^-xuCKiUjiyxjTu ^ UicL^^hrfj i-ftA-v,,: ■ 
 
 ^KjeAAj Unfi^, ^^'&^ 
 
 Glass _lZi_Jj^->_ 
 Book /i V-"iJ 
 

 CONTENTS. 
 
 Chapter 
 I 
 II 
 III 
 
 I 1. 
 
 I 3. 
 
 IV 
 
 V 
 
 VI 
 
 VII 
 
 VIII 
 
 IX 
 
 X 
 XI 
 
 I 3. 
 
 I 2. 
 I 3. 
 
 XII 
 XIII 
 XIV 
 
 XV 
 
 1 1. 
 
 I 2. 
 I 3. 
 I 4. 
 
 Paqb 
 
 Introduction 7 
 
 Historical Summary 9 
 
 Methods and Records 13 
 
 Methods and Apparatus 13 
 
 Explanation of Records and the Method of 
 
 Measurement 16 
 
 Description of Accompanying Photographs 19 
 
 The Fixation Pauses 23 
 
 The Number of Fixation Pauses and the Extent 
 
 of Page Read per Fixation 23 
 
 Motor Habits 29 
 
 Refixations 30 
 
 Is There Perception During Eye Movement? 42 
 
 The Distribution of Attention in Perception 48 
 
 Visual Processes Involved in Recollecting 57 
 
 The Duration of the Fixation Pauses 59 
 
 Relation TO THE Number of Pauses 59 
 
 The Length of Fixation and the Reaction Time 
 
 OF the Eye 61 
 
 The Span of Attention and the Duration of 
 
 Pauses 62 
 
 Experimental Isolation of the Reading Pauses... 65 
 
 The Number Span of Attention 67 
 
 The Location of the Fixation Pauses 74 
 
 The Location of Pauses in Ordinary Reading 74 
 
 Proof Reading, Effect of Articulation, etc 91 
 
 Children's Reading 9(3 
 
 The Effect of Long and Short Words 98 
 
 The Length of Text Lines and Motor Habits 99 
 
 Rapidity of Reading 116 
 
 The Fatigue op the Eyes in Reading 123 
 
 Summary and Conclusions 128 
 
 Appendix 133 
 
 Acknowledgements 135 
 
6 CONTENTS 
 
 INDEX OF PLATES 
 
 Plate Page 
 
 I Line Drawing OF Apparatus Opposite Page 13 
 
 II The Reading Movements of Seven Subjects Opposite Page 17 
 
 III Refixations Opposite Page 35 
 
 IV Tachistoscopic Reading, etc Opposite Page 55 
 
 V Reading of Numerals, Children's Reading, 
 
 ETC Opposite Page 73 
 
 INDEX OF CHARTS OF THE LOCATIONS OF FIXATION PAUSES 
 
 Chaht Page 
 
 I — II 76— 80 
 
 III — VI 81— 82 
 
 VII — XI 87— 89 
 
 XII — XVI 92— 95 
 
 XVII 97 
 
 XVIII — XXII 101 — 105 
 
 XXIII 119 
 
THE PSYCHOLOGY OF READING 
 
 CHAPTER I 
 Introduction 
 
 The motor tendency or actual movement of the eyes in response 
 to changes in the direction of thought, as in the imagining and recall 
 of different extents of space and locality, has been a particularly 
 good illustration of the ideomotor theory of mental action. The sig- 
 nificance, however, of the fact itself for a general study of the in- 
 tellectual processes involved has been but little ap-preciated. The 
 relation is well exemplified in the reading movements of the eyes, 
 their automatic and reflex character and their usual freedom from 
 voluntary control making the dependence particularly close and 
 worthy of study. These movements are not only subject to the in- 
 fluence of the direction of thought as words and phrases are read 
 and assimilated, but they are also directly concerned in the sensory 
 processes of perception. The peculiarly jerky and interrupted form 
 of movement, which is more exactly a series of alternate pauses and 
 movements, is due in part both to the requirements of apperception 
 and assimilation, and to the physiological limitations of perception. 
 This twofold relation of these movements to central activities, on the 
 one hand, and, on the other hand, as the necessary accessory to a 
 peripheral organ of sensation, gives them an intermediary position 
 between sensation and recognition and between thought and motor 
 expression, which is of particular interest for the cues or indices 
 which a study of them may give of some of the workings of the 
 mind. 
 
 If, for example, the attention pauses upon some word or phrase, 
 singles out some misspelling, passes hurriedly over a commonplace 
 of thought or a familiar phrase, or again lingers on the difficult 
 parts of the sentence, we shall find some indication, at least, of these 
 facts in the varying duration of the pauses and in peculiarities of 
 the movements of the eye. If, in the second place, there are among 
 the diversity of methods of printing and the ingenious devices and 
 arrangements of the printer some which are naturally better suited 
 to the physiological limitations of vision, it is to be expected that 
 differences in the ease and rapidity of eye movement may point to 
 
8 TEE PS7CH0L0GT OF BEADING 
 
 some of the unnecessary difficulties which the printer has imposed. 
 The reading movement will be found to have the characteristics 
 of habits acquired to meet artificial conditions, adjustable to the 
 varying extents of angular displacement of the eye, etc., but less 
 adaptable to some typographical methods and arrangements than to 
 others. 
 
 In the next place, individual peculiarities of different readers 
 also make a careful analysis and description of the eye movements 
 desirable. There are, for instance, differences in the span or dis- 
 tribution of attention which are correlated with motor peculiarities, 
 and dependent upon differences in the ease of acquirement of the 
 above noted motor habit of reaction. They are further illustrated 
 in motor differences between fast and very slow readers, and be- 
 tween children and adults. 
 
 The primary object of this investigation is, therefore, to determine 
 with exactness the form and character of the movement of the eye 
 in reading, and to define or plot the positions on the page which 
 correspond to the so-called fixation or reading pauses of the eye. 
 A more detailed and accurate analysis of the physiological factors 
 of the reading process is desirable for the reasons above assigned, 
 and is in many respects essential to future study of the psychology 
 of reading. Since it is mainly, if not solely, during the pauses 
 that the eye is susceptible to significant stimulation,^ a knowledge 
 of where these pauses are made relative to the page that is being 
 read, and what the conditioning factors are which determine their 
 location, should furnish a better basis for the study of the prob- 
 lems of reading than has yet been available. The questions, there- 
 fore, concerning the perception and recognition of words and sen- 
 tences, the typographical and more purely physiological factors 
 which impede or aid in reading, the basis of the individual differ- 
 ences in the rapidity of reading and other related problems will 
 be discussed in reference to the data thus obtained. 
 
 In Chapter II an historical summary is made of the parts of the 
 work of earlier investigators which bear directly upon the experi- 
 ments of this investigation. The more general literature of read- 
 ing will be discussed as occasion demands in relation to the special 
 problems raised and in connection with my own experiments. 
 
 ^ B. Erdmann and R. Dodge, Psychologische Untersuchungen iiber das 
 Lesen auf Experimenteller Grundlage. Halle. 1898. 
 
Digitized by the Internet Archive 
 in 2011 with funding from 
 The Library of Congress 
 
 http://www.archive.org/details/psychologyofreadOOdear 
 
PLATE I 
 
 ai^oo^j^ a3SiHfi^j;g 
 
CHAPTER II 
 
 Historical Summary 
 
 The introspective evidence of what the eye does in reading is 
 ordinarily very meager, although there is much here to show 
 that the differences among individuals are much greater than 
 has been suspected. Most of us are vaguely aware of the sensa- 
 tion of the return sweep of the eye from the end of one line to the 
 beginning of the next line, but have no feeling at all — even when 
 the attention is directed to the process — of the movement as the 
 line is read/ 
 
 Lamare and Javal- seem to have been the first to record the ob- 
 servation that the movement of the eye in reading was by little 
 jerks and not a continuous movement across the page. By means 
 of a sound attachment to the upper eye-lid they were able to count 
 roughly the number of these little jerks by the faint sound which 
 was made as the eye-ball moved against the eye-lid.^ Landolt 
 studied these movements by direct observation of the eye. "He 
 had his subjects read slowly and was thus able to count the jerks. 
 He found that an average of one and fifty-five hundredths words 
 were read per fixation at a distance of 30 cm. Reading of a for- 
 
 ^ Erdmann and Dodge (Op. cit., Chap. I) cite an instance which is 
 probably by no means common, where the movement was felt as a series 
 of rhythmically interrupted movements and the interesting suggestion 
 is made that the local sign theory of Lotze gives evidence of such a 
 highly developed motor consciousness. It is not improbable that our 
 motor theories of feelings of innervation have a similar individualistic 
 basis. 
 
 'Aiinales d'OcuUstique, 1879, Tome LXXXII, p. 252 (note). 
 
 "D'apres des recherches entreprises par M. Lamare dans notre labora- 
 toire I'oeil subit m^me plusieurs saccades dans le courant de chaque 
 ligne, environ une par 15 a 18 lettres de texte, il est probable que chez 
 le myope, a chacune de ces saccades de I'oeil repond une variation brusque 
 de I'accommodation." 
 
 No evidence has been adduced to support this assumption of Javal, 
 that there is a change of accommodation with every 'jump' of the eye. 
 This is certainly not the explanation or cause of the halting movement 
 of the eye, — as might be implied by this statement, because, of course, 
 movements of the widest extent can be made without 'halting.' The 
 explanation is, doubtless, that proposed by Erdmann and Dodge, i. e., the 
 lack of perception during movement. It may be further questioned 
 whether a change of accommodation occurs at every pause. The change 
 in accommodation after a large angular displacement of the eye must, of 
 course, be greater than after a small movement, and the time necessary 
 to make this adjustment in the former case would presumably be longer, 
 but this is not found to be the case. 
 
 " Compare Huey, American Journal of Psychology, XI, p. 286. 
 
10 THE PSYCHOLOGY OF READING 
 
 eign language required more fixations, as did the reading of dis- 
 connected words, of numbers, and of lists of proper nouns. "^ 
 
 These interesting observations were further extended and first 
 given significance by the extensive investigations of Erdmann and 
 Dodge.^ By a similar method of direct observation, the movements 
 of the reader's eye were recorded from the reflection in a mirror. 
 It was found that an average of five fixation pauses per line was 
 made by one subject (D) in reading a page from Locke's Es- 
 say, (width of line 8.3 cm.) and an average of eight pauses per 
 line was made by E in reading from Helmholtz's Physiol. Optik, 
 (width of line 12.2 cm.). The variations from this average were 
 small. The main conclusions from their varied experiments are 
 the following: A larger number of pauses are made in difficult 
 than in comparatively simple narration, and the simpler the text 
 the more nearly equal are the pauses and movements. Similarly 
 a longer time and more pauses are necessary to read a foreign lan- 
 guage than to read one's own language. The same is true when 
 the attention is directed to the text, as in proof-reading. The 
 average angle of movement between two successive pauses varied 
 from 3° 54'' to 5° for difficult or unfamiliar passages of the above 
 cited books, and from 4° 14' to 5° 36' for familiar passages. That 
 is, in either case^ the eye does not reach to the very edge of the 
 line, but the first and last fixations fall within the line. The first 
 fixation comes nearer to the beginning of the line than the last 
 fixation to the end of the line. This distance, as the angle of 
 movement shows, is lessened in the more difficult reading. 
 
 On the basis of the measurement by Dodge, by the Lamansky 
 method, of the angular velocity of the eye's movements, and on the 
 assumption that this velocity holds good for the movements of the 
 eye in reading, it is argued that from if to -|f (according 
 to the individual) of the reading time is consumed by fixations, 
 and that the recognition of letters is not possible in the remaining 
 short intervals of time, and must, therefore, take place only during 
 the fixation pause proper. This conclusion is of first importance 
 for the theory and further study of perception during reading. 
 The experiments and reasoning on which it is based will be pre- 
 sented in detail and discussed in a succeeding section, in connec- 
 tion with further measurements of the relative duration of fixation 
 pauses and movements proper. 
 
 ^ Quoted from Huey, Vol. XI, p. 286. The article of Landolt ("Archives 
 D'Opthalmologie, II (1891) pp. 385-395) was not accessible to me at the 
 time of writing. 
 
 ' Op. cit., Chap. I, pp. 36-76. 
 
 ' Op. cit., Chap. II. 
 
HISTORICAL SUMMARY 11 
 
 The first successful attempt at direct registration of the move- 
 ments of the eye and the first exact measurements of the number 
 and duration of the fixation pauses were made by Huey.^ 
 
 Huey recorded the movements of the eye by a direct attachment 
 of a plaster of Paris eye-cup and aluminum pointer (Delabarre 
 eye-cup) to the cornea of the reader's eye. The records were made 
 upon the smoked drum of a kymograph, upon which also a time 
 record was written. The weight and inconvenience of the eye-cup 
 was not believed by the writer and his subject to interfere seriously 
 with the conditions of normal reading. The average number of 
 fixations per line in 30 lines of the American Journal of Psy- 
 chology thus recorded was found for two subjects to be 4.8 (M. V. 
 .5) and 4.5 (M. V. .6). These readings were about the ordinary dis- 
 tance (33-35 cm.) from the eye. Doubling this distance did not 
 increase the number of fixations and decreasing the size of type 
 (from 10 to 8 point) had but little effect. The velocity of move- 
 ment for the average interfixation arc (the latter was found to be 
 from 3° 21' to 4° 14') is given as 41. 8o- to 48or. The mean varia- 
 tion is estimated at not more than 5%-6%. The time occupied in 
 the reading pauses is naturally found to be much more variable. 
 The average of one passage was 190.9o- with a mean variation of 
 48.6cr. The increase in the speed of reading is brought about 
 "solely, or at least mainly, by decreasing the number and duration 
 of the reading pauses." The average number of words read per 
 fixation in lines of different lengths varying from 121 mm. to 21 
 mm. show a slight advantage for the short line. The largest num- 
 ber of words (3.63) was read in a line of 60.5 mm. in length, but 
 there is very considerable variation and no attempt was made to 
 analyze further the cause for these differences. 
 
 In these experiments of Erdmann and Dodge and Huey as well 
 as in investigations of Cattell,^ Zeitler,^ and others, and more re- 
 cently of Messmer* and Becker^ with the fall chronometer and 
 tachistoscope, the differences between the extent of the field of ap- 
 perceptive attention in reading and the psychological possibilities 
 of perception have been emphasized. While it is possible, as has 
 been shown by these investigators, to see the whole of a line of 
 about the width of this page in from two to three separate expos- 
 ures, the number of fixations ordinarily made is at least double 
 that number. One of the more evident reasons for this difference 
 
 ^ American Journal of Psychology, Vols. X, XI and XII. 
 = Cattell, Phil. Stvdien, II and III, Mind, XI, Brain, VIII. 
 ^ Zeitler, Phil. Stiidien, XVI, 380-465. 
 
 * Archiv Messmer, fiir die Gesamte Psychologie, Bd. II, Heft. 2 u. 3. 
 " Becker, Zeitschrift fur Psychologie u. Phijsiologie der Sinnesorgane, Bd. 
 36, Heft, 1 u. 2. 
 
12 TEE PSYCHOLOGY OF READING 
 
 is that in reading the recognition and assimilation of words must 
 go on at the same time that the sensations are received, whereas 
 in the experiments with the fall exposure apparatus the former 
 processes are completed only after the end of the period of sensory 
 excitation. But, even thus, it is quite conceivable that the eye 
 might single out the two or three most advantageous positions in 
 the line for perception and remain there until the words read were 
 sufficiently well recognized and understood, and then proceed to 
 the next point of vantage. Such a procedure would give a regular 
 and uniform mode of reaction, and if we wished to teach our chil- 
 dren how best to read, the physiological difficulties which at pres- 
 ent receive little attention would perhaps be overcome, when we 
 had trained a series of motor habits which would vary only or 
 mainly according to the different lengths of line. That the actual 
 method of reading is quite different from this is no evidence that 
 it is the best or only possible way. It is certain that reading on 
 its physiological side at least is artificial to the extent that it is 
 dependent on conditions that have been developed with little direct 
 reference to the physiology or psychology of vision. What the 
 effect is of some of these latter conditions upon the more purely 
 cognitive and perceptive processes can perhaps be determined by 
 a more detailed analysis of the eye movements. The average 
 lengths, general location of the fixation pauses, and the average 
 rate of movement have been given in the works already outlined, 
 but before many of these questions regarding the reading process 
 can be answered with any certainty, it will be necessary to plot 
 the location of the separate pauses with greater accuracy, and to 
 compare the lengths of the individual pauses, — when the various 
 characteristics of the printed page and its subject matter are 
 varied under experimental conditions, — with the apperceptive char- 
 acter of the words being read. This has been attempted in the 
 following pages. The methods and apparatus and the limits of 
 accuracy of the measurements are discussed in the following section. 
 
CHAPTEE III 
 
 Methods and Eecords 
 
 § 1. Methods and Apparatus 
 
 The means which have been employed for registering the move- 
 ments of the eye is the photographic method of Prof. R. Dodge, and 
 the apparatus (Cf. Plate I) used is a modification of the Dodge 
 falling plate camera.^ Photographs are made by this method not 
 of the movement of the eyeball itself, as such photographs are not 
 
 DESCRIPTION OF APPARATUS. PLATE I. 
 
 Appaeatus. Line drawing, as seen from above. AA, head rest. B, 
 Dodge tailing plate box. C, extension bellows. P, perimeter. The pen- 
 cil of light irom the lantern passes through plates of blue glass (2), and 
 is reflected from the cornea into the lens of the camera (4) . 3. Book rest. 
 
 5. Electrical connections with magnet of spring pendulum. 
 
 B. JJODGE Falling Plate Box, as seen with side of box removed. 
 
 6. Plate i^older. It is supported by a piston rod which plays into cylin- 
 der of oil, and is released in its fall by opening a valve (a) in the bot- 
 tom of the cylinder. As fast as the oil is forced out, it is returned by 
 means of a vent pipe (8) into the top of the cylinder. Any lateral move- 
 ment of the plate is 'taken up' by a spring and guiding rails (7). The 
 sensitive plate is successively exposed as it passes by the aperture (13). 
 The fall of the plate breaks the electrical connection at (10) and releases 
 the spring pendulum from its magnet (11). 
 
 AA. Head Rest. See text for description. 
 
 Eklakged Record. The original record was enlarged about twenty 
 times. This drawing has been reduced in reproduction about ten times 
 so that it represents an actual enlargement of the original record of 
 about two times (cf. text). 
 
 sufficiently differentiated to admit of exact measurement, but of 
 the movements of the corneal reflection; that is, the eye itself is 
 not directly photographed, but the records on the negative are of 
 the reflection from the eccentric surface of the cornea of a bright 
 pencil of light which is thrown upon the eye from the reflecting 
 mirror of a heliostat, or from some other source of light. The 
 direct rays of an arc-light made parallel by means of intercepting 
 lenses were, because of their greater constancy and adaptability, 
 mainly used in these experiments.- The experiments could then 
 
 ^ For description of the original apparatus and methods see the Psy- 
 chological Revietp, Vol. VIII, pp. 147-151, and in more recent form in the 
 American Journal of Physiology, Vol. VIII, pp. 308-310. 
 
 ^ The light was passed through plates of blue glass, as that added to 
 the comfort of the reader and admitted the actinic rays. 
 
14 THE PSYCHOLOGY OF READING 
 
 be equally well made by daylight or in a darkened room. The 
 light was placed at an angle of 30° to the right of the primary 
 line of regard of the eye, and in the same horizontal plane as the 
 eye. The camera itself stood at a corresponding angle to the left 
 of the subject, and the pages or books to be read were then held 
 in a book rest at a convenient distance (about 30 cm.) directly in 
 front, so that the angular excursions of the eye to the right and 
 left were about equal. The reading matter also could be illumi- 
 nated either by daylight or artificial light as desired. The reading 
 was binocular, but the photographs were made of the right eye 
 only. 
 
 The records are continuous photographs of the horizontal move- 
 ments of the corneal reflection made upon a slowly falling photo- 
 graphic plate of great sensitiveness. 
 
 The position of the source of light must be so adjusted, and the 
 eye of the reader brought into such a position before the camera 
 that, if a horizontal plane be drawn from the centre of the light, it 
 will bisect the middle of the pupil, coincide with the line repre- 
 senting the horizontal meridian of the eye, and pass through a 
 horizontal slit cut in the box of the camera to admit light to the 
 sensitive film. It was not necessary in the investigation to deter- 
 mine the actual extent of the movement of the eye-ball itself, pro- 
 vided the corneal reflection might be considered a true representa- 
 tion of that movement. That this is the case for a limited area of 
 the normal eye Prof. Dodge has demonstrated in an as yet unpub- 
 lished article. (See report of the Proceedings of the Am. Psych. 
 Ass. at Phila., Dec, 1904. The Jour, of Phil. Psych, and Scient. 
 Methods, Vol. II, No. 3, or the Psych. Bulletin, Feb., 1905. He 
 has shown that in the normal eye the mesial eccentric surface of 
 the cornea is a true optical surface presenting no irregularities and 
 approximates very closely the spherical form. The arc of move- 
 ment in reading the ordinary length of line is a comparatively 
 small one, and it has been possible to keep the reflection at the cen- 
 ter of the cornea, for the shorter lines it does not pass beyond the 
 part of the cornea above the pupil and never reaches to the junc- 
 tion of the cornea with the sclerotic, where the inaccuracies are 
 marked. 
 
 The plate holder of the camera is the most important part of 
 the apparatus. It is, as just said, movable in the vertical plane, 
 and in its present form is capable of most finely adjusted and uni- 
 form rates of motion. "In the original apparatus, the plate holder 
 [6] , was attached to the rod of a piston playing in a cylindrical air 
 compressor [8] ; the plate could fall only as the air escaped from 
 the cylinder, and the velocity of the fall was regulated by the size of 
 
METHODS AND RECORDS 15 
 
 the opening through which the air escaped. Slight oscillations in 
 the velocity of the fall, owing to the elasticity of the air, suggested 
 the use of a fluid resistance. The loss of head which would nat- 
 urally result as the fluid was forced out of the cylinder, was ob- 
 viated by leading the vent pipe into the top of the cylinder, and 
 thus returning the oil as fast as it was forced out, maintaining a 
 constant volume in the cylinder and rendering the whole device 
 automatic and clean. "^ As there must be absolutely no lateral play 
 during the downward movement, a small wheel is attached to each 
 side of the holder, and they, in turn, are carefully fitted to guiding 
 rails. Any lateral motion is 'taken up' by a small spring to which 
 one of the wheels is fastened. The rate of fall is governed by the 
 amount of oil that is allowed to flow through the valve. 
 
 The time of the eye's movement is also recorded upon the fall- 
 ing plate by means of the oscillations of a spring pendulum which 
 is released from a magnet by a break in the electric current, caused 
 by the initial movement of the plate. The pendulum vibrates di- 
 rectly behind and at the edge of the horizontal slit of the camera, 
 and the time line is made by the alternate admission and 
 exclusion of light. The rate of vibration of the pendulum is ac- 
 curately determined by comparison with a similar record of the 
 vibrations of a tuning fork (100 per sec). This comparison needs 
 to be made but infrequently as long as the length of the pendulum 
 remains constant. 
 
 Only one other feature of the apparatus needs further descrip- 
 tion. The failure to exclude small head movements, as will be 
 noted later, caused considerable difficulty in the experiments. Sev- 
 eral forms of head rests and various schemes were tried, the most 
 satisfactory being secured by modifications of the Helmholtz head 
 rest and mouth piece.^ The head rest proper was cut out of a 
 single piece of heavy oak board and securely fastened to the large 
 firm table upon which the camera rested, and also to the camera 
 itself. In order to secure greater steadiness and firmness the sup- 
 port on which the camera was constructed and to which the head 
 rest was attached was cut from a heavy two inch plank. This was 
 in turn securely clamped to the table. On each of the uprights of 
 the head rest AA, two spools of solid brass were fastened, aa and ee. 
 The two upper and the two lower were connected by a strong brass 
 rod on which spools of brass were slipped (c) and (/). These 
 could be moved along the bars h and c. Through these brass 
 blocks holes were drilled at right angles to the cross rod b and c 
 
 ^ Op. cit. American Joxirnal of Physiology, VIII, p. 310. 
 ' Several of the improvements are due to the kind suggestions of Pro- 
 fessor Dodge. 
 
16 THE PSYCHOLOGY OF READING 
 
 and small rods inserted which could be moved up and down and 
 were secured by thumb screws. To the end of the upper rod {d d') 
 a cross piece or 'bridge' was fastened which could be adjusted to 
 fit over the bridge of the nose much like a pair of spectacles. On 
 the end of the lower rod was clamped a mouth piece with wax im- 
 pression. This was so made that it could reach to the back teeth 
 and be held securely. The bridge piece and mouth piece when 
 rightly adjusted served much like a vise. The whole mouth piece 
 and connections could finally be moved vertically in the slot ii and 
 securely clamped by the thumb screws at the desired height, i. e., 
 so as to bring the eye of the subject on line with the camera. Com- 
 plete adjustments in all directions were thus attained, together 
 with the necessary solidity and steadiness. A 'finder' in the form 
 of a spectacle loop {t) facilitated in securing the right position for 
 the eye, and was found very useful in working with children. The 
 'finder' could be removed when the proper position was secured. 
 The head 'rest' is somewhat needlessly complicated, due to home 
 construction. 
 
 § 2. Explanation of Records and the Method of Measurement 
 
 If the eye were held immovable during the exposure of the sen- 
 sitive plate, the record on the negative would be a single straight 
 vertical line reaching from top to bottom. Any horizontal move- 
 ment of the eye to the right or left would, on the other hand, be 
 represented by an oblique line. In the accompanying plates (See 
 Plates II and V), therefore, the short straight vertical lines repre- 
 sent the momentary pauses of the eye in reading, and the short 
 oblique or nearly horizontal lines the interconnecting eye move- 
 ments. The long oblique or nearly horizontal lines which separate 
 the vertical lines into groups of three or more are the return sweeps 
 of the eye from the end of one line to the beginning of the next. 
 
 The extent of the eye's movement was magnified on the nega- 
 tive about four times. This was accomplished in the apparatus, as 
 constructed for these experiments, by extending the bellows of the 
 camera several times the focal length of the lens. The angular 
 displacement of the eyes as well as the magnifying power of the 
 lens were thus utilized. The incidental disadvantage of increasing 
 the size of the pencil of rays reflected from the eye and thus secur- 
 ing a wide and clumsy line for measurement was overcome by ex- 
 cluding with the diaphragm all but a small bundle of rays crossing 
 through the center of the lens. A very fine line capable of exact 
 measurement was thus obtained, and at the same time the con- 
 trivance had the advantage of allowing larger units of measure- 
 
PLATE II 
 
 
 
 
 jB SlitMfc"'aa-^Ti 
 
METHODS AND RECORDS 
 
 17 
 
 ment than the actual extent of the movement of the corneal re- 
 flection. 
 
 The horizontal distances on the negative which represent the 
 arc of movement of the eye from one edge of the print to the 
 other, i. e., from the beginning to the end of the line, were deter- 
 
 DESCRIPTION OF PLATE II. 
 
 (The records are one-half the size of the original.) 
 
 Column I. Comparison of the vibrations of a Koenig tuning-fork 
 (vibrating at the rate of 100 beats per second, cf. Plate III, 2) with the 
 oscillations of the spring pendulum of apparatus which appear on the 
 right. 
 
 2. Subject T. Record No. I A. reading 'A license for promotion' etc. 
 
 T. 
 
 2B. 
 
 ' Newspaper column, 'St. Peters- 
 burg' etc. 
 Evolutionary thought etc. 
 
 ' Newspaper column, 'St. Peters- 
 burg' etc. 
 
 ' 'A license for promotion' etc. 
 
 ' Newspaper column, 'St. Peters- 
 burg' etc. 
 
 ' 'A license for promotion' etc. 
 
 ' Newspaper column, 'St. Peters- 
 burg' etc. 
 
 ' Newspaper column, 'St. Peters- 
 burg' etc. 
 
 ' Newspaper column, (second 
 reading). 
 
 ' 'The boatswain' etc., (long 
 line ). 
 
 ' 'And this old savage' etc., 
 (short line). 
 
 ' 'There was between them' etc., 
 (short line). 
 
 ' 'There was between them' etc., 
 (long line). 
 
 (a) The knife lines which are traced on the extreme right of print 
 aiark the oscillations of the spring pendulum. 
 
 (b) In records No^ 11, No. 7, and No. 15 secondary reflections from 
 the sclerotic which were not excluded in these instances may be detected 
 to the right of the main record. Note also the long preliminary fixation 
 in No. 10; the signal for beginning to read was delayed. 
 
 (c) The variation in the width of the lines of the records is due to 
 email differences in focusing. 
 
 (d) Parallel lines made with a sharp steel point and marking the edges 
 of the page read may be noted in nearly all the records. 
 
 (e) For further description, see the text. Each group of short verti- 
 cal lines represents the fixation pauses of a given text line, the longer 
 horizontal lines are return sweeps from the end of one line to the begin- 
 ning of the next line. 
 
 4. 
 
 
 H. 
 
 ' " 33B. 
 
 5. 
 
 
 H. 
 
 ' " 13B. 
 
 6. 
 
 
 S. 
 
 " 41B. 
 
 7. 
 
 
 S. ' 
 
 " 43 A. 
 
 8. 
 
 
 F. 
 
 ' " 49B. 
 
 9. 
 
 
 F. 
 
 " 49 A. 
 
 10. 
 
 
 M. 
 
 " 50 A. 
 
 11. 
 
 
 M. 
 
 ' " 50B. 
 
 12. 
 
 
 St. 
 
 " 63 A. 
 
 13. 
 
 
 St. 
 
 ' " 63B. 
 
 14. 
 
 
 E. 
 
 ' " 59B. 
 
 15. 
 
 
 E. 
 
 ' " 59 A. 
 
 mined for each passage before reading by having the subject fixate 
 for a moment first one and then the other of two bright needles 
 which were placed on either side of the page, even with the edges 
 
18 TEE PSYCHOLOGY OF READING 
 
 of the lines, and somewhat above them. The advantage of such 
 needle lines for accurate fixation has been demonstrated by earlier 
 experiments. The appearance on the negative of these two fixa- 
 tions is of two short parallel lines connected by a single horizontal 
 line of the eye sweep. The distance between them denotes, then, 
 the length of the line which is being read in terms of the arc of the 
 eye's movement, or more exactly, its corneal reflection. 
 
 The method of measurement was as follows : These parallel 
 lines were extended with a very fine steel point from the top to 
 the bottom of the negative. In order to find the location of any 
 fixation pause, it was only necessary to measure first its distance 
 on the plate from one of these parallel lines. Its distance from 
 the edge of the line on the page would then be this amount multi- 
 plied by the quotient of the length of the line divided by the total 
 distance between these parallel lines. For instance, the ratio be- 
 tween the length of a newspaper line and its representation on the 
 negative was about 11 to 1. Any measurement of the negative 
 to be put in terms of distances in the line needed simply, therefore, 
 to be multiplied by 11. The measurements, which were actually 
 thus made directly froui the plates of the earlier records, are 
 exact to an eighth of a millimeter. A more expeditious and work- 
 able method was employed for the larger number of records. Par- 
 allel lines of the same distance apart as the lengths of line of the 
 pages read were first drawn out on long strips of paper. These 
 were fastened upon a large wooden screen before a stereopticon 
 lantern. The original photographic plates were then placed in the 
 holder of the lantern, and the records enlarged to the size of the 
 drawings, that is, so that the parallel lines which had been draAvn 
 on the photographic plates and which marked, as above noted, the 
 edges of the lines of print, exactly coincided with the parallel lines 
 on the screen. The records, while thus enlarged by the lantern, 
 were carefully copied out in their entirety in a pencil drawing. An 
 illustration of a small section of an enlargement of one of the rec- 
 ords is shown on Plate I. The centers of the lines denoting the 
 pauses of the eye were then 'brought down' on these drawings by 
 the use of draughtsman's calipers and square, and located on a 
 line drawn perpendicular to the parallel lines. Finally tracings 
 of these latter lines were made on transparent paper. To deter- 
 mine at what places the reading pauses occur in the lines, these 
 transparent records have but to be placed over the lines of the 
 proper pages. The location of the fixation pauses may be studied 
 through the transpareuts, nr may, finally, be transferred to the 
 printed page as is done in Section XI. The work of measurement 
 by this method was most laborious, although not so fatiguing upon 
 
METHODS A'ND RECORDS 18 
 
 the eye as that of direct measurement from the plate, but was per- 
 haps justified by the greater accuracy secured. The enlargement 
 of record, the bringing down of the drawing, and transcription to 
 transparent paper, and finally to the page of the fixation pauses, 
 and, in the second place, the measurement and computation of the 
 duration of the pauses, as later described, demanded a minimum 
 expenditure of between five and six hours time for each record. 
 The marking of the fixation pauses on the line, save for other 
 sources of error that will be discussed immediately, is accurate 
 to the limits of a single small letter of a newspaper page. 
 
 About 150 records have been secured in the various experiments 
 described below. 
 
 § 3. Description of the Accompanying Photographs 
 
 Since the records are taken upon a falling plate, and the bot- 
 tom of the plate is thus the first part exposed as it passes by the 
 slit in the camera box, the readings are to be made from the bot- 
 tom up. Before the first line of reading there appear on some of 
 the plates the long vertical lines denoting the preliminary fixations 
 of the edges of the page. The movement of the eye is, of course, 
 from left to right. Each group of several heavy short vertical 
 iines slanting from left to right represents, to restate for purposes 
 of clearness what has been briefly described, fixations in the given 
 line, and the length of these lines is directly proportional to the 
 length of time of the fixations. 
 
 If, as stated before, the eye remained immovable during a fixa- 
 tion, the record should be a perfectly straight vertical line. It is 
 clear from an examination of the plates that this is often not the 
 case. The lines incline or bulge somewhat to the right or left 
 (more frequently to the right), and are somewhat broken by various 
 irregularities. Some of these are due to slight movements of the 
 head, which could not be entirely excluded, and in a few cases to 
 minute vibrations of the apparatus. But in by far the greater 
 number of instances, they are due to the unsteadiness of the eye 
 itself. These various factors can be pretty well differentiated. 
 The presence of head movements has in some of the later experi- 
 ments been detected by employing a secondary reflection from some 
 object affixed to the head. The reflection of light from a small 
 glass bead placed near the eye and to the right of it was first used, 
 but the most successful method has been to have the subject read 
 through spectacles of plain glass. A very small dot upon the glass, 
 which in no wise interferes with the reading, and in fact can not be 
 seen by the reader, is sufficient to cause a difference in the lumin- 
 
20 TEE PSYCHOLOGY OF BEADING 
 
 osity of the reflection which the glass throws into the lens, and any 
 movement of the head can thus be detected upon the negative. The 
 mark should be placed in the horizontal plane described above, and 
 a little to the right of the pupil. Although it is possible in this 
 way to detect the presence of head movements with some consider- 
 able accuracy, the difficulty of correction was such that, in view of 
 the fact that they do not affect the general results, no further at- 
 tempt has been made to eliminate them from the measurements. 
 The divergences from the vertical which are caused by the pecu- 
 liarities of the eye itself appear to be of two kinds, first, simply 
 unsteadiness and perhaps torsion of the eyeball caused, doubtless, 
 by the unequal tension of the eye muscles, and secondly, a very 
 gradual change of position. This latter, as noted by Huey, ap- 
 pears to be a sort of 'normal' nystagmus. If it were not for the 
 fact that it occurs in both directions, to the left as well as to the 
 right, it might be thought that in these experiments .it was 
 caused by the reflex effect of the source of illumination, which is 
 to the right of the primary line of regard of the eye. It is true 
 that the line of movement starts more often to the right than to 
 the left, but this is more probably due to the fact that the muscu- 
 lar tension or 'pull' of the eye as well as the movement of atten- 
 tion is also from left to right. In some cases these 'shifting' move- 
 ments cover as much of the line of the page as is covered by one 
 or two short changes of fixation. Their velocity is so much slower 
 than the usual interfixation movements that they possibly satisfy 
 the requirements of new fixations. They show that, quite apart 
 from the compensating movements, the eye seldom remains im- 
 movable in the same position while maintaining an objectively exact 
 fixation. In marking these peculiar fixations in the following 
 pages, the limits of their changing positions and the direction of 
 movement are defined. As will be noted later, their presence and 
 the extent of their movement is one of the marked individual dif- 
 ferences which are found in~ the way different persons move their 
 eyes in reading. 
 
 The semi-nystagmatic movement of the eye occurs in the fixa- 
 tions which are made at the beginning and the end of the reading 
 in order to determine the location of the edge of the page, and 
 thus introduces a constant source of error in all the measurements. 
 Its extent varies with different individuals. The amount can be 
 at least approximately determined by measuring the ability of the 
 subjects to maintain a long fixation of the same object during the 
 entire time of the usual reading. This should give the maximum 
 error.^ 
 
 ^ On account of this error, several secondary criteria were always taken 
 into account in determining where the line representing the edge of the 
 
METHODS AND RECORDS 21 
 
 The most satisfactory explanation of the phenomenon as a whole 
 seems to be that any one of the several retinal points lying close to 
 the fovea will satisfy equally well the requirements of what is ob- 
 jectively the same fixation. This conclusion seems also to be borne 
 out by an experiment with a series of successive fixations of the 
 same object (a brass rod). The subject was asked to fixate ac- 
 curately, look away to the side, refixate, etc. The recurring of clear 
 double images of a similar rod placed between the one fixated and 
 the subject was taken as evidence that the same objective fixation 
 was renewed. These 'shifting' movements which occurred as be- 
 fore did not interfere with the fixation or the maintenance of the 
 double images, and thus supports the conclusion that in a single or 
 in succeeding fixations of the same object within small limits the 
 same retinal point or points do not need to be brought in line with 
 the object of fixation.^ 
 
 One further peculiarity that is common in greater or less de- 
 gree and frequency in all the records is the refixations or regres- 
 sive movements. The eye has read on too rapidly and returns to 
 reread some word, or the subject 'loses the sense,' and the eye must 
 retrace its steps. This often occurs at the beginning of the lines. 
 The eyes of most readers seldom come quite out to the edge of the 
 line, but the first and last fixations fall somewhere within the line. 
 This distance may be falsely estimated, the eye takes a position in 
 its first stop too far from the edge to perceive clearly all the letters, 
 or at least finds it more convenient to move nearer to the edge 
 before going on with the reading of the line. This is a very com- 
 mon occurrence with some readers with certain lengths of line, for 
 example subject T. This error at the beginning of the line is also 
 
 page should be drawn. If, for example, there is a 'regressive' movement 
 (see below) after the first pause, and tlie second fixation is thus made to 
 the left of the first, this can be taken as good evidence that the eye had 
 not reached the edge of the page. One or more such refixations are to 
 be found in nearly all passages. Corrective movements give similar indi- 
 cations. Then the eye makes a longer excursion in some lines than in 
 others; these may be taken to indicate, when there is any doubt, an ap- 
 proach to the maximum excursion. Corrective movements in the pre- 
 liminary fixations themselves also indicate attempts at closer fixation. 
 By employing these various accessory checks and corrections, in connec- 
 tion with the preliminary fixations made at the beginning and end of a 
 passage, it is believed that the locations of the edges of the page have 
 been determined with considerable accuracy. Finally, it is to be remem- 
 bered that the relative distance apart of any two fixations is not affected 
 by the errors of preliminary fixation. 
 
 ^ For a careful study of the inaccuracies of fixation cf. The Yale Psy- 
 chological Studies, New Series, Vol. 1, No. 1, C. N. McAllister "Fixation of 
 Points in the Visual Field," published since the above was written. See 
 report of a paper on the "Fixation Pauses of the Eye during Reading," 
 read by the writer at the Philadelphia meeting (Dec. 1904) of the Ameri- 
 can Psychological Association. Psychological Bulletin, Feb., 1905. 
 
22 THE PSYCHOLOGY OF READING 
 
 unquestionably due in part to a purely physiological cause, the 
 inaccuracy of the motor innervation of the eye muscles. They are, 
 in other words, partly 'corrective movements,'^ and will be dis- 
 cussed in that connection later. 
 
 ^ Cf. 'Retinal Local Signs,' by the writer. Psychological Review, XI, 
 p. 298 ff. 
 
CHAPTER IV 
 
 The Fixation Pauses 
 
 § 1. The Number of Fixation Pauses and the Extent of Page 
 Bead per Fixation 
 
 In the classical experiments of Prof. Cattell with the fall ex- 
 posure chronometer it was determined that ''on the average con- 
 sciousness can at one time grasp four numbers, three to four let- 
 ters, two words, or a sentence composed of four words. "^ This 
 last extent of perception gives a measure of the possibilities of the 
 range of perceptions of words, which in actual reading, as is known, 
 is but seldom attained. In the following tables three or four cases 
 may be noted where eight and even nine words were read in two 
 fixations, thus covering the whole extent of a single line of a news- 
 paper page and equaling the possible extent of word perception 
 that has been thus determined. But ordinarily the amount read 
 falls far short of this. Stated in millimeters, the total extent of 
 the reader 's range has been determined as 24.49 mm. ;- in the 
 eases just cited from the accompanying tables 28.2 mm. of reading 
 matter per fixation is the amount read. On another page (see 
 Table II) with nearly double this length (97.2 mm.) of line, four 
 fixations was the smallest number made and eleven the largest 
 nrimber. Huey found, as already stated, that he made an average 
 number of 4.8 fixations per line (M. V. 0.5). As above noted,, 
 neither doubling the distance (33-35 cm.) of the page from the 
 eye, nor decreasing the size of type made any appreciable change 
 in this number. The arc of the eye movement covered 78-85.8 % 
 of the line. Change in the length of line had the greatest effect 
 upon the number of pauses. His results are given in the follow- 
 ing table in terms of the number of words read per fixation. Ac- 
 cording to this, lines of from 60.5 mm. to 43 mm. and of 21 mm. — 
 in the latter case the whole passage could be read without horizon- 
 tal movement — give the best results as regards the number of words 
 that can be read at one pause. The results, however, as the author 
 recognizes, have little significance when considered in such detail, 
 as the effect of differences in subject matter was not considered. 
 
 1 Brain, Vol. VIII, p. 304 ff. 
 
 = Huey, American Journal of PsycJwJogy, Vol. XI, p. 300 ff. 
 
24 
 
 TEE PSYCHOLOGY OF READING 
 
 They seem to show, however, that roughly speaking there will be 
 fewer pauses in the short line than in the long line. 
 
 TJie number of words read per fixation in lines of different lengtlis 
 as measured hy Huey'^ with tlie Delaharre eye-cup and attach- 
 ment 
 
 Length of Distance Words per 
 line from eye fixation 
 
 Cosmopolitan Magazine 
 
 
 
 121 mm. 33 
 
 cm. 
 
 1.83 
 
 
 
 
 121 ' 
 
 33 
 
 u 
 
 1.50 
 
 Am. Journal of Psych., 
 
 10 
 
 point 
 
 98 ' 
 
 ' 35.5 
 
 (1 
 
 1.96 
 
 
 8 
 
 11 
 
 98 ' 
 
 1 11 
 
 11 
 
 1.91 
 
 
 10 
 
 li 
 
 98 ' 
 
 33 
 
 11 
 
 1.70 
 
 
 
 ti 
 
 60.5 ' 
 
 1 11 
 
 11 
 
 3.63 
 
 
 
 u 
 
 43 ' 
 
 1 i( 
 
 11 
 
 2.44 
 
 
 
 u 
 
 37 ' 
 
 35.5 
 
 11 
 
 1.94 
 
 
 
 11 
 
 30 ' 
 
 33 
 
 It 
 
 1.58 
 
 
 
 (( 
 
 25 ' 
 
 1 K 
 
 11 
 
 2.16 
 
 
 
 (1 
 
 21 ' 
 
 1 1( 
 
 11 
 
 2.17 
 
 
 
 11 
 
 21 ' 
 
 35.5 
 
 11 
 
 3.33 
 
 1 Am. J. of Psych., Vol. XII, p. 293. 
 
 The experiments about to be reported were made upon eleven 
 subjects, eight adults and three children. The adults, as noted 
 later, were selected for the laboratory experiments by preliminary 
 tests from a group of about thirty students. Two of the children 
 were from the Speyer school and one from the New York City 
 public schools. 
 
 In the following table a distribution is given of the number of 
 fixations per line made by five different subjects in reading the 
 same column of a newspaper, and the results of a second, third 
 and fourth reading on the part of some of the subjects. An in- 
 terval of a month elapsed between the first and second readings of 
 subject H, and a week between the second and fifth. The second, 
 third, and fourth were made in immediate succession. Similarly, 
 there is a month's time between the first and the last two (second 
 and third) readings of subject T. The readings of the other sub- 
 jects were made on the same day. In Table II the number of 
 fixation pauses in the first readings of four subjects of eight lines 
 of a passage of the same size type and of about equal grade of 
 subject matter, but of nearly twice the width of line (96.5 mm.), 
 is given. 
 
THE FIXATION PAUSES 
 
 25 
 
 •<ti »0 CO lO CD CO lO 
 
 CO 
 
 CO 
 
 1 
 
 II 
 
 p 
 
 Q 
 
 <! 
 
 < 
 
 ^ 
 
 
 II 
 
 li 
 
 5 
 
 > 
 
 < 
 
 aDiOQ0l--l>-»O»O'<*iCD 
 
 ~. CO Oi t^ t-- CD CO 
 
 t~ T-l 
 
 3 
 S3 
 
 
 fl 
 
 
 
 
 lO 
 
 
 
 
 ■O 
 
 QD 
 
 (N 
 
 -4^ 
 
 
 u 
 
 
 a> 
 
 
 
 
 
 
 3 
 
 
 X 
 
 CD 
 
 
 
 ■^ CO ■* CO 'ti « ec 
 
 CD OS 
 lO LO »0 00 -^ "3 »0 
 
 eo' cq 
 
 lO 
 
 1—1 
 
 CO 
 
 eo 
 
 lO 
 
 1— 1 
 
 CO 
 
 ^ 
 
 Tf<TtiiO-<*i'*ieOC<l^rfTti^rt<COrf<TtiTttttCO-*iO 
 
 •<*icotoiC'^cO'^cO'^'^'«*'^'*cocoeo-*'*'eoeo 
 
 I I I 
 
 Tti-rtf-^COCOCOrf-rtiTfiTt^^TtiTtioOiO-* 
 
 Co" 1-1 
 
 
 p 
 
 1=1 
 
 
 O 
 
 at 
 
 
 'T' 
 
 03 
 
 0) 
 
 X 
 
 « 
 
 «fl^ 
 
 CO 
 
 -M 
 
 
 >s 
 
 a 
 
 ^ 
 
 W 
 
 2 
 
 
 p 
 
 « 
 
 o 
 
 0) 
 
 o 
 
 
 si 
 
 
 +3 
 
 p 
 
 
 
 t*j 
 
 o 
 
 J2 
 
 « 
 
 M 
 
 o 
 
 ©-ti 
 
 ^ 
 
 11 
 
 
 
 01 j3 
 
 bl 
 
 >l-4 
 
 s 
 
 § 
 
 TS 
 
 a< 
 
 2 
 
 o 
 
 a 
 
 -c 
 
 02 ^' 
 
 t ® o 
 
 CO-^»OiOTj<u3rti-<*-^iOCO-<*<'^'* 
 
 •^ CO CO •^ CO CD CO 'ti iO '* ■^ iO ■^ CO CO ■* 
 
 ICCOIM-^NCD -^rJlTjI-rtlCOCOCOCOCClC 
 
 •^•«*i-«t'^'<tiCD •^■^•«J<ioiO'**'*5^ 
 
 I 
 
 t^ CD CO iO CD I •* •^ 
 
 iO »0 lO »o 
 
 X>0005QpaOt^»OTlHl^OO-JDC500CD<X>aOCDODt^t^COOOCDCO 
 
 rtC^CO-^»OCOt'QOOSO'-iC^CO-<*iOCDt^C»050i-iiMCO'*0 
 0) 
 
 3 
 
 o 
 
 t^ 
 
 Tt< 
 
 I— 1 
 
 CD 
 
 o 
 
 CO 
 
 !M 
 
 •^ 
 
 ^ 
 
 CO 
 
 C<l 
 
 a 
 
 05 
 
 CO 
 
 l-H 
 
 ^:S 
 
 73 m 
 
 00 w 
 
 35 
 
 o| 
 
 
 
26 TEE PSYCHOLOGY OF READING 
 
 Table II 
 
 Tlie nnjnber of fixation pauses and the average number of words 
 per fixation in a long line {length of line 96.5 mm.). First 
 readings of four subjects 
 
 Lines 
 
 No. of w. 
 
 T 
 
 H 
 
 8 
 
 F 
 
 1 
 
 12 
 
 7 
 
 5 
 
 8 
 
 10 
 
 2 
 
 12 
 
 7 
 
 6 
 
 9 
 
 9 
 
 3 
 
 11 
 
 9 
 
 5 
 
 8 
 
 10 
 
 4 
 
 13 
 
 8 
 
 5 
 
 7 
 
 8 
 
 5 
 
 10.5 
 
 7 
 
 5 (7 
 
 )-8 
 
 10 
 
 6 
 
 10.5 
 
 10 
 
 4 
 
 8 
 
 
 7 
 
 10.5 
 
 7 
 
 5 
 
 8 
 
 
 8 
 
 11.5 
 
 6 
 
 5 
 
 8 
 
 
 Total 
 
 
 
 
 
 
 Av. 
 
 11.3 
 
 7.5 
 
 5 
 
 8 
 
 9.4 
 
 A.D. 
 
 
 1 
 
 .25 
 
 .25 
 
 .7 
 
 No. of w. 
 
 
 
 
 
 
 per fix'n. 
 
 
 1.5 
 
 2.3 
 
 1.4 
 
 l.Oi 
 
 The average number of pauses per line in fourteen -readings by 
 five subjects of the same newspaper column is 4.76 (A. D. 1.35) 
 fixations per line, or an average of 1.64 words per fixation (A. D. 
 .37). -This latter is a somewhat larger average than will hold in 
 general, because a larger weight is given in the latter to the read- 
 ing's of the more rapid readers, and the repeated readings are 
 included (see below). The averages of the successive readings 
 considered separately are given in Table III. They vary from 3.J 
 to 7.1 pauses per line, or an average of 1.9 to 1.0 words per fixa- 
 tion. All subjects read, as stated, the same pages. The individual 
 differences are large in view of the comparatively simple and com- 
 monplace character of the subject matter. 
 
 Table III 
 
 The average number of pauses (of Table I) distributed by read- 
 ings and subjects. 
 
 1st R.' 2d R. 3d R. 4th R. 5th R. 
 
 
 3 
 
 
 o 
 
 s 
 
 
 hi 
 o 
 
 3 
 
 o 
 
 S 
 
 
 o 
 
 s 
 
 
 o 
 
 a 
 
 ^ 
 
 
 ^K 
 
 (S 
 
 
 ^>< 
 
 ^ 
 
 Jh 
 
 & 
 
 
 ^M 
 
 
 
 ^H 
 
 o 
 
 S-i 
 
 
 
 
 
 *t-i ?^ 
 
 tfH 
 
 
 vi 
 
 
 <t-IS-i 
 
 «IH 
 
 
 tw -^ 
 
 
 d 
 
 Q 
 
 
 
 
 6 
 
 Q 
 
 
 o 
 d 
 
 
 
 
 d 
 
 Q 
 
 
 O 
 
 d 
 
 Q 
 
 ofe 
 
 QQ 
 
 » 
 
 < 
 
 IZICu 
 
 ^ 
 
 < 
 
 |2i|l, 
 
 ^ 
 
 < ^^ 
 
 ^ 
 
 < 
 
 IZiS 
 
 Z 
 
 <j 
 
 iZiPn 
 
 T 
 
 8.8 
 
 .6 
 
 1.9 
 
 3.7 
 
 .6 
 
 1.9 
 
 3.8 
 
 1.1 1.9 
 
 
 
 
 
 
 
 H 
 
 5.4 
 
 .85 
 
 1.2 
 
 3.9 
 
 1.5 
 
 1.9 
 
 3.4 
 
 1.5 2.1 
 
 3.6 
 
 1.2 
 
 2 
 
 4 
 
 .85 
 
 1.7 
 
 S 
 
 5.3- 
 5.6 
 
 .7 
 
 1.4- 
 1.3 
 
 5.5- 
 5.6 
 
 1.3- 
 1.2 
 
 
 
 
 
 
 
 
 
 
 F 
 
 
 
 
 5.1 
 
 1.0 
 
 1.4 
 
 
 
 
 
 
 
 
 
 M 
 
 7.1 
 
 1.1 
 
 1.0 
 
 6.1 
 
 1.2 
 
 1.1 
 
 
 
 
 
 
 
 
 
 »R= 
 
 Reading 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
THE FIXATION PAUSES 27 
 
 The average of five subjects in their first reading (Table I and 
 Table III) is 5.3 fixations per line or 1.3 words per fixation. The 
 average number of pauses made in a line of nearly double this 
 length is (cf. Table II) 7.45, and the average number of words 
 per fixation 1.6. In this case, therefore, more words are read per 
 fixation in the longer line. The variation among different indi- 
 viduals shows, however, that these averages as the results quoted 
 above of Huey have little significance by themselves. Subject H 
 read more words per fixation in the longer line and subject T fewer. 
 The question regarding the advantages and disadvantages of the 
 short and long line is one that can be properly discussed only in 
 connection with the other factors of individual apperceptive and 
 perhaps motor differences which have not yet been sufficiently 
 analyzed. The above comparison is at least sufficient to show that 
 the problem is more than one of the number of words that can be 
 perceived, or the number of movements made, and to call into 
 question the conclusions of Javal, Huey and others that more 
 words are read per fixation in the short line than in the long. The 
 problem of the length of text-line is made the subject of a succeed- 
 ing section. 
 
 Confining attention to the larger individual differences in the 
 number of pauses (Table I), these differences are at least roughly 
 correlated with the rate of reading, the rapid readers making 
 fewer fixations. As (except in first reading of H) the rate of fall 
 of the plate is nearly the same for all subjects, these differences 
 in the speed of reading can be approximately estimated from a 
 comparison of the number of lines read in each case (see Table I). 
 An examination of the above table also shows that, while there is 
 a tendency (compare subjects H and M) for a second, third or 
 more readings of the same passage to decrease somewhat the num- 
 ber of pauses, the effect is very variable. Subject T, for instance, 
 reads on the average about the same number of words per fixation 
 each time. When, however, a large number of readings is made 
 in immediate succession, there is a decided decrease in the number 
 (and length) of the pauses. In general the effect of repeated read- 
 ings is to give greater preciseness and accuracy of movement. In 
 the first and second readings it is in the case of some subjects 
 very difficult to decide at times how many pauses have been made, 
 because the movements and pauses of the eye are run together by 
 the above noted unsteadiness of the eye during fixation. After 
 the passage has been read several times the alternation of fixation 
 and movement becomes very regular and clearly defined. It would 
 seem that the tension in which the eye muscles must continually be, 
 when about to react to the more or less indefinite word stimuli on 
 
28 THE PSYCHOLOGY OF READING 
 
 the periphery, is much relaxed after the successive readings have 
 made the eye familiar with the peripheral stimulation. In Table 
 IV the first and fifteenth readings of the same passage by subject 
 S are compared as regards the number of pauses. The passage 
 was learned by heart before the fifteenth reading was reached, 
 and the number of fixations represent, then, the minimum needed 
 to bring the passage into view, or possibly simply to 'keep time' 
 with the visual auditori-motor processes of repetition. 
 
 Table IV 
 
 Comparison of the number of fixation pauses in the first and fif- 
 teenth readings of the same page 
 
 S|tJBJECT S 
 
 Line 1 2 3 4 5 6 7 Av. A. D. Av. no. of w. 
 
 IstR. 5 5 5-6 8-9 4 5 5 5.4 .9 1.3 
 
 15th R. 4 3 4 3 4 3 3 3.4 .5 2.1 
 
 As has just been noted, there is a wide difference in the form 
 of the eye movements between individuals. This can be best seen 
 from a comparison of the negatives themselves (see Plate II). 
 The differences can not be shown simply by the comparative num- 
 ber or duration of their fixation pauses and connecting move- 
 ments. The movements of subject T are not only rapid, and the 
 average fixations short, but they are in comparison with subject H, 
 also a fast reader,^ singularly precise. The fixations also are ex- 
 act, and the eye comes to a distinct stop, and between each pause 
 there is a clearly defined horizontal movement. The case is quite 
 different with subjects H and S. It is often difficult to tell where 
 the movement ends and the fixation begins ; during the fixation 
 the eye is seldom at rest, but is moving gradually backwards and 
 forwards. A change of position is thus gained which is often equal 
 to that of one or two of the usual connecting movements, and may 
 well aid in recognition. Two or more fixations often approximate 
 to what is introspectively thought to be the form of movements of 
 the eye, i. e., a gradual continuous change of the positions of the 
 eye along the whole line. These individual peculiarities are not 
 necessarily correlated with differences in the rate of reading. The 
 movements and fixations of subject F are quite as precise as those 
 of subject T. Subjects S, M, and D take rather middle ground 
 
 ^ Subject T was found to be the fastest reader by several different 
 tests in a group of 30 or more persons, and subject H stands near tlie 
 top of the group in rate of reading. The differences of the first and last 
 of the group in the several tests were as one to three. Subjects S and 
 F stand very near the end of the group and in the order named. 
 
THE FIXATION PAUSES 29 
 
 between the extremes. The accompanying photographs and the 
 charts of fixations, which follow later, represent these differences 
 better. 
 
 § 2. Motor Habits 
 
 An examination of the distribution of the fixation pauses in 
 Table I shows a tendency to the formation of 'short-lived' motor 
 habits. For example, in the last part of the third reading of H 
 there is a series of five lines which were read with three fixation 
 pauses each, and this is preceded by a group of four lines of four 
 fixations each. Judging from a comparison of earlier and later 
 readings of the same lines this is a purely artificial arrangement 
 of movements which the eye holds to quite arbitrarily. We shall 
 recur to tliis fact again later in discussing the causes which condi- 
 tion the positions of fixations in the line. A similar tendency is 
 to be noted in all the readings of subject T ; in his case series of 
 fours predominate, broken generally by a short series of threes, 
 and an occasional fixation of five. While such a rhythmical group- 
 ing of movements is perhaps less often secured at the very begin- 
 ning of a passage, and upon its first reading, it seems almost a 
 matter of chance whether this is the case or not. For example, 
 the first reading of subject H shows comparatively little tendency 
 to uniformity, whereas in the second reading, more than a month 
 later, the reading is performed almost wholly and from the start 
 in groups of four fixations per line. The initial series is broken 
 by the infreciuent case of the reading of two lines by one hori- 
 zontal movement, in this case of six fixations.^ The very first 
 reading of another newspaper column (not given in the tables) 
 by this subject shows a parallel case of motor habit formed from 
 the first line. After such a rhythmical series is well under way, 
 it would seem reasonable to expect that there must be some clearly 
 marked change in the character or difficulty of the succeeding 
 lines before the motor habit is altered or broken. If, in short, 
 the difficulties of word recognition, the sequence of words and 
 thought and other such apperceptive conditions remain similar, 
 there would appear to be no reason why the motor habit should not 
 persist. An examination of the table is sufficient to show, how- 
 ever, that these changes may go on quite apart from any relation 
 to the nature of the subject matter. This is best shown by the 
 fact that there is absolutely no uniformity in this respect between 
 the different subjects or the successive readings by the same sub- 
 
 ^ As the apparatus records only by chance and most imperfectly the 
 vertical movement of the eye it is not possible to say whether or not the 
 eye alternated between the two lines, but this is, doubtless, the case. 
 
30 THE PSYCHOLOGY OF READING 
 
 ject. In the third reading of T, to cite one instance, the thirteenth 
 line is the last of a group of five lines that are read in four fixa- 
 tions. In the next reading, which immediately followed, although 
 there is a similar series of five lines with four fixations preceding 
 and a group of four lines of four fixations each following, this 
 thirteenth line is now read with three fixations. There is no ap- 
 parent reason why this long rhythmical series should have been 
 broken in the second reading and not in the first. Because of the 
 instability or lack of regularity in these rhythmical series, I have 
 named them for convenience of reference the 'short lived' motor 
 habits of the eye in reading. Although the number of movements 
 and fixations remains the same in these motor series, it should, how- 
 ever, be noted that there is a wide difference in the relative distances 
 apart of the fixations within such groups. An addition or sub- 
 traction of one fixation may not, therefore, alter the general char- 
 acter of the mcvement, as it woukl, were the intertixation move- 
 ments always of the same length. 
 
 The ease of the formation of motor habits seems to be one of 
 the characteristics of rapid readers as contrasted with slower read- 
 ers. T and H are about twice as rapid readers as S, and nearly 
 three times as fast as F. The long motor series are found only in 
 the reading of the first two (T and H). The larger number of 
 lines read by them gives some advantage to the first readers, but 
 even in the fifteenth reading of a passage by S there was not as 
 marked a tendency to uniformity of innervation as in either of the 
 first named subjects in their first readings. This fact is further 
 il lustra ted in the ease of other subjects, and will be referred to^ 
 again. 
 
 § 3. Refixations 
 
 Other factors which aft'ect the process of perception are the 
 'regressive' and 'corrective' movements of the eye. For simplicity 
 of statement a somewhat arbitrary distinction will be made be- 
 tween these two types of movement. A 'regressive' movement is 
 caused by faulty perception or insufficient recognition and is always 
 made to the left. It occurs most frequently in the return sweep 
 of the eye from the end of the previous line. The eye for some 
 reason stops too far short of the edge of the line, and it is neces- 
 sary to make a supplementary movement towards the edge. Simi- 
 larly within the line the eye may have been carried on too far 
 for ease of recognition, and a short movement is made in the oppo- 
 site direction. The 'corrective' movements,^ on the other hand, 
 
 'Compare The PsycJiological Review, July-Sept., 1904, "Retinal Local 
 Signs," where these corrective movements are described by the writer. 
 
THE FIXATION PAUSES 31 
 
 are real inaccuracies in the motor innervation or functioning of 
 the eye muscles. They may occur in either direction. When they 
 occur to the left, they are not easily distinguishable from the 're- 
 gressive' movements. For example, the real reason why the eye 
 in returning from the end of one line to the beginning of the next 
 falls short may be either an error of perception, i. e., the working 
 span, so to speak, of perception or assimilation has been overesti- 
 mated, or it may simply be an inaccuracy of muscular movement, 
 i. e., an underestimation of linear magnitude. Some differentia- 
 tion between the two movements is possible. The corrective move- 
 ments as determined in the article referred to above^ are generally 
 of less extent and the intervals of pause before the corrections are 
 made are, also, very much shorter. But, again, these movements 
 may be further confused with the unsteadiness of the eye in fixa- 
 tion. When, however, there is an overestimation of the length of 
 line, it is generally to be considered as due to inaccuracy of move- 
 ment. The eye has been carried beyond the edge of the line, and a 
 short, quick movement back is made. These cases can be differ- 
 entiated with certainty. But it is doubtless true that a corrective 
 movement due to overestimation occurs just as often within the 
 line. As it would, however, simply 'correct' the first fixation in 
 the line, it is, then, as stated, not distinguishable from the un- 
 steadiness which accompanies fixation in general, if, indeed, it is 
 not the explanation of a part of this unsteadiness. Because of 
 these complications, in taking account of these 'refixations, ' only 
 those which are made to the left have been recorded." It has, 
 however, seemed advisable to distinguish between those which occur 
 at the beginning of the line, and those Avhich are found elsewhere 
 within the line. They have, therefore, been divided in the tables 
 on this basis into two classes ; first, those which are supplementary 
 to the return sweep of the eye, marked Sp. They include both the 
 positive^ corrective movements and the initial regressive movements. 
 Those marked Rg are tlie regressive movements Avithin the line. 
 The term 'refixation' is used to cover all cases. In Table V the 
 
 ^ Psychological Revicic, XI, p. 298 ff. 
 
 ^ The refixations which are made for perceptive purposes are natur- 
 ally only regressive (i. e., to the left). When a movement and pause 
 occurs to the right it marks, of course, a new fixation. 
 
 ^ In the article just referred to, these corrective movements were 
 designated as positive, which added to the primary movement in its 
 original direction, i. e , the eye had fallen short of the mark, and another 
 movement in the same direction was necessary. Those movementy in 
 which an overestimation occurred, causing a return movement in the 
 opposite direction to the initial movement were, on the other hand, des- 
 ignated 'negative.' For the reasons stated above, account is made only of 
 the 'positive' corrections. 
 
32 THE PSYCHOLOGY OF BEADING 
 
 number of refixations is arranged according to the lengths of lines 
 in which they occurred. Comparison is made of three different 
 pages. 
 
 Table V 
 
 The number of refixations in liyies of different lengths 
 
 I Page II Page III 
 
 Length of line Length of line Length of line 
 
 107 mm. 96.5 mm. 56.5 mm. 
 
 iNo. of No. of No. of No. of No. of No. of No. of No. of No. of 
 
 lines Sp Rg lines Sp Rg lines Sp Kg 
 A 
 
 ^. Subject T 16 11 11 7 74 5 
 
 " H 65 14 3 6 1 106 9 
 
 "85 4 13 3 4 60 8 3 
 
 Percentage of lines having refixations 
 
 B 8p. Rg 8p. Rg. Sp. Rg 
 
 ^. Subject T 69% 63.6% 18% 
 
 " H 21.6% 4.5% 16.6% 18% 
 
 " S 80% 23% 30.7% 11.4% 4% 
 
 Distribution of refixations in percentages of the whole number 
 according to the order of lines irrespective of their lengths 
 
 C. 1st 10-20 
 
 Line 2nd 3rd 4th 5th 6th 7th 8th 9th all other Ls. 
 
 24.07% 7.4% 5.5% 7.4% 9.2% 5.5% 12.9% 7.4% 5.5% 14.8% 
 
 ^ 'No. of lines' refers to the total number of lines read by each subject. 
 
 Under A (Table V) the number of lines read and the number 
 of supplementary and regressive movements made are arranged 
 according to subjects T, H and S. Under B, the same facts are 
 restated in the percentage of lines in which these movements oc- 
 cur; under G the refixations, etc., are redistributed accordingly as 
 they occur in the first, second, third, or fourth lines, etc. The re- 
 sults in C are stated in percentages of the whole number of cases 
 (54). As can be seen (C), the largest number are found in the 
 first line of the different passages read. This may possibly be due 
 to some peculiarity of the experiment. The eye before reading 
 made in each case a movement of the same length as the line about 
 to be read, i. e., as already described above, between the needles 
 which were used to denote the edges of the line. But there would 
 seem to be no particular reason why this return movement to the 
 beginning of the first line should not be as accurate as later, and 
 
TEE FIXATION PAUSES 33 
 
 no peculiarity was noted by the subjects. The more natural ex- 
 planation is that the eye has not yet accustomed itself to the right 
 extent of motor innervation. 
 
 My purpose in the preceding tables is first to call attention to the 
 frequency of refixation. It is also to be noted that there are some 
 differences between subjects in this respect and these differences 
 are most marked in the longer lines. It appears also true that 
 the ease and accuracy of movement and of perception, as denoted 
 by the comparative infrequency of refixations, is greater in the 
 shorter line. Subject T makes, however, a supplementary move- 
 ment so regularly in the long line, that it must be considered his 
 habitual reaction; that is, to a certain extent, instead of determin- 
 ing the length of the return sweep of the eye by the length of line 
 being read, the return sweep proper remains relatively constant, 
 whatever the length of line may be. It is then corrected or supple- 
 mented by another movement, if it is found necessary for perception. 
 This gives evidence of another interesting variation of the method of 
 performing the physiological process of reading. The inaccuracy 
 of movement and the necessity of refixation must ordinarily, ex- 
 cept possibly in the case of T, be a hindrance to the rapidity of 
 perception. That they are found somewhat more frequently in the 
 long lines may be considered one of the disadvantages of this 
 length of line. When reading aloud, regressive movements are in 
 some cases doubtless due to the eye returning to the word that is 
 being pronounced. But since they are as common in silent as in 
 oral reading, the more evident explanation is, then, that the eye has 
 run too far ahead for the thought, or some misunderstanding has 
 been met which calls for clearer perception of the words that are 
 but now reaching consciousness. The refixations, especially when 
 they occur within the line, may thus serve to indicate roughly how 
 far the perceptive processes antedate recognition and assimilation. 
 There is evidence that the corrective movements, as might be 
 expected, occur also in the vertical plane. The eye does not fol- 
 low along the line with exactness, — as, for instance, Javal thought, 
 along a line between the middle and top of the small letters. In 
 view of the inaccuracies of movement this may be considered as 
 literally a physical impossibility. A plate falling in the vertical 
 plane is evidently not at all suited to a study of discrepancies of 
 movement in that plane. The rate of movement, however, of the 
 eye upward from the horizontal is in a few cases so rapid as to 
 counterbalance and even exceed the rate of fall of the plate. An 
 
34 TEE PSYCHOLOGY OF BEADING 
 
 upward movement of the eye is reversed, on the negative, and so 
 the movement appears in the same direction as that of the record- 
 ing plate. In this way these inaccuracies in the vertical plane are 
 occasionally detected even in these records. They are deserving 
 of further study. 
 
 To test the hypothesis suggested above that the greater number 
 of refixations in the first line is due to the fact that the extent of 
 movement has not been adjusted correctly for the length of line, I 
 arranged an experiment so that the length of the return sweep of 
 the eye was varied several times during the reading. This was 
 attained by simply cutting out a page of print by lines and then 
 pasting the lines together so that the edges were indented or un- 
 even. The exact arrangement which has been set up in the same 
 size of type used may be found in the appendix. This dis- 
 tance between the lines on the original page was 2.5 mm. It has 
 been necessary to make the plate for reproduction (cf. PI. Ill) from 
 a typewritten copy in order to write in conveniently the signs 
 which denote the places where the eye pauses. The lines were 
 originally all of the same length, 88 mm. The amount of indenta- 
 tion was 26 mm. The small size of type was used purposely to 
 make it more difficult for the eye to decide just where the lines 
 began. If the apparatus were capable of registering movements 
 in the vertical plane, there is every reason to believe, as has been 
 noted, that the inaccuracies of fixation on the vertical plane would 
 be found perhaps as marked as those of the horizontal. This is 
 evident from a common experience; not infrequently we find, espe- 
 cially when the line is long, of small type, and of little interspac- 
 ing, that we have begun reading the wrong line, the eye having 
 either skipped a line or having returned to the beginning of the 
 line that has just been read. 
 
 The method of plotting the location of the pauses is described 
 later. It is sufficient at present to note that in this page a single 
 straight vertical line I denotes the place of the fixation. If the 
 eye is unsteady or moves somewhat in one direction or the other, 
 a horizontal line connected with this and drawn under the letters 
 denotes the direction and extent of the movement; i. e., the sign 
 ) = fixation and very slow movement to the right, j = move- 
 ment to the left. The fixations are numbered in the order in 
 which they are made. These numbers are put below the signs. 
 The durations of the pauses in sigma are placed above as index 
 numbers to the signs. These movements during fixation are dif- 
 
-^. 
 
 PLATE III 
 
 (REFIXATIONS) 
 
 
 = E 
 
 
 3 
 
 '-)■ 
 
 >- 
 
 
 
 'rsi 
 
 H 
 
 -a 
 
 5 g 
 
 E 
 
 n qL 
 
 ;For Description and Explanation see pp. 34-35 of Text) 
 
THE FIXATION PAUSES 35 
 
 ferentiated, as above noted, from the interfixation movements 
 proper by less velocity. Such a combination of signs as appears 
 at the beginning of the fifth line somewhat 
 incompletely describes the unsteadiness and the 
 . r « several directions of movements during fixa- 
 
 tion. What took place was as follows: The 
 eye, after a pause at A, makes a gradual for- 
 ward movement to B, pauses and moves back 
 rather rapidly to C, and from C moves back 
 very gradually toward the edge of the line, 
 and a little beyond A. 
 The results are most interesting. The eye fell too far short in 
 the first line (Fixation No. 1) necessitating a refixation nearer 
 the beginning of the line. Just the opposite 'mistake' was made 
 in the second line, that is, the extent of movement was overesti- 
 mated. 
 
 The following explanation of this error seems most in accord 
 with the facts. The length of, at least, the first return sweep is 
 estimated largely in indirect vision, when the eye is focused near 
 the end of the previous line. The peripheral estimation in this 
 case was vertically inexact to the extent that the edge of the sec- 
 ond line was overlooked, and the edge of the third line mistaken 
 for it. The error was not perceived until the first fixation was 
 made in or above the beginning of the third line. The move- 
 ment of the attention back towards the beginning of the sec- 
 ond line is shown by the direction of the gradual movement dur- 
 ing the fixation. From that position, however, a fairly clear 
 peripheral perception of the beginning of the second line could be 
 secured, and the second fixation is, therefore, made well within 
 that line. That there is still some uncertainty is shown by the 
 backward movement of the third fixation. 
 
 The error made at the beginning of the second line is corrected 
 in the third line, although there is also evidence here of overesti- 
 mation in the comparatively unusual proximity of the first fixa- 
 tion to the beginning of the line. The more accurate estimation 
 of the correct arc of movement may be due in part to another 
 factor, namely, the resident muscular sensations of the extent of 
 the angular displacement of the eye along the second line. The 
 accuracy of these sensations is not, however, believed to be as great 
 as when the extent of movement is determined by the retinal local 
 signs. The accuracy of the former is, as will be further noted, 
 greatly increased by repetition of the same arc of movement. 
 
 The angular displacement of the line of regard from the end 
 of the second line to the beginning of the third line was exception- 
 
36 THE PSYCHOLOGY OF BEADING 
 
 ally large, corresponding to about 115 mm. on the page. It is the 
 present thesis that this overestimation was due to two factors, the 
 tendency to repeat the same extent of movement as was made in 
 the previous return movement, and, secondly, the relative inac- 
 curacy of peripheral discrimination. When local discrimination is 
 more exact, i. e., when the point of stimulation lies nearer the fovea, 
 it serves as a check; on the tendency of the motor impulse to repeat 
 itself. The fifth and sixth lines are an instance in point. The 
 beginning of these lines falls well within the field of accurate local 
 discrimination, and, although there is considerable unsteadiness of 
 fixation, there is no inaccuracy of the motor impulse such as was 
 found in the fourth and again in the seventh line. The seventh 
 fixation of the fifth line is also of interest as further showing the 
 difficulty and uncertainty of movement produced by this arrange- 
 ment of lines. The eye moves beyond the edge of the line, neces- 
 sitating a correction by the eighth fixation. To repeat, is this 
 peculiar phenomenon as well as the unsteadiness of the fixations 
 at the beginning of the fifth and sixth lines due to a sort of blind 
 force of motor innervation? The peculiar unsteadiness of fixation 
 just noted, and the large error of estimation now to be noted in 
 the first fixation of the seventh line show that the force of the 
 resident sensations of the previous movement is from long habit 
 and practice very potent in determining the extent of the succeed- 
 ing movement. The exact point of fixation depends on the ac- 
 curacy of local discrimination in indirect vision, i. e., on the retinal 
 local signs. If the latter are accurate, as was the case in line five, 
 because the edge of the line could be clearly seen when the eye 
 was fixating the last word of line four, the next fixation will be 
 exact, despite the tendency of motor innervation to be governed in 
 extent of movement by the previous movement. The regressive 
 movements in the first and second fixations of both lines give evi- 
 dence of this tendency. If, on the other hand, the point that is 
 about to be fixated lies so far upon the periphery that the accuracy 
 of local discrimination is decreased, the motor impulse has full 
 sway, and is subject to correction only after it has completed 
 itself. This seems to be the case in line seven. As in the case of 
 the second and third lines, and particularly the fourth line, the 
 angular distance from the end of line six to the beginning of line 
 seven was too great to admit of accuracy in local discrimination, 
 if made, when the eye was fixating the former point. The arc of 
 movement of the previous line is repeated to a minute, with the 
 result that the first fixation fell far short of the edge of the line 
 and a supplementary movement of large extent was made neces- 
 sary. The case is nearly an exact reversal of the overestimations 
 
THE FIXATION PAUSES 37 
 
 made in the second, third and fourth lines. There was the same 
 inaccuracy of local discrimination due to extreme peripheral vis- 
 ion. But whereas the repetition of the same arc of movement in 
 lines 5 and 6 gave opportunity for acquiring a short motor tend- 
 ency or habit for a given amount of angular displacement, the 
 lack of uniformity in the extent of movement in lines 1, 2, 3 and 
 4, due to the inaccuracies of perception in both the vertical and 
 horizontal planes caused a wider extent of over estimation. In a 
 iwrd, the over estimation mid the underestimation are both ex- 
 plained hy the inaccuracy of local discriminations on the periph- 
 ery; the differences in amount were due to the acquirement in 
 one case of a tendency to a fixed extent of motor innervation, and 
 to the lack of it in the other case. 
 
 The correct extent of movement is acquired in the first or at 
 most second trial, and the succeeding movement follows automatic- 
 ally and with greater accuracy. The motor habit once acquired 
 tends to persist even when the angular extent of movement is 
 altered, as in the present experiment. The error of fixation is, 
 therefore, in this case due more to the persistence of the motor im- 
 pulse than to inaccuracy of perception. An inaccuracy due to 
 the latter cause, i. e., an error of local discrimination of peripheral 
 vision, would very infrequently, especially in movement of this 
 small angular displacement, equal in extent the distance found 
 between the first and second fixations of the seventh line. The 
 average errors determined in the experiments above referred to^ 
 were for one subject 1° 48'' and the other subject 2° 59' 24''. The 
 extent of the inaccuracy of vertical estimation assumed as the ex- 
 planation of the error made in the second line is, on the other 
 hand, quite within these limits. 
 
 The conclusion, therefore, seems warranted by the results of 
 the experiment that the larger number of refixations at the begin- 
 ning of a passage is due to the fact that the motor impulse is in 
 the process of adjustment. The experiment also justifies a second 
 conclusion, namely, that normally the extents of angular displace- 
 ment of the eye tend to become uniform, due to the formation of 
 motor habits of reaction. The errors of fixation are due in part 
 to the subversion of these motor tendencies. 
 
 The greater preponderance, therefore, of refixations in the first 
 lines of a passage, to recur to the original question, is due to the 
 inaccuracy of local discrimination unsupplemented by any resident 
 muscular sensations of the correct extent of movement. Since the 
 length of lines remained the same in the above pages, the accuracy 
 
 ' Cf. Psychological Revieiv, XI, p. 301. 
 
38 THE PSYCHOLOGY OF READING 
 
 of fixation in the succeeding lines was much abetted by the rapid 
 acquirement of a short motor habit of the given extent of move- 
 ment. 
 
 The reason why there are more refixations in long lines is not 
 due primarily to motor difficulties, but to inaccuracy of peripheral 
 discrimination. The reason why there are fewer refixations and 
 why greater accuracy of movement is attained in the short line 
 is that movement is subject to the correction of accurate vision and 
 the " guidance of motor habits. 
 
 The experiment has demonstrated that the difficulties of ac- 
 curate fixation may be increased even for the practised adult 
 reader by necessitating too frequent change of the arc of move- 
 ment. Uniformity in the length of line is doubtless, therefore, 
 even more a desideratum for beginners in reading. An examination 
 of school primers and readers will, however, convince anyone that 
 the mistake of employing many different lengths of line is unfor- 
 tunately more commonly made in these books than in perhaps any 
 other class of books. The fault arises in part from printing each 
 sentence as a single paragraph, necessitating many incomplete 
 lines ; but it is chiefly due to the insertion of pictures and drawings 
 in the margins of the pages. If the pictures are highly colored 
 this in itself, if we may give any weight to the experiments which 
 have been made concerning the reflex effect of peripheral color 
 stimuli upon the accuracy of fixation,^ may be a hindrance to ac- 
 curacy in the reading movements. It is, of course, recognized that 
 such objections as the latter may be entirely outweighed by the 
 advantages of securing and holding by this indirect means the in- 
 terest and attention of the pupil, etc. But the objectionable feat- 
 ures may be minimized. There is, however, no such evident justi- 
 fication for the large differences in the lengths of line. As will 
 be made clear when the question is taken up in a later section, the 
 acquirement of a uniform motor habit and a certain rhythmical 
 sequence of movements and fixations depends entirely on the length 
 of line, almost to the exclusion of the effect of differences in the 
 form of sentence and of subject matter. 
 
 As an illustration of the differences that may be found on 
 almost any page, I have given below the measurements of the first 
 twenty lines of a second reader in common use. 
 
 ^ Cf. for example R. MacDougall. American Journal of Physiology, 1903 
 pp. 122-130. 
 
THE FIXATION PAUSES 39 
 
 Measurements in mm. of the le^igths of line in a 'second reader' 
 Line mm. Line mm. 
 
 1 
 
 57.5 
 
 2 
 
 49.5 
 
 3 
 
 12 
 
 4 
 
 35.5 
 
 5 
 
 41 
 
 6 
 
 41 
 
 7 
 
 41 
 
 8 
 
 35 
 
 9 
 
 83.5 
 
 10 
 
 96.5 
 
 11 
 
 61 
 
 12 
 
 96 
 
 13 
 
 84 
 
 14 
 
 85 
 
 15 
 
 97 
 
 16 
 
 87 
 
 17 
 
 67 
 
 18 
 
 81 
 
 19 
 
 71 
 
 20 
 
 95 
 
 The length of line of the book in question is 101 mm. ; that of 
 the page under experiment was 88 mm. The indentations were 
 26 mm. ; in the reader there are differences of 40 mm. and more. 
 There is little justification for the diversity of motor innervation 
 which is thus produced.^ The only advantage is in the size of 
 type of the reader, which was 2 mm. in height (of small n), but, 
 as noted later, I have found many much longer lines in these read- 
 ers of but 1.5 mm. in height. 
 
 As an illustration of the extent to which this total disregard of 
 uniformity in line arrangement has been carried, a couple of 
 pages from first and second readers in common use are here re- 
 produced with the omission of the pictures on the margins and in 
 the body of the text. 
 
 An examination of the printed charts and diagrams that are 
 used in the teaching of reading and the other school subjects 
 will also reveal a large number of the more purely physiological 
 difficulties, needlessly and gratuitously put in the way of the be- 
 ginner. 
 
 ^ A small indentation of a few millimeters, for example of every other 
 line, but not of sufficient extent to affect the uniformity of the horizontal 
 movements, or even the length of line Itself, may be of a distinct ad- 
 vantage in eliminating motor inaccuracies in the vertical plane. The 
 indentation would help to differentiate the lines, and prevent their con- 
 fusion. This suggestion was made by Professor Cattell. 
 
40 the psychology of reading 
 
 Pages Illustrating Poor Line Arrangement 
 
 On his way back he finds the horse 
 looking over the fence, as if he 
 were watching to see that the work 
 is properly done. Then the gardener 
 
 thinks of the thirsty flowers and gets 
 his watering pot to sprinkle them. 
 A tree in one end of the gar- 
 
 den is the home of this little bird, 
 
 who likes to nip the pears that 
 
 grow on it. Here is a pear, 
 
 and here is an apple, and here 
 are some ripe strawberries, too. 
 
 The birds find them all; for the 
 
 gardener grows them for 
 his children. The children eat 
 them, and think they are very good. 
 The gardener has also many pretty 
 flowers in his garden: red roses and violets. 
 
THE FIXATION PAUSES 41 
 
 Pages Illustrating Poor Line Arrangement 
 
 NCE upon a time a donkey 
 heard a grasshopper chirp- 
 ing in the grass. 
 
 He was very much 
 
 pleased with the beautiful song. 
 
 **Ah !" he said to himself, 
 
 ** if I could sing like that, how happy 
 
 I should be!'' 
 
 So he bowed low to the grasshopper, 
 
 and said, ** Kind friend, what food do you 
 
 eat to make your voice so sweet?" 
 
 "I drink the evening dew," replied the 
 grasshopper. 
 
 The foolish donkey tried to live on the 
 same food, and died of hunger. 
 
 An eagle flying through the air was 
 wounded by an arrow from the bow of 
 
 a hunter. 
 
 The king of birds fell to the ground, 
 
 for the arrow had entered his heart. 
 
 As the eagle looked at the arrow, he saw that 
 
 its shaft was tipped with his own feathers. 
 
 *^How hard is my fate!" he moaned, **I 
 
 helped to wing the arrow which kills me." 
 
CHAPTER V 
 
 Is There Perception During Eye Movement? 
 
 Preliminary to the further study of the significance of the 
 fixation pauses, it is necessary to review in greater detail the evi- 
 dence (see above) upon which the conclusion was reached that the 
 eye does not see during the movements in reading. The answer 
 is of first importance to the theory of the fixation pauses. As 
 already noted, Erdmann and Dodge computed that from If to 
 If of the reading is spent in fixation, and, on the basis of 
 measurements by Dodge of the angular velocity of eye movements, 
 argued that, if the rate of movement found held good for read- 
 ing, it did not admit of perception during the interval. 
 
 Dodge found by means of a modification of the Helmholtz- 
 Lamansky method of after images the average length of time for a 
 movement through five degrees to be .015", and through 10 de- 
 grees .02'^ Estimating that the velocity of the short interfixation 
 movements must be at least as much as .02'^, the problem was to 
 determine whether a fusion of the letters — which would preclude 
 their recognition— does not result at that rate. About 12 or 13 
 letters represent in the line the average extent of the movement 
 of the eye between two pauses. Counting the separate parts of 
 the letters and the intervening spaces, the appearance represents 
 an alternation of about 50 black and white stripes. The white 
 spaces are about three times as broad as the black lines. This 
 would make the length of stimulation of each of the black lines (if 
 the velocity of movement for the whole arc is .02"), .0002", and 
 of the white spaces .0006". But according to the experiments of 
 Plateau fusion of the black and white sectors of the color wheel 
 occurs when the interval of stimulation of the single sector is as 
 long as .008". Baxt also determined by another method that the 
 time of stimulation in order to allow recognition of a single letter 
 must be .0303" and for recognition of three letters .0533". While 
 the conditions in reading are not exactly the same as the above 
 experimental conditions, the very much smaller time of stimula- 
 tion in the former case seemed to the writers to warrant the con- 
 clusion that fusion occurs during the movements of the eye, and 
 that, therefore, the recognition of words occurs only during the 
 fixation pauses. 
 
PERCEPTION DURING EYE MOVEMENT 4R 
 
 The question was also put to an experimental test by passing 
 lines of print rapidly from left to right before the eye at the same 
 rate which the eye must supposedly pass over the letters in read- 
 ing, and it was found that fusion of the letters did actually re- 
 sult. The importance of the conclusion, of course, lies in this, — 
 that it is only in the interval of rest or pause that visual percep- 
 tion is possible. 
 
 The conclusion carries with it two or three assumptions which 
 may be called in question and at least need verification; first, that 
 the angular velocity of the short movements of the change of fixa- 
 tion within the line is as rapid as the voluntary sweeps of the 
 eye; whereas it is possible that the speed of the return movements 
 of the eye to the right may be so much more rapid than that of 
 the interfixation movements proper, that perception may be ex- 
 cluded during the former but not during the latter. In the second 
 place, the possibility of considerable variations in the rate of the 
 interfixation movements themselves was not recognized. It is as- 
 sumed that the pause is a full stop of the eye and that the remain- 
 ing time must be taken up wholly by the change from one of these 
 positions of rest to the next or succeeding pause. But it is evi- 
 dent from an examination of the accompanying records that there 
 are intermediary stages between the distinct breaks or pauses in 
 the movement and the movements proper. 
 
 These facts or possibilities make an actual measurement of the 
 interfixation movements themselves desirable. This has already 
 bleen done to a certain extent in two more recent investigations.^ 
 The measurements of Huey are subject to the criticism that the 
 extra weight of the eye-cup and pointer may have retarded the 
 movements of the eye. He found the velocity of movements to 
 the right to be on an average 43. 9o- for angular movements aver- 
 aging 3° 36'. The value of the more accurate measurement of 
 Dodge and Cline is unfortunately somewhat lessened by the fact 
 that only the average of movements varying from 2° -7° are given. 
 That is, it may be of some importance to know what the velocity 
 of movement is in each case and especially when the movement is 
 of small extent. 
 
 Their results are as follows: 
 
 'Huey op. cit., Vol. XI. Dodge and Cline, The Angle Velocity of 
 Bye Movement. Psychological Review, Vol. VIII, pp. 145-157. 
 
44 THE PSYCHOLOGY OF READING 
 
 Dodge and Cline on the velocity of the eye in reading 
 
 Movements towards the right Movements towards the left. 
 Sub- 
 jects No. M. M. V. Max. Min. No. M. M. V. Max. Min. 
 A 12 23.7 3.9 36.5 17.5 5 45.6 3.7 50.1 40.9 
 B 18 21.9 3.9 29.2 15 5 40.1 2.9 43.8 35.0 
 C 10 23.2 6.1 32.8 14.2 4 37.1 1.9 39 33.4 
 
 General 
 
 Average 22.9 40.9 
 
 The average time of the three subjects for the interfixation 
 movements of from 2°-7° is 22.9(t and for the return movements 
 to the left of 12°-14° 40.7cr. It is noted that this result of the 
 inovements to the right approximates closely the original as- 
 sumption of Erdmann and Dodge regarding the rate of movement 
 {.Q2"). Compared with the angular velocity of movements made 
 in response to peripheral stimulus, the movements to the right fall 
 below the general average of eye movements of 5° (34.5o- M. V. 
 1.5), and those to the left that of the general average of eye move- 
 ments between 10° and 15°, i. e., between 41. So-, M. V. 1.4, and 
 46.7(7, M. V. 4.5. 
 
 The conclusion drawn from this fact is that "the character of 
 the eye movements in reading is not materially different, qualita- 
 tively or quantitatively, from the eye movements which are made 
 in response to peripheral stimuli as the eye looks back and forth 
 between two fixation points." 
 
 The same writer has since in another study^ after a most care- 
 ful series of experiments and many preliminary failures succeeded 
 in demonstrating experimentally that fusion does take place in 
 these latter movements. The normal fusions of colored fields were 
 secured and "a part of a page of print exposed only during 
 movement gave a perfect though shadowy series of gray bands 
 on a lighter background, in which individual letters or words were 
 absolutely unrecognizable."- Although the arc of movement em- 
 ployed in these experiments was necessarily large (50°) and the 
 velocity of movement was, therefore, proportionately great, there 
 can be but little question that the results are directly applicable 
 to the movements in reading in general. Since, however, the rate 
 of movement has not been determined for any large number of 
 interfixation movements, it seems desirable that this be done, 
 especially as these latter movements differ from the longer 
 
 ^ Dodge, "Visual Perception During Eye Movement." Psychological Re- 
 view, Vol. Vll, pp. 454-465. 
 
 ^ For further experiments and discussion concerning the nature and 
 cause of fusion, see the article just quoted, pp. 464-465 in particular, and 
 Holt, Psychological Review, Monograph Supplements, Vol. IV; and Dodge, 
 Psychological Bulletin for Aug. 15, 1905. 
 
PERCEPTION DURING EYE MOVEMENT 45 
 
 movements made in response to peripheral stimulation in one 
 way at least which may have some significance; they are of 
 much greater variability in velocity. In the course of this 
 investigation I have determined in a very large number of cases 
 the exact letters and words that are passed over by any given in- 
 terfixation movement, and at the same time measured the rate of 
 movement. With this data at hand I have repeated the calcula- 
 tions of Erdmann and Dodge, in order to offer additional evidence 
 for the operation of fusion in the short and comparatively slow 
 interfixation movements. 
 
 Although the rate of fall of the plate adopted for reasons al- 
 ready assigned is not as well suited as a more rapid fall would have 
 been for careful measurement of the interfixation movements, since 
 the eye's movement is enlarged as noted about four times by the 
 camera, and the record may then be enlarged upwards of twenty 
 times without inconvenience, giving a total enlargement of from 
 75 to 100 or more times, sufficient accuracy can be attained from 
 the records as made. In order to secure as great accuracy as pos- 
 sible, however, only the best and clearest of the interfixation move- 
 ments on a given plate were selected for measurement. This doubt- 
 less gives a greater preponderance to the slower movements, as they 
 naturally make the longer and clearer record, but this is an ad- 
 vantage, because the purpose in mind is not to determine the aver- 
 age rapidity of movement, but simply to test the validity of the 
 supposition. And it is in the slow movements, if at all, that the 
 hypothesis is open to question. 
 
 These measurements (as well as those of Dodge) are always 
 subject to one error due to the limitations of the method of regis- 
 tration. When the movement of the eye does not follow the exact 
 horizontal level, but moves upwards or downwards at the same 
 time, the extent of registration is diminished in the first case and 
 increased in the second case due to the fact that the plate is fall- 
 ing during these changes in the vertical plane. Head movements 
 in the vertical plane would cause a similar difficulty and error. 
 That this does occur can be detected in some instances, but the 
 amount of error can not be determined. But this error is as apt 
 to occur in one direction as in the other; that is, is as apt to in- 
 crease the duration of movement as to diminish it. We have, 
 therefore, selected only the slowest movements. If the rate of 
 movement in these is sufficient to cause fusion, it must occur in the 
 more rapid movements. 
 
 In Table VI the velocity of the first ten return movements to 
 the left in reading a newspaper column is given. Width of line 
 56.5 mm., app. 12° 30'. The extents of movement are stated in 
 
THE PSYCHOLOGY OF READING 
 
 terms of distance on the line, and in degrees of the angular excur- 
 sion of the eye. The results agree with these of Dodge just quoted. 
 
 
 
 Table VI 
 
 
 
 
 Velocity of movements 
 
 
 
 RETURN MOVEMENTS TO THE LEFT 
 
 
 
 <r 
 
 Extent of Line 
 Covered 
 
 Degrees of Mt. 
 
 1-2 
 
 44 
 
 45.5 mm = 
 
 10.0 
 
 2-3 
 
 54 
 
 37.25 = 
 
 8.1 
 
 3-4 
 
 35 
 
 51.5 = 
 
 11.3 
 
 4-5 
 
 50 
 
 49.5 = 
 
 10.9 
 
 5-6 
 
 42 
 
 40 = 
 
 8.8 
 
 6-7 
 
 38 
 
 45.5 = 
 
 8.0 
 
 7-8 
 
 38 
 
 49 = 
 
 10.2 
 
 8-9 
 
 41 
 
 35.5 = 
 
 7.8 
 
 9-10 
 
 48 
 
 57 = 
 
 12.5 
 
 10-11 
 
 25 
 
 63.5 = 
 
 14.0 
 
 11-12 
 
 30 
 
 47.5 = 
 
 10.4 
 
 12-13 
 
 35 
 
 46 = 
 
 10.1 
 
 13-14 
 
 38 
 
 24 = 
 
 5.2 
 
 14-15 
 
 46 
 
 45 = 
 
 9.9 
 
 15-26 
 
 35 
 Av. = 39.00- 
 
 47 = 
 
 10.3 
 
 
 Av. ^ 45.7mm. 
 
 Av. = 9.8° 
 
 
 
 Table VII 
 
 
 Showing the duration and number of letters and intervening spaces 
 
 covered by sixteen interfixation movements 
 
 No. of 
 
 (a) Letters (b) 
 
 1 14 
 
 2 10 
 
 3 9 
 
 4 12 
 
 5 8 
 
 6 5 
 
 7 13 
 
 8 7 
 
 9 6 
 
 10 4 
 
 11 11 
 
 12 6 
 
 13 5 
 
 14 8 
 
 15 7 
 
 16 4 
 
 Totals 129 304 351 
 
 Spaces 
 
 Ti 
 Movei 
 
 me of 
 :nent in"" 
 
 38 
 
 
 81 
 
 20 
 
 
 •jO 
 
 21 
 
 
 19 
 
 33 
 
 
 lo 
 
 20 
 
 
 16 
 
 11 
 
 
 22 
 
 28 
 
 
 16 
 
 15 
 
 
 32 
 
 15 
 
 
 25 
 
 11 
 
 
 25 
 
 22 
 
 
 30 
 
 14 
 
 
 22 
 
 11 
 
 
 19 
 
 22 
 
 
 22 
 
 13 
 
 
 19 
 
 10 
 
 
 15 
 
PERCEPTION DURING EYE MOVEMENT 47 
 
 Taking the totals and counting each letter as composed of two 
 black lines and counting the intervening spaces as equal in width 
 to three black lines, we have (2X129+3X304=) 1170 black and 
 white sectors, which must be passed over by the eye in 351cr, or 
 the time of stimulation of one black sector is .0003''. The result 
 is very close to the first computations of Erdmann and Dodge 
 (.0002'). The duration of stimulation is well below the threshold 
 necessary for producing a distinct visual impression. It is true 
 that if one selects the two or three movements of least duration in 
 the above table, the time of stimulation is just about the threshold. 
 But even if there were opportunity for perception during the slow- 
 est movements, the habit established in the average and rapid 
 movements of disregarding percepts during movement would doubt- 
 less carry over to the slow movements. 
 
CHAPTER VI 
 
 The Distribution of Attention in Perception 
 
 In the above computations no account has been made of the 
 slow 'shifting' movements that are very common in the records of 
 some subjects, noticeably H and S. They form an intermediate 
 stage between the fixations proper and the interfixation movements. 
 The distance traversed is often equal to that read in other cases 
 by two or more fixations and connecting movements, and, although 
 the eye has been in constant motion, the necessary perceptions 
 have been made quite as well as if the eye had made one or more 
 abrupt pauses. Since the movement is so slow— not infrequently 
 ten to twenty or more times slower than the usual movement 
 between fixations— it is fair to assume that tlie eye does see during 
 these movements and that in fact they take the place of the more 
 clearly defined fixation pauses. 
 
 In the present study, therefore, the 'shiftings' under discus- 
 sion have been grouped as fixations. The chief reason for this 
 classification is that they are generally followed by a rapid move- 
 ment of the usual velocity. The 'shifting' movement can hardly, 
 therefore, serve the purpose of a mere change of the point of re- 
 gard, and its function must be similar to that of the usual reading 
 pause. What the difi:erentiation is functionally between this grad- 
 ual movement and exact fixation must be left largely an open 
 question. Judging, however, from a study of the directions of 
 various movements, it seems evident that they follow or accompany 
 the movement of attention. 
 
 It is not believed, however, that these shifting movements nec- 
 essarily denote a successive perception of the parts over which the 
 line of regard is moving, although it may possibly mean that in 
 some cases, and it is then the exceptional mode of reaction. 
 The basis for this supposition is that an examination (cf. Charts 
 of Subject H) of the words thus read offers no explanation of why 
 these elements should be singled out for successive perception, 
 especially since in a previous or succeeding reading the fixation 
 is often made without shifting. It is by far more likely that the 
 lingering of the eye over the separate letters has no effect on the 
 simultaneity of the resulting perception. It more probably denotes 
 a central difference in the rate of assimilation. Suppose, for ex- 
 
DISTRIBUTION OF ATTENTION 49 
 
 ample, that, as a given foveal perception is completed, the attention 
 changes to the matter lying on the periphery of vision. This move- 
 ment of attention must precede that of the eye from purely physio- 
 logical reasons. The difference in the promptness of the corre- 
 sponding change in fixation may then depend on differences in 
 the intensity and concentration of the attention upon the new 
 peripheral sensation. In one case, because perchance of the slow- 
 ness of assimilation and recognition, the attention, although already 
 turned to the periphery, is not at first as 'intent' or alert to the 
 sensation. In that case the muscular tension and unbalance to- 
 wards the periphery which always marks the separation of atten- 
 tion and fixation is less acute, and the eye shifts more gradually, 
 at least at the start. When, on the other hand, attention is imme- 
 diately 'concentrated' peripherally, the muscular strain is at once 
 too great and a rapid change of fixation follows. We can, in other 
 words, conceive of a sort of rivalry of the peripheral perceptions 
 to gain the center of consciousness, and this may cause a fluctua- 
 tion of the attention from the fovea to periphery which the slower 
 movement of the eye muscles can simulate only by denoting the 
 general 'tendency' of the change of attention. When, according 
 to this view, the peripheral stimuli at once gain the center of con- 
 scious perception, there is no hesitation in the response of the eye. 
 This, however, it should be noted, marks a successive perception 
 of word wholes and phrases, and gives no confirmation to a theory 
 of a successive perception of the letters and elements of the word. 
 If, moreover, the shifting movements were due to the latter cause, 
 we should be at a loss to explain the usual type of motionless fixa- 
 tion, as will be noted further in another connection. 
 
 In this respect the hypothesis of the relation of fixation and 
 attention is at variance with the Wundtian theory, developed by 
 Zeitler.^ The latter held that the perception of a given group of 
 several Avords (during the short exposure of the tachistoscope) is 
 not simultaneous but successive. The attention moves over the 
 field of exposure from left to right, passes lightly over the unim- 
 portant parts of the words, and singles out certain more signifi- 
 cant parts, the so-called domineering letters. These serve ordi- 
 narily as the principal basis for perception. (Domineering letters 
 are (1) capitals, (2) letters that stand above the line as 1, t, d, h, 
 k, and especially such combinations as scht, and (3) in some in- 
 stances m, w; i. e., the broader letters, although neither author is 
 clear on this point.) 
 
 Messmer^ proposed a modification of this theory. The attention 
 is supposed by him to have no uniform motion but to flit back and 
 
 ^Op. cit. 
 
50 TEE PSYCHOLOGY OF READING 
 
 forth over the words, singling out first the domineering letters and 
 secondly the other letters/ Successive stages of apprehension oc- 
 cur only in words which include one or more domineering letters. 
 I quote directly from Messmer's own statement of his interesting 
 theory. 
 
 "The greater or less uniformity of general configuration (Ge- 
 sammtfiguration) does not in itself stand in direct relation to 'recog- 
 nizability.' It serves rather as a more or less powerful occasion or 
 cause (Anlass) to a unitary-innervation (Gesammtinnervation), If, 
 for example, I have seen the words physiologisch, psychologisch, 
 philosophisch, philologisch, as often as the words wimmern, iiber- 
 einstimmen, nennen, weinen, etc., I am compelled in reading the 
 first example to read by analyzing into sections (Stiicken), while 
 the words of the second group are clear to me at a glance. Why 
 is this? The words of the first group receive from their letters 
 much less assistance to a 'Gesammtcharakler,' than those of the 
 second group. One may say : the more a word possesses letters 
 of individual geometrical form, so much the more does a unitary 
 innervation run danger of being divided. On the other hand, the 
 more the letters of slight individual form (with straight lines) pre- 
 dominate, so much the greater is the impulse to a unitary inner- 
 vation. ' '2 
 
 Since, however, "the greatest unity of total image (Gesammt- 
 bildes) and the maximum vivacity (Lebhaftigkeit) are two dia- 
 metrically opposed factors," or again, since "the Gesammtinnerva- 
 tion and the certainty of recognition stand in inverse ratio to each 
 other," the most fortunate cases for recognition are those in which 
 "both these factors participate in approximately equal parts in 
 the word-image." That is to say, words like wimmern are per- 
 ceived immediately as wholes, with little or no effort and for that 
 very reason are likely to be confused with a word like weinen. 
 Words, on the other hand, like physiologisch and psychologisch 
 compel on the part of attention a prior recognition of certain more 
 prominent parts. The words are split into the elements and the 
 separate sections successively perceived, and are thus less often 
 confused. The best words for both certainty and rapidity of recog- 
 nition are those which combine the merits of both these classes, 
 i. e., part of letters that stand above the line and part of short let- 
 ters. The examples cited of this class are ' charakteristisch ' and 
 ' wissenschaftliche. ' It is not necessary here to discuss in detail 
 
 ^ I have reviewed the work of Messmer in more detail than is neces- 
 sary here in the Journal of Philosophy, Psychology and Scientific Methods, 
 Aug., 1905. 
 
 = Op. cit, pp. 224 ff. 
 
DISTRIBUTION OF ATTENTION 51 
 
 this interesting analysis, but I may record a repeated observation 
 of my own experience which I have found verified in the case of 
 several students who make much use of the words, that there are 
 no two words which I more often confuse, and apparently in direct 
 proportion to the times read than the English words psychological 
 and physiological, and a good second are the words philosophisch 
 and philologisch. My own observation is that these words are 
 read wrongly repeatedly and therefore read as wholes, and even 
 when the context makes it clear that a mistake has been made it is 
 only with the greatest effort that it is possible to analyze them 
 into their parts and tell 'which is which,'— so strong is the ten- 
 dency to read the words as wholes.^ 
 
 That the domineering letters as well as other peculiarities which 
 we note in words, as their lengths and the relation of letters, are 
 made use of as 'cues' in perception is, I think, not questioned. 
 Moreover, it is evident that some of the subjects in the experiments 
 of Messmer in tachistoscopic reading caught only small sections or 
 parts of words and that others read larger sections and whole 
 words. But I have yet to find any evidence to substantiate the 
 two hypotheses based on these facts, viz., first, that the presence 
 of several domineering letters and complexes lessens the chances 
 of the perception of the words as wholes, — this point has just been 
 discussed,— and secondly, that because some individuals perceive 
 smaller extents of words than others (in tachistoscopic reading) 
 the perception of the letters is composite ; i. e., made by a hur- 
 ried combination of the fragments of perception. To discuss this 
 latter point, it is just as logical or illogical to adopt the alternative 
 that, since some subjects perceive whole words and others only sec- 
 tions, that, therefore, the real method of the latter is analytic. 
 That is, it might be argued they really perceive the whole word or 
 much larger sections, but only a small section gets to conscious- 
 ness with sufficient vividness. Either assumption is equally objec- 
 
 ^ There is no difference in the unity and simultaneity of perception of 
 the two classes of words. Both are perceived at once as units, but 
 while nennen and weinen are similar in form, the context generally saves 
 them from confusion, whereas not infrequently any one of the words, 
 physiologisch, psychologisch, philosophisch will fit for the moment into the 
 context. The trouble comes when we have to analyze and see which is 
 the right word. So accustomed are we to perceive words as units, that 
 we find considerable difliculty in making this analysis. If there is a dif- 
 ference between these two classes of words, the difference first appears 
 w!hen we try to analyze; and of the two classes the latter seems 
 the harder to analyze. The reverse ought to be the case were they or- 
 dinarily first perceived as parts, but how much of this is due to con- 
 text and previous confusion and how much to the presence of domineer- 
 ing letters, is an open question. The difference in these two classes of 
 letters is well worth pointing out, but I doubt if the question of their 
 significance in perception has been correctly answered by Messmer. 
 
53 TEE PSYCHOLOGY OF BEADING 
 
 tionable. The only fact which has been brought out is a difference 
 in the span of attention among different individuals. And this no 
 more justifies the assumption that the method of perception is 
 synthetic in one case, than that it is analytic in the other. It jus- 
 tifies neither. What this fact does make plaiisible is that the sub- 
 ject who can read but small sections may have to combine them in 
 order to obtain wholes. The subject who seizes the larger sections 
 and whole words has them to start with and consequently has no 
 need of synthesis ; and in that case, the perception of the first sub- 
 ject may be accurate and exact, and that of the latter confused 
 and liable to error. The fallacy of Messmer is in attempting to 
 reduce to uniformity results that clearly indicate individual differ- 
 ences among his subjects. Evidently as far as tachistoscopic read- 
 ing is concerned, the individual who reads by syllables (or sec- 
 tions of domineering letters), must combine them to get words, 
 but the question still remains as to the extent to which these results 
 of tachistoscopic reading with exposures of a few thousands of a 
 second are directly applicable to reading in the -usual way. The 
 results of the present investigation will, it is hoped, offer some 
 evidence on this point. 
 
 There is, on the other hand, evidence to support the observa- 
 tion that the threshold of perception of the domineering letters, 
 to use Messmer 's terminology, is (in the tachistoscopic reading) 
 lower than that of the other letters, — although the much greater 
 time of stimulation may, as suggested, in ordinary reading make 
 this observation of comparatively little general importance. A 
 study of the tables of readings that are given by Messmer shows, 
 for instance, that in the comparatively few cases Avhere isolated 
 letters are perceived, they are usually domineering. (The follow- 
 ing letters were the only ones perceived apart from the other let- 
 ters, given in the order that they are found in Messmer 's tables 
 (pp. 202-204) i, k, h, w, g, k, g, h, b, gew, sch, her, ge, b, d, w, i, 
 ch, ge, S, nnen).^ 
 
 But this fact does not in the least support the theory that sev- 
 eral of these letters are perceived in succession before other letters. 
 In fact, the tables of Messmer point to the opposite theory; for, if 
 there were successive perception, one would possibly expect to find 
 an occasional perception of a series of domineering letters as h, t, 
 fl, etc., whereas with the exception of the above letters, and 'arts,' 
 and 'he' of 'heimwartsfliehenden,' every other partial perception 
 of a word is by syllaUes, and not by domineering letter combina- 
 tions. The original contention of Erdmann and Dodge hold^ good 
 
 '■ These are not so given by Messmer, but are collected from his tables 
 by the writer. 
 
DISTRIBUTION OF ATTENTION 53 
 
 that while some letters undoubtedly condition the total-word form 
 more than others ("Words of characteristic optical ' Gesammtf orm ' 
 are more easily recognized than those of undifferentiated (gleich- 
 formiger) configuration." Op. cit., p. 157), yet the percept itself 
 during a given exposure is unitary and simultaneous. 
 
 And, in ordinary reading, the facts brought out in the present 
 investigation seem to the writer to justify the further conclusion 
 that the foveal perception during any given fixation is simulta- 
 neous, and that when the attention is changed even to a smaller 
 part of its original field, as in the ease of a misspelled word (cf. 
 gorgeously, Chart 62B), or to a letter out of alignment, as in proof 
 reading, there is a change in the character and content of the re- 
 sulting perception. The first perception of the word just noted 
 was of its total word form, the second differed in that it was lim- 
 ited to a special scrutiny for the misspelling. Even such a small 
 change in the point of attention usually causes a change in the 
 point of fixation. This means that successive sections of the word 
 command the centre of attention. Peripheral images are less vivid 
 and clear than foveal, but they are always 'wholes.' 
 
 The extent or span of attention may also vary greatly. There 
 is reason to believe, as will be shown later, that the general char- 
 acter of the whole line or a large part of it is often perceived dur- 
 ing the first fixation. In order to accomplish this, the span of 
 attention must not only be widened, but the attention may shift 
 towards the periphery. In the latter case, it is believed that this 
 change of attention is evidenced by a gradual movement of the 
 eye towards the periphery, which, as has been pointed out, is most 
 noticeable in the first fixations. This perception is quite different 
 from that usual in the succeeding fixations as regards the definite- 
 ness of its peripheral content ; but it is at the same time simulta- 
 neous for large groups of letters and words, the object being to 
 obtain a rough impression of perhaps the length of the line and 
 its general appearance. It is hard to believe that the succession 
 proposed by these authors is not purely a matter of interpretation 
 and that the succession is one of syllables and words and not in the 
 first instance of short and domineering letters. These letters as 
 far as perception is concerned are, except when specifically attended 
 to, perceived as parts of larger wholes. When on the extreme peri- 
 phery, they are seen only in rough outline and always in combi- 
 nation. 
 
 The movement of attention supposed by Zeitler and Messmer, 
 would in the case of ordinary (non-tachistoscopic) reading pre- 
 sumably be limited to the field of the single fixation; they must, 
 therefore, recognize a more general movement of attention from 
 
64 TBE P8TCE0L00Y OF BEADING 
 
 one fixation to another. The latter change is the only one for 
 which the writer finds evidence. The former hypothesis not only 
 lacks experimental evidence, (difficult in either case to establish), 
 but there are some experimental results which tend to disprove it. 
 Becker^ has made an interesting experiment of voluntarily separat- 
 ing the attention from fixation as a test of successive apprehen- 
 sion. To repeat somewhat for purposes of clearness: 
 
 "The general theory advocated by Wundt was that the un- 
 equally favorable position of the different parts of the field of 
 view during fixation is rendered less noticeable through the fact 
 that the point of attention moves successively to those letters which 
 are distant from the poiiit of fixation; a 'searching' (Absuchen), 
 as it were, of the visual field thus occurs at each fixation. Such 
 an hypothesis seems improbable in view of the difficulty which is 
 ordinarily experienced in separating the points of attention and 
 fixation, and the experiments made tend to disprove it." 
 
 "In the first experiment [of Becker] the attention and fixa- 
 tion are centered together upon a letter lying in the primary line 
 of regard; in the second experiment the attention is separated 
 from the point fixated to a letter lying to the left of it. If a fre- 
 quent change of attention occurs during the exposure, both let- 
 ters ought to be equally well perceived in either instance. The 
 results show, however, that in the first case the letters to the left 
 are less often read, and in the second case the percentage of per- 
 ception of the left letter is twice as great as in the first experi- 
 ment."^ 
 
 The writer has also subjected the hypothesis of Zeitler and 
 Messmer to a further test. If a 'searching' by attention of the field 
 of view occurs, it seems not unlikely, in view of the observed rela- 
 tion of fixation and attention, that the action of attention should 
 have some effect on the accuracy or steadiness of fixation, which 
 might be detected. Photographs have, therefore, been made of the 
 eye while reading several phrases and sentences of four or five 
 short words exposed for 10o--100o-. 
 
 In order to register on the negative along with the reflection 
 from the cornea the exact time or point where the exposure began, 
 the electrical circuit of the fall exposure apparatus was con- 
 nected also with the spring pendulum and was closed previous to 
 exposure, thus keeping the pendulum drawn to the magnet; the 
 beginning of the exposure broke the circuit and released the 
 
 ^Op. cit. 
 
 ^ Quoted for convenience from the writer's review of the article of 
 Becker in the Journal of Philosophy, Psychology and Scientific Methods, 
 Aug., 1905. 
 
PLATE IV 
 
 1 2 
 
 B 
 
 SN M 
 
 "■^x" 
 
 '"1- 
 
 y \^ < 
 
DISTRIBUTION OF ATTENTION 55 
 
 pendulum. The vibrations of the pendulum then gave opportu- 
 nity to determine how soon after the exposure any movements oc- 
 curred, in case that should happen. The results of four records 
 are as follows. A copy of one of the plates (2 records, Nos. 4 
 and 5) is reproduced in Plate IV. 
 
 DESRIPTION OF PLATE IV. 
 
 (The records are one-third the size of the original records.) 
 A 1. Record showing the steadiness of apparatus and absence of lat- 
 eral movement in the falling plate. It is a registration of the reflection 
 of light from a gilded bead placed in the head rest. 
 
 2. Vibrations of Koenig tuning fork (vibrating 100 per second) cf. 
 PI. II. The horizontal knife lines at every 19 vibrations represent the 
 incidence of the oscillations of the spring pendulum. 
 
 3. Steadiness of fixation. Record shows the degree of steadiness in 
 fixation which it is possible at times to maintain for a short interval. 
 The oscillations of the pendulum appear on the left side of this print; 
 the light is admitted to this part of the negative only through a small 
 aperture in pendulum. 
 
 4. Record No. 54. Tachistoscopic reading. (This print was by mis- 
 take inverted, i. e., mounted upside down, the beginning of record is at 
 the top edge of page.) The arrow marks the beginning of the ex- 
 posure. Note that the unsteadiness of the eye begins a little before the 
 actual exposure. 
 
 5. Record No. 53b. Same as No. 4. 
 
 6. Effect of the recollection of memorized passages upon the steadi- 
 ness of fixation. The arrows denote the beginning of new lines (of 
 poetry) in the recitation. 
 
 Plate IV. B. Line drawings of eye movements in the visual recall of 
 memorized passages. (See text.) 
 
 In two of the records there is absolutely no evidence of any 
 movement or unsteadiness of the eyes. This is surprising in view 
 of the difficulty of maintaining steadiness. In the other two rec- 
 ords (those reproduced in Plate IV, Nos. 4 and 5) there is some 
 unsteadiness of the eye at the time of exposure. Since, however, 
 in each case the unsteadiness begins about 40o- before the begin- 
 ning of the exposure, it is clearly not due to a movement of atten- 
 tion, but simply due to the muscular tension of maintaining fixa- 
 tions previous to the exposure. The unsteadiness lasts for an equal 
 or somewhat longer time after the exposure. It is also interest- 
 ing to note that these (33B and 34A) were the last two records 
 taken. Several intervening exposures were made, and the fatigue 
 of maintaining preliminary fixation may have had something to 
 do with the increased unsteadiness. If the attention moved about 
 as argued by the above writers, its movements would often be of 
 much larger extent than the usual interfixation movement. It seems 
 unlikely that such changes should in one case cause a change of 
 fixation and in the other case produce no effect on even the steadi- 
 ness of the eye. If, on the other hand, perception is simultaneous 
 
56 TEE PSTCHOLOGT OF READING 
 
 for the given area in question, and the attention changes only 
 when its field is extended, or contracted to a smaller area of the 
 same field, we would naturally expect no change of fixation, as in 
 the present instance. 
 
 Finally, the movement of attention supposed by these authors 
 does not differ either in its purpose or, in many eases at least, in 
 its extent from that which accompanies every change of fixation. 
 In one case, however, there is always a physiological accompani- 
 ment, in the other, none. The purpose of the change during fixa- 
 tion is, according to the Wundtian theory, to equalize the unequally 
 favorable portions of the field of view ; accordingly, the change in 
 this case must not infrequently cover an extent of twenty-six let- 
 ters, — indeed the theory was invented to explain such extensive 
 perceptions,— that is, an extent of reading matter larger than that 
 covered by the average change of attention which accompanies fixa- 
 tion. It is certainly necessary to explain how such movements of 
 attention could occur without affecting the steadiness of fixation. 
 One more objection may be made by the defender of successive per- 
 ception to the above experiment, namely, that in the case of the 
 short exposure the disappearance of the objective stimulus renders 
 a change of fixation useless. It is, therefore, inhibited, although 
 the movement of the attention itself is carried out. This point is 
 discussed in the following section. 
 
CHAPTER VII 
 
 Vis'UAi Processes Involved in Recollecting 
 
 But is it true that the continued presence of the objects of 
 perception is essential in order that a change of the attention may 
 'release' the motor discharge of the eye muscles; and, if so, is the 
 inhibition of movement always successful? The results of another 
 experiment seem to the writer to offer some evidence upon this 
 question. That which we ordinarily do when we run over in 'our 
 mind's eye' the lines of a page which we have just been reading 
 or of a passage which we have committed to memory offers an 
 instance of a movement of attention over a field that is not pres- 
 ent to the visual sense, except as a memory image. The fading 
 after-images of the short exposure must furnish a much more ob- 
 jective basis for motor discharge than those recalled after an in- 
 terval of time has elapsed. Is, however, the recall attended by eye 
 movement? The experiment was made in two parts. In the first 
 case the subject, while stationed before the camera and looking 
 straight ahead at a white cardboard placed at ordinary distance 
 in lieu of a page of print, repeated a stanza of Longfellow's Evan- 
 geline which had been previously memorized. He was asked to 
 repeat it to himself in a just audible tone much as he would do 
 naturally if by himself. Of the subjects employed one was quite 
 unaware of the purpose of the experiment, the other was familiar 
 with it. (Cf. Plate IV, No. 6.) Since, however, small movements 
 or unsteadiness of the eye are not subject to voluntary control, 
 knowledge of the purpose of the experiment could hardly have any 
 effect on the results. The second division of the experiment dif- 
 fered from the first in only one particular, the subject was in- 
 structed to try to hold in mind a visual image of the page as the 
 lines were repeated. 
 
 In order to be able to tell what part of the passage corre- 
 sponded to any given point in the records, just as the last word of 
 each line was repeated by the subject, the experimenter closed or 
 opened alternately an electric switch controlling an indicator. The 
 movements of the indicator were recorded in the plate at the side 
 of the corneal reflection, as described above. A record (of the 
 subject who was unfamiliar with the object of the experiment) is 
 reproduced on Plate IV, No. 6. As may be noted there is consid- 
 
58 THE PSYCHOLOGY OF READING 
 
 erable eye movement during the recitation. The movement, more- 
 over, is not that of mere unsteadiness of the eye, but there are sev- 
 eral changes of position which closely resemble the true interfixa- 
 tion movements. In this case the subject was 'holding in mind' 
 the visual image of the page. 
 
 When the passages were merely repeated, the movement was not 
 so marked, (the motor discharge moving perhaps into other paths 
 more naturally, e. g., the auditory-motor system.) A difficulty in 
 recalling parts of the passage quite unconsciously contributed to 
 the value of one of the records. The hesitation and attempt to 
 remember accentuated the eye's movements. Several changes of 
 fixation nearly equaling in extent a movement from the end of one 
 line to the beginning of the next are made during this confusion. 
 This record and also the records of this subject when instructed to 
 keep in mind the appearance of the page can unfortunately not be 
 directly reproduced because of the faintness of the negative. A 
 careful pen drawing of the distinct parts of the records may 
 be found on Plate IV, B. The second, drawing (No. 2 in Plate) 
 (GOB) (subject trying to secure visual image of words of page 
 as he spoke them) shows not a little resemblance to the usual 
 fixation pauses and movements. Several long movements like the 
 usual return sweeps of the eye are of most interest. These long 
 uninterrupted movements are introspectively the most prominent 
 of our motor sensations. They are in fact practically the only 
 sensations of eye movement which most persons have. 
 
 Individual differences are to be expected, dependent possibly 
 on the closeness with which the subject approximates the vis- 
 ualist type. As is well known, many can recall during the recita- 
 tion of a memorized passage a pretty constant image of the gen- 
 eral appearance of the page and of an occasional word or group 
 of words. The records of the second subject are more conclusive 
 than those of the first, due doubtless to an individual difference in 
 this respect. They show that, if there is a distinct change of the 
 point of attention, a corresponding eye movement may occur even 
 if the object is only imagined. Since during the period immedi- 
 ately following the momentary stimulation of the short exposure 
 with the taehistoscope the image of the words exposed is held in 
 mind for a brief moment in a way not unlike that of the above 
 experiment, it seems likely that, if in that case, too, the changes 
 of attention, — often of wide extent, — assumed to occur upon the 
 Wundtian hypothesis, did actually take place, they could not 
 fail to have a physiological effect upon the eye's fixation similar 
 to the above described motor phenomena. 
 
CHAPTER VIII 
 
 The Duration op the Fixation Pauses 
 § 1. Relation to the Number of Pauses 
 
 The more pauses there are in a line the shorter their lengths 
 on the average, and, vice versa, the fewer the pauses the longer 
 any one pause is apt to be. The number and duration of the 
 fiauses must, therefore, be considered together in an analysis of 
 their functions. Viewed simply from the standpoint of speed of 
 reading, it is in general an advantage to read a given line with 
 the smallest possible number of pauses, because while the elimina- 
 tion of a pause increases somewhat the average duration of the 
 remaining pauses, the total time for the line is decreased, or re- 
 mains constant. In the third reading of the first line of the news- 
 paper passage by H (Chart I, p. 75), for example, the average 
 time of the four fixations was 215a; in the fourth reading three 
 fixations are made with their average leng-th 286o-, but the third 
 as a whole took the same time as the fourth (860o-) ; similarly, in 
 the second reading of the second line the total time for four fixa- 
 tions was 950(7, that for the third and fourth readings in three 
 fixations each GTOo- and 820o- respectively. 
 
 Other comparisons of this fact can be made from the Chart I 
 of fixation pauses on p. 75. When, on the other hand, we regard 
 the matter with reference to the differences of words in ease of 
 recognition and assimilation, as, for instance, if we wish to deter- 
 mine what words and sentences are from this standpoint best 
 suited to be put into an elementary school reader, the length of 
 the single fixations and not their number becomes most important. 
 To refer again to the chart (I) on p. 75, although the number of 
 fixations is about the same, the fourth line has a little advantage 
 in the time that is needed for recognition and assimilation. 
 
 The average durations of the fixation pauses of five subjects (see 
 Table VIII) in reading the same newspaper passage were (Sub- 
 ject T) 160.82O- (A. D. 36.8) ; (Subject H) 216o- (A. D. 50) ; Sub- 
 ject S) 255.5a (A. D. 87) ; (Subject F) 401.9a (A. D. 163.7) ; 
 (Subject M) 196a (A. D. 79.9). In reading a scientific passage 
 taken from the Popular Science Monthly, the average for T was 
 169.41a (A. D. 53.01) ; for H 241a (A. D. 67.3), and for Subject 
 
60 THE PSYCHOLOGY OF READING 
 
 S 253.57o- (A. D. 102.7o-)/ These differences between subjects are 
 naturally very closely correlated with their differences in the ra- 
 pidity of reading, as determined by these tests and by others given 
 independently. In the accompanying charts, pp. 75-82, the dura- 
 tion of the pauses in each line and the total and average lengths 
 have been denoted, and on pp. 77-80, several lines of the news- 
 paper column are put in immediate succession for better compari- 
 son of these apperceptive differences of the subjects. In the fol- 
 lowing table (Table VIII, A-D) the average number of fixations 
 has been given with the average and median lengths for several 
 passages. The larger number of pauses in the longer lines shows 
 that the total time per line was, of course, much longer. 
 
 Table VIII 
 
 Tlie average and median duratio7is (in a) of the fixation pauses of 
 different subjects in reading the same pages 
 
 A. A newspaper column. 
 Width of line 57 mm. Height of type 1.5 mm. 
 
 
 
 
 
 
 Av. No. of 
 
 
 Av. 
 
 
 M. 
 
 A. D. 
 
 pauses per line, 
 
 Sub. T. 
 
 160.8 
 
 
 152 
 
 36.8 
 
 3.7 
 
 " H. 
 
 216. 
 
 
 220 
 
 50 
 
 3.9 
 
 " S. 
 
 255.5 
 
 
 238 
 
 87 
 
 5.5-6 
 
 II p 
 
 401.9 
 
 
 240 
 
 163.7 
 
 5.4 
 
 " M. 
 
 196 
 
 
 160 
 
 79.9 
 
 7 
 
 
 
 B 
 
 . 'A License to teach.' 
 
 
 
 Width of line 
 
 96.5 mm. 
 
 Height of type 1 .5 
 
 mm. 
 
 
 
 
 
 
 Av. No. of 
 
 
 Av. 
 
 
 M. 
 
 A. D. 
 
 pauses per line. 
 
 Sub. H. 
 
 211.6 
 
 
 186 
 
 68.7 
 
 6. 
 
 " S. 
 
 200.7 
 
 
 186 
 
 61.0 
 
 7 
 
 " F. 
 
 295.3 
 
 
 277 
 
 91.3 
 
 8.5 
 
 " St. 
 
 275.9 
 
 
 257 
 
 67.2 
 
 5.3 
 
 C Athletics. 
 Width of line 107.5 mm. Height of type 1.5 mm. 
 
 Sub. St. 215.2 199.8 54.3 5.6 
 
 D. A Bcientific article. 
 Width of line 107.5 mm. Height of type 1.5 mim. 
 
 Av. No. of 
 A. D. pauses per line. 
 
 58.0 6.5 
 
 67.3 5.4 
 
 102.7 7.5 
 
 ^ The times given for these pauses are accurate within ten sigma. 
 For purpose Of securing as many lines of reading as possible on one 
 negative, the rate of fall of the photographic plate was made very slow. 
 The long duration of the fixations makes such determinations sufficiently 
 accurate for the purpose in hand. With a faster fall, the records can 
 be read to l<r. 
 
 
 Av. 
 
 M, 
 
 Sub. T. 
 
 169.4 
 
 150 
 
 " H. 
 
 241 
 
 218 
 
 " S. 
 
 253.5 
 
 200 
 
DURATION OF PAUSES 61 
 
 § 2. The Length of Fixation and the Reaction Time of the Eye 
 
 The length of fixation pauses must also be conditioned by the 
 reaction time of the eye. Dodge^ found the mean reaction time 
 for two subjects to be 162c7- and ITOo-. The minimum reactions 
 were as short as 140o-. As remarked by Huey,- however, "In 
 reading it may well be . . . that the reaction is somewhat . . . 
 quicker, (Huey's own determinations were longer than the above 
 due possibly to the weight of the eye cup and pointer, i. e., 206.7a 
 M. V. 20.7o- for visual stimuli) because of the heightened readiness 
 for the stimuli produced by 'associative expectancy' and, perhaps, 
 also, by the stimulus to reaction having been actually noted in the 
 remote periphery even while the eye is fixated in the preceding 
 line ; ' ' and the result of his measurements of the average length of 
 pause for one passage 190.9o-, M. V. 48.6o-; for another passage 
 108. 3cr, M. V. 32. 6(7 ; and for a third passage ISBo- support this 
 supposition. While the average times of the reading pause are, 
 however, as above determined in the experiments, in excess of the 
 minimum reaction time, they are not infrequently much shorter, 
 due doubtless to the above explanation, which indeed amounts to a 
 distinct difference in the method of experimentation. 
 
 ^ Psycliological Review. Sept., 189&. 
 ^American Journal of Psychology, XI, 295. 
 
CHAPTER IX 
 
 The Span of Attention and the Duration of Pauses 
 
 It is natural to suppose that the length of time the eye spends 
 at any one part of the line is somewhat indicative of the apper- 
 ceptive qualities of those words. This is, however, not literally 
 true. The attention is often distracted by matter lying on the ex- 
 treme periphery and by words immediately above and below in other 
 lines. It has already been argued that the eye tends to follow 
 each shift of the attention in order to bring the object nearer the 
 fovea, and that the unsteadiness of fixations is due to the acute 
 unbalance and general alertness of attention to peripheral excita- 
 tion. It is certain at any rate that within the rather long dura- 
 tion of the fixation pause the attention is, so to speak, distributed 
 over the immediate field of view. The duration of any one pause 
 need not, therefore, necessarily represent the time that is neces- 
 sary to perceive the words immediately fixated, although it may 
 do so approximately. 
 
 The first noticeable instance in point is that of the first fixa- 
 tion in each line. As may be seen from the distribution of lengths 
 of pauses according to their order in the line, the longest pause in 
 the line is very regularly found at the first fixation. Since it is 
 evident that the more difficult parts of the line do not always come 
 first, some other explanation must be sought. The fact is illus- 
 trated in the distribution in Table IX A. The median durations 
 of the first and folloAving fixations in reading several lines of a 
 newspaper column are there compared. 
 
 Table IX 
 
 The median durations {in a) of fixation pauses distributed accord- 
 ing to their order in the line 
 
 A. Newspaper column. Length of line ii6. 6 mm. 
 
 1st 2d 3d 4th 5th 6th 7th 
 
 Subject fixation fixation fixation fixation fixation fixation fixation 
 
 M AD M AD M \I) M AD M AD M AD M AD 
 
 T 190 26.4 148 16.5 151 -29.4 152 54.1 
 
 H 240 43.3 190 27.3 205 32.9 220 41.4 140 40. 
 
 S 280 62.9 306 72.9 194 127.7 238 65.1 121.5 65.5 243 61.4 144 45. 
 
 B. Scientific passage. 
 
 M AD M AD M AD M AD M AD M AD M AD 
 
 T 160 j80. 170 34.3 160 47.1 140 20. 150 27.1 155 23. 140 13.3 
 S 160 58. 200 64. 200 32. 200 20. - 200 202 190 122 370 214 
 
THE SPAN OF ATTENTION 63 
 
 It will be noted that the median durations of the first pauses 
 in the case of subjects T and H are longer than those of the succeed- 
 ing pauses. The most reasonable explanation of this fact seems to be 
 that the eye here takes a much more general survey of the line as 
 a whole than it does ordinarily in the later fixation pauses. An- 
 other explanation which suggests itself is that the attention may 
 be distributed both to the left and right in the case of the first 
 fixation. That is, as the eye seldom comes back to the very edge 
 of the line, matter lying both to the left and right of the line of 
 regard must be perceived together. This would not be the case in 
 the following fixations. The direction of attention after the first 
 pause would be always in the line of the eye's advance. It is pos- 
 sible to test this last hypothesis in the cases much more frequent 
 in some subjects than others where the fixation does actually re- 
 turn to the very edge of the line. The first fixation ought in tliese 
 cases to be shorter, that is, of the same average length of the other 
 fixations in the line. But an examination of the records shows that 
 this is not the case. As will be further noted in another connec- 
 tion it IS, therefore, more in accord with the facts to suppose that 
 in all fixations the perceptive process is generally concerned "with 
 the matter lying on either side of the point immediately fixated 
 and that when the first pause is longer than the average of the 
 other pauses in the line it denotes a more general survey of the 
 whole line. 
 
 It will be further noted by a comparison of Table A (the news- 
 paper readings) that in the case of subjects T and H there is a 
 second increase in the distribution of the length of pauses in the 
 third and fourth fixations. This may be considered to indicate 
 that the 'general survey' of the first pause covered the end of the 
 lines only superficially, and the last two fixations are, therefore, 
 somewhat longer than the second or intermediary but not as long 
 as the first. In the case of the third subject there is a partial al- 
 teration of long and short pauses. This on a basis of the previous 
 hypothesis may be taken to indicate that the extent or span of at- 
 tention is usually smaller than in the other two subjects. This, of 
 course, has already been indicated by the larger number of pauses 
 per line, but the one may not improbably be the cause in part of 
 the other. The subjects who can make at the beginning and else- 
 where a more general survey of the line than is done in the inter- 
 mediary pauses may for that reason read the remaining part of 
 the line with fewer pauses. Although the number of cases meas- 
 ured is small, it is further interesting to note that this 'wider 
 spanning' of the attention, if this is the true explanation of the 
 longer initial pauses, is made more frequently by the more rapid 
 
64 TEE PSYCHOLOGY OF READING 
 
 readers. It seems to point to another differentiation and cause of 
 the individual variations in the velocity of reading, and it might 
 be naturally expected that a certain amount of 'prevision' would 
 contribute to the grasping of larger units of thought than when 
 the attention is more contracted. As just said, it may be for this 
 reason that the rapid readers need to make fewer pauses pej^ line. 
 The facts, in a word, justify the conclusion, that the attention is 
 much more widely 'distributed' at some pauses than at others— of 
 this there is really no question, the different extents of line be- 
 tween the fixations must be explained on some such hypothesis, — 
 and, secondly, that the frequent occurrence of a longer initial pause 
 shows that the lengths of the fixations are not solely dependent 
 upon the apperceptive character of the words or phrases more im- 
 mediately attended to, but that they are presumably influenced by 
 other parts of the sentence, i. e., the assimilative attention is more 
 widely extended in some pauses than in others. The explanation 
 given above for the so-called 'shifting' movements in the direction 
 of the eye advances, is none other than this, i. e., the attention is 
 ahead and 'pulling' the eye along. 
 
 In the distribution in Table IX B of the durations of the fixa- 
 tions in a longer length of line the first fixation is not longer than 
 the average. It is to be noted here, however, that at the beginning 
 of nearly every line there is a 'refixation.' This means that in 
 the first fixation the attention is directed to the left and towards 
 the edge of the page, and cannot, clearly, then, perform the func- 
 tion of scanning the line as a whole to the right. It may also be 
 that in this case the involved style necessitated a more contracted 
 span of attention and gives little opportunity for grasping the 
 larger units of thought. If, however, the refixations are counted 
 in together with the first fixation pause, — and there is good reason 
 for doing this, since only when taken together do they represent 
 the full amount of time spent at the beginning of the line, — the 
 above discussed phenomenon holds equally good of this latter pas- 
 sage. One other factor must also be taken into account, that is, 
 the greater length of text-line in the second passage. It is quite 
 possible that a longer initial pause in a line of this length would 
 be of but comparatively little value because, first, the perception of 
 the extreme periphery would be much less distinct, and, secondly, 
 because the greater separation in the long line of those phrases 
 and clauses which belong together may make such prevision rela- 
 tively more difficult. (See Chapter XII for further elaboration of 
 this point.) The question as to which one of these factors, style, 
 length of line, etc., gives the correct solution to the problem is 
 finally discussed in Chapter XII, where new data concerning the 
 phenomenon is presented. 
 
THE SPAN OF ATTENTION 65 
 
 § 1. Experimental Isolation of the Beading Pause 
 
 Although, in view of the preceding, perception cannot be con- 
 sidered as limited to the period of fixation, valuable data has been 
 secured and may yet be expected from the experimental isolation 
 of the reading pause. A method for further investigation which 
 for some purposes has several advantages over that of the fall ex- 
 posure apparatus and the tachistoscope is the following. If words 
 and short sentences are arranged, as the following series of non- 
 sense syllables, p. 66, at such a distance apart that when the eye 
 is fixating one, the other is not within the range of clear vision, 
 there will be but a minimal disturbing influence of irrelevant mat- 
 ter, and the length of the pause will be approximately that which 
 would be made in actual reading. As the length of any pause de- 
 pends in part upon the difficulty of recognition and perception, a 
 comparison can be made of the words and phrases in this respect, 
 and, secondly, the extent of the line that may naturally be per- 
 ceived together when the eye is freed from other distracting ele- 
 ments at the periphery and in the lines above can be determined. 
 For, if the group becomes too large, a second fixation will be made. 
 As an illustration a print is made on Plate V, 1, of the length of 
 time spent by the eye in the perception and assimilation of the fol- 
 lowing series of nonsense words. The eye, as will be seen from the 
 print, followed the arrangement of words and made but five pauses 
 in each horizontal line. 
 
 Several factors which may be differentiated by further experi- 
 ments evidently enter into the assimilation or apperception of the 
 nonsense syllables. A study of these elements should throw light 
 on the process of word apperception in general. The length of the 
 pause is due in part to the sequence of letters. If that is the nor- 
 mal or more common sequence of words, such as 'werq,' 'wopi,' 
 'gero,' 'apli,' *enfa,' the association process is less interfered with; 
 such combinations as 'ciuo,' 'weao,' 'dpiu,' disappoint the associa- 
 tive expectancy and the time taken for perception is longer. A sec- 
 ond and perhaps more important element is that of the ease of pro- 
 nunciation. The syllables last mentioned, 'ciuo,' 'dpiu,' 'weao,' 
 etc., as well as others like 'fiea,' are evidently much more difficult to 
 pronounce at sight than ' wopi, ' ' gero, ' ' werq ' and even ' f hwe. ' Ar- 
 ticulation or some form of motor expression is undoubtedly one of 
 the factors which determine the length of the fixation pauses in gen- 
 eral and its influence presumably varies with different individuals, 
 the rate of reading of some apparently being not faster than some 
 form of motor expression. I have made comparison, for example, 
 of the tables of the rate of reading of twenty or more subjects as 
 
66 TEE PSYCHOLOGY OF BEADING 
 
 given by Quantz^ and Huey^ with a view to this question. The 
 differences in the speed of reading silently in the usual way among 
 one class of subjects differ in a ratio of 2.5 to 1,^ i. e., the fast- 
 est are 2.5 times as rapid as the slowest. But in reading aloud 
 they are only about a third faster. This is, of course, due in part to 
 the real physiological difference in the time of articulation, but it 
 seems also to show that the fastest readers had curtailed the motor 
 expression most in their usual method of reading. 
 
 LISTS OF NONSENSE WORDS 
 
 SFAG TUIE WERQ TIOE ERSA 
 
 ASSE GERO BERA YUIE GHEI 
 
 REWO PIEA WOPI FHWE EIHU 
 
 CIUO EI DO APLI TUG A EOPU 
 
 VOEI SAEI TIUE EUFA OIJU 
 
 RUIA WEAO DPIU BEWA VIUE 
 
 ^ 'Problems in the Psychology of Reading.' Psychological Review, Mono- 
 graph Supplements, Vol. 11. 
 
 ^'Op. cit. 
 
 ^ These calculations are not given by the authors quoted but are made 
 by the writer. 
 
CHAPTER X 
 
 The Number Span of Attention 
 
 A similar experiment has been made with numbers. .Numbers 
 for reasons which are discussed later are well suited to an experi- 
 ment of this sort, chiefly because in reading them the span of at- 
 tention is more limited to the area of the immediate fixation pause 
 than in reading of words or even of nonsense material. 
 
 The results of the experiments as given below show that it takes 
 an appreciably longer time (2101a) to read a series of twenty-four 
 numerals when grouped as fours than when grouped as threes. The 
 reading of larger groups of five and six figures demands a propor- 
 tionately longer time. The reason for this fact is, it is believed, 
 that the groups of. numerals are not read as wholes, as words 
 usually are, but by their separate digits. The process of analyzing 
 and recombining the larger groups makes a greater demand on at- 
 tention which, as denoted by the greater duration of the fixation 
 pauses, is proportionate in almost geometrical ratio to the number 
 of digits in the groups. 
 
 Why should numerals differ from words in this respect? The 
 pathological cases cited by Brandenburg, as Storrung^ has shown, 
 point to a difference in the central connections between words and 
 numerals, and as well between the names of numerals (i. e., written 
 with letters) and the figures themselves. The subjects had lost 
 completely the capacity to read words and single letters, but they 
 could read numbers 'ganz fliessend,' even up to five or more digits. 
 The explanation of the difference is that if the subject could read 
 a single number he could easily read several, as it would be 
 simply a matter of position; whereas if the subject could read let- 
 ters it would be no indication that he could read words. (The 
 now well-known cases cited by Leube and Lowenfeld, in which the 
 patients were not able to read letter by letter but only when the 
 total word was seen, show "that the Gesammtbild of the word can at 
 all events play an important part, and that reading does not need 
 to proceed letter by letter."- A series of figures, to put the mat- 
 ter another way, cannot acquire the independent significance which 
 words do, and that because the succession of letters is always the 
 
 1 Storrung, Vorlesungen ilber Psychopathologie, 165 fF. Leipsig, 1900. 
 « Storrung. Op. cit , 163, 164. 
 
68 THE PSYCHOLOGY OF BEADING 
 
 same and unchangeable in a given word, but in a series of numbers 
 the possible variation is almost infinitely large. It is to be ex- 
 pected, therefore, that in reading numbers the attention is con- 
 cerned with much smaller units than in reading words. And this 
 proves to be the case. 
 
 The method of experiment has already been indicated. The 
 results of one subject will be discussed for illustration. In the 
 first record twenty-four figures were spaced in groups of threes. 
 There were two lines of these groups. In the third line the digits 
 were printed consecutively without spacing. In the second record 
 the same nimiber of numerals was grouped by fours and a similar 
 unspaced line of twenty-four was printed below it. The third 
 record was of mixed groups of numerals of from two to six digits. 
 The original pages were typewritten. They are reproduced here in 
 as nearly similar type and spacing as possible. The photographic 
 records are reproduced in PL V, Nos. 3 and 4. No. 2 is the read- 
 ing of an unspaced line of numerals by another subject. 
 
 473 869 673 695 938 593 583 984 
 
 948 374 567 284 736 927 684 734 
 
 473829483746583928473645 
 
 748392837465738293847564 
 
 5849 9426 5748 2938 8493 75849^ 
 
 758493827364537284938475 
 8695 3847 7483 5928 5748 8493 
 
 87589382736452 7384958773 
 
 ^ This last group through a mistaike of the typewriter was made of 
 five digits instead of four. 
 
NUMBER SPAN OF ATTENTION 69 
 
 The time taken to read the two lines of eight groups of three 
 figures was practically the same (3625or in one case, 3610or in the 
 other, or an average of 452a for each group of three). The time, 
 however, taken to read the six groups of four digits was 5727o- or 
 an average for each group of 955o-. The results of the experiment 
 came as a surprise to the subjects (the results of two other sub- 
 jects and with different series were similar). Introspectively it 
 seemed to the subjects that the groups of threes and fours were 
 perceived as wholes and with equal facility. The next part of the 
 experiment further illustrates where the difficulty lies. The sub- 
 jects were instructed in reading the unspaced lines to group them 
 in reading in the first case by threes and in the second case by 
 fours. This the subjects believed that they did successfully. The 
 only difficulty noted introspectively was that the 'eye ran too far 
 ahead of the group that was to be singled out, necessitating some 
 backward movements." The record shows that this is somewhat 
 the case particularly in the grouping by fours. The result was 
 similar to that in the spaced groups. The total time taken for 
 reading the twenty-four letters by threes was 5068or, by fours 6247o-. 
 The exact series and arrangement of the third experiment is 
 given below, and below it^ the times taken to read each group. The 
 photographic record of this experiment was omitted by mistake 
 and is inserted at the end of Plate Y as No. 14. 
 
 287 4936 892 58763 
 
 689243 832 9477 28 
 
 3344 92 899 43621 
 
 93 482 998763 29 
 
 Times in a taken to remd the above numerals grouped as above 
 ' Qn next page. 
 
70 TEE PSYCHOLOGY OF BEADING 
 
 443 824 451 964 
 
 1658 410 824 410 
 
 857 284 451 1148 
 
 451 476 1895 
 
 The reason which seems to the writer to underlie these differ- 
 ences in the time taken to read two digits and six digits has al- 
 ready been suggested. It does not ordinarily take any longer 
 (see charts below) to read a long word of ten letters than it 
 does to read a short word of three, in fact, as will be seen, the smal I 
 word may cause the greater difficulty. This seems to justify but 
 one conclusion, the words are read as wholes ; and the fact that there 
 is a constantly increasing ratio of time taken in reading the larger 
 numbers shows that the latter are not so read. The innervation is 
 split up and, as said, it is first necessary to analyze and then com- 
 bine into wholes. The experiment with the unspaced line of nu- 
 merals further illustrates that under the exactly similar circum- 
 stances it is harder to pick out a group of four numbers and then 
 combine them than it is to analyze and unify a group of three. 
 
 If the attention is in these readings analytic, so to speak, this 
 should supply a fair test of the above described hypothesis that 
 the changes of the attention in perception will cause corresponding 
 movements of the eyes even within the small limits indicated. Are 
 such changes in the location of the fixation pauses to be discov- 
 ered in the records? For purposes of comparison prints of two 
 of the negatives have been reproduced on Plate V, one of the read- 
 ings by threes and the other of the groups of four digits. The 
 prints do not, unfortunately, show the minute changes as well 
 as the negatives themselves. The small shifts of the bright point 
 can be easily detected in the original records, by small differences 
 in the intensity of the impression on the negative. If the eye re- 
 mains for even an almost unmeasurable fraction of a second longer 
 in one place than in another, the film is sensitive enough to cause 
 a differentiation in the 'depth' of the impression. The general 
 penumbra tends to obscure this, but the difference can be clearly 
 detected when the record is held before a bright light. These minute 
 but exact differences are quite indiscernible in the prints, espe- 
 
NUMBER SPAN OF ATTENTION 71 
 
 ;;ially as the records are reduced in making the plates. However, 
 a careful comparison of the prints of the readings of sentences 
 and those of the numerals will show unmistakably that the fixa- 
 tion of the eye in one case is usually without noticeable shifts, 
 whereas from two to five separate and distinct breaks can in most 
 cases be detected in the fixations of the numerals. When the num- 
 ber is over four, one or more of the movements is regressive. In 
 the record representing the reading of the unspaced line of twenty- 
 four digits no less than thirty changes or shifts of the bright point 
 can be counted. This shows that the eye necessarily moved back 
 and forth somewhat in analyzing and combining the separate digits 
 into the prescribed units. And it is the contention of the writer 
 that if a similar process of analysis and comMnation were gone 
 through with hy the attention in reading the words and domineer- 
 ing letter complexes, etc., there wotdd necessarily he some similar 
 indication in the angidar displacement of the line of regard. 
 
 These differences appear in the reading of numerals only be- 
 cause the process of perception in the latter case is materially dif- 
 ferent from that which takes place in the reading of words. In 
 reading groups of three or more numerals the process of innerva- 
 tion is broken into smaller divisions;^ in the reading of familiar 
 words and phrases, it is a unit coextensive with the whole imme- 
 diate area of the fixation. Groups of numerals (except in the 
 case of familiar dates, as 1905, which are exceptions that prove the 
 rule) are so seldom made up of the same series of digits that they 
 never acquire a ' Gesamtbild, ' and are therefore recognized as totals 
 only after a successive perception of the digits has taken place. 
 
 The case is exactly similar in the reading of groups or lines of 
 consonants; where vowels are added as in nonsense words, the 
 
 ^ Although records were also made at the time the above were taken 
 to determine whether even groups of two are regularly recognized as 
 wholes, they are not accessible, and there has not been opportunity to 
 repeat them. The records of the readings of Nos. 28 (2nd line, No. 4) ; 
 92, (3rd line. No. 2); and 93 (4th line. No. 1), (see PL V, 14), as well as 
 the number of separate movements in the groups of four seem sufficient 
 evidence that the single digit is often the unit. 
 
 The results for at least these subjects evidently do not agree with the 
 conclusions of the tachistoscopic experiments quoted earlier that "on the 
 average consciousness can at one time grasp four numbers, three to four 
 letters," etc. The discrepancy is found also in the perception of words 
 but not to the same extent for the same subjects. If the interpretation 
 given above of the purpose and significance of the fixation pauses is cor- 
 rect, it must presumably be conceded that (1) the results of tachisto- 
 scopic experiments cannot be applied directly to reading, and that (2) a 
 greater dependence is put upon the visual memory image of the percep- 
 tion in the former case than in the ordinary practice. That is, either 
 (1) the actual extent of matter assimilated is greater in the short expos- 
 ures, and (2), since the text is no longer before his eyes, the subject has 
 learned from a little practice to depend more on the visual memory 
 images than is customary in ordinary reading. 
 
72 THE PSYCHOLOGY OF READING 
 
 span of attention is somewhat widened by the occasional recogni- 
 tion of a familiar syllable. 
 
 This is the significance of the difference in the time of percep- 
 tion of words of increasing complexity, as determined by Cattell, 
 and the numbers of increasing digits, as determined by Fried- 
 rich. Cattell found {Phil. Studien, III) that the recognition of 
 one short English word from a set of twenty-six took 360a, a short 
 German word 367o-, and a long English word 315a. Here was an 
 increase of but 15cr. Friedrich found (cf. Jastrow, 'The Time Rela- 
 tions of Mental Phenomena Table,' pp. 32-33) the reading time of 
 one place numbers to be 318cr (simple reaction time 186a), for three 
 place numbers 397a, and for five place 697a. The increase in the 
 complexity of the words was much greater than that of the numer- 
 als, but the time taken was shorter because the plurality was more 
 of a unit in the case of words than that of numerals, which could 
 not be grouped, for the reasons assigned. It will be shown in the 
 following charts (Chapter XI) that in many instances the long- 
 word took no more, and perchance even less time than the short, 
 because by means of abbreviated cues its real conscious significance 
 was no more complex than the short word, and not only did not 
 demand an analysis by a fluctuation of attention, but its abbrevi- 
 ated cue made its recognition even quicker than in the case of the 
 short word. The original results of Cattell for lines, figures and 
 letters illustrate and substantiate this conclusion as well as any. A 
 series of short lines is most easily grouped, secondly figures, and 
 least easily nonsense words, i. e., disconnected letters. The rate of 
 error in the letters increases most rapidly because the natural word 
 association is at every increase more difficult, whereas the analysis 
 of the larger groups of numerals is not so difficult because we are 
 accustomed to divide them into sections, and secondly, there are 
 no interfering associations as in letters. The increase in time 
 although marked is for this reason much less in the case of numbers. 
 The lines are homogeneous and for this reason more of them can 
 be perceived at one time than in the case of either letters or 
 numbers. 
 
 The question of the distribution or fluctuation of attention sim- 
 ply resolves then into one of the success or failure in forming uni- 
 tary groups. 
 
 The experiments of Hylan on the distribution of attention 
 {PsycJiological Review, 1903), reinforce this point. We perceive 
 successively that which has no connection, but as soon as connec- 
 tion appears between the isolated perceptions, they tend to be uni- 
 fied; our perception is no longer successive. It appears (Cf., p. 
 500) that the perception of isolated figures of increasing com- 
 plexity was successive, as denoted by the regular increase in the 
 time of reaction, but that with five figures this fluctuation and 
 counting process could not go on in the time given. What exped- 
 ient was then unconsciously adopted? They were grouped rhyth- 
 mically into groups of one and two, or three or four, with the re- 
 sult that the reaction time M^as shortened. This grouping was not 
 made by all subjects, and with the natural result that when the 
 numbers were successively perceived, the reaction time was uni- 
 
PLATE V 
 
NUMBER SPAN OF ATTENTION 73 
 
 formly longer. Here again this individual difference in the in- 
 ability to group into units appears, and it would be interesting to 
 know if this inability was not closely correlated with the rate of 
 reading of the different persons. 
 
 The general conclusion is that there is strictly no such thing 
 as a distribution of attention to disparate and unassociated things, 
 and that such distribution is a psychological and logical impossi- 
 bility. When things are isolated in association there is a 
 fluctuation of attention between them. When by means of 
 the many tricks of association they are grouped into a conscious 
 unity, this fluctuation is no longer necessary. The whole of the at- 
 tention is given to the process until another isolated factor is in- 
 troduced, and the question is simply again renewed whether the 
 fluctuation of attention must occur to keep them both before con- 
 sciousness, or whether the new factor can in some way be linked to 
 the previous complex. 
 
 DESCRIPTION OF PLATE V. 
 
 1. Reading of the list of nonsense words printed on page 66; 5 words 
 to a line. 
 
 2. Reading of series of unspaced numerals. 
 
 3. (No. 70 A.) Reading of lines composed of eight groups of three 
 digits (spaced in first two lines, unspaced in third line). (See Chart, 
 p. 68.) 
 
 4. (No. 70 B.) Reading of lines composed of six groups of four digits 
 (spaced in first line, unspaced in second line). (Cf. Chart, p. 68.) 
 
 5. No: 34 A.) Proof-reading. 
 
 6. (No. 35 A.) Proof-reading. 
 
 7. (No. 33 A.) Reading aloud or with articulation 'St. Petersburg,' 
 etc. Subject H. Cf. Plate II, No. 5, for silent reading. 
 
 8. (No. 42 A.) Reading aloud 'A license for promotion,' etc. Subject 
 S. Cf. Plate II, No. 6, for silent reading. 
 
 9. (No. 45 A.) Subject D. Reading from child's book. 
 10-13. Children's Reading. 
 
 10. (No. 72.) Child, age 9. 
 
 11. (No. 25.) Child, age 10. 
 
 12. (No. 21.) Child, age 11. 
 
 13. (No. 22.) Child, age 11. 
 
 14. (No. 71.) Reading lines composed of groups of two to six digits. 
 
CHAPTER XI 
 
 The Location of the Fixation Pauses 
 
 § 1. The Location of Pauses in Ordinary Beading 
 
 The experimental methods which have been hitherto employed 
 in investigating the reading process have been too inexact to plot 
 with any considerable accuracy the parts of the lines directly fix- 
 ated. It is on this account principally that little is known of the 
 reasons why the fixations occur at one point in the line rather than 
 at another. As preliminary to a discussion of these conditions, I 
 have marked for examination and comparison the locations of the 
 fixation pauses in the readings of several subjects and of several 
 kinds of subject matter and line arrangement. (See Charts I-XI.) 
 The general method of measurement has already been indicated 
 above. A vertical line, as |, drawn at the letter or spacing de- 
 notes the 'Blickpunkt' of the eye in fixation; when the fixation is 
 unsteady or shifts, a bracket, [, denotes the starting point and a 
 parenthesis, ) , the end of the movement. When the shift is from left 
 to right, the order of the signs is [), when from right to left it is 
 ( ] . When the eye simply oscillates back and forth during fixation 
 the limits of the movement are then marked by two brackets [ ] . 
 Sometimes the first fixation falls farther within the line than the 
 second (cf. Refixations and Regressive Movements). This reversed 
 order is denoted by the letters a, b, c. The duration of the pauses 
 in sigma is placed as an index number at each sign, as 
 
 1225 r330\ / "I '-^20 1230 1410 r210"l 
 lb W ) \ Jc Id |e Lf J 
 
 This series of signs is made use of to indicate the following 
 series of movements and fixations. The first fixation (a) fell short 
 of the end of the line; during the pause of ^d sigma there was a 
 gradual 'shifting' from left to right within the limits marked by 
 the brackets and parenthesis ; when the line of regard had reached 
 the point denoted by the parenthesis, there was a rapid movement 
 made back towards the left to the point marked by the straight line 
 |b; the eye paused here 225o-; a forward movement was then made to 
 the place indicated by the bracket ]c. During this pause of 220a 
 the eye gradually shifted to the left as far as the parenthesis. The 
 
LOCATION OF PAUSE 8 75 
 
 next movement brought the line of regard to |d. This fixation was 
 maintained without unsteadiness or shifting; the next advance was 
 to |e at which point the eye paused for 410o-, the sixth fixation 
 pause was made at f. Here the shifting of the eye was made now 
 to the right and now to the left, for simplicity the limits only of 
 this shifting in two directions are given, and they are marked by 
 the two brackets facing each other. With this explanation it is be- 
 lieved no difficulty will be found in interpreting the signs. 
 
 At the right side of some charts a table is given of the number 
 of pauses. The averages for the whole page are given at the end 
 of the passages. 
 
 In Chart I A several readings of a newspaper column by the 
 same subject H are arranged line by line for comparison. Only 
 three lines of the first reading are given. The second reading was 
 made a month after the first. The others were made in immediate 
 succession after the second. Chart I B shows the first and fifteenth 
 readings by another subject of this passage. 
 
 Chart II shows the location of the fixation pauses, etc., of five 
 different subjects, T, H, S, F, and M, in reading this same news- 
 paper column (of the day on which the experiments began). These 
 readings are arranged line by line for better comparison. 
 
 Charts III to VI, (pp. 81-82), are the readings of four subjects, 
 T, H, S, and St, respectively. The passage is one clipped from a 
 current number of the Popular Science Monthly. 
 
 Charts VII to XI, (pp. 87-88), are the readings of five subjects 
 of another passage from a school manual. 
 
 THE LOCATIONS OF THE FIXATION PAUSES 
 
 Successive Readings by the Same Subject of a Newspaper 
 -, » Column „ 
 
 I CHAET I A I 
 
 "5 o 
 
 el o 
 
 5A ^T. PETE|RSBUR|G, Nov.] 2.|— Th|e Ad[m)iralty 6 
 
 490 140 140 2j0 
 
 13B ST. PE[TE)RSBURG, [N)ov. 2.|— The Ad[m)iralty 4 
 
 200 160 160 340 
 
 15A |ST. PETjERSBURG, Nov. [2.—) The Adm[ira)lty 4 
 
 b 
 
 440 220 200 
 
 15B ST. PETER|SBURG, Nov. 2.— The| Adm[ira)lty 3 
 
 5A has| t|elegraphed [to th)e of[fic)ers of| the Ba|ltic 6 
 
 330 320 160 240 
 
 13B has[ telegra)phed to the| o(fficers] of the [Bal)tic 4 
 
 340 170 160 
 
 15A has] telegraphed to] the officers o[f) the Baltic 3 
 
 360 190 270 
 
 15B has teleg|raphed to th[e offi)cers of t|he Baltic 3 
 
76 THE PSYCHOLOGY OF BEADING 
 
 3 
 
 5A fle6[t, )who [wer)e (left be]hind |at Vigo [in order) 5 
 
 230 180 240 260 160 
 
 13B fleet, wli(o ]wer[e )left behi|nd at Vigo [in orde)r| 4 
 
 264 176 220 <^' 362 ''^ 
 
 15A fleet, [who) |were left behind at V[igo )in [order 4 
 
 286 176 276 
 
 IBB fleet, who wer|e left |behind at Vigo in |order 3 
 
 190 240 240 130 
 
 13B that the[y) |might testify, and| who were( on] 4 
 
 264 176 209 143 
 
 15A "^^[hat Hhey might( te]stify, and [w)ho [were on 4 
 
 276 132 242 166 
 
 15B that th[ey )might tesjtify, and w[ho) were |on 4 
 
 120 220 260 160 
 
 13B theijr (way] to St. Petersbur[g, to re)mai(n] in 4 
 
 ^ 308 ■ 242 
 
 15A the[ir way) to St. Petersbur]g, to remain in 2 
 
 264 276 176 
 
 15B their wayl to St. Petersburg,! to remain lin 3 
 
 13B P|aris. 
 15A j Paris. 
 15B IParis. 
 
 I B AND C 
 
 First and Fifteenth Reading op Same Passage by Another 
 
 Subject 
 
 CHART I B 
 
 416 442 136 340 
 
 [ST.] PETER(SB]URG, Nov|. 2.— (Th]e Admiralty 4 
 
 314 144 306 238 238 263 136 
 
 [ )|has telegra[phed) I to th[e of)ficers| of th|e Baltic 7 
 
 ' 280 '' 823 *= 176 * "^ 218 218 316 
 
 |fleet, who] w(er]e left behind [at) [Vigo i|n order 6 
 
 226 333 136 198 226 243 144 
 
 [th)at they |might| testity,| and| who[ w)ere on| 7 
 
 407 212 148 268 74 
 
 their[ way )to St.| Petersb|urg, to re|main in| 5 
 
 [Paris.) 
 
 212 249 194 397 120 129 
 
 [ Admir)al |Kaznakoff,| a m>em(b]er| of th|e 6 
 
 342 333 231 203 
 
 Admir[alt)y Council, [ )who w|as appojinted 4 
 
 268 462 333 323 
 
 yest(6r]day upon( th]e in(te]rnational com[-) 4 
 
 328 180 876 142 342 142 
 
 (mission whic|h is to i|nquire| int(o] the [N)oirth 6 
 
 Av. duration, 225.5cr; A. D., 87<r; M., 238ff. Av. 5.4 
 
 A. D. 1.1 
 
LOCATION OF PAU8ES 77 
 
 CHART I C 
 
 360 248 120 
 
 ST. P(ET]E,RSBURG, Nov[. )2.— The [Admiralty 3 
 
 237 229 189 
 
 has [te)legraphe(i t|o the officers |of the Baltic 3 
 
 678 132 148 109 
 
 fleet, w[ho) were [left behind a|t Vigo i|n order 4 
 
 200 154 207 
 
 that [they) might t|estify, and w|ho were on 3 
 
 152 258 144 174 387 
 
 their [w)ay t|o St. Pete|rsburg, to r|emain in 4 
 
 367 ^ ^ 
 
 PariCs.) 
 
 367 b » 135 75 
 
 Ad[miral) Kaznakoff, a [m) ember o|f the 3 
 
 214 214 148 
 
 Admira[lty ) Council, who was (a]ppointe|d 3 
 
 253 162 
 
 |b [y)esterday upon the international com- 2 
 
 Av. duration, 212. 3<r; A. D., 72.2<t; M., 189. 6(r. Av. 3.3 
 
 A. D. .4 
 
 CHART II 
 
 Readings by Five Subjects of the Same Newspaper Passage 
 
 342 152 152 361 
 
 T ST. PB(TE]RSBUR|G, No[v.) 2.— Th|e Admiralty 4 
 
 490 140 140 260 
 
 H ST. PB[TERSBURG, [N)ov. 2.-]-The Ad[m)iralty 4 
 
 416 306 136 136 340 
 
 S [ST. ]PETER|S(B]URG, Nov|. 2.— (Th]e Admiralty 4 
 
 670 ^' ''^1116 646 
 
 F ST. P|ETERSBUR[G, Nov). 2.— The A[dm)iralty 3 
 
 811 148 140 171 351 163 273 
 
 M (S[T. |PE]TER|SBUR[G, )No[v.) 2.— [Th)e Ad|miralty 7 
 
 a I b (2 
 
 123 162 218 
 
 T has telegraphejd to the offi|cers of the jBaltic 3 
 
 330 220 160 240 
 
 H ha[s telegra)phed to th|e o(fficers] of the [Bal)tic 4 
 
 314 144 306 238 238 263 136 
 
 S [ ) |has telegrap[he) (d ]to th[e of)ficers| of th|e Baltic 7 
 
 " 291 ^ 179 382 ^ ® 687 ** * 529 « 
 
 F |has telegr|aphe|d to the[ o)fficers of the B|altic 5 
 
 218 195 234 226 171 179 78 
 
 M |ha|s telegraphed [to) the of|fice|rs of the |Ba[lt)ic 7 
 
 b a 
 
78 THE PSYCHOLOGY OF READING 
 
 214 162 176 162 49 
 
 T fleet, w[ho) wer|e left behin|d at [V)igo in ord|er 5 
 
 230 180 240 410 
 
 H fleet, wh(o ]wer[e )left belii|nd at Vigo in[ orde)r 4 
 
 280 323 176 218 218 316 
 
 S |fleet, who| w(er]e left behind [at) |Vigo i|n order 7 
 
 43.'; 306 270 268 188 282 364 376 
 
 F |fleet, wh|o w|ere le|ft beli|ind [at) [Vigo) in oi|rder 8 
 
 366 162 168 160 120 9fi 176 88 296 
 
 M [ )fleet[, who |wer|e left| b|ehind ]at |Vigo in| order 9 
 
 196 146 117 107 97 
 
 T that t[h)ey migh|t testijfy, and whlo were] on 5 
 
 190 240 240 180 
 
 H that th|e(y ]might testify, and| who were( on] 4 
 
 226 '' 333 *• 136 198 225 243 144 
 
 S [th)at the[y )might| testify], and| who[ w)ere on| 7 
 
 446 246 294 447 188 
 
 F t[hat t)hey [m)ight t|estify, and w]ho w|ere on 5 
 
 264 104 112 136 128 144 136 
 
 M [that t|hey [might tes]tify, a(n]d who] were( o]n 7 
 
 20n 110 160 210 
 
 T their |way to |St. Pet|ersburg, to re|naain in 4 
 
 120 220 260 160 
 
 H thei|r( way] to St. Petersbur[g, to re)mai(n] in 4 
 
 ^ 407 ■ 212 148 268 74 
 
 S their[ way )to St.| Petersb|urg, to re|niain in| 5 
 
 662 706 200 282 317 200 
 
 F [the)ir way[ t)o St. P|etersburg, to| r[em)ain| in 6 
 
 '311 114*^ 139 "= 90 *I66 '^ 114 ^ 106 
 
 M [the)ir wa|y to |St. P|etersburg, to| rema[in) in| 7 
 
 170 
 
 T ^Pari|s. 
 
 270 
 
 H [Paris). 
 
 212 
 
 S [Paris.) 
 
 432 
 
 F [Paris.) 
 
 264 
 
 M [Par)is. 
 
 7 
 
 170 160 170 
 
 T ^Ad|miral Kaz|nakoff, a member o|f the 3 
 
 270 190 180 300 220 160 
 
 H ^Ad)mir|al Kazna[ko)ff, ([a) m]emb|er of the 6 
 
 ' 212'' 249 c 194 ^ 397;d 120 • 129 
 
 S Admir)al |Katnakoff,] a mem(b]er| of th]e 6 
 
 432 482 776 188 .388 
 
 F Adm)i(ra]l Kazna[k)off, a |mem|ber of the 5 
 
 264 ■ 418^ 123 = e d gjj igg 
 
 M [ )Adm[ir)al |Kaznakoff, a mem|ber of |the 6 
 
 ^ 'Paris' of line 6 was read with line 7 by all subjects. The location 
 of the fixations and their durations are marked in both 'Paris' and 'Ad- 
 miral, it not being possible to determine how the eye divided its time 
 between the two words. 
 
LOCATION OF PAUSES 79 
 
 8 
 
 160 120 IfiO 220 
 
 T Admiral|ty Cou|n|cil, who was| appointed 4 
 
 270 bl90 ^ ^=180 300 200 160 
 
 H ^Admi|ral|ty Counc[i)l, [(who) w]as |appointed 6 
 
 342 333 ^ 231 "^ ^ 203 
 
 S Admir[alt)y Council [, )who wa|s appo]inted 4 
 
 458 364 211 188 247 
 
 F [A)dniiralty Cou|ncil, wh|o wa(s] ap[p)oined 5 
 
 270 139 196 114 106 246 
 
 M Adm|iralt|y Counci|l, who] wa|s appoint|ed 6 
 
 190 170 110 
 
 T yesterda[y) upon the internation[a)l |com- 
 
 240 240 240 130 
 
 H yeste[rda)y up (on] the in[tern]ati[o)nal com- 
 
 268 462 233 323 
 
 S yest[er)day upon (th]e in (te] national com[-) 
 
 10 
 
 280 160 210 100 
 
 T mission [w)hich is to inq|uire in|to the [Nort)h 4 
 
 200 180 160 240 
 
 H mission [whic)h is| to inqu[ir)e in[to) the North 4 
 
 228 180 876 142 342 142 
 
 S |mission whic|h is to i|nquire| int(o] the [N)orth 6 
 
 11 
 
 170 130 120 160 
 
 T [Sea af)fair, started |for| Paris toda|y. His 4 
 
 240 170 220 220 
 
 H Sea afEair,[ )star|ted for Par|is to-da[y. )His 4 
 
 12 
 
 170 160 110 120 
 
 T appoint[me)nt is welc|omed as an[ i)ndication| 
 
 280 240 230 210 120 
 
 H [appoi)ntment isl welc[ome)d [as) an indilcation 
 
 13 
 
 170 100 140 130 
 
 T [that t)here will b|e a thoro[ug)h search for| the 4 
 
 220 130 200 220 240 
 
 H [that) ther(e] wil[l be a th)oroug(h] sear[ch) for the 5 
 
 14 
 
 200 120 160 100 
 
 T facts of [th)e case.[ )The Ad[m)iral won[ ) golden 4 
 
 310 200 160 160 
 
 H [fac]ts of the case. T(he] Ad[m|iral won) golden 4 
 
 a b d c 
 
 ■ Lines 7 and 8 were read with one horizontal movement by Subject H. 
 
80 TEE PSYCHOLOGY OF READING 
 
 15 
 
 180 190 120 
 
 T o[pin)ions among the members[ o)f th|e con- 3 
 
 240 200 210 170 SO 
 
 H lopinions amolng the [members) o(f] tlhe con- 5 
 
 16 
 
 190 140 180 
 
 T ve[ntion held in| Washington in 1889 t[o )for- 3 
 
 380 200 180 
 
 H vention he(ld] in Washingto[n In 18)8(9] to for- 3 
 
 17 
 
 120 130 160 
 
 T mulate rules of| the road for[ t)he se[a). He 3 
 
 290 190 
 
 H |mulate rules [of )the road fo|r the sea. He 3 
 
 18 
 
 190 I£0 140 200 
 
 T was Rus[sia)'s repre|sentative ther|e. He w|ill 4 
 
 19 
 
 200 110 100 100 
 
 T prepare the| Russian |case and will] cite a[n) 4 
 
 20 
 
 190 140 260 210 
 
 T incident [ ) reported in t|he Naval A[nn)u[a)l of 4 
 
 21 
 
 180 130 160 
 
 T 1903. [ In) that casle thel British cruiser 
 
 
 
 CHART 
 
 II 
 
 
 
 
 SUMMARY. 
 
 
 
 Av. No. 
 
 Av. 
 Duration. 
 
 M. 
 
 A.D. 
 
 T 
 
 3.8 
 
 160.8 
 
 152 
 
 36.8 
 
 H 
 
 3.9 
 
 216 
 
 220 
 
 50. 
 
 8 
 
 5.5 
 
 255 
 
 238 
 
 87. 
 
 F 
 
 5.4 
 
 401.9 
 
 240.7 
 
 163.7 
 
 M 
 
 7 
 
 196 
 
 160 
 
 79.9 
 
LOG AT am OF PAUiiEi^ 81 
 
 CHARTS III-VI 
 
 .Readings by Four Subjects op Passage prom Scientific Article 
 
 CHART III, No. 28 T 
 
 340 260 012 210 160 150 
 
 Ev[olutio]nary study and [t)hought h|ave been h[in)dere|d by the c|onfu- 6 
 
 260 HO 160 110 140 110 
 
 sio[n of tw)o unrel|ated biological| phenomena, (|1) evolution|ary pr|ogress 6 
 
 170 160 ICO 160 230 240 
 
 or vital] m[o)tion, and (2) t[h)e origina[ti)on or multipl[ic)ation| of species. 6 
 
 160 140 160 160 160 180 160 160 100 
 
 The '|origin' |of a spec[ies) is not m|ore evolutilonjary than ]any [o)th)e]r stage 9 
 
 b a 
 340 210 160 140 160 160 140 
 
 [in] its history. The| causes of the |su|bdiv[i)sion of spe|cies ar|e not causes 7 
 
 270 210 210 130 190 
 
 [o]f vital motion; the t[wo) processes] are quite] distinct. [Th)e separation 5 
 
 160 120 380 120 130 160 
 
 of] two s]pecies is not [a) focus o]f the evoluti]onary problem;] it is a mere 6 
 
 80 200 
 
 lincident of [de)]velop [mental ) history. 4 
 
 b a c 
 
 49 
 
 No. of fixations, 49. Average duration, 169.41. A. D. 53.01 (Median, 150). 
 
 CHART IV, No. 14 A H 
 
 291 148 174 227 174 148 212 
 
 [ Evo(lu]tiona[ry )study an[d tho)uglit [hav)e been[ hinde)red by the [o)nf[u-) 7 
 
 bi b2 a c 
 
 318 169 227 174 249 227 
 
 sio[n of tw) ( ]unr[elat)ed biologica[l p)henomena, (1), e[vol)uio[na)ry pr[o)gress 6 
 
 ■ I 32 
 
 394 206 276 216 167 
 
 or v[ital )motion, and (2) th[e or(igi]nat[ion )or multiplica[tion) of specie[s.) 5 
 
 bi b2 
 
 383 394 167 166 91 
 
 The 'ori[gin' of )a species is not mo[re evolutlonar[y th)an any o]the[r st)age 5 
 
 335 429 280 302 268 
 
 in its[ his)tory. The c[auses o)f the subdivisio[n of) spec[ies are no)t caus]es 5 
 
 362 218 179 173 212 
 
 of vital [ mo)tion; the two [proce)sses are[ ) quite distinct. [ )The separ]ation 5 
 
 i'66 239 176 176 
 
 of two spec[ie)s is not a fo[cus o)f the evolutionary probl[em; it) is a ]mere 4 
 
 610 237 
 
 Mncide[nt [of d)ev(elop[m)ental histor]y. 
 
 b a b a 
 
 610 237 203 139 203 296 
 
 'Se[gre[gat)ion) is t[he) principle ]or active [cau)se of the multipli[ca)tion 6 
 
 b aba 
 
 <i24 166 212 230 16.", 396 
 
 species, [but) the natu[re a)nd cause]s of evolu[tio)nary pr[ogre]ss are not ]to 6 
 
 49 
 
 ' These two lines are read together, the eyes alternating somewhat between 
 the two during the first half of the line. The first three pauses are therefore marked 
 in each line. In the page read each pause comes under the other. 
 
82 THE PSYCHOLOGY OF BEADING 
 
 CHART V, No. 43 B S 
 
 200 120 SCO 200 200 160 160 140 100 
 
 |EJvolutioiiary( s(tud]y and t|hought| have be|eii hind|ered b|y the |confus- 9 
 
 320 190 140 140 190 190 620 170 
 
 |si[on o)f Itwo unre|lated biolo|gical phe|nomena, (1) evo[lu)tionary[ )progress 8 
 
 b a c 
 
 340 120 200 230 1120 230 
 
 [or vi]tal m|otion, and (2) t|he orjigination or [mul)tipli[eation )of [s)pecies. 6 
 
 ei e2 
 80 160 200 220 280 700 120 
 
 |The 'or[igin' of| a species is not |more| evolutionary [than any) other s|tage 7 
 
 b a 
 
 200 120 200 160 200 160 730 
 
 |in its| history. The ca|uses of the subdi|vision of sp|ecies are |not [causes) 7 
 
 b a 
 
 Average pause, 253.51; A. D., 102.7; Median, 200. 
 
 CHAET VI, No. 74 A St 
 
 250 246 205 164 160 
 
 ESvolution[ary) study an|d thought hav|e been hindered] by the con|fu- 5 
 
 164 184 283 276 131 
 
 siojn of two unrelated! biological phen|omena, (1) evolu[tio)nary progr|ess 5 
 
 312 221 160 246 193 
 
 [or vita)l motion, and (2) t[he) originatio|n or multipli| cation of species|. 5 
 
 131 144 162 238 94 283 
 
 [T|he) 'origin' of[ )a species is not more[ evolu)tionary thlan any other stag[e) 6 
 
 b a 
 
 291 291 230 217 123 
 
 in[ its) history. The ca[uses) of the subdivisi|on of species a[re) not causes| 5 
 
 136 197 303 152 116 86 221 
 
 of| vi[tal) motion; the two[ p)rocesses are quite [di|stin)ct. |The separa|tion 7 
 
 b a d s 
 
 324 206 226 225 180 
 
 |of two species is not] a focus of the ev|olutionar[y ) problem; it is [a )mere 5 
 
 176 217 
 
 [ )inc|ident of[ devjelopmental history. 4 
 
 ad c b 
 
 42 
 
 The following observations may be made of the charts in gen- 
 eral : 
 
 The first and last fixations generally fall within the edges of 
 the line, i. e., at a little distance from its beginning and end. This 
 is true in less degree for some individuals than for others, and is, 
 secondly, dependent upon the character of the end words in the 
 line. 
 
 The exact point that is fixated may be in any part of the words, 
 or in the spacing between them. It does not fall predominately in 
 the first part of words, nor does it occur more frequently in the 
 first part of the sentence than in the last, and apparently pays 
 little attention to many of the laws of apperception, or the rules 
 of the rhetorician. 
 
LOCATION OF PAUSES 83 
 
 But, as has already been remarked, the exact point of fixation 
 may in itself be of little significance, and may vary within certain 
 limits withoTit affecting the process of perception. It is important 
 mainly as denoting that the attention is 'distributed' immediately 
 about it. In other words, the process of assimilation and apper- 
 ception are immediately concerned with that part of the field of 
 perception which lies about the point of fixation. There are, how- 
 ever, certain fairly well differentiated features which decide where 
 the eye is to stop, and it will be the object of the next section to 
 point out and discuss these. 
 
 Why Does the Eye Stop at the Places Indicated in the Charts? 
 
 The purpose of the repeated readings of the same passage by 
 the various subjects is to give a wider basis for comparison and 
 analysis of the evidently complex conditions which direct the 
 peculiar alternation of movements and pauses of the eye. In the 
 case of successive readings by one subject, if the apperceptive and 
 other conditions which govern the movements and pauses are 
 at all constant, it is perhaps natural to suppose that they would 
 tend to act much in the same way when a passage is re-read, and in 
 the later reading, therefore, the fixation pauses would be formed 
 in about the same places as before. The familiarity or practice 
 secured ■ in the first reading might supposedly act in one of two 
 ways. The eye might through practice come to follow a very defi- 
 nite and comparatively invariable method of advance, with marked 
 decrease in the time taken for the pause ; or, more probably, the 
 knowledge and remembrance of what was to follow would cause 
 both fewer and shorter fixation pauses. The latter effect might be 
 minimized, if the matter read were such as to be little likely to be 
 long remembered, and secondly if several weeks' time were allowed 
 to pass between the readings. The first five lines of each reading are 
 taken for comparison. The second reading was, as stated, made a 
 month after the first. The second, third and fourth readings were 
 made one after the other in the same hour. The smaller number of 
 pauses in the second reading shows that some general familiarity 
 with the word sequence was carried over from the first reading. 
 There is also a somewhat further decrease in the succeeding readings. 
 The pauses which have been dropped out or whose work is done in the 
 later readings by other pauses may be taken to represent the least 
 essential parts of the line. Where the location of the fixations re- 
 main practically the same throughout all or a majority of the read- 
 ings, we may expect to find there the parts of the line most neces- 
 sary for recognition, and the factors which condition most effect- 
 ively the places where the eye stops. 
 
84 THE PSYCHOLOGY OF BEADING 
 
 In the reading in the case of different persons, there are also 
 many factors which would seem to work for uniformity. Although 
 naturally considerable apperceptive differences are to be expected, 
 a great deal of our reading and training, through school and col- 
 lege, etc., has been similar; the fund of 'associative expectancy' in 
 the sequence of words and phrases must be fairly common, at least 
 in the daily reading of the daily newspaper and magazine, and 
 besides the many stereotyped similarities of writing, there are also 
 certain natural and rather arbitrary signs in the forms of print- 
 ing which the eye follows. If the more or less artificial matters 
 of arrangement of type, of spacing, punctuation, capitalization, 
 length of line, etc., have an appreciable influence, it would un- 
 doubtedly be to make the alternation of movements and pauses 
 more uniform. The similarities in this respect, which can be traced 
 in the reading of different persons, will strengthen the evidence 
 for the working of definite conditioning factors. 
 
 What, then, are the facts which hold good throughout? The 
 first which strikes the attention is that the short connective and 
 non-substantive words, the prepositional phrases and relative 
 clauses, make the greatest demands upon perception. This is best 
 evidenced by the fact that they necessitate the eye's coming out 
 quite to the edge of the lines. A comparison of all the readings in 
 Charts I to VI shows that the phrases 'in order that' (third line), 
 'who were on' (fourth line), 'to remain in' (fifth line), etc., in- 
 crease the angular excursion of the eye nearly to its maximum for 
 that line. In contrast, compare the distance of the last fixation in 
 the first, second, and - s e- vonm ^tmes from the edge; the words 'ad- 
 miralty,' 'Baltic,' 'appointed,' etc., extend the space between the 
 fixations. So also at the beginning of the line, for example in Chart 
 II, subject T, compare the distance of the first fixation from the 
 end in 'telegraphed' (line 2), 'admiralty' (line S) , 'yesterday' 
 (line d), with that in the fourth line 'that they might,' etc., or the 
 twelfth line 'that there will be,' etc. That these phrases are most 
 important elements for fixation is further made clear in the re- 
 peated readings of Chart I. In line four, the eye passes with but 
 a single break from the phrase 'that they might' directly to the 
 word 'and,' but makes two fixations in the phrase 'who were on.' 
 So the preposition 'in' (line 5) in the phrase 'to remain in,' re- 
 ceives surprising attention for so small a word when it is seen that 
 all of the first part of the line 'their way to St. Petersburg' is 
 grasped in a single fixation. The decrease in pauses in the re- 
 peated readings is made most easily in the lines having a few long 
 substantive words, with greater difficulty in those lines which are 
 filled with short worded clauses and relative and conjunctive 
 
LOCATION OF PAV8E8 85 
 
 \ 
 phrases. Compare the groups of lines three and four with those 
 of one and two. This point is further illustrated by the compara- 
 tive lengths of the space between any two fixations in other parts 
 of the line. For example, again, notice, in Chart II, subjects T 
 and H, the proximity and number of the pauses to the phrases ' into 
 the north' (otnth line), 'for the facts' (end of - twelfth ^ line) , 'of 
 the case' (%i|rteenth line), 'to formulate' (liJIteenth line), 'of the 
 road' and 'for the sea 'K^Mfclteenth line), 'in the naval' (aijtfliteenth 
 line), 'In that case' ( twontiow t^ line ) . And contrast, again, the 
 greater speed of perception in the words 'has telegraphed' (line 
 2), 'testify' (line 4), 'international' (line fj) , (except in the read- 
 ing of H), 'golden opinion among the members' (lines 1^ and 1^) , 
 'in Washington' (line 1^). These examples are sufficient to show 
 that the relative pronouns, prepositions, conjunctions and the aux- 
 iliary verbs are apt to divide the process of apperception into nar- 
 rower units than in the ease of nouns and substantive words, ad- 
 jectives, etc. The explanation is doubtless that these words are 
 not associated in one phrase more regularly than in another, they 
 cannot be fused into a larger apperceptive unit, as the syllables 
 into a word form, or ' phrase whole ' ; but each stands by itself, and 
 must be so perceived. 
 
 A Mmilar frequency of fixation is to be noted about the date 
 'Nov. 2' in the first line of Chart I. The reason is evident. The 
 numeral and abbreviation also stand for separate units, must be 
 perceived separately, and cannot be easily made a part of some 
 constant unit. The experiments with lines and groups of numer- 
 als have already emphasized this fact. A similar result was true 
 in the reading of lines of letters. It would, of course, naturally 
 have been much harder to group the letters. The principle in- 
 volved is probably the explanation of the small units of perception 
 in the cases of the prepositions and conjunctions, etc. Since they 
 occur now with one word and now with another, they cannot with- 
 out danger of error be fused into larger wholes, and, for that rea- 
 son, they must, except where the context gives the connection, be 
 separately perceived. When the noun or other word implies in 
 itself the nature of the following preposition, as in 'The causes of 
 the subdivion of species' (Charts III- VI) the latter becomes an 
 integral part of the unit of perception. With the word 'cnnse' we 
 expect to have 'the cause of or 'for.' The sort of conjunctive 
 word, etc., may often be similarly supplied. 
 
 The nouns and adjectives and verbs which have been cited are 
 in contrast recognized with a wider grasp of perception and as 
 wholes. Such words as (Chart II) 'telegraphed,' 'international,' 
 'Washington,' 'representative,' are perceived at a single fixation. 
 
86 THE PSTCHOLOGT OF EEADIXG 
 
 The question may arise as to why there is more than a single 
 nsation in some of these words, as, for example, is the ease 
 in the word 'international' in Chart II, line S^ subject T, 'evolu- 
 tionary,' Charts JJ^ and Y, line 1, 'subdivision,' Chart III, line 5, 
 and 'developmental,' line 8. In the first place, it will be noted 
 that these eases are veiy much the exception, especially for the 
 rapid readers or those who have usually a wider span of attention. 
 Ordinarily these words are perceived as units, the fixation pause 
 occurring either in or near them. Charts III and VI, having a 
 larger number of nouns and substantive words, give the best op- 
 portunity for making this comparison. It will be noted here that 
 each one of the long words is perceived by a single fixation. Such 
 a phrase as 'biological phenomena' (Chart III, line 2), which is 
 perceived with one fixation, is nearly equal in extent or number 
 of letters to the working range of the field of perception as de- 
 termined by the fall exposure apparatus. This is, therefore, the 
 rule. The exceptions in which a second fixation is made need 
 probably no further explanation than that in these cases some mis- 
 conception of the thought or incomplete perception made it nee- 
 essarj- to fixate the word again. As will be further noted in the 
 ease of reading aloud, the eye often follows along as the syllables 
 of a word are successively assimilated. This is particularly true 
 of some slow readers who perhaps depend more on the motor ar- 
 tieulatory system in silent reading. Compare, for example, the 
 word eq[ui)va!le nt in line 2 of Chart X (a slow reader.) In the 
 case of the slow reader the unit is often, therefore, the syllable in- 
 stead of the whole word. 
 
 Xot infrequently the apperception of the individual or his mem- 
 ory for words and phrases allows of an exceptionally large 'jump' 
 between fixation pauses. The word * St. Petersburg, ' Chart I, causes 
 one or two fixations in the first line, whereas in the fifth line it lies 
 mainly in the field of indirect vision. In another passage (Chart 
 YII, a teacher of 'Education') the long phrase 'methods and 
 principles of teaching' occurs in the fifth and ninth lines. In the 
 fifth line it was read with the usual number of fixations, in the 
 ninth with one fixation. It is probably this skipping of familiar 
 or oft recurring phrases as much as anything which gives readers 
 the impression that they are reading but one or two words to the 
 line. In view of these many possible individual apperceptive dif- 
 ferences, the cues of context and familiarity' with subject matter, 
 it is perhaps surprising that the results of the experiments have 
 even the consistency found. For this reason the slower readers 
 perhaps best illustrate the above noted facts in regard to the im- 
 
LOCATIOy OF PAUSES 87 
 
 portanee for recognition of the various parts of the sentence. 
 (Compare for example Chart X) .^ 
 
 It is then not to be expected that the necessity for more care- 
 ful fixation of the 'transitiYe' parts of speech will appear equally 
 in all subjects or passages. The general associative Trord expec- 
 tancy vrill cause phrases to be passed over that might in the ease 
 of another reader need more careful attention. 
 
 CHAETS YII-XI 
 
 Eeadin'gs of Foue Subjects of 'A License for Promotion/ etc. 
 
 CHAET Til. I A H 
 
 II. [ A) li[ce)nse for pro[mo)tiorL may [be is)sued [to) the lio[lder ) of[ a) 7 
 teach|er's licens[e )No. 1 wtio has[ lia)d experien'ce rated as[ eq)uival[eiit) to 6 
 
 t[lir)ee [y)ears o)f New Yo[rk )city pu[b)lie sctool teaching ( (i]nclu ding[ o)iie 9 
 
 b a 
 
 ye[ar) in [the c[ity) of Ne[w )York);[ w)ho passes ajn examiniation !in the 8 
 
 b a c 
 
 pjrinciples a'nd niet|hods of [teaching, ojr in lieu [the)reof produces evi- 7 
 
 denc[e )of[ ha)v[in)g succ[ess)fuIlT pur[s)ued. siince b[eginnin'g h[is) worjk as a 10 
 
 b a 
 
 teac|h[er), in some [re) cognized [in)stitutio n of learn'ing. one or! two satis- 7 
 
 'b a 
 
 facto[ry) cjourses of s[tu)dy invoilying ijn all no[t )less than jsixty hours' 7 
 
 at[te)nd[anc)e, in pri[nci)ples and methods of te[achi)ng: [and who passes 5 
 
 b a . c 
 
 an| ac-a'demic ex[am)inatio'n in one^ of several grouJps ofi subjects as 7 
 
 a c 
 
 an|nou[nce)d [by) the Boarjd of Ex[am)iners.[ — )A gi-aduati[ng) class license 8 
 
 b a 
 
 [may )be issue;d to the [ho)lder of |a promo[ti)on license who hajs t-aught for 7 
 
 a p[er)iod r[at)ed as e(iu|ivalent to five ye[a)rs of^ New York city public 5 
 
 CHAET Till. 13 A H 
 
 ~-5 19i is." Ill 167 192 
 
 [ A lice)nse for pro[mo)tion may [bel issued to [the h[old^er of' a 6 
 
 142 161 156 Ci: ess U< 161 347 
 
 [)teacher['s) license No[. 11 who ha[s )had ex[peri)ence rat[ed) as eq[uiya)lent to' S 
 
 192 167 136 142 41.5 
 
 thre[ej y)ears of New York c[it)y public scjhool teach[ing -(in)cluding one .5 
 
 409 316 ISO 105 136 142 
 
 ye[ar in th)e city of New York;)= w[ho passes) an exam[in[a)tion[ !in) the 6 
 
 * b c e f d 
 
 229 111 161 245 ISl 179 136 
 
 [pr)inciples[ an)d methods| of teaching, or in[ lie)u [ther)eof p[ro1duces[ e)vi- 7 
 
 272 ItT 51,1 3.^3 
 
 dence of ha[ving )successfullly pursued, since [begi)nning his w[ork as )a 4 
 
 19-* 217 167 179 .^03 217 
 
 teac[her, )in some r[ec)ognized [in)stituiion of [lear)ning. one or[ t)wo[ satis- 6 
 
 2.i4 167 :;.-. 145 1T9 
 
 factory c[ourses) of[ st)udy involv[ing) in all n[ot) less t[ha)n sixty hours 5 
 
 156 -210 111 1~6 eii IbT 142 
 
 attendance.! in p[ri]nciples and methods of teac'hing[ an!d w[ho) passes 7 
 
 a b c 
 
 — — ^^ 
 
 'Especially '[w)ho [h)as I ha[d'»,' line 2. with •experi[en)ce rated [a)s; 
 who [p)assi[e)s a[n).' line 4. and ^[n) lieu thereof." line 5, with 'm[eth- 
 o)ds of teachi[n)g and successfully." 
 
 - These parentheses belong in text. 
 
88 THE P ST GEOLOGY OF READING 
 
 Summary of Chart VIII 
 N., 54. Av., 211.6. A. D., 68.7. (M., 186.) 
 
 CHART IX, 41 B S 
 
 336 204 102 J50 96 180 160 
 
 [)A license f[o)r promoti|on ma[y )be issue[d )to the h|old6r of| a 7 
 
 348 174 174 ISO 262 168 132 234 114 
 
 t[eac)her's licen|se No. 1 whlo has ha|d experien|ce rated |as equ|ival|eiit| to 9 
 
 A b C D E FGl.H 
 
 288 120 166 270 228 204 90 246 
 
 [th)ree years of |New Yor|k city public |school teac|hin|g Hinc|ludin[g) one 8 
 
 A BC D GEFH 
 
 3.S0 222 144 162 180 216 
 
 [y)ear in the city of Ne|w York)^; who pa|ss6|g an exam[in)ation in t|he 6 
 
 A B D C E 
 
 246 198 126 276 372 396 
 
 prjinciples and |methods of t|eaching, or in |lieu thereo[f )prod[uces) evi- 6 
 
 .390 132 126 186 84 144 78 
 
 denc(e ]of having s|uccessfully [pur)sued, sin|ce beginn|ing his wLork) as a| 7 
 
 72 120 186 168 228 240 366 166 
 
 [te)ac|her, in [.some recog|nized institut[io)n of learning|, one or t|wo sat|is- 8 
 
 B A 
 248 174 192 204 294 222 
 
 fa|c[tory) c|ourses o|f stud[y invo)lving i|n all not less than .sixty hours '6 
 
 B A 
 
 57 
 
 N., 57. Av., 200.7. M., 186. A. D., 61.02. 
 
 ^ These parentheses belong in text. 
 - Incomplete line. 
 
 CHART X, 49 B F 
 
 641 686 214 327 390 161 233 289 289 463 
 
 _A l[i)cense for| pr|omotion |may be| issue|d to| the| holde|r of| a 10 
 
 648 302 321 214 486 23.3 246 369 366 132 182 
 
 teacher'[s l)icens[e) No. 1[ w)ho [h)as |ha[d )experi[en)ce rated [a)s eQ[ui)v|ale|nt to 11 
 
 182 2,3.3 196 ,369 214 ,327 277 369 289 220 
 
 [t)hre[e) yea[rs) of [N)ew York [ci)ty publi[c )school[ t)each|ing (in[cl)udin|g one 10 
 
 302 258 4,34 340 226 366 233 233 416 
 
 ye[ar) i)n the ci[ty) of Ne[w )York); who[ p)ass|[e)s a[n )examinatio[n )in the 8 
 
 AS C D E Fig H 
 
 365 207 629 220 233 214 132 176 239 214 
 
 pr|in[ci)ples and m[etho)ds of teachi[n)g, or i[n ]lieu th|ereo|f produ|ce& e|vi- 10 
 
 AC D E F 
 
 333 239 107 
 
 dence[ of) having] succes|ssfully pursued, since beginning his work as a 3 
 
 N., 52. Av., 295.3. M., 277.2. A. D., 91.3. 52 
 
LOCATION OF PAUSES 89 
 
 CHART XI, 72 B S 
 
 407 245 279 309 ,3B8 
 
 |II. A| license fojr promotion ma|y be issued t[o [the lio)lder of a 5 
 
 ' . ei 92 
 
 240 220 323 304 284 191 
 
 [) [)teacher's license| No. 1 who [ha)s had experience| rated as (eq]uivalent to 6 
 
 338 17fi 294 206 260 
 
 [)three years of Ne|w York city pu|blic school tjeaching ((in] eluding one 5 
 
 278 216 lie 240 216 206 
 
 y|ear in the city of New] York); who pa|ss[es a)n exa|mination in the| 6 
 
 a b dee f 
 
 394 375 293 240 336 
 
 [princip)les and me[thod)s of teaching, |or in lieu there|of produces [evi-) 5 
 
 240 254 317 403 307 
 
 d[en)ce of having s|uccessfully pursue|d, since begi|nning his work[ as a) 5 
 
 301 202 244 324 249 211 132 
 
 teajcher, in| some rec|ognized in|stitution of [learnin)g, one or [twl)o satis|- 6 
 
 b a c 
 
 296 660 348 ' 221 273 188 
 
 factory c|o[urses o]f study involv[in)g in all not[ less th)an sixty ho|urs 
 
 188 249 282 132 
 
 |[atten)dance, in[ princi]ples and meth|ods of teaching; and who passes 
 
 b a 
 
 No. of fixations, 48. Av. duration, 275.9<t, A. D. 67.2. Median, 257.0o-. 
 
 The requirements of apperception or assimilation determine, 
 as has been seen in the comparison of nouns and phrases, the gen- 
 eral extent and position of the apperceptive unit. But, as has 
 been doubtless noted, it is apparently immaterial in most cases 
 whether the eye stops at the center or towards the end of the words. 
 Is, then, the place of fixation within these limits determined by 
 chance? The experiments of the lines of numerals and letters 
 shows that the eye is dependent, through custom, on the more or 
 less artificial spacing and grouping of words, and it is by causes 
 of this sort that the fixation is governed within the limits assigned 
 by apperception. In the first line of Chart I, for instance, the 
 pause is made in the spacing after the date and numerals, and 
 before the dash. The marks of punctuation have a similar effect. 
 Note, for instance, the last pause in line if. Charts II and III, 
 the second in Une^^, Chart III, the first in line 20, or the sixth 
 pause in line^even. Chart VII. 
 
 In the case of some readers, the marks of punctuation, and the 
 like, prove greater 'obstacles' to the eye's movement than in others. 
 They often cause a break in the movement where undoubtedly there 
 would be none made for purely perceptive reasons. Bracketing, 
 quotation marks, and particularly any peculiarity in printing, may 
 do this. For instance, in pages of proof reading, where the pauses 
 are, as in reading aloud, much more frequent and closer together, 
 a false alignment, etc., will often 'halt' the eye. The proof 
 reader is on the watch for these inaccuracies : but the effect is often 
 
90 TEE PSYCHOLOGY OF READING 
 
 similar in the usual reading, although, of course, the mind is not 
 ordinarily conscious of it any more than it is of the fixations in 
 general. 
 
 Other factors which may govern fixation are the capital and 
 'domineering' letters. These latter are, in general, the letters most 
 easily perceived,^ mainly "those letters which project above the 
 body of the word and the additional middle sized letters x and z."^ 
 Examples of the latter are 'admiralty,' Chart II, line 1, 'officers,' 
 line 2, 'Kaznahoff,' Charts I and II, 'Baltic,' Chart I, line 2, and 
 'Paris,' Chart I, subjects T and S, 'Vigo,' (line 3), Chart II, line 
 3, etc. There are many of the capital and domineering letters in 
 each line, and they are naturally not all equally operative; but, 
 when the general distance of the interfixation movements is once 
 determined, they, with the other causes mentioned, may sometimes 
 govern the location within this area. They furnish, possibly, a 
 better form of peripheral stimulation than the middle sized letters, 
 i. e., they 'catch the eye' more readily. They may, in other words, 
 be supposed to condition the word complex — the 'Gesammtbild' — 
 to a greater extent than the other letters of a word.^ 
 
 The influence of some or all of these factors when in another 
 line than the one being read is also to be noted. They often dis- 
 tract the attention, but they are undoubtedly also of service in de- 
 termining the location of the fixation pauses in the second line. 
 There is plenty of evidence to show that when the eye is reading 
 in one line, the mind is vaguely conscious of matter lying in the 
 succeeding lines. For instance, the long distance of the first fixa- 
 tion in line 2, Chart II, from the edge of the page, in 'has tele- 
 graphed' is doubtless due to the fact that this word was in the 
 'margin of consciousness,' while the word 'St. Petersburg,' just 
 above it, was being read. If the words in the line below have ab- 
 solutely nothing or little to do with the sense of the line being 
 read, this view is often more of a hinder ance than a help. In the 
 case cited, however, the associative expectation was ' ready ' for such 
 a word as 'telegraphed' at the beginning of a newspaper column. 
 The reading together of lines six and seven of Chart II is doubt- 
 less, also, in a similar way due to the initial similarity of the lines, 
 i. e., 'admiral,' in line six, 'admiralty,' in line seven. 
 
 To answer, then, the question as to what determines the loca- 
 tion of the fixation pauses, it is first and principally the unit of 
 apperception. This may be small in extent, as in numerals, ab- 
 breviations, conjunctive phrases, etc. — in a word, when several 
 
 ^ See Cattell on the Legibility of Letters, in articles referred to above. 
 ^ Cf. above Chap. 
 
 ^ The statement of the possible effect of the domineering and capital 
 letters is offered somewhat tentatively. 
 
LOCATION OF PAUSES 91 
 
 small words cannot be regularly associated together into a single 
 unchanging group, or large in the cases of long Avords and very 
 familiar phrases. The location of the fixation within these limits 
 may depend in part on the more or less artificial peculiarities of 
 spacing, punctuation, and the forms of letters in printing. To 
 these latter causes must also be added the fact earlier noted, that 
 within certain limits the eye can regulate its positions, in order 
 to maintain its so-called 'short-lived motor habits.' 
 
 § 2. Proof-reading. Effect of Articulation and of Near and Far 
 Accommodation, etc. 
 
 The effect of various other factors, as that of near and far ac- 
 commodation, the reading of foreign languages, etc., upon the 
 number and duration of the fixation pauses has already been re- 
 ferred to. 
 
 The significance of the experiments in these cases is in at least 
 one particular the same, i. e., it is not the mere extent of words 
 which can be easily seen which determines the span of attention or 
 the number of fixation pauses. Although there is more of the line 
 in the field of clearest vision in far accommodation, the number of 
 pauses remains about constant. The results of readings of the 
 same page by one subject are shown in Charts XII and XIII. The 
 page was placed in the first case (No. 17) at approximately 15 
 centimeters from the cornea of the eye, and in the second case 
 (No. 18) at approximately 40 centimeters from the eye. 
 
 The efli'ect of reading of a foreign language is to increase both 
 the number and duration of the pauses. The same result may be 
 illustrated in proof reading. (Cf. Plate V, 5, 6, and Chart XIV.) 
 The page read is from a biographical directory. Names have been 
 altered but their general form and length have not been changed. 
 
 The efl^ect of reading a passage aloud is also somewhat similar. 
 Naturally greater attention must often be given to the parts of a 
 word in pronouncing it, with the result that the word innervation 
 is more often divided. The time per fixation as well as the num- 
 ber of fixations may be increased. It is somewhat a matter of in- 
 dividual difference. In Chart XV B of Record -12 A (subject S), 
 even fewer fixations are made in the first few lines than in 41 A, 
 which WitS the silent reading and immediately preceded it. The 
 time spent per line in fixation is, however, increased hy 461cr, for 
 example, in the first line. The case is somewhat reversed in Record 
 29, Chart XV A. This passage was read aloud immediately after 
 reading it silently (cf. No. 28, Chart XV A). The number and 
 time would have been somewhat increased had the reading aloud 
 
92 THE PSYCHOLOGY OF READING 
 
 preceded the silent reading. As it is, it takes l,474o- longer to 
 read the whole passage. 
 
 Fifty pauses are made in the silent reading (No. 28), 68 in the 
 reading aloud (No. 29). The average time of the pauses in the 
 latter case (No. 29) is shorter, 144.1o- (A. D, 47.6) as compared 
 with 169. 4o- (A. D. 53). Since the total time is longer, this shows 
 that in reading aloud the time is more equally distributed (see also 
 chart) in all parts of the line. It shows that the freer movement 
 of the attention 'wave' is repressed by the slow process of articu- 
 lation. In reading aloud as in reading a foreign language, the 
 attention must be given more often than ordinarily to parts of 
 words and syllables. These facts may also be noted in even the 
 seventh reading of a passage by subject H. This passage (33 A), 
 which was read aloud, may be compared with any of the earlier 
 readings above. 
 
 It is interesting to note in No. 29 the similarly large percent- 
 age of refixations at the beginning of the line, which was much in 
 evidence in the first reading (No. 28). It further supports the 
 hypothesis that it is a matter of habitual reaction for this subject 
 (T). Since it is not possible to return with any certainty to the be- 
 ginning of this long line, the movement regularly falls short, and a 
 regressive movement is made if necessary; in lines two, three, and 
 seven it was not necessary, in the other lines regressive movements 
 are made (cf. above). The charts of (1) reading by near and 
 far accommodation, (2) of proof reading, and (3) of reading aloud 
 follow (Charts XII-XVI). See also Plate V, Nos. 5, 6, 7, and 8. 
 
 CHART XII 
 
 Reading at Near Accommodation 
 No. 17 
 Summary 
 [ )B|volutiolnary stu[dy a.)nd thought h|ave bee[n) hinder|ed by |the colnfu- 9 
 
 c b a 
 
 si|on of t|wo im|related biologic|al phenomen|a, ^1) ev|olutionar|y prog|ress 8 
 
 or[ v)ital motion and H2) [ th)e or|igination or[ multi)plicati[o)n of sp[ecies.) 6 
 
 T[he 'o)rigin' of[ a) species is not mo|re evolutionary than any other stage 3 
 in its history. ' 
 
 ^ These parentheses belong in text. 
 
LOCATION OF PAUSES 93 
 
 CHART XIII 
 
 Far Accommodation 
 No. 18 
 
 E|volutio|nary study a|nd though|t have been |liindere(d by ]the |confu- 7 
 
 b a 
 
 sion o[f two) [unr(e)late]d biological| phenomen(a, V{1) (evolut]i|onary |progress 8 
 
 a b c 
 
 or v[it[al )m)otion, and ^(2)i t]he origin(ati]on or muI|tiplication of[ species.) 6 
 
 a b a 
 
 in its| history. Th|e causes of t|he subdivis]ion of s|pecies [are n)ot causes 6 
 
 of [vital) motion; the t|wo proc|esses are [qui)te distinct. Th|e separation 5 
 
 of[ tw)o [spe)cies is| not a fojcus (of th]e evolutionary problem; it is a mere 
 'ncident ofdevelopmental history. 
 
 ^ These parenthesis about the numerals belong in text. 
 
 CHAUT XIV 
 No. 34 B 
 
 166 104 165 116 lie 424 363 198 380 
 
 IVIento[r, ]P|rof. |Albe|rt |M('cDona[ld)S) 305 |E. |S[oa]p- 9 
 
 268 164 313 647 490 171 291 
 
 '[ ston)e St., I Lex|ingto)n, Ct.[ )Bi\_ology.'\ X|enia], O, 7 
 
 431 4fi0 313 189 313 183 
 
 [Aug). 5, 45.[ ]A.B, Pri|nceton, [89, )A.[M, )91, [fell)ow. 7 
 
 437 212 183 448 342 206 236 
 
 91-|92; M|unic|h, 92[93. )Pr[inc)ipal, |High |Sch, 7 
 
 389 130 407 490 342 372 
 
 Xen[ia, O,) 9|0-91; prolf. )zool.[ and) ge[ol, )TIlto|n, 6 
 
 484 264 313 207 142 236 206 389 193 
 
 93(-95;] \Uol, ICt). st\ate, |95. P.|A,A; |Bio|l. S|oc. 9 
 
 372 106 94 
 
 Ge[ology ]of( p]et|roleums Connecticut survey. 
 
 a c b 
 
 No. of fixations, 45. 
 
 ^ These parentheses belong in text. 
 
 ^ Note overestimation due doubtless to extension and heavy type of 
 first line. 
 
94 THE P87CB0L0GT OF BEADING 
 
 CHART XV A 
 Reading Aloud 
 
 No. 29 
 Sulject T 
 
 64 i 112 3 112 gg 160 96 96 168 
 
 Evolut[io)n|a|ry study an[d |t)houglit h|a|ve been [hindered (by the co[nfu)- 10 
 
 bea fc dg h i j 
 
 104 128 88 136 40 216 88 
 
 sion of t|wo uni'(elated biolo|gical plieno|mena, ^{\1) evoluti[on)ary pro|gress 7 
 
 104 112 144 128 168 128 112 
 
 or vi|tal motion|, and (2) the o|r!ginatioln or multiplica|tion of |speci|es. 7 
 
 104 112 104 128 120 128 128 § 96 152 
 
 |Th|e 'o|rigin' [o)f a speci|es is n|ot more| evolutiona[ry) than an|y ot|her[ ) stage 11 
 
 120 192 88 168 160 264 88 lo2 120 136 136 
 
 i|n[ i)ts history.| The causes| of the s[ub)d[ivi)si|on of |specie|s [ar)e n|ot causes 11 
 
 b a 
 
 160 104 160 112 S 4E0 § 1 P 
 
 0'|f v|ita|l mot|ion; the |two p[ro)cesses are qu|ite dist[in)ct. The [ae)paration 9 
 
 88 72 S 376 88 136 112 
 
 of two specie|s is not a fjocus of| the evolution [a) ry proble|m; it is| a ni|ere 7 
 
 g § 1 72 272 230 
 
 in[cid|ent o)f d|evelop[in)ental histor[y.) 
 
 c b + d a 9 f 
 
 No. of fixations, 68. Total duration, 9,804. Av., 144.1. M., 128. A. D., 47.6. 
 
 ^ Note very short pause. Probably not a real fixation, but just a 'hitch' in the 
 eye's movement. 
 
 CHART XV, B 
 Reading Aloud 
 
 No. 42 A 
 
 Subject S 
 
 442 612 363 272 
 
 |A license for pr [emotion ma)y be issu|ed to the ho[lder o)f a 4 
 
 394 279 476 496 266 224 367 96 
 
 tea|cher's licen[se) No. 1 [w)ho[ has h)ad( e]xperien|ce rated as [equ)ival|ent to 8 
 
 340 394 238 278 211 190 177 231 150 
 
 t[hree) years o(f Ne]w York] city pu|blic sc|hool tea|ching ^(|includi|ng |one 9 
 
 374 286 218 728 299 266 170 
 
 year! in the city of| New Yorik)'|; who passe [s a)n |examin|ation in t|he 7 
 
 252 292 320 510 286 962 
 
 princ(iples] and me(tho]ds of teac|hing, or in lieu [th)ereof prod(u]ces evi-[) 6 
 
 No. of fixations, 34. Total duration, 11,437.6. Av., 336.4. M., 285.6. A. D., 121.4. 
 ' These parentheses belong in text. 
 
LOCATION OF PAUSES 95 
 
 CHART XVI A 
 
 Usual Method of Reading 
 
 793 296 321 
 
 1 ST[. P)BTE:RSBURG, Nov. 2.— [)The A[d)miralty 3 
 
 460 217 211 249 
 
 2 has [t)elegraphed to tjlie o[ffic)ers of the [Ba)ltic 4 
 
 326 260 370 234 331 
 
 3 [flee)t, who |were left [be) hind at V|igo in (or]der 5 
 
 336 316 330 297 270 
 
 4 [th)at the(y ]mig[ht )testify, an[d )who [wer)e on 5 
 
 381 268 341 308 
 
 5 t[hei)r way to St.| Petersburg, [to) remain[ in) 4 
 
 365 
 
 6 [Pari)s. 
 
 365 207 241 308 234 
 
 6 [ Ad)m[ir)al K[ax)nakoff, [a ni)ember o|f the 5 
 
 187 227 294 375 
 
 7 Admi[ralty )Co[)ncil, who |was app|ointed 4 
 
 b a 
 
 376 227 201 167 
 
 8 yes[ter)day upon| the interna [tion)al com|- 4 
 
 268 207 221 268 
 
 9 missio[n )w'hie|h is to in|quire into t[he )North 4 
 
 274 194 221 201 254 
 
 10 Sea affa[ir,) star|ted for P[aris) to-d[ay. )Hi|s 5 
 
 366 294 274 
 
 11 a[pp) ointment is[ we)lcomed as an[ in)dication 3 
 
 288 140 187 241 
 
 12 that there[ w)i]l be [a) thoroug[h s)earch for t[he) 4 
 
 234 221 281 247 
 
 13 fac[ts )of[ t)he case. The Admi[ral wo)n go (Id] en 4 
 
 b a 
 
 288 160 221 187 
 
 14 opinio [ns ) among th|e membersj of the eo[n-) 4 
 
 416 164 
 
 15 vention held in Wa[shing)ton in 1889 toi| for- 2 
 
 CHART XVI B 
 
 Reajung^Aloud 
 
 323 57l 
 
 3C6 S23 fiTS 201 143 
 
 1 ST. P[ETE)RSBURG,[ N)ov. 2.)— (T]he Ad|miralty 5 
 
 351 184 238 184 173 92 151 166 
 
 2 [has )telegraph|ed t|o the| offic|ers of th[6 B)[alt)i|c 8 
 
 412 132 164 171 661 
 
 3 fleet, w|h]o wer|e left] behin|d at Vig[o, in) [ ord)er 5 
 
 az ai SI 82 
 
 280 214 166 319 164 132 
 
 4 tha[t t)hey migh|t testify,[ an)d w[ho) we[re) |on 6 
 
 198 143 148 170 164 242 
 
 5 (th]ei[r w)ay |to St.| Petersbur|g, to r[e)main in 6 
 
 149 
 
 6 [Pa)ris. 
 
 143 215 214 
 
 '7 [ ) Admiral [Kaxn)akofE, a [me)mber of [th)e 4 
 
 149 125 132 143 314 
 
 8 Admiral[ty) Co[u)ncil,| (w]ho wa[s app)ointed 5 
 
 182 166 280 204 121 326 
 
 9 ye|ster|day up [on) the int(er[nati[ona)l [com-) 6 
 
 292 247 99 209 236 242 
 
 10 m|iss[ion) which is| to [in) [quire )into th[e N)orth 6 
 
 No. of fixations, 51; total duration, ll,070cr; average duration, 217.1<r; 
 A. D., 76.7; median, 183.6(r. 
 
 * Lines 6 and 7 read in one horizontal movement. 
 
96 TEE PSTCBOLOGY OF READING 
 
 § 3. Children's Beading 
 
 What is the method of progress and what differences do chil- 
 dren show in the ease and rapidity with which they acquire the 
 mechanism of eye movements ? Do the backward readers show any 
 anomalies of mere eye movement? 
 
 It was at first my purpose to devote a chief part of the in- 
 vestigation to a study of these and similar questions. The adult 
 practice and method is, however, presumably the only criterion we 
 have, and by which we may judge of the progress of the beginner. 
 But little has been determined in regard to the individual peculiari- 
 ties of eye movement in the case of adults, and it has, therefore, 
 been necessary to confine the experiments almost wholly to the 
 latter subjects. The experiments upon children are, therefore, of- 
 fered tentatively because of the small number of cases.^ 
 
 The records which are reproduced in Plate V and Chart XVII 
 are those of three boys, of nine, ten, and eleven years of age, and 
 from the fifth, fourth, and third grades respectively. The passages 
 read were from the readers in use in the latter two grades. They 
 are presented more to illustrate the differences between adults and 
 children who have learned to read than for their value as illus- 
 trative of differences among children. What progress takes place 
 as the child learns to read can, of course, best be studied by re- 
 peated experiments upon the same individual. 
 
 The records appear, however, to represent several stages of de- 
 velopment, which differ in the number and duration of the fixa- 
 tion pauses. The ' Gesammtbild ' is evidently much more divided in 
 the case of the youngest subject. This appears in the relative fre- 
 quency of the fixation pauses. The attention span is much wider 
 in the case of the ten-year-old boy, and at times equals that of the 
 adult whose reading of the same page is also reproduced for pur- 
 poses of comparison. The more purely physiological difiiculties 
 have been fairly well mastered. The rate of movement in the re- 
 turn sweeps and in the interfixation movements is not different 
 from that of the adult. There is some unsteadiness and refixation, 
 but the mere accuracy of fixation appears as exact as that of 
 adults. The chart below shows the location and the duration of 
 the pauses, — the two factors in which there is the greatest differ- 
 ence between child and adult. The overestimation of the return 
 sweep at the beginning of the first lines of Readings B and C is of 
 interest. It is perhaps increased here because of the indentation 
 of the first line. 
 
 ^ It is the plan of the writer to supplement this part of the investiga- 
 tion by an independent study. 
 
LOCATION OF PAUSES 97 
 
 CHART XVII A 
 
 I. Children's Reading 
 
 8. N. Fourtli Grade. Age 10. Third Beading 
 
 No. 25 
 
 404 464 644 S26 332 380 714 
 
 But |the dark[ o)bjects [were) no[t [cat)t[le.) )T[hey )were 7 
 
 e d f d \ 
 
 692 296 oOO 890 720 
 
 [the Black) D|ouglas[ and) his [men, creeping) [on )liands 5 
 
 666 558 1154 
 
 and fe|et towar|d the foot of t[he castle wall. )Some 3 
 
 CHART XVII B 
 II 
 
 C. L. Third Grade. Age 11. First Reading 
 No. 21 
 
 207 493 230 487 286 1133 
 
 [ K[i)n[g Mi)das thought, a[n)d thought.] At last| |h|e| 9 
 
 abc d e fhgi- 
 
 330 £46 174 190 196 499 756 666 358 773 
 
 loo|ked| at t|he str|ang|er| and sa[id), [")I wish| that[) 11 
 
 f e 
 784 179 146 186 167 879 420 454 426 319 570 392 
 
 ev[e)r)ythin|g |that I |touc[h ma)y [tur)n [to) go|ld[.") | | | 12 
 
 a. db ce i f hjgki 
 
 CHART XVII C 
 III 
 
 C. L. Second Reading 
 No. 22 
 
 * ( 7][8 ) [lO ) 
 
 12 4 6 1° 3 6 9 11 13 12 14 
 
 ^t )|Ki|ng| |Midas t|h|oug[ht, an)d [tho)ug[h|t.) [ At last) he 14 
 
 1060 190 393 211 225 
 
 looked[ at) the | stran|ger and] said, "I| wish that 5 
 
 183 197 211 169 120 127 743 
 
 ever|yt|h|ing that| a to(uc]h may tu|rn [to gold)." . 7 
 
 2 13 
 
 281 176 190 127 169 299 169 190 232 
 
 "Are [yo)u| sure t|h|at yo|u (wo[ul)d|be [satisfied 9 
 
 1 3 24 5 6798 
 
 246 274 113 
 
 [ )[th)en?"l 
 
 2 3 1 
 
 239 127 330 169 
 
 "Yes," a[n)swered [Mi)das. "I wo[uld) ask for 7 
 
 nothing more." 
 
 I" Point to which the eye came from the 12th fixation, returning to 
 the 13th fixation after a pause of 143<r. 
 
 ^Duration of pauses in first line: (1) 491, (2) 134, (3) 141, (4) 379, 
 (5) 323, (6) 197, (7) 547, (8) 288, (9) 358, (10) 176, (11) 106, (12) 100, 
 (13) 456, (14) 414. 
 
98 TEE PSYCHOLOGY OF READING 
 
 § 4. The Effect of Long and Short Words 
 
 A comparison of the number of fixations and their duration 
 per line in passage 63 A (Chart XX A), composed mainly of short 
 words, with the same passage in Chart XXI A, which is composed 
 of rather long words, may be made as a test of the theory proposed 
 by Messmer/ viz., that a succession of small words tends to break 
 up the motor innervation and is harder to read than a passage 
 composed mainly of long words. Record 63 A shows in the first 
 place that it is not necessary to fixate every word. Were the in- 
 nervation broken into smaller units there is reason to believe that 
 fact would be indicated by an increase in the number of fixations, 
 just as the number of pauses is increased in a foreign language or 
 in difficult passages. A comparison of these charts will show that 
 in the first six lines, of equal length and size of type, 37 fixations 
 are made in the passage of long words, and 35 in the short word 
 passage. There are 79 small words and 54 long words, or an aver- 
 age of 2.3 short words per fixation as against but 1.5 long words, 
 i. e., one and one-half times as many. It is not the short words as 
 such, hut the words which cannot he easily grouped with others 
 which necessitate separate fixation. The same holds true in the 
 short line arrangements. Attention may be called to the misspelled 
 word in 62 B, line 1, 'gorgeosly.' The subject noticed the mis- 
 spelling in reading and the record shows that it was carefully 
 fixated. 
 
 ^Op. cit. 
 
CHAPTER XII 
 
 The Length of Text-Lines and Motor Habits 
 
 Very little careful experimentation has been made relative to 
 determining the proper lengths of lines of print; and as a result 
 there has been considerable difference of opinion with very little 
 evidence one way or the other. Weber/ for example, believed that 
 "up to a length of 150 mm. (6 inches), but not beyond, long lines 
 make it easier to read quickly. He requires a minimum of 100 
 mm. (4 inches), a maximum of 150 mm. (6 inches) for the normal 
 line. He wishes, therefore, that school-books should be printed 
 where possible in lines of 140-150 mm."- 
 
 Javal, on the other hand, believed 'the long lines to be the 
 reason why progressive myopia is so frequent in Germany. He con- 
 siders that with long lines, short-sighted people must exert their 
 accommodation the more frequently and strongly in the middle of 
 the lines as their eyes are focused for the ends of the lines. "^ 
 
 Cohn adds: "It seems to me that 100 mm. (4 inches) is the 
 greatest length admissable, and 90 mm. (3.6 inches) the best length 
 for lines of ordinary print where the 'n' is 1.5 mm, high. With 
 larger types a longer line of 110 mm. (4.4 inches) is allowable." 
 
 Since differences in phraseology, subject matter, size of type, 
 and length of line, etc., may each affect both the number and 
 length of pauses, it seems evident that the only method of deter- 
 mining the influence of each factor is by elimination of the others. 
 In the case of subject matter it has been noticed that a second 
 reading may often produce but little change, and, if the reading is 
 repeated after the lapse of a considerable length of time, the influ- 
 ence of repetition may be largely disregarded. The method of 
 studying the influence of the length of line on the ease and quick- 
 ness of reading, has been, therefore, to keep the subject matter, 
 the size of type, etc., and all other conditions uniform. The pas- 
 sage was simply typewritten or printed in two different lengths of 
 line. In the first reading one arrangement of the length of line 
 
 ^ A. Weber. Ueber die Augenuntersuchungen in den hoheren Schulen 
 zu Darmstadt. Referat und Memorial, erstattet der grossherzoge Minis- 
 terial-Abtheilung fiir Gesundheitspflege. Marz, 1881. 
 
 -Quoted from Cohn: 'The Hygiene of the Eye,' p. 206. 
 
 ^Cohn: Ibid, p. 205 ff. 
 
100 THE PSYCHOLOGY OF READING 
 
 was used, and, in the second reaamg, the other arrangement. Sev- 
 eral weeks intervened between the first and second readings, but, 
 in order that there might be no opportunity for practice effect, the 
 long-line passage was read first in one experiment and last in the 
 next, etc. Six of the passages read were typewritten, the height 
 of the small 7i being 2 mm., and two were set up at the printer's 
 in the different lengths of line desired. 
 
 The several records will be discussed separately. They appear 
 as Charts XVIII-iXXII, but will be cited by their original record 
 number. The passages for this experiment are with one exception 
 chosen from the plain narrative of Robinson Crusoe, in order that 
 there might be as little difference as possible in subject matter in 
 all the tests. 
 
 The first set of readings of two subjects of the same passage in 
 long and short line arrangement follows (59 A and 59 B, 68 A and 
 64 A). Subject S is, in general, a very rapid reader, subject E 
 a somewhat slower reader. Since the lines were typewritten they 
 were not all of equal length; the long line varies between 179 mm. 
 and 190 mm., the short line between 86 and 98. In the average of 
 the lines compared, the long line is just double the length of the 
 short line. 
 
 CHART XVIII A 
 Subject E 
 No. 59 A 
 
 297 192 264 260 260 210 260 167 117 
 
 There was[ )betwe|en the|m and |my ca|stle the cre|ek, whi|ch I men|tioned o|ften 9 
 
 291 248 303 363 198 111 
 
 at the first[ p)art of my story, w[he)n I landed | my cargoes| out o[f )my s|hip 6 
 
 403 163 170 290 233 333 226 81 233 19j 
 
 and this I kn[ew h)e| must| ne|ees[sa)rily swim o[ver), or th|e poor| [wre)tch [wou)ld 10 
 
 bcasd f ihg j 
 
 289 768 220 264 264 1246 138 366 
 
 be taken ther[e;) but when [the sav)age esc|aping c|am|e th|ither, |he mad[e n)othing 8 
 
 a b cedfgh 
 
 214 233 239 161 283 428 296 189 170 226 
 
 of it, t|hough| the ti|de was] then u|p; but plung(ing] in,[ swa)m th|rouglh in| about 11 
 
 b a&c ^ 
 
 189 170 94 163 220 362 264 189 396 182 
 
 thirt[y s)troke|s, or (th]ereabo|uts, lan[ded), and rjan) (on ]with |exceeding [strength 9 
 
 aa a I 
 
 846 326 266 332 637 288 230 
 
 and s[wi)ftness. Wh|en the (th]ree pursuers] came to th[e )creek, |I found |that 7 
 
 192 211 134 211 172 224 262 230 224 332 166 
 
 two [of) th|em| could (s]wi|m, bu|t the t[h)ird could| n|ot,[ )and [he,) stand |ing on the 12 
 
 bac d* *g '^ij k ' 
 
 No. of fixations, 72; median of fixation pauses, 233. Icr (A. D. 66.7); length of line, 
 179-189 mm.; total time, 11,088.8(7 (5 lines). 
 
TEXT-LINES AND MOTOR HABITS 101 
 
 CHART XVIII B 
 No. 59 B 
 
 92 123 128 149 220 202 141 132 
 
 1 There [wa)s b|e)tween| them [an)d( ni]y c|ast|le 8 
 
 b a c d g e f h 
 
 189 132 145 189 
 
 2 the cre|ek,| which |I mentio(ne]d o|ften 5 
 
 123 189 189 194 194 
 
 3 at I the f|irst pa[rt) of my[ )story,[ when) 5 
 
 b a 
 
 270 ISO 162 163 
 
 4 I landed m[y c)argoe|s out| of the |ship; 4 
 
 189 108 99 207 108 
 
 5 and thi[s )I kn|ew, |he m|ust necess|arily 5 
 
 261 199 216 112 
 
 6 swim o(ve]r, or the| poor wret|ch would] 4 
 
 270 211 163 139 
 
 7 be taken( ]there; but whe|n t|he sava|ge 4 
 
 266 221 163 171 81 
 
 8 escap(ing] came (t] hither, he m|ade| noth|ing 5 
 
 a b dee 
 
 126 135 189 198 186 139 
 
 9 of it,| thoug|h( th]e tide (w]a[s t]he(n] up; 6 
 
 b a c d f 8 
 
 211 283 135 113 
 
 10 but pljunging in, s[wam) throug|h in| about 4 
 
 Length of line, 87-96 mm. 
 
 No. of fixations, 50; average duration, 182.4a-; total time, 9,118.8(t. 
 
 CHART XI'X A 
 
 Subject 8 
 No. 68 A Second Beading 
 
 418 216 159 244 249 
 
 There was b|etween them and |my castle t|he [) creek, which I mentio[ned o)ften 5 
 
 a b d c 
 
 291 263 268 146 206 173 47 103 
 
 at [the )first pa[r)t of my story,| when I la[n)ded my| cargoes[ o)ut of m|y ship|; 8 
 
 308 230 257 207 207 
 
 [and this) I knew he mu|st necessarily [sw)im over, [o)r the poor wr[etch) would 5 
 
 197 234 211 161 234 166 165 
 
 be take[n| th)ere; but when| the savage elscaping came [th)ither, he| made n|othing 7 
 
 a d 
 
 369 233 266 247 67 108 94 
 
 of it, [tho)ugh the tide was] then up; but [plunging in, swa[m t)hrou|gh |in a|bout 7 
 
 a b c d f 9 g 
 
 139 166 292 216 279 270 297 126 
 
 thi|rty [stro)kes, |or there|abouts, [la)nded, and| ran on wit[h )exceeding s|trength 8 
 
 40 
 
 234 180 262 261 366 189 72 
 
 and| swi[ftn)ess. When |the three pursue[rs )came to t[he cre)ek, I fo|und t|hat 7 
 
 b a c d 
 
 256 180 171 247 252 144 167 
 
 two of I them [coul)d s|wim|, but the t[hir)d could not, a[nd) he, sta|nding [on )the 8 
 
 badce f g^^ 
 
 No. of pauses, 57; average duration, 213cr, A. D., 60.6. 190 mm. length of longest 
 line, 179 mm. length of shortest. 
 
102 TEE PSYCHOLOGY OF READING 
 
 CHAET XIX B 
 
 No. 64 A 
 
 321 204 204 286 
 
 1 There was b|et|ween theni| and my [ca)stle 4 
 
 dz d I 
 188 209 214 173 
 
 2 the c|reek, w|hich I men|tioned oft|en 4 
 
 239 198 198 147 
 
 3 at the| first par|t of my s|tory, w|hen 4 
 
 270 265 183 
 
 4 I I[and)ed my cargoe|s out of t[he shi)p; 3 
 
 256 168 224 112 
 
 5 and this 1 1 knew|, he must| necessa|rily 4 
 
 336 198 132 107 ' 
 
 6 swim [ove)r, or the |poor wre|tch wou|ld 4 
 
 2ta 91 178 76 
 
 7 be take[n )there; bu|t when t|he sa|vage 4 
 
 178 224 153 311 
 
 8 escapi|ng cam[e) thith|er, he made n|othing 4 
 
 b a 
 
 306 229 168 
 
 9 of it, th[oug)h the tide |was then |up; 3 
 
 265 122 229 132 
 
 10 but [pl)unging i|n, swam thr|ough in [ab)out 4 
 
 282 224 163 31 
 
 11 thirt[hy) strokes, or |thereab|outs, l|an- 4 
 
 168 132 224 183 102 
 
 12 ded, I [and )ran on wi|th excee|ding st|rength 5 
 
 b a 47~ 
 
 173 244 249 260 
 
 13 and sw[iftn[e)ss. Whe[n) tlie [th)ree pursuers 4+ 
 
 b a 
 
 No. of fixations, 47; average duration, 200. 4(r; total time, 9,419. So-. 
 
 CHART XX A 
 
 No. 63 A 
 
 240 260 192 197 163 164 
 
 The boat[swa)in was killed upon tjhe spot; t|he next [m)an was sh[o)t in t|he 6 
 
 254 240 269 278 236 269 
 
 body[, a)nd fell jus(t ]by him,| though he did| not die [unt)il an hour[ o)r two 6 
 
 288 293 269 246 163 182 
 
 a[fter); and the thi[rd )ran for i]t. At th[e )noise of th[e fi)re, [I )inimed- 6 
 
 202 206 240 216 187 604 
 
 iate[ly ad)vanced with m|y whole army,[ )which was| now eight |men; viz], my- 6 
 
 246 341 312 240 370 
 
 self, |generali[s)simo; Frida(y, ]my lieutena[nt-g)eneral; the c[ap)tain and 5 
 
 269 106 264 187 246 264 
 
 hi[s tw)o men, and the[ th)ree prisloners of war[, )whom he [had) trust[ed wit)h 6 
 
 418 264 211 221 110 164 "^ 
 
 arms. jWe came up [on) them indee[d) in the dark,[ so) that [the)y could [ no)t 6 
 
 307 278 221 192 178 
 
 see our n[umber;) and I made th[e m)an they had l[ef)t in the boa[t,) who[ was) 5 
 
 394 293 144 178 
 
 now one o[f u)s, to call th[em) by name, t[o t)ry if I [co)uld bring them to a 4 
 
 Length of line, 178-9 mm.; No. of fixations, 35 (6 lines); average duration, 244. 2cr; 
 total time, 8,548. So-. 
 
TEXT-LINES AND MOTOR HABITS 103 
 
 CHART XX B 
 
 No. 69 A 
 Passages of Long Words 
 
 244 00 117 270 117 148 
 
 1 Th|e boa|tswai[n )was k|illed upon| the sp|ot; 6 
 
 c b a d 
 
 270 239 217 276 
 
 2 [ )the next [ma)n was shot [in) the bod[y, )and 4 
 
 307 260 191 191 
 
 3 f[e)ll just by h[im,) though[ he) did not [d)ie 4 
 
 191 254 270 170 86 
 
 4 un|til[ a)n hour or two a[ft)er; and [th)e thi|rd 5 
 
 b a 
 
 3fil 69 276 140 81 
 
 5 r|an for it. |At the noi|se of the fi[re), I| 5 
 
 297 292 270 81 
 
 6 i[mme)diately advanc]ed with my[ who)le ar|my, 4 
 
 276 389 302 103 
 
 7 [wh)ich was now ei[gh)t men; viz. [m)yself,| gen- 4 
 
 276 44 83 187 281 110 
 
 8 e[rali)ssimo|; F|ri|day, ni[y) lieutenant] general^, 6 
 
 b d c 
 
 263 99 297 176 
 
 9 t[he) captai|n and his t[wo m)en, and the[ t)hree 4 
 
 292 170 165 149 
 
 10 [pr)isoners of w[ar), whom he [had) trusted [ wi)th 4 
 
 11 arms.] We came upon them indeed in the 
 
 46 
 Lengths of line, 90-106 mm.; No. of fixations, 46; average duration, 
 204.3(r; total time, 9,400.0o-. 
 
 CHART XXI A 
 No. 62 B 
 
 260 .334 236 341 422 434 484 373 
 
 T|he gorg|e|osly co|stumed impe|rial pIenipote|ntiary suf[fe)red excrujciating 8 
 
 b c e d 
 
 615 416 409 668 347 
 
 [an)guish at the rec[ol) lection of his [perso)nal thoughtle|ssness and care|less 5 
 
 496 416 415 617 336 
 
 [ne)ss. There lay[ be)fore him the rece[n)tly appointed] ambassador but |now 5 
 
 322 384 3u9 602 372 260 
 
 r|uthlessly murde|red by an hi|reling ass(assin]. Although (ther]e undoubt|ed- 6 
 
 416 384 310 329 409 397 
 
 ly| existed severa|l indicatio|ns of his persojnal innoc[ence,) what [people )of 6 
 
 378 366 180 397 360 329 174 
 
 i[nte)lligence would] hesitate t]o proclaim] the startling ]circumst]antial ev]i- 7 
 
 360 484 316 463 
 
 [dence) prepond[erou)sly] conclusi]ve. 4 
 
 a b c 
 
 41 
 
 No. of fixations, 41; average duration, 378. 2(r; total time, 15,506.0*. 
 
104 THE PSYCHOLOGY OF BEADING 
 
 CHART XiXI B 
 No. 69 B 
 
 65 265 310 245 156 
 
 1 The go|rgeously[ c)ostum[e)d imper[ial) ple|ni- 5 
 
 320 325 235 146 
 
 2 pot[en)tiary suffe[r)e<J excru|ciating a|nguish 4 
 
 285 256 £40 100 
 
 3 a[t t)he recolle[c)tion of his per[s)onal th[o)ught- 4 
 
 270 300 260 140 
 
 4 lesshess a[nd) careless|ness. There[ )lay |be- 4 
 
 285 235 280 125 
 
 5 fo|re h)im the recent|ly appointe|d ambassa|dor 4 
 
 315 260 155 125 165 
 
 6 but no[w r)uthless|ly murdered| b|y an hireli|ng 5 
 
 185 126 260 266 175 
 
 7 a[ss)asin.| Although[ t)here undoubt[ed)ly ex-| 5 
 
 350 240 200 200 
 
 8 iste[d sev)eral md|ications |of his persona[l) 4 
 
 310 295 210 
 
 9 innocenc[e,) what people p|f intelli[ge)nce 3 
 
 320 146 276 140 
 
 10 would [hesitate to |proclaim[ th)e startli|ng 4 
 
 105 176 276 80 
 
 11 cir[curas)|tantial evidence prepo[nde)rou|sly 4 
 
 276 
 
 12 conclus[ive). 1 
 
 47 
 No. of fixations, 47; average duration, 2220-; total time, 10,670(r. 
 
 CHAET XXII A 
 
 No. 68 B 
 
 309 176 215 249 206 
 
 And I this |old savage |was in the [rig)ht, for, as[ I u)nderstood 5 
 
 268 223 249 107 167 
 
 |since by other |hands, the sa|vages of that p[art) never [at) tempted 5 
 
 216 223 180 133 
 
 [to )go over to the island [a)fterwards. Th[ey )were so terri|fied 4 
 
 266 238 ISO 202 144 
 
 wi[th) the accounts g|iven by these fo[ur )men (for it se[e)ms t|hey 5 
 
 198 220 266 189 229 
 
 [did) escape the| sea), that th[ey) believed w [ho) ever went t[o t)hat 5 
 
 396 99 229 229 54 
 
 enc[hant)ed island wjould be destroyed [w)ith fire fr|om the g|ods. 5 
 
 267 235 272 286 239 
 
 T[his, )however, I |knew not, an|d therefor[e w)as under [cont)inual 5 
 
 173 145 286 319 146 
 
 ap|pre[hen)sions for a goo[d )while, and kept[ a)lways upon | my 5 
 
 340 350 302 225 802 
 
 gu[a)rd, I and all] my army; for as[ the) re w[ere) now f[our o)f us, 5 
 
 177 196 240 201 251 
 
 I wo|uld |have ventured up|on an hund[red) of them [fairly i)n So 
 
 336 278 230 201 144 
 
 the[ ope)n field at any tim|e. In a |little time,| however, |no more 5 
 
 54 
 Fixation pauses, 54. Average duration, 225cr; A. D., 50 (M., 229.5; total 
 time, 12,196.10- (ll,621o- for 51 pauses). 
 
 51 
 
TEXT-LINES AND MOTOR HABITS 105 
 
 CHART XXII B 
 No. 63 B 
 
 269 167 195 131 112 
 
 [An)d this [ol)d savage [w)as in th[e) righ[t, f)or, as 1 5 
 
 319 282 57 227 
 
 und|erstood sin[c)e by other [h)ands, [the) savages of 4 
 
 269 241 190 62 
 
 t|hat part never [a)ttempted to go o[ve)r to the i[sla)nd 4 
 
 262 195 109 169 
 
 [ a)fterwards. They [wer)e so [terrifi)ed with th[e )ac- 4 
 
 293 208 132 311 
 
 [ca)unts given by [t)hese four [me)n (for it seem[s th)ey 4 
 
 848 51 231 158 
 
 did [esc)ape the sea), th[at) th[ey) believed [whoe)ver 4 
 
 239 US 149 127 
 
 wen[t to th)at enchante[d is)land wo[ul)d be destr[oye)d 4 
 
 299 172 181 97 
 
 w[ith) fire from the g[ods.) This, howev[er), I kn[ew) 4 
 
 286 211 203 145 
 
 [not,) and therefore |was under [co)ntinual ap[preh)en- 4 
 
 246 189 233 265 
 
 so[ons) for a good w[hil)e, and kep|t always [upon m)y 4 
 
 255 296 291 ^ 163 
 
 |guard, and I and al[l )my army; for [as) there were [now) 4 
 
 348 277 220 172 
 
 (f]our of us, I wo[uld hav)e ventured |upon an [hu)n- 4 
 
 264 181 261 137 
 
 dred [of t)hem fa[irl)y in the op [en ) field at any time.| 4 
 
 53 
 No. of fixations, 53; Average duration, 206. So-; total time, 10,963.0(r. 
 
 CHART XXII C 
 No. 58 A 
 Subject E 
 
 462 167 192 229 334 130 62 
 
 And [thi)s old sa|vage was i|n the rig|ht, (fo]r, a|s I [understood 7 
 
 248 166 161 260 217 
 
 since by| other hands, [t)he sa|vages of that (part n[e)ver attempted 5 
 
 328 248 347 173 316 140 179 
 
 to go) o]ver [to) th|e island aft|rwards. Th(ey w]er|e so| terrified 7 
 
 a c b 
 
 334 210 297 216 102 372 179 
 
 wit(h ]the accou|nts gi(ven ]by the[s)e four m|en( (for ]it se|ems they 7 
 
 396 198 241 124 403 384 
 
 did e|scape the s|ea), that th[ey) believe|d wh[o)ever w[ent )to that 6 
 
 342 102 186 246 432 120 204 
 
 ench(an]ted islan|d would b|e destroy|ed wi|th fir|e fr|om the gods. 7 
 
 216 144 96 180 132 282 162 288 186 
 
 This, ho(w]|ever,| I kne|w n[o)t, and [th)erefo|re wa|s unde[r co)ntinual 9 
 
 b a 
 
 186 186 228 342 300 364 408 268 
 
 appreh|ensions| for a( g]o(od] while,| and ke|pt alw|ays upon| my 8 
 
 b a d c 
 
 300 264 150 294 222 
 
 gua|rd, I an[d)| a[ll) m|y army; for as there were now four of us, 
 
 b a a c d 
 
 Fixation pauses, 61. Average duration, 243.8(7 (M., 228); A. D., 76; total 
 time, 14,870(7-. Length of line, 113.5 mm. 
 
106 THE PSYCHOLOGY OF READING 
 
 A comparison of the charts shows that in each case more fixa- 
 tions are required in the short line in order to read the same num- 
 ber of words/— 47 as against 40 for S, (68 A and 64 A), 50 as 
 against 44 for subject E (59 A and B), or from % to % as many 
 again. A tendency to a motor habit, which is evinced by a repeti- 
 tion of the same number of fixation pauses in successive lines, is 
 evidently more marked in the short lines of each subject than in 
 the long. It is more successful in the readings of subject S than in 
 those of E. In the latter 's reading of the long line there is no in- 
 dication at all of a motor tendency. No two successive lines have 
 the same number of fixations. 
 
 The same results appear in the readings of 63 A and 69 A 
 (subject S). In this case the long line was read first. There are 
 35 fixations in the long line, 46 in the short line. These are printed 
 in Charts XX A and XX B on pages 102 and 103. Similarly, in 
 records 62 B and 69 B, which are printed in Charts XXI A and 
 XXI B on pages 103 and 104, the short line is three-fifths of 
 the long line. The latter requires 41 fixations for the passage, the 
 former 47. The span of attention is therefore somewhat smaller 
 in the short line. The increase of pauses in 69 B is doubtless, how- 
 ever, due in part to the necessity of carrying over long words di- 
 vided at the end of the line. 
 
 Cf. 'plenipote|ntiary' in long line with 'ple|ni- 
 
 pot[en)tiary 
 also 'There lay[ be) fore him' with ' There [ )lay jbe- 
 
 fo[re h)im'. 
 
 Subject E believed that in reading passages of this character 
 (as No. 59 A) he was accustomed to move his eye 'down the mid- 
 dle of the page.' That this is what was attempted, although quite 
 unsuccessfully, is perhaps shown by the first fixation of line 3, No. 
 59 A. The subject was evidently hampered by a line of this length. 
 It may be noted that in line 6 'thirty' was misspelled 'thirthy' in 
 the original paragraph, the backward movement of the first fixa- 
 tion and partial stop at (b) may denote the observance of that 
 peculiarity. Both subjects S and E had been reading some proof 
 and hence were more apt to notice these irregularities. 
 
 To summarize, more fixations are made in the short line ar- 
 rangement than in the long, e. g., 44 fixations for 5 lines of 59 A 
 
 ^ When all the lines and fixations of a passage are not taken for com- 
 parison, the number actually compared is indicated in the margin of the 
 charts. For example, the first five lines of Chart XVIII A on page 100 
 contain the same number of words as the first ten of Chart XVIII B on 
 page 101; the number of fixations in the first five lines of the first passage 
 is, therefore, added and placed in the margin after the fifth line. So also 
 in Chart XIX A, page 101, the first six lines only are needed for compari- 
 son with the short lines of XIX B on page 102. 
 
TEXT-LINES AND MOTOR HABITS 107 
 
 (long line) as compared with 50 fixations for 10 lines of 59 B, 
 (subject E) ; 40 fixations for the same passage as read by subject 
 S, (68 A) (the first six lines of the long line arrangement), and 
 47 in the short line arrangements of the same words; 35 fixations 
 in six lines of 63 A (long lines of short words) and 46 fixations in 
 the same amount of subject matter arranged in ten lines in 69 A. 
 
 In these cases the long lines are twice or nearly twice the length 
 of the short lines. In 68 B the line is less than a third as long 
 again as 63 B. The same fact holds true, but to a very much less 
 extent, the number of fixations being 51 for the long line, 53 for 
 the short. The long line is in this case of but 113.5 mm., the short 
 of 87.5 mm. As this is a printed page, the difference is, however, 
 largely counterbalanced by the larger number of words or letters 
 per line. This advantage is secured by the greater compactness of 
 the type setting and spacing. On this account the short line of 
 63 B is practically as advantageous as the longer line of No. 68 B 
 as regards the number of movements and fixations that are neces- 
 sary, and it has several advantages which are lacking in the longer 
 line; first, its length is better suited to the formation of short-lived 
 motor habits, and, secondly, it has an advantage in that it can 
 (perhaps on this very account) be read in less time. 
 
 The first four short line passages because of this dift'erence in 
 the compactness of type and spacing have too few words to the line. 
 The passages are too much divided, or cut up, and cause an un- 
 necessarily large number of fixations per line. They, as will be 
 seen, have an advantage in respect to the two desiderata, i. e., rhyth- 
 mical movements and short duration, just mentioned in the case of 
 No. 63 B, but, whereas in this latter passage the span of attention 
 (as denoted by the small number of pauses) was as extensive as 
 in the longer line, it was, so to speak, unnecessarily 'cramped' in 
 the former passages. 
 
 Table X 
 
 Total number of fixations in lines of different lengths made in 
 equal amounts of the same subject matter read by the same per- 
 S071 at intervals of twenty days. I. Long line arrangement. 
 II. Short line arrangement. 
 
 Nos. of Record 
 I II 
 
 Ratio 
 Length 
 
 No. of 
 
 Fixations 
 
 I II 
 
 Lengths 
 I 
 
 of Lines 
 II 
 
 Av. 
 Letters 
 I 
 
 No. of 
 1 per Line 
 II 
 
 68A 
 
 64A 
 
 2-1 
 
 40 
 
 47 
 
 179-190 
 
 86- 98 
 
 57 
 
 29 
 
 59A 
 
 59B 
 
 2-1 
 
 44 
 
 50 
 
 179-190 
 
 86- 98 
 
 57 
 
 29 
 
 63A 
 
 69B 
 
 5-3 
 
 35 
 
 46 
 
 179-178 
 
 97-106 
 
 52 
 
 30 
 
 62B 
 
 69B 
 
 5-3 
 
 41 
 
 47 
 
 179-183 
 
 97-106 
 
 62 
 
 38 
 
 68B 
 
 63 B 
 
 li-1 
 
 51 
 
 53 
 
 113.5 
 
 87.5 
 
 49 
 
 37 
 
108 TEE PSYCHOLOGY OF BEADING 
 
 To consider next the duration of the pauses, it will be noted 
 (cf. Table XI), that the average duration of the pauses is uni- 
 formly less in the shorter lines. When the short line is made of 
 sufficient length and printed in such type and spacing as make 
 wider fixation groups possible, and so equalize the number of 
 pauses, the total time per passage is also decreased. It is probable 
 that the ' practice effect ' carried over in some cases ; particularly 
 might this be the case on account of the peculiar style and subject 
 matter that is found in 62 B. 
 
 Table XI 
 
 Duration of fixation in long and short lines compared 
 
 Records Total Times Av. No. of Fix. Av. A. D. 
 
 Long line 68A 8481.0 212.0 40 6.6 .7 
 
 Short line 64A 9419.5 200.4 47 3.9 .31 
 
 Long line 59A (Siines) 11088.8 252.0 44 8.8 1.3 
 
 Short line 59B(10 lines) ^118.8 182.4 50 5 .8 
 
 Long line 63 A (6 lines) 8548.8 244.2 35 5.9 .23 
 
 Short line 69A(io lines) 9400.0 204.3 46 4.6 .72 
 
 Long line 62B 15506.0 378.2 41 6.1 .52 
 
 Short line 69B 10670.0 222.0 47 4.2 .45 
 
 Long line 68B 11621.1 227.9 51 4.9 .09 
 
 Short line 63B 10963.0 206.8 53 4. .1 
 
 As has been suggested, the data presented seem to the writer 
 to warrant the hypothesis that the differences in the rate of read- 
 ing in the case of the same individual and between different in- 
 dividuals depend largely, when other conditions are constant, on 
 the ease with which a regular rhythmical movement can be estab- 
 lished and sustained. The peculiarities of this movement are two, 
 first, a succession of the same number of pauses per line, and, sec- 
 ondly, a certain fairly uniform arrangement in the order of long 
 and short pauses, viz., (1) the first pause of a line longer than suc- 
 ceeding pauses, and (2) a secondary increase in the duration of 
 the pauses near the end of the line. These peculiarities are mutu- 
 ally dependent, and they are due chiefly to differences in the length 
 of the text lines. If the line is of such length that it is not possi- 
 ble to secure at the first fixation of the line a fairly definite im- 
 pression of a large part of the line, the eye must advance, so to 
 speak, more cautiously, and devote its attention more equally to 
 each section. The reasons for this are : first, that the peripheral 
 perception being less exact, there is danger of confusion with the 
 lines lying immediately above or below the one being read; sec- 
 ondly, the incidental but constant concurrent impression of words 
 
TEXT-LINE8 AND MOTOR HABITS 109 
 
 lying in the lines above and below and in fact the general eharac- 
 eristies of those lines, are of no particular value and not infre- 
 (uently distracting in the ease of the long line, but are doubtless 
 T, distinct advantage in the short line. Suppose, for example, that 
 when the eye is fixating at the beginning or end of one line, a few 
 words from the next line are caught sight of. If this happens in 
 a very long line, the words would have little or nothing to do with 
 the immediate sense, and must, therefore, be disregarded; whereas 
 in the short line there are fewer intervening words and ideas, so 
 that the peripheral vision (as a fringe of subconscious perceptions) 
 may aid in keeping the sense. So also in the ease of misunder- 
 standings, words which are connected with those being read are 
 found in greater proximity in the shorter line and can be more 
 easily referred to. In a long line the matter lying immediately 
 above is more apt to belong in another sentence. 
 
 The data upon which this preliminary statement is based is the 
 following. In Table XII the median and average durations of all 
 the first fixations of each line, all the second, third, etc., are given 
 for the first four charts of this section, i. e., 63 A — long line, sub- 
 ject S; 69 A, (same subject matter in short line arrangement) 
 59 A, and 59 B, long and short line arrangement respectively (cf. 
 Table IX above for similar distribution), subject E. It will be 
 noticed that in each case the extent of the acquirement of a series of 
 movements with the same number of fixations varies concomitantly 
 with a close approximation to the above noted distribution of the 
 time of fixation. The duration of the pauses in the long lines is more 
 nearly the same in all parts of the line. The number of pauses 
 per line is more nearly uniform in the reading of Subject S (cf. 
 68 A) than in that of subject E, and it may be observed that the 
 pauses towards the end of the line, L e., the fourth, fifth, and sixth, 
 are somewhat shorter for S. The long line is read by E with an alter- 
 nation of long and short pauses. It seems to substantiate the sug- 
 gestion made earlier that often every alternate fixation in the long 
 lines tended to have a somewhat wider scope than the others. In 
 neither case is there the constant decrease that is found in 69 A 
 (short line, subject S). 
 
 There is in this record first a long pause followed by a much 
 shorter pause, the third pause is again long, and followed by a 
 constantly decreasing duration of pauses towards the end of the 
 line. The same is true to a less extent in the short line of sub- 
 ject E (59 B), but the acquirement of a motor habit was not as 
 successful in the latter 's case, (cf. Chart XVIII B), and, there- 
 fore, the above noted succession and duration is concomitantly less 
 marked. In Table XII the distribution of time is given by fixa- 
 
110 TBiJ PSTCBOLOGY OF BEADING 
 
 tions for the long lines, and in Table XIII the distribution for the 
 short lines. The accuracy with which the median and average 
 deviations indicate the above noted tendencies may be verified by 
 a comparison of the charts themselves. 
 
 Table XII 
 
 Distrihution of the medians of all first, second, third, etc., fixa- 
 tions. I. Long line arrangement. II. Short line arrangement 
 
 SUBJECTS S AND E 
 
 I 63A 
 I 69A 
 I 59A 
 II 59B 
 
 
 1st 
 
 2d 
 
 :id 
 
 4th 
 
 5th 6th 
 
 7th 
 
 8th 
 
 9th 10th 
 
 11th 
 
 12th 
 
 
 Fix. 
 
 Fix. 
 
 Fix. 
 
 Fix. 
 
 Fix. Fix. 
 
 Fix. 
 
 Fix. 
 
 Fix. Fix. 
 
 Fix. 
 
 Fix. 
 
 M 
 
 268.8 
 
 264. 
 
 240.0 
 
 216.0 
 
 182.4 223.2 
 
 
 
 
 
 
 A. D. 
 
 51.7 
 
 44.8 
 
 35.2 
 
 25.6 
 
 6i.8 91.2 
 
 
 
 
 
 
 M 
 
 275.2 
 
 180.6 
 
 257 
 
 159.0 
 
 100.7 129.2 
 
 
 
 
 
 
 A. M. 
 
 35.3 
 
 91.6 
 
 41.0 
 
 52.7 
 
 57.8 19.2 
 
 
 
 
 
 
 M 
 
 290.6 
 
 231.1 
 
 226.3 
 
 254.9 
 
 262.5 265.1 
 
 333.1 
 
 260.0 
 
 189. 195.3 
 
 279.8 
 
 160.4 
 
 A. D. 
 
 47.9 
 
 131.6 
 
 43.1 
 
 38.4 
 
 82.7 47.2 
 
 56.2 
 
 83.9 
 
 30.1 20.1 
 
 53.0 
 
 
 
 M 
 
 189.2 
 
 180 
 
 166.5 
 
 153.0 
 
 139.5 220 
 
 132 
 
 
 
 
 
 A. D. 
 
 39.3 
 
 46.6 
 
 27.9 
 
 24.7 
 
 32.2 
 
 
 
 
 
 
 
 
 
 
 
 Table XIII 
 
 
 
 
 
 Distribution of time in long and short lines 
 table a,^ short lines 
 
 
 1st Fix. 
 
 2d Fix. 
 
 3d Fix. 4th Fix. 
 
 5th Fix. 
 
 6th Fix. 
 
 7th Fix. 
 
 69A 
 
 M 275.2 
 A. D. 35.3 
 
 180.6 
 91.6 
 
 257.0 159.0 
 41.0 52.7 
 
 100.7 
 
 57.8 
 
 129.2 
 19.2 
 
 
 69 B 
 
 M 285 
 A. D. 64.1 
 
 255 
 47.7 
 
 240.0 140 
 31.4 49.0 
 
 175.0 
 
 26.7 
 
 165.0 
 
 
 
 14A 
 
 M 255. 
 A. D. 46.3 
 
 198.9 
 43.0 
 
 198.9 147.9 
 29. 64. 
 
 102.0 
 
 
 
 
 59 B 
 
 M 189.2 
 A. D. 39.3 
 
 180 
 46.6 
 
 166.5 153.0 
 27.9 24.7 
 
 139.5 
 32.2 
 
 220. 
 
 
 132 
 
 
 
 
 TABLE 
 
 B. LONG LINES 
 
 
 
 63A 
 
 M 268.8 264.0 
 A. D. 51.7 44.8 
 
 240.0 216.0 
 35.2 25.6 
 
 182.4 223.2 
 52.8 91.2 
 
 
 
 
 62B 
 
 M 378.2 384.4 
 A. D. 72.8 43.4 
 
 359.6 396.8 
 64.7 107.3 
 
 365.8 362.7 328.6 : 
 26.9 60.4 155. 
 
 272.8 
 
 
 
 59A 
 
 M 290.6 231.1 
 
 226.3 254.9 
 
 232.5 255.1 233.1 : 
 
 260.0 189.0 
 
 195.3 279.8 166.4 
 
 
 A D. 47.9 131.6 
 
 43.1 38.4 
 
 82.7 47.2 56.2 
 
 83.9 30.1 
 
 20.1 53.0 00 
 
 58A 
 
 M 328.6 198.4 
 
 241.8 222 
 
 300 310 179.8 : 
 
 273 187 
 
 
 
 ^ The average deviations are in some cases large, but in view of the 
 many factors (as difference in subject matter, etc.), which would work 
 against uniformity, as well as the individual differences of subjects, and 
 since the tendency to concomitance is itself so marked and constant, it 
 is believed the conclusion is justified. 
 
TEXT-LINES AND MOTOR HABITS 111 
 
 The distribution of the length of time spent in the different 
 parts of a sentence depends more upon the chance location in the 
 line of print than on any characteristics of the separate parts of 
 the sentence itself. 
 
 The subtraction of several millimeters from the end of long 
 lines is a sufficient change to increase the duration of the initial 
 pauses of the given lines, and to cause a difference in the distri- 
 bution of the duration of the other pauses. The addition of several 
 letters to the ends of short lines has the exactly opposite effect. 
 Other factors, as pointed out, enter somewhat into the explana- 
 tion of this change, but the most important factor is the mere dif- 
 ference in the length of line. 
 
 In Tables XIV A and B, the total duration of all the first fixa- 
 tions, of all the second fixations, etc., in a given passage is shown. 
 The four passages are selected at random from the charts pre- 
 sented in earlier sections. Table A is of long lines. Table B of 
 short lines (i. e., newspaper) .The number of fixations in each class 
 is added as a check on the total durations. The facts are the same 
 as above illustrated. 
 
 Table XIV 
 
 Distrihution of the total duration of all the first, second, third, etc., 
 fixations in four passages 
 
 
 
 
 A. LONG LINES 
 
 
 
 
 
 No. 
 
 Rec. Passage 
 
 
 Whole 
 Sum 
 
 1st 2d 3d 
 Fix. Fix. Fix. 
 
 4th 
 Fix. 
 
 5th 
 Fix. 
 
 6th 
 Fix. 
 
 7th 8th 
 Fix. Fix. 
 
 14A Evolutionary.. 
 
 
 11811.4 
 
 2751.1 2563.5 1897.4 
 
 1797.3 
 
 1522.7 
 
 1067.4 
 
 212 
 
 32A Evoluting 
 
 
 13055.3 
 
 ] 
 
 2418.5 1929.5 2156.3 
 B. SHORT LINES 
 
 1595.3 
 
 1360.9 
 
 2023.0 
 
 571.8 
 
 d 
 
 izi Passage 
 
 >< 
 
 o 
 
 6 
 
 a 
 
 o 
 
 No. 
 
 1st Fix. 
 No. 
 2d Fix. 
 
 No. 
 3d Fix. 
 
 6 
 
 12; 
 
 fa 
 
 No. 
 6th Fix. 
 
 82B St. Petersburg 60 
 
 16629.8 
 
 15 5366.7 15 3367.8 
 
 14 3718.7 12 
 
 3085.4 
 
 4 1091.2 
 
 43 A St. Petersburg 25 5307.5 8 2469.6 8 1620.8 7 1033.1 2 284. 
 
 In 14 A, Table XIV, a succession of the same number of fixa- 
 tion pauses per line is attained in a longer line than is usual, and 
 it is interesting to note that the distribution of time in the fixa- 
 tions corresponds to that usual in the short newspaper lines, and 
 
112 THE P8TCH0L0GT OF READING 
 
 seldom in the long. That is, we find a long first fixation followed 
 by shorter ones with an occasional secondary increase towards the 
 end. 
 
 Perhaps a better method than either of the above to illustrate 
 the tendency which may be seen best by a study of the charts 
 themselves would be to divide somewhat artificially the lines into 
 thirds or fourths, and count the duration of all fixations which 
 fell within those limits. The predominance of longer fixations in 
 the first quarter and a half of the shorter lines, might thus be best 
 illustrated. A brief study of the charts will, however, be sufficient 
 for the purpose. 
 
 We may now recur, at the risk of some repetition, to the ques- 
 tion that was raised earlier in the discussion, as to whether it was 
 the effect of subject matter or merely of length of lines which 
 caused the difference in the distribution of times. Is the first 
 fixation on the average longer because the eye must needs come 
 to a full stop there from the long return movement, and because, 
 secondly, it must read in this case both to the right and tq the left 
 of the point of fixation, whereas in the succeeding pauses it need 
 read only towards the right? If the longer pauses were due sim- 
 ply to this physiological difficulty, it should occur as noticeably in 
 the long line as in the short, in fact the preponderance should be 
 in favor of the long line because of the greater momentum of the 
 return sweep. The records, as above noted, show, however, that 
 this characteristic is peculiar chiefly of the short line. 
 
 To summarize, the fixation pauses are uniformly longer at the 
 beginning and at certain other parts of the line. The reason as- 
 signed for this peculiarity is that at these places a more general 
 perception is secured of the ideas and words that follow in the 
 line. The succeeding fixations serve to amplify and fill out this 
 general perception. Finally, this expanding of the field of atten- 
 tion is made more frequently and with greater ease in the short 
 line. Since, therefore, it is easier to read when attention and fixa- 
 tion coincide, it is advisable that the arrangement of printing 
 should be so made that the distracting effect of unconnected words 
 on the periphery of vision is minimized and the more important 
 and salient parts of the sentence be found in those places where 
 the eye for more purely physiological reasons naturally dwells.^ 
 
 ' It was the intention of ttie writer to supplement this part of the 
 discussion by suggestions, which seem feasible, and certainly help to- 
 wards the legibility of sentence structure in print. Their publication, 
 however, has been delayed due to the laak of sufficient experimental 
 data to warrant their appearance at this time. 
 
TEXT-LINES AND MOTOR HABITS 113 
 
 The elementary school books may again, unfortunately, be taken 
 to illustrate the violation of these conclusions regarding the length 
 of text lines. It is not an uncommon practice to have the first 
 readers somewhat wider than is common in general in books of 
 that size, and wider than the second readers that are next read. 
 Promotion from the primary and second grades to the third or 
 fourth grade often means in very many schools a change from a 
 comparatively long line to one which is a third shorter. If the 
 long line must be used at all, it should from every point of view 
 come later. 
 
 The length of text line which is used in several pages of a 
 school primer in very common use at the present time is 120 mm. 
 This is a longer line than that of any other book on my shelves, 
 and a third longer than that of the majority of books and maga- 
 zines in general. Several of the pages in question are printed in 
 type but iy2 mm. in height, and in addition the margins of the 
 lines are indented for pictures. A part of one of the pages is 
 printed below. As can be seen from a comparison of this page 
 with the one prepared with a view to testing this difficulty (cf. 
 'Refixations'), this passage, as regards the ease of acquirement of 
 a uniform motor innervation of the eye muscles, is on the whole 
 the much more difficult passage. Although the type was smaller 
 (1 mm. for the small 'n'), and the spacing narrower in the passage 
 experimented with, the length of the line was but 97.5 mm. as com- 
 pared with 120 mm. in the following passage, and the indentations 
 were but 30 mm. and uniformly of that extent. In this case they are 
 56.5 mm. on one side and 45.5 mm. on the other. The extent of 
 movement is, in the case of four lines (supposing the eye to come to 
 the edge of the page) , reduced from 120 mm. to 74 mm., and in the 
 case of five other lines to 63 mm. ; that is, the extent of motor inner- 
 vation must be changed by from one-half to a third of its total ex- 
 tent. The extent of the return sweep from the end of one line to 
 the beginning of the next is in one case reduced to less than 17 
 mm., or about one-seventh of its usual extent. The acquirement 
 of a uniform motor habit of eye movement is manifestly impossi- 
 ble in such a line arrangement. 
 
114 THE PSYCHOLOGY OF READING 
 
 Chart of Long Lines and Faulty Line Arrangement 
 prom a school reader 
 
 By and by the Sun saw Bessie and Mary start from home, and he 
 watched them until they reached the kindergarten. 
 
 The Sun looked into the kindergarten, aod through the windows, he 
 sent bright-colored fairies dancing along the walls, around the chairs 
 and over the piano. These are the chairs and the piano. 
 
 Here the Sun loved to linger and 
 to see the happy children working and 
 playing together. 
 
 Twelve o'clock came and the little 
 ones went home. The Sun was now 
 
 very high in the sky; but he managed to peep 
 into the windows of the dining room and to 
 catch a glimpse of papa, mamma, Bessie, and 
 Mary seated at the lunch table. 
 These are the chairs and table. 
 
 Soon the station was reached, and then there 
 
 was a long, dusty ride in the carriage. It did 
 
 seem rather hot, but the Sun was doing his best 
 
 to make the flowers smile and grow beautiful. A drink, sparkling and 
 
 clear from the well, refreshed every one. 
 
 This is the well. 
 The water at the bottom of this deep well was very cold, 
 
 for, try as he would, the Sun never could catch a glimpse of it. 
 
 "Never mind," he said, "I suppose people do want some- 
 
 thing cool these warm days, and so I will content myself 
 
 by making the water in the brook sparkle and flash — 
 
 like a hundred rainbows." 
 
TEXT-LINES AND MOTOR HABITS 115 
 
 It is in the writer's belief clearly indicated by the above ex- 
 periments that one of the essentials of natural and rapid reading 
 IS that the reader's eye should at once be able to acquire a regular 
 and uniform motor habit of reaction for each line. When the 
 length of line and other conditions remain constant this tendency 
 to a uniform system of movement can be quickly developed what- 
 ever, within certain limits, the nature of the subject matter may 
 , be. But, if no two lines are of the same leng1;h, as in the case in 
 the above passage, this is not only impossible, but the uncertainties 
 and difficulties of accurate movement, already described in the case 
 of a practiced adult, and indicated by the number of misjudgments 
 of correct muscular movements, regressive movements, and refixa- 
 tions in general, must naturally lead to a more cautious mode of 
 eye movement. This will be hard to overcome later, in the case of 
 children, and may cause unnecessarily slow readers. It is, of 
 course, true that the child at the start confines his attention to 
 very small sections and groups of words, and will not for some 
 time acquire a method of more general 'prevision' of the line as 
 a whole, but that gives evidently no justification for practicing 
 him on line arrangements which even the skilled adult finds difficult 
 to compass. 
 
 The problem of the best length of text lines can naturally not 
 be answered apart from a consideration of the size of type, the 
 number of letters per line, and the spacing between lines, and 
 similar questions.^ It has not been possible except incidentally to 
 extend the investigation to these questions. As several sizes of 
 type have been used in the experiments, however, it can be said that 
 the size of type should not be so large in comparison with the 
 length of line that but few words can be put on a line. The effect 
 produced is to make the passage disconnected and to increase un- 
 necessarily the number of movements and pauses. This was true 
 of the short typewritten lines. 
 
 A line of from 75-85 mm. or about a third longer than the ordi- 
 nary newspaper line of the New York dailies with type 1.5 in 
 height (of small letter 'n) combines a good many advantages. An 
 increase of spacing between lines, as is done on the editorial pages, 
 for example, increases legibility as much as a larger size of type. 
 
 ^ Cf. for these matters Griffing and Franz, Psychological Review, 189fii 
 512-530. 
 
CHAPTER XIII 
 
 Rapidity of Reading 
 
 In order to give a somewhat wider scope to the experiments, the 
 subjects were selected with a view to detecting such individual dif- 
 ferences in the methods of reading as might exist. As the wide 
 difference in the rate of reading is perhaps the most noticeable in- 
 dividual peculiarity, this was made the basis for the selection of 
 most of the subjects. In order to get the extremes, the slowest and 
 the most rapid readers were selected from a group of about thirty- 
 people. The selections read were simple narratives which would 
 seem as likely as possible to be equally familiar in style, etc. To 
 various factors, such as differences of interest in the tests, the prac- 
 tice effects from reading in one kind of subject matter, — as novel 
 or newspaper, — factors which doubtless ought to be considered in 
 any experiments which have the question of the rate of reading as 
 the primary problem,— it did not seem necessary to devote special 
 attention. The sole purpose was to find subjects which fell within 
 the general classes of rather slow readers and rapid readers. Sev- 
 eral series -of experiments with a large variety of different subject 
 matter and under widely different conditions supported the ob- 
 servation that certain individuals read more rapidly on the average 
 in all cases, although the comparative ratio in degree of rapidity 
 varies according to subject matter. 
 
 In Table XV the number of words read in 20 seconds by the 
 three fastest readers and by the five slowest readers of the group 
 are given. The subjects are arranged in the order of rapidity of 
 reading. There were nine different passages each selected from 
 "Robinson Crusoe," and printed on a separate sheet of paper. 
 Nos. 1-3 were printed in 12-point, Nos. 11-13 in 10-point, and Nos. 
 21-23 in 8-point type. Passages 1, 11, 21 were printed in what is 
 known as modern style of type, passages 2, 12, and 22 in old style, 
 and 3, 13, and 23 in 8-point 'Scotch' style of type.^ 
 
 ^ I am indebted to Prof. Thorndike for securing these blanks from the 
 printers. 
 
RAPIDITY OF READING 117 
 
 Table XV 
 Rate of reading {No. of words read) of eight subjects 
 Rate of Reading (No. of Words Read) of 8 Subjects 
 
 
 
 12 Point 
 
 
 
 10 Point 
 
 
 
 8 Point 
 
 
 
 1 
 
 o 
 
 3 
 
 U 
 
 12 
 
 13 
 
 21 
 
 22 
 
 23 
 
 1 
 
 229 
 
 211 
 
 196 
 
 228 
 
 203 
 
 2:^5 
 
 222 
 
 217 
 
 220 
 
 2 
 
 197 
 
 188 
 
 187 
 
 190 
 
 179 
 
 179 
 
 191 
 
 185 
 
 
 
 3 
 
 165 
 
 165 
 
 142 
 
 185 
 
 
 199 
 
 182 
 
 170 
 
 180 
 
 4 
 
 111 
 
 121 
 
 10b 
 
 109 
 
 105 
 
 129 
 
 lis 
 
 100 
 
 117 
 
 5 
 
 101 
 
 94 
 
 100 
 
 111 
 
 92 
 
 108 
 
 113 
 
 99 
 
 102 
 
 6 
 
 93 
 
 105 
 
 100 
 
 101 
 
 118 
 
 95 
 
 96 
 
 84 
 
 93 
 
 7 
 
 85 
 
 105 
 
 88 
 
 95 
 
 93 
 
 101 
 
 93 
 
 96 
 
 96 
 
 8 
 
 68 
 
 64 
 
 62 
 
 63 
 
 63 
 
 67 
 
 68 
 
 61 
 
 60 
 
 The principal line of demarcation is betweeen subjects 3 and 4, 
 and the interval, as may be seen, is sufficiently large and uniform 
 for the purpose stated above. Subject 3 read on the average over 
 a third more words than subject 4. 
 
 In order to estimate v^hat effect marked differences in subject 
 matter and style might have, seven passages of a considerable range 
 of difficulty and general character were also given as a test to a 
 class of 41 students. The general result has already been indi- 
 cated. As the purpose was simply to test the validity of the 
 method of selection of the subjects for the laboratory experiments, 
 they will not be further discussed. The results of a third test may, 
 however, be of sufficient general value. The subjects were gradu- 
 ate students of very different interests; (x) a mathematician, (y) 
 a teacher in secondary schools, (z) a psychologist. They were 
 asked to read at their leisure carefully and for information, but 
 in the way and method which they would ordinarily follow, sev- 
 eral passages of considerable length and completeness, which were 
 selected with a view to differences in subject matter. The tests 
 were carried out under much more uniform conditions than is pos- 
 sible in a general class test. The passages were selected from the 
 following books and chapters: 1. Carlyle's 'Essay on Voltaire'; 
 2. Darwin's 'Animals and Plants under Domestication,' Chap, on 
 Inheritance; 3. Pearfeon's 'Grammar of Science,' Chap, on Mat- 
 ter; 4. Murray's 'Psychology,' Chap, on Perception; 5. Scott's 
 'Talisman,' narrative passage; 6. Byron's 'Childe Harold,' Canto 
 I: 7. Boswell's 'Life of Johnson.' 
 
 The subjects are denoted by x, y, z; the order or rank of pas- 
 sages in general interest to the different readers is denoted by 
 A-G, A denoting the most interesting passage for that subject. 
 None of the passages were familiar to the subjects with the excep- 
 
118 TEE PSYCHOLOGY OF READING 
 
 tion of that from the "Talisman." The times are in minutes and 
 fractions of a minute. 
 
 Table XVI 
 
 Comparison of rate of reading in various kinds of styles and sub- 
 ject matter 
 
 Carlyle 
 
 Darwin 
 
 Pearson 
 
 Murray 
 
 Scott 
 
 Byron 
 
 Bos well 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 17* C 
 
 10.5 D 
 
 19 E 
 
 13 G 
 
 12 A 
 
 4 B 
 
 20 F 
 
 25.5 B 
 
 25.5 C 
 
 
 26 A 
 
 18 F 
 
 
 33.5 D 
 
 28 C 
 
 3o.6 B 
 
 42 .\ 
 
 24 D 
 
 20 E 
 
 8 F 
 
 48 G 
 
 X 
 
 y 
 
 z 
 
 *Minutes and fractions of a minute. 
 
 The conclusion which seems warranted by the above data is 
 that one who reads rapidly in a given style and class of subject 
 matter, will read somewhat proportionately faster than a slow 
 reader whatever within certain recognized limits the nature of the 
 style and subject matter. We are, therefore, it seems, justified in 
 treating the extremes of the above group of thirty as two separate 
 classes, i. e., rapid readers, and slow readers. 
 
 It has already been indicated that the short text-line favors the 
 formation of motor habits, and that the fast readers attain to this 
 method more readily than slow readers. The evidence would fur- 
 ther seem to show that the acquirement of a rhythmical succession 
 of movements is one of the means by which the fast reader attains 
 to his greater speed in reading. 
 
 This fact is well illustrated by the distribution of time in the 
 pauses of records 73 A and 73 B, Chart XXIII, and the somewhat 
 more regular alternation of the number of movements in 73 B. 
 The subject read the passage first in his usual way, (73 A). On 
 the second reading he was asked to read as rapidly as possible con- 
 sistent with getting the sense. The charts follow: 
 
RAPIDITY OF BEADING 119 
 
 CHART XXIII 
 
 Speed Tests 
 
 No. 73 A 
 
 A. Usual Rate of Beading 
 
 469 297 172 324 
 
 1 Even athletic[s ar)e not wa[nt)ing in this eas|tern unlv[ersity). The 4 
 
 166 221 216 291 276 136 178 
 
 2 at[hle)tic club| consists |of seven se[ct)ions — ro|wing, tra|ck athletic|s, base- 7 
 
 199 216 199 263 206 
 
 3 ball, footbal[l, la)wn ten[ni)s, swim|ming, Judo| (a kin|d of wrestling), 5 
 
 276 129 189 199 167 91 
 
 4 fen|cing and arch|ery. In the |spring, when| the rosy c|loud of| cherry 6 
 
 286 194 189 216 199 69 
 
 5 blosso|ms covers the ba.nk| of the Ri[ver) Sumida, the[ r) owing c|lub h|olds 6 
 
 169 269 333 273 238 
 
 6 a re|gatta. In the autu(mn], the athletic sjection hold|s a meetingt in) the 5 
 
 270 1-.6 202 192 124 171 
 
 7 recrea|tio|n ground of th|e university. jRunning, jum]ping, hurd[le ) races, 6 
 
 b a 
 
 39 
 
 Av., 5.5 
 
 Total nuraber of fixation pauses, 39; total duration, 8,392.1cr; average duration, 
 
 215.20-; median, 199. So-, A. D., 54.3. Length of text line, 107.5 mm. Height of type, 
 
 1.5 mm. 
 
 CHART XXIII 
 
 No. 73 B 
 
 B. Most Rapid Rate of Reading 
 
 217 164 168 
 
 2 Even athletics are [no)t wanting in this eastern u|niversity. The] 3 
 
 168 168 168 106 161 
 
 2 athletic [cl)ub consists of| seven sections — rowin(g, ] track athlet|ics, ba|se- 5 
 
 280 128 284 
 
 3 ball, football, l[awn) tennis, swimmi[ng, )Judo (a kind[ of w)restling), 3 
 
 256 200 124 200 
 
 4 jfencing and archery. I|n the spring, when the| rosy cloud o[f cher)ry 4 
 
 214 132 162 148 
 
 5 blossom|s covers the bank of the Riv|er Sumida, the row|ing club h|olds 4 
 
 362 140 214 
 
 6 a regatta. In the a[utu)mn, the athletic section| holds a meetin[g in t)he 3^ 
 
 269 117 247 209 
 
 7 recreatio|n ground of the university|. Running, |jumping, hurdle |races, 4 
 
 231 174 190 209 110 
 
 8 etc.(, la]st t|he whole afternoon, and[ the) scene is as ani[ma)ted as eve|n 5 
 
 171 268 190 236 129 
 
 9 [a) Yale-Princeton 'rooter' cou[ld) wish; the[ slop)ing hillsid[e) of the arena]- 5 
 
 192 240 384 173 
 
 10 like (gr]ou[nd) is filled with cheering crowd (s, ]and the mingling of cos(-] 4 
 
 b a 
 
 281 266 170 222 210 86 
 
 11 [)tumes, color|s and gestures |add to the anima|tion of the scene [. )In the| 6 
 
 362 129 92 262 
 
 12 matt|er of cuppleme|ntal athletics, we m|ay note that swimmin[g )is given 4 
 Total number of fixations, 50; total duration, 9, 967. So-; average duration, 199. 3o-; Me- 
 dian, 190; A. D., 51.3. 
 
120 THE PSYCHOLOGY OF READING 
 
 It is interesting in the first place to note that the subject be- 
 lieved introspectively that the attempt to secure an increase in the 
 rapidity of reading was quite unsuccessful. "The idea of reading 
 the passage rapidly seemed a hindrance rather than a help." The 
 subject believed that he generally read at his maximum rate and 
 thought this was particularly true in the first reading of this pas- 
 sage, as it was a much more interesting passage than the passage 
 that had preceded it. 
 
 The records show, however, that the subject's introspection was 
 very inaccurate; not only was the absolute time of nearly every 
 pause diminished, but fewer pauses were made, and the average 
 distance of the eye 's first pause from the left edge of the page 
 noticeably increased. The saving in time in the first seven lines 
 amounts to 3189o-, or nearly one-third of the total time. The num- 
 ber of pauses is reduced from 39 to 26 or one-third. The accelera- 
 ition was, therefore, due as much to the increase of the span of at- 
 tention as to a decrease of the length of time of the stops. 
 
 The most important characteristic, however, and the J)oint of 
 present discussion is the difference in the mode of distribution of 
 the times of fixation. There is a decrease in the average duration 
 of the pauses in the second reading, hut this decrease is not made 
 equally in all parts of the line. It occurs chiefly in the last half 
 of the line, and in fact not 07dy is the time of the first fixations 
 not decreased' on the second reading, hut there is a slight increase 
 in hoth the average and total time spent in the initial fixations. 
 The total duration of the first fixations in lines one to seven in- 
 clusive is 1702. So- in the first reading and 1761.3cr in the second. 
 The distribution in Table XVII is given in the average length of 
 time spent in (1), first fixations; (2), in second fixations, etc., of 
 the passages. 
 
 Table XVII 
 
 Distrihution of the average duration of all the first, second, third, 
 etc., fixations in two readings of the same passage, 73 A heing 
 read in the usual way, 73 B heing read as rapidly as possihle 
 
 •"3 
 
 « M ^ a a a ^ 
 
 fa r g; fe *. fa fc 
 
 -s _: _: XI J3 a S3 
 
 o ^ . Q Av. Duration o o o o o o o 
 
 S t»'.S - of ^i^- P^r > > i^ >■ > > > 
 
 CKj ""l&H <i Passage, A. D. -ij <J <) «ij<j .ij <; 
 
 73A 5.5 .76 215.2 54.3 248.2 226.9 211.6 250 201.9 152.6 178.2 
 
 73B 3.7 .61 199.3 51.3 249.9 197.4 198.7 196.3 150.5 85.1 
 
RAPIDITY OF BEADING 121 
 
 In the first reading (73 A) the time of fixation is distributed in 
 each part of the line more equally than in the second reading. But 
 in the latter reading the method of motor functioning is that al- 
 ready described of a long first fixation followed by one or more 
 much shorter fixations, etc. (see above). This result was brought 
 about solely by increasing the speed of reading. It is significant that 
 this same peculiarity is, as already pointed out, also a characteristic 
 of the form of movement principally of those whom the other tests 
 have shown to be rapid readers, and that the slow readers do not or- 
 dinarily show this characteristic. The cause of the slow^ness of read- 
 ing may doubtless be for the most part central, and quite apart 
 from this peculiarity. But it is important to note that this char- 
 acteristic form of movement is the usual physiological accompani- 
 ment in rapid readers, and wanting to a greater or less extent in 
 the slow readers. 
 
 To restate the hypothesis which has been proposed in explana- 
 tion of this fact, "The rapid reader 'distributes his attention more 
 readily at the initial fixation of the line, and is enabled on this 
 account to fall more readily into a uniform habit of movement." 
 
 It should be added that the observation is not offered as the 
 sole explanation of differences in the rate of reading. An important 
 cause of slow reading may be, for example, the excess of the auditori- 
 motor accompaniment of repressed articulation, etc., which is sim- 
 ply another way of saying that the assimilation processes are 
 slower. The effect of articulating is itself to decrease ordinarily 
 the span of attention, and, therefore, as argued, make the dura- 
 tion of all pauses more uniform. 
 
 On the other hand, the experiments indicate that there is a 
 chance for improvement in the rate of reading. A habit of slow 
 methodical plodding is often easily fallen into which varies little 
 whatever the sort and importance of the reading, but which allots 
 to whatever is read, if it is of ordinary difficulty, about the same 
 amount of time and attention, line by line and sentence by sen- 
 tence. It often seems, for example, that the careful dwelling upon 
 each word and phrase, which is the daily method of the classical 
 student throughout many years of study, helps not a little in fixing 
 such a habit of slow assimilation. Long continuous application, in 
 which the attention remains uniformly of a medium intensity, is 
 seldom alert and selective, or never reaches its highest pitch or its 
 lowest, is little suited to the formation of a habit of attention 
 adapted to an age of newspapers and magazines and ephemeral 
 books. What may be called speed tests ought on this account to 
 have some place in school methods of teaching reading as well as 
 in other subjects in which a selective character and alertness of 
 attention is demanded and cultivated. 
 
122 THE PSYCHOLOGY OF BEADING 
 
 To summarize the facts that have been brought out in this and 
 previous sections, — rapidity of reading is not necessarily correlated 
 with regularity of movement and steadiness or preciseness of fixa- 
 tion. Some of the fastest and slowest readers were found equally 
 regular in movement. A wider 'spanning' of attention,— as de- 
 noted both by the greater frequency of long pauses at the begin- 
 ning of the line, and by fewer fixations per line,— is characteristic 
 of the more rapid readers. The slow readers have a narrower span 
 or working extent of attention, and a greater total arc of move- 
 ment. This may best be seen by comparing the extreme fixation 
 points in reading the lines in Chart II. Finally, quickness of 
 reading seems to be closely associated with the formation of motor 
 habits. 
 
CHAPTER XIV 
 
 The Fatigue of the Eyes in Reading 
 
 It has been possible to study the question of visual fatigue in 
 this investigation only in connecti(n with the general problem of 
 the characteristics of the eye's movement in reading. It has been 
 tested, however, under two different experimental conditions. In 
 each the assumption is made that the presence of fatigue may be 
 indicated by a decline in the velocity of movement. 
 
 The first method employed for testing this effect of fatigue 
 was to secure a record of reading late at night after the subject 
 had used his eyes steadily for a long period. The experiment was 
 carried out completely with but one subject. The records were in 
 this case taken between 10 :30 and 11 :00 P. M., after the subject 
 had spent fourtepli hours in study, between six and seven of which 
 were spent in proof reading. The proof reading was done in the 
 laboratory, and the subject simply brought in a clipping from his 
 galley of proof, took his place before the camera, which had been 
 previously adjusted, and read, at least as far as introspection or 
 observation could show, in exactly the same w^ay that the reading 
 had been done during the preceding hour. The subject had also 
 been reading proof for several days preceding, having read on the 
 previous day for about eight hours. He was also tested on other 
 reading matter. These records were then compared with similar 
 records which had been made a week previous and in the day- 
 time. Records were also taken when the subject was fresh on the 
 following morning. The conditions seemed to be very favorable 
 for detecting evidence of fatigue, provided it were present and ex- 
 erted any influence upon the rapidity of movement or upon the 
 duration of the fixation pauses. 
 
 The results are as follows: The passage was first read (34 B) 
 at 10:30 P. M., after the previous work described. It was then re- 
 read (35 A) the next morning, when the subject was fresh and 
 before any regular study had begun. Record 34 B, total time for 
 six lines, 12,944o-; average, 287.64(t; number of fixations, 45; simi- 
 larly, Record 35 A, total time, ll,441o-; average, 243.42a; number 
 of fixations, 47. The passage is, therefore, read in the morning 
 with a saving of l,503o-. If this record stood by itself, it might 
 seem that the shorter time was due to general familiarity with the 
 
124 THE PSYCHOLOGY OF READING 
 
 passage from the first reading, although proof would naturally be 
 read somewhat more uniformly than ordinary narration. How- 
 ever, this possible criticism is removed in the next record, where 
 the practice effect, if any, would tend to conceal the effect of 
 fatigue. 
 
 The results for other pages are as follows : Record 14 A, ' Evo- 
 lutionary study,' etc., first nine lines, total time, 11,811 Aa; 49 fixa- 
 tions; average duration, 241a, A. D. 67.3; Record 32 A, read five 
 days later at 10:80 P. M. in experiment for fatigue, total time, 
 12,482o-; 52 fixations; average duration (for 54 fixations), 241.8a, 
 A. D. 70.9. The latter reading, therefore, took 670.6a longer and 
 three more fixations. The average times were practically the same. 
 Any familiarity with the passage that remained after the interval 
 of five days would have the effect of decreasing the duration and 
 number of pauses. The increase, therefore, is the more noticeable.^ 
 
 Record 18b, 'St. Petersburg,' etc., second reading by H (fifteen 
 lines), total time, 13,000a; 60 fixations; average duration (for 63 
 pauses), 216a, A. D., 50; 32 B^, read five days later between 10:30 
 and 11:00 P. M., total time, 16,620a; 60 fixations; average dura- 
 tion, 277.2a, A. D., 64.1. Loss, 8,630a. Record 83 A was read im- 
 mediately after 32 B^. The subject was instructed to articulate 
 the words. A comparison of the record with a similar one (with 
 articulation), made the next morning at about 10 o'clock, shows 
 again that the evening performance was much slower. The fa- 
 miliarity secured from repeated readings would perhaps be about 
 the same in this case, 33 A having been the second consecutive 
 reading, whereas, although 36 A was the third reading, it was 
 made after an interval of nearly twelve hours. The results are: 
 
 Record 38 A (fatigue), total time for nine lines, 11,070.6a; 
 number fixations, 51 ; average, 217.1a, A. D., 76.7. Record 36 A, 
 total time for nine lines, 10,040a ; number fixations, 41 ; average, 
 244.87a. The record of the morning hour was, therefore, read in 
 1,030.6a less than that of the evening previous. 
 
 Although the experiments were made with but one subject, the 
 results seem sufficiently constant to warrant their presentation. 
 For more convenient reference the results are brought together in 
 Table XVIII. 
 
 ^ Especially is this true as on the first day the passage was read four 
 times in succession. The record here given of 14A was the first of these 
 readings, but the other readings must have given some familiarity which 
 Avould tend to make even five days later the passage less difficult of 
 reading, and conceal the effect of fatigue. An exactly similar condition 
 was true in the next records, two other readings having been made on 
 the same day and after 13b. 
 
FATIGUE OF THE E7ES IN READING 125 
 
 Table XVIII 
 
 Comparison of readings, (1) under normul conditions, (2) when 
 
 fatigued 
 
 
 No. of 
 Record 
 
 No. Of 
 Fixations 
 
 Total 
 
 Time 
 
 6 
 
 Av. 
 
 A. D. 
 
 Loss in 
 Time in F. 
 
 d 
 
 Normal 
 
 Fatigue 
 
 14A 
 
 32A 
 
 49 
 52 
 
 11811 
 
 12482 
 
 241 
 
 241.8 
 
 67.3 
 70.9 
 
 670 
 
 Normal 
 
 Fatigue 
 
 13B 
 
 32B 
 
 60 
 60 
 
 13000 
 16630 
 
 216 
 277.2 
 
 50 
 64.1 
 
 3620 
 
 Normal 
 
 Fatigue 
 
 36A 
 
 33A 
 
 41 
 
 51 
 
 10040 
 11070 
 
 244.8 
 217.1 
 
 76.7 
 
 lOSO." 
 
 Normal 
 
 Fatigue 
 
 35A 
 
 34B 
 
 47 
 45 
 
 12944 
 11441 
 
 287.6 
 243.4 
 
 
 i'm 
 
 The second method was to measure and compare the times or 
 duration of the eye's movement, (1) when reading the long lines 
 of print, and (2) when reading the same subject matter in short 
 lines. Because of the longer times and their greater uniformity, 
 the return sweeps from the end of one line to the beginning of 
 the next were selected for measurement instead of the more vari- 
 able interlixation movements. Here, again, the assumption is that 
 fatigue would be evidenced by difference in the rate of movement. 
 Four records of two subjects were chosen at random for compari- 
 son. The subject matter, type, etc., was the same in each reading; 
 the only change being that the lines in one case were on the aver- 
 age just twice the length of the other passage. The results are 
 presented in Table XIX. 
 
 The first column gives the number of the return movement (al- 
 •■.hough not the same movement for each subject) ; the second and 
 third columns the extent of the movement in millimeters of the dis- 
 tance on the page as read; the fourth and fifth the duration of the 
 movement; and the sixth and seventh columns the average time 
 taken to move the eye across 1 mm. of the page. Table XIX A 
 gives the times for the long lines. Table B for the short lines. 
 
 Table XIX A 
 
 Rate of movement of eye in the return sweeps of a long line. Rec- 
 ords 68 A and 59 A 
 
 No. of Mt. 
 
 Extent 
 of Mt. In MM. 
 
 S E 
 
 Duration 
 of Mt. in (J 
 S E 
 
 Time for 
 1 MM. of Mt. 
 
 (in(J) 
 8 E 
 
 1 
 
 169 
 
 142.5 
 
 103 
 
 130 
 
 .61 
 
 .91 
 
 2 
 
 174 
 
 140 
 
 132 
 
 126 
 
 .76 
 
 .90 
 
 3 
 
 153 
 
 142 
 
 122 
 
 138 
 
 .80 
 
 .97 
 
 4 
 
 163 
 
 147 
 
 145 
 
 176 
 
 .89 
 
 1.19 
 
 5 
 
 169 
 
 150.5 
 
 146 
 
 215 
 
 .86 
 
 1.43 
 
 6 
 
 160 
 
 142 
 
 154 
 
 208 
 
 .96 
 
 1.39 
 
 7 
 
 150 
 
 
 145 
 
 
 .97 
 
 
126 THE PSYCHOLOGY OF BEADING 
 
 Table XIX B 
 
 Short line arrangement 
 
 
 Extent 
 of Mt. in MM. 
 
 Duration 
 of Mt. (in (J) 
 
 Time f>)r 
 ot Mt. 
 
 1 MM. 
 
 (in 6) 
 
 No. of Mt. 
 1 
 
 S 
 68 
 
 E 
 
 74 
 
 110 
 
 79 
 
 1.62 
 
 E 
 
 l.(i7 
 
 2 
 
 63 
 
 67 
 
 103 
 
 91 
 
 1.63 
 
 1.36 
 
 3 
 
 71 
 
 64 
 
 98 
 
 106 
 
 1.38 
 
 1.64 
 
 4 
 
 64 
 
 76 
 
 79 
 
 108 
 
 1.23 
 
 1.63 
 
 5 
 
 68 
 
 58 
 
 109 
 
 67 
 
 lj<0 
 
 1.15 
 
 6 
 
 68 
 
 68 
 
 117 
 
 75 
 
 1.72 
 
 1.10 
 
 7 
 
 48 
 
 72 
 
 80 
 
 88 
 
 1.66 
 
 1.22 
 
 8 
 
 62 
 
 67 
 
 101 
 
 80 
 
 1.69 
 
 1.19 
 
 9 
 
 68 
 
 62. 
 
 120 
 
 108 
 
 1.76 
 
 1.72 
 
 10 
 
 76 
 
 71 
 
 135 
 
 74 
 
 1.77 
 
 1.04 
 
 11 
 
 75 
 
 82 
 
 92 
 
 97 
 
 1.23 
 
 1.19 
 
 12 
 
 74 
 
 69 
 
 130 
 
 126 
 
 1.75 
 
 1.81 
 
 13 
 
 58 
 
 
 99 
 
 
 1.70 
 
 
 A reference to the times given in the last two columns of Table 
 XIX A discloses a rather surprising fact; namely, that with two 
 exceptions every movement in the case of each subject is longer 
 than that which preceded it. The average duration of a move- 
 ment—to state the fact in another way — in the case of the first 
 three movements is for one subject nearly a third shorter than that 
 of the last three movements, and for the other subject about a 
 ■ fourth shorter.^ This evidence of the presence of fatigue in the 
 long line, if the progressive retardation of movement may be so in- 
 terpreted, is, as an examination of Table B will show, not found 
 in the short line arrayigement. Although there is considerable 
 variation, and in some parts of the records of each subject a suc- 
 cession of a few numbers of increasing duration, the average of 
 the last six movements is but slightly longer than that of the first 
 six, one-fifteenth in one case and one-forty-fourth in the other, i. e., 
 very much less than the probable error. The evidence, therefore, 
 points against the conclusion of Landolt and others that move- 
 ments through a small angle are very fatiguing. Although both 
 of these movements are of considerable extent, — it is the movements 
 through the larger angle which cause the muscular retardation. 
 The phenomenon is peculiar and unexpected, when consideration 
 is made of the large number of reading movements made in a sin- 
 gle hour. For evidently if the retardation continuously progressed 
 at the above rate, the eyes would have ceased moving before an 
 
 ^ A similar fact has already been noted by Prof. Dodge, (Angle Ve- 
 locity of Eye Movements, Psychological Revieio, VIII, p. 155) in the dura- 
 tion of long changes in the line of regard. "The last four movements in 
 a series of ten always averaged longer than the first four." 
 
FATIGUE OF THE EYES IN READING 127 
 
 hour was over. A study of several other records would tend to 
 show that this increase of time is but a more marked instance of 
 what occurs in a much less degree in the short lines. After a short 
 succession of movements of decreasing rate, a fresh start is taken 
 with a return to the normal or to an accelerated velocity of move- 
 ment. This evidently is repeated somewhat as in the familiar fa- 
 tigue curves of other muscles. The interesting fact is, however, 
 that the decline in rapidity should be of such a marked character, 
 since the recovery must in that case occur so much more fre- 
 quently. 
 
 The phenomenon is certainly worthy of further study. That it 
 is most marked in the ease of the long line doubtless indicates the 
 undesirability of this lengih of text line. Incidentally the records 
 of Table XIX also show the relation between the absolute rapidity 
 of movement in long and short lines. The angle velocity is, as was 
 to be expected, greater in the longer sweeps. The difference, of 
 course, is insignificant and negligible, when one is comparing total 
 times spent in reading these different lengths of text line (as 
 above) and has, therefore, not been taken into account. 
 
CHAPTEE XV 
 
 Summary and Concl'usions 
 
 1. The movements of the eyes in reading are found by means 
 of photographic registration to be composed of a succession of dis- 
 tinct pauses or fixations and connecting movements. The fixations 
 are the periods of peripheral stimulation and perception, and have 
 been on this account the principal object of study. 
 
 2. The average number of fixations per newspaper line varied 
 among the eight adult subjects experimented upon from 3. to 7.1, 
 or an average of from 1.9 words to 1.0 words per fixation. In a 
 line of not quite double this length from 7.5 to 9.4 fixations were 
 made, averaging from 1.5 to 1.09 words per fixation. The results 
 of other passages were similar. 
 
 3. The number of fixation pauses to the line is dependent on a 
 variety of factors, which are themselves by no means constant. 
 
 The amount read per fixation is seldom equal to the extent of 
 the field of perception. It is roughly correlated to individual dif- 
 ferences in the rates of reading, fewer pauses being made by the 
 more rapid readers. Constancy in the number of pauses for a 
 group of lines, regardless of at least small differences of subject- 
 matter and form, gives evidence of the formation of motor habits. 
 
 4. A motor habit is evidenced by a rhythmical series of the same 
 number of pauses per line, and by a uniform method of time dis- 
 tribution. The latter consists of a comparatively long initial pause 
 followed by two or more shorter ones of decreasing duration. These 
 may be followed by a somewhat longer pause near the end of the 
 line. These motor habits are most easily acquired in the shorter 
 lines, and aid materially in the rapidity of reading. They are fur- 
 ther one of the characteristics of rapid readers. 
 
 5. Inaccuracies of motor innervation or of the functioning of 
 the eye muscles are in evidence here as in other types of eye move- 
 ment. With these are to be classed the refixations proper which 
 are peculiar to the reading process. Refixations occur chiefiy at 
 the beginning of the line, and most frequently in the first or sec- 
 ond line of a passage. They are in these cases due to the inac- 
 curacy of peripheral perception and motor adjustment due to false 
 e-stimation of linear magnitude. An experiment made by varying 
 the length of the return movement by means of the indentation 
 
SUMMARY AND CONCLUSIONS 129 
 
 of every other line within the margin showed, it is believed, that 
 after the first or possibly second horizontal movement, the resident 
 muscular sensations of angular displacement govern the extent of 
 movement of the succeeding return sweep. This is the basis of 
 the motor habit. In the first line, sometimes in the second line, 
 before a motor adjustment has been made, it is necessary to depend 
 solely on the peripheral local signs. The longer the line the more 
 Inexact these will naturally be; the result is that the inaccuracies 
 of over and underestimation are much more frequent in the long 
 lines than in the short lines. The experiment shows the advisa- 
 bility of short lines of uniform length. A requisite of an automatic 
 movement is uniformity of conditions. Inaccuracy of movement 
 and the necessity of refixation must ordinarily be a hindrance to 
 the rapidity of perception. 
 
 The larger number of refixations at the beginning of a passage 
 is due to the fact that the motor impulse is in the process of ad- 
 justment. Normally the extent of movements of the eye tends to 
 become uniform due to the formation of motor habits of reaction. 
 The errors of fixation are due in part to the subversion of these 
 motor tendencies. The difficulties of accurate fixations may be in- 
 creased by too frequent change of the arc of movement. Refixa- 
 tions are found in the case of three subjects in 21.6 per cent., 69 
 per cent., and 80 per cent., respectively, of the return movements 
 in the longest line, and in but 11 per cent., 18 per cent., and 18 
 per cent, in the short lines for the same subjects; 21 per cent, of 
 these are in the first line. 
 
 6. There is a wide difference among persons in the way in 
 which they move their eyes in reading. The movements of some 
 are very regular and the fixations precise and relatively 'steady'; 
 in the case of others, fixation and movement are at times almost 
 indistinguishable, i. e., the eye passes over a considerable part of 
 the line by a movement that is very much slower than the inter- 
 fixation movement. 
 
 These slow changes of the point of regard are, however, to be 
 classed as fixations because perception may take place during their 
 progress. It is suggested that this difference of 'poise' among the 
 fixations themselves may depend on differences in the rate at which 
 the subject matter lying on the periphery gains the center of con- 
 sciousness. The shifting is due to the muscular tension following 
 upon the separation of fixation and attention. It is argued that 
 these changes or fluctuations of the attention ordinarily take place 
 from one word group or syllable to another, and are evidenced and 
 accompanied either by the usual interfixation movement, or by the 
 above described shifting movements. That there is no similar ten- 
 
130 THh PSYCHOLOGY OF BEADING 
 
 sion or movements of the eye muscles in tachistoscopic reading, goes 
 to show that there is no fluctuation of attention during or after the 
 exposure. The assimilation of the memory (not clear) image 
 must be successive as in ordinary reading. But it is a succession 
 not of letters, whether 'domineering' or other, but a succession of 
 syllables and words. The perception at each fixation pause is uni- 
 tary and simultaneous. 
 
 7. The hypothesis of lack of perception during movement pro- 
 posed by Erdmann and Dodge as an explanation of the need and 
 existence of fixation pauses, and since supported by experimental 
 evidence by the latter of these writers, has been further substan- 
 tiated by a measurement of the duration of the true interfixation 
 movement. The rapidity of these movements is ordinarily sufficient 
 to cause fusion of stimuli, and hence lack of perception. 
 
 8. The duration of the fixation pauses for four subjects in sim- 
 ple newspaper reading (the same passage being read by all) was 
 found to vary from an average of 160.8o- to 401.9cr, and averaging 
 from 3.7 to 7. fixations per line. There is a tendency for the slow 
 reader to make both more pauses and longer pauses. The initial 
 pause is usually the longest in the line. This is believed to be due 
 to a wider spanning of the attention at the beginning of the line. 
 
 9. The attention ' span ' and the extent of fixation in the reading 
 of numbers was limited in my subjects to one or two digits. This 
 is due to the fact that except in the case of familiar dates as of the 
 current year, etc., we seldom meet with the same combination of 
 numbers. For this reason a group of numbers does not become a 
 ' Gasammtbild ' as in the case of words, whose elements are relatively 
 fixed. It is further found that these minute changes of the atten- 
 tion from one digit to another are accompanied by similarly minute 
 changes of fixation, and it is held that, if such changes and fluctua- 
 tions of the attention, as are supposed by Zeitler and Messmer in 
 the reading of letters and domineering complexes of words, actu- 
 ally took place, we should expect to find some such indication in 
 the characteristics of the fixation pauses. The absence of such in- 
 dications argues that words are usually perceived simultaneously as 
 wholes and not successively by letters. 
 
 10. As regards the location in the line of the points of fixation, 
 it is found that the first and last fixations generally fall within the 
 edges of the line, i. e., at a little distance from its beginning and 
 end. This is true in less degree for the slow readers and in general 
 for readers of small or contracted attention span. It is further 
 dependent upon the character of the end words of the line. The 
 exact object of fixation is significant only as representing the point 
 about which are grouped the 'block' of letters that are simulta- 
 
SUMMARY AND CONCLUSIONS 131 
 
 neously perceived as one word or phrase complex. It more often 
 falls in the first third than at the center of a given perception area. 
 A new fixation is made for each unit of perception. The area cov- 
 ered by a given fixation depends on the size or extent of the per- 
 ception unit. It is large in the case of nouns, and adjectives, and 
 verbs, and usually small in the case of the connective parts of 
 speech, the conjunctions and prepositional phrases, the relative pro- 
 nouns, and the auxiliary verbs. Letters and words frequently and 
 regularly found together tend to be formed by long association into 
 one complex whole for which certain elements are then selected as 
 cues. Adjectives and nouns that frequently stand together, or for 
 which the general context has prepared the reader, will be read in 
 one unitary perception. So with many phrases. When words are 
 not regularly associated into a single unchanging group, but ap- 
 pear now with one word and now with another, these words, unless 
 helped out by immediate context, must be perceived separately. 
 This principle has already been illustrated in the case of numbers. 
 
 11. This general law is further illustrated in proof reading, 
 reading a foreign language, reading aloud, and in reading by near 
 and far accommodation. As shown in the reading at far accommo- 
 dation, it is not the mere extent of words which can be easily seen 
 which determines the span of attention or the number of fixation 
 pauses. In the first three cases, i. e., proof reading, etc., the atten- 
 tion must be given to smaller units, with the result that the num- 
 ber of pauses is increased. 
 
 12. The readings of three children, nine to eleven years of age, 
 illustrate the development of unitary word perception. The fre- 
 quency of fixation pauses shows the innervation to be much more 
 divided in the case of the youngest subject (boy of nine years of 
 age). The more purely physiological difficulties have been largely 
 mastered at this a^e. 
 
 13. The length of the text lines is mainly important in its effect 
 upon the formation of motor habits. The rate of reading depends 
 in part on the ease with which a regular rhythmical movement is 
 established. The peculiarities of this movement are, as noted, 
 two, — a succession of the same number of pauses per line, and a 
 distribution of the duration of the pauses, in which the first pause 
 of a line is uniformly longer than the succeeding pauses followed, 
 in some cases, by a second increase near the end of the line. Those 
 lines are best suited to rapid reading which give opportunity for 
 a wide span of attention, but which are not of such length that the 
 peripheral perceptions from the end or beginning of a line are too 
 inexact and confused to be of value in determining the general 
 character of a large part of the line. If the lines are too long the 
 
132 THE PSYCHOLOGY OF BEADING 
 
 incidental concurrent impression of words lying in the lines above 
 and below, and, in fact, the general characteristics of those lines, 
 are not infrequently distracting. These concurrent impressions 
 are doubtless often a distinct advantage in reading short lines, 
 since the matter lying above and below the point of regard is then 
 more closely connected in sense with the immediate object of at- 
 tention. Uniformity of length of line is a requisite for the forma- 
 tion of motor habits of reaction. 
 
 14. 'Speed' tests in reading further illustrate the fact that in- 
 creased speed of reading is attained by means of the above de- 
 scribed characteristic motor reaction. Fast and slow readers are 
 most clearly dififerentiated by the failure of the latter to acquire 
 this mode of movement. The rapid reader distributes his atten- 
 tion more readily, especially at the initial fixation of the line, and 
 is enabled on this account to fall more readily into a uniform 
 habit of movement. The slow reader has a narrower span or work- 
 ing extent of attention. This may be simply (in addition to longer 
 pauses) another evidence of slower perception and assimilation 
 processes. Steadiness and precision of fixation are not correlated 
 with the rate of reading. 
 
 15. The presence of fatigue in the eye muscles is indicated,— 
 (1) by a slower rate of reading after prolonged study and reading, 
 and (2) by a progressive decrease in the velocity of movement es-, 
 pecially noticeable in the long lines. 
 
APPENDIX 
 
 For the purpose of more exact comparison which may be desir- 
 able in subsequent studies, I have had printed on this and the fol- 
 lowing page the first line or two of each of the various passages 
 used for experiment. They are set up in the same size, style and 
 arrangement of type as that of the original passages. The page 
 of indented lines read in the experiment on 'Refixations' is printed 
 in full. All other pages than those given below were typewritten. 
 The last two lines of the appendix illustrate closely the style of 
 type, amount of spacing, etc., of the typewritten pages, and the 
 lengths of the typewritten lines have in each case been stated in 
 the text. 
 
 St. Petersburg, Nov. 2.— The Admiralty 
 has telegraphed to the officers of the Baltic 
 fleet, who were left behind at Vigo in order 
 that they might testify, and who were on 
 their way to St. Petersburg, to remain in 
 Paris. 
 
 II. A license for promotion may be issued to the holder of a 
 teacher's license No. 1 who has had experience rated as equivalent to 
 three years of New York city public school teaching (including one 
 
 Summary. 
 Evolutionary study and thought have been hindered by the confu- 
 sion of two unrelated biological phenomna, (1) evolutionary progress 
 or vital motion, and (2) the origination or multiplication of species. 
 
 Even athletics are not wanting in this eastern university. The 
 athletic club consists of seven sections — rowing, track athletics, base 
 ball, football, lawn tennis, swimming. Judo (a kind of wrestling). 
 
 King Midas thought, and thought. At last he 
 looked at the stranger and said, **I wish that 
 everything that I touched may turn to gold." 
 
134 THE PSYCHOLOGY OF READING 
 
 O T3 
 . O 
 
 I 
 
 03 
 CO 
 
 H --H fc 
 •H © o 
 
 CO ({{ 
 
 S * ^ 
 
 TJX! i^Bo -^ ^ ^ "o o ®0Q^ 
 
 ^11 M^ § ^ o II ^g t ^ ^ 
 
 <-»-'>^ L_,v^a3 c/)<lJaJ ^ScS'S'S bfl® 
 
 •^(U-'-' •O-*' Ui'-'b*, <"M-re<l+e. jrtX. 
 
 bc^u, j^ir':::^ ort^ aj-'S^'^Ksp 25':::^ 
 
 X^oj ^SLJc« v^a^ -S^-s^Siio -H+^o 
 
 § g « Is^ s ^ -g iris Is 8 -• - 
 
 J! ^ = 2 5c., Br<3?^ « 
 
 GO 
 
 >^ 0) iH 
 
 OQ -t' ® 
 
 O 4^ © 
 
 O 00 
 
 ;i to 
 
 0) bfl CO 
 
 £! C © 
 
 H 88 PJ 
 
ACKNOWLEDGEMENTS 
 
 This investigation was conducted in the Psychological Labora- 
 tory of Columbia University during the college year 1904-1905. 
 Every facility of the laboratory was most generously placed at my 
 disposal. For these courtesies and for much kindly criticism and 
 interest in the experiments, I am indebted to Prof. J. McKeen Cat- 
 tell and to Prof. E-. S. Woodworth. I desire also to express my 
 thanks and obligations to Prof. Edward L. Thorndike of Teachers' 
 College, and to Prof. Eaymond Dodge of Wesleyan University. 
 Prof. Thorndike has taken part in the experiments, and given 
 much valuable assistance in the measurement of the records and in 
 the class tests. I am indebted to Prof. Dodge for whatever knowl- 
 edge I may have secured of the technique and methods of the ex- 
 periments. My thanks are also due to those who have acted as 
 subjects, Mr. "Vivian A. C. Henmon, fellow in psychology; Mr. G. 
 C. Fracker, assistant in the department of psychology ; Miss J. R. 
 Seibert, assistant in psychology at Teachers' College; Mr. M. A. 
 Martin, scholar in psychology; Mr. George D. Strayer and Mr. Ed- 
 ward C. Elliott, fellows of Teachers' College; and Mr. Smith, me- 
 chanician to the laboratory. Mr. Henmpn has taken part in nearly 
 every division of the investigation and served as subject in the try- 
 ing fatigue tests. Miss Seibert has also assisted with the utmost pa- 
 tience and accuracy in the tedious measurement of a considerable 
 number of the records and in the tabulation of results. I am also 
 indebted to Miss Amy Schiissler, principal of the Speyer School, 
 for help in the study of children's reading, and to Dr. F. Kuhl- 
 mann, of the University of Wisconsin, who has kindly read the 
 proofs. 
 
THE PSYCHOLOGY OF READING 
 
 AN BXPEKIMENTATj STUDY OF 
 
 THE BEADING PAUSES AND 
 
 MOVEMENTS OF THE EYE 
 
 BY 
 
 WALTEE FENKO DEABBORN, Ph.D. 
 
 Instructor in Educational Psychologfy, ' 
 
 University of WisconsiiL< '^ \ v\ ■ \ l ij •?. 
 
 
 
 archives of 
 philosophy, psychology and scientific methods 
 
 EDITBD BY 
 
 J. MoKIJEN CATTBL.I. AND FEBDEBICK J. E. WOODBRIDGE 
 
 No. 4, MABCH, 1906 
 
 Columbia University Contributions to Philosophy and Psychology, Tol. XIV., No. 1 
 
 NEW YOBK 
 
 THE SCIENCE PRESS 
 
 't,'^^ 
 
ARCHIVES OF PHILOSOPHY 
 PSYCHOLOGY AND SCIENTIFIC METHODS 
 
 Editorial communications should be addressed to Professor J. McKeen Cattbll, 
 Garrison, N. Y. , or to Professor Fkederick J. E. Woodbridge, Columbia University, 
 New York City. 
 
 Subscriptions and advertisements should be sent to The Science Press, Sub- 
 Station 84, New York City. The subscription price is five dollars a volume contain- 
 ing between six and seven hundred pages. The numbers are as follows : 
 
 1. Measurements of Twins : Edward L. Thoendike. 50 cents. 
 
 2. Avenarius and the Standpoint of Pure Experience : Wendell T. Bush. 75 cents. 
 
 3. ThePsychology of Association: Felix Arnold. 50 cents. 
 
 4. The Psychology of Beading : Walter F. Dearborn. $1.00. 
 
 5. The Measurement of Variable Quantities : Franz Boas. 50 cents. 
 
 6. The Diurnal Course of Human Efficiency : H. D. Marsh. In press. 
 
 7. The Psychology of Mentally Deficient Children : Naomi Norsworthy. In press. 
 
 8. The Time of Perception as a Measure of DifiEerences in Sensation ; Vivian A. C. 
 
 Henmon. In press. 
 
 THE JOURNAL OF PHILOSOPHY 
 PSYCHOLOGY AND SCIENTIFIC METHODS 
 
 The contents of recent numbers include : 
 
 The Knowledge Experience and its Relationships : John Dewey. 
 
 Kant's Doctrine of the Basis of Mathematics : Josiah Royce. 
 
 The Psychical Complex called an Interest : Lucinda Pearl Boggs. 
 
 The Issue between Idealism and Immediate Empiricism : Charles M. Bakkwell. 
 
 Mental Elements of Dreams : Will S. Monroe. 
 
 Peeling and Conception : Kate Gordon. 
 
 Cognitive Experience and its Object : B. H. Bode. 
 
 Association and Atomism : Felix Arnold. 
 
 A Philosophical Confession : Harald HOffding. 
 
 A Syntactician among the Psychologists : Basil L. Gildersleeve. 
 
 The Real and the Pseudo Psychology of Religion : Irving King. 
 
 The Reeducation of an Aphasic : Shepherd Ivory Franz. 
 
 An Apparent Contradiction in the Modern Theory of Judgment : W. B. Pillsbury. 
 
 An Empirical Definition of Consciousness : Wendell T. Bush . 
 
 How Two Minds can know One Thing : William James. 
 
 Some Outstanding Problems for Philosophy : Cassius J. Keyser. 
 
 Immediate Empiricism : John Dewey. 
 
 $3 a Year (26 numbers) 15 cents a Copy. 
 
 LIBRARY OF PHILOSOPHY 
 
 PSYCHOLOGY AND SCIENTIFIC METHODS 
 
 Theory of Mental and Social Measurements : Edward L. Thorndike. $1.50. 
 Science and Hypothesis : Henri Poinoar^. Translated by George Bruce Halsted, with 
 an Introduction by Josiah Royce. $1.50. 
 
 THE SCIENCE PRESS, 
 
 Sul>Station 84, New York City. 
 
 b Ap '03 
 
SPECIES AND VARIETIES: 
 
 THEIR ORIGIN BY MUTATION 
 By HUGO DE VRIES 
 
 Professor of Botany in the University of Amsterdam. 
 
 Edited by DANIEL TREMBLY MACDOUGAL 
 
 Assistant Director of the New York Botanical Garden. 
 
 Cloth, gilt top, xviii + 847 pages. Price, postpaid, $5.00 (2Is.). 
 
 " Prof, de Vries may well be regarded as the foremost advocate of experimental evo- 
 lution, the man, moreover, who gave us the mutation theory of organic evolution. The 
 volume before us is a splendid scientific plea for the experimental study of organic 
 life." — Scientific American. 
 
 " It will be seen that Professor de Vries has made an important contribution to 
 science and thought. It is a masterly work and it is attractive because of the popular 
 style in which the experiments are described and the conclusions stated. Every student 
 of science will welcome it." — The Scotsman. 
 
 THE OPEN COURT PUBLISHING COMPANY 
 
 J322 WABASH AVENUE, CHICAGO 
 
 EDUCATIONAL REVIEW 
 
 ANALYTICAL INDEX 
 To Volumes 1 to 25, January, 1891, to May, 1903 
 
 By CHARLES ALEXANDER NELSON 
 
 Head Reference Librarian, Columbia University Library. 
 
 8vo, cloth, $3.00. By mail, $3.J2. 
 
 " The fortunate libraries and individuals who possess complete sets of the Review 
 will be grateful for this work, which now transforms a row of bound volumes into a 
 veritable encyclopedia of education, ready for consultation upon almost every imagin- 
 able modern educational problem. Analysis by topics is the leading feature of this 
 index, and the work has been intelligently performed. The work is a dictionary cata- 
 logue (author and subject), extending to upwards of two hundred double-columned 
 pages." — The Dial, Chicago, 111. 
 
 EDUCATIONAL REVIEW PUBLISHING COMPANY 
 COLUMBIA UNIVERSITY, NEW YORK 
 
 EDUCATIONAL PSYCHOLOGY 
 
 By Professor EDWARD L. THORNDIKE 
 
 In this book Professor Thorndike applies to a number of social and especially educa- 
 tional problems, the methods of exact science. The topics are treated in the light of the 
 most recent researches and with the aid of modern statistical technique. The book thus 
 provides those interested in education as a profession or as a feature of American life 
 with a sample of scientific method in this special field as well as with important infor- 
 mation which has hitherto been inaccessible. The attitude of the author, who is the 
 head of the department of Educational Psychology in Teachers College, Columbia Uni- 
 versity, and the author of numerous original contributions to dynamic psychology, is 
 that of a candid and painstaking student of the work that has been done in this field and 
 upholds rigorous ideals of scientific accuracy and logic. The book is so written and 
 illustrated as to be readable and teachable. 
 
 LEMCKE & BUECHNER, 8J2 Btoadway, New York 
 
X 
 
 Standard Boo ks on Psy chology, Etc. 
 
 By Professoi GEORGE STUART FULLERTON 
 
 Professor of Philosophy, Columbia University 
 
 A System of Metaphysics 
 
 Those who have been attracted by Professor FuUerton's articles 
 from time to time will be interested in his work as a whole ; only 
 small portions of it have so far been printed. 
 
 Cloth, B-vo,l%^XiO net (postage 19c.). 
 
 By EDWARD BRADFORD TITCHENER, Ph.D 
 
 Sage Professor of Psychology at Cornell University. 
 
 A Primer of Psychology 
 
 Outlines, methods and results of modem psychology. Cloth, $1.00 net. 
 
 An Outline of Psychology 
 
 Third Edition. Cloth, 8vo, $1.50 net. 
 
 Experimental Psychology 
 
 A MANUAL OP LABORATORY PRACTICE. 
 
 Volume I. Qualitative Experiments. In two parts. 
 
 Part I. Student's Manual. Cloth, $1.60 net- 
 
 Part II. Instructor's Manual. Cloth, $2.50 net. 
 
 Volume II. Quantitative Bxperiments. Preparing. 
 
 By JOSIAH ROYCE, Ph.D,, LL.D., Harvard University. 
 
 Outlines of Psychology 
 
 An Elementary Treatise with Some Practical Applications. 
 
 Cloth, 8vo, $1.25 net (postage 20c.). 
 
 By GEORGE M. STRATTON, Ph.D., University of California. 
 
 Experimental Psychology 
 
 And its bearing Upon Culture. Cloth, $2,00 net. Postage 13c. 
 
 Dictionary of Philosophy and Psychology 
 
 Including many of the principal Conceptions of Ethics, Logics, 
 ^Esthetics, Philosophy of Religion, Mental Pathology, Anthropol- 
 ogy, Biology, Neurology, Physiology, Economics, Political and 
 Social Philosophy, Philology, and Education, and giving a Term- 
 inology in English, French, German and Italian. 
 Written by many hands and Edited by J, MARK BALDWIN, 
 LLtD.^ Johns Hopkins University, with the co-operation and assist- 
 ance of an International Board of Consulting Editors. 
 In three volumes. Vols. I. and II. now ready, each $8 net. 
 
 When net books are ordered from the publishers, carriage is uniformly an extra charge. 
 
 THE MaeMILLRN eOMPaWY, Publishers 
 66 Fifth Avenue, New York 
 
 t\. 
 
/./ 
 
 i V