papers showed little or nothing 
 new. Their talks, therefore, could have had no effect except 
 to strengthen their earlier impressions. 
 
 The Tests. 
 
 Five passages were used in the various experiments. They 
 were as follows. 
 
 Test I. — The King Who Became Just. 
 
 " There was-once, | in the eastern-part-of Egypt,-a king, || 
 whose-reign-had long been-a course-of savage-tyranny ; | long- 
 had he ruined-the rich | and distressed-the poor. || Suddenly-he 
 changed-his course ] and ruled-so well as-to be called-the just. 1| 
 When asked-by a favorite-the reason-for this-change, | he re- 
 plied: II 'I saw-a dog, | which, soon after-it had bitten-off-the 
 leg-of a fox, II was struck-on the head-by a great-stone, | that 
 cracked-its skull. | The stone-was thrown-by a man, || who-at 
 that instant-ran-in the way-of a horse | and was trod on-and 
 lamed-forever. || A short time-after | the horse-broke-its ankle- 
 bone-between two-stones. || These-sudden-misfortunes-con- 
 vinced me | that men-are used-as they use-others.'" 
 
 This passage was selected and adapted as one suitable for 
 younger classes. It presents interesting and intelligible mate- 
 rial, and contains a variety of elements. It is so condensed as 
 to furnish abundant matter for a three-minute exercise in mem- 
 orizing. The marks indicate the analysis on the basis of which 
 the scoring of the number of ideas retained was based. The 
 hyphens separate the detailed thoughts, the score in regard to 
 
A STUDY OF MEMORY. 29 
 
 which was usually taken as the record of the excellence of the 
 memory. The perpendicular lines separate certain subtopics, 
 the score in regard to which I have compared with the score in 
 detailed ideas. Finally, the parallel perpendiculars separate 
 the main topics into which the thought of the passage was 
 analyzed. 
 
 Test 2. — Cicero. 
 
 •' Cicero, | the greatest-of the Roman-orators, || was born-at 
 Arpinum, | an obscure-country-town. || His family-was of the 
 middle class-only, | and without wealth, || yet he rose-rapidly | 
 through the ranks-of Roman-official service | until at the age-of 
 forty-six | he became-consul. || In oratory-he is | by universal 
 consent | placed side by side-with Demosthenes, | or at least- 
 close after him. || He surpassed-the great- Attic-orator | in bril- 
 liancy-and variety, | but lacked-his moral-earnestness-and con- 
 sequent-impressiveness. || He could be-humorous, -sarcastic, | 
 pathetic, -ironical,-satirical, || and when he was-malignant | his 
 mouth was-most-foul | and his bite-most-venomous. || His de- 
 livery-was impassioned-and fiery, | his voice strong, -full, -and 
 sweet, I his figure-tall, -graceful, -and impressive." 
 
 It was thought that this passage would be suitable for his- 
 tory or Latin classes of the high-school grade, and it was used 
 with such subjects. It bore on work that frequently they were 
 doing, and it offered a chance to study the influence of this 
 work, besides appealing to the teachers as embodying material 
 worthy of reflection. 
 
 These two passages were the ones used with by far the 
 greater number of subjects. The other three : one, a con- 
 densed historical sketch ; the second, a detailed description, and 
 the last, an abstract discussion, were given to the same stu- 
 dents. They represent widely different fields of thought, and 
 may be taken as affording a good basis for comparing the 
 memory for different classes of material. The passages follow. 
 
 Test J. — The History of the Jews. (From Milman's 
 * History of the Jews.') 
 
 ♦' The Jews, -without reference to-their religious-belief, -| are 
 among-the most-remarkable-people-| in the annals-of man- 
 
30 E. N. HENDERSON. 
 
 kind. II Sprung-from one-stock, | they pass-the infancy-of 
 their nation | in a state of servitude-in a foreign-country, || 
 where, nevertheless, -they increase-so rapidly-as to become-on a 
 sudden | the fierce-and irresistible-conquerors | of their native- 
 valleys-in Palestine. || There they-settle down | under a form 
 of government-and code of laws | totally unlike-those of any 
 other I rude or-civilized-community. || They sustain-a long- 
 and doubtful-conflict, ] sometimes-enslaved, -sometimes-victori- 
 ous, I with the neighboring-tribes. || At length, -united-under 
 one-monarchy, | they gradu ally-rise-to the rank j of a powerful, - 
 opulent, -and commercial people. || Subsequently, -weakened 
 by-internal-discord, | they are overwhelmed-by the vast-mon- 
 archies I which arose-on the banks-of the Euphrates, | and are 
 transplanted-into a foreign-region. || They are partially-re- 
 stored I by the generosity-or policy-of the Eastern-sovereigns, | 
 to their native-land." 
 
 Test /J.. — The Dutch Homestead. (From Irving's 'Leg- 
 end of Sleepy Hollow.') 
 
 "It was-one-of those spacious-farm-houses, j with high- 
 ridged-but lowly-sloping-roofs, | built-in the style-handed down 
 from-the first-Dutch-settlers, || the low-projecting-eaves-forming 
 a piazza-along the front | capable-of being closed up-in bad 
 weather. || Under this-were hung-flails, -harness, | various- 
 utensils-of husbandry, ] and nets-for fishing-in the neighboring- 
 river. || Benches-were built-along the side-for summer use; 
 I and a great-spinning wheel-at one end, | and a churn-at the 
 other, I showed-the various uses-to which this important-porch- 
 might be devoted. |1 From this piazza-one might enter-the 
 hall, I which formed-the center-of the mansion | and the usual- 
 place of residence. || Here-rows-of resplendent-pewter | 
 ranged-on a long-dresser | dazzled-his eyes. || In one corner- 
 stood a huge-bag of wool, | ready-to be spun ; | in another-a 
 quantity-of linsy-woolsey, | just-from the loom; || ears-of Indian 
 corn I and strings-of dried-apples-and peaches ] hung-in gay- 
 festoons-along the walls, | mingled-with the gaud-of red-pep- 
 pers." 
 
 Test ^. — The Stages in the Development of Human 
 Theory. (From Martineau's ' Comte's Positive Philosophy.') 
 
A STUDY OF MEMORY. 3 1 
 
 " From the study-of the development-of human-intelligence, 
 in all-directions, -and through all-times, | the discovery-arises- 
 of a great-fundamental-law, || to which-it is necessarily-subject, 
 I and which-has a solid-foundation-of proof, | both-in the f acts-of 
 our-organization | and in our-historical-experience. || The law- 
 is this :- I that each-of our-leading-conceptions,- | each-branch-of 
 our-knowledge,- | passes-successively-through three-different- 
 theoretical-conditions : II the Theological, -or fictitious; | the 
 Metaphysical, -or abstract; | and the Scientific, -or positive. || 
 In other words, -the human-mind, -by its nature, | employs-in its 
 progress-three-methods-of philosophizing, | the character of 
 which-is essentially-different, | and even-radically-opposed : | 
 viz., -the theological-method, -the metaphysical,-and the posi- 
 tive. I Hence-arise-three-philosophies, | or general-systems-of 
 conceptions-on the aggregate-of phenomena, | each of which- 
 excludes-the others. || The first-is the necessary-point of de- 
 parture-of the human-understanding ; | and the third-is its fixed- 
 and definitive-state. | The second-is merely-a state of transi- 
 tion." 
 
 The Classes of Subjects Tested. 
 
 The following classes of students took part in the experi- 
 ments : 
 
 1. One hundred and three pupils in Public School No. 40, 
 New York City. These were all boys varying in age from 10 
 to 16. The school includes both primary and grammar depart- 
 ments, and all pupils are ranged from beginning to finishing 
 classes in 16 grades ; grade lA being the lowest and 8B the 
 highest. The pupils taking part in the experiment were in 
 three grades as follows : 33 in grade 7B, 43 in grade 6B, and 
 27 in grade 5B. 
 
 2. Fifty-three pupils in the Erasmus Hall High School, 
 Brooklyn. Of these 37 were girls and 16 boys. The school is 
 arranged in 8 grades, beginning with grade i, and the grades 
 are divided into sections indicated by the letters A, B, etc. 
 The pupils tested were in three grades : 23 in grade i, section 
 A, 18 in grade 4, section B, and 12 in grade 8, section B. 
 
 3. Twenty-one students in a course in psychology in Colum- 
 
32 E. N. HENDERSON. 
 
 bia University. Most of them were seniors in Columbia Col- 
 lege. Four, however, were women from Barnard College. 
 
 4. Eighteen graduate students in Columbia University. 
 
 5. Seventeen students in the summer session of 1902, at Co- 
 lumbia University. Of these 4 were women. 
 
 The following abbreviations will be used to indicate the 
 different classes of subjects: P. S. 7B, P. S. 6B, and P. S. 5B 
 will signify the different grades in Public School No. 40. E. 
 H. S. 8B, E. H. S. 4B, E. H. S. lA, will indicate the various 
 classes in the Erasmus Hall High School. S. S. students will 
 be used for students in the summer session at Columbia Univer- 
 sity. The expressions college students and graduate students 
 will serve to denominate the other classes tested. 
 
 It will be seen that the 212 subjects range from the young 
 pupil, whose ability to read and write is barely sufficient to make 
 such an experiment possible, to individuals well on toward 
 middle life, and representing a high grade of scholastic attain- 
 ment. Inasmuch as most of the graduate students took part in 
 three experiments and the college students in two, I had in all 
 259 papers in which the various reproductions were complete. 
 
 This number may not seem sufficiently large for a statistical 
 investigation, but it must be remembered that the length and 
 character of the tests made them fairly reliable measures of the 
 powers of the individuals. It was not so necessary, therefore, 
 as it would have been in a shorter test, to swallow up errors re- 
 garding individual measurements by the mere mass of testimony. 
 On the other hand, only in so far as the individuals tested are 
 typical can the results be said to be universally valid. To the 
 consideration of these results let us now proceed. 
 
 The Results. 
 These I think it convenient to group in two divisions: A^ 
 the amounts of loss in the various reproductions and with the 
 various classes of subjects and tests ; B^ the character of this 
 loss in the different cases. 
 
 J. A. The Amounts 0/ Loss. 
 
 In order to deal with this matter a method of scoring was 
 necessary. This required analysis of the thought of the pas- 
 
^^: 
 
 A STUDY OF MEMORY. 33 
 
 sages — an analysis that is indicated in the passages as printed. 
 The following table gives the standard score of each passage. 
 
 Topics. Subtopics. Details. Words. 
 
 The King Who Became Just. 9 20 69 138 
 
 Cicero. 9 25 64 125 
 
 The History of the Jews. 8 26 80 163 
 
 The Dutch Homestead. 8 26 88 180 
 The Stages in the Development of 
 
 Human Theory. 7 25 93 171 
 
 It must be confessed that this analysis has in it something 
 arbitrary. To say that each of the detailed thoughts thus indi- 
 cated is equal in value to every other is manifestly absurd. 
 And this is true whether our estimate be based on relative impor- 
 tance to the thought in general or on relative difficulty of recall. 
 But it must be granted that the same objection could be raised 
 against any endeavor to compare two mental conditions quanti- 
 tatively. However, as the mind of the subject traveled over 
 the thought it was trying to reproduce, it may be conceived to 
 have rested momentarily on each of the details indicated. In | 
 general, the better memories could be expected to retain not 
 only the easily remembered details, but also the ones harder to 
 recall, whereas the poorer ones would retain only the former 
 class. In such cases the scores given can not be challenged on 
 the ground that the lack of equality between the units renders 
 the ranking of the subjects arbitrary. Placing different values 
 on the ideas or analyzing the units differently might affect the 
 ranking in cases where the loss of certain ideas is pitted against 
 that of different ones, but seldom, I am confident, could one 
 justify a valuation or an analysis so different from mine as to 
 affect materially the ranking of the student. Hence, the general 
 results of my investigation are, I conceive, not dependent on 
 the peculiarities of my scoring. 
 
 The scores given have not been diminished because of errors. 
 They are records only of what was retained. I have taken the 
 ground that the erroneous idea that contains the suggestion of 
 the true one deserves a positive rather than a negative score. It 
 indicates a thought corresponding, however inaccurately, to the 
 earlier one. Such ideas are given a part of the value of an 
 accurate memory. Some individuals, it is true, leave unex- 
 
34 E. N. HENDERSON. 
 
 pressed the hazy idea that they fear is erroneous. They might 
 suffer by comparison with cloudier minds that failed to discover 
 the presence of the fog. However, a mind that feels a certain 
 idea to be inaccurate is usually able to express the part or phase 
 of the thought that is accurate, and thus render a true account 
 of what was in the memory. 
 
 In comparing the abilities of different classes of subjects, I 
 have used for the most part the scores made in detailed thoughts 
 and in words. This was regarded as the safest basis for such 
 comparison. If the equal value of the elemental ideas is ques- 
 tioned, much more might that of the subtopics and topics into 
 which the thought of the passage was analyzed. A considera- 
 tion of the scores in larger thoughts as compared with those in 
 details is reserved for the discussion on the qualitative differ- 
 ences between the different reproductions. 
 
 The scoring of words remembered might easily become a 
 complicated matter. Doubtless, the reproducing of certain 
 words means far more power of memory than that of others. I 
 have used the following system. All words of the original that 
 were reproduced in their former contexts were scored full value. 
 Commonplace words, particularly articles, prepositions, and 
 conjunctions, were not scored when reproduced out of their 
 context. On the other hand, an unusual word was regarded as 
 remembered, even though it appeared in the wrong context. 
 ■Occasionally a word was evidently used because its sound was 
 ;somewhat like that of one in the original. A half credit was 
 here given. Words that were modified to suit changes in con- 
 struction, etc., were given partial credit also. Let us now pro- 
 ceed to the comparison of the various records.^ 
 
 Average Scores of the Different Classes. 
 
 Test. 
 
 Class. 
 
 ist Rep. 
 
 Av.Dev. 
 
 Id( 
 2d Rep. 
 
 ;as. 
 
 Av 
 
 '. Dev. 
 
 3d Rep. 
 
 Av. Dev. 
 
 Test I. 
 
 S. S. students, 
 
 53-0 
 
 6.6 
 
 48.3 
 
 
 5.6 
 
 45.8 
 
 6.5 
 
 
 P. S. 7B pupils, 
 
 49-5 
 
 6.9 
 
 46.5 
 
 
 4.8 
 
 41.9 
 
 8.2 
 
 
 P. S. 6B " 
 
 45-9 
 
 6.0 
 
 44.6 
 
 
 7-3 
 
 41-3 
 
 6.2 
 
 
 P. S. 5B " 
 
 38.3 
 
 8.3 
 
 36.3 
 
 
 8.8 
 
 32,0 
 
 7-5 
 
 * I introduce in the appendix, for the benefit of those who care to examine 
 the numerical data on the basis of which all these computations depend, a list 
 of the individual scores in the various tests. 
 
A STUDY OF MEMORY. 35 
 
 
 
 
 
 Ideas. 
 
 
 
 
 Test. 
 
 Class. 
 
 ist Rep. 
 
 Av.Dev. 
 
 2d Rep. A 
 
 V. Dev. 
 
 3d Rep. 
 
 Av. Dev 
 
 Test 2. 
 
 E. H. S. SB students, 
 
 42.6 
 
 5-1 
 
 38.0 
 
 6.6 
 
 33-5 
 
 8.8 
 
 
 E. H. S. 4B 
 
 38.0 
 
 7-5 
 
 34-9 
 
 9-3 
 
 24.9 
 
 10.6 
 
 
 E. H. S. lA 
 
 36.8 
 
 7-9 
 
 31.0 
 
 8.1 
 
 22.8 
 
 8.4 
 
 
 College students, 
 
 48.0 
 
 8.2 
 
 36.5 
 
 8.2 
 
 34-5 
 
 10.3 
 
 Test 3. 
 
 College " 
 
 37-0 
 
 4.4 
 
 27.0 
 
 5-1 
 
 24.1 
 
 7.2 
 
 
 Graduate " 
 
 42.2 
 
 10.2 
 
 34.6 
 
 10.3 
 
 29-5 
 
 10. 1 
 
 Test 4. 
 
 Graduate " 
 
 57-5 
 
 8.8 
 
 46.9 
 
 8.1 
 
 40.6 
 
 8.0 
 
 Test 5. 
 
 Graduate " 
 
 45-9 
 
 13.2 
 
 30-4 
 
 8.9 
 
 21.8 
 
 8.9 
 
 Test 3. 
 
 12 selected graduates,' 
 
 44-5 
 
 9.9 
 
 37-5 
 
 9.2 
 
 32.9 
 
 8.0 
 
 Test 4. 
 
 12 
 
 60.6 
 
 7.0 
 
 49.0 
 
 6.4 
 
 41.8 
 
 7-7 
 
 Test 5. 
 
 12 
 
 45-6 
 
 II. 7 
 
 30.1 
 
 8.6 
 
 22.4 
 
 9-5 
 
 Average Scores of the Different Ci<asses. 
 
 Words. 
 
 Test. 
 
 Class. 
 
 
 tst Rep. 
 
 Av. Dev. 
 
 2d Rep. 
 
 Av. Dev. 
 
 3d Rep. 
 
 Av. Dev, 
 
 Test I. 
 
 S. S. students, 
 
 
 77-0 
 
 13-4 
 
 59-0 
 
 13-3 
 
 52.1 
 
 13.2 
 
 
 P. S. 7B pupils, 
 
 
 70.6 
 
 14. 1 
 
 61. 1 
 
 12.2 
 
 5I.I 
 
 16.6 
 
 
 P. S. 6B " 
 
 
 639 
 
 12.3 
 
 59-7 
 
 13-9 
 
 52.1 
 
 14.I 
 
 
 P. S. 5B '« 
 
 
 53-4 
 
 13-7 
 
 47-4 
 
 II. 2 
 
 40.4 
 
 II. 2 
 
 Test 2. 
 
 E. H. S. 8B students, 
 
 76.5 
 
 9.1 
 
 64.6 
 
 11.4 
 
 55-7 
 
 17.0 
 
 
 E. H. S. 4B 
 
 
 66.3 
 
 16.4 
 
 57-9 
 
 18.7 
 
 44-1 
 
 18.8 
 
 
 E. H. S. lA 
 
 
 61.2 
 
 16.5 
 
 47.8 
 
 14.3 
 
 32.4 
 
 II-3 
 
 
 College 
 
 
 82.0 
 
 195 
 
 49-7 
 
 18.0 
 
 44.2 
 
 20.7 
 
 Test 3. 
 
 College 
 
 
 43-4 
 
 8.2 
 
 26.6 
 
 8.6 
 
 22.3 
 
 9.6 
 
 
 Graduate 
 
 
 56.5 
 
 16.5 
 
 37-8 
 
 15-5 
 
 30.5 
 
 14.0 
 
 Test 4. 
 
 Graduate 
 
 
 90.9 
 
 15.1 
 
 66.9 
 
 13-9 
 
 51-9 
 
 9-4 
 
 Test 5. 
 
 Graduate 
 
 
 63.7 
 
 21.3 
 
 390 
 
 12.8 
 
 24-5 
 
 II.O 
 
 Test 3. 
 
 12 selected graduates,^ 
 
 ■ 60.7 
 
 19.2 
 
 41.3 
 
 17.1 
 
 34-9 
 
 13-4 
 
 Test 4. 
 
 12 " 
 
 (1 
 
 96.0 
 
 12.9 
 
 71.0 
 
 12.2 
 
 53-6 
 
 10.5 
 
 Test 5. 
 
 12 
 
 <i 
 
 63.2 
 
 20.9 
 
 39-2 
 
 13.2 
 
 24.4 
 
 II. 4 
 
 Average Percentages of Loss in the Various Reproductions.^ 
 
 Ideas. Words. 
 
 Test. 
 
 Class. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 Test I. 
 
 S. S. students, 
 
 23'. I 
 
 8.8 
 
 13-6 
 
 44.2 
 
 23-3 
 
 32.3 
 
 
 P. S. 7B pupils, 
 
 28.1 
 
 6.0 
 
 15-3 
 
 48.8 
 
 13-4 
 
 27.7 
 
 
 P. S. 6B " 
 
 33-4 
 
 2.8 
 
 7.8 
 
 53-7 
 
 6.5 
 
 18.4 
 
 
 P. S. 5B " 
 
 44-5 
 
 5-2 
 
 16.4 
 
 61.3 
 
 II. 2 
 
 24-3 
 
 ' For purposes of comparison the average scores of 12 students who gave all 
 the reproductions of tests 3, 4 and 5 are given. 
 
 ^ These percentages of loss are based in the case of the first reproduction 
 on the standard score, and in the second and third reproductions on the record 
 in the first reproduction. The object is to ascertain what proportion of the 
 original passage was not learned and what proportion of what was learned was 
 not retained in the second and third reproductions. The significance of these 
 data will be discussed later. 
 
36 E. N. HENDERSON. 
 
 Ideas. Words. 
 
 Test. Class. ist Rep. 2d Rep. 3d Rep. 1st Rep. 2d Rep. 3d Rep. 
 
 Test 2. E. H. S. 8B students, 33.4 10.8 19.0 38.8 15.5 27.3 
 
 40.6 8.1 34.4 46.9 12.6 33.4 
 
 42.5 15-7 38.0 51.0 21.9 47.0 
 25.0 23.9 28.1 34.4 39.3 46.1 
 
 53.7 27.0 34.8 72.1 38.7 48.6 
 47.2 18.0 30.0 65.4 33 I 46.0 
 
 34.6 18.4 28.6 49.5 26.4 42.9 
 50.5 33-7 52-4 62.7 38.7 61.5 
 
 E. H. S. 4B 
 E.H. S. lA 
 College 
 Test 3, College 
 
 Graduate 
 Test 4. Graduate 
 Test 5. Graduate 
 
 Test 3. 12 graduates, 44.3 15.7 26.0 62.7 31.9 42.5 
 
 Test 4. 12 graduates, 31. i 19.1 31.0 46.1 26.0 44.1 
 
 Test 5. 12 graduates, 50.9 34.0 50.8 63.0 37.9 61.3 
 
 An examination of the scores in test i shows that the total 
 record of each of the different classes of students grows in 
 every reproduction better as we go from the lower to the more 
 advanced classes. This is invariably true of the ideas. In the 
 case of the words the S. S. students fall below the P. S. 6B 
 pupils in the second reproduction ; and the P. S. 6B pupils did 
 as well as they in the third reproduction. The older students 
 evidently attended less to words, or at least forgot them more as 
 compared with ideas, than did the younger ones. In test 2 the 
 more advanced classes generally score a better record in ideas 
 remembered, the exception being that the college students fall 
 below E. H. S. 8B students in the second reproduction. In 
 their word record the college students went down even below 
 the E. H. S. 4B students in the second reproduction, and below 
 the E. H. S. 8B students in the third. 
 
 It is evident that so far as positive scores are concerned, the 
 advanced students do better as a rule than those in lower 
 classes. At the same time the size of the average deviations 
 in the table of totals shows how unsafe would be a prediction 
 that an individual in a higher class will do better than one in 
 the class below. A comparison of the percentages of loss in 
 tests I and 2 shows that between the first and third reproduc- 
 tions the rate of loss is highest with the lowest class, but does 
 not steadily decrease as we approach the higher ones. P. S. 
 6B pupils suffer the least percentage of loss in test i, and E. 
 H. S. 8B students remember best in test 2. There is evidently 
 a slight tendency for the more advanced classes to remember 
 better. The heavy percentage of loss of college students in the 
 
A STUDY OF MEMORY. 37 
 
 second reproduction of test 2 is probably due to the fact that the 
 test was followed by another of the same sort and also by fur- 
 ther experiments. This caused it to produce a less profound 
 impression upon them than upon the high-school students, to 
 whom it was a decided novelty. 
 
 We can be fairly sure that the more advanced students learn 
 more quickly than those in lower grades. Is this because their 
 memories are better or because their general knowledge and 
 training makes them more readily grasp the contents of the 
 passages, and more effectively address themselves to the task 
 of committing these to memory? If the record of percentages 
 of loss may be taken as the best criterion of excellence of mem- 
 ory, we may be sure that the better average scores of the older 
 classes are not due wholly to this power. It seems to me that 
 this assumption is a safe one. Relative rate of loss will inevi- 
 tably determine in the long run the relative standing of two in- 
 dividuals in regard to their knowledge of certain common ex- 
 periences, no matter what portion of these experiences may 
 have originally effected lodgment in the minds of each. Two 
 different scores each diminishing regularly by certain percentages 
 will approach each other continually if the percentages be the 
 same. If they be different, the higher rate of loss will inevit- 
 ably bring either of the scores to be the lower one, and the 
 memory suffering such loss may certainly be regarded as the 
 weaker one. 
 
 It may be objected that the power to retain for a consider- 
 able time may be to a great extent independent of the power to 
 remember long enough to reproduce. This point I shall deal 
 with later. If, however, they are, as one might expect, the 
 same at bottom, then it seems likely from the evidence of per- 
 centages of loss, that the more advanced students learned more 
 quickly because of knowledge and skill rather than because 
 their memories were better. 
 
 In some cases the slight superiority of the older students in 
 power both to learn and to remember is due to the fact that the 
 more advanced classes are made up of students with higher na- 
 tive ability than the lower ones. This result must of course 
 follow from elimination of the weaker minds from the higher 
 
3^ E. N. HENDERSON. 
 
 grades of work. In P. S. No. 40, however, the influence of 
 such selection will doubtless be small. It will therefore be of 
 some interest to examine the data from that school, with a view 
 toward determining just how much growth there is from year to 
 year in the powers in question. 
 
 Effect of Age and Training on Power to Learn and to 
 
 Remember. 
 
 In estimating the average score of the children of the same 
 age who were tested in P. S. No. 40, I made two calculations. 
 In one all the scores are used, in the other the scores of those 
 who retained less than 50 per cent, of the original matter in the 
 first reproduction were left out. I regard these records as fre- 
 quently more a result of difficulties in reading and understand- 
 ing the matter of the test than of lack of a more general ability 
 to learn or to remember. Occasionally an individual who does 
 very poorly at first will retain all he learns. If the power to 
 read and understand were at fault, then the inclusion of such 
 subjects with those who have no such difficulties confuses the 
 factors the influence of which we are trying to determine. 
 
 Average Scores According to Age. Test i. 
 
 S. S. students. i6. 15. 14. 
 
 Age. Score. Av. Dev. Score. Av. Dev. Score. Av. Dev. Score. Av. Dev. 
 
 ist rep. 53.0 6 59+ 3+ 50+ 6 46+ 6 + 
 
 2d rep. 48.3 5+58 5 47+8 44 8 
 
 3d rep. 45-8 6+ 54+ 8+ 43+ 8+ 39 9 
 
 Average Scores According to Age. Test i {Continued'). 
 13. 12. II. 10. 
 
 Age. Score. Av. Dev. Score. Av. Dev. Score. Av. Dev. Score. Av. Dev. 
 
 istrep. 44+7 41 9+ 45+ 6 38+10 
 
 2d rep. 42 + 8 40 + 9 + 43 + 8 + 37 + 10 
 
 3d rep. 37 + 9 + 37 + 9 4i 7 + 34 10 + 
 
 Average Scores According to Age. Test i. Omitting Those Bei.ow 
 50 Per Cent. First Reproduction (15 in Number). 
 
 S. S. students. i6. 15. 14. 
 
 Age. Score. Av. Dev. Score. Av. Dev. Score. Av. Dev. Score. Av. Dev. 
 
 1st rep. 53.0 6 59+ 3+ 53+ 3+ 48 5 
 
 2d rep. 48.3 5+ 58 5 5o+ 4+ 45+ 6+ 
 
 3d rep. 45.8 6+ 54+ 8+ 46+ 6 40+ 8 
 
A STUDY OF MEMORY. 39 
 
 Average Scores According to Age. Test i. Omitting Those Below 
 50 Per Cent. First reproduction (15 in Number). {Continued.) 
 
 13- 
 
 
 12. 
 
 
 II. 
 
 10. 
 
 Age. Score. Av. Dev. 
 
 Score. 
 
 Av. Dev. 
 
 
 Score. Av. Dev. 
 
 Score. Av. Dev, 
 
 1st rep. 47 5+ 
 
 48 
 
 6 
 
 
 47 6 
 
 44+ 6 
 
 2d rep. 45 6+ 
 
 45 
 
 6 
 
 
 45 6 
 
 44+ 6 
 
 3d rep. 40 8+ 
 
 42 
 
 6+ 
 
 
 42 6+ 
 
 40 9 
 
 The ages of the pupils in 
 
 I each g 
 
 rade is indicated 
 
 as follows : 
 
 Age. 16. 
 
 15- 
 
 14. 
 
 13- 
 
 12. II. 
 
 10. 
 
 P. S. 7B. 2 
 
 7 
 
 10 
 
 9 
 
 4 
 
 
 P. S. 6B. 
 
 3 
 
 II 
 
 9 
 
 12 7 
 
 
 P. S. 5B. 
 
 
 2 
 
 8 
 
 7 3 
 
 6 
 
 Total. 2 10 23 26 23 10 6 
 
 At first sight it seems that the older students show in general 
 slight superiority over the younger ones so far as scores are 
 concerned. In the later and more reliable table, however, no 
 growth appears between 11 and 14 years of age and the slight 
 advance from 14 to 15 may easily be due to chance. In gen- 
 eral, the fewness of the cases renders the evidence decidedly 
 inconclusive, but a comparison of the two tables suggests that 
 the elimination of those who have difficulty in comprehending 
 the passage has destroyed the larger part of the superiority of 
 the older students. In the cases of S. S. students, the better 
 record is to be accounted for because they were doubtless a 
 selected class as compared with the children. Indeed, one is 
 surprised that they do not show more superiority. It will be 
 seen that they are surpassed by the children 15 and 16 years of 
 age. When we compare percentages of loss according to age, 
 we find even less evidence of any grow^th from younger to older 
 children. 
 
 Average Percentages of Loss According to Age, Third Reproduc- 
 tion, Test i. 
 
 Age. S. S. students, 16 15 14 13 12 11 10 
 
 Per cent, of loss. 14 8 13 15 14 12 10 10 
 
 It would seem that the younger pupils remember quite as 
 well as the older ones, or even better. It seems to me that we 
 are justified in suspecting that whatever inferiority they show in 
 power to learn is due to their lack of ability to read and under- 
 stand readily. An examination of the papers of younger pupils 
 
40 E. N. HENDERSON. 
 
 reveals strange misstatements, that are reproduced literally in 
 the later papers. In a comparatively poor paper by a boy lo 
 years old in P. S. 5B, we find the passage, 'whose reign had 
 long been a course of savage tyranny. Long had he ruined the 
 rich and distressed the poor,' rendered as follows: ist repro., 
 'who had reign over the rich and discresed (?) the poor in 
 dammand (?)' 2nd repro., 'which ruled over the rich in dem- 
 mand (?) of the poor' ; 3d repro., ' who ruled over the rich and 
 demanded the poor.' Certainly such mistakes cannot be attrib- 
 uted to bad memory. They are evidence of an unfamiliarity 
 with the words in the text, such as is common in this grade, and 
 has disappeared almost entirely in P. S. 7B students. 
 
 Notwithstanding the fewness of my subjects, it seems to me 
 that my results warrant the suspicion that the growth with age 
 of power to reproduce immediately series of numbers, letters, 
 words, etc., established by Bolton, Bourdon, Netschajeff and 
 others, may be due not to growth in power to remember ; it is 
 very likely the result of greater ability to understand instruc- 
 tions and the meaning of words, to more familiarity with num- 
 bers, and other similar causes. Results ranged according to 
 grades would depend on training rather than age, and the fact 
 that the children of a certain age were in my tests distributed 
 in several grades tended to eliminate the factors of ability and 
 training, and to emphasize such differences, or lack of differ- 
 ences, as were due to age alone. To separate absolutely the 
 effects of age from those of training is, of course, very difficult, 
 if not impossible, and I can claim only to have indicated the 
 probability that training is the more important of the two. 
 
 Correlation of Scores According to Ability. 
 
 Let us now take up the question of whether power to learn 
 readily correlates with power to remember what is learned. 
 Two investigations may be made, each contributing to the 
 definiteness of our knowledge in the matter. The first will 
 consist in comparing the relative score of each subject in the 
 first reproduction of a certain test with that in the later repro- 
 ductions of the same passage. The second will compare the 
 
A STUDY OF MEMORY. 41 
 
 relative percentages of loss of the different subjects in the vari- 
 ous reproductions. 
 
 The method of correlation that I have used is that employed 
 by Karl Pearson/ and in such calculations generally. The 
 percentage of correlation between the scores in two functions of 
 a number of individuals is symbolized by r. The formula for 
 
 its calculation is 
 
 Ixy 
 r = 
 
 no^a^ 
 
 Here o^ represents the standard deviation of the individuals of 
 a group from the average in one function, o^ the standard devia- 
 tion from the average in the other function, x and y represent 
 the deviations of any individual from the average scores in the 
 two functions compared. The probable error in r is found by 
 the formula 
 
 This is, of course, of importance in determining the signifi- 
 cance of estimates based on a comparatively small number of 
 cases. 
 
 Correlations of the Scores in Ideas in the Various 
 
 Reproductions . 
 
 The two most important estimates that I have made are 
 based on a comparison of all the scores of the various classes of 
 individuals who took part in test i (' The King Who Became 
 Just ') and test 2 (' Cicero '). Such a method of combining data dis- 
 regards wholly the effect of age and training, or of sex, wherever 
 the subjects are of both sexes. It ranks all individuals accord- 
 ing to their abilities in learning the passage, and compares this 
 rank with that which they attain in the later reproductions. 
 We do not compare directly power to learn and power to retain, 
 but we find in how far a person will tend to preserve his original 
 rank within the limits of time over which the tests extended. 
 Such a calculation certainly is of some significance. 
 
 In test I there were 120 who took part. Comparing their 
 
 1 ' Grammar of Science. ' 
 
42 E. N. HENDERSON. 
 
 ranking in the first reproduction with that in the second we find 
 that: 8 kept the same rank, 70 lost in rank with an average 
 loss of 6.8 grades, 42 gained in rank with an average gain of 
 4 grades. The average change in rank of all was 5.4 grades. 
 
 Comparing their ranking in the first with that in the third 
 reproduction : 7 kept the same rank, 68 lost in rank with an 
 average loss of 8 grades, 45 gained in rank with an average 
 gain of 6.5 grades. The average change in rank of all was 
 6.5 grades. 
 
 There were 62 grades in the first reproduction, 66 in the 
 second, and 63 in the third. When we reflect that if all the 
 individuals ranged in the 62 grades of the first reproduction had 
 receded to the same level in the second, there would have been 
 an average change of 31 grades, we see that 5.4 grades repre- 
 sents comparatively little shifting. Neither does the average 
 change of 6.5 in the third reproduction indicate anything start- 
 ling in the way of readjustment in ranking. It must be borne 
 in mind that the difference between two successive grades is 
 usuall}^ a half unit out of 69 possible units. Hence the average 
 shifting of from 5.4 to 6.5 grades means very little difference 
 between the scores attained by most individuals and the scores 
 that would have preserved for them their former rank. 
 
 Determining more accurately by the correlation formula the 
 amount of regression toward equality in the later reproductions, 
 we find : 
 
 Between scores in ideas, first and second reproductions, r= 96%, P.E. ^ .3%, 
 
 " " third " r=88%, P.E. = 1%. 
 
 In this calculation, the comparison took account of relative 
 differences between the scores of the various ranks, and the 
 amount of correlation is seen to be even greater than might be 
 supposed from the amount of change in rank. 
 
 Considering the scores of the 74 individuals who took part 
 in the ' Cicero ' test, we find : 
 
 Between scores in ideas, first and second reproductions, r = 87%, P.E. = 1.4%, 
 " " " " " " third " r= 75%, P.E. = 2.6%, 
 
 It is evident that the lapse of two days produces very little 
 change in the relative amounts of the given passage in the minds 
 
A STUDY OF MEMORY. 43 
 
 of the various individuals taking the tests. The lapse of a month 
 doubles the discrepancy between the percentage of perfect corre- 
 lation and the degree of correlation that actually exists. Does 
 this mean that the subjects who learned more quickly will even- 
 tually recede in their knowledge of the passage to the rank 
 of those who learned little? By no means. It is true that, if 
 the persons with scores above the average had in general gravi- 
 tated downward in rank, while those below the average had 
 climbed upward, such a regression to equal rank would have 
 been indicated. But this is not the case. The shifting in rank 
 is seen to be to a great extent among those in various neighbor- 
 hoods, as it were, in the roll. Many with good records at first 
 get a better standing in the later reproductions. So too, those 
 well down on the list often fall lower. In the midst of this 
 heterogeneity is there any general tendency to regression or the 
 opposite? I do not see how we can tell by merely correlating 
 the actual scores. We may expect these for a time to approach 
 each other in actual size. This, however, would not indicate 
 any tendency for them ultimately to become equal unless the 
 percentages of loss w^ere equal. In this case their relative rank 
 would of course be preserved, while the scores would approach 
 each other. If, however, the percentages of loss be different, 
 it is evident, as we have already noted, ^ that the higher rate of 
 loss, if kept up, will bring the score of the individual who suf- 
 fers it below that of any other whose percentage of loss is less, 
 — and this no matter how great may have been his original 
 superiority. It follows that, if we wish to determine whether 
 the general tendency is toward regression or the reverse, we 
 must correlate percentages of loss in the later reproductions with 
 the amounts of difference between the standard scores and those 
 in the first reproduction. 
 
 ^Correlation of Percentages of Loss in Ideas. 
 
 Between percentages of loss in ideas, 
 
 first and second reproductions, r= i6 per cent., P.E. = 5.9 per cent. 
 Between percentages of loss in ideas, 
 
 first and third reproductions, r= 19 per cent., P.E. = 5.9 per cent. 
 
 The percentage was reduced in the case of the third repro- 
 1 Page 37. 
 
44 E. N. HENDERSON. 
 
 duction largely because certain individuals who did very poorly 
 at first retained all or about all they had gained. From being 
 at the bottom of the scale in power to learn, they are therefore 
 suddenly elevated to the top in power to remember. Such a 
 radical difference between the abilities involved would seem ex- 
 ceeding unlikely. What is probably the explanation is that the 
 individuals in question found such difficulty in reading and 
 understanding the passage that only a few ideas could be 
 learned, and these were so simple and fundamental in the pas- 
 sage that they were not easily forgotten, in fact, were forgotten 
 by almost none of those whose scores were better. It is unfair 
 to our percentage of correlation to rank such cases as at the top 
 in power of retention. The difficulty created by them can be 
 in a great measure obviated by eliminating from the calculation 
 the scores of those who learned less than 50 per cent, of the 
 original in the first reproduction. It will be remembered that a 
 similar step was taken in comparing the scores according to age. 
 We are left with 105 sets of scores, and the values of r are as 
 follows : 
 
 Between rates of loss in ideas, first and 
 
 second reproductions, r^ 13 per cent., P.E. =5.4 per cent. 
 Between rates of loss in ideas, first and 
 
 third reproductions, r = 38 per cent., P.E. =5.2 per cent. 
 
 The positive values of r indicate a constant tendency for 
 those who learn more quickly to retain a greater percentage of 
 what they have gained. Moreover, in the third reproduction 
 this relation comes out more strongly than in the second, indi- 
 cating that the general tendency is even more apparent with 
 lapse of time. We may be justified in concluding, therefore, 
 that the scores are not tending toward the same goal, but that 
 the majority of those who have done better in learning will con- 
 tinue to retain their advantage. 
 
 Again, the elimination of the poorest scores does not raise 
 the value of r between the first and the second reproductions, 
 as it does between the first and the third. On the contrary, 
 the reverse relation holds. ^ The main reason for this is un- 
 
 ^ The value of P. E., however, makes the diflference of 3 per cent insignifi- 
 cant. 
 
A STUDY OF MEMORY. 45 
 
 questionably the much smaller average loss in the second repro- 
 duction. Those who lost nothing — or even improved^ — on 
 their second attempt to write out the passage numbered 39, 
 about a third of all ; while in the third trial only 17 retained all 
 they had acquired. The poor learners who remembered all 
 had, therefore, much less effect on the value of r for first and 
 second reproductions ; and their influence was more than 
 counterbalanced by the effect of the very considerable forget- 
 fulness that some of the poorer ones displayed on this occasion. 
 Taking independently the various classes of subjects that 
 took part in test i, we get the value of r between the rates of 
 loss in ideas in the first and third reproductions. 
 
 S. S. students, r—. 14.5 per cent., P. B. = 15 per cent. 
 P. S. 7B pupils, r = 57.8 percent., P. E. = 8.4 per cent. 
 P. S. 6B pupils, r= 4 per cent., P. E. = 10 per cent. 
 
 Omitting those scoring less than 50 per cent, of the standard in 
 the first reproduction, 
 
 P. S. 6B pupils, r = 35 percent., P. E. = 9 per cent. 
 
 Two important features may be noted here : the great ad- 
 vance in the percentage of correlation with P. S. 6B pupils when 
 the four who fell below 50 per cent, in the first records were 
 eliminated; and the excellent percentage shown by P. S. 7B 
 pupils as contrasted with that of S. S. students. The first of 
 these points need not be discussed further. We see clearly the 
 importance of having all subjects able to read and understand 
 the passage. The excellence of the correlation in the case of 
 P. S. 7B pupils is undoubtedly to be attributed to the fact that 
 few of them had these difficulties that swamped a number in 
 lower grades, and also to the general uniformity in training and 
 other conditions that prevailed there. The poorness of the cor- 
 relation with S. S. students is doubtless due to the fewness of 
 the cases, and to the great diversity among the students. The 
 classes contained both sexes, and the age ranged from young 
 manhood and womanhood to middle life. The training too was 
 
 1 Those who did better in their later papers were treated as having forgot- 
 ten nothing. No negative losses (positive gains) were considered. Their in- 
 troduction would have complicated the calculations, and such gains as appeared 
 were slight, and can be neglected without serious error. (See table of results. ) 
 
4^ E. N. HENDERSON. 
 
 very diverse. Some were college graduates, others not ; and 
 ■while all were teachers, they represented all ranks in the pro- 
 fession from teachers in the primary grades to a college pro- 
 fessor. It would have been highly desirable, however, to have 
 got enough such cases to have made certain whether or not the 
 percentage given is the average even for subjects thus diverse 
 in character. My opinion is that r is here too low, and that it 
 should, as elsewhere, range in the neighborhood of 40 per cent. 
 A consideration of the papers of all classes taking test 2 
 yields the following results : 
 
 Between rates of loss iu ideas, first and 
 
 second reproductions, ^ = 4 percent., P. E. = 9 per cent. 
 Between rates of loss in ideas, first and 
 
 third reproductions, r= 36 per cent., P. E. ^8 per cent. 
 
 Here again, it will be noted, the loss between the first and 
 second reproductions does not correlate significantly with the 
 original loss. Why should the second reproduction fail to 
 bring out the relation apparent in the third? Two explanations 
 might be given. First, it might be thought that certain indi- 
 viduals in the class discussed the passage in the interval. Such 
 discussion would probably come, if it came at all, immediately 
 after the first experiment, while the matter had the freshness of 
 novelty. Its effect might very well be to fix certain parts much 
 better in the minds of the poorer students. In many cases they 
 might discover mistakes and fill in gaps in their earlier notions 
 of the contents of the passage. The second reproduction, 
 coming close after these discussions, would profit by them. If 
 the discussions were general the poorer pupils would gain most 
 and the better ones least from them, and one might expect no 
 correlation between the losses in the two reproductions, or in- 
 deed, an inverse relation. 
 
 According to this explanation our values of r between the 
 first and second reproductions are far too small to indicate the 
 truth about the relation we are seeking. Moreover, if discus- 
 sion on the part of the subjects affected the third reproduction, 
 there too r should be similarly wrong. However, we may sup- 
 pose that the interval of a month would enable the relative abili- 
 ties of the various subjects again to assert themselves, so that 
 
A STUDY OF MEMORY. 47 
 
 the true relation would be more manifest than before. That the 
 longer interval enables the memory to show more clearly its 
 defects and excellencies we cannot question. The third repro- 
 duction is, therefore, undoubtedly a better basis for determining 
 the relative amount of forgetting than the second. 
 
 I have already indicated my belief that the subjects gained 
 very little from mutual discussion.^ The fact that some did 
 better in the second than in the first reproduction is not evidence 
 that they got additional information from others, any more than 
 the ability to call up something that on a previous occasion 
 could not be remembered is evidence that one has in the mean- 
 while consulted an authority in regard to it. Immediately after 
 studying the passage certain parts might be omitted because the 
 attention does not get turned toward them when the reproduc- 
 tion is made. The accident of the moment may bring them to 
 the front later. Often the later reproductions indicate a better 
 perspective in regard to the thought of the passage as a whole 
 than do the earlier ones, where the details and the expression 
 absorb much of the attention. Such a birdseye view may seize 
 on phases of thought that are crowded out of the earlier papers. 
 A study of the gains of later papers will convince one that such 
 explanations as I have just offered will account for them far 
 better than the hypothesis of intervening discussion. At any 
 rate, we may be sure that correlation percentages were very 
 little affected by it. The smallness of r in the case of rates of 
 loss in the first and second reproductions, I account for by the 
 notion that the general loss was so small with the second paper 
 that what might be called accidental forgetfulness had an unus- 
 ually large influence. Where amounts of loss are slight, dif- 
 ferences in them due to the mood of the subject, his physical 
 health for the day or the hour, accidental disturbances of the 
 attention, etc., count for much. These differences are in the 
 later reproduction swamped in the greater loss that is the normal 
 result of the longer lapse of time. On the other hand, the fact 
 that some correlation prevails between the losses in the two 
 earlier papers is significant of the truth of the principle that 
 those who learn quickest usually remember best. 
 
 ^ See page 28. 
 
•48 E. N. HENDERSON. 
 
 It will be remembered that the scores in the first reproduction 
 correlated better with those in the second than with those in the 
 third. This may seem inconsistent with the fact that with per- 
 centages of loss the higher value of r lies between the first and 
 third reproductions. There is, however, no contradiction. An 
 individual who did poorly at first might later remember all he 
 had learned. Such a case would reduce decidedly correlations 
 based on percentages of loss. He might, however, easily retain 
 his rank in the list of scores, especially in that of the second 
 reproduction where the amount of loss is in general small. In 
 the later paper the rank would be more likely to be affected, 
 because the greater general loss would probably result in send- 
 ing him up the list. In general, it may be said that the rates 
 of loss in the second reproduction, although they correlate poorly 
 with those in the first, are on the average so small that they are 
 swamped by the general size of the scores ; while in the third 
 reproduction the rates of loss, though more in harmony with 
 those of the first reproduction, are yet sufficiently divergent and 
 sufficiently large to modify the ranking considerably. 
 
 Correlation of Rates of Loss tn Words. 
 
 In regard to words, I have calculated only two correlation 
 percentages, one making use of all the test papers in test i, the 
 other based on all the papers of test 2. I have compared the 
 losses in the first reproduction with those in the third only, re- 
 garding this as the more significant value. The following 
 values of r were obtained : 
 
 Between rates of loss in words, first and third repro- 
 ductions, test I, r^ .88 per cent. 
 
 Between rates of loss in words, first and third repro- 
 ductions, test 2, r= 25 percent. 
 P. E. = 8.5 percent. 
 
 The lack of correlation in the case of test i is striking. The 
 percentage is improved by eliminating the fifteen subjects who 
 stood lowest in rank in ideas. These are with two exceptions 
 the lowest in rank in words. The result of this calculation 
 gives 
 
 r= 16.6 per cent., P. E. =6.5 per cent. 
 
A STUDY OF MEMORY. 49 
 
 • 
 
 The powers of learning and retaining words do not go so 
 well together, it would seem, as do those of learning and of re- 
 taining ideas. 16.6 per cent, as against 38 per cent, in test i, 
 and 25 per cent, as against 36 per cent, in test 2, are -vtalues 
 of r fairly convincing on this point. However, it must be kept 
 in mind that by many the words were learned rather inciden- 
 tally. All tried their best on ideas. If little attention w^ere 
 given to the words the amount of the score would be largely 
 dependent on the accidents of expression. Many of those who 
 learned and remembered well departed considerably from the 
 words of the original. Many poor in both these functions did 
 likewise. Those who retained expressions quite literally could 
 also be found scattered among both the good and the poor in re- 
 membering ideas. On the supposition that accident determined, 
 in the case of those who made no great effort to be literal, whether 
 the words of the original were used or not, we conclude that its 
 effects would be found in the papers of both the superior and 
 the inferior minds. Distributed in this impartial way, it reduced 
 the percentage of correlation. That the memory for words is 
 doubtless weaker with most people than the memory for ideas, 
 is an additional reason for supposing accident to be influential 
 in determining the score in words. 
 
 Comparison of Re stilts in Tests j, ^ and 5. 
 
 So far we have discussed the results of tests i and 2. The 
 other tests offer an opportunity for comparing the relative abil- 
 ities of a number of individuals in different classes of material. 
 Twelve students gave me complete sets of papers in all these 
 three tests. If w^e correlate their ranking in the different tests 
 we get the following rough values of r : 
 
 Ranking in Scores of First Reproductions. 
 
 'The History of the Jews ' and 'The Dutch Homestead,' r = 6i per cent. 
 
 ' The History of the Jews ' and ' The Stages in the Develop- 
 ment of Human Theory,' ^ = 85 per cent. 
 
 * The Dutch Homestead ' and ' The Stages in the Develop- 
 ment of Human Theory,' r=5o per cent. 
 
50 E. N. HENDERSON. 
 
 Ranking in Percentages of Loss, Third Reproductions. 
 
 ' The History of the Jews ' and ' The Dutch Homestead,' r = — 25 per cent. 
 
 ' The History of the Jews ' and ' The Stages in the Develop- 
 ment of Human Theory,' r = 54 per cent. 
 
 ' The Dutch Homestead ' and ' The Stages in the Develop- 
 ment of Human Theory,' ?'== — 9 per cent. 
 
 In the first reproduction a decided tendency for those who 
 did well in one passage to do well in the others is evident. 
 Especially is this relation close in the learning of the passages 
 on the Jews and Comte's philosophy. It is interesting to note 
 that only between these two tests is there any significant corre- 
 lation in regard to power to retain what was learned. An ex- 
 amination of the matter of the tests will suggest the reason. 
 Tests 3 and 5 deal with abstractions as contrasted with test 4 
 (' The Dutch Homestead '). Wide reading and philosophical 
 study would be apt to help one in learning and remembering 
 tests 3 and 5, while they would be of little value with test 4. 
 
 Effect 0/ Method of Learning on the Results. 
 
 To the undergraduates of Columbia College I gave test 3 
 by simply reading it to them slowly, asking for a reproduction 
 as nearly literal as possible. Examining their record, we see 
 that they did a little worse than the graduate students who had 
 the longer period for learning. That they did so well com- 
 paratively, is evidence of the great importance of the first read- 
 ing of a passage as compared with repetitions. The percentages 
 of loss in the later reproductions indicate that the longer period 
 is about equally valuable for fixing the ideas and words and for 
 enabling a more complete immediate reproduction. 
 
 A comparison of the ranking in the first reproduction of the 
 18 individuals who took both tests 2 and 3 shows a correlation 
 of about 37 per cent. When we correlate rank in percentages 
 of loss in the third reproduction, r = 34 per cent. It is probable 
 that the different conditions of giving the two tests were largely 
 responsible for the smallness of these values as compared with 
 those obtained from the graduate students in the most favorable 
 cases. 
 
A STUDY OF MEMORY. 5 1 
 
 Comparison of Standing in Tests with Teacher'' s Marks. 
 
 For two classes of E. H. S. students and the pupils in P. 
 S. No. 40, I obtained estimates of class standing from the 
 teachers in charge. The 12 E. H. S. 8B students were put in 
 three grades. The 3 who were put in grade I. stood first, fourth 
 and seventh in the first reproduction of my test. The 2 who 
 were put in grade III. stood fifth and tenth in these papers. The 
 correspondence between my records and the estimates of the 
 teacher was, therefore, rather small. The rank of E. H. S. lA 
 students in the test corresponded much more closely to the 
 teacher's marking. Thirteen were put by the teacher in the same 
 section of the class as that to which my records assigned them. 
 The test records advanced 5 one grade, reduced 4 one grade, 
 and I two grades. The class was in history, and it is, there- 
 fore, likely that excellence in memory would play a great part 
 in determining standard of scholarship. 
 
 With the P. S. No. 40 pupils the following results appear: 
 The 20 best P. S. 7B pupils were graded seriatim by the 
 teacher, the others ranked as in grade 21. Grading them simi- 
 larly according to their rank in my test I find that : Six retained 
 the same rank ; 24 gained in rank (average gain = 8.3 grades) ; 
 13 lost in rank (average loss = 9 grades) ; average change in 
 rank of total = 33.7 grades. Between the two rankings the 
 value of r = 13 per cent., P. E. = 11 per cent. 
 
 The pupils in P. S. 6B were graded seriatim by the teacher. 
 Two taking part in my test were omitted in this grading. The 
 comparisons follow : One retained the same rank ; 22 gained in 
 rank (average gain = 10.9 grades) ; 18 lost in rank (average loss 
 = 12.8 grades) ; average change in rank of the total 41 = 11.5 
 grades ; between the two rankings the value of r = 22 percent., 
 P. E. = 9 + per cent. ; P. S. 5B pupils were graded seriatim ; 
 16 gained in rank (average gain 5.7 grades); 11 lost in rank 
 (average loss 8.3 grades) ; average change in rank of total 27, 
 6.8 grades. Between the two rankings the value of r = 44 per 
 cent., P. E. = 10 per cent. 
 
 The percentage of correlation is here sufficiently great to 
 warrant saying that there is a tendency for the marking of 
 
52 E. N. HENDERSON. 
 
 teachers to correspond to the ability to learn quickly. On the 
 other hand, the tendency is evidently greatly interfered with, 
 and doubtless by such factors as good conduct, application and 
 reasoning power of the child, all of which are at least partially 
 independent of the power involved in my test. 
 
 I have not discussed the relation between teachers' marks 
 and rank in the power to retain. Suffice it to say that a glance 
 at the data indicates that the correspondence is much less than 
 between the marks and power to learn quickly. It is likely 
 that the day-by-day impressions are the dominant ones in deter- 
 mining the judgment of the teacher, and these are dependent 
 on such powers as are evinced best in the first reproduction. 
 
 It is notable that the teacher's ranking was by the majority 
 improved in my tests. Fifty-three gained in rank while only 
 forty-one lost. On the other hand, the average rank gained 
 was less than that lost. This would seem to suggest that special 
 performances, or certain excellencies of conduct had bettered 
 the standing of manypupils with the teacher. Such individuals 
 suffered loss of rank in the tests. There were few of them, 
 and they lost more, while the unnoted members of the class, or, 
 perhaps, the lazy or troublesome, had a chance to be graded on 
 mere ability and rose in rank in consequence. 
 
 Summary of Conclusions on Amounts of Loss in the Various 
 
 Tests. 
 
 The conclusions reached so far have been largely quantita- 
 tive, relating to the amounts of loss in the various cases rather 
 than to its character. They may be summed up as follows : 
 
 1. The total scores of the various classes in all reproduc- 
 tions and in both ideas and words tend to increase from the 
 lower grades to the more advanced ones. 
 
 2. There seems to be a slight increase in power to learn as 
 we go from younger to older students. The evidence indicates 
 that this increase is probably wholly dependent on greater abil- 
 ity to read and understand the passages. The older students 
 do not remember a greater percentage of what they have 
 learned. Training would, therefore, seem to affect more the 
 power to learn than the power to retain. 
 
A STUDY OF MEMORY. 53 
 
 3. Those who learn quickest retain in general a greater per- 
 centage of what they have learned. This holds true of the 
 percentages remembered in both the second and third reproduc- 
 tions, but the rule comes out more clearly in the latter case than 
 in the former. The same law holds between power to learn and 
 to retain words, although the percentage of correlation is not 
 so great as with ideas. 
 
 4. Individuals tend to keep the same rank in power to learn 
 in the various tests. In power to retain the ranking was more 
 widely variant. When the material of the two tests was simi- 
 lar it corresponded to a significant extent. 
 
 5. A single reading with concentrated attention seems nearly 
 as effective in enabling an immediate reproduction of one of 
 these passages as is three minutes of study. The longer pe- 
 riod, however, seems more effective in increasing the percentages 
 both of words and ideas retained. 
 
 6. Teachers' marks in grammar grades correlate only to a 
 small extent with the ranking in the tests. In a higher grade, 
 however, where the subject demanded powers like those tested, 
 the correspondence was found to be close. 
 
 Let us now proceed to consider the changes in the character 
 of the material in the later reproductions. 
 
 B. Character of the Changes in the Later Reproductions. 
 
 The investigation of the fortunes of the various factors that 
 enter into our experience in the subsequent history of our con- 
 sciousness, if an almost hopeless, is at least a fascinating occu- 
 pation. Certainly most of our purposes in education are dominated 
 by our theories in regard to the ebb and flow of the waters upon 
 which we as teachers cast our bread. If we are to test at all 
 the accuracy of these theories, to penetrate in the least into the 
 complexity of the modifications that memory imposes upon our 
 ideas, we must, of course, be content, at any rate in the early 
 stages of investigation, with examining small patches of experi- 
 ence with a view toward ascertaining the history of the definite 
 cultures planted therein. The reproductions that I have obtained 
 are an approximation, I take it, to a cross-section revelation of 
 successive conditions in the minds of the various individuals 
 tested with reference to the passages in question. By compar- 
 
54 E. N. HENDERSON. 
 
 ing these reproductions, therefore, we should be able to ascer- 
 tain what the fate of individual ideas may be, and the effect of 
 their destinies upon the general character of the thought of the 
 whole passage involved. 
 
 I have approached these complicated matters from four direc- 
 tions, gathering a little contributory information from each study. 
 In the first place, an examination of the words, the underscoring 
 of which indicated the confidence of the subject that they oc- 
 curred in the original, throws some light on the changes in 
 thought that accompany changes in the power to recognize. 
 Secondly, the study of answers to certain questionnaires that I 
 submitted to the graduate students on the occasion of their third 
 attempt at reproducing the various passages, revealed to me 
 the method by which their thought regarding the contents of the 
 passages unfolded at that time. The third study involved the 
 comparison of the amounts of forgetting of the detailed ideas 
 and of the topics and the subtopics into which the thought of 
 the passages was analyzed. Last and most suggestive of all 
 was the investigation into the specific history of detailed ideas 
 and of topics from reproduction to reproduction. Such a study 
 might be carried on almost indefinitely, and I can only claim to 
 have continued it far enough to be able to formulate roughly 
 the leading kinds of modification that the memories evince. 
 
 /. Recognized Words. 
 
 It will be remembered that the graduate and college stu- 
 dents and those in the summer session underscored in the 
 second and third reproductions the words which they were con- 
 fident had occurred in the original. With the college students 
 this was imperfectly done, depending apparently on whether the 
 student felt inclined to use the time and effort to proceed to the 
 task of judging how he felt regarding words. In fact, the 
 omission of any scoring cannot be taken even with graduate 
 students as a certain indication that no words were recognized. 
 
 The following table indicates the results with S. S. students. 
 They are typical of the others : 
 
A STUDY OF MEMORY. S5 
 
 Underscored Words. S. S. Students. Test i. 
 
 Rank in 
 
 Rec. W,' 
 
 Mistakes 
 
 Rec. W. 
 
 Mistakes 
 
 Rec. W. 
 
 Mistakes 
 
 Mistakes 3d 
 
 ist Rep. 
 
 2d. 
 
 2d. 
 
 3d. 
 
 3d. 
 
 2d & 3d. 
 
 2d & 3d. 
 
 used 2d. 
 
 Z 
 
 69 
 
 12 
 
 66 
 
 12 
 
 44 
 
 9 
 
 I 
 
 2 
 
 56 
 
 23 
 
 38 
 
 16 
 
 32 
 
 9 
 
 7 
 
 3 
 
 65 
 
 28 
 
 69 
 
 26 
 
 48 
 
 13 
 
 3 
 
 4 
 
 61 
 
 ID 
 
 46 
 
 8 
 
 39 
 
 4 
 
 I 
 
 5 
 
 50 
 
 30 
 
 21 
 
 13 
 
 16 
 
 8 
 
 I 
 
 6 
 
 18 
 
 9 
 
 13 
 
 9 
 
 12 
 
 3 
 
 
 
 7 
 
 35 
 
 8 
 
 29 
 
 9 
 
 28 
 
 
 
 I 
 
 8 
 
 26 
 
 15 
 
 8 
 
 3 
 
 8 
 
 2 
 
 
 
 9 
 
 — 
 
 — 
 
 15 
 
 2 
 
 — 
 
 — 
 
 2 
 
 lO 
 
 3 
 
 4 
 
 6 
 
 3 
 
 3 
 
 3 
 
 
 
 II 
 
 20 
 
 4 
 
 II 
 
 I 
 
 9 
 
 
 
 
 
 12 
 
 67 
 
 18 
 
 56 
 
 29 
 
 49 
 
 9 
 
 5 
 
 13 
 
 22 
 
 12 
 
 — 
 
 — 
 
 — 
 
 — 
 
 — 
 
 14 
 
 14 
 
 2 
 
 8 
 
 2 
 
 6 
 
 I 
 
 
 
 15 
 
 27 
 
 31 
 
 5 
 
 10 
 
 5 
 
 7 
 
 2 
 
 i6 
 
 4 
 
 2 
 
 16 
 
 13 
 
 4 
 
 
 
 5 
 
 17 
 
 17 
 
 6 
 
 20 
 
 15 
 
 14 
 
 5 
 
 6 
 
 Totals, 
 
 554 
 
 214 
 
 427 
 
 171 
 
 317 
 
 73 
 
 34 
 
 A large part of the variety shown by the various scores in- 
 dicates not so much differences in ability to recognize, as dif- 
 ferent standards in regard to the degree of familiarity that war- 
 rants the underscoring of the word. Certain common features 
 appear in the total scores of every class involved. 
 
 1. More underscoring in the second than in the third repro- 
 duction. It will be noted, however, that this rule has excep- 
 tions in the case of individual records. 
 
 2. Greater percentage of errors in the judgments in the third 
 reproduction. This rule was far more apparent with the grad- 
 uate students than in the table given. It will be seen to hold for 
 most of the individuals in that table, however. 
 
 3. Of the correctly underscored words a far larger per- 
 
 1 The meaning of the scores is as follows : J^ec. W. 2d, words accurately 
 recognized, 2d reproduction. Mistakes 2d, words not in the original but 
 underscored, 2d reproduction. Rec. IV. 3d, words accurately recognized, 3d 
 reproduction. Mistakes 3d, words not in the original but underscored, 3d repro- 
 duction. Rec. W. 2d and 3d, words accurately recognized in both 2d and 3d 
 reproductions. Mistakes 2d and 3d, words not in the original but underscored 
 in both 2d and 3d reproductions. Mistakes 3d used 2d, words not in the 
 original, used in 2d reproduction but not underscored, underscored, 3d repro- 
 duction. 
 
56 
 
 E. N. HENDERSON. 
 
 centage is identical in the two reproductions than of those in- 
 correctly underscored. 
 
 4. The errors are less frequent than the correct judgments. 
 This rule is not without exxeptions, as will be seen, in indi- 
 vidual cases. 
 
 5. In general, those who remembered better seemed to un- 
 derscore a little more freely, although not more correctly, than 
 those who remembered worse. 
 
 These results are such as we might expect. They indicate 
 that with the lapse of time there is less feeling of certainty in 
 regard to the words, and more actual error in judging them. 
 Also the words correctly underscored remain to a greater ex- 
 tent the same in both reproductions, while the erroneous judg- 
 ments concern words more than half of which are different in 
 the two cases. We may infer that with the lapse of time the 
 false judgments would come to surpass in number the correct 
 ones, partly because of forgetfulness of the original words, 
 partly because new words, once introduced, would tend to be- 
 come familiar and eventually regarded as part of the original 
 text. In this connection the last column of the table is inter- 
 esting. 
 
 If we compare the number of words correctly underscored 
 with the number occurring in the original text that are actually 
 used, we obtain the following results. The scores indicate 
 averages per individual. 
 
 Words Correctly Words of the 
 
 
 
 
 Underscored. 
 
 Original Use 
 
 a. 
 
 
 "S. S. students, 
 
 2d reproduction. 
 
 32.5 
 
 59 
 
 55 per cent 
 
 <i 
 
 
 3d 
 
 25 
 
 52 
 
 48 
 
 ( 
 
 Graduates, 
 
 Test 3, 
 
 2d 
 
 8.2 
 
 38 
 
 22 
 
 ( 
 
 
 (( (( 
 
 3d 
 
 9-3 
 
 29 
 
 32 
 
 ( 
 
 
 " 4, 
 
 2d 
 
 24.4 
 
 67 
 
 36 
 
 ( 
 
 
 << (( 
 
 3d 
 
 18 
 
 52-5 
 
 34 
 
 ( 
 
 
 " 5, 
 
 2d 
 
 12 
 
 39 
 
 32 
 
 1 
 
 
 i( (( 
 
 3d 
 
 8.2 
 
 24-5 
 
 33 
 
 ( 
 
 The difference between the number of words of the original 
 used and that underscored can not be satisfactorily accounted 
 for by supposing a tendency to underscore only striking and 
 important words, leaving out conjunctions, prepositions, articles, 
 etc. While this explanation has some significance, it must also 
 
A STUDY OF' MEMORY. 57 
 
 be said that those who underscored freely marked whole sen- 
 tences, clauses or phrases. Hence only with the less confident, 
 who have small scores, is the ratio in question affected by the 
 failure to underscore less striking words. When we compare 
 individuals, we note that the number of correct recognitions 
 seems only in the most general way related to the number of 
 original words actually used. The relation is little if any more 
 significant than that between rank in correct underscoring and 
 that in number of ideas reproduced. In the second reproduc- 
 tion No. 12 among S. S. students out of 69 original words used 
 recognizes 67. No. 10 recognizes only 3 out of 47. Ability to 
 remember ideas or to use the words of the original does not 
 correlate to any significant extent with confident and accurate 
 underscoring. 
 
 What are we to say in general about this power of recogni- 
 tion? At best it is evidently uncertain, and it grows worse 
 with lapse of time. But while the percentage of errors in 
 underscored words varies in the tests from 9 per cent, to 28 per 
 cent, of the total underscored in the second reproduction, and 
 from 20 to 30 per cent, of this in the third, the words thus 
 erroneously ' recognized ' with but few exceptions represent 
 accurately enough the thought. It is plain that the feeling of 
 familiarity with the thought has overspread the attitude toward 
 the words used to express it. On the other hand, there is an 
 increasing loss of confidence, a decrease in the amount of 
 ' feeling of recognition,' and this in turn is justified by the 
 increasing inaccuracy in the underscoring, and by the disap- 
 pearance from the reproductions of many more of the original 
 words. We note how closely the ratios of correct underscorings 
 to original words used keep to each other in the second and 
 third reproductions.^ The loss of confidence and the change in 
 expression are doubtless due to the same cause, and this cause 
 is evidently a change in the relations between the thought ele- 
 ments that enter into the passages, rather than a disappear- 
 ance of these. What is the nature of this change? Some 
 light will be thrown upon it by the testimony of the students 
 themselves. 
 
 1 Compare table, page 55. 
 
58 E. N. HENDERSON. 
 
 2. Accounts of Introspection of Graduate Students. 
 
 The graduate students were allowed to make at leisure the 
 third reproduction of the passages submitted to them. At the 
 same time they were asked to note the course of their thought as 
 they endeavored to recall the contents of these passages. The 
 following directions were given them with a view toward ren- 
 dering their reports more congruous and definite. 
 
 1. Recall the sketch of The History of the Jews ^ that you 
 memorized, noting the contents of your minds as you attempt to 
 get a complete account. Note : 
 
 {a) Where images of sight occur in recall. Indicate their 
 character. 
 
 {h) Where a ' feeling ' of the contents enters in. Indicate 
 as far as you can its nature, and in what form, words or phrase 
 it first takes definite shape. 
 
 (c) Where you hesitate and reason out what should enter in. 
 Indicate the line of reasoning and whether you are confident 
 that your results correctly reproduce the original. 
 
 {d^ In what order the various words and phrases make their 
 appearance. 
 
 Make the record a running sketch of the progress of your 
 thought as you endeavor to get as complete and literal an ac- 
 count as you can of the passage. 
 
 2. Write as good a record as you can of the contents of the 
 passage, getting the thought and where possible the words. 
 Underscore the words which you are confident occurred in the 
 original. 
 
 3. To what extent, if at all, have you thought about the 
 passage between the first and the second reproductions, and 
 between the second and last ones? 
 
 As the method of study was by reading typewritten copies 
 of the passages, visual images of these, I thought, might play 
 a prominent part in the recall. Of the eighteen individuals who 
 reported their experience in recalling the passage on ' The His- 
 tory of the Jews,' ten mention the image of the appearance of 
 the paper, and many picture a number of words written thereon. 
 
 ' For the other tests the other titles are to be substituted. 
 
A STUDY OF MEMORY. $9 
 
 Six report visual images connected with the thought of the 
 passage, as of Babylon, Palestine, armies, people wandering 
 about in Palestine, etc. Four make no report of visual 
 imagery. In the case of the passage on ' The Dutch Home- 
 stead,' fourteen out of fifteen report visual images of the ap- 
 pearance of the house. One alone pictured the words, and he 
 only to a slight extent as compared with his experience with the 
 former passage. Two note the absence of the images of print- 
 ing, etc., noticed before. In the case of the third passage, on 
 Comte's philosophical ideas of the evolution of theories, four 
 out of fourteen report visual images of the paper, etc. None 
 report any other pictures, and five note the absence of visual 
 imagery. 
 
 We have here, from the point of view of recall, two types 
 of images. One, that picturing the paper, etc., reproduces the 
 original situation of learning, with a view to recalling its asso- 
 ciates ; the other reproduces the visual aspects of the meaning 
 of the passages. One person reports an absence of images, but 
 instead a sense of muscular adjustment, strain of attention, etc., 
 which was supposed to be similar to that which accompanied 
 the learning of the passages. Here^ we have that mental atti- 
 tude which, as I have before indicated, constitutes the vaguest, 
 most general clue to the recall of any definite group of ideas. 
 With one subject it involved the image of Irving's ' Sketch 
 Book,' with another that of an earlier time when the story was 
 read. 
 
 Starting from this general attitude of attention like that 
 when the -passage was learned with its accompanying imagery 
 of paper or what not, the next ordinary feature in the recall 
 seems to be the thought of the meaning of the passage. This 
 is introduced commonly by a vague, often inarticulate sense of 
 some peculiar feature of the thought as a whole. In the case 
 of ' The History of the Jews,' it expresses itself with various in- 
 dividuals as a feeling of ' reverses of fortune,' ' a struggling 
 people,' * balanced parts,' ' harassed people,' ' an outline of 
 Jewish history divided into distinct episodes,' etc. In the recall 
 of ' The Dutch Homestead,' this second stage seems with most 
 
 'See p. 24. 
 
6o E. N. HENDERSON. 
 
 to be properly represented by the picture of the house. Gen- 
 eral ' feelings of content ' here are absent, or are confined to a 
 sense of the comfort, efficiency, etc., that characterizes the life 
 of the inhabitants of the ' homestead.' These are not, I im- 
 agine, important 'cues' in the development of the recall. 
 With the account of Comte's philosophy, the second phase of 
 recall consisted usually, I should judge, in a sense of the ' three 
 stages,' more or less definitely conceived. Many of the papers 
 got no further than an attempt to realize what these stages are. 
 They were frequently erroneously stated, but the threeness was 
 invariably present. Whenever the individual was acquainted 
 with Comte's speculations, the passage was recognized as an 
 exposition of these, and the consciousness of this constituted a 
 factor in the second phase of recall. 
 
 We now come to the consideration of the methods by which 
 this general sense of meaning develops into the details of the 
 reproduction. Two distinct lines of procedure are here appar- 
 ent. If the passage were well memorized and retained, the 
 general sense of meaning quickly unfolded itself, revealing in 
 their proper order and without much effort the details of the 
 original presentation. Here the sense of words, phrases and 
 their order was frequently of much importance. Several whose 
 memories for words were good relied largely on this mechanical 
 kind of association to carry them on. When the course of the 
 recall was checked, they simply held themselves in an expec- 
 tant attitude, hoping for the right thought or expression to rise. 
 If the detail came properly, it brought with it frequently a series 
 of the original expressions. The points in the description of 
 the Dutch homestead usually developed easily, rapidly and in 
 nearly their proper order. It was the easiest of the passages 
 to commit, which means that the thought was congruous, and 
 could be embodied in a general picture, all the details of which 
 could fairly well be in the mind at once. Hence they recurred 
 readily when the general image was once in mind. 
 
 But often, instead of waiting in a receptive mood for the 
 thought to spring up, the minds of the students went out ac- 
 tively to seek it. They reasoned in the details, either from their 
 general notion of the matters with which the passages dealt, or 
 
A STUDY OF MEMORY. 6 1 
 
 from their sense of what might fittingly be said in view of the 
 main theme. One individual worked out his reproduction of 
 ' The History of the Jews ' from his notion of the ' Divine 
 Providence' shown therein. Another remembered that the 
 Jews were a ' remarkable people ' and reasoned to the causes. 
 This procedure was especially important in the case of the pas- 
 sage on Comte's philosophy, where the * three stages ' fur- 
 nished abundant opportunity for speculation. The reasoning 
 always had for its object filling in some detail or details which 
 the larger, or vaguer, thought of the passage demanded for 
 symmetry or completeness. Its conclusions were tested by their 
 familiarity in the context, but this criterion did not always suf- 
 fice to keep out details not in the original. Even ' The Dutch 
 Homestead' furnished occasion for logic, as when the thought 
 that such a description should begin with the outside suggests 
 the roof, which is promptly recognized as an initial element in 
 the description. 
 
 We may say that the introspective accounts seem to show the 
 presence in most subjects (doubtless in all) of the following 
 steps in recall. 
 
 1. A preliminary adjustment of the attention comparable to 
 that when the passage was learned, and accompanied by im- 
 agery of the experiences then. 
 
 2. A sense of the general meaning of the passage. 
 
 3. The unfolding of the details either by waiting expect- 
 antly for certain cues to lead to results, or by reasoning from 
 them to the details, and testing these by the criterion of famili- 
 arity in the context. 
 
 The first of these phases is only externally related to the 
 contents of the passage. The sense of the meaning is, how- 
 ever, especially where the passage is not well memorized verb- 
 ally, the basis of the recall of most of the details, and it also 
 determines largely the order of their reproduction. Congruity 
 with it is, doubtless, the basis of recognition of thought, some- 
 times of words. What is its true relation to the details? Evi- 
 dently it is not transcendent of them. It is rather immanent in 
 them. Better, it is made up of them. It is a vague con- 
 sciousness of their presence, and the words we conjure up to 
 
63 E. N. HENDERSON. 
 
 express it are frequently evidence of its composite quality. It 
 is not an abstraction. It is the concentrated essence of the 
 whole. Its generality is due to the suppression of details not by 
 banishment but by absorption. It is the first thing to come up 
 in our minds, because it is the one thing that was growing while 
 every other thing was being thought, and it represents, in fact, 
 is their total effect on our consciousness. 
 
 To the study of the method by which this meaning is com- 
 posed I will now turn. It is created out of details, but they are 
 modified through this productive process. They are still further 
 modified when in a later reproduction they are evoked at the 
 bidding of the general meaning. What is the total result of 
 these changes? Is it a minute account of certain fragments of 
 the thought with forgetfulness of others, or is it a fairly complete 
 representation of the larger topics combined with a vague sense 
 of the details which contribute to them. 
 
 3. Comparative Strength of Memory for Details and 
 for Larger Topics. 
 
 I have already indicated an analysis of the passages used 
 into main topics, subtopics and details.^ This division makes 
 it possible for us to compare on two levels the extent to which 
 the more and the less general elements are remembered. We 
 note, however, that the direct ratio of percentages of loss of 
 details to that of subtopics or topics would be of little signifi- 
 cance, because the disappearance of a topic would, of course, 
 involve the loss of the details coming under it. A 50 per cent, 
 loss in topics would ipso facto involve about a 50 per cent, loss 
 in details.^ The percentage of loss of details would, therefore, 
 surely be greater than that of larger topics. A more significant 
 comparison would be based on the percentages of loss of details 
 within the subtopics that are remembered, and of subtopics 
 within the topics that were retained. The data in the following 
 table were estimated in this manner. The score in the first re- 
 production is as before the basis for the calculation of the per- 
 centages in the third. 
 
 1 See pp. 28-29, 33- 
 
 *Not necessarily exactly 50 per cent, loss in details, because diflFerent topics 
 have varying scores in regard to details. 
 
A STUDY OF MEMORY. 63 
 
 
 
 First Reproduct 
 
 ion. 
 
 Third Reproduction. 
 
 
 
 Per Cent. 
 
 Per Cent. 
 
 Per Cent. 
 
 Per Cent. 
 
 Per Cent. 
 
 Per Cem 
 
 
 
 Loss 
 
 Loss 
 
 Loss 
 
 Loss 
 
 Loss 
 
 Loss 
 
 Test. Class. 
 
 Details. 
 
 Subtopics. 
 
 Topics. 
 
 Details. 
 
 Subtopics. 
 
 Topics, 
 
 I. 
 
 S. S. students, 
 
 17.3 
 
 5-7 
 
 2.6 
 
 10.3 
 
 4.1 
 
 
 
 
 P. S. 7B pupils, 
 
 31-7 
 
 6.6 
 
 5-0 
 
 8.9 
 
 5-9 
 
 4.9 
 
 
 P. S. 6B pupils, 
 
 20.3 
 
 8.7 
 
 8.3 
 
 2.1 
 
 6.4 
 
 1.4 
 
 
 P. S. 5B pupils. 
 
 239 
 
 14.8 
 
 12.3 
 
 7.0 
 
 9-9 
 
 3-7 
 
 
 Better P. S.6B1 pupils. 
 
 20.0 
 
 8.1 
 
 4.8 
 
 2.2 
 
 6.4 
 
 1-4 
 
 
 Better P. S.5B pupils. 
 
 17.6 
 
 13.6 
 
 5.8 
 
 6.8 
 
 10.2 
 
 4.1 
 
 2. 
 
 E. H. S. lA students, 
 
 22.9 
 
 15-5 
 
 8.6 
 
 11.8 
 
 10.6 
 
 21.7 
 
 
 E. H. S. 4B " 
 
 17.8 
 
 15-3 
 
 ir.7 
 
 9-1 
 
 9.8 
 
 18.1 
 
 
 E. H. S. 8B " 
 
 39-2 
 
 ID. I 
 
 3-7 
 
 9-5 
 
 4 
 
 8.6 
 
 
 College students, 
 
 13-4 
 
 8.9 
 
 4.2 
 
 10.4 
 
 7 
 
 12.7 
 
 3. 
 
 College 
 
 31-7 
 
 20.8 
 
 15-4 
 
 14. 1 
 
 15-9 
 
 9-5 
 
 
 Graduates, 
 
 19-3 
 
 15-8 
 
 18.7 
 
 9-9 
 
 16.9 
 
 3-8 
 
 4. 
 
 Graduates, 
 
 18. 1 
 
 17-5 
 
 
 
 17.6 
 
 12.0 
 
 5-2 
 
 5- 
 
 Graduates, 
 
 25 
 
 19-5 
 
 17.8 
 
 25.6 
 
 13-3 
 
 17-4 
 
 In every case the percentage of loss of topics in the first 
 reproduction was less than that of details, and in 11 out of 12 
 cases it was less than that for subtopics. In the third reproduc- 
 tion of tests I, 3, 4 and 5, the topics suffered a smaller percen- 
 tage of loss than the details. The subtopics invariably lost a 
 smaller percentage than the details in the first reproduction, and 
 in 7 out of 12 cases in the third. In test 2, where the topics 
 lost heavily in the third reproduction, the subtopics lost less than 
 the details in 3 out of 4 cases. We may therefore say that the 
 details disappear rapidly, leaving a comparatively large per- 
 centage of the leading topics, which are in the reproductions 
 rendered by a few representative details. The idea of the pas- 
 sage becomes generalized. Of course, there is a limit to this 
 generalization, beyond which further loss of details leaves 
 nothing that can become articulate in a reproduction. Here the 
 loss in topics must equal that in details. 
 
 If we were to eliminate the subtopics from the calculation, 
 we would find that in every case the percentage of loss for 
 details would be far in excess of that for topics. When we 
 consider that in the tests there are about as many details in 
 each topic as there are topics in each test, such a comparison 
 will seem a fair one. A second tendency may be noted, which, 
 although probable, could hardly have been so confidently pre- 
 
 * Omitting those below 50 per cent, first reproduction. 
 
64 E. N. HENDERSON. 
 
 dieted as the first. In both test i and test 2 the more advanced 
 students seem usually to suffer smaller loss in topics and sub- 
 topics than the younger ones. The perfection of this generali- 
 zation is interfered with by the excellence of E. H. S. 8B stu- 
 dents (a select class) in topics, and by a large loss in topics by P. 
 S. 7B pupils in the third reproduction. The superiority of the 
 more advanced students in grasping and retaining the general 
 meaning of the passage even when they lost as large or even a 
 larger percentage of details is, however, unmistakable. Age 
 and training, especially the latter, have doubtless something 
 to do with this, although the exclusion of inferior minds from the 
 more advanced classes may partly account for it. 
 
 4. Histoj'y of the Ideas Regarding the Contents of the 
 
 Passages. 
 
 Can this same generalizing tendency be discovered in the 
 later rendition of specific ideas in the passages? The general 
 ideas, we agree, tend to be preserved better than the details. 
 But they are themselves made up of details. Some definite 
 content is necessary, or we could not recognize the topics. 
 What details remain that we may identify the larger thoughts? 
 Do they suffer transformation? What details are lost? Are 
 they utterly lost? The answer to these questions should enable 
 us to tell not only whether the memory of such passages is a 
 memory for general ideas rather than for details, but also 
 whether there is, as it were, a generalization of the details. If 
 such is the case, then we should be able to offer a description 
 of the process of generalization. 
 
 The method of my study has been to follow the fate of par- 
 ticular ideas and contents from reproduction to reproduction. 
 The three reproductions were placed side by side in parallel 
 columns. The parts of each that deal with specific portions of 
 the original were kept as far as possible on the same horizontal 
 lines. Whatever transpositions were necessary to effect this 
 arrangement were indicated. On such a paper the modifica- 
 tions in later reproductions are readily apparent. A plausible 
 explanation for most of them is equally easy. But a classifi- 
 cation of the changes by which they may be accounted for is, I 
 
A STUDY OF MEMORY. 65 
 
 have found, somewhat complicated. For the sake of avoiding 
 too much detail and confusion, I shall present a simple state- 
 ment of the general character of the modifications, as I have 
 come to conceive them. I shall then discuss one of the pas- 
 sages, pointing out characteristic transformations of each part. 
 This will be followed by a comparison of the differences be- 
 tween the modifications of the different passages, and a record 
 of the extent to which different parts of the passages were re- 
 tained. Finally, a comparison of these records in the case of 
 P. S. pupils will enable some further deductions as to the effect 
 of age and training on the generalizing power. 
 
 a. Types of Changes in the Reproductions. 
 
 These changes may be summed up under three heads : re- 
 grouping, simplification, introductions. ^ 
 
 Rcgrotiping. — The topics and subtopics which are most 
 closely allied tend to be brought together, even when the}'- were 
 treated in separate portions of the original text. The kinship 
 may be a result of some sameness in details, or may lie in the 
 general similarity of the two topics, or in their being comple- 
 mentary parts in the treatment of some general theme. Ex- 
 amples are as follows : In the account of the Egyptian king, 
 topics 4 and 9 are often combined. One is ' when asked by a 
 favorite the reason for this change, he replied ' ; the other gives 
 the moral of the story, — /. e., the reason for the change. In the 
 account of Cicero, subtopic 2 — the statement of Cicero's posi- 
 tion among Roman orators — is frequently combined with topic 
 5 — the comparison of Cicero's oratory with that Demosthenes. 
 In the outline of Jewish history, the conflicts of the Jews with 
 neighboring tribes, topic 5, and with the nations of the Eu- 
 phrates, topic 7, are often brought together. 
 
 Illustrations might be multiplied. These will serve to make 
 my meaning clear. Such transpositions might be expected, 
 and, as a matter of fact, they take place on a large scale. 
 Often a subordinate detail in one context will be reproduced in 
 another connection where it fits, and the original context, thus 
 disrupted, if it cannot be dragged after its wandering member, 
 may be forgotten. Thus the gradual rise of the Jews 'to the 
 
66 E. N. HENDERSON. 
 
 rank of a powerful, opulent and commercial nation,' topic 6, is 
 assimilated with their increase ' so rapidly as to become on a 
 sudden the fierce and irresistible conquerors of their native val- 
 leys,' topic 3. Here the contradictory or inconsistent ideas of 
 course disappear. 
 
 Sim-plijication, — The last illustration involves not only re- 
 grouping, but assimilation with condensation and modification of 
 the details involved. Thus the passage becomes simplified. 
 The general result of this process is to force to the front cer- 
 tain striking central thoughts, and to make the other details 
 more closely dependent upon them, and so less conspicuous. 
 Taking up the two specific factors in the process, let us deal 
 with (i) condensation. Two types of this appear. First, ideas 
 that are repeated receive only one expression. We have seen 
 that a similarity of ideas will frequently cause two or more sep- 
 arated contexts to coalesce. Here the similar ideas will, of 
 course, usually be expressed but once. The Jews are said to 
 be in servitude to the Egyptians, subtopic 6, to neighboring 
 tribes, subtopic 15, and to the Babylonians, subtopic 23. In 
 several cases these ideas coalesce ; e. g.^ 'sometimes enslaved 
 by the Babylonians, Assyrians, Egyptians,' etc. Again the 
 Jews are said to be * twice exiled and twice restored to their 
 •country,' thus uniting topic 2 with 7, and 3 with 8. Similarly 
 in test 2, 'orators,' subtopic 2, and 'oratory,' subtopic 11, are 
 frequently a basis for condensation of expression, and in test 4 
 ' lowly sloping roofs ' makes ' low-projecting eaves ' unnec- 
 essary. 
 
 The second type of condensation is the omission of unim- 
 portant words by substituting words for phrases, or clauses. 
 ' In the annals of mankind,' test 3, becomes ' of the world,' ' of 
 mankind,' etc. ' Placed side by side with Demosthenes, or, at 
 least, close after him,' test 2, becomes with slight loss ' placed 
 second only to Demosthenes.' ' Through the ranks of Roman 
 official service ' becomes ' through the official ranks,' Roman 
 being understood because the idea was brought out before. 
 < Suddenly he changed his course and ruled so well as to be 
 called the just,' test i, becomes 'suddenly he so changed his 
 course as to become known as the just.' 'Was struck on the 
 
A STUDY OF MEMORY. 67 
 
 head by a stone, that cracked its skull. The stone was thrown 
 by a man ' is rendered ' whose skull was broken by a stone 
 thrown by a man.' ' Asked the reason for this change ' becomes 
 ' asked why.' ' Government and code of laws totally different 
 from those of any other rude or civilized community,' test 3, 
 appears as ' most peculiar government and laws.' The phrase- 
 ology thus omitted indicates relations that stood for the special 
 line of march of the thought of the writer of the passage. But 
 the meaning in the mind of the subject is composed of all the 
 points of view gained as a result of this journey. Because they 
 are all in the mind vaguely when the passage is about to be 
 reproduced, they need not be reached by the tortuous paths by 
 following which they originally came to view. The memory 
 inevitably condenses. It is important to note that this conden- 
 sation is to a very great extent no real loss. 
 
 (2) Modification. — Regrouping and condensation frequently 
 bring about certain modifications. Often ideas having some- 
 what the same functions are ' fused ' into a new product partak- 
 ing of the nature of each of its parents. For example, ' whose 
 reign had long been a course of savage tyranny. Long had he 
 ruined the rich and distressed the poor,' test i, is rendered, 
 ' who ruled the rich tyrannically and the poor harshly.' Here 
 the number of distinct details (according to my scheme of scor- 
 ing) is reduced nearly one half, but only one significant idea — 
 that of the length of the tyranny — is, I should judge, clearly 
 lost. ' Ruling tyrannically ' when applied to the rich does fairly 
 well for 'reign,' 'ruined' and 'tyranny.' ' Harshly' gives us 
 a fair compound for 'distressed' and 'savage.' 'Course' is 
 unimportant. The sentence, ' His family was of the middle 
 class only and without wealth,' test 2, becomes in several 
 instances something like ' His family was of moderate means.' 
 Here ' middle class ' is lost but ' moderate ' and ' means ' testify 
 that the thought has not utterly failed to influence the result. In 
 the same passage ' only ' was frequently lost. By way of com- 
 pensation for one negation, another, the lack of wealth, was 
 emphasized. The family is declared to be ' by no means,' ' not 
 at all,' or ' not in the least wealthy.' In test 3 one paper fuses 
 ' sprung from one stock they pass the infancy of their nation in 
 
68 E. N. HENDERSON. 
 
 a state of servitude in a foreign country ' into ' having their 
 origin as slaves to an alien race.' ' Sprung ' and ' infancy ' 
 unite in ' origin,' while ' race ' is more akin to * stock,' an 
 omitted word, than * country' for which it is substituted. 
 
 Sometimes ideas are in consequence of appearing in a new 
 context modified so as to fit therein. Here the idea suffers in 
 its integrity, not to preserve another that would else be lost, but 
 to save itself. Cicero is said to have been ' bitter ' instead of 
 having a ' venomous bite,' the context of the latter statement 
 having disappeared. 
 
 Modifications are sometimes made in order to make the ideas 
 consistent with the experience or preconceptions of the indi- 
 vidual. The ' stages ' of Comte are converted into a variety 
 of schemes; e. g.., ' metaphysical or speculative, ethical or re- 
 ligious, and scientific or exact.' The ' favorite ' of the Egyptian 
 King becomes a * servant,' a ' friend,' ' one of his people,' etc. 
 The ' fox,' a strange animal to city children, becomes a ' pig.' 
 ' Trod on,' an unfamiliar expression in test i, is translated into 
 * trampled on,' ' run down by.' In the last instance the fusion 
 with ' ran in the way of ' and ' kicked by ' (used in an earlier 
 reproduction) is evident. ' Changed his course and ruled so 
 well, etc.,' becomes ' changed his unjust rules,' rules being a 
 prominent idea to the much governed school-child. 
 
 Of course, there is a great deal of downright omission in 
 the reproductions. It usually takes place first with the insigni- 
 ficant ideas. They are lost largely because rearrangements 
 have rendered them useless or incongruous. The notable thing 
 is the extent to which all ideas that contribute any marked share 
 to the thought which may be said to constitute the total sense of 
 the passage leave some evidence of themselves. This matter 
 I shall discuss quantitatively later. 
 
 Introductions. — One of the most evident peculiarities of 
 the reproductions is the extent to which certain expressions or 
 thoughts are made to do service as substitutes in other contexts 
 than that in which they originally occurred. Situations at all 
 familiar are described as nearly as can be in the same terms. 
 This similarity of expression leads, of course, to greater same- 
 ness of thought than existed before, and aids in the generaliza- 
 
A STUDY OF MEMORY. 69 
 
 tion of the passage. The ' tyranny ' of the Egyptian King is 
 called ' injustice ' because his later conduct made him called 
 * the just.' This change is a type of a very common mode of 
 treating this part of the passage. One subject carries the 
 thought of justice and its opposite so far that ' ruining the rich 
 and distressing the poor ' becomes * condemning the innocent 
 and oppressing the righteous.' Again ' trampled on the poor ' 
 instead of ' distressed the poor ' is doubtless influenced by the 
 later ' trod on ' by a horse. In the Cicero test the word ' rank ' 
 was frequently used in comparing the Roman with Demosthenes, 
 a context where it does not appear in the original. 
 
 The Jews are said to be ' remarkable without reference to 
 their religious belief.' Later when their ' government and 
 laws ' are mentioned, these are declared to be < remarkable ' in 
 lieu of the original characterization. The papers teem with 
 illustrations of introductions. From the point of view of the 
 context into which they force their way, they are often really 
 modifications. Certain ideas lie in the foreground of the gen- 
 eral thought which the subject is trying to sketch in his repro- 
 duction. These thrust themselves into any congruous situation 
 that the sketch presents, and usurp the place of those original 
 ideas the function of which they can perform. All this results 
 in greater similarities, which in later reproductions are a basis 
 for a new regrouping and consequent simplification. Thus the 
 generalizing process feeds itself. It spreads through the pas- 
 sage similarities more nearly perfect than those in the original, 
 and these serve to bring about new combinations in which the 
 remaining differences tend to disappear. In this fashion that 
 general picture which may be regarded as the final product 
 is evolved. 
 
 But it has not so much annihilated as devoured the distinc- 
 tions by the suppression of which it has been forced to the front. 
 It has fed on them, and its substance is not without trace of the 
 matter out of which it has been composed, as well as of the di- 
 recting ideas by which this has been disposed. Omissions do, 
 indeed, reduce the richness of content in the passage. Intro- 
 ductions tend to chase out differences, and to leave the promi- 
 nent ideas in possession of the field. Nevertheless, it must be 
 
70 E. N. HENDERSON. 
 
 remembered that the modifications and condensations of simpli- 
 fication for the most part squeeze out only the water, leaving 
 the essence of the thought. 
 
 b. Modifications in the Reproduction of 
 'The Dutch Homestead.' 
 
 Let us now take up one of the passages and run through 
 some of the characteristic changes that different parts of it 
 undergo. ' The Dutch Homestead ' I have chosen as, per- 
 haps, offering the most suggestive material for study. Some 
 account of ways in which its specific parts came to be rendered 
 appears in the column parallel to that containing the original. 
 The numbers indicate the division into subtopics, and the com- 
 ment concerning each subtopic occupies the space opposite. 
 
 Original Text. Comment of Typical Alodifications . 
 
 1. It was one of those The formal introduction 'it was 'is 
 spacious farmhouses preserved in 37 out of 45 papers. 
 
 Probably under the influence of the 
 idea of ' lowness ' later brought 
 out, ' spacious ' is rendered 
 ' wide,' ' long,' ' rambling,' etc. 
 
 2. with high ridged but ' High ' is frequently left out because 
 lowly sloping roofs, of ' lowly.' We find all adjec- 
 tives sometimes condensed into 
 one; e. g., 'peaked,' 'steep.' 
 
 3. built in the style hand- This entire statement is often con- 
 ed down from the first densed to such a form as ' old- 
 Dutch settlers, fashioned Dutch farmhouse,' 
 
 and transferred to the beginning 
 of the passage. 
 
 4. the low projecting This context is usually combined 
 eaves forming a piaz- with that of * lowly sloping 
 za along the front roofs.' The adjective ' long ' 
 
 is applied to the piazza. I con- 
 jecture the influence of the sound 
 in ' along ' plus the thought of 
 the proportions of the house. 
 
A STUDY OF MEMORY. 
 
 71 
 
 5. capable of being 
 boarded up in bad 
 weather. 
 
 6. Under this were hung 
 flails, harness, 
 
 7. various utensils of 
 husbandry, 
 
 8. and nets for fishing 
 in the neighboring 
 river. 
 
 9. Benches were built 
 along the side for 
 summer use, 
 
 10. and a great spinning 
 wheel at one end 
 
 II. and a churn at the 
 other 
 
 12. showed the various 
 uses to which this im- 
 portant porch might be 
 devoted. 
 
 13. From the piazza one 
 might enter the hall 
 
 We find ' bad weather ' mentally 
 contrasted with ' summer,' sub- 
 topic 9, and called ' winter 
 weather ' ; or the porch is de- 
 clared to be for winter use. 
 
 The word ' hung ' occurs in yet an- 
 other context, subtopic 25. As 
 a result it is introduced wherever 
 it can be. 
 
 This is a summarizing expression 
 that rarely leaves no trace on the 
 reproduction. Its influence is 
 seen in such expressions as 
 * various articles of furniture ' 
 said to be in the hall. 
 
 This is an idea so out of harmony 
 with the general thought that it 
 is more frequently left out than 
 any other subtopic. 
 
 We find the porch called a living 
 place in summer to contrast 
 with the hall, which is the ' prin- 
 cipal place of residence.* 
 
 Spinning wheel is frequently trans- 
 ferred to the hall because natur- 
 ally connected with ' wool ready 
 to be spun,' etc. 
 
 This leaves the churn at one end of 
 the porch with nothing to balance 
 it at the other. Various articles 
 are used to fill in the gap, ' settee ' 
 being one. 
 
 We find all the articles from harness 
 to churn grouped and said to 
 indicate the occupations of the 
 people dwelling here. 
 
 The hall is often called * large ' 
 from the influence of 'spacious' 
 
72 
 
 E. N. HENDERSON. 
 
 14. which formed the cen- 
 ter of the mansion 
 
 15, and the usual place of 
 residence. 
 
 16. Here rows of resplen- 
 dent pewter 
 
 17. ranged on a long 
 dresser 
 
 18. dazzled his eyes. 
 
 19. In one corner stood a 
 huge bag of wool 
 
 20. ready to be spun. 
 
 21. In another a quantity 
 of linsey woolsey 
 
 22. just from the loom. In 
 
 23. Ears of Indian corn 
 
 and 'mansion.' When called 
 * the central hall ' the next sub- 
 topic is omitted. 
 
 Subtopics 14 and 15 are often com- 
 bined ; e. g.^ 'the spacious liv- 
 ing room.' 
 
 The word ' usual,' similar in sound 
 to ' use,' doubtless aids the gen- 
 eralization of the details of sub- 
 topics 9 and 15. 
 
 I find ' resplendent ' or some equiva- 
 lent word often transferred to 
 ' red peppers,' subtopic 26. 
 
 The piazza is ' along the front,' the 
 benches are ' along the side,' 
 and later the festoons are ' along 
 the walls.' Hence the pewter 
 is often put ' along a dresser.' 
 In one case we find two dressers 
 mentioned. 
 
 The idea of 'dazzled' is usually 
 absorbed in that of ' resplendent.' 
 
 ' Huge ' and ' bag ' are used indif- 
 ferently with either ' wool ' or 
 ' linsey woolsey.' 
 
 Omitted frequently. Possibly it is 
 absorbed into the idea of the 
 spinning wheel. 
 
 The idea of ' linsey-woolsey ' is 
 somewhat unfamiliar, but it al- 
 most holds its own with that of 
 ' wool ' in the reproductions, 
 one case this is combined with 
 subtopic 20, as ' ready to be 
 used.' 
 
 The idea of 'corn' suffers some 
 transformation frequently. It 
 becomes ' dried vegetables,' 
 
A STUDY OF MEMORY. 73 
 
 ' Other dried fruits,' because of 
 the dried apples and peaches. 
 
 24. and strings of dried Almost invariably some fruit was 
 apples and peaches mentioned although not always 
 
 apples and peaches. 
 
 25. hung in gay festoons The 'festoons' are often left out, 
 along the walls, and ' gay ' unmentioned in the 
 
 general brightness. 
 
 26. mingled with the gaud The 'red peppers' are sometimes 
 of red peppers. absorbed into the dried stuff 
 
 generally. In one case they are 
 called seed-pods, perhaps com- 
 bining with corn, as several 
 speak of the latter as ' seed 
 corn.' 
 
 I have thought that the exhibition of a typical series of re- 
 productions would have the advantage of showing such modi- 
 fications in the specific context in which they occur. Thus 
 the total effect of such changes can be better seen. In the 
 columns that follow the reproductions of one student are placed 
 side by side. Each numbered section indicates the material of 
 the reproductions on the corresponding subtopic in the passage. 
 Whatever the arrangement of the subject, the treatment of each 
 subtopic is, therefore, introduced in the order of the original 
 passage. That the transpositions might be evident, however, 
 every expression that in the subject's version did not follow di- 
 rectly after the one that in our version is put just before it is 
 introduced by a number and the sign -f- . This number indi- 
 cates the section where may be found the matter that in the 
 reproduction did actually precede it. The sign -j- and a num- 
 ber following the matter of a section indicates the section 
 where we may look for the text that in the subject's version ap- 
 peared next in order. The numbers are therefore guide-posts 
 that will serve us in redintegrating the passage as it appeared 
 in the reproduction. A number introduced parenthetically in- 
 dicates the section where the word or idea preceding it may be 
 found in the original passage. 
 
74 
 
 E. N. HENDERSON. 
 
 First Reproduction. 
 
 . It was one of those 
 spacious farmhouses 
 
 . with high ridged but 
 lowly sloping roofs 
 
 , built in the style of 
 the early Dutch set- 
 tlers, 
 
 , the drooping eaves of 
 the roof formed a 
 piazza, 
 
 5. which might be closed 
 in bad weather, 
 
 6. From the rafters hung 
 flails, harness 
 
 7. and other farm uten- 
 sils, 
 
 8. nets for fishing in the 
 neighboring river. 
 
 9. Along the house were 
 ranged benches for 
 summer use 
 
 10. and a huge ( 19) spin- 
 ning wheel at one end 
 
 11. and a churn 
 
 12. showed the various 
 uses to which the pi- 
 azza might be put. 
 
 13. From the piazza one 
 might enter the hall 
 
 14. which formed the 
 center of the house 
 
 15. where the family 
 lived. 
 
 16. Here rows of re- 
 splendent pewter 
 
 17. ranged upon dressers 
 
 18. dazzled one's eyes. 
 
 19. On either hand hung 
 (25) wool 
 
 20. ready for the wheel 
 
 21. linsey-woolsey 
 
 Second Reproduction. 
 The farmhouse +3 
 
 Third Reproduction. 
 The Dutch (3) homestead 
 
 3-I- a low rambling was a large low house with 
 structure with low +4 sloping eaves. 
 
 1+ of the early Dutch 
 settlers was + 2 
 
 2+ overhanging eaves reaching almost to the 
 under the shelter (5) ground underneath which 
 of which was a piazza was a piazza +12 
 running around the 
 house +10 
 
 9+ and hanging from 11+ harness +8 
 the piazza roof (2) 
 
 nets for fishing and sets 
 of harness (6) +12 
 
 11+ benches for the use 
 of the family in sum- 
 mer, -j-6 
 
 4+ Here was placed 
 
 +11 (cf. 19) 
 
 10+ a churn -{-9 
 
 6+ showed the varied 
 
 uses to which it was 
 
 put. 
 From the piazza one 
 
 entered 
 the central hall 
 
 the living room of the 
 family. 
 
 Here 
 
 6-f and nets for fishing in 
 the river +13 
 
 12+ Here in summer time 
 the family lived and upon 
 it were placed in various 
 positions wash benches 
 +11. 
 
 21+ a spinning wheel +25 
 
 9+ chums +6 
 4+ which served a variety 
 of usages +9. 
 
 8+ Entering the house 
 
 in the middle was a large 
 hall 
 
 the living room in winter 
 for the family and a stor- 
 age place for various (7) 
 useful articles. 
 
 Here 
 
 stood the spinning wheel were placed or hung (25) 
 
 (10), wool 
 ready for the carders 
 
 linsey-woolsey in large quantities, linsey- 
 
 woolsey + 10 
 
A STUDY OF MEMORY. 75 
 
 First Reproduction. Second Reproduction. Third Reproduction. 
 
 22. just from the loom just from the loom +24 
 4-24 
 
 23. 24+ and ears of In- 26+ and Indian corn 25+ with ears of golden 
 dian corn +26 +25 corn +26 
 
 24. 22+ rows of dried 22+ rows of fruit +26 
 apples and peaches 
 
 +23 
 
 25. 23+ shone resplendent 10+ and the walls were re- 
 
 (16) along the walls. splendent +23 
 
 26. 23+ bunches of red 24-j- of brilliant red pep- 23+ and clusters of red 
 peppers. pers +23. peppers. 
 
 Characteristic regroupings, condensations, fusions and intro- 
 ductions are here evident, as is also the generalized nature of 
 the result. The score in ideas of the third reproduction, I make 
 37 + out of a possible 88. Of the 50 -f details recorded as 
 lost, only 21 that seem to me definitely contributory to the idea 
 have left no clear trace. Doubtless, evidences of many of these 
 have escaped my observation. At any rate, it is certain that 
 the passage is reduced largely by the omission of non-essential 
 elements, and that this condensation is favored or, indeed, 
 brought about by the processes we have discussed. Let us now 
 compare the extent to which these changes prevail in tests 3, 
 4 and 5. 
 
 c. Comparison of the Generalizing Processes in Tests 
 
 3, 4 and 5. 
 
 Let us note the fate of the topics in the third reproduction. 
 Collecting data from the papers of twelve individuals who took 
 all three tests, we find that in test 3 (' The History of the 
 Jews ') an average of 22.9 per cent, of topics was lost ; in test 4 
 (' The Dutch Homestead') an average of 5.2 per cent, of topics ; 
 and in test 5 (' Stages in the Development of Human Theories') 
 of 32.1 per cent. The manifest superiority of the memory for 
 the leading ideas in test 4 arises, doubtless, partly because of 
 the simplicity and interest of the thought as contrasted with the 
 more involved and less familiar ideas in the other tests, particu- 
 larly the one on Comte. The main reason, however, is, I am 
 confident, the fact that, largely because of their greater abstract- 
 ness, tests 3 and 5 were much more readily generalized through 
 
76 
 
 E. N. HENDERSON. 
 
 regrouping, condensation and the introductions that favor such 
 changes. They are filled full of repetitions of words and ideas 
 more or less similar. The same themes are in different contexts 
 several times dealt with from slightly different points of view. 
 Such ideas, whether leading or subordinate ones, tend to re- 
 ceive one expression, and thus to disrupt the passages, thereby 
 initiating all sorts of modifications. Let us notice the nature of 
 the topics in these tests, and the way in which the loss is dis- 
 tributed among them. In the following table the subjects of 
 the various topics are summarized, and a record made of the 
 number of times each is absent in the twelve papers of the third 
 reproduction. 
 
 Topics. The History of Loss, 
 the Jews. 
 
 1 Jews remarkable apart 
 from religion. 
 
 2 Origin, and early for- 
 eign servitude. 
 
 3 Rapid increase, and 
 conquest of Palestine. 
 
 4 Settlement. Character 
 of government. 
 
 5 Conflicts with neigh- 
 bors. 
 
 6 Union, and rise to 
 prosperity. 
 
 7 Discord, and conquest 
 by Babylonians. 
 
 8 Restoration to native 
 land. 
 
 The Dutch Home- 
 stead. 
 
 Loss. 
 
 Loss. 
 
 Style and external ap- 
 pearance, o 
 The porch. o 
 
 Contents of porch 
 
 hanging up. i 
 Contents of porch on 
 
 the floor. o 
 
 Position and use of 
 
 hall. o 
 
 Pewter. 2 
 
 Wool and linsey-wool- 
 sey. I 
 Corn, dried fruit, and 
 
 Comte's Phi- 
 losophy. 
 
 Method of reaching the 
 
 law. 
 
 Its proof. 
 
 4 
 7 
 
 Law as applied to con- 
 ceptions. I 
 Statement of the succes- 
 sive conditions. i 
 Law as applied to meth- 
 ods of philosophizing. 3 
 Law as applied to sys- 
 tems of philosophy. 8 
 General relation of the 
 three stages. 3 
 
 5 peppers. 
 
 Examining the record in test 3, we notice no loss in topics 2 
 and 3, little in i, and most in 5 and 8. Do our theories afford 
 any explanation of this? The Jews are on three occasions said 
 to be conquered: in topics 2, 5 and 7. Their institutions are 
 characterized in i and 4. Their rise to prosperity appears 
 in 3 and 6. Return to Palestine is mentioned in 3 and 8. It 
 will be seen, therefore, that topics i, 2 and 3 give the important 
 themes of the passage, which the later parts develop in different 
 conditions. Hence the loss in the latter part, where these 
 themes are reiterated. Topic 5, which expresses conflict, vic- 
 tory and servitude, disintegrates easily under the influence of 
 
A STUDY OF MEMORY. 77 
 
 such Other contexts as deal respectively with slavery and con- 
 quest. Topic 8, even because of its phraseology, is easily as- 
 similated with 3. 
 
 In test 4 conditions are different. The themes stand out 
 from each other, and the main factors of the homestead are in- 
 variably given. We find transpositions in the interest of unify- 
 ing the treatment of the topics, but the topics themselves do not 
 disappear by assimilation, because they are not capable of this. 
 The loss is centered entirely upon different elements in the 
 contents of the house. Twice the pewter disappears, but not 
 without leaving vestiges of the thought of the topic here or else- 
 where. A loss in this passage is far more apt to leave a sen- 
 sible gap, that insistently demands filling, than one in either 
 tests 3 or 5. 
 
 The first two topics in test 5 concern the means of discover- 
 ing the law and the method of proving it. These, it will be 
 seen from examining the passage, are not much different. The 
 means of discovery are the methods of proof. Hence the two 
 topics frequently coalesce. The central thought of the mean- 
 ing of the passage is, however, given in topics 3 and 4, one of 
 which states the law as one of three stages, and the other char- 
 acterizes these stages. Each of these topics fails to appear but 
 once in the papers under discussion, and the idea of the three 
 stages is never absent. Into this core nearly all the rest of the 
 passage can be without much loss generalized. What went be- 
 fore indicated the necessity of the law from the very nature of 
 man and from the evidence of experience. This statement of 
 the stages as necessary ones in the history of human thought 
 embraces all this. The two subsequent topics merely reiterate 
 the fundamental thought. The distinction between methods of 
 philosophizing and systems of philosophy escaped the memories 
 of most. The last topic, bringing out the mutual relationship 
 of the three stages, is better retained because it is more clearly 
 a new point. 
 
 I do not think it too much to say that the difficulty in the 
 way of remembering abstractions is not so much the trouble we 
 have in seizing them as the ease with which they coalesce into 
 unity. The ' Logik ' of Hegel is a marvellous example of the 
 
78 E. N. HENDERSON. 
 
 way our generalizing power reconciles differences in the sphere 
 of the conceptions. To keep in mind the mass of distinctions 
 therein contained, one must embody them in concrete material 
 or rely on mechanical memory. It is such practices as these 
 that make the abstractions familiar, and force them to stand out 
 in such a way as to resist the difference-annihilating power of 
 the general meaning. 
 
 On the other hand, the detail that is least consonant with the 
 general thought, least likely to spring out of an endeavor to re- 
 instate congruous elements, is the one most commonly for- 
 gotten. In test 4, 'nets for fishing,' etc., disappears in twelve 
 out of fifteen cases in the third reproduction. And this, doubt- 
 less, because we do not think of farming and fishing together- 
 But by the way of comparison, we note that ' ready to be spun ' 
 disappears in eleven out of the fifteen instances, and very likely 
 because the introduction of the spinning wheel with the context 
 has satisfied the feeling of need for such an idea. Here we 
 have not the forgetting of the incongruous, but the absorption 
 of the similar. Again, we note that in the fifteen papers * just 
 from the loom' is missing in eleven cases, 'dazzled his eyes' 
 in eleven, ' ears of Indian corn ' in nine, and ' various utensils 
 of husbandry ' in eight. ^ We may be sure, if we think, that lin- 
 sey-woolsey is a home product and ' just from the loom,' that 
 the ' resplendent pewter would dazzle the eyes.' ' The various 
 utensils of husbandry ' are like an etc. after ' flails ' and ' har- 
 ness,' and the subsequent enumeration of other articles would 
 satisfy the demand for such an idea. The corn disappears in 
 the generalization of ' dried fruit ' that blankets the last topic. 
 On the other hand, no one of the notions of the roof, the 
 porch, the flails and harness, the spinning-wheel, the churn, the 
 hall, the pewter, the dresser, the wool, the dried fruit, or the 
 festoons is absent more than three times, many only once, and 
 roof not at all. These were ideas that did not generalize and 
 at the same time were consonant with the notion of such a farm- 
 house. 
 
 In test 3 (' The History of the Jews ') the shattering of the 
 
 ^This count is based on the presence of the idea in its original form, and 
 not on any trace of its presence. 
 
A STUDY OF MEMORY. 79 
 
 subtopics under the influence of the generalizing process is so 
 extensive as frequently to destroy the sense of many of the 
 leading movements involved. If we were to sketch the thought 
 in terms of the ideas that are represented in the majority of 
 papers, it would be : The Jews were remarkable ; they were in 
 servitude ; they conquered the inhabitants of their native land ; 
 they established a peculiar government ; they grew powerful ; 
 they were again enslaved and restored to their native land. 
 Here all subtopics have been condensed so as to leave just 
 enough to characterize leading ideas. In several cases further 
 condensation is effected by assimilating the conquest of and the 
 restoration to their native land that followed the Egyptian and 
 Babylonian captivities respectively. The special peculiarities 
 of the two contexts have yielded to the generalization that the 
 likenesses invite. Modification or omission is, therefore, their 
 fate. Often enough, however, the minor points disappear be- 
 cause they are themselves similar to ideas elsewhere introduced. 
 The increase in numbers of the Hebrews in Egypt, subtopic 7, 
 is clearly brought only in one case. Elsewhere it is evidently 
 assimilated with growth in power under a monarchy, subtopic 
 18. ' Neighboring tribes,' subtopic 16, is absent in only four 
 out of twelve cases. Generalization saved instead of destroying 
 it, for it was a good substitute either for the inhabitants of the 
 ' valleys in Palestine ' or for the ' monarchies ' on the Euphrates. 
 However, these ideas suffered in their integrity, and the topic 
 to which * neighboring tribes ' belongs disappears clearly in 
 five cases out of twelve, destroyed probably because of the em- 
 ployment elsewhere of its constituent ideas. The opportunities 
 for such disintegration in the passage under discussion are so 
 numerous that any adequate description of them must be ex- 
 ceedingly tedious. Let us dismiss the subject by recalling the 
 fact that in spite of their greater loss in topics, the third repro- 
 duction of both ' The History of the Jews ' and the passage on 
 Comte suffered a larger percentage of loss in subtopics within 
 the topics that remained than did that of ' The Dutch Home- 
 stead.' 
 
 It will also be remembered that the ranking of twelve stu- 
 dents in percentages of loss, third reproduction, corresponded 
 
So E. N. HENDERSON. 
 
 fairly well in tests 3 and 5, while that in test 4 agreed to an 
 appreciable extent with neither of the others. In the first repro- 
 ductions also the rankings in tests 3 and 5 show the greatest 
 value of r. The students who handled tests 3 and 5 well con- 
 stitute, therefore, a common group, the superiority of which does 
 not seem so manifest in test 4. How shall we explain this? 
 One can scarcely doubt that it lies in the supposition that those 
 whose training in the more abstract distinctions involved had 
 been sufficiently extensive to make these stand out sharply 
 gained an advantage in tests 3 and 5 that they did not possess 
 in test 4. Of course the factor of natural ability must be con- 
 sidered, but it is hard to suppose that natural ability accounts 
 for remembering abstractions well and more concrete matter 
 poorly. Training seems to be the only satisfactory reason for 
 the difference. Nor can we suppose that this training improved 
 the power to seize the abstractions as much as it did the memory 
 for them. In the first reproduction abilities in tests 3 and 5 cor- 
 respond only a little better with each other than they do with 
 that in test 4. But the contrast is striking in the third reproduc- 
 tion. The effect of training, I conceive, is to fix distinctions of 
 an abstract character in such a way — whether by embedding 
 them in concrete material or otherwise — as to make them resist 
 more effectively the generalizations that memory inevitably sets 
 on foot. Some further contributions on this point may be gained 
 from the papers on test i, where adults and children took part. 
 
 d. Comparison of the Generalizing Process in Adults 
 
 and Children. 
 
 It will be remembered that the memory for topics was far 
 better with S. S. students than with any class of P. S., No. 40 
 pupils.^ They showed far greater power of seizing the general 
 elements of the passage, and of retaining them, even at the 
 expense of greater percentage of loss of details within the topics 
 retained, than was suffered by the younger subjects. More 
 light is thrown on the reasons for this by a consideration of the 
 topics and subtopics retained by the various classes who took 
 part in test i. The following table offers the data in regard to 
 
 1 See pp. 62-63. 
 
A STUDY OF MEMORY. 8 1 
 
 the topics. The record is of the total loss in the third repro- 
 duction as compared with the original passage. 
 
 
 27S. s. 
 
 33 P- S. 7B 
 
 43 P. S. 6B 
 
 27 P. S. 5B 
 
 
 Contents of Topics. 
 
 students. 
 
 Pupils. 
 
 Pupils. 
 
 Pupils. 
 
 Totals 
 
 I . The king and his country, 
 
 
 
 
 
 
 
 2. His tyranny, 
 
 
 5 
 
 2 
 
 7 
 
 14 
 
 3. His reform, 
 
 
 7 
 
 4 
 
 4 
 
 15 
 
 4. Question of his favorite, 
 
 
 I 
 
 4 
 
 4 
 
 9 
 
 5. Story ; dog kills fox, 
 
 
 
 2 
 
 I 
 
 3 
 
 6. Man hits dog, 
 
 
 2 
 
 4 
 
 
 6 
 
 7. Horse treads on man, 
 
 I 
 
 I 
 
 2 
 
 3 
 
 7 
 
 8. Horse is lamed, 
 
 3 
 
 4 
 
 7 
 
 4 
 
 18 
 
 9. The moral. 
 
 
 9 
 
 12 
 
 15 
 
 36 
 
 The S. S. students lost only on topics 7 and 8. These two 
 topics frequently unite. Both horse and man are lamed. The 
 horse lames the man by treading on him and himself by stepping 
 between two stones. The ideas, therefore, combine more readily 
 than any other topics in the king's story. In topic 8 the idea of 
 * two ' is the principal one not occurring before, and we notice 
 its survival in such fusions as ' team of horses,' etc. The 
 notable thing about the P. S. No. 40 records is the great loss 
 of the moral, particularly in P. S. 5B. Frequently we find 
 that the story is concluded by a simple statement that the king 
 now became just, which satisfies the feeling of need for an 
 application of the story to the change in the king's course. This 
 statement is, however, a mere repetition of what is implied in 
 topic 4, the question of the favorite. Hence it is easily omitted. 
 The abstract moral principle is with the younger subjects 
 grasped with difficulty, and, therefore, easily generalized into 
 topics 4 and 3, as the concrete outcome of the story, which they 
 have already chronicled. Similarly topic 3, the account of the 
 change, absorbs topic 2, which describes the condition before 
 the change, and both sometimes disappear in topic 4, where a 
 reason for the change is requested. The details of the king's 
 story, on the other hand, stand apart as concrete entities not 
 easily generalized. 
 
 If we examine the record of subtopics, we find that the losses 
 follow the same rule. They occur because the intelligence and 
 experience of the subject did not find them sufficiently striking 
 and distinct to prevent the resolution of similar ideas into each 
 
82 B. N. HENDERSON. 
 
 Other. Experience develops memory, not merely by bringing 
 out the relations that are the basis of system, but because it 
 emphasizes the distinctions that without the familiarity that it 
 affords would be quickly lost in generalizations. The literal- 
 ness and love of concrete in the child does not mean inability 
 to generalize. That power is an inseparable attribute of mem- 
 ory. But it means inability to keep the generalizations from 
 absorbing each other. To him the concrete alone is proof 
 against such destruction. 
 
 e. Extent of Loss of Details by Generalizing Processes. 
 
 We have already noted ^ that the downright forgetting of an 
 important contributory thought plays a comparatively small part 
 in the condensations of the later reproductions. In the 12 papers 
 in test 2 furnished by E. H. S. 8B students, I have studied the 
 fate of each of the detailed ideas. If we count an idea present 
 when there is an evident substitute present in the reproduction, 
 we shall find the percentage of loss in the first reproduction cut 
 down from 33.4 per cent, to 18 per cent., and the loss between 
 the first and the third reproductions from 19 per cent, to 8 per 
 cent. In the third reproduction 12 out of the 64 ideas of the 
 passage are everywhere represented, 14 are missing only once, 
 and 9 only twice. Only 6 are absent in more than 50 per cent. 
 of the cases. These 6 ideas are expressed in the original 
 passage by the words ' yet,' ' moral,' ' consequent,' ' most,' 
 ' strong,' ' full.' There is quite a little loss in adjectives in the 
 latter part of the passage. An examination of these will con- 
 vince us that they multiply distinctions with little difference. 
 To one who lacks concrete illustrations of each specific pecu- 
 liarity, they must give the total impression of a multiplication of 
 words, all of which amounts to eloquence, sarcasm and com- 
 manding personal appearance. Instead of a loss of 8 per cent, 
 between the first and third reproductions, we might easily be 
 justified in saying that there is none at all save in unimportant 
 turns of expression, and in adjectives that carry to the students 
 little or no additional meaning. The 47 per cent, of loss from 
 the score of the original passage that we find in the third repro- 
 
 » P. 68. 
 
A STUDY OF MEMORY. S3 
 
 duction can, therefore, be accounted for almost entirely by gen- 
 eralization and fusion, about half of which has left clear evidence 
 of its presence in the phaseology of the papers. 
 
 f. Summary of the Discussion on the Character of the 
 Changes in the Reproductions. 
 
 What are the general results of our four leading investigations 
 on this theme ? Our study of underscored words revealed a 
 developing inaccuracy, which we attributed to an irradiation of 
 the feeling of familiarity with the thought over the new words 
 that were used to express it. At the same time, there was a 
 growing loss of confidence, indicated by less underscoring and 
 keeping pace with the greater inaccuracy. As a general ex- 
 planation of the great loss of words when the ideas remained, 
 and of confidence and accuracy in recognizing them, we sug- 
 gested the evident reason of the rearrangement of the thought- 
 The effect of this would be to render many of the original 
 words useless, to compel others to be modified in form, and 
 to break up the former order of those that remained so much as 
 to destroy one set of the mechanical associations upon which 
 word memory largely depends. The more rearrangement, the 
 greater the loss of original words, and, because of the loss of 
 recognition through sense of the serial order in the words, the 
 less the confidence of the subject, and the greater his inaccuracy. 
 When the association of the word to the idea had to be relied 
 on to effect recognition, the mere fitness of a word to express a 
 thought would often be sufficient to cause it to be underscored. 
 
 We have already discussed^ the differences between memory 
 for sentences and passages and memory for series of discon- 
 nected words. The fusion of the connected thoughts into com- 
 plex notions of the total meaning of the passages was seen to 
 constitute a * grip ' enormously more effective for holding both 
 ideas and words together than the simple sense of the situation 
 of the learning, which is the only general bond where uncon- 
 nected words or thoughts are memorized in series. This gene- 
 ral sense of meaning preserves many of the details throughout 
 the period during which one is thinking of the passage. Their 
 
 ^ See pp. 22-25. 
 
84 E. N. HENDERSON. 
 
 association is thus rendered simultaneous rather than successive. 
 In recall they develop in whatever order the generalizing proc- 
 esses by which they are attacked may dictate. The answers 
 to the questionnaires indicated that, after the subject had as- 
 sumed an attitude of attention involving suggestions of the 
 situation in which the original passage was learned, the general 
 meaning rose in his mind, and he began to develop the details ; 
 easily if they were well memorized and the mechanism of word 
 associations was in effective condition, with more difficulty, and 
 only by reasoning and experimental sentences, if this mechanism 
 failed. From the testimony of the subjects, we are therefore 
 fortified in our notion of the general meaning dominating recall, 
 or better, resolving itself into the details that constitute the re- 
 call. From the physiological point of view, we doubtless think 
 such meanings with a large section of the brain, all of which 
 remains in a disturbed condition, while its special elements 
 undergo in serial opder intenser excitements corresponding to 
 the wandering of the attention. 
 
 From our study of the memory for topics and subtopics as 
 compared with that for details, we discover that with all classes 
 of subjects percentages of loss in the larger themes is at first far 
 less than that in details within the themes that are retained. 
 When the contents of the topic dwindles to a few representative 
 details, then forgetting doubtless ceases to impair these save 
 when it sweeps away all except an inarticulate sense of the 
 larger meaning. But this disappearance of the details while 
 the topics remain is not decay of the living flesh leaving only 
 the skeleton of the thought. Rather it is the resolution of the 
 elements into condensed and generalized forms. It is the body 
 compressed into the germ from which it sprang. Regrouping, 
 with generalization of similar elements, fusion with modification, 
 omission of the unimportant and the easily understood sweep 
 away far more than the omission of the incongruous. The 
 general meaning of the passage that remains, like that which 
 was the product of the details when the passage was first read, 
 is a composite and not an abstraction. It is different from its 
 prototype largely by virtue of the rearrangements that have 
 abolished repetitions. 
 
A STUDY OF MEMORY. 85 
 
 The feebler the sense of distinctions, whether from the lapse 
 of time, the ability of the subject, or his training, the more 
 ready the mind is to substitute a repetition of some idea already 
 expressed for the thought which is felt to be lacking. We may 
 count on there being some justification, however remote, for 
 such identifying of the distinct. Subtle likenesses, not con- 
 sciously perceived, are nevertheless felt, and this feeling offers 
 a path of least resistance for the struggling thought. Evident 
 repetitions not in the original appear and form a basis for new 
 regroupings and condensations. To resist such absorption the 
 detail must be characterized by points of difference more potent 
 than the identities with the rest of the passage. Herein appears 
 the effect of training. Its object is quite as much to force to the 
 front differences as to induce generalization. Indeed, since 
 generalization is so easy and distinction so difficult, probably 
 the stress of educational effort must be spent on training acute- 
 ness in perceiving the unlike. The difficulty in generalization 
 is invariably due to inability to see the distinctions upon which 
 generalization is based. The stupid miM does not fail to gen- 
 eralize, but generalizes badly. It groups together ideas widely 
 different on the basis of superficial similarities, while the factors 
 that resist such classifications, and, indeed, compel others, re- 
 main unheeded.^ 
 
 Educational Significance of the Results of Special 
 
 Research. 
 
 It is of some value to the teacher to know that ability to learn 
 correlates to a large extent with ability to remember. Judg- 
 ments of relative ability based on either of these powers are apt 
 to be valid for both. Marks on daily work will serve to deter- 
 mine the standing that would have been attained in monthly 
 tests, provided the daily mark is based on the amount learned 
 by the student in a given time, and the monthly test is preceded 
 by no special review, in which some labor more assiduously 
 than others. School marks, however, depend very largely on 
 conduct and application. Ability is often only a secondary 
 matter. The small correlation in the case of P. S. No. 40 
 
 1 See Tarde, ' Social Laws.' 
 
86 E. N. HENDERSON. 
 
 pupils between the teacher's marks and rank in the test is evi- 
 dence of this. The memory test is, indeed, not a satisfactory- 
 test of reasoning. But it is probably as near a test of general 
 ability as any could be that appealed largely to specific powers. 
 Doubtless several tests like those submitted by me would, at 
 any rate, determine relative ability in memory. It would cer- 
 tainly be a valuable thing for teachers to know of their classes 
 just what this relative ability is, and, if these tests could be sup- 
 plemented by others, determining power in the various lines 
 along which school work proceeds, it would be possible for the 
 teacher to tell what part of the results attained in the school is 
 due to industry. If a true judgment of the pupil is desirable, a 
 clear separation of ability from perseverance is equally so. 
 
 It is also worth while for the teacher to know that the power 
 to remember connected trains of thought grows with the growth 
 in clear recognition of the factors entering into them. Not age 
 but training is probably the secret of whatever greater ability 
 the older classes posses. Their notion of the separate elements, 
 particularly the abstract ones, is clear enough to prevent the de- 
 struction of these units through the generalizing processes that 
 are inherent in the act of remembering. Education aims to 
 strengthen the sense of distinctions quite as much as the con- 
 sciousness of congruity. It is true that system is the secret of 
 memory. But system rests on the discovery of distinctions un- 
 noted by the unsystematic, which constitute threads of connec- 
 tion between details. It elevates the association between these 
 details from a mere serial association by contiguity to a simul- 
 taneous association in which similarity plays a most important 
 part. Now the kind of connection that makes all these details 
 one thought may absorb them so that in recall their separate ex- 
 istence is lost. System is the secret of remembering, but also 
 of forgetting whatever it finds inconsistent or unnecessary. 
 
 The meaning into which a group of connected ideas coal- 
 esces is often spoken of as a general idea of them. The ex- 
 pression is not inappropriate. It is a composite general idea, a 
 product of the generalization of its details, whether blind or 
 intelligent, systematic or chaotic. Such a composite idea is not 
 so different from an abstraction as we might think. When we 
 
A STUDY OF MEMORY. 87 
 
 eliminate differences and are left with identities, we have ab- 
 stractions. But the consciousness of such ideas need not be a 
 barren thing. Indeed, what makes the abstraction worth while 
 is that it comes to us entangled in a mesh of experiences which 
 it serves to systematize. The most valuable of general ideas 
 is, therefore, the concretest of abstractions in the richness of the 
 material which springs into the mind attentive to it. It is in 
 such concrete generalities that education finds its most worthy 
 results. 
 
 When we reflect upon the small fraction of what we 
 learned in school that remains within our power of distinct re- 
 call, we are prone to think the time spent there to have been 
 largely wasted ; at any rate, so far as preparation for the future 
 is concerned. But we must not forget that the impressions that 
 we retain, however vague they may seem, are a product of all 
 the thought involved in their attainment. A statement that to 
 us now might seem fully adequate to indicate the present net 
 result of the subject then studied, would not, if uttered by an- 
 other, produce in us the idea we actually have. Such an idea 
 really is a product of the experience by which it arose, and it 
 could be given to us by no other agency than that experience. 
 Upon its richness depends the vitality of the idea, its effective- 
 ness in restoring the details out of which it originally sprang. 
 
88 
 
 E. N. HENDERSON. 
 
 
 
 
 APPENDIX. 
 
 
 
 
 Summer Session Students. The King Who Became Just. 
 
 Rank. 
 
 
 
 Ideas. 
 
 
 
 Words. 
 
 
 ist Rep. 
 
 Age. ist Rep. 
 
 2d Rep. 3d Rep. 
 
 ist Rep. 
 
 ad Rep. 
 
 3d Rep. 
 
 I 
 
 Adult. 67 + 
 
 64 
 
 56 
 
 114 + 
 
 89 + 
 
 72 
 
 2 
 
 
 64 
 
 52 
 
 55 + 
 
 97 
 
 60 
 
 65 + 
 
 3 
 
 
 ' 62 
 
 55 + 
 
 56 + 
 
 95 + 
 
 73 + 
 
 77 + 
 
 4 
 
 
 58 + 
 
 57 + 
 
 56 
 
 98 + 
 
 89 + 
 
 90 
 
 5 
 
 
 58 
 
 51 + 
 
 49 + 
 
 85 + 
 
 63 + 
 
 49 + 
 
 6 
 
 
 57 
 
 53 + 
 
 47 
 
 72 + 
 
 59 
 
 49 + 
 
 7 
 
 
 56 
 
 49 + 
 
 49 
 
 78 + 
 
 57 
 
 55 
 
 « 
 
 
 54 
 
 49 + 
 
 50 + 
 
 71 
 
 60 
 
 52 
 
 9 
 
 
 51 + 
 
 46 + 
 
 40 + 
 
 82 + 
 
 60 
 
 43 
 
 lO 
 
 
 ' 51 
 
 47 
 
 45 + 
 
 68 + 
 
 54 
 
 54 
 
 II 
 
 
 51 
 
 47 
 
 42 
 
 76 
 
 58 
 
 41 
 
 12 
 
 
 49 
 
 49 + 
 
 48 + 
 
 72 + 
 
 69 + 
 
 63 
 
 13 
 
 
 49 
 
 37 
 
 35 
 
 64 + 
 
 37 
 
 3" + 
 
 14 
 
 
 45 
 
 46 + 
 
 39 
 
 71 + 
 
 64 
 
 50 
 
 15 
 
 
 45 
 
 37 
 
 35 + 
 
 50 
 
 27 + 
 
 26 
 
 i6 
 
 
 42 
 
 41 + 
 
 38 + 
 
 48 + 
 
 43 
 
 34 
 
 J7 
 
 
 ' 41 
 
 36 + 
 
 35 + 
 
 58 
 
 39 
 
 33 
 
 
 P. 
 
 S. 7B PuPir,s 
 
 The King Who Became Just. 
 
 
 Rank. 
 
 
 ] 
 
 [deas. 
 
 
 
 Words. 
 
 
 3st Rep. 
 
 Ag< 
 
 :. ist Rep. 2 
 
 d Rep. 3d Rep. i 
 
 3t Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 I 
 
 U 
 
 63+ 
 
 55 57+ 
 
 108 
 
 70 
 
 85 
 
 2 
 
 It 
 
 63 
 
 63 63 
 
 
 105+ 
 
 105 
 
 105 
 
 3 
 
 '^2 
 
 61 
 
 61+ 60 
 
 
 TOO 
 
 98 
 
 99 
 
 4 
 
 15 
 
 59 
 
 51 51+ 
 
 93+ 
 
 79+ 
 
 78+ 
 
 5 
 
 i^ 
 
 57+ 
 
 55 52+ 
 
 77 
 
 69 
 
 58 
 
 6 
 
 13 
 
 57+ 
 
 54+ 48 
 
 
 So 
 
 73+ 
 
 62+ 
 
 7 
 
 13 
 
 . 57 
 
 56+ 58 
 
 
 82+ 
 
 84+ 
 
 76 
 
 8 
 
 12 
 
 57 
 
 53+ 53+ 
 
 89+ 
 
 76+ 
 
 73 
 
 9 
 
 14 
 
 \ 56 
 
 56 41 
 
 
 91 + 
 
 77+ 
 
 47+ 
 
 ID 
 
 16 
 
 56 
 
 53+ 46 
 
 
 82+ 
 
 73+ 
 
 54+ 
 
 11 
 
 M 
 
 [ 55+ 
 
 49+ 44 
 
 
 80+ 
 
 66+ 
 
 50+ 
 
 12 
 
 i« 
 
 i 54+ 
 
 52+ 48 
 
 
 76+ 
 
 67 
 
 58 
 
 13 
 
 13 
 
 53 
 
 56+ 50 
 
 
 63 
 
 67+ 
 
 65 
 
 14 
 
 I! 
 
 5 53 
 
 53+ 40 
 
 
 69+ 
 
 64 
 
 34 
 
 15 
 
 IS 
 
 '. 52+ 
 
 45+ 42+ 
 
 64 
 
 50 
 
 44 
 
 i6 
 
 i« 
 
 5 52 
 
 44+ 39 
 
 
 75+ 
 
 60-f 
 
 42+ 
 
 17 
 
 i<i 
 
 J 50 
 
 46+ 47+ 
 
 76 
 
 71 
 
 63 
 
 i8 
 
 I/! 
 
 [ 49+ 
 
 53 52 
 
 
 68+ 
 
 64 
 
 60 
 
A STUDY OF MEMORY. 89 
 
 Rank. 
 
 
 
 Ideas. 
 
 
 
 Words. 
 
 
 st Rep. 
 
 Age. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 19 
 
 13 
 
 48+ 
 
 49+ 
 
 57+ 
 
 70 
 
 68 
 
 40+ 
 
 20 
 
 15 
 
 48+ 
 
 42 
 
 38+ 
 
 59 
 
 50+ 
 
 42 + 
 
 21 
 
 14 
 
 46+ 
 
 47 
 
 36+ 
 
 64 
 
 59 
 
 37 
 
 22 
 
 13 
 
 45 
 
 40+ 
 
 30+ 
 
 64+ 
 
 52 
 
 40 
 
 23 
 
 14 
 
 45 
 
 39+ 
 
 23+ 
 
 57 
 
 45+ 
 
 25+ 
 
 24 
 
 14 
 
 45 
 
 37 
 
 33 
 
 66 + 
 
 44 
 
 41 
 
 25 
 
 13 
 
 44+ 
 
 41 
 
 34 
 
 64+ 
 
 59+ 
 
 41 
 
 26 
 
 12 
 
 43+ 
 
 41 + 
 
 39 
 
 62+ 
 
 52+ 
 
 50+ 
 
 27 
 
 13 
 
 42+ 
 
 43+ 
 
 40+ 
 
 62+ 
 
 58+ 
 
 47+ 
 
 28 
 
 — 
 
 42 
 
 40+ 
 
 30 
 
 52 
 
 50+ 
 
 32 
 
 29 
 
 12 
 
 41 
 
 30+ 
 
 26+ 
 
 47 
 
 35 
 
 26+ 
 
 30 
 
 14 
 
 38+ 
 
 38+ 
 
 36+ 
 
 48 f 
 
 39+ 
 
 41 
 
 31 
 
 14 
 
 36 
 
 32+ 
 
 22 
 
 47+ 
 
 32 
 
 21+ 
 
 32 
 
 13 
 
 35+ 
 
 31 + 
 
 26+ 
 
 44+ 
 
 36+ 
 
 29+ 
 
 33 
 
 15 
 
 26 
 
 19+ 
 
 15 
 
 30+ 
 
 17 
 
 12 
 
 
 P. S. 
 
 6B Pupii,s. The King Who Became Just. 
 
 
 Rank. 
 
 
 
 Ideas. 
 
 
 
 Words. 
 
 
 ist Rep. 
 
 Age. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep- 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 I 
 
 II 
 
 61 
 
 61 
 
 61 
 
 82 
 
 87+ 
 
 85 
 
 2 
 
 15 
 
 59 
 
 58+ 
 
 56 
 
 99+ 
 
 95 
 
 89 
 
 3 
 
 12 
 
 58+ 
 
 60 
 
 53+ 
 
 lOT 
 
 TOO 
 
 81 + 
 
 4 
 
 12 
 
 58+ 
 
 51 + 
 
 53 
 
 71 + 
 
 53 
 
 54+ 
 
 5 
 
 12 
 
 56+ 
 
 58+ 
 
 54 
 
 89 
 
 92 
 
 78+ 
 
 6 
 
 14 
 
 56+ 
 
 58+ 
 
 53 
 
 86 
 
 84+ 
 
 76+ 
 
 7 
 
 12 
 
 54 
 
 53+ 
 
 49 
 
 83 
 
 69+ 
 
 56 
 
 8 
 
 14 
 
 54 
 
 50+ 
 
 54 
 
 78+ 
 
 68 
 
 72+ 
 
 9 
 
 15 
 
 52+ 
 
 55+ 
 
 55+ 
 
 80 
 
 89+ 
 
 90+ 
 
 10 
 
 II 
 
 52+ 
 
 50+ 
 
 48+ 
 
 69+ 
 
 68 
 
 60 
 
 II 
 
 14 
 
 51 
 
 +55 
 
 49 
 
 73+ 
 
 69+ 
 
 50 
 
 12 
 
 II 
 
 51+ 
 
 54 
 
 50 
 
 68+ 
 
 73+ 
 
 66 
 
 13 
 
 13 
 
 51+ 
 
 51 + 
 
 51 
 
 71 
 
 70+ 
 
 64+ 
 
 14 
 
 13 
 
 51+ 
 
 45 
 
 41 
 
 67+ 
 
 51+ 
 
 42 
 
 15 
 
 12 
 
 51 
 
 48+ 
 
 43 
 
 67 
 
 61 + 
 
 48 
 
 16 
 
 13 
 
 50+ 
 
 44 
 
 38 
 
 77+ 
 
 55+ 
 
 37 
 
 17 
 
 14 
 
 49 + 
 
 43 
 
 42+ 
 
 67+ 
 
 54 
 
 48 
 
 18 
 
 12 
 
 49 
 
 49 
 
 41 
 
 60+ 
 
 62+ 
 
 54 
 
 19 
 
 — 
 
 49 
 
 47+ 
 
 49 
 
 66+ 
 
 57 
 
 58+ 
 
 20 
 
 14 
 
 48+ 
 
 50+ 
 
 45 
 
 68 
 
 65+ 
 
 57+ 
 
 21 
 
 13 
 
 48 
 
 49 
 
 47 
 
 68 
 
 65 
 
 59 
 
 22 
 
 12 
 
 47+ 
 
 49 
 
 41 
 
 68 
 
 70+ 
 
 51 
 
 23 
 
 14 
 
 47+ 
 
 48+ 
 
 46 
 
 62 
 
 61 + 
 
 57+ 
 
 24 
 
 14 
 
 47+ 
 
 48+ 
 
 41 
 
 70+ 
 
 75+ 
 
 55+ 
 
 25 
 
 II 
 
 46+ 
 
 45 
 
 43 
 
 62+ 
 
 62+ 
 
 60+ 
 
 26 
 
 II 
 
 46 
 
 49 
 
 37+ 
 
 51 
 
 57 
 
 45 
 
 27 
 
 15 
 
 46 
 
 44+ 
 
 41 
 
 66 
 
 53+ 
 
 53 
 
 28 
 
 13 
 
 45+ 
 
 48 
 
 47 
 
 52+ 
 
 60+ 
 
 54+ 
 
 29 
 
 13 
 
 43+ 
 
 42+ 
 
 32+ 
 
 50+ 
 
 51 
 
 32 
 
90 E. N. HENDERSON. 
 
 Rank. 
 
 
 
 Ideas. 
 
 
 
 Words. 
 
 
 St Rep. 
 
 Age. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 30 
 
 14 
 
 43 
 
 33+ 
 
 35 
 
 55 
 
 41+ 
 
 38+ 
 
 31 
 
 14 
 
 42+ 
 
 37+ 
 
 19+ 
 
 56+ 
 
 49+ 
 
 20 
 
 32 
 
 II 
 
 42 
 
 35 + 
 
 29+ 
 
 52 
 
 48 
 
 30 
 
 33 
 
 12 
 
 42 
 
 33 
 
 32+ 
 
 54 
 
 36+ 
 
 37+ 
 
 34 
 
 13 
 
 40+ 
 
 35 
 
 29+ 
 
 53 + 
 
 45 
 
 32 
 
 35 
 
 12 
 
 40 
 
 39+ 
 
 44+ 
 
 62 
 
 61+ 
 
 61 
 
 36 
 
 12 
 
 39+ 
 
 42+ 
 
 38 
 
 50 
 
 53+ 
 
 51 
 
 37 
 
 13 
 
 39 
 
 38+ 
 
 34 
 
 52 
 
 41+ 
 
 29 
 
 38 
 
 II 
 
 37+ 
 
 36+ 
 
 35+ 
 
 43 
 
 39 
 
 35 
 
 39 
 
 14 
 
 35+ 
 
 33+ 
 
 29+ 
 
 52 
 
 41 
 
 46+ 
 
 40 
 
 13 
 
 32 
 
 25+ 
 
 24 
 
 46+ 
 
 33+ 
 
 29+ 
 
 41 
 
 12 
 
 25 
 
 27+ 
 
 27+ 
 
 38 
 
 43+ 
 
 41+ 
 
 42 
 
 12 
 
 22+ 
 
 21 + 
 
 22 
 
 35 
 
 31+ 
 
 31+ 
 
 43 
 
 14 
 
 11+ 
 
 10+ 
 
 15 
 
 21 
 
 17 
 
 22+ 
 
 P. S. 5B Pupii^. The King Who Became Just. 
 
 Rank. 
 
 
 
 Ideas. 
 
 
 
 Words. 
 
 
 1st Rep. 
 
 Age. 
 
 1st Rep. 
 
 zd Rep. 
 
 3d Rep. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 I 
 
 10 
 
 57 
 
 62 
 
 59 
 
 91 
 
 97 
 
 96 
 
 2 
 
 13 
 
 53 
 
 49+ 
 
 42+ 
 
 77 
 
 55+ 
 
 51 + 
 
 3 
 
 14 
 
 52 
 
 49+ 
 
 41 
 
 73+ 
 
 67+ 
 
 52 
 
 4 
 
 12 
 
 51 
 
 42+ 
 
 37 
 
 71 + 
 
 53+ 
 
 52 + 
 
 5 
 
 13 
 
 49 
 
 47 
 
 34 
 
 75+ 
 
 63+ 
 
 33+ 
 
 6 
 
 13 
 
 47 
 
 48+ 
 
 35+ 
 
 58 
 
 45+ 
 
 43 
 
 7 
 
 II 
 
 47 
 
 45 
 
 43 
 
 65 
 
 68 
 
 59+ 
 
 8 
 
 12 
 
 43+ 
 
 43+ 
 
 42+ 
 
 52 
 
 50+ 
 
 51 
 
 9 
 
 14 
 
 42 
 
 41 + 
 
 39 
 
 37 
 
 36+ 
 
 31 + 
 
 10 
 
 10 
 
 42 
 
 41 + 
 
 36+ 
 
 56+ 
 
 56 
 
 50+ 
 
 11 
 
 10 
 
 42 
 
 40 
 
 39 
 
 64 
 
 51+ 
 
 44+ 
 
 12 
 
 13 
 
 42 
 
 39 
 
 28 
 
 63+ 
 
 53+ 
 
 43+ 
 
 13 
 
 12 
 
 42 
 
 37+ 
 
 36 
 
 59 
 
 53 
 
 48 
 
 14 
 
 II 
 
 42 
 
 31+ 
 
 32 
 
 69+ 
 
 52 
 
 48+ 
 
 15 
 
 10 
 
 38 
 
 34+ 
 
 27+ 
 
 51 
 
 40+ 
 
 29+ 
 
 16 
 
 13 
 
 38 
 
 30 
 
 20+ 
 
 54+ 
 
 32+ 
 
 19+ 
 
 17 
 
 12 
 
 37 
 
 33+ 
 
 28+ 
 
 53 
 
 53 
 
 38 
 
 18 
 
 12 
 
 32 
 
 31 + 
 
 26 
 
 44+ 
 
 45+ 
 
 38 
 
 19 
 
 II 
 
 32 
 
 29+ 
 
 33 
 
 40+ 
 
 34+ 
 
 43+ 
 
 20 
 
 13 
 
 31 
 
 34 
 
 29+ 
 
 27+ 
 
 29+ 
 
 24+ 
 
 21 
 
 — 
 
 31 
 
 30 
 
 24 
 
 44 
 
 41 + 
 
 38 
 
 22 
 
 12 
 
 31 
 
 27+ 
 
 26+ 
 
 43 
 
 41 
 
 34 
 
 23 
 
 13 
 
 29+ 
 
 28 
 
 27 
 
 34 
 
 32+ 
 
 30+ 
 
 24 
 
 13 
 
 27 
 
 24 
 
 19 
 
 41 
 
 31 
 
 23+ 
 
 25 
 
 10 
 
 23 
 
 26+ 
 
 24 
 
 38 
 
 44 
 
 35+ 
 
 26 
 
 12 
 
 22 
 
 13 
 
 14+ 
 
 27+ 
 
 13+ 
 
 14+ 
 
 27 
 
 10 
 
 21 
 
 21 + 
 
 20 
 
 30 
 
 28 
 
 25+ 
 
A STUDY OF MEMORY. 91 
 
 E. H. S. 8B Students. Cicero, 
 
 Rank. 
 
 
 Ideas. 
 
 
 
 Words. 
 
 
 1st Rep. 
 
 ist Ret 
 
 ). zd Rep 
 
 3d Rep. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 I 
 
 62 
 
 
 62+ 
 
 118+ 
 
 
 117 
 
 2 
 
 48+ 
 
 46+ 
 
 41 
 
 84+ 
 
 80 
 
 67 
 
 3 
 
 47 
 
 47+ 
 
 46+ 
 
 84+ 
 
 81 
 
 80+ 
 
 4 
 
 44 
 
 43 
 
 35 
 
 74+ 
 
 64+ 
 
 52 
 
 5 
 
 42 
 
 42 
 
 36 
 
 77 
 
 62 
 
 53 
 
 6 
 
 41+ 
 
 31 
 
 30 
 
 62 
 
 48 
 
 40+ 
 
 7 
 
 40+ 
 
 48+ 
 
 23 
 
 73 
 
 89 
 
 46 
 
 8 
 
 39+ 
 
 23 
 
 17 
 
 78 
 
 43+ 
 
 33 
 
 9 
 
 37 
 
 35+ 
 
 31 + 
 
 73 
 
 69 
 
 52+ 
 
 10 
 
 37 
 
 30 
 
 22 
 
 64+ 
 
 45+ 
 
 31 
 
 II 
 
 36+ 
 
 36+ 
 
 32 
 
 66+ 
 
 66+ 
 
 55+ 
 
 12 
 
 36 
 
 34+ 
 
 26 
 
 61 
 
 61+ 
 
 41 
 
 
 
 E. H. S. i 
 
 ^B Students. Cicero 
 
 . 
 
 
 Rank. 
 
 
 Ideas. 
 
 
 
 Words. 
 
 
 ist Rep. 
 
 1st Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 I 
 
 57 
 
 55 
 
 48 
 
 104 
 
 103 
 
 77 
 
 2 
 
 53+ 
 
 57+ 
 
 37+ 
 
 103+ 
 
 100 
 
 56 
 
 3 
 
 47 
 
 45 
 
 38+ 
 
 90+ 
 
 85 
 
 73+ 
 
 4 
 
 45 
 
 47-f- 
 
 44+ 
 
 82+ 
 
 94+ 
 
 82+ 
 
 5 
 
 45 
 
 41 
 
 28+ 
 
 81 
 
 68+ 
 
 43+ 
 
 6 
 
 43 
 
 41 
 
 21 + 
 
 69+ 
 
 57 
 
 32 
 
 7 
 
 42 
 
 40+ 
 
 33+ 
 
 74+ 
 
 60 
 
 46 
 
 8 
 
 40 
 
 34+ 
 
 32 
 
 72+ 
 
 57 
 
 44+ 
 
 9 
 
 37 
 
 25+ 
 
 26 
 
 64+ 
 
 46+ 
 
 49 
 
 10 
 
 36 
 
 30 
 
 9+ 
 
 62 
 
 54 
 
 14+ 
 
 II 
 
 34 
 
 24 
 
 12 + 
 
 57 
 
 41+ 
 
 20+ 
 
 12 
 
 33 
 
 34+ 
 
 30 
 
 55+ 
 
 57 
 
 47 
 
 13 
 
 32+ 
 
 28+ 
 
 18+ 
 
 51 
 
 42 
 
 28 
 
 14 
 
 30+ 
 
 24+ 
 
 14+ 
 
 47 
 
 ■ 42+ 
 
 20+ 
 
 15 
 
 30+ 
 
 28+ 
 
 8 
 
 47+ 
 
 34+ 
 
 9+ 
 
 16 
 
 27 
 
 24+ 
 
 II 
 
 39 
 
 37+ 
 
 13 
 
 17 
 
 26+ 
 
 18+ 
 
 15+ 
 
 42+ 
 
 22+ 
 
 23 
 
 18 
 
 24+ 
 
 25+ 
 E. H. S. I 
 
 17 
 A Students 
 
 49+ 
 i. Cicero. 
 
 39+ 
 
 23+ 
 
 Rank. 
 
 
 Ideas. 
 
 
 
 Words. 
 
 
 i9t Rep. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 ist Rep. 
 
 ad Rep. 
 
 3d Rep. 
 
 I 
 
 60 
 
 30 
 
 14 
 
 102+ 
 
 52 
 
 18 
 
 2 
 
 56- 
 
 50+ 
 
 33 
 
 103 
 
 84+ 
 
 43+ 
 
 3 
 
 48+ 
 
 48+ 
 
 44 
 
 84 
 
 79 
 
 67+ 
 
 4 
 
 47 
 
 43+ 
 
 37+ 
 
 75+ 
 
 694- 
 
 55 
 
 5 
 
 45+ 
 
 44+ 
 
 35 
 
 82 
 
 70+ 
 
 49 
 
 6 
 
 44 
 
 38 
 
 26+ 
 
 76 
 
 61 
 
 37 
 
 7 
 
 41+ 
 
 32+ 
 
 25+ 
 
 73 
 
 47+ 
 
 29 
 
 8 
 
 39 
 
 38+ 
 
 29 
 
 57 
 
 47 
 
 32+ 
 
92 E. N. HENDERSON. 
 
 Rank. 
 
 
 
 Ideas. 
 
 
 
 Words. 
 
 
 ist Rep. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 9 
 
 38 
 
 
 35 
 
 35+ 
 
 73+ 
 
 67 + 
 
 63 
 
 ■ lO 
 
 38 
 
 
 28+ 
 
 25+ 
 
 67+ 
 
 44+ 
 
 24 
 
 II 
 
 38 
 
 
 32 
 
 22 
 
 68+ 
 
 57 
 
 38+ 
 
 12 
 
 36+ 
 
 26+ 
 
 II 
 
 56 
 
 33 
 
 13+ 
 
 13 
 
 36 
 
 
 36+ 
 
 28 
 
 51 
 
 50 
 
 46 + 
 
 14 
 
 34+ 
 
 28+ 
 
 24 
 
 46+ 
 
 36+ 
 
 29+ 
 
 15 
 
 30+ 
 
 30+ 
 
 27+ 
 
 48 
 
 43+ 
 
 40+ 
 
 i6 
 
 30 
 
 
 15+ 
 
 7 
 
 57+ 
 
 26 
 
 15 
 
 17 
 
 29+ 
 
 25 
 
 14+ 
 
 49+ 
 
 37+ 
 
 23 
 
 i8 
 
 29+ 
 
 22 + 
 
 11 + 
 
 48+ 
 
 33+ 
 
 i5 
 
 19 
 
 29 
 
 
 23 + 
 
 19+ 
 
 54+ 
 
 38 
 
 27 
 
 20 
 
 28+ 
 
 29 
 
 17+ 
 
 38 
 
 42+ 
 
 25 
 
 21 
 
 24+ 
 
 21 + 
 
 11+ 
 
 40 
 
 29 
 
 15+ 
 
 22 
 
 23+ 
 
 13 
 
 13+ 
 
 34 
 
 22 
 
 22 
 
 23 
 
 19 
 
 
 21 + 
 
 12+ 
 
 29+ 
 
 29 
 
 15+ 
 
 
 
 
 C0L1.EGE Students. 
 
 Cicero. 
 
 
 
 Rank. 
 
 
 
 Ideas. 
 
 
 
 Words. 
 
 
 ist Rep 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 I 
 
 
 61 
 
 
 36+ 
 
 Ill 
 
 
 41 
 
 2 
 
 
 60 
 
 56 
 
 54 
 
 102 
 
 92 
 
 78+ 
 
 3 
 
 
 59 
 
 55 
 
 54 
 
 113 
 
 92 
 
 85 
 
 4 
 
 
 58 
 
 
 47+ 
 
 104 
 
 
 65 
 
 5 
 
 
 58 
 
 35+ 
 
 33+ 
 
 106+ 
 
 38+ 
 
 43+ 
 
 6 
 
 
 57 
 
 45 
 
 43+ 
 
 97 
 
 65 
 
 62 
 
 7 
 
 
 56+ 
 
 49+ 
 
 47 
 
 103 
 
 88 
 
 75 
 
 8 
 
 
 54+ 
 
 41 + 
 
 32 + 
 
 102 
 
 55+ 
 
 35+ 
 
 9 
 
 
 53 
 
 
 55+ 
 
 93 
 
 
 97 
 
 lO 
 
 
 49 
 
 41 
 
 41+ 
 
 93+ 
 
 65 
 
 64+ 
 
 II 
 
 
 48 
 
 35 
 
 29+ 
 
 74 
 
 39 
 
 23 
 
 12 
 
 
 46+ 
 
 39+ 
 
 37 
 
 71+ 
 
 52+ 
 
 40 
 
 13 
 
 
 45 
 
 37+ 
 
 37+ 
 
 67 
 
 46+ 
 
 39 
 
 14 
 
 
 45 
 
 33+ 
 
 32 
 
 70 
 
 41 
 
 31 + 
 
 15 
 
 
 42+ 
 
 30+ 
 
 26 
 
 69 
 
 40+ 
 
 27 
 
 i6 
 
 
 40 
 
 22+ 
 
 10 
 
 64 
 
 24+ 
 
 6 
 
 17 
 
 
 38 
 
 35 
 
 29 
 
 68+ 
 
 38 
 
 29+ 
 
 i8 
 
 
 37 
 
 24 
 
 21 
 
 67 
 
 29 
 
 17+ 
 
 19 
 
 
 34+ 
 
 20+ 
 
 9+ 
 
 50 
 
 15+ 
 
 II 
 
 20 
 
 
 34 
 
 26 
 
 24+ 
 
 49+ 
 
 34 
 
 31 
 
 21 
 
 
 30 
 
 30 
 
 24 
 
 45+ 
 
 39 
 
 26 
 
 
 COI,I,EGE 
 
 Students. The History of the Jews. 
 
 
 Rank. 
 
 Rank. 
 
 
 Ideas. 
 
 
 Words. 
 
 
 ist Rep. 
 
 Test 2. 
 
 ist Rep. 2d 
 
 Rep. 3d Rep. 
 
 ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 I 
 
 10 
 
 52+ 42+ 43+ 
 
 53+ 
 
 32 
 
 30+ 
 
 2 
 
 7 
 
 ^ 
 
 51 41+ 43+ 
 
 61 
 
 42 
 
 44 
 
 3 
 
 13 
 
 i 
 
 ^3 3 
 
 1+ 32 
 
 53 
 
 39+ 
 
 37+ 
 
 4 
 
 6 
 
 41 27+ 25 
 
 48 
 
 30 
 
 22 
 
A STUDY OF MEMORY. 93 
 
 Rank. 
 
 Rank. 
 
 
 Ideas. 
 
 
 
 
 Words. 
 
 
 ist Rep. 
 
 Test 2. I 
 
 St Rep. [ad Rep. 
 
 3d Rep. 
 
 I 
 
 st Rep. 
 
 2d Rep. 
 
 3P Rep. 
 
 5 
 
 14 
 
 39 
 
 21+ 
 
 21 
 
 
 50+ 
 
 21 + 
 
 20+ 
 
 6 
 
 19 
 
 38 
 
 17 
 
 13 
 
 
 27 
 
 15 
 
 5 
 
 7 
 
 5 
 
 36 
 
 24 
 
 29+ 
 
 
 59 
 
 32+ 
 
 34 
 
 8 
 
 8 
 
 35 
 
 34 
 
 31 + 
 
 
 49 
 
 49+ 
 
 43+ 
 
 9 
 
 20 
 
 35 
 
 27 
 
 21 + 
 
 
 43 
 
 35+ 
 
 21 + 
 
 lO 
 
 15 
 
 35 
 
 25 
 
 26+ 
 
 
 36+ 
 
 25 
 
 17+ 
 
 II 
 
 3 
 
 35 
 
 23 
 
 21 
 
 
 34 
 
 8 
 
 12 
 
 12 
 
 i8 
 
 35 
 
 23 
 
 11 + 
 
 
 39 
 
 24 + 
 
 7+ 
 
 13 
 
 12 
 
 35 
 
 22+ 
 
 19+ 
 
 
 37+ 
 
 14+ 
 
 13 
 
 14 
 
 17 
 
 33 
 
 28 
 
 23 
 
 
 36+ 
 
 35 
 
 22 
 
 15 
 
 9 
 
 32 
 
 
 25 + 
 
 
 37 
 
 
 33 
 
 i6 
 
 II 
 
 32 
 
 26+ 
 
 25 + 
 
 
 35 
 
 19 
 
 20+ 
 
 17 
 
 i6 
 
 32 
 
 27+ 
 
 4 
 
 
 46+ 
 
 26+ 
 
 2 
 
 i8 
 
 21 
 
 28 
 
 17 
 
 17 
 
 
 33 
 
 12+ 
 
 15+ 
 
 
 Graduate Students. 
 
 The History of the Jews. 
 
 
 Rank. 
 
 
 
 Ideas. 
 
 
 
 
 Words. 
 
 
 ist Rep. 
 
 ist Rep. 
 
 
 2d Rep. ■ 
 
 5d Rep. ist Rep. 
 
 2d Rep. 
 
 3d Rep. 
 
 I 
 
 67 
 
 
 48+ 
 
 48+ 
 
 115 
 
 
 64+ 
 
 53+ 
 
 2 
 
 6o 
 
 
 57+ 
 
 54+ 
 
 100 
 
 
 85 
 
 81+ 
 
 3 
 
 57+ 
 
 
 40+ 
 
 45 
 
 68 
 
 
 45 
 
 47+ 
 
 4 
 
 55 
 
 
 43+ 
 
 37 
 
 66 
 
 
 46 
 
 32 
 
 5 
 
 51+ 
 
 
 42 
 
 32+ 
 
 58 
 
 
 44 
 
 33 
 
 6 
 
 46+ 
 
 
 51 + 
 
 40+ 
 
 72 
 
 
 69+ 
 
 49+ 
 
 7 
 
 46+ 
 
 
 45+ 
 
 40 
 
 63 
 
 
 SO+ 
 
 44+ 
 
 8 
 
 46 
 
 
 33+ 
 
 32 
 
 54 
 
 
 28 
 
 27 
 
 9 
 
 44 
 
 
 40+ 
 
 27 
 
 40 
 
 
 37+ 
 
 22 
 
 lO 
 
 40 
 
 
 18+ 
 
 19+ 
 
 53 
 
 
 23 
 
 15+ 
 
 II 
 
 39+ 
 
 
 29+ 
 
 18 
 
 44+ 
 
 29 
 
 20+ 
 
 12 
 
 38 
 
 
 32 
 
 31+ 
 
 59 
 
 
 38 
 
 33 
 
 13 
 
 38 
 
 
 35 
 
 25+ 
 
 48 
 
 
 33 
 
 26+ 
 
 14 
 
 34 
 
 
 32 
 
 27 
 
 35+ 
 
 27+ 
 
 23 
 
 15 
 
 32 
 
 
 22+ 
 
 28+ 
 
 36 
 
 
 17+ 
 
 21+ 
 
 i6 
 
 26+ 
 
 
 22+ 
 
 20 
 
 40+ 
 
 13+ 
 
 17 
 
 17 
 
 25+ 
 
 
 18+ 
 
 5+ 
 
 38 
 
 
 23+ 
 
 6 
 
 i8 
 
 16 
 
 
 10+ 
 
 7 
 
 27 
 
 
 6 
 
 4 
 
 
 Graduate Students 
 
 . The Dutch Homestead. 
 
 
 Rank. 
 
 Rank. 
 
 
 Ideas. 
 
 
 
 
 Words. 
 
 ist Rep. 
 
 Test 3. 
 
 ist Rep. zd Rep, 
 
 3d Rep. 
 
 
 1st Rep. 2d Rep. 3d Rep. 
 
 I 
 
 I 
 
 77 
 
 50 
 
 37+ 
 
 
 Ill 
 
 + 60+ 
 
 40 
 
 2 
 
 2 
 
 73 
 
 50 
 
 41+ 
 
 
 131 
 
 91 
 
 60 
 
 3 
 
 12 
 
 70 
 
 59+ 
 
 50 
 
 
 96+ 70 
 
 59 
 
 4 
 
 6 
 
 64 
 
 62 
 
 56+ 
 
 
 112 
 
 104+ 
 
 84+ 
 
 5 
 
 8 
 
 62 
 
 49 
 
 40 
 
 
 98 
 
 79 
 
 53 
 
 6 
 
 5 
 
 61 
 
 45 
 
 45+ 
 
 
 102 
 
 + 62 
 
 55 
 
 7 
 
 13 
 
 60 
 
 57+ 
 
 55+ 
 
 
 98+ 81 + 
 
 78 
 
 8 
 
 II 
 
 60 
 
 54+ 
 
 47+ 
 
 
 90 
 
 73 
 
 53 
 
94 E. N. HENDERSON. 
 
 Rank 
 
 Rank. 
 
 
 Ideas. 
 
 
 
 Words. 
 
 
 ist Rep. 
 
 Test 3. 
 
 ist Rep. 
 
 ad Rep. 
 
 1st Rep. 
 
 2d Rep. 
 
 2d Rep. 
 
 3d Rep 
 
 9 
 
 i6 
 
 57 
 
 35 
 
 26+ 
 
 80 
 
 39 
 
 27 
 
 . lo 
 
 7 
 
 53 
 
 53+ 
 
 47+ 
 
 81+ 
 
 71 
 
 54+ 
 
 II 
 
 15 
 
 53 
 
 46 
 
 41 
 
 77 
 
 62+ 
 
 47+ 
 
 12 
 
 lO 
 
 49+ 
 
 37 
 
 35+ 
 
 79+ 
 
 52 
 
 45+ 
 
 13 
 
 4 
 
 47+ 
 
 41 
 
 31 
 
 77 
 
 60 
 
 39 
 
 14 
 
 14 
 
 44 
 
 40+ 
 
 29 
 
 78+ 
 
 70 
 
 47+ 
 
 15 
 
 17 
 
 32 
 
 23 
 
 25+ 
 
 51 
 
 28 
 
 35 
 
 Graduate Students. The Stages in the Development of Human 
 
 Theories. 
 
 Rank. 
 
 Rank. 
 
 
 Ideas. 
 
 
 
 Words. 
 
 
 St Rep. 
 
 Test 3- 
 
 tst Rep. 
 
 ad Rep. 
 
 3d Rep. 
 
 ist Rep. 
 
 ad Rep. 
 
 3d Rep. 
 
 I 
 
 2 
 
 73 
 
 34 
 
 49+ 
 
 128 
 
 42 
 
 79 
 
 2 
 
 5 
 
 72+ 
 
 59+ 
 
 40 
 
 lOI 
 
 79+ 
 
 37+ 
 
 3 
 
 9 
 
 68+ 
 
 41+ 
 
 12+ 
 
 90 
 
 52+ 
 
 16 
 
 4 
 
 I 
 
 53+ 
 
 204- 
 
 29 
 
 78 
 
 18 
 
 23 
 
 5 
 
 4 
 
 52 
 
 25+ 
 
 25+ 
 
 64+ 
 
 29 
 
 17+ 
 
 6 
 
 II 
 
 49 
 
 34 
 
 19+ 
 
 56+ 
 
 31 
 
 13+ 
 
 7 
 
 6 
 
 42+ 
 
 45 
 
 17 
 
 70 
 
 71 
 
 19 
 
 8 
 
 12 
 
 40+ 
 
 31+ 
 
 25+ 
 
 54 
 
 37 
 
 25 + 
 
 9 
 
 15 
 
 35 
 
 24+ 
 
 19 
 
 40+ 
 
 26 
 
 20 
 
 10 
 
 13 
 
 35 
 
 20+ 
 
 13+ 
 
 43 
 
 36+ 
 
 20 
 
 II 
 
 8 
 
 34+ 
 
 17+ 
 
 5 
 
 42 
 
 24 
 
 3 
 
 12 
 
 16 
 
 31 + 
 
 23 
 
 9 
 
 34+ 
 
 31 
 
 14 
 
 13 
 
 14 
 
 28+ 
 
 26+ 
 
 16+ 
 
 47 
 
 35+ 
 
 21 
 
 14 
 
 18 
 
 27 
 
 22+ 
 
 24+ 
 
 43 
 
 32 
 
 34 
 
 OF TH 
 
 ' f f 0,1 
 
 OF THE 
 
 '- Town 
 
VITA. 
 
 The writer was born in LaSalle County, Illinois, December 
 17, 1869; removed to California in 1880; graduated from Uni- 
 versity of California, 1890, degree Ph.B. ; teacher in Secondary 
 School, San Mateo, CaL, 1890-92; Fellow in Philosophy, 
 University of California; 1892-95; took degrees A.B. 1893, 
 A.M. 1894, University of California ; principal and teacher of 
 classics. Woodland High School, Woodland, CaL, 1895-97; 
 Instructor in Psychology and Education, California State Nor- 
 mal School, Chico, CaL, 1897-1901 ; Fellow in Education, 
 Teachers College, Columbia University, 1901-1902 ; Professor 
 of Education, Adelphi College, Brooklyn, N. Y., 1902-. 
 
AK imTXAL FINE OF 25 CENTS 
 
 WIUU INCREASE ^Of CENTS O ^^^ ^h DAY 
 DAY AND TO $1-00 ON ^H .^ ^^^^,^p£|!l 
 OVERDUE. >• .^ajrJAA= 
 
 LD21-100m-7,'39(402i 
 
VD 00712 
 
 UNIVERSITY OF CAUFORNIA LIBRARY 
 
 '^''^ItHr