Visit https://ebookgate.

com to download the full version and

explore more ebooks

Advanced particulate morphology First Edition Beddow

_____ Click the link below to download _____

https://ebookgate.com/product/advanced-particulate-
morphology-first-edition-beddow/

Explore and download more ebooks at ebookgate.com

Here are some recommended products that might interest you.
You can download now and explore!

First Language Acquisition of Morphology and Syntax Pedro

Guijarro Fuentes

https://ebookgate.com/product/first-language-acquisition-of-
morphology-and-syntax-pedro-guijarro-fuentes/

ebookgate.com

Particulate Filled Polymer Composites 2nd Edition Rothon

https://ebookgate.com/product/particulate-filled-polymer-
composites-2nd-edition-rothon/

ebookgate.com

Particulate Systems in Nano and Biotechnologies 1st

Edition Wolfgang Sigmund

https://ebookgate.com/product/particulate-systems-in-nano-and-
biotechnologies-1st-edition-wolfgang-sigmund/

ebookgate.com

Introducing Morphology 2nd Edition Rochelle Lieber

https://ebookgate.com/product/introducing-morphology-2nd-edition-
rochelle-lieber-2/

ebookgate.com
Introducing Morphology 2nd Edition Rochelle Lieber

https://ebookgate.com/product/introducing-morphology-2nd-edition-
rochelle-lieber/

ebookgate.com

Atlas of Advanced Operative Surgery First Edition Vijay P.

Khatri

https://ebookgate.com/product/atlas-of-advanced-operative-surgery-
first-edition-vijay-p-khatri/

ebookgate.com

Morphology shape and phylogeny 1st Edition Norman Macleod

https://ebookgate.com/product/morphology-shape-and-phylogeny-1st-
edition-norman-macleod/

ebookgate.com

Advanced economies and emerging markets prospects for

globalization First Edition Alves

https://ebookgate.com/product/advanced-economies-and-emerging-markets-
prospects-for-globalization-first-edition-alves/

ebookgate.com

Comparing emerging and advanced markets current trends and

challenges First Edition Goncalves

https://ebookgate.com/product/comparing-emerging-and-advanced-markets-
current-trends-and-challenges-first-edition-goncalves/

ebookgate.com
 Fine Particle Science
and Technology Series
John Keith Beddow, Editor-in-Chief
Professor, Chemical and Materials Engineering
Division of Materials
University of Iowa
Iowa City, Iowa

Advanced Particulate Morphology

Editors

John Keith Beddow T. P. Meloy

Professor, Chemical and Materials Benedum Professor
Engineering Minterals and Engineering
Division of Materials University of West Virginia
University of Iowa Morgantown, West Virginia
Iowa City, Iowa

Future Volume Topics*

Physical-Chemical Properties of Fluid-Particle Interactions
Fine Particles
Mixing of Particulate Solids
Particle Size Analysis
Bulk Solids Handling and
Instrumentation for Fine Particle Storage
Characterization
Hazards
Particle Packing
Agglomeration
On-Line Instrumentation
Education and Research
Particle Formation and
Production Minerals Science
Fluid-Particle Separation
* Subject to change.
Advanced Particulate
Morphology

Editors

John Keith Beddow T. P. Meloy

Professor, Chemical and Materials Benedum Professor
Engineering Minerals and Engineering
Division of Materials University of West Virginia
University of Iowa Morgantown, West Virginia
Iowa City, Iowa

Fine Particle Science and Technology Series

John Keith Beddow, Ph.D.
Editor-in-Chief

Proceedings of the National Science Foundation Residential Workshop on Advanced

Particulate Morphology, University of Iowa, Iowa City, August, 1977.
Grant Number ENG 77-00980.

Boca Raton London New York

CRC PressCRC Press,

is an imprint Inc.
of the
TaylorBoca Raton,
& Francis Florida
Group, an informa business
First published 1980 by CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

Reissued 2018 by CRC Press

© 1980 by CRC Press, Inc.

CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish
reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the
consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this
publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has
not been acknowledged please write and let us know so we may rectify in any future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form
by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording,
or in any information vstorage or retrieval system, without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.
com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a
not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a
photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and
explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data

National Science Foundation Residential Workshop

on Advanced Particulate Morphology, University
of Iowa, 1977.
Advanced particulate morphology.

(Fine particle science and technology ; v. 1)

Bibliography: p.
Includes index.
1. Particles— Congresses. I. Beddow, John K.
II. Meloy, T. P. III. United States. National
Science Foundation. IV. Title. V. Series.
TP156.P3N37 1977 620’ .43 79-14310
ISBN 0-8493-5781-0

A Library of Congress record exists under LC control number: 79014310

Publisher’s Note
The publisher has gone to great lengths to ensure the quality of this reprint but points out that some imperfections in the original copies
may be apparent.

Disclaimer
The publisher has made every effort to trace copyright holders and welcomes correspondence from those they have been unable to
contact.

ISBN 13: 978-1-315-89044-9 (hbk)

ISBN 13: 978-1-351-06954-0 (ebk)

Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the
CRC Press Web site at http://www.crcpress.com
 PREFACE

This volume, based on the 1977 NSF Residential Advanced Particle Morphology
Workshop, covers the principle developments in this new and rapidly evolving field.
Chapter 1 is a review of particle morphology analysis. In it is described how the
particle shape may be digitized and transformed into a set of Fourier (or Walsh) coef-
ficients. These coefficients in themselves have physical significance (for example, A3
is an indicator of triangularity), and they may also be transformed into other mor-
phology descriptors. The sets of coefficients corresponding to the shape of the particle
profile constitute information which may then be further processed through classifiers;
or the data may be treated in the form of a signature; or the data may be subject to
harmonic analysis.
Chapter 2 is a discussion of how the coefficients can be summarized so as to give
an economical, position invariant, universal signature.
Chapter 3 describes some very interesting geological examples in which particle mor-
phology analysis has been used very successfully. An important aspect of this work is
the systematic use of a set of statistical methods in harmonic analysis of particle shape
Fourier coefficients.
Chapter 4 gives an elegant description of the use of fuzzy sets in particle shape
analysis with specific reference to the Fuzzy ISODATA.
Chapter 5 is an in-depth look at some of the orthonormal spanning sets used to
transform silhouette data into coefficients. Walsh, Fourier, Haar, and the optimum
Karhunen-Loeve transforms are reviewed as to the applicability of the particle analysis.
When particle shapes are analyzed, the analysis may be viewed as part of the general
field of pattern recognition. Chapter 6 approaches particle morphology analysis from
this point of view and gives a general outline of image processing and recognition.
Particles are not isolated from one another, and particle sets or assemblies have
structure. Chapter 7 deals with an analysis of the randomly packed structure of parti-
cles. The approach is based upon certain general similarities between packing struc-
tures of molecular assemblages and fine particle sets.
The divergent fields of interest of the authors reflect the prospects for the wide
application of particle morphology analysis and include: many branches of science and
engineering concerned with fine particles, information processing, life sciences, phar-
macy, and food technology.

JKB
TMP
 THE EDITORS

John Keith Beddow received his Ph.D. in Metallurgy from Cambridge University,
England, in 1959. Currently President of the Fine Particle Society, he is a member
of the Faculty at the University of Iowa, where he heads a small research group in
fine particle science with emphasis on morphological analysis. Dr. Beddow is an
active lecturer and author. He has also been active as a Consultant in metallurgy,
powder metallurgy, and powder technology for numerous corporations. His present
research activities are in powder metallurgy and technology and particle morphol-
ogical analysis. Dr. Beddow is married, with four daughters and has resided in the
U.S. since 1966.

Thomas P. Meloy received his Ph.D. in Metallurgy from the Massachusetts Insti-
tute of Technology in 1960. Dr. Meloy has had a distinguished career in industry,
government (where he was Director of the Engineering Division of the National
Science Foundation), and in academia. Currently, he is Benedum Professor of Min-
eral Processing at the University of West Virginia where he serves as department
chairman. He leads a research group in morphological analysis of particulates in
mineral processing. Dr. Meloy is active in public service and committee work. He
is the founder and Editor of the Journal of Ocean Technology. His present research
interests are particle characterization, effect of particle shape on the behavior of
particle systems, analysis of beneficiation circuits, the optimization of beneficiation
circuits and theory of particle separation.
 CONTRIBUTORS

John Keith Beddow Keishi Gotoh

Professor Professor of Chemical Engineering
Department of Chemical and Department of Fisheries Chemistry
Materials Engineering Hokkaido University
University of Iowa Hakodate
Iowa City, Iowa Japan

Thomas P. Meloy
James C. Bezdek Benedum Professor
Associate Professor West Virginia University
Department of Mathematics Morgantown, West Virginia
Utah State University
Logan, Utah
G. Robert Redinbo
Associate Professor
Jeffrey Brown Electrical and Systems Engineering
Geology Department Department
University of South Carolina Rensselaer Polytechnic Institute
Columbia, South Carolina Troy, New York

Julius T. Tou
Robert Ehrlich
Graduate Research Professor and
Professor
Director
Geology Department
Center for Information Research
University of South Carolina
University of Florida
Columbia, South Carolina
Gainesville, Florida

Duane T. Eppler Jeffrey M. Yarus

Geology Department Geology Department
University of South Carolina University of South Carolina
Columbia, South Carolina Columbia, South Carolina
 ACKNOWLEDGMENT

The work reported in this book is one result of a Residential Research Workshop
in Advanced Particulate Morphology conducted at the University of Iowa by one
of the editors (J. K. Beddow) in August, 1977. Grateful thanks to the Particulates
Processing and Multi-Phase Flow Program of the National Science Foundation and
the program director, Morris Ojalvo, is hereby acknowledged.
 TABLE OF CONTENTS
Chapter 1
Particle Morphological Analysis 1
John Keith Beddpw

Chapter 2
Search for Signatures 85
Thomas P. Meloy

Chapter 3
Analysis of Particle Morphology Data 101
Robert Ehrlich, P. Jeffrey Brown, Jeffrey M. Yarus, and Duane T. Eppler

Chapter 4
Particle and Grain Shape Analysis with Fuzzy Sets 121
James C. Bezdek

Chapter 5
Orthogonal Transformations on Particulate Morphological Data 141
G. Robert Redinbo

Chapter 6
Pattern Recognition Methods in Particle Shape Analysis 165
Julius T. Tou

Chapter 7
Morphology of Particle Assemblies 171
Keishi Gotoh

Index 185
 ALPHABETICAL SYMBOLS

Coefficient 3 Molecular diameter

A Non-fuzzy subset d Number of features per
A Area observation
Zero Fourier coefficient d(x) Decision function
or Walsh coefficient dR Thickness of spherical
(mean radius) shell
First Fourier or Walsh D(k) A summation
coefficient (error in DCT Discrete cosine
locating centroid) transform
Second Fourier or DFT Discrete Fourier
Walsh coefficient transform
(elongation) DJJ Dow Jones Index decline
A3 Third Fourier or Walsh D L (z) Discriminant function
coefficient Df(z) Discrimination function
(triangularity) DWT Discrete Walsh
A4 Fourth Fourier or Walsh transform
coefficient (squareness) E(k) A summation
A, Fourier i"1 coefficient f/i(x) Membership function of
A, Fourier coefficient x in A
A, Intercept on signature f jvo(t) Optimum linear
curve approximation
Fourier coefficient f p (t) Periodic extension for a
AR Aspect ratio specific interval
b. Coefficient f(R)dR Probability that center
bN A constant to center distance
B Breadth between particle and its
B Bandwidth of waveform nearest neighbor lies
B {all cases of beer} between R and R + dR
B* Fourier Coefficient from the given central
BANKRt Bankruptcies increase particle
c Number of fuzzy subsets f» Sampling frequency
c Cutoff point f(t) A function
Ci, C 2 Two plain beer cartons F Finite fuzzy subset
c,-* Undetermined constants Fc Partition coefficient
Ck,r Spectrum for Fourier FFT Fast Fourier transform
series representation of FI(CO) Fourier transform of
correlation signal duration L
Spectrum for Fourier Fv Horizontal Feret
series representation of diameter
convolution F(o>) The forward Fourier
Power spectrum transform
Coefficients FWT Fast Walsh transform
Cs(k) Transform coefficients g Radial distribution
of the shift sequence function
X( ((M + S)) ) g. A grain
Convolution between G Fuzzy subset
two waveforms GNPI Gross national product
Fourier transform of decline
C <" h Planck constant
 ALPHABETICAL SYMBOLS (Continued)

h(p) Height max Maximum

har(r,m,t) Haar function Min Minimum
Hf Entropy MIN Minimum
i Subscript, one of many n An integer
I An interval n Number of features
I NxN identity matrix n Order of coefficient
I* Number of horizontal N A finite number
intercept N The cordinality of the
L Number of vertical basis set
intercept N,, Number of density of
j Arc length of j"1 vertex particles
from starting point p A person
j Kind j — type j species p Pressure
J,(u,v) A functional, within Pr Posteriori probability
group sum of squared PL Piecewise linear function
errors r A person
3m Generalized total square r 0,1,....
error r Mean radius
J^(u,v) A generalized within r iy (i) Periodic correlation
group sum of squared function
errors functional R Radius vector
Jw, A criterion function R Roughness
k An integer, i.e., R A distance in units of
0,1,2,3,.... sphere diameter, radius
k A coefficient of spherical shell
k Boltzman's constant Re Real
k 0,1,2,...N R, y (w) Transform domain
K A constant R transform A term of the FWT
KLT Karhunen-Loeve (R.(z) rank of D,(z)
transform s Slope of signature curve
I Length of perimeter s A shift
L Arc length
L Length
Shape factors
L Lumpiness
L3 Mean particle
interception three An exponent
dimensions n= 1,2,3,...a set
LDF Rule Linear discriminant Fuzzy membership
function rule function
L(t) Coefficients for DCT t Interval {0,2rt}
m 1,2,...2 r t0 Starting position
m 0,1,2,...n-1 T Length
m An integer T A matrix
m A coefficient T Absolute temperature
m Molecular mass T Thickness
m A power u, Characteristic function
Mc Set of all nondegenerate U;t Elements of U
comparators of x Ur Members of A
M,c Set of all nondegenerate y Fuzzy subset
fuzzy c-partitions of x y A matrix
 ALPHABETICAL SYMBOLS (Continued)

U(p) Characteristic function W(n,t) Walsh function

of people x Feature vector
Ur First half of data x0 A constant
sequence x, General coordinate
UNEMP Unemployed variable
v Cluster center x(t) Original waveform
v (v,,v 2 , v,) prototype x(co) Fourier transform of
v, Free volume of a single x(t)
molecule X A vector
V Total volume X c R rf an n Sample from Rrf
Vrn Vector space of all real X 4 eX A feature vector in X
(cxn) matrices X t ,£ Measured features of Xk
V, Volume of liquid phase Y A vector
V, Empty volume of the Y(co) Fourier transform of
system y(t)
Vr Second half of data Z Coordinate number
sequence Z An observation
w Pitzer's acentric factor Z,...Z t Nearest points to Z by a
w, Weighting parameter distance measure
w Augmented weight Z,,,(T) Periodic convolution
vector function
W The n"1 root of unity Zm A convolution
 GREEK SYMBOLS

oc A real number M</ U . Grade membership of u,

oc Integer ranging from 0 inF
to 3 1*1,1*1 Normally distributed
i'* Phase angle Fourier means
transform n(u,/3,y) Fuzzy membership
ocn A constant function
oc Phase angle n,,n 2 Populations
Integer ranging from 0 e Liquid density
to 3 e,<?E,£?R Constants
Any set of constants I, Covariance of vector x
inl Optimum coefficients I Covariance matrix
A constant * Bulk mean volume
Kronocker delta fraction of particles
function <J>t Boiling mean volume
Difference, interval, fraction of particles
part of *c Critical mean volume
Tolerance of measuring fraction of particles
device *,(x) Linearly independent
Potential energy bounded function of x
£(t) Error of approximation in I"
c 2 Mean square average of <t>(f) Change of shape
AX et>m Melting mean volume
Mean square error fraction of particles
criterion <J>» A set of functions
Angle 0>» Vectors (rows in
Molecular shape orthogonal matrix T)
parameter <t>*(t) A function
A A given translation V Angle of pyramid apex
A Complex numbers (compliment oc)
Mo First Fourier V Potential energy
Coefficient, (<(>,l) between one molecule
method and all others
2n/T fundamental
parameter
 MATHEMATICAL SYMBOLS

Transpose || Absolute value

U Union ((x)) Modulo N operations
PI Intersection sal Composite term from
O Dyadic operation sin/Walsh
O Bounded sum cal Composite term from
O Ring sum cosine/Walsh
O Bounded difference Wai Walsh
V Maximum WAL Walsh
A Minimum PL/1 Computer language
A Defined as RADNO A computer program
/ Union of fuzzy title
singletons ^,/u, x^ Average of x
u x Transpose of x
F' Compliment of F £ Belongs to
XxY Cartesian product * Complex conjugation
RoS Composition of R and S R" Feature space
{} A set
 Chapter 1

PARTICLE MORPHOLOGICAL ANALYSIS

J. K. Beddow

TABLE OF CONTENTS

I. Significance of Morphological Analysis-An Engineering Discipline 3

II. Fourier, Walsh, and Haar Functions 3

A. The Method of Fourier Analysis of a Profile 3
B. The Reasons for Using Walsh Functions 6
1. Arc Length 10
C. Physical Significance of 3D Walsh Functions 10
1. Cubing the Sphere 10
2. Excess Surface Area 11
3. Edge Length 11
4. Volume of Positive Blocks 12
5. Top Surface Area 13
6. Activity Corners 14
D. Other Orthogonal Functions 14

III. The Problem of Reentrants 14

A. The (R,0) Method 14
B. The (4>,i) Method 15
C. (R,0) and (0 J) compared 17
D. Simplification of Fourier Coefficients for Regular Polygons 21

IV. Translation Invariance 23

A. The Problem of Translation Invariance 23
B. The R Transform 23

V. Methods of Feature Extraction 26

A. Harmonic Analysis 26
B. Signature Analysis 26
1. Variations of on (Type I Randomness) 28
2. Variations of An and an Simultaneously 29
C. Feature-Space Representation of Particle Shape 30

VI. Deterministic, Statistical, and Fuzzy Classifiers 30

A. A Simple Decision Function for Two Pattern Classes 37
B. Features 38
C. Deterministic Trainable Pattern Classifier 38
1. Feature Selection 40
2. Stereological-Type Features 40
3. Morphologic Characteristics 42
4. Decision Function 43
5. Experimental Results: Stereological-Type Data 45
6. Experimental Results: Morphological Characteristics 48
2 Advanced Paniculate Morphology

1. Discussion: Stereological Features 50

8. Discussion: Morphological Characteristics 51
D. Statistical Classifiers 54
1. Feature Selection 55
2. Classification Rules 56
3. Linear Discriminant Function (LDF) Rule 56
4. Quadratic Discriminant Function (QDF) Rule 58
5. Fix-Hodges k-Nearest Neighbor Rule 58
6. Classification Results and Discussion 59
E. Fuzzy Classifiers 61

VII. Verbal and Geometric Descriptors 61

A. Properties of Fuzzy Sets 64
B. Linguistic Labels 68
C. Linguistic Variables 68
D. Geometric Comparator 69
E. Verbal Definitions of Complex (Fuzzy) Concepts 73

VIII. Shape-Texture Analysis of Fine Particles 74

IX. Recent Developments 78

A. Morphology and Mode of Particle Origin 78
B. Morphology and Chemical Reaction 78
References 83
 I. SIGNIFICANCE OF MORPHOLOGICAL ANALYSIS — AN
ENGINEERING DISCIPLINE

Historically, particle size has been defined by an interaction with a measuring instru-
ment such as a sieve; and particle shape, by one or two parameters such as size and
aspect ratio. Recently particle silhouettes have been analyzed by sophisticated signal
processing techniques such as Walsh and Fourier series, yielding numerical values
which describe the particle profile shape with great accuracy. Using existing laboratory
equipment it is possible to obtain particle fingerprints with which the particle can be
uniquely identified. Corresponding sets of data can be obtained for powders in the
form of signature analysis or harmonic analysis. These data can be used to determine
whether a sample is composed of powder A or powder B, is a combination of A and
B, or if A and/or B is present in powder sample. For example, if two power-plant
stacks emit particulates, it is possible to determine at a sampling station how much of
the particle sample was contributed by the first stack and how much by the second.
By using the advanced clustering and classification technique, the number of particle
shapes is determined. For the homogeneous powders we have examined thus far, all
have multimodal shape characteristics. This unexpected result has deep theoretical im-
plications both for morphological and statistical studies.
Application of the morphological analysis techniques has widespread applications
in industry for powder specification and process control, in government for forensic
and environmental problems, and in medicine for the identification of pathogenic par-
ticles.

II. FOURIER, WALSH, AND HAAR FUNCTIONS

A. The Method of Fourier Analysis of a Profile

It has been shown that a particle profile can be considered a population of points
from which a sample set of points in the form of (x,y) coordinates can be abstracted. 1 2
This sample set must be a Nyquist set, i.e., the points must be evenly spaced round
the profile and the software has to be adjusted to do this. Using the (x,y) data, the
centroid of the shape is calculated and thereafter the set of coordinates is transformed
into polar coordinates, (R,0). The function (R,0) is then described in terms of a Four-
ier series of the general form:

R(0) = A0 + J^ An cos(n9 - an) (1)

in which A n are the Fourier coefficients and <*n are the phase angles. It may be noted
also that in terms of the more standard form of the Fourier relationship:

A / 2 . 2 \ 1/2
A
n=(an+bn) (2)

<*n = tan-' b n /a n (3)

in which a n and b n are coefficients such that:

=*A0 + | (Ancos ne + bnsin nfl) (4)

 Advanced Paniculate Morphology

ORIGINAL SHAPE
CU87

IS-

1
17.6 18.4
1
19.2
r~ 1
20.8 21.6
20.0
X COORDINATE

FIGURE 1. Original shape Cu87.

An example of the result of this type of analysis is given in Figure 1* and Table 1.
The table contains a set of coefficients and phase angles for the shape described in
Figure 1. Also Figure 2 contains a regenerate of the original profile which is obtained
by simply reversing the process on the data given in Table 1. The method of obtaining
the (x,y) data is to digitize the particle profile image. For this purpose we use a Grafpen
system (Figure 3). When using this system, the operator places a profile of the subject
particle on the Grafpen table and with the stylus hand-traces the profile. The tip of
the profile emits a noise as it touches the surface being traced and this noise is picked
up on sensitive microphones. These microphones run in a continuous bank along one
side and top of the tabletop. In this way the (x,y) coordinates can be obtained. It is
possible to adjust the rate at which data is acquired and so the size of the point sample
obtained may be adjusted for each requirement. Errors associated with digitizing pro-
cedures have been fully discussed elsewhere.3 Two points should be borne in mind.
With a method in which the profile is traced, the human error is always present. This
can be significantly reduced in the case of a system such as the Quantimet® with a
FIFI attachment. Secondly, the error associated with the Fourier method can always
be reduced by taking another term in the series.
* The (*,i) method rather than the (R0) method was used here. See Section III.B.
 TABLE 1

Fourier Coefficients

Harmonic Number A. an

1 0.9180837 -0.026830
2 0.7811343 -1.919139
3 0.6038185 2.204062
4 0.5977379 1.177760
5 0.2106746 -0.880125
6 0.1914537 3.035851
7 0.5071257 -1.026711
8 0.2885617 -2.234097
9 0.3928300 -2.244235
10 0.1503933 -1.662912
11 0.1452250 3.116115
12 0.1505209 -2.837341
13 0.4840755 -1.598395
14 0.0456248 0.882792
15 0.2388835 -0.446429
16 0.1031974 1.943743
17 0.3679479 -3.021418
18 0.1361023 1.582670
19 0.1662030 2.128934
20 0.2110651 -2.249763
21 0.1499012 1.573666
22 0.0867854 2.674203
23 0.2423907 2.593346
24 0.1337547 1.733434
25 0.1888060 -1.773540
26 0.1171121 -0.422164
27 0.1184515 0.368735
28 0.2025601 0.740007
29 0.1464919 -1.462102
30 0.1449654 -0.974561
31 0.0925761 0.503672
32 0.1278262 0.361063
33 0.1011988 2.310310
34 0.0692867 2.040118
35 0.1214545 1.330347
36 0.1129320 2.927621
37 0.0784888 2.829776
38 0.0965914 0.771211
39 0.0910937 3.069534
40 0.0778255 -1.429016
41 0.0613601 -1.262012
42 0.0825907 -0.342230
43 0.0328851 2.987070
44 0.0710672 -0.856483
45 0.0336690 2.626568
46 0.0656527 -0.647202
47 0.0577465 1.744238
48 0.0996903 -2.300911
49 0.0355991 -1.061722
50 0.0286593 -1.726463

One great advantage of the Fourier method used in this way is that the individual
coefficients have real significance in terms of the shape of the profile. For example,
plots of AO + A, cos© versus 0; and of A0 + A 2 cos 20 versus 0; and of A0 + A3
cos 30 versus 0; and so on, are shown in Figure 4. The first diagram shows that A0
 Advanced Paniculate Morphology

REGENERATE SHAPE
CU87

Q
CC
O
O

I I I
16.8 18.4 19.2 21.6

X COORDINATE

FIGURE 2. Regenerate shape Cu87.

+ A, cosO represents the error in locating the centroid. Similarly, A2 indicates the
elongation or aspect ratio of the profile; the A3 term represents its triangularity; the
A4 term indicates the degree of squareness, and so on and so forth. The coefficients
thus represent the odd and even types of rotational symmetry. 4
Another advantage of representing a closed profile in this way is that the successively
higher-order coefficients decrease so that one can represent the (A n ) as a signature plot
of A, vs. n. This is referred to in Section VB. What this means is that the higher the
order of the coefficient, the finer the morphological detail it represents.5

B. The Reasons for Using Walsh Functions*

Fourier functions operate over an interval 0 through 2n. They are orthogonal, nor-
malized, and complete. The term orthogonal means that the product of any two func-
tions integrated over an interval is zero, unless they are identical. The term normalized
means that the integral of the function squared over the interval equals one. The term
complete means (among other things) that all of the functions in the set are defined.
However, there are potential advantages in using square-wave functions instead of the
more traditional Fourier and for a number of reasons. For example, it is possible to
state the Fourier form of a line (or a surface) integral but unfortunately because the
 INTERRUPT*

FIGURE 3. Graf/pen® gp3 sonic digitizer. The third generation sonic digitizer that ob-
soletes all other graphic-to-digital converters.

integrals contain unspecified terms multiplied to trigonometric functions, they cannot

be integrated in closed form and so for this reason uses for Fourier representation of
surface features appear to be limited. Alternatively, Walsh functions (which are
square-waved functions as shown in Figure 6 of Chapter 2) can be used to represent
the particle profile in just the same way that the Fourier method can be used. This is
illustrated in Figure 58 where it can be seen that the Fourier and Walsh representations
are equivalent. Furthermore, the individual coefficients have similar significance. As
shown in Figure 4, Chapter 2, A 2 in the Walsh representation indicates elongation or
aspect ratio just as does the Fourier A 2 . Similarly, in Figure 5, Chapter 2, A3 indicates
triangularity and so on. One refers to Figures 6 and 7 to begin to see the advantages
of the Walsh representation. Thus, the excess profile can be modeled as the total length
of the vertical edges and this is 8 A2 in Figure 7. The advantages of doing this sort of
arithmatic compared to the problems associated with trying to calculate surface fea-
tures using the Fourier series are obvious.
 Advanced Paniculate Morphology

0 90 180 270 360

FIGURE 4. Plots of regenerates for various As. (From Beddow, J. K., Philip, G.
C., Nasta. M. D., and Vetter, A. F., Powder Technol., 18, 19, 1977.

Referring again to Figure 6, Chapter 2, the term Walsh function was named in honor
of the inventor of the function. The terms sal and cal are composite terms made up
from sine/Walsh and cosine/Walsh, respectively. The single function wal(j.Q) can be
defined in order to relate the three functions:

Wal(2i,e) = Cal(i,0)

Wal(2i-l,0) = Sal(i,0) (5)

in which i is termed the sequency and is one half of the number of zero crossings per
second. Note from examination of the Walsh displays in Figure 6, Chapter 2 that the
intervals are not equally spaced as are those of the Fourier function. Note also that
wal(o,0) is the same as R 0 ,; sal (2,0) is the same as R 2 and so on. (The Rs are Rada-
macher functions shown in Figure 8).
 Low-pass filtering using Fourier, Walsh and Haar transformation.

FIGURE 5. Lower pass filtering using Fourier, Walsh, and Haar transfor-
mations. (From Beauchamp, K. G., Walsh Functions and their Applications,
Academic Press, London, 1975. With permission.)

SIMULATION OF PARTICLE SURFACE

7T/n

2An-2A2n

FIGURE 6. Simulation of particle surface. (From Meloy, T. P., Powder Technol.,

16(2), 233, 1977. With permission.)
10 Advanced Paniculate Morphology

F I G U R E 7. Illustrating the concept of excess edge length and its measurement.

1. Arc Length
With the treatment of Walsh functions in the line integral, the arc length can be
described by two terms:

L = 27rA0 + 8A, + 12A3 + ...

= 27rA0 + 4 X N - A n (6)
n=2

This equation seems to be reasonable, for example, if we draw out the superposition
of A2 and A3, in which we shifted the peaks o A3 with a 15° clockwise rotation in
relation to those of A 2 . But, if we shift 30°, the second term of excess edge length is
reduced into:

Similar to the result from the 30° shift, we find that shifts of 60°, 90°, 120°, 150°,
10° 210°, 240°, 270°, 300°, 330°, and 360° also have the same expression. This means
that Equation 6 is a translation variant.

C. Physical Significance of 3D Walsh Function

/. Cubing the Sphere
It is assumed that the number of square-wave peaks are in the ratio of 1/n between
the flat square surface and the sphere surface. If the circumference of the sphere has
n positive waves then there are n 2 positive waves on the flat square surface or n 2 /n on
the sphere surface.
The surface area of a sphere having a unit in radius is

A = 47rR2 =4n (R = 1)

By assumption of square waves, the total top area of positive blocks is

A/2 = IT,
 11

There are nVn positive blocks on a sphere, the top area for each block is

We assume that each block has a square base, such that the base length of the positive
block is

(7)

Each positive block on the sphere's surface has a height of An and a square base \f2u/
n on a side.

2. Excess Surface Area

There are nVre of blocks each with four sides, a height of 2An, and a base length of
\f2n/n. Since only half of the side area is attributable to any one block, the side area
for each block is

1
/2(2AnX N /27r/nX4) =

in which

4 A
2x ? = 4^7rA,2
2

4x

6x

TT n

The total side area of all blocks of all frequencies is

Side Area = 4^/2nA + A3 + 6s/2~7rA4 + 4\/T S nAn (8)

3 n=5

J. Edge Length
Each block has four vertical edge lines. Only one-fourth of them can be assigned to
a block. The edge length for each block is

y4x4X2An =
12 Advanced Paniculate Morphology

in which

2x(2A 2 ) = 4A 2

4x(2A 3 ) = 8A3

6x(2A 4 ) = 12A4

— x(2An) = -n2
7T 7T

The resulting edge length is

2 <»
Vertical Edge Length = 4A 2 + 8A3 + 12A4 +— s n (9)
T n=S

The horizontal edge length on the top of ech block is

4x(^ >
\ n I

and the total horizontal edge length is

_
Horizontal edge length = 4J2-n + TT + 6v2w + 4V2 £ (10)
3 n=5

Meloy5 proposed that a physical property of a particle can be represented by:

Physical property ocA0 + K 2 n a A^ (11)

2

in which A0 = average radius of the particle, K = constant, a, ft = integers 0-»3, n

= the number of the coefficient, and A, = the value of the coefficient. Using this
approach the set of equations shown in Table 2 has been developed. 5
The first row of functions in Table 2 do not appear to represent a physically mean-
ingful property except for A 0 , which is the mean particle diameter. It should be noted
that this mode of definition of mean particle diameter is rigorous and it is uniquivocal.
Again in Table 2, we have already calculated the form of, for example, the edge length
of Equation 9, that appears in row 2, column 3, the side area of Equation 8 which is
in row 2, column 2. A 3D simulation of the particle surface is shown in Figure 8. This
will help us to visualize the volume o the positive blocks and the top surface area.

87T2 37T2
- A 2 + — A 3 + — A4 + 2*^ An (12 )
7T

4. Volume of Positive Blocks

This term is found in Table 2, (row 2, column 1).
 13

TABLE 2

The Meloy Equations

For Morphological Characteristics

2, |nA n
A
o+knAn

From Meloy, T. P., Powder Technol., 16(2),233, 1977. With permission.

RADAMACHER
FUNCTIONS
R

R,

R3
~u~Lru~i_

FIGURES. A set of Radamacher

functions. (From Beauchamp, K.
G., Walsh Functions and their Ap-
plications, Academic Press, Lon-
don, 1975. With permission.)

5. Top Surface Area

In Table 2 (row 3, column 1) the terms may be predicted to represent excess top
surface area. As we know, the total top surface area of all positive blocks is

in which P = 2. The equation is similar to that in Table 2 (row 3, column 1).

14 Advanced Paniculate Morphology

6. Activity Corners
In the application of the square waves to the physical properties of particles,
InocAn*", one must consider the relationship between n and An, and the determination
of n when it approaches infinity. There is a linear relationship between log An and n,
and the value of An closes to zero as n nears infinity (the signature). One can choose
an upper limit of n from a meaningful physical parameter such as the unit crystal size
— 10A, or choose a limit of n to truncate the formula without the loss of significance
of the equation.
To calculate the number of corners on a particle, one has a postulate that a pyramid
with one apical corner is used for simulating a corner on a particle, and the sharper
the apex, the greater the reactivity of the corner. As n increases, the top of each block
is so small as to be regarded as a single point. So, one can transform the vertical blocks
into pyramids with one apical corner. The base angle of a pyramid, ip, is measured as

\l/ = tan"1 = tan"1

2/2n

and the equation of corners becomes:

2A - n \ 2 2k
tan'1 —?— } = H £n 4 A£ (13)
S/27T

D. Other Orthogonal Functions

One practical implication of a complete series is that one can faithfully represent a
given profile to an accuracy that is determined by the number of terms chosen to
represent that profile. There are other orthogonal series but not all are complete, and
this lack of completeness causes one to hesitate in testing their effectiveness in mor-
phological analysis. An example of this type is the Radamacher function that is a
square wave with a unit mark space ratio. The first five Radamacher functions are
given in Figure 8. By comparison with Figure 6, Chapter 2, it will be observed that
the Radaacher functions form an incomplete but true subset to the Walsh functions.
The Haar functions have the capability of representing a given function with few
constituent terms to a high degree of accof the first eight Haar functions. Along with
the Walsh and Fourier series, Haar functions are quite capable of representing compli-
cated profiles as Figure 5 illustrates. In general, as one would expect, it takes less
Fourier terms to represent a continuous type of wave form than the number of Walsh
or Haar terms needed, respectively. Similarly, for a discontinuous wave form, the
square wave sets of function are to be preferred because less of their terms will be
required to synthesize such a wave form.
One might speculate that the Haar function which is a local function, could be used
to represent features on the profile of a particle that occur infrequently more econom-
ically than either Walsh or Fourier (which are both global functions.)

III. THE PROBLEM OF REENTRANTS

A. The (R,0) Method

The (R,Q) method is very useful for representing analytic or holomorphic particle
profiles. These contain no reentrants and therefore for every value of 0 there is one,
 15

HAAR FUNCTIONS

9 axis h
0 1/2 1

FIGURE 9. Haar functions. (From Beau-

champ, K. G., Walsh Functions and their
Applications, Academic Press, London,
1975. With permission.)

and only one, corresponding value of R. The analysis may therefore be conducted
smoothly. However, if a profile contains a reentrant, then at certain values of 0, there
will be multiple values of R. An example of this type of problem is shown in Figure
10. The only way to solve this problem with the (R,0) method is to arbitrarily choose
one R value as the one that matters and ignore the others. For example, in the case of
the profile shown in Figure 10 one decision rule could be to select the largest value R
as the value to compute with. This would tend to smooth out the curve artificially and
so information would be lost from the analysis. The amount lost in this way would
vary from particle to particle.

B. The (0,1) Method6

Instead of changing (x,y) pairs into (R,0) pairs, the profile is parametrized by its
arc length and change of slope $>(() from its starting point as in Figure 11 (i.e., (x,y)
-*• (I ,<t>(f)). In order that comparisons between particles of different sizes be meaning-
ful, the total arc lengths, or perimeters are normalized by defining

(14)
"(i) 27T

(15)
16 Advanced Particulate Morphology

FIGURE 10. Reentrance point on particle profile.

(From Fong, S. T., Beddow, J. K., and Vetter, A. F.,
Powder Technol., submitted, 1979. With permission.)

L=perimeter

•starting point

FIGURE 11. Parametric representation of a profile.

(From Fong, S. T., Beddow, J. K., and Vetter, A. F.,
Powder Technol., submitted, 1979. With permission.)

The interval of interest is t = fO,2rt] for (. = [O,L].

The function <t>*(t) is then expanded into a Fourier series of the form

0*(t) = + S (Aktcos kt + Bkt sin kt)

k=l

All the morphological information will be contained in the coefficients (A t ,Bk)- As the
ordered pair (l,$(i)) is unique for every point on the profile, multi-valueness no longer
appears, and the profile is faithfully represented by (A t ,B t ).
Since all particle profiles are closed curves, they can all be approximated by poly-
gons. If this fact is exploited, simplified equations for finding \n0, At, Bt can be derived.

(16)

_1 m 27rk
Akk = — X A*, sin (17)
kTr j=1 J L
 17

Vm-1 N = no. of vertices

m = N+1

FIGURE 12. Notations of arc length and change of

shape. (From Fong, S. T., Beddow, J. K., and Vetter, A.
F., Powder Technol., submitted, 1979. With permission.)

m
1
Bt = — £ A*j cos (18)

in which , = arc length of jth vertex from starting point, A^ = change of slope at
vertex j, and m = total number of sides of the polygon. Figure 12 illustrates the mean-
ing of item notations.
A computer program written on PL/1 has been developed by the authors to calculate
coefficients /J0, A t , Bt from x,y coordinates of real particles. A subprocedure which
regenerates profiles from a set of coefficients via a completely different route has also
been written. The overall scheme is given below.

magnified (x,y) coordinates

o, A*
picture of digitizer
particles of profile program

regenerated
particle regenerate program
profile

C. (Rd) and (<t>,l) compared

By comparing the regenerated profile with the original, validity and applicability of
the method is checked visually. Figures 13 and 14 show some of these results and a
comparison of the <t>,i method with the R,0 method for both simple and complex
particle profiles, respectively. It is evident that wherever reentrance points appear, the
($1) scheme is superior to the (R,0) one, although simple (i.e., holomorphic) profiles
are adequately represented by both.
If the profile of an m-sided regular polygon is considered, it can be proved (in the
next section) that At will vanish for any k whereas Bk is nonzero only when k is a
whole number multiple of m. In addition, those coefficients that are nonzero are given
by a very simple formula.
18 Advanced Paniculate Morphology

p
CO __

O
esi _
LU

< . • -^
z
Q X^
~ ^V
<£ f \^
O o / X,
8-~ ) \^

°— ^^\ r
* ^\ ^/

q
14.0 16.0 18.0 20.0 22.0
X COORDINATE

o
CO.

p
«N _

LU
I-

—o s—**
/^ ix.
oc I v
O °- I X
O oo J \

o ^x. ^/
*~ ^\ ^/

°14.0 16.0 18.0 20.0 22.0

X COORDINATE

FIGURE 13. (A) Original shape of Cu59, (B) regenerate shape of

Cu59 from (0,<) program using 75 coefficients, (C) regenerate shape
of Cu59 from (R,Q) program using 75 coefficients. (From Fong, S.
T., Beddow, J. K., and Vetter, A. F., Powder Technol., submitted,
1979. With permission.)
 19

o
<o —

Q
IT
O
O
O

14.0 16.0 18.0 20.0 22.0

X COORDINATE

FIGURE 13C

Q
IT
O
O
O

\
26.0 28.0 30.0 32.0 34.0
X COORDINATE

FIGURE 14. (A) Original shape of Cu89, (B) regenerate shape of

Cu89 from (+,l) program using 75 coefficients, (C) regenerate shape
of Cu89 from (R,0) prooram using 75 coefficients. (From Fong, S.
T., Beddow, J. K. and Vetter, A. F., Powder Technol., submitted,
1979. Withpermisson.
20 Advanced Paniculate Morphology

q
<o_

9
tu ^—.

8 S_ \ \

P_
"26.0 28.0 30.0 32.0 34.0
X COORDINATE

FIGURE 14B

q
to _

q
«j_
nj ^-

^ o X \
O <^ _ ) )

q
°° 26.0 28.0 30.0 32.0 34.0
X COORDINATE

FIGURE 14C
 21

For an m-sided regular polygon:

Ak = 0 for any k,
_2
BKt = — for k being whole number multiples of m,
k
Bk = 0 for other ks. (19)

This simplification increases the potential utility for relating the morphological coef-
ficients to physical and chemical properties because only a small number of coefficients
exist. By measuring these properties of regular polygonal particles (probably artificial)
and relating them to the few number of coefficients obtained, a deeper insight as to
the functional relationship between morphological coefficients and the properties of
particles and their sets can be established. An example of this might be the connection
between morphological coefficients and drag coefficients of particles moving through
a fluid. This new technique is already being applied in a study of the relationship
between changes in particle morphology and the progress of a chemical reaction.

D. Simplification of Fourier Coefficients for Regular Polygons

This method has the potential of simplifying the analysis of poygon profiles as is
demonstrated below.
The equation for an irregular polygon is:

_j m
A.k = — s A«J>; sin (20)
kn j=1 J L

For an m-sided regular polygon:

-27T
A<Pj = = constant
1 m

L
2,3 = — J
m

-1 m /-2JT\ . /27Tkj\
k S I m 1 sin I m 1
* j=i \
:. A k == —
/ \ /
Putting 0 = 2nk/m we have

Ak, ==(-\w (—\™

[ k-1 — I s sisinje
Im (21)
• V / V / j=i

If m is odd:
/-1\ / - 2 W \ r
A
k =- I — I ( I [sine + sin 29 + + sin (m-l)0 + s i n m 0 ]

/ - 1 \ /-27T\
- ( — I I I [sine + sin (m-l)fl + sin 28 + sin (m-2)0 +

+ sin re + sin (m-r)e + sin me ]

22 Advanced Paniculate Morphology

Using the trigonometrical identity:

A+B „„
0!> A-B
sm A + sin B = 2sin
2 2

2 md 2-m 4—m It—m .

= — [2sin — (cos 9 + cos 0+ + cos 8 + sin me]
km 2 2 2 2

but,

me / m \ /27Tk\
in — = sin I — 1 I I
2 \2J \m )

= sin k

=0 for any integer k

Also

/2i
sin me = sin m I —
\"
= sin 2k?r

=0 for any integer k

Therefore every term within the bracket on the R.H.S. of the equation is zero.

:. Afc = 0 for any integer k

If m is even, the middle term sin m/2 is left over in Equation 1. But

—9 = 0 for any integer k

for any integer k and m

B,, = — £ A*jJ cos

k ;_, L

Substituting

-27T
A0: =

into the equation we have

 23

Bk = — [ cos0 + cos20
km

cos(m-j)0 + cosm0 ] . . . (22)

but an m-sided regular polygon has m-fold rotational symmetry. According to Equa-
tion 22, Bt will be nonzero only if k is a whole numer multiple of m or k/2 = integer,
i.e., only B m , B Zm , B 3m , etc. are nonzero. When k/m is an interger, 6 = 2nk/m will be
a whole number multiple of 2n and cos 2nn = 1 for any n.

BKk = — [1 + 1 + 1+ + 1]
km

-2
= — m
km

= — (23)
k

IV. TRANSLATION INVARIANCE 7

A. The Problem of Translation Invariance

While the discrete Fourier transform is invariant to the circular time shift of the
input signal, the Walsh transform is not invariant. What this means in practice is that
whereas a Fourier series can be used to represent a particle profile quite independently
of the origin that one starts to digitize from, this is not the case for the discrete Walsh
transform. This is shown in Table 3 and more graphically in Figure 15.8 In Table 3
the numbers represent the discrete Walsh transform of a sinusoidal wave. There are
32 of them. In part b of Table 3 the numbers represent the same wave-form but shifted
by it/2 radians. Perhaps more clearly in Figure 15 the values of the Walsh coefficients
change over the interval with a 2n radian phase shift in size as well as in sign. Table 4
gives a set of Walsh coefficients for a particle profile. Note that the effect is the same
for this closed form as it is for the wave-form. It is clearly unsatisfactory to use a
transform that gives different results depending upon where you start to take the sam-
ple points from.

B. The R Transform
Fortunately, a form of the fast Walsh transform has been developed which is invar-
iant to the cyclic shift of the input data. But there is a price that has to be paid for
this important advantage. The original signal cannot be regenerated from the output
of this transform. While this is not great disadvantage, it does compare unfavorably
with the Fourier transform. However, the other part of the price that has to be paid is
more serious, as shown in Figure 16,8 the higher-order terms that the Walsh transform
will pick up the detail with are simply not there if one uses the R transform. This can
be a serious disadvantage if one does not take it into account. We can take note that
if it was not the attractiveness of the type of modeling that the morphological charac-
teristics represent, it would not be worth it to use Walsh or R type transforms over
the Fourier type. In addition, although the Walsh transform is a little quicker to use
than is the Fourier transform, as is shown in Table 5, this cannot offset the fact that
finer detail of particle profiles will not be obtained. However, in the case of the Haar
transform, there appears to be a substantial advantage in computation time as com-
24 Advanced Paniculate Morphology

Phase advance = 0°

Phase advance = 23°

Phase advance = 46°

Phase advance = 69°

I

Phase advance = 90°

Order of (k) 15 Sequency 31

FIGURE 15. Walsh transforms of a sinusoidal

wave-form with circular time shift (N = 32).
(From Beauchamp, K. G., Walsh Functions and
their Applications, Academic Press, London,
1975. With permission.)

TABLE 3

Walsh Transform Coefficient

For a Simple Sine Waveform N = 32

0 0.663 0.063 0 0 -0.263 0.025 0

0 -0.052" -0.006 0 0 -0.126 0.013 0
0 -0.013 -0.002 0 0 0.006 0 0
0 -0.025 -0.002 0 0 -0.062 0.006 0

For the Sine Waveform Shifted by 90°

0 -0.063 0.663 0 0 0.025 0.263 0

0 0.006 -0.052 0 0 0.013 0.126 0
0 0 -0.013 0 0 0 -0.006 0
0 0.002 -0.025 0 0 0.006 0.062 0

From Beauchamp, K. G., Walsh Functions and their Applications,

Academic Press, London, 1975. With permission.

pared with the others (Table 5). The advantages and disadvantages of using the Haar
transform have yet to be investigated.
 25

TABLE 4

Illustrating the effect of cyclic shift in the case of Walsh

Coefficients 1, 3, 4, 7 and 10 for the same particle in the
case of 3 different starting positions

Starting Point
Coefficient Position 1 Position 2 Position 3

1 2.57564 2.57564 2.57564

3 0.00418 0.00111 0.00577
4 0.18598 0.19251 0.15227
7 0.00529 0.00722 0.00049
10 0.00988 0.01936 0.01330

From Lew, G., Beddow, J. K., and Vetter, A. F., Powder

Metallurgy Int., submitted, 1979. With permission.

FIGURE 16. Comparative examples of the R-transform

and the Walsh transform. (From Beauchamp, K. G., Walsh
Functions and their Applications, Academic Press, Lon-
don, 1975. With permission.)

TABLE 5

Fourier, R, and Haar compared

Data storage
Transform Time(s) (N=1024)

Fourier 9.48 4K
Walsh 1.60 3K
Walsh—Ulman(R) 1.25 4K
Haar 0.29 4K

From Beauchamp, K. G., Walsh Functions and

their Applications, Academic Press, London,
1975. With permission.
26 Advanced Paniculate Morphology

V. METHODS OF FEATURE EXTRACTION

At its present stage of development, morphological analysis of fine particles trans-

forms a set of x,y coordinates into a set of coefficients (Fourier, Walsh, Haar, etc.).
These coefficients may then be further analyzed to yield more condensed forms of
information. These latter forms may in turn be input for further analysis as required.
For convenience, we have divided the discussion of these methods of analysis into two
sections. This one and the section following.

A. Harmonic Analysis
Ehrlich and co-workers have developed a sophisticated set of techniques for the
harmonic analysis of sediments. 910 After digitizing the particle profile and subjecting
the (x,y) data to Fourier analysis, the sets of coefficients may be displayed as one of
the following:

1. Mean harmonic amplitude spectra,

2. Shape frequency distributions.

An example of a mean harmonic amplitude spectrum is given in Figure 17 for three

different sediments: Brown Mountain Granite, Piedmont Granite, and Felsic Tuff. 1 0
As can be observed, there are differences between these three sediments over a broad
range of harmonics. In cases where these differences in harmonic amplitude spectra
are insignificant, nonexisting, or just confusing, it is possible to develop shape fre-
quency distributions with a view to clearly distinguishing between morphologies. Fig-
ure 18 shows an example of this technique for the three sediments listed above using
the 14th harmonic. 10 It will be seen that there are few high values of the 14th harmonic
in the sample of Brown Mountain Granite. There are, however, high values of the
14th harmonic in the sample of the Piedmond Granite.
It is possible to develop algorithms in order to distinguish between samples of sedi-
ments with different morphologies. 10 For example, in Figure 19, there is a schematic
illustration of an algorithm for distinguishing the sediments named. As one can see,
different legs of the investigation use quite different harmonics in order to classify the
data.
Ehrlich and his co-workers 9 have developed a very interesting set of techniques for
harmonic analysis of fine particle profiles. This set consists of three techniques (these
are the main parts): The X2 test, R-mode factor analysis, Q- mode factor analysis.
In essence the X 2 test yields information concerning the most important coefficients
with respect to discriminating between fine particle sets. The X 2 test therefore serves
as a spot test. The R-mode factor analysis generates basic factors that are linearly
related to the significant coefficients determined from the X 2 test. The R-mode analysis
is therefore equivalent to a qualitative analysis. The Q-mode factor analysis yields the
proportions of particle shapes in a mixture of shapes. This is a quantitative analysis
technique. This powerful set of techniques has been used to demonstrate the origin of
sand in river estuaries and the origin of littoral sands off the coast of the U.S.
Lewis and Goldman have used the Kolmogorov statistic for comparing distribu-
tions.25 The use of this statistic in place of the X 2 will be investigated.

B. Signature Analysis
This idea has been proposed and largely developed by Meloy.50 If the set of Fourier
coefficients corresponding to a holomorphic particle profile is plotted either as: a
graph of An versus log n, or of log A n versus log n, or of log A_ versus log 1/n (that is
 27

BROWN MOUNTAIN GRANITE

PIEDMONT GRANITE

FELSIC TUFF

.002

HARMONIC NUMBER

FIGURE 17. Mean harmonic amplitude spectra for Brown Mountain granite, Piedmont
granite, and Felsic t u f f . (From Pryzgocle, R. S., Master's thesis, University of South Car-
olina, Columbia, 1976. With permission.)

14th HARMONIC

PIEDMONT GRANITE

10-
5-

<J
z

BROWN MT. GRANITE

10-
5-

I I I i r I I I
.0003 .0032 .0048 .0063 .0077 .0093 .0122
HARMONIC AMPLITUDE

FIGURE 18. Discrimination of the Piedmont from the

Brown Mountain Granite based upon shape-frequency histo-
grams at the 14th harmonic. (From Pryzgocle, R. S., Master's
thesis, University of South Carolina, Columbia, 1976. With
permission).
28 Advanced Paniculate Morphology

• PIEDMONT GNEISS

• BLOWING ROCK GNEISS

• WILSON CREEK GNEISS

• 3rd ARKOSE SAMPLE

. 2nd ARKOSE SAMPLE

• 1st ARKOSE SAMPLE

ALL SAMPLES- . FELSICTUFF

• PIEDMONT GRANITE

. BROWN MOUNTAIN GRANITE

FIGURE 19. Rock discrimination algorithm, contrasting all nine samples collected
over single lithologies using harmonics 14, 12, 6, 19, 17, 18. (From Pryzgock, R.
S., Master's thesis, University of South Carolina, Columbia, 1976. With permis-
sion).

log A). The resultant curve decreases as n increases. One example of a signature is given
in Figure 20 for an atomized copper-lead particle. Another idea behind this concept is
that particles with similar signatures should lie within the same envelope. This is illus-
trated in Figure 21. One can hypothesize, for example, that particles of the same sub-
stance with different signature envelopes may have originated by different mecha-
nisms.
In the case of a Fourier representation of a particle profile, three different types of
particle signature have been proposed each corresponding to a certain type of random
particle. These are: type I — in which only the ^s vary randomly (<* is the phase angle),
type II — in which both the Ans and °cs vary randomly, type III — in which the Ans
vary randomly. The particle signature was hypothesized as a straight line such that its
equation is 5:

(24)

in which A; is the intercept of the ln(An) axis; s is the slope of the line.

1. Variation of an (Type I Randomness)

A pseudorandom number generator was used to vary the values of an and An ran-
domly. The varieties generated by RANDNO appear to be uniformly distributed on
the interval (0,1). When called each time, this generation function returns a distinct
six-digit number between 0 and 1. If multiplied by 360, a value returned by RANDNO
can be used to convert the value of an by the equation:

'n = n + 360 X RANDNO (25)

which means that a'n varies from 0° to 360° randomly. This uniformly distributed
random a'n, along with the unchanged An, was used to regenerate the particle shapes
by the definition of type I random particle.
 29

10

1.0

0.1

0.01
0.01 0.1 1.0
LOG A

FIGURE 20. Particle signature. Log An vs. Log A, Cu-Pb, Slope =

1.39.

A set of five original profiles of atomized copper-lead particles is shown in Figure

22 and the artificial regenerates of the type I random particles are shown in Figure 23
with a commentary given in the left-hand side of Table 6.

2. Variation of An and an Simultaneously

In type II, the RANDNO function was supplied to produce the randomness of An
in addition to an. In this case, An was assumed to have a variation of 50% of itself.
Therefore, An varies within 1.0 to 1.5 of its original size as following:

An = An + 0.5 X RANDNO X An (26)

A set of type II random regenerates of the five Cu-Pb particles is given in Figure 24
with the comment in the right-hand side of Table 6.
Although these examples are interesting in that they show visually some effects due
30 Advanced Paniculate Morphology

In An

In n

FIGURE 21. Signatures of two particle

sets (A and B). (From Lew, G., Beddow,
J. K., and Vetter, A. F., Powder Metal-
lurgy Int., submitted, 1979. With permis-
sion.)

to different types of randomness, it is clear that a much more basic approach will be
needed in order to develop a useful signature theory. Fortunately, this line of inquiry
is currently being pursued.''

C. Feature-Space Representation of Particle Shape12

In order to be able to represent the complex patterns of relationships to be observed
in particle morphology analysis, it is necessary to develop new ways of representing
these relationships. A recent paper discusses a prototype system of feature-space rep-
resentation. 12 The new method is a three-dimensional plot of form and roundness. A
triaxial graph in the plane of the paper is used by the observer to plot the form char-
acteristics of the particles under study. These three characteristics are thickness,
breadth, and length, respectively. The axial lengths are scaled from zero to one such
that the following relationship holds:

(L + B + T = 1) (27)

This method of graphical representation has been used to analyze form characteris-
tics of samples of sand and SiC. The results are sets of contours as shown in Figure
25. 12 The roundness of the particles is represented in the vertical direction as shown in
Figure 26.12 Those familiar with ternary phase diagrams will recognize the inherent
difficulties which one meets when using three-dimensional plots to represent complex
systems. However, one must remember that complicated and cumbersome though the
3-D plot may be, it constitutes a distinctly simpler generalization than the plethora of
data that it represents. This new method is therefore a significant advance for particle
morphology analysis; it points a way that representational methods might go, and it
also highlights the need for sophisticated computer interfaced imaging systems in order
that more data can be more easily amassed and understood.

VI. DETERMINISTIC, STATISTICAL, AND FUZZY CLASSIFIERS

In order to pursue the desired objective of relating the properties of particulate mat-
ter to the morphological characteristics of the particles therein, we have to try to un-
 31

FIGURE 22. Five COPb

particle profiles.
Random documents with unrelated
content Scribd suggests to you:
The Project Gutenberg eBook of Women
 This ebook is for the use of anyone anywhere in the United
States and most other parts of the world at no cost and with
almost no restrictions whatsoever. You may copy it, give it away
or re-use it under the terms of the Project Gutenberg License
included with this ebook or online at www.gutenberg.org. If you
are not located in the United States, you will have to check the
laws of the country where you are located before using this
eBook.

Title: Women

Author: Booth Tarkington

Release date: April 30, 2021 [eBook #65207]

Language: English

Credits: Al Haines, Cindy Beyer & the online Project Gutenberg

team at https://www.pgdpcanada.net

*** START OF THE PROJECT GUTENBERG EBOOK WOMEN ***

 BOOKS BY
B O OT H TA R K I N G TO N

ALICE ADAMS
 BEASLEY’S CHRISTMAS PARTY
BEAUTY AND THE JACOBIN
CHERRY
CONQUEST OF CANAAN
GENTLE JULIA
HARLEQUIN AND COLUMBINE
HIS OWN PEOPLE
IN THE ARENA
MONSIEUR BEAUCAIRE
PENROD
PENROD AND SAM
RAMSEY MILHOLLAND
SEVENTEEN
THE BEAUTIFUL LADY
THE FASCINATING STRANGER AND OTHER
STORIES
THE FLIRT
THE GENTLEMAN FROM INDIANA
THE GUEST OF QUESNAY
THE MAGNIFICENT AMBERSONS
THE MAN FROM HOME
THE MIDLANDER
THE TURMOIL
THE TWO VANREVELS
WOMEN
COPYRIGHT, 1925, BY DOUBLEDAY, PAGE &
COMPANY. ALL RIGHTS RESERVED. COPYRIGHT,
1924, 1925, BY BOOTH TARKINGTON.
PRINTED IN THE UNITED STATES AT THE
COUNTRY LIFE PRESS, GARDEN CITY, N. Y.
 CONTENTS

PAGE

Preamble ......
..................................................... vii
CHAPTER

I. Mrs.....................................................
. Dodge and Mrs. Cromwell 1
II. A L.....................................................
ady Across the Street 15
III. Perversity
.....................................................
of a Telephone 24
IV. A G.....................................................
reat Man’s Wife 33
V. One.....................................................
of Mrs. Cromwell’s Daughters 47
VI. Sallie
.....................................................
Ealing 63
VII. Napoleon
.....................................................
Was a Little Man 79
VIII. Mrs.....................................................
. Dodge’s Only Daughter 90
IX. Mrs.....................................................
. Dodge’s Husband 104
X. Lily.....................................................
’s Almost First Engagement 110
XI. Mrs.....................................................
. Cromwell’s Youngest Daughter 126
XII. Her.....................................................
Happiest Hour 142
XIII. Heartbreak
..................................................... 164
XIV. Mrs.....................................................
. Dodge’s Next-Door Neighbour 172
XV. Mrs.....................................................
. Dodge Declines to Tell 182
XVI. Mrs.....................................................
. Leslie Braithwaite’s Husband 206
XVII. “Dolling
.....................................................
” 216
XVIII. Lily.....................................................
’s Friend Ada 223
XIX. Parents
.....................................................
in Darkness 246
XX. Damsel
.....................................................
Dark, Damsel Fair 254
XXI. Mrs.....................................................
. Cromwell’s Niece 263
XXII. Wallflower
..................................................... 275
XXIII. The.....................................................
Strange Mirror 290
XXIV. Transfiguration
..................................................... 297
XXV. Glamour
.....................................................
Can Be Kept 309
 XXVI. Desert
.....................................................
Sand 314
XXVII. Miraculous
.....................................................
Accident 327
XXVIII. A P.....................................................
ublic Mockery 345
XXIX. Mrs.....................................................
. Cromwell’s Oldest Daughter 362
XXX. Mrs.....................................................
. Cromwell’s Sons-in-Law 400
XXXI. The.....................................................
Anniversary Dinner 410
 PREAMBLE

“BUT why not?” Mrs. Dodge said, leading the “Discussion” at the
Woman’s Saturday Club after the reading of Mrs. Cromwell’s essay,
“Women as Revealed in Some Phases of Modern Literature.” “Why
shouldn’t something of the actual life of such women as ourselves be
the subject of a book?” Mrs. Dodge inquired. “Mrs. Cromwell’s paper
has pointed out to us that in a novel a study of women must have a
central theme, must focus upon a central figure or ‘heroine,’ and
must present her as a principal participant in a centralized conflict or
drama of some sort, in relation to a limited group of other
‘characters.’ Now, so far as I can see, my own life has no such
centralizations, and I’m pretty sure Mrs. Cromwell’s hasn’t, either,
unless she is to be considered merely as a mother; but she has
other important relations in life besides her relations to her three
daughters, just as I have others besides that I bear to my one
daughter. In fact, I can’t find any central theme in Mrs. Cromwell’s
life or my own; I can’t find any centralized drama in her life or mine,
and I doubt if many of you can find such things in yours. Our lives
seem to be made up of apparently haphazard episodes, some
meaningless, others important, and although we do live principally
with our families and friends and neighbours, I find that people I
hardly know have sometimes walked casually into my life, and
influenced it, and then walked out of it as casually as they came in.
All in all, I can’t see in our actual lives the cohesion that Mrs.
Cromwell says is the demand of art. It appears to me that this very
demand might tend to the damage of realism, which I take to mean
lifelikeness and to be the most important demand of all. So I say:
Why shouldn’t a book about women, or about a type of women, take
for its subject some of the actual thoughts and doings of women like
ourselves? Why should such a book be centralized and bound down
to a single theme, a single conflict, a single heroine? The lives of
most of us here consist principally of our thoughts and doings in
relation to our children, our neighbours, and the people who casually
walk into our lives and our children’s and neighbours’ lives and out
again. It seems to me a book about us should be concerned with all
of these almost as much as with ourselves.”
“You haven’t mentioned husbands,” Mrs. Cromwell suggested.
“Wouldn’t they——”
“They should be included,” Mrs. Dodge admitted. “But I would
have husbands and suitors represented in their proper proportion;
that is to say, only in the proportion that they affect our thoughts
and doings. In challenging the rules for centralization that you have
propounded, Mrs. Cromwell, I do not propose that all rules of
whatever nature should be thrown over. One in particular I should
hold most advisable.”
“What rule is it?” a member of the club inquired, for at this point
Mrs. Dodge paused and the expression of her mouth was somewhat
grim.
“It is that a book about women should not be too long,” Mrs.
Dodge replied. “Especially if it should be by a man, he would be wise
to use brevity as a means of concealing what he doesn’t know. And
besides,” she added, more leniently, “by brevity, he might hope to
placate us a little. It might be his best form of apology.”

WOMEN
 I
MRS. DODGE AND MRS. CROMWELL

WE LEARNED in childhood that appearances are deceitful, and

our subsequent scrambling about upon this whirling globe has
convinced many of us that the most deceptive of all appearances are
those of peace. The gentlest looking liquor upon the laboratory
shelves was what removed the east wing of the Chemical
Corporation’s building on Christmas morning; it was the stillest
Sunday noon of a drowsy August when, without even the courtesy
of a little introductory sputtering, the gas works blew up; and both
of these disturbances were thought to be peculiarly outrageous
because of the previous sweet aspects that prevented any one from
expecting trouble. Yet those aspects, like the flat calm of the
summer of 1914, should have warned people of experience that
outbreaks were impending.
What could offer to mortal eye a picture of more secure placidity
than three smiling ladies walking homeward together after a club
meeting? The particular three in mind, moreover, were in a visibly
prosperous condition of life; for, although the afternoon was brightly
cold, their furs afforded proof of expenditures with which any
moderate woman would be satisfied, and their walk led them into
the most luxurious stretch of the long thoroughfare that was called
the handsome suburb’s finest street. The three addressed one
another in the caressively amiable tones that so strikingly
characterize the élite of their sex in converse; and their topic, which
had been that of the club paper, was impersonal. In fact, it was
more than impersonal, it was celestial. “Sweetness and Light: Essay.
Mrs. Roderick Brooks Battle”—these were the words printed in the
club’s year book beneath the date of that meeting, and Mrs.
Roderick Brooks Battle was the youngest of the three placid ladies.
 “You’re all so sweet to say such lovely things about it,” she said,
as they walked slowly along. “I only wish I deserved them, but of
course, as everyone must have guessed, it was all Mr. Battle. I don’t
suppose I could write a single connected paragraph without his
telling me how, and if he hadn’t kept helping me I just wouldn’t have
been ready with any paper at all. Never in the world!”
“Oh, yes, you would, Amelia,” the elder of the two other ladies
assured her. “For instance, dear, that beautiful thought about the
‘bravery of silence’—about how much nobler it is never to answer an
attack—I thought it was the finest thought in the whole paper, and
I’m sure that was your own and not your husband’s, Amelia.”
“Oh, no, Mrs. Cromwell,” Mrs. Battle returned, and although her
manner was deferential to the older woman she seemed to be gently
shocked;—her voice became a little protesting. “I could never in the
world have experienced a thought like that just by myself. It was
every bit Mr. Battle’s. In fact, he almost as much as dictated that
whole paragraph to me, word for word. It seemed a shame for me
to sit up there and appear to take the credit for it; but I knew, of
course, that everybody who knows us the least bit intimately would
understand I could never write anything and it was all Mr. Battle.”
“My dear, you’ll never persuade us of it,” the third lady said.
“There were thoughts in your paper so characteristically feminine
that no one but a woman could possibly——”
“Oh, but he could!” Mrs. Battle interrupted with an eagerness
that was more than audible, for it showed itself vividly in her
brightened eyes and the sudden glow of pink beneath them. “That’s
one of the most wonderful things about Mr. Battle: his intellect is just
as feminine as it is masculine, Mrs. Dodge. He’s absolutely—well, the
only way I can express it is in his own words. Mr. Battle says no one
can be great who isn’t universal in his thinking. And you see that’s
where he excels so immensely;—Mr. Battle is absolutely universal in
his thinking. It seems to me it’s one of the great causes of Mr.
Battle’s success; he not only has the most powerful reasoning
faculties I ever knew in any man but he’s absolutely gifted with a
woman’s intuition.” She paused to utter a little murmur of fond
laughter, as if she herself had so long and helplessly marvelled over
Mr. Battle that she tolerantly found other people’s incredulous
amazement at his prodigiousness natural but amusing. “You see, an
intellect like Mr. Battle’s can’t be comprehended from knowing other
men, Mrs. Dodge,” she added. “Other men look at things simply in a
masculine way, of course. Mr. Battle says that’s only seeing half. Mr.
Battle says women live on one hemisphere of a globe and men on
the other, and neither can look round the circle, but from the stars
the whole globe is seen—so that’s why we should keep our eyes
among the stars! I wanted to work that thought into my paper, too.
Isn’t it beautiful, the idea of keeping our eyes among the stars? But
he said there wasn’t a logical opening for it, so I didn’t. Mr. Battle
says we should never use a thought that doesn’t find its own logical
place. That is, not in writing, he says. But don’t you think it’s
wonderful—that idea of the globe and the two hemispheres and all?”
“Lovely,” Mrs. Dodge agreed. “Yet I don’t see how it proves Mr.
Battle has a feminine mind.”
“Oh, but I don’t mean just that alone,” Mrs. Battle returned
eagerly. “It’s the thousand and one things in my daily contact with
him that prove it. Of course, I know how hard it must be for other
women to understand. I suppose no one could hope to realize what
Mr. Battle’s mind is like at all without the great privilege of being
married to him.”
“And that,” Mrs. Cromwell remarked, “has been denied to so
many of us, my dear!”
Mrs. Dodge laughed a little brusquely, but the consort of the
marvellous Battle was herself so marvellous that she merely looked
preoccupied. “I know,” she said, gravely, while Mrs. Dodge and Mrs.
Cromwell stared with widening eyes, first at her and then at each
other. “How often I’ve thought of it!” she went on, her own eyes
fixed earnestly upon the distance where, in perspective, the two
curbs of the long, straight street appeared to meet. “It grows
stranger and stranger to me how such a miracle could have
happened to a commonplace little woman like me! I never shall
understand why I should have been the one selected.”
 Thereupon, having arrived at her own gate, it was with this
thought that she left them. From the gate a path of mottled
flagstones led through a smooth and snowy lawn to a house upon
which the architect had chastely indulged his Latin pleasure in stucco
and wrought iron; and as Mrs. Battle took her way over the
flagstones she received from her two friends renewed
congratulations upon her essay, as well as expressions of parting
endearment; and she replied to these cheerfully; but all the while
the glowing, serious eyes of the eager little brown-haired woman
remained preoccupied with the miracle she had mentioned.
Mrs. Cromwell and Mrs. Dodge went on their way with some
solemnity, and were silent until the closing door of the stucco house
let them know they were out of earshot. Then Mrs. Cromwell, using
a hushed voice, inquired: “Do you suppose she ever had a painting
made of the Annunciation?”
“The Annunciation?” Mrs. Dodge did not follow her.
“Yes. When the miracle was announced to her that she should be
the wife of Roderick Brooks Battle. Of course, she must have been
forewarned by an angel that she was ‘the one selected.’ If Battle had
just walked in and proposed to her it would have been too much for
her!”
“I know one thing,” Mrs. Dodge said, emphatically. “I’ve stood
just about as much of her everlasting ‘Mr. Battle says’ as I intend to!
You can’t go anywhere and get away from it; you can hear it over all
the chatter at a dinner; you can hear it over fifty women gabbing at
a tea—‘Mr. Battle says this,’ ‘Mr. Battle says that,’ ‘Mr. Battle says this
and that’! When Belloni was singing at the Fortnightly Afternoon
Music last week you could hear her ‘Mr. Battle says’ to all the women
around her, even during that loud Puccini suite, and she treed Belloni
on his way out, after the concert, to tell him Mr. Battle’s theory of
music. She hadn’t listened to a note the man sang, and Belloni
understands about two words of English, but Amelia kept right on
Mr. Battle-says-ing him for half an hour! For my part, I’ve had all I
can stand of it, and I’m about ready to do something about it!”
 “I don’t see just what one could do,” Mrs. Cromwell said,
laughing vaguely.
“I do!” her companion returned. Then both were silent for a few
thoughtful moments and wore the air of people who have introduced
a subject upon which they are not yet quite warm enough to speak
plainly. Mrs. Cromwell evidently decided to slide away from it, for the
time being, at least. “I don’t think Amelia’s looking well,” she said.
“She’s rather lost her looks these last few years, I’m afraid. She
seems pretty worn and thin to me;—she’s getting a kind of skimpy
look.”
“What else could you expect? She’s made herself the man’s slave
ever since they were married. She was his valet, his cook, and his
washerwoman night and day for years. I wonder how many times
actually and literally she’s blacked his boots for him! How could you
expect her not to get worn out and skimpy-looking?”
“Oh, I know,” Mrs. Cromwell admitted;—“but all that was in their
struggling days, and she certainly doesn’t need to do such things
now. I hear he has twenty or thirty houses to build this year, and
just lately an immense contract for two new office buildings.
Besides, he’s generous with her; she dresses well enough
nowadays.”
“Yes,” Mrs. Dodge said, grimly. “They’d both see to that for his
credit; but if he comes in with wet feet you needn’t tell me she
doesn’t get down on her knees before him and take off his shoes
herself. I know her! Yes, and I know him, too! Rich or poor, she’d be
his valet and errand girl just the same as she always was.”
“Perhaps,” said Mrs. Cromwell. “But it seems to me her most
important office for him is the one she’s just been filling.”
“Press agent? I should say so! She may stop blacking his boots,
but she’ll never stop that. It’s just why she makes me so confounded
tired, too! She thinks she’s the only woman that ever got married!”
“Amelia is rather that way,” the other said, musingly. “She
certainly never seems to realize that any of the rest of us have
husbands of our own.”
 “ ‘Mr. Battle can’t be comprehended from knowing other men!’ ”
Here Mrs. Dodge somewhat bitterly mimicked the unfortunate
Amelia’s eager voice. “ ‘Other men look at things in simply a
masculine way!’ ‘I know how hard it must be for other women to
understand a god like my husband just from knowing their own poor
little imitation husbands!’ ”
“Oh, no,” Mrs. Cromwell protested. “She didn’t quite say that.”
“But isn’t it what she meant? Isn’t it exactly what she felt?”
“Well—perhaps.”
“It does make me tired!” Mrs. Dodge said, vigorously, and with
the repetition she began to be more than vigorous. Under the spell
of that rancour which increases in people when they mull over their
injuries, she began to be indignant. “For one thing, outside of the
shamelessness of it, some of the rest of us could just possibly find a
few enthusiastic things to say of our husbands if we didn’t have
some regard for not boring one another to death! I’ve got a fairly
good husband of my own I’d like to mention once in a while, but
——”
“But, of course, you’ll never get the chance,” Mrs. Cromwell
interrupted. “Not if Amelia’s in your neighbourhood when you
attempt it.”
“What I can’t understand, though,” Mrs. Dodge went on, “is her
never having the slightest suspicion what a nuisance it is. I should
think the man himself would stop her.”
But Mrs. Cromwell laughed and shook her head. “In the first
place, of course, he agrees with her. He thinks Amelia’s just stating
facts—facts that ought to be known. In the second, don’t you
suppose he understands how useful her press-agenting is to him?”
“But it isn’t. It makes us all sick of him.”
“Oh, it may have that effect on you and me, Lydia, but I really
wonder——” Mrs. Cromwell paused, frowning seriously, then
continued: “Of course, he’d never take such a view of it. He
instinctively knows it’s useful, but he’d never take the view of it that
——”
 “The view of it that what?” Mrs. Dodge inquired, as her friend
paused again.
“Why, that it may be actually the principal reason for his success.
When he left the firm that employed him as a draughtsman and
started out for himself, with not a thing coming in for him to do,
don’t you remember that even then everybody had the impression,
somehow, that he was a genius and going to do wonders when the
chance came? How do you suppose that got to be the general
impression except through Amelia’s touting it about? And then, when
he did put up a few little houses, don’t you remember hearing it said
that they represented the first real Architecture with a capital ‘A’ ever
seen in the whole city? Now, almost nobody really knows anything
about architecture, though we all talk about it as glibly as if we did,
and pretty soon—don’t you remember?—we were all raving over
those little houses of Roderick Brooks Battle’s. What do you suppose
made us rave? We must have been wrong, because Amelia says now
that Battle thinks those first houses of his were ‘rather bad’—he’s
‘grown so tremendously in his art.’ Well, since they were bad, what
except Amelia made us think then that they were superb? And look
at what’s happened to Battle these last few years. In spite of
Amelia’s boring us to death about him, isn’t it true that there’s
somehow a wide impression that he’s a great man? Of course there
is!”
“And yet,” Mrs. Dodge interposed, “he’s not done anything that
proves it. Battle’s a good architect, certainly, but there are others as
good, and he’s not a bit better as an architect than Mr. Cromwell is
as a lawyer or than my husband is as a consulting engineer.”
“Not a bit,” Mrs. Cromwell echoed, carrying on the thought she
had been following. “But Mr. Dodge and Mr. Cromwell haven’t had
anybody to go about, day after day for years, proclaiming them and
building up a legend about them. Nobody has any idea that they’re
great men, poor things! Don’t you see where that puts you and me,
Lydia?”
“No, I don’t.”
 “My dear!” Mrs. Cromwell exclaimed. “Why, even Battle himself
didn’t know that he was a great man until he married Amelia and
she believed he was—and told him he was—and started her long
career of going about making everybody else sort of believe it, too.”
“I think it’s simply her own form of egoism,” said the emphatic
Mrs. Dodge. “She’d have done exactly the same whoever she
married.”
“Precisely! It’s Amelia’s way of being in love—she’s a born
idolizer. But you didn’t answer me when I asked you where that puts
us.”
“You and me?” Mrs. Dodge inquired, frowning.
“Don’t you see, if she’d married my husband, for instance,
instead of Battle, everybody’d be having the impression by this time
that Mr. Cromwell is a great man? He’d have felt that way himself,
too, and I’m afraid it would give him a great deal of pleasure.
Haven’t we failed as wives when we see what Amelia’s done for her
husband?”
“What an idea!” The two ladies had been walking slowly as they
talked;—now they came to a halt at their parting place before Mrs.
Cromwell’s house, which was an important, even imposing, structure
of the type called Georgian, and in handsome conventional solidity
not unlike the lady who lived in it. Across the broad street was a
newer house, one just finished, a pinkish stucco interpretation of
Mediterranean gaiety, and so fresh of colour that it seemed rather a
showpiece, not yet actually inhabited though glamoured with
brocaded curtains and transplanted arbor vitæ into the theatrical
semblance of a dwelling in use. Mrs. Dodge glanced across at it with
an expression of disfavour. “I call the whole thing perfectly
disgusting!” she said.
 II
A LADY ACROSS THE STREET

MRS. CROMWELL also looked at the new house; then she shook
her head. “It’s painful, rather,” she said, and evidently referred to
something more than the house itself.
“Outright disgusting!” her friend insisted. “I suppose he’s there as
much as ever?”
“Oh, yes. Rather more.”
“Well, I’ll say one thing,” Mrs. Dodge declared; “Amelia Battle
won’t get any sympathy from me!”
“Sympathy? My dear, you don’t suppose she dreams she needs
sympathy! Doesn’t she show the rest of us every day how she pities
us because we’re not married to Roderick Brooks Battle?”
“Yes, and that’s what makes me so furious. But she will need
sympathy,” Mrs. Dodge persisted, with a dark glance at the new
house across the street. “She will when she knows about that!”
“But maybe she’ll never know.”
“What!” Mrs. Dodge laughed scornfully. “My dear, when a woman
builds a man into a god he’s going to assume the privileges of a
god.”
“And behave like the devil?”
“Just that,” Mrs. Dodge returned, grimly. “Especially when his
idolater has burnt up her youth on his altar and her friends begin to
notice she’s getting a skimpy look. What chance has a skimpy-
looking slave against a glittering widow rich enough to build a new
house every time she wants to have tête-à-têtes with a godlike
architect?”
“But she’s only built one,” Mrs. Cromwell cried, protesting.
 “So far!” her pessimistic companion said; then laughed at her
own extravagance, and became serious again. “I think Amelia ought
to know.”
“Oh, no!”
“Yes, she ought,” Mrs. Dodge insisted. “In the first place, she
ought to be saved from making herself so horribly ridiculous. Of
course, she’s always been ridiculous; but the way she raves about
him when he’s raving about another woman—why, it’s too ridiculous!
In the second place, if she knew something about the Mrs. Sylvester
affair now it might help her to bear a terrific jolt later.”
“What terrific jolt, Lydia?”
“If he leaves her,” Mrs. Dodge said, gravely. “If Mrs. Sylvester
decides to make him a permanent fixture. Men do these things
nowadays, you know.”
“Yes, I know they do.” Mrs. Cromwell looked as serious as her
friend did, though her seriousness was more sympathetically a
troubled one than Mrs. Dodge’s. “Poor Amelia! To wear her youth
out making a man into such a brilliant figure that a woman of the
Sylvester type might consider him worth while taking away from her
——”
“Look!” Mrs. Dodge interrupted in a thrilled voice.
A balustraded stone terrace crossed the façade of the new
house, and two people emerged from a green door and appeared
upon the terrace. One was a man whose youthful figure made a
pleasing accompaniment to a fine and scholarly head;—he produced,
moreover, an impression of success and distinction obvious to the
first glance of a stranger, though what was most of all obvious about
him at the present moment was his devoted, even tender, attention
to the woman at his side. She was a tall and graceful laughing
creature, so sparklingly pretty as to approach the contours and
colours of a Beauty. Her rippling hair glimmered with a Venetian
ruddiness, and the blue of her twinkling eyes was so vivid that a
little flash of it shot clear across the street and was perceptible to
the two observant women as brightest azure.
 Upon her lovely head she had a little sable hat, and, over a dress
of which only a bit of gray silk could be glimpsed at throat and
ankle, she wore a sable coat of the kind and dimensions staggering
to moderate millionaires. She had the happy and triumphant look of
a woman confident through experience that no slightest wish of hers
would ever be denied by anybody, herself distinctly included; and, all
in all, she was dazzling, spoiled, charming, and fearless.
Certainly she had no fear of the two observant women, neither of
their opinion nor of what she might give them cause to tell;—that
sparkle of azure she sent across the intervening street was so
carelessly amused it was derisive, like the half nod to them with
which she accompanied it. She and her companion walked closely
together, absorbed in what they were saying, her hand upon his
arm; and, when they came to the terrace steps, where a closed
foreign car waited, with a handsome young chauffeur at the wheel
and a twin of him at attention beside the door, she did a thing that
Mrs. Dodge and Mrs. Cromwell took to be final and decisive.
Her companion had evidently offered some light pleasantry or
witticism at which she took humorous offense, for she removed her
white-gloved hand from his arm and struck him several times
playfully upon the shoulder—but with the last blow allowed her hand
to remain where it was; and, although she might have implied that it
was to aid her movement into the car, the white fingers could still be
seen remaining upon the shoulder of the man’s brown overcoat as
he, moving instantly after her, took his seat beside her in the gray
velvet interior. Thus, what appeared to be a playful gesture
protracted itself into a caress, and a caress of no great novelty to
the participants.
At least, it was so interpreted across the street, where Mrs.
Dodge gave utterance to a sound vocal but incoherent, and Mrs.
Cromwell said “Oh, my!” in a husky whisper. The French car glided
by them, passing them as they openly stared at it, or indeed glared
at it, and a moment later it was far down the street, leaving them to
turn their glares upon each other.
 “That settles it,” Mrs. Dodge gasped. “It ought to have been a
gondola.”
“A gondola?”
“A Doge’s wife carrying on with a fool poet or something;—she
always has that air to me. What a comedy!”
Mrs. Cromwell shook her head; her expression was of grief and
shock. “It’s tragedy, Lydia.”
“Just as you choose to look at it. The practical point of view is
that it’s going to happen to Amelia, and pretty soon, too! Some day
before long that man’s going to walk in and tell her she’s got to step
aside and let him marry somebody else. Doesn’t what we just saw
prove it? That woman did it deliberately in our faces, and she knows
we’re friends of his wife’s. She deliberately showed us she didn’t
care what we saw. And as for him——”
“He didn’t see us, I think,” Mrs. Cromwell murmured.
“See us? He wouldn’t have seen Amelia herself if she’d been with
us—and she might have been! That’s why I say she ought to know.”
“Oh, I don’t think I’d like to——”
“Somebody ought to,” Mrs. Dodge said, firmly. “Somebody ought
to tell her, and right away, at that.”
“Oh, but——”
“Oughtn’t she to be given the chance to prepare herself for
what’s coming to her?” Mrs. Dodge asked, testily. “She’s made that
man think he’s Napoleon, and so she’s going to get what Napoleon’s
wife got. I think she ought to be warned at once, and a true friend
would see to it.”
In genuine distress, Mrs. Cromwell shrank from the idea. “Oh,
but I could never——”
“Somebody’s got to,” Mrs. Dodge insisted, implacably. “If you
won’t, then somebody else.”
“Oh, but you—you wouldn’t take such a responsibility, would
you? You—you wouldn’t, would you, Lydia?”
 The severe matron, Lydia Dodge, thus flutteringly questioned,
looked more severe than ever. “I shouldn’t care to take such a
burden on my shoulders,” she said. “Looking after my own burdens
is quite enough for me, and it’s time I was on my way to them.” She
moved in departure, but when she had gone a little way, spoke over
her shoulder, “Somebody’s got to, though! Good-bye.”
Mrs. Cromwell, murmuring a response, entered her own domain
and walked slowly up the wide brick path; then halted, turned
irresolutely, and glanced to where her friend marched northward
upon the pavement. To Mrs. Cromwell the outlines of Mrs. Dodge,
thus firmly moving on, expressed something formidable and
imminent. “But, Lydia——” the hesitant lady said, impulsively, though
she knew that Lydia was already too distant to hear her. Mrs.
Cromwell took an uncertain step or two, as if to follow and
remonstrate, but paused, turned again, and went slowly into her
house.
A kind-hearted soul, and in a state of sympathetic distress for
Amelia Battle, she was beset by compassion and perplexity during
what remained of the afternoon; and her husband and daughters
found her so preoccupied at the dinner-table that they accused her
of concealing a headache. But by this time what she concealed was
an acute anxiety; she feared that Lydia’s sense of duty might lead to
action, and that the action might be precipitate and destructive. For
Mrs. Cromwell knew well enough that Amelia’s slavery was Amelia’s
paradise—the only paradise Amelia knew how to build for herself—
and paradises are, of all structures, the most perilously fragile.
Mrs. Cromwell was the more fearful because, being a woman,
she understood that more than a sense of duty would impel Lydia to
action: Lydia herself might interpret her action as the prompting of
duty, but the vital incentive was likely to be something much more
human; for within the race is a profound willingness to see a proud
head lowered, particularly if that head be one that has displayed its
pride. Amelia had displayed hers too long and too gallingly for
Lydia’s patience;—Lydia had “really meant it,” Mrs. Cromwell
thought, recalling the fierceness of Mrs. Dodge’s “I’ve had all I can
stand of it!” that afternoon. A sense of duty with gall behind it is
indeed to be feared; and the end of Mrs. Cromwell’s anxieties was
the conclusion that Amelia’s paradise of slavery was more
imminently threatened by the virtuous Lydia than by that gorgeous
pagan, Mrs. Sylvester.
 III
PERVERSITY OF A TELEPHONE

THE troubled lady began to wish devoutly that the sight of Mrs.
Sylvester caressing Mr. Battle had not shocked her into a fluttering
and indecisive state of mind;—she should have discussed the event
more calmly with Lydia; should have argued against anything
precipitate;—and so, as soon as she could, after her preoccupied
dinner, she went to the telephone and gave Mrs. Dodge’s number.
Mr. and Mrs. Dodge were dining in town, she was informed; they
were going to the theatre afterward and were not expected to return
until midnight. This blank wall at once increased Mrs. Cromwell’s
inward disturbance, for she was a woman readily tortured by her
imagination; and in her mind she began to design terrible pictures of
what might now be happening in the house of the Battles. Until she
went to the telephone she thought it unlikely that Lydia had acted
with such promptness; but after receiving through the instrument
the information that no information was to be had for the present,
Mrs. Cromwell became certain that Mrs. Dodge had already
destroyed Amelia’s peace of mind.
She went away from the telephone, then came back to it, and
again sat before the little table that bore it; but she did not at once
put its miraculous powers into operation. Instead, she sat staring at
it, afraid to employ it, while her imaginings became more piteous
and more horrifying. Amelia had no talk except “Mr. Battle says”; she
had no thought except “Mr. Battle thinks”; she had no life at all
except as part of her husband’s life; and if that were taken away
from her, what was left? She had made no existence whatever of her
own and for herself, and if brought to believe that she had lost him,
she was annihilated.
 If the great Battle merely died, Amelia could live on, as widows
of the illustrious sometimes do, to be his monument continually
reinscribed with mourning tributes; but if a Venetian beauty carried
him off in a gondola, Amelia would be so extinct that the act of self-
destruction might well be thought gratuitous;—and yet Mrs.
Cromwell’s imagination pictured Amelia in the grisly details of its
commission by all the usual processes. She saw Amelia drown
herself variously; saw her with a razor, with a pistol, with a rope,
with poison, with a hat-pin.
Naturally, it became impossible to endure such pictures, and Mrs.
Cromwell tremulously picked up the telephone, paused before
releasing the curved nickel prong, but did release it, and when a
woman’s voice addressed her, “What number, please?” she returned
the breathless inquiry: “Is that you, Amelia?” Then she apologized,
pronounced a number, and was presently greeted by the response:
“Mr. Roderick Battle’s residence. Who is it, please?”
“Mrs. Cromwell. May I speak to Mrs. Battle?”
“I think so, ma’am.”
In the interval of silence Mrs. Cromwell muttered, “I think so” to
herself. The maid wasn’t certain;—that was bad; for it might indicate
a state of prostration.
“Yes?” said the little voice in the telephone. “Is it Mrs. Cromwell?”
Mrs. Cromwell with a great effort assumed her most smiling and
reassuring expression. “Amelia? Is it you, Amelia?”
“Yes.”
“I just wanted to tell you again what a lovely impression your
essay made on me, dear. I’ve been thinking of it ever since, and I
felt you might like to know it.”
“Thank you, Mrs. Cromwell.”
“Lydia Dodge and I kept on talking about it after you left us this
afternoon,” Mrs. Cromwell continued, beaming fondly upon the air
above the telephone. “We both said we thought it was the best
paper ever read at the club. I—I just wondered if—if Lydia called you
up to tell you so, too. Did she?”
 “No. No, she didn’t call me up.”
“Oh, didn’t she? I just thought she might have because she was
so enthusiastic.”
“No. She didn’t.”
Mrs. Cromwell listened intently, seeking to detect emotion that
might indicate Amelia’s state of mind, but Amelia’s voice revealed
nothing whatever. It was one of those voices obscured and dwindled
by the telephone into dry little metallic sounds; language was
communicated, but nothing more, and a telegram from her would
have conveyed as much personal revelation. “No, Mrs. Dodge didn’t
call me up,” she said again.
Mrs. Cromwell offered some manifestations of mirth, though she
intended them to express a tender cordiality rather than
amusement; and the facial sweetness with which she was favouring
the air before her became less strained; a strong sense of relief was
easing her. “Well, I just thought Lydia might, you know,” she said,
continuing to ripple her gentle laughter into the mouthpiece. “She
was so enthusiastic, I just thought——”
“No, she didn’t call me up,” the small voice in the telephone
interrupted.
“Well, I’m gl——” But Mrs. Cromwell checked herself sharply,
having begun too impulsively. “I hope I’m not keeping you from
anything you were doing,” she said hastily, to change the subject.
“No, I’m all alone. Mr. Battle is spending the evening with Mrs.
Sylvester.”
“What!” Mrs. Cromwell exclaimed, and her almost convivial
expression disappeared instantly; her face became a sculpture of
features only. “He is?”
“Yes. He’s finishing the interior of her new house. With important
clients like that he always interprets them into their houses you
know. He makes a study of their personalities.”
“I—see!” Mrs. Cromwell said. Then, recovering herself, she was
able to nod pleasantly and beam again, though now her beaming
was rigidly automatic. “Well, I mustn’t keep you. I just wanted to tell
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!

ebookgate.com