(Ebook) Noncommutative Analysis, Operator Theory and

Applications by Daniel Alpay, Fabio Cipriani, Fabrizio

Colombo, Daniele Guido, Irene Sabadini, Jean-Luc
Sauvageot (eds.) ISBN 9783319291147, 9783319291161,
3319291149, 3319291165 Pdf Download

https://ebooknice.com/product/noncommutative-analysis-operator-
theory-and-applications-6746948

★★★★★
4.7 out of 5.0 (17 reviews )

DOWNLOAD PDF

ebooknice.com
 (Ebook) Noncommutative Analysis, Operator Theory and
Applications by Daniel Alpay, Fabio Cipriani, Fabrizio
Colombo, Daniele Guido, Irene Sabadini, Jean-Luc Sauvageot
(eds.) ISBN 9783319291147, 9783319291161, 3319291149,
3319291165 Pdf Download

EBOOK

Available Formats

■ PDF eBook Study Guide Ebook

EXCLUSIVE 2025 EDUCATIONAL COLLECTION - LIMITED TIME

INSTANT DOWNLOAD VIEW LIBRARY

Here are some recommended products for you. Click the link to
download, or explore more at ebooknice.com

(Ebook) Biota Grow 2C gather 2C cook by Loucas, Jason; Viles,

James ISBN 9781459699816, 9781743365571, 9781925268492,
1459699815, 1743365578, 1925268497

https://ebooknice.com/product/biota-grow-2c-gather-2c-cook-6661374

(Ebook) Matematik 5000+ Kurs 2c Lärobok by Lena Alfredsson, Hans

Heikne, Sanna Bodemyr ISBN 9789127456600, 9127456609

https://ebooknice.com/product/matematik-5000-kurs-2c-larobok-23848312

(Ebook) SAT II Success MATH 1C and 2C 2002 (Peterson's SAT II

Success) by Peterson's ISBN 9780768906677, 0768906679

https://ebooknice.com/product/sat-ii-success-math-1c-and-2c-2002-peterson-
s-sat-ii-success-1722018

(Ebook) Master SAT II Math 1c and 2c 4th ed (Arco Master the SAT
Subject Test: Math Levels 1 & 2) by Arco ISBN 9780768923049,
0768923042

https://ebooknice.com/product/master-sat-ii-math-1c-and-2c-4th-ed-arco-
master-the-sat-subject-test-math-levels-1-2-2326094
(Ebook) Cambridge IGCSE and O Level History Workbook 2C - Depth
Study: the United States, 1919-41 2nd Edition by Benjamin
Harrison ISBN 9781398375147, 9781398375048, 1398375144,
1398375047
https://ebooknice.com/product/cambridge-igcse-and-o-level-history-
workbook-2c-depth-study-the-united-states-1919-41-2nd-edition-53538044

(Ebook) Noncommutative Functional Calculus: Theory and

Applications of Slice Hyperholomorphic Functions by Fabrizio
Colombo, Irene Sabadini, Daniele C. Struppa (auth.) ISBN
9783034801096, 9783034801102, 3034801092, 3034801106
https://ebooknice.com/product/noncommutative-functional-calculus-theory-
and-applications-of-slice-hyperholomorphic-functions-2114096

(Ebook) Michele Sce's Works in Hypercomplex Analysis: A

Translation with Commentaries by Fabrizio Colombo, Irene
Sabadini, Daniele C. Struppa ISBN 9783030502157, 9783030502164,
3030502155, 3030502163
https://ebooknice.com/product/michele-sce-s-works-in-hypercomplex-analysis-
a-translation-with-commentaries-22504598

(Ebook) Operator theory, systems theory, and scattering theory:

multidimensional generalizations by Daniel Alpay, Victor
Vinnikov ISBN 9783764372125, 9783764373030, 3764372125,
3764373032
https://ebooknice.com/product/operator-theory-systems-theory-and-
scattering-theory-multidimensional-generalizations-886240

(Ebook) New Directions in Function Theory: From Complex to

Hypercomplex to Non-Commutative: Chapman University, November
2019 (Operator Theory: Advances and Applications, 286) by
Editors: Daniel Alpay, Ronen Peretz, David Shoikhet, Mihaela B.
Vajiac ISBN 9783030764722, 3030764729
https://ebooknice.com/product/new-directions-in-function-theory-from-
complex-to-hypercomplex-to-non-commutative-chapman-university-
november-2019-operator-theory-advances-and-applications-286-51196770
 Operator Theory
Advances and Applications
252
Daniel Alpay
Fabio Cipriani
Fabrizio Colombo
Daniele Guido
Irene Sabadini
Jean-Luc Sauvageot
Editors

Noncommutative
Analysis,
Operator Theory
and Applications
Operator Theory: Advances and Applications
Volume 252

Founded in 1979 by Israel Gohberg

Editors:
Joseph A. Ball (Blacksburg, VA, USA)
Harry Dym (Rehovot, Israel)
Marinus A. Kaashoek (Amsterdam, The Netherlands)
Heinz Langer (Wien, Austria)
Christiane Tretter (Bern, Switzerland)

Associate Editors: Honorary and Advisory Editorial Board:

Vadim Adamyan (Odessa, Ukraine) Lewis A. Coburn (Buffalo, NY, USA)
Wolfgang Arendt (Ulm, Germany) Ciprian Foias (College Station, TX, USA)
Albrecht Böttcher (Chemnitz, Germany) J.William Helton (San Diego, CA, USA)
B. Malcolm Brown (Cardiff, UK) Thomas Kailath (Stanford, CA, USA)
Raul Curto (Iowa, IA, USA) Peter Lancaster (Calgary, Canada)
Fritz Gesztesy (Columbia, MO, USA) Peter D. Lax (New York, NY, USA)
Pavel Kurasov (Stockholm, Sweden) Donald Sarason (Berkeley, CA, USA)
Vern Paulsen (Houston, TX, USA) Bernd Silbermann (Chemnitz, Germany)
Mihai Putinar (Santa Barbara, CA, USA) Harold Widom (Santa Cruz, CA, USA)
Ilya M. Spitkovsky (Williamsburg, VA, USA)

Subseries
Linear Operators and Linear Systems
Subseries editors:
Daniel Alpay (Beer Sheva, Israel)
Birgit Jacob (Wuppertal, Germany)
André C.M. Ran (Amsterdam, The Netherlands)

Subseries
Advances in Partial Differential Equations
Subseries editors:
Bert-Wolfgang Schulze (Potsdam, Germany)
Michael Demuth (Clausthal, Germany)
Jerome A. Goldstein (Memphis, TN, USA)
Nobuyuki Tose (Yokohama, Japan)
Ingo Witt (Göttingen, Germany)
Daniel Alpay • Fabio Cipriani
Fabrizio Colombo • Daniele Guido
Irene Sabadini • Jean-Luc Sauvageot
Editors

Noncommutative Analysis,
Operator Theory and
Applications
Editors
Daniel Alpay Fabio Cipriani
Department of Mathematics Dipartimento di Matematica
Ben-Gurion University of the Negev Politecnico di Milano
Beer Sheva, Israel Milano, Italy

Fabrizio Colombo Daniele Guido

Dipartimento di Matematica Dipartimento di Matematica
Politecnico di Milano Università di Roma “Tor Vergata”
Milano, Italy Roma, Italy

Irene Sabadini Jean-Luc Sauvageot

Dipartimento di Matematica Institut de Mathématiques
Politecnico di Milano Université Pierre et Marie Curie
Milano, Italy Paris, France

ISSN 0255-0156 ISSN 2296-4878 (electronic)

Operator Theory: Advances and Applications
ISBN 978-3-319-29114-7 ISBN 978-3-319-29116-1 (eBook)
DOI 10.1007/978-3-319-29116-1

Library of Congress Control Number: 2016945037

Mathematics Subject Classification (2010): 46Hxx, 46Jxx, 46Lxx, 47Axx, 47Lxx, 58Axx

© Springer International Publishing Switzerland 2016

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.

Printed on acid-free paper

This book is published under the trade name Birkhäuser.

The registered company is Springer International Publishing AG (www.birkhauser-science.com)
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

F. Arici, F. D’Andrea and G. Landi

Pimsner Algebras and Circle Bundles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

S. Bernstein
A Fractional Dirac Operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

V. Bolotnikov
On the Sylvester Equation over Quaternions . . . . . . . . . . . . . . . . . . . . . . . . 43

J. Bourgain and D.-V. Voiculescu

The Essential Centre of the mod a Diagonalization Ideal
Commutant of an n-tuple of Commuting Hermitian Operators . . . . . . 77

P. Cerejeiras and N. Vieira

Clifford–Hermite Polynomials in Fractional Clifford Analysis . . . . . . . . 81

F. Cipriani and J.-L. Sauvageot

Negative Definite Functions on Groups with Polynomial Growth . . . . 97

F. Colombo, I. Sabadini and D.C. Struppa

An Introduction to Superoscillatory Sequences . . . . . . . . . . . . . . . . . . . . . . 105

H.S.V. de Snoo and H. Woracek

Restriction and Factorization for Isometric and Symmetric
Operators in Almost Pontryagin Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

G. Dell’Antonio
Measurements vs. Interactions:
Tracks in a Wilson Cloud Chamber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

M. Elin and D. Shoikhet

The Radii Problems for Holomorphic Mappings in J ∗ -algebras . . . . . . 181
vi Contents

U. Franz, A. Kula and A. Skalski

Lévy Processes on Quantum Permutation Groups . . . . . . . . . . . . . . . . . . . 193

D. Guido and T. Isola

New Results on Old Spectral Triples for Fractals . . . . . . . . . . . . . . . . . . . . 261

W.A. Majewski and L.E. Labuschagne

Why Are Orlicz Spaces Useful for Statistical Physics? . . . . . . . . . . . . . . . 271
Preface
This volume contains papers written by some of the speakers of the Conference
Noncommutative Analysis, Operator Theory, and Applications held in Milano from
23 to 27 June 2014, and some invited contributions. The Conference has been an
occasion for researchers in different areas to meet and to share their knowledge
and ideas. The contents of the volume reflect, we hope, the effort to find a place
where researchers from the different areas may interact. All contributed papers
represent the most recent achievements in the area as well as “state-of-the-art”
expositions.
The Editors are grateful to the contributors to this volume and to the referees,
for their painstaking and careful work.
The Editors thank the Dipartimento di Matematica, Politecnico di Milano
for hosting the Conference and for assigning to them the FARB (Fondo di Ateneo
per la Ricerca di Base). Moreover they thank the Fondazione Gruppo Credito
Valtellinese as well as Birkhäuser Basel for the financial support.

October 2015 The Editors

Operator Theory:
Advances and Applications, Vol. 252, 1–25

c 2016 Springer International Publishing

Pimsner Algebras and Circle Bundles

Francesca Arici, Francesco D’Andrea and Giovanni Landi

Abstract. We report on the connections between noncommutative principal

circle bundles, Pimsner algebras and strongly graded algebras. We illustrate
several results with examples of quantum weighted projective and lens spaces
and θ-deformations.
Mathematics Subject Classification (2010). Primary 19K35; Secondary 55R25,
46L08.
Keywords. Pimsner algebras, quantum principal bundles, graded algebras,
noncommutative geometry, quantum projective and lens spaces.

Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2. Hilbert C ∗ -modules and Morita equivalence . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1. Hilbert C ∗ -modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2. Morita equivalence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3. Self-Morita equivalence bimodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.4. Frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3. Pimsner algebras and generalized crossed products . . . . . . . . . . . . . . . . . . 9
3.1. The Pimsner algebra of a self-Morita equivalence . . . . . . . . . . . . . . . . . . 9
3.2. Generalized crossed products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.3. Algebras and circle actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.4. Six-term exact sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4. Principal bundles and graded algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.1. Noncommutative principal circle bundles and line bundles . . . . . . . . . 15
4.2. Line bundles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.3. Strongly graded algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.4. Pimsner algebras from principal circle bundles . . . . . . . . . . . . . . . . . . . . . 18
5. Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.1. Quantum weighted projective and quantum lens spaces . . . . . . . . . . . . 19
5.2. Twisting of graded algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2 F. Arici, F. D’Andrea and G. Landi

1. Introduction
This paper is devoted to Pimsner (or Cuntz–Krieger–Pimsner) algebras, focusing
on their connections with noncommutative principal circle bundles as well as with
(strongly) graded algebras.
Pimsner algebras, which were introduced in the seminal work [22], provide a
unifying framework for a range of important C ∗ -algebras including crossed prod-
ucts by the integers, Cuntz–Krieger algebras [8, 9], and C ∗ -algebras associated to
partial automorphisms [11]. Due to their flexibility and wide range of applicability,
there has been recently an increasing interest in these algebras (see for instance
[13, 24]). A related class of algebras, known as generalized crossed products, was
independently invented in [1]. The two notions coincide in many cases, in particular
in those of interest for the present paper. Here we will use a more geometrical point
of view, showing how certain Pimsner algebras, coming from a self-Morita equiva-
lence bimodule over a C ∗ -algebra, can be thought of as algebras of functions on the
total space of a noncommutative principal circle bundle, along the lines of [4, 10].
Classically, starting from a principal circle bundle P over a compact topo-
logical space X, an important role is in the associated line bundles. Given any of
these, the corresponding module of sections is a self-Morita equivalence bimodule
for the commutative C ∗ -algebra C(X) of continuous functions over X. Suitable
tensor powers of the (sections of the) bundle are endowed with an algebra struc-
ture eventually giving back the C ∗ -algebra C(P ) of continuous functions over P .
This is just a Pimsner algebra construction. By analogy then, one thinks of a self-
Morita equivalence bimodule over an arbitrary C ∗ -algebra as a noncommutative
line bundle and of the corresponding Pimsner algebra as the ‘total space’ algebra
of a principal circle fibration.
The Euler class of a (classical) line bundle has an important use in the Gysin
sequence in complex K-theory, that relates the topology of the base space X to
that of the total space P of the bundle. This sequence has natural counterparts
in the context of Pimsner algebras, counterparts given by two sequences in KK-
theory with natural analogues of the Euler class and a central role played by index
maps of canonical classes.
In order to make this review self-contained, we start in §2 from recalling the
theory of Hilbert modules and Morita equivalences, focusing on those definitions
and results that will be needed in the sequel of the paper. Then §3 is devoted to
Pimsner’s construction and to the construction of generalized crossed products.
This is followed by the six-term exact sequences in KK-theory. We next move
in §4 to noncommutative principal circle bundles and graded algebras and recall
how principality of the action can be translated into an algebraic condition on the
induced grading. This condition is particularly relevant and it resembles a similar
condition appearing in the theory of generalized crossed products. We then show
how all these notions are interconnected and can be seen as different aspects of
the same phenomenon. Finally, §5 is devoted to examples: we illustrate how theta
deformed and quantum weighted projective and lens spaces fit into the framework.
 Pimsner Algebras 3

2. Hilbert C ∗ -modules and Morita equivalence

Hilbert C ∗ -modules are crucially important in modern developments of noncom-
mutative geometry and index theory. They are a generalization of Hilbert spaces
where the complex scalars are replaced by a C ∗ -algebra. From a geometrical point
of view, they can be thought of as modules of sections of a noncommutative Her-
mitian vector bundle.
We recall here some of the definitions and results that we need later on in
the paper. Our main references for this section are [18, 23].

2.1. Hilbert C ∗ -modules

Definition 2.1. Let B be a C ∗ -algebra. A right pre-Hilbert C ∗ -module over B is
a right B-module E with a B-valued Hermitian product ·, ·B : E × E → B
satisfying, for all ξ, η ∈ E and for all b ∈ B, the conditions:
 ∗
ξ, ηB = η, ξB ,
ξ, ηbB = ξ, ηB b ,
ξ, ξB ≥ 0 , and ξ, ξB = 0 ⇐⇒ ξ = 0.
A right Hilbert C ∗ -module E over B is a right pre-Hilbert C ∗ -module which is
complete in the norm · on E defined, via the norm · B on B, by
ξ 2
= ξ, ξB B .
Finally, it is said to be full if the ideal
E, EB := SpanB {ξ, ηB | ξ, η ∈ E}
is dense in B. 

There are analogous definitions for left modules, with Hermitian product
denoted B  , ; this is taken to be B- linear in the first entry,
To lighten notations we shall write ·, · = ·, ·B whenever possible and use
the name Hilbert B-module to mean a Hilbert C ∗ -module over the C ∗ -algebra B.
The simplest example of a Hilbert B-module is the algebra B itself with respect
to the Hermitian product
a, b = a∗ b .
The module B n consists of n-tuples of elements of B, with component-wise
operations, and with Hermitian product defined by
n
(a1 , . . . , an ), (b1 , . . . , bn ) = ai , bi  . (2.1)
i=1

none has a finite set {Ei }i=1 of Hilbert B-modules,

n n
Generalizing B , whenever
one can form the direct sum i=1 Ei . It is a Hilbert B-module in a obvious way
(component-wise) with Hermitian product defined as in (2.1).
Things become subtler if {Ei }i∈I is an infinite collection of Hilbert B-mod-
ules. Indeed, one needs to replace i∈I Ei by the set of sequences (ξi ), with ξi ∈ Ei
4 F. Arici, F. D’Andrea and G. Landi


and such that i∈I ξi , ξi  converges in B. Then for ξ = (ξi ) and η = (ηi ), one gets
a complete module (hence a Hilbert B-module) with the natural inner product:

ξ, η = ξi , ηi .
i∈I

A Hilbert B-module E is finitely generated if there exists a finite collection

{η1 , . . . , ηn } of elements of E such that the B-linear span of the ηi ’s is dense
in E.It is algebraically finitely generated if every element ξ ∈ E is of the form
ξ = nj=1 ηj bj for some bj ’s in B.
A Hilbert B-module E is projective if it is a direct summand in B n for some
n. By [29, 15.4.8], every algebraically finitely generated Hilbert C ∗ -module over a
unital C ∗ -algebra is projective.
Next, one defines operators between Hilbert modules: let E, F be two Hilbert
B-modules over the same C ∗ -algebra B.
Definition 2.2. One says that an operator T : E → F is adjointable if there exists
an operator T ∗ : F → E such that
T ξ, η = ξ, T ∗ η, for all ξ ∈ E, η ∈ F . 
An adjointable operator is automatically B-linear and, by the Banach–Stein-
haus theorem, bounded. The collection of adjointable operators from E to F is
denoted by LB (E, F ). A bounded B-linear operator need not be adjointable – a
simple counterexample is the inclusion i : I → B of a proper ideal in a unital
C ∗ -algebra – thus the need for the definition. Clearly, if T ∈ LB (E, F ), then
T ∗ ∈ LB (F, E). In particular, LB (E) := LB (E, E) is a ∗-algebra; it is in fact a
C ∗ -algebra for the operator norm.
There is an important class of operators which is built from ‘finite rank’
operators. For any ξ ∈ F and η ∈ E one defines the operator θξ,η : E → F as
θξ,η (ζ) = ξη, ζ, for all ζ ∈E. (2.2)
∗
Every such θξ,η is adjointable, with adjoint := θη,ξ : F → E. The closed linear
θξ,η
subspace of LB (E, F ) spanned by {θξ,η | ξ, η ∈ E} is denoted by KB (E, F ). In
particular KB (E) := KB (E, E) ⊆ LB (E); this is a closed ideal, whose elements
are referred to as compact endomorphisms. When possible we write L(E) = LB (E)
and K(E) = KB (E).
The C ∗ -algebraic dual of E, denoted by E ∗ , is defined as the space
E ∗ := {φ ∈ HomB (E, B) | ∃ ξ ∈ E such that φ(η) = ξ, η ∀η ∈ E} . (2.3)
Thus, with ξ ∈ E, if λξ : E → B is the operator defined by λξ (η) = ξ, η, for all
η ∈ E, every element of E ∗ is of the form λξ for some ξ ∈ E.
By its definition, E ∗ := KB (E, B). One says that a module E is self-dual
if the C ∗ -algebraic dual E ∗ coincides with LB (E, B), i.e., if the module map
E  ξ → λξ ∈ LB (E, B), is surjective. If B is unital, then B n is self-dual for any
n ≥ 1. As a consequence, every finitely generated projective Hilbert C ∗ -module
over a unital C ∗ -algebra is self-dual as well.
 Pimsner Algebras 5

2.2. Morita equivalence

Given a right Hilbert B-module E, by construction, compact endomorphisms act
from the left on E. Then, by defining:
K(E) ξ, η := θξ,η ,
we obtain a natural K(E)-valued Hermitian product on E. Note this is left linear
over K(E), that is K(E) T · ξ, η = T · (K(E) ξ, η) for T ∈ K(E). Thus E is a left
Hilbert K(E)-module and by the very definition of K(E), E is full over K(E). One
easily checks the compatibility condition
K(E) ξ, ηζ = ξη, ζB , for all ξ, η, ζ ∈ E. (2.4)
In particular, the B-valued and K(E)-valued norms coincide [23, Lem. 2.30]. By
its definition, K(E) acts by adjointable operators on the right B-module E. On
the other hand, with b ∈ B and ξ, η, ζ ∈ E, one computes:
K(E) ξb, ηζ = (ξb)η, ζB = ξηb∗ , ζB = K(E) ξ, ηb∗ ζ,
that is, B acts by adjointable operators on the left K(E)-module E.
This example motivates the following definitions:
Definition 2.3. Given two C ∗ -algebras A and B, a Hilbert (A, B)-bimodule E is a
right Hilbert B-module with B-valued Hermitian product  , B , which is at the
same time a left Hilbert A-module with A-valued Hermitian product A  ,  and
such that the Hermitian products are compatible, that is,
ξη, ζB = A ξ, ηζ, for all ξ, η, ζ ∈ E. (2.5)
Note that  , B is right B-linear, while A  ,  is left A-linear. 
Definition 2.4. An (A, B)-equivalence bimodule is a Hilbert (A, B)-bimodule E
that is full both as a left and as a right Hilbert module and such that
aξ, ηB = ξ, a∗ ηB and A ξb, η = A ξ, ηb∗ , (2.6)
for all ξ, η ∈ E, a ∈ A, b ∈ B. If there exist an (A, B)-equivalence bimodule one
says that the two C ∗ -algebras A and B are Morita equivalent. 
Condition (2.6) says that A acts by adjointable operators on EB – that is
the bimodule E thought of as a right B-module –, with the adjoint of a ∈ A being
its conjugated a∗ in the C ∗ -algebra A. Similarly B acts by adjointable operators
on A E. From the considerations above on the algebra K(E) of compact endomor-
phisms and in particular the compatibility condition (2.4), it is not surprising that
the algebra K(E) has a central role for Morita equivalence:
Proposition 2.5 ([23, Prop. 3.8]). Let E be a full Hilbert B-module. Then E is
a (K(E), B)-equivalence bimodule with K(E)-valued Hermitian product given by
K(E) ξ, η = θξ,η . Conversely, if E is an (A, B)-equivalence bimodule, then there
exists an isomorphism φ : A → K(E) such that
φ (A ξ, η) = K(E) ξ, η, for all ξ, η ∈ E.
6 F. Arici, F. D’Andrea and G. Landi

Thus, two C ∗ -algebras A and B are Morita equivalent if and only if

A  KB (EB ) for a full right Hilbert B-module EB , or equivalently, if and only if
B  A K(A E) for a full left Hilbert B-module A E.
In fact, Morita equivalence is an equivalence relation, with transitivity im-
plemented by taking the interior tensor product of bimodules. Given a Hilbert
(A, B)-bimodule E and a Hilbert (B, C)-bimodule F , one can form a Hilbert
(A, C)-bimodule, the interior tensor product bimodule. We do not dwell upon the
details of the construction here while referring to [23, §3.2] for instance. We shall
illustrate a particular case of the construction later on.

2.3. Self-Morita equivalence bimodules

In the present paper we are interested in noncommutative line bundles, that is
the analogue of modules of continuous sections of a line bundle over a topological
space. We are then naturally led to the following definition.
Definition 2.6. A self-Morita equivalence bimodule for B is a pair (E, φ) with E a
full right Hilbert B-module E and φ : B → K(E) an isomorphism. 
The prototypical commutative example of a self-Morita equivalence bimod-
ule is provided by B = C(X), the C ∗ -algebra of continuous functions on a com-
pact topological space, E the C(X)-module of sections of a Hermitian line bundle
L → X and φ the trivial representation.
If (E, φ) is a self-Morita equivalence bimodule over B, the dual E ∗ as defined
in (2.3), can be made into a self-Morita equivalence bimodule over B as well.
Firstly, E ∗ is given the structure of a (right) Hilbert C ∗ -module over B via φ.
Recall that elements of E ∗ are of the form λξ for some ξ ∈ E, with λξ (η) = ξ, η,
for all η ∈ E. The right action of B on E ∗ is given by
λξ b := λξ ◦ φ(b) = λφ(b)∗ ξ ,
the second equality being easily established. The B-valued Hermitian product on
E ∗ uses the left K(E)-valued Hermitian product on E:
λξ , λη  := φ−1 (θξ,η ) ,
and E ∗ is full as well. Next, define a ∗-homomorphism φ∗ : B → L(E ∗ ) by
φ∗ (b)(λξ ) := λξ·b∗ ,
which is in fact an isomorphism φ∗ : B → K(E ∗ ). Thus, the pair (E ∗ , φ∗ ) is a
self-Morita equivalence bimodule over B, according to Definition 2.6.
 φ E of E with itself
We need to recall the notion of interior tensor product E ⊗
over B. As a first step, one considers the algebraic tensor product E ⊗alg E. It has
natural right B-module structure given by
(ξ ⊗ η)b = ξ ⊗ (ηb) , for ξ, η ∈ E , b∈B,
and a B-valued Hermitian product given, on simple tensors, by
ξ1 ⊗ ξ2 , η1 ⊗ η2  = ξ2 , φ(ξ1 , η1 ) η2  (2.7)
 Pimsner Algebras 7

and extended by linearity. This is well defined and has all required properties; in
particular, the null space N = {ζ ∈ E ⊗alg E ; ζ, ζ = 0} is shown to coincide
with the subspace generated by elements of the form
ξb ⊗ η − ξ ⊗ φ(b)η , for ξ, η ∈ E , b∈B. (2.8)
One then takes E ⊗φ E := (E ⊗alg E)/N and defines E ⊗  φ E to be the (full) Hilbert
B-module obtained by completing with respect to the norm induced by restriction
of (2.7) to the quotient E ⊗φ E. We shall simply write ξ ⊗ η to mean the element
ξ ⊗ η + N in E ⊗  φ E.
The module E ⊗  φ E is itself a self-Morita equivalence bimodule over B. To
sketch how this is the case, we need some additional facts (cf. [18, Ch.4]).
For any T ∈ L(E), the operator defined on simple tensors by ξ ⊗η → T (ξ)⊗η
extends to a well-defined bounded operator on E ⊗  φ E denoted by T ⊗ id =: φ∗ (T ).
It is adjointable with adjoint given by T ∗ ⊗ id. Next, for ξ ∈ E, the equation
Sξ (η) = ξ ⊗ η, with η ∈ E, defines an element Sξ ∈ LB (E, E ⊗  φ E) whose adjoint
is Sξ∗ (η ⊗ ζ) = φ(ξ, η)ζ, for all ξ, η, ζ ∈ E. Finally, for all b ∈ B and ξ, η, ζ1 , ζ2 ,
one computes:
Sξ φ(b)Sη∗ (ζ1 ⊗ ζ2 ) = Sξ φ(b)φ(η, ζ1 )ζ2 = ξ ⊗ φ(bη, ζ1 )ζ2
= ξbη, ζ1  ⊗ ζ2 = θξb,η (ζ1 ) ⊗ ζ2 = φ∗ (θξb,η )(ζ1 ⊗ ζ2 ).
Thus
φ∗ (θξb,η ) = Sξ φ(b)Sη∗ ,
which is in K(E ⊗  φ E) since φ(b) ∈ K(E). In fact, since φ is nondegenerate, it
follows (by using an approximate unit for B in general) that φ∗ (θξ,η ) = Sξ Sη∗ .
Furthermore K(E) is generated by elements of the form θξ,η , it follows that
φ∗ (K(E)) ⊂ K(E ⊗  φ E). In fact, since φ is an isomorphism, φ∗ is an isomorphism
as well [18, Prop. 4.7]. In particular, generating elements of K(E ⊗  φ E) of the form
θξ1 ⊗ξ2 ,η1 ⊗η2 can be written as
θξ1 ⊗ξ2 ,η1 ⊗η2 = φ∗ (θξ1 b,η1 ),
with b ∈ B uniquely defined by φ(b) = θξ2 ,η2 .
Proposition 2.7. The isomorphism
 φ E)
φ(2) := φ∗ ◦ φ : B → K(E ⊗
 φ E as a self-Morita equivalence over B.
realizes the Hilbert B-module E ⊗
The construction can be iterated and, for n > 0, one gets the n-fold interior
tensor power of E over B,

 φE ⊗
E ⊗φ n := E ⊗ φ · · · ⊗
 φ E, n-factors;
again a self-Morita equivalence bimodule over B.
Remark 2.8. One could generalize the previous construction and consider for a
C ∗ -algebra B, the collection of self-Morita equivalence bimodules over B (up to
8 F. Arici, F. D’Andrea and G. Landi

unitary equivalence). This has a natural group structure with respect to the inte-
rior tensor product. The inverse of the self-Morita equivalence bimodule (E, φ) is
the dual self-Morita equivalence bimodule (E ∗ , φ∗ ). Thinking of self-Morita equiv-
alence bimodules as line bundles, this group is the Picard group of B, denoted
Pic(B), in analogy with the classical Picard group of a space, which is the group
of isomorphism classes of line bundles with group operation given by the ten-
sor product. It was shown in [2] that the Picard group of a commutative unital
C ∗ -algebra B = C(X) is the semi-direct product of the classical Picard group of
X with the group of automorphisms of the algebra B, which is the same as the
group of homeomorphisms of X. 

2.4. Frames
Let A, B be two unital C ∗ -algebras, and let E be a finitely generated (A, B)-
equivalence bimodule. Since A  K(E), there exists a finite collection of elements
η1 , . . . , ηn ∈ E with the property that

A ηj , ηj  = 1A .
j

Equivalently, by using the isomorphism φ : A → K(E), this means that


θηj ,ηj = 1K(E) . (2.9)
j

As a consequence, one can reconstruct any element of ξ ∈ E as


ξ= ηj ηj , ξB . (2.10)
j

This motivates the following definition [25].

Definition 2.9. Let B be a unital C ∗ -algebra. A finite standard module frame for
the right Hilbert B-module E is a finite family of elements {ηi }nj=1 of E such that,
for all ξ ∈ E, the reconstruction formula (2.10) holds true. 
Remark 2.10. More generally, one could consider frames with countable elements,
with (2.10) replaced by a series
convergent in E, or equivalently (2.9) replaced by
the condition that the series j θηj ,ηj is strictly convergent to the unit of L(E)
(K(E) need not be unital). We refer to [12] for details. 
The existence of a finite frame is a geometrical condition. Indeed, whenever
one has a right Hilbert B-module E with a finite standard module frame, E is
algebraically finitely generated and projective as a right module, with projectivity
following from the fact that the algebra B is unital, with the frame explicitly
providing a projection for E. Indeed, the matrix p = (pjk ) with pjk = ηj , ηk B
is a projection in the matrix algebra Mn (B). By construction (pjk )∗ = pkj and,
using (2.10),

(p2 )jl = ηj , ηk B ηk , ηl B
k

= ηj , ηk ηk , ηl B B = ηj , ηl B = pjl .
k
 Pimsner Algebras 9

This establishes the finite right B-module projectivity of E with the isometric
identification E  pB n . Furthermore, E is self-dual for its Hermitian product.
More generally, E is finitely generated projective whenever there exist two
finite sets {ηi }ni=1 and {ζi }ni=1 of elements of E with the property that

K(E) ηj , ζj  = 1K(E) . (2.11)
j
Then, any element ξ ∈ E can be reconstructed as

ξ= ηj ζj , ξB ,
j

and the matrix with entries given by ejk = ζj , ηk B is an idempotent in Mn (B),
(e2 )jk = ejk , and E  eB n as a right B-module.

3. Pimsner algebras and generalized crossed products

In this section, we recall the construction of the Pimsner algebra [22] (a ring-
theoretic version is discussed in [5]). We also review the notion of generalized
crossed product of a C ∗ -algebra by a Hilbert bimodule, that was introduced in-
dependently in [1]. The two notions are related and in the case of a self-Morita
equivalence they actually coincide.
3.1. The Pimsner algebra of a self-Morita equivalence
In his breakthrough paper [22], Pimsner associates a universal C ∗ -algebra to every
pair (E, φ), with E a right Hilbert B-module for a C ∗ -algebra B and φ : B → L(E)
an isometric ∗-homomorphism. His work was later generalized by Katsura [17], who
removed the injectivity assumption on φ.
Guided by a geometric approach coming from principal circle bundles, we
shall not work in full generality, but rather under the assumption that the pair
(E, φ) is a self-Morita equivalence bimodule for B. Things simplify considerably
and the Pimsner algebra is represented on a Hilbert module [4].
Given a self-Morita equivalence bimodule (E, φ) for the C ∗ -algebra B, in §2.3
we described the interior tensor product E ⊗  φ E, itself a self-Morita equivalence
 φn
bimodule and, more generally, the tensor product module E ⊗ , for n > 0. To
lighten notation, we write
⎧ 
⎪ ⊗ n
⎨E φ n>0
(n)
E := B n=0 .
⎪
⎩ ∗ ⊗  φ∗ (−n)
(E ) n<0
Clearly, E (1) = E and E (−1) = E ∗ . From the definition of these Hilbert B-modules,
one has isomorphisms
K(E (n) , E (m) )  E (m−n) .
Out of them, one constructs the Hilbert B-module E∞ as a direct sum:
E∞ := E (n) ,
n∈Z
10 F. Arici, F. D’Andrea and G. Landi

which could be referred to as a two-sided Fock module. As on usual Fock spaces,

one defines creation and annihilation operators. Firstly, for each ξ ∈ E one has a
bounded adjointable operator (a creation operator) Sξ : E∞ → E∞ , shifting the
degree by +1, defined on simple tensors by:
Sξ (b) := ξ · b , b∈B,
Sξ (ξ1 ⊗ · · · ⊗ ξn ) := ξ ⊗ ξ1 ⊗ · · · ⊗ ξn , n > 0,
Sξ (λξ1 ⊗ · · · ⊗ λξ−n ) := λξ2 ·φ−1 (θξ1 ,ξ ) ⊗ λξ3 ⊗ · · · ⊗ λξ−n , n < 0.
The adjoint of Sξ (an annihilation operator) is given by Sλξ := Sξ∗ : E∞ → E∞ :
Sλξ (b) := λξ · b , b∈ B,
Sλξ (ξ1 ⊗ . . . ⊗ ξn ) := φ(ξ, ξ1 )ξ2 ⊗ ξ3 ⊗ · · · ⊗ ξn , n > 0,
Sλξ (λξ1 ⊗ . . . ⊗ λξ−n ) := λξ ⊗ λξ1 ⊗ · · · ⊗ λξ−n , n < 0;
In particular, Sξ (λξ1 ) = φ−1 (θξ,ξ1 ) ∈ B and Sλξ (ξ1 ) = ξ, ξ1  ∈ B.
Definition 3.1. The Pimsner algebra OE of the pair (E, φ) is the smallest C ∗ -
subalgebra of LB (E∞ ) containing the creation operators Sξ for all ξ ∈ E. 
There is an injective ∗-homomorphism i : B → OE . This is induced by the
injective ∗-homomorphism φ : B → LB (E∞ ) defined by
φ(b)(b ) := b · b ,
φ(b)(ξ1 ⊗ · · · ⊗ ξn ) := φ(b)(ξ1 ) ⊗ ξ2 ⊗ · · · ⊗ ξn ,
φ(b)(λξ1 ⊗ · · · ⊗ λξn ) := φ∗ (b)(λξ1 ) ⊗ λξ2 ⊗ · · · ⊗ λξn
:= λξ1 ·b∗ ⊗ λξ2 ⊗ · · · ⊗ λξn ,
and whose image is in the Pimsner algebra OE . In particular, for all ξ, η ∈ E
it holds that Sξ Sη∗ = i(φ−1 (θξ,η )), that is the operator Sξ Sη∗ on E∞ is right-
multiplication by the element φ−1 (θξ,η ) ∈ B.
A Pimsner algebra is universal in the following sense [22, Thm. 3.12]:
Proposition 3.2. Let C be a C ∗ -algebra and σ : B → C a ∗-homomorphism.
Suppose there exist elements sξ ∈ C such that, for all ξ, η ∈ E, b ∈ B and α, β ∈ C
it holds that:
αsξ + βsη = sαξ+βη ,
sξ σ(b) = sξb and σ(b)sξ = sφ(b)(ξ) ,
s∗ξ sη = σ(ξ, η),
 
sξ s∗η = σ φ−1 (θξ,η ) .
Then, there exists a unique ∗-homomorphism σ : OE → C with the property that
σ(Sξ ) = sξ for all ξ ∈ E.
The following example was already in [22].
 Pimsner Algebras 11

Example 3.3. Let B be a C ∗ -algebra and α : B → B an automorphism of B. Then

(B, α) is naturally a self-Morita equivalence for B. The right Hilbert B-module
structure is the standard one, with right B-valued Hermitian product a, bB =
a∗ b. The automorphism α is used to define the left action via a · b = α(a)b and the
left B-valued Hermitian product by B a, b = α(a∗ b). The module E∞ is isomorphic
to a direct sum of copies of B. Indeed, for all n ∈ Z, the module E (n) is isomorphic
to B as a vector space, with
E (n)  a · (x1 ⊗ · · · ⊗ xn ) −→ αn (a)αn−1 (x1 ) · · · α(xn−1 )xn ∈ B. (3.1)
The corresponding Pimsner algebra OE agrees with the crossed product B α Z.

Example 3.4. In the finitely generated projective case, occurring, e.g., when the
algebra B is unital, the Pimsner algebra of a self-Morita equivalence can be realized
explicitly in terms of generators and relations [15]. Since E is finitely generated
projective, it admits a finite frame {ηj }nj=1 . Then, from the reconstruction formula
as in (2.10), for any b ∈ B:

φ(b)ηj = ηk ηk , φ(b)ηj B .
k
Then the C ∗ -algebra OE is the universal C ∗ -algebra generated by B together with
n operators S1 , . . . , Sn , satisfying

Sk∗ Sj = ηk , ηj B , Sj Sj∗ = 1,
j

and bSj = Sk ηi , φ(b)ηj B ,
k
for b ∈ B, and j = 1, . . . , n. The generators Si are partial isometries if and only
if the frame satisfies ηk , ηj  = 0 for k = j. For B = C and E a Hilbert space of
dimension n, one recovers the original Cuntz algebra On [8]. 
Similarly to crossed products by the integers, Pimsner algebras can be natu-
rally endowed with a circle action α : S1 → Aut(OE ) that turns them into Z-graded
algebras. Indeed, by the universal property in Proposition 3.2 (with C = OE , σ = i
the injection of B into OE , and sξ := z ∗ Sξ ), the map
Sξ → αz (Sξ ) := z ∗ Sξ , z ∈ S1 ,
extends to an automorphism of OE . The degree n part of OE can then be defined
as usual, as the weight space {x ∈ OE : αz (x) = z −n x}.
3.2. Generalized crossed products
A somewhat better framework for understanding the relation between Pimsner
algebras and algebras endowed with a circle action is that of generalized crossed
products. They were introduced in [1] and are naturally associated with Hilbert
bimodules via the notion of a covariant representation.
Definition 3.5. Let E be a Hilbert (B, B)-bimodule (not necessarily full). A co-
variant representation of E on a C ∗ -algebra C is a pair (π, T ) where
12 F. Arici, F. D’Andrea and G. Landi

1. π : B → C is a ∗-homomorphism of algebras;
2. T : E → C satisfies
T (ξ)π(b) = T (ξb) T (ξ)∗ T (η) = π(ξ, ηB )
π(b)T (ξ) = T (bξ) T (ξ)T (η)∗ = π(B ξ, η)
for all b ∈ B and ξ, η ∈ E. 

By [1, Prop. 2.3], covariant representations always exist.

Definition 3.6. Let E be a Hilbert (B, B)-bimodule. The generalized crossed prod-
uct B E Z of B by E is the universal C ∗ -algebra generated by the covariant
representations of E. 
In [1, Prop. 2.9] the generalized crossed product algebra is realized as a cross-
sectional algebra (à la Fell–Doran) for a suitable C ∗ -algebraic bundle over Z.
It is worth stressing that a generalized crossed product need not be a Pimsner
algebra in general, since the representation of B giving the left action need not be
injective. However, by using the universal properties of Proposition 3.2, one shows
that for a self-Morita equivalence bimodule the two constructions yield the same
algebra. The advantage of using generalized crossed products is that a C ∗ -algebra
carrying a circle action that satisfies a suitable completeness condition, can be
reconstructed as a generalized crossed product.
3.3. Algebras and circle actions
Let A be a C ∗ -algebra and {σz }z∈S1 be a strongly continuous action of the circle
S1 on A. For each n ∈ Z, one defines the spectral subspaces
 
An := ξ ∈ A | σz (ξ) = z −n ξ for all z ∈ S1 .
Clearly, the invariant subspace A0 ⊆ A is a C ∗ -subalgebra of A, with unit whenever
A is unital; this is the fixed-point subalgebra. Moreover, the subspace An Am , meant
as the closed linear span of the set of products xy with x ∈ An and y ∈ Am , is
contained in An+m . Thus, the algebra A is Z-graded and the grading is compatible
with the involution, that is A∗n = A−n for all n ∈ Z.
In particular, for any n ∈ Z the space A∗n An is a closed two-sided ideal in A0 .
Thus, each spectral subspace An has a natural structure of Hilbert A0 -bimodule
(not necessarily full) with left and right Hermitian products:
A0 x, y = xy ∗ , x, yA0 = x∗ y, for all x, y ∈ An . (3.2)
The description via spectral subspaces allows one to characterize algebras
that are obtained as generalized crossed products in the sense of Definition 3.6, in
terms of a quite manageable necessary and sufficient condition.
Theorem 3.7 ([1, Thm. 3.1]). Let A be a C ∗ -algebra with a strongly continuous
action of the circle. The algebra A is isomorphic to A0 A1 Z if and only if A
is generated, as a C ∗ -algebra, by the fixed point algebra A0 and the first spectral
subspace A1 of the circle action.
 Pimsner Algebras 13

The above condition was introduced in [11] and is referred to as having a semi-
saturated action. It is fulfilled in a large class of examples, like crossed product
by the integers, and noncommutative (or quantum) principal circle bundles, as
we shall see quite explicitly in §4 below. In fact, this condition encompasses more
general non-principal actions, which are however beyond the scope of the present
paper.
In Theorem 3.7 a crucial role is played by the module A1 . If we assume that
it is a full bimodule, that is if
A∗1 A1 = A0 = A1 A∗1 , (3.3)
the action σ is said to have large spectral subspaces (cf. [21, §2]), a slightly stronger
condition than semi-saturatedness (cf. [4, Prop. 3.4]). Firstly, the condition above
is equivalent to the condition that all bimodules An are full, that is
A∗n An = A0 = An A∗n for all n ∈ Z.
When this happens, all bimodules An are self-Morita equivalence bimodules for
A0 , with isomorphism φ : A0 → KA0 (An ) given by
φ(a)(ξ) := a ξ, for all a ∈ A0 , ξ ∈ An . (3.4)
Combining Theorem 3.7 with the fact that for a self-Morita equivalence the
generalized crossed product construction and Pimsner’s construction yield the
same algebra, we obtain the following result.
Theorem 3.8 ([4, Thm. 3.5]). Let A be a C ∗ -algebra with a strongly continuous
action of the circle. Suppose that the first spectral subspace A1 is a full and count-
ably generated Hilbert bimodule over A0 . Then the Pimsner algebra OA1 of the
self-Morita equivalence (A1 , φ), with φ as in (3.4), is isomorphic to A. The iso-
morphism is given by Sξ → ξ for all ξ ∈ A1 .
Upon completions, all examples considered in the present paper will fit into
the framework of the previous theorem.

3.4. Six-term exact sequences

With a Pimsner algebra there come two natural six-term exact sequences in KK-
theory, which relate the KK-groups of the Pimsner algebra OE with those of the
C ∗ -algebra of (the base space) scalars B. The corresponding sequences in K-theory
are noncommutative analogues of the Gysin sequence which in the commutative
case relates the K-theories of the total space and of the base space of a princi-
pal circle bundle. The classical cup product with the Euler class is replaced, in
the noncommutative setting, by a Kasparov product with the identity minus the
generating Hilbert C ∗ -module E.
Firstly, since φ : B → K(E) ⊆ L(E), the following class is well defined.
Definition 3.9. The class in KK0 (B, B) defined by the even Kasparov module
(E, φ, 0) (with trivial grading) will be denoted by [E]. 
14 F. Arici, F. D’Andrea and G. Landi

Next, consider the orthogonal projection P : E∞ → E∞ with range

∞
Im(P ) = E (n) ⊆ E∞ .
n≥0

Since [P, Sξ ] ∈ K(E∞ ) for all ξ ∈ E, one has [P, S] ∈ K(E∞ ) for all S ∈ OE . Then,
let F := 2P − 1 ∈ L(E∞ ) and let φ : OE → L(E∞ ) be the inclusion.

Definition 3.10. The class in KK1 (OE , B) defined by the odd Kasparov module
(E∞ , φ, F ) will be denoted by [∂]. 

For any separable C ∗ -algebra C we then have the group homomorphisms

[E] : KK∗ (B, C) → KK∗ (B, C) , [E] : KK∗ (C, B) → KK∗ (C, B)
and
[∂] : KK∗ (C, OE ) → KK∗+1 (C, B) , [∂] : KK∗ (B, C) → KK∗+1 (OE , C) ,

which are induced by the Kasparov product.

These yield natural six-term exact sequences in KK-theory [22, Thm. 4.8].
We report here the case C = C. Firstly, the sequence in K-theory:
1−[E] i
K0 (B) −−−−→ K0 (B) −−−∗−→ K0 (OE )
 ⏐
⏐
[∂]⏐
⏐[∂]
 .
K1 (OE ) ←−−−− K1 (B) ←−−−− K1 (B)
i∗ 1−[E]

with i∗ the homomorphism in K-theory induced by the inclusion i : B → OE . This

could be considered as a generalization of the classical Gysin sequence in K-theory
(see [16, IV.1.13]) for the ‘line bundle’ E over the ‘noncommutative space’ B and
with the map 1 − [E] having the same role as the Euler class χ(E) := 1 − [E] of
the line bundle E.
The second sequence would then be an analogue in K-homology:

K 0 (B) ←−−−− K 0 (B) ←−− −− K 0 (OE )

1−[E] i∗
⏐ 
⏐[∂] ⏐
 [∂]⏐ .
i∗ 1−[E]
K 1 (OE ) −−−−→ K 1 (B) −−−−→ K 1 (B)
where now i∗ is the induced homomorphism in K-homology.
Gysin sequences in K-theory were given in [3] for line bundles over quantum
projective spaces and leading to a class of quantum lens spaces. These examples
were generalized in [4] to a class of quantum lens spaces as circle bundles over
quantum weighted projective lines with arbitrary weights.
 Pimsner Algebras 15

4. Principal bundles and graded algebras

Examples of Pimsner algebras come from noncommutative (or quantum) principal
circle bundles. At an algebraic level the latter are intimately related to Z-graded
∗-algebras. When completing with natural norms one is lead to continuous circle
actions on a C ∗ -algebra with Z-grading given by spectral subspaces, that is the
framework described in §3.3.
4.1. Noncommutative principal circle bundles and line bundles
We aim at exploring the connections between (noncommutative) principal circle
bundles, frames for modules as described in §2.4, and Z-graded algebras. The circle
action is dualized in a coaction of the dual group Hopf algebra. Thus, we need to
consider the unital complex algebra
O(U (1)) := C[z, z −1 ]/1 − zz −1 ,
where 1 − zz −1 is the ideal generated by 1 − zz −1 in the polynomial algebra
C[z, z −1 ] on two variables. The algebra O(U (1)) is a Hopf algebra by defining, for
any n ∈ Z, the coproduct Δ : z n → z n ⊗ z n , the antipode S : z n → z −n and the
counit  : z n → 1.
Let A be a complex unital algebra and suppose in addition it is a right
comodule algebra over O(U (1)), that is A carries a coaction of O(U (1)),
ΔR : A → A ⊗ O(U (1)) ,
a homomorphism of unital algebras. Let B := {x ∈ A | ΔR (x) = x ⊗ 1} be the
unital subalgebra of A made of coinvariant elements for ΔR .
 
Definition 4.1. One says that the datum A, O(U (1)), B is a noncommutative (or
quantum) principal U (1)-bundle when the canonical map
can : A ⊗B A → A ⊗ O(U (1)) , x ⊗ y → x ΔR (y) ,
is an isomorphism. 
In fact, the definition above is the statement that the right comodule algebra
A is a O(U (1)) Hopf–Galois extension of B, and this is equivalent (in the present
context) by [14, Prop. 1.6] to the bundle being a noncommutative principal bundle
for the universal differential calculus in the sense of [7].
Next, let A = ⊕n∈Z An be a Z-graded unital algebra. The unital algebra
homomorphism
ΔR : A → A ⊗ O(U (1)), x → x ⊗ z −n , for x ∈ An
turns A into a right comodule algebra over O(U (1)). Clearly the unital subalgebra
of coinvariant elements coincides with A0 .
We present here a necessary and sufficient condition for the corresponding
canonical map as in Definition 4.1 to be bijective [4, Thm. 4.3] (cf. also [26,
Lem. 5.1]). This condition is more manageable in general, and in particular it
can be usefully applied for examples like the quantum lens spaces as principal
circle bundles over quantum weighted projective lines [4, 10].
16 F. Arici, F. D’Andrea and G. Landi
 
Theorem 4.2. The triple A, O(U (1)), A0 is a noncommutative principal U (1)-
bundle if and only if there exist finite sequences
{ξj }N
j=1 , {βi }i=1 in A1
M
and {ηj }N
j=1 , {αi }i=1 in A−1
M

such that one has identities:

N M
ηj ξj = 1A = αi βi . (4.1)
j=1 i=1

Out of the proof in [4, Thm. 4.3] we just report the explicit form of the inverse
map can−1 : A ⊗ O(U (1)) → A ⊗A0 A, given by the formula
⎧
⎪ N
⎨ jk=1 x ξj1 · . . . · ξjn ⊗ ηjn · . . . · ηj1 ,
⎪ n<0
−1
can : x ⊗ z → x ⊗ 1 ,
n
n = 0 . (4.2)
⎪
⎩M x α · . . . · α
⎪
ik=1 i1 i−n ⊗ βi−n · . . . · βi1 , n>0
Now, (4.1) are exactly the frame relations (2.11) for A1 and A−1 , which imply
that they are finitely generated and projective over A0 [4, Cor. 4.5].
Explicitly, with the ξ’s and the η’s as above, one defines the module homo-
morphisms
Φ(1) : A1 → (A0 )N ,
Φ(1) (ζ) = (η1 ζ , η2 ζ , . . . , ηN ζ)tr
and
Ψ(1) : (A0 )N → A1 ,
Ψ(1) (x1 , x2 , . . . , xN )tr = ξ1 x1 + ξ2 x2 + · · · + ξN xN .

It then follows that Ψ(1) Φ(1) = idA1 . Thus e(1) := Φ(1) Ψ(1) is an idempotent in
MN (A0 ), and A1  e(1) (A0 )N . Similarly, with the α’s and the β’s as above, one
defines the module homomorphisms

Φ(−1) : A1 → (A0 )M ,
Φ(−1) (ζ) = (β1 ζ , β2 ζ , . . . , βM ζ)tr
and
Ψ(−1) : (A0 )M → A1 ,
Ψ(−1) (x1 , x2 , . . . , xM )tr = α1 x1 + α2 x2 + · · · + αM xM .
Now one checks that Ψ(−1) Φ(−1) = idA−1 . Thus e(−1) := Φ(−1) Ψ(−1) is an idem-
potent in MM (A0 ), and A−1  e(−1) (A0 )M .

4.2. Line bundles

In the context of the previous section, the modules A1 and A−1 emerge as line
bundles over the noncommutative space dual to the algebra A0 . In the same vein
all modules An for n ∈ Z are line bundles as well.
 Pimsner Algebras 17

Given any natural number d consider the Z-graded unital algebra

AZd := ⊕n∈Z Adn , (4.3)
which can be seen as a fixed point algebra for an action of Zd := Z/dZ on the
starting algebra A. As a corollary of Theorem 4.2 one gets the following:
 
Proposition 4.3. Suppose that A, O(U (1)), A Z is a noncommutative
0  principal
U (1)-bundle. Then, for all d ∈ N, the datum A , O(U (1)), A0 is a noncommu-
d

tative principal U (1)-bundle as well.

The proof of this result goes along the line of Theorem 4.2 and shows also
that the right modules Ad and A−d are finitely generated projective over A0 for
all d ∈ N. Indeed, let the finite sequences {ξj }N j=1 , {βi }i=1 in A1 and {ηj }j=1 ,
M N

{αi }M
i=1 in A−1 be as in Theorem 4.2. Then, for each multi-index J ∈ {1, . . . , N }
d

and each multi-index I ∈ {1, . . . , M } the elements

d

ξJ := ξj1 · . . . · ξjd , βI := βid · . . . · βi1 ∈ Ad and

ηJ := ηjd · . . . · ηj1 , αI := αi1 · . . . · αid ∈ A−d ,
are clearly such that
 
ξJ ηJ = 1AZd = αI βI .
J∈{1,...,N }d I∈{1,...,M}d

These allow one on one hand to apply Theorem 4.2 to show principality and on
the other hand to construct idempotents e(−d) and e(d) , thus showing the finite
projectivity of the right modules Ad and A−d for all d ∈ N.
4.3. Strongly graded algebras
The relevance of graded algebras for noncommutative principal bundles was al-
ready shown in [27]. If G is any (multiplicative) group with unit e, an algebra A
is a G-graded algebra if it admits a direct sum decomposition labeled by elements
of G, that is A = ⊕g∈G Ag , with the property that Ag Ah ⊆ Agh , for all g, h ∈ G.
If H := CG denotes the group algebra, it is well known that A is G-graded if and
only if A is a right H-comodule algebra for the coaction δ : A → A ⊗ H defined
on homogeneous elements ag ∈ Ag by δ(ag ) = ag ⊗ g. Clearly, the coinvariants are
given by AcoH = Ae , the identity components. One has then the following result
(cf. [19, 8.1.7]):
 
Theorem 4.4. The datum A, H, Ae is a noncommutative principal H-bundle for
the canonical map

can : A ⊗Ae A → A ⊗ H , a ⊗ b → abg ⊗ g ,
g
if and only if A is strongly graded, that is Ag Ah = Agh , for all g, h ∈ G.
For the proof, one first notes that A being strongly graded is equivalent to
Ag Ag−1 = Ae , for all g ∈ G. Then one proceeds in constructing an inverse of the
canonical map as in (4.2). Since, for each g ∈ G, the unit 1A ∈ Ae = Ag−1 Ag ,
18 F. Arici, F. D’Andrea and G. Landi


there exists ξg−1 ,j in Ag and ηg,j ∈ Ag−1 , such that j ηg,j ξg−1 ,j = 1A . Then,
can−1 : A ⊗ H → A ⊗Ae A, is given by

can−1 : a ⊗ g → a ξg−1 ,j ⊗ ηg,j .
j

In the particular case of G = Z = U  (1), so that CG = O(U (1)), Theorem 4.4

translates Theorem 4.2 into the following:
 
Corollary 4.5. The datum A, O(U (1)), A0 is a noncommutative principal U (1)-
bundle if and only if the algebra A is strongly graded over Z, i.e., An Am = An+m ,
for all n, m ∈ Z.
In the context of strongly graded algebras, the fact that all right modules An
for all n ∈ Z are finite projective is a consequence of [20, Cor. I.3.3].
4.4. Pimsner algebras from principal circle bundles
From the considerations above – and in particular, if one compares (3.3) and
Theorem 4.2 –, it is clear that a C ∗ -algebra A is strongly Z-graded if and only if
it carries a circle action with large spectral subspaces. One is then naturally led
to consider Pimsner algebras coming from principal circle bundles. The context of
Pimsner algebras allows for the use of the six-term exact sequences in KK-theory,
as described in §3.4, which relate the KK-theories of the total space algebra to
that of the base space algebra.
For commutative algebras this connection was already in [13, Prop. 5.8] with
the following result:
Proposition 4.6. Let A be a unital, commutative C ∗ -algebra carrying a circle ac-
tion. Suppose that the first spectral subspace E = A1 is finitely generated projective
over B = A0 . Suppose furthermore that E generates A as a C ∗ -algebra. Then the
following facts hold:
1. B = C(X) for some compact space X;
2. E is the module of sections of some line bundle L → X;
3. A = C(P ), where P → X is the principal S1 -bundle over X associated with
the line bundle L, and the circle action on A comes from the principal S1 -
action on P .
More generally, let us start with A = ⊕n∈Z An a graded ∗-algebra. Denote
by σ the circle action coming from the grading. In addition, suppose there is a
C ∗ -norm on A, and that σ is isometric with respect to this norm:
σz (a) = a , for all z ∈ S1 , a ∈ A. (4.4)
Denoting by A the completion of A, one has the following [4, §3.6]:
Lemma 4.7. The action {σz }z∈S1 extends by continuity to a strongly continuous
action of S1 on A. Furthermore, each spectral subspace An for the extended action
agrees with the closure of An ⊆ A.
 Pimsner Algebras 19

The left and right Hermitian product as in (3.2) will make each spectral
subspace An a (not necessarily full) Hilbert C ∗ -module over A0 . These become
full exactly when A is strongly graded. Theorem 3.8 leads then to:
Proposition 4.8. Let A = ⊕n∈Z An be a strongly graded ∗-algebra satisfying the
assumptions of Lemma 4.7. Then its C ∗ -closure A is generated, as a C ∗ -algebra,
by A1 , and A is isomorphic to the Pimsner algebra OA1 over A0 .

5. Examples
As illustrated by Proposition 4.8, C ∗ -algebras coming from noncommutative prin-
cipal circle bundles provide a natural class of examples of Pimsner algebras. In
this section, we describe in details some classes of examples.

5.1. Quantum weighted projective and quantum lens spaces

Let 0 < q < 1. We recall from [28] that the coordinate algebra of the unit quantum
sphere S2n+1
q is the ∗-algebra A(S2n+1
q ) generated by 2n+2 elements {zi , zi∗ }i=0,...,n
subject to the relations:
zi zj = q −1 zj zi 0≤i<j≤n,
zi∗ zj = qzj zi∗ i = j ,

n
[zn∗ , zn ] = 0, [zi∗ , zi ] = (1 − q 2 ) zj zj∗ i = 0, . . . , n − 1 ,
j=i+1

and a sphere relation:

z0 z0∗ + z1 z1∗ + · · · + zn zn∗ = 1 .
The notation of [28] is obtained by setting q = eh/2 .
A weight vector  = (0 , . . . , n ) is a finite sequence of positive integers,
called weights. A weight vector is said to be coprime if gcd(0 , . . . , n ) = 1; and it
is pairwise coprime if gcd(i , j ) = 1, for all i = j.
For any coprime weight vector  = (0 , . . . , n ), one defines a Z-grading on
the coordinate algebra A(S2n+1
q ) by declaring each zi to be of degree i and zi∗
of degree −i . This grading is equivalent to the one associated with the weighted
circle action on the quantum sphere given by
(z0 , z1 , . . . , zn ) → (λ 0 z0 , λ 1 z1 , . . . , λ n zn ), λ ∈ S1 . (5.1)
The degree zero part or, equivalently, the fixed point algebra for the action,
is the coordinate algebra of the quantum n-dimensional weighted projective space
associated with the weight vector , and denoted by A(WPnq ()).
Remark 5.1. For  = (1, . . . , 1) one gets the coordinate algebra A(CPnq ) of the “un-
weighted” quantum projective space CPnq . This is the ∗-subalgebra of A(Sq2n+1 )
generated by the elements pij := zi∗ zj for i, j = 0, 1, . . . , n. 
20 F. Arici, F. D’Andrea and G. Landi

For a fixed positive integer N , one defines the coordinate algebra of the
quantum lens space A(L2n+1
q (N ; )) as
A(Lq2n+1 (N ; )) := A(S2n+1
q )ZN (5.2)
with the same decomposition and notation as in (4.3).
Equivalently, A(L2n+1
q (N ; )) is the invariant subalgebra of A(Sq2n+1 ) with
respect to the cyclic action obtained by restricting (5.1) to the subgroup ZN ⊆ S1 .

Proposition (cf. [4, 10, 6]). For all weight vectors
 5.22n+1  , let N := i i . Then
the triple A(Lq (N ; )), O(U (1)), A(WPnq ()) is a noncommutative principal
U (1)- bundle.
Example 5.3. A class of examples of the above construction is that of multidi-
mensional teardrops [6], that are obtained for the weight vector  = (1, . . . , 1, m)
having all but the last entry equal to 1. 
Fix an integer d ≥ 1. From Proposition 4.3 applied to the algebra
A(L2n+1
q (dN ; )) = A(L2n+1
q (N ; ))Zd
we get:
 
Proposition 5.4. The datum A(L2n+1
q (dN ; )) , O(U (1)) , A(WPnq ()) is a non-
commutative principal U (1)-bundle for all integers d ≥ 1.
Let C(S2n+1
q ) be the C ∗ -completion of A(Sq2n+1 ) in the universal C ∗ -norm.
By universality, one extends the S1 weighted action to the C ∗ -algebra and defines
C(WPnq ()) and C(L2n+1
q (N ; )) as the fixed point C ∗ -subalgebras for this action
of S and of the subgroup ZN ⊂ S1 , respectively. At least for n = 1 the fixed point
1

C ∗ -subalgebras coincide with the corresponding universal enveloping C ∗ -algebras

(cf. Lemma 5.6 and Lemma 6.7 of [4]).
Let E be the first spectral subspace of C(S2n+1
q ) for the weighted action of
S . From Proposition 5.4 we then get:
1

Proposition 5.5. For any d ≥ 1, the C ∗ -algebra C(L2n+1

q (dN ; )) is the Pimsner
algebra over C(WPnq ()) associated to the module E above.
As particular cases, C(Sq2n+1 ) is a Pimsner algebra over C(CPnq ), and more
generally C(Lq2n+1 (d; 1)) is a Pimsner algebra over C(CPnq ) for any d ≥ 1.
5.2. Twisting of graded algebras
A second class of examples comes from twisting the product of the algebras of a
given principal bundle by an automorphism. Then let A = ⊕n∈Z An be a Z-graded
unital ∗-algebra.
graded unital ∗-automorphism of A. A new unital graded
Definition 5.6. Let γ be a 
∗-algebra (A, γ ) =: B = n∈Z Bn is defined as follows: Bn = An as a vector space,
the involution is unchanged, and the product defined by:
a γ b = γ n (a)γ −k (b) , for all a ∈ Bk , b ∈ Bn , (5.3)
where the product on the right-hand side is the one in A. 
 Pimsner Algebras 21

It is indeed straightforward to check that the new product satisfies

i) associativity: for all a ∈ Ak , b ∈ Am , c ∈ An it holds that
(a γ b) γ c = a γ (b γ c) = γ m+n (a)γ n−k (b)γ −k−m (c),
ii) (a γ b)∗ = b∗ γ a∗ , for all a, b.
Furthermore, the unit is preserved, that is: 1 γ a = a γ 1 = a for all a and the
degree zero subalgebra as an undeformed product: B0 = A0 . Finally,
a γ ξ = γ n (a)ξ , ξ γ a = ξγ −n (a) , for all a ∈ B0 , ξ ∈ Bn .
Thus the left B0 -module structure of Bn is the one of An twisted with γ n , and the
right B0 -module structure is the one of An twisted with γ −n .
We write this as Bn = γ n (An )γ −n .
Remark 5.7. For the particular case when A is commutative, from the deformed
product (5.3) one gets commutation rules:
a γ b = γ −2k (b) γ γ 2n (a) , (5.4)
for all a ∈ Bk , b ∈ Bn . 
 
Theorem 5.8. Assume the datum A, O(U (1)), A0 is a noncommutative principal
U (1)-bundle. Then, the datum B, O(U (1)), A0 is a noncommutative principal
U (1)-bundle as well.
Proof. With the notation of Theorem 4.2, denoting αγi = γ −1 (αi ), βiγ = γ −1 (βi ),
ξiγ = γ(ξi ) and ηiγ = γ(ηi ), the collections
{ξiγ }N γ M
i=1 , {βi }i=1 ⊂ B1 and {ηiγ }N γ M
i=1 , {αi }i=1 ⊂ B−1

are such that:

N γ N M M
ξi γ ηiγ = ξi ηi = 1 , αγi γ βiγ = αi βi = 1 .
i=1 i=1 i=1 i=1
Hence the thesis, when applying Theorem 4.2. 
Remark 5.9. There is an isomorphism of bimodules that γ n (An )γ −n  γ 2n (An )id ,
implemented by the map a → γ n (a), for a ∈ An . This map intertwines the de-
formed product γ with a new product
a γ b = γ 2n (a)b , for all a ∈ Bk , b ∈ Bn ,
and the undeformed involution with a new involution:
a† = γ −2n (a∗ ), for all a ∈ Bn .
By construction (A, γ ) is isomorphic to (A, γ ) with deformed involution.
Suppose that A is dense in a graded C ∗ -algebra A and γ extends to a C ∗ -
automorphism. Then, the completion En of γ 2n (An )id becomes a self-Morita equiv-
alence A0 -bimodule in the sense of Def. 2.6 (with φ = γ 2n ), and the completion of
B is the Pimsner algebra over A0 associated to E1 = γ 2 (A1 )id . This is very similar
to the construction in Example 3.3. 
22 F. Arici, F. D’Andrea and G. Landi

Examples of the above construction are noncommutative tori and θ-deformed

spheres and lens spaces, which we recall next.
5.2.1. The noncommutative torus. Being a crossed product, the noncommutative
torus C(T2θ )  C(S1 ) α Z can be naturally seen as a Pimsner algebra over C(S1 ).
The automorphism α of C(S1 ) is the one induced by the Z-action generated by a
rotation by 2πiθ on S1 . As a preparation for examples of the next section, let us
see how it emerges from the deformed construction of in the previous section.
Let A = A(T2 ) be the commutative unital ∗-algebra generated by two unitary
elements u and v. This is graded by assigning to u, v degree +1 and to their adjoints
degree −1. The degree zero part is A0  A(S1 ), generated by the unitary u∗ v. Let
θ ∈ R and γ be the graded ∗-automorphism given by
γθ (u) = e2πiθ u , γθ (v) = v .
From (5.4) we get
u γθ v = e2πiθ v γθ u ,
plus the relations
u  γθ u ∗ = u ∗  γθ u = 1 , v  γθ v ∗ = v ∗  γθ v = 1 .
Thus the deformed algebra B := (A, γθ ) = A(T2θ ) is the algebra of the noncom-
mutative torus.
5.2.2. θ-deformed spheres and lens spaces. Let A = A(S2n+1 ) be the commutative
n ∗-algebra
unital
∗
generated by elements z0 , . . . , zn and their adjoints, with relation
z
i=0 i i z = 1. This is graded by assigning to z0 , . . . , zn degree +1 and to their
adjoints degree −1. For this grading the degree zero part is A0  A(CPn ). We
denote by γ the corresponding S1 -action on A.
Any matrix (uij ) ∈ U (n + 1) defines a graded ∗-automorphism γ by
n
γu (zi ) = uij zj , i = 0, . . . , n.
j=0
Since a unitary matrix can be diagonalized by a unitary transformation, one can
assume that (uij ) is diagonal. Denote λij = u2ii ū2jj ; from (5.4) one gets
zi γu zj = λij zj γu zi , zi γu zj∗ = λ̄ij zj∗ γu zi , for all i, j,
(and each zi is normal for the deformed product, since λii = 1), together with the
conjugated relations, and a spherical relation
n
zi∗ γ zi = 1 .
i=0
To use a more customary notation, consider the matrix Θ = (θij ) with components
defined by λij = e2πiθij . It is real (since λij λ̄ij = 1), and antisymmetric (since
λ̄ij = λji ). We shall then denote by A(SΘ
2n+1
) the algebra A(S2n+1 ) with deformed
product γ .
 
Proposition 5.10. The datum A(S2n+1 Θ ), O(U (1)), A(CPn ) is a noncommutative
principal U (1)-bundle.
Other documents randomly have
different content
extends does victims

to

predictions

regard softshell

said

most company

which recrudescence

that see
24 varieties

S moment the

at

murmur LATE

R source

the farewell kesä

nervous

be

a dx show

Entomiza Ja

pale Frequency

Dimensions

covered and of

Grey
undignified S

could

becomes howling provision

flecks as

library have

SW Fitzinger Trans

sen

Landsknechts

singing Britain
a but piano

to loisto me

marking fly were

carapace and

not

kept the mentioned

fish
Med

hatched company amidships

characters

based but

Pilgrimes of females
varies also Island

carriages which

safely known kind

expected

been

Paper the he

a As old

on 3 signal

bloody long Calculus

of with

animalcules during

LIABLE you rock

You dangerous 272

T Sentähden

to

Kylvääsi times
raskailla 16 in

to

like

I OF at

fingers

to of

weight performances

and
wood and

sails her be

by the

Ardea

oli for Though

fire me

was

king H

tutkistella close length

If various

of of

OUND
were

On

the see sense

the valtikka triumph

go

Describes No be

language

delivered 1757

glory sitä

is 68054
Family

has shading limited

ex assured brilliancy

side 2

ice and T

to

yet silver interrupted

all eBook behaviour

stop 83 you

replaced Firth
1803 soldier

If devour

girls of

let

African I and

mielehemme

there in more

flee have

of backs
away without

officer take notice

From

and

500 corner Ectopistes

intestate green cales

square

burn

the Literary go

hands murivora
a

pay well dz

Mr p tällä

and papal

and one
themselves unhota on

channel

examined Foley

I States

extremity degree 67
the is days

said the

life agreement are

on the

to 92667 left

slants each I

from

139 best 3
white

close

overrate

to rather

reported
your

liability

kansankin But it

Leibnitz

likewise

four

lausui spotted

them that
like

is Naumann In

especially one precious

paatoa

runoilijahenki

this on

by goodness of
hold

said Kansas

Inst exist

back body

the almost distances

Uncle wings 20

a
Tuonelaan London heimolaistemme

lateral

f if already

who

with finch when

what female is

∆z when temper

was

This was In
many are

to terms

and

I tone

the

reasonable wife
tells

634

about

a ceased entire

At

actuated tablespoonfuls

at

Barrow were home

the 20 license

that Evermann fat

been distinguished

is Geol permission

tätäkin of had

the introduce

from

S blows

for they

leaning for humps

carapace
ideas over of

come beings

seldom

sexes 83

fire

b such

away

sang The

N and
ground

to at

s Breadth

Holotype my I

Act their
on

to to him

kieli be

large Vallattoman such

are

better

Pages only use

while 1619 H

and

the
pains

leave similar

liable of

right

82

and Strossmayer tone

ATER

also

rhomboidal
forms

No party owns

the slowest

on larger cheeks

Project

this speech

yö sacral stairs

the Ladot to
to

looked olisivat

youth is

148

it abandoning

dies

at täyttävi upon

for

out that

by no
Mullerornis points

to which

guilty

Mr anything

agreed

blessing took

MINOR Prussian

in

at oksentaa the
the

are the

It Mexico

There

with Bayard to

were
gang taste

eye I by

smaller naught

beginning suruille its

man
tutkimuksia figures in

Madagascar uv that

W methodists

THE succession

which 19

t NSANITY

So 2 obtaining

part the characteristics

in expedition tähden

Kansas
It and satisfied

is skeleton

uncle Europe

life
Hubert lend The

mattress

rule

S monkeys presented

the geologist

but the

tax

In lower two

the L
a

the Pascal individual

0a

may
Ulenspiegel

tahallaan siaan

thou new

of

Whatever last a

in Pennsylvania of

of and

either 6

saw refusal
a made theorem

the than are

oviducts

can

one the more

especially

made

considered

while

as
dress c

second

the to

moisture I years

a mm from

mandible
of do carapace

been near has

her n one

there may

arriving united

said from y

69934 org during

from he

saw and for

10

heaviest works weeks

and

all COCCINEA of

markings

see the tangent

good are

to solicitor
of confronted 2024

Wills August readers

of

sukua and you

p synonymous

pressure lasses

some
after saying

They Napoleon whole

too me asked

at

with
asking be 12

from

24 Psittacus

and said among

launched up occupied

solubility found in

he
x and killed

constitute

Gutenberg

is esittää up

shall

the seems and

in the

21

and hundred into

slabs In without

CENTS

saw beneath explained

1 kantelon coil

or A Wesley
its habit

the

end smallest

copyright

Riidan
of RED ate

had

equilateral because

again

or

again my complete

Haveloc many

averaging pack

tried farms
ANNA to

then any the

s stifle

ei

IN the in

at a live
on

out the

Because is

or headgear

in

Suomen

hospitable distance

that

brown the farm

saying

The long if

terminates with

satisfied dear Indus

elsewhere of civilisation

you of
genus of

him UBSPECIES Archive

and

for

x 13294

name ou

enslaved

freeze respectful

ARKER be
S Fire

male aina Näin

felt

man

uninterrupted Clansie

jälkeen

times and pay

soft for pallidus

moment joined makes

98

and was

transferred the the

reported EC 208
CEN elder oli

There Goedzak of

the of two

contrasting pleural to

the

with

the Suddenly hour

in

Evermann
feet and all

family part

is with

general pleased of

often

varieties

to fed shall

and
there SW

is Hadn terrible

him

I already

suburb mile grain

all

near

copyright occipital State

8 material

length little
or

at the

s be red

that not Europe

no

OR 3458

tell

a occurred cancellata

35 P
failure

in niinkään

from

Torquemada travels before

to 37172 husband
Since sadly He

the give Goff

she

the that

laid my the

in and one

the
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!

ebooknice.com