(eBook PDF) Numerical Mathematics

and Computing 7th Edition by E. Ward

Cheney
Go to download the full and correct content document:
https://ebooksecure.com/product/ebook-pdf-numerical-mathematics-and-computing-7
th-edition-by-e-ward-cheney/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

(eBook PDF) Numerical Analysis: Mathematics of

Scientific Computing 3rd

http://ebooksecure.com/product/ebook-pdf-numerical-analysis-
mathematics-of-scientific-computing-3rd/

(eBook PDF) Introduction to Probability by Mark Ward

http://ebooksecure.com/product/ebook-pdf-introduction-to-
probability-by-mark-ward/

(eBook PDF) Environmental Hydrology 3rd Edition by Andy

D. Ward

http://ebooksecure.com/product/ebook-pdf-environmental-
hydrology-3rd-edition-by-andy-d-ward/

(eBook PDF) Numerical Methods for Engineers 7th Edition

http://ebooksecure.com/product/ebook-pdf-numerical-methods-for-
engineers-7th-edition/
(eBook PDF) Business Mathematics and Statistics, 7th
Edition

http://ebooksecure.com/product/ebook-pdf-business-mathematics-
and-statistics-7th-edition/

(eBook PDF) Chemistry 7th Edition by John E. McMurry

http://ebooksecure.com/product/ebook-pdf-chemistry-7th-edition-
by-john-e-mcmurry/

(eBook PDF) Numerical Analysis 10th Edition by Richard

http://ebooksecure.com/product/ebook-pdf-numerical-analysis-10th-
edition-by-richard/

(eBook PDF) Introductory Technical Mathematics 7th

Edition

http://ebooksecure.com/product/ebook-pdf-introductory-technical-
mathematics-7th-edition-2/

(eBook PDF) Introductory Technical Mathematics 7th

Edition

http://ebooksecure.com/product/ebook-pdf-introductory-technical-
mathematics-7th-edition/
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 vi

vi Contents

Double-Precision Floating-Point Form 43

Computer Errors in Representing Numbers 46
Notation fl(x) and Backward Error Analysis 47
Historical Notes 51
Summary 1.3 51
Exercises 1.3 52
Computer Exercises 1.3 54
1.4 Loss of Significance 56
Significant Digits 56
Computer-Caused Loss of Significance 57
Theorem on Loss of Precision 59
Avoiding Loss of Significance in Subtraction 60
Range Reduction 62
Summary 1.4 63
Exercises 1.4 63
Computer Exercises 1.4 65

2 Linear Systems 69
2.1 Naive Gaussian Elimination 69
A Larger Numerical Example 71
Algorithm 72
Pseudocode 75
Testing the Pseudocode 77
Residual and Error Vectors 79
Summary 2.1 79
Exercises 2.1 80
Computer Exercises 2.1 81
2.2 Gaussian Elimination with Scaled Partial Pivoting 82
Naive Gaussian Elimination Can Fail 82
Partial Pivoting and Full (Complete) Pivoting 84
Gaussian Elimination with Scaled Partial Pivoting 85
A Larger Numerical Example 87
Pseudocode 89
Long Operation Count 92
Numerical Stability 93
Scaling 94
Variants of Gaussian Eliminations 94
Condition Number 97
Backslash Operator in MATLAB 97
Summary 2.2 97
Exercises 2.2 98
Computer Exercises 2.2 100
2.3 Tridiagonal and Banded Systems 103
Tridiagonal Systems 103
Strictly Diagonal Dominance 106

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 vii

Contents vii

Pentadiagonal Systems 107

Block Pentadiagonal Systems 108
Summary 2.3 109
Exercises 2.3 110
Computer Exercises 2.3 110

3 Nonlinear Equations 114

3.1 Bisection Method 114
Introduction 114
Bisection Algorithm 116
Pseudocode 116
Numerical Examples 117
Convergence Analysis 119
False Position (Regula Falsi) Method 121
Modified False Position Method 121
Summary 3.1 122
Exercises 3.1 123
Computer Exercises 3.1 124
3.2 Newton’s Method 125
Interpretations of Newton’s Method 126
Pseudocode 127
Illustration 128
Convergence Analysis 129
Discussion of Newton’s Method 130
Systems of Nonlinear Equations 132
Fractal Basins of Attraction 134
Summary 3.2 135
Exercises 3.2 136
Computer Exercises 3.2 138
3.3 Secant Method 142
Secant Algorithm 143
Convergence Analysis 144
Comparison of Methods 147
Hybrid Schemes 148
Fixed-Point Iteration 148
Summary 3.3 149
Exercises 3.3 149
Computer Exercises 3.3 150

4 Interpolation and Numerical Differentiation 153

4.1 Polynomial Interpolation 153
Preliminary Remarks 153
Polynomial Interpolation 154

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 viii

viii Contents

Interpolating Polynomial: Lagrange Form 155

Existence of Interpolating Polynomial 156
Interpolating Polynomial: Newton Form 157
Nested Form 158
Calculating Coefficients ai Using Divided Differences 160
Algorithms and Pseudocode 165
Vandermonde Matrix 167
Inverse Interpolation 169
Polynomial Interpolation by Neville’s Algorithm 170
Interpolation of Bivariate Functions 172
Summary 4.1 173
Exercises 4.1 174
Computer Exercises 4.1 177
4.2 Errors in Polynomial Interpolation 178
Introduction 178
Dirichlet Function 179
Runge Function 179
Theorems on Interpolation Errors 181
Summary 4.2 185
Exercises 4.2 186
Computer Exercises 4.2 187
4.3 Estimating Derivatives and Richardson Extrapolation 187
First-Derivative Formulas via Taylor Series 188
Richardson Extrapolation 190
First-Derivative Formulas via Interpolation Polynomials 194
Second-Derivative Formulas via Taylor Series 196
Noise in Computation 197
Summary 4.3 197
Exercises 4.3 198
Computer Exercises 4.3 200

5 Numerical Integration 201

5.1 Trapezoid Method 201
Definite and Indefinite Integrals 201
Trapezoid Rule: Nonuniform Spacing 202
Composite Trapezoid Rule: Uniform Spacing 203
Pseudocode 204
Error Analysis 205
Applying the Error Term 208
Recursive Trapezoid Formula 210
Multidimensional Integration 211
Remarks 212
Summary 5.1 213

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 ix

Contents ix

Exercises 5.1 214

Computer Exercises 5.1 216

5.2 Romberg Algorithm 217

Description 217
Pseudocode 218
Euler-Maclaurin Formula 220
General Extrapolation 222
Summary 5.2 224
Exercises 5.2 224
Computer Exercises 5.2 226

5.3 Simpson’s Rules and Newton-Cotes Rules 227

Basic Trapezoid Rule 227
Basic Simpson’s Rule 228
Basic Simpson’s Rule: Uniform Spacing 229
Composite Simpson’s Rule 231
An Adaptive Simpson’s Scheme 231
Example Using Adaptive Simpson Procedure 234
Newton-Cotes Rules 235
Some Closed Newton-Cotes Rules with Error Terms 235
Some Open Newton-Cotes Rules with Error Terms 236
Summary 5.3 236
Exercises 5.3 237
Computer Exercises 5.3 238

5.4 Gaussian Quadrature Formulas 239

Description 239
Change of Intervals 240
Gaussian Quadrature Rules 241
Legendre Polynomials 243
Gaussian Quadrature Nodes and Weights 244
Integrals with Singularities 246
Summary 5.4 246
Exercises 5.4 248
Computer Exercises 5.4 249

6 Spline Functions 252

6.1 First-Degree and Second-Degree Splines 252
First-Degree Splines 252
Modulus of Continuity 255
Second-Degree Splines 256
Interpolating Quadratic Splines 257
Subbotin Quadratic Splines 258

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 x

x Contents

Summary 6.1 260

Exercises 6.1 261
Computer Exercises 6.1 263
6.2 Natural Cubic Splines 263
Introduction 263
Natural Cubic Splines 265
Algorithm for Natural Cubic Splines 266
Algorithm for Solving Tridiagonal System 269
Pseudocode for Natural Cubic Splines 271
Using Pseudocode for Interpolating and Curve Fitting 272
Space Curves 273
Smoothness Property 274
Summary 6.2 276
Exercises 6.2 277
Computer Exercises 6.2 280
6.3 B Splines: Interpolation and Approximation 281
Introduction 281
Bi0 Splines 281
Bi1 and Bik Splines 282
Linear Combination of Bik Splines 283
Derivatives of B Splines 284
Integration of B Splines 285
Interpolation and Approximation by B Splines 286
Higher-Degree B Splines 287
Pseudocode and a Curve-Fitting Example 288
Schoenberg’s Process 289
Pseudocode: Schoenberg’s Process 290
Bézier Curves 291
Summary 6.3 293
Exercises 6.3 295
Computer Exercises 6.3 297

7 Initial Values Problems 299

7.1 Taylor Series Methods 299
Initial-Value Problem: Analytical versus Numerical Solution 299
An Example of a Practical Problem 301
Solving Differential Equations and Integration 302
Vector Fields 303
Taylor Series Methods 304
Euler’s Method and Pseudocode 305
Taylor Series Method of Higher Order 306
Pseudocode and Results 307
Types of Errors 307
Taylor Series Method Using Symbolic Computations 308
Summary 7.1 308

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xi

Contents xi

Exercises 7.1 309

Computer Exercises 7.1 310

7.2 Runge-Kutta Methods 311

Introduction 311
Taylor Series for f (x, y) 311
Runge-Kutta Method of Order 2 312
Runge-Kutta Method of Order 4 314
Pseudocode 315
Summary 7.2 316
Exercises 7.2 316
Computer Exercises 7.2 318

7.3 Adaptive Runge-Kutta and Multistep Methods 320

An Adaptive Runge-Kutta-Fehlberg Method 320
Pseudocode 321
An Industrial Example 323
Adams-Bashforth-Moulton Formulas 324
Stability Analysis 325
Summary 7.3 328
Exercises 7.3 328
Computer Exercises 7.3 329

7.4 Methods for First and Higher-Order Systems 331

Uncoupled and Coupled Systems 332
Taylor Series Method 332
Vector Notation 334
Systems of ODEs 334
Taylor Series Method: Vector Notation 334
Runge-Kutta Method 335
Pseudocode 336
Autonomous ODE 337
Higher-Order Differential Equations and Systems 339
Higher-Order Differential Equations 339
Systems of Higher-Order Differential Equations 341
Autonomous ODE Systems 341
Summary 7.4 342
Exercises 7.4 344
Computer Exercises 7.4 346
7.5 Adams-Bashforth-Moulton Methods 347
A Predictor-Corrector Scheme 347
Pseudocode 348
An Adaptive Scheme 352
An Engineering Example 352
Stiff ODEs and an Example 353
History of ODE Numerical Methods 355
Summary 7.5 355
Exercises 7.5 355
Computer Exercises 7.5 356

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xii

xii Contents

8 More on Linear Systems 358

8.1 Matrix Factorizations 358
LU Factorization 358
Numerical Example 359
Formal Derivation 361
Pseudocode 365
Solving Linear Systems Using LU Factorization 365
LDLT Factorization 367
Cholesky Factorization 369
Multiple Right-Hand Sides 371
Computing A−1 371
Example Using Software Packages 372
Summary 8.1 373
Exercises 8.1 375
Computer Exercises 8.1 378
8.2 Eigenvalues and Eigenvectors 380
Av Scalar Multiple of v 380
Calculating Eigenvalues and Eigenvectors 381
Mathematical Software 382
Properties of Eigenvalues 383
Gershgorin’s Theorem 385
Singular Value Decomposition 387
Numerical Examples of Singular Value Decomposition 389
Application: Linear Differential Equations 391
Application: A Vibration Problem 392
Summary 8.2 393
Exercises 8.2 394
Computer Exercises 8.2 395
8.3 Power Method 396
Mathematical Derivation 396
Power Method Pseudocode 398
Aitken Acceleration 399
Inverse Power Method 400
Example: Inverse Power Method 401
Shifted (Inverse) Power Method 401
Example: Shifted Inverse Power Method 402
Summary 8.3 402
Exercises 8.3 403
Computer Exercises 8.3 404
8.4 Iterative Solutions of Linear Systems 405
Vector and Matrix Norms 405
Condition Number and Ill-Conditioning 406
Basic Iterative Methods 408
Pseudocode 412
Convergence Theorems 414

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xiii

Contents xiii

Matrix Formulation of Iterative Methods 416

Another View of Overrelaxation 417
Conjugate Gradient Method 417
CG Pseudocode and Features 419
Summary 8.4 421
Exercises 8.4 422
Computer Exercises 8.4 423

9 Least Squares Methods and Fourier Series 426

9.1 Method of Least Squares 426
Linear Least Squares 426
Linear Example 429
Nonpolynomial Example 430
Basis Functions {g0 , g1 , . . . , gn } 431
Summary 9.1 432
Exercises 9.1 433
Computer Exercises 9.1 434
9.2 Orthogonal Systems and Chebyshev Polynomials 435
Orthonormal Basis Functions {g0 , g1 , . . . , gn } 435
Outline of Algorithm 438
Smoothing Data: Polynomial Regression 439
A Procedure for Polynomial Regression 440
Least-Squares Problem Revisited 443
Gram-Schmidt Process 444
Summary 9.2 445
Exercises 9.2 446
Computer Exercises 9.2 447
9.3 Examples of the Least-Squares Principle 447
Inconsistent Systems 447
Modified Gram-Schmidt Process 448
Use of a Weight Function w(x) 448
Nonlinear Example 449
Linear and Nonlinear Example 450
Additional Details on SVD 451
Using the Singular Value Decomposition 453
Examples 455
Orthogonal Matrices and Spectral Theorem 456
Summary 9.3 456
Exercises 9.3 457
Computer Exercises 9.3 458
9.4 Fourier Series 459
Introduction 459
Least-Squares Approximation 459
Orthogonality Properties 460

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xiv

xiv Contents

Standard Integrals 461

Trigonometric Polynomial 462
Cosine Series and Sine Series 463
Sawtooth Wave Function on [−π, π] 464
Infinite Fourier Series 465
Periodic on [−P/2, P/2] 466
Periodic on [−L , L] 466
Periodic on [0, 2L] 466
Fourier Series and Music 467
Sawtooth Wave Function on [0, 2L] 467
Fourier Series Examples 468
Euler’s and de Moivre Formulas 469
Complex Fourier Series 470
Roots of Unity 471
Discrete Fourier Transform 472
Fast Fourier Transforms 473
Mathematical Software 474
Summary 9.4 475
Exercises 9.4 478
Computer Exercises 9.4 479

10 Monte Carlo Methods and Simulation 481

10.1 Random Numbers 481
Random-Number Algorithms and Generators 482
Examples 484
Uses of Pseudocode Random 486
Summary 10.1 489
Exercises 10.1 490
Computer Exercises 10.1 490
10.2 Estimation of Areas and Volumes by
Monte Carlo Techniques 491
Numerical Integration 491
Example and Pseudocode 493
Computing Volumes 494
Ice Cream Cone Example 495
Summary 10.2 495
Exercises 10.2 496
Computer Exercises 10.2 496
10.3 Simulation 498
Loaded Die Problem 498
Birthday Problem 499
Buffon’s Needle Problem 501
Two Dice Problem 502
Neutron Shielding Problem 502

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xv

Contents xv

Summary 10.3 504

Computer Exercises 10.3 504

11 Boundary-Value Problems 507

11.1 Shooting Method 507
Introduction 507
Shooting Method 508
Shooting Method Algorithm 508
Modifications and Refinements 510
Summary 11.1 511
Exercises 11.1 512
Computer Exercises 11.1 513
11.2 A Discretization Method 513
Finite-Difference Approximations 513
The Linear Case 514
Numerical Example and Pseudocode 515
Pseudocode 515
Shooting Method in the Linear Case 516
Numerical Example Revisited 517
Pseudocode 518
Summary 11.2 520
Exercises 11.2 521
Computer Exercises 11.2 522

12 Partial Differential Equations 524

12.1 Parabolic Problems 524
Some Partial Differential Equations from Applied Problems 524
Heat Equation Model Problem 526
Finite-Difference Method 527
Pseudocode for Explicit Method 528
Crank-Nicolson Method 529
Pseudocode for the Crank-Nicolson Method 530
Alternative Version of the Crank-Nicolson Method 531
Stability 532
Summary 12.1 534
Exercises 12.1 535
Computer Exercises 12.1 536
12.2 Hyperbolic Problems 536
Wave Equation Model Problem 536
Analytic Solution 537
Numerical Solution 538

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xvi

xvi Contents

Pseudocode 539
Advection Equation 541
Lax Method 541
Upwind Method 541
Lax-Wendroff Method 542
Summary 12.2 542
Exercises 12.2 543
Computer Exercises 12.2 543
12.3 Elliptic Problems 544
Helmholtz Equation 544
Finite-Difference Method 544
Gauss-Seidel Iterative Method 548
Numerical Example and Pseudocode 549
Finite-Element Methods 551
More on Finite Elements 555
Summary 12.3 557
Exercises 12.3 558
Computer Exercises 12.3 559

13 Minimization of Functions 561

13.1 One-Variable Case 561
Unconstrained and Constrained Minimization Problems 561
One-Variable Case 562
Unimodal Functions F 563
Fibonacci Search Algorithm 563
Golden Section Search Algorithm 566
Quadratic Interpolation Algorithm 568
Summary 13.1 571
Exercises 13.1 571
Computer Exercises 13.1 572
13.2 Multivariate Case 573
Taylor Series for F: Gradient Vector and Hessian Matrix 573
Alternative Form of Taylor Series 575
Steepest Descent Procedure 576
Contour Diagrams 577
More Advanced Algorithms 577
Minimum, Maximum, and Saddle Points 579
Positive Definite Matrix 580
Quasi-Newton Methods 580
Nelder-Mead Algorithm 580
Method of Simulated Annealing 581
Summary 13.2 582
Exercises 13.2 583
Computer Exercises 13.2 585

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xvii

Contents xvii

14 Linear Programming Problems 587

14.1 Standard Forms and Duality 587
First Primal Form 587
Numerical Example 588
Transforming Problems into First Primal Form 590
Dual Problem 590
Second Primal Form 592
Summary 14.1 593
Exercises 14.1 594
Computer Exercises 14.1 596
14.2 Simplex Method 597
Vertices in K and Linearly Independent Columns of A 597
Simplex Method 599
Summary 14.2 600
Exercises 14.2 600
Computer Exercises 14.2 601
14.3 Inconsistent Linear Systems 601
ℓ1 Problem 601
ℓ∞ Problem 604
Summary 14.3 605
Exercises 14.3 607
Computer Exercises 14.3 607

Appendix A Advice on Good Programming Practices 608

A.1 Programming Suggestions 608
Case Studies 611
On Developing Mathematical Software 615

Appendix B Representation of Numbers in Different Bases 616

B.1 Representation of Numbers in Different Bases 616
Base β Numbers 617
Conversion of Integer Parts 617
Conversion of Fractional Parts 619
Base Conversion 10 ↔ 8 ↔ 2 620
Base 16 622
More Examples 622
Summary B.1 623
Exercises B.1 623
Computer Exercises B.1 624

Appendix C Additional Details on IEEE Floating-Point Arithmetic 626

C.1 More on IEEE Standard Floating-Point Arithmetic 626

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xviii

xviii Contents

Appendix D Linear Algebra Concepts and Notation 629

D.1 Elementary Concepts 629
Vectors 629
Matrices 631
Matrix-Vector Product 634
Matrix Product 634
Other Concepts 636
Cramer’s Rule 638

Answers for Selected Exercises 639

Bibliography 657

Index 665

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xix

Preface

Our basic objective is to acquaint students of science and engineering with the potentiali-
ties of using computers for solving numerical problems that may arise in their professions.
A secondary objective is to give students an opportunity to hone their skills in computer
programming and problem solving. A final objective is to help students arrive at an under-
standing of the important subject of errors that inevitably arise in scientific computing as
well as learning methods for detecting, predicting, and controlling them.
Much of science today involves complex computations built upon mathematical soft-
ware systems. The users may have little knowledge of the underlying numerical algorithms
used in these problem-solving environments. By studying numerical methods, one can be-
come a more informed user and be better prepared to evaluate and judge the accuracy of the
results. Students should study mathematical algorithms to learn not only how they work, but
also how they can fail! Critical thinking and constant skepticism are traits we want students
to acquire. An extensive numerical calculation, even when carried out by state-of-the-art
software, may need to be subjected to independent verification, if possible.
We have tried to achieve an elementary style of presentation since we want this book
to be accessible to students who may not be advanced in their formal study of mathematics
and computer sciences. Toward this end, we have provided numerous examples and figures
for illustrative purposes and fragments of pseudocode, which are informal descriptions of
computer algorithms.
Believing that most students at this level need a survey of the subject of numerical
mathematics and computing, we have presented a wide diversity of topics, including some
rather advanced ones that play an important role in scientific computing. We recommend
that the reader have at least a one-year study of calculus, plus some basic knowledge of
matrices, vectors, and ordinary differential equations.

Seventh Edition Features

Following suggestions and comments from the reviewers and based on our experience in
teaching this material, we have revised and enhanced the entire book.
• Some chapters have been combined and the order of others changed. Also, the Linear
Systems chapter has been moved earlier because it is a topic that arises frequently and
needs to be presented sooner.
• A new section on Fourier Series has been added because it arises in various engineering
applications. The section on lower and upper sums in numerical integrations has been
removed.
• Now displayed in double columns, the problems have been renamed as exercises to clarify
that they are intended for practice and for further learning.
• In an effort to make the new edition more student friendly, margin notes have been added
as well as additional figures, tables, examples, and exercises.

xix

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xx

xx Preface

Suggestions for Use

Numerical Mathematics and Computing, Seventh Edition, can be used in a variety of ways,
depending on the emphasis the instructor prefers and the inevitable time constraints. Exer-
cises have been supplied in abundance to enhance the book’s versatility. They are divided
into two categories: Exercises and Computer Exercises. In the first category, there are more
than 800 exercises in analysis that require pencil, paper, and possibly a calculator. In the
second category, there are approximately 500 computer exercises involving programming
and using a computer. Students are asked to solve some exercises using advanced software
systems such as MATLAB, Mathematica, or Maple. While students may be asked to write
their own computer codes, they can often follow a model or example to assist them, but in
other cases they must proceed on their own from a mathematical description.
In some of the Computer Exercises, there is something to be learned beyond simply
writing code—a moral, if you like. For example, this can happen if the exercise being solved
and the pseudocode provided are somehow mismatched. Some Computing Exercises are
designed to give the students experience in using, mathematical software.
The Student Solution Manual is sold as a separate publication. Also, teachers who adopt
the book can obtain the Instructor Solution Manual from the publisher or via the Instructor
Companion website. Sample programs based on the pseudocode have been coded in several
programming languages and are on the textbook website.
To access course materials and companion resources, please visit the websites on p. iv;
e.g., www.cengagebrain.com. At the CengageBrain.com home page, one can search
using either the International Standard Book Number (ISBN), the title of the book, or the
last name of one of the authors. This takes you to the product page, where free companion
resources can be found. The authors have established the following website:
www.ma.utexas.edu/CNA/NMC7/
The arrangement of chapters reflects our own view of how the material might best un-
fold for a student new to the subject. In some cases, there may be little mutual dependence
among the chapters, and the instructor can order the sequence of presentation in various
ways. Certainly, instructor may omit some sections and chapters for want of time.
Our own recommendations for courses based on this text are as follows:
• A one-term course carefully covering Chapters 1 through 7 (possibly omitting Sections
4.2, 6.3, and 7.3–5, for example), followed by a selection of material from the remaining
chapters as time permits.
• A one-term survey rapidly skimming over almost all of chapters and omitting some of
the more advanced material.
• A two-term course carefully covering all chapters.

Student Research Projects

Throughout there are some Computer Exercises designated as Student Research Projects.
These are opportunities for students to explore topics beyond the scope of the textbook.
Many of these involve application areas for numerical methods. These projects usually
include computer programming and numerical experiments. A favorable aspect of these
assignments is to allow students to choose a topic of interest to them, possibly something
that may arise in their future profession or their major study area. For example, any topic
suggested may be delved into more deeply by consulting other texts and references. In

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xxi

Preface xxi

preparing such a project, the students have to learn about the topic, locate the significant
references (books, research papers, and websites), do the computing, and write a report that
explains all this in a coherent way. Students can avail themselves of mathematical software
systems such as MATLAB, Maple, or Mathematica, as well as doing their own computer
programming in whatever programming language they prefer.

Acknowledgments
We have profited from advice and suggestions kindly offered by a large number of col-
leagues, students, and users of the previous editions.
We are grateful to have had the opportunity and privilege of teaching classes in scientific
computing, numerical analysis, and various other topics in the Department of Mathematics
and the Department of Computer Sciences of The University of Texas at Austin. Without
their support and the use of their computing facilities, writing this book would not have
been possible. In particular, we thank Maorong Zou and Margaret Combs for help with
computer issues.
Recently, the second author taught classes in the Department of Petroleum Engineering
and Geosystems Engineering of The University of Texas at Austin and in the Applied
Mathematics and Computer Sciences Division of the King Abdullah Univeristy of Science
and Technology in Saudi Arabia and wishes to thank them both.
Valuable comments and suggestions were made by our colleagues and friends. In par-
ticular, we miss our friend David Young who was always very generous with suggestions
for improving the accuracy and clarity of the exposition in previous editions. Some parts
of those editions were typed with great care and attention to detail by Sheri Brice, Katy
Burrell, Kata Carbone, Margaret Combs, and Belinda Trevino. Aaron Naiman was partic-
ularly helpful in sharing material he prepared for his courses.
We wish to acknowledge the reviewers who have provided detailed critiques for this
new edition: Eugino Aulisa, Texas Tech University; Erin Bach, University of Wiscon-
sin, Madison; Marcin Bownik, University of Oregon; Olga Brezhneva, Miami Univer-
sity; George Grossman, Central Michigan University; Luke Olson, University of Illinois at
Urbana-Champaign; Ronald Taylor, Wright State University.
Reviewers from previous editions were Krishan Agrawal, Eric Back, Neil Berger,
Thomas Boger, Marcin Bownik, Olga Brezhneva, Jose E. Castillo, Charles Collins, Charles
Cullen, Elias Y. Deeba, F. Emad, Gentil A. Estévez, Terry Feagin, Jose Flores, Leslie
Foster, Bob Funderlic, Mahadevan Ganesh, William Gearhart, Juan Gil, John Gregory,
George Grossman, Bruce P. Hillam, Patrick Lang, Ren Chi Li, Wu Li, Xiaofan Li, Vania
Mascioni, Bernard Maxum, Edward Neuman, Roy Nicolaides, Luke Olson, Amar Raheja,
J. N. Reddy, Daniel Reynolds, Asok Sen, Ching-Kuang Shene, William Slough, Ralph
Smart, Thiab Taha, Ronald Taylor, Jin Wang, Stephen Wirkus, Marcus Wright, Quiang Ye,
Tjalling Ypma, and Shangyou Zhan.
Many individuals took the time to write us with suggestions and criticisms. We are
grateful to the following individuals and others who have send us e-mails concerning the
textbook or solution manuals: A. Aawwal, Nabeel S.Abo-Ghander, Krishan Agrawal, Roger
Alexander, Husain Ali Al-Mohssen, Kistone Anand, Keven Anderson, Vladimir Andrije-
vik, Jon Ashland, Hassan Basir, Steve Batterson, Neil Berger, Adarsh Beohar, Bernard
Bialecki, Jason Brazile, Keith M. Briggs, Carl de Boor, Jose E. Castillo, Fatih Celiker,
Debao Chen, Ellen Chen, Hwen Chin, Edmond Chow, Lloyd Clark, John Cook, Brad
Copper, Roger Crawfis, Charles Cullen, Antonella Cupillari, Jonathan Dautrich, James
Arthur Davis, Tim Davis, Elias Y. Deeba, Suhrit Dey, Alan Donoho, Jason Durheim, Wayne

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xxii

xxii Preface

Dymacek, John Eisenmenger, Fawzi P. Emad, Paul Enigenbury, Terry Feagin, Leslie Foster,
Peter Fraser, Richard Gardner, Mohamad El Gharamti, John Gregory, Katherine Hua Guo,
Scott Hagerup, Kent Harris, Scott Henry, Bruce P. Hillam, Tom Hogan, Jackie Hohnson,
Christopher M. Hoss, Jason S. Howel, Kwang-il In, Victoria Interrante, Sadegh Jokar, Erni
Jusuf, Jason Karns, Jacob Y. Kazakia, Grant Keady, Achim Kehrein, Jacek Kierzenka,
Daniel Kopelove, S. A. (Seppo) Korpela, Andrew Knyazev, Gary Krenz, Jihoon Kwak,
Kim Kyungjin, Minghorng Lai, Patrick Lang, Kevin Lee, Wu Li, Grace Liu, Wenguo Liu,
Stacy Long, Mark C. Malburg, Igor Malkiman, P. W. Manual, Hamidreza Mashalyekh,
Peter McNamara, Juan Meza, F. Milianazzo, Milan Miklavcic, Sue Minkoff, George Minty,
Baharen Momken, Justin Montgomery, Ramon E. Moore, Harunrashid Muhammad, Aaron
Naiman, Asha Nallana, Edward Neuman, Durene Ngo, Roy Nicolaides, Jeff Nunemacher,
Valia Guerra Ones, David Parker, Tony Praseuth, Rolfe G. Petschek, Terri Prakash, Mi-
haela Quirk, Helia Niroomand Rad, Jeremy Rahe, Frank Roberts, Frank Rogers, Simen
Rokaas, Hossein Roodi, Robert S. Raposo, Chris C. Seib, Granville Sewell, Keh-Ming
Shyue, Daniel Somerville, Nathan Smith, Mandayam Srinivas, Alexander Stromberger,
Xingping Sun, Thiab Taha, Hidajaty Thajeb, Joseph Traub, Phuoc Truong, Vincent Tsao,
Bi Roubolo Vona, David Wallace, Charles Walters, Kegnag Wang, Layne T. Watson, Andre
Weideman, Perry Wong, Richard Fa Wai, Yuan Xu, and Rick Zaccone.
It is our pleasure to thank those who helped with the task of preparing the new edition.
The staff of Cengage Learning and associated individuals have been most understanding
and patient in bringing this book to fruition. In particular, we thank Shaylin Walsh, Alison
Eigel Zade, Charu Khanna, and Christine Sabooni for their efforts on behalf of this project.
Some of those who were involved with previous editions were Seema Atwal, Craig Barth,
Carol Benedict, Stacy Green, Jeremy Hayhurst, Janet Hill, Cheryll Linthicum, Gary Ostedt,
Merrill Peterson, Bob Pirtle, Sara Planck, Elizabeth Rammel, Ragu Raghavan, Elizabeth
Rodio, Anne Seitz, and Marlene Thom.
We offer our heartfelt gratitude to Victoria Cheney, Joyce Pfluger, and Martha Wells.
We would appreciate any comments, questions, suggestions, or corrections that readers
may wish to communicate to us using the e-mail address below.

Ward Cheney & David Kincaid

kincaid@ices.utexas.edu

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Cheney-1010099 03714˙fm˙ptg01˙hires April 3, 2012 14:46 xxiii

Dedication

In memory of
our friend and colleague

David M. Young, Jr.

(“Dr. SOR”)
(1923–2008)

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
Another random document with
no related content on Scribd:
Cimbex she finds a similar arrangement, but there are ten chambers,
and no aorta.

The dorsal vessel is connected with the roof of the body by some
short muscles, and is usually much surrounded by fat-body into
which tracheae penetrate; by these various means it is kept in
position, though only loosely attached; beneath it there is a delicate,
incomplete or fenestrate, membrane, delimiting a sort of space
called the pericardial chamber or sinus; connected with this
membrane are some very delicate muscles, the alary muscles,
extending inwards from the body wall (b, Fig. 72): the curtain formed
by these muscles and the fenestrate membrane is called the
pericardial diaphragm or septum. The alary muscles are not directly
connected with the heart.

Fig. 72.—Dorsal vessel (c), and alary muscles (b), of Gryllotalpa (after
Graber); a, aorta. N.B.—The ventral aspect is here dorsal, and
nearly the whole of the body is removed to show these parts.

Fig. 73.—Diagram of transverse section of pericardial sinus of

Oedipoda coerulescens. (After Graber, Arch. Mikr. Anat. ix.) H,
heart; s, septum; m, muscles—the upper suspensory, the lower
alary.

It has been thought by some that delicate vessels exist beyond the
aorta through which the fluid is distributed in definite channels, but
this does not appear to be really the case, although the fluid may
frequently be seen to move in definite lines at some distance from
the heart.

There is still much uncertainty as to some of the details of the action

of the heart, and more especially as to the influence of the alary
muscles. The effect of the contraction of these must be to increase
the area of the pericardial chamber by rendering its floor or septum
less arched, as shown in our diagram (Fig. 73), representing a
transverse section through the pericardial chamber, H being the
dorsal vessel with m its suspensory muscles, and s its septum, with
m the alary muscles. The contraction of these latter would draw the
septum into the position of the dotted line, thus increasing the area
of the sinus above; but as this floor or septum is a fenestrated
structure, its contraction allows fluid to pass through it to the
chamber above; thus this arrangement may be looked on as a
means of keeping up a supply of fluid to the dorsal vessel, the
perforated septum, when it contracts, exerting pressure on the
tissues below; these are saturated with fluid, which passes through
the apertures to the enlarged pericardial chamber.

Some misconception has prevailed, too, as to the function of the

pericardial chamber. This space frequently contains a large quantity
of fat-body—pericardial tissue—together with tracheae, and this has
given rise to the idea that it might be lung-like in function; but, as
Miall and Denny[58] have pointed out, this is erroneous; the tissues in
Insects have their own ample supplies of air. It has also been
supposed that the alary muscles cause the contraction of the heart,
but this is not directly the case, for they are not attached to it, and it
pulsates after they have been severed. It has been suggested that
the contractions of this vessel are regulated by small ganglia placed
on, or in, its substance. However this may be, these contractions
vary enormously according to the condition of the Insect; they may
be as many, it is said, as 100 or more in a minute, or they may be
very slow and feeble, if not altogether absent, without the death of
the Insect ensuing.
The expulsion of the blood from the front of the dorsal vessel seems
to be due to the rhythm of the contraction of the vessel as well as to
its mechanical structure. Bataillon says,[59] confirming an
observation of Réaumur, that at the period when the silkworm is
about to change to the chrysalis condition, the circulation undergoes
periodical changes, the fluid moving during some intervals of about
ten minutes' duration in a reversed direction, while at other times the
blood is expelled in front and backwards simultaneously, owing
apparently to a rhythmical change in the mode of contraction of the
dorsal vessel.

As the dorsal vessel consists of a number of distinct chambers, it

has been suggested that there is normally one of these for each
segment of the body; and it appears that the total number is
sometimes thirteen, which is frequently that of the segments of the
body without the head. The number of chambers differs, however,
greatly, as we have previously stated, and cannot be considered to
support the idea of an original segmental arrangement of the
chambers. The dorsal vessel, though in the adult a single organ,
arises in the embryo from two lateral, widely separated parts which
only in a subsequent stage of the embryonic development coalesce
in the median line.

Fat-Body.

In discussing the tracheae we remarked on the importance of their

function and on their abundant presence in the body. Equally
conspicuous, and perhaps scarcely less important in function, is the
fat-body, which on opening some Insects, especially such as are in
the larval stage, at once attracts attention. It consists of masses of
various size and indefinite form distributed throughout the body,
loosely connected together, and more or less surrounding and
concealing the different organs. The colour varies according to the
species of Insect. This fat-body is much connected with fine tracheal
twigs, so that an organisation extending throughout the body is thus
formed. It may be looked on as a store of nutritious matter which
may be added to or drawn on with great rapidity; and it is no doubt
on this that many of the internal parasites, so common in the earlier
stages of Insects' lives, subsist before attacking the more permanent
tissues of their hosts. There is some reason to suppose that the fat-
body may have some potency in determining the hunger of the
Insect, for some parasitised larvae eat incessantly.

The matter extracted from the food taken into the stomach of the
Insect, after undergoing some elaboration—on which point very little
is known—finds its way into the body-cavity of the creature, and as it
is not confined in any special vessels the fat-body has as unlimited a
supply of the nutritive fluid as the other organs: if nutriment be
present in much greater quantity than is required for the purposes of
immediate activity, metamorphosis or reproduction, it is no doubt
taken up by the fat-body which thus maintains, as it were, an
independent feeble life, subject to the demands of the higher parts of
the organisation. It undoubtedly is very important in metamorphosis,
indeed it is possible that one of the advantages of the larval state
may be found in the fact that it facilitates, by means of the fat-body,
the storage in the organisation of large quantities of material in a
comparatively short period of time.

A considerable quantity of fat tissue is found in the pericardial sinus,

where it is frequently of somewhat peculiar form, and is spoken of as
pericardial cells, or pericardial tissue. Some large cells, frequently of
pale yellow colour, and containing no fat, are called oenocytes by
Wielowiejski. They are connected with the general fat-body, but are
not entirely mingled with it; several kinds have been already
distinguished, and they are probably generally present. The
phagocytes, or leucocytes, the cells that institute the process of
histolysis in the metamorphosis of Muscidae, are a form of blood
cell; though these cells are amoeboid some writers derive them from
the fat-body. The cells in the blood have no doubt generally an
intimate relation with the fat-body, but very little accurate information
has been obtained as to these important physiological points, though
Graber has inaugurated their study.[60]

Organs of Sex.

The continuation of the species is effected in Insects by means of

two sexes, each endowed with special reproductive organs. It has
been stated that there are three sexes in some Insects—male,
female, and neuter; but this is not correct, as the so-called neuters
are truly sexed individuals,—generally females,—though, as a rule,
they are not occupied with the direct physiological processes for
continuing the species.

The offspring is usually produced in the shape of eggs, which are

formed in ovaries. These organs consist of egg-tubes, a cluster of
which is placed on each side of the body, and is suspended,
according to Leydig[61] and others, to the tissue connected with the
heart by means of the thread-like terminations of the tubes.

Fig. 74.—Sex organs of female of Scolia interrupta (after Dufour); a,

egg-tubes; b, oviducts; c, poison glands; d, duct of accessory
gland (or spermatheca); e, external terminal parts of body.

The number of egg-tubes varies greatly in different Insects; there

may be only one to each ovary (Campodea), but usually the number
is greater, and in the queen-bee it is increased to about 180. In the
Queens of the Termitidae, or white ants, the ovaries take on an
extraordinary development; they fill the whole of the greatly
distended hind-body. Three thousand egg-tubes, each containing
many hundred eggs, may be found in a Queen Termite, so that, as
has been said by Hagen,[62] an offspring of millions in number is
probable. There is considerable variety in the arrangements for the
growth of the eggs in the egg-tubes. Speaking concisely, the tubes
may be considered to be centres of attraction for nutritive material, of
which they frequently contain considerable stores. Next to the
terminal thread, of which we have already spoken, there is a greater
or smaller enlargement of the tube, called the terminal chamber; and
there may also be nutriment chambers, in addition to the dilatations
which form the egg-chambers proper. Korschelt[63] distinguishes
three principal forms of egg-tubes, viz. (1) there are no special
nutriment chambers, a condition shown in Figure 74; (2) nutriment
chambers alternate with the egg-chambers, as shown in our Figure
of an egg-tube of Dytiscus marginalis; (3) the terminal chamber
takes on an unusual development, acting as a large nutriment
chamber, there being no other special nutriment chambers. This
condition is found in Rhizotrogus solstitialis. The arrangements as to
successive or simultaneous production of the eggs in the tubes
seem to differ in different Insects. In some forms, such as the white
ants, the process of egg-formation (oogenesis) attains a rapidity that
is almost incredible, and is continued, it is said, for periods of many
months. There is no point in which Insects differ more than in that of
the number of eggs produced by one female. The egg-tubes are
connected with a duct for the conveyance of the eggs to the exterior,
and the arrangements of the tubes with regard to the oviduct also
vary much. An interesting condition is found in Machilis (see Fig. 94,
p. 188), where the seven egg-tubes are not arranged in a bunch, but
open at a distance from one another into the elongated duct. The
two oviducts usually unite into one chamber, called the azygos
portion or the uterus, near their termination. There are a few Insects
(Ephemeridae) in which the two oviducts do not unite, but have a
pair of orifices at the extremity of the body. Hatchett-Jackson has
recently shown[64] that in Vanessa io of the Order Lepidoptera, the
paired larval oviducts are solid, and are fixed ventrally so as to
represent an Ephemeridean stage; that the azygos system of ducts
and appended structures develop separately from the original
oviducts, and that they pass through stages represented in other
Orders of Insects to the stage peculiar to the Lepidoptera. Machilis,
according to Oudemans, is a complete connecting link between the
Insects with single and those with paired orifices.

There are in different Insects more than one kind of diverticula and
accessory glands in connexion with the oviducts or uterus; a
receptaculum seminis, also called spermatheca, is common. In the
Lepidoptera there is added a remarkable structure, the bursa
copulatrix, which is a pouch connected by a tubular isthmus with the
common portion of the oviduct, but having at the same time a
separate external orifice, so that there are two sexual orifices, the
opening of the bursa copulatrix being the lower or more anterior. The
organ called by Dufour in his various contributions glande sébifique,
is now considered to be, in some cases at any rate, a spermatheca.
The special functions of the accessory glands are still very obscure.

Fig. 75.—Egg-tube of Dytiscus marginalis; e.c, egg-chamber; n.c,

nutriment chamber; t.c, terminal chamber; t.t, terminal thread.
(After Korschelt.)

The ovaries of the female are replaced in the male by a pair of

testes, organs exhibiting much variety of form. The structure may
consist of an extremely long and fine convoluted tube, packed into a
small space and covered with a capsule; or there may be several
shorter tubes. As another extreme may be mentioned the existence
of a number of small follicles opening into a common tube, several of
these small bodies forming together a testis. As a rule each testis
has its own capsule, but cases occur—very frequently in the
Lepidoptera—in which the two testes are enclosed in a common
capsule; so that there then appears to be only one testis. The
secretion of each testis is conveyed outwards by means of a slender
tube, the vas deferens, and there are always two such tubes, even
when the two testes are placed in one capsule. The vasa deferentia
differ greatly in their length in different Insects, and are in some
cases many times the length of the body; they open into a common
duct, the ductus ejaculatorius. Usually at some part of the vas
deferens there exists a reservoir in the form of a sac or dilatation,
called the vesicula seminalis. There are in the male, as well as in the
female, frequently diverticula, or glands, in connexion with the sexual
passages; these sometimes exhibit very remarkable forms, as in the
common cockroach, but their functions are quite obscure. There is,
as we have already remarked, extreme variety in the details of the
structure of the internal reproductive apparatus in the male, and
there are a few cases in which the vasa deferentia do not unite
behind, but terminate in a pair of separate orifices. The genus
Machilis is as remarkable in the form of the sexual glands and ducts
of the male as we have already mentioned it to be in the
corresponding parts of the female.

Fig. 76.—Tenthredo cincta. a, a, testes; b, b, vasa deferentia; c, c,

vesiculæ seminales; d, extremity of body with copulatory
armature. (After Dufour.)

Although the internal sexual organs are only fully developed in the
imago or terminal stage of the individual life, yet in reality their
rudiments appear very early, and may be detected from the embryo
state onwards through the other preparatory stages.

The spermatozoa of a considerable number of Insects, especially of

Coleoptera, have been examined by Ballowitz;[65] they exhibit great
variety; usually they are of extremely elongate form, thread-like, with
curious sagittate or simply pointed heads, and are of a fibrillar
structure, breaking up at various parts into finer threads.

External Sexual Organs.—The terminal segments of the body are

usually very highly modified in connexion with the external sexual
organs, and this modification occurs in such a great variety of forms
as to render it impossible to give any general account thereof, or of
the organs themselves. Some of these segments—or parts of the
segments, for it may be dorsal plates or ventral plates, or both—may
be withdrawn into the interior, and changed in shape, or may be
doubled over, so that the true termination of the body may be
concealed. The comparative anatomy of all these parts is especially
complex in the males, and has been as yet but little elucidated, and
as the various terms made use of by descriptive entomologists are of
an unsatisfactory nature we may be excused from enumerating
them. We may, however, mention that when a terminal chamber is
found, with which both the alimentary canal and the sexual organs
are connected, it is called a cloaca, as in other animals.

Parthenogenesis.

There are undoubted cases in Insects of the occurrence of

parthenogenesis, that is, the production of young by a female without
concurrence of a male. This phenomenon is usually limited to a
small number of generations, as in the case of the Aphididae, or
even to a single generation, as occurs in the alternation of
generations of many Cynipidae, a parthenogenetic alternating with a
sexual generation. There are, however, a few species of Insects of
which no male is known (in Tenthredinidae, Cynipidae, Coccidae),
and these must be looked on as perpetually parthenogenetic. It is a
curious fact that the result of parthenogenesis in some species is the
production of only one sex, which in some Insects is female, in
others male; the phenomenon in the former case is called by
Taschenberg[66] Thelyotoky, in the latter case Arrhenotoky;
Deuterotoky being applied to the cases in which two sexes are
produced. In some forms of parthenogenesis the young are
produced alive instead of in the form of eggs. A very rare kind of
parthenogenesis, called paedogenesis, has been found to exist in
two or three species of Diptera, young being produced by the
immature Insect, either larva or pupa.

Glands.

Insects are provided with a variety of glands, some of which we have

alluded to in describing the alimentary canal and the organs of sex;
but in addition to these there are others in connexion with the outer
integument; they may be either single cells, as described by Miall in
Dicranota larva,[67] or groups of cells, isolated in tubes, or pouches.
The minute structure of Insect glands has been to some extent
described by Leydig;[68] they appear to be essentially of a simple
nature, but their special functions are very problematic, it being
difficult to obtain sufficient of their products for satisfactory
examination.

CHAPTER V

DEVELOPMENT

EMBRYOLOGY–EGGS–MICROPYLES–FORMATION OF EMBRYO–VENTRAL
PLATE–ECTODERM AND ENDODERM–SEGMENTATION–LATER STAGES–
DIRECT OBSERVATION OF EMBRYO–METAMORPHOSIS–COMPLETE AND
 INCOMPLETE–INSTAR–HYPERMETAMORPHOSIS–METAMORPHOSIS OF
INTERNAL ORGANS–INTEGUMENT–METAMORPHOSIS OF BLOWFLY–
HISTOLYSIS–IMAGINAL DISCS–PHYSIOLOGY OF METAMORPHOSIS–
ECDYSIS.

The processes for the maintenance of the life of the individual are in
Insects of less proportional importance in comparison with those for
the maintenance of the species than they are in Vertebrates. The
generations of Insects are numerous, and the individuals produced
in each generation are still more profuse. The individuals have as a
rule only a short life; several successive generations may indeed
make their appearances and disappear in the course of a single
year.

Although eggs are laid by the great majority of Insects, a few species
nevertheless increase their numbers by the production of living
young, in a shape more or less closely similar to that of the parent.
This is well known to take place in the Aphididae or green-fly Insects,
whose rapid increase in numbers is such a plague to the farmer and
gardener. These and some other cases are, however, exceptional,
and only emphasise the fact that Insects are pre-eminently
oviparous. Leydig, indeed, has found in the same Aphis, and even in
the same ovary, an egg-tube producing eggs while a neighbouring
tube is producing viviparous individuals.[69] In the Diptera pupipara
the young are produced one at a time, and are born in the pupal
stage of their development, the earlier larval state being undergone
in the body of the parent: thus a single large egg is laid, which is
really a pupa.

The eggs are usually of rather large size in comparison with the
parent, and are produced in numbers varying according to the
species from a few—15 or even less in some fossorial Hymenoptera
—to many thousands in the social Insects: somewhere between 50
and 100 may perhaps be taken as an average number for one
female to produce. The whole number is frequently deposited with
rapidity, and the parent then dies at once. Some of the migratory
locusts are known to deposit batches of eggs after considerable
intervals of time and change of locality. The social Insects present
extraordinary anomalies as to the production of the eggs and the
prolongation of the life of the female parent, who is in such cases
called a queen.

The living matter contained in the egg of an Insect is protected by

three external coats: (1) a delicate interior oolemm; (2) a stronger,
usually shell-like, covering called the chorion; (3) a layer of material
added to the exterior of the egg from glands, at or near the time
when it is deposited, and of very various character, sometimes
forming a coat on each egg and sometimes a common covering or
capsule for a number of eggs. The egg-shell proper, or chorion, is
frequently covered in whole or part with a complex minute sculpture,
of a symmetrical character, and in some cases this is very highly
developed, forming an ornamentation of much delicacy; hence some
Insects' eggs are objects of admirable appearance, though the
microscope is of course necessary to reveal their charms. One of the
families of butterflies, the Lycaenidae, is remarkable for the complex
forms displayed by the ornamentation of the chorion (see Fig. 78, B).

Fig. 77.—Upper or micropylar aspect of egg of Vanessa cardui. (After

Scudder.)

The egg-shell at one pole of the egg is perforated by one or more

minute orifices for the admission to the interior of the spermatozoon,
and it is the rule that the shell hereabouts is symmetrically sculptured
(see Fig. 77), even when it is unornamented elsewhere: the
apertures in question are called micropyles. They are sometimes
protected by a micropyle apparatus, consisting of raised processes,
or porches: these are developed to an extraordinary extent in some
eggs, especially in those of Hemiptera-Heteroptera (see Fig. 78, C).
Some of these peculiar structures have been described and figured
by Leuckart.[70] The purpose they serve is quite obscure.

Fig. 78.—Eggs of Insects: A, blowfly (after Henking); B, butterfly,

Thecla (after Scudder); C, Hemipteron (Reduviid).

Formation of Embryo.

The mature, but unfertilised, egg is filled with matter that should
ultimately become the future individual, and in the process of
attaining this end is the seat of a most remarkable series of changes,
which in some Insects are passed through with extreme rapidity. The
egg-contents consist of a comparatively structureless matrix of a
protoplasmic nature and of yolk, both of which are distributed
throughout the egg in an approximately even manner. The yolk,
however, is by no means of a simple nature, but consists, even in a
single egg, of two or three kinds of spherular or granular
constituents; and these vary much in their appearance and
arrangement in the early stages of the development of an egg, the
yolk of the same egg being either of a homogeneously granular
nature, or consisting of granules and larger masses, as well as of
particles of fatty matter; these latter when seen through the
microscope looking sometimes like shining, nearly colourless,
globules.
 Fig. 79.—Showing the two extruded polar bodies P1, P2 now nearly
fused and reincluded, and the formation of the spindle by junction
of the male and female pronuclei. (After Henking.)

The nature of the matrix—which term we may apply to both the

protoplasm and yolk as distinguished from the minute formative
portions of the egg—and the changes that take place in it have been
to some extent studied, and Kowalewsky, Dohrn,[71] Woodworth,[72]
and others have given some particulars about them. The early
changes in the formative parts of the mature egg have been
observed by Henking in several Insects, and particularly in
Pyrrhocoris, his observations being of considerable interest. When
the egg is in the ovary and before it is quite mature,—at the time, in
fact, when it is receiving nutriment from ovarian cells,—it contains a
germinal vesicle including a germinal spot, but when the egg is
mature the germinal vesicle has disappeared, and there exists in its
place at one portion of the periphery of the egg-contents a cluster of
minute bodies called chromosomes by Henking, whom we shall
follow in briefly describing their changes. The group divides into two,
each of which is arranged in a rod or spindle-like manner, and may
then be called a directive rod or spindle. The outer of these two
groups travels quite to the periphery of the egg, and there with some
adjacent matter is extruded quite outside the egg-contents (not
outside the egg-coverings), being in its augmented form called a
polar or directive body. While this is going on the second directive
spindle itself divides into two groups, the outer of which is then
extruded in the manner we have already described in the case of the
first polar body, thus completing the extrusion of two directive bodies.
The essential parts of the bodies that are successively formed during
these processes are the aggregates, called chromosomes; the
number of these chromosomes appears to be constant in each
species; their movements and dispositions are of a very interesting
character, the systems they form in the course of their development
having polar and equatorial arrangements. These we cannot further
allude to, but may mention that the extrusion of the directive bodies
is only temporary, they being again included within the periphery of
the egg by the growth and extension of adjacent parts which meet
over and thus enclose the bodies.

The arrangements and movements we have briefly alluded to have

been limited to the unfertilised condition of the egg (we should rather
say, the fertilising element has taken no part in them), and have as
their result the union of the chromosomes existing after the extrusion
of the two polar bodies, into a small body called the female
pronucleus or egg-nucleus (Eikern), while the position of the
movements has been an extremely minute portion of the egg near to
its outer surface or periphery. The introduction of a sperm, or male,
element to the egg through the micropyle gives rise to the formation
of another minute body placed more in the interior of the egg, and
called the sperm-nucleus. The egg-nucleus, travelling more into the
interior of the egg, meets the sperm-nucleus; the two amalgamate,
forming a nucleus or body that goes through a series of changes
resulting in its division into two daughter-bodies. These two again
divide, and by repetitions of such division a large number of nuclei
are formed which become arranged in a continuous manner so as to
form an envelope enclosing a considerable part (if not quite the
whole) of the egg-mass. This envelope is called the blastoderm, and
together with its contents will form the embryo. We must merely
allude to the fact that it has been considered that some of the nuclei
forming the blastoderm arise directly from the egg-mass by a
process of amalgamation, and if this prove to be correct it may be
admitted that some portions of the embryo are not entirely the result
of division or segmentation of combined germ and sperm-nuclei.
Wheeler states[73] that some of the nuclei formed by the first
differentiation go to form the vitellophags scattered throughout the
yolk. We should also remark that, according to Henking, the
blastoderm when completed shows at one part a thickening,
immediately under which (i.e. included in the area the blastoderm
encloses) are the two polar bodies, which, as we have seen, were
formed by the germinating body at an earlier stage of its activity. Fig.
79 represents a stage in the development of Pyrrhocoris, showing
the interior of the egg after a body has been formed by the union of
the sperm and egg-nuclei; this body is about to undergo division or
segmentation, and the equatorial arrangement where this will take
place is seen. The two polar bodies P1, P2, after having been
excluded, are nearly reincluded in the egg.

The Ventral Plate.

The next important change after the formation of the blastoderm is

the partial detachment of a part of its periphery to become placed in
the interior of the other and larger portion. The way in which this
takes place will be gathered from the accompanying diagrammatic
figures taken from Graber: a thickened portion (a b) of the
blastoderm becomes indrawn so as to leave a fold (c d) at each point
of its withdrawal, and these folds afterwards grow and meet so as to
enclose the thickened portion. The outer envelope, formed in part by
the original blastoderm and in part by the new growth, is called the
serosa (e f), the inner layer (g) of the conjoined new folds being
termed the amnion: the part withdrawn to the interior and covered by
the serosa and amnion is called the ventral plate, or germinal band
(Keimstreif), and becomes developed into the future animal. The
details of the withdrawal of the ventral plate to the interior are very
different in the various Insects that have been investigated.
 Fig. 80.—Stages of the enclosure of the ventral plate: A, a, b, ventral plate;
B, c, d, folds of the blastoderm that form the commencement of the
amnion and serosa; C, e, f, part of the serosa; g, amnion.

One of the earliest stages in the development is a differentiation of a

portion of the ventral plate into layers from which the future parts of the
organisation will be derived. This separation of endoderm from ectoderm
takes place by a sort of invagination, analogous with that by which the
ventral plate itself is formed. A longitudinal depression running along the
middle of the ventral plate appears, and forms a groove or channel, which
becomes obliterated as to its outer face by the meeting together of the
two margins of the groove (except on the anterior part, which remains
open). The more internal layer of the periphery of this closed canal is the
origin of the endoderm and its derivatives. Subsequently the ventral plate
and its derivatives grow so as to form the ventral part and the internal
organs of the Insect, the dorsal part being completed much later by
growths that differ much in different Insects; Graber, who has specially
investigated this matter, informing us[74] that an astonishing
multifariousness is displayed. It would appear that the various modes of
this development do not coincide with the divisions into Orders and
Families adopted by any systematists.

We should observe that the terms ectoderm, mesoderm, and endoderm

will probably be no longer applied to the layers of the embryo when
embryologists shall have decided as to the nature of the derived layers,
and shall have agreed as to names for them. According to the
nomenclature of Graber[75] the blastoderm differentiates into Ectoblast
and Endoblast; this latter undergoing a further differentiation into
Coeloblast and Myoblast. This talented embryologist gives the following
table of the relations of the embryonic layers and their nomenclature, the
first term of each group being the one he proposed to use:—
Nussbaum considers[76] that "there are four layers in the cockroach-
embryo, viz. (1) epiblast, from which the integument and nervous system
are developed; (2) somatic layer of mesoblast, mainly converted into the
muscles of the body-wall; (3) splanchnic layer of mesoblast, yielding the
muscular coat of the alimentary canal; and (4) hypoblast, yielding the
epithelium of the mesenteron."

Fig. 81.—Early stages of the segmentation of a beetle (Lina): A,

segmentation not visible, 1 day; B, segmentation of head visible; C,
segmentation still more advanced, 2¼ days; PC, procephalic lobes; g1,
g2, g3, segments bearing appendages of the head; th, thorax; th1, th2,
th3, segments of the thorax; a1, a2, anterior abdominal.

Turning our attention to the origin of the segmentation, that is so marked

a feature of Insect structure, we find that evidence of division or
arrangement of the body into segments appears very early, as shown in
our Figure of some of the early stages of development of Lina (a beetle),
Fig. 81. In A the segmentation of the ectoderm has not commenced, but
the procephalic lobes (P C) are seen; in B the three head segments are
distinct, while in C the thoracic segmentation has occurred, and that of
the abdomen has commenced. Graber considers that in this species the
abdomen consists of ten segmental lobes, and a terminal piece or telson.
According to Graber[77] this is not a primitive condition, but is preceded by
a division into three or four parts, corresponding with the divisions that will
afterwards be head, thorax, and abdomen. This primary segmentation, he
says, takes place in the Hypoblast (Endoderm) layer of the ventral plate;
this layer being, in an early stage of the development of a common
grasshopper (Stenobothrus variabilis), divided into four sections, two of
which go to form the head, while the others become thorax and abdomen
respectively. In Lina the primary segmentation is, Graber says, into three
instead of four parts. Graber's opinion on the primary segmentation does
not appear to be generally accepted, and Wheeler, who has studied[78]
the embryology of another Orthopteron, considers it will prove to be
incorrect. When the secondary segmentation occurs the anterior of the
two cephalic divisions remains intact, while the second divides into the
three parts that afterwards bear the mouth parts as appendages. The
thoracic mass subsequently segments into three parts, and still later the
hind part of the ventral plate undergoes a similar differentiation so as to
form the abdominal segments; what the exact number of these may be is,
however, by no means easy to decide, the division being but vague,
especially posteriorly, and not occurring all at once, but progressing from
before backwards.

The investigations that have been made in reference to the segmentation

of the ventral plate do not at present justify us in asserting that all Insects
are formed from the same number of embryonic segments. The matter is
summarised by Lowne, to the effect that posterior to the procephalic lobes
there are three head segments and three thoracic segments, and a
number of abdominal segments, "rarely less than nine or more than
eleven." It will be seen by referring to Figure 81 that the segmentation
appears, not simultaneously, but progressively from the head backwards;
this of course greatly increases the difficulty of determining by means of a
section the real number of segments.
 Fig. 82.—Embryo of a moth (Zygaena) at the fifth day (after Graber): am,
amnion; s, serosa; p, procephalic lobes; st, stomodaeum; pr,
proctodaeum; g1, g2, g3, the mouth parts or head appendages; th1, th2,
th3, appendages of the thoracic segments; a1-a10, abdominal
segments; s.g, salivary gland.

The later stages in the development of Insects are already proved to be

so various that it would be impossible to attempt to follow them in detail;
but in Fig. 82 we represent a median section of the embryo of Zygaena
filipendula at the fifth day. It shows well some of the more important of the
general features of the development at a stage subsequent to those
represented in Fig. 81, A, B, C. The very distinct stomodaeum (st) and
proctodaeum (pr) are seen as inflexions of the external wall of the body;
the segmentation and the development of the ventral parts of the embryo
are well advanced, while the dorsal part of the embryo is still quite
incomplete.

The method of investigation by which embryologists chiefly carry on their

researches is that of dividing the egg after proper preparation, into a large
number of thin sections, which are afterwards examined in detail, so as to
allow the arrangement to be completely inferred and described. Valuable
as this method is, it is nevertheless clear that it should, if possible, be
supplemented by direct observation of the processes as they take place
in the living egg: this method was formerly used, and by its aid we may
still hope to obtain exact knowledge as to the arrangements and
rearrangements of particles by which the structures develop. Such
questions as whether the whole formative power in the egg is absolutely
confined to one or two small centres to which the whole of the other egg
contents are merely, as it were, passive accessories, or whether an egg is
a combination in which some portion of the powers of rearrangement is
possessed by other particles, as well as the chromosomes, in virtue of
their own nature or of their position at an early period in the whole, can
scarcely be settled without the aid of direct observation of the processes
during life.

The importance of the yolk is recognised by most of the recent writers.

Nussbaum states (loc. cit.) that "scattered yolk-cells associate themselves
with the mesoblast cells, so that the constituents of the mesoblast have a
twofold origin." Wheeler finds[79] that amoeboid cells—he styles them