Download Full Version ebook - Visit ebooknice.

com

(Ebook) Reliability and Statistical Computing:

Modeling, Methods and Applications by Hoang Pham
(editor) ISBN 9783030434113, 3030434117

https://ebooknice.com/product/reliability-and-statistical-
computing-modeling-methods-and-applications-11035302

Click the button below to download

DOWLOAD EBOOK

Discover More Ebook - Explore Now at ebooknice.com

 Instant digital products (PDF, ePub, MOBI) ready for you
Download now and discover formats that fit your needs...

Start reading on any device today!

(Ebook) Safety and Reliability Modeling and Its

Applications (Advances in Reliability Science) by Mangey
Ram, Hoang Pham ISBN 9780128233238, 0128233230
https://ebooknice.com/product/safety-and-reliability-modeling-and-its-
applications-advances-in-reliability-science-34383582

ebooknice.com

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

Viles, James ISBN 9781459699816, 9781743365571,
9781925268492, 1459699815, 1743365578, 1925268497
https://ebooknice.com/product/biota-grow-2c-gather-2c-cook-6661374

ebooknice.com

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

II Success) by Peterson's ISBN 9780768906677, 0768906679

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

ebooknice.com

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

Alfredsson, Hans Heikne, Sanna Bodemyr ISBN 9789127456600,
9127456609
https://ebooknice.com/product/matematik-5000-kurs-2c-larobok-23848312

ebooknice.com
(Ebook) Methods and Applications for Modeling and
Simulation of Complex Systems: 18th Asia Simulation
Conference, AsiaSim 2018, Kyoto, Japan, October 27–29,
2018, Proceedings by Liang Li, Kyoko Hasegawa, Satoshi
https://ebooknice.com/product/methods-and-applications-for-modeling-
Tanaka ISBN 9789811328527, 9789811328534, 9811328528,
and-simulation-of-complex-systems-18th-asia-simulation-conference-
asiasim-2018-kyoto-japan-october-2729-2018-proceedings-7329132
9811328536
ebooknice.com

(Ebook) Vagabond, Vol. 29 (29) by Inoue, Takehiko ISBN

9781421531489, 1421531488

https://ebooknice.com/product/vagabond-vol-29-29-37511002

ebooknice.com

(Ebook) Classical Hopf Algebras and Their Applications

(Algebra and Applications, 29) by Pierre Cartier, Frédéric
Patras ISBN 9783030778446, 3030778444
https://ebooknice.com/product/classical-hopf-algebras-and-their-
applications-algebra-and-applications-29-34794516

ebooknice.com

(Ebook) Master SAT II Math 1c and 2c 4th ed (Arco Master

the SAT Subject Test: Math Levels 1 & 2) by Arco ISBN
9780768923049, 0768923042
https://ebooknice.com/product/master-sat-ii-math-1c-and-2c-4th-ed-
arco-master-the-sat-subject-test-math-levels-1-2-2326094

ebooknice.com

(Ebook) Cambridge IGCSE and O Level History Workbook 2C -

Depth Study: the United States, 1919-41 2nd Edition by
Benjamin Harrison ISBN 9781398375147, 9781398375048,
1398375144, 1398375047
https://ebooknice.com/product/cambridge-igcse-and-o-level-history-
workbook-2c-depth-study-the-united-states-1919-41-2nd-edition-53538044

ebooknice.com
Springer Series in Reliability Engineering

Hoang Pham Editor

Reliability
and Statistical
Computing
Modeling, Methods and Applications
Springer Series in Reliability Engineering

Series Editor
Hoang Pham, Department of Industrial and Systems Engineering, Rutgers
University, Piscataway, NJ, USA
Today’s modern systems have become increasingly complex to design and build,
while the demand for reliability and cost effective development continues.
Reliability is one of the most important attributes in all these systems, including
aerospace applications, real-time control, medical applications, defense systems,
human decision-making, and home-security products. Growing international
competition has increased the need for all designers, managers, practitioners,
scientists and engineers to ensure a level of reliability of their product before release
at the lowest cost. The interest in reliability has been growing in recent years and
this trend will continue during the next decade and beyond.
The Springer Series in Reliability Engineering publishes books, monographs and
edited volumes in important areas of current theoretical research development in
reliability and in areas that attempt to bridge the gap between theory and application in
areas of interest to practitioners in industry, laboratories, business, and government.
**Indexed in Scopus**
Interested authors should contact the series editor, Hoang Pham, Department
of Industrial and Systems Engineering, Rutgers University, Piscataway,
NJ 08854, USA. Email: [email protected], or Anthony Doyle,
Executive Editor, Springer, London. Email: [email protected].

More information about this series at http://www.springer.com/series/6917

Hoang Pham
Editor

Reliability and Statistical

Computing
Modeling, Methods and Applications

123
Editor
Hoang Pham
Department of Industrial and Systems
Engineering
Rutgers University
Piscataway, NJ, USA

ISSN 1614-7839 ISSN 2196-999X (electronic)

Springer Series in Reliability Engineering
ISBN 978-3-030-43411-3 ISBN 978-3-030-43412-0 (eBook)
https://doi.org/10.1007/978-3-030-43412-0
© Springer Nature Switzerland AG 2020
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface

We’re living in an era of fast and unpredictable change. Billions of people are
connected to each other through their mobile devices and the Internet of Things
(IoT). Data is being collected and processed like never before. The era of AI
through reliability and statistical machine computing as well as intelligent and
recommender systems with almost all applications and service industry has expe-
rienced a dramatic shift in the past two decades to a truly global industry, known as
the Industry 4.0. The forces that have driven this change are still at play and will
continue. Most of the products which affect our daily lives are becoming even more
complex than ever. This book, consisting of 18 chapters, aims to address both
research and practical aspects in reliability and statistical computing with emphasis
on the applications. Each chapter is written by active researchers and experienced
practitioners in the field aiming to connect the gap between the theory and practice
and to trigger new research challenges in reliability and statistical computing of
recent complex products and the customer needs in practices.
In chapter “Fatigue Life Distribution Estimation”, it discusses an empirically
based methodology for estimating the cumulative distribution functions for fatigue
life that incorporates available fatigue life data for various stresses given the applied
load. Chapter “Reliability Improvement Analysis Using Fractional Failure”
describes an approach to improve product reliability during development by using
fractional failure analysis method incorporating failure fix effectiveness during each
testing and failure fix phase. The chapter also describes a new product introduction
reliability improvement process with fractional failure analysis. Chapter “Modelling
Innovation Paths of European Firms Using Fuzzy Balanced Scorecard” discusses an
approach for modeling innovation paths of European firms and its strategic per-
formance measurement by integrating fuzzy balanced scorecard. The chapter also
provides empirical evidence for the effectiveness of the approach based on a large
dataset of European firms. Chapter “Innovation Environment in Europe—Efficiency
Analysis Case Study” discusses ways to assess the effectiveness, performance, and
productivity of comparable production units within EU28 countries based on their
innovation performance measured by European Commissions’ European
Innovation Scoreboard 2017.

v
vi Preface

In chapter “Application of Artificial Intelligence in Modeling a Textile Finishing

Process”, it discusses an application of three machine learning techniques such as
extreme learning machine, support vector regression, and random forest in mod-
eling a textile finishing process to predict the color properties of ozonated
reactive-dyed cottons. Chapter “Developing Alert Level for Aircraft Components”
illustrates the development of an alert level for a helicopter air-conditioning system
(ACS) incorporating the temperature, humidity, and airflow inside the helicopter for
the purpose of improving operational performance.
Reliable network is an important problem in many real-world applications such as
traffic planning, computer network planning, and power transmission planning.
Chapter “Computation in Network Reliability” discusses procedures and algorithms
to compute the reliability of complex networks. The chapter also presents some
numerical calculations of some networks to illustrate the proposed algorithms.
Chapter “Integrating Sentiment Analysis in Recommender Systems” discusses the
use of hybrid deep learning approach for sentiment analysis task and matrix fac-
torization for recommender systems. The chapter also presents an experiment based
on Amazon food reviews and its performance of a recommendation system based on
root mean square errors score. Chapter “Crowdsourcing Platform for Collecting
Cognitive Feedbacks from Users: A Case Study on Movie Recommender System”
presents a crowd sourcing-based recommendation platform that can collect and share
cognitive feedbacks from users for a case study on movie recommender system.
Aiming to create a safety culture that would help to reduce the driver casualty
involved in traffic accidents, chapter “A DD-SHELL HF Model for Bus Accidents”
presents an analysis about the causes of the accident with respect to both human
factors and organizational lapses using the combination of the aviation dirty dozen
factor approach and the SHELL model. Chapter “Development of MI-ANFIS-BBO
Model for Forecasting Crude Oil Price” studies the adaptive neuro-fuzzy inference
system (ANFIS) with consideration of parameter optimization using the
biogeography-based optimization (BBO) algorithm and mutual information (MI)
technique for forecasting crude oil price. The approach combines the strengths of
fuzzy logic, neural network, and the heuristic algorithm to detect the trends and
patterns in crude oil price field data and improve the quality of forecasting. Chapter
“Obtaining More Specific Topics and Detecting Weak Signals by Topic Word
Selection” presents a method to automatically evaluate specific topics with respect
to their usefulness for applications like the detection of new innovations and detect
weak signals produced by latent Dirichlet allocation. The chapter also discusses the
quality of a topic in terms of both coherence and specificity by using context
information from the document corpus.
In chapter “An Improved Ensemble Machine Learning Algorithm for Wearable
Sensor Data Based Human Activity Recognition”, it presents an improved machine
learning method for wearable sensor data based human activity recognition (HAR)
using convolution neural networks (CNN) and long short-term memory (LSTM)
networks. The proposed method shows an improved performance of these machine
learning methods for HAR. Chapter “Average Failure Rate and Its Applications of
Preventive Replacement Policies” discusses the average failure rate, which is based
Preface vii

on the conditional failure probability and the mean time to failure, given that the
unit is still survival at the mission arrival time where the mission arrival time
follows a gamma distribution and the failure time of the unit has a Weibull dis-
tribution. The chapter also presents various optimization maintenance models
including age replacement policies and the periodic policies with minimal repairs.
Chapter “Optimal Maintenance Models of Social Infrastructures Considering
Natural Disasters” discusses several optimal maintenance models of social infras-
tructures considering delay, multiple degradation levels, modified costs which
depend on damage levels, and disaster recovery which minimizes the expected cost
rates. The cost factors considered in the model include maintenance periods,
maintenance time delay, wide variety of preventive maintenance costs and degra-
dation levels, and natural disaster distribution.
In chapter “Optimal Checkpoint Intervals, Schemes and Structures for
Computing Modules”, it studies a high-reliability computing system with redun-
dancy techniques and recovery methods to prevent failure occurrences considering
two checkpoint schemes in which their respective interval times are periodic and
random. The chapter discusses the optimal checkpoint interval policies and struc-
tures with computing modules for periodic and random checkpoint models. Chapter
“Data Envelopment Analysis as a Tool to Evaluate Marketing Policy Reliability”
describes the data envelopment analysis design and its applications for effectiveness
evaluation of company marketing strategies. The data envelopment analysis
approach is a useful tool that can provide suggestions on the optimal marketing
strategies to achieve superior performance. Obtaining efficient and reliable software
under resource allocation constraint is an important goal of software management
science. Chapter “Computational Intelligence Approaches for Software Quality
Improvement” discusses the usage of artificial immune algorithms for software
testing and recent development paradigms of computational intelligence for soft-
ware quality improvement.
Postgraduates, researchers, data scientists, and engineers will definitely gain
great knowledge with the help of this book in the areas of reliability engineering
and safety, applied statistics, machine learning and recommender systems, and its
applications. The material is proposed for graduate and advanced undergraduate
level students.
I acknowledge Springer for this opportunity and professional support.
Importantly, I would like to thank all the chapter authors and reviewers for their
availability for this work.

Piscataway, NJ, USA Hoang Pham

January 2020
Contents

Fatigue Life Distribution Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

D. Gary Harlow
Reliability Improvement Analysis Using Fractional Failure . . . . . . . . . . 17
Mingxiao Jiang and Feng-Bin Sun
Modelling Innovation Paths of European Firms Using Fuzzy
Balanced Scorecard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Petr Hájek, Jan Stejskal, Michaela Kotková Stříteská, and Viktor Prokop
Innovation Environment in Europe—Efficiency Analysis
Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Viktor Prokop, Jan Stejskal, Petr Hájek, and Michaela Kotková Stříteská
Application of Artificial Intelligence in Modeling a Textile
Finishing Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Zhenglei He, Kim Phuc Tran, Sébastien Thomassey, Xianyi Zeng,
and Changhai Yi
Developing Alert Level for Aircraft Components . . . . . . . . . . . . . . . . . . 85
Wai Yeung Man and Eric T. T. Wong
Computation in Network Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Shin-Guang Chen
Integrating Sentiment Analysis in Recommender Systems . . . . . . . . . . . 127
Bui Thanh Hung
Crowdsourcing Platform for Collecting Cognitive Feedbacks
from Users: A Case Study on Movie Recommender System . . . . . . . . . 139
Luong Vuong Nguyen and Jason J. Jung
A DD-SHELL HF Model for Bus Accidents . . . . . . . . . . . . . . . . . . . . . . 151
Kelvin K. F. Po and Eric T. T. Wong

ix
x Contents

Development of MI-ANFIS-BBO Model for Forecasting

Crude Oil Price . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Quang Hung Do
Obtaining More Specific Topics and Detecting Weak Signals
by Topic Word Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Laura Kölbl and Michael Grottke
An Improved Ensemble Machine Learning Algorithm
for Wearable Sensor Data Based Human Activity Recognition . . . . . . . 207
Huu Du Nguyen, Kim Phuc Tran, Xianyi Zeng, Ludovic Koehl,
and Guillaume Tartare
Average Failure Rate and Its Applications of Preventive
Replacement Policies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
Xufeng Zhao, Jiajia Cai, Satoshi Mizutani, and Toshio Nakagawa
Optimal Maintenance Models of Social Infrastructures
Considering Natural Disasters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Takumi Kishida, Kodo Ito, Higuchi Yoshiyuki, and Toshio Nakagawa
Optimal Checkpoint Intervals, Schemes and Structures
for Computing Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
Kenichiro Naruse and Toshio Nakagawa
Data Envelopment Analysis as a Tool to Evaluate Marketing
Policy Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
Zaytsev Dmitry and Kuskova Valentina
Computational Intelligence Approaches for Software Quality
Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
Grigore Albeanu, Henrik Madsen, and Florin Popențiu-Vlădicescu
Editor and Contributors

About the Editor

Dr. Hoang Pham is a Distinguished Professor and former Chairman (2007–2013)

of the Department of Industrial and Systems Engineering at Rutgers University,
New Jersey. Before joining Rutgers, he was a Senior Engineering Specialist with
the Boeing Company and the Idaho National Engineering Laboratory. He has been
served as Editor-in-Chief, Editor, Associate Editor, Guest Editor, and board
member of many journals. He is the Editor of Springer Book Series in Reliability
Engineering, the Editor of World Scientific Book Series on Industrial and Systems
Engineering, and has served as Conference Chair and Program Chair of over 40
international conferences. He is the author or coauthor of 6 books and has published
over 185 journal articles and edited 13 books including Springer Handbook in
Engineering Statistics and Handbook in Reliability Engineering. He has delivered
over 40 invited keynote and plenary speeches at many international conferences.
His numerous awards include the 2009 IEEE Reliability Society Engineer of the
Year Award. He is a Fellow of the Institute of Electrical and Electronics Engineers
(IEEE) and the Institute of Industrial Engineers (IIE).

Contributors

Grigore Albeanu “Spiru Haret” University, Bucharest, Romania

Jiajia Cai College of Economics and Management, Nanjing University of
Aeronautics and Astronautics, Nanjing, China
Shin-Guang Chen Tungnan University, New Taipei City, Taiwan
Zaytsev Dmitry International Laboratory for Applied Network Research, National
Research University Higher School of Economics, Moscow, Russia

xi
xii Editor and Contributors

Quang Hung Do Faculty of Information Technology, University of Transport

Technology, Hanoi, Vietnam
D. Gary Harlow Mechanical Engineering and Mechanics, Lehigh University,
Bethlehem, PA, USA
Michael Grottke GfK SE, Global Data Science, Nürnberg, Germany
Zhenglei He GEMTEX – Laboratoire de Génie et Matériaux Textiles, ENSAIT,
Lille, France
Bui Thanh Hung Data Analytics & Artificial Intelligence Laboratory, Faculty of
Engineering and Technology, Thu Dau Mot University, Thu Dau Mot, Binh
Duong, Vietnam
Petr Hájek Faculty of Economics and Administration, University of Pardubice,
Pardubice, Czech Republic
Kodo Ito Department of Social Management Engineering, Graduate School of
Engineering, Tottori University, Tottori, Koyama-cho Minami, Japan
Mingxiao Jiang Minneapolis, MN, USA
Jason J. Jung Department of Computer Engineering, Chung-Ang University,
Seoul, Korea
Ludovic Koehl Ecole Nationale Supérieure des Arts et Industries Textiles,
GEMTEX Laboratory, Roubaix, France
Takumi Kishida Department of Social Management Engineering, Graduate
School of Engineering, Tottori University, Tottori, Koyama-cho Minami, Japan
Michaela Kotková Stříteská Faculty of Economics and Administration,
University of Pardubice, Pardubice, Czech Republic
Laura Kölbl Friedrich-Alexander-Universität Erlangen-Nürnberg, Nürnberg,
Germany
Henrik Madsen Danish Technical University, Lyngby, Denmark
Wai Yeung Man Department of Mechanical Engineering, The Hong Kong
Polytechnic University, Hong Kong, China
Satoshi Mizutani Department of Business Administration, Aichi Institute of
Technology, Toyota, Japan
Toshio Nakagawa Department of Business Administration, Aichi Institute of
Technology, Toyota, Aichi, Japan
Kenichiro Naruse Nagoya Sangyo University, Yamanota Araicho, Owariasahi,
Aichi, Japan
Editor and Contributors xiii

Huu Du Nguyen Division of Artificial Intelligence, Dong-A University, Danang,

Vietnam
Luong Vuong Nguyen Department of Computer Engineering, Chung-Ang
University, Seoul, Korea
Kelvin K. F. Po Department of Mechanical Engineering, The Hong Kong
Polytechnic University, Hong Kong, China
Florin Popențiu-Vlădicescu University Politehnica of Bucharest Bucharest,
Romania; Academy of Romanian Scientists, Bucharest, Romania
Viktor Prokop Faculty of Economics and Administration, University of
Pardubice, Pardubice, Czech Republic
Jan Stejskal Faculty of Economics and Administration, University of Pardubice,
Pardubice, Czech Republic
Feng-Bin Sun Tesla, Inc., Palo Alto, CA, USA
Guillaume Tartare Ecole Nationale Supérieure des Arts et Industries Textiles,
GEMTEX Laboratory, Roubaix, France
Sébastien Thomassey GEMTEX – Laboratoire de Génie et Matériaux Textiles,
ENSAIT, Lille, France
Kim Phuc Tran ENSAIT & GEMTEX, Roubaix, France
Kuskova Valentina International Laboratory for Applied Network Research,
National Research University Higher School of Economics, Moscow, Russia
Eric T. T. Wong Department of Mechanical Engineering, The Hong Kong
Polytechnic University, Hong Kong, China
Changhai Yi National Local Joint Engineering Laboratory for Advanced Textile
Processing and Clean Production, Wuhan Textile University, Wuhan, China
Higuchi Yoshiyuki Faculty of Symbiotic Systems Science, Fukushima University,
Kanayagawa, Fukushima, Japan
Xianyi Zeng ENSAIT & GEMTEX, Roubaix, France
Xufeng Zhao College of Economics and Management, Nanjing University of
Aeronautics and Astronautics, Nanjing, China
Fatigue Life Distribution Estimation

D. Gary Harlow

Abstract Modeling fatigue life is complex whether it is applied to structures or

experimental programs. Through the years several empirical approaches have been
utilized. Each approach has positive aspects; however, none have been acceptable for
every circumstance. On many occasions the primary shortcoming for an empirical
method is the lack of a sufficiently robust database for statistical modeling. The mod-
eling is exacerbated for loading near typical operating conditions because the scatter
in the fatigue lives is quite large. The scatter may be attributed to microstructure,
manufacturing, or experimental inconsistencies, or a combination thereof. Empirical
modeling is more challenging for extreme life estimation because those events are
rare. The primary purpose herein is to propose an empirically based methodology
for estimating the cumulative distribution functions for fatigue life, given the applied
load. The methodology incorporates available fatigue life data for various stresses or
strains using a statistical transformation to merge all the life data so that distribution
estimation is more accurate than traditional approaches. Subsequently, the distribu-
tion for the transformed and merged data is converted using change-of-variables to
estimate the distribution for each applied load. To assess the validity of the proposed
methodology percentile bounds are estimated for the life data. The development of
the methodology and its subsequent validation is illustrated using three sets of fatigue
life data which are readily available in the open literature.

Keywords Cycle dependent parameters · Data fusion · Fatigue life

transformation · Weibull distribution

Nomenclature

α, β, γ Weibull cdf parameters

αs significance

D. Gary Harlow (B)

Mechanical Engineering and Mechanics, Lehigh University, 19 Memorial Drive West, Bethlehem,
PA 18015–3085, USA
e-mail: [email protected]

© Springer Nature Switzerland AG 2020 1

H. Pham (ed.), Reliability and Statistical Computing, Springer Series
in Reliability Engineering, https://doi.org/10.1007/978-3-030-43412-0_1
2 D. Gary Harlow

AD Anderson–Darling goodness of fit test

cdf cumulative distribution function
cv coefficient of variation
ε strain range
σ stress range
FLT Fatigue Life Transformation
F(•) cdf
KS Kolmogorov–Smirnov goodness of fit test
MLE maximum likelihood estimation
NA arbitrary normalization constant
Nf cycles to failure
n sample size
m number of different values of applied stress
p percentile
s sample standard deviation
sA arbitrary normalization constant
S–N stress–number of cycles
t time, cycles
W(α, β, γ) three parameter Weibull cdf
y1/2 median
ȳ sample average

1 Introduction

A difficulty with fatigue life data is the characterization of its variability, which can
be several orders of magnitude [1], especially for loading near operating conditions.
The variability is attributable to experimental error, as well as material microstruc-
ture or processing. Thus, estimation and prediction of fatigue life is challenging. A
key concern is the characterization of the cumulative distribution function (cdf) for
fatigue life, given an applied load, which may be either the stress range σ or the
strain range ε. The lower tail portion of the cdf which depicts high reliability is espe-
cially critical; however, that is where scatter is more pronounced, and sample sizes
are smaller. Empirically modeling fatigue life has been considered numerous times.
A simple internet search for statistical fatigue life modeling yields in excess of 10
million citations. A relatively recent work is [2], in which the authors incorporate sta-
tistical analysis with traditional stress-life, strain-life, or crack propagation models.
While empiricism is used, the thrust is to incorporate as much physically motivated
modeling as possible. More frequently, investigators attempt to fit a stress-life (S-N)
curve through the data, especially the medians for given σ or ε. A nice review
of such practices is [3]. Another example of statistical modeling of fatigue data is
contained in [4] in which over two chapters are devoted statistical methodologies.
Other examples of statistical stress-life analysis may be found in [5, 6]. While there
Fatigue Life Distribution Estimation 3

are many more papers, books, and conferences publications on statistical fatigue
modeling, these are representative. In spite of all these references, statistical fatigue
analysis is still an open area of investigation.
Herein, an empirically based approach for accurately estimating the fatigue life
cdf, given σ or ε is proposed. The methodology merges fatigue life data using a
statistical transformation for the estimation. The statistical technique increases the
sample size by merging fatigue data for more precise assessment. This is necessitated
because there is often large variability in S-N data, and the sample sizes are small.
Validation of the modeling is essential, especially for prediction of life outside of the
range of experimental observations. The validity of the methodology is evaluated by
considering percentile bounds estimated for the S-N data. The development of the
methodology and its subsequent validation is illustrated using three different fatigue
life datasets.
A fundamental issue in fatigue life estimation is the choice of an underlying
cdf. The cdf used in the ensuing analyses is a three-parameter Weibull cdf. A fairly
recent example of a traditional statistical S-N methodology using the three-parameter
Weibull cdf is [7] where fatigue of structural and rolling contact problems are con-
sidered. There seems to be a need for more experimental data to enhance modeling
in almost all fatigue analyses. This is addressed in [8] by using normalization for the
fatigue life data so that all the data are merged. The normalized data are then modeled
with a three-parameter Weibull cdf. Even though the intention in [8] is similar to the
emphasis herein, the methodology is somewhat different.

2 Fatigue Life Data

Fatigue life data are most often presented on an S-N plot which shows the fatigue
data for a given load. The load is typically stress or strain. Thus S-N can represent
stress-life or strain-life. An additional way in which the fatigue data are presented is
on a probability plot. Both of these representations will be used subsequently. Three
different sets of fatigue life data are considered for the proposed method.
The first set considered is one of the special cases taken from [9]. Fatigue testing
was conducted on 2024-T4 aluminum alloy specimens. The fatigue tests were per-
formed on rectangular specimens with dimensions of 110 mm long, 52 mm wide, and
1 mm thick, and with a center cut circular hole of radius 5 mm. Holes were cut using
standard procedures with a lathe, and burrs were removed by polishing techniques.
Testing was conducted in laboratory environment where temperatures of 295–297 K
(approximately 22–24 °C) and relative humidities of 50–56% were observed. Con-
stant amplitude tests were performed at a frequency of 30 Hz on a single machine
with a single operator in order to minimize experimental error. Because of the exten-
siveness of this data, they have been used in a variety of analyses; e.g., [10–12].
These data are summarized in Table 1. A total of 222 specimens were tested using
eight different values for σ. The specimens were tested to fracture. The sample
coefficients of variation (cv) are nearly the same, approximately 9%, for the larger
4 D. Gary Harlow

Table 1 Statistical summary of fatigue life data for 2024-T4 specimens [9]
Load, σ (MPa) Size (m) Average (x̄) Standard deviation (s) cv (%)
255 21 18,200 1760 9.6
235 30 28,700 2500 8.7
206 30 59,400 4230 7.1
177 30 146,000 12,600 8.6
157 30 264,000 22,600 8.6
137 30 519,000 96,200 18.5
127 30 1,710,000 1,090,000 63.8
123 21 4,530,000 2,660,000 58.7

values of σ. When σ is 137 MPa, the cv is about double, and for the two smaller
values of σ the scatter increases significantly. The fatigue life data are plotted on
an S-N graph in Fig. 1 in the traditional linear versus logarithm S-N format. As σ
is reduced, the increase in the scatter in life is apparent. Modeling the increasing
variability for decreasing σis the challenge for accurate fatigue life prediction.
The second set of fatigue data to be considered is data collected at room temper-
ature for ASTM A969 hot dipped galvanized sheet steel with a gauge thickness of
1.78 mm [13]. ASTM A969 is a cold-rolled, low carbon, extra deep drawing steel.
This steel is very ductile and soft, and it is age resistant. The automotive industry uses
it in applications where severe forming is required, e.g., inner door components, dash
panels, body side components, and floor pans with spare tire tubs. Fatigue tests for
the ASTM A969 specimens were conducted using a triangular waveform at 25 Hz.
The fatigue tests were terminated, i.e., designated as a failure, when the tensile load
dropped by 50% of the maximum load. A total of 69 specimens were tested to failure.
The data are summarized in Table 2. The cv given ε is more scattered than those
for the 2024-T4 data. The smallest value of ε has the largest scatter, but the second

Fig. 1 Fatigue life data for

260
2024-T4 specimens [9]
, MPa

240

220

200
stress range

180

160

140

120

10 4 10 5 10 6 10 7
number of cycles, N
Fatigue Life Distribution Estimation 5

Table 2 Statistical summary of fatigue life data for ASTM A969 specimens [13]
Load, ε (mm/mm) Size (m) Average (x̄) Standard deviation (s) cv (%)
0.0080 3 4,100 815 19.8
0.0060 12 14,000 3,800 27.2
0.0050 6 34,600 7,030 20.3
0.0040 12 56,500 8,860 15.7
0.0030 12 107,000 12,900 12.1
0.0024 6 199,000 29,800 15.0
0.0020 11 499,000 68,300 13.7
0.0018 7 1,030,000 391,000 37.8

largest ε is rather large as well. This behavior is seen graphically on the S-N plot in
Fig. 2. Consequently, these data are not as statistically well behaved as the 2024-T4
data. These data have been used to investigate other types of fatigue modeling [14,
15].
The third set to assess is 9Cr-1Mo steel which were collated from a round-robin test
program and were reported in [16]. This steel is creep strengthen, and it is frequently
used in thermal power plants to improve the energy efficiency of the power plant
by increasing operating temperatures and pressures. Specifically, 9Cr-1Mo is often
used for steam generator components of both fossil fired and nuclear power plants.
The material from which the data were generated was a single cast, rolled plate
with a nominal tensile strength of 623 MPa [16]. A total of 130 specimens were
tested to failure. The data are summarized in Table 3. The cv given ε is even more
scattered than the above datasets. In fact, there does not seem to be any discernible
pattern. The data are shown on Fig. 3. The reason for the unusual statistical behavior
would require more in depth analysis than is provided in [16]. Usually round-robin
testing has considerably more variability in results because testing conditions and

Fig. 2 Fatigue life data for 2e-2

ASTM A969 specimens [13]
, (mm/mm)

1e-2
7e-3
5e-3
strain range,

3e-3

2e-3

1e-3
10 3 10 4 10 5 10 6 10 7
cycles to failure, Nf
6 D. Gary Harlow

Table 3 Statistical summary of fatigue life data for 9Cr-1Mo specimens [16]
Load, ε (mm/mm) Size (m) Average (x̄) Standard deviation (s) cv (%)
0.021 5 472 88 18.8
0.020 32 542 189 34.9
0.019 8 572 345 60.2
0.012 34 1,260 357 28.4
0.011 6 1,290 475 36.9
0.006 37 4,820 2,000 41.6
0.005 8 12,700 7,700 60.8

Fig. 3 Fatigue life data for 3e-2

9Cr-1Mo specimens [16]
, (mm/mm)

2e-2

1e-2
strain range,

7e-3

5e-3

3e-3
102 103 104 105

cycles to failure, Nf

methodologies are not consistent. Nevertheless, the data will serve as an excellent
case for the proposed modeling approach.

3 Data Fusion for Fatigue Life Analysis

The following is a purely empirical method to improve fatigue life modeling. Because
fatigue data are usually limited in number for relatively few different loading con-
ditions, modeling is crude. A methodology that has been developed to account for
uncertainty for static properties [17] is adapted for fatigue life data. The basis of
the approach is a linear transformation of a collection of experimental observations
{y j : 1 ≤ j ≤ n} into another set of values {z j : 1 ≤ j ≤ n} so that both sets have
the same average and sample standard deviation. Let

z j = ay j + b. (1)
Fatigue Life Distribution Estimation 7

The choices of a and b in Eq. (1) are easily determined by simple algebra to be
the following:
sA sA
a= and b = N A − ȳ, (2)
sy sy

where ȳ is the average and sy is the sample standard deviation for {y j : 1 ≤ j ≤ n},
and N A and sA are arbitrary values chosen for normalization.
For fatigue data the life times are distributed over several orders of magnitude
that the procedure in Eqs. (1) and (2) is applied to the natural logarithm of the
life times. Let m be the number of different values of applied stress or strain, i.e.,
{σk : 1 ≤ k ≤ m} or {εk : 1 ≤ k ≤ m}. Given σk or εk the associated life
times are {Nk, j : 1 ≤ j ≤ n k } where nk is its sample size. Let

yk, j = ln(Nk, j ) (3)

be the transformed life times. Substituting Eq. (2) into Eq. (1) leads to the following:
sA
z k, j = (yk. j − ȳk ) + N A . (4)
s y,k

Thus, the averages and sample standard deviations of {yk, j : 1 ≤ j ≤ n k } and

{z k, j : 1 ≤ j ≤ n k } are equal to each other. The next step is to merge all the
transformed zk,j values from Eq. (4) for 1 ≤ j ≤ nk and 1 ≤ k ≤ m. The purpose
in using the merged values is to have a more extensive dataset for estimation of the
cdf. This is especially critical for estimating the extremes of the cdf more accurately.
Subsequently, an appropriate cdf F Z (z) is found that characterizes the merged data.
It is assumed that this cdf also characterizes the subsets {z k, j : 1 ≤ j ≤ n k } of the
merged set. With this assumption and the linear transformation in Eq. (4), the cdfs
for {yk, j : 1 ≤ j ≤ n k } F y,k (y) can be derived from F Z (z) as follows:

sA
Fy,k (y) = FZ ( (y − ȳk ) + N A ). (5)
s y,k

The approach is designated as the Fatigue Life Transformation (FLT). Recall that
the above methodology is applied to natural logarithms. In order to make observations
on the actual fatigue lives, the values must be changed back to actual cycles.

4 Flt Analysis for 2024-T4 Fatigue Life Data

To evaluate the effectiveness of the proposed FLT methodology, the fatigue life data
summarized in Table 1 and shown on Fig. 1 is considered. Recall that the FLT is
8 D. Gary Harlow

applied to the natural logarithm of the fatigue data; see Eq. (3). The arbitrarily cho-
sen values for N A and sA are 26 and 1, respectively. The rather large value for N A
was chosen to assure that zk,j in Eq. (4) is positive. The 222 FLT data are shown
on Fig. 4, where the axes are labeled to be easily read. Each set of data for a given
σk are transformed using FLT. These transformed data are well grouped so that it
is reasonable to merge them. Figure 5 shows the entire 222 FLT values merged into
a common sample space. The FLT merged data contain approximately 7–10 times
more data than those for each given σk . Thus, estimation for the cdf is necessar-
ily more accurate which results in a better characterization of its lower tail. Also,
notice that the cycles are transformed using the FLT procedure; they are not actual
cycles to failure, i.e., they are not equivalent to the data shown on Fig. 1. The solid
line is the maximum likelihood estimation (MLE) for a three–parameter Weibull cdf
W(α, β, γ), where α is the shape parameter, β is the scale parameter, and γ is the
location parameter. The form of W(α, β, γ) is

F(x) = 1 − exp{−[(x − γ )/β]α }, x ≥ γ . (6)

Fig. 4 FLT fatigue life data 0.999

for 2024-T4 specimens given 0.990
0.900
σ [9] FLT data
0.500
probability

= 255 MPa
= 235 MPa
0.100 = 206 MPa
0.050 = 177 MPa
= 157 MPa
= 137 MPa
0.010 = 127 MPa
0.005 = 123 MPa

0.001
1010 1011 1012 1013
transformed cycles

Fig. 5 Merged FLT fatigue 0.999

life data for 2024-T4 0.990 FLT - ln(cycles)
0.900 W( , , ) MLE
specimens [9]; Weibull MLE
0.500
probability

0.100
0.050 W( , , ) MLE
= 2.894
= 2.986
0.010 = 23.327
0.005 KS = 0.043
AD = 0.292

0.001
1010 1011 1012 1013
transformed cycles
Fatigue Life Distribution Estimation 9

Graphically, the fit is excellent. The Kolmogorov-Smirnov (KS) and Anderson-

Darling (AD) goodness of fit test statistics are 0.043 and 0.292, respectively. Both of
which indicate that the MLE is acceptable for any significance level αs less than 0.3.
Consequently, the MLE W(α, β, γ) cdf is an excellent representation of the merged
FLT data. The MLE estimated parameters are α̂ = 2.894; β̂ = 2.986; and γ̂ =
23.327.
The three–parameter Weibull cdf W(α, β, γ) shown in Eq. 6 was selected for
consideration because it has become a very popular cdf to represent fatigue data
since its namesake used it for that purpose [18]. Two popular resources for the
Weibull cdf, which advocate its use and contain examples of its applications, are [19,
20]. The primary reason for its choice, however, is because there is a minimum value
represented by γ. Typically, any collection of fatigue life data is spread over two or
three orders of magnitude. Consequently, a nonzero minimum value is required to
appropriately represent the fatigue data.
Now, it is assumed that the MLE estimated parameters for the FLT merged data
are acceptable for each of its subsets {z k, j : 1 ≤ j ≤ n k }. The cdf F y,k (y) for each
given σk can be determined using Eq. (5) as follows:
sA
Fy,k (y) = 1 − exp{−[( (y − ȳk ) + N A − γ̂ )/β̂]α̂ }. (7)
s y,k

Recall that the arbitrary constants N A and sA are 26 and 1, respectively. Equa-
tion (7) can be rewritten to put it into the standard Weibull cdf form W(α, β,
γ);
s y,k s y,k α̂
Fy,k (y) = 1 − exp{−[(y − [ ȳk + (γ̂ − N A )( )])/β̂( )] }, (8)
sA sA

where the shape parameter α̂ is the same for each individual cdf, but the scale
parameter β̂k and location parameter γ̂k are
s y,k s y,k
β̂k = β̂( ); and γ̂k = ȳk + (γ̂ − N A )( ), (9)
sA sA

which are explicitly dependent on the sample parameters for the fatigue life data
given σk and the arbitrary constants N A and sA . Recall that the FLT is for ln(N f );
consequently, the range of values is considerably smaller than that for N f . Table 4
contains the FLT cdf W(α, βk , γk ) parameters for each given σk . Figure 6 shows
the fatigue life data plotted on two parameter Weibull probability paper with the
corresponding FLT cdfs W(α, βk , γk ). Graphically, these cdfs appear to fit the data
well. Indeed, the KS goodness of fit test indicates that all these cdfs are acceptable
for any αs less than 0.25. The AD test, which focuses on the quality of the fit in
the tails, is more discriminating. The cdfs when σk equals 127, 177, 206, 235, or
255 MPa are acceptable for any αs less than 0.25. When σk equals 123 or 157 MPa,
however, the cdfs are acceptable for any αs less than 0.05. When σk equals 137, the
AD test implies that the cdf is not acceptable. Although it is not obvious on Fig. 6
10 D. Gary Harlow

Table 4 FLT Weibull

Load, σ (MPa) α̂ β̂k γ̂k
parameters for natural
logarithm of fatigue life data 255 2.894 0.280 9.556
for 2024-T4 specimens [9] 235 2.894 0.258 10.032
206 2.894 0.214 10.799
177 2.894 0.261 11.657
157 2.894 0.247 12.259
137 2.894 0.491 12.705
127 2.894 1.678 12.690
123 2.894 1.972 13.375

Fig. 6 FLT Weibull cdfs for 0.999

fatigue life data for 2024-T4 0.900
Probability of Failure

specimens [9] 0.750

0.500
0.250 = 123 MPa
= 127 MPa
0.100 = 137 MPa
0.050 = 157 MPa
= 177 MPa
= 206 MPa
0.010
= 235 MPa
0.005 = 255 MPa
FLT - W( , , )
0.001
1e+4 1e+5 1e+6 1e+7 1e+8
cycles to failure, Nf

because the cycles scale is so large, both the upper and lower tail of the FLT cdf are
sufficiently different from the life data. Even so, the overall deduction is that the FLT
transformation is acceptable for these fatigue life data. Again, the KS test supports
that conclusion, and there is only one value for σk , 137 MPa, for which the AD test
indicates otherwise.
Another way to assess the quality of the proposed FLT methodology is to consider
percentiles p of the estimated cdfs. The percentiles are given by

y p = γ̂k + β̂k [− ln(1 − p)]1/α̂ , (10)

which are computed from Eqs. (8) and (9). One of the most common percentiles that
is considered is the median y1/2 , i.e., p is 0.5. Figure 7 is an S-N graph, identical
to Fig. 1, where the solid line is the FLT estimated median, and the dashed lines
are the FLT estimated 99% percentile bounds. The 99% bounds are very tight while
encapsulating the entire set of data for each σk . They also follow the trend in the
S–N data in that they are very narrow when the life data have very little variability,
but they are broader when the life data have larger variability. This lends credence
to the FLT approach for the 2024-T4 fatigue data.
Fatigue Life Distribution Estimation 11

Fig. 7 Fatigue life data for

260
2024-T4 specimens [9] with
fatigue life data
FLT percentile bounds 240

, MPa
FLT estimated median
FLT estimated
220 99% percentile bounds
200

stress range
180

160

140

120

104 105 106 107

number of cycles, N

5 FLT Analysis for ASTM A969 Fatigue Life Data

The second applications of the FLT method is for the ASTM A969 fatigue data.
Again, the arbitrarily chosen values for N A and sA are 26 and 1, respectively. Figure 8
shows the entire 69 FLT values merged into a common sample space. The solid line
is the MLE W(α, β, γ), Eq. (6). The KS and AD goodness of fit test statistics are
0.047 and 0.524, respectively. The MLE is acceptable according to the KS test for
any significance level αs less than 0.3. For the AD test, however, it is acceptable only
for αs less than 0.2 because there is some deviation between the data and the MLE
in the lower tail. Even so, the FLT data are well represented by the MLE W(α, β, γ)
cdf. The MLE estimated parameters are α̂ = 3.057; β̂ = 2.964; and γ̂ = 23.289.
Using Eqs. (7)–(9), Fig. 9 has the ASTM A969 data with the FLT cdfs W(α, βk ,
γk ). Graphically, the FLT cdfs appear to characterize the data well. In fact, the KS
test indicates that all these cdfs are acceptable for any αs less than 0.3. The AD test,
however, implies that the FLT cdfs are marginal, at best. Clearly, the FLT cdfs are

Fig. 8 Merged FLT fatigue 0.999

life data for ASTM A969 0.990
FLT - ln(cycles)
specimens [13]; Weibull 0.900 W( , , ) MLE
MLE
0.500
probability

0.250

0.100 W( , , ) MLE
0.050 = 3.057
= 2.964
0.025
= 23.289
0.010 KS = 0.047
AD = 0.524
0.005
0.003
10 11 12 13
10 10 10 10
transformed cycles
12 D. Gary Harlow

Fig. 9 FLT Weibull cdfs for 0.99 (mm/mm)

fatigue life data for ASTM 0.95 0.0018
A969 specimens [13] 0.90 0.0020
0.0024
0.80 0.0030
0.0040

probability
0.60 0.0050
0.50 0.0060
0.0080
0.40 FLT cdfs
0.30
0.20

0.10
0.07
0.05
103 104 105 106 5x106
cycles to failure, Nf

Fig. 10 Fatigue life data for 2e-2

total strain range
ASTM A969 specimens [13] FLT median
, (mm/mm)

with FLT percentile bounds FLT 99% bounds

1e-2

7e-3
5e-3
strain range,

3e-3

2e-3

1e-3
103 104 105 106 107
cycles to failure, Nf

not as accurate in the tails. No doubt, larger samples for each εk would help with
characterization of the extremes. Using Eq. (10), Fig. 10 is an S-N graph, identical
to Fig. 2, with the addition of the FLT estimated median, and the FLT estimated 99%
percentile bounds. The 99% bounds are very tight, and the all the data are within the
bounds for each εk . They are somewhat jagged because they follow the pattern of
the S-N data. The analysis is not as crisp as that for the 2024-T4 data; however, there
seems to be merit in using the FLT approach for the ASTM A969 fatigue data.

6 FLT Analysis for 9Cr-1Mo Fatigue Life Data

The third application of the FLT method for the 9Cr-1Mo fatigue data does not
perform very well. Figure 3 shows unusually large scatter for the higher values of
εk which is a good test for the FLT methodology. The arbitrary scaling factors
are the same; N A and sA are 26 and 1, respectively. Figure 11 shows the 130 FLT
merged values with the MLE W(α, β, γ). The KS and AD goodness of fit test statistics
Fatigue Life Distribution Estimation 13

Fig. 11 Merged FLT fatigue 0.999

life data for 9Cr-1Mo 0.990
FLT - ln(cycles)
specimens [16]; Weibull 0.900 W( , , ) MLE
MLE
0.500

probability
0.250

0.100 W( , , ) MLE
0.050 = 2.497
= 2.475
0.025
= 23.769
0.010 KS = 0.028
AD = 0.300
0.005
0.003
10 11 12 13
10 10 10 10
transformed cycles

are 0.028 and 0.300, respectively. The MLE is acceptable according to the KS and
AD tests for any significance level αs less than 0.3. The merged data are very well
characterized by this cdf. The corresponding MLE estimated parameters are α̂ =
2.497; β̂ = 2.475; and γ̂ = 23.769.
The FLT cdfs W(α, βk , γk ), Eqs. (7)–(9), for 9Cr-1Mo are shown on Fig. 12.
Graphically, the FLT cdfs appear to be acceptable, at least for the cases with more
data. Clearly, when εk is 0.019, the fit is borderline. The KS test indicates that all
these cdfs are acceptable for any αs less than 0.3. On the other hand, the AD test
indicates that none of the FLT cdfs are acceptable. The compressed graphical scale
for the cycles to failure masks the poor fit of the FLT cdfs to the tails of the data.
Figure 13 is the S-N graph, Fig. 3, with the FLT estimated median, and the FLT
estimated 99% percentile bounds. Because the FLT cdfs are not very representative
of the fatigue data, the 99% bounds are erratic. All the data are within the bounds
for each εk . That is somewhat positive. They are quite jagged because they follow
the S-N data pattern. One of the difficulties in modeling these data are that there are
three values for εk that have several replicates, but the other values have only a few.
Also the scatter in the data when εk is 0.020 appears to be larger than expected.

Fig. 12 FLT Weibull cdfs 0.99 (mm/mm)

for fatigue life data for 0.95
0.90 0.005
0.006
9Cr-1Mo specimens [16] 0.80 0.011
0.60 0.012
probability

0.50 0.019
0.40 0.020
0.30 0.021
FLT cdfs
0.20
0.10
0.07
0.05

0.01
102 2x10 2 5x10 2 103 2x10 3 5x10 3 104 2x10 4 5x10 4
cycles to failure, Nf
14 D. Gary Harlow

Fig. 13 Fatigue life data for 3e-2

total strain range
9Cr-1Mo specimens [16] FLT median

, (mm/mm)
with FLT percentile bounds 2e-2 FLT 99% bounds

1e-2

strain range,
7e-3

5e-3

3e-3
102 103 104 105
cycles to failure, Nf

Since there is no explanation for this in [16], the reason would purely speculation.
The analysis is marginal for the 9Cr-1Mo data. There may be merit in using the FLT
approach for some insight, but conclusions would need to be made very cautiously.

7 Observations and Conclusions

Three sets of fatigue life data were considered for the proposed FLT method. These
datasets were selected because of they have replicate data for several applied stress
or strain ranges, σk or εk , respectively. The primary reason for the proposed FLT
method is to accurately model the statistical nature of fatigue life data. Specifically,
the estimation of underlying cdfs is crucial. It is well known that fatigue life data have
rather large amounts of variability particularly for applied loads similar to typical
operating conditions. Modeling these data is very challenging. Associated with this
is that many fatigue life data sets have relatively few choices for σk or εk , and
each choice has limited observations.
The FLT introduced in this paper attempts to help improve fatigue life modeling
by using a statistically based transformation to merge data, thereby increasing the
effective sample size. The FLT approach transforms the fatigue life data for each
given σk or εk so that the averages and standard deviations are the same. Subse-
quently, the data are merged, and a suitable cdf is statistically estimated for the entire
collection. The cdf for each given σk or εk is obtained by standard change-of-
variables methods using the cdf that characterizes the entire transformed and merged
data.
Using the 2024-T4 fatigue life in [9], the FLT is very promising partly because
there are eight different values for σk and a total of 222 data being considered. A
three–parameter Weibull cdf W(α, β, γ) is an excellent representation of the merged
FLT data, and a W(α, β, γ) is also appropriate characterization for the underlying cdfs
given σk . Consequently, there is assurance that the lower tail behavior is adequately
Fatigue Life Distribution Estimation 15

modeled because of the methodology. Additionally, the validation for the approach is
the computed FLT percentiles for the S-N data. The computed FLT 99% percentiles
fully encompass all the S-N data, but more importantly, the bounds are quite tight.
The conclusion from this analysis is that the FLT methodology is warranted.
To corroborate this conclusion, two other sets of fatigue life data were considered;
ASTM A969 [13] and 9Cr-1Mo [16]. In both cases, the merged FLT data are well
characterized by a W(α, β, γ). The KS test indicates that the transformed W(α, β, γ) is
also suitable for the underlying cdfs given εk , but the AD test is more discriminating.
The tail behavior of the underlying cdfs given εk are marginally acceptable at best,
if at all. The 99% percentiles for the ASTM A969 S-N data are quite good. They
are encompass the data, and they are tight. For the 9Cr-1Mo S-N data, however, the
percentile lines are not very regular. They do encompass the data, but the data have
so much variability that little is gained by the analysis.
Based on these three examples, the FLT approach should be employed for fatigue
life data analysis when an empirical method is desired. The FLT method excellent for
one of the cases, 2024-T4, acceptable for another case, ASTM A969, and marginal
for the other case, 9Cr-1Mo. As with all empirical analyses, caution must be exer-
cised when it is implemented. Limited applied loads with limited replicate data for
each load hinders accurate modeling for any method including the FLT. As with
all empirical methods, the more data there is, the better the accuracy will be. The
example which was the worst, 9Cr-1Mo, seems to be poor because there is overly
large scatter in the data for the higher loading conditions coupled with applied loads
with only a few replicates. Again, the FLT methodology should be implemented with
care.
Many sets of experimental fatigue life data contain censoring. This will be inves-
tigated in the future. In this case the cdf estimation is more advanced, especially for
a three-parameter cdf. In principle the FLT methodology should be similar except
for the adjustment for censoring. All things considered, the proposed FLT approach
has sufficient promise that further investigation and analysis is certainly warranted.
The overarching observation is that the FLT approach is useful if the fatigue data are
reasonably well behaved.

References

1. Harlow DG, Wei RP, Sakai T, Oguma N (2006) Crack growth based probability modeling of
S-N response for high strength steel. Int J Fatigue 28:1479–1485
2. Castillo E, Fernández-Canteli A (2009) A unified statistical methodology for modeling fatigue
damage. Springer Science and Business Media B.V., Berlin
3. Schneider CRA, Maddox SJ (2003) Best practice guide on statistical analysis of fatigue data.
International Institute of Welding, IIW-XIII-WG1-114-03, United Kingdom
4. Collins JA (1993) Failure of materials in mechanical design: analysis, prediction, prevention,
2nd edn. Wiley-Interscience Publications, New York
5. Little RE, Ekvall JC (eds) (1981) Statistical analysis of fatigue data. ASTM STP744, American
Society for Testing and Materials, Philadelphia
16 D. Gary Harlow

6. Zheng X-L, Jiang B, Lü H (1995) Determination of probability distribution of fatigue strength

and expressions of P–S–N curves. Eng Fracture Mech 50:483–491
7. Shimizu S (2005) P–S–N/P–F–L curve approach using three-parameter Weibull distribution for
life and fatigue analysis of structural and rolling contact components. Tribol Trans 48:576–582
8. Fernández–Canteli A, Blasón S, Correia JAFO, de Jesus AMP (2014) A probabilistic
interpretation of the miner number for fatigue life prediction. FratturaedIntegritàStrutturale
30:327–339
9. Shimokawa T, Hamaguchi Y (1985) Relationship between fatigue life distribution, notch con-
figuration, and S-N curve of a 2024–T4 Aluminum Alloy. Trans ASME J Eng Mater Technol
107:214–220
10. Harlow DG (2005) Probability versus statistical modeling: examples from fatigue life
prediction. Int J Reliab Qual Saf Eng 12:1–16
11. Harlow DG (2011) Generalized probability distributions for accelerated life modeling. SAE
Int J Mater Manuf 4:980–991
12. Sakai T, Nakajima M, Tokaji K, Hasegawa N (1997) Statistical distribution patterns in
mechanical and fatigue properties of metallic materials. Mater Sci Res Int 3:63–74
13. IF_DDQ_HDG70G_Strain_Life_Fatigue.xls www.autosteel.org (2004). Automotive Applica-
tions Council, Steel Market Development Institute, Auto/Steel–Partnership Fatigue Data
14. Harlow DG (2014) Low cycle fatigue: probability and statistical modeling of fatigue life. In:
Proceedings of the ASME 2014 pressure vessels and piping conference, PVP2014–28114,
Anaheim, CA, 20–24 July 2014, V06BT06A045. ISBN 978–0–7918–4604–9
15. Harlow DG (2017) Confidence bounds for fatigue distribution functions. In: Proceedings of the
ASME 2017 pressure vessels and piping conference, PVP2017–65416, Waikoloa, HI, 16–20
July 2017. V01AT01A011. ISBN 978–0–7918–5790–8
16. Thomas GB, Varma RK (1993) Evaluation of low cycle fatigue test data in the BCR/VAMAS
intercomparison programme. Commission of the European Communities, BCR Information
Applied Metrology, Brussels, Luxembourg
17. Harlow DG (2007) Probabilistic property prediction. Eng Fract Mech 74:2943–2951
18. Weibull EHW (1951) A statistical distribution function of wide applicability. ASME J Appl
Mech 18:293–297
19. Abernethy RB (2006) In: Abernethy RB (ed) The new Weibull handbook, 5th edn. 536 Oyster
Road, North Palm Beach, Florida, pp 33408–34328
20. Rinne H (2009) The Weibull distribution. CRC Press, Chapman-Hall, Taylor and Francis Group,
Boca Raton

D. Gary Harlow is a Professor in Mechanical Engineering and Mechanics at Lehigh University in

Bethlehem, PA. Since 2008 he has been the department chair. He received his BA in Mathemat-
ics (1973) from Western Kentucky University and his MS (1976) and Ph.D. in Applied Proba-
bility and Stochastic Processes (1977) from Cornell University. His research has primarily cen-
tered on scientifically and mechanistically based probability modeling and statistical analyses.
Specifically, his research has included modeling of failure processes (creep, creep-fatigue, very
high cycle fatigue, corrosion pitting, and corrosion fatigue) in materials (aluminum alloys, steels,
and composites); stochastic fracture mechanics; mechanical and system reliability; applications of
stochastic processes; and methods in applied statistics for materials characterization. His research
has been sponsored throughout his career by the NSF, FAA, DoD, and DARPA. He is a fellow of
ASME.
Exploring the Variety of Random
Documents with Different Content
the terrible wear and tear of mind which, from the coast to Uvira,
had never been absent. I felt the proud consciousness of having
done my best, under conditions from beginning to end the worst and
the most unpromising, and that whatever future evils Fate might
have in store for me, that it could not rob me of the meed won by
the hardships and sufferings of the past.
Several Arab merchants were preparing to return coastwards for
the “Mausim” (monsoon), or Indian trading-season, which, at
Zanzibar, includes the months of December, January, and February,
and they were not unwilling to avail themselves of my escort. But
several reasons detained me at Kazeh. Some time was required to
make preparations for the long down march. I had not given up the
project of returning to the seaboard viâ Kilwa. Moreover, it was
judged advisable to collect from the Arabs details concerning the
interesting countries lying to the north and south of the line
traversed by the Expedition. As has been mentioned in Chap. XI., the
merchants had detailed to me, during my first halt at Kazeh, their
discovery of a large Bahr—a sea or lake—lying fifteen or sixteen
marches to the north; and from their descriptions and bearings, my
companion had laid down the water in a hand-map forwarded to the
Royal Geographical Society. All agreed in claiming for it superiority of
size over the Tanganyika Lake. I saw at once that the existence of
this hitherto unknown basin would explain many discrepancies
promulgated by speculative geographers, more especially the
notable and deceptive differences of distances, caused by the
confusion of the two waters.[8] Remained only to ascertain if the
Arabs had not, with the usual Oriental hyperbole, exaggerated the
dimensions of the Northern Lake.
[8] Mr. Erhardt, for instance, “Memoir on the Chart of East and Central
Africa, compiled by J. Erhardt and J. Rebmann, London, 1856,” announces
the “existence of a Great Lake, called in the south Niandsha (Nyassa), in
the north Ukerewe, and on the coast Niasa and Bahari ya Uniamesi,” makes
the distance through Dschaga (Chhaga) and the Masai plains only fifty-nine
marches.
 My companion, who had recovered strength from the repose and
the comparative comfort of our head-quarters, appeared a fit person
to be detached upon this duty; moreover, his presence at Kazeh was
by no means desirable. To associate at the same time with Arabs
and Anglo-Indians, who are ready to take offence when it is least
intended, who expect servility as their due, and whose morgue of
colour induces them to treat all skins a shade darker than their own
as “niggers,” is even more difficult than to avoid a rupture when
placed between two friends who have quarrelled with each other.
Moreover, in this case, the difficulty was exaggerated by the Anglo-
Indian’s complete ignorance of Eastern manners and customs, and
of any Oriental language beyond, at least, a few words of the
debased Anglo-Indian jargon.
I have dwelt upon this subject because my companion has
thought proper to represent (in Blackwood, Oct. 1859) that I was
“most unfortunately quite done up, but most graciously consented to
wait with the Arabs and recruit health.” This is far from being the
fact. I had other and more important matter to work out. Writing
from the spot (Unyanyembe, 2nd July 1858, and published in the
Proceedings of the Royal Geographical Society, 24th Jan. 1859) my
companion represents the case somewhat differently. “To diminish
the disappointment, caused by the short-coming of our cloth, in not
seeing the whole of the Sea Ujiji, I have proposed to take a flying
trip to the unknown lake, while Captain Burton prepares for our
return homewards.”
On the 30th June the subject was brought forward in the presence
of Said bin Salim and the Baloch. The former happily lodged at
Kazeh, felt loath to tear himself from the massive arms of his
charmer Halimah. He finessed as usual, giving an evasive answer,
viz. that he could not decide till the last day, and he declined to
influence the escort, who afterwards declared that he had done all in
his power to deter them from the journey. In vain my companion
threatened him with forfeiture of his reward after he returned to
Zanzibar; in vain my companion told him that it was forfeited.[9] He
held firm, and I was not over-anxious in influencing him, well
knowing that though the Baloch, a stolid race, might prove
manageable, the brain of the Machiavellian Arab, whose egregious
selfishness never hesitated at any measure calculated to ensure its
gratification, was of a somewhat too heavy metal for the article
opposed to it. That Said bin Salim attempted to thwart the project I
have no doubt. The Kirangozi, and the fifteen porters hired from his
village with the tempting offer of five cloths per man, showed an
amount of fear and shirking hardly justified by the real risks of
treading so well known a tract. The Jemadar and his men at first
positively refused their escort, but the meaning word “Bakhshish”
slipping in reassured me. After informing them that in case of
recusancy their rations should be stopped, I inquired the amount of
largesse expected. The ten efficient men composing the guard
demanded fifteen cloths a piece, besides one porter each to carry
their matchlocks and pervanents. The number of the porters was
reduced, the cloth was procured from an Arab merchant, Sayf bin
Said el Wardi, at an expense of one hundred dollars, made payable
by draught upon Ladha Damha of Zanzibar: at the same time, the
Baloch were warned that they must option between this and the
reward conditionally promised to them after return.[10] Their bad
example was followed by the old and faithful servant “Bombay,” who
required instant dismissal unless he also received cloth before the
journey: he was too useful to my companion as interpreter and
steward to be lightly parted with. But the granting his claim led to a
similar strike and menace on the part of the bull-headed slave
Mabruki, who, being merely a “headache” to me, at once “got the
sack” till he promised, if pardoned, to shake off his fear, and not to
be naughty in future. By dint of severe exertion my companion was
enabled to leave Kazeh on the 10th July.
[9] I transcribe the following words from my companion’s paper
(Blackwood, October 1859): “I urged that it was as much his (Said bin
Salim’s) duty as mine to go there; and said, unless he changed his present
resolution, I should certainly recommend the Government not to pay the
gratuity which the consul had promised him on condition that he worked
entirely to our satisfaction, in assisting the Expedition to carry out the
Government’s plans.”
 [10] So my report printed in the Proceedings Roy. Geog. Soc. loco cit.
“Our asses, thirty in number, all died, our porters ran away, our goods were
left behind; our black escort became so unmanageable as to require
dismissal; the weakness of our party invited attacks, and our wretched
Baloch deserted us in the jungle, and throughout have occasioned an
infinity of trouble.”

I proceed to recount the most important portion of the information

—for ampler details the reader is referred to the Journal of the Royal
Geographical Society—collected during my halt at Kazeh from
various sources, Arab and African, especially from Snay bin Amir,
concerning—

THE NORTHERN KINGDOMS: KARAGWAH, UGANDA, AND UNYORO.

The extensive and hitherto unknown countries described in this

chapter, being compact despotisms, resembling those of Ashanti and
Dahomey more than the semi-monarchies of Unyamwezi and Urundi,
or the barbarous republics of Uvinza and Ujiji, are designated the
Northern Kingdoms. It is regrettable that oral information, and not
the results of actual investigation, are offered to the reader
concerning regions so interesting as the Southern Tanganyika, the
Northern Kingdoms, and the provinces south of Unyanyembe. But
absolute obstacles having interfered, it was judged advisable to use
the labours of others rather than to omit all notice of a subject which
has the importance of novelty, because it lacked the advantages of a
regular exploration.
Informants agree in representing the northern races as superior in
civilisation and social constitution to the other tribes of Eastern and
Central Africa. Like the subjects of the Kazembe, they have built
extensive and regular settlements, and they reverence even to
worship a single despot, who rules with a rigour which in Europe
would be called barbarity. Having thrown off the rude equality of
their neighbours, they recognise ranks in society; there is order
amongst men, and some idea of honour in women; they add to
commerce credit, without which commerce can hardly exist; and
they hospitably entertain strangers and guests. These accounts are
confirmed by the specimens of male and female slaves from
Karagwah and Uganda seen at Unyanyembe: between them and the
southern races there is a marked physical difference. Their heads are
of a superior cast: the regions where the reflective faculties and the
moral sentiments, especially benevolence, are placed, rise high; the
nose is more of the Caucasian type; the immoderate masticating
apparatus which gives to the negro and the lower negroid his
peculiar aspect of animality, is greatly modified, and the expression
of the countenance is soft, kindly, and not deficient in intelligence.
From Unyanyembe to Kibuga, the capital of Uganda, are fifty-three
stages, which are distributed into four crucial stations of Usui,
Karagwah, dependent Unyoro, and Uganda. A few remarks
concerning each of these divisions may not be unacceptable.
Between Unyanyembe and Usui are sixteen long, or nineteen
short, stages. Though the road is for the most part rough and hilly,
the marches can scarcely be reduced below ten statute, or six
rectilinear geo. miles per diem; in fact, the geographer’s danger
when making these estimates is, that of falling, through fear of
exaggeration, into the opposite and equally incorrect extreme. The
general direction of the line leading from Kazeh, in Unyanyembe, to
Karagwah, pointed out by Snay bin Amir, bore 345° (corrected
332°); the length of the nineteen marches would be about 115 geo.
miles. The southern frontier of Usui may, therefore, be safely placed
in S. lat. 3° 10′.
The route from Kazeh to Usui falls at once westward of the line
leading to the Nyanza Lake; it diverges, however, but little at first, as
they both traverse the small districts of Ulikampuri, Unyambewa, and
Ukuni. Usonga, crossed in five short marches, is the first
considerable district north of Unyanyembe. Thence the road enters
the province of Utumbara, which is flanked on the east by Usambiro,
and on the west by Uyungu, governed by the Muhinda Sultan,
Kanze. Utumbara, as has been mentioned, was lately plundered, and
Ruhembe, its chief, was slain, by the predatory Watuta. In Utumbara
and Usambiro the people are chiefly the Wafyoma, a tribe of
Wanyamwezi: they are a commercial race, like the Wajiji—trafficking
in hoes and ivory; and their present Sultan, Mutawazi, has often
been visited by the Arabs. Uyofu, governed by Mnyamurunda, is the
northern boundary of Unyamwezi, after which the route enters the ill
famed territory of Usui.
Usui is traversed in seven marches, making a sum of twenty-six
from Kazeh. According to the former computation, a total march of
about 156 geo. miles would place the southern frontier of Karagwah
in S. lat. 2° 40′. The road in several parts discloses a view of the
Nyanza Lake. Usui is described as a kind of neutral ground between
the rolling plateau of Unyamwezi and the highlands of Karagwah: it
is broken by ridges in two places—Nyakasene the fourth, and
Ruhembe the seventh stage, where mention is also made of a small
stream. From this part of the country a wild nutmeg is brought to
Kazeh by caravans: the Arabs declare that it grows upon the well-
wooded hills, and the only specimen shown was heavy and well
flavoured, presenting a marked contrast to the poor produce of
Zanzibar island.
The Wasúí, according to the Arabs, are not Wanyamwezi. They are
considered dangerous, and they have frequently cut off the route to
caravans from Karagwah. Their principal sultan, a Muhinda named
Suwarora, demands exorbitant blackmail, and is described as
troublesome and overbearing: his bad example has been imitated by
his minor chiefs.
The kingdom of Karagwah, which is limited on the north by the
Kitangure or Kitangule River, a great western influent of the Nyanza
Lake, occupies twelve days in traversing. The usual estimate would
thus give it a depth of 72, and place the northern limit about 228
rectilinear geo. miles from Kazeh, or in S. lat. 1° 40′. But the
Kitangure River, according to the Arabs, falls into the Nyanza
diagonally from south-west to north-east. Its embouchure will,
therefore, not be distant from the equator. The line of road is thus
described: After ascending the hills of Ruhembe the route, deflecting
eastward, pursues for three days the lacustrine plain of the Nyanza.
At Tenga, the fourth station, the first gradient of the Karagwah
mountains is crossed, probably at low levels, where the spurs fall
towards the lake. Kafuro is a large district where merchants halt to
trade, in the vicinity of Weranhánjá, the royal settlement, which
commands a distant view of the Nyanza. Nyakahanga, the eighth
stage, is a gradient similar to that of Tenga; and Magugi, the tenth
station, conducts the traveller to the northernmost ridge of
Karagwah. The mountains are described as abrupt and difficult, but
not impracticable for laden asses: they are compared by the Arabs to
the Rubeho chain of Usagara. This would raise them about 4000 feet
above the mean level of the Unyamwezi plateau and the Nyanza
water, and about 8000 feet above this sea. Their surface, according
to the Arabs, is alternately earth and stone, the former covered with
plantains and huge timber-trees, the latter bare, probably by reason
of their altitude. There are no plains, bush, or jungle, but the deep
ravines and the valleys intersecting the various ridges drain the
surface of the hills, and are the sites of luxuriant cultivation. The
people of Karagwah, averse to the labour of felling the patriarchs of
the forest, burn “bois de vache,” like the natives of Usukuma. North
of Magugi, at Katanda, a broad flat extends eastwards: the path
thence descends the northern counterslope, and falls into the alluvial
plain of the Kitungure River.
Karagwah is thus a mass of highlands, bounded on the north by
dependent Unyoro, on the south by Usui, eastward by the tribes of
Wahayya and Wapororo, upon the lacustrine plain of the Nyanza; on
the south-west it inosculates with Urundi, which has been described
as extending from the north-eastern extremity of the Tanganyika
Lake. Its equatorial position and its altitude enable it to represent
the Central African prolongation of the Lunar Mountains. Ptolemy
describes this range, which he supposes to send forth the White
Nile, as stretching across the continent for a distance of 10° of
longitude. For many years this traditional feature has somewhat
fallen into discredit: some geographers have changed the direction
of the line, which, like the Himalayas, forms the base of the South
African triangle from east and west to north and south, thus
converting it into a formation akin to the ghauts or lateral ranges of
the Indian peninsula; whilst others have not hesitated to cast ridicule
upon the mythus. From the explorations of the “Mombas Mission” in
Usumbara, Chhaga, and Kitui, and from the accounts of Arab visitors
to the lands of Umasai and the kingdom of Karagwah, it appears
that from the fifth parallel of S. lat. to the equator, an elevated mass
of granite and sandstone formation crosses from the shores of the
Indian Ocean to the centre of Tropical Africa. The vast limestone
band which extends from the banks of the Burramputra to those of
the Tagus appears to be prolonged as far south as the Eastern Horn,
and near the equator to give place to sandstone formations. The line
is not, however, as might be expected from analogy with the
Himalayan, a continuous unbroken chain; it consists of insulated
mountains, apparently volcanic, rising from elevated plains, and
sometimes connected by barren and broken ridges. The south-
eastern threshold of the Lunar cordillera is the highland region of
Usumbara, which may attain the height of 3000 or 4000 feet above
sea-level. It leads by a succession of mountain and valley to Chhaga,
whose apex is the “Æthiopian Olympus,” Kilima-Ngao. From this
corner-pillar the line trends westward, and the route to Burkene
passes along the base of the principal elevations, Doengo Engai and
Endia Siriani. Beyond Burkene lies the Nyanza Lake, in a huge gap
which, breaking the continuity of the line, drains the regions
westward of Kilima-Ngao, whilst those to the eastward, the Pangani
and other similar streams, discharge their waters to the south-east
into the Indian Ocean. The kingdom of Karagwah prolongs the line
to Urundi, upon the Tanganyika Lake, where the south-western
spurs of the Lunar Mountains form a high continuous belt. Mr.
Petherick, of Khartum, travelling twenty-five marches, each of
twenty miles (?), in a south-south-western and due-southerly
direction from the Bahr el Ghazal, found a granitic ridge rising, he
supposes 2000 to 2500 feet above the plain, near the equator, and
lying nearly upon the same parallel of latitude, and in about 27° E.
long. Beyond that point the land is still unexplored. Thence the
mountains may sink into the great Depression of Central Africa, or,
deflected northwards of the kingdom of Uropua, they may inosculate
with the ridge which, separating the northern negroid races of
Islamised Africa from their negro brethren to the south, is popularly
known, according to Denham and Clapperton, as el-Gibel Gumhr,—
Jebel Kamar,—or Mons Lunæ.
 The high woody hills of Karagwah attract a quantity of rain. The
long and copious wet monsoon divides the year into two seasons—a
winter of seven or eight, and a summer of four or five months. The
Vuli, or lesser rains, commence, as at Zanzibar, with the Nayruz
(29th of August); and they continue with little intermission till the
burst of the Masika, which lasts in Karagwah from October to May or
June. The winds, as in Unyamwezi, are the Kaskazi, or north and
north-east gales, which shift during the heavier falls of rain to the
Kosi, the west and south-west. Storms of thunder and lightning are
frequent, and the Arabs compare the down-pour rather to that of
Zanzibar island than to the scanty showers of Unyamwezi. The
sowing season at Karagwah, as at Msene and Ujiji, begins with the
Vuli, when maize and millet, the voandzeia, various kinds of beans
and pulse, are committed to the well-hoed ground. Rice being
unknown, the people depend much upon holcus: this cereal, which
is sown in October to prepare for the Masika in November, has, in
the mountains, a short cane and a poor insipid grain of the red
variety. The people convert it into pombe; and they make the wine
called mawa from the plantains, which in several districts are more
abundant than the cereals. Karagwah grows according to some,
according to others imports from the northern countries, along the
western margin of the Nyanza Lake, a small wild coffee, locally
called mwámí. Like all wild productions, it is stunted and
undeveloped, and the bean, which, when perfect, is about the size
of a corking-pin’s head, is never drunk in decoction. The berry
gathered unripe is thrown into hot water to defend it from rot, or to
prevent its drying too rapidly—an operation which converts the husk
to a dark chocolate colour—the people of this country chew it like
tobacco, and, during visits, a handful is invariably presented to the
guest. According to the Arabs, it has, like the kishr of Yemen,
stimulating properties, affects the head, prevents somnolency,
renders water sweet to the taste, and forms a pleasant refreshing
beverage, which the palate, however, never confounds with the taste
of the Mocha-berry. In Karagwah a single khete of beads purchases
a kubabah (from 1 lb. to 2 lbs.) of this coffee; at Kazeh and Msene,
where it is sometimes brought by caravans, it sells at fancy prices.
Another well-known production of all these regions is the
mt’hípít’hípí, or Abrus precatorius, whose scarlet seeds are converted
into ornaments for the head.
The cattle is a fine variety, with small humps and large horns, like
that of Ujiji and Uviva. The herds are reckoned by Gundu, or
stallions, in the proportion of 1 to 100 cows. The late Sultan Ndagara
is said to have owned 200 Gundu, or 20,000 cows, which late civil
wars have reduced to 12,000 or 13,000. In Karagwah cattle forms
wealth, and everywhere in Africa wealth, and wealth only, secures
defenders and dependants. The surplus males are killed for beef;
this meat, with milk in its various preparations, and a little of the fine
white hill-honey, forms the food of the higher classes.
The people of Karagwah, who are not, according to South African
fashion, called Wakaragwah, are divided into two orders—Wahuma
and Wanyambo—who seem to bear to each other the relation of
patron and client, patrician and plebeian. The Wahuma comprises
the rich, who sometimes possess 1000 head of cattle, and the
warriors, a militia paid in the milk of cows allotted to their temporary
use by the king. The Wanyambo—Fellahs or Ryots—are, it is said,
treated by the nobles as slaves. The men of Karagwah are a tall
stout race, doubtless from the effect of pure mountain-air and
animal food. Corpulence is a beauty: girls are fattened to a vast bulk
by drenches of curds and cream thickened with flour, and are duly
disciplined when they refuse. The Arabs describe them as frequently
growing to a monstrous size, like some specimens of female Boers
mentioned by early travellers in Southern Africa. Fresh milk is the
male, sour the female beverage. The complexion is a brown yellow,
like that of the Warundi. The dress of the people, and even of the
chiefs, is an apron of close-grained mbugu, or bark-cloth, softened
with oil, and crimped with fine longitudinal lines made with a batten
or pounding club. In shape it resembles the flap of an English
saddle, tied by a prolongation of the upper corners round the waist.
To this scarcely decent article the chiefs add a languti, or Indian-T-
bandage of goat’s skin. Nudity is not uncommon, and nubile girls
assume the veriest apology for clothing, which is exchanged after
marriage for short kilts and breast coverings of skin. Both sexes
wear tiara-shaped and cravat-formed ornaments of the crimson
abrus-seed, pierced and strung upon mondo, the fine fibre of the
mwale or raphia-palm. The weapons are bows and arrows, spears,
knobsticks, and knives; the ornaments are beads and coil-bracelets,
which, with cattle, form the marriage settlement. The huts are of the
conical and circular African shape, with walls of stakes and roofs so
carefully thatched that no rain can penetrate them: the villages, as
in Usagara, are scattered upon the crests and ridges of the hills.
The Mkámá, or Sultan of Karagwah, in 1858, was Armanika, son
of Ndagara, who, although the dignity is in these lands hereditary,
was opposed by his younger brother Rumanika. The rebel, after an
obstinate attack, was routed by Suna, the late despot of Uganda,
who, bribed by the large present of ivory, which was advanced by
Musa Mzuri of Kazeh, then trading with Armanika, threw a large
force into the field. Rumanika was blinded and pensioned, and about
four years ago peace was restored. Armanika resides in the central
district, Weranhanja, and his settlement, inhabited only by the royal
family, contains from forty to fifty huts. He is described as a man
about thirty to thirty-five years old, tall, sturdy, and sinewy-limbed,
resembling the Somal. His dress is, by preference, the mbugu, or
bark-cloth, but he has a large store of fine raiment presented by his
Arab visitors: in ornaments he is distinguished by tight gaiters of
beads extending from knee to ankle. His diet is meat and milk, with
sometimes a little honey, plantains, and grain: unlike his subjects, he
eschews mawa and pombe. He has about a dozen wives, an
unusually moderate allowance for an African chief, and they have
borne him ten or eleven children. The royal family is said to be a
race of centagenarians; they are buried in their garments, sitting and
holding their weapons: when the king dies there is a funeral feast.
Under the Mkama is a single minister, who takes the title of
Muhinda, and presides over the Wakungu, elders and headmen,
whose duty it is to collect and to transmit to the monarch once every
month his revenues, in the shape of slaves and ivory, cattle and
provisions. Milk must be forwarded by proprietors of cows and herds
even from a distance of three days’ march. Armanika is an absolute
ruler, and he governs without squeamishness. Adulterers are
punished by heavy fines in cattle, murderers are speared and
beheaded, rebels and thieves are blinded by gouging out the eyes
with the finger-joints of the right-hand, and severing the muscles.
Subjects are forbidden to sell milk to those who eat beans or salt,
for fear of bewitching the animals. The Mkama, who lives without
state or splendour, receives travellers with courtesy. Hearing of their
approach, he orders his slaves to erect four or five tents for shelter,
and he greets them with a large present of provisions. He demands
no blackmail, but the offerer is valued according to his offerings: the
return gifts are carefully proportioned, and for beads which suit his
taste he has sent back an acknowledgment of fifty slaves and forty
cows. The price of adult male slaves varies from eight to ten fundo
of white, green, or blue porcelain-beads: a woman in her prime
costs two kitindi (each equal to one dollar on the coast), and five or
six fundo of mixed beasts. Some of these girls, being light-coloured
and well favoured, sell for sixty dollars at Zanzibar. The merchants
agree in stating that a European would receive in Karagwah the
kindest welcome, but that to support the dignity of the white face a
considerable sum would be required. Arabs still visit Armanika to
purchase slaves, cattle, and ivory, the whitest and softest, the
largest and heaviest in this part of Central Africa. The land is rich in
iron, and the spears of Karagwah, which are, to some extent,
tempered, are preferred to the rude work of the Wafyoma. Sulphur
is found, according to the Arabs, near hot springs amongst the
mountains. A species of manatus (?) supplies a fine skin used for
clothing. The simbi, or cowrie (Cypræa), is the minor currency of the
country: it is brought from the coast by return caravans of
Wanyamwezi.
The country of Karagwah is at present the head-quarters of the
Watosi, a pastoral people who are scattered throughout these Lake
Regions. They came, according to tradition, from Usingo, a mountain
district lying to the north of Uhha. They refuse to carry loads, to
cultivate the ground, or to sell one another. Harmless, and therefore
unarmed, they are often plundered, though rarely slain, by other
tribes, and they protect themselves by paying fees in cattle to the
chiefs. When the Wahinda are sultans, the Watosi appear as
councillors and elders; but whether this rank is derived from a
foreign and superior origin, or is merely the price of their presents,
cannot be determined. In appearance they are a tall, comely, and
comparatively fair people; hence in some parts every “distinguished
foreigner” is complimented by being addressed as “Mtosi.” They are
said to derive themselves from a single ancestor, and to consider the
surrounding tribes as serviles, from whom they will take concubines,
but to whom they refuse their daughters. Some lodges of this people
were seen about Unyanyembe and Msene, where they live by selling
cattle, milk, and butter. Their villages are poor, dirty, and
unpalisaded; mere scatters of ragged round huts. They have some
curious practices: never eat out of their own houses, and, after
returning from abroad, test, by a peculiar process, the fidelity of
their wives before anointing themselves and entering their houses.
The Arabs declare that they are known by their black gums, which
they consider a beauty.
The last feature of importance in Karagwah is the Kitangure River
on its northern frontier. This stream, deriving its name from a large
settlement on its banks, according to some travellers flows through a
rocky trough, according to others it traverses a plain. Some, again,
make it thirty yards, others 600, and even half a mile, in breadth. All
these statements are reconcileable. The river issues from Higher
Urundi, not far from the Malagarazi; but whilst the latter, engaged in
the Depression of Central Africa, is drawn towards the Tanganyika,
the former, falling into the counterslope, is directed to the north-east
into the Nyanza Lake. Its course would thus lie through a mountain-
valley, from which it issues into a lacustrine plain, the lowlands of
Unyoro and Uganda. The dark and swift stream must be crossed in
canoes even during the dry season, but, like the Malagarazi, about
June or at the end of the rains, it debords over the swampy lands of
its lower course.
From the Kitangure River fifteen stations conduct the traveller to
Kibuga, the capital district of Uganda, and the residence of its
powerful despot. The maximum of these marches would be six daily,
or a total of ninety rectilinear geographical miles. Though there are
no hills, the rivers and rivulets—said to be upwards of a hundred in
number—offer serious obstacles to rapid travelling. Assuming then,
the point where the Kitangure River is crossed to be in S. lat. 1° 14′,
Kibuga may be placed in S. lat. 0° 10′. Beyond Weranhanja no
traveller with claims to credibility has seen the Nyanza water. North
of Kibuga all is uncertain; the Arabs were not permitted by Suna, the
last despot, to penetrate farther north.
The two first marches from the Kitangure River traverse the
territory of “dependent Unyoro,” so called because it has lately
become subject to the Sultan of Uganda. In former times Unyoro in
crescent-shape, with the cusps fronting eastwards and westwards,
almost encompassed Uganda. From dependent Unyoro the path,
crossing a tract of low jungle, enters Uganda in the concave of the
crescent. The tributary Wahayya, under Gaetawa, their sultan, still
extend to the eastward. North of the Wahayya, of whose territory
little is known, lies “Kittara,” in Kinyoro (or Kiganda?), a word
interpreted to mean “mart,” or “meeting-place.” This is the region
which supplies Karagwah with coffee. The shrub is propagated by
sowing the bean. It attains the height of five feet, branching out
about half-way; it gives fruit after the third, and is in full vigour after
the fifth year. Before almost every hut-door there is a plantation,
forming an effective feature in the landscape of rolling and wavy hill,
intersected by a network of rivers and streams: the foliage is
compared to a green tapestry veiling the ground; and at times,
when the leaves are stripped off by wind and rain, the plant appears
decked with brilliant crimson cherry-like berries. The Katonga River,
crossed at Kitutu, is supposed to fall into the Nyanza, the general
recipient of the network of streams about Karagwah. This diagonality
may result from the compound incline produced by the northern
counterslope of the mountains of Karagwah and the south-westward
depression necessary to form and to supply the lake. The Katonga is
a sluggish and almost stagnant body of considerable breadth, and
when swollen it arrests the progress of caravans. Some portions of
the river are crossed, according to the Arabs, over a thick growth of
aquatic vegetation, which forms a kind of matwork, capable of
supporting a man’s weight, and cattle are towed over in the more
open parts by cords attached to their horns. Four stations lead from
the Katonga River to Kibuga, the capital district of Uganda.
Kibuga is the residence of the great Mkámá or chief of Uganda.
Concerning its population and peculiarities the Arabs must be
allowed to tell their own tale. “Kibuga, the settlement, is not less
than a day’s journey in length; the buildings are of cane and rattan.
The sultan’s palace is at least a mile long, and the circular huts,
neatly ranged in line, are surrounded by a strong fence which has
only four gates. Bells at the several entrances announce the
approach of strangers, and guards in hundreds attend there at all
hours. They are commanded by four chiefs, who are relieved every
second day: these men pass the night under hides raised upon
uprights, and their heads are forfeited if they neglect to attend to
the summons of the king. The harem contains about 3000 souls—
concubines, slaves, and children. No male nor adult animal may
penetrate, under pain of death, beyond the Barzah, a large vestibule
or hall of audience where the king dispenses justice and receives his
customs. This palace has often been burned down by lightning: on
these occasions the warriors must assemble and extinguish the fire
by rolling over it. The chief of Uganda has but two wants with which
he troubles his visitors—one, a medicine against death; the other, a
charm to avert the thunderbolt: and immense wealth would reward
the man who could supply either of these desiderata.”
Suna, the great despot of Uganda, a warlike chief, who wrested
dependent Unyoro from its former possessor, reigned till 1857. He
perished in the prime of life and suddenly, as the Arabs say, like
Namrud, whilst riding “pickaback”—the state carriage of Central
Africa—upon a minister’s shoulders, he was struck by the shaft of
the destroyer in the midst of his mighty host. As is the custom of
barbarous and despotic races, the event was concealed for some
months. When the usual time had expired, one of his many sons,
exchanging his heir-elective name “Sámunjú” for Mtesa, became
king. The court usage compels the newly elected chief to pass two
years in retirement, committing state affairs to his ministers; little,
therefore, is yet known of him. As he will certainly tread in the
footsteps of his sire, the Arabs may again be allowed to describe the
state and grandeur of the defunct Suna; and as Suna was in fact the
whole kingdom of Uganda, the description will elucidate the
condition of the people in general.
“The army of Uganda numbers at least 300,000 men; each brings
an egg to muster, and thus something like a reckoning of the people
is made. Each soldier carries one spear, two assegais, a long dagger,
and a shield, bows and swords being unknown. When marching the
host is accompanied by women and children carrying spare
weapons, provisions, and water. In battle they fight to the sound of
drums, which are beaten with sticks like those of the Franks: should
this performance cease, all fly the field. Wars with the Wanyoro, the
Wasoga, and other neighbours are rendered almost chronic by the
policy as well as the pleasure of the monarch, and there are few
days on which a foraging party does not march from or return to the
capital. When the king has no foreign enemies, or when the
exchequer is indecently deficient, he feigns a rebellion, attacks one
of his own provinces, massacres the chief men, and sells off the
peasantry. Executions are frequent, a score being often slain at a
time: when remonstrated with concerning this barbarity, Sana
declared that he had no other secret for keeping his subjects in awe
of him, and for preventing conspiracies. Sometimes the king would
accompany his army to a battue of game, when the warriors were
expected to distinguish themselves by attacking the most ferocious
beasts without weapons: even the elephant, borne down by
numbers, yielded to the grasp of man. When passing a village he
used to raise a shout, which was responded to by a loud flourish of
horns, reed-pipes, iron whistles, and similar instruments. At times he
decreed a grand muster of his soldiery: he presented himself sitting
before his gate, with a spear in the right hand, and holding in the
left the leash of a large and favourite dog resembling an Arab suluki
or greyhound. The master of the hounds was an important
personage. Suna took great pleasure in witnessing trials of strength,
the combatants contending with a mixture of slapping and pushing
till one fell to the ground. He had a large menagerie of lions,
elephants, leopards, and similar beasts of disport, to whom he would
sometimes give a criminal as a ‘curée:’ he also kept for amusement
fifteen or sixteen albinos; and so greedy was he of novelty that even
a cock of peculiar or uniform colour would have been forwarded by
its owner to feed his eyes.”
Suna when last visited by the Arabs was a “red man,” aged about
forty-five, tall, robust, and powerful of limb, with a right kingly
presence and a warrior carriage. His head was so shaven as to leave
what the Omani calls “el Kishshah,” a narrow crest of hair like a
cock’s comb, from nape to brow; nodding and falling over his face
under its weight of strung beads, it gave him a fierce and formidable
aspect. This tonsure, confined to those about the palace,
distinguishes its officers and inmates, servile as well as free, from
the people. The Ryots leave patches of hair where they please, but
they may not shave the whole scalp under pain of death, till a royal
edict unexpectedly issued at times commands every head to shed its
honours. Suna never appeared in public without a spear; his dress
was the national costume, a long piece of the fine crimped mbugu or
bark-cloth manufactured in these regions, extending from the neck
to the ground. He made over to his women the rich clothes
presented by the Arabs, and allowed them to sew with unravelled
cotton thread, whereas the people under severe penalties were
compelled to use plantain fibre. No commoner could wear domestics
or similar luxuries; and in the presence, the accidental exposure of a
limb led, according to the merchants, to the normal penalty—death.
Suna, like the northern despots generally, had a variety of names,
all expressing something bitter, mighty, or terrible, as, for instance,
Lbare, the Almighty (?); Mbidde and Purgoma, a lion. He could not
understand how the Sultan of Zanzibar allowed his subjects
treasonably to assume the name of their ruler; and besides
mortifying the Arabs by assuming an infinite superiority over their
prince, he shocked them by his natural and unaffected impiety. He
boasted to them that he was the god of earth, as their Allah was the
Lord of Heaven. He murmured loudly against the abuse of lightning;
and he claimed from his subjects divine honours, which were as
readily yielded to him as by the facile Romans to their emperors. No
Mgándá would allow the omnipotence of his sultan to be questioned,
and a light word concerning him would have imperilled a stranger’s
life. Suna’s domestic policy reminds the English reader of the African
peculiarities which form the groundwork of “Rasselas.” His sons,
numbering more than one hundred, were removed from the palace
in early youth to separate dungeons, and so secured with iron collars
and fetters fastened to both ends of a long wooden bar that the
wretches could never sit, and without aid could neither rise nor lie.
The heir-elective was dragged from his chains to fill a throne, and
the cadets will linger through their dreadful lives, unless wanted as
sovereigns, until death release them. Suna kept his female children
under the most rigid surveillance within the palace: he had, however,
a favourite daughter named Nasuru, whose society was so necessary
to him that he allowed her to appear with him in public.
The principal officers under the despot of Uganda are, first, the
Kimara Vyona (literally the “finisher of all things”): to him, the chief
civilian of the land, the city is committed; he also directs the kabaka
or village headmen. The second is the Sakibobo or commander-in-
chief, who has power over the Sáwágánzí, the life-guards and slaves,
the warriors and builders of the palace. Justice is administered in the
capital by the sultan, who, though severe, is never accused of
perverting the law, which here would signify the ancient custom of
the country. A Mhozi—Arabised to Hoz, and compared with the Kazi
of el Islam—dispenses in each town criminal and civil rights. The
only punishments appear to be death and mulcts. Capital offenders
are beheaded or burned; in some cases they are flayed alive; the
operation commences with the face, and the skin, which is always
much torn by the knife, is stuffed as in the old torturing days of Asia.
When a criminal absconds, the males of his village are
indiscriminately slain and the women are sold—blood and tears must
flow for discipline. In money suits each party begins by placing
before the Mhozi a sum equivalent to the disputed claim; the object
is to prevent an extensive litigiousness. Suna used to fine by fives or
tens, dozens or scores, according to the offender’s means; thus from
a wealthy man he would take twenty male and twenty female slaves,
with a similar number of bulls and cows, goats and kids, hens and
even eggs. One of his favourites, who used constantly to sit by him
on guard, matchlock in hand, was Isa bin Hosayn, a Baloch
mercenary of H. H. Sayyid Said of Zanzibar. He had fled from his
debtors, and had gradually wandered to Uganda, where the favour
of the sovereign procured him wealth in ivory, and a harem
containing from 200 to 300 women. “Mzagayya,”—the hairy one, as
he was locally called, from his long locks and bushy beard—was not
permitted, nor probably did he desire, to quit the country; after his
patron’s death he fled to independent Unyoro, having probably
raised up, as these adventurers will, a host of enemies at Uganda.
Suna greatly encouraged, by gifts and attention, the Arab
merchants to trade in his capital; the distance has hitherto
prevented more than half-a-dozen caravans travelling to Kibuga; all
however came away loudly praising his courtesy and hospitality. To a
poor trader he has presented twenty slaves, and an equal number of
cows, without expecting any but the humblest return. The following
account of a visit paid to him in 1852, by Snay bin Amir, may
complete his account of the despot Uganda. When the report of
arrival was forwarded by word of mouth to Suna, he issued orders
for the erection of as many tents as might be necessary. The guest,
who was welcomed with joyful tumult by a crowd of gazers, and was
conducted to the newly-built quarters, where he received a present
of bullocks and grain, plantains and sugar-canes. After three or four
days for repose, he was summoned to the Barzah or audience hall,
outside of which he found a squatting body of about 2000 guards
armed only with staves. Allowed to retain his weapons, he entered
with an interpreter and saluted the chief, who, without rising,
motioned his guest to sit down in front of him. Suna’s only cushion
was a mbugu; his dress was of the same stuff; two spears lay close
at hand, and his dog was as usual by his side. The Arab thought
proper to assume the posture of homage, namely, to sit upon his
shins, bending his back, and, with eyes fixed on the ground—he had
been cautioned against staring at the “god of earth,”—to rest his
hands upon his lap. The levee was full; at a distance of fifty paces
between the king and the guards sat the ministers; and inside the
palace, so placed that they could see nothing but the visitor’s back,
were the principal women, who are forbidden to gaze at or to be
gazed at by a stranger. The room was lit with torches of a gummy
wood, for Suna, who eschewed pombe, took great pleasure in these
audiences, which were often prolonged from sunset to midnight.
The conversation began with a string of questions concerning
Zanzibar, the route, the news, and the other staple topics of
barbarous confabulation; when it flagged, a minister was called up
to enliven it. No justice was administered nor present offered during
the first audience; it concluded with the despot rising, at which
signal all dispersed. At the second visit Snay presented his blackmail,
which consisted of ten cotton cloths, and one hundred fundo of
coral, and other porcelain beads. The return was an offering of two
ivories and a pair of serviles; every day, moreover, flesh and grain,
fruit and milk were supplied without charge; whenever the wish was
expressed, a string of slave-girls presently appeared bending under
loads of the article in question; and it was intimated to the “king’s
stranger” that he might lay hands upon whatever he pleased,
animate or inanimate. Snay, however, was too wise to avail himself
of this truly African privilege. During the four interviews which
followed, Suna proved himself a man of intelligence: he inquired
about the Wazungu or Europeans, and professed to be anxious for a
closer alliance with the Sultan of Zanzibar. When Snay took leave he
received the usual present of provisions for the road, and 200
guards prepared to escort him, an honour which he respectfully
declined: Suna offered to send with him several loads of elephants’
tusks as presents to H. H. the Sayyid; but the merchant declined to
face with them the difficulties and dangers of Usúí. Like all African
chiefs, the despot considered these visits as personal honours paid
to himself; his pride therefore peremptorily forbade strangers to pass
northwards of his capital, lest the lesser and hostile chiefs might
boast a similar brave. According to Snay, an European would be
received with distinction, if travelling with supplies to support his
dignity. He would depend, however, upon his ingenuity and good
fortune upon further progress; and perhaps the most feasible plan to
explore the water-shed north of the Nyanza Lake would be to buy or
to build, with the permission of the reigning monarch, boats upon
the nearest western shore. Suna himself, had, according to Snay,
constructed a flotilla of matumbi or undecked vessels similar in
shape to the Mtope or Muntafiyah—the modern “Ploiaria Rhapta” of
the Sawahili coast from Lamu to Kilwa.
Few details were given by the Arabs concerning the vulgar herd of
Waganda: they are, as has been remarked, physically a finer race
than the Wayamwezi, and they are as superior in character; more
docile and better disciplined, they love small gifts, and show their
gratitude by prostrating themselves before the donor. The specimens
of slaves seen at Kazeh were, however, inferior to the mountaineers
of Karagwah; the complexion was darker, and the general
appearance more African. Their language is, to use an Arab phrase,
like that of birds, soft and quickly spoken; the specimens collected
prove without doubt that it belongs to the Zangian branch of the
great South-African family. Their normal dress is the mbugu, under
which, however, all wear the “languti” or Indian-T-bandage of
goatskin; women appear in short kilts and breast-coverings of the
same material. Both sexes decorate their heads with the tiara of
abrus-seeds alluded to when describing the people of Karagwah. As
sumptuary laws impede the free traffic of cloth into Uganda, the
imports are represented chiefly by beads, cowries, and brass and
copper wires. The wealth of the country is in cattle, ivory, and
slaves, the latter often selling for ten fundo of beads, and the same
sum will purchase the Wasoga and Wanyoro captives from whom the
despot derives a considerable portion of his revenues. The elephant
is rare in Uganda; tusks are collected probably by plunder from
Usoga, and the alakah of about ninety Arab pounds is sold for two
slaves, male or female. The tobacco, brought to market in leaf, as in
Ujiji, and not worked, as amongst the other tribes, is peculiarly
good. Flesh, sweet potatoes, and the highly nutritious plantain,
which grows in groves a whole day’s march long, are the chief
articles of diet; milk is drunk by women only, and ghee is more
valued for unction than for cookery. The favourite inebrients are
mawa and pombe; the latter is served in neatly carved and coloured
gourds, and the contents are imbibed, like sherry cobbler, through a
reed.
From Kibuga the Arabs have heard that between fifteen and
twenty marches lead to the Kivira River, a larger and swifter stream
than the Katonga, which forms the northern limit of Uganda, and the
southern frontier of Unyoro. They are unable to give the names of
stations. South of Kivira is Usoga, a low alluvial land, cut by a
multitude of creeks, islets, and lagoons; in their thick vegetation the
people take refuge from the plundering parties of the Waganda,
whose chief built, as has been told, large boats to dislodge them.
The Wasoga have no single sultan, and their only marketable
commodity is ivory.
On the north, the north-west, and the west of Uganda lies,
according to the Arabs, the land of Independent Unyoro. The slaves
from that country vaguely describe it as being bounded on the
north-west by a tribe called Wakede, who have a currency of
cowries, and wear tiaras of the shell; and the Arabs have heard that
on the north-east there is a “people with long daggers like the
Somal,” who may be Gallas (?). But whether the Nyanza Lake
extends north of the equator is a question still to be decided. Those
consulted at Kazeh ignored even the name of the Nyam-nyam; nor
had they heard of the Bahri and Barri, the Shilluks on the west, and
the Dinkas east of the Nile, made familiar to us by the Austrian
Mission at Gondokoro, and other explorers.
The Wanyoro are a distinct race, speaking a language of the
Zangian family: they have suffered from the vicinity of the more
warlike Waganda, who have affixed to the conquered the
opprobrious name of widdu or “serviles;” and they have lost their
southern possessions, which formerly extended between Karagwah
and Uganda. Their late despot Chawambi, whose death occurred
about ten years ago, left three sons, one of whom it is reported has
fallen into the power of Uganda, whilst the two others still rule
independently. The county is rich and fertile, and magnificent tales
are told concerning the collections of ivory, which in some parts are
planted in the ground to pen cattle. Slaves are cheap; they find their
way to the southern markets viâ Uganda and Karagwah. Those seen
at Kazeh and Kirira, where the Arab traders had a large gang,
appeared somewhat inferior to the other races of the northern
kingdoms, with a dull dead black colour, flattish heads, brows
somewhat retreating, prominent eyes, and projecting lower jaws.
They were tattooed in large burnt blotches encircling the forehead,
and in some cases the inferior excisors had been extracted. The
price of cattle in Unyoro varies from 500 to 1000 cowries. In this
country ten simbi (Cypræa) represent one khete of beads; they are
the most esteemed currency, and are also used as ornaments for the
neck, arms, and legs, and decorations for stools and drums.
During my companions’ absence much of my spare time was
devoted to collecting specimens of the multitudinous dialects into
which the great South African family here divides itself. After some
months of desultory work I had learned the Kisawahili or coast
language, the lingua Franca of the South African coast: it is the most
useful, because the most generally known, and because, once
mastered, it renders its cognates as easy of acquirement as Bengali
or Maharatti after Hindostani. The principal obstacle is the want of
instructors and books—the Kisawahili is not a written language; and
the elementary publications put forth in Europe gave me the
preliminary trouble of composing a grammar and a vocabulary. Said
Bin Salim, though bred and born amongst the Wasawahili, knew but
little of the tongue, and his peculiarities of disposition rendered the
task of instruction as wearisome to himself as it was unsatisfactory
to me. My best tutor was Snay Bin Amir, who had transferred to the
philology of East Africa his knowledge of Arabic grammar and
syntax. With the aid of the sons of Ramji and other tame slaves, I
collected about 1500 words in the three principal dialects upon this
line of road, namely the Kisawahili, the Kizaramo—which includes
the Kik’hutu—and the Kinyamwezi. At Kazeh I found a number of
wild captives, with whom I began the dreary work of collecting
specimens. In the languages of least consideration I contented
myself with the numerals, which are the fairest test of independence
of derivation, because the most likely to be primitive vocables. The
work was not a labour of love. The savages could not guess the
mysterious objects of my inquiry into their names for 1, 2, and 3;
often they started up and ran away, or they sat in dogged silence,
perhaps thinking themselves derided. The first number was rarely
elicited without half an hour’s “talkee-talkee” somewhat in this
style:—
 “Listen, O my brother! in the tongue of the shores (Kisawahili) we
say 1, 2, 3, 4, 5”—counting the fingers to assist comprehension.
“Hu! hu!” replies the wild man, “we say fingers.”
“By no means, that’s not it. This white man wants to know how
thou speakest 1, 2, 3?”
“One, two, three what? sheep, or goats, or women?”—expressing
the numerals in Kisawahili.
“By no means, only 1, 2, 3 sheep in thine own tongue, the tongue
of the Wapoka.”
“Hi! Hi! what wants the white man with the Wapoka?”
And so on till patience was almost impossible. But, like the Irish
shay-horse of days gone by, their tongues once started often
hobbled on without halting. The tame slaves were more tractable,
yet even in their case ten minutes sufficed to weary out the most
intellectual; when the listless and incoherent reply, the glazed eye
gazing at vacancy, and the irresistible tendency to gape and yawn,
to nod and snooze, evidenced a feeble brain soon overworked. Said
Bin Salim would sit staring at me with astonishment, and ejaculate,
like Abba Gregorius, the preceptor of Ludolph, the grammarian
philologist and historian of Æthiopia, “Verily in the coast-tongue
words never take root, nor do they bear branches.”
The rest of my time was devoted to preparations for journeying.
The Fundi’s tent, which had accompanied us to Uvira, was provided
with an outer cover. The Sepoys’ “pal,” brought from Zanzibar,
having been destroyed by the ill-treatment of the villain Kannena, I
made up, with the aid of a blackguard Baghdadi, named ’Brahim, a
large tent of American domestics, which having, however, but one
cloth, and that of the thinnest, proved a fiery purgatory on the
down-march eastwards. The canvas lug-sail was provided with an
extra double cloth, sewn round the top to increase its dimensions: it
thus became a pent-shaped affair, twelve feet long, eight broad, and
six feet high—seven would have been better,—buttoned at the foot,
which was semicircular, and in front provided with blue cotton
curtains, most useful against glare and stare. Its lightness, combined
with impenetrability, made it the model of a tent for rapid marching.
It was not, however, pegged down close to the ground, as some
explorers advise, without the intervention of ropes; in these lands, a
tent so pitched would rot in a week. The three tents were fitted with
solid male bamboos, and were provided with skin-bags for their
pegs, which, unless carefully looked after, disappear almost daily.
The only furniture was a kitanda or cartel: some contrivance of the
kind, a “Biddulph,” or an iron bed-frame, without joints, nuts, or
screws, which are sure to break or to be lost, is absolutely necessary
in these lands, where from Kaole to Uvira every man instinctively
attempts to sit and to sleep upon something that raises him above
the ground. Moreover, I have ever found the cartel answer the
threefold purpose of bed, chair, and table; besides saving weight by
diminishing the quantity of bedding required.
To the task of tent-making succeeded tailoring. We had neglected
to provide ourselves with the loose blanket suits, served out to
sailors on board men-of-war in the tropics: they are most useful in
passing through countries where changes of climate are sudden and
marked. Besides these, the traveller should carry with him an ample
store of flannels: the material must be shrunk before making up
shirts, otherwise it will behave as did the Little Boy’s mantle when
tried by the frail fair Guinever. A red colour should moreover be
avoided, the dye soon turns dark, and the appearance excites too
much attention. Besides shirt and trousers, the only necessary is a
large “stomach-warmer” waistcoat, with sleeves and back of similar
material, without collar—which renders sleeping in it uneasy—and
provided with four flapped pockets, to contain a compass and
thermometer, a note-book, and a sketch-book, a watch and a
moderate-sized knife of many uses. The latter should contain
scissors, tweezers, tooth-pick, and ear-pick, needle, file, picker, steel
for fire, turnscrew, watch-spring-saw, clasp blade, and pen blade: it
should be made of moderate dimensions, and for safety be slung by
a lanyard to the button-hole. For the cold mornings and the noon-
day heats, I made up a large padded hood, bound round the head
like the Arab Kufiyah. Too much cannot be said in favour of this
article, which in eastward travel defends the eyes from the fiery
glare, protects, when wending westwards, the carotids against the
solar blaze, and, at all times, checks the intrusive staring of the
crowd. I reformed my umbrella, ever an invaluable friend in these
latitudes, by removing the rings and wires from the worm-eaten
stick, and by mounting them on a spear, thus combining with shelter
a staff and a weapon. The traveller should have at least three
umbrellas, one large and water proof—white, not black—in the
shape of those used by artists; and two others of moderate size, and
of the best construction, which should be covered with light-coloured
calico, as an additional defence against the sun. At Kazeh I was
somewhat deficient in material: my lazy “Jack of all trades,”
Valentine, made, however, some slippers of green baize, soled with
leather, for me, overalls of American domestics for my companion,
and various articles of indigo-dyed cotton for himself and his fellow-
servant, who presently appeared tastefully rigged out like Paul and
Virginia in “Bengal blue.”
The minor works were not many. The two remaining
portmanteaus of the three that had left the Coast were cobbled with
goatskins, and were bound with stout thongs. The hammocks, of
which half had disappeared, were patched and provided with the
Nara, or Indian cotton-tape, which in these climates is better than
either reims or cord. To save my eyes the spectacle of moribund
fowls, suspended to a porter’s pole, two light cages were made after
the fashion of the country, with bent and bound withes. The metal
plates, pots, and pans were furbished, and a damaged kettle was
mended by a travelling tinker: the asses’ saddles and halters were
repaired, and, greatest luxury of all, a brace of jembe or iron hoes
was converted into two pairs of solid stirrups, under the vigilant eye
of Snay bin Amir. A party of slaves sent to Msene brought back fifty-
four jembe, useful as return-presents and blackmail on the down-
march: they paid, however, one cloth for two, instead of four. Sallum
bin Hamid, the “papa” of the Arabs, sold for the sum of forty dollars
a fine half-bred Zanzibar she-ass and foal—there is no surer method
of procuring a regular supply of milk on Eastern journeys. My black
and white beads being almost useless, he also parted with, as a
peculiar favour, seventeen or eighteen pounds of pink-porcelains for
forty dollars, and with a Frasibah of coffee, and a similar quantity of
sugar for eighty dollars, equal to sixteen pounds sterling. On the
14th July the last Arab caravan of the season left Unyanyembe,
under the command of Sayf bin Said el Wardi. As he obligingly
offered to convey letters and any small articles which I wished to
precede me, and knowing that under his charge effects were far
safer than with our own people, I forwarded the useless and
damaged surveying instruments, certain manuscripts, and various
enclosures of maps, field and sketch-books, together with reports to
the Royal Geographical Society.
This excitement over I began to weary of Kazeh. Snay bin Amir
and most of the Arabs had set out on an expedition to revenge the
murder of old Silim—an event alluded to in a former page, and the
place had become dull as a mess-dinner. Said bin Salim, who was ill,
who coughed and expectorated, and sincerely pitied himself because
he had a cold, became more than usually unsociable: he could enjoy
nothing but the society of Brahim, the bawling Baghdadi, and the
crowd of ill-flavoured slavery that flocked into the vestibule. My
Goanese servant, who connected my aspect with hard labour,
avoided it like a pestilence. Already I was preparing to organise a
little expedition to K’hokoro and the southern provinces, when
unexpectedly,—in these lands a few cries and gun-shots are the only
credible precursors of a caravan,—on the morning of the 25th
August reappeared my companion.
At length my companion had been successful, his “flying trip” had
led him to the northern water, and he had found its dimensions
surpassing our most sanguine expectations. We had scarcely,
however, breakfasted, before he announced to me the startling fact,
that he had discovered the sources of the White Nile. It was an
inspiration perhaps: the moment he sighted the Nyanza, he felt at
once no doubt but that the “Lake at his feet gave birth to that
interesting river which has been the subject of so much speculation,
and the object of so many explorers.” The fortunate discoverer’s
conviction was strong; his reasons were weak—were of the category
alluded to by the damsel Lucetta, when justifying her penchant in
favour of the “lovely gentleman,” Sir Proteus:—
 “I have no other but a woman’s reason.
I think him so because I think him so;”[11]

and probably his sources of the Nile grew in his mind as his
Mountains of the Moon had grown under his hand.
[11] The following extract from the Proceedings of the R. Geographical
Society, May 9, 1859, will best illustrate what I mean:—
Mr. Macqueen, f.r.g.s., said the question of the sources of the Nile had
cost him much trouble and research, and he was sure there was no
material error either in longitude or latitude in the position he had ascribed
to them, namely, a little to the eastward of the meridian of 35°, and a little
northward of the equator. That was the principal source of the White Nile.
The mountains there were exceedingly high, from the equator north to
Kaffa Enarea. All the authorities, from east, west, north, or south, now
perfectly competent to form judgments upon such a matter, agreed with
him; and among them were the officers commanding the Egyptian
commission. It was impossible they could all be mistaken. Dr. Krapf had
been within a very short distance of it; he was more than 180 miles from
Mombas, and he saw snow upon the mountains. He conversed with the
people who came from them, and who told him of the snow and exceeding
coldness of the temperature. The line of perpetual congelation, it was well
known, was 17,000 feet above the sea. He had an account of the navigation
of the White Nile by the Egyptian expedition. It was then given as 3° 30′ N.
lat. and 31° E. long. At this point the expedition turned back for want of a
sufficient depth of water. Here the river was 1370 feet broad, and the
velocity of the current one-quarter of a mile per hour. The journals also
gave a specific and daily current, the depth and width of the river, and
every thing, indeed, connected with it. Surely, looking at the current of the
river, the height of the Cartoom above the level of the sea, and the distance
thence up to the equator, the sources of the Nile must be 6000 or 8000 feet
above the level of the sea, and still much below the line of the snow, which
was 6000 or 8000 feet farther above them. He deeply regretted he was
unable to complete the diagram for the rest of the papers he had given to
the Society, for it was more important than any others he had previously
given. It contained the journey over Africa from sea to sea, second only to
that of Dr. Livingstone. But all the rivers coming down from the mountains
in question, and running south-eastward, had been clearly stated by Dr.
Krapf, who gave every particular concerning them. He should like to know
what the natives had said was to the northward of the large lake? Did they
say the rivers ran out from or into the lake? How could the Egyptian officers
be mistaken?
 Captain Speke replied. They were not mistaken; and if they had pursued
their journey 50 miles farther, they would undoubtedly have found
themselves at the northern borders of this lake.
Mr. Macqueen said that other travellers, Don Angelo for instance, had
been within one and a half degree of the Equator, and saw the mountain of
Kimborat under the Line, and persisted in the statement, adding, that
travellers had been up the river until they found it a mere brook. He felt
convinced that the large lake alluded to by Captain Speke was not the
source of the Nile: it was impossible it could be so, for it was not at a
sufficiently high altitude.
The paper presented to the Society, when fully read in conjunction with
the map, will clearly show that the Bahr-el-Abied has no connection with
Kilimanjaro, that it has no connection whatever with any lake or river to the
south of the Equator, and that the swelling of the river Nile proceeds from
the tropical rains of the northern torrid zone, as was stated emphatically to
Julius Cæsar by the chief Egyptian priest Amoreis 2000 years ago.
In nearly 3° N. lat. there is a great cataract, which boats cannot pass. It
is called Gherba. About half-way (50 miles) above, and between this
cataract and Robego, the capital of Kuenda, the river becomes so narrow as
to be crossed by a bridge formed by a tree thrown across it. Above Gherba
no stream joins the river either from the south or south-west.

The main argument in favour of the Lake representing the great

reservoir of the White River was, that the “principal men” at the
southern extremity ignored the extent northward. “On my inquiring
about the lake’s length the man (the greatest traveller in the place)
faced to the north, and began nodding his head to it; at the same
time he kept throwing forward his right hand, and making repeated
snaps of his fingers endeavoured to indicate something
immeasurable; and added, that nobody knew, but he thought it
probably extended to the end of the world.” Strongly impressed by
this valuable statistical information, my companion therefore placed
the northern limit about 4°-5° north lat., whereas the Egyptian
expedition sent by the late Mohammed Ali Pacha, about twenty
years ago, to explore the Coy Sources, reached 3° 22′ north lat. It
therefore ought to have sailed fifty miles upon the Nyanza lake. On
the contrary, from information derived on the spot, that expedition
placed the fountains at one month’s journey—300 to 350 miles—to
the south-east, or upon the northern counterslope of Mount Kenia.
Whilst marching to the coast, my companion—he tells us—was
assured by a “respectable Sowahili merchant, that when engaged in
traffic some years previously to the northward of the line, and the
westward of this lake, he had heard it commonly reported that large
vessels frequented the northern extremity of these waters, in which
the officers engaged in navigating them used sextants and kept a
log, precisely similar to what is found in vessels on the ocean. Query,
could this be in allusion to the expedition sent by Mohammed Ali up
the Nile in former years?” (Proceedings of Royal Geographical
Society, May 9, 1859.) Clearly, if Abdullah Bin Nasib, the Msawahili
alluded to, had reported these words, he merely erred; the Egyptian
expedition, as has been shown, not only did not find, they never
even heard of a lake. But not being present at the conversation I am
tempted to assign further explanation. My companion, wholly
ignorant of Arabic, was reduced to depend upon “Bombay,” who
spoke an even more debased dialect than his master, and it is easy
to see how the blunder originated. The Arabic bahr and the
Kisawahili bahari are equally applicable in vulgar parlance to a river
or sea, a lake or a river. Traditions concerning a Western Sea—the to
them now unknown Atlantic—over which the white men voyage, are
familiar to many East Africans; I have heard at Harar precisely the
same report concerning the log and sextants. Either, then, Abdullah
Bin Nasib confounded, or my companion’s “interrupter” caused him
to confound the Atlantic and the Lake. In the maps forwarded from
Kazeh by my companion, the River Kivira was, after ample inquiry,
made a western influent of the Nyanza Lake. In the map appended
to the paper in Blackwood, before alluded to, it has become an
effluent, and the only minute concerning so very important a
modification is, “This river (although I must confess at first I did not
think so) is the Nile itself!”
Beyond the assertion, therefore, that no man had visited the
north, and the appearance of sextants and logs upon the waters,
there is not a shade of proof pro. Far graver considerations lie on the
con. side: the reports of the Egyptian expedition, and the dates of
the several inundations which—as will presently appear—alone
suffice to disprove the possibility of the Nyanza causing the flood of
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

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

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebooknice.com