Distribution Theory Applied to

Differential Equations 1st Edition Adina

Chirila
Visit to download the full and correct content document:
https://textbookfull.com/product/distribution-theory-applied-to-differential-equations-1s
t-edition-adina-chirila/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Introduction to Complex Theory of Differential

Equations Savin

https://textbookfull.com/product/introduction-to-complex-theory-
of-differential-equations-savin/

An Invitation to Applied Mathematics Differential

Equations Modeling and Computation 1st Edition Carmen
Chicone

https://textbookfull.com/product/an-invitation-to-applied-
mathematics-differential-equations-modeling-and-computation-1st-
edition-carmen-chicone/

Applied Stochastic Differential Equations 1st Edition

Simo Särkkä And Arno Solin

https://textbookfull.com/product/applied-stochastic-differential-
equations-1st-edition-simo-sarkka-and-arno-solin/

Applied Differential Equations with Boundary Value

Problems 1st Edition Vladimir A. Dobrushkin

https://textbookfull.com/product/applied-differential-equations-
with-boundary-value-problems-1st-edition-vladimir-a-dobrushkin/
Theory of Differential Equations in Engineering and
Mechanics 1st Edition Kam Tim Chau

https://textbookfull.com/product/theory-of-differential-
equations-in-engineering-and-mechanics-1st-edition-kam-tim-chau/

Introduction to Partial Differential Equations Peter J.

Olver

https://textbookfull.com/product/introduction-to-partial-
differential-equations-peter-j-olver/

From Ordinary to Partial Differential Equations 1st

Edition Giampiero Esposito (Auth.)

https://textbookfull.com/product/from-ordinary-to-partial-
differential-equations-1st-edition-giampiero-esposito-auth/

Ordinary Differential Equations and Boundary Value

Problems Volume I Advanced Ordinary Differential
Equations 1st Edition John R Graef

https://textbookfull.com/product/ordinary-differential-equations-
and-boundary-value-problems-volume-i-advanced-ordinary-
differential-equations-1st-edition-john-r-graef/

Elementary differential equations Second Edition

Roberts

https://textbookfull.com/product/elementary-differential-
equations-second-edition-roberts/
Adina Chirilă
Marin Marin
Andreas Öchsner

Distribution
Theory Applied
to Differential
Equations
Distribution Theory Applied to Differential
Equations
Adina Chirilă Marin Marin Andreas Öchsner
• •

Distribution Theory Applied

to Differential Equations

123
Adina Chirilă Marin Marin
Department of Mathematics Department of Mathematics
and Computer Sciences and Computer Sciences
Transilvania University of Braşov Transilvania University of Braşov
Braşov, Romania Braşov, Romania

Andreas Öchsner
Faculty of Mechanical Engineering
Esslingen University of Applied Sciences
Esslingen am Neckar, Baden-Württemberg
Germany

ISBN 978-3-030-67158-7 ISBN 978-3-030-67159-4 (eBook)

https://doi.org/10.1007/978-3-030-67159-4
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Switzerland AG 2021
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Initial Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Test Functions and Regularization . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Seminorms and Locally Convex Spaces . . . . . . . . . . . . . . . . . . 11
2.3.1 Locally Convex Spaces . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.2 Convex and Balanced Sets . . . . . . . . . . . . . . . . . . . . . 13
2.3.3 Absorbing Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4 Duals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4.1 Reflexive Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.5 The Inductive Limit Topology . . . . . . . . . . . . . . . . . . . . . . . . . 20
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3 Convex and Lower-Semicontinuous Functions . . . . . . . . . . . . . . . . 25
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2 Convex Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.3 Lower Semicontinuous Functions . . . . . . . . . . . . . . . . . . . . . . . 28
3.4 Convexity and Lower Semicontinuity . . . . . . . . . . . . . . . . . . . . 30
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4 The Subdifferential of a Convex Function . . . . . . . . . . . . . . . . . . . . 37
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.2 The Conjugate Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.3 The Additivity of the Subdifferential . . . . . . . . . . . . . . . . . . . . 52
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

v
vi Contents

5 Evolution Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.2 The Resolvent and the Yosida Approximation . . . . . . . . . . . . . 61
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6 Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........ 73
6.1 Fundamental Spaces in the Theory of Distributions . . ........ 73
6.1.1 On Some Properties of the Spaces C m ðXÞ
and C1 ðXÞ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6.2 The Space of Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.3 The Dual of C 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6.4 The Derivative of a Distribution . . . . . . . . . . . . . . . . . . . . . . . 87
6.5 Distributions as Generalized Functions . . . . . . . . . . . . . . . . . . . 90
6.6 On Some Spaces of Distributions . . . . . . . . . . . . . . . . . . . . . . . 96
6.7 The Primitive of a Distribution . . . . . . . . . . . . . . . . . . . . . . . . 96
6.7.1 Structure Theorems . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.8 Extras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
6.8.1 Higher-Order Primitives . . . . . . . . . . . . . . . . . . . . . . . 109
6.8.2 The Local Structure of Distributions . . . . . . . . . . . . . . 112
6.9 Convolutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
6.9.1 The Direct Product of Distributions . . . . . . . . . . . . . . . 117
6.9.2 Convolution of Distributions . . . . . . . . . . . . . . . . . . . . 121
6.9.3 Convolution of Functions and Distributions:
Regularization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
6.9.4 Convolution Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
7 Tempered Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
7.1 The Schwartz Space of Infinitely Differentiable Functions
Rapidly Decreasing at Infinity . . . . . . . . . . . . . . . . . . . . . . . . . 133
7.2 Tempered Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
7.3 The Fourier Transform in SðRn Þ . . . . . . . . . . . . . . . . . . . . . . . 140
7.3.1 The Inverse Fourier Transform . . . . . . . . . . . . . . . . . . 142
7.3.2 Properties of the Fourier Transform . . . . . . . . . . . . . . . 144
7.4 Fourier Transform of Tempered Distributions . . . . . . . . . . . . . . 145
7.5 The Fourier Transform of a Distribution with Compact
Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
7.6 The Product of a Distribution by a C 1 Function . . . . . . . . . . . 151
7.7 The Space of Multipliers of S 0 ðRn Þ . . . . . . . . . . . . . . . . . . . . . 152
7.8 Some Results on Convolutions with Tempered Distributions . . . 156
7.9 The Paley-Wiener-Schwartz Theorem . . . . . . . . . . . . . . . . . . . . 160
7.10 A Result on the Fourier Transform of a Convolution
of Two Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Contents vii

8 Differential Equations in Distributions . . . . . . . . . . . . . . . . . . . . . . 167

8.1 Ordinary Differential Equations . . . . . . . . . . . . . . . . . . . . . . . . 167
8.1.1 Linear Differential Equations with Constant
Coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
8.1.2 An Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
8.2 Partial Differential Equations . . . . . . . . . . . . . . . . . . . . . . . . . . 181
8.2.1 The Direct Product . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
8.2.2 Hyperbolic Partial Differential Equations . . . . . . . . . . . 183
8.2.3 Parabolic Partial Differential Equations . . . . . . . . . . . . 185
8.2.4 Elliptic Partial Differential Equations . . . . . . . . . . . . . . 186
8.2.5 The Cauchy Problem . . . . . . . . . . . . . . . . . . . . . . . . . 188
8.2.6 An Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
9 Sobolev Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
9.1 The Sobolev Space H 1 ðXÞ . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
9.2 The Sobolev Space H m ðXÞ . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
9.3 The Sobolev Space W k;p ðXÞ . . . . . . . . . . . . . . . . . . . . . . . . . . 203
9.4 The Sobolev Spaces H s ðRn Þ . . . . . . . . . . . . . . . . . . . . . . . . . . 206
9.5 Besov Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
9.5.1 The Nonhomogeneous Littlewood-Paley
Decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
9.5.2 Definition and Properties . . . . . . . . . . . . . . . . . . . . . . 217
9.5.3 The Homogeneous Littlewood-Paley Decomposition
and the Homogeneous Besov Spaces . . . . . . . . . . . . . . 222
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
10 Variational Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
10.1.1 The Stokes System . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
10.1.2 The Elasticity System . . . . . . . . . . . . . . . . . . . . . . . . . 232
10.1.3 The Plate Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
10.2 The Approximation of Variational Problems . . . . . . . . . . . . . . . 237
10.2.1 The Galerkin Method . . . . . . . . . . . . . . . . . . . . . . . . . 241
10.2.2 The Finite Element Method . . . . . . . . . . . . . . . . . . . . 242
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
11 On Some Spaces of Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . 249
11.1 The Spaces DLp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
11.2 The Space O0C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
viii Contents

12 On Some Differential Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . 263

12.1 Local and Pseudolocal Operators . . . . . . . . . . . . . . . . . . . . . . . 263
12.2 Hypoelliptic Partial Differential Operators . . . . . . . . . . . . . . . . 266
12.3 Existence of Fundamental Solutions . . . . . . . . . . . . . . . . . . . . . 267
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Chapter 1
Introduction

Abstract We give a brief motivation for the concept of distributions. We explain

why this concept is useful in applied mathematics. We present some landmarks in
the history of this subject. Then we present the main topics and results in the theory
of distributions.

1.1 Initial Remarks

The concept of distributions is a relatively recent one since it was used for the first
time in 1951 in a study of the French mathematician L. Schwarz [43]. Distributions
are a generalization of the concept of functions, hence some mathematicians call
them today “generalized functions”. This generalization was motivated by practical
situations. For instance, a boundary value problem from the theory of differential
equations (or even a Cauchy problem) has a solution if the right-hand side of the
problem is at least of class C 1 . But in the phenomena modelled by these problems
these conditions are not satisfied. In most of these phenomena, the functions are not
even continuous, very often having discontinuities of the second kind.
We will see that the same problem, formulated in the context of the theory of
distributions will not impose restrictions on the right-hand side, these being auto-
matically satisfied. We anticipate that every distribution is differentiable (in a sense
that we will discuss and which does not differ too much from the differentiation of
a function) of every order. In the context of the theory of distributions the regularity
conditions are weaker. For example, the functions do not need to be continuous [1,
2, 4].
The theory of distributions, as it is structured today, has the disadvantage that it
employs very sophisticated mathematical concepts, see for example [13, 18, 20, 21,
31, 35, 42]. Therefore, this theory is less accessible to the researchers in mathematics.
Most of the concepts used in the theory of distributions are at the intersection of other
subjects of mathematics.
The fundamentals of the theory of distributions are based on advanced topics in
functional analysis [3, 5, 6, 9, 24, 30, 36, 37, 47, 49], topology, mathematical physics
[7, 10, 11, 14, 16, 17, 19, 22, 23, 25, 32, 34, 38–40, 45, 46, 48] differential geometry

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 1

A. Chirilă et al., Distribution Theory Applied to Differential Equations,
https://doi.org/10.1007/978-3-030-67159-4_1
2 1 Introduction

and so on [8, 15, 26–29, 33, 44]. We may consider that the origins of this theory
coincide with Sobolev’s trial to define generalized solutions for hyperbolic partial
differential equations in the 1940s. Sobolev’s idea was employed by L. Schwarz in
the 1960s, when the concept of distributions was used for the first time. In 1971
Schwarz [43] published a book in which he laid the foundations of the unified theory
of distributions by using the method of linear functionals. Afterwards, Sikorski [41]
introduced a theory of distributions by using fundamental sequences of continuous
functions. This method is similar to that used by Dedekind [12] to introduce the real
numbers as cuts in the set of natural numbers.
The way in which Schwarz introduced the concept of distributions seems more
natural and for this reason we use it in the sequel. Since the contributions of Schwarz
intersect with the ones of Sobolev, many concepts and results from the theory of
distributions use the names of these two mathematicians, even though the theory
as it is today used, is almost completely different from that initiated by these two
scientists.
The main qualitative improvement expected from this theory is the simplification
of the mathematical theory in order to make it more accesible. Starting with the
initiators of this theory, the concepts have become more complex.

References

1. R.A. Adams, J.J.F. Fournier, Sobolev Spaces (Elsevier, 2003)

2. J.J. Alibert, G. Bouchitte, Non-uniform integrability and generalized Young measures. J. Con-
vex Anal. 4(1), 129–147 (1997)
3. H.W. Alt, Lineare Funktionalanalysis (Springer, Berlin Heidelberg, 2006)
4. L. Ambrosio, N. Gigli, G. Savare, Gradient Flows in Metric Spaces and in the Space of
Probability Measures (Birkhäuser, Basel, 2005)
5. V. Barbu, T. Precupanu, Convexity and Optimization in Banach Spaces, 4th edn. (Springer,
2012)
6. J. Borwein, Convex Functions: Constructions, Characterizations and Counterexamples (Cam-
bridge University Press, New York, 2010)
7. R.I. Bot, S.M. Grad, G. Wanka, Fenchel’s duality theorem. J. Optim. Theory Appl. 132, 509–
515 (2007)
8. D. Braess, Finite Elemente: Theorie, schnelle Löser und Anwendungen in der Elastizitätstheorie
(Springer, 2007)
9. H. Brezis, Functional Analysis, Sobolev Spaces and Partial Differential Equations (Springer,
2011)
10. M.G. Crandall, Semigroups of nonlinear contractions and dissipative sets. J. Funct. Anal. 3,
376–418 (1969)
11. C. De Lellis, L. Szekelyhidi, Jr., The Euler equations as a differential inclusion. Ann. Math.
(2) 170(3), 1417–1436 (2009)
12. R. Dedekind, Stetigkeit und irrationale Zahlen (Braunschweig University, 1872)
13. J.J. Duistermaat, J.A.C. Kolk, Distributions: Theory and Applications (Birkhäuser, 2010)
14. D.G. Ebin, J. Marsden, Groups of diffeomorphisms and the motion of an incompressible fluid.
Ann. Math. 2(92), 102–163 (1970)
15. H. Elman, D. Silvester, A. Wathen, Finite Elements and Fast Iterative Solvers: With Applications
in Incompressible Fluid Dynamics, 2nd edn. (Oxford University Press, 2014)
References 3

16. L.C. Evans, Partial Differential Equations, 4th edn. (American Mathematical Society, 2008)
17. I. Fonseca, G. Leoni, Modern Methods in the Calculus of Variations: L p Spaces (Springer,
2007)
18. F.G. Friedlander, M. Joshi, Introduction to the Theory of Distributions, 2nd edn. (Cambridge
University Press, 1999)
19. D. Gilbarg, N.S. Trudinger, Elliptic Partial Differential Equations of Second Order, 2nd edn.
(Springer, 2001)
20. D.D. Haroske, H. Triebel, Distributions, Sobolev Spaces, Elliptic Equations (European Math-
ematical Society, 2008)
21. L. Hörmander, The Analysis of Linear Partial Differential Operators I: Distribution Theory
and Fourier Analysis, 2nd edn. (Springer, 1990)
22. T. Kato, Nonstationary flows of viscous and ideal fluids in R3 . J. Funct. Anal. 9, 296–305
(1972)
23. T. Kato, On the eigenfunction of many-particle systems in quantum mechanics. Comm. Pure
Appl. Math. 10, 151–177 (1957)
24. Y. Komura, On linear topological spaces. Kumamoto J. Sci. A 5, 148–157 (1962)
25. L. Lichtenstein, Grundlagen der Hydromechanik (Springer, 1929)
26. M. Marin, H. Arabnia, Equations of Evolution (Elliot & Fitzpatrick, Athens, USA, 2010)
27. M. Marin, A. Öchsner, Complements of Higher Mathematics (Springer, 2018)
28. M. Marin, A. Öchsner, Essentials of Partial Differential Equations (Springer, 2019)
29. S.D. Micu, Introduction to Finite Element Method (Romanian), Publications of the Centre for
Nonlinear Analysis and its Applications
30. G. Minty, Monotone operators in Hilbert space. Duke Math. J. 29, 341–346 (1962)
31. D. Mitrea, Distributions, Partial Differential Equations, and Harmonic Analysis (Springer,
2013)
32. L. Onsager, Statistical hydrodynamics. Nuovo Cimento (9), 6(Supplemento, 2(Convegno Inter-
nazionale di Meccanica Statistica)), 279–287 (1949)
33. A. Öchsner, Computational Statics and Dynamics: An Introduction Based on the Finite Element
Method (Springer, Singapore, 2020)
34. N.S. Papageorgiou, Weak solutions of differential equations in Banach spaces. Bull. Austral.
Math. Soc. 33, 407–418 (1986)
35. I. Richards, H. Youn, Theory of Distributions: A Non-technical Introduction (Cambridge Uni-
versity Press, 1995)
36. R.T. Rockafellar, Convex Analysis, No. 28 (Princeton University Press, Princeton, N.J., 1970)
37. W. Rudin, Functional analysis (Tata-McGraw Hill Publishing, 1974)
38. V. Scheffer, An inviscid flow with compact support in space-time. J. Geom. Anal. 3(4), 343–401
(1993)
39. A.I. Shnirelman, On the nonuniqueness of weak solution of the Euler equation. Comm. Pure
Appl. Math. 50(12), 1261–1286 (1997)
40. R.E. Showalter, Hilbert space methods for partial differential equations. Electron. J. Diff. Eqns.,
Monograph 01 (1994)
41. R. Sikorski, The elementary theory of distributions. Bull. Acad. Pol., Warszawa (1957)
42. R. Strichartz, A Guide to Distribution Theory and Fourier Transforms (CRC Press, 1994)
43. L. Schwartz, Theorie des Distributions (Herman & Cie, Paris, 1971)
44. E. Stein, Singular Integral and Differentiability Properties of Functions (Princeton University
Press, 1970)
45. L. Szekelyhidi, Weak solutions of the Euler equations: non-uniqueness and dissipation.
Journees Equations aux derivees partielles, Roscoff, 1-5 juin 2015
46. L. Szekelyhidi Jr., E. Wiedemann, Young measures generated by ideal incompressible fluid
flows. Arch. Rational Mech. Anal. (2012)
47. H. Triebel, Theory of Function Spaces (Birkhäuser, 1983)
48. E. Wiedemann, Existence of weak solutions for the incompressible Euler equations. Ann. Inst.
H. Poincare Anal. Non Lineaire 28(5), 727–730 (2011)
49. L.C. Young, Lecture on the Calculus of Variations and Optimal Control Theory (American
Mathematical Society, 1980)
Chapter 2
Preliminaries

Abstract This chapter presents some concepts that are used in the next chapters. One
section is dedicated to seminorms and locally convex spaces, as well as to convex,
balanced and absorbing sets. Another section presents duals and reflexive spaces.
Finally, the inductive limit topology is presented.

2.1 Introduction

Let X be a set. We denote by 2 X the set of all subsets of X . This includes the empty
set.

Definition 2.1.1 Let X be a set and T be a subset of 2 X . Then the pair (X, T ) is a
topological space if it satisfies the following properties:
(i) ∅ ∈ T , X ∈ T ;
(ii) if U ∈ T , V ∈ T , then U ∩ V ∈ T ;
(iii) if  is an arbitrary set and Uλ ∈ T for all λ ∈ , then ∪λ∈ Uλ ∈ T .

Then we call T a topology on X [6].

Definition 2.1.2 A topological space (X, T ) is called Hausdorff if for different

points there exist disjoint open neighbourhoods, that is,

x = y ⇒ ∃Ux , U y ∈ T such that x ∈ Ux , y ∈ U y and Ux ∩ U y = ∅. (2.1.1)

Proposition 2.1.1 Let (X, T ) be a Hausdorff space. Then every sequence has at
most one limit point.

Definition 2.1.3 We call a pair (X,d) a metric space if d : X × X → [0, ∞) has the
three properties below:
(i) d(x, y) = 0 ⇔ x = y;
(ii) d(y, x) = d(x, y), ∀x, y ∈ X ;
(iii) d(x, y) ≤ d(x, z) + d(z, y), ∀x, y, z ∈ X .

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 5

A. Chirilă et al., Distribution Theory Applied to Differential Equations,
https://doi.org/10.1007/978-3-030-67159-4_2
6 2 Preliminaries

Then a map d : X × X → [0, ∞) satisfying these three properties (of definiteness,

symmetry and the triangle inequality) is called a metric on X [2, 7].

Definition 2.1.4 Let (X, d) be a metric space. Then a sequence x : N → X is called

a Cauchy sequence if

∀ε > 0, ∃k0 ∈ N such that ∀k, j ≥ k0 , d(x j , xk ) < ε. (2.1.2)

Definition 2.1.5 Let (X, d) be a metric space. Then it is called complete if every
Cauchy sequence converges.

Definition 2.1.6 We call a pair (X,  · ) a normed space if X is a K vector space

and  ·  : X → [0, ∞) is a map satisfying the three properties below:
(i) x = 0 ⇒ x = 0;
(ii) αx = |α|x, ∀α ∈ K, x ∈ X .
(iii) x + y ≤ x + y.

Then a map  ·  : X → [0, ∞) satisfying these three properties (of definiteness,

homogeneity and the triangle inequality) is called a norm on X [2, 7].

Definition 2.1.7 Let (X,  · ) be a normed space. Then it is called a Banach space
if it is complete under the metric d : X × X → [0, ∞) defined by

d(x, y) = x − y, ∀x, y ∈ X. (2.1.3)

Definition 2.1.8 Let X be a K vector space. We call a map (x, y) → (x, y) X from
X × X to K a sesquilinear form if for all x, y, x1 , y1 , x2 , y2 ∈ X and all α ∈ K, the
following holds true:
(i) (αx, y) X = α(x, y) X , (x, αy) X = ᾱ(x, y) X ;
(ii) (x1 + x2 , y) X = (x1 , y) X + (x2 , y) X , (x, y1 + y2 ) X = (x, y1 ) X + (x, y2 ) X .

Definition 2.1.9 Let X be a K vector space. Then a sesquilinear form which is

symmetric, i.e.,
(i) ∀x, y ∈ X, (y, x) X = (x, y) X
and positive definite, i.e.,
(ii) ∀x ∈ X, (x, x) X ≥ 0 and (x, x) = 0 ⇔ x = 0
is called a scalar product.

Definition 2.1.10 Let X be a K vector space. The pair (X, (·, ·) X ), where (·, ·) X is
a scalar product, is called a pre-Hilbert space.

Definition 2.1.11 A pre-Hilbert space (X, (·, ·) X ) which is complete under the norm
induced by the scalar product is called a Hilbert space.

Below we present the theorem of Hahn-Banach.

2.1 Introduction 7

Theorem 2.1.1 Let X be an R vector space. We assume that

(i) p : X → R is sublinear, that is, ∀x, y ∈ X

p(x + y) ≤ p(x) + p(y),

(2.1.4)
p(αx) = α p(x), ∀α ≥ 0;

(ii) f : Y → R is linear, where Y is a subspace of X ;

(iii) f (x) ≤ p(x), ∀x ∈ Y .
Then there is F : X → R linear with

F(x) = f (x), ∀x ∈ Y,
(2.1.5)
F(x) ≤ p(x), ∀x ∈ X.

Below we present another version of the Hahn-Banach theorem

Theorem 2.1.2 Let X be a normed K vector space. Let Y be a subspace of X with

the norm induced by X . Then for every y  ∈ Y  , there exists x  ∈ X  such that

x  = y  on Y,
(2.1.6)
x   X  = y  Y  .

2.2 Test Functions and Regularization

Definition 2.2.1 Let  ⊂ Rn be open. Let f be a complex-valued function defined

on . We call support of f , and denote it by supp f , the closure in  of the set

{x ∈  : f (x) = 0}. (2.2.1)

Observation 2.2.1 The support of f is the smallest relative closed set outside of
which f is identically zero.

By definition, a test function on  is a C ∞ function on  which has compact

support in . By Cc∞ () we denote the vector space of all test functions on .
Note that this vector space endowed with an appropriate topology is important in the
definition of distributions on .
We introduce the function
 1
e |x|2 −1 if |x| < 1
β(x) = . (2.2.2)
0 if |x| ≥ 1

This is a C ∞ function for which the support is the closed unit ball. We can divide
by a constant, namely Rn β(x)d x. Then we obtain another C ∞ function, which we
denote by α. Its support is the closed unit ball and we have
8 2 Preliminaries

α(x)d x = 1. (2.2.3)
Rn

Then we introduce ∀ε > 0 the function

x 
αε (x) = ε−n α . (2.2.4)
ε

Note that αε (x) ∈ Cc∞ (Rn ). Moreover, its support is Bε (0), the closed ball with
center at the origin and radius ε. We have

αε (x)d x = 1. (2.2.5)
Rn

Note that the family of test functions (αε )ε>0 can be used to regularize discon-
tinuous functions, like integrable ones or L p functions [1, 4, 5]. In other words, we
can show that such functions can be approximated by test functions.
Definition 2.2.2 A function f defined in  is locally integrable if f is integrable
(in the sense of Lebesgue [3]) on every compact subset K ⊂ .
We denote by χ K the characteristic function of K , which is equal to 1 on K and
0 outside of K . Hence, f is locally integrable in  if, for every compact set K ⊂ ,
the product f · χ K is integrable on .
Definition 2.2.3 We consider that u is a locally integrable function on Rn . Then the
convolution of u and αε is defined by
 
u ε (x) = u(x − y)αε (y)dy = u(y)αε (x − y)dy. (2.2.6)
Rn Rn

Note that other possible notations are (u ∗ αε )(x) or (αε ∗ u)(x).

Theorem 2.2.1 Let u be a locally integrable function in Rn . We have
1. The convolution u ε is a C ∞ function in Rn .
2. If u has compact support K (and in the case u is integrable in Rn ), the support
of u ε is contained in the ε-neighbourhood of K .
3. If u is a continuous function, then u ε converges uniformly to u on compact subsets
of Rn .
4. If u ∈ L p (Rn ), 1 ≤ p < +∞, then u ε → u in L p (Rn ).
Proof Step 1 We want to show that u ε ∈ C ∞ (Rn ). Note that both integrals in (2.2.6)
are being evaluated over compact subsets of Rn . Hence, it suffices to apply the
classical theorem about differentiation inside the integral sign.
Step 2 By definition, the ε-neighbourhood of K is the set

Kε = Bε (x) (2.2.7)
x∈K
2.2 Test Functions and Regularization 9

of all closed balls with center x ∈ K and radius ε. If x ∈ Rn and d(x, K ) > ε, then
x∈/ K ε so that αε (x − y) = 0 for all y ∈ K . Hence, the second integral in (2.2.6) is
equal to zero. We deduce that the support of u ε is contained in K ε .
Step 3 Let u be a continuous function. We assume that L is an arbitrary fixed
compact subset of Rn . We evaluate

u ε (x) − u(x) = [u(x − y) − u(x)] αε (y)dy. (2.2.8)
Rn

Note that u is uniformly continuous on L. Therefore, given σ > 0, there is δ > 0

such that
|u(x − y) − u(x)| < σ (2.2.9)

for all x ∈ L and for all |y| < δ. By considering that ε ≤ δ, we obtain

|u ε (x) − u(x)| ≤ |u(x − y) − u(x)| αε (y)dy < σ, (2.2.10)
Rn

for all x ∈ L. This implies that u ε → u, uniformly on L, as ε → 0.

Step 4 Let u ∈ L p (Rn ), 1 ≤ p < ∞. We recall that u can be approximated in L p
by continuous functions with compact support. Moreover, if u ∈ L p , then u ε ∈ L p .
This can be derived from Minkowski’s inequality [3] in its integral form
 1
p
  p 1
p
||u ε || p = |u ε (x)| d x
p
= u(x − y)αε (y)dy dx ≤
Rn Rn Rn
  1
p
≤ |u(x − y)αε (y)| d x p
dy = (2.2.11)
Rn Rn
  1
p
= αε (y) |u(x − y)| d x p
dy = ||u|| p .
Rn Rn

Given σ > 0, we consider a continuous function v with compact support, such that
σ
||u − v|| p < . (2.2.12)
3
It follows from inequalities (2.2.11) and (2.2.12) that
σ
||u ε − vε || p ≤ ||u − v|| p < . (2.2.13)
3
Then we evaluate

||u ε − u|| p ≤ ||u ε − vε || p + ||vε − v|| p + ||v − u|| p . (2.2.14)

10 2 Preliminaries

Note that v is a continuous function with compact support. Hence, by step 3, vε → v

uniformly on Rn . Therefore, we have vε → v in L p . If we choose ε so small that
σ
vε − v p < , (2.2.15)
3
then every one of the three terms at the right-hand side of the inequality (2.2.14) is
smaller than σ/3. This implies that ||u ε − u|| p < σ. 
Definition 2.2.4 The family of test functions (αε )ε>0 is called a regularizing family
of functions in Rn .
If we consider that ε = 1j , then the sequence of functions

α j (x) = j n α( j x), j = 1, 2, . . . , (2.2.16)

is called a regularizing sequence of functions.

Corollary 2.2.1 Let  ⊂ Rn be open. Then Cc∞ () is a dense subspace of L p (),
1 ≤ p < +∞.
Corollary 2.2.2 Let K ⊂  be compact. Then there is a function φ ∈ Cc∞ () such
that 0 ≤ φ ≤ 1 and φ = 1 on K .
Proof Without loss of generality, we consider that  is bounded. Let δ be the distance
from K to the boundary of  and let K δ/3 be the δ/3-neighbourhood of K . We denote
by χδ/3 the characteristic function of K δ/3 . Clearly, the function

φ = χδ/3 ∗ αδ/3 , (2.2.17)

satisfies the required properties. 

Let K ⊂ Rn be compact and V be an arbitrary neighbourhood of K . Similary, we
can prove that there is a function φ ∈ Cc∞ (Rn ) such that 0 ≤ φ ≤ 1, φ is equal to 1
on a neighbourhood of K and supp φ ⊂ V .
Corollary 2.2.3 Let A, a ∈ R such that 0 < a < A. We denote by B A and B A−a
concentric balls of radius A and A − a, respectively. Then there is a function φ ∈
Cc∞ (Rn ) such that:
(i) supp φ ⊂ B A ,
(ii) φ(x) = 1 on B A−a ,
(iii) for all p ∈ N n , |∂ p φ(x)| ≤ C( p, n)a −| p| , ∀x ∈ Rn .
Proof We consider that χ is the characteristic function of the concentric ball with
radius A − (2a/3). Then we define
  
1 x−y
φ(x) = χ ∗ αδ (x) = αδ (x − y)dy = α dy,
B A−(2a/3) δn B A−(2a/3) δ
(2.2.18)
2.2 Test Functions and Regularization 11

with δ = a/3. Clearly, supp φ ⊂ B A and φ = 1 on B A−a . For all 1 ≤ j ≤ n, we

obtain  
δ −1 ∂α x − y
∂ j φ(x) = n dy. (2.2.19)
δ B A−(2a/3) ∂x j δ

Therefore, we have
 y 
δ −1
|∂ j φ(x)| ≤ ∂jα dy = δ −1 ∂ j α(t)dt ≤ C( j, n)a −1 . (2.2.20)
δn Rn δ Rn

Condition (iii) follows in a similar way. 

Corollary 2.2.4 Let K ⊂ Rn be compact. Let ( j )1≤ j≤k be a finite open covering
of K . Then there are functions φ j ∈ Cc∞ (Rn ) such that 0 ≤ φ j ≤ 1, 1 ≤ j ≤ k and
k
j=1 φ j = 1 on a neighbourhood of K .

Proof We denote by K̊ j the interior of K j . Then we can find compact sets (K j )1≤ j≤k
such that K j ⊂  j and
 k
K ⊂ K̊ j . (2.2.21)
j=1

For every j, let ψ j ∈ Cc∞ ( j ) be such that 0 ≤ ψ j ≤ 1 and ψ j = 1 on K j . We

consider

φ1 = ψ1 , φ j = ψ j (1 − ψ1 ) · · · (1 − ψ j−1 ), j = 2, . . . , k. (2.2.22)

Clearly, the functions (φ j )1≤ j≤k satisfy the required properties. 

Note that the functions (φ j )1≤ j≤k are called a C ∞ partition of unity subordinated
to the covering ( j )1≤ j≤k of K .

2.3 Seminorms and Locally Convex Spaces

Let K be a field, for example, the field of real numbers R or the field of complex
numbers C. Let E be a vector space over K . We consider that R+ is the set of all
nonnegative real numbers.

Definition 2.3.1 A map p : E → R+ is a seminorm on E if it satisfies

(i) p(x + y) ≤ p(x) + p(y), ∀x, y ∈ E;
(ii) p(λx) = |λ| p(x), ∀x ∈ E, ∀λ ∈ K .
Moreover, p is a norm if it satisfies
(iii) p(x) = 0 ⇔ x = 0.
12 2 Preliminaries

Example 1 Let p ∈ R, 1 ≤ p < ∞. We consider the following norm on Rn

  1p

n
x = (x1 , . . . , xn ) ∈ R → x =
n
|xi | p
∈ R+ . (2.3.1)
i=1

Moreover,
x∞ = sup |xi | (2.3.2)
1≤i≤n

defines a norm on Rn .
Example 2 Let 1 < p < ∞. We consider that L p (I ) is the space of functions f
defined on I and such that | f | p is integrable on I in the Lebesgue sense [3]. We
define the following seminorm on L p (I )

 b
1
p

 f p = | f (t)| p dt . (2.3.3)
a

2.3.1 Locally Convex Spaces

Let E be a vector space over K . A topology defined on E is compatible with the

vector space structure of E if the maps

(x, y) ∈ E × E → x + y ∈ E
(2.3.4)
(λ, x) ∈ K × E → λx ∈ E

are continuous. A vector space endowed with a compatible topology is called a

topological vector space [6]. The topology of a topological vector space can be
characterized by a fundamental system (or basis) of neighbourhoods of the origin.
In a topological space, a collection of sets is called a fundamental system of neigh-
bourhoods of a point if every set of the collection contains the point, the intersection
of any two sets of the collection contains a set which belongs to the collection and
every neghbourhood of the point contains a set of the collection.
For locally convex spaces, a fundamental system of neighbourhoods of the origin
can be characterized by a family of seminorms.
Let p be a seminorm on E. An open ball with center x0 ∈ E and radius r > 0 is
the set
B(x0 , r ) = {x ∈ E| p(x − x0 ) < r }. (2.3.5)

Moreover, a closed ball with center x0 ∈ E and radius r > 0 is the set

B̄(x0 , r ) = {x ∈ E| p(x − x0 ) ≤ r }. (2.3.6)

2.3 Seminorms and Locally Convex Spaces 13

A neighbourhood of x0 is a set V containing a ball with center x0 .

Let x0 ∈ E be arbitrary. Note that the set of all open (closed) balls with center x0
defines a fundamental system of neighbourhoods of x0 of a topology on E which is
compatible with the vector space structure of E.
Let a ∈ E be fixed and λ ∈ K , λ = 0 be fixed. The following maps

x → a + x,
(2.3.7)
x → λx

are called translation and homothety and are homeomorphisms on E. Hence, it suf-
fices to have a fundamental system of neighbourhoods of the origin in order to have
a fundamental system of neighbourhoods of a point a ∈ E.
Let ( pi )i∈I be a family of seminorms defined on E. We define for every x0 ∈ E,
ε ∈ R, ε > 0 and F a finite part of I

V (x0 , ε, F) = {x ∈ E| pi (x − x0 ) < ε, i ∈ F}. (2.3.8)

The set V (x0 , ε, F) is the intersection of the balls with center x0 and radius ε,
corresponding to the seminorms pi with i ∈ F.
The family of all sets V (x0 , ε, F) when ε runs through the set of all positive real
numbers and F runs through all finite subsets of I defines a fundamental system of
neighbourhoods of x0 of a topology on E compatible with the vector space structure
of E. The space E endowed with this topology is called a locally convex topological
vector space.
If E is a locally convex topological vector space and ( pi )i∈I is a family of semi-
norms defining its topology, then E is a Hausdorff space [3] if and only if to any pair
x, y ∈ E with x = y we associate a seminorm pk such that pk (x) = pk (y).

2.3.2 Convex and Balanced Sets

Definition 2.3.2 Let E be a vector space over K . A subset A of E is convex if,

given two points x, y ∈ A, the segment αx + β y with α, β ∈ R+ and α + β = 1 is
contained in A.

Example 1 The whole space E is convex. The empty set is convex.

Example 2 Balls are convex sets.
Example 3 Segments are convex sets.
Example 4 If (Ai )i∈I is a family of convex sets of E, then the intersection A =
∩i∈I Ai is a convex set.

Definition 2.3.3 Let A be a subset of E. The smallest convex subset of E containing

A is called the convex hull of A.
14 2 Preliminaries

Note that, given a set A, its convex hull is always well defined. Clearly, it suffices
to consider the intersection of the family of all convex sets which contain A. Note
that this family is nonempty because it contains the whole space E.
The convex hull of A will be denoted by (A). This can be characterized as the set
of all x ∈ E which can be represented as a finite sum x = i∈F αi xi with xi ∈ A,
αi ∈ R+ , i∈F αi = 1, where F is a finite set of indices which depend on x.

Definition 2.3.4 Let A be a subset of E. Then it is called balanced if λA ⊂ A for

all λ ∈ K such that |λ| ≤ 1.

Example 1 Balls with center at the origin are balanced sets.

Example 2 Let E = R. The set [0, 1] is convex, but it is not balanced.
Let A be a subset of E. We assume that it is convex and balanced. Then for every
pair of points x, y ∈ A we have αx + β y ∈ A for all α, β ∈ K with |α| + |β| ≤ 1.
Note that the intersection of a family of convex and balanced subsets of E is a
convex and balanced subset of E.

Definition 2.3.5 Let A be a subset of E. The smallest balanced convex subset b (A)
of E containing A is called the balanced convex hull of A.

Note that the balanced convex hull of A is the intersection of all balanced convex
sets containing A. Moreover, we have the following characterization

x ∈ b (A) ⇔ x = αi xi , (2.3.9)
i∈F

where xi ∈ A, i∈F |αi | ≤ 1 and F is a finite set of indices.

2.3.3 Absorbing Sets

Definition 2.3.6 Let V be a subset of a vector space E. V is called absorbing if

given x ∈ E, there is λ ∈ R, λ > 0 such that λx ∈ V .

Example 1 The set {−1, 0, 1} is an absorbing set on the real line R.

Example 2 Balls with center at the origin are absorbing sets. More generally,
every neighbourhood of the origin in a topological vector space is an absorbing set.
Let E be a vector space endowed with a seminorm p. The unit ball

U = {x ∈ E| p(x) < 1} (2.3.10)

is a balanced convex set. Moreover, it is absorbing.

We can show that if V is a balanced convex and absorbing subset of E, then

q(x) = inf{λ ≥ 0|x ∈ λV } (2.3.11)

2.3 Seminorms and Locally Convex Spaces 15

is a seminorm on E. Clearly, q is well defined on E and q : E → R+ because V is

absorbing. If x, y ∈ E, let λ > 0 and μ > 0 be such that x ∈ λV and y ∈ μV . Since
V is convex, we deduce that
 
λ μ
x + y ∈ λV + μV = (λ + μ) V+ V ⊂ (λ + μ)V. (2.3.12)
λ+μ λ+μ

Hence, we obtain
q(x + y) ≤ q(x) + q(y). (2.3.13)

Since V is balanced, we deduce that

q(λx) = |λ|q(x). (2.3.14)

Therefore, we conclude that q is a seminorm on E.

Moreover, we have

{x ∈ E|q(x) < 1} ⊂ V ⊂ {x ∈ E|q(x) ≤ 1}. (2.3.15)

Theorem 2.3.1 Let E be a topological vector space. The following conditions are
equivalent:
(i) E is locally convex;
(ii) there is a fundamental system of convex neighbourhoods of the origin;
(iii) there is a fundamental system of absorbing balanced convex neighbourhoods
of the origin.

Proof Step 1. We show that (i) implies (ii). If the topology of E is defined by a
family of seminorms ( pi )i∈I , then the sets

V (F, ε) = {x ∈ E| pi (x) ≤ ε, i ∈ F} (2.3.16)

with F a finite part of I and 0 < ε < 1, form a fundamental system of convex
neighbourhoods of the origin.
Step 2. We show that (iii) implies (i). Let V be an absorbing balanced convex
subset of E which belongs to a fundamental system of neighbourhoods of the origin.
For every such V we have a corresponding seminorm pV . Hence, the family of
seminorms obtained in this way defines the topology of E.
Step 3. We show that (ii) implies (iii). Let V be a convex neighbourhood of the
origin. It suffices to prove that the set

U = ∩|λ|=1 λV (2.3.17)

is an absorbing balanced convex neighbourhood of the origin. Note that the map
(μ, x) → μx is continuous at the point (0, 0). Hence, there are ε > 0 and V  a
neighbourhood of zero in E such that
16 2 Preliminaries

μx ∈ V, ∀ |μ| ≤ ε, ∀x ∈ V  . (2.3.18)

Or there is a neighbourhood of zero W such that

μW ⊂ V, ∀ |μ| ≤ 1. (2.3.19)

Hence, we have
μW ⊂ V, ∀ |μ| = 1. (2.3.20)

This implies that W ⊂ λV , ∀ |λ| = 1. Therefore, U is a neighbourhood of zero in

E.
It is clear that U is a convex set since it is the intersection of convex sets. Moreover,
it is absorbing. We will show that U is balanced. If x ∈ U , then the segment [0, x]
is contained in U , that is

λx ∈ U, ∀ λ such that 0 ≤ λ ≤ 1. (2.3.21)

By the definition of U , if x ∈ U , then λx ∈ U , ∀ |λ| = 1. Hence, if μ = 0 and |μ| ≤ 1,

then we have μ
μx = |μ| x ∈ U. (2.3.22)
|μ|

This shows that U is balanced. 

In the sequel, we give examples of locally convex spaces.

Example 1 A seminormed space is a locally convex space. In particular, a normed
space is a locally convex space.
Example 2 Let X be a locally compact topological space. We consider that C(X )
is the space of all complex-valued continuous functions on X . Moreover, we consider
that K is the collection of all compact subsets of X . Let us consider the family of
seminorms ( pk )k∈K with pk ( f ) = supt∈K | f (t)|. This family of seminorms defines
a Hausdorff locally convex topology on C(X ), see [3].
Note that a sequence ( f j ) j of functions of C(X ) converges to zero in this topology
if and only if f j (x) converges to zero uniformly on each compact subset K of X .
Therefore, this locally convex topology is called the topology of uniform convergence
on compact subsets of X .
Let X =  ⊂ Rn be open. Then such a topology can be defined by a sequence
of seminorms. Clearly, it suffices to consider an increasing sequence (K j ) j∈N of
compact subsets contained in  whose union is  and the corresponding sequence
of seminorms ( p K j ) j∈N . Therefore, the topology of C() is defined by a countable
fundamental system of neighbourhoods of the origin. Then the Hausdorff locally
convex space C() is a metrizable space. Furthermore, C() is a complete space
because the uniform limit on compact subsets of  of continuous functions is a
continuous function. Therefore, the topology just defined on C() is the natural
topology of C().
2.3 Seminorms and Locally Convex Spaces 17

Definition 2.3.7 A Frechet space is a Hausdorff locally convex, metrizable and

complete space [3].
p
Example 3 Let  ⊂ Rn be open. Let 1 ≤ p < ∞. We consider that L loc () is
the space of classes of measurable functions [1, 4, 5] such that on every compact
subset K of  we have  1
p
| f (x)| p d x < ∞. (2.3.23)
K

p
The elements of L loc () are called pth power locally integrable functions. We con-
sider the seminorm  1
p
pk ( f ) = | f (x)| p d x . (2.3.24)
K

The family of seminorms ( pk ), where k runs over all compact subsets of , defines a
p
locally convex topology on L loc (). Such a topology can be defined by a countable
p
family of seminorms. Then we can prove that L loc is a Frechet space.

2.4 Duals

Let E be a vector space over a field K . Its algebraic dual E ∗ is the vector space of all
linear maps x ∗ : E ∗ → K . We denote by x ∗ (x) or (x, x ∗ ) the value of x ∗ on x ∈ E.
Let E be a topological vector space over a field K . Then the dual E  of E is the
subspace of E ∗ consisting of all continuous linear maps (or functionals) on E.
For every x ∗ ∈ E ∗ we set
 
px ∗ (x) = x, x ∗ . (2.4.1)

Then px ∗ defines a seminorm on E and the family ( px ∗ )x ∗ ∈E ∗ defines a locally convex

topology on E, which we denote by σ(E, E ∗ ).
When E is a topological vector space, the topology σ(E, E  ) defined on E by the
family of seminorms ( px  )x  ∈E  is called the weak topology of E. Clearly, the weak
topology is coarser than the given topology of E and also coarser than σ(E, E ∗ ).
Similarly, we can consider the weak topology σ(E  , E) on the dual E  of E. Then
a sequence (x j ) converges weakly to zero in E  if and only if, for every x ∈ E, the
sequence (x j (x)) converges to zero in x. Hence, the weak topology on E  coincides
with the pointwise convergence topology.
Moreover, on the dual E  of a topological vector space E we can consider the
strong topology of E  , which is also a locally convex topology.

Definition 2.4.1 Let E be a topological vector space. Let V be a neighbourhood of

zero in E. Then A ⊂ E is bounded if there is λ > 0 such that λA ⊂ V .
18 2 Preliminaries

Let E be a locally convex topological vector space. Then every neighbourhood

of zero contains a balanced neighbourhood of zero. Therefore, A is bounded in E
if to every neighbourhood of zero V there is ε > 0 such that λA ⊂ V for all |λ| ≤
ε. Note that every topological vector space has a fundamental system of balanced
neighbourhoods of zero.
Example 1 Finite subsets of E are bounded sets.
Example 2 In a seminormed space, balls are bounded sets.
Example 3 Every relatively compact subset A of a locally convex space E is
bounded. Let V be a neighbourhood of zero in E. Then there is an open neighbour-
hood of zero W in E such that W + W ⊂ V and μW ⊂ W for all |μ| ≤ 1. Note that
A is relatively compact. Then we can find a finite subset (x j )1≤ j≤ p of elements of
A such that the open sets (x j + W )1≤ j≤ p cover A. Note that the set (x j )1≤ j≤ p is
bounded in E. Then we can find λ ∈ R, 0 < λ < 1 such that λx j ⊂ W , 1 ≤ j ≤ p.
Then, we have
p
λA ⊂ ∪ j=1 λ(x j + W ) ⊂ W + W ⊂ V. (2.4.2)

Therefore, A is bounded.

Definition 2.4.2 Let B be a subset of a topological vector space E. The polar set
B ◦ of B is the subset of E  defined as
 
B ◦ = {x  ∈ E  | x, x  ≤ 1, ∀x ∈ B}. (2.4.3)

Let A ⊂ E be bounded. We will prove that its polar set A◦ is an absorbing,

balanced and convex subset of E  . Let x  , y  ∈ A◦ and α, β ∈ R+ such that α + β =
1. Then we obtain
     
x, αx  + β y  ≤ α x, x  + β x, y  ≤ 1. (2.4.4)

Therefore, A◦ is convex.
Let x  ∈ A◦ and λ ∈ K , |λ| ≤ 1. Then
   
x, λx  = |λ| x, x  ≤ 1. (2.4.5)

Therefore, we have λx  ∈ A◦ . Hence, A◦ is balanced.

We consider z  ∈ E  and let V be a neighbourhood of zero in E defined as
 
V = {x ∈ E| x, z  ≤ 1}. (2.4.6)

Note that A ⊂ E is bounded. Then there is λ > 0 such that λA ⊂ V . Therefore,

we have    
x, λz  = λx, z  ≤ 1, ∀x ∈ A. (2.4.7)

This implies that A◦ is an absorbing set in E  .

2.4 Duals 19

Hence, to every bounded set A of E we can associate the following seminorm on

E
p A◦ (x  ) = inf{λ ≥ 0|x  ∈ λA◦ }. (2.4.8)

Let B denote the family of all bounded subsets of E. Then the family of semi-
norms ( p A◦ ) A∈B defines a Hausdorff locally convex topology on E  , called the strong
topology of E  . Note that a sequence (x j ) converges strongly to zero in E  if and
 
only if the sequence x j (x) converges uniformly to zero on every bounded set of
E. Hence, the strong topology on E  is also called the topology of uniform conver-
gence on bounded sets of E. We denote by E b the dual E  endowed with the strong
topology.
Example Let E be a normed space. Let E  be its dual endowed with the norm
 
x   E  = sup x, x  . (2.4.9)
x E ≤1

Then E  is a Banach space [3]. Therefore, the strong topology on E  coincides with
the one defined by the above norm.

2.4.1 Reflexive Spaces

Let E  be the dual of the Banach space E  . We consider the following norm
 
x   E  = sup x  , x  (2.4.10)
x   E  ≤1

on E  . Therefore, E  becomes a Banach space.

Every element x ∈ E defines a continuous linear functional x̃ on E 
 
x̃(x  ) = x, x  , ∀x  ∈ E  . (2.4.11)

Note that the map x → x̃ is an isometry from E into E  . Moreover, E is called a

reflexive space if this isometry is onto.
We consider that E is a Hausdorff topological vector space and that E b is its strong
dual. The dual of E b is the bidual E  of E. Then we consider the map x ∈ E →
x̃ ∈ E  . If this map represents an isomorphism from E onto E b , then the topological
vector space E is reflexive.
Example 1 The n-dimensional Euclidean space Rn is reflexive. Moreover, finite-
dimensional Hausdorff topological vector spaces are reflexive.
Example 2 Hilbert spaces are reflexive spaces.
Example 3 The Banach spaces L p (), 1 < p < ∞ are reflexive. The spaces
L () and L ∞ () are not reflexive.
1
20 2 Preliminaries

2.5 The Inductive Limit Topology

In the sequel, we consider the inductive limit topology.

We consider that (E i )i∈N is an increasing sequence of locally convex spaces such
that the identity map E i → E i+1 is continuous for every i. Moreover, we define
∞
E = ∪i=1 Ei . (2.5.1)

Then we consider on E the finest locally convex topology for which the identity map
E i → E is continuous for every i = 1, 2, . . .. This topology represents the inductive
limit topology of E defined by the subspaces E i . Moreover, the space E endowed
with this topology is called the inductive limit of the spaces (E i )i∈N .
Let V be a convex set. We want V to be a neighbourhood of zero in the inductive
limit topology. Then it is necessary and sufficient to consider that every intersec-
tion V ∩ E i is a neighbourhood of zero in E i for all i = 1, 2, . . .. Moreover, by
considering all convex hulls  ∞ 
V =  ∪i=1 Vi (2.5.2)

we obtain a fundamental system of neighbourhoods of the origin in E. Note that

every Vi runs over a fundamental system of convex neighbourhoods of zero in E i ,
i = 1, 2, . . ..

Proposition 2.5.1 We consider that E is the inductive limit of (E i )i∈N and that F is
any locally convex space. Then a linear map u : E → F is continuous if and only if
the restriction u i of u is continuous from E i into F for all i.

Proof Step 1. Let u be continuous. Since the identity map E i → E is continuous

by definition, every restriction u i is continuous.
Step 2. We assume that every u i is a continuous map from E i into F. Let U be a
convex neighbourhood of zero in F. Then there is a convex neighbourhood of zero
Vi in E i such that u i (Vi ) ⊂ U . Hence,
 ∞ 
V =  ∪i=1 Vi (2.5.3)

is a neighbourhood of zero in E. In fact, u(V ) ⊂ U . This implies that u is continuous

from E into F. 

Lemma 2.5.1 We assume that F is a locally convex space and G is a closed subspace
of F. Let V be a balanced convex open neighbourhood of zero in G. We consider
x ∈ F such that x ∈
/ G. Therefore, there is a balanced convex open neighbourhood
of zero W in F such that x ∈
/ W and W ∩ G = V .

Proof Note that G is closed in F. Then there is a balanced convex open neighbour-
hood of zero V0 in F such that

(x + V0 ) ∩ G = ∅ and V0 ∩ G ⊂ V. (2.5.4)
2.5 The Inductive Limit Topology 21

We consider that W is the balanced convex hull of V ∪ V0 . Note that W is open.

We will show that W ∩ G = V . In fact, we have W ∩ G ⊃ V . Let w ∈ W ∩ G.
Therefore, we have
w = αv + βv0 , (2.5.5)

where v ∈ V , v0 ∈ V0 and |α| + |β| ≤ 1. If β = 0, there is nothing to show. Hence,

we consider that β = 0. Then we have v0 ∈ V0 ∩ G ⊂ V and w ∈ V . We assume by
contradiction that x ∈ W . Therefore, we have x = y + z, where y ∈ G and z ∈ V0 .
So, we obtain
y = x − z ∈ (x + V0 ) ∩ G, (2.5.6)

which is impossible. 

Theorem 2.5.1 Let E be a vector space. We assume that E is the union of an

increasing sequence (E i )i∈N of locally convex spaces with the following properties:
(i) the identity map E i → E i+1 is continuous for every i;
(ii) the topology induced by E i+1 on E i coincides with the topology of E i for every
i;
(iii) E i is a closed subspace of E i+1 for every i.
Therefore, the following hold true:
(iv) the inductive limit topology of E induces on every E i its original topology;
(v) a subset A is bounded in the inductive limit E if and only if there is an index
j such that A is contained and it is bounded in E j .

Proof Step 1. We will prove that the topology induced by E on every E i coincides
with the given topology of E i . Let Vi be a balanced convex neighbourhood of zero
in E i . It suffices to prove that there is a neighbourhood of zero V in E such that

Vi = V ∩ E i , ∀i. (2.5.7)

Note that there is a sequence (Vi+k ), k = 0, 1, 2, . . . of balanced convex neighbour-

hoods of zero in E i+k such that

Vi+k−1 = Vi+k ∩ E i+k−1 , k = 1, 2, . . . . (2.5.8)

We consider
V = ∪∞
k=0 Vi+k . (2.5.9)

Note that V is a neighbourhood of zero in E and Vi = V ∩ E i .

Step 2. We consider A ⊂ E to be bounded. We assume by contradiction that there
is no index i such that A is contained in E i . Hence, there is an increasing sequence
of indices (i n ) and a sequence (xn ) of elements of E such that

xn ∈ A ∩ E in and xn ∈
/ E in−1 . (2.5.10)
22 2 Preliminaries

Note that there is a sequence (Vn ) of balanced convex open neighbourhoods of zero
in E in such that
xn ∈
/ nVn and Vn ∩ E in−1 = Vn−1 . (2.5.11)

We consider V = ∪∞
n=1 Vn . Hence, V is a neighbourhood of zero in E such that

V ∩ E in = Vn and xn ∈
/ nV. (2.5.12)

Finally, we have a contradiction with the assumption that A is bounded in E. 

Example 1 Let  ⊂ Rn be open. We consider (K i ) an increasing sequence of

compact subsets of  such that  = ∪i K i . Moreover, we consider E = Cc () to
be the space of all complex-valued continuous functions defined in  and with
compact support. Then we consider E i = Cc (; K i ) to be the subspace of Cc ()
which consists of all functions having support in K i . We have

Cc () = ∪i Cc (; K i ). (2.5.13)

We consider on Cc (; K i ) the topology of uniform convergence on K i . This topology

is locally convex. Moreover, it is defined by the seminorm

p K i ( f ) = sup | f (x)|. (2.5.14)

x∈K i

Note that p K i is a norm on Cc (; K i ). Furthermore, the space Cc (; K i ) endowed

with this norm is a Banach space. Note that the embedding

Cc (; K i ) → Cc (; K i+1 ) (2.5.15)

is continuous. Moreover, Cc (; K i ) is a closed subspace of Cc (; K i+1 ).

Then we consider on Cc () the inductive limit topology of the spaces Cc (; K i ).
This topology is the natural topology of Cc ().
Note that a sequence (ϕ j ) converges to zero in Cc () if and only if the following
conditions hold true:
(i) There is K ⊂  compact such that supp ϕ j ⊂ K , ∀ j;
(ii) the sequence (ϕ j ) converges to zero uniformly on K .

Definition 2.5.1 A continuous linear functional on Cc () is a Radon measure on

, see [3].

We will denote by M() the space of all Radon measures on . This is the
topological dual of Cc (). Note that a linear map

μ : Cc () → C (2.5.16)

is a Radon measure if and only if

Another random document with
no related content on Scribd:
 There are fully eight varieties of American dogs, several of which
are natives of the far North. The lagopus, or isatis, a native of
Spitzbergen and Greenland, extends over all the Arctic regions of
America and Asia, and is found in some of the Kurile Islands. Dogs
are employed to draw sledges in Newfoundland and Canada; and
the Eskimo dogs, used for this purpose by the Arctic explorers, are
famous for their strength, their docility, and power of endurance.
They were mute, until they learned to bark from European dogs on
board the discovery ships.
 CHAPTER VI.
ICELAND AND THE ICELANDERS.

ust within the Arctic region, but nearly on the limits of what
geographers call the Atlantic Ocean, lies an island which,
since its colonization in the ninth century, has not ceased to
excite the interest of the explorer and the man of science.
Iceland—which measures about 300 miles at its greatest length,
from east to west, and about 200 miles at its greatest breadth, from
north to south—is situated in lat. 63° 23’-66° 33’ N., and long. 13°
22’-24° 35’ W.; at a distance of 600 miles from the nearest point of
Norway, 250 from the Faröe Isles, 250 from Greenland, and above
500 miles from the northern extremity of Scotland. As early as the
eighth Christian century it was discovered by some European
emigrants; though, indeed, the Landnana Book, one of the earliest of
the island-records, asserts that they found the memorials of a yet
earlier settlement in various Christian relics, such as wooden
crosses, which appeared to be of Irish origin. At all events, the first
really successful attempt at colonization was made by Ingolf, a
Norwegian, who planted himself and his followers at Reikiavik in
874. In the following century a somewhat extensive immigration took
place of Norwegians who resented the changes of polity introduced
by Harold Haarfager, and all the habitable points on the coast were
occupied by about 950 a.d. Fifty years afterwards, though not
without much opposition, Christianity was legally established, and
the bishoprics of Holar and Skalholt were founded. The government
assumed the character of an aristocratic republic, with a popular
assembly, called the Althing, meeting every summer in the valley of
Thingvalla. Commerce was encouraged, and the Icelanders early
distinguished themselves by the boldness of their maritime
enterprise, and the extent of their ocean fisheries.
About the year 932 they discovered Greenland, and about 986 a
portion of the North American coast, which they called “Vineland.”
They did not confine their voyages to the north, but sent their ships
even as far south as the Mediterranean. From 1150 to 1250 is rightly
considered the most flourishing period of Icelandic literature and
commerce. After the conquest of the island by Haco VI. of Norway,
much of the old spirit seemed to die out. When Norway was united to
Denmark in 1380, Iceland was included in the bond, and it is still
regarded as a dependency of the latter kingdom. In 1540 it
embraced the principles of Lutheran Protestantism. Its population at
one time numbered 100,000, but it gradually diminished until, in
1840, it was reduced to 57,094; but a slow increase has taken place
of late years, and it now amounts to about 70,000. The language
spoken is the old Norse.
Iceland is a fifth part larger than Ireland, and its superficial area is
estimated at 39,207 square miles. Not more than 4000 miles,
however, are habitable, all the rest being ice and lava; for the island
seems to be little more than a mass of trachyte, snow-shrouded and
frost-bound, resting on a sea of fire. It consists of two vast parallel
table-lands, the foundations of ranges of lofty mountains, most of
which are active volcanoes; and these table-lands strike across the
centre of the island, from north-east to south-west, at a distance from
one another of ninety to one hundred miles. Their mountainous
summits are not pyramidal, as is generally the case in Europe, but
rounded like domes, as in the Andes of South America. Their sides,
however, are broken up by precipitous masses of tufa and
conglomerate, intersected by deep ravines of the gloomiest
character. They are covered with a thick shroud of ice and snow, but
in their wombs seethe the fiery elements which ever and anon break
forth into terrible activity. The eastern table-land and its mountain
range is the most extensive, and contains Oërafa, the culminating
point of Iceland. It is visible from a great distance at sea, like a white
cloud suspended above the island. Its height is 6426 feet, and it
springs from a vast mountain-mass; no fewer than 3000 square
miles being perpetually burdened with ice and snow, at an altitude
varying from 3000 to 6000 feet.
 AN ICELANDIC LANDSCAPE.
A very considerable portion of the island is occupied by the large
glaciers which descend from the mountains, like frozen torrents,
pushing forward into the lowlands, and even to the margin of the
sea. These act as almost impassable barriers to communication
between the various inhabited districts.
We have spoken of the two ranges of table-lands as about ninety
to one hundred miles apart. The interspace forms a low broad valley,
which opens at either extremity on the sea—an awful waste, a region
of desolation, where man is utterly powerless; where the elements of
fire and frost maintain a perpetual antagonism; where blade of grass
is never seen, nor drop of water; where bird never wings its way, and
no sign of life can be detected. It seems a realization of Dante’s
“circle of ice” in the “Inferno.” The surface consists of lava streams,
fissured by innumerable crevices; of rocks piled on rocks; of dreary
glaciers, relieved by low volcanic cones. It is supposed that some
remote portions of the inaccessible interior are less barren, because
herds of reindeer have been seen feeding on the Iceland moss that
fringes the borders of this dreary region. But there is no reason to
believe that it can ever be inhabited by man.

MOUNT HEKLA, FROM THE VALLEY OF HEVITA.

The extremities of the valleys, where they approach the ocean,
are the principal theatres of volcanic activity. At the northern end the
best-known volcano is that of Hekla, which has attained a sinister
repute from the terrific character of its eruptions. Of these six-and-
twenty are recorded, the last having occurred in 1845–46. One
lasted for six years, spreading devastation over a country which had
formerly been the seat of a prosperous colony, burying the fields
beneath a flood of lava, scoriæ, and ashes. During the eruption of
September 2, 1845, to April 1846, three new craters were formed,
from which columns of fire sprang to the height of 14,000 feet. The
lava accumulated in formidable hills, and fragments of scoriæ and
pumice-stone, weighing two hundredweight, were thrown to a
distance of a league and a half; while the ice and snow which had
lain on the mountain for centuries were liquefied, and poured down
into the plains in devastating torrents.
But the eruption of another of these terrible volcanoes, the
Skaptá Jokul, which broke out on the 8th of May 1783, and lasted
until August, was of a still more awful character. At that time the
volcanic fire under Europe must have raged most violently, for a
tremendous earthquake shattered a wide extent of Calabria in the
same year, and a submarine volcano had flamed fiercely for many
weeks in the ocean, thirty miles from the south-west cape of Iceland.
Its fires ceased suddenly; a series of earthquakes shook the
island; and then Skaptá broke forth into sudden and destructive
activity.
For months the sun was hidden by dense clouds of vapour, and
clouds of volcanic dust were carried many hundreds of miles to sea,
extending even to England and Holland. Sand and ashes, raised to
an enormous height in the atmosphere, spread in all directions, and
overwhelmed thousands of acres of fertile pasturage. The
sulphurous exhalations blighted the grass of the field, and tainted the
waters of river, lake, and sea, so that not only the herds and flocks
perished, but the fish died in their poisoned element.
The quantity of matter ejected by the rent and shivered mountain
was computed at fifty or sixty thousand millions of cubic yards. The
molten lava flowed in a stream which in some places was twenty to
thirty miles in breadth, and of enormous thickness; a seething,
hissing torrent, which filled the beds of rivers, poured into the sea
nearly fifty miles from its points of eruption, and destroyed the fishing
on the coast. Some of the island-rivers were heated, it is said, to
ebullition; others were dried up; the condensed vapour fell in whirls
of snow and storms of rain. But dreadful as was the eruption itself,
with its sublime but awful phenomena, far more dreadful were its
consequences. The country within its range was one wide ghastly
desert, a fire-blighted wilderness; and, partly from want of food,
partly owing to the unwholesome condition of the atmosphere, no
fewer than 9336 men,[10] 28,000 horses, 11,461 cattle, and 190,000
sheep, were swept away in the short space of two years. Even yet
Iceland has scarcely recovered from the blow.
At the northern end of the great central valley the focus of
igneous phenomena is found in a semicircle of volcanic heights
which slope towards the eastern shore of the Lake Myvatr. Two of
these are very formidable,—namely, Leirhnukr and Krabla on the
north-east. After years of inaction, they suddenly broke out with
tremendous fury, pouring such a quantity of lava into the Lake
Myvatr, which measures twenty miles in circuit, that the water was in
a state of ebullition for many days. On the sides of Mount Krabla,
and at the base of this group of mountains, are situated various
caldrons of boiling mineral pitch, the ruined craters of ancient
volcanoes; and from their depths are thrown up jets of the molten
matter, enveloped in clouds of steam, and accompanied by loud
explosions at regular intervals.
But the most singular phenomena in this singular country, where
frost and fire are continually disputing the pre-eminence, are the
Geysers, or eruptive boiling springs. These all occur in the trachytic
formation, are characterized by their high temperature, by holding
siliceous matter in solution, which they deposit in the form of
siliceous sinter, and by evolving large quantities of sulphuretted
hydrogen gas.
Upwards of fifty geysers have been counted in the space of a few
acres at the southern end of the great valley. Some are constant,
some periodical, some stagnant, some only slightly agitated. The
grandest and most celebrated are the Great Geyser and Strokr,
thirty-five miles north-west from Hekla. These, at regular intervals,
hurl into the air immense columns of boiling water, to the height of
one hundred feet, accompanied by clouds of steam and deafening
noises. In the case of the Great Geyser, the jet issues from a shaft
about seventy-five feet deep, and ten in diameter, which opens into
the centre of a shallow basin, about one hundred and fifty feet in
circumference. The basin is alternately emptied and filled: when
filled, loud explosions are heard, the ground quivers, and the boiling
water is forced upwards in gigantic columns. Thus the basin is
emptied, and the explosions cease until it is refilled.
 Messrs. Descloiseaux and Bunsen, who, according to Mrs.
Somerville, visited Iceland in 1846, found the temperature of the
Great Geyser, at the depth of seventy-two feet, to equal 260° 30’ F.
prior to a great eruption, reduced, after the eruption, to 251° 30’ F.;
an interval of twenty-eight hours passing in silence.

THE GREAT GEYSER.

About one hundred and forty yards distant is the Strokr (from
stroka, to agitate), a circular well, forty-four feet deep, with a tube
eight feet wide at its mouth, diminishing to little more than ten inches
at a depth of twenty-seven feet. The surface of the water is in
constant ebullition, while at the bottom the temperature exceeds that
of boiling water by about twenty-four degrees. It appears, from
experiments made by Donny, that water, long boiled, becomes more
and more free from air, and that thus the cohesion of the particles is
so much increased, that when the heat is sufficiently increased to
overcome that cohesion, the production of steam is so considerable
and so instantaneous as to induce an explosion. In this circumstance
M. Donny finds an explanation of the phenomena of the Geysers,
which are in constant ebullition for many hours, until, being almost
purified from air, the intense internal or subterranean heat
overcomes the cohesion of the particles, and thus an explosion
takes place.
Lord Dufferin describes an eruption which he witnessed on the
occasion of his visit to the Geysers, but for which he waited three
days. Like pilgrims round some ancient shrine, he says, he and his
friends kept patient watch; but the Great Geyser scarcely deigned to
vouchsafe the slightest manifestation of its latent energies. Two or
three times they heard a sound as of subterranean cannonading,
and once an eruption to the height of about ten feet occurred. On the
morning of the fourth day a cry from the guides made them start to
their feet, and with one common impulse rush towards the basin.
The usual underground thunder had already commenced. A violent
agitation was disturbing the centre of the pool. Suddenly a dome of
water lifted itself up to the height of eight or ten feet, then burst, and
fell; immediately after which a shining liquid column, or rather a
sheaf of columns, wreathed in robes of vapour, sprang into the air,
and in a succession of jerking leaps, each higher than the last, flung
their silver crests against the sky. For a few minutes the fountain
held its own, then all at once appeared to lose its ascending power.
The unstable waters faltered, drooped, fell, “like a broken purpose,”
back upon themselves, and were immediately sucked down into the
recesses of their pipe.
The spectacle was one of great magnificence; but no description
can give an accurate idea of its most striking features. The
enormous wealth of water, its vitality, its hidden power, the illimitable
breadth of sunlit vapour, rolling out in exhaustless profusion,—these
combine to impress the spectator with an almost painful sense of the
stupendous energy of nature’s slightest movements.
The same traveller furnishes a very humorous account of the
Strokr (or “churn”).
It is, he says, an unfortunate Geyser, with so little command over
its temper and stomach, that you can get a rise out of it whenever
you like. Nothing more is necessary than to collect a quantity of sods
and throw them down its funnel. As it has no basin to protect it from
these liberties, you can approach to the very edge of the pipe, which
is about five feet in diameter, and look down at the boiling water
perpetually seething at the bottom. In a few minutes the dose of turf
just administered begins to disagree with it; it works itself up into “an
awful passion;” tormented by the qualms of incipient sickness, it
groans and hisses, and boils up, and spits at you with malicious
vehemence, until at last, with a roar of mingled pain and rage, it
throws up into the air a column of water forty feet high, carrying with
it all the sods that have been thrown in, and scattering them, scalded
and half-digested, at your feet. So irritated has the poor thing’s
stomach become by the discipline it has undergone, that even long
after all foreign matter has been thrown off, it continues retching and
sputtering, until at last nature is exhausted, when, sobbing and
sighing to itself, it sinks back into the bottom of its den.
The ground around the Geysers, for about a quarter of a mile,
looks as if it had been “honeycombed by disease into numerous
sores and orifices;” not a blade of grass grew on its hot, inflamed
surface, which consisted of unwholesome-looking red livid clay, or
crumpled shreds and shards of slough-like incrustations.

A region, corresponding in character to the desert mountain-

mass we have been describing, stretches westward from it to the
extremity of the ridge of the Snaefield Syssel, and terminates in the
remarkable cone of Snaefield Jokul.
 The island coasts exhibit a singularly broken outline, and the
deep lochs or fiords, like those of Norway, only less romantic, dip
into the interior for many miles, and throw off numerous branches.
These fiords are wild and gloomy; dark, still inlets, with precipices on
either side, a thousand feet in height, and the silence unbroken, save
by the occasional wash of the waters, or the scream of a solitary
ocean-bird. Inland, however, they assume a gentler character: they
end in long narrow valleys, watered by pleasant streams, and bright
with pasture. In these bits of Arcadia the inhabitants have built their
towns and villages.
In the valleys on the north coast, which are adorned by clumps of
willow and juniper, the soil is comparatively fertile; but the most
genial scenery is found on the east, where, in some places, the
birch-trees reach a height of twenty feet, and are of sufficient size to
be used in house-building. The fuel used by the Icelanders is the
drift-wood which the Gulf Stream brings from Mexico, the Carolinas,
Virginia, and the River St. Lawrence.
In the south of the island the mean temperature is about 39°; in
the central districts, 36°; in the north it rarely rises above 32°, or
freezing-point. Thunder-storms, though rare in high latitudes, are not
uncommon in Iceland; a circumstance which is due, no doubt, to the
atmospheric disturbances caused by the volcanic phenomena.
Hurricanes are frequent, and the days are few when the island is
free from sea-mists. At the northern end the sun is always above the
horizon in the middle of summer, and under it in the middle of winter;
but absolute darkness does not prevail.
One of the most interesting places in Iceland is Thingvalla, where
of old the “Althing,” or supreme parliament, was wont to hold its
annual assemblies, under the “Logmathurman,” or president of the
republic.
It is nothing more than a broad plain on the bank of the River
Oxerá, near the point where the swift waters, after forming a noble
cascade, sweep into the Lake of Thingvalla. Only a plain; but the
scenery around it is indescribably grand and solemn. On either side
lies a barren plateau, above which rises a range of snowy
mountains, and from the plateaus the plain is cut off by deep
chasms,—that of Almanna Gja on the east, and the Hrafna Gja on
the west. It measures eight miles in breadth, and its surface is
covered by a network of innumerable fissures and crevices of great
depth and breadth. At the foot of the plain lies a lake, about thirty
miles in circumference, in the centre of which two small crater-
islands, the result of some ancient eruption, are situated. The
mountains on its south bank have a romantic aspect, and that their
volcanic fires are not extinct is shown by the clouds of vapour
evolved from the hot springs that pour down their rugged sides. The
actual meeting-place of the Althing was an irregular oval area, about
two hundred feet by fifty, almost entirely surrounded by a crevice so
broad and deep as to be impassable, except where a narrow
causeway connected it with the adjacent plain, and permitted access
to its interior. At one other point, indeed, the encircling chasm is so
narrow that it may possibly be cleared at a leap; and the story runs
that one Flosi, when hotly pursued by his enemies, did in this way
escape them; but as falling an inch short would mean sure death in
the green waters below, the chasm may be regarded as a tolerably
sure barrier against intruders.
The ancient capital of the island was Skalholt, where, in the
eleventh century, was founded the first school; an episcopal seat; the
birthplace of a long line of Norse worthies, Islief the chronicler,
Gissur the linguist, and Finnur Johnson the historian. But its glories
have passed away; its noble cathedral has ceased to exist; and
three or four cottages alone perpetuate the name of the once
flourishing city.
The present capital is Reikiavik, to which, in 1797, were
transferred the united bishoprics of Stoolum and Skalholt. It consists
of a collection of wooden sheds, one story high, rising here and there
into a gable end of greater pretensions, extending along a tract of
dreary lava, and flanked at either end by a suburb of turf huts. On
every side of it stretches a dreary lava-plain, and the gloom of the
scorched and ghastly landscape is unrelieved by tree or bush. The
white mountains are too distant to serve as a background to the
buildings, but before the door of each merchant’s house, facing the
sea, streams a bright little pennon; and as the traveller paces the
silent streets, whose dust no carriage-wheel has ever desecrated,
the rows of flower-pots peeping out of the windows, between white
muslin curtains, at once convince him that, notwithstanding their
unostentatious appearance, within each dwelling reign “the elegance
and comfort of a woman-tended home.”
The prosperity of Reikiavik is chiefly due to its excellent harbour,
and to the fish-banks in its neighbourhood, which supply it with an
important commercial staple. In the summer and early autumn it is
much visited by tourists, who start from thence to admire the
wonders of Hekla, Skapta, and the Geysers; but its busiest time is in
July, when the annual fair draws thither a crowd of fisher-folk and
peasants. From a distance of forty and fifty leagues they come, with
long trains of pack-horses, their stock-fish slung loose across the
animals’ backs, and their other wares packed closely in boxes or
bags of reindeer-skin.

HARBOUR OF REIKIAVIK.
 The Icelander is honest, temperate, hospitable, possessed with a
fervent spirit of patriotism, and strongly wedded to the ancient
usages. He is also industrious; and though his industry is but scantily
remunerated, he earns enough to satisfy his simple tastes. In the
interior his chief dependence is on his cattle; and as grass is the
main produce of his farm, his anxiety during the haymaking season
is extreme. A bad crop would be almost ruin. He is, however, wofully
ignorant of agriculture as a science, does but little for the
improvement of the soil, and employs implements of the most
primitive character. The process of haymaking in Iceland is thus
described:—
The best crops are gathered from the “tún,” a kind of home-park
or paddock, comprising the lands contiguous to the farmstead—the
only portion of his demesne to which the owner gives any special
attention, and on the improvement of which he bestows any labour.
This “tún” is enclosed within a wall of stone or turf, and averages an
extent of two or three acres, though sometimes it reaches to ten. Its
surface is usually a series of closely-packed mounds, like an
overcrowded graveyard, with channels or water-runs between, about
two feet deep. Hither every person employed on the farm, or whom
the farmer can engage, resorts, with short-bladed scythe and rake,
and proceeds to cut down the coarse thick grass, and rake it up into
little heaps.
Afterwards the mowers hasten to clear the neighbouring hill sides
and undrained marshes.
This primitive haymaking, so unlike the systematic operation
which bears that name in England, is carried on throughout the
twenty-four hours of the long summer day. The hay, when sufficiently
dried, is made up into bundles, and tied with cords and thongs, and
packed on the back of ponies, which carry it to the clay-built stalls or
sheds prepared for it. And a curious sight it is to see a long string of
hay-laden ponies returning home. Each pony’s halter is made fast to
the tail of his predecessor; and the little animals are so
overshadowed and overwhelmed by their burdens, that their hoofs
and the connecting ropes alone are visible, and they seem like so
many animated haycocks, feeling themselves sufficiently made up,
and leisurely betaking themselves to their resting-places.
During the protracted winter the Icelander, of course, can attend
to no out-of-door labour, and passes his time within his hut, which, in
many parts of the island, is not much superior to an Irish “cabin.”
The lower part is built of rough stones up to a height of four feet,
and between each course a layer of turf is placed, which serves
instead of mortar, and helps to keep out the cold. The roof, made of
any available wood, is covered with turf and sods. On the southern
side the building is ornamented with doors and gable-ends, each of
which is crowned by a weathercock. These doors are the entrances
to the dwelling-rooms and various offices, such as the cow-shed,
store-house, and smithy. The dwelling-rooms are connected by a
long, dark, narrow passage, and are separated from each other by
strong walls of turf. As each apartment has its own roof, the building
is, in effect, an aggregate of several low huts, which receive their
light through small windows in the front, or holes in the roof, covered
with a piece of glass or skin. The floors are of stamped earth; the
fireplace is made of a few stones, rudely packed together, while the
smoke escapes through a hole in the roof, or through a cask or
barrel, with the ends knocked out, which acts as chimney.
In some parts of the island lava is used instead of stones, and
instead of wood the rafters are made of the ribs of whales. A horse’s
skull is the best seat provided for a visitor. Too often the same room
serves as the dining, sitting, and sleeping place for the whole family,
and the beds are merely boxes filled with feathers or sea-weed.
There are, however, a few houses of a superior character, in which
the arrangements are not much unlike those of a good old-fashioned
English farm-house; the walls being wainscotted with deal, and the
doors and staircase of the same material. A few prints and
photographs, some book-shelves, one or two little pictures, decorate
the sitting-room, and a neat iron stove, and massive chests of
drawers, furnish it sufficiently.
From the houses we turn to the churches. In Reikiavik the church
is a stone building, the only stone building in the town; but this is
exceptional: most of the churches are not much better than the
houses. We will be content, therefore, with a visit to the Reikiavik
sanctuary, which is a neat and unpretending erection, capable of
accommodating three or four hundred persons. The Icelanders are
not opposed to a “decent ritual,” and the Lutheran minister wears a
black gown with a ruff round his neck. The majority of the
congregation, here as everywhere else, consists of women; some
few dressed in bonnets, and the rest wearing the national black silk
skull-cap, set jauntily on one side of the head, with a long black
tassel drooping to the shoulder, or else a quaint mitre-like structure
of white linen, almost as imposing as the head-dress of a Normandy
peasant. The remainder of an Icelandic lady’s costume, we may add,
consists of a black bodice, fastened in front with silver clasps, over
which is drawn a cloth jacket, gay with innumerable silver buttons;
round the neck goes a stiff ruff of velvet, embroidered with silver
lace; and a silver belt, often beautifully chased, binds the long dark
wadmal petticoat round the waist. Sometimes the ornaments are of
gold, instead of silver, and very costly.
Towards the end of the Lutheran service, the preacher descends
from the pulpit, and attiring himself in a splendid crimson velvet
cope, turns his back to the congregation, and chants some Latin
sentences.
Though still retaining in their ceremonies, says a recent traveller,
a few vestiges of the old religion, though altars, candles, pictures,
and crucifixes yet remain in many of their churches, the Icelanders
are stanch Protestants, and singularly devout, innocent, and pure-
hearted. Crime, theft, debauchery, cruelty are unknown amongst
them; they have neither prison, gallows, soldiers, nor police; and in
the manner of their lives mingles something of a patriarchal
simplicity, that reminds one of the Old World princes, of whom it has
been said, that they were “upright and perfect, eschewing evil, and in
their hearts no guile.”
In the rural districts, if such a phrase can properly be applied to
any part of Iceland, the church is scarcely distinguishable from any
other building, except by the cross planted on its roof. It measures,
generally, from eight to ten feet in width, and from eighteen to
twenty-four feet in length; but of this space about eight feet are
devoted to the altar, which is divided off by a wooden partition
stretching across the church, just behind the pulpit. The communion-
table is nothing more than a small wooden chest or cupboard, placed
at the end of the building, between two small square windows, each
formed by a single common-sized pane of glass. Over the table is
suspended a sorry daub, on wood, intended to represent the Last
Supper. The walls, which are wainscotted, are about six feet high;
and stout beams of wood stretch from side to side. On these are
carelessly scattered a number of old Bibles, psalters, and loose
leaves of soiled and antiquated manuscripts. The interior of the roof,
the rafters of which rest on the walls, is also lined with wood.
Accommodation, in the shape of a few rough benches, is provided
for a congregation of thirty or forty souls.
Poor as are the churches, the pastors are still poorer. The best
benefice in the island is worth not much more than £40 per annum;
the average value is £10. The bishop himself does not receive more
than £200. The principal support of the clergy, therefore, is derived
from their glebe-land, eked out by the small fees charged for
baptisms, marriages, and funerals.
Such being the case, the reader will not be surprised to learn that
the Icelandic clergy live miserably and work hard. They assist in the
haymaking; they hire themselves out as herdsmen; they act as the
leaders of the caravans of pack-horses which carry the produce of
the island to the ports, and return loaded with domestic necessaries;
and they distinguish themselves as blacksmiths, as veterinarians,
and shoers of horses.
Dr. Henderson gives an interesting and graphic account of a visit
he paid to one of these “poor parsons,” John Thorlukson, who, while
supporting himself by drudgery of this painful kind, translated
Milton’s “Paradise Lost” and Pope’s “Essay on Man” into Icelandic.
“Like most of his brethren, at this season of the year,” says Dr.
Henderson, “we found him in the meadow assisting his people at
haymaking. On hearing of our arrival he made all the haste home
which his age and infirmity would allow, and bidding us welcome to
his lowly abode, ushered us into the humble apartment where he
translated my countrymen into Icelandic. The door was not quite four
feet in height, and the room might be about eight feet in length by six
in breadth. At the inner end was the poet’s bed; and close to the
door, over against a small window, not exceeding two feet square,
was a table where he committed to paper the effusions of his muse.
On my telling him that my countrymen would not have forgiven me,
nor could I have forgiven myself, had I passed through this part of
the island without paying him a visit, he replied that the translation of
Milton had yielded him many a pleasant hour, and often given him
occasion to think of England.”
It is true that this passage was written some fifty-five years ago,
but the condition of the clergy of Iceland has not much improved in
the interval.

Travelling in Iceland, even under the more favourable conditions

brought about by a constant influx of tourists, is not to be achieved
without difficulty and discomfort. Not only is the country destitute,
necessarily, of inns and the usual arrangements for the convenience
of travellers, but much, very much, depends upon the weather. With
a bright sky overhead, it is possible to regard as trivial and unworthy
of notice the small désagréments which, in bad weather, develop into
very serious annoyances. The only mode of travelling is on
horseback, for as there are no roads, carriages would be useless;
while the distances between the various points of interest are too
great, the rivers too violent, and the swamps too extensive for
pedestrian tours to be undertaken. Even the most moderate-minded
tourist requires a couple of riding-horses for himself, a couple for his
guide, and a couple of pack-horses; and when a larger company
travels, it expands into a cavalcade of from twenty to thirty horses,
tied head to tail, which slowly pick their way over rugged lava-beds
or dangerous boggy ground.
It is one thing, as Lord Dufferin remarks, to ride forty miles a day
through the most singular scenery in the world, when a glorious sun
brings out every feature of the landscape into startling distinctness,
transmuting the dull tormented earth into towers, domes, and
pinnacles of shining metal, and clothes each peak in a robe of many-
coloured light, such as the “Debatable Mountains” must have been in
Bunyan’s dream; and another to plod over the same forty miles, wet
to the skin, seeing nothing but the dim gray bases of the hills, which
rise you know not how, and care not where. “If, in addition to this,
you have to wait, as very often must be the case, for many hours
after your own arrival, wet, tired, hungry, until the baggage-train, with
the tents and food, shall have come up, with no alternative in the
meantime but to lie shivering inside a grass-roofed house, or to
share the quarters of some farmer’s family, whose domestic
arrangements resemble in every particular those which Macaulay
describes as prevailing among the Scottish Highlanders a hundred
years ago; and if, finally, after vainly waiting for some days to see an
eruption which never takes place, you journey back to Reikiavik
under the same melancholy conditions, it will not be unnatural that,
on returning to your native land, you should proclaim Iceland, with
her geysers, to be a sham, a delusion, and a snare!”
There are no bridges in Iceland; no bridges, except, indeed, a
few planks flung across the Bruera, and a swing-bridge, or kláfe,
which spans the Jokülsa; and, as is still the case in some parts of the
Scottish Highlands, the traveller must ford the streams, which are
always rapid, and sometimes inconveniently deep. The passage of a
river is, therefore, a formidable enterprise, as may be inferred from
the experiences of Mr. Holland and other travellers.
The guide leads the way, and the caravan follow obediently in his
wake, stemming, as best they can, the swift impetuous torrent. Often
the boiling water rises high against the horse’s shoulders, and
dashes clouds of spray in the face of the riders. The stream is so
furiously fast that it is impossible to follow the individual waves as
they sweep by, and to look down at it almost makes one dizzy. Now,
if ever, is the time for a firm hand, a sure seat, and a steady eye: not
only is the current strong, but its bed is full of large stones, which the
horse cannot see through the dark waters; and should he fall, the
torrent will carry you down to the sea, whose white breakers are