Visit https://ebookfinal.

com to download the full version and

explore more ebooks

A Whistle Stop Tour of Statistics Everitt

_____ Click the link below to download _____

https://ebookfinal.com/download/a-whistle-stop-tour-of-
statistics-everitt/

Explore and download more ebooks at ebookfinal.com

Here are some suggested products you might be interested in.
Click the link to download

Encyclopaedic Companion to Medical Statistics 2nd Edition

Brian S. Everitt

https://ebookfinal.com/download/encyclopaedic-companion-to-medical-
statistics-2nd-edition-brian-s-everitt/

Medical Statistics from A to Z A Guide for Clinicians and

Medical Students 2nd Edition B. S. Everitt

https://ebookfinal.com/download/medical-statistics-from-a-to-z-a-
guide-for-clinicians-and-medical-students-2nd-edition-b-s-everitt/

A Quantitative Tour of the Social Sciences Andrew Gelman

https://ebookfinal.com/download/a-quantitative-tour-of-the-social-
sciences-andrew-gelman/

A handbook of statistical analyses using Stata 3rd ed

Edition Brian S. Everitt

https://ebookfinal.com/download/a-handbook-of-statistical-analyses-
using-stata-3rd-ed-edition-brian-s-everitt/
Riddles Of Existence A Guided Tour Of Metaphysics 1st
Edition Earl Conee

https://ebookfinal.com/download/riddles-of-existence-a-guided-tour-of-
metaphysics-1st-edition-earl-conee/

Start a revolution stop acting like a library Bizzle

https://ebookfinal.com/download/start-a-revolution-stop-acting-like-a-
library-bizzle/

Combinatorics A Guided Tour 1st Edition David R. Mazur

https://ebookfinal.com/download/combinatorics-a-guided-tour-1st-
edition-david-r-mazur/

Ordinary Differential Equations A Brief Eclectic Tour

David A. Sánchez

https://ebookfinal.com/download/ordinary-differential-equations-a-
brief-eclectic-tour-david-a-sanchez/

The Seven Hills of Rome A Geological Tour of the Eternal

City Grant Heiken

https://ebookfinal.com/download/the-seven-hills-of-rome-a-geological-
tour-of-the-eternal-city-grant-heiken/
A Whistle Stop Tour of Statistics Everitt Digital Instant
Download
Author(s): Everitt, Brian S
ISBN(s): 9781439877494, 1439877491
Edition: Online-Ausg
File Details: PDF, 2.63 MB
Year: 2011
Language: english
 Statistics

A Whistle-Stop Tour of Statistics

A Whistle-Stop Tour of Statistics A Whistle-Stop Tour of Statistics
A Whistle-Stop Tour of Statistics introduces basic probability and
statistics through bite-size coverage of key topics. A review aid and
study guide for undergraduate students, it presents descriptions of key
concepts from probability and statistics in self-contained sections.

Features
• Presents an accessible reference to the key concepts in
probability and statistics
• Introduces each concept through bite-size descriptions and
presents interesting real-world examples
• Includes lots of diagrams and graphs to clarify and illustrate
topics
• Provides a concise summary of ten major areas of statistics
including survival analysis and the analysis of longitudinal data

Written by Brian Everitt, the author of over 60 statistical texts, the book
shows how statistics can be applied in the real world, with interesting
examples and plenty of diagrams and graphs to illustrate concepts.

Brian S. Everitt
Everitt

K13590

K13590_Cover.indd 1 11/2/11 9:22 AM

 A Whistle-Stop Tour of Statistics

K13590.indb 1 03/11/11 4:52 PM

This page intentionally left blank
 A Whistle-Stop Tour of Statistics
Brian S. Everitt

K13590.indb 3 03/11/11 4:52 PM

CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2012 by Taylor & Francis Group, LLC

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

No claim to original U.S. Government works

Version Date: 20111025

International Standard Book Number-13: 978-1-4398-7749-4 (eBook - PDF)

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

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

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

Trademark Notice: Product or corporate names may be trademarks or registered trademarks,

and are used only for identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com

and the CRC Press Web site at

http://www.crcpress.com
 To all my long-suffering doubles partners at the Edward Alleyn Tennis
Club

K13590.indb 5 03/11/11 4:52 PM

This page intentionally left blank
 Contents

Preface xi

1 Some Basics and Describing Data 1

1.1 Population, Samples and Variables 1
1.2 Types of Variables 2
1.3 Tabulating and Graphing Data: Frequency Distributions,
Histograms and Dotplots 3
1.4 Summarizing Data: Mean, Variance and Range 8
1.4.1 Measures of Central Tendency 8
1.4.2 Measures of Variability 9
1.5 Comparing Data from Different Groups Using Summary
Statistics and Boxplots 10
1.6 Relationship between Two Variables, Scatterplots and
Correlation Coefficients 12
1.7 Types of Studies 15
1.8 Summary 17
Suggested Reading 18

2 Probability 19
2.1 Probability 19
2.2 Odds and Odds Ratios 21
2.3 Permutations and Combinations 22
2.4 Conditional Probabilities and Bayes’ Theorem 24
2.5 Random Variables, Probability Distributions and
Probability Density Functions 26
2.5.1 Random Variable 26
2.5.2 Probability Distribution 26
2.5.3 Probability Density Function 28
2.6 Expected Value and Moments 31
2.7 Moment-Generating Function 36
2.8 Summary 39
Suggested Reading 41

vii

K13590.indb 7 03/11/11 4:52 PM

 viii Contents

3 Estimation 43
3.1 Point Estimation 43
3.2 Sampling Distribution of the Mean and the Central
Limit Theorem 44
3.3 Estimation by the Method of Moments 46
3.4 Estimation by Maximum Likelihood 48
3.5 Choosing Between Estimators 50
3.6 Sampling Distributions: Student’s t, Chi-Square
and Fisher’s F 51
3.7 Interval Estimation, Confidence Intervals 54
3.8 Summary 57
Suggested Reading 58

4 Inference 59
4.1 Inference and Hypotheses 59
4.2 Significance Tests, Type I and Type II Errors, Power and
the z-Test 60
4.3 Power and Sample Size 64
4.4 Student’s t-Tests 66
4.5 The Chi-Square Goodness-of-Fit Test 72
4.6 Nonparametric Tests 75
4.7 Testing the Population Correlation Coefficient 76
4.8 Tests on Categorical Variables 78
4.9 The Bootstrap 84
4.10 Significance Tests and Confidence Intervals 86
4.11 Frequentist and Bayesian Inference 87
4.12 Summary 89
Suggested Reading 92

5 Analysis of Variance Models 93

5.1 One-Way Analysis of Variance 93
5.2 Factorial Analysis of Variance 96
5.3 Multiple Comparisons, a priori and post hoc Comparisons 102
5.4 Nonparametric Analysis of Variance 106
5.5 Summary 107
Suggested Reading 109

6 Linear Regression Models 111

6.1 Simple Linear Regression 111
6.2 Multiple Linear Regression 113
6.3 Selecting a Parsimonious Model 117
6.4 Regression Diagnostics 122

K13590.indb 8 03/11/11 4:52 PM

 Contents ix

6.5 Analysis of Variance as Regression 126

6.6 Summary 132
Suggested Reading 132

7 Logistic Regression and the Generalized Linear Model 133

7.1 Odds and Odds Ratios 133
7.2 Logistic Regression 135
7.3 Generalized Linear Model 141
7.4 Variance Function and Overdispersion 146
7.5 Diagnostics for GLMs 146
7.6 Summary 147
Suggested Reading 148

8 Survival Analysis 149

8.1 Survival Data and Censored Observations 149
8.2 Survivor Function, Log-Rank Test and Hazard Function 150
8.3 Proportional Hazards and Cox Regression 158
8.4 Diagnostics for Cox Regression 161
8.5 Summary 162
Suggested Reading 164

9 Longitudinal Data and Their Analysis 165

9.1 Longitudinal Data and Some Graphics 165
9.2 Summary Measure Analysis 170
9.3 Linear Mixed Effects Models 171
9.4 Missing Data in Longitudinal Studies 177
9.5 Summary 179
Suggested Reading 180

10 Multivariate Data and Their Analysis 181

10.1 Multivariate Data 181
10.2 Mean Vectors, Variances, Covariance and Correlation
Matrices 182
10.3 Two Multivariate Distributions: The Multinomial
Distribution and the Multivariate Normal Distribution 184
10.4 The Wishart Distribution 186
10.5 Principal Components Analysis 187
10.6 Summary 191
Suggested Reading 192

Index 193

K13590.indb 9 03/11/11 4:52 PM

This page intentionally left blank
 Preface

According to my Penguin English Dictionary, whistle-stop, used before a

noun, means ‘consisting of brief stops in several places’ and this whistle-stop
tour of statistics does just that, with the ten ‘stops’ being ten major areas
of statistics. In A Whistle-Stop Tour of Statistics quintessential accounts of
the topics which are the subject of each part are given with the summaries
at the end of each part collecting together the most important formulae,
etc. The book is intended as a quick source of reference and as an aide-
memoir for students taking A-level, undergraduate or postgraduate statistics
courses. The numerous examples included in the ‘tour’ may also be helpful
to instructors on such courses by providing their students with small data sets
with which to work. The book was partially suggested by the two excellent
‘Companions’ by A.C. Fischer-Cripps:

Fischer-Cripps, A.C. (2003) The Physics Companion, Taylor & Francis,

New York.
Fischer-Cripps, A.C. (2005) The Mathematics Companion, Taylor &
Francis, New York.

xi

K13590.indb 11 03/11/11 4:52 PM

This page intentionally left blank
 Some
Basics and
Describing
1
Data

1.1 PoPuLATIon, SAMPLES

AnD VARIABLES
Population: The set of all ‘individuals’ (often humans, but not always so) that
are of interest in a study; for example, all men born in Essex in 1944, all
women between the ages of 20 years and 60 years living in San Francisco, all
companies in Paris with fewer than 20 employees.
Sample: A subset of the ‘individuals’ in a population; for example, 1000 men
born in Essex in 1944, 100 women between the ages of 20 and 60 living in
San Francisco, 10 companies in Paris with fewer than 20 employees. Samples
may be taken in various ways but in most cases we assume random sampling,
meaning that each individual in the population has the same probability of
being sampled.
Data: Information collected on members of the sample by measuring, count-
ing or observing the value (or values) of some variable (variables) on each
sample member where a variable is any quantity that may vary from individual
to individual; for example, weight of each of the 1000 Essex men, number of
sexual partners for each of the 100 San Francisco women, annual turnover for
each of the 10 Paris companies.

K13590.indb 1 03/11/11 4:52 PM

 2 A Whistle-Stop Tour of Statistics

Nearly all statistical analysis is based on the principle that one collects
data on the members in the sample and uses this information to make some
inferences about the population of interest. Samples are needed because it is
rarely possible to study the whole population. The relation between sample
and population is subject to uncertainty and we use probability concepts to
quantify this uncertainty (see Chapters 2, 3 and 4).

1.2 TyPES of VARIABLES

Four types of variable may be distinguished:

Nominal (categorical) variables: Variables that allow classification

with respect to some property; examples are marital status, sex and
blood group. The categories of a nominal scale variable have no
logical order; numbers may be assigned to categories but merely as
convenient labels for these categories.
Ordinal variables: Variables that have one additional property over a
nominal scale variable, namely, a logical ordering of categories; now
the numbers assigned to categories indicate something about the
amount of a characteristic possessed but not about the differences
in the amount. Examples of such variables are ratings of anxiety and
depression and assessments of IQ.
Interval variables: Variables possessing a further property over ordi-
nal scales, and that is that equal differences on any part of the
scale reflect equal differences in the characteristic being measured.
The zero point for such scales does not represent the absence of
the characteristic the scale is used to measure. Examples of inter-
val scale variables are temperatures measured on the Celsius (C) or
Fahrenheit (F) scales.
Ratio variables: Variables that differ from interval scale variables in
having a true zero point that represents the absence of the character-
istic measured by the scale. Examples are temperature on the Kelvin
(K) scale and weight.

The different types of measurement scales may require the application of dif-
ferent types of statistical methods for valid conclusions to be made.

K13590.indb 2 03/11/11 4:52 PM

 1 • Some Basics and Describing Data 3

1.3 TABuLATInG AnD GRAPhInG

DATA: fREquEncy DISTRIBuTIonS,
hISToGRAMS AnD DoTPLoTS
The construction of informative tables is often one of the first steps in trying
to understand a data set.

ExamPlE 1.1
Eye colour collected on a sample of 22,361 children in Aberdeen,
Scotland. Eye colour is a categorical variable. How can we usefully tabu-
late the data?

Solution
Simply count the number of children in each category of the eye colour
variable:

EyE colour
BluE lIght MEdIuM dark
Count of 2978 6697 7511 5175
children
Percentages 13.3 29.9 33.6 23.1
NB: If only percentages are given, the size of the sample
on which they are based must also be quoted.

ExamPlE 1.2
The starting positions of a sample of 144 winners of eight-horse horse
races on circular tracks were recorded where starting position is num-
bered 1 to 8 with position 1 being closest to the rail on the inside of the
track. So the original data consisted of a series of 144 numbers from 1 to
8, i.e., 1,1,3,2,4,......4,3,8. What type of variable is starting position and
how can the data be tabulated to make them more transparent?

Solution
If the starting positions are equally spaced out from the fence, then start-
ing position is an interval variable but one that is discrete rather than

K13590.indb 3 03/11/11 4:52 PM

 4 A Whistle-Stop Tour of Statistics

continuous. If the starting positions are not equally spaced out from the
fence then the variable is ordinal only. To tabulate the data construct a
table giving the number of winners in each starting position.

startINg PosItIoN
1 2 3 4 5 6 7 8
Number 29 19 18 25 17 10 15 11
of wins

ExamPlE 1.3
The heights in millimetres of a sample of 169 men have been collected.
What is a useful way of tabulating them?

Solution
Here we can count the number (frequency) of men falling into each of
a number of intervals for height to give a frequency distribution table.

class
INtErval FrEquENcy
1550–1599 5
1600–1649 12
1650–1699 36
1700–1749 55
1750–1799 35
1800–1849 16
1850–1899 9
1900–1949 1

Tables can often be usefully represented by various graphics:

Bar chart: A graphical representation of data classified into a number

of categories. Equal-width rectangular bars are constructed over
each category with height equal to the observed frequency of the
category. See Figure 1.1 for an example. Bar charts are often used
but it is doubtful that they provide any advantage over the corre-
sponding table of frequencies.

K13590.indb 4 03/11/11 4:52 PM

 1 • Some Basics and Describing Data 5

30

25
Percentage (n = 22,361)

20

15

10

0
Blue Light Medium Dark
Eye Colour

fIGuRE 1.1 Bar chart of eye colour counts.

Pie chart: A widely used graphical technique for representing rela-

tive frequencies associated with the observed values of a categorical
variable. The chart consists of a circle divided into sectors whose
sizes are proportional to the quantities (often percentages) that they
represent. An example is given in Figure 1.2. Such graphics are pop-
ular in the media but have little advantage over the tabulated data,
particularly when the number of categories is small.
Dotplot: A graphic with horizontal line for each category of a cat-
egorical variable and a dot on each giving either the category fre-
quency or the value of some other numerical quantity associated
with the category. Examples are shown in Figures 1.3 and 1.4.
Dotplots are generally more effective displays than both bar charts
and pie charts.
Histogram: A graphical representation of a frequency distribution table
in which class frequencies are represented by the areas of rectangles
centred on the class interval. If the latter are all of equal length then
the heights of the rectangles are proportional to the observed class
frequencies. An example is shown in Figure 1.5.

K13590.indb 5 03/11/11 4:52 PM

 6 A Whistle-Stop Tour of Statistics

Light

Blue

Medium
Dark

fIGuRE 1.2 Pie chart of eye colour frequencies.

Starting position 8

Starting position 7

Starting position 6

Starting position 5

Starting position 4

Starting position 3

Starting position 2

Starting position 1

10 15 20 25
Frequency of Winners

fIGuRE 1.3 Dotplot of horse race winner frequency by starting position.

K13590.indb 6 03/11/11 4:52 PM

 1 • Some Basics and Describing Data 7

Crane drivers

Woodworkers

Leather

Electrical

Textile

Printing

Sales

Farmers

Clerical

Managers

Professional

50 60 70 80 90 100 110
SMR

fIGuRE 1.4 Dotplot of standardized mortality rates (SMR) for lung cancer in
several occupational groups.

60

50

40
Frequency

30

20

10

0
1550–1599 1600–1649 1650–1699 1700–1749 1750–1799 1800–1849 1850–1899 1900–1949

Height (mm)

fIGuRE 1.5 Histogram of heights of 169 men.

K13590.indb 7 03/11/11 4:52 PM

 8 A Whistle-Stop Tour of Statistics

1.4 SuMMARIzInG DATA: MEAn,

VARIAncE AnD RAnGE
More concise summaries of data than tables and graphs are obtained by calcu-
lating a small number of numerical characteristics of the data; such numerical
characteristics are termed statistics of the data or sample. Emphasis here will
be on statistics for interval or ratio scale variables.

1.4.1 Measures of central Tendency

mean: Given a sample of n values of a variable, x1, x2, x3, …, xn, the mean
value of the variable x– is

∑x i

x= i =1
n

If the variable values are summarized in a frequency distribution table with k

classes with midpoints x1, x2, …, xk and frequencies f1, f 2, …, f k then the mean
is calculated as

∑fx i i

x= i =1
k

∑f
i =1
i

median: For a sample of variable values that are skewed (lack of symme-
try in the frequency distribution of the data) the mean may not be a useful
measure of central tendency; in such cases the median may be more useful.
The median is simply the value in the ranked variable values that divides
the sample into two parts each with 50% of the variable values. When there
is an odd number of observations the median is the middle value; when the
number of observations is even the median is calculated as the average of
the two central values. For symmetric data the sample mean and median
will be approximately equal.
mode: The most frequently occurring value in a sample.

K13590.indb 8 03/11/11 4:52 PM

 1 • Some Basics and Describing Data 9

ExamPlE 1.4
What are the mean and the median for the following two data sets?

Set1: 3.1,5.2,2.1,6.3,7.8,4.1,4.0,8.8,0.6
Set2: 1.1,3.2,4.2,0.7,5.0,5.5,4.6,9.0,15.0,18.0

Solution
Set1: Mean = 4.67, median = 4.10
Set2: Mean = 6.63, median = 4.80

1.4.2 Measures of Variability

Variance: Given a sample of n values of a variable, x1, x2, …, xn, the variance
s2 is given by

s2 =
1
n ∑(x
i =1
i − x )2

(NB: The variance, s2, is often defined with a denominator of n − 1 rather

than n for reasons that will be made clear in Chapter 3.)
Standard deviation: The square root of the variance, gives a measure of vari-
ability in the same units as the original variable values.
Range: The difference between the largest and variable values in the sample.
Quick to calculate but the range is generally not a very useful measure of
variability.
inter-quartile range: The difference between the third and first quartiles of a
sample of variable values where quartiles are the values that divide the sample into
four sub-sets each containing an equal number of observations. More useful than
the range because it is less affected by extreme observations very different from
the bulk of the values in the sample; observations often referred to as outliers.

ExamPlE 1.5
What are the variance, standard deviation, range and inter-quartile
range for the following sample of suicide rates per million for 11 cities in
the United States?

K13590.indb 9 03/11/11 4:52 PM

 10 A Whistle-Stop Tour of Statistics

cIty suIcIdE ratE

New York 72
Los Angeles 224
Chicago 82
Philadelphia 92
Detroit 104
Boston 71
San Francisco 235
Washington 81
Pittsburgh 86
St. Louis 102
Cleveland 104

Solution
Variance = 3407.9, standard deviation = 58.4, range = 164, 1st
quartile = 81.5, 3rd quartile = 104, inter-quartile range = 22.5.
The extreme suicide rates for San Francisco and Los Angeles
(what’s so bad about this cities?) tend to distort all the measures of
variability except the inter-quartile range.

The quartiles of a data set divide the data into four parts with equal numbers
of observations. The quartiles are the 25th, 50th and 75th percentiles of the
data where the Pth percentile of a sample of n observations is that value of
the variable with rank (P/100)(1 + n); if this is not an integer, it is rounded to
the nearest half rank.

1.5 coMPARInG DATA fRoM

DIffEREnT GRouPS uSInG SuMMARy
STATISTIcS AnD BoxPLoTS
Many studies involve the comparison of a sample of values on some variable of
interest in different groups, for example, in men and women, in different coun-
tries, in different medical diagnostic groups. For example, Table 1.1 (known as
a contingency table) allows a comparison of the smoking habits of people in
different occupations.

K13590.indb 10 03/11/11 4:52 PM

 1 • Some Basics and Describing Data 11

TABLE 1.1 Cross-classification of smoking habits and occupation

sMokINg occuPatIoN (PErcENtagEs
charactErIstIc showN)
BluE total
daIly avEragE collar ProFEssIoNal othEr saMPlE sIzE
20+ 44.1 6.2 49.7 8951
10–19 42.7 5.8 51.5 2589
1–9 38.6 7.6 53.8 1572
Former smoker 30.9 11.0 58.1 8509
Never smoked 30.5 11.8 58.7 9694

The table demonstrates that there is a far smaller percentage of profes-

sional workers who are heavy smokers than for other types of worker.
Comparisons of grouped data can also be in terms of summary statistics
and graphics. For example, Table 1.2 gives the weights (in grammes) of 10 items
as measured on two kitchen scales and Table 1.3 gives the values of various sum-
mary statistics for the two weighing machines
Boxplot (box-and-whisker plot): A useful graphic for comparing data sets
which displays the distribution of the bulk of the observations and any extreme
values. The graphic is based on the five number summary of the data, i.e., the
minimum value, lower quartile, median, upper quartile and maximum value.
For the weights measured by two scales the boxplots are shown in Figure 1.6.
The plot clearly shows the similarity of most characteristics of the two samples
of weights.

TABLE 1.2 Weights in grammes

of 10 items as measured by two
kitchen scales
ItEM scalE a scalE B
1 300 320
2 190 190
3 80 90
4 20 50
5 200 220
6 550 550
7 400 410
8 610 600
9 740 760
10 1040 1080

K13590.indb 11 03/11/11 4:52 PM

 12 A Whistle-Stop Tour of Statistics

TABLE 1.3 Summary statistics for the weights provided by two kitchen scales
staNdard INtEr-quartIlE
MEaN dEvIatIoN raNgE raNgE
scale a 413 321.6 1020 402.5
scale B 427 324.7 1030 390.0

1000

800
Weights (grammes)

600

400

200

0
Scale A Scale B

fIGuRE 1.6 Boxplots for weights as measured by two kitchen scales.

1.6 RELATIonShIP BETwEEn Two

VARIABLES, ScATTERPLoTS AnD
coRRELATIon coEffIcIEnTS
univariate data: Data containing the values of a single variable for a sample
of individuals.
Bivariate data: Data with the values of two variables for each member of the sam-
ple. Assessing the relationship between the pair of variables becomes important.
Scatter diagram (scatterplot): Simply a graph in which the values of one
variable are plotted against those of the other; a very useful graphic for study-
ing the relationship between two variables.

K13590.indb 12 03/11/11 4:52 PM

 1 • Some Basics and Describing Data 13

ExamPlE 1.6
Construct a scatterplot of suicide rates for various cities in the United
States (see Example 1.5) and the percentage unemployed for the same
cities given below.

cIty (%uNEMPloyEd)
Ny la chI. PhIl. dEt. Bos. sF wash. PItts. st.l. clE.
3.0 4.7 3.0 3.2 3.8 2.5 4.8 2.7 4.4 3.1 3.5

Solution
A scatterplot of the two variables is shown in Figure 1.7. The plot shows
that there is some relationship between %unemployed and suicide rates
with higher values of the former being loosely associated with higher val-
ues of the latter; the suicide rates for SF and LA are very different from
the rates for the other cities, and their unemployment percentages are a
little higher than the percentages of the other cities.

SF
LA

200
Suicide Rate (per million)

150

100 St.L. Clev. Det.

Phil.
Pitts.
Wash. Chi.
Bos. NY

2.5 3.0 3.5 4.0 4.5

Percentage Unemployed

fIGuRE 1.7 Scatterplot of suicide rate vs. %unemployed for 11

cities in the United States.

K13590.indb 13 03/11/11 4:52 PM

 14 A Whistle-Stop Tour of Statistics

Sample correlation: The sample correlation is a measure of how closely the

points in a scatter plot of the sample values of two variables lie to a straight
line. The relationship between two variables can often be usefully quantified
by calculating a measure of their correlation, i.e., a measure of the strength and
the direction of the linear relationship between two random variables.
Pearson’s correlation coefficient (product moment correlation coefficient):
The most commonly used correlation coefficient; usually denoted by r and for
n pairs of variable values ( x1 , y1 ),( x 2 , y2 ),…,( x n , yn ) given by

∑ (x − x )( y − y )
i i

r= n
i =1
n

∑i =1
( xi − x )2 ∑ (y − y)
i =1
i
2

The coefficient r can take values between –1 and 1 with the numerical value
indicating the strength of the linear relationship and the sign indicating the
direction of the relationship. The plots in Figure 1.8 show scatter plots of sev-
eral data sets and the associated values of r. The plots in Figure 1.8 (g) and
(h) demonstrate that r is only useful for assessing the strength of the linear
relationship between two variables.
Spearman’s rho: A correlation coefficient that uses only the rankings of the
observations. If the ranked values of the two variables for a sample of individu-
als are ai and bi with di = ai – bi, then Spearman’s rho is given by

6 ∑di =1
2
i

rho = 1 −
n −n
3

For the suicide and %unemployed variables in 11 cities in the United States given
earlier, Pearson’s correlation coefficient is 0.81 and Spearman’s rho is 0.88.

K13590.indb 14 03/11/11 4:52 PM

 1 • Some Basics and Describing Data 15

r = –0.05 r = 0.48 r = 0.72 r = –0.83

6 16
10
14
5 0
8
12
Y 4 Y 6 Y Y
10
–5
3 4 8

2 6
2 –10

1 3 5 1 3 5 1 3 5 1 3 5
X X X X
(a) (b) (c) (d)

r=1 r = –1 r=0 r = 0.93

12 –1 35
30 200
10 –2
25
–3 100
8
Y Y Y 20 Y
6 –4 15 0

10
4 –5 –100
5
2 –6 0
1 3 5 1 3 5 –6 –2 2 6 –6 –2 2 6
X X X X
(e) (f ) (g) (h)

fIGuRE 1.8 Scatterplots and correlation coefficients for eight data sets.

1.7 TyPES of STuDIES

The four main types of study used in scientific research are surveys, experi-
ments, observational studies and quasi-experiments.
• Surveys: Asking questions has been found to be a remarkably effi-
cient way to obtain information from and about people whether via
written questionnaires, in-person interviews, telephone conversa-
tions or the Internet. But non-response can cause problems and the

K13590.indb 15 03/11/11 4:52 PM

 16 A Whistle-Stop Tour of Statistics

questions asked need to elicit accurate responses. One of the most

famous surveys of the 20th century was that conducted by Alfred
Charles Kinsey into human sexual behaviour.
• Experiments: The essential feature of an experiment is the large
degree of control in the hands of the experimenter which can, in
general, allow unambiguous conclusions to be made about some
treatment or intervention of interest. The clinical trial is the most
common form of experiment with groups of participants formed by
randomization.
• Observational: Studies in which the researcher does not have the
same amount of control as in an experiment; in particular, people
cannot be assigned at random to receive the procedures or treat-
ments whose effect it is desired to discover, instead members of
naturally occurring groups are studied. An example for which an
observational study rather than an experiment is necessary is in an
investigation of the relationship between smoking and blood pres-
sure; people cannot be randomly assigned to be ‘smokers’ or ‘non-
smokers’ but the blood pressure of people who smoke and those
who do not can be recorded and compared. Observational studies
are less powerful than experiments because there is more ambiguity
in the interpretation of results. There are various types of obser-
vational studies, for example, case-control studies, cross-sectional
studies and cohort (prospective) studies; the essential features of
each type are illustrated in Figure 1.9
• Quasi-experiments: Such studies resemble experiments but are weak
on some of the characteristics; in particular, the ability to manipu-
late the groups to be compared is not under the investigator’s con-
trol. But interventions can be applied to naturally occurring groups.
Cross-sectional
study (survey) Disease events
Case- Risk factors Cohort
control
study
study

Time

Disease All Baseline

(yes/no) data risk factors Follow-up

fIGuRE 1.9 Schematic comparison of the three major types of observational

studies.

K13590.indb 16 03/11/11 4:52 PM

 1 • Some Basics and Describing Data 17

An example is provided by a study investigating the effectiveness of

different methods of teaching mathematics, in which the different
methods are given to the members of different classes in the school.

1.8 SuMMARy
Population: The complete set of individuals of interest.
Sample: A sub-set of the individuals that make up a population.
Variable types: Nominal/categorical, ordinal, interval, ratio.
Bar chart: A graphic for displaying the frequencies of the categories
of a categorical variable.
Pie chart: A graphic for displaying the relative frequencies associated
with the observed values of a categorical variable.
Dotplot: A graphic for displaying quantitative data which are labelled
in some way.
Histogram: A graphical representation of a frequency distribution
table in which class frequencies are represented by the areas of rect-
angles centred on the class interval.

∑x i

Mean: x = i =1
n

∑ (x − x )
1
Variance: s 2 = i
2
n i =1

Range: xmax – xmin.

inter-quartile range: Difference between the first and third quartiles
of a sample.
Percentile: The Pth percentile of a sample of n observations is that
value of the variable with rank (P/100)(1 + n); if this is not an integer,
it is rounded to the nearest half rank.
Boxplot: A useful graphic for comparing the characteristics of samples
of variable values from different populations.
Scatterplot: A graphic displaying how two variables are related.

K13590.indb 17 03/11/11 4:52 PM

 18 A Whistle-Stop Tour of Statistics

∑ (x − x )( y − y )
i i

Pearson’s correlation coefficient: r = n

i =1
n

∑i =1
( xi − x ) 2 ∑ (y − y)
i =1
i
2

Spearman’s rho: Pearson’s correlation coefficient based on the ranks

of the two variables.
types of study: Experiments, e.g., clinical trials; observational, e.g.,
case-control studies, quasi-experimental.

SuGGESTED READInG
Altman, DG (1991) Practical Statistics for Medical Research, Chapman & Hall, London.
Daly, F, Hand, DJ, Jones, MC, Lunn, AD and McConway, KJ (1995) Statistics, Addison
Wesley, Wokingham.
Everitt, BS and Palmer, CR (2010) Encyclopaedic Companion to Medical Statistics,
Wiley, Chichester.
Freedman, D, Pisari, R and Purves, R (2007), Statistics, 4th Edition, W.H. Norton and
Co., New York.

K13590.indb 18 03/11/11 4:52 PM

 Probability

2.1 PRoBABILITy
2
Events: The results of experiments or observations. Compound and simple
events need to be distinguished. For example, consider the age of a person,
x. Each particular value of x represents a simple event, whereas the statement
that a person is in her fifties describes the compound event that x lies between
50 and 60. Every compound event can be decomposed into simple events; a
compound event is an aggregate of certain simple events.
Probability: A measure of that event’s uncertainty. By convention probabili-
ties are measured on a scale from zero to one, with zero corresponding to an
event that is impossible and one to an event that is certain.
Classical definition of probability: If there are n equally likely possibilities
of which s result in the event of interest (often referred to as a success) then the
probability of the event, p, is given by

p = s/n

with the probability that the event does not happen (often denoted a failure)
being 1 − p.

ExamPlE 2.1
What is the probability of a year, which is not a leap year, having 53 Sundays?

Solution
A non-leap year of 365 days consists of 52 complete weeks and 1 day
over. This odd day may be any one of the 7 days of the week and there is
nothing to make one more likely than another. Only one will lead to the
result that the year will have 53 Sundays; consequently, the probability
that the year has 53 Sundays is simply 1/7.

19

K13590.indb 19 03/11/11 4:52 PM

 20 A Whistle-Stop Tour of Statistics

alternative definition of probability: The limit of the relative frequency of a

success in a large number of trials, n, as n tends to infinity; explicitly,

s
p = Lim
n→∞ n

where s is the number of successes in the n trials.

ExamPlE 2.2
Suppose 1,000,000 children are born in a year, with 501,000 of them
being boys. The relative frequency of boys is 501,000/1,000,000 = 0.501.
Here the denominator is large and we can justifiably claim that the prob-
ability of a male child is 0.501.
Now suppose that in a small village somewhere, 10 children are born
in a year of which 7 are boys. The relative frequency of male babies is
7/10 = 0.7 but claiming that this is also the probability of a male baby is
now not justified because the number of babies is too small.

Subjective probability: The classical and relative frequency definitions of

probabilities cannot be used in all situations where probability statements are
required; for example, the question, ‘what is the probability that life exists
on a planet circling the nearest star, Proxima Centauri?’ The classical defini-
tion cannot provide an answer and there are no relative frequency data avail-
able that might help. In such circumstances we can only express a subjective
opinion about the required probability, i.e., assign a subjective probability. We
might use experience, intuition or a hunch to arrive at a value but the assigned
probability will be subjective and different people can be expected to assign
different probabilities to the same event. Subjective probabilities express a
person’s degree of belief in the event and are often referred to as personal
probabilities.
addition rule for probabilities: Often the events for which we require
probabilities are made up of a number of simpler events and so we need to
consider how to combine the probabilities of these simpler events to arrive
at the probability of the more complex event. For mutually exclusive events
(events that cannot occur together), E1, E2,…, Er, we have the following addi-
tion rule:

Pr(E1 or E2 or … or Er ) = Pr(E1) + Pr(E2) + …Pr(Er)

K13590.indb 20 03/11/11 4:52 PM

 2 • Probability 21

ExamPlE 2.3
What is the probability of rolling a fair die (one in which all 6 faces are
equally likely) and getting an even number?

Solution
Pr(2 or 4 or 6) = Pr(2)+Pr(4) + Pr(6) = 1/6 + 1/6 + 1/6 = 1/2

(NB: When events are not mutually exclusive the addition rule no longer holds;
for example, for two events E1 and E2 that can occur together the probability
that either E1 or E2 occurs is given by

Pr(E1 or E2) = Pr(E1) + Pr(E2) – Pr(E1 and E2).)

multiplication rule for probabilities: Now consider events that can happen
together but which are independent, i.e., the occurrence or otherwise of any
event does not affect the occurrence or otherwise of any other event. Here the
probability that all events E1, E2,…,Er occur simultaneously is found from the
following multiplication rule:

Pr( E1 and E2 and …and Er ) = Pr( E1 )Pr( E2 )…Pr( Er )

ExamPlE 2.4
What is the probability of getting a triple 6 when rolling three fair dice?

Solution
Pr(triple 6) = Pr(6 with die 1)Pr(6 with die 2)Pr(6 with die 3) = 1/6 × 1/6
× 1/6 = 1/216

2.2 oDDS AnD oDDS RATIoS

odds and probability: Gamblers (and some statisticians) prefer to quantify
their uncertainty about an event in terms of odds rather than probabilities

K13590.indb 21 03/11/11 4:52 PM

 22 A Whistle-Stop Tour of Statistics

although the two are actually completely synonymous. The exact relationship
between odds and probability is as follows:

• An event with odds of ‘F to 1 in favour (odds-on)’ has probability

F/(F + 1),
• An event with odds of ‘A to 1 against’ has probability 1/(A + 1),
• An event with probability P implies that the odds in favour are
P/ (1 – P) to 1, whereas the odds against are 1/P – 1.

ExamPlE 2.5
What are the odds for an event with probability 1/5?

Solution
The odds against are 1/0.2 – 1 = 4 to 1 against. This simply expresses
the fact that the probability of the event not occurring is four times that
of it occurring.

ExamPlE 2.6
What are the odds for an event with probability 4/5?

Solution
The odds in favour are 0.8/0.2, i.e., 4 to 1; here the probability of the event
occurring is four times that of it not occurring.

2.3 PERMuTATIonS AnD

coMBInATIonS
Permutations: If r objects are selected from a set of n total objects (and not
replaced after each object is selected) any particular ordered arrangement of
the r objects is called a permutation. The total number of possible permuta-
tions of r objects selected from n objects is denoted npr and given by

n!
n pr =
(n − r )!

where n! (read n factorial) is given by n! = n(n–1)...3 × 2 × 1 with 0! defined as one.

K13590.indb 22 03/11/11 4:52 PM

 2 • Probability 23

The number of permutations of all the objects taken together, npn, is n!

Combinations: If r objects are selected from n total objects, and if the order of
the r objects is not important, then each possible selection of r objects is called
a combination, denoted by ncr and given by
n!
n cr =
r !(n − r )!
(ncr is often called a binomial coefficient.)
The numbers of permutations and combinations of r objects in a total n
objects are often useful when assigning probabilities to events.

ExamPlE 2.7
The four letters s, e, n and t are placed in a row at random. What is the
chance of their standing in such order to form an English word?

Solution
The four letters can be arranged in 4! = 24 different permutations, all of
which are equally likely. Only four of these arrangements, sent, nest, nets,
tens, produce an English word so the required probability is 4/24 = 1/6.

ExamPlE 2.8
A bag contains five white and four black balls. If balls are drawn out of
the bag one by one, what is the probability that the first will be white, the
second black and so on alternately?

Solution
Assuming white balls are all alike and black balls are all alike, the number
of possible combinations of ways in which the nine balls can be drawn from
the bag is 9!/(4! × 5!) = 126. The balls are equally likely to be drawn in any of
these ways; only 1 of the 126 possibilities is the alternate order required, so
the probability of white, followed by black, followed by white, etc. is 1/126.

ExamPlE 2.9
Four cards are drawn from a pack of 52 cards. What is the probability that
there will be one card of each suit?

K13590.indb 23 03/11/11 4:52 PM

 24 A Whistle-Stop Tour of Statistics

Solution
Four cards can be selected from the pack in 52!/(4! × 48!) = 270,725 ways.
But four cards can be selected to be one of each suit in only 13 × 13 × 13 ×
13 = 28,561 ways. So the required probability is 28,651/270,725, which is
just a little over 1/10.

2.4 conDITIonAL PRoBABILITIES

AnD BAyES’ ThEoREM
Conditional probability: The probability that an event A occurs given the
outcome of some other event B on which A is dependent; usually written as
Pr(A|B). It is not, of course, necessary for Pr(A|B) to be equal to Pr(B|A). For
example, Pr(spots|suffering from measles) is likely to be very different from
Pr(suffering from measles|spots). If Pr(A|B) = Pr(A) then the events A and B are
independent. The probability that both of two dependent events A and B occur is
Pr(A and B) = Pr(A) × Pr(B|A)

ExamPlE 2.10
When rolling two dice what is the probability that the number on the die
rolled second is larger than the number on the die rolled first?
Solution
FIrst dIE sEcoNd dIE hIghEr IF ProBaBIlIty that
scorE rEsult Is sEcoNd dIE hIghEr
1 2 or 3 or 4 or 5 or 6 1/6 + 1/6 + 1/6 + 1/6 + 1/6
2 3 or 4 or 5 or 6 1/6 + 1/6 + 1/6 + 1/6
3 4 or 5 or 6 1/6 + 1/6 + 1/6
4 5 or 6 1/6 + 1/6
5 6 1/6
6 Cannot have a higher score 0
Pr(first die shows a 1 and second die is higher) = 1/6 × 5/6 = 5/26. Now
find corresponding probabilities for different first die scores and then add
to give the result as 15/36.

K13590.indb 24 03/11/11 4:52 PM

 2 • Probability 25

Bayes’s theorem: A theorem relating conditional probabilities:

Pr( B|A)Pr( A)
Pr( A|B) =
Pr( B)

ExamPlE 2.11
A cab was involved in a hit-and-run accident at night. Two cab compa-
nies, the Green and the Blue, operate in the city. The following facts are
known:

• 85% of the cabs in the city are Green and 15% are Blue.
• A witness identified the cab as Blue. The court tested the reliability of
the witness under the same circumstances that existed on the night of
the accident and concluded that the witness correctly identified each
one of the two colours 80% of the time and failed 20% of the time.

What is the probability that the cab involved in the accident was actually
Blue?

Solution
Letting A be the event that a cab is Blue and B the event that the witness
says he sees a Blue cab then, equating percentages to probabilities, what
we know is

Pr(A) = 0.15, Pr(B|A) = 0.80.

The unconditional probability of the witness saying that he saw a Blue cab is

Pr(B) = 0.15 × 0.80 + 0.85 × 0.20 = 0.29

Applying Bayes’ theorem we can now find what we require, namely, the
probability that the cab is Blue given that the witness says it is Blue, Pr(A|B),

Pr( B|A) Pr( A) 0.80 × 0.15

Pr( A|B) = = = 0.411
Pr( B) 0.29

So the probability that the cab is actually Blue given that the witness iden-
tifies the cab as Blue is less than a half. Despite the evidence the eyewit-
ness offers, the hit-and-run cab is more likely to be Green than Blue! The
report of the eyewitness has, however, increased the probability that the
offending cab is Blue from its value of 0.15 in the absence of any evi-
dence (known as the prior probability) to 0.41 (known as the posterior
probability).

K13590.indb 25 03/11/11 4:52 PM

 26 A Whistle-Stop Tour of Statistics

Conditional probabilities are of particular importance in medical diagno-

ses, i.e., using medical tests to diagnose the condition of a patient. The sensitiv-
ity and specificity of such a test are defined as
Sensitivity = Pr(Test positive|Patient has the disease)
Specificity = Pr(Test negative|Patient does not have the disease)

2.5 RAnDoM VARIABLES,

PRoBABILITy DISTRIBuTIonS AnD
PRoBABILITy DEnSITy funcTIonS

2.5.1 Random Variable

A rule that assigns a unique number to each outcome of an experiment; these
numbers are called the values of the random variable. The value of a random
variable will vary from trial to trial as the experiment is repeated. Random
variables can be either discrete, meaning that they can only take specific
numerical values, or continuous, meaning that they can take any value within
a continuous range or interval. An example of a discrete random variable is
the number of heads obtained when a coin is tossed 10 times; here the ran-
dom variable can take values 0,1,2,3,...,10. An example of a continuous random
variable is age at death in humans. Random variables are usually denoted by
capital letters, for example, X, with particular outcomes being represented by
the corresponding lower case letter, here, x.

2.5.2 Probability Distribution

If X represents a discrete random variable which can take on m values,
x1, x2,…, xm , the probability distribution (sometimes probability function)
of X is a function f where f(x i) = Pr(X = x i), i = 1,2…m, 0 ≤ f(x) ≤ 1 and
m
∑ i=1 f(xi) = 1.
(NB: f is used below to denote a variety of different probability functions.)

ExamPlE 2.12
What is the probability distribution for the number of heads obtained in
n independent tosses of a fair coin?

K13590.indb 26 03/11/11 4:52 PM

 2 • Probability 27

Solution
The random variable, X, here takes the values 0,1,2,...,n. The probability
of a head on any toss is ½ as is the probability of a tail. The probability
of obtaining r heads in the n tosses is found by multiplying the prob-
ability of r heads and n − r tails, namely, (0.5)r × (0.5)(n−r) by the number
of combinations of r heads in n tosses, ncr, to give the required probability
function as

Pr( X = r ) = f (r ) = n cr × (0.5)r × (0.5)( n −r )

n!
= (0.5)n , r = 0,1, 2,… n
r !(n − r )!

Binomial distribution: The result in the example above can be general-

ized to a situation where there are n independent trials with the probability
of a success on each trial being p, to give what is known as the binomial
distribution:

n!
f (r ) = pr (1 − p)( n−r ) , r = 0,1, 2,…n
r !(n − r )!

Poisson distribution: Another important probability distribution for a discrete

random variable, X, which arises as a limiting form of the binomial distribu-
tion as n → ∞ and p → 0 with λ = np is the Poisson distribution:

λ xe− λ
Pr( X = x ) = f ( x ) = , x = 0,1, 2…, ∞
x!

The Poisson distribution is also important in its own right for assigning
probabilities to rare events of some sort occurring randomly in time or
space.

ExamPlE 2.13
How do the probabilities given by the Poisson distribution compare with
those from a binomial distribution with n = 60 and p = 0.05?

K13590.indb 27 03/11/11 4:52 PM

Exploring the Variety of Random
Documents with Different Content
 Yünnan is surrounded by pretty places for excursions, so we set
off one fine morning in our chairs to visit a metal temple about five
miles distant, called “Gin Tien”—namely, “Golden Temple.” As we
passed through the sweet-scented bean-fields we saw many children
enjoying swings, a sight we had not met before in any other part of
the country. Soon we reached the foot of the hills and ascended
through woods filled with a delicious aromatic scent; but the trees
were quite unfamiliar to us, and whether it was from them or the
brushwood that the scent came we could not determine. The temple
was beautifully situated on the hillside, and the courts rose one
above the other, with long flights of steps leading from one to
another. At the top of the first flight was an archway surmounted by
a temple containing a small wooden “god of literature.” The design
on the cover of “The Face of China” is the god of literature: in one
hand he holds the brush and in the other a tablet, and he stands on
one foot on a fish. Doolittle explains it thus: “There are two stars
which the Chinese profess to have discovered to have the
supervision of the affairs of this world relating to ‘literature and the
pencil.’ One of these, Kue Sing, is said to be the fifteenth star of the
twenty-eighth constellation, answering to parts of Andromeda and
Pisces.”
A miscellaneous collection of gods lined the sides of the court in
open corridors, but they were much neglected and in a ruinous-
looking condition. At the top of the steps was a terrace on which
was a fine camellia tree in full bloom, and other shrubs, while
chattering groups of white-and-black starlings lent animation to the
scene. Another flight of steps led to a court in which was the fine
copper temple, painted black and gold, standing on a platform made
of the celebrated Tali marble. A metal flag and bells that tinkled in
the wind hung from an adjoining flagstaff, and another camellia tree
was a sheet of pink blossom standing in glorious contrast behind the
temple close to a gateway. This gateway led into another courtyard,
where there were rooms which could be hired by any one who
wishes to spend a few days there. Europeans often make use of
these hill temples in the hot weather, even inhabiting sometimes the
same rooms as the gods, of whom a new use is made as pegs on
which to hang clothes in lieu of wardrobes—a proceeding which in
nowise shocks the Chinese worshippers.
We were by no means the only visitors to the temples. There
was quite an array of chairs waiting in the courtyards. Some of the
people were gambling, others having their midday meal, others lying
on couches smoking opium or admiring the view; but of worship
there was no sign whatever. While I sketched some women came to
look on, and had a little conversation (strictly limited, owing to my
ignorance). They carried beautiful orchids which they said they had
gathered in the mountains. In fact, Gin Tien is to Yünnan Fu what
Richmond is to London.
Another interesting excursion which we made was to the Rock
Temples above the lake of Yünnan. There is a canal, about two miles
long, leading directly from the city to the lake, and our host sent to
hire a boat the day previous to our excursion. The arrangement
made was that we were to have a crew of four men, in order to
convey us as rapidly as possible to our destination; but when we got
on board we found that our four men were represented by a woman
and her three boys, aged approximately sixteen, ten, and three. We
remonstrated, but it was so comic that we could do nothing but
laugh; and finally she hired a man to come and row, paying him
about twopence a day, whereas we paid her five shillings. Arrived at
the farther side, we had a steep climb through pine woods to the
temples, which are impartially Buddhist and Taoist. They must have
been hewn out of the cliffs with an immense amount of labour and
cost, for the approach to the upper ones was through winding
galleries cut in the solid rock. The gods themselves are in shrines cut
in the rock, and at the top of all is a little temple dedicated to the
“god of literature,” which was also carved out of stone; and there
were other gods carved above the entrance. From the little platform
in front of it there is a marvellous view of the lake and plain
stretched far below, where fishing-boats looked like insects, and over
which floated the shadows of the clouds.
 On our way down our servant had prepared tea at a Buddhist
temple, where we sat on stools (on a platform) at a low table. At an
adjoining table there was a large family party of men and boys also
having a meal, the ladies and girls of the family taking theirs in an
inner room. We could not help admiring the charming sets of
baskets in which they had brought their provisions, and we found
them very friendly and talkative. They had many questions to ask of
us, and informed us that they were jewellers in the city, finally
suggesting that we should all go home together! We felt that this
would be far too slow a process when we saw the ladies with their
tiny feet laboriously toiling downwards, with the help of walking-
sticks to steady them. So we made our excuses and hastened back,
as the sun was already getting low.
We dedicated our last day at Yünnan to shopping, for the shops
were beginning to reopen after their long inaction for the New Year.
This place was in former times a happy hunting-ground for bronzes,
but there are not many to be had now, and none of any value, while
all the prices have gone up, many of them a hundred per cent.
Copper work is the special industry of Yünnan, though all the copper
is supposed to go direct to Peking. There are two families who for
many generations have had the monopoly of making beautiful little
copper boxes inlaid with silver. The work is very fine, and some of
the designs are particularly attractive. Skin boxes are a speciality of
this place, and we found it necessary to get some in which to carry
our purchases; we also added a coolie to our party, as the loads
carried here are not allowed to exceed eighty pounds per man. If we
could fly, how quickly should we reach Bhamo!—only 360 miles
through the air, instead of 967 miles by the road, with a total ascent
of 26,000 feet. This is a computation in Hosie’s book, “Three Years in
Western China,” but I think the distances are decidedly over-
estimated. When we left Yünnan Fu our party numbered twenty-four,
and our chairs looked much more dignified than on arrival. The poles
were all carefully bandaged with bright-blue cotton like a mandarin’s
chair, because of the winds, as our head coolie informed us that
otherwise the poles were apt to crack. We were told to expect high
winds all the rest of our journey through this province, for they are
prevalent at this time of year. The prospect sounded discouraging,
for the sun was hot, and we were obliged to wear large hoods, as
the sun and the wind together had nearly skinned our faces.
However, like all our previous information about the journey, the
difficulties proved much less serious than we expected. In fact, so
far we had had nothing to complain of beyond the inevitable
disagreeables one encounters on travelling away from the beaten
track. As we left the city we noticed a curious mingling of the past
and present at the city gate: on the one side a dismantled cannon
made by, or under the direction of, the French Jesuit Fathers, and
stamped with the Christian symbol; on the other side of the gate, a
notice-board warning passengers to keep to the right side of the
road.
Leaving the city, we soon reached the mountains, and day by
day skirted the upper part of them; sometimes plunging down deep
into the valleys, especially for our resting-places at night. The people
seem a sturdy, solid race, but through the greater part of the
province which we have traversed, and especially round the capital,
they are greatly disfigured by goitre. Every day we see scores of
people (even quite young children) suffering from this disease. The
women do a large share of all the hard work, carrying heavy loads,
despite their small feet; the loads are fastened on by broad bands
passed round the forehead, like those of Newhaven fishwives. These
bands are frequently run through holes in a big wooden collar worn
both by men and women. Some of these collars have pretty little bits
of carving on them.
On the roads we met innumerable droves of pack-animals,
mostly laden with blocks of salt. The pack-men have special inns
where they put up, which are nothing more than stables, and scores
of animals can be accommodated in them. Despite the badness of
the roads and the rough way in which they are hustled along, we
have not seen a single beast with broken knees. They are allowed to
rest free from loads or saddles at midday, and to roll in the dust at
pleasure. The loads are fastened on to a framework which fits into
the saddle and so avoids the necessity of being adjusted on the
animal itself. There are regular camping grounds for the pack-
animals all along the road, and they seem the best tempered beasts
imaginable. The leaders usually wear bright red rosettes on their
heads, often with mirrors in them, and also the Government loads
have brilliant flags attached, which give them a picturesque look.
Some of them wear the long tails of the Amherst pheasant fastened
between their ears, and look as proud of themselves as a
fashionable London lady with the huge plumes now in fashion. Some
of them wear bells, which are necessary so as to herald their
approach on these narrow, winding, and precipitous highways.
 FELLOW-TRAVELLERS

Every day we were more enchanted with the beauty of the

country and the delights of spring. The banks are carpeted with
primulas, and the hill slopes bright with rose-coloured camellias,
scarlet azalea, white and crimson rhododendrons, yellow jessamine,
clematis, begonias, and numberless flowering shrubs, many of which
we have never seen before. This is the part of the world from which
the majority of our flowering shrubs have originally come. It would
be a paradise for botanists, and makes one long for more knowledge
of many subjects, so as to be able to enjoy the journey still more
and profit by it more thoroughly.
The second day after leaving Yünnan Fu, there was an
earthquake, and it is a mystery why the front of our room did not
entirely collapse. The inns here are really superior, but on this
occasion the outer wall of our room happened to be constructed in
sections at all sorts of angles, none of them what they were meant
to be, and with extensive gaps between. It was quite impossible to
shut the doors, and there was no pretence of a fastening, so we had
put up a curtain in order to obtain a small measure of privacy.
Happily, the people did not seem so inquisitive as they used to be
when I was travelling in China fifteen years ago. In fact, we rarely
see eyes peering through holes in the paper windows. Glass
windows are still unknown in the inns, except in an occasional one in
Shantung.
At Lu Feng Hsien we had an amusing experience. As we were
resting after our evening meal there was suddenly a great noise of
drums, and we were told that the dragon lantern was in the street.
It turned out to be a sort of entertainment given by a cash shop
next door, and not only was there a very bedraggled-looking dragon
about twenty feet long, but also fish lanterns and sundry fireworks.
The men carrying the sections of the dragon leapt about like demons
as the shower of “golden rain” (fireworks) was turned full upon
them, and the dragon withed with unwonted energy. The drums
never ceased for a moment, so that it was rather a relief when the
show came to an end by the exhaustion of the internal illumination
of the dragon. We were stopping at an inn just outside the city wall,
and when we left the next morning we crossed a fine suspension-
bridge with an imposing archway at each end of it. The chief
magistrate of the district happened to be travelling on the same road
with us, and sent word ahead that we were to be accommodated in
a charming inn that day, having invited us to stop at his Yamen in
the city the previous night. When magistrates are travelling they
always send to engage an inn beforehand, and a little official flag is
then hung outside to show that the inn is full. Mr. Ku suggested that
we should go in for an official flag, but we feared lest complications
should arise.

TOMB OF A PHILOSOPHER AMONG RICE-FIELDS

The fresh New Year mottoes put up on the doorposts of our

room stated encouragingly that “all cultured people inhabit this
room,” and “the courtyard is full of chairs and carts,” but, true to
Chinese incongruity, our coolies filled one of the rooms and the pigs
occupied the background!
Day after day new flowers and birds appear on our pathway—
white camellias, daphne, dog-roses, a flight of brilliant green parrots,
long-tailed tits, seagulls; though what they are doing out here we
cannot imagine. At one village where we spent a night the
magistrate sent word that the people were in a somewhat disturbed
condition, so he would send a special watch to guard us. We strolled
out into the fields to try and get a sketch of the large flocks of
cranes feeding in the rice and bean fields, but they seemed
disturbed and would not let us get anywhere near them, flying away
screeching loudly. As we got back to the village everything looked as
peaceful as possible, and the guard had duly arrived. They seemed
to think it part of their duty to share our room, one settling down to
a comfortable smoke, the other helping to shell the beans for our
evening meal. When we thought they had sufficiently studied us and
our surroundings we invited them to go outside, and they soon had
a cheerful fire blazing in the courtyard, where they remained all
night. We got infinite amusement out of the naïve ways of our
coolies and the soldiers. A heavy storm of rain, for instance, came
on while we were halting at a village, and immediately one of my
carriers came and sat upon the ground beside my chair in order to
share the benefit of my umbrella. He had not the slightest idea, of
course, that I might not wish for his close neighbourhood, for fear of
participating in more than the shelter of an umbrella.
One thing seems strange as one travels day by day from one
end of this great empire to the other, and that is the utter absence
of any landed gentry; never a country seat or any house larger than
a farmer’s, and never a garden of any kind for the cultivation of
flowers except within the cities; cottage gardens are unknown here.
Pots in a courtyard show a certain love of flowers, and the poorest
coolie will stop to gather a handful of camellias to decorate his load,
or a flower to stick behind his ear. Rich people all love to live in big
towns, where they are close to their associates.
One day the head coolie came in with our suit-cases in a state of
great agitation, dragging with him a frightened-looking creature
whose horse he said had pushed the luggage into the stream. To our
dismay, the luggage was dripping with water, and the culprit had
been hauled along to see the extent of the damage. Our usual good
luck, however, had followed us; though a new silk dressing-gown
was soaked with mud and water, my sketches (next to it) had only
mud traces on their backs, and nothing else was hurt. The man said
he was very sorry, and evidently expected we should charge
damages. He protested that he was only a poor farmer and had no
means of payment. No doubt the head coolie, who is responsible for
any loss and is bound to make good any breakage caused by the
carelessness of the carriers, would have extorted damages from him,
but as we did not, he told him to kotow; our servants sternly
repeated the command, and an interested crowd of spectators
watching the show added their injunctions, so that when the man
grovelled in the dust and knocked his forehead on the ground, we
were sorely tempted to laugh. The tragi-comic effect was irresistibly
funny. It was necessary, however, to impress our men with the
heinousness of the offence, lest our curios should come into greater
danger. A diversion was caused by the entrance of several coolies
begging me to look at their sprains, gatherings, &c., so we dismissed
the poor farmer and set to work with our out-patient department. At
this stage of the journey the coolies were very apt to give out a
little, as the strain began to tell upon them. The weather grew
warmer daily, and the crops seemed almost to grow visibly before
our eyes. Fields of mustard in the plains were dazzlingly yellow and
sweet-scented, and the poppies and wild roses were coming into
bloom. We were struck with the beauty of various kinds of wild vines
and the enormous leaves of Senecio wilsonianus. There are a great
many varieties of these vines and of the roses, and on the eastern
side of Yünnan we found particularly sweet-scented white banksia
roses. Mr. Wilson, who spent a considerable time in studying the
flora of China, discovered no fewer than 2000 new varieties of
plants, and Messrs. Veitch of Chelsea have a most interesting
collection of the plants which he brought back. The fact that he was
able to bring back over 5000 specimens seems almost incredible to
any one who knows the difficulties of transport. But perhaps the
most striking of all the flowers that we met growing profusely in this
region was the Jasminum primulinum, a large, brilliant jasmine of
which there were the most magnificent hedges. If only we could
have stayed a little longer we should have been able to see far more
of the shrubs in blossom, as everywhere we noticed they were full of
promise.
 VILLAGE SCREEN WALL

CARRYING CHAIR

The architecture is more ornate in this district than in the centre

of China, many of the houses in the villages being elaborately
decorated, and having beautiful wooden carved screens and
balustrades in front of them. Some of the shops had most knowing
little beasts carved at the end of the counter, and the signboards,
which were formed into Chinese lanterns, were extremely
picturesque. At the entrance of one of the villages there was the
most grotesque design on the large screen which is frequently to be
seen facing the entrance gate of villages. This particular one had a
gigantic creature (a cross between a cat and a toad) hanging on to a
fleeing mouse, whose hind leg had been stretched out to an
abnormal length in its endeavour to escape. Every village of any size
has such a screen (called in Chinese, “shelter”), standing about
fifteen feet high and twenty-five feet in length. Some of the designs
on them are really artistic, but as a rule their best quality is the
spirited action of the various beasts which are represented, of which
the one I have sketched (from memory) is a good example. The
chair below it shows what we travelled in for some 1200 miles; the
most comfortable way of seeing the country that can be imagined.
Under the seat we carried all our small impedimenta, while the
coolies made use of the back for theirs.
 CHAPTER XXI
Tali Fu

T
HE first glimpse of the Tali lake with snow-capped mountains
running along its western bank is very beautiful. It is a typical
Chinese landscape, namely, a “mountain and water picture”;
but at this time of year (the beginning of March) the winds are so
high that the distant range of snow mountains to the north is
invisible, and during the five days that we spent in the
neighbourhood we did not once catch a glimpse of them. At the
south of the lake is a large and important village, Siakwan, much
more important in some respects than the prefectural town of Tali Fu
itself, as it lies on the high-road from Yünnan Fu to Teng Yueh. All
the trade caravans pass directly through Siakwan without branching
off to Tali. At the markets, which are frequently held, there may be
seen various aboriginal tribes such as the Miaos, the Lolos, the Ming
Chiaos, and the Shans. The place is noted for being one of the
windiest spots to be found anywhere, and we found it unpleasant
both coming and going to Tali. At Tali the wind was most riotous at
nights, and came in great gusts, dismantling the walls of our rooms
and covering my bed with a little shower of photographs and texts,
and nearly blowing us out of bed. The lake is always dangerous on
account of these winds, and when we visited the shore, about an
hour’s walk from the city, there were only a few little fishing-boats to
be seen close inland. The villages along it are mainly inhabited by
Ming Chiaos, who speak a language of their own. It is but few
comparatively of these villagers who can speak or even understand
Chinese, and those who learn it do so for the sake of trading in the
city. The women and girls wear their hair dressed rather differently
from the Chinese, and have a different type of face. Also, they do
not bind their feet, nor wear such elegant shoes. They seemed quite
friendly, but were rather too inquisitive, which prevented our
enjoying our picnic as much as we should have done.
TALI FU
 I made a sketch of a little bridge with a picturesque archway in
the shape of a tower leading on to it, behind which may be seen the
mountains towering up to a great height. These mountains are very
little known, and a European who went up to visit a temple some
hours distant got lost, and was three days before he was able to find
his way back; when he returned he was quite exhausted for lack of
food, and his clothes were almost torn to pieces by the bushes
through which he had been obliged to force his way. There are very
few inhabitants amongst these mountains, and they are of a fierce
and uncertain character.
The Tali Fu women were the most pleasant and responsive we
met in the province, and some looked decidedly intelligent. They did
beautiful embroidery of various kinds, and we were glad of the
opportunity to order straw hats from them with which to protect
ourselves from the sun. We were not glad, however, to see them
being brought home on the heads of the woman and child; though
this seems to be considered quite the correct thing in China. There
was no alternative between having these particular hats or going
without any, which was too risky. We are quite convinced that the
wearing of a “bandeau” inside the hat must have originated in China,
where the use of it is universal from the lowest to the highest. The
poorest coolie has a basket bandeau in the huge, pancake-like straw
hat which shelters him equally from sun or rain. The size of these
hats is two feet in diameter, and the bandeaux raise them so high
above the head that the effect is extremely comic, and they have to
be held on by an elaborate system of strings at the sides and back,
sometimes ornamented by the gayest of tassels, and with bright-
green strings tied under the chin. The favourite colour of a bandeau
is magenta, and the material of which it is made is satin. The
portrait of myself at the beginning of this volume shows our
complete travelling costume, together with the Buddhist pilgrim stick
and the horn spectacles which were formerly a distinctive mark of
the “scholar.” On returning home we found the hats were hardly an
exaggeration of the size worn in London. Liu had procured himself a
different variety of straw hat (see page 211) and a bright-green
waterproof cover for it.
Few foreign ladies visit Tali Fu, so our arrival created a mild
excitement in the place. An official lady (the highest but one in the
city) took the opportunity of paying a call on the missionary nurse
who was our hostess. She had assisted at the operation performed
by the doctor on the lady’s daughter, who was suffering from a
disfigurement of the face. The lady called at the moment lunch had
been served, so of course it had to be indefinitely postponed (the
Chinese calls not unfrequently last for several hours). She was
gorgeously attired in a heliotrope skirt embroidered with magenta
and green, which had come from Chengtu, and a fur-lined plum-
coloured coat. She wore gold pins galore in her hair, and a large
amber bead; gold tooth-picks and ear-picks hung from a button, and
she used these with French aplomb. Our clothes were examined with
much minuteness; she even pinched my toe to see if it could really
extend to the full length of my shoe, so I took it off to satisfy her on
the subject. Perhaps it was this that encouraged her to begin to
undress my friend down the back in order to see how her clothes
were put on; but the unfastening of the blouse and camisole still left
it a mystery how the skirt could be made to encompass so small a
waist—a problem which I fear will never be solved, for Chinese skirts
are completely open down one side and are merely folded round the
person. On the opposite side they also are open up to the
waistband, though this does not show, owing to the upper part of
the dress hanging over it. We spent some time drinking tea, and the
lady talked impartially with her hostess and her hostess’s servants,
her own remaining in the room all the time and also drinking tea.
Our ages had to be discussed, and as it appeared that mine almost
coincided with her own, I learnt the interesting fact that I was born
in the cycle of the dog. Owing to my fair hair I have always been
considered by the Chinese to be well over eighty years of age, both
now and on my previous visit. This is an extremely useful illusion, as
it renders me worthy of great reverence, and made it much more
proper for me to be travelling about and sitting by the high-road
sketching.
The call still dragged on, and we next discussed the price of our
clothes and the places we had visited, why we were unmarried, and
many other topics of equal interest. After about an hour of this
conversation the caller expressed a wish to see the bedrooms, and
having to pass through the dining-room in order to get there, she
perceived with surprise that the table was laid for a meal. The white
table-cloth was a new object to her, and one of deep interest. She
begged to be allowed to watch us eat, so of course she was invited
to join the belated meal. While this was being brought in she went
upstairs, and having inspected everything carefully, she tried the bed
(poor hostess!), and asked for some “flea medicine,” alias Keating.
We sat down to lunch, and the lady sent her servant to bring
from her sedan-chair a beautifully coloured silk serviette, which she
fastened to a button at the neck of her coat with a gold clasp. She
got on very well with a fork instead of chopsticks, but found the
combined use of knife and fork as difficult as we did that of
chopsticks. According to Chinese etiquette, all the company use their
own implements to help themselves, and the dish was conveniently
near to the lady! When she had no further use for the chicken bones
she handed them over her shoulder to the maid, to be thrown into
the courtyard; the more natural course, which would have been
adopted had she been at home, would have been to put them under
the table. Carpets are not in use even in the houses of the wealthy,
except on special occasions such as the New Year, so that the floors
are treated in the same way as the ground out of doors! Any “plat”
which was not to the lady’s liking was also handed to the maid to
eat, and it struck us as quite a new use for servants. When the table
boy was not otherwise employed, she questioned him minutely as to
household arrangements, the buying of food, &c., as if his mistress
were not present, and the information she received seemed to
interest her vastly. She continually pressed us to eat, and did full
justice to the meal herself, ladling the gravy out of the dish at
intervals with her own spoon, despite the fact that she was on her
way to a feast. I presented her with a piece of English ribbon and
explained its use, which happily coincided with what she had already
heard of English dress; but I was disappointed that she did not
follow the Chinese custom of presenting me with something in
return. My curiosity to see what her gift would be was justly
punished. Finally she left, seemingly well pleased with her visit,
which would not have been considered at all a long one in “the
middle kingdom,” although we had found it somewhat exhausting.
 CHAPTER XXII
Tali Fu to Teng Yueh

W
E left Tali Fu on a stormy-looking morning, accompanied by
the usual windy gusts, and after about a couple of miles
we reached the famous temple of the goddess of mercy,
the Buddhist Kwanyin. Our coolies went in to offer incense and to
invoke the goddess, as they did on their arrival at Tali Fu; for she is
a great favourite with the coolie class. There is a beautiful little
shrine perched on a big rock in a sort of small tank, which the
goddess (under the guise of an old woman) is supposed to have
brought here. In the courtyard beyond is a Buddhist temple, and in
one of the side courts a temple to a famous general, containing a
statue of him much more life-like than any we have yet seen. The
head coolie took this opportunity of consulting the goddess as to his
mother’s welfare—he has not seen her for over ten years—but the
answer was both unsatisfactory and vague! Our head coolie is a
quaint individual, but decidedly attractive. When we gave him a tael
for a pork-feast for the men he kept back part of it, and they were
extremely angry. In order to appease them, he not only had to pay
up, but also to give them a pork-feast on his own account. He wears
a most beautiful jade bangle, and is always immaculately clean to
look at, however dust-stained the rest of us may be. Some
Chinamen have that delightful quality of always looking clean and
keeping tidy.
When we were ready to start again the rain came down in
torrents, so the men put on their hats and every one demurred to
going on. There was no time to waste, however, so we insisted on
setting out, and after a time the rain stopped and the wind soon
dried our wet clothes. Turning down the valley after passing through
Siakwan, we entered the defile through a heavy gateway, which is
part of the fortifications, through which the wind tore like a
hurricane, snatching off the black scarf I had wound round my hair
in Chinese fashion, for it was impossible to keep on our gigantic
hats. We made our way through an impressive gorge, following the
river till we came to our halting-place at night. We had learnt at Tali
that the official rest-houses are open to travellers when not occupied
by officials, so we sent on to engage rooms in one, and certainly
gained both in quality and quantity, as we occupied a sort of large
barn with pillars in it, which looked as if it had formerly been used as
a stable. The next night we were not so lucky, and had the most
riotous party of rats in the loft above us that I have ever
encountered. Their revels brought down no inconsiderable portion of
the ceiling on our heads, and finally the rats themselves came down
in a sort of stampede upon us, showing no respect whatever for the
British face or form. The early dawn was further made hideous by a
chorus of geese, pigs, fowls, and mules, a horrid discord from which
we hurried away into the lovely dew-drenched country. Crossing a
suspension-bridge, we began a long, steep climb through the pine
woods, and the miseries undergone were soon forgotten in the ever
fresh beauties of the day. The long range of snow-capped mountains
lay behind us, and for three hours we mounted higher and higher,
coming to large trees of brilliant crimson rhododendrons, and various
kinds of roses climbing high up amongst them.
 SUSPENSION BRIDGE

As we sat in the shade near our coolies, taking a little rest, a

stranger coolie came and sat down in his hat beside us, to the great
indignation of one of our men, who promptly administered a kick in
the rear. He returned the kick as soon as our man turned his back,
but otherwise quite acquiesced in his dislodgment. A much dirtier
man of our own then came and occupied the seat, perhaps in order
to prevent any other intruder from doing so. The journey from Tali to
Bhamo was one, if possible, of more beauty than any former day’s
travel of this beautiful tour. The hedges were ablaze with blossom,
and the Pyrus japonica was wonderful. The size and colour of the
blossom astonished us, and also the size of the plant, which is more
like a tree than a shrub. The air was sweet with jasmine and orange
blossom and other shrubs of which we did not know the names. The
ascents and descents grew longer and more precipitous. We saw a
new variety of bridge of great interest, entirely composed of creeper.
The bridge was like a hammock, but fully a hundred feet long, and
very loosely woven together; the only part of it not made of creeper
was a narrow plank along the centre for walking on, and the upright
poles at each end. There was a creeper hand-rail, but it would have
been of very little use to hold on to it, as it was only attached at the
ends.
The seventh day after leaving Tali Fu, we crossed the Mekong, a
most picturesque river sunk in a narrow gorge and spanned by a fine
suspension-bridge. As I sat and sketched it in the early morning the
air was fragrant with wild orange blossom and jessamine, and
handsome plants of orange-coloured abutilon reared themselves by
the roadside. My sketching of the suspension-bridges might naturally
strike the eye of the engineer as incorrect, owing to the slightness of
the curve; but the fact is that they are tightened up after
construction with very great labour, by means of windlasses, till they
are (comparatively speaking) horizontal. On the board above the
entrance of the bridge was inscribed, “Built by the god.”
 BRIDGE MADE OF CREEPERS

The mountains are part of the great Himalayan range, and the
spurs run almost directly from north to south, so that our road from
east to west was always up and down. On reaching our halting-place
for the night we found all the inns occupied, except one where there
was a room but no beds. After much altercation, they agreed at one
of the inns to turn out some of the coolies and to give us their room;
but on inspection the look of it was so unalluring that I went to see
what the bedless room was like, and found a large barn over a
stable, which contained nothing but a family altar and some fodder.
This seemed infinitely preferable to us, especially as it had two large
windows looking on to the mountains. We spread our oilcloth on the
floor and our mattresses on it, and passed a far more comfortable
night than many that we had had, being undisturbed by cats or rats.
Only a night or two before, a cat had raided our larder, which we
thought we had made quite secure, overturning the pan, getting the
lid off, and departing with our tender young chicken, of which one
leg only was left in the pan. We always had to be very careful in
packing up all food before going to bed, but that mattered little, as
we took none in the morning before starting. Indeed, we found we
required much less food on this journey than elsewhere, and two
light meals a day were amply sufficient for our needs. At midday we
always lunched al fresco, being set down outside the village where
the men were to get their meal.
The next day after the theft we were unable to get a chicken of
any sort, and the following day we had to put up with a prehistoric
beast. In fact, our hands became horny from wrestling with tough
fowls. Sometimes unexpected dainties turned up, and certainly we
had a novel surprise on arriving at the town of Yung Chang Fu.
While sitting at supper, a visiting-card was brought in from the
magistrate, with a tray containing two fine ducks, two tins of
sardines, and a bottle of House of Commons whisky; two lively fowls
also were deposited on the floor. We sent our thanks to the
magistrate for his kindness, and accepted the fowls, but returned
the remainder of his gift according to custom. Our servant had vainly
been trying to get a chicken for us, so the gift was most opportune,
and we should have liked to thank the polite magistrate in person.
 After dinner we sent for the head coolie, as this was one of the
places where we paid the wages, and were amused to see him
secrete thirty-six ounces of silver up his sleeve as if it were nothing.
The total cost of hire for eighteen coolies from Yünnan to Bhamo, a
journey which occupied thirty-three travelling days, was a little under
£34.
For the first and last time in the whole of our six months’
journey, we met a European on the road, and he kindly gave us
permission not only to stay at his house at Teng Yueh, but gave us
the keys of his store-room; having heard previously of our coming,
he had given instructions to the caretaker to be on the look-out for
us. In any case, we should not have been likely to miss him on the
high-road, except at a spot where they have made a new road to
shorten the distance, and where Liu took the opposite road from the
one we did when he went ahead to get rooms.
Three days after we had crossed the Mekong, we had a stiffer
climb down to the Salween, which is 2300 feet above sea-level. It
has a shocking reputation for malaria, but at this season there was
nothing to fear. However, I offered a dose of quinine to whoever
wished to have it, and every one eagerly applied. The only member
of the party who was unable to take quinine tried what is frequently
found to be a good alternative, cinnamon. The Salween has a two-
span bridge, connected in the centre by a fine block of masonry. The
village close by seems to be inhabited by Shans, and the coolies
took a long rest before beginning the toilsome ascent. We climbed
up for about two hours, and were thankful for a night’s rest before
climbing the remaining three. We had sent one of our escort ahead
in good time to secure the only inn, which was of primitive
construction, and of which, as not infrequently happens, the
partitions were so flimsy as to allow of a fairly good view from one
room into another.
The scenery next day was particularly fine, and from this point
the vegetation became semi-tropical. The monkeys barked and
chattered, but kept out of sight, and we spent a long day in climbing
down from the eight-thousand-foot elevation to which we had
climbed so laboriously on the previous evening. The following day
we reached Teng Yueh in a tearing wind. In fact, we have found the
wind the most fatiguing element in our travels.
Teng Yueh (or Momein, as it was formerly called) boasts an
English consulate, and we were kindly entertained there by the
Consul’s arrangement, although he himself was absent. His delightful
Chinese major-domo received us with his best English welcome, and
delighted us by at once asking “Missee want hot bath?” “Other
Missee want hot bath?” and promptly went to get them ready. He
proved to be the most attentive of servants, and told us we were to
ask for whatever we wanted. Naturally, the first requisite seemed a
washerwoman, and he promised to send “catchee female,” and the
next morning he introduced an elegant native laden with jewelry. To
a further request that he would procure a small piece of mutton for
us to take on our journey, he said, “No can catchee mutton, but goat
very good.”
The neighbourhood of Teng Yueh is picturesque, surrounded by
mountains, and the town itself is more open and clean than most.
The people have pleasant manners and are friendly to the foreigner.
There is a new Custom House and post-office, and houses for the
European staff, which are not yet all completed. These stand out
conspicuously, surrounded by their walls, and we saw them many
miles away as we descended from the mountains. We felt that this
was our last chance of getting Chinese things, so we spent the
afternoon shopping, but were not altogether successful. The jade for
which the place is celebrated we found more expensive than we had
expected, and there did not seem much choice. The threatening look
of the sky made us procure waterproof hat-covers at one penny
apiece, and they proved quite efficacious in dispersing the clouds.
There were pine-apples, bananas, tomatoes, and various vegetables
in the market; and we also succeeded in buying postage-stamps, of
which we had found the post-office of Tali Fu practically empty. We
had some difficulty, in fact, in getting them at most of the post-
offices in the province of Yünnan.
SHAN WOMAN
 And now we started on our last Chinese journey, after one day’s
rest, feeling each day more regretful at the thought that it would so
soon be ended. The days grew hotter and the wind stronger, and the
road not nearly so beautiful, and the inns decidedly worse, yet still
we looked back with wistful eyes to the “Flowery Land” (so justly
named) we had passed through. The country seemed mainly
inhabited by Shans, a fine, well-built race, very clean-looking, and
the women picturesquely dressed. Most of them were entirely
dressed in black, with large silver ornaments, and they looked very
attractive with jaunty bunches of jasmine stuck in their tall black
turbans. Their villages looked quite different from the Chinese
villages. The inns look miserable, but are less abject than they
appear; and when, to my surprise, in the dim light I saw through the
airy bamboo screen (which formed the wall) a portly pig arise from
my side, I thought how often one had passed much less comfortable
nights in pleasanter surroundings! I think the pig must have taken
up her residence there when I was asleep.
From Teng Yueh to Bhamo the route lies all the way along the
river-side of the Taping, and the first part of it is most tiresome, as it
is constantly in the shingly bed of the river itself, where the heat of
the sun is reflected from the stones. Our coolies’ clothing grew daily
less, and even the chilly creature who wore an eighteenpenny skin
coat discarded it. In fact, the coolies began to look like the
conventional pictures of John the Baptist with one bare arm exposed
to view. One of our attentive boys always presented us with our hats
when he thought it time to put them on in the morning on account
of the heat. We tied them very securely to our chairs ever since one
of the coolies thought fit to carry my friend’s hat on his own head.
There was an amusing theatrical performance of an elementary
kind held close to our inn in a small village. It was a sort of votive
offering as a thanksgiving for the recovery of a sick person. The
stage was in the middle of the road. It consisted in a table and a
small bench in front of it, on which stood the principal figure,
representing the Emperor. He recited a good deal under the large
flowing beard which covered his mouth, and which he occasionally
pretended to pull. As it was attached with string round his ear, a
slight tweak would have detached it. Before him stood three gaudily
dressed figures, who made a few dance steps at intervals, and
turned round; they also recited now and then. Their main
occupation, however, seemed to be fanning themselves and
expectorating. When they got tired they withdrew behind the table
for tea, and the Emperor twitched off his beard and imperial clothes
in the twinkling of an eye, with an air of infinite relief. The clothes
were thrown into a large basket which seemed to contain all the
stage properties. The spectators were really the most interesting
part of the show, as several of the women had evidently got on their
best clothes, and fine silver ornaments all over their high turbans,
and massive necklaces and bangles. We were beginning at last to
get away from the prevalent plague of goitre, and it was a relief to
see clean, healthy people again.
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!

ebookfinal.com