HUE on the last date stamped below 
 
 UNj\/-'r " BRANCH 
 
 CALlFORNfA 
 L/BRARY 
 
 L-'S ANGELES. CAUF.
 
 FREEHAND PERSPECTIVE 
 AND SKETCHING
 
 FREEHAND 
 PERSPECTIVE 
 AND SKETCHING 
 
 PRINCIPLES AND METHODS OF 
 EXPRESSION IN THE PICTORIAL 
 REPRESENTATION OF COMMON 
 OBJECTS, INTERIORS. BUILDINGS 
 AND LANDSCAPES 
 
 BY 
 
 DORA MIRIAM NORTON 
 
 INSTRUCTOR IN PERSPECTIVE, SKETCHING 
 AND COLOR, PRATT INSTITUTE, BROOKLYN 
 
 FOURTH EDITION 
 
 BROOKLYN 
 PUBLISHED BY THE AUTHOR 
 
 1916 
 
 S9 18
 
 N 
 
 Copyright, 1908 
 
 By Doua Miriam Norton 
 
 C139 
 
 THE UNIVERSITY PRESS, CAMBRIDGE, V. S. A.
 
 
 TO THE 
 
 MEMORY OF WALTER SMITH 
 
 FIRST DIRECTOR OF THE MASSACHUSE'lTS NORMAL ART SCHOOL 
 INSPIRING CRITIC AND JUDICIOUS FRIEND 
 
 THIS BOOK IS DEDICATED 
 
 WITH THE WISH THAT IT MAY HELP OTHERS AS ITS 
 AUTHOR HAS BEEN HELPED 
 
 D. M. N. 
 
 iz/rs'
 
 PREFACE 
 
 REVISED FOR THE SECOND EDITION 
 
 THIS book presents essentially the course of study in Free- 
 hand Perspective and Sketching as developed during its 
 teaching at Pratt Institute since the founding of the 
 institute in 1887. It consists of a series of illustrated exercises 
 with explanatory text, so covering the subject that students who 
 follow the course as directed acquire the power to draw with ease 
 and intelligence, not only from objects, but from memory and 
 from descriptions. The principles and methods here set forth 
 have been taught by the author for some years in the above school, 
 and have been found practically effective in that direction. In 
 revising it the author has drawn upon its use as a text-book 
 for large classes, as well as in other directions. 
 
 As offered to the public this course is intended to form a text- 
 book for classes in high, normal, and technical schools and in 
 colleges ; also as a book of reference for supervisors and teachers 
 of drawing, for draughtsmen and artists whose training in per- 
 spective needs to be supplemented, and for the instruction of 
 students so situated that personal art teaching is beyond their 
 reach. Since manuals for the teaching of drawing to children 
 already exist, the methods here presented are primarily such as 
 have been found effective with maturer minds. Its relation 
 to the teaching of children is thus like that of a grammar to 
 the ** language lessons ' ' of the primary schools. From it the 
 teacher, whether of children or of adults, may select material for 
 courses according to age or aims in study. 
 
 In the case of older students, though perspective books excellent 
 in certain directions have been published, it has been found diffi- 
 cult to direct inquirers to anything at once applicable to immedi- 
 ate use and comprehensive enough to give a working knowledge of
 
 PREFACE 
 
 the subject. For several years, therefore, the need which this 
 book is intended to meet has been increasingly felt. In the hope 
 that it may pass on to others the aid received in the past it is 
 sent forth. 
 
 The author gladly acknowledges indebtedness to many sources 
 in the making of this volume. Although the naming of all would 
 be impossible in this brief space some are so preeminent that 
 mention cannot be forborne. The experiences of teaching the 
 subject under the care of Mr. Walter Smith, then Director of 
 the Massachusetts Normal Art School, and later with Mr. Walter 
 Scott Perry, Director of the School of Fine and Applied Arts of 
 Pratt Institute, himself a teacher of great originality and force in 
 the subject for some years, have been most fruitful in the accumu- 
 lation of subject matter for this course. The author's drawings 
 have been so largely and sympathetically supplemented by the 
 work of Mr. Ernest W. Watson, now teaching in the Institute, 
 as to merit appreciation beyond that due the ordinary illustrator. 
 In the bringing out of the book the practical advice of Mr. 
 C. Franklin Edminster, many years an instructor in the same 
 school, and the critical taste of Mr. Henry Lewis Johnson, editor 
 of Tlie Printing Art, to whose suggestion is owing the form in 
 which the book appears, have been of great value. Of these aids, 
 and of others not mentioned, it is a pleasure to here express a 
 grateful appreciation. 
 
 D. M. N. 
 
 Brooklyn, June 24, 1910.
 
 CONTENTS 
 
 Page 
 
 Introduction xi 
 
 Chapter 
 
 I. General Directions 1 
 
 II. Pencil Measurement and the Picture Plane 4 
 
 III. The Ellipse 8 
 
 IV. A Cn-INDER AND A CYLINDRICAL ObJECT 12 
 
 V. An Object above the Eye and the Cone Principle 18 
 
 VI. A Cream Jug 20 
 
 VII. A Time Study 24 
 
 VIII. A Group of Cylindrical Objects 26 
 
 IX. Cylindrical Objects jGrouped with Fruit 29 
 
 X. A Group of OsjECTS^-wtOM Memory or Invention 31 
 
 XI. The Cylinder Cone and Ball Grouped — A Problem for Original^ 
 
 Study 34 
 
 XII. The Study of Straight Line Objects 36 
 
 XIII. Drawing the Book in Two Positions 43 
 
 XIV. The Book with a Cylindrical Object 45 
 
 XV. The Cylinder and Rectangular Block ^ A Problem for Original 
 
 Study 48 
 
 XVI. The Further Study of Straight-Line Objects — A Cube at Angles 
 
 WITH the Picture Plane 49 
 
 XVII. The Cube in Two Different Positions 53 
 
 XVIII. A Book at Angles to the Picture Plane 58 
 
 \ XIX. Two Books at Different Angles to the Picture Plane .... 61 
 -^XX. The Actual Center of the Circle and Measurement into the 
 
 Picture by Parallel Lines 63 
 
 XXL Books with a Cylindrical Object 67 
 
 XXII. The Study and Drawing of a House 69 
 
 XXIII. A Building from the Photograph or a Print 81 
 
 XXIV. Type Forms Helpful in Understanding the House — The Square 
 
 Frame 85 
 
 XXV. The Square Pyramid and Square Plinth 88 
 
 XXVI. The Square Frame Leaning on the Rectangular Block — A Prob- 
 lem FOR Original Study 9j 
 
 ix
 
 CONTENTS 
 
 Chapter Page 
 
 XXVII. Cylindrical Objects when not Vertical 92 
 
 XXVIII. A Group of Flower Pots 95 
 
 XXIX. The Circular Frame in a Square Frame 96 
 
 XXX. A Round Window 100 
 
 XXXI. The Clock — A Problem 102 
 
 XXXII. The Arch • 103 
 
 XXXIII. Interiors — A Room Parallel to the Picture Plane .... 105 
 
 XXXIV. Interiors Continued— A Room at Angles to the Picture Plane 110 
 XXXV. Further Studies of Interiors 114 
 
 XXXVI. A Chair 118 
 
 XXXVII. The Hexagonal Plinth in Two Positions 121 
 
 XXXVIII. Interior with a Tiled Floor 126 
 
 XXXIX. The Hexagonal Prism and Frame 128 
 
 XL. The Triangular Prism and Frame — A Problem for Original 
 
 Study 131 
 
 XLI. The Study op Parallel Perspective 132 
 
 XLII. A Street from the Photograph 137 
 
 XLIII. Exceptions to the Use of the Flat Picture Plane 139 
 
 XLIV. Shadows 143 
 
 XLV. Out-of-doors Work 154 
 
 SOLUTIONS OF PROBLEMS , I6I 
 
 INDEX 171
 
 INTRODUCTION 
 
 FREEHAND Perspective teaches those few principles 
 or truths which govern the appearance of things to 
 the eye, and the application of these principles to the 
 varied conditions encountered in drawing. Strictly speaking, 
 there are but two foundation truths in perspective, namely: 
 
 First. Things appear smaller in proportion to their dis- 
 tance from the eye. A house ten rods distant can be 
 wholly seen through one pane of glass (Fig. 8, 
 Ch. II). 
 
 Second. The eye can see surfaces in their true 
 shape only tvhen placed at 7'ight angles to the direc- 
 tion in which the eye looks, or, generally speaking, 
 parallel to the face. When not so placed they ap- 
 pear lessened in one dimension, that is, either nar- 
 rowed or shortened, in proportion as they are 
 turned away from the face or tend to coincide 
 with the direction of seeing. This apparent change of shape is 
 Foreshortening. The cylinder top held at right angles to the 
 direction of seeing appears as a circle (A in Fig. 1). When 
 turned away from this direction (as at B), it appears nar- 
 
 -^ 
 
 n rowed, or foreshortened. So the pencil seen its 
 
 /'^'-^'^ full length at A in Fig. 2 appears foreshortened 
 
 /y\ y\ when held as in B. All the phenomena of free- 
 
 ' A 'I B hand perspective, however complicated and per- 
 
 ^^^' ^ plexing, may be simplified by referring to one 
 
 or both of these principles. 
 
 One great obstacle to the ready mastery of these prin- 
 ciples is our knowledge of the actual shapes of objects. For 
 
 xi
 
 FREEHAND PERSPECTIVE 
 
 instance, we hnoiv the top of a cylinder (B, Fig. 1) to be 
 in fact a circle, and therefore we tend to mentally see a circle, 
 though it is just as truly a fact that the top can only appear 
 to the eye as a circle when the cylinder is held so as to lose 
 sight of all other parts of it, as at A. Consequently, the first 
 aim and benefit in studying perspective is the learning to see; 
 that is, to know what is the image really presented to the eye. 
 Therefore no step should ever be passed without clearly see- 
 ing the appearance under consideration. And in all drawings 
 the final test must be the eye; for, unless the drawing looks 
 right, it is not right. All rules and tests are only means to 
 this end. 
 
 Furthermore, the right study of perspective, which is think- 
 ing and drawing in perfect coordination, enables the student 
 to draw objects singly or combined or in unfamiliar positions, 
 without having them in sight. Also he should be able to 
 draw an object which he has never seen if a description of it 
 can be supplied. That this last is quite possible any prac- 
 tical artist will agree. The writer recalls hearing a popular 
 illustrator ask in a company of friends, " Does any one know 
 what a cider press is like? " adding that he must put one 
 in an illustration with no chance to see the thing itself. No 
 doubt of the sufficiency of a description was expressed, and in 
 this case it must suffice — a not uncommon situation. Hence 
 the necessity of memory work and dictation problems, such as 
 form part of this course of study. 
 
 Finally, it is not intended that in later practical work drawings 
 should be actually constructed by the explanatory methods here 
 given. These exercises should be drawn as directed, since only 
 by the actual experience of doing it can their principles be mas- 
 tered,, but a rigid clinging to these methods in practice would 
 result in very little art. Freehand Sketching means dratving by 
 the trained eye and judgment^ only using constructive methods to 
 test new or doubtful points. It is to make such sketching valu- 
 able by a foundation of definite knowledge that these methods 
 
 xii
 
 INTRODUCTION 
 
 are given. The trained artist draws a vase in his flower study, or 
 a round tower in a landscape with no distinct recalling of ellipse 
 laws, feeling only joy in the living curves as they spring out 
 under his hand. But he would labor long and wearily over their 
 shaping had he not this foundation knowledge, which he uses 
 almost unconsciously. 
 
 xiu
 
 Chapter I 
 
 GENERAL DIRECTIONS 
 
 MATERIALS. — Any paper having a fine and fairly soft 
 texture can be used. It should produce an even 
 grain in both vertical and horizontal pencil strokes. 
 Pencil exercises such as those reproduced in this book are 
 usually drawn on paper of quarter imperial size (11" x 15"), 
 on which at least an inch and a half of margin is allowed. 
 This is a good size for the student's drawings, whether copied 
 from these exercises or drawn from objects. Have two 
 pencils, one fairly soft (as No. 2 Faber, SM Dixon, or 
 2 B Koh-i-noor) , and a harder one ; also a good eraser. 
 
 Line Practice. — Cut the pencil like the illustration 
 (Fig. 3), and rub on practice paper ^ till a broad line, 
 
 firm at the edges 
 and transparent 
 (that is, with the 
 grain of the paper 
 slightly showing 
 through it) can be 
 made. Sit erect, 
 with the paper directly 
 in front, and have the 
 desk top inclined, or use 
 a drawing board (Fig. 4), 
 that the paper may be as 
 nearly as possible parallel with the face. Hold the pencil almost 
 flat, as in the illustration (Fig. 5), and as loosely as is consistent 
 
 ^ Save spoiled sheets for this. Practice paper should be like that on which drawings are 
 made. 
 
 Fig. 3 
 
 Fig. 4
 
 FREEHAND PERSPECTIVE 
 
 Fig. 
 
 with a steady control. For horizontal lines use position A, 
 Fig. 5, moving the pencil from left to right; for vertical lines 
 use position B, moving from the top downward. Practice 
 vertical, horizontal, or oblique lines persistently; moving the 
 hand freely from the shoulder, not resting it on the wrist or 
 
 elbow. If the muscles acquire an 
 unpleasant tension, relax by dropping 
 the hands at the sides and loosely 
 shaking them. Unfamiliar or diffi- 
 cult exercises should be first carefully 
 sketched with a thin, light line. If 
 wrong, draw over without erasing 
 until a satisfactory form is obtained. 
 Erase the incorrect part, and ren- 
 der expressively (Ch. IV). But after 
 the composition of the exercise is 
 planned, such straight lines as mar- 
 gins, cylinder sides, and many ellipses may be drawn in full at 
 once. And as the student gains in skill, more and more of the 
 work should at the first touch be put on the paper as it is 
 intended to remain. Exact knowledge is to be acquired only 
 that artistic interpretations may be expressed with ease and 
 certainty. 
 
 Models for "Work. — Objects in common use have been chosen 
 for most of these exercises. Geometric solids are assigned only as 
 needed for the clearer elucidation of perspective truths. Neces- 
 sary models, as the cylinder, the cube, and others, should be made 
 by the student as directed. For forms (as the hexagonal frame) 
 too complicated to be easily made, the well-known wooden 
 models have been used. But after thorough mastery of the 
 simpler forms, most of the later lessons can be understood with- 
 out models. 
 
 Placing of Models. — All objects for study should be placed so 
 as to present their vertical surfaces in nearly their true shape to 
 the student. Thus if the model is to be near, as on the table 
 
 2
 
 GENERAL DIRECTIONS 
 
 at which the student sits, it is better to raise it a few inches 
 (Fig. 4). This will not be necessary if it can be placed four or 
 five feet distant. If the study is seen too much from the top, 
 the perspective will be unpleasantly violent, as in a photograph 
 where the camera has been pointed too much downward. 
 
 The Table Line. — To indicate a supporting surface under the 
 objects a horizontal line (A, B in Fig. 6) is used. It stands for 
 the back edge of the table or other horizontal support- 
 ing surface, and is called the Table Line. It should be 
 represented as further back than any portion of the 
 study. As will be observed later, it need not be used 
 if the supporting surface is otherwise suggested, as by 
 a cast shadow (Fig. 34). 
 
 All Work Freehand. — All work is to be done freehand, that 
 is, with no ruling, and no measuring other than by the eye 
 and pencil. 
 
 Fig. 6
 
 Chapter II 
 
 PENCIL MEASUREMENT AND THE 
 PICTURE PLANE 
 
 PENCIL Measurement. — Before studying the exercises which 
 follow, the beginner should become familiar with Pencil 
 Measurement. Place a book upright directly in front of 
 the eye. With one eye shut and the arm at full length (to ensure 
 a uniform distance from the eye) measure on the pencil held hori- 
 zontally the apparent width of the book. Then turning the pen- 
 cil, compare this dis- 
 tance with its height 
 (Fig. 7). (It is bet- 
 ter to take the smaller 
 distance first, and to 
 measure it into the 
 larger.) Compare the 
 proportions so found 
 with those obtained 
 by actual measure- 
 ment of the book. 
 But always get the 
 pencil measurement 
 first, for this compels the eye to do all that it can unaided 
 before showing by actual measurement how much better it can 
 learn to do. 
 
 Now turn the book away a little, and compare this new ap- 
 pearance of the width with the height (Fig. 12). 
 
 The Picture Plane. — Here we must learn to keep tlie pencil 
 parallel tvith the face in order that the pencil measurement 
 may be reliable. For this, go to the window, and stand facing 
 
 4 
 
 Fig. 7
 
 PENCIL MEASUREMENT, ETC. 
 
 WINDOV(/ USED AS 
 
 PCCTt/BE 
 
 , Plane 
 
 z 
 
 
 o 
 
 
 P 
 
 
 
 e 
 
 a 
 
 
 
 
 
 
 
 
 J 
 
 •0 
 
 
 
 a 
 
 
 
 
 2 
 
 
 
 
 <J 
 
 
 B. Plan of A 
 
 A. Showing use 
 
 OF WINDOW AS 
 PICTURE PUANE 
 
 the glass, so the face is parallel with it. Choose some object 
 seen through the window, as another house, and resting the 
 pencil against the glass measure its width and compare that 
 with its height 
 (Fig. 8). 
 
 Observe that 
 if the outline of 
 the house could 
 be traced by the 
 pencil on the 
 glass it would 
 form correctly 
 the apparent 
 shape of that 
 house. 
 
 This leads us 
 toseethata//j!;fr- 
 spective drcming 
 may be regarded 
 as placing on 
 paper the equiva- 
 lent of such a tracing on the glass. It will therefore be apparent 
 at once that pencil measurement, to be correct, must be taken 
 with the pencil held as if laid on such a pane of glass; or in 
 
 other words, 07i a plane parallel ivith and in 
 front of the face. This imaginary transparent 
 plane is called the Picture Plane, and is a 
 most important factor in all freehand draw- 
 ing. Thus, by turning or revolving the 
 pencil on the glass in front of the face, 
 that is, by revolving the pencil in the picture 
 plane, it can he made to cover the appearance of any possible line or 
 direction. For example, the sloping gable edge of the outside 
 house, though retreating from the eye and therefore foreshort- 
 ened, can be covered by the revolving pencil (Fig. 9), thus giving 
 
 5 
 
 Fig. 8 
 
 .•iflHH 
 
 Fig. 9
 
 FREEHAND PERSPECTIVE 
 
 the appearance or picture of its direction. Its apparent or fore- 
 shortened length can also be taken on the pencil and compared 
 with any other dimension, as the height of the nearest corner. 
 The essential requirement is that the pencil shall constantly lie flat 
 on this pane of glass; that is, on the picture plane. 
 
 We have therefore, in the use of pencil measurement on the 
 picture plane, a ready and accurate means of ascertaining any 
 direction or any proportionate dimension seen by the eye. It 
 cannot give us actual sizes, as the length of the gable in feet ; but 
 it will tell us how long the slanting line representing the gable 
 must be drawn in proportion to other parts of the house. In this 
 case, for instance, the sloping edge appears three-fourths of the 
 gable width. The difficulty in using this valuable aid with exact- 
 ness lies in the beginner's trouble in keeping the pencil always in 
 his invisible picture plane. Any distance between the eye and 
 the object may be assumed for the picture plane. But for 
 accuracy this assumed distance must be kept the same while 
 comparing sizes. This is easily done by sitting erect and 
 measuring at arm's length, supporting the elbow with the other 
 hand if needful. The student should then mentally see the pic- 
 ture plane, recalling that it is vertical^ or parallel with the face 
 when looking at the middle of the objects to be drawn. That is, 
 it is at right angles to what we may call the Central Direction 
 of Seeing. 
 
 The Central Direction of Seeing. — This extends from the eye to 
 the center of the objects observed, while the face and the picture 
 plane are parallel to each other and at right angles to it. The 
 picture plane may then be thought of as a transparent vertical 
 plane pierced in its middle by the direction of seeing. 
 
 We have said the central direction of seeing is at right angles 
 to the face. Since the face is generally vertical, the direction of 
 seeing is generally horizontal (A in Fig. 14, Ch. III). The com- 
 monest exception is that of being directed slightly downward (B 
 in same Fig.). In this case it cannot be at right angles to the 
 picture plane. It will, however, always appear at right angles to 
 
 6
 
 PENCIL MEASUREMENT, ETC. 
 
 it when looked at from above. That is, it is at right angles from side 
 
 to side, and in a plan will always be shown at right angles, as in 
 
 Fig. 8. 
 
 Return now to the seat (Fig. 7), and try pencil measurement 
 
 on the turned book. Imagine as clearly as possible the trans- 
 parent picture plane at arm's length, 
 
 on which the pencil may be revolved, 
 
 but through which it must never be 
 
 thrust. Starting with the pencil erect 
 
 (Fig. 10) drop it directly over to the 
 
 left (Fig. 11), watching carefully to 
 
 keep it from leaning back or forward. 
 
 Let another person help by turning the 
 
 book away while you measure it and at 
 
 the same time Fig. lo 
 
 keep the pencil from following it back- 
 ward as it is turned away (Fig. 12). 
 Thus as the book is turned, the pencil, 
 if it remains on the picture plane, shows 
 ^ the book to appear narrower or be fore- 
 shortened. What 
 
 is now sought for 
 is that which the 
 eye really sees as the width, not what the 
 mind knows it to be. It is of great im- 
 portance to distinguish sharply between 
 actual facts of form and size and the per. 
 spective appearance of them as presented 
 to the eye. Fig. iz 
 
 An excellent object for practice is a door. Stand facing a 
 closed door, and take its proportions by pencil measurement. 
 Then let some one open it, and observe the apparent decrease 
 in width. 
 
 For further consideration of the picture plane see Chapters 
 XXXIV, XLI, and XLIII. 
 
 7 
 
 Fig. 11
 
 H 
 
 Chapter III 
 THE ELLIPSE 
 
 AVING learned that the book cover and door appear 
 foreshortened in proportion as they are near to co- 
 inciding with the direction in which they are seen, 
 
 Fig. 12a 
 
 we naturally look for the same change in the Circle. 
 
 Making a Cylinder. — Fold 
 
 the long edges of a piece ^^ 
 of stiff paper (A in Fig. 
 12a) and roll it into a cylin- 
 der, tucking one end of the 
 paper under the folds of 
 the other (B in Fig. 12a). 
 
 Seeing the Ellipse. — Holding the cylinder vertically, 
 as in A, Fig. 13, and with one eye closed, raise it 
 slowly till on a level with the eye. The top now 
 appears as a straight line (B, Fig. 13). It is so fore- 
 shortened that its surface is entirely lost to sights 
 leaving only its edge visible. Now, keeping the 
 cylinder vertical, lower it till the eye sees into it 
 perhaps half an inch. Observe carefully the shape 
 formed 
 by the 
 top. 
 Turn it 
 so the 
 top appears as a cir- 
 cle (A in Fig. 14), 
 then, iiolding it ver- 
 tically again (as at 
 B), compare men- 
 tally the apparent shapes as the top is placed in the two different 
 positions. Now (keeping it always vertical) raise and lower the 
 
 8 
 
 Fig. 14
 
 THE ELLIPSE 
 
 Fig. 15 
 
 cylinder slowly, and note how the form of the top changes, 
 appearing rounder as it is lowered. 
 
 Symmetry of the Ellipse. — This peculiar shape, varying in round- 
 ness between the straight line and the circle, represents the ap- 
 pearance of the circle seen ohliqueli/, and is the Ellipse, one of the 
 most beautiful, spirited, and subtle of curves. While the circle 
 is formed by a curve bending equally 
 in all parts, the outline of the ellipse 
 is constantly changing in the degree b( 
 of its curvature. From the middle 
 of each side (A, A in Fig. 15) this 
 curvature increases smoothly to the 
 ends (B, B). Thus the ellipse may be divided by lines through 
 the middle of its sides and ends into four duplicate curves or 
 quarters. These lines are known as the Long and Short Diiime- 
 ters. On these two lines the ellipse must be symmetrical, what- 
 ever the proportion of the diameters to each other; that is, 
 whatever the roundness of the ellipse. 
 
 Testing the Ellipse. — A test useful to determine the correctness 
 of a drawing of the ellipse is sighting with one eye along the long 
 diameter. If the ellipse is perfect it will appear foreshortened to 
 a circle having a diameter equal to the short diameter of the ellipse. 
 But there is no test of the ellipse like the ellipse itself as seen in 
 objects. The student should compare his drawing of ellipses with 
 the rhythmically varying curves which compose ellipses as seen in 
 real objects, correcting and comparing till the eye is satisfied. If 
 this be faithfully done, the time will be short before 
 ellipses, often deemed a bugbear of freehand draw- 
 ing, become a pleasure instead of a penance. 
 
 Roundness of Ellipses According to Position. — Since 
 the top ellipse appears rounder as it is dropped below 
 the eye level, it must be concluded that could the 
 bottom be fully seen it would appear as a rounder 
 ellipse than that of the top. Place the cylinder on the table and 
 trace around the bottom with a pencil. Move the cylinder to one 
 
 9 
 
 Fig. 16
 
 FREEHAND PERSPECTIVE 
 
 side and compare the shape of this traced ellipse with that of the 
 top ellipse (Fig. 16). Also compare both with that part of the 
 cylinder bottom which can be seen. There is no difficulty in 
 perceiving that the ellipses in a vertical cylinder below the eye 
 are rounder as they are farther below the eye level. 
 
 Now, keeping the cylinder vertical, raise it slowly. When the 
 bottom ellipse reaches the level of the eye, it appears as a straight 
 line (A in Fig. 17), like the top ellipse when at the 
 same height. When the cylinder is moved on 
 above the eye, the bottom becomes an ellipse (B), 
 which as we raise it farther above the eye level 
 appears rounder. We perceive that it appears 
 rounder or less foreshortened in proportion as it is 
 farther from coinciding with the direction in which 
 the eye looks to see it, as was the case with the 
 book cover in Chapter II. Furthermore, if the 
 cylinder be turned horizontally and held at the level 
 of the eye with its length parallel to the picture 
 plane, and one end be brought in front of the eye, 
 we shall again see this circular end as a straight line (B in Fig. 18), 
 because it coincides with the direction of seeing. If the cylinder 
 be moved horizontally to one side, still keeping its length parallel 
 with the picture plane (A in Fig. 18), 
 the ellipse appears to widen exactly 
 as when the cylinder was held verti- 
 cally and moved above or below the 
 eye level. The circular top appears 
 as a circle only when its surface is 
 at right angles to the direction of 
 seeing (A, Fig. 14). When oblique to this direction, as at B, it 
 appears as an ellipse, or foreshortened circle. The ellipse is 
 plainly, therefore, an illustration of the second great principle, 
 that of Foreshortening. 
 
 Practice of Ellipses. — The student should now practice drawing 
 ellipses, both vertical and horizontal, until they can be formed 
 
 10 
 
 Fig. 18
 
 THE ELLIPSE 
 
 with ease and exactness. Mark the extreme points (A, A, B, B 
 Fig. 15) first taking care to have B B exactly opposite the middle 
 of A A. Hold the pencil for drawing ellipses as directed m 
 Chapter I for straight lines, using a position of the hand that 
 will brmg the pencil at right angles to the long diameter. If the 
 elhpse is horizontal, begin it a little to the left of the middle of 
 the upper side, drawing to the right first. If vertical, begin below 
 the middle of the left side, and draw up. Make the whole outline 
 with one movement, first carrying the pencil evenly several times 
 over the paper without touching it, to gain confidence and cer- 
 tainty of movement. 
 
 11
 
 Chapter IV 
 A CYLINDER AND A CYLINDRICAL OBJECT 
 
 THE student should draw this exercise, following carefully 
 the directions given. After doing so he should draw a 
 cylindrical object of his own choosing, putting in practice 
 the principles taught in this chapter. 
 
 KVSSMSafi^-^VAC*. 
 
 Fig. 19 
 
 Planning the Drawing. — The Design, or Composition, or Decora- 
 tive Arrangement of the exercise, which is that kind of beauty 
 secured by a harmonious and artistic relating of the work and its 
 spaces, is to be considered first in all drawings, and should always 
 be kept in mind. For this exercise (Fig. 19) we first consider 
 how to place most effectively in a drawing these two separated 
 
 12
 
 CYLINDER AND CYLINDRICAL OBJECT 
 
 objects, a cylinder whose height is twice its width and some 
 simple cylindrical object (in this case a rose jar). To this end 
 after drawing the margin the extreme points in the boundaries 
 of the objects are lightly indicated on the paper (Fig. 20), taking 
 care that the spaces between them and the margin are such as to 
 
 give an agreeable and interesting division of the 
 
 inclosed surface. The continuous table line be- 
 j hind them indicates that they stand on the 
 
 same surface, and thus links them together. 
 
 The size of the space between them, being no 
 
 ^°" ^^ more than that between either and the side 
 
 margin, also helps unite them; and the position of the ornament 
 
 on the jar, near the middle of the sheet, attracts the eye to the 
 
 center in comparison with the whole. 
 
 Drawing the Cylinder. — For this the paper cylinder model used 
 in Chapter III is placed as shown in the illustration (Fig. 19), 
 that is, a little below the eye level, and at least six times its 
 height from the eye. The apparent proportion of the top (that 
 is, if the width of the ellipse appears to be one third, one fourth 
 or some other part of its length) should be carefully judged by 
 the eye and then tested by pencil measurement. Four points for 
 this ellipse should then be lightly marked, and it should be drawn 
 through these points as previously directed. The bottom ellipse 
 is sketched directly under the upper and in the same way, re- 
 membering that it must be of the same length, but rounder. The 
 back or invisible part of each ellipse is left light. The straight 
 lines for the sides must be tangential to the ends of the ellipses, 
 so they will join with perfect smoothness. If they do not thus 
 join, the ellipse is the part most likely to be wrong. 
 
 Now is the time to put the drawing back by the paper model 
 and compare the two. Look longest at the model, glancing briefly 
 at the drawing ; the aim being always to form in the mind a clear 
 image of the model's true shape, and to correct the work by it. 
 The student should ask himself if the cylinder in his drawing 
 appears to press evenly on the ground like the model. The com- 
 
 13
 
 FREEHAND PERSPECTIVE 
 
 monest error is that of bending the outline of the partially visible 
 ellipse too much at A in Fig. 21, and not increasing the curvature 
 toward its ends (B, B), thus making the curve 
 more circular than elliptical, and causing the cyl- 
 inder in his drawing to look as if it would rock 
 on its base, instead of resting firmly on every 
 part of it. Sketching the ellipse entire (C, C) i* 
 
 Corr^c//<ys ^^ ^^^ ^^ SVLGII Si CaSC. 
 
 ^^^' ^^ Next a pencil line is drawn around the paper 
 
 cylinder half way between the top and bottom. Points (E, E, 
 Fig. 21) are then marked on the drawing, one half way between 
 the fronts, and the other half way between the backs of the 
 ellipses. These points should be tested after marking and made 
 correct, but never measured till the eye has been made to do its 
 utmost. Observe that these marks give a short diameter for the 
 middle ellipse half way in size between those of the upper and 
 lower ellipses. 
 
 In the same way the two other lines around the cylinder may 
 be made on the model and represented in the drawing. 
 
 Now another paper cylinder, of the same length as the first 
 but only two thirds its diameter, must be made, and placed within 
 the first. Have the space between them even all the way around, 
 so that the circular tops of the two cylinders form concentric cir- 
 cles. They appear as ellipses. Observing carefully the space be- 
 tween these ellipses, the student easily sees that it appears widest 
 between the ends, and a little wider between the front sides than 
 between the back sides. As we 
 made the inner cylinder two thirds 
 of the diameter of the outer, the 
 horizontal space between the' ends 
 
 Fig ^^ 
 
 of the two ellipses will each be act- 
 ually made one sixth of the length of the outer ellipse. They will 
 also appear as sixths, because the ends of the ellipse are equally 
 distant from the eye. The ends of the inner ellipse (C, C) are 
 marked by light vertical lines. For its front and back we divide 
 
 14
 
 CYLINDER AND CYLINDRICAL OBJECT 
 
 the width of the outer one also into sixths, but as these sixths 
 are in perspective or at varying distances from the eye, they are 
 perspective sixths. That is, they appear successively smaller as they 
 recede from the eye. This perspective division is here made wholly 
 by the eye (though later another method is given). The per- 
 spective middle (point G) is first marked on the short diameter, 
 making the near half considerably larger than the far one. Each 
 perspective half is then divided into perspective thirds, after 
 ^ which the six divisions are tested to see if they are successively 
 smaller, as directed above. Draw the inner ellipse, making its 
 ends tangential to the vertical lines (C, C), and exactly opposite 
 the middle of its short diameter. It will now be found, if we 
 draw the long diameter of each of these two ellipses where it 
 must always be, in the apparent middle from front 
 to hack, that the long diameter of the inner one 
 falls higher on the paper than that of the outer 
 one. From this we conclude that neither long 
 diameter represents the actual diameter of the 
 circle. Fig. 23 shows the plan of a circle with 
 its true diameter, A A. The eye at x sees B B, a 
 line connecting two tangentials, as the longest line 
 in that circle. It therefore becomes the long diam- 
 eter of the ellipse which the eye in that position 
 sees. Meanwhile the actual diameter appears both 
 shorter and farther back than B B, because far- 
 ther away. That part of the circumference back ^^^- -^ 
 of B B, though actually larger, is so foreshortened as to appear 
 exactly like the part in front, producing the symmetry which is 
 the wonderful and unfailing characteristic of the ellipse. 
 
 Since the inner circle is smaller, the eye can see farther round 
 it, as shown in Fig. 23. This furnishes another reason for its long 
 diameter falling farther back, and agrees with the fact that the 
 really even space between the two circles appears greatest in front. 
 
 The Rose Jar. — For the second object proceed as with the 
 cylinder, drawing lightly all of the ellipses entire first. Should 
 
 15
 
 FREEHAND PERSPECTIVE 
 
 Fig. 2J. 
 
 any fall at the same height as one on the cylmder, it must be 
 made of the same roundness, since the two objects are shown 
 by the table line to be on the same surface, and are equally near 
 
 the eye. Compare the size of the ellipses 
 with the extreme width of the space occu- 
 pied on the paper by the jar. Compare 
 also the lengths of the top and bottom 
 ellipses, and the length of each with the 
 extreme width of the jar. Observe that 
 the sides of its short cylindrical neck slope 
 outward slightly toward the body of the jar. 
 The Shoulders and Base of the Jar. — Before 
 drawing the side outlines, hold the jar verti- 
 cally at arm's length, and with the top on a 
 level with the eye. Mark the point (A in 
 Fig. 2-i) where the body and neck boundaries meet. Holding this 
 point, lower the object till the top appears as a fairly round ellipse. 
 It will be plain that we now see a portion of surface beyond the end 
 of the ellipse, and more than half way over its shoulder. The 
 boundary line which marks the limit of our seeing has moved back 
 on the shoulder, so that it passes out of 
 sight behind the neck. A little experi- 
 menting shows that the surface visible be- 
 yond the end of the ellipse is in exact yl 
 proportion to the roundness of the ellipse. 
 Now place a sheet of paper on the 
 table, and first holding the jar so its base 
 is on a level with the eye (Fig. 25), mark 
 the extreme point of the base, B. Lower 
 the jar slowly, till the bottom rests on the 
 paper. Mark point B on the paper and 
 then the points (C, C) where the side 
 boundaries now appear to meet the bottom ellipse. Trace 
 around the bottom and lift aside the object. It will be found 
 that the projecting mass of the jar, being nearer the eye than 
 16
 
 CYLINDER AND CYLINDRICAL OBJECT 
 
 the ellipse, had hidden from sight more than half of it. It is 
 also evident that on the lower and receding part of the jar the 
 boundary line advances, so that we see less than half of the sur- 
 face, instead of more, as at the top. If we now trace on the jar 
 its boundary line and turn it around, the tracing will be seen to 
 cross the object obliquely (Fig. 26). With these facts 
 in mind, complete the drawing of the rose jar. 
 
 Tangential Joinings. — All meetings of boundary lines 
 with ellipses must he tangential. That is, they must 
 touch so that if smoothly continued they would not 
 cut the ellipse. 
 Artistic Rendering. — The jar should be first drawn with thin 
 light lines, corrected to accuracy, and afterward rendered with sig- 
 nificance. For though outlines are entirely conventional, never 
 being seen in nature, yet they may not only be made to inarl^: off 
 beautiful and interesting shapes, but by their character to sug- 
 gest other truths and qualities for the enhancement of charm. 
 Thus in the rose jar the front edge is shown to be nearer than 
 the back by the heavier line, and the rounded thickness of the 
 top is indicated by the absence of nearly all of the inner ellipse 
 at the back and of the outer one at the front. The sides are 
 drawn with a little lighter lines at the top, and though firm 
 enough to clearly present the shape of the jar, are lighter than 
 the front part of the top, because representing a part of the jar 
 further from the eye. 
 
 The ornament may help to express the rounded form of the 
 jar by its foreshortened shape as it nears the boundary, and by 
 the greater clearness and emphasis of that portion of it most 
 thrust forward. Its outlines, emphasized on one side, with the 
 other side light or lost, and the detail shown in the side lines of 
 the jar, indicate it to be in relief. The expression of color in 
 places is used to strengthen the projection of the jar. 
 
 These remarks, however, must not be understood as rules. 
 They are but suggestions for the incitement of the student to 
 
 use his own artistic judgment. 
 2 17
 
 Chapter V 
 
 AN OBJECT ABOVE THE EYE AND THE 
 CONE PRINCIPLE 
 
 T 
 
 HE electric lamp shade here shown (Fig. 27) is vertical 
 and above the eye, and its ellipses therefore increase in 
 roundness toward its top. The student should draw this 
 
 exercise, and make 
 another study from 
 some, object similarly 
 placed. 
 
 Dra-wing the Object. 
 I — Proceed as in the 
 \ previous exercise. 
 \ Observe that the slop- 
 \ ing side boundary 
 I lines of the shade join 
 I the ellipses in front 
 \ of their ends. The 
 \ flaring shade is like 
 I the lower half of the 
 rose jar reversed. Its 
 smaller part is far- 
 thest from the level 
 of the eye, as was the 
 base of the rose jar, 
 and we therefore do 
 not see half way 
 round it. To make 
 this clear, hold a cone^ 
 in various positions 
 
 Fig. 27 
 
 ^ One may be made from paper (Fig. 28). 
 18
 
 OBJECT ABOVE THE EYE, ETC. 
 
 Fig. 28 
 
 (that is, above and below the eye), and with apex up and down 
 (Fig. 29). 
 
 The button through which the cord is drawn forms an 
 oblique ellipse. But by turning Fig. 27 so as to bring this 
 ellipse horizontal, the button will be found symmetrical 
 on its axis (see Ch. XXVII); and as it is arched or 
 thickened in its middle, the space in front of the holes 
 for the cord appears a good deal wider 
 than that back of them. Each side of 
 the button is a very flat modified cone. 
 Notice in the outline of the cylinder the 
 slight depression marking where the key 
 enters its side. 
 
 It may be now noted that ivhen the cone apex 
 (that is, its decrease of diameter) is nearer- the 
 eye than its base, we see more than half way round 
 it. Conversely, we see less than half ivay round 
 tvhen the apex is farther from the eye^. This char- 
 acteristic of curved or sloping surfaces in cylindrical objects 
 may be termed the Cone Principle. Broadly speaking, it is the 
 expression^ hy outline merely, of Belief, or Solidity, or TJiird 
 Dimension. The rose jar in Chapter IV, the cream jug^in Chapter 
 VI, and indeed all cylindrical objects with flaring or bulging 
 sides, are examples. More advanced applications of this principle 
 are found in the drawing of such natural objects as trees and 
 mountains, also in drawing the human face and figure. 
 
 Fig. 29 
 
 19
 
 T 
 
 Chapter VI 
 A CREAM JUG 
 
 HE Model. — Provide an object similar to the cream pitcher 
 
 here shown (Fig 
 should be made. 
 
 30), from which the student's drawing 
 If inexperienced he will be helped by 
 
 „____ first making a copy 
 
 , from this example. 
 I The Handle. — Place 
 the jug a little below 
 the eye (according to 
 the directions in Ch. 
 I). Draw the cylindri- 
 cal body entire first as 
 if it had neither handle 
 nor spout, and with 
 light lines (Fig. 31). 
 Then hold the model 
 with its center at the 
 level of the eye and 
 with the handle in 
 profile (A in Fig. 32). 
 Observe that a center 
 line for the joining of 
 the handle with, the 
 model would fall in 
 the curved boundary 
 line or profile of the 
 jug. Turn it to bring 
 the handle directly in 
 front, when this same center line {x in Fig. 32) will appear 
 straight and vertical. Now, turn the jug slowly back, bringing 
 
 20 
 
 t — 
 
 Fig. 30
 
 A CREAM JUG 
 
 Fig. 31 
 
 Fig. 32 
 
 the handle again into the boundary line. It is apparent that as 
 the handle revolves, its center line of joining changes in appear- 
 ance from a straight line in front through a succession 
 of curves that increase in roundness till at last it coin- 
 cides again with the profile of the jug (C in Fig. 32). 
 These curves are lines such as would be produced on 
 the surface by cutting vertically through the jug cen- 
 
 ^ ter, as an apple is 
 
 halved ; and may be named 
 Profile Lines or Profiles. 
 
 Replace the model on the 
 table and revolve the handle 
 to the side again, when it 
 will be seen that the ends 
 of these profile curves rest on the top and bottom ellipses 
 of the jug (Fig. 33). And the side boundary of 
 the jug does not now coincide with the profile 
 at the side, as it did (A in Fig. 32) when the jug 
 was held at the eye level. This is because of 
 the change in the position of the boundary. 
 As the jug is placed below the eye the bound- 
 ary advances from A to B, recedes from B to 
 C, and advances again from C to D, in accord- 
 ance with the cone principle (Ch. V). 
 In Fig. 34 is shown by a dotted tracing of this 
 boundary how it actually differs from 
 the profile curves in Fig. 33. In 
 sketching a profile curve, therefore, allowance must be 
 made, as shown in Fig. 33. Note how x, x, x^ the 
 points of the smallest diameter, fall in an ellipse at that 
 height ; also the points ?/, y, y, of the greatest diameter. 
 Fig. 34 q^^^ ^j ij^^g^ profiU curvcs, shapcd according to its 
 nearness to the boundary of the object, should he sketched as a 
 guide for the attachment of the handle. 
 
 In the same way we observe the shape of the handle itself to vary 
 
 21 
 
 Fig. 33
 
 FREEHAND PEKSPECTIVE 
 
 Fig. 35 
 
 Fig. 36 
 
 according to position, from a profile vieiv at the side (A in Fig. 32) 
 to a view of its outer surface (B in Fig. 32). 
 
 The Spout. — Looking directly into the jug from above (Fig. 
 35), we note that the spout is directly opposite the handle, so that 
 a horizontal line through the middle of both would 
 pass through the center of the circular top. We there- 
 fore mark the perspective middle of the top ellipse (0 
 in Fig. 36) (that is, making the nearest half larger), 
 draw a line through it from the center of the handle top, 
 and mark the end of the spout on this line. For the 
 width of the spout, set off perspective distances from this line 
 either way on the top edge of the pitcher, remembering that the 
 half nearest the end of the ellipse is much 
 the more foreshortened; and that the dif- 
 ference is greater the more the top is fore- 
 shortened. From these points to the tip 
 of the spout straight lines may be sketched 
 as guides for drawing the edges, which 
 
 may be straight but usually curve both upward 
 and sidewise. The profile or center line for the 
 spout is sketched like that for the handle (A in 
 Fig. 36). 
 
 The Foot. — The drawing of the foot also needs 
 some explanation, though covered by the cone 
 principle of Chapter V. In profile it would ap- 
 pear as at A in Fig. 37, with the circles as straight 
 lines; and the student should raise his model to 
 the eye level and observe it thus. On lowering 
 the model these seeming straight lines appear as ellipses (B, 
 Fig. 37), and the lower part of the side boundary lines of both the 
 pitcher and its foot move forward of the ends of these ellipses 
 till tangential joinings are made at C and D. The upper part 
 of the boundary of the foot moves back, joining the upper ellipse 
 at E. In consequence, this side outline of .the foot (E D)is a 
 little lengthened, making its curve less round than in profile. 
 
 22 
 
 Fig. 37
 
 A CREAM JUG 
 
 the other with the 
 A 
 
 The lower half of the jug, as we can now see, is a modified ver- 
 tical cone with its apex down (A, Fig. 38). The foot is modified 
 from two cones; one with the apex up 
 apex down (B, Fig. 38). See Chapters 
 ly and V. 
 
 The Ornament. — The principle fol- 
 lowed in suggesting the perspective of 
 the ornament will be readily seen from 
 
 the illustration (Fig. 39). The 
 
 curved guide lines are parts 
 
 ti\f !<• ! of profiles similar to those for 
 h'^^i placing the handle and spout. 
 
 Fig. 38 
 
 The student will now begin to understand that it is 
 possible to recognize and suggest the solid rounding 
 surface of the object by every line and touch upon, it. 
 To this end that part of the ornament nearest the 
 eye is more emphasized in the final drawing. And 
 looking carefully at the object, we see that besides 
 its foreshortening, that part of the ornament near 
 the boundary is less distinct, and is often lost in the reflec- 
 tions from its surroundings. 
 
 \^'f 
 
 K 
 
 m 
 
 Fig. 39 
 
 23
 
 Chapter VII 
 A TIME STUDY 
 
 HAVING- carefully studied the principles of cylindrical 
 objects, it is now best to take a specified time, as fifteen 
 minutes, for the more free drawing of such an object, 
 choosing a simple one at first. Proceed as before, except that 
 most of the measuring and testing must be omitted. This leaves 
 
 Fig. 40. A Time Study. 
 
 time to draw slowly and thoughtfully, making the unaided eye do 
 all that is possible. Study the general shape, looking long at the 
 object, and moving the pencil several times, without marking 
 over the paper where the lines are to be drawn to acquire confi- 
 dence and certainty of touch. Require yourself to work with 
 
 24
 
 A TIME STUDY 
 
 no erasing (except of such construction lines as may show when 
 the drawing is done), and to stop when the time is up. It will 
 be found a valuable exercise to draw in this way the same object 
 several times. After one drawing is done, carefully examine 
 and test it to find the errors, but do not correct them on that 
 drawing. Instead, make those points right in your next attempt 
 at the same object. 
 
 Observe in Fig. 39 how the effect of glass is given by a few 
 lines selected from those many graceful curves of delicate dark 
 and light which appear in the object; also the sketching of its 
 high lights, or window reflections, and the wavy distortion of 
 lines seen through it. The straight lines give a firmness to the 
 composition which is needed, since the bowl consists wholly of 
 curved lines. 
 
 2.5
 
 Chapter VIII 
 A GROUP OF CYLINDRICAL OBJECTS 
 
 FIG. 44 is an exercise in the grouping of cylindrical objects 
 agreeably and appropriately together. The student is 
 advised to first draw this example, using carefully the 
 explanations given. After that, he should arrange and draw 
 another group of two cylindrical objects. 
 
 Making the Composition. — In composing this second group, 
 experiments should be made with a number of objects, combin- 
 ing them in different ways. A Finder, which is a 
 card having a small rectangular opening cut in it 
 (Fig. 41), will greatly assist in judging the pictorial 
 
 Fig. 41 effcct of a compositiou, especially in a rectangular 
 margin. The student should look through it at his arrangement 
 with one eye, letting its edge take the place of a margin, and 
 moving it back and forth till the 
 place is found where it makes the 
 group look best. Little trial or 
 *' thumb-nail" sketches (Figs. 42 
 and 43) should also be made to 
 determine the best arrangement. 
 
 In Fig. 44, for example, we 
 observe that the objects are such as might naturally be placed 
 together, and are placed in positions that are not unusual. Next 
 their shapes make a pleasant relief or contrast to each other with- 
 out harsh or awkward opposition ; one being tall and slender and 
 the other lower and round. Yet the teapot is not so low nor 
 wide but that it echoes in some degree the dominant height of 
 the candlestick, thus aiding harmony. Its spout is allowed to 
 
 26
 
 A GROUP OF CYLINDRICAL OBJECTS 
 
 project across the candlestick, thus contributing to the unity of 
 the composition. The leaning bowl, by passing behind both, 
 also strengthens 
 unity, and by its 
 lighter and more 
 interrupted lines 
 furnishes a tran- 
 sition, or connec- 
 tion, between the 
 nearer objects 
 and the white 
 paper. These 
 results might be 
 secured by other 
 groupings. But 
 had the candle- 
 stick been in 
 front, for in- 
 stance, its pro- 
 jection above and 
 below the teapot 
 would have been 
 so nearly equal as 
 to seem uninter- 
 esting (Fig. 42). 
 Yet we could 
 have remedied 
 this somewhat by 
 placing the can- 
 dlestick a greater 
 
 distance in ad- Fig. 44 
 
 vance, or raising 
 
 the teapot handle. Or the cover could have been placed on the 
 ground in front (Fig. 43) — indeed, many possibilities will be 
 suggested by a little study. 
 
 27
 
 FREEHAND PERSPECTIVE 
 
 Drawing the Group. — In drawing this exercise, observe that 
 though the bottom ellipses of the two objects are at the same 
 level the nearer appears slightly rounder (Solutions of Problems, 
 p. 162). Care must be taken in placing these ellipses to allow 
 for the bulging of the teapot sides. Remember that in propor- 
 tion as the bottoms are drawn foreshortened so must all spaces 
 on the table be regarded as foreshortened. Note also that all 
 ellipses in the candlestick which are nearer the eye level than the 
 top of the teapot will be less round than those of the teapot. 
 
 The Teapot Ears. — In placing these, an ellipse may be used as 
 
 a guide (A in Fig. 45) . The middle points of the two ears should 
 
 A ^--:5^"~~-v, ^® ^^ ^ ^^^^^ passing through the perspective (that is, 
 
 ^V^^^S^w actual) center (o) of this ellipse, as were the handle 
 
 'y^^-- — ^^;\>1 and nose of the pitcher in Chapter VI. The 
 
 J3 ^,j'^2_^^ cover is arched (B in Fig. 45), so that it conceals 
 J— — i. the back of its elliptical edge. This arched shape 
 
 PROFILE VIEW .,.., • ^ o o • 
 
 6How.NGARCHOFcovtR jg distuictly socu 111 tiiG form of its top boundary 
 
 and is a very different shape from the ellipse. 
 
 But this arched boundary does not fall in the actual middle of 
 
 the cover (since we are looking down on it), but a little beyond 
 
 that. The knob is in the actual middle. 
 
 The rendering of two things is more complicated and inter- 
 esting than that of one alone. As the candlestick is farther away 
 than the teapot, its lines are made lighter, and in places are quite 
 lost. The lines of the glass rim, or hoheche are thinner, more 
 interrupted and more smoothly sweeping. The leaning bowl 
 may be omitted at this time, if found difficult. The principles 
 of its construction are given later (Ch. XXVII). If omitted it 
 will be found necessary to make the farther lines of the candle- 
 stick lighter yet, to serve in place of the bowl as a transition. 
 
 28
 
 Chapter IX 
 
 CYLINDRICAL OBJECTS GROUPED 
 WITH FRUIT 
 
 A N example of grouping is here given in which part of the 
 /\ group is cut by the margin, while the apples illustrate 
 -A- ^ the combination of natural forms with cylindrical ob- 
 jects. As in the preceding exercise, the student should compose 
 
 Fig. 46 
 
 a group corresponding to this exercise and draw it ; and if inex- 
 perienced should draw this before making his original one. 
 
 Study of the Group. — In locating the pitcher on the paper, see 
 that its base is far enough from the dish for the two objects to clear 
 
 29
 
 FREEHAND PERSPECTIVE 
 
 each other. Observe the generally elliptical shape of the curves 
 in the glass pitcher ; and how the edge of the plate is seen warped 
 and interrupted through it. The plate is made subordinate, as 
 
 forming part of the background for 
 the other two objects. Its position, 
 appearing in its actual shape as a 
 simple circle, contributes to the de- 
 sired effect of quietness or subor- 
 dination, as does its being .cut by 
 the margin line, and its lighter and 
 
 \ 
 
 \ 
 
 
 PESWtCTIVE. 
 
 X-Y roRESHORTENED. 
 
 Fig. 47 slightly interrupted lines. 
 
 Since it is standing vertically, it must be supported by a 
 vertical surface behind it. Consequently the table line (Ch. I) 
 must be placed only far enough on the paper above the lowest 
 point of the plate edge to express the foreshortened necessary 
 distance of this point from the wall behind it (Fig. 47). 
 
 30
 
 Chapter X 
 
 A GROUP OF OBJECTS FROM MEMORY 
 OR INVENTION 
 
 THIS example illustrates the drawing of objects from 
 invention or memory. The student may sketch this 
 exercise as directed; then should invent or draw from 
 memory one of his own arrangement, making small trial 
 sketches as in Chapter YIII, and using the best of these in his 
 
 :-;;>:>^,---fC==CXAM\\ 
 
 _ -..J 
 
 Fig. 48 
 
 final composition. Should his memory not be clear enough 
 for this, it may be refreshed as often as necessary by study 
 of the objects he chooses to draw, the only condition being 
 that the drawing he done ivithout the object in view. 
 
 31
 
 FREEHAND PERSPECTIVE 
 
 Drawing the Above Study. — In this exercise the Japanese 
 luncheon carrier is placed first. Its ellipses are sketched in 
 full, whether entirely seen or not. The bowl-shaped top, being 
 slightly inclined, is drawn on a leaning axis (A B in Fig. 49). 
 But it is perfectly symmetrical on this axis (Ch. XXVII) . This 
 symmetry should be tested in the drawing by turning it to 
 bring the axis vertical, when any error is easily detected. (Ch. 
 XXVIII.) 
 
 The Flat Dish. — It was desired to draw the flat dish as it 
 would appear if touching the luncheon carrier. Its height (x y) 
 is therefore measured upon the front of that object from its lower 
 
 edge, and an ellipse of 
 the proper roundness 
 drawn at that height. 
 The top ellipse of the 
 dish would touch the 
 other object somewhere 
 in this ellipse, and so 
 was drawn tangential 
 ?i to it. To obtain the 
 bottom ellipse of the 
 dish, this same height, 
 increased to allow for 
 its slightly greater near- 
 ness to the eye, was measured downward from the dish top. 
 But as the sides of the dish are flaring, this measuring was done 
 from the estimated true middle (0 in Fig. 49) of the top of the 
 dish, giving O' for the true center of the lower ellipse. The foot 
 is like a very short cylinder. The flaring sides of the dish are 
 drawn tangentially from the rim (F, F) to the upper ellipse of 
 the foot. 
 
 The Ornament. — In drawing the ornament on the luncheon 
 carrier the explanation in Chapter VI is recalled. On the cover 
 the band of fret decoration appears narrowed at its front, and 
 widest at the ends. It is a modification of the cylinder top in 
 
 32 
 
 Fig. 49
 
 OBJECTS FROM MEMORY 
 
 Chapter IV. Note the foreshortening in its details, and how the 
 Hnes of the fret express the curving form of the cover. It will 
 be seen that the stripes on the object and some lines of the fret 
 follow the profile lines mentioned in Chapter VI. 
 
 The Fan. — Like the plate in Chapter IX, the fan is purposely 
 placed so that it is not foreshortened. Therefore the two points 
 (G, Gr) at which it rests on the table appear, as they actually are, 
 in a horizontal line. It also appears in its true shape, symmetrical 
 on an axis passing through its handle (H, H). It is more easily 
 drawn entire first, erasing later the part not needed. 
 
 "^ 
 
 33
 
 Chapter XI 
 
 THE CYLINDER CONE AND BALL GROUPED 
 —A PROBLEM FOR ORIGINAL STUDY 
 
 GENERAL Conditions for Perspective Problems. — Problems 
 are to the student both a test of his comprehension of 
 the subject thus far, and an exercise by which the subject 
 becomes firmly fixed in his mind. To this end the drawings 
 must be made tvitliout the models in sight, though they should be 
 studied, and if necessary even sketched in the required positions 
 before drawing. If the student is at loss to recall 
 their appearance while engaged in work they may 
 be studied as often as needed ; provided only that 
 neither the models nor sketches of them are be- 
 fore the student as the drawing is made. It can- 
 ' ncTUBc'^ Pu^^4E )' not be expected that any object should be drawn 
 ' ' ' until opportunity has been given for its thorough 
 
 ^^^- ^0 study ; but on the other hand it is not mastered 
 
 until it can be correctly drawn from unaided knowledge and 
 memory. The stated dimensions are important, giving training 
 in the expression of proportion, though drawings need not be 
 full size. 
 
 Drawings should of course be made without assistance, and 
 without referring to the explanations in the back of this 
 book. When the student has under the required conditions 
 made his drawing, he may then test his work by consulting 
 the explanation. 
 
 Conditions of this Problem. — In this problem the cylinder and 
 cone are to be 4" in diameter by 8" high, and the ball 4'' in 
 
 34
 
 CYLINDER CONE AND BALL 
 
 diameter. The group is to be drawn as if resting on a surface 
 which is twice the cylinder height below the eye, and at least six 
 times its height distant. The cylinder stands on one end and the 
 cone on its base, touching the cylinder and a little in front of it 
 at one side. The ball also touches the cylinder, and is a little 
 more in front of it on the other side. The plan (Fig. 50) will 
 make this clearer. 
 
 S5
 
 Chapter XII 
 
 THE STUDY OF STRAIGHT-LINE 
 OBJECTS 
 
 A Book with Back Parallel with the Face 
 
 FOR this study provide a book, two long pencils, and three 
 yards of fine twine, also paper for sketching. Choose 
 a book of interesting appearance; a somewhat worn, 
 leather-bound book is best. Place it well back on the table in 
 
 front of you and below the 
 eye with its back next to 
 and parallel with the pic- 
 ture plane, and its ends 
 equally distant from you 
 (Fig. 51). 
 
 The Book Below the Eye. 
 
 — Two surfaces are visible,, 
 the back and one cover. 
 Count the edges seen (seven), 
 then decide how many of 
 these are actually horizontal.^ 
 If the 
 
 Fig. 51 
 
 book is 
 placed as directed, its back, being parallel with 
 the picture plane, will be seen in its true shape 
 if traced upon it. Lifting the cover till it 
 is vertical (Fig. 52), we see that the cover 
 also now appears in its actual form 
 
 ^ It may not at first be realized that the ends of the cover are horizontal, as M'ell as its 
 sides. But as they are contained in a horizontal surface (in this case the cover), they also 
 must be horizontal. Their perspective appearance must be distinguished from their actual 
 position. 
 
 36 
 
 Fig. 52 
 
 But as we drop it slowly
 
 STRAIGHT-LINE OBJECTS 
 
 back till horizontal, we observe that the further edge seems to 
 grow shorter because moving from the eye, and that the whole 
 cover becomes foreshortened or narrowed from front to back, 
 
 like the circular ends of the cylinder in 
 Chapter IV. If (as with the house in 
 Ch. II) , a pane of glass were standing in place 
 of the imaginary picture plane, a tracing of 
 the cover on that would be a true perspec- 
 How to draw on the paper such a 
 Stand the pen- 
 
 FiG. 53 
 
 V tive of it. 
 perspective is our problem. 
 
 Fig. 54 
 
 cils against the nearest corners of the cover 
 (Fig. 53) ; then closing one eye, and keeping the other exactly 
 opposite the middle of the book, incline the 
 pencils toward each other (being careful 
 not to lean them back or forward) until 
 they appear to lie just along the retreat- 
 ing ends of the cover (Fig. 54). Let 
 another person hold a ruler against the 
 pencils, moving it down until its edge seems 
 to coincide with the further edge of the 
 
 cover (Fig. 55). Now the pencils and the 
 ruler together picture the apparent shape of 
 the cover, and we plainly see how the ap- 
 parent shortening of the back edge (caused by 
 its greater distance from us) makes the ends 
 appear to converge toivard each other. The 
 question now is : Can the law of that conver- 
 gence be so determined that it may be applied in any drawing! 
 
 The Converging Book Ends. — Substitute for the pencils the 
 string slipped under the cover to the back, and using one eye as 
 before, bring the ends together so that the strings will appear to 
 exactly coincide with the ends of the cover as did the pencils. 
 (Be sure to keep the string vertically over the front edge of the 
 book, not letting it fall back or forward.) The pencil may now 
 
 37 
 
 Fig. 55 
 
 4^zm
 
 FREEHAND PERSPECTIVE 
 
 be taken in the other hand, and slipped down on the string, to 
 form again the shape of the foreshortened cover (A in Fig. 59). 
 
 Still holding the string as before, raise the book and string 
 a few inches, keeping the book level and the string taut (A in 
 
 Fig. 56 
 
 Fig. 56). The string does not now cover the book ends, and 
 the joining must be brought lower (as in B) that it may do so. 
 If the book is raised more, the joining is yet 
 nearer to the book, as in C; until when the 
 book cover is at the level of the eye (Fig. 57) 
 the string and the book cover both disappear 
 in, or coincide with, the upper edge of the book. 
 Now, starting with the position last shown, 
 (Fig. 57) hold the thumb and finger firmly at 
 that place on the eye level (this can be done 
 by noting a point behind it on the wall), and 
 let the book drop slowly. Keep it exactly 
 horizontal, and let the string slip through the 
 stationary thumb and finger, so that their meet- 
 ing point remains at the eye level. If this is carefully done, it 
 will be seen that as the book descends, the string continues 
 to cover the converging ends (as in C and B, Fig. 56). At the 
 same time the cover appears to grow wider, and its ends more 
 and more nearly vertical. 
 
 38 
 
 Fig. 57
 
 STRAIGHT-LINE OBJECTS 
 
 These experiments should also be tried with the lower 
 cover, holding the book above the eye (Fig. 58), and raising and 
 lowering it. 
 
 From the foregoing study it is easily perceived that, provided 
 we Jceej) the book horizontal, the point toward ivhich its ends appear to 
 converge remains always at the level of the eye. 
 We have therefore only to sketch the eye level 
 at its right height compared with some measure- 
 ment on the object and mark the point of con- 
 vergence in the right place on it, to be able to 
 use it for drawing these converging lines. 
 
 We have also found that the horizontal booh 
 covers (like the cylinder top in Ch. IV) appear 
 foreshortened according as they approacli the eye 
 level, whether above or below it (Figs. ^6 
 and 58). 
 
 Sketching the Book. — The book may now be replaced as at first. 
 Then, holding the strings, as before, take the pencil as in B, Fig. 
 
 Fig. 58 
 
 59, that the thumb nail may be used as a sliding gauge. With it 
 measure the length of the back of the book on its upper near 
 edge and compare its length with the vertical distance from this 
 edge to where the strings join (C in Fig. 59). (In this case it 
 takes one and one fourth of the book length to reach the joining 
 of the strings.) Now the back of the book may be sketched in, 
 the point of convergence (the joining of the string) marked on 
 
 39
 
 FREEHAND PERSPECTIVE 
 
 Fig. 60 
 
 the paper one and a half book lengths above its middle, and lines 
 
 drawn from the upper corners of the back to this point. On 
 these lines the ends of the cover are to be marked 
 off. The perspective or apparent width of the 
 cover may be found by measuring it with a pen- 
 cil held vertically as in Fig. 60, and comparing 
 this dimension with the length of the book. In 
 this case the apparent width is one fourth of the 
 book length. 
 
 The Level of the Eye. — This will be found of 
 the greatest importance in all drawings. It should 
 
 be carefully marked in the drawing as soon as the position of the 
 
 objects on the paper give a basis for locating it. At first, another 
 
 person may assist (Fig. 61), but a 
 
 little practice will enable the student 
 
 to find it for himself. The top of 
 
 a pencil, held vertically over the 
 
 objects, will appear as a straight 
 
 line when at the height of the eye 
 
 (Fig. 62). Or if any part of the 
 
 study is as high as the eye, the eye 
 
 level will be where any horizontal Fig. ei 
 
 surface or any receding horizontal lines appear as straight lines. 
 
 See Fig. 63. 
 
 Parallel Lines. — By holding one string down on 
 
 the near end of a margin line on the book this line 
 will be seen to converge to the same point with 
 the two ends (Fig. 64). By placing a second book 
 on and parallel to the first, we can show that all 
 lines parallel until the fir^st two converging ones ivill 
 appear to converge ivith them to the same point. An 
 Fig. 62 important deduction from this is that parallel lines 
 
 appear to converge to the same point. 
 
 It is also evident that since the whole book cover is fore- 
 shortened from front to back, the margins will be foreshortened 
 
 40
 
 STRAIGHT-LINE OBJECTS 
 
 ill the same direction. And we find that the side margins are> 
 foreshortened in length, but not in width; while the front and' 
 back margins are foreshortened in width, and 
 the back one more than the front. This fol- 
 lows the principle of the top of the hollow 
 cylinder in Chapter IV. 
 
 The Vanishing Point. — We see that the 
 book ends seem to converge in proportion ^^^- ^^ 
 
 as the back edge of the cover appears shorter. If a second 
 book like this were placed back of, and touching it, its front 
 
 edge would appear of 
 the same length as the 
 back of this, and its back 
 edge shorter, while its 
 ends would converge in 
 a line with those of the 
 first book (Fig. 65). This 
 can be imagined as re- 
 peated infinitely, each 
 book appearing smaller 
 ^^^- 6^ than the one before it, 
 
 and the cover ends all falling in the same con- 
 verging lines, until a point would be all that could 
 rei3resent the last . book. .The 
 row of books might be said to 
 vanish in this point, which is 
 therefore called the Vanishing 
 Point of such lines as converge 
 toward it, as do the ends of 
 the book cover. Vanishing 
 Points, like the Level of the 
 Eye, play a most important part in the study 
 of perspective. 
 A familiar example of vanishing lines, as those which appear 
 to vanish, or converge perspectively, are called, is found in a 
 
 41
 
 FREEHAND PERSPECTIVE 
 
 receding railroad track (Fig. 66). The ties appear shorter 
 as they are successively farther from the eye; and the rails 
 appear and converge, till the whole track, if it could be seen 
 for a long enough distance, might seem to disappear, or vanish 
 in a point. 
 
 
 42
 
 Chapter XIII 
 DRAWING THE BOOK IN TWO POSITIONS 
 
 T 
 
 HE student may copy this example, but in any case 
 should place a book successively in these positions and 
 draw from that; having it high enough or far enough 
 
 from the eye, to see it 
 in a normal position as 
 explained on page 2. 
 He should also make 
 drawings from mem- 
 ory of a book in both 
 positions, expressing 
 them as artistically 
 as possible. 
 
 In the first position 
 on this sheet, the 
 eye level falls off the 
 paper; and may be 
 marked for use on a 
 piece of paper fastened 
 to the drawing (Fig. 
 68). See that the table 
 line is high enough on 
 the paper to clear the 
 lower back corners of 
 the book. 
 
 It will be observed 
 that the back of the 
 book is not quite flat 
 but slightly curved — a modification of the cylinder, 
 be understood by holding the cylinder horizontally 
 
 43 
 
 L 
 
 Fig. 67 
 
 This will 
 (Fig. 69).
 
 Fig. 68 
 
 FREEHAND PERSPECTIVE 
 
 The curve opposite the eye is seen as a straight line, since it 
 coincides with (or lies in a plane passing through) the direction 
 
 of seeing. The farther these lines are from 
 coinciding with this direction (in this case 
 to right and left) the more apparent is their 
 curvature. 
 
 For the second position in this exercise 
 the book is opened and turned around so that 
 its ends are parallel with the picture plane. 
 They may therefore be drawn in their true 
 shape like the back of the book. The sides 
 and all lines parallel with them now vanish 
 to VP ^ (on the eye level directly in front of 
 the student). Note that points A, B, and C, 
 where the book rests on the horizontal table (Fig. 68), are in a 
 straight line that is parallel to the picture 
 plane, and therefore drawn in its true 
 direction, which is horizontal. Observe 
 the projection of the covers beyond the 
 leaves, and that it extends backward at 
 D and E. The thickness of the covers 
 must be recognized, though the wearing off of the edges and 
 corners may obliterate their sharpness. Since the right and left 
 corners of the book are equally distant from the eye care must 
 be taken that the covers are drawn of equal width. The clasps 
 must be long enough to allow of their being fastened when the 
 book is closed. Their ends are in a line converging to VP. 
 
 The table line, being a subordinate element, should be so 
 placed that most of its length is covered. Avoid anything which 
 would tend to emphasize it, as making it coincide with the back 
 corners of the book. 
 
 1 Used as an abbreviation for the vanishing point. 
 
 Fig. 69 
 
 44
 
 Chapter XIV 
 THE BOOK WITH A CYLINDRICAL OBJECT 
 
 THIS exercise (Fig. 70) combines a book in one of the 
 two positions previously studied with a cylindrical ob- 
 ject. The student may draw this example or not, 
 according to his proficiency; but should compose and sketch 
 a similar group, arranging and making trial sketches of several 
 
 Fig. 70 
 
 compositions. Observe that the extreme points of the book 
 must be equidistant from the eye, as in Chapters XII and 
 XIII. But as soon as we place another object with the book, 
 the two must be considered together as forming one group or 
 picture. 
 
 45
 
 FREEHAND PERSPECTIVE 
 
 PUN OF AJ 
 
 9HOWINO 
 
 PLANE 
 
 TO BOOK 
 
 This brings us to reflect that whatever the number of objects 
 we include in our picture, it is always drawn tvith the eye directly 
 opposite the picture as a whole, so that the center of seeing is in 
 the middle of the group from side to side. We also 
 recall that the picture plane is always at right 
 angles (viewed from above) to the direction of see- 
 ing. So, if the cylindrical object is placed on, or 
 in front of the book (as in Figs. 70 and 71), the 
 central direction of seeing the picture is not 
 changed ; and the picture plane continues parallel 
 to one set of lines in the book as in the preceding 
 exercise with the book alone. (See plan, Fig. 71.) 
 If, on the other hand, the cylindrical object is 
 placed at the side and the picture thus enlarged 
 in one direction only (Fig. 72), the direction of 
 seeing is immediately thereby 
 moved to correspond, and the 
 picture plane moves with it. 
 The book will cease to be equi- 
 distant at its ends from the 
 picture plane and cannot be drawn as previ- 
 ously studied. This subject is considered 
 more fully in Chapters XXXIV and XLI. 
 It would of course be possible to add objects 
 to the book equall}^ at both 
 sides (as in Fig. 73), but dan- 
 ger of stiffness in such an 
 arrangement must then be 
 remedied by some such device as the string of 
 beads, making a more complicated study than 
 is desirable at present. 
 For this exercise, therefore, place the cylindrical object some- 
 where within the extreme points from side to side of the 
 book. 
 
 It will be observed that a cylindrical object is always placed so 
 
 46 
 
 Fig. 71 
 
 PlANOrX 
 
 OSJECTSX 
 PLACED \ 
 
 M 
 ^ 
 
 /the SAKIE 
 /with the bOOK 
 /at TKt RiOHT. 
 
 TIltOEMTIM 
 13 MOVED, \ 
 
 J 
 
 /of the PICTUOC 
 /and TMt PICTvnC 
 
 TDTTIB ' 
 
 /loNOSH PARALLtl. 
 /SOOK. 
 
 
 \L 
 
 
 
 Fig. 
 
 72 
 
 Fig. 73
 
 THE BOOK WITH A CYLINDRICAL OBJECT 
 
 that a part of its base is seen, if only a very small part. For 
 this reason, it is not put behind the book unless the foot 
 can be left partly visible. The reason for this precaution is 
 the uncertain effect produced by a study in which it is not 
 observed.
 
 Chapter XV 
 
 THE CYLINDER AND RECTANGULAR BLOCK 
 - A PROBLEM FOR ORIGINAL STUDY 
 
 FOR general directions see Chapter XI. 
 The Models. — The rectangular block is 4" square by 8" 
 long. It may be made of cardboard, cut as in the diagram 
 (Fig. 74), and glued,^ like the cube in Chapter XVI. Or two 
 cubes, made as there directed, may be used in its place. The 
 cylinder is 4" by 8", and has a circle about its middle. 
 
 
 
 
 
 _-/6,» 
 
 
 
 <■ 
 
 -• ^ 
 
 
 t 
 
 i 
 si 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ,-.! 
 
 
 
 Fig. 74 
 
 Fig. 75 
 
 Positions. — The block lies on one long face, its long edges par- 
 allel with the picture plane. The cylinder stands on one base in 
 front of the block, touching it at its middle (Fig. 75) The 
 models rest on a surface three times the height of the block below 
 the eye, and are four feet distant. 
 
 1 The light lines indicate where it is scored and bent for the edges of the block. The 
 quarter-inch projections are laps for fastening. 
 
 48
 
 Chapter XVI 
 
 THE FURTHER STUDY OF STRAIGHT- 
 LINE OBJECTS — A CUBE AT ANGLES 
 WITH THE PICTURE PLANE 
 
 THE Model. — For this study make a cube, four inches on a 
 side, from cardboard cut as in the ilhistration (A, Fig. 76). 
 Pass a string under one edge and out of adjacent corners 
 (B, Fig. 76) before glueing together. 
 
 Study of the Subject. — Turn the cube SO the string comes from 
 the upper front corners, and place it 
 as the book was placed in Chapter 
 XII (Fig. 77). Now, holding a front 
 corner of the cube firmly, revolve 
 the cube on that corner, bringing 
 the side x into sight (Fig. 78). The 
 moment the cube begins to revolve, 
 
 A 
 
 16;^- 
 
 the front, ^, 
 
 begins 
 
 to be turned 
 away, ceasing to be 
 parallel with the pic- 
 ture plane, and tend- 
 ing toward coinciding 
 with the direction of seeing 
 
 :^i 
 
 Fig. 76 
 
 In proportion as it 
 is turned away, its right edge (H) becomes shorter, 
 so that its upper and lower edges (E and F) appear 
 to converge. The cube may be revolved until 
 these edges (and their parallel, D) in their turn 
 converge directly in frout (Fig. 79), as A and B did 
 at first. Then the side x^ becoming parallel with the picture 
 plane, will in turn be seen in its true shape, while its top and 
 bottom edges appear horizontal. 
 
 4 49 
 
 Fig. 77
 
 FREEHAND PERSPECTIVE 
 
 Fig. 78 
 
 
 Now turn the cube slowly back to the position of Fig. 78, 
 and with the stnngs find the converging point of A and B, 
 
 Figure 80 shows the cube in this position, 
 
 and the vanishing of A, B, and C by the use 
 
 of three strings. The lines at right angles 
 
 to them, which in Fig. 79 vanished directly 
 
 in front, here (in Fig. 80) vanish so far to the 
 
 right that the strings cannot reach their 
 
 vanishing point. 
 
 If the cube is now revolved in the opposite direction, this 
 
 vanishing point (which we may call VP2 ^) again moves 
 
 inward, as seen in Fig. 81. 
 
 It will be now readily seen that though any set of 
 parallel horizontal lines (as A, B, and C) are directed 
 more to the right or left, according as the cube is turned, 
 
 they are never actually raised 
 or lowered. Hence their van- 
 ishing point does not move 
 up or down, but is always ^'^- '^^ 
 found on the eye level. Wo may 
 therefore conclude that receding hori- 
 zontal lines always vanish in the eye 
 level. 
 
 We also confirm what was observed 
 in Chapter XII, that parallel lines vanish 
 to the same point. 
 
 Let us now study the effect on the 
 shape of its faces of revolving the cube. 
 In Figure 77 the front face, «/, appears 
 in its true shape, while lines at right 
 angles to this face (as A and B) vanish 
 directly in front, and the sides x and y 
 are invisible. As the cube is revolved 
 (Fig. 78) so that x comes into sight, so y is turned away, or 
 
 1 In distinction to that already found. Vanishing points are numbered in order of finding. 
 
 50 
 
 Fig. 80 
 
 £yc Cc^&l 
 
 // ' 
 
 
 
 / 
 
 
 / 
 
 Fig. 81
 
 STUDY OF STRAIGHT-LINE OBJECTS, ETC. 
 
 ArjAloT A CufiE. 
 
 A.B,AA<oC ARE. 
 TOO/TEEO Fob 
 Tut WIDin OF 
 
 X; AND D, E 
 
 ANO F ABE 
 AJOT yiEEP 
 ENOUOM FOB 
 THE roRE- 
 
 OF Y- 
 
 G-CoWECTIeH 
 
 OF A, 8Y 
 CHANGINO THE 
 WIDTH OFTHa 
 
 THEomecnoN 
 
 OF THE VANI/H- 
 ^ ec f?ioriT. 
 
 B— Correction 
 
 OF A, BY CMANO- 
 IklG THE _fuANTOF 
 THE VANI/HINO 
 
 eoosf, whem the 
 wicth of the ■ 
 
 51 D£.^ 1; FOl/NOTO 
 flS «I&MT. 
 
 Fig. 82 
 
 foreshortened. As x widens, and its horizontal edges (A and B) 
 seem less steep, the other side narrows, and its horizontal edges 
 (E and F) appear more steep. Steep- 
 ness of the horizontal edges, therefore, 
 goes ivith foreshortened surfaces. Good 
 judgment on this point is very impor- 
 tant, as the cube is the basis for later 
 estimates of foreshortened surfaces. 
 For this reason much space has 
 been given to its study. It should 
 be drawn with great care till thor- 
 oughly mastered. 
 
 The Recession of Horizontal Surfaces Toward the Eye Level. — It 
 will be interesting here to place several cubes in a receding row, 
 and see how the vanishing lines, being all included in one or 
 the other of two sets, will vanish accordingly to one or the other 
 
 of two vanishing points. Taking out 
 every second cube (Fig. 83), we find the 
 vanishing of those left to be unaltered. 
 We also perceive that the table on 
 which all rest seems to rise as it re- 
 cedes, apparently tending to vanish or 
 merge itself in the line marking the 
 eye level. 
 
 Looking at the tops of the cubes, all 
 situated in one horizontal plane, and recalling the horizontal 
 surfaces in previous drawings (as the book covers and the 
 cylinder ends) we conclude that all horizontal surfaces appear 
 to approach the level of the eye as they recede. This is seen 
 to be true whether they are below the eye or above it. The 
 vertical distance between receding horizontal planes must ap 
 pear less as it is farther from the eye, till at an infinite dis- 
 tance it would be entirely lost, and the parallel planes would 
 vanish in a line (the eye level) as parallel lines vanish in a 
 point. 
 
 51 
 
 Fig. 83
 
 FREEHAND PERSPECTIVE 
 
 The Eye Level. — The eye level, or level of the eye, is not 
 actually a line ; but a height, or invisible horizontal plane, which 
 may be said to extend indefinitely. Thus if the student's eye is 
 five feet above the ground, his eye level passes through and 
 includes every point at that height. But as each one's eye level 
 is " edge to" him, it would (if visible) always appear to Mm as a 
 line, hence it is always drawn as a line. 
 
 52
 
 Chapter XVII 
 THE CUBE IN TWO DIFFERENT POSITIONS 
 
 THIS exercise should first be drawn from the objects, and 
 then from memory, according to the general directions 
 for memory work (Ch. XI). Two drawings on one 
 sheet, showing the cube in different positions, are to be made. 
 They should be represented as of the same size, which may be 
 
 r-- — 
 
 / 
 
 Fig. 8-1 
 
 done by making their nearest vertical edges of the same length 
 and at the same height on the paper, and using the same eye 
 level for both. 
 
 Position of Models. — For the first drawing, place the card- 
 board cube so that its front faces are equally turned away, or 
 make angles of forty-five degrees with the picture plane (A in 
 
 5«i
 
 FREEHAND PERSPECTIVE 
 
 Fig. 85, also plan). Notice that its upper back corner will then 
 appear exactly behind the upper front one, the vertical sides 
 
 Fig. 85 
 
 of equal width and the side corners opposite each other and 
 equidistant from the center. In the second position the cube is 
 turned so its right face makes an angle of 
 sixty degrees with the picture plane (B in 
 Figs. 85 and 86). 
 
 Making the Drawing. — Fasten the paper 
 in its place on the desk or have its posi- vuv^ or a plah of b 
 tion so marked that it can be accurately ^'^" ^^ 
 
 returned to the same place. Draw the margin lines and lightly 
 mark the extreme points for the two cubes (Ch. IV). Note that 
 
 in the first position (A, Fig. 85) ^the cube 
 occupies slightly more space, both horizon- 
 tally and vertically. Since the cube is a 
 type solid the lines in its final rendering 
 are simple and firm, only varying slightly 
 in thickness to suggest distance. Begin 
 the first cube with the easiest part, which is its nearest vertical 
 edge. This is parallel with the picture plane, and so is drawn 
 in its true position. As soon as this line is placed mark the 
 eye level (in this case it falls off the paper) finding its height 
 as directed for the book. The numbers on the diagram (Fig. 87) 
 
 54 
 
 DlAGPAM 
 SHcM/lNG A 
 CONVENIENT 
 ORDER FOR 
 DRAWING THE 
 LINES OF THE 
 CUliC, AMD OF 
 RECTANGULAR 
 OBJECTS IN GENERAL 
 
 Fig. 87
 
 THE CUBE IN TWO DIFFERENT POSITIONS 
 
 give the order iu which not only cubes, but rectangular ob- 
 jects generally, should be drawn. Get the direction of lines 
 2, 2 by pencil measurement (Ch. II) with espe- 
 cial care, as their meeting with the eye level 
 determines the vanishing points. Hold the pen- 
 cil vertically in front of, or even touching the 
 nearest end of line 2 (Fig. 88). Then keeping 
 it parallel with the picture plane 
 (that is, not receding as the line 
 does, but resting in an imaginary 
 vertical plane) revolve it down- 
 ward to the right until it seems to cover line 2 
 (Fig. 89). Holding it thus, with the other hand 
 slip the paper (on which the 
 drawing has been started) up 
 vertically behind it till the pen- 
 cil touches the upper end of 
 the vertical line already drawn, 
 
 and lies on 
 
 Fig. 88 
 
 Fig. 89 
 
 the paper, 
 sh o win g 
 
 Fig. 90 
 
 the direction line 2 should take (Fig. 90). 
 (This puts the paper in the position of 
 the picture plane.) Draw this first line 2, 
 and mark its vanishing point on the eye 
 level (VPl). The direction of the other 
 line 2 could be found in the same way 
 but as in this case they make equal angles 
 with the picture plane, their vanishing 
 points will be equidistant from the center, 
 
 and VP2 can therefore be so located, and the second line 2 drawn 
 
 to it. Lines 3, 3 are then drawn (recalling that parallel lines 
 
 converge to the same vanishing point). 
 
 For lines 4, 4 compare the apparent width of a near vertical 
 
 face (A in Fig. 91) with the front vertical line (B in Fig. 91). 
 
 55 
 
 Fig. 91
 
 FREEHAND PERSPECTIVE 
 
 (This front line, being seen in its actual position and unfore- 
 shortened, is the best for use as a unit of measurement.) Mark to 
 right and left from line 1 in the drawing the proportionate dis- 
 tance so found (in this case two thirds of line 1) and draw lines 
 
 4, 4. From their upper extremities 
 draw lines 5, 5 to their respective 
 vanishing points. 
 
 For the second drawing place the 
 
 ^ ? ' \ lui ^^^® ^^ directed, and proceed as 
 
 ^L^^^,^^.' ill with the first cube. In this case 
 
 I "// 
 
 Fig. 92 
 
 yP2 falls so far away that it can- 
 not be shown in the illustration 
 (B, Fig. 85). But we know that 
 it must fall somewhere in the eye 
 level. (It will be so found in the 
 illustration, if tested.) At this stage a string pinned to VP4 will 
 aid in detecting errors of vanishing, and will also make real the 
 fact that these lines must vanish precisely to their oivn vanishing 
 point. 
 
 A Valuable Testing Method. — After this the following far more 
 speedy and convenient method of testing should be acquired: 
 With one eye closed hold the drawing close to the eye level, and 
 turn it so that 
 one set of van- 
 ishing lines are 
 directed to the 
 open eye (Fig. 
 92). Push the 
 drawing back 
 or forward as 
 needed till the 
 
 eye occupies the place of the vanishing point for the lines in 
 question. Now sight back over this set of converging lines, 
 when it will be found that any failing to properly vanish are 
 quickly seen and easily noted for correction. A little expe- 
 
 5() 
 
 Fig. 93
 
 THE CUBE IN TWO DIFFERENT POSITIONS 
 
 rieuce is needed to do this successfully, but it is well worth 
 the trouble. 
 
 Testing Before a Class. — An impressive method of demonstra- 
 ting the vanishing of lines when teaching a class is the following. 
 Draw a long horizontal line on the blackboard and mark it " Eye 
 Level." Tack each pupil's drawing in turn on the blackboard so 
 that the blackboard eye level coincides with the eye level of the 
 pupil's drawing. With a long ruler follow out one of the vanish- 
 ing lines (Fig. 93), and find its vanishing point on the blackboard 
 eye level. Holding the ruler at this vanishing point as a pivot, 
 swing it over the other lines of the set that should vanish to that 
 point. The test is convincing, even to children; and helps 
 greatly to form a standard of accuracy. 
 
 It should always be remembered, however, that such measur- 
 ing is only for testing^ never for drawing the lines. 
 
 57
 
 Chapter XVIII 
 
 A BOOK AT ANGLES TO THE PICTURE 
 
 PLANE 
 
 THE student may copy this example but must in any case 
 draw from a book similarly placed ; and finally make a 
 correct and spirited drawing of the same from memory. 
 The position of this book is like that of the last cube (Ch. 
 XVII). In studying this position begin with the book directly in 
 
 r 
 
 '«aaya<ii>^DeaaiCfflagia»aas:!Wai/.j; 
 
 »rn*irmj«»n'Wt(i*(-'<^icnr2:3r:iHSH| 
 
 Fig. 94 
 
 front as in Chapter XII. Note the convergence of its ends ; then 
 turning it slowly into the required position for drawing (Figs. 94 
 and 95), observe how the ends change in their convergence and 
 how their vanishing point moves to the right on the eye level 
 as the book is turned. Look also at the long edges of the 
 book and see how, at the first movement of revolving it, they 
 
 58
 
 A BOOK AT ANGLES TO PICTURE PLANE 
 
 cease to appear horizontal, and vanish toward a point which, 
 though at first infinitely distant, must nevertheless fall on the 
 eye level. 
 
 Drawing the Book. — Sketch the margin lines, and plan a good 
 position of the book in relation to the inclosed space. Mark the 
 height of the eye level as soon as a dimension (as xy. Fig. 95) by 
 
 Fig. 95 
 
 which it can be estimated is decided on. Find the direction of the 
 book edges (corresponding to lines 2, 2 in the cube in Ch. XVI) 
 with especial care. Sketch in the book with delicate lines, pro- 
 ceeding in the order observed when drawing the cube, and 
 correcting where necessary. 
 
 Artistic Expression. — Finally, the subject should be rendered 
 artistically. To accomplish this, the line is adapted to the qual- 
 ity of that portion on which it is used. Certain features may be 
 selected for use to augment interest ; as the curving ridges, the 
 ornament, and the title space on the back, or even the worn 
 corners. But having expressed in these details the point intended 
 (as a worn corner by the shape of its boundary line) take care to 
 do no more. It is wearisome, for instance, to see lines on these 
 corners to represent the separation into layers caused by wear. 
 Lines also produce a dark color, while worn corners are generally 
 light; and are also undesirable places for the use of dark 
 spots. 
 
 59
 
 FREEHAND PERSPECTIVE 
 
 As the vertical edges of the cube are drawn vertical because 
 parallel with the picture plane, so the corners of the book must be 
 made vertical in the drawing, as they are in reality. For in- 
 stance, points C and D being in a vertical line, must be so placed 
 in the drawing. The same is true of the curves on the back of 
 the book. 
 
 At this point the student readily sees that all vertical lines 
 (since the picture plane is vertical) will he parallel to the picture 
 plane, and must invariably be draivn as they actually are, or vertical. 
 
 60
 
 Chapter XIX 
 
 TWO BOOKS AT DIFFERENT ANGLES TO 
 THE PICTURE PLANE 
 
 B 
 
 EGIN the study of this subject by placing the books as 
 in Fig. 97. Observe that in this position there is but 
 one vanishing point for the two objects, the ends of 
 
 Fig. 96 
 
 the books being all parallel, and their other horizontal edges 
 parallel with the picture plane. Now turn the whole group, 
 as in Fig. 98, and see that we have two vanishing points, 
 
 61
 
 FREEHAND PERSPECTIVE 
 
 EYiUvU- 
 
 one for the ends and the other for the long edges of the 
 books. 
 
 Now revolve the upper book a little more (Fig. 99), so 
 that its horizontal edges cease to be parallel 
 to those of the other, 
 and it will have its 
 own points of con- 
 vergence (VPS and ^-^vf 
 VP4). Its length ap- ^ 
 pears lessened, and its 
 ends longer, for this Fig. 98 
 
 change. The shortened edges vanish more steeply, and those 
 which have become longer appear less steep. We find, as 
 would be expected, that when Imes cease to he parallel, their 
 vanishing points are different. 
 
 Fig. 97 
 
 •*• TjCf A J<, 
 
 Fig. 99 
 
 62
 
 Chapter XX 
 
 THE ACTUAL CENTER OF THE CIRCLE AND 
 MEASUREMENT INTO THE PICTURE BY 
 PARALLEL LINES 
 
 PRELIMINARY Study. — Does the eye see half way round the 
 cylinder I The question is best answered by experiment. 
 Holding the cylinder vertically and rather near (to 
 more easily see the facts), mark on it the points where the side 
 boundaries appear to meet the top (A and B in Fig. 101). It will 
 
 Fig. 100 
 
 be found that they are actually less than half way from the front 
 to the back (Fig. 102). Yet the pencil has marked what the eye 
 mw as the greatest dimension. As shown in Fig. 103 this appar- 
 ent greatest dimension (A B) forms the long diameter of the ellipse 
 
 63
 
 FREEHAND PERSPECTIVE 
 
 Showing apparent middue from 
 
 FffONT To BACK. OR LONG DIAMETER 
 
 Fig. 101 
 
 in the perspective view. It is evident, therefore, that the eye does 
 not see half way round the cylinder, and (as seen in Ch. IV) that 
 
 the long diameter of the ellipse is 
 not an actual diameter of the circle, 
 while that portion of the circumfer- 
 ence beyond the long diameter (A B) 
 is actually more than half of the circle, 
 the part in front of A B appearing equal 
 to it only because nearer to the eye. 
 
 The actual position of the apparent 
 greatest dimension (the long diameter) 
 changes with the position of the observer. 
 The plan (Fig. 102) shows that C D would appear as the greatest 
 dimension if the eye should be at 2. This also may be seen by 
 experiment (as in Fig. 101). 
 
 Planning the Exercise. — In placing this exercise observe that 
 the perspective of the concentric square and circles is made 
 much larger than the geometric diagram, 
 to show more clearly the perspective 
 details. 
 
 Drawing the Circles. — When the square 
 has been drawn in perspective (like the 
 top of the cube in Fig. 77) its actual 
 center (o in Fig. 103) is found at the 
 crossing of its diagonals, as in the geo- 
 metric diagram above. In the diagram 
 the ends of its diameters mark the points 
 (C, D, E, and F) where the circle touches 
 the square, and they will do the same in 
 the perspective. The diameters pass through the true center and 
 one is parallel to the picture plane. It can therefore be drawn 
 in the perspective in its actual direction, giving two points 
 (c and d). The other diameter, being parallel with the receding 
 sides of the square, vanishes with them in the eye level directly 
 in front (at VPl) giving points e and /. Now, though the actual 
 
 61 
 
 R-ANOFABCVC, 
 5HO"0t/lN& THE 
 
 ACTyAu Place 
 
 OF THE LON& 
 DIAMETER, A B
 
 ACTUAL CENTER OF CIRCLE, ETC. 
 
 Geometric diagram. 
 
 diameter of the circle touches the square in c and d, the ellipse 
 
 appears longest at a part nearer than c and d, which seems to 
 
 be exactly half way between e and /. Through this half-way 
 
 point {x) the long diameter can 
 
 be drawn ; making it longer than 
 
 c-d, and yet not quite touching 
 
 the square. The ellipse is then 
 
 easily sketched through these six 
 
 points (a, 6, c, d, e, and/), making 
 
 it symmetrical on a-b and e-f. 
 
 For the other ellipses the points 
 where they cross the actual diam- 
 eter of the circle [c-d) are marked 
 by lines from 1, 2, 3, and 4 which 
 vanish in VPl, giving four points 
 (9, 10, 11 and 12), two for each 
 of the smaller ellipses. 
 
 Measuring Distances into the 
 Picture. — For the front and back 
 points of these ellipses, line ef 
 must be divided into six perspec- 
 tivelij equal parts, as EF in the 
 diagram is divided into six actu- 
 ally equal parts. This can be done, and in practice usually is 
 done, by the eye (as for the cylinder in Ch. IV), noting that 
 the true center (o), already known, is one point of division. But 
 the use of the diagonal for such distances is simple and often 
 a convenience. Thus it is easy to see that in the diagram the 
 vertical lines from 1, 2, 3, and 4 cut the diagonals proportionately 
 to the divisions on GH^ in this case into six equal parts. These 
 divisions can in turn be transferred to EF by horizontal lines 
 from the points on the diagonal HI, giving 5, 6, 7, and 8, the four 
 
 1 Students of geometry will recognize in this the problem 
 of dividing a line proportionately by means of parallel lines cross- 
 ing a triangle. 
 
 5 65 
 
 Perspective representation of above. 
 Fig. 103 
 
 Fig. 104
 
 FREEHAND PERSPECTIVE 
 
 points needed. In the perspective the method is the same, using 
 lines perspectively parallel to e-f — that is, the lines already 
 drawn from 1, 2, 3, and 4 to VPl. This use of the diagonal 
 occurs further on, as for the steps in Chapter XXII. 
 
 A Second Method. — The vanishing point of the diagonal can 
 also be used to obtain these points. Thus the diagram shows 
 that lines from 1, 2, 9, and 10 parallel with the diagonal GJ will 
 mark on EF the same divisions. In the perspective these lines 
 will appear perspectively parallel to the diagonal — that is, drawn 
 to the same vanishing point. Since they are horizontal, that 
 vanishing point will be on the eye level. Therefore the diagonal 
 G-J can be carried out to the eye level to find its vanishing point 
 (VP2) to which the parallel lines are drawn. 
 
 The principle to be remembered for use is: Whatever meas- 
 urements can he obtained geometrically by the use of actually parallel 
 lines, can be obtained in perspective by the use of perspectively parallel 
 lines. 
 
 It must, however, be noted that these are only relative meas- 
 urements. A first distance into the picture — the foreshortened 
 width of the square in this case — is determined freehand by 
 past experience (as with the cube). Mechanical perspective 
 gives methods of obtaining this first distance, the position of the 
 eye and the picture plane being given. It can also be obtained 
 from a side view, by using the same data. Both these methods 
 are too complicated for common use in freehand work. Such 
 proportions are so easily estimated by recalling the cube that it 
 is better to rely on a trained judgment for them. 
 
 66
 
 Chapter XXI 
 BOOKS WITH A CYLINDRICAL OBJECT 
 
 T 
 
 HE student should take this exercise as previous ones, 
 copying first if he needs to do so, then composing and 
 sketching a similar study, and finally making a drawing 
 
 Fig. 105 
 
 from memory. For both of the latter several different arrange- 
 ments of objects, with trial sketches, should be made; and the 
 best chosen to use in the final drawing. 
 
 The Finder (Ch. VIII) should be used to compare the 
 effect of different compositions, also the effect of cutting out 
 
 67
 
 FREEHAND PERSPECTIVE 
 
 compositions from larger ones by different margins. Note 
 that in Fig. 106, with a tall object, the books are turned so 
 
 that their horizontal dimensions are 
 not great enough to neutralize the 
 dominant vertical effect. In Fig. 
 107, on the other hand, the long 
 horizontal dimensions of the books 
 and the low flat dish harmonize 
 very well, and this arrangement 
 
 Fig. 106 
 
 Fig. 107 
 
 necessitates a marginal rectangle longer from side to side. 
 The books and dish alone make a good simple arrangement, 
 but the tray may be added if desired. 
 
 Fig. 108 
 
 68
 
 Chapter XXII 
 THE STUDY AND DRAWING OF A HOUSE 
 
 M 
 
 ODEL for the Study. — Make an equilateral triangular 
 prism from cardboard cut as in the diagram (Fig. 110), 
 and place it on the top of two cubes. Put a box or 
 
 Fig. 109 
 
 books on the table under this model, raising it so that the level 
 of the eye will fall one-fourth way up on the cubes (Fig. Ill) . 
 Place the model about sixteen inches from the eye, and turn it 
 so its long edges will make angles of thirty degrees with the pic- 
 ture plane (Fig. 119) . It may now be regarded as the type form 
 of a house, seen (in proportion to its size) from an ordinary posi- 
 
 69
 
 FREEHAND PERSPECTIVE 
 
 8!4-^ 
 
 Fig. 110 
 
 tion for viewing a house. By aid of the imagination, it may- 
 be regarded as a house of two stories, with a front door in the 
 
 middle of a side, the box top taking 
 the place of the ground. 
 
 This exercise should be first drawn 
 in thin, light lines, studying the dia- 
 grams, and following the directions. 
 The construction lines should then be 
 erased, and the drawing rendered as 
 shown in Fig. 109. 
 
 Drawing the Exercise. — Begin with 
 the nearest vertical edge of the house. 
 The model was placed so that the 
 level of the eye should be one-fourth way up the height of its 
 rectangular part because the eyes of a person standing might 
 be about five feet above the 
 ground, and the height of a 
 two-story house at its eaves 
 about twenty feet. Mark the 
 eye level on the paper there- 
 fore, one fourth of the height 
 of the nearest edge from its 
 bottom, and take the direction 
 of lines A and B (Fig. 112) 
 to determine the two vanishing points, exactly as was done 
 with lines 2, 2 in the cube (Ch. XVII) although, being above 
 the eye, they appear to tend downward. Draw the two lower 
 horizontal lines and the side vertical lines as those of the cube 
 were drawn. 
 
 To construct the roof recall that the end of our model is 
 an equilateral triangle (Fig. 113) with its apex over the center 
 of the house end. Draw the diagonals of this square house end 
 and carry up a vertical line of indefinite length from its center, 
 on which the apex of the gable is to be marked. The actual 
 roof height of our small model may be found by this diagram ; 
 
 70 
 
 Fig. Ill
 
 THE STUDY AND DRAWING OF A HOUSE 
 
 but as that makes a roof steeper than is usual, we will set it 
 off less in the drawing, that is, making EC (Fig. 112) but a 
 
 Fig. 112 
 
 little more than CD. (Since these distances are in the same 
 vertical line, and so at the same distance from the picture plane 
 they are seen and drawn in their true proportions to each 
 other.) 
 
 The sloping sides of the gable may then be drawn 
 to the house corners, and the ridgepole to VP2. The 
 gable apex on the other end may be found by drawing 
 a vertical line from the center (x) of the invisible end 
 to cut the ridgepole. Its slanting sides are drawn to 
 complete the blocking-in lines thus far of the house. 
 
 71 
 
 c ^ 
 
 E.ND View 
 Fig. 113
 
 FREEHAND PERSPECTIVE 
 
 Oblique Vanishing Lines. — We have said little about the slop- 
 ing end lines of the roof. But now, looking again at our 
 model (Fig. Ill), we see that the ridgepole, R, because it is 
 farther away than the eaves, appears shorter, so that the 
 slanting ends (F and Gr) of the front surface of the roof appear 
 to converge upward. Turning to the drawing, we find in con- 
 firmation of this that (if the drawing has been carefully made) 
 these lines do thus converge. Now let us search for the general 
 truth governing that convergence. These slanting ends are not 
 horizontal, so that we should not expect them to tend toward the 
 eye level ; and we observe that they do not. But they are actu- 
 ally parallel to each other and therefore must vanish or appear 
 to converge to the same point. Hoiv to find that point is the 
 question. 
 
 Put some books in the place of the house model, and arrange 
 them so that their edges vanish like the house edges. Now raise 
 
 the upper cover (A, 
 ^'' ~ Fig. 114), and observe 
 
 that its ends 1 and 2, 
 though still parallel 
 with each other like 
 the ends of the roof, 
 have ceased to be 
 parallel with the 
 other book ends. 
 They therefore no 
 longer vanish toward 
 VPl, but to a higher 
 point. We have not, 
 however, turned these 
 edges to right or left, but have simply lifted their farther ends 
 or revolved them in parallel vertical planes. Therefore their 
 vanishing point cannot move to right or left; but as they are 
 revolved, must appear to move directly uptvard, or in a vertical line 
 passing through VPl. This continues until the cover becomes 
 
 72 
 
 Fig. 114
 
 THE STUDY AND DRAWING OF A HOUSE 
 
 vertical; when its ends appear in their true position and cease 
 to vanish, like all vertical lines. 
 
 Place strings under the cover, as in Chapter XII. Holding 
 the strings with the left hand as in the illustration (B in Fig. 
 114) raise the cover with the right (keeping the strings parallel 
 to the picture plane). By this experiment their convergence 
 toward a point in the vertical line from VPl is more plainly 
 shown. Since these slanting book ends are neither horizontal 
 nor vertical but oblique to both directions, their vanishing point 
 or that of any set of oblique lines, may be distinguished as an 
 Oblique Vanishing Point, or OVP. 
 
 By revolving the upper cover farther, or opening the lower 
 cover and using the string (A in Fig. 115), oblique vanishing 
 points below the 
 eye level may be 
 determined. And 
 the apparent di- 
 rection of oblique 
 lines can be found 
 with the pencil ex- 
 actly as that of 
 any line (B in 
 Fig. 115). 
 
 Vanishing Traces. 
 — By turning one 
 of these illustra- 
 tions around, to bring the eye level vertical, as in Fig. 116, it 
 will be seen that the line containing OVPl and 0VP2 serves 
 a purpose similar to that of the eye level. We note that 
 the surface formed l)y the visible ends of the hooJcs appears to 
 recede, or vanishes, toward this line. If a larger book be placed 
 against the other ends, the surface of the larger book, being 
 parallel to the visible ends of the other books, will be found to 
 vanish toward the same line. It may be concluded that all sur- 
 faces parallel to the hook ends in this case will vanish in this line, 
 
 73 
 
 Fig. 115
 
 FREEHAND PERSPECTIVE 
 
 exactly as all horizontal surfaces appear to vanish toward the 
 eye level. We may call this line a Vanishing Trace. The eye 
 level is such a vanishing trace for all horizontal surfaces. See 
 
 note, Chapter XI, in Solu- 
 tions of Problems. 
 
 First replacing the house 
 model as in Fig. Ill, we now 
 turn to the drawing and test 
 these oblique lines (P and Gr 
 in Fig. 112). If correctly 
 drawn they will be found 
 to converge toward a point 
 (OVPl) directly above VPl. 
 At once use is made of this 
 point for drawing the ends 
 of the roof projection (sug- 
 gested in the model by pin- 
 ning cardboard as in Fig. 
 117). These edges are par- 
 allel with the corresponding roof edges, like the book margins; 
 so their width can be set off on the upper line of the house (B, 
 Figs. 112 and 117) to right and left, 
 (points X and y) remembering that 
 the nearer distance appears a little 
 greater. Through these points draw --^ 
 lines vanishing to OVPl. A similar 
 projection is measured downward on a 
 continuation of the oblique gable edge 
 F beyond its lower end (z) ; and through 
 this point a line parallel to B (that is, vanishing with it in VP2) 
 forms the eaves. (The estimation of these last measurements by 
 the eye forms an important part of the student's training and 
 should be carefully thought out. Thus the eaves projection from 
 line B forward is more foreshortened than the gable projection 
 from F to the left; and distances should be set off accordingly.) 
 
 74 
 
 Fig. 116 
 
 I E 
 
 Fig. 117
 
 THE STUDY AND DRAWING OF A HOUSE 
 
 BAC»\ VIEW 
 
 End View 
 
 Fig. 118 
 
 The lower oblique vanishing point (0VP2) is used for the 
 projections on the back slope of the roof. Continue the back 
 oblique roof edge (line H) to meet the vanishing trace through 
 VPl, giving 0VP2. Draw line K from the near end (I) of the 
 ridgepole to 0VP2, and cut it by a line from the nearest eaves 
 corner to VPl. Where this line cuts the oblique edge K will be 
 the eaves corner (M) for the far side of the roof ; and a line from 
 it to VP2 fornas the eaves on that side. 
 
 There is another way of getting the projections on the 
 further slope of the roof, which is useful in case 0VP2 falls too 
 far away to be conveniently used. 
 Turning the model we see that the 
 invisible line (L in back view, Fig. 
 118), if carried to the edge, ends 
 in O, horizontall}^ opposite x (end 
 view). A line from x through O 
 would therefore vanish in VPl. 
 Hence, to obtain 0, line L is carried forward indefinitely, and 
 cut by a line from x to VPl. The desired edge is then drawn 
 from point I through O indefinitely, and cut by a line from J to 
 VPl, giving the corner, M. 
 
 The "L" Part of the House. — The plan (Fig. 119) shows its 
 position. Its width is marked off on the farther (invisible) 
 
 end of the house, and it is drawn 
 as was the main part of the 
 house. Note the less steep slope 
 of the porch roof (Fig. 112), so 
 that its oblique lines are not 
 parallel with those of the other 
 roofs but have another vanishing 
 point, 0VP3, lower in the same 
 vertical line. 
 
 Windows and Doors. — The windows and doors may be marked 
 on the model (Fig. 120). It will be easily seen that their top 
 and bottom edges are all parallel to the horizontal lines of the 
 
 75 
 
 Fig. 119
 
 FREEHAND PERSPECTIVE 
 
 side on which they are located and therefore vanish to the same 
 point. Mark their heights on the nearest vertical edge of the 
 house, and draw lines (as P, in Fig. 112), thence to the vanishing 
 
 -D- 
 
 DD D DD 
 
 fEONT VIEW 
 
 Fig. 120 
 
 points. On these lines their perspec- 
 tive widths are to be set off. Begin 
 with the door. Find the middle of 
 the house front by its diagonals, and 
 make the near half of the door a 
 little wider than the far one. Check 
 this by seeing that the remaining 
 distances (from the door to the front corners of the house) are 
 also perspectively equal, that is, the near one larger. Mark the 
 sides of the windows in the same way. Remember that since 
 the space between the near window and the near corner is con- 
 siderably nearer to us than that between the far window and the 
 far corner, more difference should be made in their size than 
 between the halves of the door. 
 
 The width of the windows on the end of the house should be 
 to that of the front ones as the right face of the second cube in 
 Chapter XVII (Fig. 84) is to the right one. The height of the 
 windows in the " L " is made the 
 same perspectively by carrying their 
 measurements from the right front 
 corner of the main house on lines 
 vanishing to VPl. These lines lie 
 on the invisible end of the main 
 house, and from where they reach the 
 "L" are continued along its front by 
 lines running to VP2. 
 
 The Chimney. — To better visualize this part of the house, cut 
 and fold cardboard as in Fig. 121. Get the slope of lines 1, 2, 3, 
 and 4 by laying the cardboard against the apex of the gable 
 and marking around it. Stand this model on the roof in its 
 middle, and after marking on the roof around it, cut out the 
 space so marked and push the chimney down through the open- 
 
 76 
 
 --4- IN 
 
 Fig. 121
 
 THE STUDY AND DRAWING OF A HOUSE 
 
 iug until it projects the proper distance above the roof. Lay a 
 pencil on the roof against the chimney (Fig. 122) , and move it to 
 the left without changing its direction till it coincides with the 
 gable edge. This shows the gable 
 edge and the oblique line where the 
 chimney passes through the roof to 
 be actually parallel. This oblique 
 line, therefore, has the same vanish- 
 ing point as that of the gable line, 
 which is OVPl. The top of the 
 chimney front and the line below 
 it (AB, Fig. 123) are parallel to the 
 eaves and ridgepole. Turn the model (Fig. 120, end view) and 
 see that the top edge of the chimney is parallel with the hori- 
 zontal lines on the house end, which we have already drawn ^ to 
 
 Fig. 122 
 
 Fig. 123 
 
 VPl. A pencil held horizontally and moved slowly up in front 
 of the model will help to see this as will marking the chimney in 
 the model off into bricks (Fig. 124). 
 
 Drawing the Chimney. — Continue on the roof the center line 
 used for the door (that is, vanish a line from its top to OVPl), 
 
 77 
 
 ^
 
 FREEHAND PERSTECTIVE 
 
 r^w^ra^s^ 
 
 a!& 
 
 J 
 
 A- Plan or Cmimney. 
 g- "Profile smowino orna- 
 mental. BAND AT TOP. 
 
 Fig. 124 
 
 and mark down from the ridgepole on this line half the thickness 
 of the chimney (judged by the eye). Draw a line through this 
 
 point toward YP2, and on it set off to right 
 and left perspectively equal distances for 
 the breadth of the chimney (AB). Draw 
 line C to OVPl. Where it crosses the ridge- 
 pole (D) is the middle of the chimney from 
 front to back. Make the far half of the 
 chimney proportionately as much smaller 
 than the near half as the far half of the 
 house end is smaller than its near half. 
 The projecting band at the top of the chimney is shown in 
 plan and profile in Fig. 124. Its perspective is drawn as are 
 projecting book covers. Be careful to represent the backward 
 projection on the farther side. 
 
 The Steps. — For these the detail drawing (Fig. 125) is first 
 made. The height under the threshold of the door is a little 
 less than two feet, or about 
 one third of the height of 
 the eye — enough for three 
 steps. Divide the vertical 
 line under the near edge of 
 the door therefore into three 
 equal parts, and draw lines 
 of indefinite length to VPl 
 through the four points of 
 division. On the lower line, 
 B, mark off the proper dis- 
 tance (as four feet), which 
 
 may be estimated by comparison with the windows on the end 
 of the house (their width being parallel with these lines, and 
 usually about three feet). Divide this distance into perspective 
 halves. A vertical line from the near end of line B, cutting line 
 E in Point 2, completes the rectangle, 1-2-3-4, the middle of 
 which can be found by its diagonals, giving the perspective 
 
 78
 
 THE STUDY AND DRAWING OF A HOUSE 
 
 ^ 
 
 \^ 
 
 c> 
 
 \/ 
 
 <» 
 
 .^B 
 
 /■^ 
 
 B 
 
 Fig. 126 
 
 halves required. (See A, Fig. 126.) The further half is for the 
 wide top step. The near half of the rectangle can be divided 
 again in the same way for the two lower steps. Where the 
 vertical line from the near end of B cuts 
 line C is the upper near corner of the lower 
 step. A vertical line through O will mark its 
 width on C, and continued to cut D forms 
 the nearest front corner of the second step. 
 
 Another method of sketching the steps 
 is shown in Fig. 126. When the first step 
 has been drawn its diagonal is continued 
 through 6, cutting line D in 7, and forming 
 the diagonal for the second step, which is completed by con- 
 tinuing line D to cut a vertical from 6 in 8. This can be con- 
 tinued for as many steps as needed. The diagonal can also be 
 used as a test for steps drawn by the first method. 
 
 The long edges of the steps vanish in VP2, and are cut alter- 
 nately by lines vanishing in VPl and vertical lines. 
 
 The Dormer "Window. — This is constructed in principle like 
 the gable of the roof. The detail drawing (Fig. 127) should be 
 
 carefully studied, and drawn sepa- 
 rately if desired, before sketching 
 the window on the house. 
 
 On the center vertical line of the 
 house front continued upward, mark 
 the height of the dormer from line 
 B. (In this case it is not so high 
 as the main house.) Through this 
 point (S) the dormer ridgepole is 
 drawn to VPl, and cut by the 
 oblique middle line on the roof (in point T). The width (1-2) 
 of the dormer is then marked perspectively to right and left on 
 line B. Through these points (1 and 2) the " valleys," or meeting 
 lines of the dormer with the main roof, are drawn from the roof 
 end (T) of the dormer ridgepole to the edge of the eaves (points 
 
 79 
 
 Fig. 127
 
 FREEHAND PERSPECTIVE 
 
 U and V) . From these points the edges of the dormer roof pro- 
 jection run parallel respectively to A and C. 
 
 Oblique Lines in the Dormer. — These two lines A and C, 
 though oblique to line B, are in the same vertical plane (as the 
 gable lines F and H in the main house are in the same plane 
 with line A in Figs. Ill and 112). Therefore draw a second 
 vanishing trace for oblique lines vertically through VP2, and 
 continue A upward until it cuts this trace in 0VP4, to which 
 draw the edge D. The other oblique edge (E) vanishes in the 
 same vertical below VP2. 
 
 When experience has been acquired, such oblique lines can 
 be satisfactorily drawn without actually finding their vanishing 
 points. Such convergences are generally estimated in practical 
 work. But estimates are much more valuable when made 
 with a knowledge of methods by which they can be definitely 
 determined. 
 
 80
 
 Chapter XXIII 
 
 A BUILDING FROM THE PHOTOGRAPH OR 
 
 A PRINT 
 
 T 
 
 HE example given in Fig. 128 is from the old church of 
 San' Apollinare in Classe, near Ravenna. 
 
 The beginner may draw this as a preparation for his 
 
 Fig. 128 
 
 next work, which should be the drawing of a building or part of 
 one from a print of his own selection. 
 
 Making a Selection. — This choice should be carefully made, 
 care being taken to secure unity, or an appearance of one whole 
 
 6 81
 
 FREEHAND PERSPECTIVE 
 
 Fig. 129 
 
 thing having a center of interest and parts which are subor- 
 dinated, or catch the eye less quickly. It should be well placed 
 
 in its rectangle (Chs. VIII and XXI). 
 For instance, the tall tower in Fig. 129 
 needs a margin that is longest vertically, 
 and quite narrow, to produce a harmony 
 of lines. The smaller buildings with it 
 give variety, and by a contrast which is 
 not too great emphasize its height, be- 
 ing subordinated that the tower may 
 remain prominent in the composition. 
 
 In Fig. 130, on the other hand, 
 the long, 
 low mass 
 of farm 
 buildings 
 set well 
 back into 
 the pic- 
 ture re- 
 quires a rectangle that is longer horizontally. 
 
 Some of the different selections that may be made from 
 
 one print (Fig. 131) are 
 shown in Fig. 132. The 
 beginner can by such 
 means obtain an example 
 simple enough to be with- 
 in his powers and often 
 a better composition. 
 
 Drawing from the Print. 
 — As soon as the place 
 of the building on the 
 paper is fixed, the Jeml 
 of the eye must he deter' 
 mined and marked, and the vanishing points of the principal sets 
 
 'H'iTi 
 
 Fig. 130 
 
 ^■■^■^■KT^^ 
 
 
 [ ^'^% ffi JVMBKl 
 
 
 1 
 
 f ■ 
 
 k "1 
 
 BHHr 
 
 ^^^K|s£|^|u^^ 
 
 Fig. 131
 
 BUILDING FROM PHOTOGRAPH, ETC. 
 
 of horizontal lines must he found on that. It is of course easier 
 
 for the beginner to use such vanishing points as are near enough 
 
 to be marked. But the 
 
 student must fully un- 
 derstand that a point too 
 
 far away to be marked 
 
 can be mentally located, 
 
 and the lines drawn 
 
 toward it with closely 
 
 approximated accuracy. 
 
 The essential thing is to 
 
 have the position of such a 
 
 point dearly thought out, 
 
 — even, for instance, as 
 
 specifically as that it is 
 
 " the width of the board," 
 
 or " three times " that, 
 
 distant. The power to 
 
 do this accurately grows 
 
 rapidly, and can be 
 
 attained by students 
 
 of moderate ability. It 
 
 is one object of this 
 
 study. 
 
 Rendering from the Print. — As more complex sketches are 
 
 made, certain parts may be expressed in color (that is, covered 
 
 with a tone of pencil lines), as was 
 done with the title space of the 
 books in Chapter XIX. The door- 
 way, windows, and shaded sides of 
 the buildings in this exercise (Fig. 
 131), are examples of this. Such 
 
 use of color is intended sometimes to attract the eye to the most 
 
 important or interesting parts, or to bring out the beauty of such 
 
 details as the majestic forms of the trees in A, Fig. 132. 
 
 83 
 
 Fig. 132
 
 FREEHAND PERSPECTIVE 
 
 The Comparative Simplicity of Perspective. — By experience in 
 mentally grouping each new vanishing line with the set to which 
 it belongs, the perspective of apparently difficult studies becomes 
 simple. In Fig. 133 a seemingly complex group of buildings is 
 shown to need but four vanishing points for nearly all of its 
 lines. 
 
 84
 
 Chapter XXIV 
 
 TYPE FORMS HELPFUL IN UNDERSTAND- 
 ING THE HOUSE ^ — THE SQUARE FRAME 
 
 T 
 
 HE Model. — The model for the square frame is six inches 
 on a side, and one inch square in section. Looked at 
 from the front, 
 
 it appears as two con- 
 centric squares one 
 inch apart (Fig. 135). 
 It is placed with one 
 set of long edges ver- 
 tical, and the other 
 horizontal and mak- 
 ing angles of sixty 
 degrees to the left 
 with the picture plane 
 (Fig. 136). 
 
 In considering its 
 shape it may be first 
 regarded as a Plinth, 
 or one-inch rectangu- 
 lar slice from a six- 
 inch cube, and there- 
 fore one sixth of the 
 cube in thickness 
 (Fig. 137). 
 
 ^ Some of the geometric 
 solids here and later given may 
 be omitted at the discretion of 
 the teacher. Those selected 
 for study should be such as to 
 supply any deficiencies in the 
 student's mastery of the subject. 
 
 
 iaia»soKi^^Viz:xsSiS¥i^»^¥iSs:^->^S:^:S:S»ij,a:iVSi^Ui, 
 
 Fig. 134 
 
 85
 
 FREEHAND PERSPECTIVE 
 
 
 A ,/ 
 
 
 
 z 
 
 Y 
 
 
 B 
 
 
 TRONT 
 
 Fig. 135 
 
 Fig. 13G 
 
 Drawing the Model. — The lines of this solid may be drawn 
 and its proportions established in the same manner as those of 
 the cube (p. 54). Remember to place the eye 
 level at once after drawing the first vertical edge. 
 Like the cube, these type forms should be lightly 
 sketched first and later rendered with the firm, 
 simple lines appropriate to them. Being more 
 complicated than the cube, their visible edges 
 may, if necessary, be strengthened (that is, be drawn as they 
 would finally appear) as soon as determined, to avoid confusion. 
 
 The inner square of the frame may ne:!rt be marked 
 out on the plinth (Fig. 137). To do this place points 
 one sixth of the front vertical edge from each end, 
 and from them vanish lines to VPl. In these lines 
 the edges A and B of the inner square must lie. By 
 the front view (Fig. 135), we perceive that the corners 
 of the inner square lie in the diagonals of the outer 
 
 one. One diagonal, C, will mark two corners 
 {x and y) of the inner square. Its other two 
 corners are found by drawing the vertical edges 
 of the inner square, from corner x down to line 
 B and from y up to A. 
 
 If this inner square were cut out, leaving a 
 frame, parts of the inner thickness of the frame could then 
 be seen. Of this inner thickness, line D (Fig. 138) lies in the 
 back surface of the frame, parallel 
 to B, and at actually the same 
 height. Hence it can be started 
 at a point [z) on the right hand 
 vertical edge of the frame, obtained 
 by drawing line E from the near 
 end of B to VP2. This may be 
 called carrying line B " around 
 the corner." From this point z^ D is drawn to YPl. The 
 lower inner edge (F) at the back is parallel to the outer thick- 
 
 86 
 
 Fig. 137 
 
 Fig. 138
 
 TYPE FORMS HELPFUL, ETC 
 
 ness edges, and therefore can be drawn to VP2, cutting off line 
 D. A vertical line from where D and F meet completes the 
 inner thickness. 
 
 Tests. — The correctness of this drawing can be tested by 
 adding the invisible portions, shown 
 by dotted lines in Fig. 138. Thus 
 if line A be carried around the cor- 
 ner, giving the invisible edge H, the 
 inner invisible edge I should cut it in 
 line Gr continued. 
 
 The Application of Type Form 
 Principles. — The application of the 
 foregoing work to the drawing of 
 such parts of the house as windows and doors may be seen 
 in Fig. 139, where the inner edges of door and window frames 
 converge with the set of horizontal lines at right angles to the 
 door. For instance, lines A and B converge with the dormer 
 eaves and other lines tending to VP2, and line C vanishes with 
 the set to VPl. 
 
 Fig. 139 
 
 87
 
 Chapter XXV 
 
 THE SQUARE PYRAMID AND SQUARE 
 
 PLINTH 
 
 T 
 
 HE Models. — The plinth is two inches high and six inches 
 square : the pyramid four inches square at base and eight 
 inches high. The models can be made (Fig. 141). If 
 
 made, note that in 
 order to secure the 
 required height in 
 the completed pyra- 
 mid, the length [xij 
 in Fig. 141) of each 
 triangular side piece 
 of the pyramid pat- 
 tern is measured 
 from xy in Fig. 142, 
 where the true 
 length of a side face 
 is shown. The face 
 x-y-3 in Fig. 142 
 leans back, making 
 x-o foreshortened. 
 
 Position. — The 
 plinth rests on one 
 square face, with its 
 sides at angles of 
 thirty and sixty 
 degrees with the 
 picture plane (Fig. 
 ,.! 142). The pyramid 
 Fig. 140 stands on the plinth, 
 
 88
 
 SQUARE PYRAMID AND SQUARE PLINTH 
 
 T 
 i 
 
 A. Pattern 
 
 FOR MAKING 
 PYRAMID 
 
 «\rith its base parallel to and equidistant from the edges of the 
 plinth top. 
 
 Drawing the Models. — Proceed 
 with the plinth as with the cube, 
 remembering that its height is 
 
 but one third as 
 
 much as the cube 
 
 in proportion to 
 
 its breadth. 
 For the pyra- 
 
 mid base, one 
 
 perspective sixth 
 
 must be marked 
 
 from each end 
 
 on line AB (Fig. 
 
 143). To get 
 
 these points (1 and 2) a diagonal of the side 
 
 ABCD can be used (as in Ch. XX). The vertical 
 
 edge, AC, being unforeshortened, is first divided 
 
 into six actually equal parts. Lines from the up- 
 
 f- 
 
 B. Pattern for 
 
 MAKINO PLINTH 
 
 16 in. 
 
 Fig. 141 
 
 B. [Plan 
 
 Fig. 149 
 
 EVg LIVCl - 
 
 ii£7 
 
 per and lower division points 
 
 (8 and 9) to VPl transfer these 
 
 divisions proportionately to 
 
 the diagonal, AD. From the 
 
 diagonal they are transferred 
 
 by vertical lines to the edge 
 
 AB. (See also Fig. 144.) 
 Or a diameter through the 
 center (O) 
 
 will divide f^^- i*^ 
 
 AB into perspective halves at 3, when each 
 half can be divided into thirds by the eye. 
 One method can loe used to prove the other. 
 From points 1 and 2 draw lines to VP2. Where they cut a 
 
 diagonal, as AE, will be two corners (4 and 5) of the pyramid. 
 
 89 
 
 Geometric view or 
 
 SIDE piF PLINTH . 
 Fig. 144
 
 FREEHAND PERSPECTIVE 
 
 Lines to VPl through these points will give the two other corners 
 
 (6 and 7), and complete the base of the pyramid. 
 
 The apex of the pyramid will be vertically over the center of 
 
 its base, point O. On a vertical line from O must be measured 
 
 the perspective height of the pyramid. Its 
 actual height is four times that of the plinth. 
 The nearest corner of the plinth is convenient 
 to use, hence from A the vertical line is con- 
 tinued, and four times AC is measured on it, 
 giving AG for the pyramid height as it would 
 appear at that point. If now this height, AGr, 
 could be moved back on the diagonal AE to O 
 it would appear to shorten as moved. Its top 
 would describe an actually horizontal line 
 above AE, that is, a line parallel to it, conse- 
 quently vanishing to the same point. There- 
 fore continue the diagonal AE to the eye level, 
 giving VP3, and draw^ the parallel line from 
 G to VPS, which will mark on the vertical 
 from the desired perspective height at x. 
 Complete the pyramid by drawing its oblique 
 edges to the corners of its base. 
 
 Applications. — The difficulty of making 
 
 Fig. 145 
 
 a church spire or a tower (Fig. 
 145) " stand true " will be readily 
 recognized. The use of the di- 
 agonals (AB and CD) will aid 
 in placing its axis and apex. 
 
 Any upright rectangular ob- 
 ject with equally sloping sides is 
 easily constructed on the same 
 principle, as the grape basket in 
 Fig. 145a. If such an object is 
 leaning, the meeting of its cor- 
 ners continued can still be used 
 as there shown. 
 
 6PAPr-BA«K»T 
 
 Fig. 143a 
 
 90
 
 Chapter XXVI 
 
 THE SQUARE FRAME LEANING ON THE 
 
 RECTANGULAR BLOCK — A PROBLEM 
 
 FOR ORIGINAL STUDY 
 
 THE Models. — These have ah^eady been described on pages 
 48 and 85, respectively. 
 Position — The block rests on one long face, with its 
 square ends making angles of sixty degrees with the picture plane. 
 
 Plak 
 
 Fig. 146 
 
 The frame leans against the block, equidistant from its ends, 
 and with a distance equal to half the width of the block between 
 its lower edge and the block. 
 
 91
 
 Chapter XXFII 
 
 Fig. 1-17 
 
 CYLINDRICAL OBJECTS WHEN NOT 
 VERTICAL 
 
 A LTHOUGrH in Chapter III the cylinder held horizontally 
 /% was mentioned, we have only studied cylindrical objects 
 
 ^ JL when vertical. In this position they have been found 
 
 symmetrical, the ellipses and the axis (which is the middle 
 
 from end to end) being at right angles to each other. To study 
 
 them in other positions begin with the 
 cylinder model held horizontally, with its 
 middle on the eye level, and its ends 
 equally distant (Fig. 147). It will be read- 
 ily seen that the ends now appear at right 
 angles to the axis, as when the object was 
 vertical. Turning it a little so the right 
 
 end can be seen (still keeping it horizontal and at the eye level), 
 
 it will be observed that the apparent directions of the axis and 
 
 ellipses are unchanged. (The axis being 
 
 a horizontal line and at the eye level 
 
 remains apparently horizontal, and the 
 
 ellipses still appear vertical. The further 
 
 ellipse has become a little shorter and 
 
 rounder, and the side boundaries, like all 
 
 parallel receding horizontal lines, appear 
 
 to converge to the eye level.) 
 
 Now lower the model, keeping it turned away, till it rests a 
 
 foot below the eye on some horizontal support (as the box in 
 
 Fig. 149). 
 
 The side boundaries and the axis, being below the eye, vanish 
 
 upward to a point on the level of the eye. They will continue 
 
 92 
 
 Fig. 148
 
 CYLINDRICAL OBJECTS NOT VERTICAi. 
 
 If WJ I I- 
 
 UMmMliliiuif 
 
 to vanish to the eye level, whether below or above the eye (B 
 
 in Fig. 149), as long as the cylinder is kept horizontal. The 
 
 ellipses should now be exam- 
 ined to see if in this position 
 
 they appear, as formerly, to be 
 
 at right angles to the axis. Do 
 
 this first with the head erect 
 
 as usual, looking with care, 
 
 and deciding mentally. Then 
 
 try inclining the head (in this 
 
 case of A, Fig. 149, to the right 
 
 and downward) to bring the 
 
 face in relation to the model 
 
 as it would be if both were 
 
 vertical. Two pencils held in 
 
 the shape of a letter T, held in 
 
 front of the cylinder (as in Fig. 
 
 150) 
 will 
 help 
 
 make sure that the axis of the cylin- 
 der and the long diameters of its 
 ellipses unmistakably appear at right 
 angles to each other. 
 
 To understand how this can be 
 the case, 
 hold the 
 
 cylinder again at the eye level, as 
 
 in Fig. 151, and with a pencil mark 
 
 on it the points (A and B) where 
 
 the side boundaries meet the ends 
 
 of the ellipse. Lower the cylinder 
 
 again (Fig. 152) when it will be seen that these points are not 
 
 now at the ends of the ellipse, and that the line AB is not now 
 
 its long diameter. It has now a new long diameter, CD, at right 
 
 93 
 
 .iiiMim 
 
 LUJlU 
 
 m- 
 
 Fig. 149 
 
 Fig. 151
 
 FREEHAND PERSPECTIVE 
 
 cTBKps. 
 
 angles to the axis in its new apparent direction; and also new 
 side boundary lines from C and D toward the vanishing point. 
 
 As the long diameter is 
 always at right angles to the 
 axis, it must change its position 
 when the object is moved so that 
 its axis appears changed in 
 direction. We see therefore 
 that tlie long diameter is movable. 
 These experiments may be 
 tried with other cylindrical ob- 
 jects, as a tumbler, or the flower pots in the next chapter. The 
 leaning dish in Chapter 
 YIII, and the tilted cover 
 in Chapter X are examples. 
 It will invariably be found 
 that, provided the circular de- 
 tails of a 
 
 Fig. 152 
 
 
 Fig. 1o4 
 
 cylindrical ^^^- ^^^ 
 
 object are actually at right angles to its axis, they 
 
 ivill appear so whatever the position of the object. 
 
 Consequently, cylindrical objects alivays appear 
 
 symmetrical. 
 
 Tests. — A drawing of such an object may 
 be tested by turning it to bring the object in 
 question vertical (Fig. 156, Ch. XXVIII), when 
 errors in symmetry will be more apparent. 
 
 A wheeled vehicle (Fig. 153) is a common 
 illustration of this principle; also a clock (Ch. 
 XXXI) and the round arches in Chapter XXXII. 
 Others will readily occur to the student. 
 Flowers (Fig. 154) are striking examples, and 
 many awkward drawings of flowers are so be- 
 cause drawn in ignorance of this beautiful and 
 simple principle of the symmetry of the cylinder. 
 
 94
 
 Chapter XXVIII 
 A GROUP OF FLOWER POTS 
 
 S with previous examples, drawing this exercise is optional 
 with the student, according to his proficiency. But he 
 should compose a similar group, that is, having in it at 
 
 Fig. 155 
 
 least one cylindrical object not vertical. 
 And he should take especial pains to se- 
 cure the symme- 
 try of such non- 
 vertical objects. 
 
 The illustra- 
 tion (Fig. 156) 
 shows how this 
 symmetry may be 
 tested by turning Fig. ise 
 
 the group. 
 
 95 
 
 A.
 
 Chapter XXIX 
 
 THE CIRCULAR FRAME IN A SQUARE 
 
 FRAME 
 
 T 
 
 HIS is an example of rectangular and cylindrical forms 
 in the same object. The explanation should be carefully- 
 studied, and the exercise drawn unless the student is 
 
 ^ experienced. 
 
 The Circular Frame. — 
 . After the square frame 
 I (Ch. XXIV) is drawn, 
 I look at the model from 
 f the front (Fig. 158), and 
 \ note that the outer sur- 
 f face of the ring touches 
 I the square frame at four 
 I points only — where the 
 i diameters of the square 
 5 cross it (A, B, C, and D). 
 i These diameters, being 
 \ parallel respectively to 
 \ the sides of the square, 
 \ are represented in the 
 I perspective by a vertical 
 t line, and a line vanish- 
 j ing in VPl, both pass- 
 j ing through the true 
 i center O, at crossing of 
 ^ the diagonals. Through 
 these four points the 
 outer edge of the cir- 
 cular frame must pass. 
 
 96 
 
 Fig. 1.57
 
 CIRCULAR FRAME IN A SQUARE FRAME 
 
 
 1 - 
 
 A 
 
 
 -c 
 
 ^G 
 
 s^ 
 
 8- 
 
 V^ 
 
 J, 
 
 
 1 
 
 
 Disregarding at first the opening in it, the circular frame may- 
 be temporarily thought of as a slice from the cylinder, one inch 
 thick, and therefore a very short cylinder, with 
 an axis only one inch long. As placed within 
 the square frame (Figs. 157 and 159) its axis and 
 side boundaries are parallel with the short edges 
 of the square, and its actual centre and that of 
 _ _._.. the latter coincide. 
 
 Hence we may draw its 
 axis from this center (0) 
 to VP2. Its circular outer edge will 
 be seen as an ellipse at right angles to 
 this axis, and passing through the four 
 points A, B, C, and D, previously found. 
 As the short diameters of such ellipses 
 always appear to lie in a line with the 
 axis of 
 the 
 
 Front view. 
 Fig. 158 
 
 Fig. 159 
 
 axis 
 
 ,.,.^ 
 
 Top AS SEEN SY EYE, 
 
 lookino at a . 
 Fig. 160 
 
 the object (Fig. 160), 
 line ivill give the apparent direc- 
 tion of the short diameter of the 
 ellipse. (Here the beginner is 
 advised to turn the paper round, 
 bringing the axis as a vertical 
 line, the better to secure the 
 symmetry of the cylindrical part 
 
 of the model. See Fig. 161.) 
 
 To obtain the length of this short 
 diameter, slightly curved lines, perpen- 
 dicular to the short diameter, are 
 sketched from B to the right and from 
 C to the left, giving x and ij for the 
 extreme front and back of the ellipse. 
 Now (since an ellipse is always symmet- 
 rical) the longest dimension or long diameter of this ellipse is not 
 on the vertical line through A and D (Fig. 159), but is on a line at 
 
 Fig. 161 
 
 97 
 
 S
 
 FREEHAND PERSPECTIVE 
 
 right angles to x ?/, and through its apparent middle from front 
 to hack; making one end fall in front of D, and the other back 
 of A. It will also be a little in front of O, the true center of 
 the circle, and a little longer than A D, but not touching the 
 frame. Mark lightly and accurately this apparent middle and 
 sketch the long diameter through it. Mark the ellipse ends 
 by sketching rounded curves, from A back and from D for- 
 ward, making them symmetrical on the long diameter, and 
 equidistant from the axis. Complete the ellipse by connecting 
 these ends and sides, correcting if necessary, till the ellipse is 
 perfect. 
 
 For the inner ellipse proceed as with the top of the cylinder 
 in Chapter IV, remembering to make its proportions as shown 
 in Fig. 158 ; also that the perspective halves of the short diameter 
 are already found by the true center, O. Or four points for the 
 inner circle (e, e',/and/. Fig. 161a) may be found by the method 
 given for concentric circles in Chapter XX. Thus points E and 
 F are set off their proportionate height on the nearest vertical 
 edge xy, and lines vanished from them to VPl. Where these lines 
 cross AD are e and /, the two points for the inner circle corre- 
 sponding to A and D in the outer circle. For the points on its 
 horizontal diameter (e' and /', corresponding to C and B in 
 the outer circle) vertical lines are 
 drawn from the points (1 and 2) 
 where the vanishing lines cross the 
 diagonal. 
 
 The inner ellipse is then con- 
 structed as was the outer one. ^j^ Fig. leia 
 ' The side boundary lines of the circular frame may now be 
 drawn to VP2, when there will remain only its inner thickness to 
 draw. The edge of this is an inner circle on the back of the 
 frame, actually like the inner one in front. Draw lines (E and F) 
 from the ends of the front inner ellipse to VP2. These may be 
 called side boundary lines of the cylindrical opening ; and their 
 convergence measures the smaller length of the desired back el- 
 lipse, which lies with its ends in these lines as does the front one. 
 
 98 
 
 
 K 
 
 / 
 
 c 
 
 fc 
 
 X' 
 
 
 />« 
 
 
 S 
 
 ^ k 
 
 J 
 
 ^' 
 
 K. 
 
 'y 
 
 
 / 
 
 
 b 

 
 CIRCULAR FRAME IN A SQUARE FRAME 
 
 The round arch (Fig. 162, also Ch. XXXII) is an interesting 
 application of this principle. Errors in drawing these and 
 kindred forms (B in Fig. 162) so common with beginners are 
 easily avoided when these principles are understood. 
 
 
 ^° J- 
 
 A- Showing application of methods. 
 
 B. WRONO. INNER BACK ELUPSE 
 MADt Too 5H0RT, AS IF ACTUALW 
 SMALLER, INSTEAD OF A DUPLICATE. 
 
 Fig. 162 
 
 99
 
 T 
 
 Chapter XXX . 
 A ROUND WINDOW 
 
 HIS exercise (Fig. 163) and the methods of sketching it 
 should be carefully studied, and if necessary it should 
 be drawn. After this the student should sketch a 
 
 similar example 
 k from a building or 
 
 photograph. 
 
 Having drawn 
 ^'' ' the straight-line part 
 
 i i > of the exercise, the 
 
 round window is 
 '^-- " ( ^\ \\^ next to be consid- 
 
 ^ ered. This is actu- 
 
 ally a cylindrical 
 <^P'' • ^ opening in the wall. 
 
 Being above the eye, 
 Affc.., .''^M^''" the circles of thewin- 
 
 ~^ 
 
 dow appear as slant- 
 ing ellipses like the 
 cylinder ends in Fig. 
 .^i, 149 (Ch. XXVII). 
 
 ^''^ ^ To make sure that 
 
 the slant of these 
 ellipses shall agree 
 with the straight-line 
 [ part of the building, 
 
 U the axis of this cylin- 
 
 ,j X drical window is 
 
 '' ^'-h used. This axis is 
 
 S^Ji actually at right an- 
 
 ^^^- ^^^ gles to the- wall, and 
 
 100
 
 A ROUND WINDOW 
 
 is therefore parallel to the lines already vanishing in VPl. 
 
 Hence the apparent middle of the window ellipse (O in Fig. 
 
 164) is first located, and the axis is i 
 
 drawn through it to VPl, extend- 
 ing forward indefinitely. The long 
 
 diameter (AB) is sketched at right 
 
 angles to the axis, and the short 
 
 diameter (CD) is set off on the 
 
 axis line. The ellipse is then drawn 
 
 through these four points. The 
 
 inner ellipses of the window are 
 
 shorter as 
 well as far- 
 ther back 
 than the 
 outer one. 
 
 For the 
 partial ellip- 
 ses of the 
 
 quatrefoil, the actual center of the inner 
 ellipse (1 in Fig. 165) must be marked, 
 and through it a vertical line (EF) and a 
 horizontal one (GH, vanishing in VP2) 
 drawn. The real centers of the quatrefoil 
 circles fall each on one or the other of 
 these lines. Their long diameters (a little 
 in front of these real centers) are parallel 
 
 with the other long diameters of the window. 
 
 This quatrefoil is especially an example of objects which 
 
 should in practice be sketched freehand first, and afterward tested 
 
 by the constructive methods here given. (See Introduction.) 
 
 A. FRONT VIEW 
 
 Fig. 164 
 
 '-^^ 
 
 PERSPecTive 
 Fig. 165 
 
 101
 
 T 
 
 f*^ 
 
 T" 
 
 ) 
 
 7 
 
 f 
 \ 
 
 M 
 
 1 
 
 n 
 
 \ 
 
 1 
 
 3 
 
 , 
 
 % 
 
 1 
 
 
 11 
 
 j 
 
 
 t 
 
 
 s 
 
 
 ;i 
 
 
 i 
 
 
 f 
 
 Chapter XXXI 
 THE CLOCK A PROBLEM 
 
 HIS example (Fig. 166) is given as an aid in rendering, 
 
 though it may be drawn first if desired. 
 
 The Model. — Any clock containing rectangular forms 
 
 — ' - "" '■■■-- and the usual cir- 
 
 cular face will serve 
 as the model. It 
 should be placed 
 above the eye. 
 
 Conditions. — Un- 
 like proJolems in 
 general, this draw- 
 ing may be made 
 from the object, 
 but must be done 
 without assistance. 
 The aim is to test 
 the student's abil- 
 ity to apply the 
 principles taught 
 in the last few 
 chapters. 
 
 Fig. 166 
 
 102
 
 Chapter XXXII 
 THE ARCH 
 
 THESE arches from the cloister of St. Paul's Without 
 Gates, at Rome, also illustrate the symmetry of 
 cylinder, and can be drawn by the same method 
 
 the 
 
 the 
 
 as 
 
 Fig. 167 
 
 the round window in the last chapter. (See Ch. XXX.) They 
 are semi-cylinders and their openings are semicircles (Fig. 169). 
 The semicircles are sketched on a horizontal line (A) which, 
 
 103
 
 FREEHAND PERSPECTIVE 
 
 being above the head in this instance, vanishes downward (Fig. 
 168). The true centers of the semicircles are on this line, and from 
 these centers the axes are 
 drawn, vanishing with 
 other lines to VPl. The 
 joints of the stones form- 
 ing each arch, being lines 
 really tending to meet in 
 its true center, are so 
 drawn in perspective. 
 
 Pointed arches, and 
 other modified forms can 
 be readily drawn on the 
 same principles here 
 used. 
 
 The student should 
 draw this exercise unless experienced, when he should instead 
 select a print of artistic interest, illustrating the same principle, 
 and make from it a careful and expressive sketch. In either 
 case, he should follow his first drawing with another involving 
 the use of this principle, from a building. 
 
 Fig. 168 
 
 Front Viemv 
 Fig. 169 
 
 104
 
 Chapter XXXIII 
 
 INTERIORS 
 
 A ROOM PARALLEL TO THE 
 PICTURE PLANE 
 
 THE Cube as a Model. — With a penknife loosen one face of 
 the cardboard cube, and turn it back or take it off. Place 
 it within a foot of the eye, with the opening parallel to 
 
 
 the picture plane, and the eye level a little less than two thirds 
 of the way up. If desired, the v^indows, doors, rug, and pictures 
 may be marked with a pencil on the inside of the cube. It now 
 serves to illustrate the room shown in Figs. 170 and 172, as the 
 cubes and prism illustrated a house in Chapter XXII. In this 
 room the floor, ceiling, side walls, and all details on their surfaces 
 
 105
 
 FREEHAND PERSPECTIVE 
 
 (as the window, the side door, and the rug) are foreshortened; 
 while the side edges converge to VPl directly in front. The back 
 wall and all surfaces parallel to it (as the end of the table and 
 one side of the stool), being parallel to the picture plane, appear 
 in their true shape. 
 
 Directions. — This example (Fig. 170) should be drawn by the 
 student if a beginner. After this the end of a room, also with its 
 farther wall parallel to the picture plane, should be drawn from 
 memory or invention.^ A hall, a 
 kitchen, a street car, or a piazza w^ill 
 be recognized as especially adapted 
 to such views. 
 
 The Apparent Width of the Sides. 
 — To aid in estimating this, recall the 
 appearance of the most foreshortened 
 side of the cube in Chapter XVI. 
 This estimate may then be tested 
 by pencil measurement of the cube 
 model. 
 
 The Pictures. — Although the pic- 
 ture on the right wall is inclined 
 slightly forward, its sides are still parallel to the picture plane ; 
 and therefore appear in their true direction and shape. With 
 the picture on the back it is not so. Its top is slightly nearer to 
 us than its lower edge ; and must therefore appear longer, mak- 
 ing the sides appear to converge downward (though almost 
 imperceptibly) . To determine the direction of this convergence, 
 revolve it forward on its lower edge in imagination until hori- 
 zontal (Fig. 171). It will at once^,be seen that the sides (A and 
 B) in this position are parallel with the lines already vanishing in 
 
 ^ It may be asked why memory oi- inventive drawings should be advised before study from 
 a room. But in this case we meet a subject of which all have something in memory; and 
 drawing from memory when possible (when one has something remembered) is not only far 
 pleasanter, but much less laborious. The object or place itself presents to a beginner a confus- 
 ing mass of detail, much of it not needed for the drawing. The student must make later many 
 drawings from the place to accumulate knowledge ; but will always do his most free and 
 individual work from this knowledge, not directly from the object. 
 
 106 
 
 Fig. 171
 
 ROOM PARALLEL TO PICTURE PLANE 
 
 VPl. If now the picture should be slowly revolved upward, the 
 converging point for the sides would descend as the picture rises. 
 As the picture is not moved sidewise at all, this point of con- 
 vergence (OVPl) can only move downward in a vertical line 
 from VPl like, the roof ends in Chapter XXII. When the 
 picture is returned to its original position, it varies but slightly 
 from the vertical ; consequently OYPl is too far away to locate 
 and the vanishing of its 
 sides must be estimated. 
 Make sure it is slight 
 enough, and toward a 
 point vertically under 
 VPl. 
 
 An Open Door. — So 
 far we have found but 
 one vanishing point on 
 the eye level. If we 
 begin to open the far- 
 ther door, its horizontal 
 lines will instantly ac- 
 quire a vanishing point, 
 but at an infinite dis- 
 tance. In proportion as the door swings toward being parallel 
 with the sides of the room (that is, with the direction of seeing) , 
 this vanishing point will move inward. When the door becomes 
 quite parallel with the walls of the room, this point will coincide 
 with VPl. If the door is swung still farther back, this point 
 moves on toward the left. The apparent width of the door 
 thus opened may be measured by an ellipse on the floor 
 (Fig. 172) representing the path of its near corner as it swings 
 in a circle. The short diameter of this ellipse is proportioned 
 to the foreshortening of the floor in which it lies. It can 
 also be found by the use of lines parallel to the floor diagonal 
 (Fig. 173). From this diagram it is seen that since the floor is a 
 square, its diagonal is a line at 45° with its sides. By lines at 45°, 
 
 107 
 
 Fig. m
 
 FREEHAND PERSPECTIA^E 
 
 distances into the picture can be measured equal to others paral- 
 lel to the picture plane (Fig. 103, p. 65). In Fig. 173 (and per- 
 spectively in Fig. 172) AC is made equal to AB by drawing BC at 
 
 c 45°, or parallel to the floor diagonal.^ 
 
 I /"'^ I ^^>/^| It will also be noted that the 
 
 thickness edges of the door, being 
 horizontal and at right angles to 
 its top and bottom, have their 
 own vanishing point upon the eye 
 level (VP4). Also the door knob 
 is cylindrical, and its axis is paral- 
 
 ~SSZSZZZZZZZZZZZZZ2ZZZZZA 
 
 A 
 
 iM—-\' 
 
 \^J^ 
 
 
 
 
 
 ,0^ 
 
 6i 
 
 n 
 
 lei to these edges. 
 
 Fio. 173 
 
 The Stool. — The proportion of 
 the stool is found in the same 
 manner as the height of the pyra- 
 mid in Chapter XXV. The room 
 may be considered as nine feet high, and the stool as approxi- 
 mately eighteen inches, or one sixth the height of the room. 
 Mark any point (as H in Fig. 172) where it is desired to place it 
 on the floor of the room. Its height cannot be compared directly 
 with that of the room here, for we cannot determine where a 
 vertical line from H will touch the ceiling. Therefore imagine 
 the stool moved from point H in a straight line to any place on 
 the front edge of the room, as I, where its height (IJ) can be 
 measured by that of the room. If now it were moved back on this 
 same line (IH) , its top would move in a horizontal line directly 
 over IH, that is, actually parallel to it, or vanishing in the same 
 point on the eye level. Both lines (HI and one from J) may 
 therefore be carried to this point VP5. The height of the stool, 
 when placed at any point on the line from 1 to VP5 will be 
 the vertical distance (as at H) between these lines. 
 
 The near side of this stool is now drawn in its true shape, and 
 the parts at right angles to this side found by vanishing lines 
 to VPl. 
 
 * The ellipse is made horizontal, as explained in Chapter XLIII. 
 
 108 , ^
 
 ROOM PARALLEL TO PICTURE PLANE 
 
 Finally, to improve the composition some of the ceiling and 
 a little of the floor are cut oif , as shown by the dotted lines in 
 Fig. 172. This gives a more generally favorable shape (Fig. 170) 
 to the inclosnre, and keeps the oblique lines from running to its 
 corners, which should always be avoided. (See p. 116.)^ 
 
 * See Chapter XLII for fiu-ther coDsideratiou of a room parallel with the picture plane. 
 
 109
 
 Chapter XXXIV 
 
 INTERIORS CONTINUED — A ROOM AT 
 ANGLES TO THE PICTURE PLANE 
 
 THE Model. — The cube model may be prepared for illus- 
 trating this study by removing a side adjacent to the 
 opening made for the previous study. 
 
 Position. — The 
 ■j::=^e-'- v-'^" ^^-^ -s eye level and the dis- 
 
 tance from the eye 
 are the same as in 
 the last chapter. 
 Place the model so 
 that its receding 
 faces make very un- 
 equal angles with the 
 picture plane (B in 
 Fig. 175). Both 
 sides and the top 
 and bottom are now 
 foreshortened and 
 their horizontal lines 
 vanish respectively 
 in VPl and VP2 
 (Fig. 177). 
 
 Selection of Sub- 
 ject, and Use of the 
 Picture Plane. — It will 
 be seen that the room 
 in Fig. 174 is the same as in the previous chapter. The differ- 
 ence is in the selection of subject-space, and in the consequent 
 
 110 
 
 Fig. 174
 
 ROOM AT ANGLES TO PICTURE PLANE 
 
 relation of the subject-matter within that space to the picture 
 plane. Thus in the last study, the extreme points {x and y) of the 
 back wall of the 
 room are equally 
 distant from the 
 picture plane as 
 well as from the 
 eye (A in Fig. 175). 
 Therefore we could 
 not, without ab- 
 surdity, vanish the 
 lines on this wall 
 in either direction 
 (A in Fig. 176), 
 
 ^ Picture 
 
 Plane 
 
 J 
 
 \ z 
 
 
 / 
 
 \ ° 
 
 
 / 
 
 \ \- 
 
 o 
 
 / 
 
 \ ^ 
 
 z 
 
 
 \ ^ 
 
 ■ui , 
 
 / 
 
 \ ^ 
 
 u / 
 •^ / 
 
 
 
 
 
 \ < 
 
 ^ / 
 
 
 \ 1^ 
 
 o / 
 
 
 \ t- 
 
 
 
 
 
 
 
 
 A 
 
 Fig. 175 
 
 still less in both 
 
 (B in Fig. 176). 
 
 The only way to 
 
 give a truthful 
 
 impression of the 
 
 back wall is to draw it in its true shape (C in Fig. 176), as was 
 
 done in the last chapter. 
 
 But in the present ex- 
 ercise (Fig. 171) part of 
 the room is left out. The 
 central direction of seeing 
 is therefore moved to the 
 right, and with it the pic- 
 ture plane is turned (B, 
 Fig. 175). Consequently 
 the wall, xy, which was 
 before parallel to the pic- 
 ture plane, now recedes 
 from it. Hence the height 
 of the room at the corner, 
 being farther into the pic- 
 
 FiG. 176 
 
 111
 
 FREEHAND PERSPECTIVE 
 
 ture and from the picture plane, appears less than at the right 
 and left; and all horizontal lines on both walls appear to 
 converge. 
 
 In this drawing the rectangle incloses the parts of most 
 interest and cuts off the awkward outer lines (Fig. 177). An 
 
 Fig. 177 
 
 alternate selection is indicated by dotted lines. This brings us 
 to the consideration of a new point, namely: — 
 
 What to Include in a Picture. — This cannot be all that it is 
 possible to see from any point, for the head can be turned to 
 see all parts of the horizon circle. Such a view when painted 
 forms a panorama, which is a continuous cylindrical picture 
 surrounding the spectator. 
 
 It is evident that the legitimate picture must include, or cut 
 out from what it is possible to see, only such a space as can be 
 perceived by the eye in a single effort of seeing. It should 
 leave out whatever cannot be seen without turning the head, 
 or even noticeably moving the eyes. It is generally understood 
 
 112
 
 ROOM AT ANGLES TO PICTURE PLANE 
 
 that sixty degrees of the horizon circle is the most that 
 be taken, and that usually thirty degrees is better. The 
 that the greatest dimension of the 
 selected space (whether height or 
 width) shall not exceed the artist's 
 distance from that dimension. 
 
 It will be seen by the diagram 
 (Fig. 178) that this is equivalent 
 to not exceeding sixty degrees in 
 the picture. It is also apparent 
 that it does not prevent the in- 
 clusion of objects nearer than that 
 greatest width if they come into 
 the picture space, as the table and 
 part of the rug in this illustration. 
 
 In conclusion it should be 
 noted : 
 
 First, — Tlie picture plane is dif- 
 ferent for each neiv picture selection. 
 
 Second. — The picture angle 
 should not he over sixty degrees^ and 
 is letter less. 
 
 3. the pa[zt selected fob a picture . 
 Fig. 178 
 
 should 
 rule is 
 
 118
 
 Chapter XXXV 
 FURTHER STUDIES OF INTERIORS 
 
 THE student should copy carefully Fig. 179, always deter- 
 mining the level of the eye, and locating the vanishing 
 points either actually or mentally. Note that the em- 
 phasis of contrast and interest is concentrated on the old- 
 
 
 i ~ ' r<,5\ 
 
 ^ 
 
 114 
 
 Fig. 179
 
 FURTHER STUDIES OF INTERIORS 
 
 fashioned desk while the details on the wall beyond it are very- 
 quiet. The settle comes forward more (that is, has more con- 
 trast and emphasis) than the wall, but 
 less than the desk. The floor is pur- 
 posely quiet in detail to aid in con- 
 centration of interest. 
 
 Fig. 180 
 
 Fig. 181 
 
 (Figs. 180, 181, 
 artistic value of 
 principal sets of 
 
 and 
 
 the 
 
 The chair in the 
 extreme right of 
 the fore-ground 
 is subordinated 
 also. 
 
 Original Work. — Following the 
 above copy several interiors should be 
 drawn from the place. The finder 
 should be used, and thumb-nail sketches 
 182) should be made to test the 
 selection. Avoid equal angles in the 
 
 should be drawn 
 
 vanishing lines. If both 
 
 at forty-five degrees, or nearly that, th^ 
 
 composition would have a stiff effect. 
 
 The contrast between surfaces turned 
 
 away much and others turned away 
 
 little is generally pleasing. Should it 
 be impossible to avoid 
 equal angles of van- 
 ishing, relieve the stiff- 
 ness by using details 
 (as the sofa and win- 
 dow in Fig. 183) that 
 
 are widely different in effect. The rug, 
 being parallel with the sofa, assists further in 
 overcoming the monotony. It is undesir- 
 able to show the floor and ceiling as occupy- 
 ing equal space in the picture; and it is 
 
 usually better not to show both floor and ceiling. (Compare 
 
 Fig. 182 with the same subject in Fig. 181.) A subject may 
 
 115 
 
 Fig. 183
 
 FREEHAND PERSPECTIVE 
 
 Fig. 184 
 
 often be improved by cutting it with a different margin 
 (Fig. 185). 
 
 Care should be taken to place the principal oblique lines 
 of the study in such a relation to the 
 margin or inclosing rectangle that they 
 do not conspicuously point to the picture 
 corners. The rectangle which bounds a 
 composition is an orderly conventional 
 shape, and 
 hence unno- 
 ticeable, leav- 
 ing the atten- 
 tion to be 
 concentrated 
 on the picture. 
 Lines to its corners direct the atten- 
 tion there, and defeat this end. 
 
 Vignetting. — It is not necessary p^^ jg^ 
 
 that an inclosing margin should always be 
 
 used. The drawing 
 may be vignetted, or 
 blended off into the 
 white paper (Fig. 
 186). This is less 
 easy to do as the 
 tendency of the ir- 
 regular outer edge 
 is to contrast sharply 
 with the white paper, 
 detracting from the 
 effect of the more 
 important central 
 parts. The edge details therefore must be carefully " quieted," 
 or rendered inconspicuous. 
 
 Interesting effects are often produced with a partial margin 
 
 116 
 
 
 Fig. 186
 
 FURTHER STUDIES OF INTERIORS 
 
 line (Fig. 187). This is advisable if the center of interest is 
 too near the edge of the paper to blend off well. 
 
 So many principles of perspective and of artistic rendering 
 are included in the drawing of interiors that they form a most 
 important division of the subject. In figure compositions they 
 are constantly used, especially by the illustrator. 
 
 Fig. 187 
 
 117
 
 T 
 
 Chapter XXXVI 
 A CHAIR 
 
 HE chair in Fig. 188, like most chairs is different from 
 the stool in our first interior study (Ch. XXXIII) in 
 that the sides are not parallel to each other (see plan, 
 
 — .„ _^.- :_ Fig. 189). Also 
 
 t the seat is a 
 trifle lower 
 in the back, 
 so that lines 
 B, B (Fig. 190) 
 slope down- 
 ward toward 
 the back. But 
 the chair is 
 symmetrical on 
 its center line 
 (front view, 
 Fig. 189), there- 
 fore the hori- 
 zontal lines in 
 its front and 
 back are par- 
 allel, having a 
 common van- 
 ishing point on 
 the eye level. 
 
 Drawing the 
 Chair. — After 
 
 J having planned 
 
 its place on the 
 
 118 
 
 Fig. 188
 
 A CHAIR 
 
 L I \ 
 
 PLAN OF 5CAT 
 
 paper and located the eye level, the next step is to take the direc- 
 tion of the front and side of the seat (lines A and B, Fig. 190). 
 Leaving line A for the present we carry line B out, obtaining 
 VPl. The feet, being on the horizontal floor, fall in lines vanish- 
 ing to the eye level — on the right to VPl, and on the left in lines 
 E and F. But as the chair sides are not parallel, line E, being 
 turned away from the picture plane more than F, appears steeper, 
 and its vanishing point, VP2, is nearer ; 
 
 than VPS, to which line F vanishes. | 
 
 The near side of the seat, line A, being 
 in the same vertical plane with line E 
 (directly over it) must vanish where it 
 will cut the vanishing trace of this ver- 
 tical plane, which is a^ vertical line 
 through the vanishing point of line E. 
 (Compare the roof, p. 72.) Carrying out 
 line A, therefore, to this vertical trace 
 gives its vanishing point, OVPl, to 
 which we vanish line Gr, which is parallel with A. 
 
 In the same way, line F, vanishing to a point on the eye level 
 a little further away than VP2, gives the location of a vertical 
 trace for 0VP2, in which the other (invisible) side of the seat 
 must vanish. The crossing of these invisible seat edges (H and I) 
 and of the lines C and F aid in shaping the further leg. 
 
 raoNT vitw 
 
 Fig. 189 
 
 Fig. 190 
 
 As the seat is evenly slanted, 0VP2 is on a level with OVPl. 
 
 All edges parallel to the front and back converge to VP2. 
 The curving front of the seat and the curves of the back are 
 drawn so their ends rest on a line to VP2. A rocking chair is 
 
 119
 
 FREEHAND PERSPECTIVE 
 
 simply a chair (A, Fig. 191) placed on rockers (B, Fig. 191). The 
 weight of its back generally tips it backward when at rest, hence 
 
 the usual appearance of a rocking chair is 
 that of an ordinary chair plus its rockers. 
 In B, Fig. 191, lines A and B, which above 
 are actually horizontal, here actually in- 
 cline downward at the back; and must 
 vanish to a point below the eye level (like 
 A, G and H in Fig. 190). The seat and 
 arms are more foreshortened, and the back 
 less erect. But the horizontal lines C, D, 
 E, F and their parallels, remain actually 
 horizontal, and vanish to the same point 
 as in A. The directions of the slanting 
 back, the legs and the arm supports should 
 be carefully taken, remembering that their 
 ends and ornamental details (as G, H and 
 D) fall in lines to VP2. In sketching 
 straight guide lines for the rockers, recall 
 that they are farther apart in front, to agree with the sides of the 
 seat. The lines where the cylindrical rungs enter the legs are 
 actually modified circles, hence they will be seen as shapes modi- 
 fied from the elliptical, and should be carefully considered and 
 drawn. 
 
 Following this, the student should choose and draw at least 
 one other piece of furniture. Selection should be made of 
 something interesting in itself ; that is, well designed and con- 
 structed, and agreeable in association. In chairs the old-fash- 
 ioned rush or splint bottom ones, the wooden rockers of our 
 grandmothers, the beautiful examples by Chippendale, Sheraton 
 and others of the colonial period, and good examples of modern 
 mission shapes may be mentioned as among those satisfactory 
 for study. If one is fortunate enough to get a really fine old 
 cradle, it is a most instructive subject, as is an antique desk or 
 a tall clock. 
 
 Fig. 191 
 
 120
 
 T 
 
 /- 
 
 > 
 
 k 
 
 
 1 
 
 r 
 
 / 
 
 Chapter XXXVII 
 
 THE HEXAGONAL PLINTH IN TWO 
 POSITIONS 
 
 HE student should draw this example from the objects, 
 following the directions here given; and should then 
 sketch them from memory, as by the general directions 
 . in Chapter XI. 
 The Model. ^ 
 
 This can be 
 
 made from card- 
 board according 
 
 to the diagram 
 
 (Fig. 193). Tor , ' \.__.^.,_.,„__^^' 
 
 the first posi- 
 tion it should 
 
 be placed about 
 
 three feet from ' '^^^ 
 
 the eye and 
 
 nine inches be- 
 low it, and with 
 
 two vertical \ ^' 
 
 faces parallel 
 
 with the picture 
 
 plane. 
 
 The Greomet- ,, , ^ , 
 
 ric Hexagon. — ] " . | \/ 
 
 This should be 
 
 constructed 
 
 first. Divide 
 
 the line AB 
 
 (Fig. 194) in \ --^^.^___^_._ ^ I 
 
 halves, drawing f,^ 192 
 
 121
 
 FREEHAND PERSPECTIVE 
 
 ■•^ h -^ 
 
 Fig. 193 
 
 a perpendicular at the point of division. Measure the distance 
 AB, taken on the pencil, from B against the perpendicular. 
 
 Where it falls (at O) will be one corner 
 of an equilateral triangle (AOB) which 
 will form one sixth of the desired hexa- 
 gon. Sketch vertical lines of indefinite 
 length from A and B, and cut them by 
 continuing the sides of the equilateral 
 triangle (AO and 
 BO) to D and C, 
 then draw DC. The 
 constructive rec- 
 tangle ABCD, which 
 we have now com- 
 pleted, will be al- 
 ways essential in 
 drawing the perspective of the hexagon. 
 Its diagonals (A and B) will form two di- 
 agonals of the hexagon, and its center (0) 
 will be the center of the hexagon. The 
 other diagonal, EF, is drawn through O parallel to the rectangle 
 ends, half on each side of the center. Its length should be tested 
 by that of the diagonals already found. The other four sides 
 complete the geometric hexagon ABCDEF. 
 
 Drawing the Hexagonal Top. — The rectangle ABCD (Fig. 195) 
 is drawn first. Although it is actually nearly twice as long as 
 its width it will be found to appear less than half as long. The 
 diagonals of the rectangle will in perspective, as in the geo- 
 metric view, form two diagonals of the hexagon. The other 
 diagonal, EF, is set off on a line of indefinite length through 
 their crossing. Looking at the diagram, it is observed that the 
 sides of the rectangle (AC and BD) and its middle (0) divide 
 this diagonal (EF) into four equal parts at x, and y. In the 
 perspective drawing (Fig. 195), the two middle fourths (xO and 
 Oy) are seen to be already measured. Since these fourths are 
 
 122 
 
 Fig. 194
 
 HEXAGONAL PLINTH IN TWO POSITIONS 
 
 all equally distant from the picture plane, they appear equal, and 
 are so set off from x and y. The hexagonal top is completed by 
 drawing its last four sides. 
 
 The Thickness of the Plinth. — The front face of this thick- 
 ness, being parallel to the picture plane, is drawn in its true 
 shape. The lower edges of the other two faces are parallel to 
 the receding horizontal edges (AE and BE) above them, and 
 will therefore vanish with these edges respectively to VP2^and 
 VPS. Vertical lines downward from E and F will complete the 
 plinth. 
 
 VPl 
 
 VPI 
 
 fVf LEl/EL 
 
 Fig. 195 
 
 A Test for Vanishing Lines. — The other two diagonals and 
 the receding back edges of the hexagon are in reality each 
 parallel to one or the other of the two sets of vanishing lines 
 just drawn. Therefore when carried out to the eye level, 
 they should meet respectively in VPI and VP2 if the drawing 
 is correct. 
 
 The Hexagonal Plinth Slightly Turned. — For this drawing turn 
 the model so its front face will make an angle of thirty degrees 
 with the picture plane (B in Fig. 196). Draw the constructive 
 rectangle, ABCD ; taking with the pencil the direction of the front 
 edge and of the imaginary left side (AC in Fig. 196)^ of the rec- 
 tangle. Take especial pains to have these lines correct in direction 
 
 1 Corresponding to the receding edges (2, 2) first drawn in sketching the cube. 
 
 123
 
 FREEHAND PERSPECTIVE 
 
 and length, as an error here causes a particularly unpleasant 
 representation. In vanishing the back edge with the front one be 
 careful to keep it also tending upward. 
 
 CYE LEVEL 
 
 iPlOTVRtl PlONC 
 R- PI-AM or HEXAGON 8£EN IN A. 
 
 Fig. 196 
 
 A. Wrono. Lines I, 2, 3 
 
 AND 4 DO NOT CONVERGE 
 TO THE EYE LEVEL. 
 
 The vanishing point (VP2) of these front and back edges 
 is so far away that their convergence will naturally be only 
 
 jestimated. Hence the necessity of locating 
 definitely in mind what cannot he seen — 
 that is, the vanishing point, not only of 
 these two lines, but of their parallels, 
 the diagonal EF and the lower edge of 
 the front face. 
 
 In this position the last diagonal, EF, 
 is in perspective, and therefore its four 
 actually equal divisions will appear de- 
 creasing in size, or perspectively equal. 
 Consequently, if the work so far done is 
 correct it will be found that the space 
 xO is (almost imperceptibly) greater than 
 Oy^ because a little nearer. Hence E:c should be set off a little 
 larger than xO, and y¥ a little smaller than Oy. Similar cases, 
 as the cylinder top in Chapter IV, and the concentric circles in 
 Chapter XX are readily recalled. 
 
 Testing the Drawing. — The test used for the other drawing 
 of the plinth is equally effective here. It is more needed here, 
 
 124 
 
 B. Wrong-, lines J, 2,3 
 
 AND 4 ARE TOO &TEE.P. 
 
 hexagon appears tiltep. 
 Fig. 197
 
 HEXAGONAL PLINTH IN TWO POSITIONS 
 
 since the vanishing point for one set of lines has not been 
 actually found. But if these lines have been carefully thought 
 out as to direction, the errors will be found encouragingly 
 slight. 
 
 The test of placing the eye at the vanishing point (Ch. XVH), 
 to sight back along the lines which should converge to them, is 
 especially applicable here. 
 
 125
 
 T 
 
 
 <mM»"g "'""' « " >'" ' 
 
 ni rm 
 
 INTERIOR WITH A TILED FLOOR 
 
 HE plan of this floor is shown in Fig. 199. As drawn in 
 the example, the foreshortening of the tiles is proportioned 
 to that of parallel surfaces, such as the receding " treads " 
 
 or tops of the steps. 
 From the vanish- 
 ing of their edges, 
 we may judge 
 these treads to be 
 foreshortened about 
 one half. Conse- 
 quently lines paral- 
 lel to the staircase 
 edges (vanishing in 
 VPl) must be fore- 
 shortened as much. 
 The tile adjacent 
 to the lowest step 
 is the best to begin 
 with (since it is the 
 same distance into 
 the picture). As with 
 the previous hex- 
 agons the rectangle 
 ABCD is drawn 
 first. And since 
 much depends on 
 the correctness of 
 this first rectangle, 
 it is worth while to 
 
 ■iW ■jatnuttf iirii<».j 
 
 J 
 
 Fig. 198 
 
 126
 
 INTERIOR WITH A TILED FLOOR 
 
 take especial pains with it. Observe that as the tiles are here 
 placed, it is the width (AD) of this rectangle which is 
 parallel with the receding staircase edges. 
 Thus these tiles are three and a half inches 
 on their edges, making the rectangle width 
 actually half the height of a seven-inch step. In 
 the drawing this width, being foreshortened as 
 
 much as the 
 
 steps, ap- 
 pears only 
 one fourth 
 as wide as 
 the height of the step. The 
 actual length of the rectangle 
 is seen by the diagram to be 
 one eighth less than twice its 
 
 Fig. 199 
 
 Fig. 200 
 
 width; or (what is the same thing) one eighth less than the 
 
 height of the step. Being slightly turned e^pView of^tep 
 
 away, it will appear a very little shorter in 
 
 comparison than that. In this case it was 
 
 made one sixth less than the height of the 
 
 step. 
 
 It will be readily seen how if the rectangle 
 proportions are right, the lines of this first 
 hexagon, when carried out forward and back, will give points for 
 the other tiles, making them fall harmoniously into their proper 
 perspective. 
 
 Fig. 201 
 
 127
 
 Chapter XXXIX 
 THE HEXAGONAL PRISM AND FRAME 
 
 T 
 
 HIS exercise may be drawn from the objects, if they are 
 at hand. If they cannot readily be had, the drawing 
 juay be made from these directions, using the cardboard 
 
 plinth made for 
 Chapter XXXVII. 
 The objects should 
 then be drawn 
 from memory. 
 
 The Prism. — Be- 
 ing the simpler to 
 draw, this object 
 should be taken 
 first, though it 
 should be placed 
 at the bottom of 
 the sheet, on ac- 
 count of its greater 
 horizontal dimen- 
 sions. 
 
 This model is 
 eight inches long, 
 and the diameter 
 of its hexagonal 
 bases is four inches. 
 It is placed so that 
 the bases make 
 angles of sixty de- 
 grees with the pic- 
 ture plane. The 
 
 Fig. 202 
 
 128
 
 THE HEXAGONAL PRISM AND FRAME 
 
 Fig. 203 
 
 dotted lines in Fig. 203 show how the cardboard plinth may be 
 placed in the same position as a help in study. 
 
 The Nearest Vertical Hexagonal Base. — Sketch the construc- 
 tive rectangle previously used, noting that its width is fore- 
 shortened as much as the most foreshort- 
 ened side of the cube in Chapter XVII. 
 Set off the third diagonal of the hexagon 
 (EF) perspectively on a line vanishing to 
 VPl through the rectangle center (as on 
 page 124). 
 
 The Long Edges of the Prism. — Take 
 the direction of the nearest upper edge, giving VP2. Vanish 
 the other long edges with it, and set off on the one first drawn 
 its apparent length (AGr). This can be easily estimated by re- 
 calling the cube. That is, its actual length as given is twice that 
 of the diameter of its base (the near vertical line AB). AGr will 
 therefore be as long as two cubes placed side to side.^ 
 
 The Further Base. — For the horizontal top edge of the further 
 base draw a line parallel to the same line in the near one (that 
 is, vanishing in VPl), which gives GH. For the corner corre- 
 sponding to B in the near base drop a vertical 
 from Gr, giving J. The upper oblique edge 
 (GI) is parallel to AF in the near base, and 
 their vanishing point is OVPl, vertically 
 above VPl. Another oblique line, from the 
 nearest point, I, to the lowest, J, completes 
 the prism. 
 
 The Hexagonal Frame. — This model is three 
 inches on a side, and is one inch square in sec- 
 tion (Fig. 204). It stands on one rectangular 
 face with its hexagonal faces at an angle of 
 thirty degrees with the picture plane (Fig. 205) . For the outer 
 hexagon and the outer thickness proceed as in the prism. 
 
 Face view or fpame 
 
 Section of fi?ame 
 Fig. 204 
 
 XVI. 
 
 This method is given ia a slightly different form under Solutions of Problems, Chapter 
 9 129
 
 FREEHAND PERSPECTIVE 
 
 For the inner hexagon we may first study the actual shape in 
 Fig. 204, where it is seen that the vertical frame thickness can 
 be conveniently carried across to measure it at C^ on the 
 nearest vertical, BC. We know this thickness to be one inch, 
 which is more than one sixth and less than one fifth of the 
 
 vertical CB. If one fifth of CB (CO in 
 Fig. 205) be found, and three quarters of 
 this {Cy) be taken, it will serve the purpose. 
 Mark the same distance from B up (point 
 2). From these points {y and b) vanish the 
 horizontal lines of this inner hexagon with 
 their parallels to VPl. The corners of the 
 inner hexagon are on the diagonals of the 
 outer one, so the crossings of the diagonals 
 by these two vanishing lines give four corners of the inner 
 hexagon (1, 2, 3, and 4). From 2 an oblique line vanishes down- 
 ward with its parallels to 0VP2, marking point 5 on the hori- 
 zontal diagonal EF. Another from 3, also vanishing in 0VP2, 
 is drawn from 3 upward to cut EF in point 6. Lines from 6 to 
 1 and from 2 to 5, complete the inner hexagon ; and should, if 
 the drawing is correct, vanish with those parallel to them in 
 OVPl. 
 
 The inner edges of the thickness which are visible vanish 
 from 4 and 5 to VP2, being parallel to the outer thickness edges. 
 For the visible part of the further inner hexagon, a line of the 
 near inner hexagon, as 4-3, may be carried " around the corner " 
 and back, as in the square frame (Ch. XXIV). From the point 
 where this vanishing line (7-8) cuts the line from 4 to VP2, an 
 edge (8-9) vanishes with its parallels (CF and others) to OVPl. 
 This point is even further away than 0VP2, so that the con- 
 vergence of its vanishing lines must be slighter. Where line 8-9 
 crosses the one from 5 to VP2 it meets the last visible edge of 
 this back inner hexagon, a line vanishing in 0VP2. 
 
 130
 
 Chapter XL 
 
 THE TRIANGULAR PRISM AND FRAME — 
 PROBLEM FOR ORIGINAL STUDY 
 
 THE MODELS. — The prism is eight inches long, and its 
 triangular ends are four inches on a side. The frame is 
 six inches on a side, and one inch square in section. 
 See diagrams. Fig. 206. 
 
 A TEIANQULAR FACE 
 OF FPAME 
 
 B. 5IC)E VIEW 
 OF FPAME 
 
 Positions. — The objects are 
 placed four feet away, and one 
 below the eye, and are separated, 
 as were the models in the last 
 chapter. The prism rests on 
 one long face, with its long edges 
 making angles of thirty degrees 
 with the picture plane. The 
 frame rests on a rectangular face 
 with its triangular faces at thirty 
 degrees with the picture plane. 
 
 Arrangement on the Sheet. — The drawings are to be placed on 
 one sheet. The position of the paper (whether with its long 
 edges horizontal or vertical) and the placing of the drawings 
 on the sheet must be such as to produce the most agreeable and 
 satisfactory effect. 
 
 C END OF PPISM 
 
 d 51 de view of pei3m 
 Fig. 206 
 
 131
 
 Chapter XLI 
 THE STUDY OF PARALLEL PERSPECTIVE 
 
 F 
 
 Fig. 207 
 
 ROM Chapters XXXIV and XXXV it is seen that inte- 
 riors follow the law of the cube. This, however, leads to 
 what may seem an inconsistency. Why, it may be asked, 
 
 does the table in Fig. 207 differ from 
 the cube in Fig. 208^ In the cube 
 B was made shorter than A because 
 farther into the picture. But in 
 the table B was not drawn shorter 
 than A. 
 
 The answer to this is that the 
 table was not studied 
 alone, as was the cube. 
 It was part of a pic- 
 ture in which the dominant part (the back of the 
 room) was parallel to the picture plane. Having 
 drawn the side of the table parallel to the sides 
 
 of the room, it is absurd (Fig. 209) 
 to draw its end otherwise than 
 parallel with the back of the room. 
 Fig. 207 satisfies the eye and gives 
 a true impression of the room and 
 its contents. Could the room be 
 erased, leaving the table alone, it 
 would present the error shown in 
 Fig. 210. Here it forms the whole 
 picture and its picture plane makes 
 an angle with its ends (see plan in Fig. 210), hence it must be 
 drawn as below. 
 
 132 
 
 Fig. 208 
 
 THE TABLE I5WB0N0 
 
 Fig. 209
 
 STUDY OF PARALLEL PERSPECTIVE 
 
 COCTSECTION 
 
 Fig. 210 
 
 It is undeniable that in Fig. 207 B is farther from the eye 
 than A. But since drawing that corner smaller produces the 
 false impression seen in Fig. 
 209, we are guided by the dis- 
 tance of these points not from 
 the eye, but from the picture 
 plane. The picture plane simply 
 forms the best means of attain- 
 ing our fundamental object — a 
 truthful representation. Hence 
 the necessity of determining 
 the limits of the picture (Fig. 
 211) and of clearly fixing in mind reSviSn orp™E plane 
 
 TO TABLE. WHEN TABLE 
 
 the central direction of seeing '^ ^'-°''^ 
 and the picture plane. 
 
 Under some conditions, sur- 
 faces may even be drawn in 
 their true shape when not quite 
 
 parallel to the picture plane. In Fig. 211 
 none of the vertical surfaces, as A, B, and C, 
 are exactly parallel with the picture plane. 
 In Fig. 212 this is shown by the conver- 
 gence of the lines at right angles to these 
 surfaces ; their vanishing point being a little 
 out of the center of the picture. Yet if 
 all these vertical surfaces are drawn in per- 
 spective (A in Fig. 212) the result is mis- 
 leading or impossible, and the eye protests. 
 But the drawing is perfectly satisfactory in 
 B, Fig. 212. 
 
 A convincing illustration of dominant 
 surfaces parallel to the picture plane is the 
 familiar form of a bureau. With an un- 
 broken top (A in Fig. 213) it is easily drawn like the book 
 and cube. If now the middle of the upper drawer is cut out, the 
 
 133 
 
 'i^''^^/'^//////'/. 
 
 PLAN OF 
 INTERIOR 
 jHOWN IN 
 
 GuRE ais.. 
 
 Fig. 211
 
 FREEHAND PERSPECTIVE 
 
 A UntpuE drawing 
 
 OF VIEW FROM X, IN P(./\N 
 
 B CORRECT DRAWING 
 or TME SAME VIEW. 
 
 Fig. 212 
 
 remaining small ones are seen to occupy positions similar to that 
 of the table in Fig. 1. 
 
 The Street. — The street is another example of these con- 
 ditions. Viewed from the middle of a crosswalk {x in plan, 
 
 Fig. 214) the fronts 
 of the houses pre- 
 sent to the beholder 
 a perspective like 
 that of the interior 
 in Fig. 207. They 
 vanish to the center 
 of the picture, and 
 surfaces at right 
 angles to them are 
 drawn in their true 
 shape. This is done 
 even if the con- 
 vergence is not 
 toward the exact middle of the picture (Fig. 215), provided it does 
 
 not fall to right or left of the house fronts. 
 
 In this view, although the beholder has 
 passed on to y, the conditions are still 
 like those of the interior in Fig. 211. 
 
 But if instead of using both sides 
 of the street for our picture, we choose 
 one of the corners, the picture plane for 
 this forms a different angle with the 
 principal surfaces (Fig. 214), and the view 
 must be drawn as shown in Fig. 216. 
 
 It may therefore be concluded, that 
 171 any picture having a domhiant imrt 
 parallel with the picture plane and conse- 
 quently clraivn in its true shape^ all por- 
 tions of that picture which are parallel tvith the picture plane must 
 also he drawn in their true shape. 
 
 134 
 
 i^^^crr 
 
 ■<p 
 
 .-4P- 
 
 # J ^^ 
 
 Fig. 213
 
 STUDY OF PARALLEL PERSPECTIVE 
 
 Also, even such dominant parts as are not quite parallel 
 with the picture plane must 
 be drawn in their true shape 
 in certain cases ivliere draiving 
 them in perspective produces 
 false or misleading results. 
 Finally, it is of great import- 
 ance to include in the pic- 
 ture only what can easily he 
 ^een. 
 
 Parallel perspective, as ■^^^^ 
 work under such conditions 
 is called, involves no depart- 
 ure in principle from free- 
 hand perspective in general. 
 It is merely an adaptation 
 of perspective methods to 
 certain conditions in the 
 subject.^ 
 
 Space has been given here 
 to a somewhat extended con- 
 sideration of the subject. 
 
 Plan OF the 
 
 ^TEEET 6HOWN 
 IN PEESPECTIVE 
 
 Fig. 214 
 
 
 -Ti 
 
 VIEW or JTEEET FBOM Y. 
 
 Fig. 215 
 1 The terms " parallel " and " angular " perspective, though used for lack of better ones, 
 
 are therefore far from satisfactory. 
 
 135
 
 FREEHAND PERSPECTIVE 
 
 because the confusion concerning it that frequently exists is 
 deemed unnecessary. It has been found that students may be 
 easily led to distinguish when such conditions are present, after 
 which there is no difficulty in dealing with them. 
 
 view of coenee fbom y 
 Fig. 216 
 
 136
 
 Chapter XLII 
 
 A STREET FROM THE PHOTOGRAPH 
 
 THIS exercise (Fig. 217) may be drawn first if judged best, 
 noting carefully the changes made in rendering from 
 the photograph shown in Fig. 218. The student should 
 then select a print 
 
 street and 
 
 drawing 
 
 of a 
 make a 
 
 from it. All 
 sketches should 
 be thoroughly | 
 thought out, hav- 
 ing the level of 
 the eye carefully 
 placed, and all the 
 vanishing points 
 located, either 
 actually or men- 
 tally. It will 
 probably be neces- 
 sary to correct 
 some distortions 
 of the camera (see 
 Ch. XLIII). 
 
 This drawing 
 should be followed 
 by studies from the 
 street and from 
 memory. Notes 
 and sketches made 
 
 137 
 
 Fig. 217
 
 A STREET FROM THE PHOTOGRAPH 
 
 by the student at the place chosen may be used to help this 
 memory work. 
 
 Fig. 218 
 Entrance to the Ponte Vecchio, Florence 
 
 188
 
 Chapter XLIII 
 
 EXCEPTIONS TO THE USE OF THE FLAT 
 PICTURE PLANE 
 
 IT will be observed that in photographs the circular tops of 
 columns near the edges of the picture often appear as slant- 
 ing ellipses (Fig. 219). And all who have an acquaintance 
 with mechanical 
 perspective will 
 recall that in cer- 
 tain problems the 
 ellipses of cylin- 
 ders do not work 
 out at right an- 
 gles to the axis 
 (Fig. 220). While 
 the eye sees ob- 
 jects pictured on 
 the inside of the 
 spherical eyeball, 
 the camera forms 
 its pictures, and 
 mechanical per- 
 spective projects its problems on a flat surface. Therefore the 
 camera cannot wholly reproduce objects as seen by the eye, 
 and certain results obtained by mechanical perspective are 
 untrue representations.^ 
 
 ^ The photographic error has been recognized, and a camera is now made in which a 
 clockwork attachment brings each part of the plate in turn directly facing the part of the 
 subject it is to receive, and gives horizontal ellipses to columns wherever placed in the 
 picture. 
 
 139 
 
 Fig. 219 
 Cloisters of the Monastery of San Martino, Naples
 
 FREEHAND PERSPECTIVE 
 
 As for mechanical perspective, though useful in many cases, 
 it has sometimes obscured the real aim of representative drawing. 
 
 It has even been taught that the 
 flat picture plane should be used 
 for all representative work as in 
 mechanical perspective, logically 
 to the end, regardless of any pro- 
 test of the eye as to its results. 
 To this error it is sufficient reply 
 to say that the aim of freehand 
 
 
 - 
 
 ^^=- 
 
 ^ 
 
 — 
 
 
 
 . r 
 
 1 I 
 
 c1l 
 
 f| 
 
 ^ 
 
 1 
 
 1 
 
 
 
 f 
 
 1 
 
 t" 
 
 Jit* 
 
 ^ 
 
 
 Cylinders as found by mechanical perspective, 
 Side cvlindebs vntsue as REPREseivTATiONa. 
 
 Fig. 220 
 
 perspective is the draiving of objects 
 as they appear; and that the eye never sees a column as in Fig. 
 219, nor a cylinder as the outer 
 
 ones in Fig. 220. When, there- 
 fore, the use of the flat picture 
 plane produces an untrue draw- 
 ing, it is evident that an excep- 
 tion must be made in that case. 
 
 The Cylindrical Picture Plane. 
 — Looking at Fig. 220, we find 
 that the middle cylinder, which 
 does appear right to the eye, 
 extends equally each side of the 
 central direction of seeing, so 
 that the picture plane is parallel 
 to the apparent breadth of the 
 cylinder. By drawing the other 
 cylinders as if each had such a 
 central direction of seeing and 
 such a picture plane of its 
 own (A, Fig. 221) a result is 
 obtained that appears true to 
 the eye (B, Fig. 221). 
 
 In other words, cylindrical 
 
 A. Plan 
 Showing 
 
 Vi>E OF5PEC- 
 JAL PICTURE 
 PLANES- THE 
 EQUIVALENT OF 
 THE CylINDRi- 
 CAU PICTURE. 
 PLANE . 
 
 EYE 
 
 objects, however placed, should be 
 
 B. Appearancp of group at a, drawn by 
 
 \}i>e OF special -THATiS, CYHNDRlCflU, PLANED 
 
 FiQ. 221 
 
 140
 
 EXCEPTIONS TO FLAT PICTURE PLANE 
 
 ^ " Side viev/ 
 
 ' Showing the u4E or 
 
 PICTURE PLANES INCLINED 
 FROM THE VERTICAL-THftr 15, 
 
 THE Sphericau picti;re pw\n&. 
 Fig. 222 
 
 drawn as if for those objects alone, the picture plane was bent or rolled 
 into a cylindrical picture plane. But this does not apply to the 
 straight-line portions of the 
 picture (as the block in Figs. 
 220 and 221), nor to the plac- 
 ing of the cylindrical parts, nor 
 to their height. These must 
 be determined in the ordinary 
 way, by using the flat picture 
 plane. We onlj^ abandon the 
 flat picture plane where we 
 cannot otherwise produce a 
 representation which the eye 
 will accept as true. 
 
 The Spherical Picture Plane. — Another exception occurs in a 
 vertical direction. Thus, the only outline that will truthfully 
 
 represent a ball to the eye is a 
 circle. To obtain that, we must 
 regard its special central direction 
 of seeing as directed to its middle, 
 not only from side to side (as in 
 case of the cylinder), but from 
 top to bottom also. If the ball is 
 above or below the eye therefore, 
 its special picture plane is slanted 
 accordingly (Fig. 222). In this 
 case the picture plane (again /or 
 such ohjecfs alone) may be called 
 a Spherical picture plane. 
 
 An example of its application 
 is the case of a model posed 
 higher than the student who is 
 drawing (Fig. 223). The head is 
 foreshortened vertically, and the forehead appears smaller in pro- 
 portion than the lower and nearer features. At the same time the 
 
 141 
 
 •''^•^iDI§J^ 
 
 Fig. 223
 
 FREEHAND PERSPECTIVE 
 
 window beyond that model is drawn on the usual flat picture 
 plane; that is, with its vertical lines vertical, as always. 
 
 These distinctions will be found not only necessary, but 
 natural and easy to make ; ^ especially if care is taken to include 
 in the picture space only what the eye can see ivithout noticeably 
 moving the eyehaUs (Ch. XLI). The picture plane should be 
 regarded as limited to what will cover this selected space, and we 
 have no concern with what lies outside of that. 
 
 When working from a photograph therefore, as must often be 
 done, such camera distortions as the columns in Fig. 219 should 
 be corrected to agree with what is pictured by the eye. And in 
 freehand work only such truths of mechanical perspective should 
 be used as produce results which the eye confirms as true repre- 
 sentations. Where the eye and a train of reasoning are in 
 conflict, the reasoning should be scanned for errors. Unless a 
 drawing looks right, it may safely be pronounced not right. It 
 may look right, and still be wrong ; but if the eye refuses to be 
 satisfied, it is certainly wrong. 
 
 1 So natural and easy, in fact, that space for this explanation is hardly needed, except to 
 guard against false reasoning in the subject. 
 
 142
 
 Chapter XLIV 
 SHADOWS 
 
 WHILE it is unnecessary for the mastery of freehand 
 sketching to study this subject exhaustively, there 
 are a few simple facts which have been found funda- 
 mentally useful in practice, and which may be easily understood. 
 
 Fig. 221 
 
 To that end the student should follow these explanations 
 carefully, making experiments and sketches as needed. He 
 should then compose and draw a group similar to Fig. 224, 
 also should make other studies involving the use of the truths 
 here developed. 
 
 Light may be regarded as composed of an infinite number 
 of rays. From a lamp they extend outward in all directions, 
 
 143
 
 FREEHAND PERSPECTIVE 
 
 forming what may be called a sphere of light. The shadow of 
 
 the apple on the right of the lamp in Fig. 225 extends toward the 
 
 right; that of the book 
 on the left in an almost 
 opposite direction. The 
 sun, on the contrary, is 
 so much larger than the 
 earth, and its rays Lave 
 traveled such an incon- 
 ceivable distance, that to 
 us they are parallel, as 
 are the paths of fall- 
 Fi«- 225 ing raindrops. Fig. 226 
 
 illustrates this familiar truth. (This, of course, is also true of 
 
 moonlight.) There are therefore two classes, of shadows: those 
 
 cast by the sun, and those 
 
 produced by near light, as a 
 
 lamp. Under those formed 
 
 by the sun may be studied 
 
 first: 
 
 Small Objects in a Room. — 
 
 If a shadow box be placed 
 
 near and a little back of 
 
 the window,^ as shown in 
 
 Fig. 227, the shadow edge 
 
 (A-6) cast by the vertical 
 
 box edge AB will lie on the 
 
 floor of the box in a line actually parallel to that of the vertical 
 
 hat-pin. The shadow of a vertical vase (Fig. 228) also casts a 
 
 shadow in the same direction. (That is, line C, the shadow of 
 
 its vertical axis will be parallel with the shadows of the vertical 
 
 line AB on the box floor, and on the horizontal book cover.) The 
 
 1 In this case the window is larger than the box ; so that as far as the box is concerned the 
 rays of hght are parallel. As will be seen later in this chapter, the diffused light from a win- 
 dow causes radiating shadows in the room itself. 
 
 144 
 
 Fig. 226
 
 SHADOWS 
 
 shadow (F^^) of the vertical book corner (FD) will be parallel 
 with these lines. If we push the box back or forward 
 they all change direc- 
 
 EYE LEVEL 
 
 tion, becoming more 
 nearly parallel with 
 the picture plane 
 as the box moves 
 forward, and vanish- 
 ing more steeply if 
 we put it further 
 back of the window. 
 But they always Fig. 227 
 
 remain actually parallel to each other. 
 
 In the same way we see in Fig. 227 that the shadow of the 
 horizontal edge BH and of a horizontal hat-pin (EF) are actually 
 parallel to each other, on both the back and the floor of the 
 box. Also in Fig. 228 the shadow (b-e) falling on the hori- 
 zontal book cov- 
 er from the hori- 
 zontal box edge 
 (BE) is parallel 
 to the shadow 
 (d-g) falling on 
 the horizontal 
 box surface from 
 the horizontal 
 line Da. 
 
 Wemaythere- 
 FiG. 228 fore say that the 
 
 shadows' of actually parallel lines ivill lie actually parallel to each 
 other on the same surface or on surfaces parallel with each other. 
 
 But in perspective, parallel lines vanish; and if they are 
 
 horizontal lines they vanish to the level of the eye. We should 
 
 therefore expect these parallel shadows to also vanish thus, and we 
 
 find they do vanish, in the same manner as any lines or objects. 
 
 10 145
 
 FREEHAND PERSPECTIVE 
 
 Fig. 229 
 
 We next observe (in Fig. 228) that the vertical vase casts 
 a vertical shadow on the vertical back of the box. Then we 
 recall that the horizontal edges BE and DGr cast on horizontal 
 
 surfaces horizontal 
 shadow edges h-e and 
 d-g. As these edges 
 vanish, their shadows 
 vanish with them to 
 the same point, as any 
 parallel lines would. 
 
 It thus appears that 
 when the receiving sur- 
 face is pm'allel to the 
 object or line casting the 
 shadow, the shadow will also he parallel to the object or line. 
 
 This brings us to consider how to find the extent of shadows. 
 If the shadow box is lowered from its usual place on the table 
 to the floor, the shadows will be found shorter (Fig. 229). The 
 light, falling more steeply, cuts off the shadows nearer the 
 objects. When the box is lifted back to the table (Fig. 228) the 
 shadows will be seen to lengthen. 
 With a light ruler or '' straight- 
 edge " (Fig. 230), take the actual 
 direction from the hat-pin top (D) 
 to its shadow (d) on the floor of 
 the box. Keeping the ruler in the 
 same actual direction, move it to 
 the left till it grazes the top corner (B) of the box. It will be 
 found also to mark the shadow (h) of point B. 
 
 The evident truth is that the direction of the light-ray from any 
 point in the ohject 7narhs the same point in its shadow. Therefore 
 to find the shadow of any point, as of the other hat-pin head 
 (E, Fig. 227) we have only to draw the light-ray from E to where 
 it strikes the receiving surface. 
 
 But since in this case the window is nearer than the box, the 
 
 146 
 
 Fig. 230
 
 SHADOWS 
 
 light-rays are receding slightly, hence must appear to converge 
 a little, like any parallel receding lines. Therefore to draw them, 
 we first find their vanishing point. Imagine one of these rays 
 (as D-^, Fig, 227) dropped vertically to the floor of the box (as 
 the gable edge in Ch. XXII was dropped). It would then lie 
 in the horizontal shadow line (C-^) directly under it, and would 
 vanish in VPS. When lifted again to its former oblique position 
 (D-^) its vanishing point would have moved down in a vertical line 
 from VPS, and become OVPl. All other light-rays in this illustra- 
 tion (as Bb and E-e) appear to converge to this vanishing point. 
 
 Hotv to find where the light ray mid the receiving surface 
 meet is the iiext consideration. In the 
 case of D-^ we had a vertical line, DC, 
 and its shadow, C-^, cut by the light-ray, 
 T>-d. From E we can imagine a vertical 
 line, similar to DC, dropped to the floor 
 of the box. (We can find where this 
 vertical line will touch the floor by a vertical from F to the box 
 edge MJ at point K, and a vanishing line from VPl through 
 that point, K, to cut the vertical from E, in L.) From L a line 
 actually parallel to the shadow Q-d (vanishing in VPS) cuts the 
 light-ray in the desired point, e. This is essentially the way most 
 shadow points are found: — hy a vertical line from the point on the 
 object to the receiving surface^^an^ from that a shadow line on the 
 receiving surface to cut the light-ray. In other words we pass an 
 imaginary vertical plane through the point and the light-ray. 
 Fig. 2S1 shows a simple application of these principles. When 
 the cube has been drawn, the direction and length of the shadow 
 edge Ba maybe assumed (or taken if drawing from the object). 
 This gives the direction of the light-ray A-a. The shadow a-c 
 vanishes with AC till cut by the light-ray from C, and c-d 
 vanishes with CD. These points can be tested by vanishing the 
 light-rays from the other corners to OVPl, thus completing the 
 several vertical planes as above mentioned. 
 
 The shadow of the left horizontal box edge (BH) falls partly 
 
 147
 
 FREEHAND PERSPECTIVE 
 
 ou the back of the box in a slantmg Hue which may be thus deter- 
 mined. Take the book out (Fig. 227), when it will be seen that 
 the near part of the shadow, beginning at h, vanishes on the floor of 
 the box to yPl till it reaches the box edge in point i. The shadow 
 must start at H, hence H-i is the line in question. Many shadow 
 lines can be found thus, — 1)1) locating any two points in the line. 
 
 Looking again at the shadow 
 of the vase on the back of the 
 box, we observe that the shadow 
 of its horizontal circular top f all- 
 
 ^.■5HOWiNOr A 
 Veeticau plane 
 at eioht anoue5 to 
 
 CENTER or^hAPOW 
 
 Perspective 
 
 or ABOVE. 
 
 FACE VIEW OF 
 VERTICAL PART 
 OF 5HAPOW. 
 
 Fig. 232 Fig. 233 
 
 ing on the back of the box is not a horizontal curve. To 
 understand this, we will begin with the shadow of any vertical 
 cylindrical object on a horizontal surface, as in Fig. 232. It will 
 be found actually symmetrical. In perspective it will be fore- 
 shortened (B in Fig. 232) ; and lines marking its horizontal details 
 (as AB, CD, and EF) will vanish, as would any parallel'horizontal 
 lines. If now we move this object toward a vertical surface, 
 placed so that, vieived from above it mahes right angles ivith the light 
 rays (as shown in A, Fig. 233), the shadoiv on the vertical surface 
 
 148
 
 SHADOWS 
 
 A. JnowiNG- 
 VERTICAL Plane. 
 
 ATUNEQOAU ANGLES 
 TO 5MAD0\A/ CENTEE 
 
 B Pei?spe:ctive of above 
 — shadow distorted 
 
 Fig. 234 
 
 also ivill he actuaUy symmetrical; though it may appear fore- 
 shortened (B in Fig. 233), as in this case. 
 Now if the receiving surface be 
 
 turned so it is not at right angles 
 
 (viewed from above) with the 
 
 light (Fig. 234), we get what we 
 
 observed in Fig. 228, — an actu- 
 ally one-sided shadow. The reason 
 
 for this distortion is made clear 
 
 from the plan in Fig. 234. The 
 
 descending light-ray from y has 
 
 farther to travel before striking 
 
 the receiving surface, hence its 
 
 shadow, y, is lower than the 
 
 shadow from z. Such variations 
 
 of original shapes are of the same 
 
 nature in producing beauty (and 
 
 consequent enjoyment) as theme 
 
 variations in music. Thus the bottle with shoulders (Fig. 235) 
 
 acquires a charm from the proximity of its interestingly altered 
 
 shadow-self which it cannot have alone. 
 
 For the shadows of curves vertical 
 ' planes through several points (as x, y, 
 and z) are taken and the curve then 
 sketched freehand. The use of this 
 method is also shown in drawing the 
 horizontal hat-pin in Fig. 227 and the 
 the vase shadow in Fig. 228. But as 
 soon as the underlying truths are 
 ^^^- ^^"^ clearly understood, the actual taking 
 
 of points is seldom needed. 
 
 So far the shadows have fallen on flat surfaces, but the 
 
 shadow of the vertical box edge in Fig. 228 falls partly on the 
 
 curved book back ; and on this it forms a vertical curve — that is, 
 
 with its ends in a vertical line. The curve is sketched by the 
 
 149
 
 FREEHAND PERSPECTIVE 
 
 eye, though it could be constructed by points. In Fig. 229 the 
 shadow on the book back is cast by a horizontal edge and is 
 
 therefore an oblique curve. In this 
 case its upper point (M) is found by 
 imagining the book cover continued 
 until it cuts the back of the box in a 
 line from N vanishing with the long 
 box edges. Where the shadow line 
 HI cuts this line (point o) will be 
 the farther end of the shadow line on the book cover. This line 
 will vanish in VPl, and where it cuts the upper edge of the book 
 back will be M, the upper end of the curve. 
 
 The book in Fig. 236 shows the use of points when the edge 
 casting the shadoiv is itself oblique. A vertical line from D to the
 
 SHADOWS 
 
 A. Plan, J MOWING 
 
 T?At)IATINQ RAVJ. 
 
 (%P^ POINT O, THE 
 V""^ CrAS TLAMC 
 
 edge 1-2, and a line from that point (E) to C, constructs one 
 vertical plane. The light-ray from point A cuts the shadow- 
 direction of AB at a. The shadow of CD travels from C through 
 a to the edge 1-2, 
 and from there to 
 D. 
 
 Shadows on a 
 House. — The truths 
 thus developed ap- 
 ply to out-of-door 
 work, as shown in 
 the house (Fig. 237). 
 Here one new con- 
 dition is met, — the 
 shadow of the ver- 
 tical dormer edge 
 falls on the oblique 
 surface of the roof, 
 and hence has an 
 oblique vanishing 
 point. This vanish- 
 ing point is easily 
 found, as we already 
 have two points in 
 the vanishing trace 
 of theroof, — OVPl 
 and VP2. The line 
 containing these 
 points is the vanishing trace, not only of the roof, hid of the infi- 
 nite plane containing the roof. (It therefore can be drawn as long 
 as needed, being really infinite in length.) So we have only to 
 draw the trace from OVPl to VP2, and mark 0VP3 on it verti- 
 cally over YP3, exactly as we marked OVPl over VPl. 
 
 The shadow of the bush is an instance of the ease with which 
 shadow laws are applied to natural objects. The shadow is sketched 
 
 151 
 
 Perspective 
 
 OF ABOVE 
 
 Fig. 238
 
 FREEHAND PERSPECTIVE 
 
 freehand ; but with much greater certainty for knowing that its 
 center must fall on the ground in the direction of VPS, and that it 
 can extend no farther than its meeting with the ray of light, HI. 
 Shadows from a Lamp. — The radiating rays from an artificial 
 light can all be contained in an infinite number of radiating ver- 
 tical planes through the light itself. Some of these radiating 
 planes are seen in the plan (OA, OB and others in Fig. 238). 
 
 These radiating 
 planes are used 
 instead of the 
 parallel vertical 
 planes previously 
 explained. Other- 
 wise the methods 
 are the same as 
 with light from 
 
 the sun. Thus in 
 the shadow of the 
 stool the light-ray 
 from O through 
 D gives d where 
 cut by a line on 
 the ground di- 
 rectly under it 
 ^'^- ~^^ (from through 
 
 E). The shadow of Gh falls on the floor in the direction of 
 o-H till it reaches the wall. On the vertical wall, the shadow 
 of the vertical GrH is also vertical. It is ended by the light-ray 
 from O through Gr. The shadow of the edge GI will be parallel 
 to it, and like it will appear as a vanishing line to VPl. The 
 near part (d-j) of the shadow of DI will be parallel to DI, and 
 will vanish to VPl till it reaches the wall at J. A line from J 
 to I completes the shadow of DI. 
 
 Shadows in an Interior. — These are partly like the lamp 
 shadows. For instance the shadow on the couch in Fig. 239 
 
 152
 
 SHADOWS 
 
 extends in an almost opposite direction from that of the chair. 
 These shadows are produced by the diffused dayhght radiating 
 from the window. On the other hand a patch of sunlight falling 
 through the window would follow the laws of sunlight generally. 
 The edges a-l) and c-d vanish with AB and CD, while points a 
 and & are marked by the meeting of light-rays from A and B 
 with shadows of verticals from A and B. In this case the light 
 comes from beyond the window, hence the light-rays recede up, 
 and appear to converge or vanish in that direction. 
 
 153
 
 Chapter XLV 
 OUT-OF-DOORS WORK 
 
 A LTHOUGH the same perspective principles apply to out- 
 /\ of-doors work the conditions of the study vary, and some 
 -A. ■■^ cases need explanation. 
 
 Vanishing Points. — In drawing the house (Ch. XXII), we 
 placed ourselves proportionately in relation to the small cube 
 as we should naturally be in relation to the real house. Thus 
 the sixteen inches of distance from the eye, or four times the 
 
 51OE VIEW or i-iousr SMOwiMG two positions op eye 
 
 Fig. 240 
 
 height of the cube, was equivalent to only four times the height 
 of a twenty-foot house, or eighty feet — less than five rods. At 
 this short distance the vanishing of the lines is very decided ; but 
 at a half mile from the same house, those lines appear nearly 
 horizontal. The reason for it is seen in Fig. 240. When the eye 
 is near the house (at x) the apparent difference in length between 
 the edges AB and CD is greater than when the eye is at y. 
 This is shown on picture plane 1 by al) and ccl, and on picture 
 plane 2 by a'l) and c'cT. The horizontal edges of the house 
 as seen by the eye from x would therefore vanish more steeply, 
 causing the vanishing point to fall nearer, as shown in A, Fig. 
 241. As seen from y, the horizontal edges are less steep ; there- 
 fore in B the vanishing points fall much farther away. 
 
 It follows, therefore, that the greater the distance of the eye 
 
 154
 
 OUT-OF-DOORS WORK 
 
 from an object , the farther to right and left will the vanishing 
 points fall. When the house is a half mile away they fall so far 
 to left and right that its horizontal lines appear almost level. 
 Hence the beginner in landscape work, accustomed only to near 
 objects, is sometimes puzzled, because distant houses seem to 
 have no perspective. And the landscape artist who has " no 
 
 .aJ^^^.,, 
 
 SHAPE OF HOUSE AS 
 5CEN FROM X 
 
 B SHAPE OF HOUSE AS 
 
 SEEN reoM Y, 
 
 Fig. 241 
 
 trouble with houses " in the distance may shrink from attempting 
 them in the foreground. 
 
 Size of Objects Seen. — The image formed on the retina of the 
 eye is always exceedingly small, and with distant objects becomes 
 microscopic. All mental picturing of the size of objects pro- 
 ceeds from our mental knowledge of their actual dimensions. 
 Size judged from seeing alone can be but a matter of comparison. 
 This is easily proved by asking two persons how large the moon 
 appears to them. Here we have an object whose real size and 
 distance are so great as to be no guide in comparison with other 
 objects and it will probably appear of a different size to each 
 person. There is consequently no such thing as the drawing 
 of objects '' the size they appear." Size in draiving is merely 
 relative; and the scale on which a drawing is made is ivholJy a 
 matter of choice. We may choose to make a drawing what is 
 termed " actual size," but this means that we regulate its size by 
 a mental knowledge obtained either from measuring in the ordi- 
 nary way, or by putting our sketch back by the object to compare 
 them by the eye. 
 
 The absolute size of objects varies so much, also, that unless 
 the picture contains something the size of which is well known 
 
 155
 
 FREEHAND PERSPECTIVE 
 
 and but little variable, we cannot be sure of the sizes repre- 
 sented. The human fig- 
 ure serves best for such 
 a standard, but some 
 objects always adjusted 
 to the human figure in 
 size, as steps, and often 
 doors, will answer in its 
 place. 
 
 TJie size of objects ac- 
 cording to their distance 
 iiito the picture is impor- 
 tant in out-of-doors work 
 also. Here the indis- 
 pensable picture plane 
 becomes again useful. 
 In Fig. 242, for instance, 
 the gondolier must not 
 be too large for the 
 buildings. Lines drawn 
 from his head and a 
 Fig. 242 poiut ou the watcr di- 
 
 rectly under it to the eye level will contain between them his 
 heisfht above the water 
 
 all the way to their van- 
 ishing point. If we 
 wish to know, for in- 
 stance, whether the door 
 on the left is large 
 enough, we have only 
 to draw horizontal lines 
 from its top and from a 
 point directly under it on 
 the plane of the water ^^^- -^^ 
 
 continued. Where this water line would cut the ^ater line from 
 
 156
 
 OUT-OF-DOORS WORK 
 
 the figure to the vanishing point a vertical line is erected, on 
 which the two heights can be compared. 
 
 Reflections. — If a mirror be 
 laid on a table, and a cup placed 
 on it (Fig. 243), the reflection 
 will appear precisely like the 
 cup reversed, with its bottom 
 resting against the bottom of 
 the real cup. The reflection 
 will not present to the eye the 
 same shape as the real cup, 
 for besides being reversed it is 
 farther below the eye, making 
 its inside invisible while its 
 base is covered by that of the 
 real cup. We can also see 
 farther around on its flaring 
 
 -7-x 
 
 A. Perspective- 
 
 SHOVl/lNG STAKE 
 FORESHORTENED 
 
 B. Side view of stake - showing how 
 its foreshortening- occurs. 
 
 Fig. 244 
 
 surface, because its decrease of diameter is toward the eye level, 
 while in the actual object it is away from the eye level. 
 
 Now since it is like the cup reversed we see that any point 
 (as A) in the cup, must be reflected directly under itself. So a 
 stake, thrust into a pool of still water (Fig. 244), will produce 
 a reflection like itself reversed ; ^ and each point in the reflec- 
 tion will be directly under the same point in the real stake. 
 
 ^ Tn this case appearing longer than the real stake, as explained a few pages later. 
 
 157
 
 FREEHAND PERSPECTIVE 
 
 It is therefore evident that in case of reflections on a horizon- 
 tal surface J the image formed must he vertically under the reality. 
 
 Consequently, as long as the 
 reflecting surface remains 
 horizontal, reflections on it 
 cannot be thrown to one 
 side, but must be shown di- 
 rectly under the real objects, 
 even if the reflecting surface 
 be broken (as in Fig. 242). 
 And if in drawing reflections 
 we represent them out of 
 the vertical (Fig. 245) the 
 reflecting surface (in this 
 case the water) appears to be 
 sloping, like rapids in a river. 
 We may therefore take as 
 our rule that reflections are 
 invariably like the reflected 
 object reversed on the reflecting 
 plane. 
 
 When the Object is Sep- 
 arated from the Reflecting 
 Surface. — In Fig. 246 the bungalow is separated from the reflecting 
 
 WRONG. Reflections not ver- 
 tically UNDER OBJECTS RE- 
 FLECTED. WATER APPEARS 
 
 sloping- not leueu. 
 
 Fig. 245 
 
 Ihls ctistonce (E f) further into the 
 picture thar\ tb>e bocJ'Wou3e., 
 
 Fig. 216 
 158
 
 OUT-OF-DOORS WORK 
 
 surface (the water) by a high bank. But by using a water line 
 of the boathouse (which stands parallel to it and directly on the 
 water), the points (A, B, and C) where the bungalow edges 
 continued would strike the plane of the water can be closely 
 approximated. From there the points (a, b, c, and d) for the 
 reflection of the bungalow are measured vertically. 
 
 Even if we had not the boathouse to give parallel vanishing lines 
 on the water, the necessary points (A, B, and C) could be esti- 
 mated with sufficient accuracy after a little experience. The main 
 thing to remember is that it is on the reflecting surface or 07i 
 its plane continued that the object is reversed in its reflection. 
 
 Reflections on Vertical Surfaces. — With reflections on vertical 
 surfaces the problem is very simple. In Fig. 247 the box appears 
 reversed as far back 
 of the mirror surface 
 (the thickness of its 
 frame) as it actually 
 stands in front of it. 
 
 Length of the Re- 
 flection. — The verti- 
 cal length of the re- 
 flection, while the 
 reflecting surface is 
 unbroken (as in Fig. 
 243) is actually the 
 same as that of the real subject. This does not mean that the 
 reflection will always appear of the same vertical length as the 
 object, as that depends on its position and on the location of 
 the point from which it is viewed. In Fig. 244 the stake is 
 seen from a higher point and leans toward the beholder. It is 
 consequently seen foreshortened, as the roundness of its top 
 indicates. The reflection, being reversed, appears practically in 
 its true length. A point (x) on the surface of the water directly 
 under the top, appears lower than where the stake enters the 
 water, because nearer the eye. Cases like the familiar " silver 
 
 159 
 
 Fig. 247
 
 FREEHAND PERSPECTIVE 
 
 path " of the moon in rippUng water, or like Fig. 242, where 
 the reflections of upright objects appear lengthened vertically 
 as well as broken, are caused by the many curved surfaces of the 
 waves on which successive bits of the reflection fall. 
 
 Use of the Finder. — Nowhere will the finder (Ch. YIII) be 
 of more use than in out-of-doors work. The difference in distance 
 between the near and far objects in a landscape is so great, 
 that the beginner finds it hard to realize how much difference 
 he must make in size. The finder serves as a measuring unit 
 for these differences, besides being invaluable as an aid in 
 selection. 
 
 160
 
 SOLUTIONS OF PROBLEMS 
 
 T 
 
 CHAPTER XI — PAGE 34 
 
 THE CYLINDER CONE AND BALL 
 
 HE Cone. — After the cylinder is drawn, the base of the cone is 
 next placed. This is actually a circle, and of the same size as 
 the cylinder base. Its position will be clear from the plan (Fig. 
 
 249, p. 34). In per- 
 spective it is best 
 placed by its true 
 center. If the cone 
 were moved on the 
 ground around and 
 touching the cylin- 
 der, this center 
 would describe a 
 circle, twice the 
 diameter of the 
 cylinder base, and 
 equally distant at 
 every point from 
 the cylinder. This 
 circle is sketched in 
 perspective (Fig. 
 250) as an ellipse 
 (see Chs. IV and 
 XX). The true 
 center for the base 
 of the cone is placed 
 on this ellipse (at 
 0). From this 
 true center a ver- 
 
 11 
 
 :^^^s^'r:^ *:rjtoL«fSs2t!i3»p;«;2s^«*sss*i:.n-ts: 
 
 ►•«-^«Mt»ii^>Sti^jifo:&2iS»itjK«a<ii;S>;iC-,i?ia««iSi;^^ 
 
 ZXSajJUSVl.flt.gX.is:^., 
 
 
 Fig. 248 
 
 161
 
 FREEHAND PERSPECTIVE 
 
 Fig. 249 
 
 tical line of indefinite length may be erected, on which to set off the 
 axis of the cone. Its height, being actually the same as that of 
 the cylinder, will appear slightly greater because nearer the eye. At the 
 same time its apex (F), being nearer, cannot appear 
 quite so high on the paper as even the nearest edge 
 of the cylinder top.^ Through its lower end (0) the 
 real diameter of its circular base passes. Being at 
 the same distance into the picture as the axis, and 
 like it parallel with the picture plane, it appears 
 in its true proportion to the axis (one half). It is 
 therefore so set off, equally on each side of 0, 
 giving AB. If right, this true diameter should measure a little greater 
 than the diameter of the cylinder. The base of the cone, though actually 
 of the same size as the cylinder, will ajyjiear both 
 a little larger and a little rounder, because nearer. 
 The short diameter of this base (CD) is therefore 
 set off greater than that of the cylinder base, 
 remembering that as is the real center, DO 
 must be larger than CO. The long diameter (the 
 light line in front of AB) is next drawn, exactly 
 in the middle of CD and a very little longer than 
 AB. The ellipse is then sketched through the 
 
 four points A, B, C, and D, taking care to have it touch the base 
 of the cylinder, and to make the greatest length not on AB, but on 
 the long diameter. The cone is completed by drawing its sides from 
 the apex tangentially to this elliptical base. 
 
 The Ball. — If the ball be rolled about and touching the cylinder it 
 will follow the same path as the center of the cone base, so that its 
 resting point will always be somewhere in the ellipse representing that 
 path (its center being always vertically above the resting point). We 
 should therefore mark some point in the large ellipse (in this case x") 
 for the resting point of the ball. If we stoop to bring the eye nearly 
 
 ^ r>eiiig actually of the same height, they lie in the same horizontal plane. This plane, 
 being below the level of the eye, appears to recede upward, as the table does. This will be 
 better understood if a sheet of paper is laid on the tops of the two objects, when it can be seen 
 that it appears to recede upward. The following chapter will further illustrate this truth. 
 
 162 
 
 Fig. 250
 
 SOLUTIONS OF PROBLEMS 
 
 to the table level, and look at the ball, we shall see it resting on this 
 spot. But if we return to the point from which the group was to be 
 viewed, we shall find this point hidden by the projecting mass of the 
 ball. The circle which represents the boundary of the ball is therefore 
 drawn with its lower edge a little below x, and its center vertically over 
 that point. Being nearer the eye, its diameter is greater than AB. 
 
 CHAPTER XV 
 
 THE CYLINDER AND THE RECTANGULAR BLOCK 
 
 PAGE 48 
 
 The Block. — This should be drawn first 
 parallel with the pic- 
 ture plane, it will 
 appear in its true 
 shape, and the block 
 ends will vanish in 
 VPl like the book 
 ends in Chapter XH. 
 In setting off the 
 apparent width of 
 the top, we remem- 
 ber that it is actually 
 narrower in propor- 
 tion than the book 
 cover. 
 
 The Cylinder. — 
 The cylinder rests 
 against this block 
 (side view, Fig. 252), 
 so we can measure 
 the height of its back 
 (AB, Fig. 253) actu- 
 ally, making it twice 
 the height of the block 
 
 Since its front face is 
 
 Fig. 251 
 
 163
 
 FREEHAND PERSPECTIVE 
 
 51DE VIEW or GCOUP. 
 
 Fig. 252 
 
 front. The lower base is actually the same in width as the block, but be- 
 ing nearer the eye it will appear larger. Just how much can be easily 
 determined. Draw the invisible lines of the block (the dotted lines in 
 Fig. 253), and carry the di- 
 agonal of one half (x) for- 
 ward in a line of indefinite 
 length. Cut this by line y 
 of the invisible edge of the 
 block. From the point (C) 
 so found draw line s to the 
 right, to cut another invisi- 
 ble edge continued. This constructs another 
 rectangle the actual size of the side of the block, 
 but nearer, hence appearing larger.* The middle 
 of the front of this rectangle (point D) will be 
 the front of the base of the cylinder. Its back 
 will be A, and its long diameter, EF, can be set off on a line sketched 
 half way between this front and back, marking E and F half way 
 between AD and the ends of the construction rectangle. Through these 
 four points the bottom ellipse is drawn. 
 
 For the top ellipse a similar rectangle can be constructed directly 
 above it. This will give a much more foreshortened ellipse, as would be 
 expected. The back and front of the middle ellipse are drawn as 
 directed in Chapter IV. 
 
 Fig. 253 
 
 CHAPTER XXVI — PAGE 91 
 
 THE SQUARE FRAME LEANING ON A REC- 
 TANGULAR BLOCK 
 
 The Rectangular Block. — This solid is drawn in the same manner 
 as the book similarly placed. Recall that it is equal to two cubes, as 
 shown in Fig. 255, where a diagonal (AB) of the horizontal top face 
 
 1 This use of the diagonal for measuring will be found in Chapters XXII, XXV, and 
 
 others. 
 
 164
 
 SOLUTIONS OF PROBLEMS 
 
 of the first cube is carried to its vanishing point VPS, from which a 
 perspectively parallel line is drawn back to C, cutting the back edge 
 
 Fig. 254 
 
 in D, which becomes the farthest corner 
 of the second cube, and of the whole block. v${^, 
 A line through D, vanishing in VPl gives 
 the corner E. 
 
 The Frame. — For this the loiver face 
 only may be first considered. This face 
 rests on AE in points one eighth of AE 
 from either end, which can be easily found 
 by the method used on page 79 for the steps. (The vertical edge 
 AH is divided into eighths, and lines from points 1 and 2 vanished 
 to VP2. Where they cross the diagonal Al verticals are erected, 
 marking points F and G.) 
 
 Now by placing a card against one of the side thicknesses of the 
 
 165 
 
 PERSPECTIVE VIEW. 
 THE LI IVES ABAND CD 
 
 are actually parallel 
 am0van16hin vp3. _. / 
 
 Fig. 255
 
 FREEHAND PERSPECTIVE 
 
 Fig. 256 
 
 frame. 
 
 frame (Fig. 255) we readily see that these thicknesses are vertical, and 
 that their vertical planes (one of which is indicated by the card) 
 
 are parallel to the block ends, and cut 
 the block side A E I H in vertical lines 
 (FJ and GK, already drawn), and also cut 
 the ground in horizontal lines through J 
 and K, parallel to the horizontal lines of 
 the block ends. In these horizontal ground 
 lines rest the lower corners of the leaning 
 We therefore vanish the lines through J and K to VPl, 
 continuing them forward indefinitely. The distance JL (Fig. 257) 
 is actually equal to half the block width ; that is, to J3, and can 
 be easily measured by a line through J parallel to H3 (vanishing 
 with it to VP4). From where this line meets the continuation of the 
 edge HN (in point 4) a line parallel to HI (vanishing in VP2) gives 
 the corners L and M, and the ground edge of the frame. 
 
 The leaning edges of the square are drawn from L and M through 
 F and G, and will be found to vanish in OVPl. They vanish a little less 
 (have a more distant vanishing point) than the horizontal edge LM, and 
 should therefore be made slightly longer, but shorter than if standing 
 erect from L. We can check our estimate by comparison with LO, 
 the height of the frame at that point (EH). To obtain this, the height 
 of the block at J is carried for- 
 ward to P on the vertical (by 
 lines vanisliing in VPl) and one 
 half of LP added at the top, 
 making LO. 
 
 The Thickness Edges. — 
 These edges must vanish sharp- 
 ly (or have a near vanishing 
 point) because the edges at 
 right angles to them (the ones 
 to OVPl) are foreshortened but 
 little. Hence 0VP2 is placed 
 but little below the group. To this point these short edges are drawn, 
 
 IGG 
 
 -X -TiVPi: — -^ 
 
 T^^HSi-* 
 
 Fig. 257
 
 
 Fig. 258 
 
 SOLUTIONS OF PROBLEMS 
 
 carrying them forward of the corners indefinitely for a short distance. 
 
 The foreshortening of these edges (actually one sixth of the long edges of 
 
 the square) will be much greater than that of the 
 
 long edges ; and a corner is marked accordingly 
 
 at Q. From this corner a horizontal edge van- 
 ishes to VP2, and a long oblique one to OVPl, 
 
 giving respectively R and S where each crosses 
 
 a thickness edge. From S the other horizontal 
 
 edge vanishes to VP2 and from R the other long oblique one to OVPl, 
 
 completing the square. 
 
 The Inner Square. — This is drawn as on page 86. 
 The frame width can be measured perspectively on 
 LT by continuing the edge TS to cut the con- 
 tinuation of LO in point 6, dividing L5 into six 
 actually equal parts and vanishing lines to 0VP2 
 
 THl^LiN^E'l'^HreHll°na as? froui the first and last of these division points, eriv- 
 
 VANISH IN bvP* '^RE HERE SEEN ■«• ^ <=> 
 
 To«.AeTUAuuy.«.RALLx.. -^^^ Q ^^^ ^^ ^j^^ ^^^^^^ nccded. 
 
 Fig. 259 
 
 CHAPTER XXXI 
 PAGE 102 
 
 THE CLOCK 
 
 The dial of the clock is expressed 
 by the methods used for the circu- 
 lar frame in the square frame (Ch. 
 XIX). After the rectangular por- 
 tion of the clock has been sketched, 
 an enclosing square for the dial (AB 
 CD) is drawn and the clock axis 
 vanished from the actual center of 
 this square. The axis is continued 
 forward indefinitely (upward to left) 
 to give the direction of the short di- 
 ameter (which always apjjears to lie in line with the axis, 
 actually at right angles to it — p. 97, Fig. 160) . For the ends (G 
 
 1G7 
 
 Fig. 260 
 
 though 
 andH)
 
 FREEHAND PERSPECTIVE 
 
 of the short diameter, slightly curved lines are sketched up to the right 
 from E and in reverse direction from F, giving the front and back of the 
 ellipse. The long diameter is then sketched, of an indefinite length, 
 midway between G and H and at right angles to the axis. It is then cut 
 by curves from I to the right, and from J to the left, being careful to 
 keep K and L equally distant from the axis. The remainder of the el- 
 lipse may then be drawn, correcting the curves as needed ; and even 
 moving K and L if necessary for symmetry, but not changing points E, 
 F, I or J. The inner ellipse is drawn in the same way, as explained on 
 page 98. 
 
 CHAPTER XL -PAGE 131 
 
 THE TRIANGULAR PRISM AND FRAME 
 
 The Triangular Frame. 
 
 — This solid is easily con- 
 structed by the use of the 
 cube (see dotted lines). 
 The length of AC is 
 found by the diagonal 
 (ED) of one side of this 
 imaginary cube, DF being 
 made equal to DG (p. 
 165). The steps in Chap- 
 ter XXII illustrate this 
 method. The end view 
 shows how the triangle 
 is related in shape to the 
 square face of the cube. 
 Its vertical center line is 
 located by the diagonals 
 of the square, and the 
 height of its apex is meas- 
 ured to X on the near verti- 
 cal edge of the cube (AE). 
 The Triangular Frame. — I 
 
 This is also readily drawn Fig. 261 
 
 168
 
 SOLUTIONS OF PROBLEMS 
 
 by the help of one face of the cube. After the triangular outline is 
 sketched, the height of one lower bar (one sixth of the height of the cube) 
 is marked upward from C, giving point E, and the lower edge of the 
 inner triangle is vanished through E. Where BF, drawn so as to divide 
 AD perspectively, crosses this lower edge is a corner of the inner triangle. 
 
 Fig. 262 
 
 Fig. 263 
 
 Through this corner (G) another edge of the inner triangle vanishes to 
 OVPl. The other edge is drawn toward 0VP2. The thicknesses are 
 found as in the square frame (Ch. XXIV). 
 
 169
 
 INDEX 
 
 Aims of perspective, drawing objects as 
 they appear, 140 
 learning to see, xii 
 
 to acquire artistic judgment, 80, 83, 
 120 
 Apparent size of objects, according to dis- 
 tance, xi, 156 
 in out-of-door work, 155 
 in relation to other parts of picture, 
 
 108 
 relative only, 6, 155 
 Arch, errors in drawing, 99 
 
 pointed and other forms, 104 
 round, 99, 103-104 
 Artistic judgment, 80, 83, 120 
 Artistic rendering, book, 59 
 buildings, 83 
 glass, 25, 28, 30 
 rose jar, 17 
 
 Background, subordination of objects in, 
 fan, 33 
 
 leaning bowl, 26 
 
 plate, 30 
 Bases of cylindrical objects, see Foot 
 
 always partly visible, 47 
 
 location on horizontal surfaces, 29, 30, 
 32, 161-162 
 Baskets, 90 
 Benefit of perspective study, xi 
 
 acquiring of artistic judgment, 80, 83, 
 120 
 
 learning to see correctly, xii 
 Book, artistic rendering of, 59 
 
 at angles to picture plane, 58-60 
 
 back of, 44 
 
 clasps of, 44 
 
 cover thickness, 44 
 
 in two positions, 43, 44 
 
 margins, 40, 41 
 
 projection of covers, 44 
 
 use of pencils to show convergence of 
 lines, 37 
 
 use of strings, 38, 39 
 
 vertical edges, 60 
 
 with back parallel to face, 38-42 
 with cylindrical object, 46-47 
 
 Books, two, at different angles to the pic- 
 ture plane, 61, 62 
 with a cylindrical object, 67, 68 
 
 Boundary, movable, 16, 17, 21 
 tangential to ellipses, 13, 17 
 
 Buildings, camera distortions in, 139, 142 
 few vanishing points for, 84 
 from photograph or print, 81 
 house, the, 69-80; see House 
 round arch, the, 99, 103 
 round window, the, 100, 101 
 spire or tower, 90 
 type forms useful in, 85, 88 
 
 Camera distortions, 139, 142 
 Carrying lines "around a corner," 86, 130 
 Central direction of seeing, alluded to, 10, 
 140 
 explained, 6 
 
 moves with changed picture center, 46 
 Chair, the study of, 118-120 
 Circle, actual center of, 63; see Ellipse 
 concentric circles, 14, 63-66 
 location of its center in ellipse, 15, 65 
 only position in which seen as circle, 
 
 xii 
 see obliquely, 9 
 Circular frame within square frame, 96-99 
 application of its principles, 99; see 
 Cylindrical objects not vertical 
 Clock, 102, 167 
 Color, in buildings, 83 
 of book, 59 
 on rose jar, 17 
 Composition, cylinder and cylindrical ob- 
 ject, 12 
 cylindrical objects grouped, 26; with 
 
 books, 68 
 in selecting from interior, 112, 115; 
 from photograph of building, 81 
 Concentric circles, 14 
 
 with square, 63-66 
 Cone model, 18 
 
 171
 
 INDEX 
 
 Cone principle, 19 
 Cover of teapot, 28 
 Cream jug, foot, 22 
 
 handle, 20 
 
 ornament, 23 
 
 spout, 22 
 
 study of, 20-23 
 Cube, at 45° with picture plane, 53, 56 
 
 at 30° and 60° with picture plane, 50 
 
 making the drawing, 54 
 
 order of drawing edges, 54 
 
 proportions used in estimating other 
 objects, 51, 86, 129, 164-165, 
 167-168 
 
 recession of horizontal surfaces to eye 
 level, 51 
 
 relation of foreshortening to vanish- 
 ing of edges, 50, 51 
 
 study of, 48-52 
 
 taking direction of edges with pencil, 
 55 
 
 tests of vanishing lines by string and 
 by eye, 56; on blackboard, 57 
 Cylinder, errors in ellipses of, 14 
 
 hollow, the, 14, 15 
 
 inner cylinder, 14 
 
 models for, 8, 14 
 
 perspectively equal divisions, 15 
 
 position of model, 13 
 
 roundness of ellipses, 14-16 
 
 sides tangential to ellipses, 13 
 
 study of, 12-15 
 
 symmetry of ellipses, 15 
 
 true diameter of circle, 15 
 Cylindrical objects grouped, 26-28 
 Cylindrical objects not vertical, 92-94 
 
 application of principle, 94 
 
 other examples, button on cord, 19; 
 circular frame, 96-99; clock, 
 102; flower pots, 95; leaning 
 bowl, 27; luncheon carrier, 32; 
 round arches, 99, 103-104; round 
 window, 100, 101 
 
 symmetry of appearance, 93-94 
 
 test for drawing of, 94, 95 
 Cylindrical objects with fruit, 29, 30 
 Cylindrical picture plane, 140-141 
 
 Desk, 179 
 
 Diagonals, use for measuring, concentric 
 circles, 65, 66 
 door in room, 108 
 
 square frame, 86 
 square plinth, 89 
 Drawing from a description, xii; 
 Problems 
 
 see 
 
 Ears of teapot, 28 
 
 Ellipse, at right angles to axis in cylindrical 
 objects, 93-94 
 
 common errors in, 14 
 
 diameters of, 9, 15 
 
 drawn entire first, 14, 15 
 
 from concentric circles, 14, 63-66 
 
 measurement on its diameters, 14, 15 
 
 position of hand in drawing, 11 
 
 practicing, 10, 11 
 
 roundness according to position, 9, 
 10, 13-15 
 
 study of, 8-11 
 
 symmetry, 9, 15 
 
 tangential to boundary lines, 13, 
 15, 17 
 
 test of shape, 9 
 
 true diameter of circle, 15 
 
 varying curvature of boundary line, 9 
 Exceptions to the use of the flat picture 
 plane, 139-142 
 
 cylindrical picture plane, 140-141 
 
 spherical picture plane, 141-142 
 Eye level, explained, 52 
 
 finding, 39, 40 
 
 importance, 40 
 
 way of using, 39 
 
 Fan, 33 
 
 Finder, 26, 67, 160 
 Flower pots, 95 
 
 Foot of cylindrical objects, at least partly 
 visible, 47 
 
 of cream jug, 22, 23 
 
 of rose jar, 16, 17 
 Foreshortening, xi, 10 
 Foundation truths of perspective, two, xi 
 Freehand sketching defined, xii 
 Freehand work entirely, 3 
 Fruit grouped with cylindrical objects, 
 
 29-30 
 Furniture, chair, 118 
 
 clock, 102, 167 
 ( desk, 179 
 
 good examples, 120 
 
 rocking chair, 118 
 
 proportioning to other objects, 108 
 
 172
 
 INDEX 
 
 Geometric solids, omitted at the teacher's 
 
 discretion, 85, note 
 Geometric measurements, obtained per- 
 
 spectively, 66, 108 
 Glass, bowl, 25 
 pitcher, 30 
 
 Handle, cream jug, 20-22 
 Hexagonal plinth, application of study, 
 126-127 
 
 test for, 124-125 
 
 two positions, 121-125 
 Hexagonal prism and frame, 128-130 
 
 estimating length of prism, 129 
 Horizontal surfaces foreshortened, 37, 40, 
 
 41 
 Horizontal surfaces recede to eye level, 51 
 Horizontal vanishing edges, 37-39, 50 
 House, 69-80 
 
 chimney, 76 
 
 dormer window, 79 
 
 eaves projections, 74 
 
 "L"part, 75 
 
 model, 69 
 
 porch, 75 
 
 roof, 70-73 
 
 steps, 78 
 
 windows and doors, 75-76 
 How much to include in the picture, 112 
 
 Interiors, at angles to picture plane, 110- 
 
 113 
 ceiling, little or none shown, 115 
 door in an interior, 107 
 from memory, 106 and note 
 further studies of, 114-116 
 lines must not point to corners, 109, 
 
 116 
 parallel to picture plane, 105-109, 
 
 132, 133 
 picture on wall, 106 
 relation of subject-space to picture 
 
 plane. 111 
 selection of subject-space, 110 
 stool, 108 
 with tiled floor, 126-127 
 
 Knob of teapot cover, 28 
 
 Lamp shade, 18, 19 
 
 Line, directions for drawing, 1, 2 
 
 expressive, 17 
 
 texture of, 1 
 
 Lines of the picture must not make equal 
 angles, 115 
 nor run to corners of margin, 109, 116 
 
 Margin of picture, 1, 13 
 
 cutting the group, 27, illus., 28, 30 
 moving, to improve picture, 116 
 partial, 116 
 Margins, of the book, 40, 41 
 Materials, pencil and paper, 1 
 
 models, 2; see Models 
 Measuring, by the diagonal, 65, 86, 89, 
 107-108 
 distance into the picture, 65-66 
 height within the picture, 90, 108 
 only relative, 6, 66 
 Measurements obtained geometrically can 
 be so obtained perspectively, 66, 
 107-108, 122-123, 164-165, 167- 
 168 
 Mechanical perspective, value alluded to, 
 140 
 errors of, 139 
 correction of, 140 
 limitations of, 142 
 Memory work, conditions of, 31 
 from interiors, 106 
 group of objects from, 31-32 
 less laborious, note, 106 
 necessity for, xii 
 specially advised, 43, 53, 58, 67 
 Methods, their subsequent use in practical 
 
 work, xii, 101 
 Models, in general, 2 
 
 making, cone, 18, 19; cylinder, 8; 
 cube, 49; rectangular block, 48; 
 hexagonal plinth, 121; triangular 
 prism, 131 
 position for drawing, 2 
 
 Oblique vanishing lines, chair, 119 
 
 dormer, 80 
 
 hexagonal frame, 129, 130 
 
 Hght rays, 147 
 
 roof of house, 72 
 
 shadow on roof, 151 
 Obstacle to mastery of perspective, xii 
 Ornament, constructive principles of, 23 
 
 on Japanese luncheon carrier, 32, 33 
 
 rendering of, 17 
 
 use in composition, 13 
 Out-of-doors work, 154-160 
 
 173
 
 INDEX 
 
 greater distance of vanishing points, 
 
 154-155 
 reflections, 156-160; see Reflections 
 size of objects, 155 
 
 Parallel retreating lines, convergence 
 
 of, 40, 50 
 Parallel retreating horizontal lines, meet- 
 ing at eye level, 50, 93 
 Parallel perspective, bureau, 134 
 interiors, 105-109, 132-133 
 street, 134, 136-137 
 term "parallel" unsatisfactory, 135, 
 note 
 Pencil measurement, difficulties of, 6 
 essential requirements for, 6 
 gives relative size only, 6 
 of a door, 7 
 of the book, 39, 40 
 of the cube, 55, 56 
 of the ellipse, 13 
 on a window-pane, 5 
 on the picture plane, 7 
 study of, 4-7 
 Picture plane, different for each picture, 
 46, 113 
 exceptions to use of flat picture planes, 
 
 139-142 
 may be assumed at any distance, 6 
 method of using, 7 
 position relative to central direction 
 
 of seeing, 6 
 relation to group of objects, 46 
 relation to subject in drawing in- 
 terior, 110-112 
 study of, 4-7 
 Position, for drawing, 1 
 
 of hand for ellipses, 1 1 ; for lines, 1-2 
 of models, 2 
 Practical use of methods, xii, lOl 
 Practice, of ellipses, 11 
 
 of lines, 1, 2 
 Principles of perspective, two founda- 
 tion, xi 
 Problems, clock, 102 
 
 conditions, general, 34; special, 102 
 cylinder, cone and ball, 34, 161-163 
 reasons for giving, xii 
 rectangular block and cylinder, 48, 
 
 163 
 square frame leaning on block, 91, 
 
 164 
 triangular prism and frame, 131, 167 
 
 Profile lines, 21, 22 
 
 QUATREFOIL, 101 
 
 Railroad track, illustrates vanishing 
 
 lines, 41, 42 
 Reflections, lengthened by waves, 160 
 
 length of vertical, 159, 160 
 
 on a horizontal surface, 156-158 
 
 on a vertical surface, 159 
 
 when separated from reflecting sur- 
 face, 158 
 Rose jar, artistic rendering, 17 
 
 foot, 16, 17 
 
 ornament on, 17 
 
 shoulders, 16 
 
 study of, 15-16 
 
 tangential joinings, 17 
 Rocking chair, 118 
 Round arch, 99, 103 
 Round window, 100-101 
 
 San' Apollinare, Church of, 81 
 Selection for picture, from photograph, 81 ; 
 see Composition 
 
 from interior, 112, 115 
 Shadows, 143-153 
 
 cast by an oblique edge, 150-151 
 
 cast by parallel rays (sun, moon), 
 144-151 
 
 cast by rays from lamp, 143, 152 
 
 distorted, 148-149 
 
 in an interior, 152-153 
 
 in a shadow-box, 144-150 
 
 located by imaginary vertical planes, 
 147 
 
 located by two points, 148 
 
 of a cube, 147 
 
 of curves, 149 
 
 of natural objects, 151-152 
 
 on an obHque surface, 151 
 
 on a house, 151 
 
 on curved surfaces, 149-150 
 
 vanishing of light rays, 146-147 
 
 vanishing of shadow-directions, 145- 
 146 
 Shoulders of cylindrical objects, 16; re- 
 lation to cone principle, 19 
 Solutions of problems, 161-168 
 
 cylinder and rectangular block, 163 
 
 cylinder, cone and ball, 161 
 
 square frame leaning on rectangular 
 block, 164 
 
 174
 
 INDEX 
 
 triangular prism and frame, 167 
 Spherical picture plane, 141-142 
 Spout, of cream jug, 22 
 Square frame, 85-87 
 
 test of, 87 
 
 application of study, 87 
 
 leaning on rectangular block, 
 164 
 
 91, 
 
 Table line, explained, 3 
 
 high enough on paper, 43 
 
 position with plate on edge, 30 
 
 significance in composition, 13 
 
 subordination of, 44 
 Taking direction of vanishing edges with 
 
 pencil, 55 
 Teapot, cover, 28; ears, 28 
 
 knob, 28 
 
 study, 26-28 
 Tests, by blackboard for vanishing lines, 
 57 
 
 by eye, for vanishing lines, 56 
 
 cylindrical objects not vertical, 95 
 
 hexagonal plinth, 124 
 
 square frame, 87 
 
 the ellipse, 9 
 
 the eye, a final test, xii, 142 
 Thumb-nail sketches, interiors, 115 
 
 still-life objects, 26, 68 
 Tiled floor, 109, 126 
 Time study, glass bowl, 24, 25 
 Triangular prism and frame, 131, 167 
 Two books, at difl^erent angles, 61-62 
 
 with cylindrical object, 67-68 
 
 Vanishing lines, " converging," 37-38 
 
 example of, railroad, 41-42 
 
 oblique, 73; see Oblique vanishing 
 lines 
 
 taking direction of, with pencil, 55 
 
 tests of, 56, 57 
 Vanishing of parallel planes, 52, 73-74, 
 
 162, note 
 Vanishing points, abbreviation of, 44 
 
 numbering, 50 
 
 oblique, 73; see Oblique vanishing 
 lines 
 
 use without marking, 56, 80, 124 
 Vanishing traces, 73, 151 
 Vertical lines drawn vertical, 60 
 Vignetting, 116 
 
 175

 
 ^
 
 UNIVERSITY OF CALIFORNIA, LOS ANGELES 
 
 THE UNIVERSITY LIBRARY 
 
 This book is DUE on the last date stamped below 
 
 ^ 
 
 }S^ 
 
 
 1^\W' 
 
 Form L-!> 
 2r,m-10,'4i<2491)
 
 UC SOUTHERN REGIONAL LIBRARY FACILITY 
 
 iiiiin II iiii| II 
 
 AA 001 
 
 32316 
 
 '-/BRAfi^ 
 
 ( ^ A PLEA«^ DO NOT REMOVE 
 THIS BOOK CARD^ 
 
 ^^^^l•LIBRARY(9/r 
 
 
 University Research Library