Analysis Geometry And Topology Of Elliptic Operators Papers In Honor Of Krzysztof P Wojciechowski Matthias Lesch

jeknicrozh38 6 views 77 slides May 21, 2025
Slide 1
Slide 1 of 77
Slide 1
1
Slide 2
2
Slide 3
3
Slide 4
4
Slide 5
5
Slide 6
6
Slide 7
7
Slide 8
8
Slide 9
9
Slide 10
10
Slide 11
11
Slide 12
12
Slide 13
13
Slide 14
14
Slide 15
15
Slide 16
16
Slide 17
17
Slide 18
18
Slide 19
19
Slide 20
20
Slide 21
21
Slide 22
22
Slide 23
23
Slide 24
24
Slide 25
25
Slide 26
26
Slide 27
27
Slide 28
28
Slide 29
29
Slide 30
30
Slide 31
31
Slide 32
32
Slide 33
33
Slide 34
34
Slide 35
35
Slide 36
36
Slide 37
37
Slide 38
38
Slide 39
39
Slide 40
40
Slide 41
41
Slide 42
42
Slide 43
43
Slide 44
44
Slide 45
45
Slide 46
46
Slide 47
47
Slide 48
48
Slide 49
49
Slide 50
50
Slide 51
51
Slide 52
52
Slide 53
53
Slide 54
54
Slide 55
55
Slide 56
56
Slide 57
57
Slide 58
58
Slide 59
59
Slide 60
60
Slide 61
61
Slide 62
62
Slide 63
63
Slide 64
64
Slide 65
65
Slide 66
66
Slide 67
67
Slide 68
68
Slide 69
69
Slide 70
70
Slide 71
71
Slide 72
72
Slide 73
73
Slide 74
74
Slide 75
75
Slide 76
76
Slide 77
77

About This Presentation

Analysis Geometry And Topology Of Elliptic Operators Papers In Honor Of Krzysztof P Wojciechowski Matthias Lesch
Analysis Geometry And Topology Of Elliptic Operators Papers In Honor Of Krzysztof P Wojciechowski Matthias Lesch
Analysis Geometry And Topology Of Elliptic Operators Papers In Honor Of Kr...


Slide Content

Analysis Geometry And Topology Of Elliptic
Operators Papers In Honor Of Krzysztof P
Wojciechowski Matthias Lesch download
https://ebookbell.com/product/analysis-geometry-and-topology-of-
elliptic-operators-papers-in-honor-of-krzysztof-p-wojciechowski-
matthias-lesch-973550
Explore and download more ebooks at ebookbell.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.
Contemporary Mathematics 350 Noncompact Problems At The Intersection
Of Geometryanalysisand Topology 2004th Edition Abbas Bahri
https://ebookbell.com/product/contemporary-mathematics-350-noncompact-
problems-at-the-intersection-of-geometryanalysisand-topology-2004th-
edition-abbas-bahri-230629534
Perspectives In Analysis Geometry And Topology On The Occasion Of The
60th Birthday Of Oleg Viro 1st Edition Selman Akbulut Auth
https://ebookbell.com/product/perspectives-in-analysis-geometry-and-
topology-on-the-occasion-of-the-60th-birthday-of-oleg-viro-1st-
edition-selman-akbulut-auth-2489208
Sobolev Maps To The Circle From The Perspective Of Analysis Geometry
And Topology Haim Brezis
https://ebookbell.com/product/sobolev-maps-to-the-circle-from-the-
perspective-of-analysis-geometry-and-topology-haim-brezis-37579316
Data Science Foundations Geometry And Topology Of Complex Hierarchic
Systems And Big Data Analytics 1st Edition Fionn Murtagh
https://ebookbell.com/product/data-science-foundations-geometry-and-
topology-of-complex-hierarchic-systems-and-big-data-analytics-1st-
edition-fionn-murtagh-6762502

Topological Recursion And Its Influence In Analysis Geometry And
Topology Chiuchu Melissa Liu
https://ebookbell.com/product/topological-recursion-and-its-influence-
in-analysis-geometry-and-topology-chiuchu-melissa-liu-48137526
Winding Around The Winding Number In Topology Geometry And Analysis
1st Edition John Roe
https://ebookbell.com/product/winding-around-the-winding-number-in-
topology-geometry-and-analysis-1st-edition-john-roe-54792354
Analysis Geometry And Quantum Field Theory Clara L Aldana Maxim
Braverman
https://ebookbell.com/product/analysis-geometry-and-quantum-field-
theory-clara-l-aldana-maxim-braverman-6704136
Nonlinear Analysis Geometry And Applications Proceedings Of The Second
Nlagabirs Symposium Cap Skirring Senegal January 2530 2022 Diaraf Seck
https://ebookbell.com/product/nonlinear-analysis-geometry-and-
applications-proceedings-of-the-second-nlagabirs-symposium-cap-
skirring-senegal-january-2530-2022-diaraf-seck-46517662
Nonlinear Analysis Geometry And Applications Proceedings Of The Third
Nlagabirs Symposium Diaraf Seck
https://ebookbell.com/product/nonlinear-analysis-geometry-and-
applications-proceedings-of-the-third-nlagabirs-symposium-diaraf-
seck-57374890

/
Af
*•«.
,***"">*'
i&i$!*v
,«**^
J*
Analysis, Geometry and
Topology of Elliptic Operators
Editors Bernhelm BooR-Bavnbek
Slawomir Klimek
Matthias Lesch
Weiping Zhang

Analysis, Geometry and
Topology of Elliptic Operators

This page is intentionally left blank

Analysis, Geometry and
Topology of Elliptic Operators
Editors
Bemhelm BooR-Bavnbek
Roskilde University, Denmark
Slawomir Klimek
IUPUI, USA
Matthias Lesch
Universitat Bonn, Germany
Weiping Zhang
Nankai University, China
\jjp World Scientific
NEW JERSEY • LONDON • SINGAPORE • BEIJING • SHANGHAI • HONG KONG • TAIPEI • CHENNAI

Published by
World Scientific Publishing Co. Pte. Ltd.
5 Toh Tuck Link, Singapore 596224
USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601
UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library.
ANALYSIS, GEOMETRY AND TOPOLOGY OF ELLIPTIC OPERATORS
Copyright © 2006 by World Scientific Publishing Co. Pte. Ltd.
All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means,
electronic or mechanical, including photocopying, recording or any information storage and retrieval
system now known or to be invented, without written permission from the Publisher.
For photocopying of material in this volume, please pay a copying fee through the Copyright
Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to
photocopy is not required from the publisher.
ISBN 981-256-805-0
Printed in Singapore by World Scientific Printers (S) Pte Ltd

Contents
Preface ix
Part I. On the Mathematical Work of Krzysztof P.
Woj ciechowski
Selected Aspects of the Mathematical Work of Krzysztof P.
Wojciechowski 3
MATTHIAS LESCH
Gluing Formulae of Spectral Invariants and Cauchy Data Spaces 23
JINSUNG PARK
Part II. Topological Theories
The Behavior of the Analytic Index under Nontrivial Embedding 41
DAVID BLEECKER
Critical Points of Polynomials in Three Complex Variables 63
LIVIU I. NICOLAESCU
Chern-Weil Forms Associated with Superconnections 79
SYLVIE PAYCHA and SIMON SCOTT
Part III. Heat Kernel Calculations and Surgery
Non-Laplace Type Operators on Manifolds with Boundary 107
IVAN G. AVRAMIDI
Eta Invariants for Manifold with Boundary 141
XIANZHE DAI

vi Contents
Heat Kernels of the Sub-Laplacian and the Laplacian on Nilpotent
Lie Groups 173
KENRO FURUTANI
Remarks on Nonlocal Trace Expansion Coefficients 215
GERD GRUBB
An Anomaly Formula for L2-Analytic Torsions on Manifolds with
Boundary 235
XIAONAN MA and WEIPING ZHANG
Conformal Anomalies via Canonical Traces 263
SYLVIE PAYCHA and STEVEN ROSENBERG
Part IV. Noncommutative Geometry
An Analytic Approach to Spectral Flow in von Neumann Algebras 297
MOULAY-TAHAR BENAMEUR, ALAN L. CAREY, JOHN PHILLIPS,
ADAM RENNIE, FYODOR A. SUKOCHEV, and KRZYSZTOF P.
WOJCIECHOWSKI
Elliptic Operators on Infinite Graphs 353
JOZEF DODZIUK
A New Kind of Index Theorem 369
RONALD G. DOUGLAS
A Note on Noncommutative Holomorphic and Harmonic Functions
on the Unit Disk 383
SLAWOMIR KLIMEK
Star Products and Central Extensions 401
JOUKO MlCKELSSON
An Elementary Proof of the Homotopy Equivalence between the
Restricted General Linear Group and the Space of Fredholm
Operators 411
TlLMANN WURZBACHER
Part V. Theoretical Particle, String and Membrane
Physics, and Hamiltonian Dynamics
T-Duality for Non-Free Circle Actions
ULRICH BUNKE and THOMAS SCHICK
429

Contents vii
A New Spectral Cancellation in Quantum Gravity 467
GlAMPIERO ESPOSITO, GUGLIELMO FUCCI, ALEXANDER
KAMENSHCHIK, and KLAUS KIRSTEN
A Generalized Morse Index Theorem 493
CHAOFENG ZHU

This page is intentionally left blank

Preface
On May 20-22, 2005, a workshop was held at Roskilde University in
Denmark to honour Krzysztof P. Wojciechowski on his 50th birthday. This
volume collects the papers of that workshop.
The purpose of the volume is twofold. The more obvious one is to ac­
knowledge and honour Krzysztof Wojciechowski's contributions over the
last 20-25 years to the theory of elliptic operators. Lesch's write up goes
over many of Krzysztof's achievements, highlighting those insights that
were particularly influential in shaping the direction of the theory. It is
supplemented by Park's review of recent work pioneered by Wojciechowski.
As our second purpose, we also hope to offer younger researchers and grad­
uate students a snapshot of the current state of affairs. The proceedings
contain a mix of review and research papers, both reflecting on the past
and looking into the future. We obviously do not attempt to speak for the
whole, vast area of the theory of elliptic operators. Most papers in these
proceedings are, in one way or another, studying objects and techniques
that have interested Krzysztof: spectral invariants, cutting and pasting,
boundary value problems, heat kernels, and applications to topology, ge­
ometry and physics.
The modern theory of elliptic operators, or simply elliptic theory, has
been shaped by the Atiyah-Singer index theorem created some 40 years ago.
The Atiyah-Singer index theory expanded the scope of ellipticity to consider
relations with and applications to topology. The notion of index acquired a
dual personality, both analytical and topological. Consequently, wherever
topological invariants appear, one is now tempted to see if the analytical
aspects can be developed to interpret the invariant. In other words, analysts
are always on the lookout for topological or geometrical invariants hoping
to find operators behind them. Developments in topology are therefore of
special interest to elliptic theorists. Bleecker's paper revisits some aspects
of the so called embedding proof of the Atiyah-Singer index theorem. The
contributions of Bunke and Schick on T-duality and Nicolaescu's survey of
singularities of complex surfaces detail some topological theories of potential
interest to analysts and possible applications of analytical methods.
IX

x Preface
Heat kernel techniques are at the heart of another one of the several
proofs of the Atiyah-Singer index theorem. Different tools and techniques
have been developed and are continuing to be developed to understand
heat kernels and related spectral functions in a variety of situations. Two
problems stand out: to describe and compute variations of heat kernels
with respect to parameters and to calculate asymptotics of heat kernels -
like functions of operators. These have been the central technical issues for
much of Krzysztof Wojciechowski's work. As the scope of elliptic theory
increases, so is the variety of contexts for heat kernel calculations which
will undoubtedly occupy the interest of people in the future. The papers
of Avrimidi on heat kernels of non-Laplace operators, of Furutani on heat
kernels on nilpotent Lie groups, of Grubb on expansions of zeta-like func­
tions, and of Paycha and Rosenberg on canonical traces all fall into this
category.
Since the original papers on index theory, elliptic theory has continued
to develop. More areas of mathematics, other than topology, have started
influencing its progress. More and more objects of a similar nature to
index have been investigated. For one thing, index is a very simple spectral
invariant, and an important branch of elliptic theory looks at other spectral
invariants and their geometrical and topological significance. We need to
mention here some invariants that have particularly interested Krzysztof:
the eta invariant, spectral flow, analytic torsion and infinite dimensional
determinants. But there are many other invariants such as Seiberg-Witten
invariants and elliptic genus. We expect that this list is not complete and
that the future will bring more analytic invariants with topological and
geometrical applications.
In the spirit of topological surgery theory, a major effort was undertaken
to study elliptic operators and their spectral invariants using "cutting and
pasting". This naturally leads to the problem of how to set up an elliptic
theory on manifolds with boundary. This is the subject that Krzysztof has
devoted most of his mathematical efforts shaping. The papers of Dai on
eta invariants, of Ma and Zhang on L2-torsion, and Park's review of gluing
formulas for zeta determinants, as well as the contribution of Lesch, give
the state of the art for at least some of the questions in this area.
Beside topology, the operator theory and operator algebras have been
and will in the future be a driving force in the development of elliptic theory.
What started with the analysis of a single Fredholm operator on a mani­
fold, acquired greater depth and importance by considering whole spaces
of operators. With the invention of operator K-theory, elliptic theory is
evolving in a more abstract, algebraic fashion. Ellipticity is now defined
not just for (pseudo) differential operators and not just on manifolds with
or without boundary or even with corners. The proper context for the

Preface xi
study of ellipticity is noncommutative differential geometry. Noncommuta-
tive geometry aims to consider discrete spaces as well as noncommutative
objects on equal footing with topological spaces. Moreover, there is a du­
ality which runs even deeper with the modern interpretation of an elliptic
operator as a K-cycle over a C*-algebra. It seems quite possible, and even
likely, that such more algebraic trends will constitute the mainstream of
elliptic theory in the future. Operator-theoretic contributions to this vol­
ume include papers by Benameur et al. on spectral flow in von Neumann
algebras, by Douglas on a new kind of index theorem, by Klimek on a non­
commutative disk, by Mickelsson on star products and central extensions
and by Wurzbacher on homotopy calculations for some spaces of operators,
while Dodziuk explores elliptic theory in a discrete setting.
Theoretical particle, string and membrane physics have and will con­
tinue to provide major motivation for elliptic theory. As the world of el­
ementary particles continues to expand, one naturally suspects that the
so-called elementary particles are not so elementary any more. Some of
the current theories develop the idea that the basic structures of the uni­
verse are not point-like but rather stringy- or membrane-like. Such objects
would naturally live in dimensions higher than our 4 dimensional world.
To write down laws for such objects one is lead to modern global analy­
sis involving arbitrary dimensional manifolds and operators on them. Of
course new structures and new ideas also appear, such as supersymmetry,
conformal symmetry, mirror symmetry and anomalies. Many exciting new
mathematical questions arise. Several papers in this volume follow this line
of reserach: Bunke and Schick on very general mirror symmetry, Esposito
et al. on quantum gravity, Paycha and Rosenberg on conformal anom­
alies, Paycha and Scott on superconnections, Zhu on symplectic functional
analysis and Hamiltonian dynamics.
With its intricate theory, powerful methods and variety of applications,
the theory of elliptic operators should stay in the forefront of mathematics
for long years to come. The fact that this has been the case in the re­
cent past, is due in a nontrivial way to the work and insights of Krzysztof
Wojciechowski.
Acknowledgements. We thank the authors for their contributions, the
Non-commutative Geometry Center and the PDE Network of The Danish
Science Research Council for financial support, Roskilde University for hos­
pitality, and Benjamin Himpel (Bonn University) and the Staff of World
Scientific Publishing Company for encouragement and help in preparing
this volume.
The Editors

Part I
On the Mathematical Work of
Krzysztof P. Wojciechowski

This page is intentionally left blank

Analysis, Geometry and Topology
of Elliptic Operators, pp. 3-22
© 2006 World Scientific Publishing Co.
SELECTED ASPECTS OF THE MATHEMATICAL WORK
OF KRZYSZTOF P. WOJCIECHOWSKI
MATTHIAS LESCH
Mathematisches Institut
Beringstr. 1
53115 Bonn, Germany
lesch@math. uni-bonn. de
Dedicated to Krzysztof P. Wojciechowski on his 50th birthday
To honor and to please our friend Krzysztof P. Wojciechowski I will review the
milestones of his mathematical work. This will at the same time be a tour of
Analysis and Geometry of Boundary Value Problems. Starting in the 80s I will
discuss the spectral flow and the general linear conjugation problem, the Calderon
projector and the topology of space of elliptic boundary problems. The theme of
the 90s is the eta invariant. The paper with Douglas was fundamental for estab­
lishing spectral invariants for manifolds with boundary and for the investigation of
the behavior of spectral invariants under analytic surgery. This was so influential
that many different proofs of the gluing formula for the eta-invariant were pub­
lished. Finally turning to the new millennium we will look at the zeta-determinant.
Compared to eta this is a much more rigid spectral invariant which is technically
challenging.
2000 Mathematics Subject Classification. Primary 58J32; Secondary 58J30, 58J28,
58J52, 35P99
1. Introduction
1.1. The framework and the problem
To begin with let us describe in general terms the problems to which
Krzysztof P. Wojciechowski has contributed so much in the last 25 years.
Let X be a compact smooth Riemannian manifold with boundary X =
dX. Furthermore, let E, F be hermitian vector bundles over X and let
D : r°°(X, E) —^ T°°(X, F) (1)
Supported partially by Sonderforschungsbereich/Transregio 12 "Symmetries and Uni­
versality in Mesoscopic Systems" (Bochum-Duisburg/Essen-Koln-Warszawa)
3

4 Matthias Lesch
be an elliptic differential operator: T°°(X, E) denotes the spaces of smooth
sections of the bundle E.
In this situation some natural questions occur:
1. What are appropriate boundary conditions for D on X?
This question is absolutely fundamental since without imposing bound­
ary conditions we cannot expect D to have any reasonable spectral theory.
A boundary condition is given by a pseudo-differential operator
P:r°° (£,£)—*r°°(£,£) (2)
of order 0.a The realization Dp of the boundary condition given by P is
the differential expression D acting on the domain
dom(Dp) := {u e L\{X,E) | P(«|E) = 0}. (3)
Since D is elliptic what one should expect naturally for P to be "ap­
propriate" is that elliptic regularity holds. That is if Du £ L^(X,E) h is of
Sobolev order s > 0 and if P(w|S) = 0 then u £ L2s+d{X, E) is already of
Sobolev order s + d, where d denotes the order of D.
2. What is the structure of the space of all (nice) boundary conditions
and how do spectral invariants of Dp depend on the boundary condition?
These problems are the Leitfaden of Krzysztof P. Wojciechowski's work.
If we are given a realization Dp of a nice boundary value problem we can do
spectral theory and study the basic spectral invariants of Dp. We will see
that the question in the headline leads to interesting and delicate analytical
problems. Let us specify the kind of spectral invariants we mean here.
The most basic spectral invariant of the Fredholm operator Dp is its
index
ind Dp = dim ker Dp — dim coker Dp. (4)
More rigid (and analytically more demanding) spectral invariants are
derived from the heat trace
tr(e-"*)= £ e-"', (5)
Aespec£>p\{0}
aOne could think of more general definitely nonlocal boundary operators, but in this
paper we will content ourselves to pseudo-differential boundary conditions.
bWe denote the space of sections of E which are of Sobolev order s by L^(X, E).

Aspects of the mathematical work of Krzysztof P. Wojciechowski 5
where Dp is now assumed to be self-adjoint, via Mellin transforms. The
most important examples are the 77-invariant
1 r°°
v{Dp) = [fik^l *("1)/2tr(D*c""*)*
and the (-determinant
logdetc (£>) = _ —
as
1 f°°
s=0
s=0
(6)
(7)
The existence of these invariants is highly non-trivial since it depends
on the meromorphic continuation of the right hand side of (6) and (7).
In the following sense the index is the least rigid and the (-determinant
is the most rigid of these three invariants. In order not to get into too much
technicalities assume for the moment that J9(s)a<s<6 is a smoothly varying
family of elliptic operators on a closed manifold.0
The index is insensitive to small perturbations of the operator. Hence
indD(s) will not depend on s at all. The variation of the 77-invariant is
easy to understand. First of all the reduced 77-invariant
v{D(s)) = hdimkei D(s) + T?(D(S)) (8)
has only integer jumps and the total number of jumps equals the spec­
tral flow of the family D(s) over the interval [a,b]. The variation of
77(Z?(s))modZ is local in the sense that ^(?7(.D(s)modZ) is the integral
of a density which is a local expression in terms of the coefficients of the
operator and its derivatives, cf. Gilkey [12], Sec. 1.13. The variation of the
(-determinant is more complicated and depends on global data.
It is therefore most natural that the early work of Krzysztof P. Woj­
ciechowski dealt with problems related to the index. The paper [11] with
Douglas is a landmark since it is the starting point of a whole decade seeing
a lot of papers focusing on the 77-invariant and the (-determinant. I was
told that it came as an almost unbelievable surprise for the mathematical
community when 77-function and 77-invariant for Dirac operators on com­
pact manifolds with boundary were established in [11], since until then the
77-invariant was only established for closed manifolds and considered solely
as a natural correction term associated to index problems on manifolds with
boundary and living exclusively on the boundary.
cHere smoothly varying means that all coefficients depend smoothly on the parameter.

6 Matthias Lesch
The paper [11] already contained one of the major analytical tools which
has been refined and exploited ever since: the adiabatic method (see Section
3.1 below).
There is a variant of the problems mentioned above which I would like
to point out. Suppose that M is a closed manifold which is partitioned by
a separating hypersurface S C M. I. e. there are compact manifolds with
boundary Y, X such that d
M = Y U2 X. (9)
After having chosen appropriate boundary conditions PX,PY for D
on X, Y we have three versions of D: DPX,DPY and the essentially self-
adjoint operator D on the closed manifold M. In a sense we have nD =
Dpy UDpx" and it is natural to ask how the spectral invariants of D, Dpx,
and DpY are related. Krzysztof P. Wojciechowski and his collaborators
have provided us with spectacular results on this problem.
1.2. The basic framework of boundary value problems for
Dirac type operators
Let us be a bit more specific now and describe the basic set-up of boundary
value problems for Dirac type operators as we understand it today.
Let X and D be as before. We assume that D is an operator of Dirac
type. That is in local coordinates
d2
D2 = -gtj(x)ITa,nkE—— h lower order terms. (10)
This is the most general notion of Dirac operator. The leading symbol of
D
<jD(x,df)=i[DJ]x, feC°°(X), (11)
induces a Clifford module structure on E. That is we put for v G TXX e
c{v):=-<jD{x,vb). (12)
i
Then c(v)2 = —g(v,v) and hence by the universal property of the Clif­
ford algebra, c extends to a section of the bundle Hom(CI(TX,g), EndE)
This is a situation which is typical for surgery theory in which we would have E =
Sk X Sl, Y = Sk x Dl, where Sk denotes the unit sphere in mk+1 and Dk denotes the
unit disc in Rfc.
eThe Riemannian metric provides us with the "musical" isomorphisms \> : TXM —• T*M
and jj = b_1.

Aspects of the mathematical work of Krzysztof P. Wojciechowski 7
of algebra-homomorphisms between the bundle of Clifford-algebras
Cl(TX, g) and the endomorphism bundle End E. This gives E the structure
of a Clifford-module.
If we choose a Riemannian connection V on £ we can form the Dirac
operator Z)v on E which is locally given by
In the terminology of Boofi-Bavnbek and Wojciechowski [7] such oper­
ators are called "generalized Dirac operators". The operators Dv and D
obviously have the same leading symbol, hence
D = Dv + V (14)
with VeT°°(X, End E).
Next we have to take the boundary of X into account. We fix a dif-
feomorphism from a collar U of the boundary onto N := [0,e) x X. Then
we may choose a unitary transformation 4> from L2(U, E) onto the product
Hilbert space L2([0,e),L2(E,F)). The operator $D$_1 which, by slight
abuse of notation, will again be denoted by D then takes the form
D\N = J(-^ + B(x))+V(x) (15)
where J 6 T°°(T,,EndE) is a unitary reflection (J2 = -I, J* = -J),
V € Coo([0,e),roo(£,End£)) and {B{x))0<x<e is a smooth family of first
order formally self-adjoint differential operators on the closed manifold E
(called the tangential operator).
Replacing B(x) by B(x) + J~1V(x) be obtain alternatively
D\N = J(^ + B(x)) (16)
at the expense that now B(x) has only self-adjoint leading symbol.
We emphasize that J is independent of x and that (15) holds for all
operators of Dirac type (Briining and Lesch [9], Lemma 1.1). The repre­
sentation (15) of a generalized Dirac operator is crucial for the geometry of
their boundary value problems. In the existing literature, one could some­
times get the impression that for (15) to hold one needs that D is the Dirac
operator of a Riemannian connection on E as in (13) or even a compatible
Dirac operator.
Furthermore, for many results to be presented below only the following
properties of D will be needed:

8 Matthias Lesch
(1) D is first order formally self-adjoint elliptic,
(2) D has the form (15) near the boundary,
(3) D has the unique continuation property.
Properties of Dirac operators which are related to Clifford algebras will
more or less play no role.
D is formally self-adjoint. That is for sections f,g £ T°°(X, E) we have
Green's formula
(Df, g) - (/, Dg) = - J (/, g)E. dvol(z). (17)
In order to obtain an unbounded self-adjoint operator in L2(X, E) we have
to impose appropriate boundary conditions.
For a pseudo-differential orthogonal projectionf
P:L2(£,£)-^L2(£,£)
we define Dp to be the differential expression D acting on the domain (3).
Definition 1.1. 1. In the notation of (15) we abbreviate BQ := B(0)
and denote by P+(BQ) the orthogonal projection onto the positive spectral
subspace of Bo- This is a pseudo-differential operator of order 0. Its
principal symbol is denoted by <7p+(Bo).
2. The boundary condition defined by P is called well-posed if for
each £ G T*T, \ {0} the principal symbol <7p(£) of P maps range0>+(BO)(£)
bijectively onto range <Tp(£).
This is Seeley's definition of well-posedness [22]. If P is well-posed then
Dp has nice properties.
Proposition 1.1. Let P be well-posed. Then Dp is a Fredholm operator
with compact resolvent. Moreover it is regular in the sense that if a dis­
tributional section u of E satisfies Du £ L2(X,E) and P(u|£) = 0 then
ueL2a+1(X,E),s>0.
It turns out that for Dirac type operators this notion of regularity al­
ready characterizes the class of well-posed boundary conditions as was
shown by Briining and Lesch [9].
So far we have basically presented the status of affairs from the point
of view of classical elliptic theory.
fThis is not a big loss of generality. It can be shown that if the boundary operator
has closed range then the boundary condition may be represented by an orthogonal
projection.

Aspects of the mathematical work of Krzysztof P. Wojciechowski 9
2. The early work on spectral flow and the general linear
conjugation problem
[25, 4, 5]
The early papers [25, 4, 5] (in part with Boofi) on the general linear
conjugation problem are fundamental for our todays understanding of the
structure of boundary value problems of Dirac type operators. The linear
conjugation problem is the natural generalization of the classical Riemann
Hilbert problem to elliptic operators (cf. [7], Sec. 26).
Consider a partitioned manifold M — Y Us X as in (9) and let
D =
0 £>_
D+ 0
(18)
be a super-symmetric Dirac operator. That is the bundle E = E+ © E~~ is
Z^-graded and D is odd with respect to this grading.
In a collar N = (—e, e) x £ of £ we write D in the form (16) and hence
we get for D+
D+=a{-^ + B(x)) (19)
where a £ T°°(E,~Hom(E+, E~)) is unitary (and independent of x) and
(B(x))-e<x<£ is a smooth family of elliptic differential operators with self-
adjoint leading symbol.
Furthermore, let $ € r°°(S, Aut(E)) be a unitary bundle automor­
phism8 of E which is even with respect to the grading. Multiplication
by $ is a pseudo-differential operator of order 0 which we denote by
the same letter. We assume that $ commutes with the leading symbol
of B(x). As a consequence the operator $5$-1 — B is of order 0 and
$P+(B(x)) — P+(B(x))§ is of order —1 and thus acts as a compact oper­
ator on L2(£,£+).
We introduce a local boundary value problem by letting the differential
expression D+ act on
dom(Dt) :={{u1,u2)£Ll{Y,E+)®L21(X,E+) | Ul|£ = $u2|£}. (20)
From Green's formula (17) on derives
(D%y = Da_*a" (21)
gKrzysztof P. Wojciechowski originally treated more generally $'s which cover a diffeo-
morphism of E. Then multiplication by $ is a Fourier integral operator.

10 Matthias Lesch
and thus
rj4>©o-*<r*
D* 0
(22)
_£)* 0
One can show that £)*®°'*<J jg a realization of a local elliptic boundary
value problem. Introducing the Cauchy data spaces
N(D+, X) := {u|£ I u £ L?/2(E, £+), D+u = 0} (23)
we find
ind D% = dim((<i>N(D+ ,X))n N(D-, Y))
-dim((J$*J*N(D-,X))nN(D„,Y)). '
Before we can state the main result on the linear conjugation problem
we need to elaborate a bit more on the Cauchy data spaces.
2.1. Calderon projector and the smooth self-adjoint
Grassmannian
Definition 2.1. The (orthogonalized) Calderon projector C(D,X) is the
orthogonal projection onto the Cauchy data space N(D,X).
There is a little subtlety here. The natural construction of the Calderon
projector via the invertible double (cf. [7], Sec. 12) gives a pseudo-
differential (in general non-orthogonal) projection onto the Cauchy data
space. It is an orthogonal projection if D is in product form (cf. (36)
below) near the boundary. Of course, for any projection there is an orthog­
onal projection with the same image and using the results of Seeley [24] it
follows that
Proposition 2.1. The orthogonalized Calderon projector C(D,X) is a
pseudo-differential operator of order 0. Its leading symbol coincides with
the leading symbol aP+(B0) °fP+{Bo)-
The pseudo-differential properties of the Calderon projector had been
developed by Calderon [10] and Seeley [23]. In [6] we will show that the or­
thogonalized Calderon projector can be constructed from a natural bound­
ary value problem on the disconnected double XJJX. For brevity we will
address the orthogonalized Calderon projector just as Calderon projector.
The in my view most important observation of the papers [4, 5] is the
fact that the Cauchy data spaces are Lagrangian. To explain this note that
on the Hilbert space L2(E,E) we have the symplectic form
w(f,g):=-(Jf,g). (25)

Aspects of the mathematical work of Krzysztof P. Wojciechowski 11
This claim may be somewhat bewildering since L2(E, E) is firstly a complex
vector space and secondly infinite-dimensional. Nevertheless, w is a non-
degenerate skew-adjoint sesqui-linear form and it turns out that it makes
perfectly sense to talk about Lagrangians, symplectic reductions, Maslov
indices etc. The only difference is that, due to the infinite-dimensionality,
Fredholm conditions come into play. This is a fascinating story and an
elaboration would definitely need more space. For some basics cf. Kirk and
Lesch [14], Sec. 6. We state explicitly what Lagrangians are in L2(X,E).
Lemma 2.1. A subspace L C L2(X,E) is Lagrangian if and only if L1- =
J(L).
The following is basically a consequence of Green's formula (17).
Proposition 2.2. A realization Dp of a boundary condition is a symmet­
ric operator if and only if range P is an isotropic subspace of L2(X,E).
Moreover, if P is well-posed then Dp is self-adjoint if and only if range P
is Lagrangian.
The following Theorem was proved first in [4]:
Theorem 2.1. Let X be a compact Riemannian manifold with boundary
and let D be a Dirac type operator on X. Then the Cauchy data space of
N(D,X) is a Lagrangian subspace of L2(X,E) with respect to the symplec­
tic structure (25) induced by Green's form.
This theorem is not only beautiful. It is of fundamental importance.
We are now able to describe spaces of well-posed boundary value problems
as Grassmannian spaces:
Definition 2.2. Let V be the space of all pseudo-differential orthogonal
projections acting on L2(Y,,E).
The pseudo-differential Grassmannian Gri(Bo) is the space of P € V
such that
P - P+(B0) is of order - 1. (26)
The space of P G V such that the difference P — P+(B0) is smoothing is
denoted by Groo(Bo)-
Finally the self-adjoint (smooth) pseudo-differential Grassmannian
Gr*(i?o) is the space of P £ Grp(_B0),p £ {l,°o}, whose image is addi­
tionally Lagrangian.

12 Matthias Lesch
Since P+(B) and C(D,X) have the same leading symbol (26) may be
replaced by
P - C(D, X) is of order - 1. (27)
Hence P and C(D, X) also have the same leading symbol and thus it is
obvious from the Definition 1.1 that the boundary condition given by P is
well-posed.
Furthermore, since the difference of any two elements P,QG Gri (Bo)
is compact they form a Fredholm pair, that is
PQ : range Q —> range P (28)
is a Fredholm operator. The index of this Fredholm operator is denoted by
ind(P,Q). We have
ind(P, Q) = dim(kerP n range Q) - dim(rangeP n kerQ). (29)
2.2. The main theorem on the general linear conjugation
problem
We are now in a position to state the main result on the general linear
conjugation problem.
Theorem 2.2. The index of the linear conjugation problem (20) is given
by
indZ3* = md(I-C(£>+,Y),$C(D+,X)$-1)
= indD + md(C(D+,X) - &C(D+, Y))
= indD + ind(P+(P0) - $P_(P0))-
There would be much more to say. This index theorem is related to a
lot. It is a generalization of the classical Riemann Hilbert problem on the
complex projective line. It is related to the spectral flow and to the index
of generalized Toeplitz operators.
I will not go into that. But let me say that the papers [25, 4, 5] contain
much more. They provide a comprehensive presentation of the spectral flow
and its topological meaning, Fredholm pairs, and the construction of the
Calderon projector. Also it is proved that P+ (Bo) is a pseudo-differential
operator.

Aspects of the mathematical work of Krzysztof P. Wojciechowski 13
3. The ^-invariant
[11, 26, 27, 16, 28]
Let us start with some general remarks on rj- and ^-functions. Let T be
an unbounded self-adjoint operator in the Hilbert space H. Assume that
T has compact resolvent such that the spectrum of T consists of a sequence
of eigenvalues
l^i I < I ^21 < • • • (repeated according to their finite multiplicity)
with |An| —> oo. If An satisfies a growth condition
\Xn\>Cna, (30)
for some a > 0 then we can form the holomorphic functions
JL |ArssignA (31)
A£specT\{0}
and
= ti(T\T\-s~1), Res>i:
C(T;a):= £ X~s
AespecT\{0}
= tr(T-s), Res>i.
v a
(32)
If T is non-negative then (,(T;s) is also a Mellin transform similar to the
first equality in (31)
c(T;s) = rWo *s"ltr(e~t7*)d*- (33>
For general T the function £(T; s) can still be expressed in terms of Mellin
transforms using the formula
C(T; *) = \ (C(T2; 5/2) + V(T; s)) + e~^\ «(r2; s/2) - ^(T; s)). (34)
Up to a technical point the existence of a short time asymptotic ex­
pansion of tr(Te~tT ),tr(e_tT ) and the meromorphic continuation of the
functions £(T;s),rj(T;s) is equivalent (cf. Briining and Lesch [8], Lemma
2.2, for the precise statement).

14 Matthias Lesch
If T is an elliptic operator on a closed manifold then it follows from the
celebrated work of Seeley [24] that r)(T; s), ((T; s) extend meromorphically
to C with a precise description of the location of the poles and their residues.
If rj(T; s) is meromorphic at least in a half plane containing 0 one defines
the r\ -invariant of T as
(35)
= constant term in the Laurent expansion at 0
=:V(T;0).
In many situations one can even show that r)(T\ s) is regular at 0. The
77-invariant was introduced in the celebrated work of Atiyah, Patodi and
Singer [1, 2, 3] as a boundary correction term in an index formula for
manifolds with boundary.
We return to manifolds with boundary and consider again a compact
Riemannian manifold X with boundary dX = S and a formally self-adjoint
operator of Dirac type acting on the hermitian vector bundle E.
Prom now on we assume that D is in product form near the boundary.
That is in the collar N = [0, e) x £ of the boundary D takes the form
D\N = J^+B) (36)
with J, B as in (15) and such that B is independent of x. The formal
self-adjointness of D and B then implies
JB + BJ = 0. (37)
The next Theorem guarantees the existence of the /^-invariant and the
(^-determinant on the smooth self-adjoint Grassmannian.
Theorem 3.1. [28] For P e Gr^(B) the functions T](DP;S),C{DP\S) ex­
tend meromorphically to a half-plane containing 0 with poles of order at
most 1. Furthermore, 0 is not a pole and ((DP;0) is independent of P.
Let me say a few words about the strategy of proof. As pointed out be­
fore we have to prove short time asymptotic expansions for ti(DPe~tDp)
and tx{e~tDp). Duhamel's principle11 allows to separate the interior con­
tributions and the contributions coming from the boundary. Namely, let
hA big word for something very simple: the method of variation of the constant for first
order inhomogeneous ordinary differential equations.

Aspects of the mathematical work of Krzysztof P. Wojciechowski 15
(f 6 Co°([0, e)) be a cut-off function with <p = 1 near 0. Extend ip by 0 to
a smooth function on X.
Let D be any elliptic extension of D to a closed manifold1 and let Dpto
be the model operator J(-^ + B) on the cylinder [0, oo) x £ with boundary
condition P at {0} x S. Then
tv(DPe-tDh =tT(ipDP0e-tDr.°) +
~ ~2 (38)
tr((l- <p)De-tD)+0(tK), t^0+
for any K > 0.
By local elliptic analysis the second term in (38) has a short time asymp­
totic expansion [12], Lemma 1.9.1. So one is reduced to the treatment of the
model operator Dp,o- For the Atiyah-Patodi-Singer problem P = P+{B)
—tD2
there are explicit formulas for e p+'° from which the asymptotic expan­
sion can be derived using classical results on special functions. Finally, for
P e Gr^0(5) the operator Dpfi can be treated as a perturbation of the
APS operator DP+to [28].
A completely different approach by Grubb [13] leads to the generaliza­
tion of Theorem 3.1 to all well-posed boundary value problems.
3.1. The adiabatic limit
Let us explain the result of [11, 26, 27] on the adiabatic limit of the -q-
invariant. We start with a partitioned manifold M = Y Us X. Then we
stretch the neck by putting
XR = [0,R]xZU{R}xSX,
YR = [-R,0}xEU{_R}xEY,
MR = YR U{0}xS XR.
Denote by rj(D,MR) the reduced 77-invariant of D on MR and by
rj{Dp,XR) the reduced ^-invariant of Dp on XR.
Theorem 3.2. We have
lim fj(D,MR) = lim fj{DI_P {B),YR)
H—KX> H—>oo /QQ
+ lim 7J(DP+{B),XR) mod Z.
'The existence of such a D is not essential for the following result but it simplifies the
exposition. For Dirac type operators we can choose D to be the invertible double.

16 Matthias Lesch
We should be a bit more specific about the meaning of P+ (B) here. The
positive spectral projection of B is Lagrangian if and only if B is invertible.
If B is not invertible then one has to fix a Lagrangian subspace of the null
space of B. So whenever a Lagrangian is needed we choose P+(B) such
that
i(0,oo)
(B) < P+(B) < (B).
That this is possible follows from the Cobordism Theorem (cf. [11] or Lesch
and Wojciechowski [16]).
In [11] it was shown that the 77-invariant makes sense for generalized
Atiyah-Patodi-Singer boundary conditions, i.e. for -Dp+(B). Moreover, it
was shown that limfl_>oo fJ{^P+(B)^R) exists. The limit can be interpreted
as the 77-invariant of the operator D on the manifold with cylindrical ends
XOQ. The full strength of Theorem 3.2 was proved in [26, 27]. In fact the
(mod Z reductions) of the ingredients of formula (39) do not depend on Z
as was observed by W. Miiller [18]. In this way we obtain the gluing for­
mula for the 77-invariant for the boundary condition P+(B). The following
generalization to all P e Gr^, (B) is worked out in [28].
Theorem 3.3. Let M = Y Us X be a partitioned manifold and let D be
a Dirac type operator which is in product form in a collar of S. Then for
f/(D, M) = fj(DP, X) + fiiDj-p, Y) mod Z. (40)
There is even a formula ii I — P is replaced by a general Q € Gr^ (—B).
This is an extension of a formula for the variation of the 77-invariant under
a change of boundary condition from [16], cf. also Theorem 4.1 below.
Because of its importance let us look briefly at the method of proof.
The first observation is that the heat kernel of the model operator D =
J(-^ + B) on the cylinder KxS is explicitly known since D2 is just a
direct sum of one—dimensional Laplacians — -^ + b2. Let £cyi(t; x, y) be this
cylinder heat kernel. Furthermore, denote by £n{t;x,y) the heat kernel of
D on the stretched manifold MR.
Next one chooses ^-dependent cut-off functions (j>j,R,tpj,R,j — 1,2, as
follows:
JO if \x\ < 3R/7,
1 1 if |x| > 4R/7,
-4>l,R = 1-^2, R-

Aspects of the mathematical work of Krzysztof P. Wojciechowski 17
Finally, choose 4>J,R such that 4>J^RI\>^R = V'j.fi- Then paste the heat kernel
SR on MR and the cylinder heat kernel to obtain the kernel
Qii{t; x, y) = <t>\,R{x)ecy\{t; x, y)jji,R{y) + 4>2,R{x)£R(t; x, y)i/>2,R(y). (41)
Then Duhamel's principle yields
£R(t)=QR(t)+SR#CR(t), (42)
where # is a convolution and CR is an error term.
It seems that not much is gained yet. The point is that Douglas and
Wojciechowski [11] could show that in the adiabatic limit the error term is
negligible in the following sense:
Theorem 3.4. There are estimates
\£R{t;x,y)\ < Clt-dimX/2ec2te-c3d2(a;,!/)A)
WR^CR^t-x^WKc^e-^/'
with ci,C2, C3 independent of R.
Note that this result is much more than e.g. (38). For the rj- and
^-determinant the full heat semigroup contributes. It is astonishing that
nevertheless in the adiabatic limit the full integrals from 0 to 00 in (6) and
(7) split into contributions coming from the cylinder and from the interior
of the manifold.
4. The relative ^-invariant and the relative (^-determinant
[16, 21]
Recall from Theorem 3.1 that for P 6 Gr^(5) the C-function ((DP; s)
is regular at 0. One puts
detc^:=(eXP(-C'(i^0))< °*^*. (43)
[O, 0e spec Dp.
In view of (34) and Theorem 3.1 a straightforward calculation shows for
Dp invertible
dekDP =exP(z|(C(^;0)-r?(DF)) - ±?(D%;0)). (44)
We emphasize that the regularity of T](Dp;s) and ((DP;s) at s = 0 is
essential for (44) to hold. (44) shows that the 77-invariant is related to the
phase of the ^-determinant and that in general
(detcD)2^detc(£)2).

18 Matthias Lesch
The natural question which arises at this point is
Problem 4.1. How does detc(DP) depend on P e Gr^(B)?
The answer to this problem has a long history. Since the only joint
paper of Wojciechowski and myself deals with an aspect of the problem I
take the liberty to add a few personal comments. In 1992 I was a Postdoc
at University Augsburg. At that time the paper [11] had just appeared
and the gluing formula for the ^-invariant was in the air. Still much of
our todays understanding of spectral invariants for Dirac type operators on
manifolds with boundary was still in its infancy. When Gilkey visited he
posed a special case of the Problem 4.1. If the tangential operator is not
invertible there is no canonical Atiyah-Patodi-Singer boundary condition
for D. The positive spectral projection of B is not in Gr^c(B). Rather one
has to choose a Lagrangian subspace V C ker B and put
Pv •= l(o,oc)(B)+Uv,
where Hy denotes the orthogonal projection onto V. Then Py e Gr^(B).
The boundary condition given by Py is called a generalized Atiyah-Patodi-
Singer boundary condition. Gilkey asked how the eta-invariant depends on
V.
I did some explicit calculations on a cylinder which let me guess the cor­
rect formula. However, I did not know how to prove it in general. Somewhat
later Gilkey sent me a little note of Krzysztof dealing with the same prob­
lem. He urged us to work together. I was just a young postdoc and I felt
honored that Krzysztof, whose papers I already admired, quickly agreed.
Except writing papers with my supervisor this was my first mathematical
collaboration. It was done completely by fax and email; Krzysztof and I
met for the first time more than a year after the paper had been finished.
In [16] Krzystof and I proved a special case of the following result. The
result as stated is a consequence of the Scott-Wojciechowski Theorem as
was shown in [14], Sec. 4. The Scott-Wojciechowski Theorem will be
explained below.
Theorem 4.1. Let P,Q € Gr^(5). Then
rj(Dp) - 7j(DQ) = logdetF($(P)$(Q)*) modZ. (45)
If P or Q is the Calderon projector then (45) is even an equality [14].

Aspects of the mathematical work of Krzysztof P. Wojciechowski 19
The general answer to Problem 4.1 given by Scott and Wojciechowski
[21] is just beautiful. To explain their result we need another bit of notation.
Recall that J defines the symplectic form on L2(T,,E) (25). Let
E = Ei@ E-i
be the decomposition of E into the eigenbundles of J. If P 6 Gr^ (B) then
L = range PcL2(T,,E)
is Lagrangian and from Lemma 2.1 one easily infers that the restrictions
of the orthogonal projections H±i = -^ (i ± J) onto E±i map L bijectively
onto L2(£,.E±j) and
^(P):=U.io(Ui\Ei)-1
is a unitary operator from L2(E,Ei) onto L2(E,E-i). For P we then have
the formula
P=U J W). (46)
For P,Q £ Gr^(B) the operator 3>(P)*$(Q) — J is smoothing and hence
$(P)*3>(<5) is of determinant class.
With these preparations, the Scott-Wojciechowski theorem reads as fol­
lows.
Theorem 4.2. Let P e Gr^(S) an«f Ze< C(D,X) be the orthogonalized
Calderon projector. Then
de,t(Crf = WDCMW'+WWW). (47)
5. Adiabatic decomposition of the £—determinant
[15, 19, 20]
When the gluing formula for the ^-invariant had been established it
was Krzysztof's optimism that eventually lead to a similar result for the
^-determinant. The author has to admit that he was an unbeliever: I could
not see why a reasonable analytic surgery formula for the ^-determinant
should exist. Well, I was wrong. A fruitful collaboration of J. Park and
Krzysztof P. Wojciechowski eventually proved that the adiabatic method,
which originally had been developed in the paper [11], was even strong
enough to prove an adiabatic surgery formula for the ^-determinant.

20 Matthias Lesch
Consider again the adiabatic setting MR,XR,YR as in (3.1). In order
not to blow up the exposition too much I will not present the result in its
most general form. Rather I will make the following technical assumptions:
(1) The tangential operator B is invertible.
(2) The L2-kernel offlonlU [0, oo) x £ and Y U [0, oo) x £ vanishes.
Then the adiabatic surgery theorems for the Laplacians read as follows:
Theorem 5.1. Let A±tRtd be the Dirichlet extension of D2 on XR,YR
resp.; DR denotes the operator D on XR. Then
Iim detcf R = J^BI.
fl-»oo detc A+>fi>ddetc A_iiJid V
Theorem 5.2. Let D+tRtp+, D-tRtp_ be the operator D with Atiyah-
Patodi-Singer boundary conditions on XR,YR resp. Then
lim n2 det<f^ n2 = 2-<'(*2,o).
fl^°o detc£)j)fljn>detcDiRn<
These technical assumptions mentioned above were removed in Park
and Wojciechowski [20]. For details the reader should consult loc. cit.
Finally, the "adiabatic" results on the zeta-determinants obtained by
Park and Wojciechowski are not adiabatic any more. Loya and Park [17]
showed that most of those results (and more) are true without stretching.
Krzysztof P. Wojciechowski did have different (and charming) ideas how
to remove stretching of the cylinders. Unfortunately, his serious illness did
not allow him to fill all the details and finish the paper.
References .
1. M. F. Atiyah, V. K. Patodi, and I. M. Singer, Spectral asymmetry and Rie-
mannian geometry I, Math. Proc. Cambridge Philos. Soc. 77 (1975), 43-69.
2. M. F. Atiyah, V. K. Patodi, and I. M. Singer, Spectral asymmetry and Rie-
mannian geometry II, Math. Proc. Cambridge Philos. Soc. 78 (1975), 405-
432.
3. , Spectral asymmetry and Riemannian geometry III, Math. Proc.
Cambridge Philos. Soc. 79 (1976), 71-99.
4. B. Boofl and K. P. Wojciechowski, Desuspension and splitting elliptic symbols
I, Ann. Global Anal. Geom. 3 (1985), 337-383.
5. , Desuspension and splitting elliptic symbols II, Ann. Global Anal.
Geom. 4 (1986), 349-400.
6. B. Boofi-Bavnbek, M. Lesch, and C. Zhu, In preparation.

Aspects of the mathematical work of Krzysztof P. Wojciechowski 21
7. B. BooC-Bavnbek and K. P. Wojciechowski, Elliptic boundary problems for
Dirac operators, Birkhauser, Basel, 1993.
8. J. Briining and M. Lesch, On the eta-invariant of certain non-local boundary
value problems, Duke Math. J. 96 (1999), 425-468, dg-ga/9609001.
9. , On boundary value problems for Dirac type operators: I. Regularity
and self-adjointness, J. Funct. Anal. 185 (2001), 1-62, math.PA/9905181.
10. A. Calderon, Boundary value problems for elliptic equations, Outlines of the
joint Soviet-American symposium on partial differential equations (Novosi­
birsk), 1963, pp. 303-304.
11. R. G. Douglas and K. P. Wojciechowski, Adiabatic limits of the r)-invariants
the odd-dimensional Atiyah-Patodi-Singer problem, Comm. Math. Phys.
142 (1991), 139-168.
12. P. Gilkey, Invariance theory, the heat equation, and the Atiyah-Singer index
theorem, 2. ed., CRC Press, Boca Raton, 1995.
13. G. Grubb, Trace expansions for pseudodifferential boundary problems for
Dirac-type operators and more general systems, Ark. Mat. 37 (1999), 45-
86.
14. P. Kirk and M. Lesch, The r\-invariant, Maslov index, and spectral flow for
Dirac-type operators on manifolds with boundary, Forum Math. 16 (2004),
553-629, math.DG/0012123.
15. S. Klimek and K. P. Wojciechowski, Adiabatic cobordism theorems for ana­
lytic torsion and rj-invariant, J. Funct. Anal. 136 (1996), 269-293.
16. M. Lesch and K. P. Wojciechowski, On the rj-invariant of generalized Atiyah-
Patodi-Singer boundary value problems, Illinois J. Math. 40 (1996), 30-46.
17. P. Loya and J. Park, On the gluing problem for the spectral invariants of
dirac operators, To appear.
18. W. Miiller, Eta invariants and manifolds with boundary, J. Differential Geom.
40 (1994), 311-377.
19. J. Park and K. P. Wojciechowski, Adiabatic decomposition of the £-
determinant of the Dirac Laplacian I. The case of invertible tangential oper­
ator, Comm. Partial Differential Equations 27 (2002), 1407-1435, (with an
Appendix by Y. Lee).
20. , Scattering theory and adiabatic decomposition of the (-determinant
of the Dirac Laplacian, Math. Res. Lett. 9 (2002), 17-25.
21. S. G. Scott and K. P. Wojciechowski, The (-determinant and Quillen de­
terminant for a Dirac operator on a manifold with boundary, Geom. Funct.
Anal. 10 (1999), 1202-1236.
22. R. T. Seeley, Topics in pseudo-differential operators, C.I.M.E., Conference
on pseudo-differential operators 1968, Edizioni Cremonese, Roma, 1969, pp.
169-305.
23. , Singular integrals and boundary value problems, Amer. J. Math. 88
(1966), 781-809.
24. , Complex powers of an elliptic operator, Proc. Sympos. Pure Math.
10 (1967), 288-307.
25. K. P. Wojciechowski, Spectral flow and the general linear conjugation problem,
Simon Stevin 59 (1985), 59-91.

22 Matthias Lesch
26. , The additivity of the rj-invariant. The case of an invertible tangential
operator, Houston J. Math. 20 (1994), 603-621.
27. , The additivity of the n-invariant. The case of a singular tangential
operator, Comm. Math. Phys. 109 (1995), 315-327.
28. , The ^-determinant and the additivity of the n-invariant on the
smooth, self-adjoint Grassmannian, Comm. Math. Phys. 201 (1999), 423-
444.
Received by the editors September 14, 2005; revised January 5, 2006

Analysis, Geometry and Topology
of Elliptic Operators, pp. 23-38
© 2006 World Scientific Publishing Co.
GLUING FORMULAE OF SPECTRAL INVARIANTS AND
CAUCHY DATA SPACES
JINSUNG PARK
School of Mathematics, Korea Institute for Advanced Study
207-43, Cheongnyangni 2-dong, Dongdaemun-gu
Seoul 130-722, Korea
[email protected]. kr
Dedicated to Krzysztof P. Wojciechowski on his 50th birthday
We review the gluing formulae of the spectral invariants - the ^-regularized deter­
minant of a Laplace type operator and the eta invariant of a Dirac type operator.
In particular, we explain the crucial role of the Cauchy data spaces in these gluing
formulae.
2000 Mathematics Subject Classification. Primary 58J28; Secondary 58J52
1. Introduction
In this article, we survey the gluing formulae of the spectral invariants
- the ^-regularized determinant of a Laplace type operator and the eta
invariant of a Dirac type operator. After these spectral invariants had
been originally introduced by Ray and Singer [30] and Atiyah, Patodi, and
Singer [1] respectively, these invariants have been studied by many people
in many different parts of mathematics and physics. Here we discuss the
gluing formulae of these spectral invariants. These formulae have been
proved independently by several authors using different techniques. For
nice introductions to this subject, we refer to Bleecker and Boofi-Bavnbek
[3] and Mazzeo and Piazza [21] where the reader can find many technical
details and ideas of proofs. Therefore, instead of repeating the details of
these introductions, we explain one principle which holds for all the known
gluing formulae of the spectral invariants. This principle also enabled us to
get a new proof of the gluing formulae of the eta invariant of a Dirac type
operator and simultaneously to prove the gluing formula of the ^-regularized
determinant of a Dirac Laplacian [17], [18]. We hope that this article would
be helpful in the understanding of gluing formulae of the spectral invariants
23

24 Jinsung Park
and other related gluing problems in similar situations.
Now let us review briefly the history of this subject: gluing problems
of the spectral invariants. This will help the reader to understand the
motivation of this article.
First, it is appropriate to begin with mentioning the pioneering work
of Wojciechowski. In his paper with Douglas [10], they found a striking
formula, which states that the eta invariant of a Dirac type operator over
a manifold with boundary converges to a local expression as the cylindrical
length near the boundary is getting longer and longer (the adiabatic limit).
Although they did not formulate the gluing formula of the eta invariant in
their paper [10] explicitly, this result suggested the existence of the gluing
formula of the eta invariant.
The work of Douglas and Wojciechowski [10] stimulated many mathe­
maticians working around the eta invariant, so after their paper appeared,
during the last 15 years the gluing formula of the eta invariant has been
proved independently and using different techniques by Hassell, Mazzeo,
and Melrose [12], Wojciechowski [34], Bunke [6], Miiller [23], Briining and
Lesch [5], Kirk and Lesch [14], Park and Wojciechowski [27] and many oth­
ers. Although their proofs are different from each other, they altogether
used the generalized Atiyah-Patodi-Singer spectral projection to impose the
boundary conditions for the Dirac operator over manifold with boundary.
Among the aforementioned works, the formula of Kirk and Lesch is the
most complete in the sense that their formula has no integer ambiguity
(Bunke's formula also holds without the integer ambiguity) and they show
the origin of the integer contribution in their proof. In fact, they needed to
use the Calderon projector for the boundary condition to formulate their
formula. Hence, this seems to suggest that the Calderon projector might
be the natural projection in the gluing formula instead of the generalized
Atiyah-Patodi-Singer spectral projection.
We can also see such a suggestion from the adiabatic decomposition
formula of the (^-regularized determinant of the Dirac Laplacian proved by
Park and Wojciechowski [25], [26], [27]. In their formula, the adiabatic limit
of the ratio of the ^-regularized determinants of the Dirac Laplacians over
the original manifold and decomposed submanifolds is mainly described in
terms of the scattering matrices of the corresponding Dirac operators over
manifolds with cylindrical ends, which are obtained by attaching half in­
finite cylinders to the decomposed manifolds with boundaries. Here, we
can regard the scattering matrix for a noncompact manifold with cylin­
drical end as corresponding to the Calderon projector for a manifold with

Gluing formulae of spectral invariants and Cauchy data spaces 25
boundary.
Following this suggestion - the use of the Calderon projector, Loya and
Park [17] could find a new proof of the gluing formula of the eta invariant of
a Dirac type operator, which also provides us, simultaneously, with the glu­
ing formula of the (-regularized determinant of a Dirac Laplacian. Actually,
these formulas are not two different formulas, but just two aspects - phase
and modulus - of one unified formula. To state their formulae, they needed
to introduce an operator U which is defined by the Cauchy data spaces of
the restricted Dirac operators over the decomposed manifolds. The proof
in [17] can be easily employed for more general situations, for instance, for
noncompact manifolds [18]. (Technically, the proof in [17] does not use the
fact the variation of the eta invariant is locally computable, which holds
only for compact manifolds.)
We can also see such a suggestion in the gluing formula of the (-
regularized determinant proved in the end of 80's by Burghelea, Fried-
lander, and Kappeler (BFK) [7]. In (a special case of) their formula, the
ratio of the (-regularized determinants of Laplace type operators over the
original manifold and decomposed submanifolds is mainly described by the
(-regularized determinant of a certain operator 72. over the cutting hyper-
surface, and this operator TZ has an expression in terms of the Cauchy data
spaces of the restricted Laplace operators over decomposed manifolds. As
we will explain, the operators U and 1Z can be understood under the fol­
lowing principle: The gluing formulae of the spectral invariants are mainly
described by the difference of the Cauchy data spaces.
Now let us explain the structure of this article. In Section 2, we first
review the gluing formula of BFK [7] and one of its generalization [16]. In
Section 3, we explain how the operator 1Z can be understood in terms of
the Cauchy data spaces. In Section 4, we review the gluing formulae of the
spectral invariants of a Dirac type operator proved by Loya and Park [17].
Finally, in Section 5, we explain the operator U, which plays the crucial role
in the gluing formulae of the spectral invariants of a Dirac type operator,
and its underlying meaning compared with 71.
The author gives his sincere thanks to Paul Loya for his helpful com­
ments and the referee for corrections and suggestions, all of which consid­
erably improved this article.

26 Jinsung Park
2. Gluing formula of the ^-regularized determinant of a
Laplace type operator
In this section, we review the gluing formula of the ^-regularized deter­
minant of a Laplace type operator. This type of formula was first proved
by Burghelea, Friedlander and Kappeler (BFK) in [7] where they call this
a Mayer-Vietoris type formula. Although their formula holds for a more
general situation, that is, more general differential operators and general
local elliptic boundary conditions, here we just restrict our discussion to a
Laplace type operator and the Dirichlet boundary condition.
Now let us explain the BFK formula in more detail. Let M is a compact
manifold and Y is a hypersurface in M, which decomposes M into two
submanifolds Af_ and M+ (here we assume that M_ is the left side manifold
and M+ is the right side manifold). Hence we have
M = M_UM+, F = M.nM+.
Let us consider a Laplace type operator over M,
AM:H2(M,E)—>L2(M,E),
where E is a Hermitian vector bundle over M. For the restrictions of AM
to M_ and M+, we impose the Dirichlet boundary conditions so that we
obtain
AM± := AM|M± : dom(AM±) := {<t> £ #2(M±,£) |7o(0) = 0}
^L2{M±,E)
where 70 : H2(M±,E) -> H?(Y,E0) (E0 := E\Y) denotes the restriction
map to Y. Now let us recall that the (-function of AM is defined by
((s,AM):=— ts-1[Tr(e-*A^)-dimkerAM]^ (1)
1 (s) Jo
for 5?5 ^> 0 and this has the meromorphic extension over C with s = 0 as
a regular value. Then the (^-regularized determinant of AM is defined by
detcAM := exp(-C'(0, AM)). (2)
The (-regularized determinant of AM±, det^AM± is defined in a similar
way as det^ AM- NOW a natural question in this circumstance is the relation
of det^AM with det^AM_, det^AM+, and the BFK formula gives us the
answer to this question.
To explain their formula, we need to introduce an operator 1Z acting
on C°°{Y,E0), which is denned as follows: First, given / £ C°°(Y,E0),

Gluing formulae of spectral invariants and Cauchy data spaces 27
let us denote by <j>± the (assumed to be unique) solutions of the Dirichlet
problems for the restrictions of AM to M±:
AM(f>i = 0 over M± \ Y, <f>±\Y = /•
Then, the operator 72. is defined by
7lf := (ducf>_)\Y - (ducf>+)\Y
where u is the normal variable to Y such that ±du is the inward directional
derivative for M±.
Remark 2.1. For 72. to be well defined, we need the condition that the
Dirichlet problem of AM|M± has a unique solution. This is the case for
the Laplace-Beltrami operator acting on the space of fc-forms (see Remark
3.1). We always assume that the Laplace type operator AM satisfies this
condition in this article.
It is known that 72. is a nonnegative pseudodifferential operator over Y
of order 1, hence its £-regularized determinant is well defined. Now we are
ready to state the BFK formula.
Theorem 2.1. [7] When ker AM = {0}, we have
, /6t^M A = C(Y) • dekK (3)
detcAM+ • detf AM_
where C(Y) is a constant depending only on the symbols of AM, AJK±, 72.
over Y.
The BFK formula in (3) describes the ratio of det^AM and det,jAM+ •
detf AM_ in terms of det^72. modulo the constant C(Y), which can be con­
sidered as data near Y. Note that although the operator 72. is denned
over Y, 72. contains global information over M via the null solutions of the
restrictions of AM to M±.
Remark 2,2. By definition of 72., ker AM = {0} implies that ker 72. = {0}.
Hence, under this condition, all the operators occurring in (3) have trivial
kernels. Without this condition, we have an additional term on the right
side of (3).
Remark 2.3. When we assume that AM has the following product form
over a collar neighborhood U = Y x [—1, l]u of Y,

28 Jinsung Park
where u denotes the variable of the normal direction to Y and Ay is a
Laplace type operator over Y, we can obtain the exact value of C(Y) as in
19], [15], [28],
C(F) = 2-c(°'Ay)-fly. (4)
Here ((s,Ay) is the f-function of Ay and hy '•= dimker Ay.
For a noncompact manifold M, the operator e~x^M is not of trace class.
Hence the ^-regularized determinant cannot be defined as in compact case.
But, in this case, one can use the 6-trace of Melrose [22] or the relative
trace of Miiller [24] instead of using the ordinary trace of e~tAM. For the
C-regularized determinant defined by the 6-trace or relative trace over a
noncompact manifold, its gluing formulae have been proved by Hassell and
Zelditch [13] for the decomposition of M = K2 into a compact smooth
domain and its complement, and by Carron [9] for the general noncompact
case. Here we just explain one generalization of the BFK formula in (3)
for a noncompact manifold X with cylindrical end. The manifold X with
cylindrical end has the following decomposition,
X = N Uy Z
where N is a manifold with boundary Y and Z = Y x [0, oo)u. We may
assume there is a collar neighborhood W = Y x [—1,0]„ of Y within N.
We consider a Laplace type operator Ax acting on C°° (X, E) where E is
a Hermitian vector bundle over X. We also assume product structures of
the Riemannian metric of X and the Hermitian metric of E over
WUy Z^Y x [-l,oo)u.
Finally we assume the following expression of Ax over W Uy Z,
d2
Ax|wuyZ =--j-^ + Ay (5)
where Ay is a Laplace type operator over Y. As before, we impose the
Dirichlet boundary conditions for the restrictions of Ax to A^ and Z and
denote by Ajv, A% the resulting operators. Then the relative £-function
for (Ax, Az) is defined by
C(s, Ax, AZ) := |^y ( £ + £ )ts~l Tr(e-'A* - e~^)dt.
Here the integral fQ • dt has a meromorphic extension from 5Rs S> 0 to C and
the integral Jx • dt has a meromorphic extension from Ks <C 0 to C. The

Gluing formulae of spectral invariants and Cauchy data spaces 29
resulting meromorphic extension of C(s, Ax, Az) is regular at s — 0. (In
the above definition of ((s, Ax, Az), the relative trace Tr(e~'Ax - e~'Az)
contains the zero eigenvalues of Ax, but these are cancelled out after taking
the sum of JQ • dt and J1 • dt.) Then the relative ^-regularized determinant
is defined by
detc(AX) Az) := exp ( - C(0, Ax, Az)).
To explain the generalization of the BFK formula to manifolds with cylin­
drical end, we need to introduce some more notations. Let {v,j} be an
orthonormal basis for the kernel of Ax on L2(X, E) and let {Uj} be a ba­
sis of the 'extended L2-somtions' (bounded solutions of AxUj = 0) such
that at oo on the cylinder, {Uj(oo)} are orthonormal in L2(Y,Eo) where
EQ := E\Y- Let Vj = Uj\y and Vj = Uj\y be the restrictions of Uj and Uj,
respectively, to the hypersurface {0} x Y. It can be shown that the sections
{VJ,VJ} are linearly independent in L2(Y, Eo), therefore both operators
3 3
are nonnegative linear operators on the finite-dimensional vector space V =
sp&n{vj, Vj} c L2(Y,Eo)- Since the set {VJ, Vj} is a linearly independent
set spanning V, the operator
L + L-.V —+V
is positive. In particular, det(L + L) is nonzero. Now we can state the main
result of [16].
Theorem 2.2. [16] The following equality holds,
detc(Ax,Az) = 2_c(0,Ay)-hy detc7^
detcAiv det(L + L)
where £(s, Ay) is the ^-function of Ay and hy = dimker Ay.a
Remark 2.4. Originally the formula (6) was given in terms of the 6-trace
in [16]. We refer to [16] for this and an elementary introduction to the
6-trace. Since we do not assume any condition on the kernel of Ax, we
have the additional term det(i + L) on the right side of (6) (see Remark
aAfter [16] was completed, we learned that this result was also proved in "Regularized
determinants of Laplace type operators, analytic surgery and relative determinants" by
J. Muller and W. Muller.

30 Jinsung Park
2.2). By the product structure in (5), we could obtain the explicit form of
the constant C{Y) = 2-«0.Av)->>y as m (4).
3. The operator TL and the Cauchy data spaces of Laplace
type operator
The discussion in this section holds for a more general situation, but we
just restrict our concern to the closed manifold M. The main purpose of
this section is to investigate the operator 1Z in terms of the Cauchy data
spaces of AM\M± •
Now we recall the operator AM over M which is decomposed into
M-,M+ by the hypersurface Y. The trace map 7 is defined by
y(<f>) = (0|y,(^)|y) : H2(M,E) —+ fff(Y,E0) © Hi(Y,E0)
where EQ := E\y. Here u denotes the normal variable for the collar neigh­
borhood
U = Yx [-1,1],,
of Y and ±<9„ is the inward directional normal derivative to M±. Then
the restriction A± :— AM\M± determines the Cauchy data space H(A±)
defined by
W(A±) := { (f,g) G C°°(Y,E0) ® C°°(Y,Eo) \ ^ € C°°(M±,E)
such that AM4> = 0 on M±\Y and 7(1/1) = (f,g) }.
Hence, H{A±) consists of the pair of the Dirichlet and Neumann data of
A± over M±.
The Dirichlet-to-Neumann operator A/± over C°°(Y, Eo) is denned to
be the map sending / G C°°(Y, EQ) to the corresponding Neumann data
g G C°°(Y,E0) such that (f,g) G H(A±). Note that the well-definedness
of Af± (kerA/± = {0}) is equivalent to the condition that the operator A±
with the Dirichlet (Neumann) boundary condition has no null solution.
Remark 3.1. Let us consider the case of the Laplace-Beltrami operator
(d + d*)2 acting on the space of A:-forms over a manifold with boundary.
The operator (d + d*)2 with the Dirichlet boundary condition has no null
solution. Indeed, a null solution of (d + d*)2 would also be a null solution
of d + d* by the Green formula. But, this is impossible by the unique con­
tinuation theorem for d+d* [2]. Hence, the Dirichlet-to-Neumann operator
for the Laplace-Beltrami operator is well defined.

Gluing formulae of spectral invariants and Cauchy data spaces 31
Under the condition assumed in Remark 2.1, the operator A/± defines its
graph which is exactly H(A.±). We refer to [29], [20] for some more detailed
explanations about the Dirichlet-to-Neumann operator and its application
to the relative formula of the Dirichlet/Neumann Laplacians.
Now, recalling the definition of TZ and A/±, we can easily see that
TZ = N_-N+:Cco (Y, Eo) —• C°° (Y, E0). (7)
Hence, we can see that the operator TZ describes the difference of two
Cauchy data spaces Ti(A±) in the sense that
(f,Kf) = (f,/f-f)-(f,M+f) for (/,A/±/)GW(A±).
In conclusion, the gluing formula of det^Ajv? is mainly described by the
difference of the Cauchy data spaces H{A±).
4. Gluing formula of the spectral invariants of a Dirac type
operator
In this section, we discuss the gluing formulae [17] of the spectral invariants
of Dirac type operators, that is, the eta invariant of a Dirac type operator
and the ^-regularized determinant of a Dirac Laplacian.
Let V be & Dirac type operator acting on C°°(M, S) where M is a
closed compact Riemannian manifold of arbitrary dimension and S is a
Clifford bundle over M. Let Y be an embedded hypersurface in M which
decomposes M into two submanifolds M_ and M+. Hence we have
M = M_ U M+, Y = M_ n M+.
We assume all geometric structures are of product type over a tubular
neighborhood U == [—1,1] u x Y of Y where the Dirac operator takes the
product form over U,
V\u = G(du + DY),
where G is a unitary operator on So '•— S\y and Dy is a Dirac type operator
over Y satisfying
G2 = -Id and DYG = -GDy .
Recall that the eta function of T> and the zeta function of V2 are defined
through the heat operator e~tV by
1 r°°
^'^ := F7HH / ^ ^(Ve-™2) dt,
11 2 ) Jo
((s,V2) := -i- / t-1['Ir(e-«'a) -dimker©] dt,
1 Ks) Jo

32 Jinsung Park
which are defined a priori for ?Rs ^> 0 and extend to be meromorphic
functions on C that are regular at s = 0. The eta invariant and the reduced
eta invariant of T> are defined by
77(2?) := 77(0,13), 77(D) := 1(^(0,23) + dimkerP),
and the ^-determinant of V2, det^D2 is defined as in (2) using ((s,V2).
By restriction, V induces Dirac type operators T>± over M±. For these
operators, we choose orthogonal projections V± over L2(Y, So) that provide
us with well-posed boundary conditions for T>± in the sense of Seeley [33].
Then the following operators
VP± : dom(PP±) - L2 (M±, S) (8)
where
dom(2?P±) := {$ G Hl(M±,S) | P±f>|y) =0},
share many of the analytic properties of V; in particular, they are Fredholm
and have discrete spectra, but are not necessarily self-adjoint. Amongst
such projectors are the (orthogonalized) Calderon projectors C± [8], [32]
which are projectors defined intrinsically as the unique orthogonal projec­
tors onto the closures in L2(Y, So) of the infinite-dimensional Cauchy data
spaces of T>±:
W(2?±) := { / G C°°(y, So) I 3^ G C°°(M±, 5) such that
V(f> = 0 on M±\y and <f>\Y = f }. (9)
To state our main theorem, we recall that the Calderon projectors C±
have the matrix forms
with respect to the decomposition
L2(Y,S0) = L2(Y,S+)®L2(Y,S-),
where S± C So are the (ii)-eigenspaces of G. The maps
K±:L2(Y,S+)^L2(Y,S-) (11)
are isometries, so that

Gluing formulae of spectral invariants and Cauchy data spaces 33
is a unitary operator over L2(Y, S~), which is moreover of Fredholm deter­
minant class. The last property follows easily from the fact that the differ­
ences of C± and the Atiyah-Patodi-Singer spectral projections are smooth­
ing operators; we refer to [31] for the details. We denote by U the restriction
of U to the orthogonal complement of its (—l)-eigenspace. We also put
L:=^Vj®v* (12)
where hu = dim ker 2? and Vj = Uj\y with the orthonormal basis {UJ} of
ker V. Then L is a positive operator on the finite-dimensional vector space
(kerX>)|y. We are now ready to introduce the main result of [17].
Theorem 4.1. [17] The following gluing formulae hold:
detcP2 „_„„.n?.wfc„,, . „_a , , ^Id + U + U-1*
detcX>2_ . dekV2+
= 2-c(o,^)-v(deti)-2det^£i±^±^l))
fj(V) - fj(Vc_) - f,{Vc+) = ^~. Log detF U mod Z,
where C(s,Dy) is the (-function of DY, hy = dim ker Dy, det^ denotes
the Fredholm determinant, and Log the principal value logarithm.
We can replace the Calderon projector C± in the gluing formulae in
Theorem 4.1 by other orthogonal projection in the smooth, self-adjoint
Grassmannian Gr^XJCD±), which consists of orthogonal projections V± such
that V± — C± are smoothing operators and
GV± = (Id -V±)G.
Let Vi G Gr^CV-) and V2 e Gr*oc(V+). Then the eta invariant of VVi
and the ^-regularized determinant of V2?. are well defined by the results of
Grubb [11] and Wojciechowski [35]. The orthogonal projections V\ and V%
determine maps K\ and K2 as in (10), and we define
Ui := K-Kf1, U2:=K2K+1, U12 := -KiK^1 over L2(Y, S~).
As before, let us denote by U the restriction of Ui to the orthogonal com­
plement of its (—l)-eigenspace. We define the operator L\ over the finite-
dimensional vector space ran(C_) n ran(Id — V) by
Li = -P1GTZZ1GP1 (13)
where 72— is the BFK operator for the double of (7J>_)2 (see chapter 9 of
[4] for the double construction), and Pi is the orthogonal projection onto

34 Jinsung Park
ran(C_) n ran(Id — V). Then L\ is a positive operator [19]. We define
Li in a similar way. We can now state the general gluing formulae for the
spectral invariants of Dirac type operators.
Theorem 4.2. [17] The following general gluing formulae hold:
A J^' =2-C(°^)-^(detL)-Met,(2Id + ^ + ^1)
detcZ7^ • detcp2,2 v ' \ 4 /
•J](detLi) -2detF( 4 j '
fj(V) - fj(VVl) - fj(VV2) = -i-LogdetFt/12 modZ.
Z7TI
Remark 4.1. As in Theorem 2.2, we can generalize Theorem 4.2 to non-
compact manifolds with cylindrical end. We refer to [18] for this result and
its proof.
The gluing formula of the eta invariant in Theorem 4.2 (when Vi are the
generalized Atiyah-Patodi-Singer spectral projections) has the same form
as (or its reduced form modulo Z of) the gluing formulae proved by Has-
sell, Mazzeo, and Melrose [12], Wojciechowski [34], Bunke [6], Miiller [23],
Briining and Lesch [5], Kirk and Lesch [14], Park and Wojciechowski [27]. In
this formula, the data given by the Calderon projector or the Cauchy data
spaces are cancelled so these data do not appear in the gluing formula. This
is the main reason that the important role of the Cauchy data spaces in the
gluing formula of the eta invariant has not been noticed before the work of
Kirk and Lesch [14]. But, in the gluing formula of the ^-regularized deter­
minant in Theorem 4.2, these terms appear via U with the additional terms
Ui when we impose boundary conditions given by orthogonal projections
other than the Calderon projectors C±. Comparing the gluing formulae in
Theorems 4.1 and 4.2, we can see that the Calderon projectors are the most
natural projections for the gluing formulae of the spectral invariants since
imposing the boundary conditions by these projections makes the formulae
the simplest possible.
5. The operator U and the Cauchy data spaces of a Dirac
type operator
In this section, we investigate the operator U, which appears both gluing
formulae of the ^-regularized determinant and the eta invariant in Theorem

Gluing formulae of spectral invariants and Cauchy data spaces 35
4.1. The purpose of this section is to explain that the unitary operator U
also describes the difference of the Cauchy data spaces H(D±) as 1Z did the
difference of H(Aj-).
As before, let us consider the operator T> over the closed manifold M
which is decomposed into M-,M+ by the hypersurface Y. (The following
discussion holds for more general situations, but we just restrict our concern
to the closed manifold M.) Now let us recall that H(D±) consists of the
boundary values of the null-solutions of V± over M± (see the definition (9))
and the (orthogonalized) Calderon projectors C± are the unique orthogonal
projectors onto the closures of H{T)±), W(X>±) in L2(Y,S0).
To explain the underlying meaning of U, let us explain how we could
derive the definition of the operator U by modelling on the operator V,.
First, let us recall the expression of TZ in (7) with A/± and that A/±
determines H(A±) as its graph. So, our task is to find the operator whose
graph is H(D±). Noting that the Calderon projector C± has the matrix
form in (10), we can see that K± determines (the closure in L2(Y, So) of)
W(P-t) as its graph.
Now the question is how we define U using K±. We here observe that
H(T>±) is a Lagrangian subspace in L2(Y, So) with respect to the symplectic
form (G , ). Hence, there is the unitary operator over L2(Y,S~) which
transforms H(V+) to H(D-), that is, describes the difference of them.
From this reasoning, we can see that the operator K_K^X does this since
(x, K_X) = (x, (/t_/t^1)«;_(-a;) for (X,K±X) £H(D±).
But, we actually need to find the unitary operator which transforms the
Cauchy data space of V*^. to H(V-), where W+ is the reflection of the Dirac
type operator V+ to the manifold M^, which is the left side manifold on the
double of M+. This is because M+(M-) is a right (left) side manifold and
we have to compare the Cauchy data spaces over the left side manifolds
to measure the true difference of the Cauchy data spaces. Recalling the
double construction in chapter 9 of [4], we can see that the corresponding
Calderon projector C\ on the left side manifold M+ to C+ on M+ is given
by
+ 2 \-K+ Id /
Therefore, the operator — K+ determines the Cauchy data space W(D^_)
as its graph. In conclusion, we can see that U = —K_K^1 is the correct
operator measuring the true difference of the Cauchy data spaces.

36 Jinsung Park
The interesting point is that we can obtain the modulus and the phase
data from U via the following equality:
where the principal logarithm of the first factor U, LogU describes the
gluing formula of the eta invariant and the second modulus part describes
the gluing formula of the ^-regularized determinant in Theorem 4.1. In fact,
this is not a simple coincidence but follows from the deep relation between
the eta invariant of the Dirac operator and the ^-regularized determinant
of the Dirac Laplacian as the proof of Theorem 4.1 shows. We refer to [17]
for the details of its proof.
References
1. M.F. Atiyah, V.K. Patodi, and I.M. Singer, Spectral asymmetry and Rieman-
nian geometry. I, Math. Proc. Cambridge Philos. Soc. 77 (1975), 43-69.
2. C. Bar, Zero sets of solutions to semilinear elliptic systems of first order,
Invent. Math. 138 (1999), 183-202.
3. D. Bleecker and B. Boofi-Bavnbek, Spectral invariants of operators of Dirac
type on partitioned manifolds, Aspects of Boundary Problems in Analysis and
Geometry, 1-130, Oper. Theory Adv. Appl., 151, Birkhauser, Basel, 2004.
4. B. Boofi-Bavnbek and K. P. Wojciechowski, Elliptic Boundary Problems for
Dirac Operators, Birkhauser Boston Inc., Boston, MA, 1993.
5. J. Briining and M. Lesch, On the r)-invariant of certain nonlocal boundary
value problems, Duke Math. J. 96 (1999), 425-468.
6. U. Bunke, On the gluing formula for the n-invariant, J. Differential Geom. 41
(1995), 397-448.
7. D. Burghelea, L. Friedlander, and T. Kappeler, Mayer-Vietoris type formula
for determinants of differential operators, J. Funct. Anal. 107 (1992), 34-65.
8. A.-P. Calderon, Boundary value problems for elliptic equations, 1963 Outlines
Joint Sympos. Partial Differential Equations (Novosibirsk, 1963) pp. 303-304
Acad. Sci. USSR Siberian Branch, Moscow
9. G. Carron, Determinant relatif et la fonction Xi, Amer. J. Math. 124, no. 2
(2002), 307-352.
10. R. G. Douglas and K. P. Wojciechowski: Adiabatic limits of the n-invariants.
The odd-dimensional Atiyah-Patodi-Singer problem, Comm. Math. Phys. 142
(1991), 139-168.
11. G. Grubb, Poles of zeta and eta functions for perturbations of the Atiyah-
Patodi-Singer problem, Comm. Math. Phys. 215 (2001), 583-589.
12. A. Hassell, R. R. Mazzeo, and R. B. Melrose, Analytic surgery and the accu­
mulation of eigenvalues, Comm. Anal. Geom. 3 (1995), 115-222.
13. A. Hassell and S. Zelditch, Determinants of Laplacians in exterior domains,
IMRN 18 (1999), 971-1004.

Gluing formulae of spectral invariants and Cauchy data spaces 37
14. P. Kirk and M. Lesch, The eta invariant, Maslov index, and spectral flow
for Dirac-type operators on manifolds with boundary, Forum Math. 16 (2004),
553-629.
15. Y. Lee, Burghelea-Friedlander-Kappeler's gluing formula for the zeta-
determinant and its applications to the adiabatic decompositions of the zeta-
determinant and the analytic torsion, Trans. Amer. Math. Soc. 355, no. 10
(2003), 4093-4110.
16. P. Loya and J. Park, Decomposition of the zeta-determinant for the Laplacian
on manifolds with cylindrical end, Illinois J. Math. 48, no. 4 (2004), 1279-1303.
17. P. Loya and J. Park, On the gluing problem for the spectral invariants of
Dirac operators, Adv. Math., to appear.
18. P. Loya and J. Park, On the gluing problem for Dirac operators on manifolds
with cylindrical ends, J. Geom. Anal. 15 (2005), 285-319.
19. P. Loya and J. Park, The comparison problem for the spectral invariants of
Dirac type operators, Preprint, 2004.
20. P. Loya and J. Park, £-determinants of Laplacians with Neumann and Dirich-
let boundary conditions, J. Phys. A, 38 (2005), 8967-8977.
21. R. Mazzeo and P. Piazza, Dirac operators, heat kernels and microlocal anal­
ysis. II. Analytic surgery, Rend. Mat. Appl. (7) 18, no. 2 (1998), 221-288.
22. R. B. Melrose, The Atiyah-Patodi-Singer Index Theorem, A.K. Peters,
Wellesley, 1993.
23. W. Miiller, On the L -index of Dirac operators on manifolds with corners of
codimension two. I, J. Differential Geom. 44 (1996), 97-177.
24. W. Miiller, Relative zeta functions, relative determinants and scattering the­
ory, Comm. Math. Phys. 192 (1998), 309-347.
25. J. Park and K. P. Wojciechowski, Adiabatic decomposition of the £-
determinant of the Dirac Laplacian. I. The case of an invertible tangential
operator. With an appendix by Yoonweon Lee, Comm. Partial Differential
Equations 27 (2002), no. 7-8, 1407-1435.
26. J. Park and K. P. Wojciechowski, Scattering theory and adiabatic decompo­
sition of the (^-determinant of the Dirac Laplacian, Math. Res. Lett. 9 (2002),
no. 1, 17-25.
27. J. Park and K. P. Wojciechowski, Adiabatic decomposition of the
^-determinant and Scattering theory, Michigan Math. J., to appear,
math.DG/0111046 .
28. J. Park and K. P. Wojciechowski, Adiabatic decomposition of the zeta-
determinant and the Dirichlet to Neumann operator, J. Geom. Phys. 55
(2005), 241-266.
29. J. Park and K. P. Wojciechowski, Agranovich-Dynin formula for the zeta-
determinants of the Neumann and Dirichlet problems, Spectral geometry of
manifolds with boundary and decomposition of manifolds, 109-121, Contemp.
Math., 366, Amer. Math. Soc, Providence, RI, 2005.
30. D. B. Ray and I. M. Singer, R-torsion and the Laplacian on Riemannian
manifolds, Adv. Math. 7 (1971), 145-210.
31. S. G. Scott and K. P. Wojciechowski, The £-determinant and Quillen de­
terminant for a Dirac operator on a manifold with boundary, Geom. Funct.

38 Jinsung Park
Anal. 10, no. 5 (1999), 1202-1236.
32. R. T. Seeley, Singular integrals and boundary value problems, Amer. J. Math.
88 (1966), 781-809.
33. R. T. Seeley, Topics in pseudo-differential operators, Pseudo-Diff. Operators
(C.I.M.E., Stresa, 1968), 1969, pp. 167-305.
34. K. P. Wojciechowski, The additivity of the rj-invariant. The case of a singular
tangential operator, Comm. Math. Phys. 169 (1995), 315-327.
35. K. P. Wojciechowski, The ^-determinant and the additivity of the rj-invariant
on the smooth, self-adjoint Grassmannian, Comm. Math. Phys. 201, no. 2
(1999), 423-444.
Received by the editors September 15, 2005 ; Revised January 4, 2006

Exploring the Variety of Random
Documents with Different Content

SZENT SZERAFIN MEGTÉRÉSE.
Élt Firenze városában egy érsek, Seraphino nevü. Apja után
tulajdonképpen Calzolaio-nak hívták, de fiatal korában feljebbvalói és
társai a mennyei szelídségért, mely lelkében lakott, és szép,
kedvesen jóságos arczáért, mely a szeráfokra emlékeztetett,
Seraphinónak nevezték el; és ez a név rajta maradt, mert – ámbátor
sokáig élt – soha se cselekedett olyat, a mi erre a névre ráczáfolt
volna.
Seraphino egyike volt a legerényesebb embereknek, a kik valaha
e földi tereken bolyongtak.
Abban az időben Firenze városát polgárháborúk dúlták. Mindenki
kivette részét a harczból: fiatalok és öregek, férfiak és asszonyok, a
papok épp úgy, mint a világiak. Erőszakosságok, gyilkosságok,
rokonnak rokon ellen támadása: a köznapi dolgok közé tartoztak.
Ebben a sötét korszakban Seraphino volt az egyetlen, a kinek
szívében épen élt Szent János rövid tanítása: »Fiaim, szeressétek
egymást!« Seraphino a békét hirdette, a béke művén fáradozott, és
életének egy nevezetes része abban telt el, hogy folytonosan
könyörgött a körülötte viaskodóknak: hagyjanak fel az örökös
küzdelemmel és éljenek istennek tetsző életet!
Ő maga soha se bántott senkit; mindenkivel szemben nyájas volt,
és nem vétett a légynek se. Szerette társait, szerette minden
felebarátját, haragosainak – mert mégis voltak haragosai –
megbocsátott. Az igazságtalanságot elszenvedte visszatorlás, panasz
és szemrehányás nélkül; a maga kis körében fölmerült ellentéteket
elsimítani, kiegyenlíteni törekedett. Az emberi életet mindenekfölött
drágának tartotta; sőt azt vallotta, hogy az állatokat se szabad
bántani, mert ők is isten teremtései. A mi nagy szó volt abban a
korban, mely ugyancsak megszokta a vérontás látványát.

Seraphinónak ehhez a jó tulajdonságához más erények is
járultak. Szerény volt és kivánságok nélkül való; lelkében nem talált
helyet se a kevélység, se a nagyravágyás. Hűségesen
engedelmeskedett minden isteni és emberi törvénynek; őszintén
alázatos volt, és érdemei nem tették elbizakodottá. Végül – a mire
akkortájt se özönlött a példa – tiszta életet élt.
Úgy látszott, nem csupán a haragot zárta ki a szívéből, hanem
minden indulatot, minden szenvedélyt.
Szóval Seraphino ékes volt minden erénynyel, főképpen minden
keresztény erénynyel.
Azért nem szorúl magyarázatra, hogy – bár távol esett tőle a
nagyravágyás – már negyven éves korában érsekké lett. Nemcsak
hogy nem törte magát ez után a nagy tisztség után: minden áron el
akarta hárítani magától, mert a feladatot súlyosnak, erejét
meghaladónak és mások jogos várandóságának tartotta. De hiába
szabódott; vállalnia kellett.
A nép örült ennek a választásnak. Seraphinót mindenki tisztelte,
és azok, a kik unták már a sok háborút és belső villongást, azt
remélték, hogy az új, békésebb állapotoknak Seraphino erős oszlopa
lesz.
Érsekségének mindjárt a legelső évében Seraphinónak még egy,
eddig nem ismert erénye tudódott ki. Kiderült, hogy nemcsak
jóságos, békeszerető és kifogástalan életű, hanem egyszersmind
ékesszavú.
Silvestrot, az új gonfalonieret, gyönyörű beszéddel üdvözölte, a
mely mindenkit elragadott. A beszéd oly bátor volt, hogy még
Silvestrónak is tetszett. De nem csupán bátor, hanem tartalmas is.
Magasztos gondolkozásra valló, bölcs és megszívlelésre méltó
tanácsok foglaltattak benne; gyönyörű útmutatás arra nézve, hogy
az új gonfaloniere hogyan teljesítse tisztjét. Ezeket a bölcs
tanácsokat ugyan Silvestro soha se fogadta meg, de ha a beszédnek
nem is volt foganatja, ez nem kisebbíthette Seraphinónak az

érdemét. Nem a hasznáért kell becsülni a szépet; a mi szép,
önmagában is szép, szép akkor is, a mikor haszon, eredmény, siker
nem kiséri.
E sokat igérő kezdet után Seraphino érsek egy vagy két éven át
többször is szólt a firenzei néphez. És minél többször szólt, annál
általánosabban csodálták ékesszólásának művészetét. A második
alkalomtól fogva mindannyiszor a földi lét hívságairól és a másvilág
örömeiről beszélt. Ez a fenséges tárgy sohase lelt hivatottabb
magyarázóra. Seraphino olyan meggyőzően ecsetelte a földi élet
nyomorúságait, és olyan lángoló lelkesedéssel, annyi rajongással, oly
fennen szárnyaló képzelettel azokat a gyönyörűségeket, a melyekre
a másvilágon tehetünk szert, ha idelenn szabályszerűen élünk, hogy
az embernek kedve kerekedett nyomban meghalni, mert ki ne
vágyott volna rá, hogy minél hamarább legyen része abban a
tökéletes megtisztultságban, a melynek örömeit Seraphino eladdig el
se képzelt színekkel tudta lefesteni?!
És ha Firenze polgárai akkoriban igen sűrűn küldözgették
egymást a másvilágra, ez meglehet, nem egyszer tisztára
jóindulatból és felebarátságból történt. Azoknak is, a kik inkább csak
hirtelen indulatból vagy egyszerűen gonoszságból öltek, némi
mentségéül szolgált, hogy gonoszságukat nem tetézték
hitetlenséggel s lelkiismeretükön valamelyest könnyítettek azzal a
tudattal, hogy embertársukat csupán az egyik és éppen a
hitványabbik életétől fosztják meg.
Természetesen Seraphinót nem lehet felelőssé tenni azért, mert
kortársai az ő legtisztább szándékait rosszra magyarázták és rosszra
használták. Közös sorsa ez minden tiszta szándéknak és öröktől
fogva, mert a világi rossz öröktől való.
Nem lehet felelőssé tenni annál kevésbbé, mert Firenze
lakosainak csak igen kevés alkalmuk volt arra, hogy az érsek
ékesszólását félre magyarázzák.
É

Érsekségének harmadik évétől fogva Seraphino nem hallatta
szavát; egyszer se, soha többé. Hirtelen kivonta magát mindenből, a
mi a nyilvánosság előtt folyt le, s halvány, jóságos arcza örökre
eltünt a prédikáló székből. Híre terjedt, hogy nagyon beteg; s az,
hogy hetekig feküdt, igaznak bizonyult. Nemsokára lábra állott
ugyan, de ékesszólásának csodálói ezután is hiába várták. Mindentől
és mindenkitől visszavonultan élt a »halmok« magaslatán, és csak
ritkán látták, a mint papjai kiséretében, botjára támaszkodva, olykor-
olykor kisétált a sorompó felé. A kik sétautján találkoztak vele, fehér
arczának láttára, aggodalmasan suttogták:
– A szent ember nagyon beteg!… A szent ember nem fog sokáig
élni!…
És imádkoztak az életéért.
Pedig voltaképpen nem volt súlyosabb baja és az élete nem
forgott nagyobb veszedelemben, mint azoké, a kik imádkoztak érte.
De azt hitte, hogy valami lassan ölő, titokzatos kórság környékezi,
úgy érezte, hogy testi ereje rohamosan hanyatlik, és ez az érzés
nagyon megriasztotta.
Első ifjúságában ugyanis beteges és rendkívül gyönge testalkatú
volt. Gyakran vesződött kisebb bajokkal; koronkint lappangó lázak
gyötörték és egyszer olyan csudálatos betegségbe esett, mely
társait, a jó testvéreket, valósággal megijesztette. Nem tudott
magáról és félrebeszélt. A jó testvérek rémületükben elhivatták
Ruggiero mestert, a ki mindenkinél jobban ismerte a csillagok
járását, az emberi szervezet titkait és a füvek gyógyító erejét. És a
kis Seraphino, vagy az orvosságtól, vagy a sok fejcsóválástól, jobban
lett. De mielőtt föllábadt, egy napon, a mikor azt hitték, hogy alszik,
kihallgatta Ruggiero mesternek Filippo testvérrel való beszélgetését.
– Attól félek – szólt Filippo testvér – hogy szegény testvérünk
nem lesz hosszú életü.
– Bizony, nem sok reménysége van rá – felelt Ruggiero mester. –
Mindenből az látszik, hogy fiatalon fog meghalni, de ha nagyon

vigyáz magára, ha mindenekben kiméli magát, megérheti, míg érsek
lesz belőle.
Ez a beszélgetés nem ment ki a Seraphino fejéből. Nem szólt
senkinek, de attól fogva őrizkedett a megfázástól; gondja volt rá,
hogy a munkában való elmerüléssel meg ne erőltesse magát; távol
tartotta magától az izgalmakat, és ha valaki olyat cselekedett vagy
mondott, a mi mást haragra lobbantott volna, Seraphino megfékezte
indulatát és így szólt magában:
– Nem, nem… azért is másra gondolok!… mert ha felindulok:
ezzel is megrövidítem az életemet!…
Néha a rosszban is van valami jó, és nem lehetetlen, hogy
Seraphino lelkének tökéletes épségét éppen ennek a mindig
óvatosságra intő, elűzhetetlen, leküzdhetetlen, sötét gondolatnak
köszönhette. Mert ifjúkorában őt is megkörnyékezte a kisértés, de
Seraphino így szólt magában:
– Nem, nem… a kisértés a gyehennához vezet!… A bűn megöli a
lelket… megöli a lelket, és azonkívül aláássa az egészséget is!…
S ez a meggondolás erőssé tette a csábítással szemben.
Ruggiero mester intelme Seraphinot végig kisérte egész életén.
Nem mult nap, hogy ne gondolt volna rá; elhatározásait, majdnem
minden lépését ez irányította; nyomot hagyott a szavain, rá
nehezedett egész gondolkozására. Később, a mint mult az idő, és
még folyvást életben találta magát: a kellemetlen emlék elveszítette
erejét; néha, kivált tavasz idején, a napfény óráiban, fölszabadúltnak
hitte magát, mint az elitélt, a ki a jó jelekből azt reméli, hogy
kegyelmet kapott. De később is megtörtént, jóllehet ritkábban, hogy
az éj sötétjében hirtelen felriadt, s virradatig küzködött a
felejthetetlen mementóval. Gondolatainak szövevényéből
akaraterejének minden megfeszítésével se tudta kiirtani; akármivel
foglalkozott, a mementó ott vigyorgott elméjének munkája mögött.

Mikor érsekké lett, az új hivatalával kapcsolatos elfoglaltságok
egy kissé elszórakoztatták. A köztisztelet sokféle megnyilatkozása jól
esett neki és elevenebbé tette; új tennivalói, az ezekkel járó gond, a
több emberrel való érintkezés egy kissé elterelték figyelmét a
közelében borongó sötétségről, a melyről azelőtt alig tudta levenni a
szemét. De ez nem sokáig tartott. Új környezete egy-két évi
megszokás után nem szórakoztatta többé; betelvén legtitkosabb
óhajtása is, nem volt, a mit czélul tűzhetett volna maga elé;
elfoglaltságát megunta és rábízta alantasaira; tehette, a mit akart,
tehát: csak gondolkozott, töprengett, tépelődött.
És ekkor egyszerre, váratlanul, újra meglepték azok a lappangó
lázak, a melyek ifjúságát megrontották. Több, mint húsz esztendő
múlt el azóta; egy jó harmadrész emberélet. Már elérte azt a kort, a
melyben az erősek közül is sokan kidőlnek; elérte, hogy érsekké lett,
megélt annyi esztendőt, a mennyit Ruggiero mester a legjobb esetre
jósolt, a mennyinél többet a csillagok járása nem igért neki,
akárhogyan vigyáz, akárhogyan kiméli magát.
Seraphino megrémült. Mert akármilyen csodaszépen tudta
ecsetelni a másvilág örömeit, azért mégse kivánkozott a másvilágra.
A mikor a lopva járó, alattomos láz egy éjszaka ismét felköltötte
álmából és szívének vad dobogásából meggyőződött róla, hogy
csakugyan beteg, a sötétségben összetette kezét és így imádkozott:
– Rex regum, dominus dominantium! Tudom, hogy ez a földi élet
hitványság, siralom; csak pitvara az igazi életnek; keserves föltétel, a
melyet a kiválasztottaknak elengedsz; megpróbáltatás, a melyen
túlesni boldogság. Tudom, hogy gyarlóság, ostobaság, főbenjáró
vétek: ragaszkodni ehhez a hívsághoz; mea culpa, mea maxima
culpa! És mégis könyörgök: hosszabbítsd meg földi életemet! Az
üdvösség, melyet a jóknak adsz, véges értelmünkkel föl nem
fogható; ellenben ez a hitványság közel esik hozzánk, már megvan,
szívünk szerint való, és valamint az örökélet, azonképpen ez is a te
rendelésed! Érdemetlen szolgád vagyok; sokat vétkeztem, ha
tettekkel nem is, legalább gondolatban. De a hit, a tiszta szándék, a

jóakarat nem hiányzott szívemből; iparkodtam, hogy minél inkább
megtisztuljak, gondosan ápoltam a Benned való bizodalmat és igaz
lélekkel, buzgón hirdettem a Te igéidet! Rex regum, dominus
dominantium! Esengek, hosszabbítsd meg földi létemet, a melylyel
összeforrottam, a melynél egyebet értelmem nem bir felfogni! Avagy
ha máskép parancsoltad volna: erősíts meg hitemben, azonképpen,
mint szegény szolgád rajta volt, hogy megerősítse hitökben a
gondjára bizottakat, megtérítse a tévelygőket és felvilágosítsa a
magukkal meghasonlottakat!
A láz és aggodalom ágyba döntötték. Ijedelmében meghagyta,
hogy hívják ágyához Firenze legnagyobb tudósait. Ruggiero mester
már régen meghalt, de tanítványokat hagyott, a kiknek annak idején
megmagyarázta volt, hogy miképpen kell megfoltozni a mások
életét.
A tudósok eljöttek, megnézték a beteget, és nem valami kisebb
emberről lévén szó, hanem magáról az érsekről: igen-igen csóválták
a fejöket.
Akárhogyan csóválták a fejöket, Seraphino meggyógyult.
Meggyógyult, de csak azért, hogy ezentúl már csupán a remegésnek
éljen.
A tudósok a csillagok járásából azt olvasták ki, hogy az érseknek
kimélnie kell magát. Az érsek tehát kimélte magát. Tenyerének
vizsgálatából kiderült, hogy a szellemi megerőtetés tette beteggé; az
érsek tehát megfogadta, hogy többé nem fog szónokolni. Szemének
fényességéből látták, hogy óvakodnia kell a templom hidegétől;
arczának színéből kitűnt, hogy a bőjtölés ártalmára van. Az érsek
tehát kikérte a betegeket megillető fölmentést, s tartózkodott úgy a
böjttől, mint a misemondástól.
És ettől fogva csak az egészségét ápolta. Nyáron is meleg
ruhában járt; gyógyító italokat, kenőcsöket, elixireket használt,
továbbá, nemcsak hogy szónokolni nem mert, a közönséges
beszédet is kerülte. Ha éppen kellett: suttogva szólt, hogy meg ne

erőtesse a mellét. Ha a levegőre ment: testőrökkel vette körül
magát, a kik megóvják az utonállóktól és más gonosz emberektől.
Ha az étel nem ízlett neki, elszomorodott, és imádkozott, hogy
egyébre gondoljon.
Ezután se utasította el a szegényeket, a kik alamizsnáért
folyamodtak hozzá; de a kuruzslókat, a kik ápolták, aranynyal fizette.
Multak az esztendők; megöregedett. És egész élete abban telt el,
hogy folyton-folyvást oltalmazta földi lételét. Nem csinált egyebet,
csak óvakodott a másvilágtól: napról-napra, reggeltől estig, egyedül
azon mesterkedett, hogy minél később kerüljön oda.
Már jóval túl járt hatvanadik évén, a mikor harmadszor is
megbetegedett. Halálos reszketés fogta el, homlokáról patakzott a
verejték, és még a suttogás is elakadt a torkán. Mindenből úgy
látszott, hogy meghal. A tudósok már nem is csóválták a fejöket;
csak fölemelték szemöket az égre. Papjai összegyűltek ágya körül és
sirtak.
Seraphino tudta kötelességét. Összeszedte minden erejét, és
érczes, tisztán hallható hangon szólt hozzájok:
– Ne sírjatok! Az Úr azt rendelte, hogy elhagyjam ezt a siralom
völgyét… legyen áldott szent akarata! Ne sírjatok, hiszen az örök
életbe költözöm át! Hallgassátok meg utolsó szavamat:
Gondjaitokba, buzgó gondjaitokba ajánlom anyaszentegyházunkat,
lelketekre kötöm, hogy tántoríthatatlan hűséggel, teljes odaadással
szolgáljátok! Lángoló buzgalommal terjesszétek az üdvözítő hitet,
óvjátok a hívek erkölcsét, világosítsátok föl a tévelygőket és
imádkozzatok bűnös lelkemért!
De bensejének legmélyén így könyörgött:
– Rex regum, dominus dominantium! Mindenhatóságodnak
zsámolyánál, esengve, rimánkodva könyörgök: hosszabbítsd meg
nyomorult életemet még egy évvel… egy hónappal… egy héttel…
egy nappal… egy órával!…

Aztán érezte, hogy öntudata elhagyja és azt hitte, hogy meghal…
De nem halt meg, csak elveszítette az eszméletét.
S mialatt papjai a haldoklókért szóló imát mormolták körülötte,
Seraphino, lázas álmában azt képzelte, hogy immáron egy szűk és
nagyon homályos czellában fekszik, melyben csak annyi világosság
derengett, a mennyit egy ajtóhasadékon át a kívül égő nagy tűz
sugározhat be a sötét szobába. Arrafelé nézett, a honnan a kevés
világosság derengett, s a hogy a szeme jobban hozzászokott a
sötétséghez, látta, hogy a szűk czella csak háromfalú, az egyik
oldalán nyilt, de kívül is vakhomály ül a közelben és a tűzfény csak
valahonnan messzebbről jöhet.
– Hol vagyok? – suttogta Seraphino.
Mintegy feleletül megzendült a vezeklők kara. Nem tudta kitalálni,
honnan hallatszanak a borzadályt keltő hangok, de tisztán hallotta
ezt a szót:
– Miserere!
Megrázkodott. Valami különös nyugtalanság, majd vad félelem
fogta el.
És a láthatatlanok kara mind erősebben, mind rémesebben zúgta
a korált… harsonaszó hallatszott, a mely egyre siketítőbben
hangzott, átviharzotta a czellát és majdnem megrepesztette a fülét…
Kétségbeesetten orditott föl… s a hangok orkánja hirtelen elült.
A csenddel visszanyerte az eszét is; valamelyest megnyugodott.
Tehát a másvilágon van. Az igazat megvallva, egy kissé meg volt
lepetve. No lám, csakugyan nem szűnt meg teljesen; a végesből
csakugyan a végtelenbe lépett át! Nagyot lélekzett; egy kő esett le a
szívéről. Akármilyen határtalan a bizalom, a bizonyosság mégis csak
jól esik. Odalenn a legigazabbakat is nyugtalanná teszi az, hogy
vaksötétbe kell ugraniok. A mit a gondviselés eltakart a földi szem
elől, ismeretlenségével mégis csak izgatja egy kevéssé a halandó
képzeletét, akármilyen erősen bízzék is az égi irgalmasságban. S a

milyen izgató az ismeretlentől való oktalan félelem: olyan vég nélkül
megnyugtató a földöntúli béke, a lélek előtt megnyilt bizonyosság.
Milyen oktalan volt, hogy annyit rettegett a megígért, várt, remélt,
de ismeretlen világtól! És valami nagyon édes érzés járta át.
Hanem a mint tovább nézelődött maga körül, öröme egyre
halványabbra vált; a vakhomály, melyen tekintete nem birt keresztül
hatolni, megdöbbentette, s a végtelen csend félelemmel töltötte el.
Azután valami eddig még nem ismert szorongás szállotta meg. Úgy
tetszett neki, mintha köröskörül rémes árnyékok imbolyognának,
mintha égő öntudatát jéghideg fuvallat járná át, s mintha lelkének
mélyén egy szörny ütné fel a fejét. Újra rettegés fogta el; kiáltani
szeretett volna, de nem tudott; mozdulni akart, de nem birt. És
hirtelen átsajgott rajta a rémítő tudat:
– A hol vagyok, az nem az igazak örök hazája! Az a tűzfény a
tisztító tűz fénye!
Tekintete oda meredt, a hol a kevés világosság derengett, és
ekkor úgy rémlett neki, mintha a tűzfény erősödnék. Ez a látvány
megriasztotta; nem akart oda nézni többé, de nem tudta elvenni a
tekintetét. A czella nyílása pedig egyre világosabb lett. A tűzfény
egyre sűrűbb kévékben ömlött a sötét falak felé; a nyílás egészen
kivilágosodott; rettenetes, tűrhetetlen hőséget terjesztő fényesség
öltögette be vörös és fehéren izzó, útálatos lángnyelveit a padló, a
falak és a boltozat felé; végül egy lángtenger hömpölygött odakünn.
Könyörögni akart, hogy ne kelljen látnia ezt a rémséget. De
elfelejtette összes imádságait, és érezte, hogy már gondolkozni se
bir. Csak megkövesült tekintetében maradt még némi élet.
És látta, hogy a lángtengerben ördögök futkosnak, a hátukon
nagy puttonnal, a melyben a lelkeket vitték. Nehezen czipelték
terhöket, és igen fáradtak lehettek, mert pántlika-alakú, vérvörös
nyelvök ocsmány szájukból egész hosszúságában kilógott. Egy ördög
éppen a czella nyílása mellett vonszolta terhét, s a puttonban ülő

kárhozott bevigyorgott Seraphinóra. Seraphino dermedten bámult rá:
megismerte Bonifácz püspököt.
Olyan iszonyat szállotta meg, a minőt a földön soha se ismert.
Érezte, hogy teljesen elveszti az eszét és oktalanságában menekülni
akart. De nem birt fölemelkedni fekvőhelyéről; mocczanni se tudott.
Ekkor eszébe jutott Jézus Krisztus neve. Segítségül híván az isten
fiának végtelen kegyelmét, kétségbeesésében még egy utolsó
erőfeszítést tett; és tagjai most már engedtek akaratának, teste
összerázkodott, térdének fagya fölengedett, törzse megmozdult, feje
felszökkent fekvőhelyéről, és hátrafelé fordult, a menekvést keresve.
Csak ekkor látta, hogy valaki mozdulatlanul áll fekvőhelye
mögött: egy hófehér lepelbe öltözött, gyönyörű arczú ifju, a ki
végtelen nyugalommal nézett rá. Az arcza csodaszép volt, de a
tekintete oly merő, hogy Seraphino ezt a tekintetet nem birta
kiállani.
Visszabujt helyére, hogy ne lássa a szelíd és mégis vérfagyasztó
szempárt, összekuczorodott, a mennyire tudott és fogvaczogva
kérdezte:
– Angyal vagy?
– Én a Halál vagyok – szólalt meg amaz a feje mögött. – Én
hoztalak ide, hogy nemsokára átadjalak az örök Igazságnak.
– Hol vagyok? – rebegte Seraphino.
– A pokolban – felelt a Halál.
– Mennyei irgalmasság! – nyöszörgött Seraphino – Miért hoztál
ide? Csak nem maradok itt?!
– Itt maradsz örökkön – szólt a Halál – innen nincs visszatérés!
– Istennek szent anyja, könyörögj érettem! – nyögte Seraphino –
hiszen szándékaim tiszták valának!

– Szándékaid – szólt a Halál – benn voltak az égi serpenyőben,
de porszemnek itéltettek.
Seraphino kétségbeesetten tördelte a kezét.
– Quid tunc miser sum dicturus?! – tépelődött. – Mea culpa, mea
maxima culpa! A serpenyőben talán úgy találtatott, hogy
bizodalmam nem volt elég erős. De hiszen hűségesen hittem,
követtem, hirdettem a szent tanítást!… Életemet imádkozással
töltöttem el és nem szűntem meg könyörögni, hogy a mennyei
irgalmasság erősítse meg bizodalmamat!… Cselekedetem igaz volt,
mert mindent elkövettem, hogy megszilárdítsam másoknak a hitét,
szándékom tiszta, mert nem volt egyéb igyekezetem, mint hogy
egészen megtisztuljak bizodalmamban, még gondolatom is szenny
nélkül való, mert végre is hívő voltam, jóllehet ez nekem nehezebb
volt, mint másnak!
– Nem azért kerültél ide – szólalt meg újra a Halál. – Ha netalán
vétkeztél ily gondolatokkal, bűnöd magától bocsánatot nyert. Az égi
irgalmasság nem itéli el azt a lelket, a mely tökéletlenségében,
oktalanságában, vakságában, nem éri föl észszel az ő végtelenségét,
mindenható voltát és mindenütt jelenvalóságát. A midőn a
megismerést elzárta a halandó elől, már előre megbocsátotta az
emberi értelem minden megtévedését.
– De hát akkor miért kerültem ide?! – kínlódott Seraphino a
kétségek között. – Hiszen nem voltak halálos bűneim! Nem
vétkeztem a tíz parancsolat ellen a hét főbűntől tiszta maradtam,
égbekiáltó vétket nem követtem el! Miért jutottam mégis a pokolba?
Mi volt a bűnöm?
– A legnagyobb bűn! – felelt a Halál. – A bűnök anyja!… az a
bűn, a melyben az összes bűnök bennfoglaltatnak: az önzés!
– Az ember eredendő bűnben fogantatik – mentegetőzött
Seraphino – de Jézus Krisztus mindnyájunkat megváltott!

– Nyiljék meg valahára a szemed! – szólt a Halál. – A te vétked
nem az eredendő bűn volt, hanem a legnagyobb vétek: az életnek,
az Isten mindenhatóságából kisugárzó erőnek gonosz eltékozlása. Az
életben egy megbízást kaptál az Istentől: te ezzel a megbízással
visszaéltél. Életednek rendeltetése volt: és te elvontad életedet a
rendeltetésétől. Életednek egy czélt kellett volna szolgálnia: te
életedet elsikkasztottad ettől a czéltól, és csak arra használtad fel,
hogy minél hosszabbra nyújtsd, mert jól esett élned. Erényesnek
képzelted magad, mert látszatra megtartottad a szabályokat. De vak
voltál és nem láttad át, hogy mindenki annyit ér, a mennyit
másoknak tesz, – a mennyi áldozatra, a mennyi önfeláldozásra
képes, – a hányszor koczkára teszi jólétét és egész életét az
egyetemesért. Azt hitted, hogy látod az Istent; és nem láttad azt,
hogy az Isten nem óvja az egyén életét, csak az egyetemes életet
óvja. Azt hitted, hogy istenfélő vagy és ügyet se vetettél a törvények
törvényére. Hogy az istenségnek mik a czéljai az egyetemes élettel:
nem nyilatkoztatja ki a halandónak, nem mondta meg neked sem;
de megmondta, beleírta a lelkedbe, hogy az élet nem tulajdonod,
csak véges birtokod; hogy nem a te czéljaidra kaptad, hanem
rendeltetése van; kinyilatkoztatta előtted is, mert szíved megsugta,
mire való az életed; tudtad, miért, mi végre születtél s hogy a földi
pálya csak arravaló, hogy: közremunkálj az Ő czéljaira! Te nem
vétkeztél a szabályok ellen és értéktelen holmikat nem kapartál félre
magadnak, de eltulajdonítottad, elsikkasztottad, elprédáltad a
legbecsesebbet: azt, a mi egyedül isteni eredetű; nem loptál
felebarátjaidtól, de megloptad az Istent!
Ebben a pillanatban a lángtenger becsapott a czellába. A
vérvörös nyelvű, kaján pofájú ördögök berohantak, megragadták az
érseket, és kihurczolták a rettenetes tűzbe. Előbb nyársra húzták, és
testét, a melyet a földön annyira óvott a hidegtől, a tűzben
megpörkölték, aztán beledobták egy forró olajjal teli, roppant
medenczébe, a melyben ezer meg ezer kárhozott lélek visított és
üvöltött… Seraphino elordította magát, és fölébredt.
* * *

A verejték csorgott a homlokáról, de betegsége elmúlt. Még
aznap fölkelt, és környezete álmélkodva látta, hogy frissebb és
egészségesebb, mint valaha.
De még jobban csodálkozott mindenki, a mikor kitűnt, hogy az
érseket mennyire megváltoztatta az a betegség, a melyből csak
valamely isteni csoda gyógyíthatta ki.
Seraphino attól a naptól fogva, a mikor talpra állott, egészen más
ember volt, mint előbb.
Az első dolga az volt, hogy elbocsátotta összes kuruzslóit, és az
aranyakat, a melyekkel azelőtt szolgálataikat fizette, a szegények
között osztotta szét.
Aztán visszatért a városba, és megjelent az emberek között, hogy
segítségökre legyen.
Éppen akkor a pestis tizedelte meg Firenze lakóit. Seraphino
sorra járta a betegeket; ápolta, s vigasztalta őket, és feladta nekik az
utolsó szentséget. Később az Arno kiáradt; az ősz érsek csónakon
járta be az elárasztott városrészt, kinyújtotta megerősödött kezét a
vízbefúlók felé és partra szállította őket. Ellenség támadta meg a
várost; Seraphino beállt a harczolók közé. Tűzvész ütött ki, mely
elhamvasztotta Firenze legnépesebb épületeit; az érsek elsőnek
jelent meg a szerencsétlenség helyén, elsőnek rohant be az égő
palotákba és megmentett kilencz gyermeket.
Százszor tette koczkára életét, de a veszedelem, melyet
mindenütt keresett, elbújt előle. És száz évig élt.
Az emberek pedig, midőn látták, mire képes az az aggastyán, a
kit hatvan éves koráig haldoklónak ismertek, igen elcsodálkoztak.
Némelyek így suttogtak:
– Milyen erősnek kell lennie annak a hitnek, a mely készséget ad
ekkora önfeláldozásra!

Mások meg így:
– Ebben a hitben bizonynyal isteni erő van!
Akkortájt sokak lelkét hitetlenség szállotta meg, az a betegség, a
mely ragadósabb és gyilkosabb a pestisnél. Seraphino Istenhez
térítette őket, mert a hitetlenek, látván, hogy az ősz ember mind a
két karján siránkozó angyalokat hurczol ki a tűzből, önkéntelenül így
kiáltottak föl:
– Mégis van másvilág!

OSTOBA ÁGOST HALÁLA.
I.
Ostoba Ágost nem esett le a trapezről; kis fekete lova, a berber
Kara Musztafa, nem rugta homlokon; az a kártya-alkotmány se
szakadt le alatta, a melynek legtetején mint karnevál-király trónolt.
Ostoba Ágosttal az történt, hogy a mint egy szép tavaszi délután
csöndesen üldögélt az Artista-kávéházban, s unatkozva olvasta az
amerikai Művészélet híreit, egyszerre csak lefordult a székéről s
odavágta a fejét a padlóra.
Az Artista-kávéház délután mindig tele van emberrel. Ostoba
Ágost körül, a baleset előtt, egy sereg mindenféle fajta művész
zsibongott. Mellette egy csapat leány ült, a kik Bing impresszáriót, a
zsiványok királyát szidták, vagy a Bing hirdetéseitől elvakultan az
újvilág kincseiről beszéltek, álmodoztak.
Az asszonynép sikoltozva ugrált szerte, a mikor a szalonkabátos
hústömeg oda esett a kivágott czipős, formás lábak elé. A férfiak
ellenben oda rohantak, s a kávés, a kinek első gondolata az volt,
hogy az ijesztő látvány sokat árthat üzlete jóhírének, ordítva
intézkedett, hogy a holtra vált embert vigyék a kisebbik külön
szobába.
Ez azonban nem történhetett az óhajtott gyorsasággal, s az
utczáról sokan meglátták a jelenetet. Egész horda tódult be a
kávéházba, s a csaposlegények, a kik Ostoba Ágostot az ablaktalan
fülkébe átszállították, alig tudtak maguknak utat törni. Valahonnan
egy orvos is előkerült.
A külön szoba ajtaját becsukták, s míg a tíz ember közt szorongó
orvos a gutaütöttet élesztgette, a rejtelmesen zárt ajtó körül ijedt,

de azért nem kevésbbé lármás tanácskozás indult meg. Ez a
tanácskozás nem arra irányult, hogy a hirtelen érdekessé vált
szalonkabátba miképpen lehetne új életet önteni, hanem hogy a
hasonló sorstól hogyan óvja magát, a ki még ráér óvakodni.
Mindenki azon tanakodott, hogy mi okozhatta a szerencsétlenséget,
mintha különös események és egyéni hibák nélkül soha se halna
meg az ember.
Csak azok hallgattak, a kik tegnap este még vígan pofozkodtak
vele a porondon. De ezekről az elfehéredett arczokról szintén ezt a
kérdést lehetett leolvasni:
– Mi történhetett vele? Tiz évvel ezelőtt még gondtalan
biztonsággal ugrott egyik trapezről a másikra, sohase szédült. És
most, tessék, nem tud megállni a lábán! Mi történhetett vele?
Mikor a külön szoba ajtaja kinyilt, ez a szó futott végig a
kávéházon:
– Él!
Mindenki megkönnyebbedett s mikor szájról-szájra járt, hogy az
orvos a beteg fölépülését reméli, a kávéház visszakapta rendes
képét. Este, a czirkuszban, még sokat beszéltek arról, hogy kár volt
Ágostnak ifjú éveiben oly sűrűn iddogálni, de ezzel le is tárgyalták az
esetet.
A következő napok mintha az orvosnak adtak volna igazat. Ágost
jobban lett, hat hét mulva elhagyhatta a kórházat, megint tudott
járni és beszélni. De méltóztassék elhinni nekem, hogy akármit
mondott az orvos, Ostoba Ágost végérvényesen meghalt ott, az
Artista-kávéház külön szobájában.
II.
A tanár, miután alaposan megvizsgálta Agostot, így szólt hozzá:

– Tartsa meg azt az életrendet, a melyet megszabok. Ne egyék
húst, ne egyék tésztát, ne egyék fűszerest, ne egyék czukrost, ne
egyék sokat, ne egyék keveset. Ne egyék levest, a leves fölösleges
teher a gyomornak. Ne egyék főtt húst, az nem tápláló. Ne egyék
olyat, a mi izgatja, s ne egyék jóllakásig. Nem szabad bort inni, nem
szabad szivarozni. Tartózkodjék minden izgalomtól, értse meg:
minden izgalomtól. Mozogjon, de ne erőltesse meg magát; el ne
tespedjen, de valahogy eszébe ne jusson ugrándozni. A fő: a jó
levegő. Ne járjon kávéházba, feküdjék le korán, s tartózkodjék sokat
a szabadban. De vigyázzon, mert az esti levegő nem tesz jót önnek.
Menjen délvidékre, minél előbb, de akárhol lesz, hat óra után csukja
be az ablakot. A fő az, hogy: mozogjon, de ne mozogjon sokat.
– Igenis – szólt Ágost.
Egy kicsit megbutult a sok jótanácstól; hirtelen nem tudta, hogy
mit szabad hát csinálnia? Hogy némiképpen kiigazodjék,
megkoczkáztatta ezt a kérdést:
– És a czirkusz?
A tanár mosolygott.
– Arról egyelőre nem lehet szó.
Ez az egyelőre folyton ott hangzott Ágostnak a fülében.
Megértette, hogy nemsokára meg fog halni, s egyébre nem tudott
gondolni.
Őrült félelem fogta el. Napokat töltött folytonos rettegésben, s
folyton az utczát járta, mintha, a míg talpon van, nem érné utól a
nyomában járó veszett kutya. Néha a sok járásban halálosan
elfáradt, s csak mikor nem birta tovább, vette észre, hogy áthágta az
orvos rendeletét. Ilyenkor azt képzelte, hogy ezzel a sok-sok ezer
lépéssel csak még közelebb jutott a halálhoz; figyelni kezdte magán
káros cselekedete hatását, s szíve vad kalapálása úgy elrémítette,
hogy ordítani szeretett volna.
Mihelyt fölébredt, az volt az első gondolata:

– Még élek.
De a második már az:
– Egy nappal kevesebb van hátra.
Sajnálta az időt, a melyet alvással töltött el. Azalatt nem élt; azzal
is kevesebb. Éjjel pedig kétségbeejtette, hogy nem tud elaludni.
Segítségért kiabált volna, ha nem szégyenli magát; s felköltötte a
szomszédait: tegyenek vele valamit, hogy el tudjon aludni.
Ágost szeretett volna még sokáig élni. Borzadva gondolt a sírra, a
koporsóra, a férgekre. Mért nem hagyják a halottakat a szabadban,
a napon? Mért nem adnak nekik legalább akkora helyet, mint egy
vasúti kocsi? Mért nem védik őket tudományos szerekkel az undok
állatoktól, a míg csak egy parány van belőlük? De a legborzasztóbb
mégis az: nem lenni!… A nap süt az égen, a virágok nyílnak, a
lányok szeretkeznek, az utczákon vidám zsivaj közepett aczélizmú
fiatal fiúk biczikliznek, a villámos csilingel, az emberek sietnek a
czirkuszba, él ló, kutya, hernyó, légy, csak ő nem él, s azzal, hogy ő
nem él, úgy nem törődik senki, mint a hogy nem törődnek egy
eltaposott bogárral, a mely volt, vagy nem volt, mindegy!…
Ötvennyolcz éves volt, de vannak kilenczven éves emberek is.
Mért nem jutott neki is annyi, mint ezeknek a szerencséseknek?
Irigyelte az öregeket: mennyivel többet értek el, mint ő!… S irigyelte
a fiatalokat: ezeknek még milyen sok van hátra!… És úgy nézett az
emberekre, mintha megrabolták volna a legféltettebb kincsétől.
Mióta tudta, hogy nagybeteg, nem szeretett senkit. Ha férfival
kellett beszélnie, az járt az eszében, hogy ennek a szavai elcsennek
valamit az idejéből; ennek könnyü, de neki, szegény fejének! Ha
lányt látott, gondolatában gyűlölettel kérte számon a jövő szerelmeit,
a csókokat, a melyeket ez a leány akkor ad, mikor ő már por, hamu,
föld, semmi.
Haragudott az emberekre és kerülte őket.

Egyszer az orvos, a ki észrevette, hogy az állapot a szokottnál is
rohamosabban romlik, azt mondta neki:
– Ne emészsze magát. Szórakozzék egy kicsit, járjon emberek
közé.
Sietett engedelmeskedni, de olyan félénken lépett a beszélgetők
körébe, mintha bocsánatot kért volna tőlük, hogy születni
bátorkodott. Alázatosan hallgatta őket s bár nem érdekelte, a mit
mondottak, csak akkor vált meg tőlük, a mikor magára hagyták.
Beléjök ragaszkodott, mintha a körülötte levő életben rejlenék az ő
élete is, s elevenségük, mozgékonyságuk, függetlenségük láttán
mindenkinél alábbvalónak érezte magát. Összehasonlította magát
velök, s arra az eredményre jutott, hogy őt már talán a hülyeség
környékezi. Gyógyító szernek használta a társaságot, de ennek a
gyógyító szernek a használata után még többet szenvedett, mint
előbb.
A czirkuszba nem igen mert elmenni, mert ahányszor arra járt, az
irigység és emlékei nagyon fölizgatták: mindig az érütését számlálva
tért onnan haza. Tudta, hogy sohase léphet föl többé, de
beletörődött ebbe is, csakhogy valamivel tovább élhessen. Volt egy
kis megtakarított pénze; mindezt rászánta, hogy a mennyire lehet,
meghosszabbíthassa az életét. Elment délvidékre, s hat óra után
becsukta az ablakot, lefeküdt korán, mozgott, de nem mozgott
sokat, nem ivott, nem szivarozott, szabály szerint evett, nem gondolt
egyébre, csak az egészségére, s miután minden orvosi rendeletet
pontosan megtartott, a nagy vigyázat izgalmai következtében a
második gutaütés sokkal korábban következett el, mint az orvos
várta.
III.
A harmadik gutaütés alkalmával Ostoba Ágost halálát már az
orvosnak is el kellett ismernie. A halottkém kiállította a bizonyítványt,
az anyakönyvvezető följegyezte az esetet, az ujságok néhány
Á

jóakaró sort szenteltek a kedvelt pojácza emlékének, s Ostoba Ágost
akárhogy kapálódzott előbb, alaposan belenyugodott sorsába.
Ostoba Ágost ekkor már senkinek nem hiányzott. Családja nem
volt, s a czimborái régen megbarátkoztak azzal a gondolattal, hogy
Ostoba Ágostról le kell mondaniok. A legtovább Kara Musztafa
nyihogott gazdája után, mert a kis berber csődörrel más nem tudott
bánni, csak a bohócz, s mikor Ágost megbetegedett, a tudós
lovacska leszorult a porondról. De azóta Kara Musztafa is
beletörődött a megváltozott életbe, s Kara Musztafa számára Ágost
nem akkor halt meg, a mikor gazdája szíve nem vert többet.
A czirkuszban helyét már régen betöltötték, s az új Ostoba Ágost
egy év óta közkedveltségnek örvendett. Mégis a temetés után az
Artista-kávéházban sokat beszéltek a régi pályatársról.
A fiatalabb emberek úgy találták, hogy végre is eleget élt. Az
öregebbek: hogy nem folytatott észszerű életmódot.
S mindenki tudott róla egy-egy kis esetet, mely élénken
jellemezte, mennyire félt mindig a haláltól.
De erről senki se tudott többet, mint az egyik istállómester, a ki
Ágostot betegségében a leggyakrabban látogatta meg. Ez az
istállómester volt az, a kit hajdan, mikor Ágost még harsogó
derültségeket fakasztott, a fárasztó lovaglások szüneteiben, a
közönség nagy vigasságára, addig pofozott, míg nem az
istállómester dőlt ki, hanem Ágost esett össze a fáradtságtól.
Az istállómester sok részletet tudott a beteg pojácza
halálfélelméről, s beszélgetés közben utánozni kezdte azokat a
vadállati hangokat, melyeken a bohócz, mikor már a végét járta s a
betegségtől bárgyuvá lett, részvevő látogatója előtt elbőgte a
haláltól való irtózását.
Az istállómester olyan mulatságosan utánozta ezeket a hangokat,
hogy mindenki nevetett.

Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.
More than just a book-buying platform, we strive to be a bridge
connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.
Join us on a journey of knowledge exploration, passion nurturing, and
personal growth every day!
ebookbell.com