Download the Full Version of textbook for Fast Typing at textbookfull.

com

Analyzing Health Data in R for SAS Users 1st

Edition Monika Maya Wahi

https://textbookfull.com/product/analyzing-health-data-in-r-
for-sas-users-1st-edition-monika-maya-wahi/

OR CLICK BUTTON

DOWNLOAD NOW

Download More textbook Instantly Today - Get Yours Now at textbookfull.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

SAS for R users : a book for budding data scientists First

Edition Ohri

https://textbookfull.com/product/sas-for-r-users-a-book-for-budding-
data-scientists-first-edition-ohri/

textboxfull.com

Clinical trial data analysis with R and SAS Second Edition

Chen

https://textbookfull.com/product/clinical-trial-data-analysis-with-r-
and-sas-second-edition-chen/

textboxfull.com

Analysis of Correlated Data with SAS and R Mohamed M.

Shoukri

https://textbookfull.com/product/analysis-of-correlated-data-with-sas-
and-r-mohamed-m-shoukri/

textboxfull.com

Analyzing High-Dimensional Gene Expression and DNA

Methylation Data with R 1st Edition Hongmei Zhang (Author)

https://textbookfull.com/product/analyzing-high-dimensional-gene-
expression-and-dna-methylation-data-with-r-1st-edition-hongmei-zhang-
author/
textboxfull.com
SAS certification prep guide base programming for SAS 9
Fifth Edition. Edition Sas Sas Sas

https://textbookfull.com/product/sas-certification-prep-guide-base-
programming-for-sas-9-fifth-edition-edition-sas-sas-sas/

textboxfull.com

Analyzing Data through Probabilistic Modeling in

Statistics 1st Edition Dariusz Jakóbczak

https://textbookfull.com/product/analyzing-data-through-probabilistic-
modeling-in-statistics-1st-edition-dariusz-jakobczak/

textboxfull.com

Survival Analysis with Interval Censored Data A Practical

Approach with Examples in R SAS and BUGS 1st Edition Kris
Bogaerts
https://textbookfull.com/product/survival-analysis-with-interval-
censored-data-a-practical-approach-with-examples-in-r-sas-and-
bugs-1st-edition-kris-bogaerts/
textboxfull.com

Infographics Powered by SAS:: Data Visualization

Techniques for Business Reporting 1st Edition Travis
Murphy
https://textbookfull.com/product/infographics-powered-by-sas-data-
visualization-techniques-for-business-reporting-1st-edition-travis-
murphy/
textboxfull.com

Clinical Data Quality Checks for CDISC Compliance Using

SAS 1st Edition Sunil Gupta (Author)

https://textbookfull.com/product/clinical-data-quality-checks-for-
cdisc-compliance-using-sas-1st-edition-sunil-gupta-author/

textboxfull.com
Analyzing Health Data in R
for SAS Users
Analyzing Health Data in R
for SAS Users

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

© 2018 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

Printed on acid-free paper

International Standard Book Number-13: 978-1-4987-9588-3 (Hardback)

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

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmit-
ted, 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 organization that provides licenses and
registration for a variety of users. For organizations that have been granted a photocopy license by the CCC,
a separate system of payment has been arranged.

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

Library of Congress Cataloging‑in‑Publication Data

Names: Wahi, Monika, author. | Seebach, Peter, author

Title: Analyzing health data in R for SAS users / Monika Wahi and Peter
Seebach
Description: Boca Raton : CRC Press, 2017. | Includes bibliographical
References, Identifiers: LCCN 2017021131 | ISBN 9781498795883 Subjects: LCSH:
Bioinformatics. | Medical informatics. | R (Computer program language) | SAS
(Computer file) Classification: LCC QH324.2 .W34 2017 | DDC 610.285--dc23
LC record available at https://lccn.loc.gov/2017021131

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com

and the CRC Press Web site at

http://www.crcpress.com
Contents

Preface ......................................................................................................................ix
Acknowledgments .................................................................................................xi
Authors ................................................................................................................. xiii

1. Differences between SAS and R .................................................................1

Structure of Program.......................................................................................1
Installation of PC Version: SAS versus R .................................................2
Licensing Differences: SAS versus R ........................................................ 3
SAS Components versus R Packages .......................................................3
What is RStudio? .........................................................................................4
Maintaining Current Versions in SAS versus R .....................................4
SAS versus R User Communities .............................................................. 6
SAS versus R User Interfaces ..................................................................... 7
Code Documentation and Metadata: SAS versus R ............................... 9
Handling of Data ........................................................................................... 12
A Focus on SAS Data Handling .............................................................. 12
Comparison to R Data Handling ............................................................ 13
Basic Differences in Code Syntax: R versus SAS .................................. 15
SAS Formats and Labels versus R Approaches .................................... 17
SAS versus R—What to Choose? .................................................................22
Why R Can Be a Difficult Choice for Public Health Efforts................22
Considerations When Choosing R versus SAS ..................................... 24
Optional Exercises ......................................................................................... 25
All Sections ................................................................................................ 25
Questions ............................................................................................... 25
Answers ................................................................................................. 25

2. Preparing Data for Analysis....................................................................... 27

Reading Data into R....................................................................................... 27
Importing Data .......................................................................................... 27
Checking the Dataset after Reading It in .............................................. 28
Checking Data in R........................................................................................ 33
Statistics on Continuous Data in R ......................................................... 33
Visualizing Continuous Data in R.......................................................... 37
Statistics on Categorical Data in R: One Variable ................................. 40
Statistics on Categorical Data in R: Crosstabs.......................................44
Editing Data in R............................................................................................ 48
Trimming off Unneeded Variables ......................................................... 48
Applying Qualification Criteria through Subsetting Datasets .......... 50
Creating Grouping Variables................................................................... 53

v
vi Contents

Creating Indicator Variables for Two-Level Categories....................... 61

Creating Indicator Variables for Three-Level Categories .................... 62
Creating Indicator Variables for Multilevel Ordinal Categories ........ 63
Creating Indicator Variables for Multilevel Nominal Categories ...... 65
Creating Missing Flags............................................................................. 67
Preparing Binary Outcome Variable ...................................................... 68
Planning a Survival Analysis Dataset with Time-to-Event
Variables ..................................................................................................... 68
Developing the Survival Dataset ............................................................ 70
Dealing with Dates ................................................................................... 82
Recoding and Classifying Continuous Variables................................. 88
Recoding a Continuous Outcome Variable ........................................... 89
Data Validation in R ...................................................................................... 92
Bivariate Relationships between Continuous Variables...................... 93
Bivariate Relationships between Categorical and
Continuous Variables ............................................................................... 99
Bivariate Relationships between Categorical Variables .................... 106
Power Calculations ................................................................................. 108
Write Out Analytic File .......................................................................... 115
Optional Exercises ....................................................................................... 116
Section “Reading Data into R” .............................................................. 116
Questions ............................................................................................. 116
Answers ............................................................................................... 116
Section “Checking Data in R” ............................................................... 117
Questions ............................................................................................. 117
Answers ............................................................................................... 118
Section “Editing Data in R” ................................................................... 122
Questions ............................................................................................. 122
Answers ............................................................................................... 124
Section “Data Validation in R” .............................................................. 127
Questions ............................................................................................. 127
Answers ............................................................................................... 128

3. Basic Descriptive Analysis ....................................................................... 133

Making “Table 1”—Categorical Outcome ................................................ 133
Structure of Categorical Table 1 ............................................................ 133
SAS Approaches to Categorical Table 1 Structure.............................. 136
SAS Approaches to Table Presentation Using Excel .......................... 141
SAS Bivariate Categorical Tests ............................................................. 142
The Table Command in R ...................................................................... 143
R Approaches to Categorical Table 1 .................................................... 145
Approaches to Automating Table Generation in R ............................ 153
R Bivariate Statistical Tests .................................................................... 155
Contents vii

Making “Table 1”—Continuous Outcome ............................................... 156

Structure of Continuous Table 1 ........................................................... 156
SAS Approaches to Continuous Table 1 .............................................. 156
Continuous Bivariate Statistical Tests in SAS ..................................... 163
R Approaches to Continuous Table 1 ................................................... 165
Continuous Bivariate Statistical Tests in R .......................................... 171
Descriptive Analysis of Survival Data...................................................... 176
Summary Statistics and Plots on Time Variable................................. 176
Generating and Plotting Survival Curves ........................................... 177
Bivariate Tests of Survival Curves ........................................................ 185
Optional Exercises ....................................................................................... 189
Section “Making ‘Table 1’—Categorical Outcome” ........................... 189
Questions ............................................................................................. 189
Answers ............................................................................................... 189
Section “Making ‘Table 1’—Continuous Outcome”........................... 190
Questions ............................................................................................. 190
Answers ............................................................................................... 190
Section “Descriptive Analysis of Survival Data” ............................... 193
Questions ............................................................................................. 193
Answers ............................................................................................... 194

4. Basic Regression Analysis ........................................................................ 197

This Book’s Approach ................................................................................. 197
Selection of Modeling Approach .......................................................... 197
Selection of Manual Approach .............................................................. 199
Operationalizing the Stepwise Selection Process .............................. 200
Prespecifying Hypotheses and Avoiding Fishing ............................. 201
Linear Regression and ANOVA ................................................................. 203
Preparing to Run Linear Regression .................................................... 203
Linear Regression Modeling and Model Fit Statistics ....................... 205
Selecting the Final Linear Regression Model ..................................... 211
Considerations in Improving the Final Model ................................... 214
Considering Collinearity ....................................................................... 214
Adding Interactions ................................................................................ 215
Goodness-of-Fit Statistics....................................................................... 217
Linear Regression Model Presentation ................................................ 218
Plot to Assist Interpretation ...................................................................225
Logistic Regression ...................................................................................... 228
Estimates Produced by Logistic Regression ....................................... 229
Logistic Regression Considerations ..................................................... 229
Introduction to Logistic Regression Modeling ................................... 231
Logistic Regression Modeling and Model Fitting .............................. 232
Selecting the Final Logistic Regression Model ................................... 239
viii Contents

Logistic Regression Model Presentation ............................................. 249

Plot to Assist Interpretation ................................................................... 255
Survival Analysis Regression .................................................................... 256
Selecting a Parametric Distribution in Survival Analysis ................ 257
Selecting a Semiparametric Distribution for Survival Analysis ...... 262
Introduction to Survival Analysis Regression Modeling ................. 264
Survival Analysis Regression Modeling and Model Fitting ............ 265
Parametric Survival Analysis................................................................ 265
Semiparametric Survival Analysis ....................................................... 270
Issues to Consider in Survival Analysis .............................................. 274
Selecting the Final Survival Analysis Model ...................................... 275
Survival Analysis Model Presentation ................................................ 275
A Note about Macros................................................................................... 276
Optional Exercises ....................................................................................... 278
Section “This Book’s Approach” ........................................................... 278
Questions ............................................................................................. 278
Answers ............................................................................................... 278
Section “Linear Regression and ANOVA” .......................................... 278
Questions ............................................................................................. 278
Answers ............................................................................................... 279
Section “Logistic Regression” ............................................................... 280
Questions ............................................................................................. 280
Answers ............................................................................................... 280
Section “Survival Analysis Regression” .............................................. 281
Questions ............................................................................................. 281
Answers ............................................................................................... 282
Section “A Note about Macros” ............................................................ 285
Questions ............................................................................................. 285
Answers ............................................................................................... 285

References ........................................................................................................... 287

Index ..................................................................................................................... 297
Preface

When I, Monika Wahi, gave a presentation at the Effective Applications of the

R Language (EARL) Conference in Boston in November 2015, there I attended
a panel discussion of the event leaders, although it was a very open dis-
cussion, with the audience enthusiastically participating. The goals of the
R Consortium were delineated, which led to the consideration of how to
promote the increased usage of the R language by statisticians. After all,
R is open source software, so it does not have a marketing department per se.
Someone in the audience brought up with the specific topic of healthcare
analytics, asking the attendees why R is not used more often in healthcare.
After a pause, someone pointed out the domination of SAS in the indus-
try, which crowds out R. But then, another person speculated that the Food
and Drug Administration (FDA) would only accept analysis in SAS (a myth
which we debunked in Chapter 1). It became clear that there seemed to be
many barriers to promoting R among healthcare analysts.
After further consideration, this was ironic, because both R and SAS are
very extensive languages, but what we do in healthcare analytics on a day-to-
day basis is relatively straightforward. We are not modeling any spaceship
trajectories, or building complex economic models, or predicting weather or
the outcome of sports games. It occurred to me that as we generally do the
same things over and over again in healthcare analytics, and because most
people in healthcare analytics do these things in SAS, a book that focuses
only on explaining how to do what we normally do in SAS in R would help
healthcare analysts who may want to use R also.
Therefore, this book is aimed at the healthcare analyst who is a SAS user
looking to learn R. Chapter 1: “Differences between SAS and R” is written
mainly for those who are interested in the backstory of how SAS and R are
run differently. Readers who want to get immediately to coding should
skip to Chapter 2: “Preparing Data for Analysis”, which describes editing
and validating an analytic dataset in R rather than using SAS data steps.
Chapter 3: “Basic Descriptive Analysis” will explain how to use the analytic
dataset developed to produce a basic descriptive analysis (often presented
as Table 1 in manuscripts). Finally, Chapter 4: “Basic Regression Analysis”
covers linear and logistic regression and basic survival analysis. As R pro-
duces such lovely plots, an insert with 16 color plots is included in the center
of this book.
We hope this book encourages you to try R for some healthcare analysis
tasks.

Monika Wahi and Peter Seebach

ix
Acknowledgments

The number of people who have helped us with this book is large, and we
are eternally grateful. But after controlling for age, relevant skills, and educa-
tion, it turns out that specifically no one individual contributed significantly
more to help us with this book than another. Our conclusion is that you are
all amazing, and we thank you very much for your support!

xi
Authors

Monika Wahi, MPH, CPH obtained a bachelor of science in costume design

and textiles and clothing (with a concentration in Journalism) from the College
of Human Ecology, University of Minnesota, St. Paul, Minnesota, and, then
went on to complete her masters in public health from the University of
Minnesota School of Public Health, Minneapolis, Minnesota. After serving
in several different scientific and administrative roles at Hennepin County
in Minnesota, a nonprofit Alzheimer’s research institute in Florida, and at
the U.S. Army at a site in the Greater Boston area, she struck out on her own
to build her consulting business, DethWench Professional Services (DPS).
Since 2012, she has been serving as a lecturer at the Labouré College, Milton,
Massachusetts, teaching classes in the U.S. healthcare system and statistics.
Monika has also led to the expansion of DPS to serving an international
clientele through providing online educational material and services as well
as public health research consulting.

Peter Seebach was raised by mathematicians, and never fully recovered.

He earned his bachelor’s degree in psychology from St. Olaf College,
Northfield, Minnesota, and then went on to apply it to a career as a software
developer and writer. He enjoys computers, writing, and writing about com-
puters. He is an outlier, and therefore tends to throw off the average.

xiii
1
Differences between SAS and R

This chapter is meant to help the SAS user conceptualize the important dif-
ferences between R and SAS that will affect the work of the healthcare ana-
lyst who knows SAS but is looking to learn how to use R. The first section
describes important differences in the structures of the SAS and R programs.
This leads to the discussion in the second section, which focuses on how
these differences in structure affect the differences in data handling between
the two programs. The third section of this chapter contextualizes the choice
between using R versus SAS for a healthcare analytics project and provides
a guide for selecting which software to use. Optional practice exercises are
included in the fourth section.

Structure of Program
Perhaps the most important difference between SAS and R is the structure
of how the program is built and maintained. To begin to explain this differ-
ence, we will start with considering how different the download and install
process is between PC SAS and R. Next, we will cover how these differences
are also reflected in the differences in the way licensing is maintained in the
two programs. Third, the differences between SAS components and their
parallel in R, called R packages, will be discussed, and after that, differences
between SAS and R in approach to maintaining the most current version in
a production environment will be explained. The differences between the
activities of SAS and R user communities will be described, followed by the
differences between the SAS and R user interfaces. Finally, some thoughts on
the principles of organizing code, metadata, and documentation in SAS and
R are presented.

1
2 Analyzing Health Data in R for SAS Users

Installation of PC Version: SAS versus R

When using a typical PC SAS license from a university, the data analyst
has to download and install the program on his or her Windows personal
computer (PC), as the current operating system (OS) for Macintosh is not
supported.* To do this, the analyst is provided access to either many CDs
that her computer will tell you to put in or take out of her CD/DVD drive
during the installation process, or alternatively, an extremely large setup
file that takes a very long time to download (plan for hours, not minutes).
The analyst is also provided a small text file with unique license information
for her institution’s license.
Once the analyst has access to these setup files, a setup file is run on the
analyst’s PC, and this setup file takes a while to extract. The installation
must be monitored because the user has to click through many menus. This
is because these SAS institutional licenses are very extensive and include
many add-ons and components. This heavy-handed installation process is
designed to make sure the analyst has access to all the components to which
she is entitled under the license (which, at a university, tends to be a large
volume).
However, when working in SAS, the analyst may encounter the rare case
in which she is using an analytic function that is not included in the SAS
license components, and this throws up an error message. This is a confus-
ing situation because the error message does not usually point to a miss-
ing component; it simply rejects the code as being incorrect in some way, so
troubleshooting can be challenging.
Because R is open source, meaning that R developers are volunteers and
make the code and documentation for how R runs readily available on the
Internet, there is no need for the advanced functions to control licensing
that SAS employs. This makes the way R is distributed different from
SAS. Anyone can go to the public website called “The Comprehensive
R Archive Network” (CRAN) (https://cran.r-project.org) [1] and down-
load the latest version of R for Mac or Windows and install the core
program. The user interface (UI) looks slightly different on Mac versus
Windows, but the differences are minor.† The R installation file is small
relative to the SAS installation file, and once the user downloads this file
and runs it, the setup wizard is clear and easy, making for a quick instal-
lation process.

* Per SAS, SAS versions developed since Macintosh OS X came out are not supported on OS
X. A modification of SAS called JMP (pronounced “jump”) has been developed for Mac and
Windows users alike, although it has greatly reduced functionality. SAS also has a web-based
version that can run in Mac and Windows browsers which is mainly used in teaching rather
than production environments.
† The examples given in this book refer to the Windows version of R.
Differences between SAS and R 3

Licensing Differences: SAS versus R

The exact components of the SAS institutional license make are communi-
cated to SAS during the installation process at the step where the user is
asked by the setup wizard to reference the small text file provided with the
setup files. As stated earlier, the list of components included in a SAS insti-
tutional license are usually large. When SAS negotiates enterprise licenses
with universities, they are set up such that a large volume of components is
included, and the student or faculty pays only a negligible fee or nothing to
obtain this licensed version, provided they can prove their status with the
university. This is because SAS wants to promote use and learning of all its
components at universities.
Importantly, SAS prices their licenses for non-university businesses dif-
ferently. As an example, an independent nonprofit research institute on
the campus of a state university in Florida contacted SAS and asked if the
research institute could use the university license. SAS did not agree, and
instead, prepared a license for PC SAS strictly for the research institute. This
license only had one seat, and only had the base component of SAS (“base
SAS”) and the basic engine that runs regression functions (“SAS STAT”). This
one-seat license cost the nonprofit approximately $10,000 per year in 2007, so
the nonprofit could not add extra seats. Hence, researchers who learned SAS
at the university and then went on to be hired as scientists at the nonprofit
research institute were not able to use their extensive knowledge of all the
components of SAS due to the prohibitively expensive licensing approach.
Because R is open source, there is no licensing fee, so R is perfectly suited
for public health work in low- to middle-income (LME) countries, nonprof-
its, and businesses with a low profit margin. This book can hopefully help
bridge the gap between SAS and R for SAS-experienced healthcare analysts
who are priced out of the market by this situation.

SAS Components versus R Packages

Base SAS, the core of SAS, is a rather large program. As mentioned before, the
main R program that is downloaded from CRAN is much smaller than the SAS
setup program, and therefore, downloading it is fast and installing it is very easy.
The drawback is that just about anything the user tries to do once installing R will
require an outside component not in the base program, called an R “package.”
Because R is not licensed for purchase, it is much more efficient for each
user to simply build their version of R by installing the packages needed.
Admittedly, this can be a daunting task, as some packages are based on other
packages, so these all need to be installed. For example, to make a Kaplan–
Meier plot in R, the user must install the “survival” package, as well as the
“KMSurv” package [2]. More recently, packages are designed to automati-
cally install packages on which they are dependent, so this problem does not
occur as frequently anymore.
4 Analyzing Health Data in R for SAS Users

Luckily, as with the base R program, all the packages are free and are easy
to install from within the program. In the native R UI, there is a menu with
only seven selections, and “packages” is one of them. If the user chooses the
“packages” menu and selects “load package,” the UI presents a list of pack-
ages that can be installed, and the user just needs to select the correct ones
and install them. Packages can also be installed using commands, but the
load package menu makes the process extremely easy.

What is RStudio?
This book gives guidance on using the PC version of R in Windows as it
appears with its native UI. R’s native UI is typically sufficient to be used by
healthcare analysts to develop statistical models. However, there is also an
integrated development environment (IDE) that can be used called RStudio.
This is supported by the R Consortium, which is a collaboration between the
R Foundation (which maintains CRAN), RStudio (a collection of R develop-
ers working on the IDE), and other big tech companies such as Microsoft and
Google [3]. Like R, RStudio is also open source and free to individual users.
RStudio is different than R in that it is an IDE and includes a source code
editor, build automation tools, and a debugger. RStudio can be run as a desk-
top or server version, so it is used at universities in programs that teach pro-
gramming in IDEs [4]. It is an excellent tool for deploying Shiny, which is
an R package that interfaces R with the web and turns R analyses into web
applications [5].
Generally, the capabilities afforded by RStudio are required for deploying
web applications, but for hypothesis-driven healthcare analytics, RStudio
can be overkill. For example, RStudio has several windows that are associ-
ated specifically with the IDE, and would not appear in R. Therefore, unless
the analyst needs an IDE, using R rather than RStudio is preferred.

Maintaining Current Versions in SAS versus R

When using a university SAS license, there is a month in the year that the
license expires because these are set up as yearly licenses. When the license
expires, the user can still open the SAS program, but it throws up an error
message indicating the user will need an updated text file with a new license,
and does not let the user unlock the program until this is loaded. At this
point, the user must obtain the new text file from the university, and from
within SAS, load this license, thus unlocking the program. Although SAS
does update its base program with new version from time to time, it does not
do it often, so renewing the license is much more common than installing a
new version of SAS.
Part of the reason SAS does not update its base program often has to do
with how it determines what to put in updates. Independent SAS program-
mers typically develop macros (canned code procedures) in SAS macro
Differences between SAS and R 5

language and make these available on the Internet. These are not held in a
central repository but posted all over the Internet, and are also highlighted
in SAS white papers presented for regional as well as national and interna-
tional user groups [6–8], which are also not available in a central indexed
repository (although they are posted on SAS-sponsored websites and are
easy to find through a search on the web).
If a particular macro becomes popular with SAS users, some may choose
to write a peer-reviewed article about it [9], fostering discussion of the macro
in the SAS community. If SAS receives enough requests to include the macro
as a main function in SAS, and can verify the functioning of the macro, SAS
may choose to include it in its next build. This is a decision made by SAS
on a business level, not by the SAS user community, so opinions may differ
between the user community and the business as to what macros should be
included as procedures, or “PROCs,” in new versions of SAS.
R does not have the constraints SAS has with respect to innovation and
change management. First, its base program is very lean, which means it
is not difficult to update and disseminate. For this reason, new versions of
R can be released quite often (at least once per year). The penalty for the
R user is that if she wants to update her core version of R, she will have to
reload all the R packages she has been using to carry on with her analy-
sis. An advantage is that the newest version is always available on the
CRAN site for free, as are the packages, and this means that the only loss
in efficiency is through time required to update the R core program and
reload the packages. Also, R developers have created packages available
on CRAN that can be loaded and configured to run automated functions
to help R shops keep their base R and all its packages updated on a regular
basis.
SAS macros that are not available within the program are documented all
over the Internet by individual programmers, just like R packages that are
not official or are under development. However, macros built into SAS as
commands or “PROCs” are well-documented in SAS’s online and paper help
files. Similarly, R’s published packages are well-documented on the CRAN
website [10], as standardized comprehensive documentation is required to
be approved to have a package published to CRAN.
When SAS updates its program, it includes updates to its core, new mac-
ros built in as PROCs, and updates to old PROCs and functions as well. This
means that production operations dependent on SAS can effectively plan
for change management. Conversely, because R updates its core regularly,
and the authors of their packages also update their packages idiosyncrati-
cally, maintaining the most up-to-date version of R is actually rather chal-
lenging in a production environment. Independent tools, functions, and
routines have been developed to assist with this task [11], but it is impor-
tant to emphasize that those running R in production must be proactive in
maintaining the most up-to-date versions of R and packages being used in
production.
6 Analyzing Health Data in R for SAS Users

It must be noted that R has a different method than SAS in deciding what
packages to include on CRAN. Like with SAS, R users can prepare and pub-
lish unverified packages for R online [12,13], and adventurous R users can
test them. If the package is deemed worthy by many, the author may apply
to have the package included on the CRAN site (and thus have it show up
in the “load package” menu). But depending upon the package, installing
a new package may effectively require devoting 15 min or more to loading
different packages on which a new package might be based, and testing the
results to make sure the new package can actually run. When this happens,
the R user may nostalgically muse about how easy it is to run macros and
PROCs in SAS by comparison, but once the packages are sufficiently loaded
in R and running well, this nostalgia generally passes.

SAS versus R User Communities

SAS is known for its expansive support for its user community. SAS orga-
nizes “SUGs,” or SAS User Groups, and provides logistic and financial
support for conferences requested by user groups [6–8,14,15]. These confer-
ences produce an extensive set of non-peer-reviewed white papers that pro-
mote the use of all of SAS’s various components by helping the user deal
with the unnecessary complexity of SAS. These white papers are unfor-
tunately of varying quality; some are well-written and helpful, whereas
others are poorly written and confusing, and there seems to be little in the
way of consistent standards as to what is accepted as a SAS white paper.
Nevertheless, SAS users are provided a fertile opportunity to thrive aca-
demically with SAS given the many opportunities offered to publish white
papers and present them at meetings of SUGs, as well as the vast amount
of documentation available online that results from the publication of these
white papers, which effectively facilitates the adoption of SAS’s more com-
plex functions.
On the other hand, R, which is an open-source software and therefore
does not have a conference or marketing department, has a different land-
scape when it comes to user groups and community gatherings. Meetup.
com, an online platform designed with tools meant to be used to build com-
munities through the scheduling of informal meetings or “Meetups,” has
been used extensively by the R community to gather old hands as well as
early adopters together informally for mutual presentations [16]. Unlike
with SUGs, these Meetups are community-initiated and supported, so they
are held in cheaply available spaces, and often garner little or no industry
financial support.
Happily, this trend is changing, as R is piquing interest with big play-
ers, most notably Microsoft, which purchased Revolution Analytics in 2015.
Revolution Analytics was an R shop that specialized in improving data han-
dling in R so as to enable it to process larger datasets [17]. In addition, larger
scale gatherings, such as the Effective Applications of the R Language (EARL)
Differences between SAS and R 7

conference, which is now held twice per year on two different continents, is
sponsored by various members of industry, including Mango Solutions, a
long-time big-ticket R analytics consulting group founded in 2002, as well as
Microsoft itself [18].

SAS versus R User Interfaces

The PC SAS interface provides a birds-eye-view of many of SAS’s functions
all at once. On the left panel, the user can toggle between a list of results
(output) and an explorer window that will show the contents of “work” (the
temporary directory where SAS puts datasets while it is running) as well as
other libraries mapped with the LIBNAME statement. Through this panel,
the user can easily open and view the datasets (in *.sas7bdat format) in the
mapped libraries.
In the main panel, users can display multiple windows at once. The user
can open multiple code windows (which can be saved as *.sas files), display
the log window (which shows the log of executed statements as well as error
messages, and can be saved as a *.log file), as well as view an output file
that can be displayed in hypertext markup language (HTML) and saved
separately.
R’s interface runs differently. Like SAS, the R PC interface allows for multi-
ple windows to be opened at once, but the function of the windows is differ-
ent. When R is launched, only one window is opened automatically, and this
is called the R console. This window is where statements will be executed,
log statements will be recorded, and tabular and numeric output will appear.
Next, the user can open one or more code editing windows. R has a menu
with only five choices: File, Edit, Packages (described earlier), Windows, and
Help. Choosing File—New Script from the menu will open a new code edit-
ing window. If the users saves a file from this window (Choose File—Save
As when the code window is selected), it will be saved as an *.R file.* Please
note that it is easy to open *.R files in Notepad or another word processing
program without needing to load R.
Once the user has a code window open, she can prepare code and run it
in the console by either copy/pasting it from the code editing window into
the console and pressing “enter,” or by highlighting the code in the code
window, right-clicking and choosing “Run line or selection” (also Control-R),
and this will transfer the code to the console and run it. Or, she can simply
type code directly into the console and hit “enter,” and the code will run.
An important point to be made here is that although R has an extensive
array of statements that could be used to do the equivalent of mapping a
LIBNAME in SAS, probably the easiest way to set a singular default directory

* Our experience is that occasionally, the *.log extension used in SAS is already designated to
default to another application in Windows (thus prompting a dialogue box during installa-
tion). So far, we have not encountered extension conflicts in R.
8 Analyzing Health Data in R for SAS Users

is to run the R program, select the console, then using the menu, choose
File—Change Dir. This will allow the user to navigate to the directory on her
PC or server where most of her R project is stored, and this will make loading
code from that directory much simpler throughout the session. Please note
that this designation will end at the end of the R session. Also, please take
notice of the fact that if the code window is selected and the user chooses
File—Open Script, the browser will default to the last code directory used,
but if the console is selected, the same operation will bring the user to the
directory designated during the File—Change Dir function.
The console is the place where the following appear: log messages, error
messages, and non-graphical output. For example, if the user requests a fre-
quency table, or conducts a calculation, or asks for the number of observa-
tions in a dataset, after the line of code is run in the console (displayed in red
text), this information will be reported subsequent to the code (displayed in
blue text). The console will continue to fill up with a history unless the user
chooses Edit—Clear Console. Clearing the console from time to time may be
helpful for the user involved in a large project, and might be seen as equiva-
lent to clearing the log file in SAS.
Because so many different types of messages appear in the console when
R code is run, saving “log files” is not as straightforward in R as it is in SAS.
(Please notice that the menus change depending upon whether the console
or the code window is highlighted.) Other options exist in the console’s set
of menus, but it may serve the user best if the user wants to save elements of
the session (such as error messages for later troubleshooting, spontaneously
formed code in the console, or log messages) to select the console window
and choose Edit—Select All, and then choose Edit—Copy.* Next, the user
can switch to a word processing environment (such as Notepad or Microsoft
Word) and choose Paste or Control−V.
At this point, it is important for the user to edit the pasted text to remove
the elements of the output not wanted to be saved in the manually edited
“log file.”† Because the R log files require some hand-editing, there are both
pros and cons. An advantage in R, unlike in SAS, is that the user can easily
remove parts of the text that are unimportant to her (such as confirmation
of reading in a data file, or reports of successful execution of simple com-
mands). It is also easy to annotate these parts to provide insight into the doc-
umentation being kept. The disadvantage is that this effort constitutes work
that is generally not done when SAS log files are saved. In reality, however,
anyone who has dug through SAS log files to troubleshoot the execution of
a large batch of code might not consider this feature of R a disadvantage!

* Please note that the menu must be used to execute the Select All command, but after that,
Control-C can be used in lieu of Edit – Copy command.
† Users familiar with SPSS will consider this reminiscent of the output window in SPSS, which
is easy to save in an SPSS analysis, but hard to understand after the fact if the file is not some-
how reduced, rearranged, or annotated.
Differences between SAS and R 9

This is because actually organizing a log file into what is important and
annotating it builds in efficiency later when troubleshooting may be neces-
sary, especially when a programming team is involved. Otherwise, it may
not be helpful to save the log file in the first place.
Finally, R handles graphical output files somewhat similarly to PC SAS in that
it opens and displays graphics in a separate window. The window automatically
opens when graphics requests are run in the console, and the window is titled
R Graphics. Once this window is open, the user can select it (which changes the
menu options) and choose Save As, where the user is presented a list of formats
to choose from, including the popular *.pdf, *.png, and *.jpg. If *.jpg is chosen, the
user is offered the selection between 50%, 75%, and 100% quality.
R’s handling of graphics represents a strong advantage over SAS, because in
SAS, the output function for graphics always involves the rigid and complex
“output delivery system,” or ODS, which produces files that generally need
to be post-processed in another program to be made even minimally pre-
sentable for journal publication or even presentation in an informal meeting.
In SAS, the alternative to this requires advanced expertise on the part of the
programmer to set options in SAS to modify the output as SAS commands
are executed so the resulting graph does not need extensive post-processing.
With R, this entire complexity is avoided, and even a basic graphics editor
such as Paint can be used to manually add detail to *.png and *.jpg graph-
ics files saved from R in this way. Further, as will be demonstrated later in
this book, but most specifically in Chapter 2, little expertise is required to
add elements through programming code to graphical output in R, such as
adding labels to x and y-axes, whereas these functions require much more
programming ability when done in SAS.

Code Documentation and Metadata: SAS versus R

A classic common challenge in any statistical program when working with large
datasets to complete a complex, long-term health data analytics project involv-
ing an interdisciplinary team, a lot of data, a lot of code, and a lot of time, is
entropy.* A part of this entropy that can drive the programmer to distraction is
having trouble keeping track of the meaning, content, and method of generation
of both the native variables (the variables present in the original data file used in
the analysis) and the new variables the programmer creates through code that
edits data. For this reason, in both SAS and R as well as any other statistical program,
the following modern process for maintaining code files is recommended:

1. Code files should be relatively short and focused on only one func-
tion (e.g., reading in a dataset, running histograms, running fre-
quencies, running regressions).

* A close colleague has referred to this phenomenon as “biostatistics paper hell.”

10 Analyzing Health Data in R for SAS Users

2. Code files should have a numerical prefix that causes the code files
to line up in the order in which they should be run, followed by a
shorthand label indicating the function of that code (e.g., 100_Read
in data, 105_Run summary statistics, 110_Create age group variable).
3. Advanced users can designate logic and naming conventions with
respect to these prefixes (code starting with 00n indicates pre-
processing, code starting with 10n indicates code that edits data,
code starting with 20n indicates analytic code that does not edit
data, code starting with 70n indicates exploratory code).
4. Prefixes should not be named in increments of 1 to allow for insertion
of code in between code files at a later date (e.g., naming code “100_
read in data,” with the next code being named “105_create age group
variable” will allow for the possibility of insertion of code “103_*”
after the fact, in case minor post-processing is found to be required
after reading the data in).

An analogy can be made to making a movie. Although viewers perceive the

movie as happening in chronological order, the director and actors are aware
that the scenes are actually not shot in chronological order. As a data “director,”
the programmer may develop “scenes” (programming commands) for later
in the “movie” (such as regression code), only to find that to make the story
straight and clear, she has to redo scenes from earlier in the movie (such as
generating a new variable in the analytic dataset). This way, if the age group
variable she created has an error or needs to be changed, she can go back to
the code earlier in the “movie,” edit this code properly, then start from the
beginning and run the whole “movie” of code from the beginning to the pres-
ent state to make sure the story hangs together after the edit.
Along with this principle, the importance of prolific and well-organized
comments in code, in both SAS and R as well as any other statistical program,
cannot be overstated. In SAS, commenting is done in between bookends of
/* and */ (e.g., /*This is a comment*/) or by simply using * at the beginning
of the line (e.g., *This is a comment). R uses the second approach to com-
menting (e.g., #This is a comment), but not the first. Therefore, although old
hands at SAS might be familiar with SAS banners that precede the code in
code files, these are not used in R.
This discarding of banners before code actually represents an innovation;
modernly, SAS banners (as well as banners in other statistical code) are to
be strongly discouraged in health data analysis. Instead, the information that
would have belonged in the SAS banners of the 1980s and 1990s programming
styles should now be placed instead in an outside program that is more accessi-
ble (such as Microsoft Word or Excel) and follow an organized standard. Banners
contain information that is human-generated, such as notes on interpretation of
variables or variances from standards in programming, not information that can
be automatically generated, such as a list of field names and attributes.
Differences between SAS and R 11

Many long-term SAS programmers have not learned this new principle,
and this inhibits clear communication and discussion about the nature of
the data, code, and analysis, as non-programmer managers and subject mat-
ter experts in other fields in healthcare and science cannot easily view this
metadata reported in the SAS banners. Hence, today, SAS banners should
be avoided in code, and R code should also not include banners, and should
only include prolific, carefully-worded and carefully-placed comments.
While some programmers shun excessive commenting, given the modular
approach recommended to developing code files, many code files are short,
so the comments do not complicate comprehension of the code.
Additionally, datasets themselves should not be documented in actual
code anymore, and all dataset metadata should be developed and made
available in easily accessible word processing and spreadsheet applications
(such as Microsoft Word and Excel) or else in the less-recommended PDF
format.* This means that SAS “labels” and “formats” should no longer be
used, and instead, this information should appear in a data dictionary docu-
ment, preferably in a spreadsheet format. An excellent example of publicly
available metadata that is presented in optimal format is the set of metadata
developed for the US Military Health System Data Repository, the MDR [19].†
The MDR’s use of this approach to metadata ensures that those receiving
MDR data files can study them adequately before attempting to read them
into any statistical software. This excellent set of metadata also provides
researchers from all fields necessary documentation to aid them in develop-
ing intelligent and informed data requests prior to receiving the data.
SAS programmers have historically embedded the metadata of variable
meanings, meanings of various categories in categorical variables, or infor-
mation about complex code functions that require explanation in their actual
SAS files, not in external metadata. This means that the actual file of SAS
code that assigns labels to variables in a dataset, and SAS FORMAT code
that labels the meanings of various levels of a categorical variable, would
sometimes be the only documentation of what these variables mean. This
effectively limits the ability of an interdisciplinary team to use the metadata
and complicates the interdisciplinary communication required to complete a
health analytics project successfully. It also essentially quashes the prospect
of group troubleshooting if multiple programmers are using different types
of statistical software.
On the other hand, using the methods described here of code organiza-
tion, naming conventions, and metadata maintained external to the statisti-
cal program, the programmer can not only reduce the time and complexity
associated with troubleshooting but also make sure that such a “movie” can

* PDF format is harder to navigate, copy, and paste, compared to Microsoft Word and Excel.
† For an excellent example, the reader is encouraged to download the M2 data dictionary from
the MDR web site, which is formatted in Microsoft Excel, and browse the different tabs to
better understand the optimal format of metadata described in this section.
12 Analyzing Health Data in R for SAS Users

be watched and understood by an outside programmer. This is especially

important for scientific writers who may be lucky enough to have their work
accepted in a top-tiered journal, such as the Journal of the American Medical
Association (JAMA) or the New England Journal of Medicine (NEJM). If this
occurs, the journal requests the code and data from the analyst and hires an
independent statistician to verify the results from the analysis. Following
these coding and documentation conventions can make this transfer of
knowledge accurate and easy. If the programmer finds herself is such a situ-
ation, without much complexity, she can provide the metadata files as well
as an organized compendium of code to the verifying statistician with an air
of professionalism while leaving little room for miscommunication during
the transfer process.

Handling of Data
This section begins with a short description of some issues with SAS data
handling that must be understood to make an informed comparison with
R data handling. Next, features of R data handling will be described and
contrasts between SAS and R data handling will be highlighted. Third, basic
differences in SAS and R code syntax will be explained, followed by a dis-
cussion about how to approach the thought of SAS macros when developing
code in R. Finally, an explanation of approaches in R equivalent to SAS labels
and SAS formats will be described.

A Focus on SAS Data Handling

In order to effectively contrast SAS data handling with R data handling, it
is extremely important to gain at least a working knowledge of the issues
inherent in SAS data handling. This is because SAS data handling is par-
ticularly complex, and therefore is the topic of many SUG-initiated white
papers [6,14].
The main reason SAS data handling is so commonly discussed that it is
more complex than that of other modern programs, and relies on the user
developing extremely efficient code using SAS’s commands. SAS’s current
yet historical “data step” language (which is used for data editing in SAS) is
unnecessarily restrictive and complex simply due to lack of modern innova-
tion which is now possible.* Even in the current version, SAS executes data

* The monopoly status of SAS has relieved much pressure on the company to improve the
product in this way. Also, those who are data step masters may feel their positions threat-
ened by users of simpler data editing software, so they inadvertently collude with SAS by
continuing to study, receive certifications in, and promote the use of SAS data step language.
Differences between SAS and R 13

steps by reading row by row (observation by observation) and executing

code in the data step to that particular row. SAS then walks to the next quali-
fying row and performs the set of code again, being stuck in an implied loop
[20]. This means that any indexing in SAS is most efficiently done by simply
sorting the SAS dataset by the indexing variable before executing a function.
In other words, when executing data step code on records where a field is a
certain value, sorting by that field will make the execution of the SAS code
go faster. Sorting may seem easy, but the sort function itself is inefficient
and can take enormous amounts of time depending on the dataset and SAS
implementation [21].
Aside from the complexity of the data step code, the second main reason
that data step functions are discussed ad nauseum in SAS white papers is
that to overcome the inefficiency of these data step functions, the programmer
needs to craft optimal code to ensure data management functions execute in
a reasonable timeframe as is generally required of a “procedural” language.†
Crafting this optimal code requires much study of data step language and
function. Compare this situation to that with structured query language
(SQL), which is a “declarative” rather than “procedural” language. This
means that the user declares what she wants returned in her query output
with relatively simple code, and a query optimizer running in the back-
ground automatically generates an evidence-based optimized execution
plan based on known dataset information such as the performance of prior,
recent queries. Thus, the query executes according to an optimal plan devel-
oped on-the-fly as a custom algorithm from the query optimizer, taking into
account all information known by the optimizer about the query at the time
(such as location of certain types of data within the dataset being queried).
When the process is described this way, one does not need to make a leap
to realize how much more efficient developing data editing code is in SQL,
which relies on a well-designed optimizer, when compared to programmer-
developed data step language in SAS, which is more prone to inefficiencies.
In reaction to this expanding realization among SAS users, SAS developed
PROC SQL, which is not really a PROC but a “language within a language.”
While adopting the characteristics of SQL and being much more intuitive and
easier to use than SAS data step language, SAS PROC SQL is prohibitively
inefficient in processing and cannot be used on large datasets. This is due
to an inefficient optimizer that results in extremely slow-executing queries,
leading disappointed SAS users back to the data step drawing board [22].

Comparison to R Data Handling

Before making a direct comparison between R’s and SAS’s handling of datas-
ets, it is necessary to take one step back and look at the “forest” of R and SAS

† Note that the nomenclature of “PROCs” in SAS point to its procedural processing feature.
14 Analyzing Health Data in R for SAS Users

rather than the “trees” of specific functions in R and SAS. SAS was developed
expressly for data analysis, meaning importing large datasets and running sta-
tistical analyses on them. SAS has certainly developed other functions—such
as data editing commands, commands to execute calculations, and commands
to set global variables—but these were all developed with the endgame of a
statistical analysis in mind. For this reason, although SAS can execute functions
like defining global variables, SAS’s core processes largely revolve around the
manipulation of datasets and executing statistical calculations on them.
R, on the other hand, was not designed expressly to handle datasets.
Although it is excellent at this function, it is also used for complex math-
ematical calculations, and is deployed extensively in the field of engineer-
ing. Instead of having the separate, specially-designed functions expressly
for handling datasets that SAS has—data steps, PROCs, macros, and other
SAS functions—R has generic functions for handling not only datasets but
also all the other types of objects in R [20]. Although SAS users tend to think
of programming in terms of dataset editing and analysis, it is important to
adopt a different perspective when approaching R for these functions. It is
more useful to consider R as a multifaceted program with an extensive set
of functions for the user to apply, and the health data analyst chooses to use
the functions helpful for dataset editing and analysis. This book is a guide to
choosing these functions.
Here are examples as to where this difference plays out. In SAS, health
analysts work with datasets and also arrays—which are a defined group of
columns in a SAS dataset. In R, there are other objects. The main R object data
analysts use is a “data frame,” which is similar to a SAS dataset. However, in
R, it is possible to make a “vector,” which is a single column of variables all
of the same type, a “matrix,” and a “list” (collection of objects) [20]. This ulti-
mately is an advantage if the R user learns how to make use of these other
objects to assist in data analysis, as they are not available to the SAS user.
SAS can be used in an object-oriented way, but this is not typically done
in health analytics. However, R is completely object-oriented—which is why
it is so much easier to create and invoke objects in R compared to SAS [20].
This difference also sheds light on how R processes indexes differently than
SAS. When R is used for data analysis, the main object is a dataset—which
should be coerced to be a “data frame” object in R. The first column of the
R data frame needs to be a primary key, or a column where all the numbers
are unique. This column is designated as the “row.names” column in R. In
a published dataset, there is usually a column already existing that has this
attribute that can be used as the row.names column. However, if there is not
one, R generates one on loading, which is similar to the OBS or observation
count generated by SAS when loading a dataset.
These differences are reflected in basic differences in how code is formed
in R versus SAS, and those impact how the programmer should approach
designing code. The next sections will describe some high-level differences
in thought that need to be applied while designing a project in R versus SAS.
Exploring the Variety of Random
Documents with Different Content
 CHAPTER V
THE WOODWIND FAMILY

The Woodwind; the reed; the flute; the piccolo; the oboe; the cor anglais; the
bassoon; the double-bassoon; the clarinet; the basset-horn; the bass-clarinet.

The Woodwind Family consists of instruments that may be

described as wooden tubes, or pipes, through which the performer
blows, stopping the holes in these pipes with his fingers in order to
get various notes. Some of these are furnished with reeds and some
are without. It is easy for us to tell the difference when we look at
the Orchestra. The flutes are held horizontally and have no reeds. All
the reed instruments are held by the player in a straight line,
perpendicularly. The Reed Family is divided into two groups: the
oboe group, furnished with a double reed; and the clarinet group,
furnished with a single reed. This reed, single or double, placed in
the mouthpiece of the instrument, is the “speaking” part. Without it,
the instrument could not be played. The reed corresponds to the
sound-post of the violin.
The reed is made of the outer layer of a certain kind of grass that
grows in the south of Europe. Most of it is obtained from Fréjus on
the Mediterranean. The reed is very difficult to fit and the player is
very particular about it. If anything goes wrong with the reed, the
instrument makes a dreadful noise that is called the couac, or quack.
It is even worse than the wolf[19] on a stringed instrument.
In all woodwind instruments the embouchure is important. The
embouchure is a certain arrangement of the lips by which the
performer throws into the instrument all the breath that comes
through the mouth without losing any of it and without giving the
slightest hissing sound.
 THE FLUTE
If we listen attentively to any piece of orchestral music, we will
notice that the voice of the flute is rarely silent. Very often it doubles
the first violins in the melody, running along with them smoothly and
sweetly. Sometimes it plays an unobtrusive part of its own and every
now and then bursts out into a lovely and elaborate solo, when its
clear, silvery, liquid notes sound deliciously cool against the warm,
vibrant strings. The flute is one of the most agile and flexible
instruments in the whole Orchestra. The flute is the nightingale, the
thrush, the lark, the oriole, the mocking-bird of the Orchestra. It
warbles.
The voice of the flute is gentle; it is ethereal; it is heavenly; it is
pure; it is sweet; and it is soothing. Therefore, composers make use
of it for poetic and tender sentiment; for scenes of a religious
nature; and to suggest beautiful dreams. It is both graceful and
poetic and it induces reverie.
“To most persons,” Lavignac writes, “as to myself, the ethereal,
suave, transparent timbre of the flute, with its placidity and its poetic
charm, produces an auditive sensation similar to the visual
impression of the color blue, a fine blue, pure and luminous as the
azure of the sky.”
The flute is a long tube made in three pieces, or joints, as they are
called. The head is one-third the length of the tube; the body carries
the keys that produce the scale of D major; and, lastly, comes the
foot joint, or tail-joint. The flute is cylindrical and is made of wood or
silver. In the silver flute the head-joint alone is slightly conical. In the
side of the head there is a large opening, less than an inch below
the cork, and across this opening the performer blows his breath. On
the lower part of the flute are six holes to be stopped at will by the
first three fingers of each hand; and three or four levers on the
lowest joint furnish additional notes below the regular scale of the
instrument.
 The performer holds the instrument transversely, sloping
downward against the lower lip with the hole, through, or across,
which he is to blow turned slightly outward, so that the stream of
wind—the “air-stream” it is called—shall strike against the outer
edge of this hole. The left hand takes the position nearest the
player’s mouth. Four open keys are closed by the first, second and
third fingers and thumb placed at the back of the instrument. The
little finger touches an open key, G-sharp, or A-flat. On the right
hand joint are three open keys for the first, second and third fingers,
with the accessory, or “shake,” keys. The right little finger takes the
closed key of D-sharp and the two open keys of C-sharp and C. The
G-sharp key is open in some flutes, but generally G-sharp closed key
is used by flute-players.
 FIRST FLUTE, SYMPHONY SOCIETY OF NEW YORK

George Barrère

The flute has no reed. Instead of a reed the “air-stream” from the
player’s lips, thrown against the sharp edge of the hole obliquely,
produces the sound-waves.
 The principle is that each note comes independently out of a
separate hole and speaks independently, just as if the rest of the
tube were cut off. All keys are open with the exception of G-sharp,
E-flat keys—and also the two small trill-keys.
Formerly the flute had no keys, or levers. It merely had finger-
holes; but between the years 1832 and 1847 Theobald Boehm, a
German, by following some experiments made by Captain Gordon of
Charles X’s Swiss Guards, worked away until he developed a system
of keys, manipulated by means of levers. His invention was so
successful that the player has now command of more holes; and, by
means of this system, it therefore became possible to play in every
key.
The flute stands in this one scale of D-major, so the only way to
get higher notes depends upon the breath and lips of the player.
“This is the eternal question,” says George Barrère, “playing upper
octaves does not require mere blowing as we can play forte in lower
octave and pianissimo in the upper. The real means of playing upper
is lips. It is not a secret; but how many flute-players ignore it,
making the flute the most disagreeable instrument to hear!” A good
embouchure, as the whole manipulation of the mouth is called, is
essential to artistic flute-playing. Moreover, the fingers must be
raised at equal heights—and not too high.
The player takes a calm, firm, easy, and often graceful, attitude
before his desk. Good flute-players also learn a great proportion of
their music by heart.
Staccato notes and ornamental passages are produced by “single
tonguing,” and “double tonguing,” and “triple tonguing.” For different
effects the player makes an effort to pronounce certain consonants,
k or t for example; but instead of pronouncing them he blows them
off his tongue in a little kind of explosion. But all this is done quickly
and with ease by a virtuoso.
The tones of the first octave are rather faint; those of the second
octave, produced by exactly the same fingering as those of the first
and with a stronger blowing of the breath, are stronger; and those
of the third octave, also produced by the same fingering, are more
penetrating.
Boehm’s explanation is worth quoting to help us understand the
production of tone. He says: “The open air-column in a flute’s tube is
exactly comparable to a stretched violin string. As the string is set
into vibration by the bow, the air-column in the flute is set into
vibration by the blowing of the performer’s breath and management
of the lips. As the clear quality of tone of a violin depends upon the
proper handling of the bow, so the pure quality of tone of a flute
depends upon the direction of the ‘air-stream’ blown against the
edge of the mouth-hole.
“Each octave requires a different direction of the ‘air-stream’; and,
by increasing the force of the breath, the tone is increased. By ‘over-
blowing,’ each tone can be made to break into higher tones.”
Older composers seem to have cared very little for the flute. They
did not have the modern improved Boehm flute. They found that the
performer often played out of tune. Cherubini said: “The only thing
in the world that is worse than one flute is two.” Many agreed with
him. However, Haydn wrote a trio for flutes in his oratorio of The
Creation and Handel wrote a beautiful obbligato for it in the aria,
“Sweet bird that shun’st the noise of folly,” in Il Penseroso, where it
imitates the bird. Bach wrote six Sonatas for the flute.
Handel’s aria, “O ruddier than the Cherry,” in Acis and Galatea,
now played on the piccolo, was originally written for the flute.
Mozart wrote two concertos for the flute and one for flute, harp
and orchestra. It is very evident in the Magic Flute.
A solo passage in Beethoven’s Leonora Overture, No. 3, is very
famous. In Beethoven’s Pastoral Symphony it impersonates the
nightingale.
Mendelssohn loved the flute dearly. It is very important in his
Midsummer Night’s Dream music. It plays lovely sustained chords in
the Overture, a beautiful part in the Nocturne and the Scherzo
contains one of the most celebrated passages ever written. He also
gave it an exquisite obbligato in the quartet O Rest in the Lord in the
oratorio of Elijah.
Wagner has a fine part for the flute throughout the Meistersinger;
it is important in the Largo of Dvořák’s New World Symphony, where
it plays with the oboe; Liszt made it conspicuous in his Hungarian
Rhapsodie, No. 2; and it sings in the Morning of Grieg’s Peer Gynt
Suite.
Berlioz and Tschaikowsky both played the flute themselves and,
naturally, their works are full of beautiful melodies for this
instrument. Berlioz calls for two flutes and harp for the Cris des
Ismaëlits in L’Enfance du Christ. Tschaikowsky’s symphonies are a
delight to the flute-player. An exceptionally striking use of the flute is
in the Danse des Mirlitons and Danse Chinoise in the Nut-cracker
Suite.
Richard Strauss, who always goes a little farther than anybody
else, has in the “Windmill” number of the Don Quixote Variations
called for the “flutter tongue,” a new way of rolling the tongue. The
name describes it.
Last, but not least, we must recall Gluck. What could be more
beautiful than his use of the flute in Armide, unless it is to be found
in the music of Orfeo? All through that beautiful opera the plaintive,
tender voice of the flute is conspicuous. Not only does it play
melodies for the enchanting ballets and minuets, but its wailing
notes tell us of the grief of Orpheus for his adored Eurydice; and
when we arrive in the Elysian Fields with Orpheus its pure and
ethereal voice, heard in a solo of ravishing beauty, lifts us out of the
everyday world we live in and transports us into a realm of blissful
peace and enchanting beauty.
 FREDERICK THE GREAT PLAYING THE FLUTE

With his Orchestra at Sans Souci

In early days the flute was played by holding it straight in front

and not horizontally as shown in the picture facing page 74. The
German, Quantz, did much to bring the horizontal flute into fashion.
One of his most enthusiastic pupils was Frederick the Great, King of
Prussia, who is shown in the picture facing this page playing a flute
concerto with his orchestra at Sans Souci, taken from an engraving
by Chodowieki. The King’s favorite greyhounds are the only listeners.
Franz Benda is the first violin and Christian Friedrich Fasch, who
succeeded Philipp Emanuel Bach (son of J. S. Bach), is playing the
harpsichord.
Modern compositions usually call for two flutes and a piccolo.

THE PICCOLO
 The piccolo is the little flute. Properly, it should be spoken of as
the piccolo flute, for just as we have seen in the case of the
violoncello the word ’cello means little or small, so the word piccolo
is an adjective and not a noun. However, people speak of it simply
as the piccolo. The piccolo plays the upper octave of the flute. It is
less than half the length of the flute and it lacks the “foot-joint.” Its
compass is over two octaves. Almost every piccolo player can play
high B and even C. The music for the piccolo is always written in the
Treble Clef, an octave below the real pitch, that is to say an octave
below the real sound of the notes. The fingering and technique are
exactly the same as for the flute, so anything that can be played on
the flute can be played on the piccolo.
It should be noted here that two-thirds of the compass of the flute
plays within the compass of a high soprano; now, the piccolo, on the
other hand, is nearly always playing in a register higher than that of
any human voice. It is the most acute and piercing of all instruments
in the orchestra; for even the corresponding notes produced by
harmonics on the violin are far less shrill and penetrating. The
piccolo rarely plays in its lower register. Its second octave is bright
and joyous; but in this we hardly distinguish it from the flute. What
we do notice are the piercing upper notes in quick runs, in chromatic
passages and wild screams. Sometimes too, the piccolo can be made
to utter something of a diabolical nature.
The piccolo is often used to brighten the upper notes of the other
members of the Woodwind Family in all kinds of combinations. This
method of using the piccolo might be likened to brightening up an
article with gold leaf. We might say that the piccolo sometimes adds
a sort of gilt edge to the melody.
Berlioz liked to use it in this way for additional ornamentation, so
we hear a great deal of this kind of piccolo gilt-edging, as we might
call it, in his works. Berlioz gave it a great deal of thought. “In pieces
of a joyous character,” he wrote, “the sounds of the second octave
are suitable in all their gradations; while the upper notes are
excellent fortissimo for violent and tearing effects: in a storm, for
instance, or in a scene of fierce, or infernal, character. Thus the
piccolo flute figures incomparably in the fourth movement of
Beethoven’s Pastoral Symphony—now alone and displayed above the
low tremolo of violins and basses, imitating the whistlings of a
tempest whose full force is not yet unchained—now on the higher
notes still, together with the entire mass of the Orchestra. Gluck in
the tempest of Iphigénie en Tauride has known how to make the
high sounds of the piccolo flute in unison grate still more roughly by
writing them in a succession of sixths, a fourth above the first
violins. The sound of the piccolo flutes issuing out in the upper
octave, produces, therefore, a succession of elevenths with the first
violins, the harshness of which is here of the very best effect.
“In the chorus of the Scythians, in the same opera, the two
piccolo flutes double in the octave the little grouped passages of the
violins. These whistling notes mingled with the ravings of the savage
troop, with the measure and incessant din of the cymbals and
tambourine make us shiver.
“Everyone has remarked the diabolic sneer of the two piccolo
flutes in thirds in the drinking-song of Freischütz. It is one of
Weber’s happiest orchestral inventions.
“Spontini in his magnificent bacchanalian strain in the Danaïdes
(since become an orgy chorus in Nurmahal) first conceived the idea
of uniting a short piercing cry of the piccolo flutes to a stroke of the
cymbals. The singular sympathy, which is thus created between
these very dissimilar instruments, had not been thought of before. It
cuts and rends instantaneously, like the stab of a poignard. This
effect is very characteristic—even when employing only the two
instruments mentioned; but its force is augmented by an abrupt
stroke of the kettledrums joined to a brief chord of all the other
instruments.
“Beethoven, Gluck, Weber and Spontini have thus made ingenious
use—no less original than rational—of the piccolo flute. But when I
hear this instrument employed in doubling in triple octave the air of
a baritone, or casting its squeaking voice into the midst of a religious
harmony, or strengthening and sharpening—for the sake of noise
only—the high part of the orchestra, I think it a stupid method of
instrumentation.
“The piccolo flute may have a very happy effect in soft passages;
and it is mere prejudice to think that it should only be played loud.
Sometimes it serves to continue the scale of the large flute by
following up the latter and taking high notes beyond the flute’s
command. The passing from one instrument to the other may then
be easily managed by the composer in such a way as to make it
appear that there is only one flute of extraordinary compass.”
Handel used an instrument that corresponded in his day to the
piccolo in his wonderful accompaniment to the bass song, “O ruddier
than the Cherry,” in Acis and Galatea, where he gives it a pastoral
character. He also makes it play an obbligato in the aria, “Hush ye
Pretty Warbling Choir,” in the same cantata. He does the same thing
again in the aria, Auguelletti che cantate, in Rinaldo. Meyerbeer
gives it much to do in his infernal waltz in Robert le Diable; and in
Marcel’s song “Piff Paff” in Les Huguenots it adds brilliancy to the
martial effect. Beethoven has a striking place for the piccolo in the
finale of his Egmont Overture; Verdi makes it heard in Iago’s
drinking-song in Otello; it is conspicuous in the grotesque dances of
the dolls in the ballet of Coppelia by Delibes; Wagner uses it in his
storms, in the Ride of the Walküre, and in all his fire-music in the
Nibelungen Ring; Strauss gives a peculiar trill for it in Till
Eulenspiegel; and Berlioz gives it full play in his Carnaval Romain,
op. 9; and in the Minuet of the “Will o’ the Whisps” of his Damnation
of Faust he calls for three piccolos.
Therefore, we might characterize the piccolo as the imp, or
demon, of the Orchestra, or the flash of lightning, or the darting
flame, or the whistling wind.

THE OBOE
 The Oboe, like the violin, comes from a family of long ancestry. It
goes back to ancient Egypt, Assyria and Greece. In the Middle Ages
this family was known as the Bombardo, Bombardino, Bombardi, or
Chalumeau. The Germans called this family Pommers, which seems
to be a corruption of Bombardi.
“The Bombardo,” writes Carl Engel, “was made of various sizes
and with a greater or smaller number of finger-holes and keys. That
which produced the bass tones was sometimes of enormous length
and was blown through a bent tube like the bassoon, the invention
of which it suggested. The smallest instrument, called chalumeau
(from calamus, a reed) is still occasionally to be found among the
peasantry in the Tyrol and some other parts of the Continent. The
Germans call it Schalmei and the Italians piffero pastorale. In
England it was formerly called shawm, or shalm.”
The type of these instruments was a conical tube of wood with a
bell at one end and a bent metal tube at the other containing a
double reed mouthpiece. There was a quartet of them; and the
oboe, or hautbois (high-wood) was the treble. There was also an
oboe d’amore[20] and an oboe di caccia, hunting oboe (from which
the cor anglais is supposed to have been derived), and there were
many others. Old writers refer to them as chalumeau and schalmey
and shawm; and in such a general and confused way that it is hard
to know just which special instrument they are talking about. These
old oboes are called for in Bach’s scores; but they began to drop out
of use in his time. We know, however, that chalumeau was the
instrument of the old reed band that always played the melody and
that from it sprang the oboe of to-day. The instrument went through
many changes before it reached its present condition; but none of
these affected the family voice—the penetrating, roughish twang.
The Bombardino-Schalmey voice still persists: it is like some other
famous family traits—the Bourbon nose and the Hapsburg lip for
instance—it is hard to suppress. However, it is this peculiar voice
that makes the oboe such a desirable member of the orchestra.
 “The timbre is thin and nasal, very piercing in its forte passages,
of exquisite refinement in its piano passages; harsh and of bad
quality in its very high and very low notes. The oboe is artless and
rustic in its expression; it is pastoral and melancholy; if it is gay, its
gayety is frank and almost excessive and exaggerated; but its
natural tone is of a gentle sadness and a resigned endurance. It is
unrivalled in depicting simple, rural sentiments of any kind, and on
occasion can even become pathetic.”[21]
 FIRST OBOE, SYMPHONY SOCIETY OF NEW YORK

Henri De Busscher

The oboe is the most elaborate and complicated of all the reed
instruments. The mechanical changes are due to Apollon Marie Rose
Barret (1804-1879), a remarkable French oboe-player, aided by a
French instrument-maker named Triébert. Historically and musically
the oboe is the most important member of the reed band. It is first
of all a melodic instrument; or, in other words, its tone quality is
what it is especially valued for and not for brilliant passages. It can
call up pastoral scenes and it can express innocence, grief, pathos
and gentle gayety.
The oboe is a wooden pipe, or tube, with conical bore widening
out gradually until it forms a small bell, shaped something like the
flower of a morning-glory, or convolvulus. At the opposite end it has
a small metal tube, or mouthpiece, called “staple,” to which the reed
(consisting of two blades of thin cane) is attached by means of
silken threads. Along the wooden pipe are two “speaker keys,”
worked by metal rods called “trackers.” This reed is the speaking
part of the instrument.
The oboe is made in three pieces,—the head-piece, bottom and
bell-joints. The player first screws the joints of his instrument
together so that the finger-holes are in a straight line, and then he
puts the reed in the head-piece. The first, second and third fingers
of each hand are used to cover the holes. The whole instrument
rests on the thumb of the right hand. The little fingers and the
thumb of the left hand are used for the keys. The fingers are always
placed over the finger-holes ready to close them when necessary.
The fundamental scale is obtained by opening and shutting the holes
pierced laterally in the pipe and these are governed by a mechanism
called “speaker keys.”
The scale of the oboe begins on the middle C and is chromatic.
The instrument, being conical, “over-blows” an octave. It runs up to
the extreme treble G. But although it is a soprano, it is a “middle
compass” instrument. Music for it is written from F on the first space
of the Treble Clef to D in octave above.
The fingering resembles that of the flute; but, owing to the reed in
his mouth, the player can only use single tonguing.
 The player puts the reed between his lips, taking care that his
teeth do not touch the mouthpiece. Then he places his tongue
against the open part of the reed, presses the reed with his lips,
draws his tongue gently backwards and blows a stream of air into
the oboe, managing his breath as if for singing. Sometimes he
pronounces the syllable doo and sometimes that of too, according to
the effect he wants to get.
The double reed in the player’s mouth is the sound-producer. The
air-column inside the pipe acts as a resonating medium,
strengthening the vibrations of the reed by vibrations of its own.
As the player is obliged to take his lips from the mouthpiece to
exhale, he cannot perform long sustained passages without pauses.
While the oboe does not require as much breath for blowing as
some instruments do, the difficulty is to exhale and refill the lungs so
as to go on with the rest of the work.
The notes are produced by holes, some open, others closed by
keys raised by means of levers. The oboe, like the flute, is an octave
instrument, that is to say it “over-blows” the octave. The oboe
possesses notes sufficient for an octave, or more, with chromatic
intervals. The next octaves are obtained by means of cross-fingering
and from the octave keys which do not give out an independent note
of their own but determine a node in the column of air and so raise
the pitch of any other note an octave. The oboe is a “non-
transposing instrument” and sounds the note written.
“It is possible to play on this instrument chromatic scales and
arpeggio passages; legato and staccato; leaps; cantabile passages;
sustained notes; diminuendo and crescendo; grace notes and
shakes.”
“The oboe,” writes Berlioz, “is especially a melodic instrument. It
has a pastoral character full of tenderness,—indeed I might say
timidity. Candor, artless grace, soft joy, or the grief of a fragile being
suits the oboe’s accents: it expresses them admirably in its cantabile.
 “A certain degree of agitation is also within its powers of
expression; but care should be taken not to urge it into utterance of
passion—a rash outburst of anger, threat, or even heroism; for then
its small acid-sweet voice becomes uneffectual and absolutely
grotesque.
“Gluck and Beethoven understood marvellously well the use of this
valuable instrument. To it they owe the profound emotions excited
by several of their finest pages. I have only to quote from Gluck’s
Agmemnon’s air in Iphigénie en Aulide ‘Can the harsh Fates?’ These
complaints of an innocent voice, these continued supplications, ever
more and more appealing, what instrument could they suit so well
as an oboe? And the celebrated burden of the air of Iphigénie en
Tauride, ‘O Unhappy Iphigénie!’
“Beethoven has demanded more from the joyous accent of the
oboe. Witness the solo of the Scherzo of the Pastoral Symphony and
also that of the Scherzo of the Ninth Symphony, also that in the first
movement of the Symphony in B-flat. But he has no less felicitously
succeeded in assigning them sad, or forlorn, passages. These may
be seen in the minor solo of the second return of the first movement
of the Symphony in A-major in the episodical Andante of the finale
to the Eroica Symphony; and above all, in the air of Fidelio, where
Florestan, starving to death, believes himself, in his delirious agony,
surrounded by his weeping family and mingles his tears of anguish
with the broken sobs of the oboe.”
In Beethoven’s Pastoral Symphony the oboe impersonates the
quail and in Haydn’s Seasons it imitates the crowing of a cock in a
long and difficult passage. Perhaps the most beautiful use of the
oboe in all music is in Gluck’s opera of Orfeo, in which it plays an
exquisite minuet with the flute and a beautiful ballet with the violin.
Schubert uses it charmingly in the second movement of his
Symphony in C-major.

COR ANGLAIS
 The cor anglais, or English horn, differs slightly in appearance
from the oboe; but these differences help us to identify it. In the
first place, it ends in a kind of ball; and in the second place, there is
a bent crook at the other end that holds the mouthpiece containing
the double reed. It is supposed that the word Anglais is a corruption
of the word anglé, meaning bent; for in olden times this instrument
was bent at an obtuse angle in the middle of the tube. It is,
therefore, more correct to call it cor anglais than English horn. The
English have had nothing whatever to do with the development of
the instrument.
The cor anglais is nothing more or less than the alto, or tenor,
oboe. It has the same scale and compass as the oboe; but it stands
in the key of F, a fifth below that of the oboe. It is, however, unlike
the oboe, a “transposing instrument,” that is to say, the music does
not represent the real sounds. In the case of the cor anglais the
music is written in a key a fifth above the real sounds. Any good
oboe player can play the cor anglais, because the technique and
fingering are practically the same.
“Its tone,” says Lavignac, “is essentially sad, melancholy,
sorrowful. The cor anglais exactly suits the expression of mental
suffering, which is, therefore, especially characteristic of it.” “Its
quality of tone,” says Berlioz, “less piercing, more veiled and deeper
than that of the oboe, does not so well as the latter lend itself to the
gayety of rustic strains. Nor could it give utterance to anguished
complainings. Accents of keen grief are almost beyond its powers. It
is a melancholy, dreamy, and rather noble voice, of which the
sonorousness has something vague and remote about it which
renders it superior to all others in exciting regret and reviving
images and sentiments of the past when the composer desires to
awaken the secret echo of tender memories. In compositions where
the prevailing impression is that of melancholy the frequent use of
the cor anglais hidden in the midst of the great mass of instruments
is perfectly suited.”
 The cor anglais has been called “an oboe in mourning.” Perhaps
that will give the best idea of its sorrowful voice.
The cor anglais came directly from the alto pommer of the
Schalmey-Pommer family.[22] Most probably the oboe di caccia, or
hunting oboe, was its immediate ancestor. A very good reason for
thinking this the case is because in Rossini’s Overture to William Tell
the “Ranz des vaches” (calling the cows) was originally given to the
oboe di caccia, which was still in use in Rossini’s time; and when the
oboe di caccia became obsolete, the part was taken by the newer
cor anglais.
The cor anglais and the oboe assumed their modern appearance
about the same time. Both instruments were much changed in
construction and mechanism during the last hundred years; but both
instruments kept the old family voice, which has a curious harsh
quality combined with plaintiveness.
Beethoven wrote a Trio for two oboes and cor anglais, op. 29. The
French composers made it popular. Meyerbeer has it play an
obbligato to the aria “Robert, toi que j’aime,” in Robert le Diable;
Berlioz made it important in his Symphonie Fantastique; and it
appears in Dvořák’s New World Symphony, having a melody in the
Largo with accompaniment of strings con sordini. Strauss gives it
prominence in Heldenleben.
 COR ANGLAIS, SYMPHONY SOCIETY OF NEW YORK

Attilio Bianco

Of its famous solos none is so haunting as the plaintive part in Act

III of Tristan and Isolde. Here the long, sad melody heard on the
Shepherd’s pipe is entrusted to the saddest voice in the orchestra,—
the cor anglais.

THE BASSOON
The bassoon is the bass of the oboe group, holding the same
place in this family that the violoncello does in the String Family. It is
a descendant of the old bass pommer, the bass of the Schalmey
Family; but in the various transformations that took place between
1550 and 1600 the characteristic Schalmey family voice disappeared
in the bassoon. The tone-color of the bassoon is quite unlike that of
the oboe and that of the cor anglais, although it is played with a
double reed.
The bassoon is a pipe, or tube, eight feet long conically bored and
turned back upon itself so as to reduce its length to about four feet.
The instrument consists of five pieces: (1) the bell; (2) the bass, or
long joint; (3) the double joint; (4) the wing; and (5) the crook, a
small curved tube of metal which holds the mouthpiece with the
double reed. The bottom of the instrument is stopped by a flattened
oval cork. The pipes meet at the double joint and turn upward. The
holes are pierced obliquely so as to bring them within reach of the
player’s fingers. There are three holes in the wing-joint and three
others in the front of the double joint, to be closed by the first three
fingers of each hand. A single hole on the back of the double joint is
for the thumb of the right hand. The little finger of the right hand
touches two keys; and a series of interlocking keys is on the bass, or
long, joint producing the lowest notes of the scale for the left thumb
to work.
The player holds the instrument diagonally in the hollow of his two
hands, with the left hand uppermost at the level of his breast, and,
of course, nearest the bell of the bassoon. The right hand is placed
below and behind his right thigh. The double joint of the bassoon
rests against the player’s knee. The bell of the instrument points
upward.
 The bassoon stands in the key of G-major and plays an octave
lower than the oboe. Its compass is three octaves and a half, the
lowest note being B-flat. The music for it is written in the Bass Clef
and in the Tenor Clef for the highest notes. Like the flute and oboe,
its deep notes are its fundamental tones; those of its middle register
are second harmonics; and those of its highest register are third,
fourth and fifth harmonics. The fingering is the same for all octaves.
The higher notes are produced by “over-blowing,” so that the air-
column in the instrument vibrates differently according to the way
the player directs his breath.
“Its lowest tones,” writes Lavignac, “are solemn and pontifical, like
an organ pedal. Its medium register has a sweet sonority of some
richness but little strength; and its high register has the most
expression, but is painful, distressed and dejected. At the same time
this instrument has comic possibilities. In the medium, and lower
registers certain staccato notes which have been often used have a
certain grotesqueness bordering on awkwardness.”
“The bassoon was first used,” says Dr. Stone, “in Cambert’s
Pomone, Paris, 1671; but it has gradually risen to the position of a
tenor, or even alto, frequently doubling the high notes of the
violoncello, or the lower register of the viola. The cause of the
change is evidently the greater use of bass instruments, such as
trombones and ophicleides, in modern orchestral scores on the one
hand and the improvements in the upper register of the bassoon
itself on the other. There is a peculiar sweetness and telling quality
in these extreme sounds, which has led to their being named ‘vox
humana notes.’ We have good evidence even in Haydn’s time that
they were appreciated; for in the graceful Minuet of his Military
Symphony we find a melody reaching to the treble A. Haydn uses it
as one of the most prominent voices of his orchestra.”
Until Mozart’s time the bassoon was little else but an instrument
for doubling the bass of the Strings; but Mozart did great things with
it. He even went so far as to write a Concerto for it. It is important in
his operas, particularly Don Giovanni; in his Requiem; and in his
Symphonies.
After Mozart fixed its place in the Orchestra, Beethoven brought it
forward and made its part so conspicuous and so elaborate that the
performers had to set to work to improve their technique.
“Beethoven never failed to employ it largely, reinforcing it, in some
cases, by the double-bassoon. The First Symphony is remarkable for
the assignment of subject, as well as counter-subject, in the slow
movement to first and second bassoons working independently; both
afterwards joining with the two clarinets in the curious dialogue of
the trio between strings and reeds. The Second Symphony opens
with a prominent passage for the bassoon in unison with the bass
strings; in the Adagio of the Fourth Symphony is an effective figure
exhibiting the great power of staccato playing possessed by the
bassoon; in the first movement of the Eighth Symphony, it is
employed with exquisite humor and in the Minuet of the same
Symphony it is entrusted with a melody of considerable length.
Perhaps the most remarkable passage in Beethoven’s writing for this
instrument occurs in the opening of the Finale of the Ninth, or
Choral Symphony, where the theme of the movement, played by
violoncellos and violins in unison is accompanied by the first bassoon
in a long independent melody of the greatest ingenuity and
interest.”[23]
Cherubini gave the bassoon a solo in his opera of Medea; Gluck
gave it a solo in some of his dance music in Orfeo; Rossini opens his
Stabat Mater with it; and Weber gave it much to do in his operas.
Weber wrote a Concerto for it and also an Andante and Hungarian
Rondo. Mendelssohn also was fond of it. Dr. Stone has well summed
up his use of this instrument as follows: “Mendelssohn shows some
peculiarity in dealing with the bassoon. He was evidently struck not
only with the power of its lower register, a fact abundantly illustrated
by his use of it in the opening of the Scotch Symphony and with the
trombones in the grand chords of the Overture to Ruy Blas, but he
evidently felt with Beethoven the comic and rustic character of its
tone. This is abundantly shown in the music to the Midsummer
Night’s Dream, where the two bassoons lead the quaint Clown’s
March in thirds and still farther on in the Funeral March, which is
obviously an imitation of a small country band, consisting of clarinet
and bassoon, the latter ending unexpectedly and humorously on a
solitary low C. In the Orchestra the bassoon also suggests the
braying of Bottom. It is worth notice how the acute ear of the
musician has caught the exact interval used by the animal without
any violation of artistic propriety.”
 BASSOON, SYMPHONY SOCIETY OF NEW YORK

Ugo Savolini

Modern composers have delighted in exhibiting the telling qualities

of the bassoon. A notable example is in Tschaikowsky’s Pathetic
Symphony and in the waltz movement of his Fifth Symphony. In his
Marche Slave it is very effective in unison with the violas.
Brahms shows it off well in his C-minor Symphony; Strauss in his
Heldenleben, Till Eulenspiegel and Don Juan; and Elgar in his Pomp
and Circumstance March and Variations III and IX of the Enigma.
Wagner gets desolation and sorrow out of it; and, occasionally,
humor; and Humperdinck makes comic use of it in Hänsel and
Gretel, where it frequently comments on what is happening on the
stage.
The bassoon gives long sustained notes, shakes and staccato
notes, which are “dry” and grotesque. The English and French name
bassoon and basson refer to its pitch in the bass; but the Italians
and Germans call it Fagotto and Fagott, because they think in shape
it resembles a bundle of sticks, or fagots.
To-day there are usually three bassoons in the Orchestra,—the
first and second bassoon and the double-bassoon.

THE DOUBLE-BASSOON
The double-bassoon is an octave below the bassoon. It doubles
the bass of the bassoon as the double-bass doubles the violoncello.
The double-bassoon is a conical wooden pipe of hard wood—often
maple—more than sixteen feet long and doubled back four times on
itself. The crook, or mouthpiece, into which the double reed is
fastened, is much like that of the bassoon; but the metal bell points
downwards.
Though the instrument is not a transposing one, the music is
written an octave higher than it sounds to avoid the use of ledger
lines. Its compass is from the middle C to the deep sixteen-foot C.
The double-bassoon was used in the Orchestra in Handel’s time.
Haydn calls for it in The Creation; Brahms, in his C-Minor Symphony;
Mendelssohn, in his Hebrides Overture; and Beethoven reinforces
the march in the Finale of his Fifth Symphony with it. He assigns it a
leading part in the Ninth Symphony.

THE CLARINET
The clarinet is also a descendant of the Bombardino-Chalumeau
Family of which, as we have seen, there were so many members.
The great difference between the ancestor of the oboe and the
ancestor of the clarinet was that the oboe’s ancestor was conical in
bore and played with a double reed and the clarinet’s ancestor was
cylindrical in bore and played with a single reed. That fact was the
parting of the ways and was destined to make all the difference in
the world. It would seem at first that the tone of the two old
chalumeaux (the one double-reeded and the other single-reeded)
was at first much alike; but as time went on and the single-reeded
chalumeau developed into the modern clarinet, the old rough, reedy
voice disappeared for a rich, warbling voice that has more of the bird
in it than of the reed.
“The clarinet is one of the most beautiful voices in the orchestra,”
Lavignac thinks. “It is the richest in varied timbres of all the wind
instruments. It possesses no less than four registers, perfectly
defined; the chalumeau, which contains the deepest notes and
recalls the old rustic instrument of that name; the medium, warm
and expressive; the high, brilliant and energetic; and the very high,
biting and strident. All these registers, thanks to the progress of
manufacture, are able to melt into one another in the happiest
manner possible and furnish a perfectly homogeneous scale. Almost
as agile as the flute, as tender as, and more passionate than, the
oboe, the clarinet is infinitely more energetic and richer in color.”
About 1690 Johann Christopher Denner of Nuremberg added the
twelfth key. He bored a small hole nearer the mouthpiece on a
chalumeau type of instrument, and made a key to it to be
manipulated by the thumb of the left hand. By this he increased the
compass of the instrument by more than an octave. It may be said
that from this date the clarinet came into existence. From the crude
instrument of two keys and seven holes has evolved the present-day
clarinet with seventeen keys and twenty-one holes, of which seven
are covered directly by the fingers and the others by the keys.
The clarinet is a cylindrical piece of wood, or a tube, about two
feet long, ending in a bell. It is made in sections: (1) mouthpiece;
(2) barrel joint; (3) left-hand, or upper joint; (4) right-hand, or lower
joint; (5) bell. The lowest note is emitted through the bell. The right-
hand thumb supports the instrument.
The reed is flat and the mouthpiece is curved backwards to allow
of vibration.
The reed is carefully thinned at the point where it vibrates against
the curved table of the mouthpiece. The vibration is caused by the
air pressure against the reed, thus engendering sound. The air-
column in the instrument is shortened, or lengthened, by the
opening, or closing, of the holes and keys, emitting high, or low,
sounds accordingly. The reed vibrates through the action of the air.
The lips of the player merely encompass the reed and mouthpiece,
slightly pressing the reed.
Again, to quote from Lavignac: “This instrument, the richest in
compass and in variety of timbre of all the wind instruments, is
subject to a very special and very curious law. Its tube is absolutely
cylindrical and open; and its column of air is set in vibration by a
single, flexible reed. Now a peculiarity in pipes of this construction is
that the vibrating segment forms, not at the middle point, but at the
end where the reed is, so the mode of subdivision of the air-column
is the same as if the pipe were stopped. The clarinet has, therefore,
only the harmonics of unequal numbers, which renders its fingering
very different from that of the flute, the oboe and the bassoon. It
would seem that this might place it below them. On the contrary,
this instrument lends itself with admirable suppleness to the
expression of all sentiments which the composer may wish to entrust
to it.
 CLARINET, SYMPHONY SOCIETY OF NEW YORK

Gustav Langenus

“Its compass, the greatest possessed by any wind-instrument

chromatically, has a great deal to do with giving it this richness of
expression; but the diversity of timbre belonging to its lower, middle
and higher registers must be regarded as the real superiority of the
clarinet.”
Berlioz says: “The clarinet is little appropriate to the Idyl. It is an
epic instrument, like horns, trumpets and trombones. The voice is
like that of heroic love. This beautiful soprano instrument, so ringing,
so rich in penetrating accents, when employed in masses, gains,
when employed as a solo instrument, in delicacy, evanescent
shadowings and mysterious tenderness what it loses in force and
powerful brilliancy. Nothing so virginal, so pure as the tint imparted
to certain melodies by the tone of a clarinet, played in the medium
by a skilful performer. It is the one of all the wind instruments which
can best breathe forth, swell, diminish and die away. Thence the
precious faculty of producing distance, echo, the echo of echo, and a
twilight sound. What more admirable example could I quote of the
application of some of these shadowings than the dreamy phrase of
the clarinet accompanied by a tremolo of stringed instruments in the
midst of the Allegro of the Overture to Freischütz! Does it not depict
the lonely maiden, the forester’s fair betrothed, who, raising her
eyes to heaven, mingles her tender lament with the noise of the
dark woods agitated by the storm? O Weber! Beethoven, bearing in
mind the melancholy and noble character of the melody in A-major
of the immortal Andante in his Seventh Symphony, and in order the
better to render all that this phrase contains at the same time of
passionate regret, has not failed to consign it to the medium of the
clarinet. Gluck, for the ritornello of Alceste’s air, ‘Ah, malgré moi,’ had
at first written it for the flute; but perceiving that the quality of tone
of this instrument was too weak and lacked the nobility necessary to
the delivery of a theme imbued with so much desolation and
mournful grandeur, gave it to the clarinet.
“Neither Sacchini, nor Gluck, nor any of the great masters of that
time availed themselves of the low notes of the instrument. I cannot
guess the reason. Mozart appears to be the first who brought them
into use for accompaniments of a serious character, such as that of
the trio of masks in Don Giovanni. It was reserved for Weber to
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!

textbookfull.com