An Introduction To Viscosity Solutions

for Fully Nonlinear PDE with

Applications to Calculus of Variations
in L 1st Edition Nikos Katzourakis
(Auth.)
Visit to download the full and correct content document:
https://textbookfull.com/product/an-introduction-to-viscosity-solutions-for-fully-nonline
ar-pde-with-applications-to-calculus-of-variations-in-l-1st-edition-nikos-katzourakis-aut
h/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Introduction to Simple Shock Waves in Air With

Numerical Solutions Using Artificial Viscosity Seán
Prunty

https://textbookfull.com/product/introduction-to-simple-shock-
waves-in-air-with-numerical-solutions-using-artificial-viscosity-
sean-prunty/

Calculus of Variations An Introduction to the One

Dimensional Theory with Examples and Exercises 1st
Edition Hansjörg Kielhöfer (Auth.)

https://textbookfull.com/product/calculus-of-variations-an-
introduction-to-the-one-dimensional-theory-with-examples-and-
exercises-1st-edition-hansjorg-kielhofer-auth/

Multivariable Calculus with MATLAB With Applications to

Geometry and Physics Ronald L. Lipsman

https://textbookfull.com/product/multivariable-calculus-with-
matlab-with-applications-to-geometry-and-physics-ronald-l-
lipsman/

An Introduction to Cryptocurrencies The Crypto Market

Ecosystem Nikos Daskalakis

https://textbookfull.com/product/an-introduction-to-
cryptocurrencies-the-crypto-market-ecosystem-nikos-daskalakis/
Stochastic Processes and Calculus An Elementary
Introduction with Applications 1st Edition Uwe Hassler

https://textbookfull.com/product/stochastic-processes-and-
calculus-an-elementary-introduction-with-applications-1st-
edition-uwe-hassler/

An Introduction to Intermediate and Advanced

Statistical Analyses for Sport and Exercise Scientists
1st Edition Nikos Ntoumanis

https://textbookfull.com/product/an-introduction-to-intermediate-
and-advanced-statistical-analyses-for-sport-and-exercise-
scientists-1st-edition-nikos-ntoumanis/

An Introduction To Statistical Learning With

Applications In R Gareth James

https://textbookfull.com/product/an-introduction-to-statistical-
learning-with-applications-in-r-gareth-james/

Calculus of Variations 1st Edition Filip Rindler

https://textbookfull.com/product/calculus-of-variations-1st-
edition-filip-rindler/

Introduction to Statistics and Data Analysis With

Exercises Solutions and Applications in R 1st Edition
Christian Heumann

https://textbookfull.com/product/introduction-to-statistics-and-
data-analysis-with-exercises-solutions-and-applications-in-r-1st-
edition-christian-heumann/
SPRINGER BRIEFS IN MATHEMATICS

Nikos Katzourakis

An Introduction
To Viscosity Solutions
for Fully Nonlinear
PDE with Applications
to Calculus of
∞
Variations in L
SpringerBriefs in Mathematics

Series editors
Krishnaswami Alladi, Gainesville, USA
Nicola Bellomo, Torino, Italy
Michele Benzi, Atlanta, USA
Tatsien Li, Shanghai, People’s Republic of China
Matthias Neufang, Ottawa, Canada
Otmar Scherzer, Vienna, Austria
Dierk Schleicher, Bremen, Germany
Benjamin Steinberg, New York, USA
Yuri Tschinkel, New York, USA
Loring W. Tu, Medford, USA
G. George Yin, Detroit, USA
Ping Zhang, Kalamazoo, USA

SpringerBriefs in Mathematics showcases expositions in all areas of mathematics

and applied mathematics. Manuscripts presenting new results or a single new result
in a classical field, new field, or an emerging topic, applications, or bridges between
new results and already published works, are encouraged. The series is intended for
mathematicians and applied mathematicians.
BCAM SpringerBriefs
Editorial Board
Enrique Zuazua
BCAM—Basque Center for Applied Mathematics & Ikerbasque
Bilbao, Basque Country, Spain

Irene Fonseca
Center for Nonlinear Analysis
Department of Mathematical Sciences
Carnegie Mellon University
Pittsburgh, USA

Juan J. Manfredi
Department of Mathematics
University of Pittsburgh
Pittsburgh, USA

Emmanuel Trélat
Laboratoire Jacques-Louis Lions
Institut Universitaire de France
Université Pierre et Marie Curie
CNRS, UMR, Paris

Xu Zhang
School of Mathematics
Sichuan University
Chengdu, China

BCAM SpringerBriefs aims to publish contributions in the following disciplines:

Applied Mathematics, Finance, Statistics and Computer Science. BCAM has
appointed an Editorial Board, who evaluate and review proposals.
Typical topics include: a timely report of state-of-the-art analytical techniques,
bridge between new research results published in journal articles and a contextual
literature review, a snapshot of a hot or emerging topic, a presentation of core
concepts that students must understand in order to make independent contributions.
Please submit your proposal to the Editorial Board or to Francesca Bonadei, Executive
Editor Mathematics, Statistics, and Engineering: francesca.bonadei@springer.com

More information about this series at http://www.springer.com/series/10030

Nikos Katzourakis

An Introduction To Viscosity
Solutions for Fully Nonlinear
PDE with Applications to
Calculus of Variations in L∞

123
Nikos Katzourakis
Department of Mathematics and Statistics
University of Reading
Reading
UK

ISSN 2191-8198 ISSN 2191-8201 (electronic)

ISBN 978-3-319-12828-3 ISBN 978-3-319-12829-0 (eBook)
DOI 10.1007/978-3-319-12829-0

Library of Congress Control Number: 2014954594

Springer Cham Heidelberg New York Dordrecht London

© The Author(s) 2015

This work is subject to copyright. All rights are reserved 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. Exempted from this legal reservation are brief
excerpts in connection with reviews or scholarly analysis or material supplied specifically for the
purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the
work. Duplication of this publication or parts thereof is permitted only under the provisions of
the Copyright Law of the Publisher’s location, in its current version, and permission for use must always
be obtained from Springer. Permissions for use may be obtained through RightsLink at the Copyright
Clearance Center. Violations are liable to prosecution under the respective Copyright Law.
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.
While the advice and information in this book are believed to be true and accurate at the date of
publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for
any errors or omissions that may be made. The publisher makes no warranty, express or implied, with
respect to the material contained herein.

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

To my beloved wife
Preface

This set of notes corresponds to the lectures of a postgraduate short course given by
the author at the BCAM—Basque Centre for Applied Mathematics at Bilbao of
Spain in early May 2013.
The purpose of these notes is to give a quick and elementary, yet rigorous,
presentation of the rudiments of the so-called theory of Viscosity Solutions which
applies to fully nonlinear first- and second- order Partial Differential Equations
(PDE). For such equations, particularly for second- order ones, solutions generally
are nonsmooth and standard approaches in order to define a “weak solution” do not
apply: classical, strong almost everywhere (a.e.), weak, measure-valued and dis-
tributional solutions either do not exist or may not even be defined. The main
reason for the latter failure is that, the standard idea of using “integration-by-parts”
in order to pass derivatives to smooth test functions by duality, is not available for
non-divergence structure PDE.
The name of this theory originates from the “vanishing viscosity method”
developed first for first- order fully nonlinear PDE (Hamilton-Jacobi PDE). Today,
though, it comprises an independent theory of “weak” solutions which applies to
fully nonlinear elliptic and parabolic PDE and in most cases has no or little relation
to the idea of adding a viscosity term. The formal notions have been introduced by
P.L. Lions and M.G. Crandall in the early 1980s for first- order PDE, following
preceding contributions of L.C. Evans. The extension to the case of second- order
PDEs came around the 1990s by H. Ishii and P.L. Lions.
Interesting PDE to which the theory applies arise in Geometry and Geometric
Evolution (Monge-Amperé PDE, Equations of Motion by Mean Curvature),
Optimal Control and Game Theory (Hamilton-Jacobi-Bellman PDE, Isaacs PDE,
Differential Games) and Calculus of Variations in Lp and L1 (Euler-Lagrange PDE,
p-Laplacian, Aronsson PDE, 1-Laplacian).
Due to the vastness of the subject, a drastic choice of material is required for a
brief and elementary introduction to the subject of Viscosity Solutions. In the case
at hand, our main criterion has been nothing more but personal taste. Hence, herein,
we shall restrict ourselves to the second- order degenerate elliptic case, focusing in
particular on applications in the modern field of Calculus of Variations in L1 .

vii
viii Preface

An inspection of the Table of Contents gives an idea about the organisation of

the material. A rather immediate observation of the expert is that, unlike most
standard texts on Viscosity Solutions where Uniqueness and Comparison form the
centre of gravity of the exposition, herein they appear rather late in the presentation
and are not overemphasised. This shift of viewpoint owes to that the author is
directed mostly towards extensions of the theory to the vector case of systems. In
this realm, the primary focus switches to existence methods, while comparison and
uniqueness are not true in general, not even in the most ideal cases.
Throughout these notes, no previous knowledge is assumed on the reader’s
behalf. Basic graduate-level mathematical maturity suffices for the reading of the
first six chapters, which is the general theory. For the next two chapters which
concern applications to Calculus of Variations, some familiarity with weak deriv-
atives and functionals is assumed, which does not go much deeper than the defi-
nitions. The last chapter collects, mostly without proofs, extensions and perhaps
more advanced related topics which we were not able to cover in detail in this
introduction to the subject.
There exist several excellent expository texts in the literature on Viscosity
Solutions and Calculus of Variations in L1 . In particular, we would like to point
out [CIL, C2, L1, Ko, C1, B, G, MR, Dr]. However, all the sources we are aware of
are either of more advanced level, or have a different viewpoint. We believe that the
main contribution of the present text is elementary and is mostly addressed to
students and non-experts. This “textbook style” is reflected also in the fact that
references do not appear in the main text. Most of the results are not optimal, but
instead there has been a huge, and hopefully successful effort for the main ideas to
be illustrated as clearly as possible. We hope that these notes serve as a suitable first
reading on the theory of Viscosity Solutions.

Reading, UK, May 2014 Nikos Katzourakis

Acknowledgments

I am indebted to Enrique Zuazua for inviting me to write this Brief for Springer.
Special thanks are due to Juan Manfredi for his share of expertise on the subject
of the present notes during several personal communications. I would like to thank
Tristan Pryer for the careful reading of an earlier version of this manuscript. His
suggestions certainly made the material more readable. I also warmly thank
Federica Dragoni for her remarks and suggestions. Last but not least, I am grateful
to the referees of this monograph whose numerous suggestions improved the
quality of the presentation.

ix
Contents

Part I General Theory

1 History, Examples, Motivation and First Definitions. . . . . . . . . . . . 3

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2 Second Definitions and Basic Analytic Properties

of the Notions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

3 Stability Properties of the Notions and Existence via

Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

4 Mollification of Viscosity Solutions and Semiconvexity . . . . . . . . . . 49

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

5 Existence of Solution to the Dirichlet Problem via

Perron’s Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

6 Comparison Results and Uniqueness of Solution to the Dirichlet

Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

Part II Applications

7 Minimisers of Convex Functionals and Existence of Viscosity

Solutions to the Euler-Lagrange PDE . . . . . . . . . . . . . . . . . . . . . . 91
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

xi
xii Contents

8 Existence of Viscosity Solutions to the Dirichlet Problem

for the ‘-Laplacian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

9 Miscellaneous Topics and Some Extensions of the Theory . . . . . . . 111

9.1 Fundamental Solutions of the 1-Laplacian . . . . . . . . . . . . . . . . 111
9.1.1 The 1-Laplacian and Tug-of-War Differential Games . . . . 113
9.1.2 Discontinuous Coefficients, Discontinuous Solutions . . . . . 115
9.1.3 Barles-Perthame Relaxed Limits (1-Sided Uniform
Convergence) and Generalised 1-Sided Stability . . . . . ... 116
9.1.4 Boundary Jets and Jets Relative to Non-open Sets. . . . ... 118
9.1.5 Nonlinear Boundary Conditions . . . . . . . . . . . . . . . . ... 119
9.1.6 Comparison Principle for Viscosity Solutions Without
Decoupling in the x-variable . . . . . . . . . . . . . . . . . . . ... 121
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 123
 Part I
General Theory
Chapter 1
History, Examples, Motivation and First
Definitions

What is it all about?

Viscosity Solutions form a general theory of “weak” (i.e. non-differentiable)

solutions which applies to certain fully nonlinear Partial Differential Equations
(PDE) of 1st and 2nd order.

Let u : Ω ⊆ Rn −→ R be a function in C 2 (Ω), n ≥ 1. In the standard way, C k (Ω)

denotes the space of k-times continuously differentiable functions over the domain
Ω.
Consider the PDE

F(·, u, Du, D 2 u) = 0, (1.1)

where
∂
Du = (D1 u, . . . , Dn u), Di ≡ , i = 1, . . . , n (Gradient vector),
∂ xi
 i=1,...,n
D 2 u = Di2j u j=1,...,n (Hessian matrix),

and
F : Ω × R × Rn × S(n) −→ R

is the function defining the equation1 whose arguments are denoted by

F(x, r, p, X ), (x, r, p, X ) ∈ Ω × R × Rn × S(n).

1 It is customary to call F the “coefficients” and we will occasionally follow this convention. This
terminology is inherited from the case of linear equations, where F consists of functions multiplying
the solution and its derivatives. We will also consistently use
 the notation “·” for the argument of x,
namely F(·, u, Du, D 2 u)|x = F x, u(x), Du(x), D 2 u(x) . We will not use the common clumsy
notation F(x, u, Du, D 2 u).
© The Author(s) 2015 3
N. Katzourakis, An Introduction To Viscosity Solutions for Fully Nonlinear PDE
with Applications to Calculus of Variations in L ∞ ,
SpringerBriefs in Mathematics, DOI 10.1007/978-3-319-12829-0_1
4 1 History, Examples, Motivation and First Definitions

F is always assumed to be continuous, unless stated otherwise. The set S(n)

denotes the symmetric n × n matrices:
  
n×n 
S(n) := X = (X i j )i=1,...,n
j=1,...,n ∈ R  X i j = X ji .

The 2nd order PDE (1.1) trivially includes the case of fully nonlinear 1st order PDE
(Hamilton–Jacobi PDE):
F(·, u, Du) = 0. (1.2)

The PDE (1.1) (and hence (1.2)) is our basic object of study. As usually, here and
subsequently,

the modifier “fully nonlinear” stands for “F may not linear in any of its
arguments”, but of course linear equations are not excluded. We will say that
an equation is genuinely fully nonlinear, if is nonlinear with respect to the
highest order derivatives.

If the partial function

X → F(x, r, p, X )

is linear, the PDE is typically called quasilinear and hence F can be written in the
form
F(x, r, p, X ) = Ai j (x, r, p)X i j + B(x, r, p) (1.3)

for a matrix function

A : Ω × R × Rn −→ Rn×n

and a scalar function B : Ω ×R×Rn −→ R. The summation in i, j in (1.3) is tacitly

assumed without explicitly writing the sum (“Einstein’s convention”). By denoting
the Euclidean inner product of the matrix space Rn×n by “:”, that is

A : B := Ai j Bi j = tr(A B),

the general quasilinear PDE can be written as

A(·, u, Du) : D 2 u + B(·, u, Du) = 0. (1.4)

1 History, Examples, Motivation and First Definitions 5

Historical comment

Viscosity Solutions were first introduced in the 1980s by Crandall and Lions
for F(·, u, Du) = 0 as a

uniqueness criterion,

in order to select one of the infinitely-many strong a.e. Lipschitz solutions of

the Dirichlet problem for F(·, u, Du) = 0.

Example 1 (Non-uniqueness of strong solutions) The Dirichlet problem for the

1-dimensional Eikonal PDE:

|u |2 − 1 = 0, on (−1, +1),
u(±1) = 0,

admits infinitely many Lipschitz continuous solutions u : (−1, +1) → R, with

|u |2 = 1 a.e. on (−1, +1). Indeed, one of these solutions is the distance from the
endpoints of the interval:

x − 1, on (−1, 0),
u 1 (x) :=
1 − x, on (0, 1),

which solves |u |2 = 1 on (−1, 0) ∪ (0, 1). Then, by reflection of the graph of u 1

with respect to the horizontal line u = 1/2, the function
⎧
⎪
⎪ x − 1, on (−1, −1/2),
⎨
−x, on (−1/2, 0),
u 2 (x) :=
⎪
⎪ x, on (0, 1/2),
⎩
1 − x, on (0, 1),

is also a solution of |u |2 = 1 on (−1, −1/2) ∪ (−1/2, 0) ∪ (0, 1/2) ∪ (1/2, 1). By

reflection of the graph of u 2 with respect to the line u = 1/4, we construct a new
solution u 3 , and so on. See Fig. 1.1.
In many problems, strong a.e. solutions of Hamilton–Jacobi PDE were con-
structed by utilising the Vanishing Viscosity method.

The idea of the Vanishing Viscosity method: In order to solve the PDE

F(·, u, Du) = 0,

we approximate by solutions of the elliptic PDE

εΔu ε + F(·, u ε , Du ε ) = 0,
6 1 History, Examples, Motivation and First Definitions

Fig. 1.1 The folding construction of Example 1

and investigate the limit of (u ε )ε>0 as ε → 0. The hope then is that we can obtain
some sort of estimates which do not collapse as ε → 0 that will allow to pass to a limit
u ε → u which “solves” the original PDE (“Δ” of course stands for the Laplacian).

Viscosity Solutions originate from the “Vanishing Viscosity method”, but

should not be confused with the latter!!!
In most cases, especially for 2nd order PDE, Viscosity Solutions have nothing
to do with adding an “εΔu” term.

Historical comment

For 2nd order PDE, problems more fundamental than uniqueness arise: “sin-
gular solutions” appear, which can not be rigorously justified, because

F(·, u, Du, D 2 u) = 0

does not support integration by parts!!!

The meaning of the above statement is that we can not “pass derivatives to test
functions” by integration by parts in order to have, for example, a function which is
only (weakly) once differentiable as solution of a 2nd order equation.

Remark 2 (Digression into divergence structure PDE) For the 2nd order PDE
 
Di Ai (·, u, Du) + B(·, u, Du) = 0, (1.5)

where u : Ω ⊆ Rn → R, we have the option to define weak solutions by duality,

that is by taking “integration by parts” as definition when testing against smooth
1 History, Examples, Motivation and First Definitions 7

functions with compact support:


−Ai (·, u, Du) Di ψ + B(·, u, Du)ψ = 0,
Ω

for all ψ ∈ Cc∞ (Ω).

In particular, divergence structure PDE are always quasilinear and not genuinely
fully nonlinear. Indeed, by assuming that A, B are C 1 and distributing derivatives in
(1.5), we have

Ai p j (·, u, Du)Di2j u + Ai r (·, u, Du)Di u

+ Ai xi (·, u, Du) + B(·, u, Du) = 0. (1.6)

Hence, the dependence in D 2 u is linear. In (1.6), the subscripts p j , r, xi denote

differentiation of the coefficients A, B with respect to the respective arguments.

Remark 3

Viscosity Solutions form a duality-free

“NONLINEAR DISTRIBUTION THEORY”,
whose ideas are needed even for linear PDE!!!

j=1,...,n
Example 4 Let A = (Ai j )i=1,...,n : Ω ⊆ Rn −→ S(n) be a map with values non-
negative matrices, such that either A has a vanishing eigenvalue, or A is not locally
Lipschitz continuous, then the linear PDE

Ai j (x)Di2j u(x) = 0,

may not have any non-trivial solutions (the affine ones) in all standard senses! Explicit
A’s will be given later.
On the other hand, standard elliptic estimates (Schauder) imply that if A is both
Hölder continuous C α and strictly positive, then the PDE has only smooth C 2,α
solutions.
This problem is the same as the problem to interpret (and solve) rigorously the
equation
H = δ,

where δ is the Dirac “function” and H = χ(0,∞) is a discontinuous solution, without

having Distributions (Generalised Functions) and Measures at hand.
8 1 History, Examples, Motivation and First Definitions

However, there are limitations: Viscosity Solutions apply to degenerate elliptic and
degenerate parabolic PDE. The appropriate ellipticity notion will be given after we
will many classes of equations to which our theory applies.

Example 5 (PDE to which Viscosity Solutions apply)

(1) An all-important equation: the ∞-Laplacian For u ∈ C 2 (Ω) and Ω ⊆ Rn , the
infinity-Laplacian is
Δ∞ u := Du ⊗ Du : D 2 u = 0, (1.7)

where the expression in (1.7) is understood as the double sum Di u D j u Di2j u. The
geometric meaning of (1.7) is that the rank-one tensor product matrix Du ⊗ Du is
pointwise normal to D 2 u in Rn×n . Equation (1.7) is the fundamental PDE of Calculus
of Variations in L ∞ (the analogue of the Euler-Lagrange PDE), when considering
variational problems for the so-called Supremal functionals, the simplest of which is

E ∞ (u, Ω ) := Du L ∞ (Ω ) , Ω  Ω. (1.8)

Here Ω is an open set compactly contained in Ω and u is locally Lipschitz contin-

uous, that is, in the Sobolev space
  

1,∞
Wloc (Ω) := u ∈ L ∞
loc (Ω)  ∃ Du a.e. on Ω and Du ∈ L ∞
loc (Ω) n
.

The study of such problems has been initiated by Aronsson in the 1960s, but their
rigorous study started in the 1990s by utilising the theory of Viscosity Solutions.
Both (1.7) and (1.8) arise also in the geometric problem of Optimisation of Lip-
schitz Extensions of functions, Differential games, (implicitly) in Motion by Mean
Curvature ...
The ∞-Laplacian is the formal limit of the p-Laplacian as p → ∞. The
p-Laplacian is  
Δ p u := Di |Du| p−2 Di u = 0,

and Δ p u = 0 is the Euler-Lagrange PDE of the p-Dirichlet functional

E p (u, Ω) := |Du| p .
Ω

The ∞-Laplacian is a quasilinear, degenerate (elliptic) PDE in non-divergence form,

and will be our primary example in these notes.
Hard-to-interpret “singular solutions” of the ∞-Laplacian:
(a) by rewriting (1.7) as

1 
Δ∞ u = Di u Di |Du|2 = 0, (1.9)
2
1 History, Examples, Motivation and First Definitions 9

we see that every classical (i.e. C 1 ) solution of the Eikonal PDE

|Du|2 − 1 = 0,

formally solves (1.9), but not (1.7) (When we write (1.7) expanded in the form of
(1.8), we have the product of distribution with a C 0 function, which is not well
defined) !!!
(b) In two dimensions, the ∞-Laplacian takes the form

Δ∞ u = (u x )2 u x x + 2u x u y u x y + (u y )2 u yy = 0

(in x − y notation), when applied to a function u = u(x, y), u : R2 → R. By

separating variables and looking for solutions of the form u(x, y) = f (x) + g(y), a
simple calculation leads to the saddle “solution”

u(x, y) = x 4/3 − y 4/3

which is merely C 1,1/3 (R2 ) Hölder continuous and lacks 2nd derivatives along the
axes x = 0 and y = 0 !!!
In both (a) and (b), it might be that the singular (not twice differentiable)
“solution” is the unique C 1 solution of (an appropriately formulated) minimisa-
tion problem for the supremal functional (1.8). The latter makes perfect sense if only
1st derivatives exist, at least a.e. on the domain, with no need for reference to 2nd
derivatives.

(2) More examples

(a) Linear 2nd order PDE (static, or 1st order in time) with C 0 coefficients:

Ai j (x)Di2j u(x) + Bk (x)Dk u(x) + c(x)u(x) = 0,

−u t + Ai j (x)Di2j u(x) + Bk (x)Dk u(x) + c(x)u(x) = 0.

(b) Quasilinear 2nd order divergence PDE (static, or 1st order in time) with A in
C 1 and B in C 0 :
 
Di Ai (·, u, Du) + B(·, u, Du) = 0,
 
−u t + Di Ai (·, u, Du) + B(·, u, Du) = 0.

(c) Quasilinear 2nd order non-divergence PDE (static, or 1st order in time) with
A, B in C 0 :

Ai j (·, u, Du)Di2j u + B(·, u, Du) = 0,

−u t + Ai j (·, u, Du)Di2j u + B(·, u, Du) = 0.
10 1 History, Examples, Motivation and First Definitions

This case includes (a) and (b) as special instances.

(d) Hamilton–Jacobi–Bellman type and Isaacs type equations (static, or 1st order
in time):
 
sup Aiaj (·)Di2j u + B(·, u, Du) = 0,
a∈A
 
−u t + sup Aiaj (·)Di2j u + B(·, u, Du) = 0,
a∈A

and
 
inf sup Aiabj (·)Di2j u + B(·, u, Du) = 0,
b∈B a∈A
 
−u t + inf sup Aiabj (·)Di2j u + B(·, u, Du) = 0,
b∈B a∈A

where {Aiaj }a∈A and {Aiabj }a∈A

b∈B
are families of linear coefficients. Both classes of PDE
are fully nonlinear.
(e) Functions of the eigenvalues of the Hessian (static, or 1st order in time): the
Hessian D 2 u is (pointwise) a real symmetric n × n matrix. By the Spectral Theorem,
the spectrum σ (D 2 u) consists of n real eigenvalues, which we place in increasing
order:  
σ (D 2 u) = λ1 (D 2 u), . . . , λn (D 2 u) , λi ≤ λi+1 .

Then, consider the PDE of the form

 
G ·, u, Du; λ1 (D 2 u), . . . , λn (D 2 u) = 0,
 
−u t + G ·, u, Du; λ1 (D 2 u), . . . , λn (D 2 u) = 0.

This class consists of fully nonlinear PDE.

Interesting special cases: n
(i) take G(l1 , . . . , ln , p, r, x) := i=1 li − f (r ). Then we obtain the nonlinear
Poisson equation:
Δu = f (u).

(ii) Take G(l1 , . . . , ln , p, r, x) := i=1

n l − f (x). Then we obtain the Monge-
i
Amperé equation
det(D 2 u) = f.

(f) Obstacle problems and Gradient Constraint Problems: this includes fully non-
linear PDE of the general form
 
max F(D 2 u), u − f = 0,
 
min F(D 2 u), |Du| − f = 0.
1 History, Examples, Motivation and First Definitions 11

Below we give the appropriate notion of ellipticity, which is necessary for the devel-
opment of the theory.
Definition 6 (Degenerate Ellipticity)2 The PDE
 
F ·, u, Du, D 2 u = 0

is called degenerate elliptic, when the coefficient

F : Ω × R × Rn × S(n) −→ R

satisfies the weak monotonicity

X ≤ Y in S(n) =⇒ F(x, r, p, X ) ≤ F(x, r, p, Y ), (1.10)

for all (x, r, p) ∈ Ω × R × Rn .

Remark 7
(α) We recall that, by definition, the inequality A ≥ 0 in the matrix space S(n) ⊆
Rn×n means that σ (A) ⊆ [0, ∞), that is all the eigenvalues of the symmetric
matrix A are non-negative:

A : q ⊗ q = Ai j qi q j ≥ 0, q ∈ Rn .

Of course, A ≤ B means B − A ≥ 0. This definition is weak enough to include

the case of Hamilton–Jacobi PDE F(·, u, Du) = 0 which has no dependence
on 2nd derivatives. Obviously, “degenerate parabolic PDE”
 
−u t + F ·, u, Du, D 2 u = 0

are defined to be degenerate elliptic in the previous sense, since derivative in

“time” is of first order.
In these notes we will focus on the elliptic case. All the definitions, the
techniques and the arguments in the proofs extend to the “parabolic” case with
minor modifications.
(β) Degenerate ellipticity in the examples:
(a) the ∞-Laplacian is degenerate elliptic. Indeed, by defining

F∞ ( p, X ) := X : p ⊗ p,

2 In other texts, ellipticity is defined with the opposite inequality. This choice of convention that

we make is more convenient, since we consider non-divergence operators like Δ∞ . We do not use
integration by parts, hence there is no reason to consider “minus the operator”. According to our
convention, Δ and Δ∞ are elliptic, while for text using the opposite convention, −Δ and −Δ∞ are
elliptic.
12 1 History, Examples, Motivation and First Definitions

we have F∞ ( p, X ) − F∞ ( p, Y ) = (X − Y ) : p ⊗ p ≤ 0, when X − Y ≤ 0.
In the examples (b)–(d), the PDE is degenerate elliptic when the coefficient
matrix of D 2 u is non-negative.
For (e), the PDE is degenerate elliptic when for each i = 1, . . . , n, the func-
tion li → G(x, r, p; l1 , . . . , ln ) is non-decreasing. In particular, the Monge-
Amperé PDE is degenerate elliptic when we restrict ourselves to the cone of
convex functions u and we also assume f ≥ 0.
In (f), the PDE is degenerate elliptic when F is.

Now we have enough material in order to proceed to the motivation and the
definition of the central notion of “weak solutions” we will use in these notes.
The pedagogical idea behind Viscosity Solutions is to:

use the Maximum Principle in order to “pass derivatives to smooth test func-
tions” in a nonlinear fashion, without duality!

Motivation of the Definition Suppose that u ∈ C 2 (Ω) is a classical solution of the

PDE  
F x, u(x), Du(x), D 2 u(x) = 0, x ∈ Ω, (1.11)

and that (1.11) is degenerate elliptic, that is, F satisfies (1.10).

Assume further that at some x0 ∈ Ω, u can be “touched from above” by some
smooth function ψ ∈ C 2 (Rn ) at x0 . By the latter, we mean that the difference has a
vanishing (local) maximum in a neighbourhood of x0 (see Figs. 1.2 and 1.3):

u − ψ ≤ 0 = (u − ψ)(x0 ), on a ball Br (x0 ) ⊆ Ω. (1.12)

Fig. 1.2 A 3D illlustration of the upper test function

1 History, Examples, Motivation and First Definitions 13

Fig. 1.3 A 2D illlustration of

the upper test function

By (1.12) and Calculus, at the maximum point x0 we have vanishing gradient and
non-positive hessian:

D(u − ψ)(x0 ) = 0, in Rn (1.13)

D (u − ψ)(x0 ) ≤ 0, in S(n).
2
(1.14)

By using that u is a solution, we have

 
0 = F x0 , u(x0 ), Du(x0 ), D 2 u(x0 )
   
= F x0 , ψ(x0 ), Dψ(x0 ), D 2 u(x0 ) by (1.12), (1.13)
   
≤ F x0 , ψ(x0 ), Dψ(x0 ), D 2 ψ(x0 ) by (1.11), (1.14) .

Hence, we conclude that if u is a solution of (1.11), then

⎫
u − ψ ≤ 0 = (u − ψ)(x0 ) ⎬  
on a ball Br (x0 ) ⊆ Ω, ⇒ F x0 , ψ(x0 ), Dψ(x0 ), D 2 ψ(x0 ) ≥ 0.
⎭
x0 ∈ Ω, ψ ∈ C 2 (Rn )
(1.15)
Similarly, we conclude that if u is a solution of (1.11), then
⎫
u − φ ≤ 0 = (u − φ)(y0 ) ⎬  
on a ball Br (y0 ) ⊆ Ω, ⇒ F y0 , φ(y0 ), Dφ(y0 ), D 2 φ(y0 ) ≤ 0. (1.16)
⎭
y0 ∈ Ω, φ ∈ C 2 (Rn )

The crucial observation now is that

the implications (1.15) and (1.16) have been derived by using that u ∈ C 2 (Ω)
is a classical solution of the PDE (1.11), but DO NOT DEPEND ON THE
EXISTENCE OF (Du, D 2 u) !!!
14 1 History, Examples, Motivation and First Definitions

In particular, (1.15) and (1.16) make sense when u is merely continuous. Hence, this
leads to the idea that we can extract (Du, D 2 u) from the PDE, by testing against
smooth “touching” functions, and using the Maximum Principle together with the
ellipticity of the PDE. Consequently, we have the next definition:

Definition 8 (Viscosity Solutions)3 Let u ∈ C 0 (Ω), Ω ⊆ Rn , and consider the

degenerate elliptic PDE (see (1.10))
 
F ·, u, Du, D 2 u = 0. (1.17)

(a) We say that u is a Viscosity Subsolution of the PDE (or a Viscosity Solution of
F ·, u, Du, D 2 u ≥ 0) when
⎫
u − ψ ≤ 0 = (u − ψ)(x0 ) ⎬  
on a ball Br (x0 ) ⊆ Ω, ⇒ F x0 , ψ(x0 ), Dψ(x0 ), D 2 ψ(x0 ) ≥ 0.
⎭
x0 ∈ Ω, ψ ∈ C 2 (Rn )
(1.18)

(b) We say that uis a Viscosity Supersolution of the PDE (or a Viscosity Solution of
F ·, u, Du, D 2 u ≤ 0) when
⎫
u − φ ≤ 0 = (u − φ)(y0 ) ⎬  
on a ball Br (y0 ) ⊆ Ω, ⇒ F y0 , φ(y0 ), Dφ(y0 ), D 2 φ(y0 ) ≤ 0. (1.19)
⎭
y0 ∈ Ω, φ ∈ C 2 (Rn )

(c) We say that u is a Viscosity Solution, when it is both a Viscosity Subsolution and
a Viscosity Supersolution.

In words, the definition means that

when we touch from above (respectively, below) by a smooth test function at

a point, then, the test function is a subsolution of the equation (respectively,
supersolution) at that point.

3 In other texts Viscosity sub/super solutions are defined with the opposite inequalities. The direc-

tion of the inequalities corresponds to the choice of convention in the ellipticity notion. However,
no confusion should arise for the readers because the definition is essentially the same in both
cases: when it comes down to writing the inequalities for, say, Δ∞ , in either choice of conventions
we have
Dψ(x) ⊗ Dψ(x) : D 2 ψ(x) ≥ 0
when ψ touches u from above at x ∈ Ω and

Dφ(y) ⊗ Dφ(y) : D 2 φ(y) ≤ 0

when φ touches u from below at y ∈ Ω.

1 History, Examples, Motivation and First Definitions 15

Remark 9
(a) It might look strange to the reader that the definition is actually a criterion:
whenever we can touch from either above or below, then we have an appropri-
ate inequality. At points where no upper or lower touching functions exist, the
candidate for solution automatically solves the PDE at these points.
(b) It is not a priori clear that the definition is not void, that is that we can indeed
find such touching functions at “sufficiently many” points on Ω. If this were not
the case for a continuous candidate solution u, this u would automatically be a
solution since we would have nothing to check.
(c) Splitting the notion of solution to 1-sided halves is essential. If we had assumed
that we can simultaneously touch from above and below at any point, then as
we will see in the next chapter, we would not really define a new truly “weak”
solution, since for this hypothetical notion 1st derivatives would have to be
Lipschitz continuous at the touching point.
(d) It is not clear yet that the notion is compatible with classical solutions. That is,
we do not yet know whether the family of C 2 Viscosity Solutions and the family
of classical solutions of a degenerate elliptic PDE coincide.

All the issues raised in the remark, will be answered in the next Chapter, where
we also investigate another equivalent definition and exploit its analytic properties.
This alternative point of view is based on the possibility to define “pointwise weak
derivatives (Du, D 2 u)”, without reference to “test functions”. Note that

merely C 0 functions can be interpreted as solutions of fully nonlinear PDE!

The fact that we can drop all derivatives from the PDE, has the following out-
standing implication that we will see in following chapters:

VISCOSITY SOLUTIONS PASS TO LIMITS UNDER MERELY C 0 CON-

VERGENCE, NO CONTROL ON ANY DERIVATIVES IS REQUIRED !!!

In the above, “C 0 convergence” stands for locally uniform convergence, namely,

the convergence arising from the natural topology of the space C 0 (Ω). Also,
16 1 History, Examples, Motivation and First Definitions

the Dirichlet problem

  
F ·, u, Du, D 2 u = 0, in Ω,
u = b, on ∂Ω,

has a unique continuous Viscosity Solution, under natural assumptions on

F, Ω, b!

The latter solution might be nowhere differentiable on Ω, and hence (Du, D 2 u)

might have no classical sense! Such strong convergence and existence results hold
only for the linear theory of Distributions (Generalised Functions) for PDE with
constant coefficients.

Remarks on Chapter 1 Viscosity solutions were introduced first by Crandall and

Lions in [CL83] as a uniqueness criterion for 1st order PDE. The essential idea
regarding passage to limits was observed earlier by Evans in [Ev78, Ev80]. As usu-
ally, the original point of view of the notions was slightly different. Crandall-Evans-
Lions [CEL84] have definitions similar to the ones we use today. The viscosity
notions were later extended to the 2nd order case by Ishii in [Is89]. Aronsson was
the first to consider L ∞ variational problems in [Ar65]. The ∞-Laplacian made
its first public appearance in his subsequent papers [Ar67, Ar68]. Aronsson him-
self was also the first to observe in [Ar84, Ar86] “singular solutions” which at the
time were hard-to-justify (visosity theory had not been develop yet). Battacharya-
DiBenedetto-Manfredi were the first to consider the ∞-Laplacian in the Viscosity
sense in [BDM89]. Background material on divergence structure PDE (although
remotely needed for this chapter) can be found in standard references like the text-
book of Evans [E4]. For PDE theory of classical solutions to nonlinear elliptic equa-
tions, the standard reference is Gilbarg-Trudinger [GT98]. For regularity estimates
of viscosity solutions to fully nonlinear equations, the interested reader may consult
(the advanced book of) Cabré-Caffarelli [CC95].

References

[Ar65] G. Aronsson, Minimization problems for the functional supx F(x, f (x), f (x)). Arkiv
für Mat. 6, 33–53 (1965)
[Ar67] G. Aronsson, Extension of functions satisfying Lipschitz conditions. Arkiv für Mat.
6, 551–561 (1967)
[Ar68] G. Aronsson, On the partial differential equation u 2 xu x x + 2u x u y u x y + u 2 yu yy = 0.
Arkiv für Mat. 7, 395–425 (1968)
[Ar84] G. Aronsson, On certain singular solutions of the partial differential equation u 2 xu x x +
2u x u y u x y + u 2 yu yy = 0. Manuscripta Math. 47, 133–151 (1984)
References 17

[Ar86] G. Aronsson, Construction of singular solutions to the p-Harmonic equation and its
Limit equation for p = ∞. Manuscripta Math. 56, 135–158 (1986)
[BDM89] T. Bhattacharya, E. DiBenedetto, J. Manfredi, Limits as p → ∞ of Δ p u p = f and
related extremal problems, Rend. Sem. Mat. Univ. Poi. Torino Fascicolo Speciale,
Nonlinear PDE’s (1989)
[CC95] L.A. Caffarelli, X. Cabré, Fully Nonlinear Elliptic Equations, AMS, Colloquium Pub-
lications 45, Providence (1995)
[CEL84] M.G. Crandall, L.C. Evans, P.L. Lions, Some properties of viscosity solutions of
Hamilton-Jacobi equations. Trans. Amer. Math. Soc. 282, 487–502 (1984)
[CIL92] M.G. Crandall, H. Ishii, P.L. Lions, User’s guide to viscosity solutions of 2nd order
partial differential equations. Bull. AMS. 27, 1–67 (1992)
[CL83] M.G. Crandall, P.L. Lions, Viscosity solutions of Hamilton-Jacobi equations. Trans.
Am. Math. Soc. 277, 1–42 (1983)
[E4] L.C. Evans, Partial differential equations. Graduate studies in mathematics. Am. Math.
Soc. (1991)
[Ev78] L.C. Evans, A convergence theorem for solutions of nonlinear second-order elliptic
equations. Indiana Univ. Math. J. 27, 875–887 (1978)
[Ev80] L.C. Evans, On solving certain nonlinear partial differential equations by accretive
operator methods. Israel J. Math. 36(225), 220–247 (1980)
[GT98] D. Gilbarg, N. Trudinger, Elliptic Partial Differential Equations of Second Order,
Classics in Mathematics, reprint of the, 1998th edn. (Springer, Heidelberg, 1998)
[Is89] H. Ishii, On uniqueness and existence of viscosity solutions of fully nonlinear second-
order elliptic PDE’s. Comm. Pure Appl. Math. 42, 1–45 (1989)
Chapter 2
Second Definitions and Basic Analytic
Properties of the Notions

In the previous chapter we defined a notion of non-differentiable “weak” solutions

which applies to degenerate elliptic PDE of the general form
 
F ·, u, Du, D 2 u = 0 (2.1)

that makes sense when u is merely in C 0 (Ω), Ω ⊆ Rn . As always, the “coefficients”

F = F(x, r, p, X ) is a continuous and possibly nonlinear function defined on Ω ×
R × Rn × S(n). We recall that the appropriate degenerate ellipticity condition we
need to impose is

X ≤ Y in S(n) =⇒ F(x, r, p, X ) ≤ F(x, r, p, Y ). (2.2)

In this chapter we will handle the issues raised at the end of the previous chapter, the
most important of which were
(a) is the definition of Viscosity Solutions non-void? Namely, is it true that any
continuous function can be touched from above and below by a smooth function at
“sufficiently many” points of the domain, in order to uniquely determine a solution
on open subsets?
(b) Are Viscosity Solutions compatible with classical solutions? Namely, a twice
differentiable Viscosity Solutions of (2.1) is the same as a classical solution?
(c) Why is it necessary to split the notion to two halves?
Were the answer “no” to (a), then the definition would be trivial since, if such a
hypothetical continuous function existed, would be a solution to ALL equations in
the Viscosity sense!
Were the answer “no” to (b), the definition would obviously be naive. If one was
to prove existence of a “weak” solution which happens to be twice differentiable,
then this solution should definitely solve the PDE pointwise.
Concerning (c), as we shall see a little later, if we were to require that we can
touch from above and below by smooth functions simultaneously at the same points,
then we would not go in a class much weaker than that of strong solutions.
We begin by introducing an alternative but equivalent definition.

© The Author(s) 2015 19

N. Katzourakis, An Introduction To Viscosity Solutions for Fully Nonlinear PDE
with Applications to Calculus of Variations in L ∞ ,
SpringerBriefs in Mathematics, DOI 10.1007/978-3-319-12829-0_2
20 2 Second Definitions and Basic Analytic Properties of the Notions

Motivation for the Definition The idea is to somehow

DEFINE POINTWISE GENERALISED DERIVATIVES (Du, D 2 u) FOR

NON SMOOTH FUNCTIONS BY MEANS OF THE MAXIMUM PRIN-
CIPLE !

We begin by slightly relaxing the classical Calculus definition of derivatives. The

ideas is simple–we just take as definition the Taylor expansion at a point. This makes
no difference for 1st derivatives, but is essential for 2nd (and higher order) derivatives.

Definition 1 (Pointwise Derivatives as Taylor Expansions) Let u ∈ C 0 (Ω) and

x ∈ Ω ⊆ Rn . We say that u is twice differentiable at x if there exist p ∈ Rn and
X ∈ S(n) such that

1
u(z + x) = u(x) + p · z + X : z ⊗ z + o(|z|2 ), (2.3)
2
as z → 0. In such case, we call p the gradient of u at x and denote it by Du(x) and
we call X the hessian of u at x and denote it by D 2 u(x).
In the standard way, we shall say that u is once differentiable at x when

u(z + x) = u(x) + p · z + o(|z|), (2.4)

as z → 0, and then p will be denoted by Du(x).

Remark 2 We use “·” to denote the Euclidean inner product in Rn . The meaning
of (2.3) is that there exists a continuous increasing function σ ∈ C 0 [0, ∞) with
σ(0) = 0 such that
 
 
u(y) − u(x) − p · (y − x) − 1 X : (y − x) ⊗ (y − x) ≤ σ(|y − x|)|y − x|2 ,
 2 
(2.5)
for y ∈ Ω. If such a pair (Du(x), D 2 u(x)) exists, then it is unique. Equation (2.4) is
nothing but the classical notion of 1st derivative.

The new ingredient of (2.3) is that we defined D 2 u(x) at x, although the gradient
Du may not exist as a function near x. In classical calculus, D 2 u(x) is defined by
differentiating the gradient map Du : Ω ⊆ Rn −→ Rn at x.
Another random document with
no related content on Scribd:
The Project Gutenberg eBook of The young
volcano explorers
 This ebook is for the use of anyone anywhere in the United
States and most other parts of the world at no cost and with
almost no restrictions whatsoever. You may copy it, give it away
or re-use it under the terms of the Project Gutenberg License
included with this ebook or online at www.gutenberg.org. If you
are not located in the United States, you will have to check the
laws of the country where you are located before using this
eBook.

Title: The young volcano explorers

Or, American boys in the West Indies

Author: Edward Stratemeyer

Illustrator: A. B. Shute

Release date: September 3, 2023 [eBook #71552]

Language: English

Original publication: Boston: Lothrop, Lee & Shepard Co, 1902

Credits: David Edwards, David E. Brown, and the Online

Distributed Proofreading Team at https://www.pgdp.net

*** START OF THE PROJECT GUTENBERG EBOOK THE YOUNG

VOLCANO EXPLORERS ***
Down came the youth, head first.—Page 145.
 Pan-American Series

THE
YOUNG VOLCANO
EXPLORERS
OR
AMERICAN BOYS IN THE WEST INDIES

BY
EDWARD STRATEMEYER
Author of “Lost on the Orinoco,” “With Washington in the
West,”
“American Boys’ Life of William McKinley,” “Old Glory Series,”
“Ship and Shore Series,” “Bound to Succeed Series,” etc.

ILLUSTRATED BY A. B. SHUTE

BOSTON:
LOTHROP, LEE & SHEPARD CO.
1930
Copyright, 1902, by Lee and Shepard
Entered at Stationers’ Hall, London

All rights reserved

The Young Volcano Explorers

PRINTED IN U.S.A.
 PREFACE
“The Young Volcano Explorers” is a complete story in itself, but
forms the second volume of a line issued under the general title of
the “Pan-American Series.”
When I began this series of tales I had in mind to acquaint my young
readers with some of the sights to be seen in the three Americas,—
especially such portions as lie outside of the United States. In the
first volume, called “Lost on the Orinoco,” I told of the sight-seeing
and adventures of five American lads, who, in company with their
academy professor, who is also a great traveler and hunter, journey
to Venezuela, our sister Republic on the north-east coast of South
America. The boys visit several principal cities, inspect cocoa and
coffee plantations as well as gold and silver mines, and then explore
the mighty river already named.
In the present volume the scene is shifted from Venezuela to the
West Indies, that group of islands of which Porto Rico has already
become the property of the United States. Because of the recent
Spanish-American War, and the still more recent volcanic
disturbances in Martinique and St. Vincent, these islands are of
unusual interest to us. In this book the boys and their instructor sail
from Venezuela to Jamaica, stopping at Kingston, and then go to
Havana, Cuba. In Cuba several places of importance are visited, and
then the trip is continued to Hayti and to Porto Rico, where the party
travel overland from San Juan to Ponce. At the latter city word is
received that the fathers of two of the boys have gone to St. Pierre,
Martinique, and the party start for that point, only to encounter the
effects of the volcanic eruption when still far at sea. But Martinique is
visited, nevertheless, and later on St. Vincent also, and in spite of
the many dangers, all ends happily.
In penning this tale I have, as usual, tried to be as accurate as
possible when giving historical or geographical details. The latest
and best American and Spanish authorities have been consulted,
and, in the case of the disasters at Martinique and St. Vincent, I have
read with care the reports of all who suffered and escaped, and of
those who have since visited these spots of interest.
Once again I wish to thank the many thousands who have perused
my former works. May the present volume fulfil their every
expectation.
Edward Stratemeyer.
 CONTENTS
CHAPTER PAGE
I. Introducing the Boys 1
II. News from Home 10
III. Off for Jamaica 20
IV. A Talk About a Meal 30
V. The Defense of January Jones 42
VI. An Interruption to Sight-Seeing 53
VII. The Joke on Captain Sudlip 62
VIII. Sight-Seeing in Havana 72
IX. The Disappearance of Hockley 82
X. The Old Convent 93
XI. A Strange Story 103
XII. About Cuba and Tobacco Raising 114
XIII. A Cuban Baseball Game 123
XIV. Winning a Victory 132
XV. Hockley Takes a Bath and Gives One 141
XVI. Good-bye to Cuba 152
XVII. A Joke on Shipboard 161
XVIII. Hayti, Sugar Making, and Another Joke 172
XIX. A Talk About Porto Rico 183
XX. An Adventure in the Mountains 193
XXI. Across Porto Rico on Horseback 204
XXII. Hockley in Trouble 213
XXIII. The Bully is Humbled 223
XXIV. Something About Earthquakes and Volcanoes 233
XXV. A Collision at Sea 243
XXVI. The Lumber Raft 256
XXVII. Stone Dust and Boiling Water 265
XXVIII. The Eruption of Mont Pelee 274
XXIX. The Destruction of St. Pierre 285
XXX. Looking for the Missing Ones 296
XXXI. Dangerous Volcano Exploring 306
XXXII. The Fate of Captain Sudlip 316
XXXIII. A Happy Meeting—Conclusion 325
 PAN
AMERICAN
SERIES
THE YOUNG VOLCANO
EXPLORERS
 CHAPTER I
INTRODUCING THE BOYS

“Here comes the mail boy, Darry. Now for letters and newspapers
from home. There ought to be plenty of letters for all of us.”
“Don’t be too sure Frank. The mails in Venezuela are mighty slow,
especially now when there is another revolution on.”
“Oh, I don’t believe this one-horse revolution will have anything to do
with the mails,” put in a third boy, as he joined his chums, who were
lounging in the hallway of a spacious hotel in Caracas, the capital
city of Venezuela. “They have them too often, you know.”
“Hush, don’t call it a one-horse revolution, Beans,” returned Darry
Crane, in a warning voice. “The people who live here think it’s just as
important as any revolution that ever was—and it probably is, to
them.”
“There goes the professor for letters now!” burst out Frank Newton.
“Oh, I do hope he gets lots of them!” Frank was always more anxious
for letters from home than anyone else. “Yes, here he comes with a
handful, and an armful of papers in the bargain.”
The coming of the mail always attracted a crowd of patrons of the
hotel, and soon the boys found themselves surrounded by those
anxious to get their letters and papers. Looking, they saw the
gentleman who had their mail wave his hand to them and disappear
in the direction of the hotel courtyard and they speedily followed.
“What have you for me, Professor?” was the question asked by one
and all, and now two more boys hurried up, making five in all who
waited eagerly for news from home.
It was an interesting group, and while the mail matter is being
distributed, we will take the opportunity of introducing each individual
to the reader.
The oldest boy present was Mark Robertson. He was a lad of
seventeen, and was the son of a dry goods importer. His father
owned an interest in several mills in England and Scotland and
made semi-yearly trips across the Atlantic, and the family were well-
to-do.
When at home Mark lived on Madison Avenue in New York city, and
directly opposite to him lived Frank Newton, another of the boys of
the group. Although Frank was a year younger than Mark, the two
were warm chums. Frank’s father was a banker, and if he was not a
millionaire he was certainly well provided for financially.
The liveliest boy in the crowd was Dartworth Crane, always called
Darry for short. He was but fifteen, the son of a rich Chicago cattle
dealer, and to him life was one long, sunshiny holiday. It was very
hard for Darry to take anything seriously, and his good nature was as
spontaneous as it was catching.
“Darry would make a cow laugh,” said Frank, more than once, and
the others agreed with him. As said before, Mark was Frank’s closest
chum, but Darry was no mean second.
The fourth lad of the group was a tall, well-built individual of sixteen,
with a high forehead and a thick mass of curly hair. This was Samuel
Winthrop, generally called “Beans,” because he had been born and
brought up in Boston. Sam was the son of a well-to-do widow of the
Back Bay district. He was a studious, observant young fellow,
seldom, however, given to “airing his knowledge,” and he and Mark
were as friendly as were Frank and Darry.
The fifth youth in the crowd was a tall, lank individual of about Mark’s
age, with a white freckled face and reddish hair. His name was
Jacob Hockley, and he was the son of a millionaire lumber dealer of
Pennsylvania. His manner was varied, at times exceedingly “bossy,”
as the others termed it, and then again exceedingly sour and
morose. The latter mood had won for him the nickname of “Glummy”
or “Jake the Glum,” and although he objected strenuously to being
called such a name, yet it clung to him in spite of everything. Hockley
had plenty of money and spent it freely, but even this failed to make
him any close friendships.
“Glummy thinks money is everything,” said Mark in speaking of the
matter one day. “But sooner or later he is bound to learn that there
are some things that even money can’t buy.” And Mark was right.
True friendship is never a matter of dollars and cents.
For several years all these boys had attended a boarding academy
located among the hills of New Hampshire. Lakeview Academy, as it
was called, was presided over by Professor Amos Strong, a kindly
and well educated gentleman, who had in years gone by been a
great traveler and hunter. Professor Strong had often told the lads
about his hunting expeditions in various parts of the globe, and
through these stories a plan had originated to visit Central and South
America, the expedition to be under the personal supervision of the
professor himself.
At first Professor Strong could not see his way clear to leaving the
academy, but a fire came and destroyed the place, and at the same
time the professor’s brother, also a teacher, left the faculty of
Harvard. It was then arranged that the school building should be
rebuilt under the directions of the brother, who was afterward to
assume control of the institution. This would give Professor Strong
the liberty he desired and which he, in secret, greatly craved. For
many years a wanderer on the face of the earth, binding himself
down to steady teaching had proved rather irksome to him.
After a good deal of discussion it was decided that the party of six
should first visit Venezuela, and in the first volume of this series,
entitled “Lost on the Orinoco,” I related the particulars of the journey
from New York to La Guayra, the nearest seaport, and told of the
sight-seeing and adventures while visiting Caracas, the capital,
Macuto, the fashionable summer resort, the great Gulf of Maracaibo,
and other points of interest, including cocoa and coffee plantations
and gold and silver mines, and also a never-to-be-forgotten journey
up that immense river, the Orinoco, the second largest stream in
South America.
The boys had had adventures in plenty, and becoming lost on the
Orinoco had almost proved a serious happening for Mark and Frank,
who had wandered away in a jungle that seemed to have no end. All
of the party had met more than one wild animal, and a squall on
Lake Maracaibo had come close to sending them all to the bottom.
It was not to be expected that four such whole-souled lads as Mark,
Frank, Darry and Sam could get along smoothly with such an over-
bearing and peculiar youth as Jake Hockley. They were sorry that
the bully was along, and it was not long before there was a bitter
quarrel and some of the boys came to blows with Hockley. This was
stopped by Professor Strong, who said they must do better in the
future.
With no special friend in the party, and with a strong desire to be
“sporty,” and to do things which were not permitted by the professor,
Hockley struck up an acquaintance with one Dan Market, a man
from Baltimore, whose reputation was none of the best. This Markel
succeeded in getting all of Hockley’s money away from him, and it
was only through a discovery made by Mark and Frank that Markel
was arrested and the money was recovered. But Markel had
escaped, and what had become of him nobody knew. Getting back
his money and also a watch which had been taken, had made
Hockley friendly to Frank and Mark for the time being, and the bully
was also friendly to Sam and Darry, for they had saved him from
being crushed to death by a boa constrictor, having shot and killed
the hideous reptile just in the nick of time.
It had been the intention of Professor Strong to take the boys from
Venezuela to Brazil, but while the party was resting at a mining town
called Castroville, there came in a report that the tropical fevers were
raging in the latter republic, and that it was likely Venezuela, Brazil,
Colombia, and some other countries would soon be mixed up in
revolutions and wars, and it was then decided by a general vote that
they should move northward again and visit the West Indies, taking
in Jamaica, Cuba, Porto Rico, and other important islands.
“That will just suit me,” said Mark, when this decision was reached. “I
want to see where the battles of the Spanish-American War were
fought and also what sort of a place Porto Rico, our new possession,
is.”
“And I want to get out of a country that grows boa constrictors,”
came from Hockley. “Ugh, they’re awful. Professor, are there such
snakes in the West Indies?”
“I never heard of any, Jacob. But they have poisonous reptiles, such
as are to be found in all tropical climates.”
“Well, I won’t care so much for them, if they aren’t six or seven yards
long,” grumbled the tall youth.
“That’s Glummy all over,” whispered Frank to Darry. “Every time he
speaks of that constrictor he tacks a yard or two on to the length.”
“Never mind—it’s only a snake story, you know,” returned light-
hearted Darry. “They don’t count, for nobody believes them. We’ll
never get credit for killing anything more than a snake as long as
your arm.”
The journey from Castroville to Caracas overland had been
uneventful. They had stopped at numerous plantations and small
towns on the way, and they had seen sheep and wild horses without
number. They had also done considerable hunting, and each of the
boys could now boast of being a creditable shot. If there was an
exception it was Hockley, who, in spite of all the instructions given
him by the professor, would take little pains in the handling of his
gun.
On arriving at Caracas they put up at the same hotel which had been
their stopping place on first visiting the capital. All had long before
sent letters home speaking of their change of plans, and they were
now anxious to ascertain how their parents would view the matter,
even though, on leaving home, everything had been left to Professor
Strong personally, he being fully competent to judge of what was
best for all.
 CHAPTER II
NEWS FROM HOME

Letters from home are always welcome, and doubly so when one
has been out of reach of the mail for a long time. As Professor
Strong held up nearly a dozen written communications, the lads fairly
pounced upon them, and for the time being the newspapers were
forgotten.
“Here’s one for Mark and two for Frank,” cried Darry. “Where is
mine?”
“Here you are,” said Frank, holding out three. “And here is one for
you, Jake.”
“And here’s another for Jake,” added Sam. “And two for myself.”
“No, boys, these two are my own!” laughed Professor Strong,
holding them back. “You mustn’t expect all the good things, you
know.”
“Only one for me!” murmured Mark. “And Darry has three. That’s
hardly fair.”
“You want to send your folks a bottle of ink and some pens,” replied
Darry. “Never mind; I’ll tell you what’s in mine, after I read ’em.”
“Don’t say a word,” burst out Frank. “Look at that letter, twelve or
fifteen pages, and this one of mine has only four pages. Mark
needn’t complain.”
“Here’s a photograph,” came slowly from Sam. “My mother! It was
nice of her to send that.” And he gazed at the picture affectionately.
Soon all were sitting in the hammocks and easy chairs at hand
reading the letters received. There was all sorts of news from home,
of more or less interest. Mark’s father was just home from a trip to
England and Scotland, and he and Mr. Newton were contemplating a
trip together, for business as well as pleasure. Sam’s mother had left
Boston to visit relatives up in Maine, and hoped her only son would
do his best to keep out of peril and mischief. Darry’s letters told how
his father had gone to Colorado to look up a big cattle deal, and
there was also a letter from a sister who was just leaving home to go
to boarding school. Hockley had word that his father had lost one
saw mill by fire but had bought two mills to take its place, and that
the elder Hockley was now president of a new Consolidated Lumber
Company.
“Tell you what, pop’s getting there,” said Hockley, when he told this
news to the others. “He’ll be the richest lumber dealer in the country
before he gets through,” and he fairly swelled with pride over the
announcement.
“I wonder where my father and yours will go,” said Frank to Mark,
after the letters had been read, not once but several times. “It’s
queer neither my letter nor yours tells that.”
“I guess they hadn’t made up their minds when the letters were
written. I once heard father speak of coming down to Cuba and
Jamaica. It would be queer if they did come down and we met them.”
“That would be just all right,” was the ready return, but as Mark
spoke he never dreamed of the terrible circumstances which was to
make that meeting a reality. Could he have looked ahead it would
have caused him more than one shudder.
It had been determined that they should take the train for La Guayra
on the following morning. There was to be a sailing of a steamer for
Kingston and Havana one day later, and Professor Strong had
already telegraphed ahead for the necessary accommodations.
“We can take our last look around Caracas to-day,” announced
Professor Strong. “Have you boys any place in mind that you would
like to visit?”
A general discussion arose. While it was at its height a gentleman
who had driven to the hotel in his carriage came in and rushed up to
the group.
“My own very dear friend, Amos Strong, once more,” he said, with a
strong Spanish accent. “I received word last night that you and the
boys had come back to Caracas. So I made up my mind I must see
you all again. And how have you been?” And he shook hands
cordially.
“We are all right, Morano,” was the professor’s reply, as the boys
crowded around to speak to the newcomer, who was a teacher at the
University at Caracas and an old college friend of Professor Strong.
“But we’ve had lots of adventures,” put in Darry.
“To be sure, Master Darry—you could not keep out of them. You see
I have not forgotten how one of my horses once ran away with you,”
and Enrique Morano laughed.
Besides being an instructor at the Caracas University, Enrique
Morano was the owner of a large plantation just outside of the city,
which the party had once visited, much to the delight of all hands.
Now Morano insisted that he be allowed to send for his largest
carriage and take them for a drive to such points of interest as they
elected to see.
This was just what the boys wished, and inside of an hour the
carriage was on hand and they piled in. Enrique Morano himself
drove, with Professor Strong beside him. Soon the city itself was left
behind and they were bowling along over a smooth highway in the
direction of Antimano, situated some miles to the westward.
It was a perfect day and the boys enjoyed the sights greatly, as they
passed plantation after plantation. The roadway was lined at some
points with beautiful tropical trees, and flowers were by no means
lacking.
“Tell you what, Señor Morano keeps good horse-flesh,” remarked
Darry as they spun along. “This is better than a drive in Lincoln
Park.”
“Or Central Park either,” added Frank.
“Oh, my father keeps just as good horses,” came from Hockley. “He’s
got a trotter that can beat all creation for stepping out on a smooth
road like this.”
So far they had passed but few turnouts on the highway. But now
they saw approaching a carriage with a single seat, drawn by a team
of horses which were covered with foam. On the seat sat two men,
both evidently Americans.
“Reckless drivers,” murmured Enrique Morano, as the other carriage
came closer. “Be careful there!” he called out, in Spanish, and pulled
sharply to one side.
“Clear the track!” shouted one of the men, in English, and the tone of
voice showed that he had been drinking. “We don’t get out of the
way for no sun-baked native!” And he cracked his whip loudly.
The carriage came closer and it was only by a few inches that it
missed striking Enrique Morano’s turnout. As it swept by the boys
got a good look at the occupants.
“Dan Markel!” almost shouted Hockley and Mark, in a breath.
“That was Markel, just as sure as you are born,” came from Frank.
“Who would have expected to meet him here?”
“We ought to go after him,” continued Hockley.
“You evidently know the man,” said Enrique Morano, bringing his
team to a halt and gazing inquiringly at the boys.
“He is the man who robbed Hockley,” explained Professor. “He was
captured once, down on the Orinoco, but he escaped.”
“Ah, I see. Yes, he should be caught. And the other man, what of
him?”
He looked at the boys again, but all shook their heads. They could
not remember having seen the individual before.
“Guess he’s another victim,” was Mark’s comment. “I don’t believe
Markel had any money of his own. He has struck up an
acquaintance with some newcomer and is doing the sponging act.”
“Or else he is spending the proceeds of another robbery,” said Sam.
“Do you wish to go after the man?” asked Enrique Morano.
“Oh, well, you might as well let him go,” yawned Hockley. “If we
make another complaint and he is locked up, we may have to stay
here as witnesses against him.” Hockley would have liked to see
Dan Markel behind the bars but he was afraid that the rascal might
tell of some things which would prove discreditable to both of them.
“I’d really like to know if that other man was a victim,” mused Frank,
as they continued on their way. “If he is, he ought to be warned.”
There the matter was dropped, and for the time being Dan Markel
was forgotten. When they came back to Caracas at nightfall they
looked in several directions for the rascal but could catch no sight of
him. The truth of the matter was that Markel had recognized several
of the boys on the instant of passing them, and although partly
intoxicated he still had wit enough left to keep hidden.
The party were to leave for La Guayra at nine in the morning, and
the boys were up long before that time, taking a last look at their
surroundings. Then came a good breakfast, and soon they were on
the cars and winding over and around the mountains which separate
the capital city of Venezuela from the seacoast.
“Here is where we had to get off and walk,” said Mark, as the train
rushed on. “Do you remember that, Frank?”
“To be sure I do. And I remember how you got lost in a hole under
the cliffs, too.”
“Yes, and not far away is the spot where you and Hockley pitched
into each other,” whispered Mark, with a sly look at the bully, who sat
just ahead. He did not think the lank youth heard, but he was
mistaken.
“Raking up old sores, eh?” growled Hockley, swinging around and
with his face very red. “I thought all that was to be dropped.”
“It is to be dropped, too, Jake,” answered Mark, quickly.
“Oh, yes,” came with a sneer. “You’re dropping it fast enough.”
“But I didn’t really mean anything, Jake,” pleaded Mark. “It—er—just
came to my mind, that’s all. It’s past and gone now.”
“You can’t humbug me, Mark Robertson! You’re laughing in your
sleeve because you think Frank got the best of me in that fight. But
let me tell you I would have come out on top if Professor Strong
hadn’t come up and stopped us.”
“Maybe you would not have come out on top,” said Frank, dryly, for
his temper was rising. “I fancy I had the best of it by a good deal.
Anyway, your teeth——”
“Oh, let that old quarrel drop,” came from Darry, who had caught the
latter part of the conversation. “We’re out for a good time, and let us
have it.”
“I’m willing to let it drop,” said Mark, readily.
“I wouldn’t have said a word, only Hockley—” began Frank.
“That’s it, blame everything on me!” howled the bully, his anger
getting the better of him. “I knew that pretended friendliness of yours
wouldn’t last. You are all down on me and you know it. But I’ll show
you a trick or two before we’re done—you see if I don’t!”
Professor Strong had gone to the end of the car for a drink of water
and to gaze for a moment out of the doorway. Now he returned to his
seat near the boys, and the talk came to a sudden end.
 CHAPTER III
OFF FOR JAMAICA

“He is the same old Hockley,” murmured Mark to Frank, when he got
the chance. “His friendliness was all put on.”
“No, I don’t think that, exactly,” returned Frank. “I think he meant
well, but he’s one of the kind who won’t let matters rest. I suppose it
galls him to think I had the better of that fight, and some day he’ll try
to square accounts.”
“In that case, Frank, you’ll have to be on guard.”
“Oh, don’t worry; I’ll keep my eyes open.”
As my old readers know, La Guayra is only a small seaport, located
on a stretch of land between the water and the high cliffs of the
mountains. It is a dirty, ill-smelling place, and nobody lives there who
can help it.
“I’m glad we haven’t to stop here long,” said Sam. “It smells like dead
fish and oil mixed. Where is the steamer?”
An hour later found them on board of the craft, an old-fashioned, tub
sort of an affair named the Chester. She was an English boat
devoted to the carrying trade between Trinidad, La Guayra,
Kingston, and other points in the Caribbean Sea. Her captain was
named Jason Sudlip, and he was a burly fellow, with a reddish face
and black, piercing eyes.
“This boat ought to be called the Chestnut, instead of the Chester,”
remarked Darry, after he and the others had made a tour of
inspection. “She’s old enough to vote twice over. It’s a wonder she
hasn’t gone to pieces long ago.”
“Better not let the captain hear you talk like that,” came from Sam.
“We’re lucky to get passage, so I was told. Steamers for Kingston
and Havana are scarce.”
“She is no such steamer as brought us down from New York, that’s
sure,” put in Mark. “But knocking around as we are, we’ll have to
take what comes.”
Their baggage had been brought on board and placed in their
staterooms, and now the boys started in to make themselves
comfortable, for the journey to Kingston would last about four days,
and that to Havana three or four days more.
“What a stuffy hole this is,” grunted Hockley, as he gazed around the
stateroom which had been assigned to him and Darry. “I don’t see
why the professor didn’t pick out something worth living in. I can’t
sleep here.”
“The professor said he had done the best he could, Jake.”
“It’s a beastly shame. This ain’t fit for a mule to stall in.”
“Oh, it isn’t as bad as that. However, if you don’t like it, you can
complain to the professor.”
Hockley would not do this and only continued to growl, until Darry
grew so sick of hearing him that he escaped to the deck and there
joined Sam and Frank.
“No two ways about it, Glummy has got a bilious attack,” he
announced. “He has had the sore head ever since we got to
Caracas. Now he’s kicking about the stateroom; says it isn’t fit for a
mule. Of course it isn’t the finest in the world, but it’s not as bad as
that.”
“It’s a pity he is in such a humor,” returned Sam. “And after
everything was going so swimmingly, too. But I think I can explain it,
in part at least.”
“Then do so by all means, Beans,” cried Darry and Frank, in a
breath.
“I learned it by accident, when we came on board. I picked up part of
a letter Hockley had torn up. I didn’t know it was his at the time. It
read to the effect that he couldn’t have any more money at present,
that Mr. Hockley was going to send the funds direct to Professor
Strong. Jake evidently wanted money very much, and his father’s
refusal to give him some has upset him.”
“That would do it, for Glummy always wants to cut a dash with his
rocks,” said Darry. “But he needn’t take it out on us. I’ve got to room
with him, unfortunately, but I shan’t stand much of his cutting up.”
As only three staterooms were to be had, Amos Strong had taken
Sam in with himself and given another room to Frank and Mark. This
just suited the two New York boys and they at once proceeded to
make themselves as comfortable as the limited means permitted.
As the steamer was a small one and carried but few passengers, the
help was also limited, and both the table and the staterooms were
looked after by a tall, gaunt negro who rejoiced in the name of
January Jones. The colored man was from Alabama and had
shipped some months before. He was thoroughly good natured, but
one would never have guessed this from his face, which was usually
as solemn as that of an owl.
“Be pleased to help yo’ young gen’men wid yo’ things,” he
announced, on presenting himself. “Do whatever yo’ wants me to,
sah.”
“Thanks, but everything is about done,” said Mark, who was
addressed. Then he added: “Are you the porter?”
“Yes, sah, I’se de portah an’ de waitah, sah.”
“Two in one, eh? All right, George, we’ll call on you later.”
“Yes, sah—January Jones, sah.”
“January Jones!” came from Frank. “That’s an odd name. How did
they come to name you January?”
“Ole massa did dat, sah—afo’ de wah, sah. Called me January
because I was boahn in July, sah—said eberything was goin’ wrong,
wid that wah comin’ on, an’ things was turnin’ ’round, sah.”
At this Mark and Frank laughed outright.
“He was certainly a humorist, January,” said Mark.
“Massa was a cotton plantah, sah—had a big plantation on de ribber,
sah. He’s dead an’ gone now, an’ so is de missus.”
“Well, then he was a planter as well as a humorist.”
“Didn’t nebber hear him say nuffin ’bout bein’ in de humorist
business, sah. But I was a little piccaninny dem days an’ didn’t know
nuffin. Den I can’t do nuffin fo’ you?”
“Not now, but we’ll want you later,” answered Mark, and January
Jones shuffled off to call on the others.
“That’s what I call a character,” said Mark, after another laugh. “I
reckon we can have some fun with January before this trip comes to
an end.”
Early in the morning the Chester left the harbor of La Guayra, with its
many ships of all nations and its strange native fishing smacks. The
boys and the professor stood at the rail and remained there until land
became a mere blur in the distance.
“Farewell to Venezuela!” cried Sam, taking off his hat and making a
profound bow. “Our trip was full of pleasure and full of excitement.
May the coming trip please us equally.”
“And be less full of peril,” added Professor Strong. “Too much
excitement is worse than none at all.”
“Oh, Professor, we’ve got to have some excitement,” said Darry.
“Now that it is all over I think everything was just about right.”
“But I am under orders to bring you back home safe and sound,”
returned Amos Strong, with a peculiar smile. “And how can I do that
if you insist on placing your head in the lion’s mouth?”
“I haven’t seen a lion yet,” said Darry, with a vacant stare. “Where is
he and where’s his mouth?”
There was a general laugh, in which even the professor joined, for
he loved a joke as well as any of them.
“You know I was speaking metaphorically, Dartworth. Of course I
cannot look ahead, but I sincerely trust that you do what you can to