0% found this document useful (0 votes)
14 views46 pages

Computer Arithmetic and Validity Theory Implementation and Applications 2ed. Edition Ulrich Kulisch

The document promotes various ebooks available for download at ebookname.com, including titles on computer arithmetic, experimental design, and biophysics. It highlights the importance of reliable computing and interval arithmetic in modern scientific computing, referencing the development of the IEEE P1788 standard. The foreword and preface discuss the evolution of computer arithmetic and the need for improved reliability in numerical computations.

Uploaded by

morseyvho
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
14 views46 pages

Computer Arithmetic and Validity Theory Implementation and Applications 2ed. Edition Ulrich Kulisch

The document promotes various ebooks available for download at ebookname.com, including titles on computer arithmetic, experimental design, and biophysics. It highlights the importance of reliable computing and interval arithmetic in modern scientific computing, referencing the development of the IEEE P1788 standard. The foreword and preface discuss the evolution of computer arithmetic and the need for improved reliability in numerical computations.

Uploaded by

morseyvho
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 46

Endless Ebook, One Click Away – Start Downloading at ebookname.

com

Computer arithmetic and validity Theory


implementation and applications 2ed. Edition
Ulrich Kulisch

https://ebookname.com/product/computer-arithmetic-and-
validity-theory-implementation-and-applications-2ed-edition-
ulrich-kulisch/

OR CLICK HERE

DOWLOAD EBOOK

Browse and Get More Ebook Downloads Instantly at https://ebookname.com


Click here to visit ebookname.com and download ebook now
Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Experimental design unified concepts practical


applications and computer implementation First Edition
Bowerman

https://ebookname.com/product/experimental-design-unified-
concepts-practical-applications-and-computer-implementation-
first-edition-bowerman/

Arithmetic and Logic in Computer Systems 1st Edition Mi


Lu

https://ebookname.com/product/arithmetic-and-logic-in-computer-
systems-1st-edition-mi-lu/

Computer Arithmetic and Verilog HDL Fundamentals 1st


Edition Joseph Cavanagh

https://ebookname.com/product/computer-arithmetic-and-verilog-
hdl-fundamentals-1st-edition-joseph-cavanagh/

Epithelial Cell Culture Protocols Second Edition Amanda


Simcox (Auth.)

https://ebookname.com/product/epithelial-cell-culture-protocols-
second-edition-amanda-simcox-auth/
Alternative Exchanges Second Hand Circulations from the
Sixteenth Century to the Present 1st Edition Laurence
Fontaine

https://ebookname.com/product/alternative-exchanges-second-hand-
circulations-from-the-sixteenth-century-to-the-present-1st-
edition-laurence-fontaine/

Vision The Approach of Biophysics and Neurosciences


Proceedings of the International School of Biophysics
Casamicciola Napoli Italy 11 16 October on Biophysics
and Biocybernetics v 11 1st Edition Carlo Musio
https://ebookname.com/product/vision-the-approach-of-biophysics-
and-neurosciences-proceedings-of-the-international-school-of-
biophysics-casamicciola-napoli-italy-11-16-october-on-biophysics-
and-biocybernetics-v-11-1st-edition-carl/

Using Evidence of Student Learning to Improve Higher


Education 1st Edition George D. Kuh

https://ebookname.com/product/using-evidence-of-student-learning-
to-improve-higher-education-1st-edition-george-d-kuh/

The Case for Israel 1st Edition Alan Dershowitz

https://ebookname.com/product/the-case-for-israel-1st-edition-
alan-dershowitz/

Physics in the Arts Complementary Science 1st Edition


P.U.P.A. Gilbert

https://ebookname.com/product/physics-in-the-arts-complementary-
science-1st-edition-p-u-p-a-gilbert/
Science of Pain 1st Edition Allan I Basbaum

https://ebookname.com/product/science-of-pain-1st-edition-allan-
i-basbaum/
De Gruyter Studies in Mathematics 33

Editors
Carsten Carstensen, Berlin, Germany
Nicola Fusco, Napoli, Italy
Fritz Gesztesy, Columbia, USA
Niels Jacob, Swansea, United Kingdom
Karl-Hermann Neeb, Erlangen, Germany
Ulrich Kulisch

Computer Arithmetic and


Validity

Theory, Implementation, and Applications

2nd Edition

De Gruyter
Mathematical Subject Classification 2010: Primary: 06, 65, 68, 94; Secondary: 65G, 65Y,
68M, 94C.

Additional keywords: order, semimorphism, rounding invariant structures, floating-point arith-


metic, interval arithmetic, exact dot product, reliable computing, computing with guarantees,
enclosure methods.

ISBN 978-3-11-030173-1
e-ISBN 978-3-11-030179-3
Set-ISBN 978-3-11-030180-9
ISSN 0179-0986

Library of Congress Cataloging-in-Publication Data


A CIP catalog record for this book has been applied for at the Library of Congress.

Bibliographic information published by the Deutsche Nationalbibliothek


The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie;
detailed bibliographic data are available in the internet at http://dnb.dnb.de.

© 2013 Walter de Gruyter GmbH, Berlin/Boston

Typesetting: P TP-Berlin Protago-TEX-Production GmbH, www.ptp-berlin.de


Printing and binding: Hubert & Co. GmbH & Co. KG, Göttingen
Printed on acid-free paper

Printed in Germany

www.degruyter.com
Lasset uns am Alten,
so es gut ist, halten
und dann auf dem alten Grund
Neues schaffen Stund um Stund.

House inscription in the Black Forest.


Foreword to the second edition

When the first edition of this book was published interval arithmetic or computing with
guaranties was a subject for specialists. But only a few months after publication of the
book the IEEE Computer Society founded a new standard committee IEEE P1788
for interval arithmetic in August 2008. It says: The aim of the standard is to improve
the availability of reliable computing in modern hardware and software environments
by defining the basic building blocks needed for performing interval arithmetic. This
event has moved the contents of this book right into the center of interest and of sci-
entific computing. Work on the standard began immediately and is still in progress. A
major number of scientists from all over the world are participating in its development.
The work is now approaching its end. So there is great hope that future processors will
better and uniquely support interval arithmetic and computing with guarantees.
Following the contents and the intentions of the book, the author has prepared and
presented several motions to the standard committee IEEE P1788. They have all been
accepted. These motions include the basic arithmetic operations for floating-point
intervals, the comparison and lattice operations, an exact dot product and complete
arithmetic, rounding control for floating-point operations and from complete format
to floating-point and interval formats, denotations for the fundamental spaces of in-
terest, and for the central operations of Kaucher arithmetic. This is an extension of
conventional interval arithmetic by so-called improper intervals with great potential
for applications in computer geometry.
In the older and classic books of interval arithmetic by R. E. Moore [457, 458],
G. Alefeld and J. Herzberger [28, 29], or E. Hansen [219, 222] real intervals are de-
noted by upper case letters and the bounds by lower case letters where the lower bound
carries an index 1 and the upper bound an index 2, for instance, A D Œa1 , a2 . This no-
tation is well established in interval arithmetic and it is used throughout this book. We
mention this here because a couple of years ago a group of scientists recommended
denoting a real interval by a bold face lower case letter, the lower bound by under-
lining and the upper bound by overlining the letter, for instance, a = Œa, a. Although
the draft of the standard P1788 currently uses this notation, we do not follow this rec-
ommendation in this book and prefer the older notation. There are good reasons for
doing this: It is difficult to use the new notation in handwritten texts. Furthermore, the
contents of the book are not, unlike the standard P1788, restricted to real intervals.
It also considers intervals over the complex numbers and intervals over vectors and
matrices of real and complex numbers. Here the older notation allows to stay closer
to conventional notations of mathematics. The index notation also is more suited for
viii Foreword to the second edition

proving several theorems in the book. Further, an improper interval is just a pair of
real numbers where the first component may be greater than the second. Here, under-
and overlining the components may be a constant source of confusion. After all, num-
bering the components needs fewer key strokes than under- and overlining.
The labor of preparing and writing the new edition of the book has been enormous.
Minor changes had to be made throughout the entire book. The development of the
standard IEEE P1788 brought different notations and the time and communications
with members of the standard committee P1788 brought new insights. Thus large por-
tions of several chapters had to be newly written.
I wish to express my deep appreciation to those many colleagues who have com-
mented on my contributions to the development of the standard IEEE P1788. Although
I did not follow his suggestions completely, I owe particular thanks to Stefan Siegel
for his help rewriting Chapter 7 on hardware support for interval arithmetic. Klaus
Braune again was a great help whenever problems with the LATEX system appeared. I
also again thank Bo Einarsson for proofreading great portions of the book. My thanks
also go to the publisher for asking me in time to prepare a second edition of the book.
So the work could be done without getting under too high pressure.

Karlsruhe, November 2012 Ulrich W. Kulisch


Preface

In general, a computer is understood to be an engine that deals with data. These are
stored by binary digits or bits. The data can represent all kinds of objects such as
whole numbers, letters of an alphabet, the books of a library, the accounts of a bank,
the inhabitants of a town, the text of a law, or the features of a disease. If the com-
puter is functioning correctly, then its processing is free of error. A computer that is
programmed properly is therefore generally considered to be an infallible tool.
It is quite ironical that this view is often not justified when the digital computer is
used for its original purpose – numerical or scientific computation. This book treats
the arithmetic processing of real numbers in computers. It describes how the computa-
tional error can be reduced and avoided by special design of the computer’s arithmetic.
A real number like , for instance, consists of an infinite b-adic expansion, an infi-
nite string of digits. On a computer, however, only finite parts of these expansions can
be represented. Thus arithmetic for real numbers can only be approximated. For nu-
merical or scientific computing, floating-point numbers and floating-point arithmetic
have been used since the early days of electronic computers. A floating-point number
takes the form m  2e . Here m is the mantissa or significand and e is the exponent.
Computers designed and built by Konrad Zuse, the Z3 (1941) and the Z4 (1945),
were among the first computers that used the binary number system and floating-point
for number representation. In the Z4, a floating-point number was represented by a
32 bit word and arithmetic was used in a manner very similar to the data format single
precision on today’s computers. Over the years computer technology was permanently
improved. This permitted an increase of the word size.
In 1985 a standard for floating-point arithmetic was internationally adopted. It is
known as the IEEE 754 floating-point arithmetic standard. It specifies two data formats
for single precision of 32 bits and double precision of 64 bits. The format double
precision uses 53 bits to represent the mantissa and 11 bits for the exponent. A revision
of the standard IEEE 754, published in 2008, added another word size of 128 bits.
The IEEE floating-point arithmetic standard is undoubtedly thorough, consistent,
and well defined. It has been widely accepted and has been used in almost every proces-
sor developed since 1985. This has greatly improved the portability of floating-point
programs. Until today the most used floating-point format is double precision. It corre-
sponds to about 16 decimal digits. Floating-point arithmetic has been used successfully
in the past. Every computer user is trained and familiar with all details of floating-
point arithmetic including all its exceptions like underflow, overflow, 1, C1, NaN
(not a number), 0, C0, and so on. Seventy years of extensive use of floating-point
x Preface

arithmetic makes many users believe that this is the only reasonable way of using the
computer for scientific computing.
In the 1950s computers were able to perform about 100 floating-point operations
per second (flops). For comparison: With a mechanical or electrical calculator a man
or woman can execute about 1 000 arithmetic operations somewhat reliably in a day.
A computer now could do this in 10 seconds. So the step to the electronic computer
really meant a revolution in computing speed. In order to judge the reliability of com-
puted results, error estimates were derived for frequently used numerical methods in
the 1950s and 1960s. These are built upon estimates of the error of each single arith-
metic operation. However, already in 1965 computers were on the market (CDC 6600)
that performed 105 flops. At these speeds a conventional error analysis becomes ques-
tionable. It easily can be shown that for a system of two linear equations with two
unknowns computers even today deliver a result of which possibly not a single digit is
correct. Such results strongly suggest trying to use the computer more for computing
close bounds of the solution instead of murky approximations.
Dramatic advances in computer technology, in memory size, and in speed have
been made since 1985. Arithmetic speed has gone from megaflops (106 flops), to gi-
gaflops (109 flops), to teraflops (1012 flops), to petaflops (1015 flops), and it is already
approaching the exaflops (1018 flops) range. This is not just a gain in speed. A qual-
itative difference goes with it. At the time of the megaflops computer a conventional
error analysis was recommended in every numerical analysis textbook. Today the PC
is a gigaflops computer. For the teraflops or petaflops computer conventional error
analysis is no longer practical. An avalanche of numbers is produced when a petaflops
computer runs. If the numbers a petaflops computer produces in one hour were to be
printed (500 on one page, 1 000 on one sheet, 1 000 sheets 10 cm high) they would
need a pile of paper that reaches from the earth to the sun1 and back. In 2012 the
fastest computers are listed to perform about 10 petaflops. With these the pile would
grow from the earth to the sun in less than 3 minutes. This is hard to imagine. Light
takes about 8 minutes for this distance.2
Computing indeed has already reached astronomical dimensions! With increasing
speed, problems that are dealt with become larger and larger. Extending the word size
cannot keep up with the tremendous increase in computer speed. Computing that is
continuously and greatly speeded up calls conventional computing into question. Even
with quadruple or extended precision arithmetic the computer somehow remains an
experimental tool.
The capability of a computer should not just be judged by the number of operations
it can perform in a certain amount of time without asking whether the computed result
is correct. It should also be asked how fast a computer can compute correctly to 3, 5, 10
or 15 decimal places for certain problems. If the question were asked that way, it would

1 The distance between the earth and the sun is about 150  106 km.
2 The speed of light is ca. 300 000 km/sec.
Preface xi

very soon lead to better computers. Mathematical methods that give an answer to this
question are available for very many problems. Computers, however, are at present
not designed in a way that allows these methods to be used effectively. Computer
arithmetic must move strongly towards more reliability in computation. Instead of the
computer being merely a fast calculating tool it must be developed into a scientific
instrument of mathematics.
Mathematical analysis has designed algorithms that deliver highly accurate and
completely verified results. These algorithms compute with intervals of real numbers.
Intervals bring the continuum on the computer. An interval between two floating-point
bounds represents the continuous set of real numbers between these bounds.
This book deals with computer arithmetic in a more general sense than usual, and
shows how the arithmetic and mathematical capability of the digital computer can be
enhanced in a quite natural way. The work is motivated by the desire and the need to
improve the accuracy of numerical computing and to control the quality of computed
results.
As a first step towards achieving this goal, the accuracy requirements for the elemen-
tary floating-point operations as defined by the IEEE arithmetic standard, for instance,
are extended to the customary product spaces of computation: the complex numbers,
the real and complex intervals, the real and complex vectors and matrices, and the real
and complex interval vectors and interval matrices. All computer approximations of
arithmetic operations in these spaces should deliver a result that differs from the correct
result by at most one rounding. For all these product spaces this accuracy requirement
leads to operations which are distinctly different from those traditionally available on
computers. This expanded set of arithmetic operations is taken as a definition of what
is called advanced computer arithmetic in the book.
Central to this treatise is the concept of semimorphism. It provides a mapping princi-
ple between the mathematical product spaces and their digitally representable subsets.
The properties of a semimorphism are designed to preserve as many of the ordinary
mathematical laws as possible. All computer operations of advanced computer arith-
metic are defined by semimorphism.
The book has three antecedents:

(I) Kulisch, U. W., Grundlagen des numerischen Rechnens – Mathematische Be-


gründung der Rechnerarithmetik, Bibliographisches Institut, Mannheim, Wien,
Zürich, 1976, 467 pp., ISBN 3-411-015617-9.

(II) Kulisch, U. W. and Miranker W. L., Computer Arithmetic in Theory and Prac-
tice, Academic Press, New York, 1981, 249 pp., ISBN 0-12-428650-X.

(III) Kulisch, U. W., Advanced Arithmetic for the Digital Computer – Design of Arith-
metic Units, Springer-Verlag, Wien, New York, 2002, 139 pp., ISBN 3-211-
83870-8.
xii Preface

The need to define all computer approximations of arithmetic operations by semi-


morphism goes back to the first of these books. By the time the second book had been
written, early microprocessors were on the market. They were made with a few thou-
sand transistors, and ran at 1 or 2 MHz. Arithmetic was provided by an 8-bit adder.
Floating-point arithmetic could only be implemented in software. In 1985 the IEEE
binary floating-point arithmetic standard was internationally adopted. Floating-point
arithmetic became hardware supported on microprocessors, first by coprocessors and
later directly within the CPU. Of course, all operations of advanced computer arith-
metic can be simulated using elementary floating-point arithmetic. This, however, is
rather complicated and results in unnecessarily slow performance. A consequence of
this is that for large problems the high quality operations of advanced computer arith-
metic are hardly ever applied. Higher precision arithmetic suffers from the same prob-
lem if it is simulated by software.
During the course of the book it is shown that all operations of advanced computer
arithmetic can be provided on the computer if, beyond conventional floating-point
arithmetic, two additional features are made available by fast hardware:

I. fast hardware support for interval arithmetic and

II. a fast and exact multiply and accumulate operation or, equivalently, an exact scalar
product.

Fast hardware circuitries for I. and II. are developed in Chapters 7 and 8, respectively.
This additional computational capability is gained at very modest hardware cost. I. and
II., in particular, are basic ingredients of what is called validated numerics or verified
computing. With simple hardware circuitry, interval operations can be made as fast
as floating-point operations. The simplest and fastest way for computing a scalar or
dot product is to compute it exactly. By pipelining, it can be computed in the time
the processor needs to read the data, i.e., it comes with utmost speed. Besides being
more accurate, the new computer operations I. and II. greatly speed up the operations
of advanced computer arithmetic. This would boost both the speed of a computation
and the accuracy of its result. Advanced computer arithmetic opens the door to very
many additional applications. Compared with conventional computer arithmetic all
these applications are extremely fast.
This book has three parts. Part 1, of four chapters, deals with the theory of computer
arithmetic, while Part 2, also of four chapters, treats the implementation of arithmetic
on computers. Part 3, of one chapter, illustrates by a few sample applications how
advanced computer arithmetic can be used to compute highly accurate and mathemat-
ically verified results.

Part 1: The different kinds of computer arithmetic for the customary product spaces of
computation follow an abstract mathematical pattern and are just special realizations
Preface xiii

of it. In the book the basic mathematical properties are extracted into an axiomatic
approach to computer arithmetic. Abstract mathematical concepts are suited to focus
the understanding on the essentials.
Frequently mathematics is seen as the science of structures. Analysis carries three
kinds of structures: an algebraic structure, an order structure, and a topological or
metric structure. These are coupled by certain compatibility properties, for instance:
a  b ) a C c  b C c.
It is well known that floating-point numbers and floating-point arithmetic do not
obey the rules of the real numbers R. However, rounding is a monotone function.
So the changes to the order structure are minimal. This is the reason why the order
structure plays a key role for an axiomatic approach to computer arithmetic.
All the product spaces under consideration carry an order structure with respect to
which they are complete lattices. The computer representable subsets can be charac-
terized as complete sublattices. A rounding is defined as a monotone mapping of the
basic set into the computer representable subset. Arithmetic in the latter is defined by
a monotone, antisymmetric rounding applied to the result of the operation carried out
in the basic set. In case of interval spaces, the rounding is additionally upwardly di-
rected, i.e., it maps the correct result of the operation in the powerset onto the least
upper bound in the subset with respect to set inclusion as the order relation. The map-
ping principle just described is called a semimorphism. It is close to a homomorphism
between ordered algebraic structures.
Definition of computer arithmetic by semimorphism in the product spaces under
consideration does not necessarily directly lead to operations that are executable on
the computer. It is, however, shown in the book that in all cases the operations can be
reduced to computer executable operations via isomorphisms. These isomorphisms are
to be established in the mathematical spaces in which the actual computer operations
operate. This requires a careful study of the structure of these spaces. Their properties
are defined as invariants with respect to semimorphisms. These concepts are developed
in the first three chapters of the book.
The fourth chapter deals with interval arithmetic. In most of the preceding literature,
interval arithmetic is defined and considered for closed and bounded real intervals.
See, for instance the well-known books by R. E. Moore [457, 458], by G. Alefeld
and J. Herzberger [28, 29], or by E. Hansen [219, 222]. This leads to well defined and
well-known formulas for the arithmetic operations. Overflow and underflow, however,
can occur and need special treatment. Another approach is taken by SUN’s interval
Fortran [725]. Here, intervals are considered over the extended real numbers R :D
R [ f1, C1g and 1 and C1 are permitted as elements of unbounded intervals.
Then interval operations are to be defined for operations like 1  1, 0  1, or 1=1
which leads to unusual and unsatisfactory operations.
This book avoids the difficulties of both approaches. Interval arithmetic just deals
with sets of real numbers. A real interval is defined as a closed and connected set of real
xiv Preface

numbers.3 Since 1 and C1 are not real numbers, they cannot be elements of a real
interval. They can only be bounds of an interval. Formulas for the arithmetic operations
for bounded real intervals are well established in conventional interval arithmetic. It
is shown in the book that these formulas can be extended to unbounded real intervals
by a continuity principle. As a result, it is further shown in the book that for a bound
1 or C1 in an interval operand the bounds for the resulting interval can easily be
obtained from the formulas for bounded intervals by applying well-established rules
of real analysis for computing with 1 and C1. This new approach to real interval
arithmetic leads to an algebraically closed calculus which is totally free of exceptions.
It remains free of exceptions if the operations are mapped on a floating-point screen
by the monotone, upwardly directed rounding.

Part 2: In Part 2 of the book, basic ideas for the implementation of advanced computer
arithmetic are discussed under the assumption that the data are floating-point numbers.
Algorithms and circuits are developed which realize the semimorphic operations in
the various spaces mentioned above. The result is an arithmetic with many desirable
properties, such as high speed, optimal accuracy, theoretical describability, closedness
of the theory, and ease of use.
Chapters 5 and 6 consider the implementation of elementary floating-point arith-
metic on the computer for a large class of roundings. The final section of Chapter 6
contains a brief discussion of all arithmetic operations defined in the product sets men-
tioned above as well as between these sets. The objective here is to summarize the def-
inition of these operations and to point out that they all can be performed as soon as an
exact scalar product is available in addition to the operations that have been discussed
in Chapters 5 and 6.
Floating-point operations with directed roundings are basic ingredients of interval
arithmetic. But with their isolated use in software, interval arithmetic is too slow to
be widely accepted in the scientific computing community. Chapter 7 shows, in par-
ticular, that with very simple circuitry interval arithmetic can be made practically as
fast as elementary floating-point arithmetic. To enable high speed, the case selections
for interval multiplication (9 cases) and division (14 cases including division by an in-
terval that includes zero) are done in hardware where they can be chosen without any
time penalty. The lower bound of the result is computed with rounding downwards and
the upper bound with rounding upwards by parallel units simultaneously. Also the ba-
sic comparisons for intervals together with the corresponding lattice operations and
the result selection in more complicated cases of multiplication and division are done
in hardware. There they are executed by parallel units simultaneously. The circuits
described in this chapter show that with modest additional hardware costs, interval
arithmetic can be made almost as fast as simple floating-point arithmetic. Such high

3 In real analysis a set of real numbers is called closed if its complement is open.
Other documents randomly have
different content
INITIAL LETTER BY VESALIUS
(From the “Fabrica”, 1543)

114
CHAPTER TWELFTH
Contemporary Anatomists

Shortly after the publication of the Fabrica, great activity was


manifested in anatomic research, and numerous opponents and
critics of Vesalius appeared in the arena of science. The criticism of
such men as Jacobus Sylvius and John Dryander, while it was of a
violent type, was of much less importance than was that of
Eustachius, Columbus and Fallopius. Vesalius was not without his
partisans, of whom Ingrassias and Cannanus are worthy of mention.

Bartholomeus Eustachius
Eustachius was born at San Severino, a small city near Salernum,
about the year 1520. He studied anatomy in Rome and made
remarkable progress in this science. In the year 1562, as he informs
us in his Opuscula Anatomica, he was professor of medicine in the
Collegio della Sapienza at Rome. Like many other men of genius,
Eustachius died in poverty. In August, 1574, having been called by
the illness of Cardinal Rovere to Fossombrone, Eustachius died upon
the journey.

To Eustachius posterity is indebted for a series of splendid


copperplate engravings which were designed to illustrate the
anatomy of the human body. These plates, the handiwork of
Eustachius, and the first anatomical illustrations wrought in 115
copper, were completed in 1552, only nine years after the
first impression of the book of Vesalius. Unfortunately for himself,
and worse for medical science, Eustachius was unable to publish
them. If this magnificent atlas of anatomy could have been
published when completed, the anatomical discoveries of the
eighteenth century would have come two hundred years earlier.
Unfortunately the entire text of the work is lost. For one hundred
and thirty-eight years the Eustachian plates remained either in the
family of Pinus, an intimate friend of the anatomist, or were buried
in the Papal Library at Rome. When discovered they were presented
by Pope Clement XI. to his physician, Lancisi, who published them
with notes of his own, at Rome, in 1714. In 1740 they were issued
under the direction of Cajetan Petrioli. Four years later the edition by
Albinus appeared, which was republished in 1761. The anatomical
writings of Eustachius were published during his lifetime, in 1564. It
is upon his Tabulae Anatomicae that the fame of this wonderful man
is founded. If this work had been published in 1552, Eustachius
would have divided with Vesalius the honor of founding human
anatomy. The victim of circumstances, his name has been
overshadowed by that of Vesalius, to whom in some respects he was
superior. Deprived during life of his merited honors, Eustachius has
been awarded a goodly share of posthumous fame.

116
BRAIN AND NERVES BY EUSTACHIUS
(Reduced one-half)

117
MUSCLES BY EUSTACHIUS
(Reduced one-half)

Eustachius was the first anatomist to describe, with any degree of


accuracy, the tube which bears his name. We can truly say he
discovered it, since Alcmaeon dissected only the lower animals, and
was not an accurate observer, as his view that goats breathe
through the ears, amply testifies. Eustachius discovered the tensor
tympani and stapedius muscles, the modiolus and membranous
cochlea, and the stapes. The honor of the discovery of the stapes is
claimed for no less than five renowned anatomists, namely,
Fallopius, Ingrassias, Columbus, Colladus, and Eustachius. It is
unnecessary to discuss this disputed claim to priority. The truth
seems to be that the stapes was discovered by both Ingrassias and
Eustachius, each independently of the other. In 1546 Ingrassias
publicly demonstrated the little bone of the ear in his lectures 118
at Naples. Fallopius, after learning from an eyewitness that
Ingrassias had actually discovered and named the ossicle,
relinquished his claim to the discovery. Columbus and Colladus filed
their information at too late a date. Eustachius, as previously stated,
finished his anatomical plates in 1552. His seventh plate shows,
among other subjects, the auditory ossicles—malleus, incus and
stapes—and tensor tympani muscle. These objects are delineated as
taken from a human subject, and also from a dog.

Eustachius discovered the origin of the optic nerves, and the sixth
cerebral nerves. He gives excellent pictures of the corpora olivaria
and corpora pyramidalia; of the stylo-hyoid muscle; of the deep
muscles of the neck and throat; of the suprarenal capsules, and of
the thoracic duct. He also described the ciliary muscle. Eustachius
was the first anatomist who accurately studied the teeth and the
phenomena of the first and second dentition. In his researches he
employed magnifying glasses, maceration, exsiccation, and various
methods of injection.
Realdus Columbus
The first anatomical treatise containing an account of the lesser, or
pulmonary circulation, was the monumental work, De Re Anatomica,
libri xv., written by Realdus Columbus and sumptuously published at
Venice in the year 1559. This, however, was not the first printed
account of the lesser circulation. Six years prior to the publication of
the book of Columbus, the unfortunate Servetus, in a theological
treatise, described correctly the course of the blood in its 119
transit through the lungs. Tried for heresy, Servetus was
burned, together with all obtainable copies of his book. Although it
had been printed, the work was suppressed; hence it follows that
Columbus was the first to publish the great discovery. Of the life of
this anatomist we know but little. Born at Cremona, a small Milanese
village, the year of his birth is unknown. He died in 1559, while his
book was being printed. A few copies were finished before his
demise, since a copy belonging to the late Dr. George Jackson Fisher,
of Sing Sing, N.Y., contains the author’s own dedication to Pope Paul
IV., while in other exemplars, the dedication has been written by the
two sons of Columbus, and is addressed to “Pio IIII., Pont. Max”.
This prelate, on the death of Paul IV., on August 18, 1559, became
the head of the Church.

Some writers have held that the discovery of the lesser circulation
was not made by Columbus independently of Servetus, but that a
copy of the book of Servetus had drifted into Italy and had been
read by Columbus. There is no direct evidence to support this view.
When Vesalius was called to Madrid as physician to Charles the Fifth,
Columbus, in 1544, succeeded him in the University of Padua; two
years later he filled the anatomical chair at Pisa, and in 1546, Pope
Paul IV. called him to Rome. Here he spent the later years of his life,
engaged in teaching anatomy and in writing his book. For forty years
Columbus pursued his anatomical studies, and in that period he
dissected an unusually large number of bodies. Fourteen subjects
passed under his scalpel in a single year.
120
TITLE-PAGE OF COLUMBUS’S ANATOMY
(Reduced one-half)

Columbus frequently made experiments upon living animals. He was


the first to use dogs for such purposes, preferring them to swine.
Book XIIII. of the work of Columbus is upon the subject of
vivisection, De viva sectione. In this he tells us how to employ living
dogs in demonstrating the movements of the heart and brain, 121
the action of the lungs, etc. Columbus was the first anatomist
who demonstrated experimentally that the blood passes from the
lungs into the pulmonary veins. “When the heart dilates”, says
Columbus, “it draws natural blood from the vena cava into the right
ventricle, and prepared blood from the pulmonary vein into the left;
the valves being so disposed that they collapse and permit its
ingress; but when the heart contracts, they become tense, and close
the apertures, so that nothing can return by the way it came. The
valves of the aorta and pulmonary artery opening, on the contrary,
at the same moment, give passage to the spirituous blood for
distribution to the body at large, and to the natural blood for
transference to the lungs”.

Like Servetus, Columbus held to the idea of “spiritus”. Harvey was


the first physiologist who recognized the circulation as purely a
movement of blood. All before him assumed the existence of a
mixture of air and blood. Columbus, pupil and prosector of Vesalius,
like his great master, denied the existence of foramina in the cardiac
septum.

Gabriel Fallopius

122
GABRIEL FALLOPIUS
Gabriel Fallopius (1523-1562), of Modena, was a noted Italian
anatomist. In his twenty-fifth year he was made professor of
anatomy at Pisa. Although the span of his life was short, he will be
remembered always as the discoverer of the tubes which bear his
name. According to Fisher, Fallopius “described the ear more
minutely than had ever before been done. He discovered the little
canal along which the facial nerve passes after leaving the auditory;
it is still called the aquaeductus Fallopii. He demonstrated the fact of
the communication of the mastoid cells with the cavity of the
tympanum; and also described the fenestrae rotunda and ovalis. In
the treatment of diseases of the ear, he used an aural speculum, and
employed sulphuric acid for the removal of polypi from the meatus.
In some of his supposed discoveries he had long been anticipated;
for example, the tubes which bear his name were known and
accurately described by Herophilus, over three hundred years before
the Christian era, and also by Rufus of Ephesus, of whom 123
Galen speaks as the best anatomist of the second century.
Rufus refers to two varicose and tortuous vessels passing from the
testes (as the ovaries were called) to the cavity of the uterus.
Fallopius, however, gave a full account of their course, position, size
and structure. He cut into them and found them hollow, gave them
the name of tubae seminales, and posterity attached his name to
them, and in time came to a better comprehension of their true
function. This is not the only instance in the history of anatomical
discovery where the name of a person, not its discoverer, has been
given to an organ. Allusion has been made to Fallopius as a botanist;
a genus of plants, Fallopia, has been named in honor of him”.

Fallopius was appointed professor of anatomy at Pisa, in the year


1548; and later, at the instance of the Grand Duke of Tuscany,
Cosimo I., he received a professorship at Padua, as successor to
Vesalius. Besides the chair of anatomy and surgery and of botany, he
also held the office of superintendent of the new botanic garden in
that city. Fallopius remained in Padua to the day of his death, which
occurred in 1562. He was very properly succeeded by his favorite
pupil, Fabricius ab Aquapendente, who had been for some time
previously his anatomical demonstrator. His collected works, as
published in Venice, 1606, embrace twenty-four treatises distributed
in three folio volumes. Only one of his works was published during
his lifetime, namely, his Observationes Anatomicae, Venice, 1561,
which is considered one of his most valuable books, containing, as it
does, most of his discoveries and his animadversions on the works of
other anatomists.

This was written as a supplement to the anatomy of Vesalius, 124


for it follows the same order, passes upon the same subjects,
corrects the inaccuracies of the Vesalian treatise, and supplies what
is wanting. Throughout the work Fallopius treats Vesalius with great
respect, and never mentions him without an honorable title. Vesalius
wrote an answer to this work, entitled, Observationum Fallopii
examen, in which he acknowledges the courtesy of Fallopius, but, as
argument progresses, appears to be out of temper.

After the death of Fallopius it was thought that no successor except


Vesalius could be found competent to fill his place. Accordingly
Vesalius was chosen. The news of his appointment reached him
while he was returning from a pilgrimage to Jerusalem.
Unfortunately he was shipwrecked and perished, otherwise history
would have afforded an example of the master filling the chair of the
pupil.

John Philip Ingrassias


Ingrassias, who lived between the years 1510-1580, was a graduate
of the celebrated Paduan School. He described minutely the anatomy
of the ear, including the tympanum, fenestrae rotunda and ovalis,
the cochlea, the semi-circular canals, and the tensor tympani
muscle. His admiring pupils caused his portrait to be painted and
placed in the Neapolitan School, with this inscription:—“To Philip
Ingrassias, of Sicily, who, by his lectures, restored the science of true
Medicine and Anatomy in Naples, his pupils have suspended this
portrait as a mark of grateful remembrance”. Ingrassias was a
voluminous writer, his chief work being a treatise on 125
osteology, which was published twenty-three years after his
death. When the plague depopulated Palermo, in 1575, his devotion
was such as to earn for him the title of the Sicilian Hippocrates. Few
men have been more earnest workers in medical science. If his fame
as an anatomist has not equalled that of others, the cause is to be
sought in the multiplicity of competitors, not in lack of zeal and
ability.
INGRASSIAS
126
CHAPTER THIRTEENTH
Commentators and Plagiarists

Medical history furnishes numerous examples of literary theft. In


many instances an entire set of anatomical plates has been pirated
by unscrupulous publishers. In a few cases both text and plates have
been appropriated by medical authors. The most notorious example
of this form of theft was furnished by William Cowper (1666-1709),
an English surgeon and anatomist, who, having secured three
hundred copies of Bidloo’s set of one hundred and five anatomical
[26]
plates, in 1697 issued the work as his own. Cowper added a few
original illustrations to the book.

Vesalius suffered severely at the hands of the plagiarists. Pirated


editions of the Tabulae Anatomicae were printed in several cities,
chiefly in Germany. As regards the Fabrica, we may say that it has
been the fountain from which many anatomical writers have derived
practically all of their illustrations and much of their text.

The fame of the Fabrica soon spread throughout Europe. It was


published in Germany, in Holland and in England. An epitome of its
contents was issued in Latin, in 1545, by Thomas Geminus, or
Gemini, under the title:—Compendiosa totius Anatomiae delineatio,
aere exaratum per Thomam. Geminum. It contained forty of the
Vesalian plates, cut in copper, and was the first book issued 127
in England in which the roller printing process was employed.
It was dedicated to Henry the Eighth, and was embellished with
“one of the earliest and most curious of all extant engraved title-
pages”.

In 1553, Geminus issued a second edition, in which the text was


translated into English. This edition was dedicated to Edward the
Sixth, with a commendatory note, “To the gentill readers and
Surgeons of Englande”. Six years later the third English edition
appeared, which was inscribed to Queen Elizabeth. It contains the
first published portrait of the Queen. She is shown upon the
engraved title-page, and, strange to say, above her is another
queenly figure, with a pen in her right hand, a wreath on her left,
her foot resting on the globe, and styled Victoria.

Another English work on anatomy, which is filled with poor imitations


of Vesalius’s illustrations, is the Microcosmographia of Helkiah
Crooke, or Crocus, who was “Professor in Anatomy and Chirurgery”.
Its chief value rests in an elaborately engraved title-page, a part of
which shows Crooke giving a demonstration in anatomy in the
presence of the “Worshipfull Company of Barber-Chirurgeons”, in
London, early in the seventeenth century.

John Banister of Nottingham, in 1578, borrowed a few Vesalian


woodcuts for use in The Historie of Man, sucked from the sappe of
the most approved Anatomists and published for the Utilitie of all
Godly Chirurgians within this Realme.

Most of the host of translators, epitomizers, commentators and


imitators of Vesalius have passed into oblivion. A few of these 128
persons have possessed enough of individuality to deserve
recognition.
Juan Valverde di Hamusco, a Spaniard who was born about the year
1500, studied anatomy at Padua and later at Rome. His book,
Historia de la Composicion del Cuerpo Humano, was published at
Rome in 1556. It contains forty-two copperplates and an engraved
title-page. Although the author says he has used only the Vesalian
plates, his work contains several plates which are not to be found in
Vesalius’s writings. For example, Valverde shows a muskelmann with
his skin held in his right hand, the left grasping a dagger which may
have been used in the skinning process. Other original drawings
show the abdomen and intestines, a pregnant woman with the
abdomen opened, and illustrations of the superficial veins.

Valverde was physician to Cardinal Juan de Toledo, Archbishop of


Santiago, to whom the work is dedicated. The illustrations were
drawn by Gaspar Becerra and were engraved by Nicholas Beatrizet.
Valverde’s book went through several editions. It forms a landmark
in the medical history of Spain—a country which, for many years,
was behind other states of Europe in matters of science.

To name the list of anatomical writers who have derived their artistic
inspiration from the Fabrica would require much more space than is
at our disposal. It must suffice to say, that, for a period of two
centuries, nearly all treatises on anatomy contained illustrations
which were taken from the writings of Vesalius. With few exceptions,
these reproductions were little better than caricatures of the original
figures.

Of the numerous editions of the Fabrica there are three 129


which are highly prized, namely, the first one, 1543; the
second, issued in 1555, containing eight hundred and twenty-four
pages, with many changes in the text; and the 1725 edition of the
collected writings of Vesalius. The last named is a huge volume
which was published at Leyden under the supervision of Boerhaave
and Albinus, with the illustrations cut in copper by Jan
[27]
Wandelaar .
It contains the Fabrica, the Epitome, the Epistola de Radicis Chynae,
various anatomical treatises of a controversial character, and the
Chirurgia Magna which has been wrongly attributed to Vesalius.
Morley says of this book:—“After his death a great work on surgery
appeared, in seven books, signed with his name, and commonly
included among his writings. There is reason, however, to believe
that his name was stolen to give value to the book, which was
compiled and published by a Venetian, Prosper Bogarucci, a literary
crow, who fed himself upon the dead man’s reputation”.

130
CHAPTER FOURTEENTH
The Court Physician

Vesalius, having finished the Fabrica, intended to write a work on the


practice of medicine which should be based on pathology. He makes
mention of this in the preface of the Fabrica, and in numerous places
in the body of the book he describes the pathologic appearances
which he found in dissection.

Returning to Padua after a year’s absence, he found that the


University for which he had strenuously labored was a very hotbed
of opposition. His former pupil and friend, Realdus Columbus, who
was now lecturing on anatomy at Padua, had turned against him.
How deeply Vesalius was wounded by the man whom he had made,
can be appreciated only by those who have been placed in similar
circumstances. The controversy between Columbus and Vesalius was
of a bitter and personal character.

On all sides the views of Vesalius were attacked, and the defenders
of Galen joined hands with men like Columbus in an effort to
besmirch the great anatomist. Disgusted with such treatment,
Vesalius, early in 1544, went to Pisa. Here he conducted a course in
anatomy. Leaving Pisa, he went to Bologna where he made some
special dissections upon two bodies. About this time he declined a
chair in the University of Pisa which was tendered to him by 131
direction of Cosimo de’ Medici. Tired of the apparently
useless effort to make men see the truth, sick of disputes and
arguments, persecuted by members of his own profession, in a fit of
passion Vesalius threw his manuscripts into the fire and ended his
career as a scientist. “Thus”, says Morley, “he destroyed a huge
volume of annotations upon Galen; a whole book of Medical
Formulae; many original notes upon drugs; the copy of Galen from
which he lectured, covered with marginal notes of new observations
that had occurred to him while demonstrating; and the paraphrase
of the books of Rhazes, in which the knowledge of the Arabians was
collated with that of the Greeks and others”.
CHARLES THE FIFTH

While in this frame of mind it is not surprising that he should have


accepted the appointment of Archiatrus to Charles the Fifth of Spain.
The great Emperor was now at the zenith of his fame. His kingdom,
which reached from South America to the Zuyder Zee, was well
under control, but the monarch already contemplated the abdication
of the throne in favor of his son Philip, who is known in history as
Philip the Second.

Vesalius left Italy and took up his residence at Madrid. He 132


was now in his thirtieth year. As Archiatrus he accompanied
the Emperor in the fourth French war, in which he gained his first
experience as a military surgeon. He also acted as physician to
Charles and to the members of the imperial household. The war
ended in September 1544. In January, 1545, Charles went to
Brussels, and remained in the Netherlands for many months.
Vesalius was now in his native country, and in April, 1546, he visited
the graves of his ancestors at Nymwegen and Wesel. In the same
year he published a new edition of his treatise on the China root.

On the twenty-fifth day of October, 1555, amid a scene of pomp and


splendor, in the presence of the assembled representatives of the
Netherlands, Charles formally surrendered to his son all his
territories, jurisdiction and authority in the Low-Countries. This was
the first of a series of acts by which the Emperor gradually
relinquished the reins of power, in order to spend his remaining days
in a cloister. Philip thus became the heir to a vast dominion. Vesalius
was continued in office as Archiatrus by the new Emperor. From both
Charles and Philip, Vesalius received many marks of honor. It was he
who rescued Charles from what was thought to be a mortal disease.
At a later date, when Philip’s unfortunate son, Don Carlos, received a
severe injury to the head, and after the treatment of the Spanish
physicians had failed, it was Vesalius who saved his life by an
operation. These cures, and the accurate prediction of the death-day
of Maximilian d’Egmont, placed the fame of Vesalius at high tide.

133
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

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

Let us accompany you on the journey of exploring knowledge and


personal growth!

ebookname.com

You might also like