Download the Full Version of textbook for Fast Typing at textbookfull.

com

Symmetry in Complex Network Systems Connecting

Equivariant Bifurcation Theory with Engineering
Applications In

https://textbookfull.com/product/symmetry-in-complex-
network-systems-connecting-equivariant-bifurcation-theory-
with-engineering-applications-in/

OR CLICK BUTTON

DOWNLOAD NOW

Download More textbook Instantly Today - Get Yours Now at textbookfull.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Network Medicine Complex Systems in Human Disease and

Therapeutics Joseph Loscalzo

https://textbookfull.com/product/network-medicine-complex-systems-in-
human-disease-and-therapeutics-joseph-loscalzo/

textboxfull.com

An Algebraic Approach to Symmetry with Applications to

Knot Theory David Edward Joyce

https://textbookfull.com/product/an-algebraic-approach-to-symmetry-
with-applications-to-knot-theory-david-edward-joyce/

textboxfull.com

Bifurcation in Autonomous and Nonautonomous Differential

Equations with Discontinuities Marat Akhmet

https://textbookfull.com/product/bifurcation-in-autonomous-and-
nonautonomous-differential-equations-with-discontinuities-marat-
akhmet/
textboxfull.com

Individual Differences in Language Learning: A Complex

Systems Theory Perspective Carol Griffiths

https://textbookfull.com/product/individual-differences-in-language-
learning-a-complex-systems-theory-perspective-carol-griffiths/

textboxfull.com
Optimization of Complex Systems: Theory, Models,
Algorithms and Applications Hoai An Le Thi

https://textbookfull.com/product/optimization-of-complex-systems-
theory-models-algorithms-and-applications-hoai-an-le-thi/

textboxfull.com

Multicomponent and Multiscale Systems Theory Methods and

Applications in Engineering 1st Edition Juergen Geiser
(Auth.)
https://textbookfull.com/product/multicomponent-and-multiscale-
systems-theory-methods-and-applications-in-engineering-1st-edition-
juergen-geiser-auth/
textboxfull.com

Complex Dynamical Systems in Education Concepts Methods

and Applications 1st Edition Matthijs Koopmans

https://textbookfull.com/product/complex-dynamical-systems-in-
education-concepts-methods-and-applications-1st-edition-matthijs-
koopmans/
textboxfull.com

Action Research in Software Engineering Theory and

Applications Miroslaw Staron

https://textbookfull.com/product/action-research-in-software-
engineering-theory-and-applications-miroslaw-staron/

textboxfull.com

Emergent Behavior in Complex Systems Engineering A

Modeling and Simulation Approach 1st Edition Saurabh
Mittal
https://textbookfull.com/product/emergent-behavior-in-complex-systems-
engineering-a-modeling-and-simulation-approach-1st-edition-saurabh-
mittal/
textboxfull.com
Understanding Complex Systems

Visarath In
Antonio Palacios

Symmetry in
Complex Network
Systems
Connecting Equivariant Bifurcation
Theory with Engineering Applications
Springer Complexity
Springer Complexity is an interdisciplinary program publishing the best research and
academic-level teaching on both fundamental and applied aspects of complex systems—
cutting across all traditional disciplines of the natural and life sciences, engineering, economics,
medicine, neuroscience, social and computer science.
Complex Systems are systems that comprise many interacting parts with the ability to
generate a new quality of macroscopic collective behavior the manifestations of which are
the spontaneous formation of distinctive temporal, spatial or functional structures. Models
of such systems can be successfully mapped onto quite diverse “real-life” situations like
the climate, the coherent emission of light from lasers, chemical reaction-diffusion systems,
biological cellular networks, the dynamics of stock markets and of the internet, earthquake
statistics and prediction, freeway traffic, the human brain, or the formation of opinions in
social systems, to name just some of the popular applications.
Although their scope and methodologies overlap somewhat, one can distinguish the
following main concepts and tools: self-organization, nonlinear dynamics, synergetics,
turbulence, dynamical systems, catastrophes, instabilities, stochastic processes, chaos, graphs
and networks, cellular automata, adaptive systems, genetic algorithms and computational
intelligence.
The three major book publication platforms of the Springer Complexity program are the
monograph series “Understanding Complex Systems” focusing on the various applications
of complexity, the “Springer Series in Synergetics”, which is devoted to the quantitative
theoretical and methodological foundations, and the “Springer Briefs in Complexity” which
are concise and topical working reports, case studies, surveys, essays and lecture notes of
relevance to the field. In addition to the books in these two core series, the program also
incorporates individual titles ranging from textbooks to major reference works.

Editorial and Programme Advisory Board

Henry Abarbanel, Institute for Nonlinear Science, University of California, San Diego, USA
Dan Braha, New England Complex Systems, Institute and University of Massachusetts, Dartmouth, USA
Péter Érdi, Center for Complex Systems Studies, Kalamazoo College, Kalamazoo, USA and Hungarian Academy
of Sciences, Budapest, Hungary
Karl Friston, Institute of Cognitive Neuroscience, University College London, London, UK
Hermann Haken, Center of Synergetics, University of Stuttgart, Stuttgart, Germany
Viktor Jirsa, Centre National de la Recherche Scientifique (CNRS), Université de la Méditerranée, Marseille,
France
Janusz Kacprzyk, System Research, Polish Academy of Sciences, Warsaw, Poland
Kunihiko Kaneko, Research Center for Complex Systems Biology, The University of Tokyo, Tokyo, Japan
Scott Kelso, Center for Complex Systems and Brain Sciences, Florida Atlantic University, Boca Raton, USA
Markus Kirkilionis, Mathematics Institute and Centre for Complex Systems, University of Warwick,
Coventry, UK
Jürgen Kurths, Potsdam Institute for Climate Impact Research (PIK), Potsdam, Germany
Ronaldo Menezes, Department of Computer Science, Florida Institute of Technology, Melbourne, FL, USA
Andrzej Nowak, Department of Psychology, Warsaw University, Warsaw, Poland
Hassan Qudrat-Ullah, School of Administrative Studies, York University, Toronto, ON, Canada
Linda Reichl, Center for Complex Quantum Systems, University of Texas, Austin, USA
Peter Schuster, Theoretical Chemistry and Structural Biology, University of Vienna, Vienna, Austria
Frank Schweitzer, System Design, ETH Zürich, Zürich, Switzerland
Didier Sornette, Entrepreneurial Risk, ETH Zürich, Zürich, Switzerland
Stefan Thurner, Section for Science of Complex Systems, Medical University of Vienna, Vienna, Austria
Understanding Complex Systems
Founding Editor: S. Kelso

Future scientific and technological developments in many fields will necessarily

depend upon coming to grips with complex systems. Such systems are complex in
both their composition—typically many different kinds of components interacting
simultaneously and nonlinearly with each other and their environments on multiple
levels—and in the rich diversity of behavior of which they are capable.
The Springer Series in Understanding Complex Systems series (UCS) promotes
new strategies and paradigms for understanding and realizing applications of
complex systems research in a wide variety of fields and endeavors. UCS is
explicitly transdisciplinary. It has three main goals: First, to elaborate the concepts,
methods and tools of complex systems at all levels of description and in all scientific
fields, especially newly emerging areas within the life, social, behavioral, economic,
neuro- and cognitive sciences (and derivatives thereof); second, to encourage novel
applications of these ideas in various fields of engineering and computation such as
robotics, nano-technology and informatics; third, to provide a single forum within
which commonalities and differences in the workings of complex systems may be
discerned, hence leading to deeper insight and understanding.
UCS will publish monographs, lecture notes and selected edited contributions
aimed at communicating new findings to a large multidisciplinary audience.

More information about this series at http://www.springer.com/series/5394

Visarath In Antonio Palacios
•

Symmetry in Complex
Network Systems
Connecting Equivariant Bifurcation Theory
with Engineering Applications

123
Visarath In Antonio Palacios
Space and Naval Warfare Systems Center Department of Mathematics, Nonlinear
San Diego, CA Dynamical Systems Group
USA San Diego State University
San Diego, CA
USA

ISSN 1860-0832 ISSN 1860-0840 (electronic)

Understanding Complex Systems
ISBN 978-3-662-55543-9 ISBN 978-3-662-55545-3 (eBook)
DOI 10.1007/978-3-662-55545-3
Library of Congress Control Number: 2017947684

© Springer-Verlag GmbH Germany 2018

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.
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, express 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.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer-Verlag GmbH Germany
The registered company address is: Heidelberger Platz 3, 14197 Berlin, Germany
To our wives, Christine and Irene, and our
sons, Beredei, Reynard and Daniel, and to
our beloved parents, for their love and
support throughout these years.
Preface

The seminal work by Lorenz in 1963 [264], and later by May in 1976 [273–275],
has led scientists and engineers to recognize that nonlinear systems can exhibit a
rich variety of dynamic behavior. From simple systems, such as the evolution of
single species [314], an electronic or biological oscillator [423, 424], to more
complex systems, such as chemical reactions [33], climate patterns [153], bursting
behavior by a single neuron cell [208], and flocking of birds [333, 393], Dynamical
Systems theory provides quantitative and qualitative (geometrical) techniques to
study these and many other complex systems that evolve in space and/or time.
Regardless of the origins of a system, i.e., Biology, Chemistry, Engineering,
Physics, or even the Social Sciences, dynamical systems theory seeks to explain the
most intriguing and fundamental features of spatio-temporal phenomena.
In recent years, systems made up of individual units coupled together, either
weakly or tightly, have gained considerable attention. For instance, the dynamics of
arrays of Josephson junctions [18, 99, 100, 160], central pattern generators in
biological systems [85, 226, 227], coupled laser systems [328, 419], synchroniza-
tion of chaotic oscillators [315, 426], collective behavior of bubbles in fluidization
[163], the flocking of birds [393], and synchronization among interconnected
biological and electronic nonlinear oscillators. These are only a few representative
examples of a new class of complex dynamical systems or complex networks. The
complexity arises from the fact that individual units cannot exhibit on their own the
collective behavior of the entire network. In other words, the collective behavior is
the exclusive result of the mutual interaction that takes place when multiple units
are interconnected in some fashion.
In most cases, three factors are normally considered when studying the collective
behavior of a complex system. Mainly, the internal dynamics of each individual
unit or cell, the topology of cell connections, i.e., which cells communicate with
each other, and the type of coupling. More recently, a fourth factor has gained
further attention—symmetry. It is well-known that symmetry alone can restrict the
type of solutions of systems of ordinary– and partial differential equations, which
often serve as models of complex systems. So it is reasonable to expect that certain
aspects of the collective behavior of a complex system can be inferred from the

vii
viii Preface

presence of symmetry alone. In fact, the work by Golubitsky [145, 146, 149] lays
down the theoretical foundations for a model-independent analysis to understand,
and predict, the behavior of a dynamical system using, mainly, the underlying
symmetries of the system while separating the fine details of the model. While this
approach has been widely successful in explaining computer simulations and
experimental observations of many different spatio-temporal phenomena, it has
found limited use in the conceptualization and development of nonlinear devices
even though many of those systems are, inherently, symmetric. And while many
works have been dedicated to study the symmetry-preserving phenomenon of
synchronization [25, 315, 317, 382, 426], significantly less is known about how one
can exploit the rich variety of collective patterns that can emerge via
symmetry-breaking bifurcations, such as heteroclinic cycles [63].
Over the past 17 years, we and other colleagues and students have been
attempting to bridge the current gap between the theory of symmetry-based
dynamics, equivariant bifurcation theory, and its application to developing non-
linear devices. At the beginning, around the year 2000, we were interested in
developing new methods to manipulate frequency in arrays of nonlinear oscillators
for antenna devices. Collaborators from the U.S. Navy had already shown [166]
that small frequency perturbations applied to the end points of a chain of nonlinear
oscillators can lead to a change in the direction of the radiation pattern. That is, they
demonstrated that beam steering was possible without mechanically rotating an
antenna. The next puzzle that we had to solve was to manipulate the collective
frequency of the array over a broad band without changing the internal frequency of
each individual oscillator. But just when we were about to solve this problem, we
were steered, no pun intended, into developing a new class of highly sensitive,
low-power and low-cost, magnetic- and electric field sensors. Theoretical work for
this new class of sensors started around 2002. The fundamental principles were
twofold: to exploit coupling-induced oscillations in symmetric networks to generate
self-induced oscillations, thus reducing power consumption; and to exploit
symmetry-breaking effects of heteroclinic cycles to enhance sensitivity. As a
starting point, we chose fluxgate magnetometers as individual units, because their
behavior is governed by a one-dimensional autonomous differential equation.
Consequently, based on the fundamental theory of ODEs, it follows that in the
absence of any forcing term the one-dimensional dynamics of the individual units
cannot produce oscillations. But when the fluxgates are coupled then the network
can, under certain conditions that depend on the coupling strength, oscillate. This
configuration could demonstrate to skeptics that self-induced oscillations can
indeed be engineered. In practice, the network would still need, of course, a min-
imum amount of energy to kick it off of its trivial equilibrium state and get the
oscillations going. Overall, we were able to show that, under certain conditions, the
sensitivity response of an array of weakly coupled fluxgate sensors can increase by
four orders of magnitude while their cost could be simultaneously reduced to a
fraction of that of an individual fluxgate sensor. This technology matured around
2005 with design, fabrication and deployment.
Preface ix

In 2006, we extended the work on magnetic fields to electric field sensors. These
sensors are also governed by one-dimensional, overdamped, bistable systems of
equations. We conducted a complete bifurcation analysis that mirrors that of the
fluxgate magnetometer and, eventually, translated the research work into a micro-
circuit implementation. This microcircuit was intended to be used for measuring
minute voltage or current changes that may be injected into the system. The con-
ceptualization of these sensors employs the model-independent approach of
Golubitsky’s theory for the study of dynamical systems with symmetry, while the
development of laboratory prototypes takes into account the model-specific features
of each device which, undoubtedly, may impose additional restrictions when we
attempt to translate the theory into an actual experiment. For instance, a sensor device
that measures magnetic flux, as oppose to electric field signals, may limit the type of
coupling functions that can be realized in hardware. In other words, not every ide-
alization of a network-based structure can be readily implemented in the laboratory.
This and other similar restrictions need to be kept in mind while reading this book.
Around that same 2006 year, we started, in tandem, to the work on electric field
sensors, theoretical studies of networks of Superconducting Quantum Interference
Devices (SQUIDs). The work was suspended until 2009 when we returned to
explore in greater detail the response of networks of non-uniform SQUID loops.
The technology matured by 2012 with applications to antennas and communication
systems. Around that same period, 2007–2009, we went back to the study of
multi-frequency oscillations in arrays of nonlinear oscillators. In fact, we were able
to develop a systematic way to manipulate collective frequency through cascade
networks. The work matured in 2012 with the modeling, analysis, design and
fabrication of the nonlinear channelizer. This is an integrated circuit made up of
large parallel arrays of analog nonlinear oscillators, which, collectively, serve as a
broad-spectrum analyzer with the ability to receive complex signals containing
multiple frequencies and instantaneously lock-on or respond to a received signal in
a few oscillation cycles. Again, the conceptualization of the nonlinear channelizer
was based on the generation of internal oscillations in coupled nonlinear systems
that do not normally oscillate in the absence of coupling. Between 2007–2011, we
investigated various configurations of networks of coupled vibratory gyroscopes.
The investigations showed that networks of vibratory gyroscopes can mitigate the
negative effects of noise on phase drift. But the results were, mainly, computational
and applicable only to small arrays. Finally, between 2012–2015, we developed the
necessary mathematical approach to study networks of arbitrary size. This work
showed the nature of the bifurcations that lead arrays of gyroscopes, connected
bidirectionally, in and out of synchronization. The results were applicable to net-
works of arbitrary size.
In the past few years, previous works have led us into new topics. Networks of
energy harvesters, which, interestingly, are governed by ODEs that resemble those
of vibratory gyroscopes. This feature highlights again the model-independent nature
of the analysis of differential equations with symmetry. In the year 2011, in par-
ticular, we started a new project to study the collective behavior of spin-torque
nano-oscillators. The motivation for this work is a conjecture by the 2007 Nobel
x Preface

Laureate, Prof. Albert Fert, about the possibility that synchronization of

nano-oscillators could produce substantial amounts of microwave power for prac-
tical applications. Determining the regions of parameter space of stable synchro-
nized solutions was a very challenging problem due to the nature (non-polynomial
form) of the governing equations. Finally, this year we overcame the major diffi-
culties by exploiting, again, equivariant bifurcation theory. And the most recent
project that we started in 2016 is about networks of coupled oscillators for
improving precision timing with inexpensive oscillators, as oppose to atomic
clocks.
Along the way, several patents were approved by the U.S. Patent Office for the
works related to these projects, including:
2007 U.S. Patent # 7196590.
Multi-Frequency Synthesis Using Symmetry Methods in Arrays of Coupled
Nonlinear Oscillators.
2008 U.S. Patent # 7420366.
Coupled Nonlinear Sensor System.
2009 U.S. Patent # 7528606.
Coupled Nonlinear Sensor System for Sensing a Time-Dependent Target Signal
and Method of Assembling the System.
2011 U.S. Patent # 7898250.
Coupled Fluxgate Magnetometers for DC and Time-Dependent (AC) Target
Magnetic Field Detection.
2011 U.S. Patent # 8049486.
Coupled Electric Field Sensors for DC Target Electric Field Detection.
2012 U.S. Patent # 8049570.
Coupled Bistable Microcircuit for Ultra-Sensitive Electric and Magnetic Field
Sensing.
2012 U.S. Patent # 8212569.
Coupled Bistable Circuit for Ultra-Sensitive Electric Field Sensing Utilizing
Differential Transistors Pairs.
2015 U.S. Patent # 8994461.
Sensor Signal Processing Using Cascade Coupled Oscillators.
2015 U.S. Patent # 9097751.
Linear Voltage Response of Non-Uniform Arrays of Bi-SQUIDS.
2016 Under review. Navy Case: 101427.
Enhanced Performance in Coupled Gyroscopes and Elimination of Biasing
Signal in a Drive-free Gyroscope.
2016 Under review. Navy Case: 101950.
Arrays of Superconducting Quantum Interference Devices with Self Adjusting
Transfer to Convert Electromagnetic Radiation into a Proportionate Electrical
Signal to Avoid Saturation.
2016 Under review. Navy Case: 102297.
2D Arrays of Diamond Shaped Cells Having Multiple Josephson Junctions.
2016 Under review. Navy Case: 103829.
Network of Coupled Crystal Oscillators for Precision Timing.
Preface xi

None of these projects would have been possible without the active participation
of students, joint work with collaborators, and the financial support from various
sources. We would like to thank each of the students first: John Aven [21, 22],
Jeremmy Banning [26], Katherine Beauvais [30], Susan Berggren [36, 159], Bernard
Chan, Nathan Davies [91], Scott Gassner [132, 133], Mayra Hernandez [168], Habib
Juarez, Tyler Levasseur, Patrick Longhini [261, 262], Daniel Lyons [266, 267],
Antonio Matus [272], Derek Moore, Loni Olender, Steven Reeves [331] Norbert
Renz [332], Richard Shaffer [359], Brian Sturgis-Jensen, James Turtle [398, 399],
Huy Vu [407], Sarah Wang, Bing Zhu [433]. Special acknowledgement and thanks
to Patrick Longhini, he was the first student that got involved in the work through his
Master and, later on, Ph.D. thesis. He continues to be an extremely valuable asset to
multiple ongoing projects. Collaborators include: Bruno Ando (Univ. of Catania),
Marcio De Andrade (SPAWAR), Salvatore Baglio (Univ of Catania, Italy), Peter
Blomgren (SDSU), Donald Bowling (NAWC), Pietro-Luciano Buono (Univ. of
Ontario Institute of Technology, Canada), Adi Bulsara (SPAWAR), Lowell Burnett
(QUASAR), Juan Carlos Chaves (HPTi), Ricardo Carretero (SDSU), Anna Leese de
Escobar (SPAWAR), Jocirei Dias Ferreira (Federal Univ. of Mato Grosso, Brazil),
Hugo Gonzalez-Hernadez (Instituto Tecnologico de Monterrey), Frank Gordon
(SPAWAR), Takachi Hikihara (Kyoto Univ., Japan), Calvin Johnson (SDSU), Andy
Kho (SPAWAR), Daniel Leung (SPAWAR), John F. Lindner (College of Wooster),
Norman Liu (SPAWAR), Joseph M. Mahaffy (SDSU), LT Jerome McConnon
(SPAWAR), Brian K. Meadows (SPAWAR), Oleg Mukhanov (HYPRES), Joseph
Neff (SPAWAR), Suketu Naik (Weber State Univ.), Martin Nisenoff (M. Nisenoff
Associates), Georgy Prokopenko (HYPRES), Wouter-Jan Rappel (UCSD), LT
Sarah Rice (SPAWAR), Robert Romanofsky (NASA), Vincenzo Sacco (Univ.
Catania), Benjamin Taylor (SPAWAR), Edmond Wong (SPAWAR), Yongming
Zhang (QUASAR).
Many thanks to the Chaos Group at the Oak Ridge National Laboratory, Stuart
Daw, Charles Finney, and Sreekanth Pannala, for very stimulating discussions. We
also wish to acknowledge very fruitful interactions with John Angus and Ali Nadim
who served as committee members for many of the Ph.D. theses that derived from
related projects. Special thanks to Pietro-Luciano Buono with whom we have
collaborated extensively in recent years to apply advance methods from equivariant
bifurcation theory. One of the authors, Antonio Palacios, wishes to thank Marty
Golubitsky, in particular, for his mentorship and guidance to learn from him (during
a postdoctoral appointment) the principles and methods for studying dynamical
systems that posses symmetry.
We also wish to acknowledge the financial support provided by several agencies
to conduct the necessary research work that serves as the foundation of some of the
technologies discussed in this book, including: Army Research Office, Department
of Defense, Department of Energy, the National Science Foundation, the National
Security Agency, the Office of Naval Research, the San Diego Foundation, and the
Space and Naval Warfare Center, San Diego. We wish to acknowledge the con-
tinuous support of Dr. Michael Shlesinger from ONR.
xii Preface

The book is intended for a broad audience. For engineers who might be inter-
ested in applying ideas and methods from dynamical systems with symmetry and
equivariant bifurcation theory to design and fabricate novel devices. For mathe-
maticians and physicists who might be interested in translational research work to
extrapolate fundamental research theorems into practical applications. And for
scientists from many disciplines, viz. Biology, Chemistry, Computer Science,
Geology, etc., who might be interested in the interplay between theory and real-life
applications from the general field of nonlinear science.
The book is organized as follows. In Chap. 1 we present fundamental ideas of
complex networks and bistability, which is a common feature of many sensor
devices; and then we dedicate a few sections to introduce basic ideas, methods and
examples in the analysis of differential equations (ODEs and PDEs) with symmetry.
One particular class of solutions that rarely appears in generic versions of systems
of differential equation are heteroclinic cycles. These types of solutions are, how-
ever, generic features of systems with symmetry. We exploit these cycles to
enhance sensitivity and, thus, we dedicate a section to explain what they are and
how they can be found. The book is then organized in two parts. Part I, Chap. 2
through Chap. 6 is dedicated to translational research work that already led to
mature technologies. These technologies include networks of fluxgate magne-
tometers; arrays of micro-electronic electric field sensors; networks of SQUIDs;
cascade arrays of nonlinear oscillators for multi-frequency generators; and a special
chapter in honor of the theoretical work by Pietro-Luciano and Marty Golubitsky: a
device realization of a Central Pattern Generator network of the animal gaits studied
by them. Part II, Chap. 7 through Chap. 10 include, mainly, theoretical works that
have not yet mature into actual device realizations. The technologies that may
derive from these works are part of ongoing efforts.

San Diego, USA Visarath In

2017 Antonio Palacios
Contents

1 A Unifying Theme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Complex Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Bistability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Self-oscillating Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4 The Role of Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5 Symmetry-Breaking Bifurcations . . . . . . . . . . . . . . . . . . . . . . . . 13
1.6 Coupled Cell Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.7 Heteroclinic Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.8 Representative Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2 Coupled-Core Fluxgate Magnetometer . . . . . . . . . . . . . . . . . . . . . . . 37
2.1 Fluxgate Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.2 Modeling Single-Core Dynamics . . . . . . . . . . . . . . . . . . . . . . . . 41
2.3 Coupled Single-Domain System . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.4 Frequency Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.5 Sensitivity Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.6 Alternating Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.7 AC Field Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
2.8 Target Signal Contamination . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
2.9 Effects of Nonhomogeinities . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
2.10 Effects of Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
2.11 Laboratory Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
3 Microelectric Field Sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
3.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
3.2 Circuit Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
3.3 Analysis and Bifurcation Diagrams. . . . . . . . . . . . . . . . . . . . . . . 110
3.4 Numerical and Experimental Results . . . . . . . . . . . . . . . . . . . . . 111
3.5 Period and Residence Times Response . . . . . . . . . . . . . . . . . . . . 118
3.6 SPICE Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

xiii
xiv Contents

4 Superconductive Quantum Interference Devices (SQUID) . . . . . . . . 127

4.1 History of Superconductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
4.2 The Josephson Effect and SQUID Technology . . . . . . . . . . . . . . 131
4.3 Phase-Space Dynamics of DC SQUID . . . . . . . . . . . . . . . . . . . . 132
4.4 Chimera States in Non-locally Coupled Arrays. . . . . . . . . . . . . . 138
4.5 The DC Bi-SQUID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
4.6 Serial Bi-SQUID Array. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
4.7 Parallel Bi-SQUID Array . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
4.8 Design, Fabrication, and Evaluation . . . . . . . . . . . . . . . . . . . . . . 154
5 Frequency Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
5.1 Frequency Up-Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
5.2 Experiments on Frequency Up-Conversion . . . . . . . . . . . . . . . . . 176
5.3 Frequency Down-Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
5.4 Experiments on Frequency Down-Conversion . . . . . . . . . . . . . . 191
5.5 Large Frequency Downconversion Ratios . . . . . . . . . . . . . . . . . . 199
5.6 Nonlinear Channelizer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
5.7 Experimental Setup of Nonlinear Channelizer . . . . . . . . . . . . . . 218
6 ANIBOT: Biologically-Inspired Animal Robot . . . . . . . . . . . . . . . . . 229
6.1 Central Pattern Generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
6.2 CPG Network Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
6.3 Analog Fitzhugh–Nagumo Neuron Circuit . . . . . . . . . . . . . . . . . 232
6.4 Patterns and Locomotion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
6.5 Leg Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
7 Gyroscope Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
7.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
7.2 History of Navigation Systems . . . . . . . . . . . . . . . . . . . . . . . . . . 238
7.3 Evolution of Gyroscopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
7.4 Vibratory Gyroscopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
7.5 Bi-Directionally Coupled Ring . . . . . . . . . . . . . . . . . . . . . . . . . . 249
7.6 Unidirectionally Coupled Ring . . . . . . . . . . . . . . . . . . . . . . . . . . 264
7.7 Drive-Free Gyroscope System . . . . . . . . . . . . . . . . . . . . . . . . . . 277
7.8 Hamiltonian Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
8 Energy Harvesting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
8.1 State of the Art . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
8.2 Single Magnetostrictive Energy Harvesting Model . . . . . . . . . . . 299
8.3 Coupled Energy Harvester System . . . . . . . . . . . . . . . . . . . . . . . 302
8.4 Computational Bifurcation Results . . . . . . . . . . . . . . . . . . . . . . . 305
8.5 Hamiltonian Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
8.6 Experimental Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
Contents xv

9 Spin Torque Nano Oscillators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317

9.1 The Giant Magnetoresistance (GMR) Effect . . . . . . . . . . . . . . . . 317
9.2 Spin Torque Nano Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
9.3 Landau-Lifshitz-Gilbert-Slonczewski (LLGS) Model . . . . . . . . . 321
9.4 The Synchronization Challenge . . . . . . . . . . . . . . . . . . . . . . . . . 323
9.5 Series Array . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
9.6 Complex Stereographic Projection . . . . . . . . . . . . . . . . . . . . . . . 328
9.7 Hopf Bifurcation Curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
9.8 Nonlinear Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
9.9 Locking into Synchronization . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
10 Precision Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
10.1 History of Precision Timing Devices . . . . . . . . . . . . . . . . . . . . . 339
10.2 Crystal Oscillators. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
10.3 Two-Mode Oscillator Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
10.4 Coupled Crystal Oscillator System . . . . . . . . . . . . . . . . . . . . . . . 348
10.5 Averaging and Symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
10.6 Linearization and Spectrum at the Origin . . . . . . . . . . . . . . . . . . 357
10.7 Stability and Bifurcation Results . . . . . . . . . . . . . . . . . . . . . . . . . 360
10.8 Numerical Continuation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366
10.9 Phase Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
10.10 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
Chapter 1
A Unifying Theme

Nonlinear systems can behave very robustly, yet they can also exhibit high sensitivity
to small perturbations [205, 206, 277, 404, 420]. This dichotomy is a consequence of
the parameters that govern the behavior of a given system. In regions where changes in
parameters do not yield radically different behavior the system is expected to behave
robustly. However, in regions that contain a bifurcation point, high-sensitivity can
ensue. Close to the onset of oscillations, in particular, a small perturbation can drasti-
cally alter the behavior of a nonlinear oscillator, it can either destroy the oscillations
completely or it can change their characteristics such as frequency and amplitude. For
instance, temperature variations can reset a circadian oscillator, see Fig. 1.1 while
noise [277] can induce an exogenous oscillator to oscillate in regions where only
steady-state behavior would dominate the system dynamics absent the noise. Similar
sensitivity features can be used for signal amplification in electrical, mechanical and
optical systems [213]. Exploring those features is the subject of this chapter and of
the entire book.

1.1 Complex Networks

Over the past seventeen years, we and other colleagues and students have been
developing, through theory and experiments, a new paradigm that combines ideas
and methods from the theory of Nonlinear Dynamical Systems with Symmetry in
Mathematics, Physics, and Engineering, for performance enhancement of nonlinear
devices. A unifying theme and goal of this paradigm is to demonstrate, theoretically
and experimentally, that collective behavior, which is uniquely produced by inter-
connected nonlinear devices, can be exploited to develop novel complex network
based systems that can outperform the function of their individual counterparts.
After all, many sensory systems [225] in animals are controlled by clusters of neu-
ron cells, located in the central nervous system and physically interconnected to
© Springer-Verlag GmbH Germany 2018 1
V. In and A. Palacios, Symmetry in Complex Network Systems,
Understanding Complex Systems, DOI 10.1007/978-3-662-55545-3_1
2 1 A Unifying Theme

Fig. 1.1 Endogenous circadian oscillators of Drosophila and Mammals. Source Journal of Cell
Science 119, 4793–4795 (2006)

produce, somehow, an optimal response, which in many cases outperforms those

of humans. A dog’s sense of smell is said to be a thousand times more sensitive
than that of humans, thanks to the collective behavior of more than 220 olfactory
receptor cells in its nose, see Fig. 1.2(left), while the temporal resolution of the com-
pound eye (Drosophila melanogaster) of a fruitfly eye is approximately ten times
better than that of humans due to the collective output of 6400 photoreceptor cells,
tightly bound to one another, see Fig. 1.2(right). These biologically-inspired ideas
are model-independent, so we hope they can stimulate further development of other
sensory systems.
The collaborative work has lead to a broad class of nonlinear devices been fabri-
cated, including: highly-sensitive, low power, miniaturized magnetic- and electric-
field sensors; networks of SQUID (Superconducting Quantum Interference Devices)
for the development of a sensitive, low noise, significantly lower Size, Weight and
Power (SWaP) antenna integrated with Low-Noise Amplifier (LNA); arrays of gyro-
scopes with reduced phase drift for improved navigation; frequency up/down convert-
ers; a broad-spectrum analyzer with the ability to receive complex signals containing
multiple frequencies and instantaneously lock-on or respond to a received signal in a
few oscillation cycles; networks of energy harvesters with increased power output; an
inexpensive precision timing device through coupled nonlinear oscillators circuits;
and networks of spin-torque nano-oscillators for generating microwave signals at the
nano scale. Figure 1.3 shows a composite snapshot of the multiple projects outlined
above. To provide readers with a description of the translational work, involving, con-
1.1 Complex Networks 3

Fig. 1.2 Biological sensors made up of coupled cell systems. (Left) A dog’s smell is thousand
times more sensitive than that of humans thanks to the collective behavior of more than 220 olfac-
tory receptor cells. (Picture courtesy of http://www.faithfool.wordpress.com). (Right) The temporal
resolution of a firefly is ten times better than in humans due to the collective output of approxi-
mately 6400 photoreceptor cells tightly coupled to one another. Picture courtesy of “Research Penn
State” [306]

Fig. 1.3 Representative projects with a unifying theme: to exploit the collective behavior of complex
network systems to model, analyze and design advanced engineering systems. From top-right, in
counter-clockwise direction, coupled nonlinear oscillators for antennas; highly-sensitive coupled-
core fluxgate magnetometer; networks of vibratory gyroscopes with optimal phase drift; thousands
of superconductive loops for communication systems; arrays of spin-torque nano oscillators for
microwave signal generation; and networks of energy harvesting systems
4 1 A Unifying Theme

ceptualization, derivation of mathematical models, analysis, design and fabrication

of these technologies, is the main goal of this book.
The technologies that are derived from the proposed network-based configuration
can be effectively described as hybrid, discrete-continuous, nonlinear dynamical
system: a discrete number of individual components, in which each unit is governed
by a continuous model of nonlinear differential equations, physically coupled to form
the network structure. These networks can also be described as complex networks in
the sense that their response cannot be achieved by individual units, i.e., only by the
mutual interaction of the individual components.
Although a direct relation to biological systems has yet to be demonstrated in more
detail, we show proof of concept that a network-based approach can indeed lead to
significant performance enhancements for certain classes of sensors. In particular:
fluxgate magnetometers, electric-field sensors, and SQUID sensors. But the emphasis
of the proposed approach on magnetic and electric-field sensors is not exhaustive. The
basic principles of cooperative behavior can be applied to enhance the performance of
a wide range of nonlinear systems, such as antennas and radar systems; gyroscopes,
energy harvesters; precision timing devices and microwave signal generators. By
cooperative we mean patterns of collective behavior. For instance, we show that
complete synchronization, in which all units oscillate in-phase with the same wave
form and amplitude, can lead to significant reduction in the phase drift that affects
the performance of gyroscopes. On the other hand, heteroclinic cycles are found
to be ideal patterns to increase sensitivity in networks of fluxgate magnetometers.
Similarly, traveling wave patterns can be exploited to improve the performance of
precision timing devices. Which pattern of behavior is better suited for a particular
application is a model dependent issue that has to be studied on a case-by-case
basis. However, the existence of these, and many other related patterns of collective
behavior are governed, mainly, by a common feature, symmetry.
In all the systems that are described in this book, our objectives include to inves-
tigate the following model-independent issues:
(i) To investigate, analytically, computationally, and experimentally, the interplay
between the discrete and continuous characteristics of the network dynamics.
(ii) To determine the fundamental limit of performance enhancement, e.g., reduc-
tion in phase drift or increase in sensitivity, that can be achieved with a network-
based device.
(iii) To develop design rules for nonlinear devices.
To accomplish these goals, we draw on methods from dynamical systems and
bifurcation theory in systems with symmetry to address some of the following, model-
dependent, fundamental issues:
• What is the optimal network architecture that can produce the highest-level of
performance enhancements at the lowest cost and with the lowest power con-
sumption?
• How can a network be built in hardware? What are the advantages and disad-
vantages of different hardware implementations, and how do they compare to
simulations in software?
1.1 Complex Networks 5

• How can a network be programmed and controlled to operate in realistic (noisy)

environments?
Since the majority of the technologies that are discussed in this book employ
nonlinear oscillator components, we present next a brief introduction of basic ideas,
concepts and principles.

1.2 Bistability

Many complex systems, natural and artificial ones, exhibit oscillatory behavior, i.e.,
cyclic behavior that repeats at regular intervals. Examples include: the rhythmic light
pulses of fireflies [53, 111], see Fig. 1.4, the electrical activity of neuron cells that
make up central pattern generators in biological systems [63, 83, 85, 148, 226,
227, 414], the patterns of lights produced by arrays of coupled lasers [328, 419],
voltage variations in modern communication systems [315, 426], the growth and
decay of population sizes between competing species [273–275], bubble formation
and evolution in fluidization and mixing processes [163], and variations in phase and
current in arrays of Josephson junctions [18, 99, 100, 160] in quantum physics.
In the absence of noise, the underlying cyclic fluctuations in a given system can
arise from individual units that oscillate on their own, also known as endogenous
or self-excited oscillators, or from exogenous units that oscillate only when they are
externally driven or coupled together. Circadian rhythms, which regulate the daily
cycle of many living organisms, plants, and animals, for instance, are endogenously

Fig. 1.4 Complex interactions among fireflies can lead them to coordinate the rhythmic flashing
lights produced by each individual firefly. Collectively, the swarm can then achieve synchronization
and oscillate in unison. Source National Geographic
6 1 A Unifying Theme

generated. In fact, the first endogenous circadian oscillation to be observed was the
movement of the leaves of Mimosa pudica, a plant studied by the French scientist
Jean-Jacques dÒetous de Mairan.1
In addition, bistability—the property that allows a system to rest in either of two
states—underlies the basic oscillatory behavior of many other natural and artificial
systems. States may include typical invariant sets, such as equilibrium points, periodic
and quasi-periodic solutions, and chaotic attractors. In the absence of an external
stimulus, the state variable x(t) of a bistable system will relax to one of the invariant
sets, and it will remain in that state unless it is switched or forced to another state.
It is in this sense that the system exhibits “memory.” Which invariant set the system
will relax to depends typically on the set of initial conditions. All bistable systems
employ some form of energy source as the underlying principle that allows them to
switch between states. The source of energy is due typically through external forcing
or through the coupling mechanism.
For instance, dynamic sensors [150, 210, 252, 326, 334], operate as exoge-
nous oscillators with nonlinear input-output characteristics, often corresponding to
a bistable potential energy function of the form

dx
= −∇U (x), (1.1)
dt

where x(t) is the state variable of the natural system or artificial device, e.g., mag-
netization state, and U is the bistable potential function. Examples include: fluxgate
magnetometers [45, 139], ferroelectric sensors [27], and mechanical sensors, e.g.,
acoustic transducers made with piezoelectric materials. Fig. 1.5(top) illustrates the
case of a double-well potential function U (x) = −ax 2 + bx 4 , whose minima are
located at ±xm and the height of the potential barrier between the two minima is
labeled by U0 .
Without an external excitation (periodic forcing or noise), the state point x(t) of
the exogenous oscillator described by Eq. (1.1) will rapidly relax to one of two stable
attractors, which correspond to the minima of the potential energy function U (x).
In the presence of an external periodic forcing term f (t), with frequency ω, the
state variable in U (x + f (t)) can be induced to oscillate periodically (with a well-
defined waveform) between its two stable attractors −xm and +xm , as is illustrated in
Fig. 1.5(bottom). The forcing term is also known as biasing signal in the engineering
literature.
Standard Detection Mechanism. To detect a small target signal (dc or low-
frequency), typically the standard, spectral-based [323, 335, 336, 338], readout
mechanism is employed. Assume ε to be the target signal. When ε = 0, the power
spectral density contains only the odd harmonics of the bias frequency ω. But when
ε > 0, the potential energy function U (x + f (t) + ε) is skewed, resulting in the
appearance of even harmonics; the response at the second harmonic 2ω is then used

1 Source: Wikipedia https://en.wikipedia.org.

1.2 Bistability 7

Fig. 1.5 (Top) Bistable

Potential
U (x) = −ax 2 + bx 4 .
(Bottom) Switching between
wells of a potential function
can be achieved by a
sufficiently large biasing

U( x, t )
signals (or noise) greater
than the potential barrier

U
0

−x x
m m

to detect and quantify the target signal, as is shown in Fig. 1.6. The standard readout
mechanisms has some drawbacks, however. Chief among them is the requirement of
a large onboard power to provide a high-amplitude, high-frequency, bias signal. The
feedback electronics can also introduce their own noise floor into the measurement
process, and finally, a high-amplitude, high-frequency, bias signals often increase
the noise floor of the system.
In biological systems, bistability is a key feature for understanding and engineer-
ing cellular functions such as: storing and processing information by the human brain
during the decision-making process [6]; regulation of the cell cycle [368, 400]; sporu-
lation, which controls the timing and dynamics of dramatic responses to stress [406];
design and construction of synthetic toggle switches [131]; and in gene regulatory
networks responsible for embryonic stem cell fate decisions [78]. In chemical sys-
8 1 A Unifying Theme

0 0
10 10

−2 −2
10 10
PSD (f)

PSD (f)
−4 −4
10 10

−6 −6
10 10

−8 −8
10 10
0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1
f f

Fig. 1.6 Power Spectrum Decomposition of the oscillations in a bistable overdamped system
subject to a periodic forcing. (Left) In the absence of an external signal the PSD shows only even
harmonics. (Right) In the presence of an external signal, however, the PSD exhibits odd as well as
even harmonics. Typically, the first even harmonic is used as a detection mechanism

tems bistability is central to the analysis of relaxation kinetics [422]. In mechanical

systems, bistable mechanisms are commonly employed in the design and fabrications
of Micro-Electro-Mechanical-Systems (MEMS) versions of relays, valves, clips, and
threshold switches [175, 325]. In electronics, hysteresis and bistability are combined
to design and fabricate Schmitt trigger circuits, which convert analog input signals
to digital output signals [353]. In neuroscience, bistability is at the core of Hopf
models [86, 178, 180], which describe the input/output response of neurons through
differential equations of the form

d Vi
τi = −Vi + g(Vi ), (1.2)
dt
where τi is a suitable time constant that controls how quickly unit(neuron) i responds
to a stimulus, Vi is the output (typically voltage) of unit i, and g is the activation
function, which normally represents a saturation nonlinearity property of neurons.
Alternatively, the phenomenon of Stochastic Resonance [127, 128, 171, 204, 206,
277] can also serve as a mechanism for generating oscillations, see Fig. 1.7. Briefly
speaking, stochastic resonance refers to noise-induced transitions which, due to the
nonlinear nature of a system dynamics, become synchronized with the period of an
external force. A distinctive feature of stochastic resonance is a rapid increase in
a system’s output Signal-to-Noise Ratio (SNR) under weak coupling followed by
a slower decrease in SNR for stronger coupling. At intermediate noise intensities,
the system exhibits maximum SNR. This feature is inherently due to the nonlinear
nature of a system’s dynamics and cannot be reproduced by linear systems.
1.3 Self-oscillating Networks 9

Fig. 1.7 (Left) Motion of a particle on a double-well potential function. In the absence of noise,
the system dynamics quickly settles into an equilibrium point. Which equilibrium point is selected
depends on the initial conditions. (Right) In the presence of noise, the system dynamics now lingers
intermittently between the two equilibrium states of the deterministic system, independently of
initial conditions

1.3 Self-oscillating Networks

A fundamental idea in the new paradigm that we have developed for performance
enhancement in nonlinear systems is simple: make exogenous oscillators behave as
endogenous ones. That is, we seek to minimize power consumption, and simulta-
neously enhance output response, by replacing force-driven exogenous oscillators
with self-oscillating networks. The paradox can be resolved through symmetry. This
point deserves a little more explanation.
Many systems are known to oscillate only when they are driven by an external
force. However, when they are connected in some fashion, the symmetry of the
resulting topology of connections, i.e., which units are coupled with each other,
and the nonlinear characteristics of each individual unit, can be exploited to induce
the interconnected network to generate a collective pattern of oscillation via an
appropriate coupling function, see Fig. 1.8.
From a mathematical standpoint, the choice of coupling function can be any type
of function, leading to a wide range of network solutions. From an engineering
standpoint, the coupling function is restricted, however, by the type of system or
technology being used. This is, in other words, a model-dependent feature of the
system. For instance, fluxgate magnetometers coupled through magnetic flux are
restricted to unidirectional coupling since directing magnetic flux in both directions
can be extremely complicated to achieve. Mechanical gyroscopes can be coupled,
however, bidirectionally through a series of mass-spring systems.
An immediate advantage of the self-oscillating network approach is the cost reduc-
tion that can be achieved by eliminating power-driven components which are typi-
cally very expensive. A more subtle advantage is to exploit the collective pattern of
oscillation to enhance output response. Consider for instance a sensor device (mag-
netic or electric) whose behavior (as is described earlier) is governed by a potential
10 1 A Unifying Theme

Sum Signal: c = [3.0, 3.5, 2.5]

1

0.5

Amplitude
1 0

−0.5

−1
0 200 400 600 800
5 2 5
Time
10

PSD
0
10

4 3 10
−5

0 0.2 0.4 0.6 0.8 1

Frequency

Fig. 1.8 (Left) Representative example of self-oscillating networks. Individually, each unit cannot
oscillate. (Right) Under certain coupling topologies, and the internal dynamics of each unit, the
network can, collectively, oscillate

function of the form (1.1). A self-oscillating network would eliminate the need of
the large onboard power that is required for a biasing signal to overcome the energy-
barrier of the system and thus induce the required oscillations. Furthermore, the
network oscillations can yield (under certain conditions) wave forms whose char-
acteristics (frequency, amplitude, and phase) can exhibit higher sensitivity to very
small target signals.
The symmetry-based approach to study nonlinear systems, while nonstandard, is
not entirely new among the mathematics community. However, there is much less
familiarity with the techniques of symmetry-breaking bifurcation, and applications
of equivariant bifurcation theory, developed by Golubitsky and Stewart [146, 147,
149], as they apply to the engineering, design and fabrication, of complex systems.
For this reason, we dedicate the next few sections to introduce, first, the mathemati-
cal formalism to describe symmetry-breaking bifurcations; networks architecture of
interconnected devices, through the coupled cell formalism and, then, a more com-
prehensive discussion of the existence of heteroclitic cycles in nonlinear systems.

1.4 The Role of Symmetry

A general approach for the analysis of complex network systems has been to derive
a detailed model of its individual parts, connect the parts and note that the system
contains some sort of symmetry, then attempt to exploit this symmetry in order
to simplify numerical computations. This approach can result in very complicated
models that are difficult to analyze even numerically.
Our approach in this book is, however, aimed at promoting a unifying theme
for the use of symmetry in a systematic way. First, determine the conditions for
1.4 The Role of Symmetry 11

the existence and stability of collective patterns of behavior of a complex system.

Then perform translational work to transfer theorems and principles into prototypes
devices that conform, as close as possible, to a particular model. For instance, the
phenomenon of coupling-induced oscillations that regulates the dynamics of the
newly postulated coupling-based sensor devices is dictated, mainly, by the group Z N ,
of cyclic permutations of N objects, i.e., units connected with a preferred direction
or directionally. It is in this sense that the ideas and methods of our approach are
device-independent so that similar principles can be readily applied to understand,
and hopefully, to enhance the performance of a wide variety of sensor devices so
long as the symmetry conditions are satisfied.
A few more details of how symmetry can appear in governing equations, such as
in systems of differential equations, are now in order.
Definition 1 Symmetry is a geometrical concept that describes the set of transfor-
mations that leave an object unchanged.
In complex systems with continuous nonlinear behavior, the objects are the gov-
erning equations, which typically consist of systems of ordinary differential equations
(ODEs) or partial differential equations, and the transformations are the changes in
the underlying variables that leave the equations unchanged. More formally, consider
the following system of ODEs

dx
= f (x, λ), (1.3)
dt

where x ∈ Rn , λ ∈ R p is a vector of parameters and f : Rn × R p → R is a smooth

function. Let γ be a particular time-independent transformation in Rn . Direct substi-
tution of the transformed variable γx into (1.3) yields γ ẋ = f (γx, λ). Consequently,
for the system of ODEs (1.3) to remain unchanged, f must commute with γ. In other
words, f (γx) = γ f (x). In practice, the set of all transformations that commute with
f forms a group Γ . We then arrive to the following formal definition.
Definition 2 A system of ODEs such as (1.3) is said to have Γ -symmetry if

f (γx, λ) = γ f (x, λ), (1.4)

for all x ∈ Rn and for all γ ∈ Γ .

Equation (1.4) also implies that f is Γ -equivariant. But more importantly, it
implies that if x is a solution of (1.3) then so is γx(t) for all γ ∈ Γ . In fact, the
collection of points γx(t), for all γ ∈ Γ forms a set called the group orbit of Γ :

Γ x = {γ x : γ ∈ Γ }.

Furthermore, the concept of group orbit applies to any point x(t), not just equi-
librium solutions.
12 1 A Unifying Theme

Fig. 1.9 Circuit realization

of a Van der Pol oscillator.
The governing equations
exhibit reflectional
symmetry with respect to the
state variables

Example 1 As an example, consider the van der Pol circuit depicted in Fig. 1.9. I L
and IC are the currents across the inductor L and capacitor C, respectively. I R is
the current across two resistors R1 and R2 located inside the rectangle labeled R in
which F(V ) = −V /R1 + V 3 /(3R22 ).
The dynamics of the circuit shown in Fig. 1.9, after rescaling, is governed by the
following second order scalar ODE

d2V dV
2
− δ( p − V 2 ) − ω 2 V = 0, (1.5)
dt dt
√
where δ = 1/(R2 C), p = R2 /R1 , ω = 1/ LC. After a change of variables, we can
rewrite the model equation (1.5) as a first order system of the form
 
dx x3
= δ px − + ωy
dt 3 (1.6)
dy
= −ωx,
dt

where x(t) = V (t). We can then find two transformations that leave this system
unchanged: the identity transformation γ1 = id, where γ1 (x, y) → (x, y), and a
second transformation, which can be described abstractly as γ2 = −1, so that
γ2 (x, y) → (−x, −y). The identity transformation is always a symmetry of any
system, while the second transformation γ2 corresponds to a reflection through the
origin in the phase space R2 . Furthermore, it can be shown that γ1 and γ2 are the
only transformations that leave (1.6) unchanged. Together, γ1 and γ2 form the group
Z2 = {γ1 , γ2 } of symmetries of the Van der Pol oscillator (1.6).

But individual solutions can also exhibit symmetry. For instance, consider equi-
librium solutions. A solution xe of Eq. (1.3) is an equilibrium or steady-state if and
only if
f (xe ) = 0.

The symmetries of equilibrium points of a Γ -equivariant ODE form a subgroup

of Γ , which we define next.
1.4 The Role of Symmetry 13

Definition 3 Let xe represent an equilibrium or steady-state solution of a Γ -

equivariant system of ODEs. The symmetries of xe form the isotropy subgroup Σ of
Γ , which is defined by

Σxe = {γ ∈ Γ : γ · xe = xe } . (1.7)

1.5 Symmetry-Breaking Bifurcations

Equivariant systems of ODEs always posses a trivial solution x0 whose isotropy

subgroup is the entire group of symmetries of the model equations [146, 147, 149].
That is, Σx0 = Γ . But as parameters are varied the system can exhibit a new solution
with less symmetry. That is, Σx ⊂ Γ . It is then said that the system has undergone a
symmetry-breaking bifurcation. We consider in this chapter two types of symmetry-
breaking bifurcations, steady-state and Hopf bifurcations. The former case leads to
new equilibrium solutions while the latter to periodic oscillations. We describe next
each of these two cases.
Steady-State Bifurcations. The iconic picture in Fig. 1.10 of the milk drop coronet
by the pioneering work on speed photography by Harold E. Edgerton illustrates
best the phenomenon of steady-state symmetry-breaking bifurcation. The pool of
milk in its unperturbed or trivial state is symmetric under arbitrary rotations and
reflections on a plane, which form the orthogonal group O(2). The perturbation
by the droplet breaks, however, the O(2) symmetry of the trivial solution and it
induces a crown-like shape with lesser symmetry. The 24-sided polygon that appears
by joining the individual clumps now has D24 -symmetry, where D N is the dihedral
group of symmetries of an N -gon.
Example 2 Consider now Euler’s column buckling experiment, illustrated in
Fig. 1.11. It consists of an elastic beam subjected by a compressive force. Upon
a critical value of the compressive force the beam deforms into one of two buck-
led states, either to the right or to the left. Which state actually appears depends
on model-dependent features such as material imperfections or thermal fluctuations.
According to Bernoulli–Euler beam theory, a mathematical model for the angle θ(t)
between the undeformed road and the tangent of the deformed rod is

E I θ (x) + P sin θ(x) = 0, (1.8)

where x is the material coordinate, E is the elastic modulus, I is moment of inertia,

P is the compressive force and L is the length of the beam. Equation (1.8) pos-
sesses reflectional symmetry θ → −θ, which is described by the group Γ = Z2 .
The unperturbed unbuckled state is the trivial solution with Γ -symmetry. Assuming
boundary conditions θ(0) = θ(L) = 0, two nontrivial solutions are ± sin(πx/L).
The isotropy subgroup of these solutions is the trivial group 1. It can be shown that
these two solutions emerge through a pitchfork bifurcation off of the trivial solution,
14 1 A Unifying Theme

Fig. 1.10 This iconic

picture of the milk drop
illustrates the phenomenon
of symmetry-breaking
bifurcations. An unperturbed
pool of milk is invariant
under arbitrary rotations and
reflections on a plane, which
form the orthogonal group
O(2). The crown-like shape
that emerges under the
perturbation by the droplet is
a 24-sided polygon whose
symmetries are described by
the dihedral group D24 .
Source Harold E. Edgerton,
Milk Drop Coronet, 1957.
2010 Massachusetts Institute
of Technology

Fig. 1.11 Euler beam

experiment. An elastic beam
is subjected to a compressive
force. Upon reaching a
certain threshold value of the
compressive force, the trivial
solution, unbuckled state,
losses stability and a buckled
state, right or left, emerges
through a pitchfork
bifurcation. Source
Wikipedia
Exploring the Variety of Random
Documents with Different Content
 COCOA-NUT SOUP.

Pare the dark rind from a very fresh cocoa-nut, and grate it down
small on an exceedingly clean, bright grater; weigh it, and allow two
ounces for each quart of soup. Simmer it gently for one hour in the
stock, which should then be strained closely from it, and thickened
for table.
Veal stock, gravy-soup, or broth, 5 pints; grated cocoa-nut, 5 oz., 1
hour. Flour of rice, 5 oz.; mace, 1/2 teaspoonful; little cayenne and
salt; mixed with 1/4 pint of cream: 10 minutes.
Or: gravy-soup, or good beef broth, 5 pints: 1 hour. Rice flour, 5
oz.; soy and lemon-juice, each 1 tablespoonful; finely pounded
sugar, 1 oz.; cayenne, 1/4 teaspoonful; sherry, 2 glassesful.
Obs.—When either cream or wine is objected to for these soups, a
half-pint of the stock should be reserved to mix the thickening with.
 CHESTNUT SOUP.

Strip the outer rind from some fine, sound Spanish chestnuts,
throw them into a large pan of warm water, and as soon as it
becomes too hot for the fingers to remain in it, take it from the fire, lift
out the chestnuts, peel them quickly, and throw them into cold water
as they are done; wipe, and weigh them; take three quarters of a
pound for each quart of soup, cover them with good stock, and stew
them gently for upwards of three quarters of an hour, or until they
break when touched with a fork; drain, and pound them smoothly, or
bruise them to a mash with a strong spoon, and rub them through a
fine sieve reversed; mix with them by slow degrees the proper
quantity of stock; add sufficient mace, cayenne, and salt to season
the soup, and stir it often until it boils. Three quarters of a pint of rich
cream, or even less, will greatly improve it. The stock in which the
chestnuts are boiled can be used for the soup when its sweetness is
not objected to; or it may in part be added to it.
Chestnuts, 1-1/2 lb.: stewed from 2/3 to 1 hour. Soup, 2 quarts;
seasoning of salt, mace, and cayenne: 1 to 3 minutes. Cream, 3/4
pint (when used).
 JERUSALEM ARTICHOKE, OR PALESTINE SOUP.

Wash and pare quickly some freshly-dug artichokes, and to

preserve their colour, throw them into spring water as they are done,
but do not let them remain in it after all are ready. Boil three pounds
of them in water for ten minutes; lift them out, and slice them into
three pints of boiling stock; when they have stewed gently in this
from fifteen to twenty minutes, press them with the soup, through a
fine sieve, and put the whole into a clean saucepan with a pint and a
half more of stock; add sufficient salt and cayenne to season it, skim
it well, and after it has simmered for two or three minutes, stir it to a
pint of rich boiling cream. Serve it immediately.
Artichokes, 3 lbs., boiled in water: 10 minutes. Veal stock, 3 pints
15 to 20 minutes. Additional stock, 1-1/2 pint; little cayenne and salt
2 to 3 minutes. Boiling cream, 1 pint.
Obs.—The palest veal stock, as for white soup, should be used for
this; but for a family dinner, or where economy is a consideration
excellent mutton-broth, made the day before and perfectly cleared
from fat, will answer very well as a substitute; milk too may in part
take the place of cream when this last is scarce: the proportion of
artichokes should then be increased a little.
Vegetable-marrow, when young, makes a superior soup even to
this, which is an excellent one. It should be well pared, trimmed, and
sliced into a small quantity of boiling veal stock or broth, and when
perfectly tender, pressed through a fine sieve, and mixed with more
stock and some cream. In France the marrow is stewed, first in
butter, with a large mild onion or two also sliced; and afterwards in a
quart or more of water, which is poured gradually to it; it is next
passed through a tammy,[26] seasoned with pepper and salt, and
mixed with a pint or two of milk and a little cream.
26. Derived from the French tamis, which means a sieve or strainer.
 COMMON CARROT SOUP.

The most easy method of making this favourite English soup is to

boil some highly coloured carrots quite tender in water slightly salted,
then to pound or mash them to a smooth paste, and to mix with them
boiling gravy soup or strong beef broth (see Bouillon) in the
proportion of two quarts to a pound and a half of the prepared
carrots; then to pass the whole through a strainer, to season it with
salt and cayenne, to heat it in a clean stewpan, and to serve it
immediately. If only the red outsides of the carrots be used, the
colour of the soup will be very bright; they should be weighed after
they are mashed. Turnip soup may be prepared in the same manner.
Obs.—An experienced and observant cook will know the
proportion of vegetables required to thicken this soup appropriately,
without having recourse to weights and measures; but the learner
had always better proceed by rule.
Soup, 2 quarts; pounded carrot, 1-1/2 lb.; salt, cayenne: 5
minutes.
 A FINER CARROT SOUP.

Scrape very clean, and cut away all blemishes from some highly-
flavoured red carrots; wash, and wipe them dry, and cut them into
quarter-inch slices. Put into a large stewpan three ounces of the best
butter, and when it is melted, add two pounds of the sliced carrots,
and let them stew gently for an hour without browning; pour to them
then four pints and a half of brown gravy soup, and when they have
simmered from fifty minutes to an hour, they ought to be sufficiently
tender. Press them through a sieve or strainer with the soup; add
salt, and cayenne if required; boil the whole gently for five minutes,
take off all the scum, and serve the soup as hot as possible.
Butter, 3 oz.; carrots, 2 lbs.: 1 hour. Soup, 4-1/2 pints: 50 to 60
minutes. Salt, cayenne: 5 minutes.
 COMMON TURNIP SOUP.

Wash and wipe the turnips, pare and weigh them; allow a pound
and a half for every quart of soup. Cut them in slices about a quarter
of an inch thick. Melt four ounces of butter in a clean stewpan, and
put in the turnips before it begins to boil; stew them gently for three
quarters of an hour, taking care that they shall not brown, then have
the proper quantity of soup ready boiling, pour it to them, and let
them simmer in it for three quarters of an hour. Pulp the whole
through a coarse sieve or soup strainer, put it again on the fire, keep
it stirred until it has boiled three minutes or four, take off the scum,
add salt and pepper if required, and serve it very hot. Turnips, 3 lbs.;
butter, 4 oz.: 3/4 hour. Soup, 2 quarts: 3/4 hour. Last time: three
minutes.
 A QUICKLY MADE TURNIP SOUP.

Pare and slice into three pints of veal or mutton stock or of good
broth, three pounds of young mild turnips; stew them gently from
twenty-five to thirty minutes, or until they can be reduced quite to
pulp; rub the whole through a sieve, and add to it another quart of
stock, a seasoning of salt and white pepper, and one lump of sugar:
give it two or three minutes’ boil, skim and serve it. A large white
onion when the flavour is liked may be sliced and stewed with the
turnips. A little cream improves much the colour of this soup.
Turnips, 3 lbs.; soup, 5 pints: 25 to 30 minutes.
 POTATO SOUP.

Mash to a smooth paste three pounds of good mealy potatoes,

which have been steamed, or boiled very dry; mix with them by
degrees, two quarts of boiling broth, pass the soup through a
strainer, set it again on the fire, add pepper and salt, and let it boil for
five minutes. Take off entirely the black scum that will rise upon it,
and serve it very hot with fried or toasted bread. Where the flavour is
approved, two ounces of onions minced and fried a light brown, may
be added to the soup, and stewed in it for ten minutes before it is
sent to table.
Potatoes, 3 lbs.; broth, 2 quarts: 5 minutes. (With onions, 2 oz.) 10
minutes.
 APPLE SOUP.

(Soupe à la Bourguignon.)
Clear the fat from five pints of good mutton broth, bouillon, or shin
of beef stock, and strain it through a fine sieve; add to it when it
boils, a pound and a half of good cooking apples, and stew them
down in it very softly to a smooth pulp; press the whole through a
strainer, add a small teaspoonful of powdered ginger and plenty of
pepper, simmer the soup for a couple of minutes, skim, and serve it
very hot, accompanied by a dish of rice, boiled as for curries.
Broth, 5 pints; apples, 1-1/2 lb.: 25 to 40 minutes. Ginger, 1
teaspoonful; pepper, 1/2 teaspoonful: 2 minutes.
 PARSNEP SOUP.

Dissolve, over a gentle fire, four ounces of good butter, in a wide

stewpan or saucepan, and slice in directly two pounds of sweet
tender parsneps; let them stew very gently until all are quite soft,
then pour in gradually sufficient veal stock or good broth to cover
them, and boil the whole slowly from twenty minutes to half an hour;
work it with a wooden spoon through a fine sieve, add as much stock
as will make two quarts in all, season the soup with salt and white
pepper or cayenne, give it one boil, skim, and serve it very hot. Send
pale fried sippets to table with it.
Butter, 4-1/2 oz.; parsneps, 2 lbs.: 3/4 hour, or more. Stock, 1
quart; 20 to 30 minutes; 1 full quart more of stock; pepper, salt: 1
minute.
Obs.—We can particularly recommend this soup to those who like
the peculiar flavour of the vegetable.
 ANOTHER PARSNEP SOUP.

Slice into five pints of boiling veal stock or strong colourless broth,
a couple of pounds of parsneps, and stew them as gently as
possible from thirty minutes to an hour; when they are perfectly
tender, press them through a sieve, strain the soup to them, season,
boil, and serve it very hot. With the addition of cream, parsnep soup
made by this receipt resembles in appearance the Palestine soup.
Veal stock or broth, 5 pints; parsneps, 2 lbs.: 30 to 60 minutes.
Salt and cayenne: 2 minutes.
 WESTERFIELD WHITE SOUP.

Break the bone of a knuckle of veal in one or two places, and put it
on to stew, with three quarts of cold water to the five pounds of meat;
when it has been quite cleared from scum, add to it an ounce and a
half of salt, and one mild onion, twenty corns of white pepper, and
two or three blades of mace, with a little cayenne pepper. When the
soup is reduced one-third by slow simmering strain it off, and set it
by till cold; then free it carefully from the fat and sediment, and heat it
again in a very clean stewpan. Mix with it when it boils, a pint of thick
cream smoothly blended with an ounce of good arrow-root, two
ounces of very fresh vermicelli previously boiled tender in water
slightly salted and well drained from it, and an ounce and a half of
almonds blanched and cut in strips: give it one minute’s simmer, and
serve it immediately, with a French roll in the tureen.
Veal, 5 lbs.; water, 3 quarts; salt, 1-1/2 oz.; 1 mild onion; 20 corns
white pepper; 2 large blades of mace: 5 hours or more. Cream, 1
pint; almonds, 1-1/2 oz.; vermicelli, 1 oz.: 1 minute. Little thickening if
needed.
Obs.—We have given this receipt without any variation from the
original, as the soup made by it—of which we have often partaken—
seemed always much approved by the guests of the hospitable
country gentleman from whose family it was derived, and at whose
well-arranged table it was very commonly served; but we would
suggest the suppression of the almond spikes, as they seem
unsuited to the preparation, and also to the taste of the present day.
 A RICHER WHITE SOUP.

Pound very fine indeed six ounces of sweet almonds, then add to
them six ounces of the breasts of roasted chickens or partridges,
and three ounces of the whitest bread which has been soaked in a
little veal broth, and squeezed very dry in a cloth. Beat these
altogether to an extremely smooth paste; then pour to them boiling
and by degrees, two quarts of rich veal stock; strain the soup
through a fine hair sieve, set it again over the fire, add to it a pint of
thick cream, and serve it, as soon as it is at the point of boiling.
When cream is very scarce, or not easily to be procured, this soup
may be thickened sufficiently without it, by increasing the quantity of
almonds to eight or ten ounces, and pouring to them, after they have
been reduced to the finest paste, a pint of boiling stock, which must
be again wrung from them through a coarse cloth with very strong
pressure: the proportion of meat and bread also should then be
nearly doubled. The stock should be well seasoned with mace and
cayenne before it is added to the other ingredients.
Almonds, 6 oz.; breasts of chickens or partridges, 6 oz.; soaked
bread, 3 oz.; veal stock, 2 quarts; cream, 1 pint.
Obs. 1.—Some persons pound the yolks of four or five hard-boiled
eggs with the almonds, meat, and bread for this white soup; French
cooks beat smoothly with them an ounce or two of whole rice,
previously boiled from fifteen to twenty minutes.
Obs. 2.—A good plain white soup maybe made simply by adding
to a couple of quarts of pale veal stock or strong well-flavoured veal
broth, a thickening of arrow-root, and from half to three quarters of a
pint of cream. Four ounces of macaroni boiled tender and well-
drained may be dropped into it a minute or two before it is dished,
but the thickening may then be diminished a little.
 MOCK TURTLE SOUP.

To make a single tureen of this favourite English soup in the most

economical manner when there is no stock at hand, stew gently
down in a gallon of water four pounds of the fleshy part of the shin of
beef, or of the neck, with two or three carrots, one onion, a small
head of celery, a bunch of savoury herbs, a blade of mace, a half-
teaspoonful of peppercorns, and an ounce of salt. When the meat is
quite in fragments, strain off the broth, and pour it when cold upon
three pounds of the knuckle or of the neck of veal; simmer this until
the flesh has quite fallen from the bones, but be careful to stew it as
softly as possible, or the quantity of stock will be so much reduced
as to be insufficient for the soup. Next, take the half of a fine calf’s
head with the skin on, remove the brains, and then bone it[27]
entirely, or let the butcher do this, and return the bones with it; these,
when there is time, may be stewed with the veal to enrich the stock,
or boiled afterwards with the head and tongue. Strain the soup
through a hair-sieve into a clean pan, and let it drain closely from the
meat. When it is nearly or quite cold, clear off all the fat from it; roll
the head lightly round, leaving the tongue inside, or taking it out, as
is most convenient, secure it with tape or twine, pour the soup over,
and bring it gently to boil upon a moderate fire; keep it well skimmed,
and simmer it from an hour to an hour and a quarter; then lift the
head into a deep pan or tureen, add the soup to it, and let it remain
in until nearly cold, as this will prevent the edges from becoming
dark. Cut into quarter-inch slices, and then divide into dice, from six
to eight ounces of the lean of an undressed ham, and if possible,
one of good flavour; free it perfectly from fat, rind, and the smoked
edges; peel and slice four moderate-sized eschalots, or if these
should not be at hand, one mild onion in lieu of them. Dissolve in a
well-tinned stewpan or thick iron saucepan which holds a gallon or
more, four ounces of butter; put in the ham and eschalots, or onion,
with half a dozen cloves, two middling-sized blades of mace, a half-
teaspoonful of peppercorns, three or four very small sprigs of thyme,
three teaspoonsful of minced parsley, one of lemon thyme and winter
savoury mixed, and when the flavour is thought appropriate, the very
thin rind of half a small fresh lemon. Stew these as softly as possible
for nearly or quite an hour, and keep the pan frequently shaken: then
put into a dredging box two ounces of fine dry flour, and sprinkle it to
them by degrees; mix the whole well together, and after a few
minutes more of gentle simmering, add very gradually five full pints
of the stock taken free of fat and sediment, and made boiling before
it is poured in; shake the pan strongly round as the first portions of it
are added, and continue to do so until it contains from two to three
pints, when the remainder may be poured in at once, and the pan
placed by the side of the fire that it may boil in the gentlest manner
for an hour. At the end of that time turn the whole into a hair-sieve
placed over a large pan, and if the liquid should not run through
freely, knock the sides of the sieve, but do not force it through with a
spoon, as that would spoil the appearance of the stock. The head in
the meanwhile should have been cut up, ready to add to it. For the
finest kind of mock turtle, only the skin, with the fat that adheres to it,
should be used; and this, with the tongue, should be cut down into
one inch squares, or if preferred into strips of an inch wide. For
ordinary occasions, the lean part of the flesh may be added also, but
as it is always sooner done than the skin, it is better to add it to the
soup a little later. When it is quite ready, put it with the strained stock
into a clean pan, and simmer it from three quarters of an hour to a
full hour: it should be perfectly tender, without being allowed to
break. Cayenne, if needed, should be thrown into the stock before it
is strained; salt should be used sparingly, on account of the ham,
until the whole of the other ingredients have been mixed together,
when a sufficient quantity must be stirred into the soup to season it
properly. A couple of glasses of good sherry or Madeira, with a
dessertspoonful of strained lemon-juice, are usually added two or
three minutes only before the soup is dished, that the spirit and
flavour of the wine may not have time to evaporate; but it is
sometimes preferred mellowed down by longer boiling. The
proportion of lemon-juice may be doubled at will, but much acid is
not generally liked. We can assure the reader of the excellence of
the soup made by this receipt; it is equally palatable and delicate,
and not heavy or cloying to the stomach, like many of the elaborate
compositions which bear its name. The fat, through the whole
process, should be carefully skimmed off. The ham gives far more
savour, when used as we have directed, than when, even in much
larger proportion, it is boiled down in the stock. Two dozens of
forcemeat-balls, prepared by the receipt No. 11, Chap. VIII., should
be dropped into the soup when it is ready for table. It is no longer
customary to serve egg-balls in it.
27. This is so simple and easy a process, that the cook may readily accomplish it
with very little attention. Let her only work the knife close to the bone always,
so as to take the flesh clean from it, instead of leaving large fragments on.
The jaw-bone may first be removed, and the flesh turned back from the edge
of the other.

First broth:—shin, or neck of beef, 4 lbs.; water, 4 quarts; carrots,

2 or 3; large mild onion, 1; celery, small head; bunch savoury herbs;
mace, 1 large blade; peppercorns, 1/2 teaspoonful; cloves, 6; salt, 1
oz.: 5 hours or more, very gently. For stock: the broth and 3 lbs. neck
or knuckle of veal (bones of head if ready): 4 to 5 hours. Boned half-
head with skin on and tongue, 1 to 1-1/4 hour. Lean of undressed
ham, 6 to 8 oz. (6 if very salt); shalots, 4, or onion, 1; fresh butter, 4
oz.; cloves, 6; middling-sized blades of mace, 2; peppercorns, 1/2
teaspoonful; small sprigs of thyme, 3 or 4; minced parsley, 3 large
teaspoonsful; minced savoury and lemon-thyme mixed, 1 small
teaspoonful (thin rind 1/2 small lemon, when liked): 1 hour. Flour, 2
oz.: 5 minutes. Stock, full five pints; flesh of head and tongue, 1-3/4
to 2 lbs.: 3/4 of an hour to 1 hour (salt, if needed, to be added in
interim). Good sherry or Madeira, 2 wineglassesful; lemon-juice, 1 to
2 dessertspoonsful; forcemeat-balls, 24.
Obs. 1.—The beef, veal, bones of the head, and vegetables may
be stewed down together when more convenient: it is only necessary
that a really good, well flavoured, and rather deeply-coloured stock
should be prepared. A calf’s foot is always an advantageous addition
to it, and the skin of another calf’s head[28] a better one still.
28. Country butchers, in preparing a calf’s head for sale in the ordinary way, take
off the skin (or scalp), considered so essential to the excellence of this soup,
and frequently throw it away; it may, therefore, often be procured from them
at very slight cost, and is the best possible addition to the mock turtle. It is
 cleared from the head in detached portions with the hair on, but this may
easily be removed after a few minutes’ scalding as from the head itself, or
the feet, by the direction given in Chapter of Sweet Dishes. In London it is
sold entire, and very nicely prepared, and may be served in many forms,
besides being added to soup with great advantage.

Obs. 2.—A couple of dozens mushroom-buttons, cleaned with salt

and flannel, then wiped very dry, and sliced, and added to the ham
and herbs when they have been simmered together about half an
hour, will be found an improvement to the soup.
Claret is sometimes added instead of sherry or Madeira, but we do
not think it would in general suit English taste so well. From two to
three tablespoonsful of Harvey’s sauce can be stirred in with the
wine when it is liked, or when the colour requires deepening.
 OLD-FASHIONED MOCK TURTLE.

After having taken out the brain and washed and soaked the head
well, pour to it nine quarts of cold water, bring it gently to boil, skim it
very clean, boil it if large an hour and a half, lift it out, and put into the
liquor eight pounds of neck of beef lightly browned in a little fresh
butter, with three or four thick slices of lean ham, four large onions
sliced, three heads of celery, three large carrots, a large bunch of
savoury herbs, the rind of a lemon pared very thin, a dessertspoonful
of peppercorns, two ounces of salt, and after the meat has been
taken from the head, all the bones and fragments. Stew these gently
from six to seven hours, then strain off the stock and set it into a very
cool place, that the fat may become firm enough on the top to be
cleared off easily. The skin and fat of the head should be taken off
together and divided into strips of two or three inches in length, and
one in width; the tongue may be carved in the same manner, or into
dice. Put the stock, of which there ought to be between four and five
quarts, into a large soup or stewpot; thicken it when it boils with four
ounces of fresh butter[29] mixed with an equal weight of fine dry
flour, a half-teaspoonful of pounded mace, and a third as much of
cayenne (it is better to use these sparingly at first, and to add more
should the soup require it, after it has boiled some little time); pour in
half a pint of sherry, stir the whole together until it has simmered for
a minute or two, then put in the head, and let it stew gently from an
hour and a quarter to an hour and a half: stir it often, and clear it
perfectly from scum. Put into it just before it is ready for table three
dozens of small forcemeat-balls; the brain cut into dice (after having
been well soaked, scalded,[30] and freed from the film), dipped into
beaten yolk of egg, then into the finest crumbs mixed with salt, white
pepper, a little grated nutmeg, fine lemon-rind, and chopped parsley
fried a fine brown, well drained and dried; and as many egg-balls,
the size of a small marble, as the yolks of four eggs will supply. (See
Chapter VIII). This quantity will be sufficient for two large tureens of
soup; when the whole is not wanted for table at the same time, it is
better to add wine only to so much as will be required for immediate
consumption, or if it cannot conveniently be divided, to heat the wine
in a small saucepan with a little of the soup, to turn it into the tureen,
and then to mix it with the remainder by stirring the whole gently after
the tureen is filled. Some persons simply put in the cold wine just
before the soup is dished, but this is not so well.
29. When the butter is considered objectionable, the flour, without it, may be
mixed to the smoothest batter possible, with a little cold stock or water, and
stirred briskly into the boiling soup: the spices should be blended with it.

30. The brain should be blanched, that is, thrown into boiling water with a little
salt in it, and boiled from five to eight minutes, then lifted out and laid into
cold water for a quarter of an hour: it must be wiped very dry before it is fried.

Whole calf’s head with skin on, boiled 1-1/2 hour. Stock: neck of
beef, browned in butter, 8 lbs.; lean of ham, 1/2 to 3/4 lb.; onions, 4;
large carrots, 3; heads of celery, 3; large bunch herbs; salt, 2 oz. (as
much more to be added when the soup is made as will season it
sufficiently); thin rind, 1 lemon; peppercorns, 1 dessertspoonful;
bones and trimmings of head: 8 hours. Soup: stock, 4 to 5 quarts;
flour and butter for thickening, of each 4 oz.; pounded mace, half-
teaspoonful; cayenne, third as much (more of each as needed);
sherry, half pint: 2 to 3 minutes. Flesh of head and tongue, nearly or
quite 2 lbs.: 1-1/4 to 1-1/2 hour. Forcemeat-balls, 36; the brain cut
and fried; egg-balls, 16 to 24.
Obs.—When the brain is not blanched it must be cut thinner in the
form of small cakes, or it will not be done through by the time it has
taken enough colour: it may be altogether omitted without much
detriment to the soup, and will make an excellent corner dish if
gently stewed in white gravy for half an hour, and served with it
thickened with cream and arrow-root to the consistency of good
white sauce, then rather highly seasoned, and mixed with plenty of
minced parsley, and some lemon-juice.
 GOOD CALF’S HEAD SOUP.

(Not expensive.)

Stew down from six to seven pounds of the thick part of a shin of
beef with a little lean ham, or a slice of hung beef, or of Jewish beef,
trimmed free from the smoky edges, in five quarts of water until
reduced nearly half, with the addition, when it first begins to boil, of
an ounce of salt, a large bunch of savoury herbs, one large onion, a
head of celery, three carrots, two or three turnips, two small blades
of mace, eight or ten cloves, and a few white or black peppercorns.
Let it boil gently that it may not be too much reduced, for six or seven
hours, then strain it into a clean pan and set it by for use. Take out
the bone from half a calf’s head with the skin on (the butcher will do
this if desired), wash, roll, and bind it with a bit of tape or twine, and
lay it into a stewpan, with the bones and tongue; cover the whole
with the beef stock, and stew it for an hour and a half; then lift it into
a deep earthen pan and let it cool in the liquor, as this will prevent
the edges from becoming dry or discoloured. Take it out before it is
quite cold; strain, and skim all the fat carefully from the stock; and
heat five pints in a large clean saucepan, with the head cut into small
thick slices or into inch-squares. As quite the whole will not be
needed, leave a portion of the fat, but add every morsel of the skin to
the soup, and of the tongue also. Should the first of these not be
perfectly tender, it must be simmered gently till it is so; then stir into
the soup from six to eight ounces of fine rice-flour mixed with a
quarter-teaspoonful of cayenne, twice as much freshly pounded
mace, half a wineglassful of mushroom catsup,[31] and sufficient
cold broth or water to render it of the consistence of batter; boil the
whole from eight to ten minutes; take off the scum, and throw in two
glasses of sherry; dish the soup and put into the tureen some
delicately and well fried forcemeat-balls made by the receipt No. 1,
2, or 3, of Chapter VIII. A small quantity of lemon-juice or other acid
can be added at pleasure. The wine and forcemeat-balls may be
omitted, and the other seasonings of the soup a little heightened. As
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

textbookfull.com