THE BOOK WAS DRENCHEDUNIVERSAL LIBRARY OU_164819 Adve] IWSYSAINNDOVER BOOKS ON ADVANCED MATHEMATICS Mathematical Analysis of Electrical and Optical Wave-Motton, Harry Bateman $1.60 Numerical Integration of Differential Equations, A. Bennett, W. Milne, Harry Bateman $135 Almost Periodic Functions, A. Besicouitch $175 Non-Euclidegn, Geometry, R. Bonola $195 Introduction to Bessel Functions, F Bowman $1 35 Theory, of Groups of Finite Order, W. Burnside $2.45 Theory of Probability, W. Burnside $100 An Elementary Treatise on Fourier’s Sertes and Spherical, Cylindrical, and Ellipsordal Harmonics, W. E. Byerly $175 Contributions to the Theory of Transfinite Numbers, G. Cantor $1.25 Introduction to the Theory of Groups of Finite Order, R.D. Carmichael $2.00 The Theory of Numbers and Diophantine Analysis, R. D. Carmichael $135 Introduction to the Theory of Fourter’s Sertes and Integrals, H. S. Carlslaw $2.00 Introduction to the Theory of Numbers, L. E Dickson $1.65 Algebraic Theones, 1. E Dickson $150 Linear Groups, L. E. Dickson $1.95 The Taylor Sertes, Paul Dienes $275 Mathematical Tables of Elementary and Some Higher Mathematical Functions, Herbert B, Dwight $1.75 Asymptotic Expansions, A. Erdélyi $135 Computational Methods of Linear Algebra, V N. Faddeeva $1.95 The Foundations of Euchdean Geometry, HG. Forder $200 The Phase Rule and Its Applications, A. Findlay $245 The Theory of Differential Equations, A. R Forsyth 6 volumes bound as 3, clothbound, the set $15.00 An Introduction to Fourter Methods and the Laplace Transformation, Philp Franklin $175 A Course in Mathematical Analysts, Edouard Goursat Three volume set $5.00 Lectures on Cauchy’s Problem, J. Hadamard $1.75 Elliptic Integrals, Harris Hancock $1.25 Lectures on the Theory of Elliptic Functions, Harris Hancock $2.55 Vector and Tensor Analysis, G. E. Hay $1.75 Theory of Functions of a Real Variable and Theory of Fourter’s Series, E.W. Hobson Two volume set, $6.00 (continued on inside back cover)COORDINATE GEOMETRYCOORDINATE GEOMETRY BY LUTHER PFAHLER EISENHART PROFESSOR OF MATHEMATICS PRINCETON UNIVERSITY DOVER PUBLICATIONS, INC. NEW YORK * NEW YORKCopyright, 1939, by Ginn and Company All Rights Reserved This new Dover edition first published in 1960 is an unabridged and unaltered republication of the First Edition. This book is republished by permission of Ginn and Company, the original publisher of this text. Manufactured in the United States of America Dover Publications, Ine. 180 Varick Street New York 14, New YorkPreface ” The purposes and general plan of this book are set forth in the Introduction. Its practicability as a text has been tested and proved through use during two years in freshman courses in Princeton University. Each year it has been revised as a result of suggestions not only by the members of the staff but also by the students, who have shown keen interest and help- fulness in the development of the project. An unusual feature of the book is the presentation of co- ordinate geometry in the plane in such manner as to lead readily to the study of lines and planes in space as a generalization of the geometry of the plane; this is done in Chapter 2. It has been our experience that the students have little, if any, difficulty in handling the geometry of space thus early in the course, in the way in which it is developed in this chapter. We have also found that students who have studied deter- minants in an advanced course in algebra find for the first time in the definition and frequent use of determinants in this book an appreciation of the value and significance of this subject. In the preparation and revisions of the text the author has received valuable assistance from the members of the staff at Princeton — in particular, Professors Knebelman and Tucker and Messrs. Tompkins, Daly, Fox, Titt, Traber, Battin, and Johnson ; the last two have been notably helpful in the prepa- ration of the final form of the text and in the reading of the proof. The Appendix to Chapter 1, which presents the relation between the algebraic foundations of coordinate geometry and axioms of Hilbert for Euclidean plane geometry, is due in the main to Professors Bochner and Church and Dr. Tompkins, particularly to the latter, at whose suggestion it was prepared and incorporated. The figures were drawn by Mr. J. H. Lewis. It remains for me to express my appreciation of the courtesy and cooperation of Ginn and Company in the publication of this book. LUTHER PFAHLER EISENHART iContents ow INTRODUCTION CHAPTER 1 Points and Lines in the Plane SECTION . The Equation of the First Degree in x and y 2. Cartesian Coordinates in the Plane 3. Distance between Two Points - Direction Numbers and Di- rection Cosines - Angle between Directed Line Segments = _ . Internal and External Division of a Line Segment a . An Equation of a Line - Parametric Equations of a Line a . Direction Numbers and Direction Cosines of a Line - Angle between Two Lines 7. The Slope of a Line 8. Directed Distance from a Line to a Point 9. Two Equations of the First Degree in x and y - Determinants of the Second Order 10. The Set of Lines through a Point 11. Oblique Axes 12. The Circle 13. Résumé - Line Coordinates ” CHAPTER 2 Lines and Planes in Space + Determinants 14. Rectangular Coordinates in Space 15. Distance between Two Points - Direction Numbers and Di- rection Cosines of a Line Segment - Angle between Two Line Segments v PAGE 1 7 20 7 16Contents SECTION 16. Equations of a Line - Direction Numbers and Direction Co- 17. 18. 19. 20. 21. sines of a Line - Angle of Two Lines An Equation of a Plane The Directed Distance from a Plane to a Point - The Dis- tance from a Line to a Point Two Equations of the First Degree in Three Unknowns - A Line as the Intersection of Two Planes Two Homogeneous Equations of the First Degree in Three Unknowns Determinants of the Third Order - Three Equations of the First Degree in Three Unknowns 22. Three Homogeneous Equations of the First Degree in Three Unknowns 23. Equations of Planes Determined by Certain Geometric Con- ditions - Shortest Distance between Two Lines 24. The Configurations of Three Planes 88 31 . Miscellaneous Exercises - The Sphere 26. 27. Determinants of Any Order Solution of Equations of the First Degree in Any Number of Unknowns - Space of Four Dimensions “o CHAPTER 3 Transformations of Coordinates . Transformations of Rectangular Coordinates . Polar Coordinates in the Plane ). Transformations of Rectangular Coordinates in Space . Spherical and Cylindrical Coordinates vi PAGE 82 88 95 98 104 106 114 119 123 127 130 137 149 154 166Contents CHAPTER 4 The Conics - Locus Problems SECTION PAGE 32. A Geometric Definition of the Conics 171 33. The Parabola 174 34. Tangents and Polars 177 35. Ellipses and Hyperbolas 182 36. Conjugate Diameters and Tangents of Central Conics 191 37. Similar Central Conics - The Asymptotes of a Hyperbola - Conjugate Hyperbolas 196 38. The Conics as Plane Sections of a Right Circular Cone 201 39. Equations of Conics Whose Axes Are Parallel to the Co- ordinate Axes 203 40. The General Equation of the Second Degree - Invariants 208 41. The Determination of a Conic from Its Equation in General Form 215 42. Center, Principal Axes, and Tangents of a Conic Defined by a General Equation 221 43. Locus Problems 229 a CHAPTER 5 The Quadric Surfaces 44, Surfaces of Revolution The Quadric Surfaces of Revolution 239 45. Canonical Equations of the Quadric Surfaces 243 46. The Ruled Quadrics 248 47. Quadrics Whose Principal Planes Are Parallel to the Co- ordinate Planes 252 48. The General Equation of the Second Degree in x, y, and z - The Characteristic Equation - Tangent Planes to a Quadric 256 VilContents SECTION PAGE 49. Centers - Vertices - Points of Symmetry 264 50. The Invariants J, J, D, and A 267 51. Classification of the Quadrics 271 APPENDIX TO CHAPTER 1 279 INDEX 293 viiiIntroduction ao Coordinate geometry is so called because it uses in the treat- ment of geometric problems a system of coordinates, which associates with each point of a geometric figure a set of num- bers — coordinates — so that the conditions which each point must satisfy are expressible by means of equations or inequali- ties ordinarily involving algebraic quantities and at times trigonometric functions. By this means a geometric problem is reduced to an algebraic problem, which most people can handle with greater ease and confidence. After the algebraic solution has been obtained, however, there. remaiys its geometric inter- pretation to be determined ; for, thg preblem is geometric, and algebra is a means to its solution, not, the end” This method was introduced by René Descartes in La Géométrie, published in 1636; accordingly coordinate geometry is sometimes called Cartesian geometry. Before the time of Descartes geometric reasoning only was used in the study of geometry. The advance in the development of geometric ideas since the time of Descartes is largely due to the introduction of his method. Geometry deals with spatial concepts. The problems of physics, astronomy, engineering, etc. involve not only space but usually time also. The method of attack upon these problems is similar to that used in coordinate geometry. However, geometric problems, because of the absence of the time element, are ordinarily simpler, and consequently it is advisable that the student first become familiar with the methods of coordinate geometry. The aim of this book is to encourage the reader to think mathematically. The subject matter is presented as a unified whole, not as a composite of units which seem to have no relation to one another. Each situation is completely analyzed, and the various possibilities are all carried to their conclusion, for frequently the exceptional case (which often is not pre- sented to a student) is the one that clarifies the general case — the idea epitomized in the old adage about an exception and a ixIntroduction rule. Experience in analyzing a question fully and being careful in the handling of all possibilities is one of the great advantages of a proper study of mathematics. Examples included in the text are there for the purpose of illustration and clarification of the text; there is no attempt to formulate a set of patterns for the reader so that the solution of exercises shall be a matter of memory alone without requiring him to think mathematically. However, it is not intended that he should develop no facility in mathematical techniques; rather these very techniques will have added interest when he under- stands the ideas underlying them. The reader may at first have some difficulty in studying the text, but if he endeavors to master the material, he will be repaid by finding coordinate ge- ometry a very interesting subject and will discover that mathe- matics is much more than the routine manipulation of processes. Since, as has been stated, coordinate geometry involves the use of algebraic processes in the study of geometric problems and also the geometric interpretation of algebraic equations, it is important that the reader be able not only to use algebraic processes but also to understand them fully, if he is to apply them with confidence. Accordingly in Chapter 1 an analysis is made of the solution of one and of two equations of the first degree in two unknowns; and in Chapter 2 are considered one, two, and three equations of the first degree in three unknowns. In order that the discussion be general and all-inclusive, literal coefficients are used in these equations. It may be that most, if not all, of the reader’s experience in these matters has been with equations having numerical coefficients, and he may at first have some difficulty in dealing with literal coefficients. At times he may find it helpful in understanding the discussion to write particular equations with numerical coefficients, that is, to give the literal coefficients particular numerical values, thus supplementing illustrations of such equations which appear in the text. However, in the course of time he will find it unneces- sary to do this, and in fact will prefer literal coefficients because of their generality and significance. In the study of equations of the first degree in two or more unknowns determinants are defined and used, first those of the xIntroduction second order and then those of the third and higher orders, as they are needed. Ordinarily determinants are defined and studied first in a course in algebra, but it is a question whether one ever appreciates their value and power until one sees them used effectively in relation to geometric problems. The geometry of the plane is presented in Chapter 1 in such form that the results may be generalized readily to ordinary space and to spaces of four and more dimensions, as is done in Chapter 2. Some of the exercises are a direct application of the text so that the reader may test his understanding of a certain subject by applying it to a particular problem and thus also acquire facility in the appropriate techniques; others are of a theoretical character, in the solution of which the reader is asked to apply the principles of the text to the establishment of further theorems. Some of these theorems extend the scope of the text, whereas others complete the treatment of the material in the text. Of general interest in this field are A. N. Whitehead’s Introduction to Mathematics, particularly Chapter 8, and E. T. Bell’s Men of Mathematics, Chapter 3.o COORDINATE GEOMETRY CHAPTER 1 Points and Lines in the Plane1. The Equation of the First Degree in x and y In his study of algebra the reader has no doubt had experi- ence in finding solutions of equations of the first degree in two unknowns, x and y, as for example x —-2y+3=0. Since we shall be concerned with the geometric interpretation of such equations, a thorough understanding of them is essential. We therefore turn our attention first of all to a purely algebraic study of a single equation in x and y, our interest being to find out what statements can be made about a general equation of the first degree, which thus will apply to any such equation without regard to the particular coefficients it may have. Ac- cordingly we consider the equation (1.1) ax + by+c=0, where a, b, and c stand for arbitrary numbers, but definite in the case of a particular equation. We say that a value of x and a value of y constitute a solution of this equation if the left- hand side of this equation reduces to zero when these values are substituted. Does such an equation have a solution what- ever be the coefficients? The question is not as trivial as may appear at first glance, and the method of answering it will serve as an example of the type of argument used repeatedly in later sections of this book. We consider first the case when a in (1.1) is not equal to zero, which we express by a 0. If we substitute any value whatever for y in (1.1) and transpose the last two terms to the right-hand side of the equation, which involves changing their signs, we may divide through by a and obtain the value r= wre, This value of x and the chosen value of y satisfy the equation, as one sees by ‘substitution; hence any value of y and the resulting value of x constitute a solution. Since y may be given any value and then x is determined, we say that when a= 0 there is an endless number of solutions of the equation. The above method does not apply when a =0, that is, to the équation Ox+by+c=0, since division by zero is not an allowable process. The reader may have been told that it is allowable, and that any number, 3Points and Lines in the Plane [Chap. 1 for example 2, divided by zero is infinity ; but infinity defined in this manner is a concept quite different from ordinary num- bers, and an understanding of the concept necessitates an ap- propriate knowledge of the theory of limits. If now 6 # 0, the above equation may be solved for y; that is, y=—c/b. For this equation also there is an endless number of solutions, for all of which y has the value —c/b and x takes on arbitrary values. Usually the above equation is written by +c=0, which does not mean that x is equal to zero (a mistake fre- quently made), but that the coefficient of x is zero. If b 0, no matter what a is, equation (1.1) may be solved axtt, for y with the result y = — the value of y corresponding to any choice of x being given by this expression. Hence we have the theorem [1.1] Am equation of the first degree in two unknowns in which the coefficient of at least one of the unknowns is not equal to zero has an endless number of solutions. There remains for consideration the case when a =0 and 6=0; that is, the equation Ox+0y4+c=0. Evidently there are no solutions when c ~ 0, and when c=0 any value of x and any value of y constitute a solution. The reader may say that in either case this is really not an equation of the first degree in x and y, and so why consider it. It is true that one would not start out with such an equation in formu- lating a set of one or more equations of the first degree to ex- press in algebraic form the conditions of a geometric problem, but it may happen that, having started with several equations, and carrying out perfectly legitimate processes, one is brought to an equation of the above type; that is, equations of this type do arise and consequently must be considered. In fact, this situation arises in § 9, and the reader will see there how it is interpreted. However, when in this chapter we are deriving theorems con- cerning equations of the first degree in x and y, we exclude the degenerate case when the coefficients of both x and y are zero. 4sec.) The Equation of the First Degree in x and y Any solution of equation (1.1) is also a solution of the equation k(ax + by +c) =0, where & is any constant different from zero. Moreover, any solution of this equation is a solu- tion of (1.1); for, if we are seeking the conditions under which the product of two quantities shall be equal to zero and one of the quantities is different from zero by hypothesis, then we must seek under what conditions the other quantity is equal to zero. Accordingly we say that two equations differing only by a con- stant factor are not essentially different, or are not independent, or that the two equations are equivalent. In view of this dis- cussion it follows that a common factor, if any, of all the co- efficients of an equation can be divided out, or the signs of all the terms of an equation can be changed, without affecting the solutions of the equation — processes which the reader has used even though he may not have thought how to justify their use. It should be remarked that the values of the coefficients in equation (1.1) are the important thing, because they fix y when x is chosen and vice versa. This is seen more clearly when we take a set of values x, y, and seek the equation of which it is a solution; this is the inverse of the problem of finding solutions of a given equation. Here the subscript 1 of x; and yi has nothing to do with the values of these quantities. It isa means of denoting a particular solution of an equation, whereas x and y without any subscript denote any solution whatever. If x; and 4; is to be a solution of (1.1), the coefficients a, b, and c must be such that (1.2) ax, +by, +c=0. On solving this equation for c and substituting in (1.1), we obtain the equation (1.3) ax + by — ax, — by, = 0, which is of the form (1.1), where now c is — (ax; + by), a con- stant for any values of a and 8, x; and y; being given constants. When equation (1.3) is rewritten in the form (1.4) a(x — m) + b(y— 91) =0, it is seen that x = x; and y = y; is a solution of (1.3) whatever be @ and 6. Since a and b can take any values in (1.4), we see 5Points and Lines in the Plane [Chap. 1 that an equation of the first degree in x and y is not completely determined (that is, a, b, and c are not fixed) when one solution is given. Suppose then that we require that a different set of quan- tities, x. and y2, be also a solution, the subscript 2 indicating that it is a second solution. On replacing x and y in (1.4) by %2 and y2, we obtain (1.5) a(x2 — 11) + b(y2 — 1) = 0. Since we are dealing with two different solutions, either x2 # x1 or yo yi; we assume that x2 x, and solve (1.5) for a, with the result (1.6) q=—p¥@rn, X2— M1 On substituting this value of a in (1.4), we obtain bf (x — 1) B= + (y—yJ]=0. we % This equation is satisfied if 6 = 0, but then from (1.5) we have a(x2—%1) =0. Since x2 x1, we must have a = 0, contrary to the hypothesis that ¢ and b are not both equal to zero. There- fore since 6 cannot be zero, it may be divided out of the above equation, and what remains may be written (1.7) yy = B= (x — m), xX2— Multiplying both sides by x2 — x, the resulting equation may be written (1.8) (2 — )x — (%2 — m1) + (X21 — M12) = 0, which is of the form (1.1), since 1, 91, x2, and ye are fixed numbers. Thus far we have assumed that x2 x1; if now x2 = x we cannot have also y2 = 4, since the two solutions are different, and consequently from (1.5) we have b= 0. In this case equa- tion (1.4) becomes a(x — x1) =0. We observe that this equa- tion differs only by the constant factor a from x—x,=0. And the same is true of the form which (1.8) takes when 6sec] The Equation of the First Degree in x and y we put x2 = x1, namely, (y2—91)(x— x1) =0. But we have remarked before that a common factor of all the coefficients does not affect the solutions of the equation. Accordingly we have the theorem [1.2] Equation (1.8) is an equation of the first degree in x and y which has the two solutions x, y; and x2, yo. We say “an equation” and not ‘the equation” because any constant multiple of equation (1.8) also has these solutions; in this sense an equation is determined by two solutions to within an arbitrary constant factor. As a result of the discussion leading up to theorem [1.2] we have the theorem [1.8] Although an equation of the first degree in x and y admits an endless number of solutions, the equation is determined to within an arbitrary constant factor by two solutions, that ts, by two sets of values of x and y; for the solutions x1, y1 and x2, yz the equation is equivalent to (1.8). Note. In the numbering of an equation, as (1.5), the number pre- ceding the period is that of the section in which the equation appears, and the second number specifies the particular equation. The same applies to the number of a theorem, but in this case a bracket is used instead of a parenthesis. EXERCISES 1, What values must be assigned to a, b, and c in equation (1.1) so that the resulting equation is x =0; so that it is equivalent to this equation? 2. Find two solutions of the equation @) 2x-3y+6=0, and show that equation (1.8) for these two solutions is equivalent to equation (i). 8. Show that the equation x — 2 y + 3=0 does not have two dif- ferent solutions for both of which x = 2; find an equation of the first degree in x and y for which this is true. 7Points and Lines in the Plane (Chap. 1 4. Show that it follows from (1.1) and (1.8) that G:b:C=Yo— Yi k1 — X22 21 — M2, where x, yi and x2, y2 are solutions of (1.1). 5. Criticize the following statements: a. In obtaining solutions of an equation ax + by +c=0, x may take any value if ¢ = 0, and only in this case. b. x —3=0, as an equation in x and y, has the solution x = 3, y = 0, and this is the only solution. 2. Cartesian Coordinates in the Plane Having studied equations of the first degree in two unknowns, we turn now to the geometric interpretation of the results of this study. This is done by the introduction of coordinates, which serve as the bridge from algebra to geometry. It is a bridge with two-way traffic; for also by means of coordinates geo- metric problems may be given alge- braic form. This use of coordinates was Descartes’s great contribution to mathematics, which revolution- ized the study of geometry. As basis for the definition of coor- dinates, we take two lines perpen- dicular to one another, as A’A and 4’ ‘ B’B in Fig. 1, which are called the 7 x-axis and y-axis respectively ; their Fig. 1 intersection O is called the origin. Suppose that x, and y; are a pair of numbers. If x; is positive, we lay off on OA from 0 a length OC equal to x; units and draw through C a line parallel to B’B; if x; is negative, we lay off from 0 a length equal to— x, units on 0A’, and draw through the point so determined a line parallel to B’B. Then, starting from O we lay off on OB a length OD equal to y; units if 9; is positive, or on OB’ a length — y; units if y; is negative, and through the point so determined draw a line parallel to A’A. These lines so drawn meet in a point Pi, which is called the graph of the pair x, 1; we say that P, is the point (m, 91), 8 a Pi (2m)Sec. 2] Cartesian Coordinates in the Plane and x, y are called the coordinates of P,; also x: is called the abscissa of P; and ; the ordinate. Evidently x: is the distance of P; from the y-axis (to the right if x: is positive, to the left if x, is negative) and y, is the distance of P; from the x-axis. In specifying a line segment OC, the first letter O indicates the point from which measurement begins, and the last letter C the point to which measurement is made. Accordingly we have CO = — OC, because for CO measurement is in the direc- tion opposite to that for OC. (The question of sign may be annoying, but in many cases it is important; there are also cases when it is not important, and the reader is expected to discriminate between these cases.) If we take two points Ci(m1, 0) and C2(x2, 0) on the x-axis, we see that the magnitude and sign of the segment C,C2 is x2 — x, since Ci\C2 = OC2 — OC. When C; and Cz lie on the same side of 0, OC2 — OC; is the dif- ference of two lengths; when they are on opposite sides of 0, it is the sum of two lengths. (The reader does not have to worry about this, since x2 — x, takes care of all of it.) CiC2 is called the directed distance from C, to C2. We observe that the coordinate axes divide the plane into four compartments, which are called quadrants. The quadrant formed by the positive x-axis and positive y-axis is called the first quadrant; the one to the left of the positive y-axis (and above the x-axis), the second quadrant; the ones below the negative x-axis and the positive x-axis, the third and fourth quadrants respectively. By definition the projection of a point upon a line is the foot of the perpendicular to the line from the point. Thus C and D in Fig. 1 are the projections of the point P; upon the x-axis and y-axis respectively. The projection of the line segment whose end points are P; and P: upon a line is the line segment whose end points are the projections of P; and P2 on the line. Thus in Fig. 1 the line segment OC is the projection of the line seg- ment DP, upon the x-axis. Two points P; and Pe are said to be symmetric with respect to a point when the latter bisects the line segment PiP2; sym- metric with respect to a line when the latter is perpendicular to the line segment P,P2 and bisects it. Thus the points (3, 2) 9Points and Lines in the Plane (Chap. 1 and (— 3, —2) are symmetric with respect to the origin, and (3, 2) and (3, — 2) are symmetric with respect to the x-axis. The reader should acquire the habit of drawing a reasonably accu- rate graph to illustrate a problem under consideration. A carefully made graph not only serves to clarify the geometric interpretation of a problem but also may serve as a valuable check on the accuracy of the algebraic work. Engineers, in particular, often use graphical methods, and for many purposes numerical results obtained graphi- cally are sufficiently accurate. However, the reader should never for- get that graphical results are at best only approximations, and of value only in proportion to the accuracy with which the graphs are drawn. Also in basing an argument upon a graph one must be sure that the graph is really a picture of the conditions of the problem (see Chapter 3 of W. W. R. Ball’s Mathematical Essays and Recrea- tions, The Macmillan Company). EXERCISES 1. How far is the point (4, — 3) from the origin? What are the coordinates of another point at the same distance from the origin? How many points are there at this distance from the origin, and how would one find the coordinates of a given number of such points? 2. What are the coordinates of the point halfway between the origin and (— 4, 6)? What are the coordinates of the point such that (— 4, 6) is halfway between it and the origin? 3. Where are the points in the plane for which x > y (that is, for which x is greater than y)? Where are the points for which x? + y? < 4? 4. Where are the points for which 0< x=1 and 0< y=1 (the symbol = meaning “less than or equal to’’)? Where are the points for which O 0, —1 | he |. On the other hand, if we have equations (4.5) and (4.6) with fy and he differing in sign, and | hi | < | ho |, this means that x and y are the coordinates of a point P lying below P; on the line through P; and Pe, as the reader will see by considering the above definition of P for the case | A: | < | he |, and drawing a figure. In accordance with custom we say that in these cases P(x, y) divides the line segment P;P2 externally in the ratio My/h2; and that when P is a point of the segment, it divides the segment P,P2 internally. Accordingly we have the theorem [4.2] The equations — fox + Iix2, — fay t+ hyo (4.8) . ho+h y ho+h give the coordinates of the point P dividing the line segment P\P2, with end points Pi(x,, 91), P2(%2, y2), in the ratio Iy/h2, internally when h, and hz have the same sign and externally when hy and hz have opposite signs; in either case P lies nearer P; or P2 according as | h| is less or greater than | he |. EXERCISES 1, Find the coordinates of the points dividing the line segment between the points (7, 4) and (5, — 6) in the ratios 2/3 and — 2/3. 2. Find the coordinates of the two points which trisect the line segment between the points (3, — 2) and (— 3, 1). 8. In what ratio does the point (3, — 2) divide the line segment between the points (5, 2) and 6, 4)? 19Points and Lines in the Plane [Chap. 1 4. Find the coordinates of the point P(x, y) on the line through the points P,(5, — 3) and P2(2, 1) such that P2 bisects the segment PiP. 5. Show that the medians of the triangle with vertices (— 1, 2), (3, 4), and (5, 6) meet in a point by finding for each median the coordinates of the point twice as far from the corresponding vertex as from the opposite side, distances being measured along the median. 6. Is there a point on the line through Pi(x:, 91) and Pa(x2, ye) dividing the line segment PP2 in the ratio h/hz = — 1? 7. Where are the points in the plane for which |x|> 2 and ly1<3? 8. Show that the points (x, 91) and (x2, y2) are collinear with the origin if and only if their coordinates are proportional. 9. Show that the points (x1, 91), (xz, y2), and (xs, ys) are collinear when there are three numbers &,, 2, ks, all different from zero, such that ki + ke+ ks =0, kin + kore + kox3=0, hiyn + hove + hays = 0. 10. Interpret the following statements: a. For h, and he both positive x defined by (4.8) is the weighted average of x; and x2 with the respective weights he and /, and accord- ing as fz is greater than /; or less than h, the weighted average is nearer x; than x2 or vice versa. b. x and y given by (4.8) are the coordinates of the center of gravity of masses hz at P; and hy; at Po. 11. What are the coordinates of the center of gravity of masses my, ma, and ms at the points (x1, 1), (x2, ¥2), and (x3, ys); of m masses at n different points? (Use mathematical induction.) 5. An Equation of a Line. Parametric Equations of a Line We have seen how any two numbers x and y define a point in the plane, called the graph of the ordered pair of numbers. Reversing the process, we have that any point in the plane is the graph of an ordered pair of numbers, the coordinates of the point, obtained by drawing through the point lines parallel to the coordinate axes and taking for these numbers the magnitudes of the intercepted segments with appropriate signs, as OC and OD in Fig. 1. An equation in x and y imposes a restriction on x and y, so that only one of these may take 20Sec. 5] An Equation of a Line arbitrary values and then the other is determined by the equa- tion. The graph of an equation is the locus (place) of all points whose coordinates are solutions of the equation, in other words, the locus of the graphs of ordered pairs of numbers satisfying the equa- tion. Each of the points of this locus possesses a geometric property common to all the points of the locus, and no other points; this property is expressed algebraically by the equation. Just what this geometric property is in the case of a particular equation is one of the interesting questions of coordinate geom- etry. The converse problem is that of finding an equation whose solutions give the coordinates of every point of a locus when the locus is defined geometrically. All of this will become clearer as we proceed to take up particular equations and particular loci. Since it requires an ordered pair of numbers to define a point in the plane and any point in the plane is so defined, we say that a plane is two-dimensional. A line is one-dimensional, since any point on the line is defined by one number, for example, its distance from a fixed point on the line. Similarly, since any point on a curve is defined by one number, for example, its distance measured along the curve (say, with a piece of string) from a point of the curve, we say that any curve is one- dimensional. The graph of an equation in x and y is some kind of curve, or a line; for, as remarked above, only one of the unknowns may be chosen arbitrarily, the other being determined by the equation, and thus the equation picks out from the two- dimensional set of points in the plane a one-dimensional set. We consider now the graph of an equation of the first de- gree in x and y. When one takes such an equation, and, having obtained a number of solutions after the manner discussed in §1, plots their graphs, one observes that they seem to lie on a straight line. In fact, one may have been told that all one has to do is to plot two solutions and draw a straight line through the two points, and that this is the graph of the equa- tion in the sense that the graph of every solution is a point of the line and every point of the line is the graph of a solution. In order to prove that this is a correct statement, we consider first the inverse problem of finding an equation of a line. 21Points and Lines in the Plane [Chap. 1 Consider a line which is not parallel to either axis, and choose upon it two particular points, P, of coordinates x, y1 and Pz of coordinates x2, ye. We denote by P of coordinates x, y any point on the line — meaning that it may be placed anywhere on the line while keeping P, and Pe fixed; in this sense we speak of P as a general or representative point, and use x and y without subscripts to distinguish a representative point from a particular point. Through Pi, P, P2 we draw lines par- allel to the axes, forming similar triangles, as shown in Fig. 6. Since the triangles are similar, corresponding sides are pro- portional, and we have PiQ QP. P, 5.1 = : 2 (2y2) G1) BQ: = OP: P(x) ee It should be observed that the two segments in each ratio are taken in the same sense. We have P:}Q=x—n, P\Q2=%2—%m, QP=Y—-MN, ? Q2P2 = y2— 1. (Although Fic. 6 Fig. 6 has been drawn with all the coordinates positive, the reader will readily verify that these relations hold equally well when the figure is so placed with reference to the axes that some or all of the numbers %1, 91, x, Y, X2, Yo are negative.) Consequently (5.1) is equivalent to the equation 52 2a JOM, (6.2) X2— M1 Ya Pi(xuys) Q(t.y) Qa(za.91) Although this result has been derived for the case when P lies on the segment P,P2, by drawing suitable figures the reader can assure himself that equation (5.2) holds when P lies on the line but outside the segment P:P2. Equation (5.2) is reducible to (5.3) O2— nx — (x2 — mi) + (x29 — M12) = 0. This equation is satisfied by the coordinates of every point on the line. Moreover, from the form (5.2) of the equation it follows that any solution of the equation leads back to the ratios (5.1), and consequently gives the coordinates of a point on the line. 22Sec. 5] An Equation of a Line Although equation (5.3) was derived for a line inclined to the coordinate axes, that is, when x2 ~ x, and yo # 1, it ap- plies to the cases when the line is parallel to either axis. In fact, if the line is parallel to the x-axis, we have yo = yi, %2 ¥ X1, in which case equation (5.3) is equivalent to 6.4) y-n=0. Also, when x2 = x1, v2 ¥ 91, equation (5.3) is equivalent to (5.5) x—%=0, which is an equation of a line parallel to the y-axis. These results follow also from (5.2), if we adopt the principle that when in an equation of two ratios either term of a ratio is equal to zero, the other term also is equal to zero. Thus if y2 = y1, we must have y—y:=0. This does not say anything about the value of the other ratio; this value is determined by the values of the terms of this ratio. Hence we have the theorem [5.1] Equation (5.2), or (5.3), ts an equation of the line through the points (x1, y1) and (x2, y2). Since x1, #1, ¥2, and ye are fixed numbers, equations (5.2) and (5.3) are of the first degree in x and y, as are also (5.4) and (5.5). Consequently we have the theorem [5.2] An equation of a straight line is of the first degree in x and y. We say ‘an equation,” and not ‘the equation,” because, if we have an equation of a line, so also is any constant multiple of this equation an equation of the line, since any solution of either equation is a solution of the other. Having shown that in an equation of any line the coeffi- cients of x and y are not both simultaneously equal to zero, we remark that if, as stated in § 1, in deriving theorems in this chapter we exclude from consideration the degenerate case of equations of the first degree for which the coefficients of both x and y are zero, we are not thereby restricting the considera- tion of all the lines in the plane. 23Points and Lines in the Plane [Chap. 1 We shall prove the converse of Theorem [5.2], namely, [5.3] The graph of any equation of the first degree in x and y is a straight line. We consider the general equation of the first degree (6.6) ax+by+c=0 in which not both a and 6 are equal to zero. For a=0, b¥ 0, equation (5.6) is reducible to the form (5.4) (with a change of notation), which is an equation of a line parallel to the x-axis. For a # 0, b= 0, equation (5.6) is reducible to the form (5.5), which is an equation of a line parallel to the y-axis. In §1 it was shown that when a ~ 0, b = 0, equation (5.6) can be put in the form (1.8) in terms of two of its solutions x, y; and x2, ye. But equation (1.8) is the same as equation (5.3). Hence any equation (5.6) can be given one of the forms (5.3), (5.4), or (5.5), and the theorem is proved. Returning to the consideration of equation (5.2), we re- mark that, when the line is not parallel to either axis, for each point on the line the two ratios have the same value, this value, say t, depending upon the values of x and y. If we put each of the ratios equal to ¢ and solve the resulting equations for x and y, we obtain (5.7) x=umtie2—m), y=ntt(y2-n). Conversely, for each value of ¢ the values of x and y given by (5.7) are such that equation (5.2) is satisfied, as is seen by sub- stitution, and consequently these values of x and y are the coordinates of a point on the line. When the line is parallel to the x-axis, equations (5.7) hold, the second equation reducing to y=; and similarly for a line parallel to the y-axis. Thus the single equation (5.2) of the line has been replaced by the two equations (5.7) through the introduction of an auxiliary variable #, called a parameter. Accordingly equations (5.7) are called parametric equations of the line. Ordinarily, as will be seen later, a curve in the plane is defined by one equation in x and y, but sometimes it is convenient to define it by two equations in x, y, and a parameter ; an example of this is found 24Sec. 5] Parametric Equations of a Line in mechanics when the coordinates of a moving particle are expressed in terms of time as parameter. When the equations of a curve are in parametric form, it may be possible to obtain a single equation in x and y by eliminating the parameter from the two equations. Thus, if we solve the first of (5.7) for ¢ and substitute the result in the second, we get an equation reducible to (5.2). Again, consider the locus of a point whose coordinates are given by x = ??, y = 2, as ¢ takes on all values. Eliminating t, we get y? = 4x, which the reader who has plotted curves in his study of algebra will identify as an equation of a parabola. The use of a parameter in defining a line, or a curve, em- phasizes the fact that a line, or a curve, is one-dimensional in that any point on it is specified by the appropriate value of a single variable, a parameter. In fact, in plotting the graph of a line, or a curve, defined by parametric equations one assigns different values to the parameter and plots the points having as coordinates the values of x and y thus obtained; it is not necessary first to eliminate the parameter from the two equa- tions and then plot the resulting equation in x and y (see Ex. 11). When equations (5.7) are written in the form (6.8) x=(l-O)utte, y=(1-)n+ty2, we see that y is equal to the sum of the same multiples of 1 and ye as x is of x, and x2. We say that x and y have the same linear and homogeneous expressions in x,, x2 and yi, y2 respec- tively, meaning that every term in each expression is of the first degree in these quantities. Accordingly we have [5.4] Any point of a line is expressible linearly and homogene- ously in terms of two fixed points of the line. Any two points of the line can be used for P, and Ps, that is, as the basis for writing the equations of the line in the form (5.8). The above theorem is the geometric equivalent of the algebraic statement [5.5] Any solution of an equation ax + by +c =0 is expressible linearly and homogeneously in terms of any two solutions. 25Points and Lines in the Plane [Chap. 1 EXERCISES 1. Obtain equations of the lines determined by the points: a. (— 1, 8) and (4, — 2). b. (3, 2) and (3, — 1). 2, Show that the points (4, — 3), (2, 0), (— 2, 6) lie on a line. 8. Find c so that the points (c, — 2), (3, 1), and (— 2, 4) shall be collinear, that is, lie on a line. 4. Where are the points for which x > 0, y>0, andx+y<1? Where are the points for which 2x —3y> 2? 5. Show that the equation of the second and third ratios in (4.3) is reducible to equation (5.3). What does this mean? 6. Show that (4.8) are of the form (5.8) with t= hy/(i1 + he). 7. In what ratio is the line segment with end points (2, 3) and (3, — 1) divided by the point of intersection of the line segment and the line 3x— 2 y—6=0? 8. Show that the equation a(x — x1) + b(y — 91) =0 for each set of values of @ and b is an equation of a line through the point (x, 91). For what values of a@ and 6 is it an equation of a line parallel to the y-axis ; an equation of the line through the point (x2, y2)? 9. Show that the line ax + by + c= 0 meets the x-axis and y-axis in the points P;(— c/a, 0) and P2(0, — c/b) respectively; when c x 0, the lengths OP; and OP2, that is, —c/a@ and —c/b, are called the x-intercept and y-intercept respectively of the line. 10. Show that when an equation of a line is written in the form Zyvn ath 1, where g and hare constants, g and hare the x- and y-intercepts respec- tively. When can an equation of a line not be given this form, which is called the equation of the line in the intercept form? 11. Plot the curve with the parametric equations x = 2 2, y=3t; also the curve x= 5cost, y=5sin#, using a table of natural sines and cosines. What is the equation of the curve in each case when the parameter is eliminated? 12. Show that, if the expressions (4.8) for x and y are substituted in an equation ax + by+c=0, the value of f/he obtained from the result is the ratio in which the segment P,P is divided by the point of intersection of the line ax + by + c= 0 and the line through P; and P2, The reader should apply this method to Ex. 7. 26sec. 6} Direction Numbers and Direction Cosines of a Line 6. Direction Numbers and Direction Cosines of a Line. Angle between Two Lines The equation (6.1) “ X— eH _ YN %2— kh Yen of the line through the points (x1, v1) and (x2, y2), obtained in § 5, expresses the condition that the direction numbers of the segments P,P and P;P2 are proportional when P is any point on the line. In fact, in Fig. 6 the sensed segments P:Q and QP give the direction numbers of P,P, and P:Q2 and QeP» the di- rection numbers of P:P2. Thus equation (6.1) is the algebraic statement of the characteristic property of a line that any two segments having an end point in common have the same direc- tion. If we say that direction numbers of any segment of a line are direction numbers of the line, it follows that a line has an endless number of sets of direction numbers, but that the numbers of any set are proportional to those of any other set. The equation (6.2) x= _Y—N, 4 0 which is satisfied by x: and y,, being of the first degree in x and y, is an equation of a line through the point (x1, 1) and with direction numbers u and »v. In order to find a segment of which u and v are direction numbers, we have only to find x2 and y2 from the equations X2—- =U, Yo—y=0. Then P;P2 is a segment with end points Pi(x1, y1) and P2(x2, y2) having u and » as direction numbers; and in terms of x2 and yz equation (6.2) becomes equation (6.1). Since parallel lines have the same direction by definition, it follows that [6.1] If (x1, 91) and (x2, y2) are any two points on a line, x2— 1 and y2— 41 are direction numbers of the line, and also of any line parallel to it. In defining the direction cosines of a line segment in § 3 we assigned sense to a segment, making the convention that for 27Points and Lines in the Plane [Chap. 1 a segment not parallel to the x-axis upward along the segment is positive and downward negative; and for a segment parallel to the x-axis the positive sense is to the right along the seg- ment. Accordingly the positive sense along a line is upward, that is, y increasing, when the line is not parallel to the x-axis; and to the right, that is, x increasing, when the line is parallel to the x-axis. Since all segments of a line have the same direction, the direction cosines of all segments are the same; we call them the direction cosines of the line. Accord- ingly the direction cosines \, u of a line are the cosines of the angles between the positive direction of the line and the positive directions of the x- and y-axes respectively. From Theorem [8.2] we have [6.2] The direction cosines d, u of a line satisfy the equation (6.3) M+ w=1; bis never negative; when u>0, -1 ab. What are the slopes of these lines? 10. Draw the graph of a line and indicate the various quantities appearing in equations (6.1), (6.2), (6.5), and (7.2). 8. Directed Distance from a Line to a Point When a point Pi(x, 1) is on the line (8.1) ax +by+c=0, the expression ax, + by; + is equal to zero, as we have seen. When P, is not on the line, this expression has a value different from zero. One might question whether perhaps this value has something to do with the distance from the line to the point. We shall answer this question by showing that this distance is ax + by +c, . . oiela LE and give a meaning to the algebraic sign of the resulting number. When a=0 and 50 in (8.1), that is, when the line is parallel to the x-axis, the value of y for each point on the line is —c/b, and consequently the distance of Pi from the line is 1 — (— c/b), that is, 9: + c/b, and this distance is positive or negative according as P; lies above or below the line; we 35Points and Lines in the Plane (Chap. 1 get this expression for the distance when we put a =0 in the expression at the close of the preceding paragraph. When 6=0 and a+ 0, that is, when the line is parallel to the y-axis, the distance is x, -++¢/a, which is positive or negative according as P, lies to the right or left of the line; this expres- sion follows from the one of the preceding paragraph on putting b= 0. We take up next the case when the line is inclined to the x-axis, in which case a ~ 0, b # 0. In § 6 we made the conven- tion that the direction upward along a line inclined to the axes is positive, and showed that the direction cosine yu is always positive, and that A is positive or negative according as the line is inclined as in Fig. 7 or in Fig. 8. In consequence of this result, and of Theorems [6.9] and [6.4], we have that the direction cosines of a line perpendicular to the line ax + by+c=0 are given by a 6 2) N= Nae ee where ¢ is +1 or —1 so that eb > 0 when b¥ 0, and ea>0 when b = 0. We consider first the case when the line (8.1) is inclined asin Fig. 10, where Pi(x1, y1) isa point above the line and P2(x2, yz) is the point in which the perpendicular through P; to the line meets it. We denote by d the distance from Pz to Pi, d Sy—> being a positive number, since it is measured in the positive direction Fic. 10 of the line P2P;. Comparing this line in Fig. 10, whose direction cosines are \ and yw, with the line P:P in Fig. 7, we see that in place of (6.5) we have Pi (21,91) (8.3) Mm—x=d, N—-Jo= dy, from which and (8.2) we have da db CN BeNOR mae 36