LaTeX - Companion PDF

This is an updated version of chapter 8 of the LATE X
Companion reflecting changes in the AMS-LATEX

distribution that made parts of this chapter obsolete. It is
based in structure and content on the original
documentation in the Companion with obsolete parts
corrected but otherwise unchanged and should not be
considered as part of a general revision of this book.
Chapter 8 of "The LaTeX Companion", updated for AMS-LaTeX version 1.2 (Sep. 1st 1997).
Copyright © 1994-97 by Addison Wesley Longman, Inc. All rights reserved.
C H A P T E R 8
Higher Mathematics
Basic LATEX offers a high level of mathematical typesetting capabilities. How-

ever, when complex equations or other mathematical constructs have to be input
repeatedly, it is up to you to define new commands or environments to ease the
burden of typing. The American Mathematical Society (AMS), recognizing that
fact, has sponsored the development of extensions to TEX, known as AMS-TEX.
They make the preparation of mathematical compuscripts less time-consuming
and the copy more consistent.
Recently these extensions were ported to LATEX in the form of a set of
packages known as “AMS-LATEX” [?]. As some parts of AMS-TEX had to do
with mathematics fonts the corresponding LATEX packages went into a separate
distribution called “AMSFonts”, rather than into AMS-LATEX.
8.1 The AMS-LATEX Project

AMS-TEX was originally released for general use in 1982. Its main strength is
that it facilitates mathematical typesetting, while producing output that sat-
isfies the high standards of mathematical publishing. It provides a predefined
set of natural commands such as \matrix and \text that make complicated
mathematics reasonably convenient to type. These commands incorporate the
typesetting experience and standards of the American Mathematical Society, to
handle problematic possibilities, such as matrices within matrices or a word of
text within a subscript, without burdening the user.
AMS-TEX lacks certain useful LATEX features such as automatic numbering
that adjusts to addition or deletion of material being the primary one. Nor does
224 Higher Mathematics
it have the laborsaving abilities of LATEX for preparing indexes, bibliographies,

tables, or simple diagrams. These features are such a convenience for authors
that the use of LATEX spread rapidly in the mid-1980s (a reasonably mature
version of LATEX was available by the end of 1983), and the American Mathe-
matical Society began to be asked by its authors to accept electronic submissions
in LATEX.
Thus, the AMS-LATEX project came into being in 1987 and three years later
AMS-LATEX version 1.0 was released. The conversion of AMS-TEX’s mathe-
matical capabilities to LATEX, and the integration with the NFSS, were done by
Frank Mittelbach and Rainer Schöpf, working as consultants to the AMS, with
assistance from Michael Downes of the AMS technical support staff.
The most often used packages are amsmath (from AMS-LATEX) and amssymb
(from the AMSFonts distribution). To invoke them in a document you write, e.g.,
\usepackage{amsmath} in the usual way. Installation and usage documentation
is included with the packages. For amssymb the principal piece of documentation
is the AMSFonts User’s Guide (amsfndoc.tex); for amsmath it is the AMS-
LATE X User’s Guide (amsldoc.tex).1
8.2 Fonts and Symbols in Formulae

8.2.1 Mathematical Symbols
(L 42–47) Tables 8.2 on the next page to 8.11 on page 227 review the mathematical symbols
available in standard LATEX. You can put a slash through a LATEX symbol by
(L 44) preceding it with the \not command, for instance.
u 6< v or a 6∈ A $u \not< v$ or $a \not\in \mathbf{A}$
Tables 8.12 on page 227 to 8.19 on page 229 show the extra math symbols of
the AMS-Fonts, which are automatically available when you specify the amssymb
package.2 However, if you want to define only some of them (perhaps because
your TEX installation has insufficient memory to define all the symbol names),
you can use the amsfonts package and the \DeclareMathSymbol command, which
is explained in section 7.7.6.
1 The AMS distribution also contains a file diff12.tex which describes differences between
version 1.1 and 1.2 of AMS-LATEX. Note in particular that in versions 1.0 and 1.1 of AMS-
LATEX, which predated LATEX 2ε , the amsmath package was named “amstex” and included some
of the font-related features that are now separated in the amssymb and amsfonts packages.
2 Note that the Companion uses Lucida math fonts which contain the standard LATEX and
AMS symbols but with different shapes compared to the Computer Modern math fonts.
8.2 Fonts and Symbols in Formulae 225
â \hat{a} á \acute{a} ā \bar{a} ȧ \dot{a} ă \breve{a}

ǎ \check{a} à \grave{a} ~a \vec{a} ä \ddot{a} ã \tilde{a}
Table 8.1: Math mode accents (available in LATEX)
α \alpha β \beta γ \gamma δ \delta \epsilon

ε \varepsilon ζ \zeta η \eta θ \theta ϑ \vartheta
ι \iota κ \kappa λ \lambda µ \mu ν \nu
ξ \xi o o π \pi $ \varpi ρ \rho
% \varrho σ \sigma ς \varsigma τ \tau υ \upsilon
φ \phi ϕ \varphi χ \chi ψ \psi ω \omega
Γ \Gamma ∆ \Delta Θ \Theta Λ \Lambda Ξ \Xi
Π \Pi Σ \Sigma Υ \Upsilon Φ \Phi Ψ \Psi
Ω \Omega
Table 8.2: Greek letters (available in LATEX)
± \pm ∩ \cap \diamond ⊕ \oplus

∓ \mp ∪ \cup 4 \bigtriangleup \ominus
× \times ] \uplus 5 \bigtriangledown ⊗ \otimes
÷ \div u \sqcap / \triangleleft \oslash
∗ \ast t \sqcup . \triangleright \odot
? \star ∨ \vee \lhda \bigcirc
◦ \circ ∧ \wedge \rhda † \dagger
• \bullet \ \setminus \unlhda ‡ \ddagger
· \cdot o \wr \unrhda q \amalg
a
Not predefined in NFSS. Use the latexsym or amssymb package.
Table 8.3: Binary operation symbols (available in LATEX)
≤ \leq,\le ≥ \geq,\ge ≡ \equiv |= \models ≺ \prec

\succ ∼ \sim ⊥ \perp \preceq \succeq
' \simeq | \mid \ll \gg \asymp
k \parallel ⊂ \subset ⊃ \supset ≈ \approx ./ \bowtie
⊆ \subseteq ⊇ \supseteq ∼
= \cong 1 \Join < \sqsubset
= \sqsupset 6 = \neq ^ \smile v \sqsubseteq w \sqsupseteq
.
= \doteq _ \frown ∈ \in 3 \ni ∝ \propto
= = ` \vdash a \dashv < < > >
Table 8.4: Relation symbols (available in LATEX)
← \leftarrow ←− \longleftarrow ↑ \uparrow

⇐ \Leftarrow ⇐= \Longleftarrow ⇑ \Uparrow
→ \rightarrow −→ \longrightarrow ↓ \downarrow
⇒ \Rightarrow =⇒ \Longrightarrow ⇓ \Downarrow
↔ \leftrightarrow ←→ \longleftrightarrow l \updownarrow
⇔ \Leftrightarrow ⇐⇒ \Longleftrightarrow m \Updownarrow
7 → \mapsto 7−→ \longmapsto % \nearrow
←- \hookleftarrow ,→ \hookrightarrow & \searrow
( \leftharpoonup * \rightharpoonup . \swarrow
) \leftharpoondown + \rightharpoondown - \nwarrow
Table 8.5: Arrow symbols (available in LATEX)
.. ..
... \ldots ··· \cdots . \vdots . \ddots ℵ \aleph
0 \prime ∀ \forall ∞
√ \infty ~ \hbar ∅ \emptyset
∃ \exists ∇ \nabla \surd 2 \Boxa 4 \triangle
3 \Diamonda ı \imath  \jmath ` \ell ¬ \neg
> \top [ \flat \ \natural ] \sharp ℘ \wp
⊥ \bot ♣ \clubsuit ♦ \diamondsuit ♥ \heartsuit ♠ \spadesuit
0 \mhoa < \Re = \Im ∠ \angle ∂ \partial
a
Not predefined in NFSS. Use the latexsym or amssymb package.
Table 8.6: Miscellaneous symbols (available in LATEX)
P Q ` R H
T \sum S \prod F \coprod W \int V \oint
J \bigcap N \bigcup L \bigsqcup U \bigvee \bigwedge
\bigodot \bigotimes \bigoplus \biguplus
Table 8.7: Variable-sized symbols (available in LATEX)
\arccos \cos \csc \exp \ker \limsup \min \sinh

\arcsin \cosh \deg \gcd \lg \ln \Pr \sup
\arctan \cot \det \hom \lim \log \sec \tan
\arg \coth \dim \inf \liminf \max \sin \tanh
Table 8.8: Log-like symbols (available in LATEX)
↑ \uparrow ⇑ \Uparrow ↓ \downarrow ⇓ \Downarrow

{ \{ } \} l \updownarrow m \Updownarrow
b \lfloor c \rfloor d \lceil e \rceil
h \langle i \rangle / / \ \backslash
| | k \|
Table 8.9: Delimiters (available in LATEX)
   
 \rmoustache  \lmoustache  \rgroup  \lgroup
 w 

 \arrowvert w \Arrowvert 
 \bracevert
Table 8.10: Large delimiters (available in LATEX)
f
abc \widetilde{abc} c
abc \widehat{abc}
←− −→
abc \overleftarrow{abc} abc \overrightarrow{abc}
abc \overline{abc} abc \underline{abc}
z}|{
abc \overbrace{abc} abc
|{z} \underbrace{abc}
√ √n
abc \sqrt{abc} abc \sqrt[n]{abc}
abc
f0 f’ xyz \frac{abc}{xyz}
Table 8.11: LATEX math constructs
z \digamma κ \varkappa i \beth k \daleth ‫ג‬ \gimel

Table 8.12: AMS Greek and Hebrew (available with amssymb package)
p \ulcorner q \urcorner x \llcorner y \lrcorner

Table 8.13: AMS delimiters (available with amssymb package)
V \Rrightarrow \rightsquigarrow ⇔ \leftleftarrows

\leftrightarrows W \Lleftarrow \twoheadleftarrow
\leftarrowtail " \looparrowleft \leftrightharpoons
x \curvearrowleft \circlearrowleft \Lsh
\upuparrows \upharpoonleft \downharpoonleft
( \multimap ! \leftrightsquigarrow \rightleftarrows
⇒ \rightrightarrows \twoheadrightarrow \rightarrowtail
# \looparrowright
\rightleftharpoons y \curvearrowright
\circlearrowright \Rsh \downdownarrows
\downharpoonright \upharpoonright,\restriction
Table 8.14: AMS arrows (available with amssymb package)
8 \nleftarrow 9 \nrightarrow : \nLeftarrow

; \nRightarrow = \nleftrightarrow < \nLeftrightarrow
Table 8.15: AMS negated arrows (available with amssymb package)
5 \leqq 6 \leqslant 0 \eqslantless

. \lesssim / \lessapprox u \approxeq
l \lessdot ≪ \lll,\llless ≶ \lessgtr
Q \lesseqgtr S \lesseqqgtr + \doteqdot,\Doteq
: \risingdotseq ; \fallingdotseq v \backsim
w \backsimeq j \subseteqq b \Subset
< \sqsubset 4 \preccurlyeq 2 \curlyeqprec
- \precsim w \precapprox C \vartriangleleft
E \trianglelefteq \vDash \Vvdash
` \smallsmile a \smallfrown l \bumpeq
m \Bumpeq = \geqq > \geqslant
1 \eqslantgtr & \gtrsim ' \gtrapprox
m \gtrdot ≫ \ggg,\gggtr ≷ \gtrless
R \gtreqless T \gtreqqless P \eqcirc
$ \circeq , \triangleq ∼ \thicksim
≈ \thickapprox k \supseteqq c \Supset
= \sqsupset < \succcurlyeq 3 \curlyeqsucc
% \succsim v \succapprox B \vartriangleright
D \trianglerighteq \Vdash p \shortmid
q \shortparallel G \between t \pitchfork
∝ \varpropto J \blacktriangleleft ∴ \therefore
\backepsilon I \blacktriangleright ∵ \because
Table 8.16: AMS binary relations (available with amssymb package)
≮ \nless \nleq
\nleqslant
\nleqq \lneq \lneqq
\lvertneqq \lnsim \lnapprox
⊀ \nprec \npreceq \precnsim
\precnapprox \nsim . \nshortmid
- \nmid 0 \nvdash 2 \nvDash
6 \ntriangleleft 5 \ntrianglelefteq * \nsubseteq
( \subsetneq \varsubsetneq $ \subsetneqq
& \varsubsetneqq ≯ \ngtr \ngeq
\ngeqslant \ngeqq \gneq
\gneqq \gvertneqq \gnsim
\gnapprox \nsucc \nsucceq
\succnsim \succnapprox \ncong
/ \nshortparallel ∦ \nparallel 2 \nvDash
3 \nVDash 7 \ntriangleright 4 \ntrianglerighteq
+ \nsupseteq # \nsupseteqq ) \supsetneq
! \varsupsetneq % \supsetneqq ' \varsupsetneqq
Table 8.17: AMS negated binary relations (available with amssymb package)
u \dotplus r \smallsetminus e \Cap,\doublecap

d \Cup,\doublecup Z \barwedge Y \veebar
[ \doublebarwedge \boxminus \boxtimes
\boxdot \boxplus > \divideontimes
n \ltimes o \rtimes h \leftthreetimes
i \rightthreetimes f \curlywedge g \curlyvee
\circleddash ~ \circledast } \circledcirc
\centerdot | \intercal
Table 8.18: AMS binary operators (available with amssymb package)
~ \hbar } \hslash M \vartriangle

O \triangledown \square ♦ \lozenge
s \circledS ∠ \angle ] \measuredangle
@ \nexists 0 \mho ` \Finv
a \Game k \Bbbk 8 \backprime
∅ \varnothing N \blacktriangle H \blacktriangledown
\blacksquare \blacklozenge F \bigstar
^ \sphericalangle { \complement ð \eth
\diagup \diagdown
Table 8.19: AMS miscellaneous (available with amssymb package)
8.2.2 Names of Math Font Commands

The list of math font commands provided by the AMS packages is shown in
table 8.20 on the next page, where for each case an example is shown. In addition,
the math font commands of table 7.4 on page 183 can be used.
In the amsmath package, \boldsymbol is to be used for individual
bold math symbols and bold Greek letters—everything in math except
for letters (where one would use \mathbf). For example, to obtain a
bold ∞, or \boldsymbol{\infty}, \boldsymbol{+}, \boldsymbol{\pi}, or
\boldsymbol{0}.
Since \boldsymbol takes a lot of typing, you can introduce new commands
for bold symbols to be used frequently:
\newcommand{\bpi}{\boldsymbol{\pi}}
\newcommand{\binfty}{\boldsymbol{\infty}}
B∞ + πB1 ∼ B∞ + πB1 \[ B_\infty + \pi B_1 \sim
\mathbf{B}_{\binfty} \boldsymbol{+}
\bpi \mathbf{B}_{\boldsymbol{1}}
\]
For those math symbols where the command \boldsymbol has no effect
because the bold version of the symbol does not exist in the currently available
fonts, there exists a command “Poor man’s bold” (\pmb), which simulates bold
\mathbb Blackboard bold alphabet, e.g., $\mathbb{NQRZ}$ gives: NQRZ (not available
in amsmath, need to load amssymb).
\mathfrak Euler Fraktur alphabet, e.g., $\mathfrak{E}=\mathfrak{mc}^2$ gives: E =
mc2 (not available in amsmath, need to load amssymb).
\boldsymbol Used to obtain bold numbers and other nonalphabetic symbols, as well as bold
Greek letters (defined in amsbsy).
\pmb “Poor man’s bold,” used for math symbols when
H bold versions don’t exist in the
available fonts, e.g., $\pmb{\oint}$ gives: and $\pmb{\triangle}$ gives:
4 (defined in amsbsy).
\text Produce normal text with correct text-spacing in the current font used outside
math, e.g., $E=mc^2\quad\text{(Einstein)}$ gives: E = mc2 (Einstein)
(defined in amstext).
Table 8.20: Font commands available in mathematics with the AMS packages
by typesetting several copies of the symbol with slight offsets. This procedure
must be used for the extension and large operator symbols from the cmex font,
as well as the AMS extra math symbols from the msam and msbm fonts.
\[ \frac{\partial w}{\partial u}

∂w ∂u \pmb{\Bigg\vert}

∂u ∂v \frac{\partial u}{\partial v} \]
P Q
With large operators and extension symbols (for example, and ) \pmb
does not currently work very well because the proper spacing and treatment of
limits is not preserved. Therefore, the TEX operator \mathop needs to be used
(see table 7.13 on page 213).
XY XY \[ \sum_{j<P}
λR(ri ) λR(xj ) \prod_\lambda \lambda R(r_i) \qquad
j<P λ xj λ \mathop{\pmb{\sum}}_{x_j}
\mathop{\pmb{\prod}}_\lambda \lambda R(x_j)
\]
To make an entire math formula bold (or as much of it as possible, depending

on the available fonts), use \boldmath preceding the formula.
The sequence \mathbf{\hat{A}} produces a bold accent character over the
A. However, combinations like \mathcal{\hat{A}} will not work in ordinary
LATEX because the \mathcal font does not have its own accents. In the amsmath
package the font change commands are defined in such a way that accent charac-
ters will be taken from the \mathrm font if they are not available in the current
font (in addition to the \mathcal font, the \mathbb and \mathfrak fonts don’t
contain accents).
8.3 Compound Symbols, Delimiters, Operators 231
8.3 Compound Symbols, Delimiters, Operators

This section3 presents the math commands that are available through the
amsmath package, which supplements LATEX in the area of compound symbols,
large delimiters, etc. In the examples, amsmath’s alignment environments are
used. In principle a detailed understanding of how they work is not necessary at
this stage, but an interested reader can turn to section 8.5 for more information.
8.3.1 Multiple Integral Signs

\iint, \iiint, and \iiiint give multiple integral signs, with the spacing be-
tween them nicely adjusted, in both text and display style. \idotsint gives two
integral signs with dots between them.
ZZ \begin{gather}
µ(u, v) du dv (8.1) \iint\limits_V \mu(u,v)\,du\,dv \\
\iiint\limits_V \mu(u,v,w)\,du\,dv\,dw \\
Z Z ZV \iiiint\limits_V \mu(t,u,v,w)\,dt\,du\,dv\,dw\\
µ(u, v, w) du dv dw (8.2) \idotsint\limits_V \mu(u_1,\dots,u_k)
ZZZZ V \end{gather}
µ(t, u, v, w) dt du dv dw (8.3)
V
Z Z
··· µ(u1 , . . . , uk ) (8.4)
V
8.3.2 Over and Under Arrows

Some extra over and under arrow operations are available. (In standard LATEX
one has \overrightarrow and \overleftarrow.)
−−−−−−→ \begin{align*}
ψδ (t)Et h = ψδ (t)Et h \overrightarrow{\psi_\delta(t) E_t h} &
−−−−−−→
←−−−−−− =\underrightarrow{\psi_\delta(t) E_t h} \\
ψδ (t)Et h = ψδ (t)Et h
←−−−−−− \overleftarrow{\psi_\delta(t) E_t h} &
←−−−−→
ψδ (t)Et h = ψδ (t)Et h =\underleftarrow{\psi_\delta(t) E_t h} \\
←−−−−→
\overleftrightarrow{\psi_\delta(t) E_t h} &
=\underleftrightarrow{\psi_\delta(t) E_t h}
\end{align*}
These
R arrows all scale properly in subscript sizes, as seen in the following inte-
gral −→
uv
vt dt, which was coded as $ \int_{\overrightarrow{uv}} vt\,dt $.
3 Some material in this and the following sections is reprinted from the electronic document
testmath.tex (distributed with AMS-LATEX) with permission of the American Mathematical
Society.
8.3.3 Dots
Ellipsis dots should almost always be typed as \dots. Positioning (on the base-
line or centered) is automatically selected according to whatever follows the
\dots. If the next character is a plus sign, the dots will be centered; if it’s a
comma, they will be on the baseline. These default dot placements provided by
the amsmath package can be changed if different conventions are wanted.
If the dots fall at the end of a math formula, the next character will be
something like \end or \) or $, which does not give any information about how
to place the dots. If that is the case, you must help by using \dotsc for “dots
with commas,” or \dotsb for “dots with binary operators/relations,” or \dotsm
for “multiplication dots,” or \dotsi for “dots with integrals.” In the example
below, low dots are produced in the first instance and centered dots in the others,
with the spacing on either side of the dots nicely adjusted.
A series H1 , H2 , . . . , a regional sum H1 +H2 + A series $H_1,H_2,\dotsc$,

· · · , an orthogonal product H1 H2 · · · , and an a regional sum $H_1+H_2+\dotsb$, an
infinite integral orthogonal product $H_1H_2\dotsm$, and
Z Z an infinite integral
··· \[\int_{H_1}\int_{H_2}\dotsi\].
H1 H2
8.3.4 Accents in Math

The following accent commands automatically position double accents correctly:
\begin{gather*}
´ ¯ ˘ ˇ
Á B̄ C̆ Ď \Acute{\Acute{A}} \qquad\Bar{\Bar{B}} \qquad
¨ ` ˆ \Breve{\Breve{C}} \qquad\Check{\Check{D}} \\
Ë Ḟ˙ G̀ Ĥ
\Ddot{\Ddot{E}} \qquad\Dot{\Dot{F}} \qquad
~
I˜
˜ J~ \Grave{\Grave{G}} \qquad\Hat{\Hat{H}} \\
\Tilde{\Tilde{I}} \qquad\Vec{\Vec{J}}
\end{gather*}
This double accent operation is complicated and tends to slow down the
processing of a LATEX file. If the document contains many double accents, you
can load the amsxtra package. It defines the \accentedsymbol command, which
you can use in the preamble of your document to help speed things up. It stores
the result of the double accent command in a box register for quick retrieval.
\accentedsymbol is used like \newcommand:
ˆ
This is a double hat Â and this δ̄˙ a delta with \accentedsymbol{\Ahathat}{\Hat{\Hat A}}
a bar and a dot. \accentedsymbol{\dbardot}{\Dot{\Bar \delta}}
This is a double hat $\Ahathat$ and this
$\dbardot$ a delta with a bar and a dot.
8.3.5 Superscripted Accents

Some accents have a wide form: typing $\widehat {xy},\widetilde{xy}$ pro-
duces xcy, x
fy. Because these wide accents have a certain maximum size, the
amsxtra package introduces a different notation to handle extremely long expres-
\ amsxtra has the following control sequences
sions: (AmBD)b instead of AmBD.
to achieve this easily:
\begin{gather}
(AmBD)b (AmBD)∨ (8.5) (AmBD)\sphat \qquad (AmBD)\spcheck \\
∼ . (AmBD)\sptilde \qquad (AmBD)\spdot \\
(AmBD) (AmBD) (8.6)
.. ... (AmBD)\spddot \qquad(AmBD)\spdddot \\
(AmBD) (AmBD) (8.7)
(AmBD)\spbreve
(AmBD)˘ (8.8) \end{gather}
8.3.6 Dot Accents

\dddot and \ddddot are available to produce tripled and quadrupled dot accents
in addition to the \dot and \ddot accents already available in LATEX:
... ....
Q R $ \dddot{Q} \qquad \ddddot{R} $
8.3.7 Roots
In ordinary LATEX the placement of root indices is sometimes not good. With
amsmath the commands \leftroot and \uproot allow the adjustment of the
position of the root. Positive arguments to these commands will move the root
index to the left and up respectively, while a negative argument will move them
right and down. The units of increment are quite small, which is useful for such
adjustments. In the example below, the root index β is moved 2 units to the left
and 4 units up.
√
β
√
β
\[ \sqrt[\beta]{k} \qquad
k k \sqrt[\leftroot{2}\uproot{4}\beta]{k} \]
8.3.8 Boxed Formulae

The command \boxed puts a box around its argument, similar to \fbox, except
that the contents are in math mode:
\[ \boxed{W_t - F \subseteq
Wt − F ⊆ V (Pi ) ⊆ Wt V(P_i) \subseteq W_t} \]
8.3.9 Extensible Arrows

\xleftarrow and \xrightarrow produce arrows that extend automatically to
accommodate unusually wide subscripts or superscripts. The text of the sub-
script or superscript are given as an optional and mandatory argument, respec-
tively:
\[0 \xleftarrow[\zeta]{\alpha}
α ∂0 α(b)
0 ←− F × 4[n − 1] −−−−→ E ∂0 b F\times\triangle[n-1]
ζ
\xrightarrow{\partial_0\alpha(b)}
E^{\partial_0b}
\]
8.3.10 \overset, \underset, and \sideset

LATEX provides \stackrel for placing a superscript above a binary rela-
tion. amsmath introduces somewhat more general commands, \overset and
\underset. These can be used to place one symbol above or below another
symbol, independently of whether it is a relation or something else. The in-
put \overset{*}{X} will place a superscript-size ∗ above the X; \underset
performs the parallel operation that one would expect.
∗ a \[ \overset{*}{X} \qquad
X X X \underset{*}{X} \qquad
∗ b
\overset{a}{\underset{b}{X}} \]
There is also a command called \sideset that serves a rather special pur-
pose: it puts symbols
P in the
Q subscript and superscript positions of large operator
symbols such as and . A prime example is the case when you want to put
a prime on a sum symbol. If there are no limits above or below the sum, you
could just use \nolimits:
X0 \begin{equation}
En . (8.9) \sum\nolimits’ E_n.
\end{equation}
But if you want not only the prime but also limits on the sum symbol,
things are not so easy. Suppose you want to add a prime on the sum symbol in
an expression, like
X \begin{equation}
nEn (8.10) \sum_{n<k,\;n\ \mathrm{odd}}nE_n
n<k, n odd \end{equation}
then you can use \sideset like this: \sideset{}{’}\sum_{...}nE_n. The extra
pair of empty braces is explained by the fact that \sideset has the capability
of putting an extra symbol or symbols at each corner of a large operator.
2 Y4 X0 \[
Ei βx \sideset{_1^2}{_3^4}\prod_k \qquad
1 3
k 0≤i≤m \sideset{}{’}\sum_{0\le i\le m} E_i\beta x
\]
8.3.11 The \smash Command

The plain TEX command \smash retains the contents of a box but annihilates
its height and depth. The amsmath package introduces the optional arguments
t and b with the \smash command. \smash[t]{...} annihilates only the top of
the box contents, retaining the bottom part, while \smash[b]{...} annihilates
the bottom part and keeps the top.
√ p \[ X_j=(1/\sqrt{\smash[b]{\lambda_j}})X_j’
Xj = (1/ λj )Xj0 Xj = (1/ λj )Xj0 \qquad
X_j=(1/\sqrt{\lambda_j})X_j’ \]
The previous example shows how the \smash command was used to limit
the depth of the radical, which otherwise extends to encompass the depth of the
subscript (right-hand formula in the baove example).
8.3.12 The \text Command

The main use of the \text command, which is also available separately with the
amstext package, is for words or phrases in a display. It is similar to the LATEX
command \mbox in its effects, but has a couple of advantages. If you would like
a word or phrase of text in a subscript, you can type
..._{\text{word or phrase}}
which, apart from having a more descriptive name, is also slightly easier to enter
than the equivalent \mbox, since the correct size is automatically chosen:
..._{\mbox{\scriptsize word or phrase}}
\[ \mathbf{y}=\mathbf{y}’ \quad
y = y0 if and only if yk0 = δk yτ (k) \text{if and only if} \quad
y’_k=\delta_k y_{\tau(k)} \]
8.3.13 Operator Names

Math functions such as log, sin, and lim are traditionally set in ro-
man type to help avoid confusion with single math variables, set in math
italic. The more common ones have predefined names: \log, \sin, \lim,
and so forth (see table 8.8 on page 226). New ones, however, come
up all the time in mathematical papers. The amsmath package provides
\DeclareMathOperator and \DeclareMathOperator* for producing new func-
tion names that will have the same typographical treatment. For instance,
\DeclareMathOperator{\xxx}{xxx} produces xxx in the proper font and au-
tomatically adds proper spacing on either side when necessary, so that you get
A xxx B instead of AxxxB. Examples of definitions of operator names are shown
below (the \, in the definition of \esssup adds some space; see table 8.21 on
page 252):
Input text
\DeclareMathOperator*{\esssup}{ess\,sup}
\DeclareMathOperator{\meas}{meas}
\newcommand{\abs}[1]{\lvert#1\rvert}
\newcommand{\norm}[1]{\lVert#1\rVert}
\begin{align*}
\norm{f}_\infty & =
\esssup_{x\in R^n}\abs{f(x)} \\
\meas_1\{u\in R_+^1\colon f^*(u)>\alpha\} & =
\meas_n\{x\in R^n\colon\abs{f(x)}\geq\alpha\}\qquad\forall\alpha>0.
\end{align*}
kf k∞ = ess sup|f (x)|

x∈Rn
1
meas1 {u ∈ R+ : f ∗ (u) > α} = measn {x ∈ Rn : |f (x)| ≥ α} ∀α > 0.
Output text
The starred form \DeclareMathOperator* is like \DeclareMathOperator;

the only difference is the placement of subscripts and superscripts, as seen in the
example above. In order to make the use of the vertical bar notation more flexi-
ble, amsmath defines the new commands \lvert, \rvert, \lVert, and \rVert,
which are comparable to LATEX’s \langle and \rangle.
With amsmath the following operators are predefined: \varlimsup,
\varliminf, \varinjlim, and \varprojlim. Here’s what they look like in use:
\begin{gather}
lim n→∞ Q(un , un − u# ) ≤ 0 (8.11) \varlimsup_{n\rightarrow\infty}
lim n→∞ |an+1 | / |an | = 0 (8.12) \mathcal{Q}(u_n,u_n-u^{\#}) \le 0 \\
λ ∗ \varliminf_{n\rightarrow\infty}
−→(mi ·)
lim ≤0 (8.13)
\left|a_{n+1}\right|/\left|a_n\right| = 0 \\
←− p∈S(A) Ap ≤ 0
lim (8.14) \varinjlim (m_i^\lambda\cdot)^* \le 0 \\
\varprojlim_{p\in S(A)}A_p \le 0
\end{gather}
8.3.14 \mod and Its Relatives

Commands \mod, \bmod, \pmod, and \pod are provided by the amsopn package
to deal with the rather special spacing conventions of “mod” notation. \bmod
and \pmod are available in LATEX, but with amsopn the spacing of \pmod will
adjust to a smaller value if it is used in a nondisplay formula. \mod and \pod
are variants of \pmod preferred by some authors; \mod omits the parentheses,
whereas \pod omits the “mod” and retains the parentheses.
\begin{equation}
gcd(k, l mod k) (8.15) \gcd (k , l \bmod k)
\end{equation}
\begin{align}
u≡v+1 (mod n2 ) (8.16) u & \equiv v + 1 \pmod{n^2} \\
u≡v+1 mod n2 (8.17) u & \equiv v + 1 \mod{n^2} \\
u≡v+1 (n2 ) (8.18) u & \equiv v + 1 \pod{n^2}
\end{align}
8.3.15 Fractions and Related Constructions

In addition to \frac (available in LATEX), amsmath provides \dfrac and
\tfrac as convenient abbreviations for {\displaystyle\frac ... } and
{\textstyle\frac ... }.
\[ \frac{1}{k}\log_2 c(f)\quad
1 1
log2 c(f ) k
log2 c(f ) \tfrac{1}{k}\log_2 c(f) \]
k
and
q r
1 1 $ \sqrt{\frac{1}{k}\log_2 c(f)}\quad
and k
log2 c(f ) log2 c(f ). \sqrt{\dfrac{1}{k}\log_2 c(f)} $.
k

For binomial expressions such as nk the amsmath packages defines the com-
mands \binom, \dbinom, and \tbinom.
! \begin{equation}
k k−1 k
k−2 \binom{k}{1}2^{k-1}+\tbinom{k}{2}2^{k-2}
2 + 2
2 (8.19)
1 \end{equation}
! and
k
k k−2 $\binom{k}{1}2^{k-1}+\dbinom{k}{2}2^{k-2}$.
and 1
2k−1 + 2 .
2
\binom, and its variants \dbinom and \tbinom, as well as \frac and its
variants \dfrac and \tfrac are implemented using the generalized fraction com-
mand \genfrac, which has six parameters.
\genfrac{ldelim}{rdelim}{thick }{style}{num}{denom}
The first two parameters ldelim and rdelim are the left and right delimiters,
respectively. The third parameter thick allows you to override the line thickness
(for instance \binom uses this to set the line thickness to zero, i.e., invisible). If
this argument is left empty, the line thickness defaults to “normal”. The fourth
parameter is the mathematics style override. It can take integer values in the
range 0–3 to select, respectively, \displaystyle, \textstyle, \scriptstyle,
and \scriptscriptstyle. Finally, the fifth argument num is the numerator,
while the sixth denom is the denominator of the fraction.
To illustrate, here is how \frac, \tfrac, and \binom might be defined.
\newcommand{\frac}[2]{\genfrac{}{}{}{}{#1}{#2}}
\newcommand{\tfrac}[2]{\genfrac{}{}{}{1}{#1}{#2}}
\newcommand{\binom}[2]{\genfrac{(}{)}{0pt}{}{#1}{#2}}
Other examples are the following re-implementation of TEX’s fraction primitives.
* + \renewcommand{\over}[2]{%
n+1 n+1 \genfrac{}{}{}{}{#1}{#2}}
n n \renewcommand{\overwithdelims}[2]{%
\genfrac{\langle}{\rangle}{}{}{#1}{#2}}
! \[ \over{n+1}{n}\qquad\overwithdelims{n+1}{n} \]
n+2 n+2 \renewcommand{\atop}[2]{%
n n \genfrac{}{}{0pt}{}{#1}{#2}}
\renewcommand{\atopwithdelims}[2]{%
\genfrac{(}{)}{0pt}{}{#1}{#2}}
" #
n−3 n−3 \[ \atop{n+2}{n}\qquad\atopwithdelims{n+2}{n} \]
\renewcommand{\above}[2]{%
n n
\genfrac{}{}{1pt}{}{#1}{#2}}
\renewcommand{\abovewithdelims}[2]{%
\genfrac{[}{]}{1pt}{}{#1}{#2}}
\[ \above{n-3}{n}\qquad\abovewithdelims{n-3}{n} \]
Of course, if you want to use a particular notation implemented with

\genfrac repeatedly throughout your document you will do yourself (and your
publisher) a favor if you define a meaningful command name with \newcommand
as an abbreviation for that notation, as in the examples above.
8.3.16 Continued Fractions

A continued fraction can be obtained as follows:
\begin{equation}
1 \cfrac{1}{\sqrt{2}+
(8.20)
√ 1 \cfrac{1}{\sqrt{3}+
2+
√ 1 \cfrac{1}{\sqrt{4}+
3+ \cfrac[l]{1}{\sqrt{5}+
√ 1
4+ \cfrac[r]{1}{\sqrt{6}+\dotsb
√ 1
5+ √ }}}}}
6 + ··· \end{equation}
Left or right positioning of any of the numerators is achieved by using the

optional argument [l] or [r] with the \cfrac command.
8.3.17 Big-g-g-g Delimiters

In order to better control the sizes of math delimiters, basic TEX introduces four
commands \big, \Big, \bigg and \Bigg, which produce ever larger versions of
the delimiter specified as parameter. These commands can be used with any of
the delimiters that can follow the \left or \right command (see tables 8.9, 8.10,
and 8.13 on page 227). Moreover, for each of the four commands above, three
variants exist for use as an opening symbol (e.g., \bigl), as a binary relation
(e.g., \Bigm), or as a closing symbol (e.g., \Biggr).4 Whereas, with basic TEX,
the sizes of these delimiters are fixed, with the amsmath package the sizes adapt
to the size of the surrounding material.
Z tε \[
Ey Lx,yx (s) ϕ(x) ds \biggl(\mathbf{E}_{y}\int_0^{t_\varepsilon}
0 L_{x,y^x(s)}\varphi(x)\,ds \biggr)
\]
Z tε {\Large
\[
Ey Lx,yx (s) ϕ(x) ds \biggl(\mathbf{E}_{y}\int_0^{t_\varepsilon}
0 L_{x,y^x(s)}\varphi(x)\,ds \biggr)
\]}
4 See table 7.13 on page 213 for a discussion of the various math symbol types.
8.4 Matrix-Like Environments and

Commutative Diagrams
8.4.1 The cases Environment
“Case” constructions can be produced using the cases environment.
( \begin{equation}
0 if r − j is odd, P_{r-j}=
Pr−j =
r! (−1)(r−j)/2 if r − j is even. \begin{cases}
(8.21) 0& \text{if $r-j$ is odd},\\
r!\,(-1)^{(r-j)/2}& \text{if $r-j$ is even}.
\end{cases}
\end{equation}
Notice the use of \text and the embedded math.
8.4.2 The Matrix Environments

The matrix environments are similar to LATEX’s array, except they do not have
an argument specifying the format of the columns. Instead, a default format is
provided: up to 10 centered columns. The examples below show how to use the
matrix environments matrix, pmatrix, bmatrix, vmatrix, and Vmatrix:
\begin{gather*}
0 1 0 −i 1 0 \begin{matrix} 0 & 1 \\ 1 & 0\end{matrix} \quad
1 0 i 0 0 −1
\begin{pmatrix}0 &-i \\ i & 0\end{pmatrix}\quad
a b 1 0

\begin{bmatrix}1 & 0 \\ 0 &-1\end{bmatrix}\\
c d 0 1 \begin{vmatrix}a & b \\ c & d\end{vmatrix}\quad
\begin{Vmatrix}1 & 0 \\ 0 & 1\end{Vmatrix}
\end{gather*}
The maximum number of columns is determined by the counter

MaxMatrixCols, which you can change using LATEX’s standard counter com-
mands. For example, suppose you have a large matrix with 19 or 20 columns,
then you can do something like this:
\begin{equation}
\setcounter{MaxMatrixCols}{20}
A=\begin{pmatrix}
...&...&...&...&...&...&...&...&...&...&...&...& ... \\
... \\
...
\end{pmatrix}
\end{equation}
\setcounter{MaxMatrixCols}{10}
8.4 Matrix-Like Environments and Commutative Diagrams 241
As counters are global in LATEX, you might want to reset the value of
MaxMatrixCols to its default value of 10 after finishing your wide matrix, since
with a high value, LATEX must work a lot harder to typeset a matrix.
To produce a small matrix suitable for use in text, use the smallmatrix
environment.
To show the effect of the matrix on To show the effect of the matrix on surrounding
surrounding lines inside a paragraph, we put lines inside a paragraph, we put it here:
it here: ( ac db ) and follow it with enough text \begin{math}
to ensure that there is at least one full line \left( \begin{smallmatrix}
below the matrix. a&b\\ c&d
\end{smallmatrix} \right)
\end{math}
and follow it with enough text to ensure that
there is at least one full line below the matrix.
A row of dots in a matrix, spanning a given number of columns, can be

obtained with the command:
\hdotsfor[spacing-factor ]{number }
The spacing of the dots can be varied by using the optional parameter
spacing-factor , for example, \hdotsfor[1.5]{3}. The number in square brack-
ets multiplies the spacing between the dots; the normal value is one.
Input text
\[ W(\Phi)= \begin{Vmatrix}
\dfrac\varphi{(\varphi_1,\varepsilon_1)} &0&
\hdotsfor{2} &0 \\
\dfrac{\varphi k_{n2}}{(\varphi_2,\varepsilon_1)} &
\dfrac\varphi{(\varphi_2,\varepsilon_2)} &0&\dots&0\\
\hdotsfor{5} \\
\dfrac{\varphi k_{n1}}{(\varphi_n,\varepsilon_1)} &
\dfrac{\varphi k_{n2}}{(\varphi_n,\varepsilon_2)}&\dots&
\dfrac{\varphi k_{n\,n-1}}{(\varphi_n,\varepsilon_{n-1})}&
\dfrac{\varphi}{(\varphi_n,\varepsilon_n)}
\end{Vmatrix}\]
ϕ

0 ............... 0
(ϕ1 , ε1 )
ϕkn2 ϕ

(ϕ , ε ) (ϕ , ε ) 0 ... 0
W (Φ) = 2 1 2 2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ϕkn1 ϕkn2 ϕkn n−1 ϕ

(ϕ , ε ) (ϕ , ε ) . . . (ϕ , ε ) (ϕ , ε )
n 1 n 2 n n−1 n n
Output text
8.4.3 The \substack command

The \substack command can be used to typeset several lines as a subscript or
superscript, using \\ as the row delimiter. This command can be used anywhere
an ordinary subscript or superscript can be used.
X \begin{equation}
P (i, j) (8.22) \sum
0≤i≤m _{\substack{0\le i\le m\\ 0<j<n}}
0<j<n
P(i,j)
\end{equation}
Each line can be left-adjusted instead of centered by using the subarray

environment.
X \begin{equation}
P (i, j) (8.23) \sum_{\begin{subarray}{l}
i∈Λ i\in \Lambda\\ 0<j<n
0<j<n
\end{subarray}}
P(i,j)
\end{equation}
8.4 Matrix-Like Environments and Commutative Diagrams 243
8.4.4 Commutative Diagrams

The commutative diagram commands of AMS-TEX are not included in the
amsmath package, but are available as a separate package, amscd. This con-
serves memory for users who do not need commutative diagrams. The picture
environment can be used for complex commutative diagrams, but for simple
diagrams without diagonal arrows the amscd commands are more convenient.5
j
\DeclareMathOperator{\End}{End}
S WΛ ⊗ T −−−−→ T \[\begin{CD}
 
  S^{\mathcal{W}_\Lambda}\otimes T
y yEnd P @>j>> T \\
(S ⊗ T )/I (Z ⊗ T )/J @VVV @VV{\End P}V\\
(S\otimes T)/I @= (Z\otimes T)/J
\end{CD}\]
A similar result, which does not look quite as good, can be produced in
ordinary LATEX by:
j
\[\begin{array}{ccc}
S WΛ⊗ T −→ T
 S^{\mathcal{W}_\Lambda}\otimes T &
 
y yEnd P \stackrel{j}{\longrightarrow} &
(S ⊗ T )/I = (Z ⊗ T )/J T \\
\Big\downarrow & &
\Big\downarrow\vcenter{%
\rlap{$\scriptstyle{\mathrm{End}}\,P$}}\\
(S\otimes T)/I & = &
(Z\otimes T)/J
\end{array}\]
When using the amscd package, you will obtain longer horizontal arrows and
improved spacing between elements of the diagram.
In the CD environment the commands @>>>, @<<<, @VVV, and @AAA give (re-
spectively) right, left, down, and up arrows. For people with keyboards lacking
the angle brackets the notations @))) and @((( are available as alternatives.
For the horizontal arrows, material between the first and second > or <
symbols will be typeset as a superscript, and material between the second and
third will be typeset as a subscript. Similarly, material between the first and
second, or second and third, A’s or V’s of vertical arrows will be typeset as left
or right “sidescripts.” This was used in the first example above to place the
operator “End P ” to the right of the arrow.
The final example again shows the use of \DeclareMathOperator.
5 Much more extensive commutative diagram packages are Kristoffer Rose’s XY-pic
system [?], Paul Taylor’s Commutative Diagram package [?], and the Diagram 3 system by
Francis Borceux [?].
\begin{equation*}
cov(L) −−−−→ non(K) −−−−→ cf(K)
 x x \DeclareMathOperator{\add}{add}
  
y   \DeclareMathOperator{\cf}{cf}
\DeclareMathOperator{\cov}{cov}
add(L) −−−−→ add(K) −−−−→ cov(K) \DeclareMathOperator{\non}{non}
\begin{CD}
\cov(\mathcal{L}) @>>> \non(\mathcal{K})
@>>> \cf(\mathcal{K})\\
@VVV @AAA @AAA \\
\add(\mathcal{L}) @>>> \add(\mathcal{K})
@>>> \cov(\mathcal{K})\\
\end{CD}
\end{equation*}
8.5 Alignment Structures for Equations

The amsmath package defines several environments for creating multiline display
equations. They perform similarly to LATEX’s equation and eqnarray environ-
ments. The following structures are discussed in the next sections.
align align* alignment at a single place

flalign flalign* spaced-out variants of the above
alignat alignat* alignment with space control
equation equation* one-line formula
gather gather* combining formula without alignment
multline multline* multiline equation (one equation number)
split splitting long formulas
Some of these multiline display environments allow you to align parts of

the formula. In contrast to the original LATEX environments eqnarray and
eqnarray*, the structures implemented by the amsmath package use a different
concept for marking the alignment points: while eqnarray is similar to an array
environment with a {rcl} preamble and therefore uses two ampersand characters
surrounding the part that should be aligned, in the amsmath structures you
should mark the alignment point (or points in alignat, for example) only with
a single ampersand character, placing it to the left of the character that should
be aligned with previous or following lines.
The amsmath structures give correct spacing around the alignment points,
while the eqnarray environment produces extra spaces depending on the param-
eter settings for array. The difference can be seen clearly in the next example,
where we typeset the same equation using the equation, align, and eqnarray
environments; ideally all three should produce the same result, but the eqnarray
environment comes out too wide.
8.5 Alignment Structures for Equations 245
\begin{equation}
x2 + y 2 = z 2 (8.24) x^2+y^2 = z^2
\end{equation}
\begin{align}
x2 + y 2 = z 2 (8.25) x^2+y^2 &= z^2 \\ x^3+y^3 &< z^3
3 3 3
x +y <z (8.26) \end{align}
\begin{eqnarray}
x^2+y^2 &=& z^2 \\ x^3+y^3 &<& z^3
x2 + y 2 = z2 (8.27) \end{eqnarray}
3 3
x +y < z3 (8.28)
8.5.1 Equation Groups without alignment

The gather environment is used for two or more equations, when no alignment
desired among them. Each one is centered separately between the left and right
margins.
\begin{gather}
(a + b)2 = a2 + 2ab + b2 (8.29) (a + b)^2 = a^2 + 2ab + b^2\\
(a + b) · (a − b) = a2 − b2 (8.30) (a + b) \cdot (a - b) = a^2 - b^2
\end{gather}
More examples are shown in section 8.7.3 on page 259.
8.5.2 Equation Groups with alignment

The align environment is used for two or more equations when vertical align-
ment is desired (usually binary relations such as equal signs are aligned). The
term “equation” is used rather loosely here to mean any math formula that is
intended by an author to be a self-contained subdivision of the larger display,
usually, but not always, containing a binary relation.
\begin{align}
x2 + y 2 = 1 x3 + y 3 = 1 (8.31) x^2 + y^2 & = 1 &
p p
x = 1 − y2 x = 3 1 − y3 x^3 + y^3 & = 1 \\
(8.32) x & = \sqrt{1-y^2} &
x & = \sqrt[3]{1-y^3}
\end{align}

With the align environment the material is spread out uniformly over the
lines. If you want to control the space between equation columns then you can
use an alignat environment. It has one required argument, for specifying the
number of “align” structures. For an argument of n, the number of ampersand

characters per line is 2n − 1 (one ampersand for alignment within each align
structure, and ampersands to separate the align structures from one another).
The special environment flalign is a form of the align environment with
added space between the component align structures.
\begin{align}
L1 = R1 L2 = R2 (8.33) L_1 & = R_1 & \qquad L_2 & = R_2 \\
L3 = R3 L4 = R4 (8.34) L_3 & = R_3 & \qquad L_4 & = R_4
\end{align}
\begin{alignat}{2}
L1 = R1 L2 = R2 (8.35) L_1 & = R_1 & \qquad L_2 & = R_2 \\
L3 = R3 L4 = R4 (8.36) L_3 & = R_3 & \qquad L_4 & = R_4
\end{alignat}
\begin{flalign}
L1 = R1 L2 = R2 (8.37) L_1 & = R_1 & \qquad L_2 & = R_2 \\
L3 = R3 L4 = R4 (8.38) L_3 & = R_3 & \qquad L_4 & = R_4
\end{flalign}
\begin{flalign*}
L1 = R1 L2 = R2 L_1 & = R_1 & \qquad L_2 & = R_2 \\
L3 = R3 L4 = R4 L_3 & = R_3 & \qquad L_4 & = R_4
\end{flalign*}
8.5.3 Split Equations without Alignment

The multline environment is a variation of the equation environment used for
equations that do not fit on a single line. The first line of a multline will be at
the left margin and the last line at the right margin except for an indention on
both sides whose amount is equal to \multlinegap. The value of \multlinegap
can be changed using LATEX’s \setlength and \addtolength commands. If
multline contains more than two lines, any lines other than the first and last
will be centered individually within the display width (unless option fleqn is in
effect). It is, however, possible to force a line to the left or the right with the
\shoveleft and \shoveright commands.
8.5 Alignment Structures for Equations 247
\begin{multline}
First line of equation \text{First line of equation} \\
\text{Centered Middle line} \\
Centered Middle line
\shoveright{\text{Right Middle line}} \\
Right Middle line \text{Other centered Middle} \\
Other centered Middle \shoveleft{\text{Left Middle line}} \\
Left Middle line \text{Last line of equation}
\end{multline}
Last line of equation (8.39)
8.5.4 Split Equations with Alignment

Like multline, the split environment is for single equations that are too long to
fit on a single line and hence must be split into multiple lines. Unlike multline,
however, the split environment provides for alignment among the split lines,
using an ampersand to mark alignment points, as usual. In addition (unlike the
other amsmath equation structures) the split environment provides no number-
ing because it is intended to be used only inside some other displayed equation
structure, such as equation, align, or gather. These outer environments will
provide the numbering.
\begin{equation}
(a + b)4 = (a + b)2 (a + b)2 \begin{split}
= (a2 + 2ab + b2 )(a2 + 2ab + b2 ) (8.40) (a+b)^4 &= (a+b)^2 (a+b)^2 \\
= a4 + 4a3 b + 6a2 b2 + 4ab3 + b4 &= (a^2+2ab+b^2)(a^2+2ab+b^2) \\
&= a^4+4a^3b+6a^2b^2+4ab^3+b^4 \\
\end{split}
\end{equation}
When the tbtags option is specified, the equation number for the split
environment will be put on the last (resp. first) line if the equation number is on
the right (resp. left). By default, the centertags option is in effect, putting the
equation number centered vertically on the height of the split, provided there
is enough room for it.
\begin{equation}
(a + b)3 = (a + b)(a + b)2 \begin{split}
= (a + b)(a2 + 2ab + b2 ) (a+b)^3 &= (a+b) (a+b)^2 \\
= a3 + 3a2 b + 3ab2 + b3 &= (a+b)(a^2+2ab+b^2) \\
(8.41)
&= a^3+3a^2b+3ab^2+b^3 \\
\end{split}
\end{equation}
8.5.5 Alignment Environments as Parts of Displays

In addition to the split environment, there are some other equation alignment
environments that do not constitute an entire display. They are self-contained
units that can be used inside other formulae, or set side by side. The environ-
ment names are: aligned, gathered, and alignedat. These environments take
an optional argument to specify their vertical positioning with respect to the
material on either side. The default alignment is centered ([c]), and its effect
is seen in the following example.
\begin{equation*}
x2 + y 2 = 1 (a + b)2 = a2 + 2ab + b2 \begin{aligned}
p
x = 1 − y2 (a + b) · (a − b) = a2 − b2 x^2 + y^2 & = 1 \\
x & = \sqrt{1-y^2}
\end{aligned} \qquad
\begin{gathered}
(a + b)^2 = a^2 + 2ab + b^2 \\
(a + b) \cdot (a - b) = a^2 - b^2
\end{gathered}
\end{equation*}
The same mathematics can now be typeset using different vertical alignments
for the environments.
\begin{equation*}
x2 + y 2 = 1 \begin{aligned}[b]
p
x = 1 − y2 (a + b)2 = a2 + 2ab + b2 x^2 + y^2 & = 1 \\
(a + b) · (a − b) = a − b
2 2 x & = \sqrt{1-y^2}
\end{aligned} \qquad
\begin{gathered}[t]
(a + b)^2 = a^2 + 2ab + b^2 \\
(a + b) \cdot (a - b) = a^2 - b^2
\end{gathered}
\end{equation*}
8.5.6 Vertical Spacing and Page Breaks in Equation

Structures
You can use the \\[dimension] command to get extra vertical space between
lines in all the amsmath displayed equation environments, as is usual in LATEX.
Unlike eqnarray, the amsmath environments do not allow page breaks between
lines, unless \displaybreak or \allowdisplaybreaks is used. The reason
8.6 Miscellaneous 249
for this is that page breaks in such situations should receive individual at-
tention from the author. \displaybreak must go before the \\ where it is
supposed to take effect. Like LATEX’s \pagebreak, \displaybreak takes an
optional argument between zero and four denoting the desirability of the page
break. \displaybreak[0] means “it is permissible to break here” without en-
couraging a break; \displaybreak with no optional argument is the same as
\displaybreak[4] and forces a break.
There is also an optional argument for \allowdisplaybreaks. This com-
mand obeys the usual LATEX scoping rules. The normal way of limiting its scope
is to put {\allowdisplaybreaks at the beginning and } at the end of the de-
sired range. Within the scope of an \allowdisplaybreaks command, the \\*
command can be used to prohibit a page break, as usual.
8.5.7 The \intertext Command

The \intertext command is used for a short interjection of one or two lines of
text in the middle of a display alignment. Its salient feature is the preservation
of alignment, which would not be possible if you simply ended the display and
then started it up again afterwards. \intertext may only appear immediately
after a \\ or \\* command.
\begin{align}
A1 = N0 (λ; Ω0 ) − φ(λ; Ω0 ), (8.42) A_1&=N_0(\lambda;\Omega’) -
A2 = φ(λ; Ω0 )φ(λ; Ω), (8.43) \phi(\lambda;\Omega’), \\
A_2&=\phi(\lambda;\Omega’)
and finally \phi(\lambda;\Omega), \\
\intertext{and finally}
A_3&=\mathcal{N}(\lambda;\omega).
A3 = N (λ; ω). (8.44)
\end{align}
Here the words “and finally” fall outside the display at the left margin.
8.6 Miscellaneous
This section discusses amsmath commands that have not been introduced yet,
and it gives a list of the document class files that come with the AMS-LATEX
distribution.
8.6.1 Equation Numbers

Each environment, except for split, has both starred and unstarred forms,
where the unstarred forms have automatic numbering, using LATEX’s equation
counter. The number on any particular line can be suppressed by putting \notag
before the \\. You can also override it with a tag of your own design using
\tag{label } \tag*{label }
where label can be any arbitrary text to be used to number the equation.
The starred form, \tag*, causes the label to be typeset without any anno-
tations like parentheses that might otherwise be added by the document class.
\tag and \tag* can also be used in the starred versions of all the amsmath
alignment environments.
\begin{gather}
x2 + y 2 = z 2 (8.45) x^2+y^2 = z^2 \label{eq:r2} \\
3 3 3 x^3+y^3 = z^3 \notag \\
x +y =z
x^4+y^4 = r^4 \tag{$*$} \\
x4 + y 4 = r 4 (∗)
x^5+y^5 = r^5 \tag*{$*$} \\
5
x +y =r 5 5
∗ x^6+y^6 = r^6 \tag{\ref{eq:r2}$’$}
x6 + y 6 = r6 (8.450 ) \end{gather}
Notice the use of the \label and \ref commands in the previous example
to allow subnumbering of equations.
When leqno is specified as an option to the amsmath package, the equation
number will be printed at the left side of the equation (by default, with amsmath,
it comes out at the right).
\begin{equation}
(8.46) sin2 η + cos2 η = 1 \sin^2\eta + \cos^2\eta = 1
\end{equation}
8.6.2 Resetting the Equation Counter

In LATEX, if you want to have equations numbered within sections—that is, have
equation numbers (1.1), (1.2), . . . , (2.1), (2.2), . . . , in sections 1, 2, and so
forth—you would probably redefine \theequation:
\renewcommand{\theequation}{\thesection.\arabic{equation}}
But now you have to reset the equation number by hand at the beginning of
each new section or chapter. To make this a little more convenient, amsmath pro-
vides a command \numberwithin. To have equation numbering tied to section
numbering, with automatic reset of the equation counter, the command is
\numberwithin{equation}{section}
As the name implies, \numberwithin can be applied to other counters be-

sides the equation counter, but the results may not be satisfactory in all cases
because of potential complications. Normal LATEX methods should be used where
available, for example, in \newtheorem.6
To make cross-references to equations easier, an \eqref command is pro-
vided. This automatically supplies the parentheses around the equation num-
ber, and adds an italic correction before the closing parenthesis, if necessary. To
refer to an equation that was labeled with the label e:baset, the usage would
be \eqref{e:baset}.
8.6.3 Subordinate numbering sequences

The amsmath package provides also a subequations environment to make it
easy to number equations in a particular group with a subordinate numbering
scheme. For example
\begin{subequations}
...
\end{subequations}
causes all numbered equations within that part of the document to be numbered
(4.9a) (4.9b) (4.9c) . . . , if the preceding numbered equation was (4.8). A \label
command immediately following \begin{subequations} produces a \ref of the
parent number 4.9, not 4.9a. The counters used by the subequations environ-
ment are parentequation and equation. They can be set by the LATEX com-
mands \addtocounter, \setcounter, \value, etc.. Moreover, the style of the
subordinate numbers, are controlled using standard LATEX methods (see Section
A.1.3). For example, redefining \theequation as follows will produce roman
numerals.
\begin{subequations}
\renewcommand{\theequation}{\theparentequation \roman{equation}}
...
8.6.4 Fine-Tuning Spacing in Math Mode

Although TEX generally does a good job of spacing elements of formulae inside
mathematics, it is sometimes necessary to fine-tune the position of one or two
of those elements. Therefore, the spacing commands shown in table 8.21 on the
next page are provided. Both the spelled-out and abbreviated forms of these
commands are robust, and they can also be used outside of math.
6 See also the discussion of the \@addtoreset command on page 23.
Positive space Negative space

Abb. ex. Spelled out Abb. ex. Spelled out
\, xx \thinspace \! xx \negthinspace
\: xx \medspace xx \negmedspace
\; xx \thickspace xx \negthickspace
x x \quad
x x \qquad
Table 8.21: The mathematical spacing commands
For allow you to further fine-tune the spacing in math expressions the com-
mand \mspace is defined. Its only argument is a LATEX length expressed in
‘math units’. One math unit, or mu, is equal to 1/18 em (see also table A.1 on
page 476). Thus, to get a negative \quad you could write \mspace{-18.0mu}.
8.6.5 A Few Points to Note

(L 151–52) Many of the commands added by the amsmath package are fragile and will need
to be \protected in commands with “moving arguments.”
With the various alignment environments available in the amsmath package,
the eqnarray environment is no longer needed. Furthermore, since it does not
prevent overlapping of the equation numbers with wide formulae, as most of the
amsmath alignments do, using the amsmath alignments seems better. amsmath
reimplements the LATEX equation environment as a one-line gather environ-
ment, and adds an unnumbered version, equation*, for symmetry. Note, how-
ever, that the command \verb might not work in the alignment environments.
\nonumber is interchangeable with \notag; the latter seems slightly prefer-
able, for consistency with the name \tag.
8.6.6 Options and Sub-Packages to the amsmath Package

A few options are recognized by the amsmath package and the classes provided
by AMS-LATEX.7 They affect the positioning of math operator limits or \tags.
centertags (default) The text of the tag of a split environment is vertically

centered with respect to its total height.
tbtags “Top-or-bottom tags”. The text of the tag of a split environment is
placed level with the last (resp. first) line, if numbers are on the right
(resp. left).
7 This is only true for the LATEX 2ε release of AMS-LATEX. Older versions of AMS-LATEX
realize these options as sub-packages.
intlimits Like sumlimits, but for integral symbols.

nointlimits (default) Opposite of intlimits.
namelimits (default) Like sumlimits, but for certain “operator names” such as
det, inf, lim, max, min, that traditionally have subscripts placed un-
derneath when they occur in a displayed equation.
nonamelimits Opposite of namelimits.
sumlimits (default) Place subscripts and superscripts of summation symbols
above and below, in displayed
Q equations.
` N L This option also affects other
symbols of the same type— , , , , and so forth—but excluding
integrals (see intlimits).
nosumlimits Place subscripts and superscripts of summation-type symbols to
the side, even in displayed equations.
The following three options are usually global document options and are
thus set on the \documentclass command. They are, however, also recognized
when the amsmath package is loaded with the \usepackage command.
leqno Place equation numbers on the left.

reqno Place equation numbers on the right (default).
fleqn Position equations at a fixed indent from the left margin rather than
centered in the text column.
The AMS-LATEX distribution consists of a set of components, which can be

loaded independently with the \usepackage command. The single most note-
worthy package is probably amsmath, but the others can be used individually.
Note that the amsbsy, amsopn, and amstext packages are included automatically
when you use the amsmath package.
amsmath Defines extra environments for multiline displayed equations, plus a

number of other enhancements for math.
amsbsy Defines the \boldsymbol and \pmb (poor man’s bold) commands.
amsopn Provides \DeclareMathOperator for defining new “operator names”
like \sin and \lim.
amstext Provides a \text command for typesetting a fragment of text inside a
display.
Other packages, providing supplementary functionality, should be loaded

explicitly. Only parts of these packages are described in the present chapter.
They are mentioned here for completeness.
amscd Defines some commands for easing the generation of commutative dia-
grams by introducing the CD environment (see Section 8.4.4). There is
no support for diagonal arrows.
amsintx Provides more descriptive command syntax for integrals and sums (not
released yet).
amsthm Provides a proof environment and extensions for the \newtheorem
command.
amsxtra Provides certain odds and ends such as \fracwithdelims and
\accentedsymbol (see Section 8.3.4).
upref Makes \ref print cross-reference numbers always in an upright/roman
font regardless of context.
Finally, there are a few packages which come with the AMSFonts distribution.
amsfonts defines the \mathfrak and \mathbb commands and sets up the fonts
msam (extra math symbols A), msbm (extra math symbols B, and black-
board bold), eufm (Euler Fraktur), extra sizes of cmmib (bold math
italic and bold lowercase Greek), and cmbsy (bold math symbols and
bold script), for use in mathematics.
amssymb defines the names of all the math symbols available with the AMS
fonts collection. This package loads the amsfonts package.
eufrak Set up the Fraktur letters.
eucal Makes \mathcal use the Euler script instead of the usual Computer
Modern script letters.
All these packages recognize the psamsfonts option, which will use the Y&Y/Blue
Sky Research version of the AMSFonts collection (which is free available on
CTAN).
8.6.7 AMS-LATEX Document Classes

The AMS-LATEX package comes with a pair of document classes called amsart
and amsbook, corresponding to LATEX’s article and book. They are primarily
designed to prepare manuscripts for submission to the AMS, but there is nothing
to prohibit their use for other purposes. With these class files the amsmath
package is automatically included, so that you can start your document simply
with \documentclass{amsart} or \documentclass{amsbook}.
8.7 Examples of Multiple-Line Equation Structures 255
8.7 Examples of Multiple-Line Equation

Structures
On the following pages we show a lot of real-life examples of the alignment
environments discussed earlier. The lines indicating the margins around the
typeset examples are not part of the environments but have been added to make
the marginal spacing stand out clearly.
8.7.1 The split Environment

The split environment is not an independent environment but should be used
inside something else, such as equation or align.
If there is not enough room for it, the equation number for a split will be
shifted to the previous line when equation numbers are on the left; the number
shifts down to the next line when numbers are on the right.
When you do not want an equation number, use the equation* environment.
Z tε
fh,ε (x, y) = εEx,y Lx,yε (εu) ϕ(x) du
Z 0
= h Lx,z ϕ(x)ρx (dz)

Z tε Z
1
+h Ey Lx,yx (s) ϕ(x) ds − tε Lx,z ϕ(x)ρx (dz)
tε
0 Z tε Z tε
1
+ Ey Lx,yx (s) ϕ(x) ds − Ex,y Lx,yε (εs) ϕ(x) ds
tε 0 0
(8.47)
This was produced by the following input (the TEX command \phantom is used
to leave a space equal to the width of its argument):
\begin{equation}
\begin{split}
f_{h,\varepsilon}(x,y)
&= \varepsilon \mathbf{E}_{x,y} \int_0^{t_\varepsilon}
L_{x,y_\varepsilon(\varepsilon u)} \varphi(x) \, du \\
&= h \int L_{x,z} \varphi(x) \rho_x (dz) \\
& \quad +h \biggl[ \frac{1}{t_\varepsilon} \biggl(
\mathbf{E}_{y} \int_0^{t_\varepsilon} L_{x,y^x(s)} \varphi(x)\,ds
-t_\varepsilon \int L_{x,z} \varphi(x) \rho_x(dz) \biggr) \\
& \phantom{{=}+h\biggl[}+\frac{1}{t_\varepsilon}
\biggl( \mathbf{E}_{y} \int_0^{t_\varepsilon} L_{x,y^x(s)}
\varphi(x) \,ds - \mathbf{E}_{x,y} \int_0^{t_\varepsilon}
L_{x,y_\varepsilon(\varepsilon s)}
\varphi(x) \,ds \biggr) \biggr] \\
\end{split}
\end{equation}
If the option centertags is included in the options list of the amsmath package,
the equation numbers for split environments will be centered vertically on the
height of the split, as shown in the example below.
Z ( Z a Z θ )
T
dθ
|I2 | = ψ(t) u(a, t) − c(ξ)ut (ξ, t) dξ dt
0 γ(t) k(θ, t) a
Z (8.48)
e−1,0 ◦ e
≤ C6 f Sa,− W2 (Ω, Γl ) |u| → W2A (Ω; Γr , T ) .
Ω
This is produced by the following input:
\begin{equation}
\begin{split}
|I_2|
&=\left| \int_{0}^T \psi(t)
\left\{ u(a,t)-\int_{\gamma(t)}^a \frac{d\theta}{k(\theta,t)}
\int_{a}^\theta c(\xi)u_t(\xi,t)\,d\xi \right\} dt
\right| \\
&\le C_6 \left| \left|
f\int_\Omega
\left| \widetilde{S}^{-1,0}_{a,-} W_2(\Omega,\Gamma_l)
\right|
\right|
\left|
|u|\overset{\circ}\to W_2^{\widetilde{A}}
(\Omega;\Gamma_r,T)
\right| \right|.
\end{split}
\end{equation}
One further example involving split and align. To obtain unnumbered

equations use the align* environment instead.
Z

|I1 | = gRu dΩ

Ω
"Z Z 2 #1/2
x
≤ C3 g(ξ, t) dξ dΩ
Ω a
(8.49)
"Z ( Z 2 ) #1/2
x
1
× u2x + cut dξ cΩ
Ω k a

−1,0 ◦ e
≤ C4 f Sea,− W2 (Ω, Γl ) |u| → W2A (Ω; Γr , T ) .
Z ( Z a Z θ )
T
dθ
|I2 | = ψ(t) u(a, t) − c(ξ)ut (ξ, t) dξ dt
0 γ(t) k(θ, t) a
Z
(8.50)
e−1,0 ◦ e
≤ C6 f Sa,− W2 (Ω, Γl ) |u| → W2A (Ω; Γr , T ) .
Ω
The input for the above formulae is:
\begin{align}
\begin{split}
|I_1| &= \left| \int_\Omega gRu \,d\Omega \right| \\
&\le C_3 \left[ \int_\Omega \left( \int_{a}^x
g(\xi,t) \,d \xi \right)^2d \Omega \right]^{1/2} \\
&\quad\times \left[ \int_\Omega \left\{ u^2_x + \frac{1}{k}
\left( \int_{a}^x cu_t \, d\xi \right)^2 \right\}
c \Omega \right]^{1/2} \\
&\le C_4 \left| \left| f \left| \widetilde{S}^{-1,0}_{a,-}
W_2(\Omega,\Gamma_l) \right| \right|
\left| |u| \overset{\circ} \to W_2^{\widetilde{A}}
(\Omega;\Gamma_r,T) \right| \right|.
\end{split}\label{eq:A} \\
\begin{split}
|I_2| &= \left| \int_{0}^T \psi(t) \left\{ u(a,t)
-\int_{\gamma(t)}^a \frac{d\theta}{k(\theta,t)}
\int_{a}^\theta c(\xi) u_t(\xi,t) \,d \xi \right\} dt
\right| \\
&\le C_6 \left| \left| f \int_\Omega
\left| \widetilde{S}^{-1,0}_{a,-}
W_2(\Omega,\Gamma_l) \right| \right|
\left| |u| \overset{\circ} \to W_2^{\widetilde{A}}
(\Omega;\Gamma_r,T) \right| \right|.
\end{split}
\end{align}
8.7.2 The multline Environment

Numbered version:
Z b Z b
[f (x)2 g(y)2 + f (y)2 g(x)2 ] − 2f (x)g(x)f (y)g(y) dx dy
a a
Z b Z b Z b Z b
= g(y)2 f 2 + f (y)2 g 2 − 2f (y)g(y) f g dy (8.51)
a a a a
This was obtained with the lines shown below.
\begin{multline}\label{eq:E}
\int_a^b \biggl\{ \int_a^b [ f(x)^2 g(y)^2 + f(y)^2 g(x)^2 ]
-2f(x) g(x) f(y) g(y) \,dx \biggr\} \,dy \\
=\int_a^b \biggl\{ g(y)^2 \int_a^b f^2 + f(y)^2
\int_a^b g^2 - 2f(y) g(y) \int_a^b fg \biggr\} \,dy
\end{multline}
An unnumbered version of the above is obtained with the same input, except
the multline environment is replaced by multline*.
Z b Z b
2 2 2 2
[f (x) g(y) + f (y) g(x) ] − 2f (x)g(x)f (y)g(y) dx dy
a a
Z b Z b Z b Z b
2 2 2 2
= g(y) f + f (y) g − 2f (y)g(y) f g dy
a a a a
And now an unnumbered version numbered with a \tag* command.

Z b Z b
2 2 2 2
[f (x) g(y) + f (y) g(x) ] − 2f (x)g(x)f (y)g(y) dx dy
a a
Z b Z b Z b Z b
2 2 2 2
= g(y) f + f (y) g − 2f (y)g(y) f g dy [a]
a a a a
This was generated with:
\begin{multline*}\tag*{[a]} ... \end{multline*}
This is the same display, but with \multlinegap set to zero. Notice that the
space on the left of the first line does not change, because of the equation number,
while the second line is pushed over to the right margin.
Z b Z b
[f (x)2 g(y)2 + f (y)2 g(x)2 ] − 2f (x)g(x)f (y)g(y) dx dy
a a
Z b Z b Z b Z b
= g(y)2 f 2 + f (y)2 g 2 − 2f (y)g(y) f g dy [a]
a a a a
{\setlength{\multlinegap}{0pt}
\begin{multline*}\tag*{[a]} ... \end{multline*}}
8.7.3 The gather Environment

Numbered version with \notag on the second line:
D(a, r) ≡ {z ∈ C : |z − a| < r}, (8.52)

seg(a, r) ≡ {z ∈ C : =z = =a, |z − a| < r},
c(e, θ, r) ≡ {(x, y) ∈ C : |x − e| < y tan θ, 0 < y < r}, (8.53)
[
C(E, θ, r) ≡ c(e, θ, r). (8.54)
e∈E
\begin{gather}
D(a,r) \equiv \{ z \in \mathbf{C}: |z-a|<r \}, \\
\operatorname{seg}(a,r) \equiv \{ z \in \mathbf{C}:
\Im z = \Im a, \ |z-a|<r\}, \notag \\
c(e,\theta,r) \equiv \{ (x,y) \in \mathbf{C}:
|x-e|<y \tan \theta, \ 0<y<r \}, \\
C(E,\theta,r) \equiv \bigcup_{e \in E}c(e,\theta,r).
\end{gather}
8.7.4 The align Environment

Numbered version:
γx (t) = (cos tu + sin tx, v), (8.55)
γy (t) = (u, cos tv + sin ty), (8.56)

α β
γz (t) = cos tu + sin tv, − sin tu + cos tv . (8.57)
β α
This was produced using the following input:
\begin{align}
\gamma_x(t) &= (\cos tu + \sin tx, v), \\
\gamma_y(t) &= (u, \cos tv + \sin ty), \\
\gamma_z(t) &= \left( \cos tu + \frac\alpha\beta \sin tv,
- \frac\beta\alpha \sin tu + \cos tv \right).
\end{align}
Unnumbered version:
γx (t) = (cos tu + sin tx, v),

γy (t) = (u, cos tv + sin ty),

α β
γz (t) = cos tu + sin tv, − sin tu + cos tv .
β α
This was generated using the following construct:
\begin{align*} ... \end{align*}
8.7.5 Using the align and split Environments within

gather
When using the align environment within the gather environment, one or the
other, or both, should be unnumbered (using the * form), since having numbering
for both the outer and inner environment would not be meaningful.
Automatically numbered gather with split and align*:
X
ϕ(x, z) = z − γ10 x − γmn xm z n
m+n≥2
X (8.58)
−1
= z − Mr x− M r −(m+n) xm z n
m+n≥2
ζ 0 = (ξ 0 )2 ,
ζ 1 = ξ0ξ1
Here the split environment gets a number from the outer gather environment;
numbers for individual lines of the align* are suppressed because of the star.
\begin{gather}
\begin{split}
\varphi(x,z)
&= z - \gamma_{10} x - \sum_{m+n\ge2} \gamma_{mn} x^m z^n \\
&= z - M r^{-1} x - \sum_{m+n\ge2} M r^{-(m+n)} x^m z^n
\end{split} \\[6pt]
\begin{align*}
\zeta^0 &= (\xi^0)^2, \\
\zeta^1 &= \xi^0 \xi^1
\end{align*}
\end{gather}
Shown below, is the *-ed form of gather with the non-*-ed form of align.
X
ϕ(x, z) = z − γ10 x − γmn xm z n
m+n≥2
X
−1
= z − Mr x− M r −(m+n) xm z n
m+n≥2
ζ 0 = (ξ 0 )2 , (8.59)
1 0 1
ζ =ξ ξ (8.60)
The latter was produced with the following construct:
\begin{gather*}
\begin{split} ...\end{split} \\[6pt]
\begin{align}...\end{align}
\end{gather*}
8.7.6 Using the alignat Environments

Numbered version:
Vi = vi − qi vj , Xi = xi − qi xj , Ui = ui , for i 6= j ; (8.61)
X
Vj = vj , X j = xj , Uj uj + qi ui . (8.62)
i6=j
This example was obtained with the commands below:
\begin{alignat}{3}
V_i &= v_i - q_i v_j, & \qquad X_i &= x_i - q_i x_j,
& \qquad U_i &= u_i, \qquad \text{for $i\ne j\,$;}\label{eq:B} \\
V_j &= v_j, & \qquad X_j &= x_j,
& \qquad U_j & u_j + \sum_{i\ne j} q_i u_i.
\end{alignat}
Unnumbered version:
Vi = vi − qi vj , Xi = xi − qi xj , Ui = ui , for i 6= j ;
X
Vj = vj , X j = xj , Uj uj + qi ui .
i6=j
\begin{alignat*}{3} ... \end{alignat*}
The most common use for alignat is for things like
x=y by (8.49) (8.63)

x0 = y 0 by (8.61) (8.64)
x + x0 = y + y 0 by Axiom 1. (8.65)
This example was obtained with the commands below:
\begin{alignat}{2}
x &= y && \qquad \text{by (\ref{eq:A})}\label{eq:C} \\
x’ &= y’ && \qquad \text{by (\ref{eq:B})}\label{eq:D} \\
x + x’ &= y+y’ && \qquad \text{by Axiom 1.}
\end{alignat}
The expanded version, flalign:
x=y by (8.63) (8.66)

x0 = y 0 by (8.64) (8.67)
x + x0 = y + y 0 by Axiom 1. (8.68)
\begin{flalign} ... \end{flalign}
8.8 Extensions to the theorem Environment

AMS-LATEX comes with the amsthm package, which extends LATEX’s
\newtheorem command. Rather than describe amsthm (see, for instance, the
section “Proclamations” in Grätzer’s book [?] for more details) we will give
some details about the theorem package, developed by Frank Mittelbach [?]. It
(L 58, 174) also offers an extension of the LATEX theorem mechanism by allowing the layout
of theorems to be manipulated by specifying a style.
In the present context the word “theorem” is used for any kind of labeled
enunciations, often set off from the main text by extra space and a font change.
Theorems, corollaries, conjectures, definitions, and remarks are all instances of
“theorems.” The header of these structures is composed of a label (such as
Theorem or Remark) and a number, which serializes an item in the sequence
of items with the same label.
Often it is necessary, in order to satisfy the requirements of different math-
ematics journals, to customize the layout of the theorem environment. Addi-
tionally, different formats may be needed to differentiate the “sort of theorem”:
e.g., remarks and definitions are set in roman, while italic is employed for main
theorems.
8.8 Extensions to the theorem Environment 263
8.8.1 Defining New Theorem Environments

As in the original LATEX version, the command \newtheorem defines a new
“theorem-like structure.” Two required arguments name the new environment
and give the text to be typeset with each instance of the new environment, while
an optional argument determines how the environment is enumerated:
\newtheorem{env-name}{label-text}
The above \newtheorem command defines the env-name environment and

its printed name will be label-text. It uses its own counter.
\newtheorem{env2-name}[env-name]{label-text2 }
The above \newtheorem command defines the env2-name environment, and

its printed name will be label-text2 . It uses the same counter as theorem set
env-name.
\newtheorem{env3-name}{label-text3 }[section]
The above variant defines the env3-name environment and its printed name
is label-text3 . Its counter is enumerated within the counter section, that is, with
every new \section the enumeration starts again with one, and the enumeration
is composed from the section number and the theorem counter itself.
\theoremstyle{style}
The \theoremstyle command can define the layout of various, or all, the-
orem sets. It should be noted that any theorem set defined by \newtheorem is
typeset in the \theoremstyle that is current at the time of the definition.
Thus, the following
\theoremstyle{break} \newtheorem{Cor}{Corollary}
\theoremstyle{plain} \newtheorem{Exa}{Example}[section]
leads to the result that the set Cor is formatted in the style break, while the
set Exa and all the following ones are formatted in the style plain, unless an-
other \theoremstyle follows. Since the definitions installed by \newtheorem
are global, you can also limit \theoremstyle locally by grouping braces.
\theorembodyfont{font-declarations}
The choice of the font for the theorem body is completely independent of the
chosen \theoremstyle; this has proven to be very advantageous. For example,
{\theorembodyfont{\rmfamily} \newtheorem{Rem}{Remark}}
plain Emulates the original LATEX definition, except that ad-

ditionally the parameters \theorempreskipamount and
\theorempostskipamount are used.
break In this style, the theorem header is followed by a line break.
marginbreak The theorem number is set in the margin, and there is a line
break as in break.
changebreak Like break, but with header number and text interchanged.
change Header number and text are interchanged, without a line break.
margin The number is set in the left margin, without a line break.
Table 8.22: List of existing theorem styles

All styles (except plain) select \normalfont\slshape as the default for
\theorembodyfont.
defines a theorem set Rem, which will be set in \rmfamily in the cur-
rent layout (which in our example is plain). As with \theoremstyle, the
\theorembodyfont chosen is that which is current at the time of \newtheorem.
If \theorembodyfont is not specified or you define \theorembodyfont{}, then
the font used will be defined by \theoremstyle.
\theoremheaderfont{font-declarations}
It is also possible to customize the font used for the theorem headers. This
is, however, a global declaration and, therefore, there should be at most one
\theoremheaderfont command in the preamble. If it is actually necessary to
have different header fonts, you will have to define new theorem styles (substi-
tuting the desired font).
Two additional parameters affect the vertical space around the theorem en-
vironments: \theorempreskipamount and \theorempostskipamount define, re-
spectively, the spacing before and after such an environment. These parameters
apply to all theorem sets and can be manipulated with the ordinary length
macros. They are rubber lengths, and therefore can contain plus and minus
parts. These parameters are set using the \setlength command.
The commands to define theorem sets, as described in this section, can only
be placed in the document preamble or in a package file.
Theorem styles, which exist to date, are shown in table 8.22
8.8.2 Examples of the Definition and Use of Theorems

Suppose that the preamble contains the declarations:
\theoremstyle{break} \newtheorem{Cor}{Corollary}
\theoremstyle{plain} \newtheorem{Exa}{Example}[section]
8.8 Extensions to the theorem Environment 265
{\theorembodyfont{\rmfamily} \newtheorem{Rem}{Remark}}
\theoremstyle{marginbreak} \newtheorem{Lem}[Cor]{Lemma}
\theoremstyle{change}
\theorembodyfont{\itshape} \newtheorem{Def}[Cor]{Definition}
\theoremheaderfont{\scshape}
Then the typical examples below show the typeset output resulting from their
use.
Corollary 1 \begin{Cor}
This is a sentence typeset in the theorem This is a sentence typeset in the theorem
environment Cor. environment \Lenv{Cor}.
\end{Cor}
Example 8.8.1 This is a sentence typeset \begin{Exa}

in the theorem environment Exa. This is a sentence typeset in the theorem
environment \Lenv{Exa}.
\end{Exa}
Remark 1 This is a sentence typeset in the \begin{Rem}

theorem environment Rem. This is a sentence typeset in the theorem
environment \Lenv{Rem}.
\end{Rem}
2 Lemma (Ben User) \begin{Lem}[Ben User]

This is a sentence typeset in the theorem This is a sentence typeset in the theorem
environment Lem. environment \Lenv{Lem}.
\end{Lem}
3 Definition (Very impressive Definition) \begin{Def}[Very impressive Definition]

This is a sentence typeset in the theorem en- This is a sentence typeset in the theorem
vironment Def. environment \Lenv{Def}.
\end{Def}
The last two examples show the effect of the optional argument to a theorem
environment (it is typeset in parentheses right after the label).
8.8.3 Special Considerations

The theorem header and body are implemented as a single unit. This means that
the \theoremheaderfont will inherit characteristics of the \theorembodyfont
if the NFSS is being used. Thus, if, for example, \theorembodyfont is \itshape
and \theoremheaderfont is \bfseries the font selected for the header will have
the characteristics “bold extended italic.” If this is not desired you should set
it to something like \theoremheaderfont{\normalfont\bfseries}. That is,
you should supply all the necessary font information explicitly. See chapter 7 for
more details about how to do that.
8.9 Mathematical Style Parameters

This section explains how you can globally control the style of your mathematical
formulae, and how you can modify the size of certain (sub)formula elements.
8.9.1 Controlling the Size of Characters

Letters and mathematical symbols sometimes get smaller when they appear in
fractions, superscripts, or subscripts. In fact, TEX has eight different styles in
which it can treat formulae, namely:
D, D0 \displaystyle formulae displayed on lines by themselves

T, T 0 \textstyle formulae embedded in the text
S, S 0 \scriptstyle formulae used as super- or subscripts
SS, SS 0 \scriptscriptstyle second- and higher-order super- or subscripts
The accented symbols represent the so-called cramped styles, which are sim-
ilar to the normal styles except that exponents are not raised so much. TEX also
uses three different type sizes for mathematics, namely: text size, script size,
and scriptscript size.
A formula set inside text (between a $ pair, or between $...$) is typeset
using text style (style T ). A formula on a line by itself, e.g., entered between
\[...\], will be typeset in display style (style D). The size of the different parts
of a formula can be determined according to the following scheme:
A symbol in style will be typeset in (example)

D, D0 , T, T 0 text size (text size)
S, S 0 script size (script size)
SS, SS 0 scriptscript size (scriptscript size)
The kind of style used in mathematics formulae is as follows:
8.9 Mathematical Style Parameters 267
style superscript subscript numerator denominator

D S S0 T T0
D0 S0 S0 T0 T0
T S S0 S S0
T0 S0 S0 S0 S0
S, SS SS SS 0 SS SS 0
S 0 , SS 0 SS 0 SS 0 SS 0 SS 0
The last two columns describe the style used in the numerator or denomina-
tor of a fraction. An example of the various styles can be seen in the continued
fraction below (see also section 8.3.16):
\normalsize
\[ b^0 + \frac{a^1}{b_1 +
a1
b0 + 2
\frac{a^2}{b_2 +
b1 + a a3 \frac{a^3}{b_3}}}
b2 + b
3 \]
In the formula above the b of b0 is in style D, with the 0 in style S; the a

and b of a1 and b1 are in style T and T 0 , respectively, with the exponent 1 in
style S and the subscript 1 in style S 0 ; the a and b of a2 and b2 are both in style
S 0 , with the exponent and subscript in style SS 0 ; finally everything in a3 and b3
is in style SS 0 .
You can give a nicer look to the above example by deciding which style is
to be used in each case. Note that to save typing, we define the abbreviation \D
for the \displaystyle command.
\newcommand{\D}{\displaystyle}
1 \normalsize
a
b0 + \[ b^0 + \frac{a^1}{\D b_1 +
a2 \frac{a^2}{\D b_2 +
b1 +
a3 \frac{a^3}{b_3}}}
b2 + \]
b3
8.9.2 LATEX Math Style Parameters

Because LATEX uses much of the mathematical machinery from TEX, we briefly (L 170)
describe the mathematical style parameters that LATEX uses to typeset formulae.
All these are length parameters which you can redefine with the \setlength or
\addtolength commands (see section A.1.4 on page 474). Moreover, two stan-
dard options, leqno and fleqn, control the numbering and alignment of formulae. (L 82)
The option fleqn causes formulae to be aligned on the left, a fixed distance from
the left margin (see \mathindent below), instead of being centered.
The option leqno causes formula numbers to appear on the left instead of at
the right (see section 8.6.6 on page 252).
In the list of mathematics style parameters below, all lengths (except
\jot and \arraycolsep) are rubber lengths. With the option fleqn, the four
displayskip lengths are made equal to the list defining length \topsep, to which
the value of \partopsep is added if the display starts a paragraph (see figure 3.5
on page 64). The four parameters \abovedisplay... and \belowdisplay...
below depend on the current font size. For this reason they cannot be modified
in the preamble of the document using \setlength, but they must be changed
by modyfying \normalsize, etc.
\arraycolsep This gives half the width of the horizontal space between
columns in an array environment (default value 5pt, see also section 5.3.2).
\jot This is the extra vertical space that is added between rows in an eqnarray
or eqnarray* environment (default value 3pt).
\mathindent This defines the indentation from the left margin of displayed
formulae for the fleqn option (the default value is equal to the indentation
of a first level list, i.e., 2.5em, and is defined by the option fleqn).
\abovedisplayskip This specifies the extra space left above a long displayed
formula, except with the option fleqn, where \topsep is used. A long formula
is one that lies closer to the left margin than does the end of the preceding
line (default value 12pt plus 3pt minus 9pt).
\belowdisplayskip This specifies the extra space left below a long displayed
formula, except with the option fleqn, where \topsep is used (default value
12pt plus 3pt minus 9pt).
\abovedisplayshortskip This specifies the extra space left above a short dis-
played formula, except with the option fleqn, where \topsep is used. A
short formula is one which starts to the right of where the preceding line
ends (default value 0pt plus 3pt).
\belowdisplayshortskip This specifies the extra space left below a short dis-
played formula, except with the option fleqn, where \topsep is used (default
value 7pt plus 3pt minus 4pt).

LaTeX - Companion PDF

Uploaded by

Copyright:

Available Formats

LaTeX - Companion PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

LaTeX - Companion PDF

Uploaded by

Copyright:

Available Formats

This is an updated version of chapter 8 of the LATE X

Companion reflecting changes in the AMS-LATEX

Basic LATEX offers a high level of mathematical typesetting capabilities. How-

8.1 The AMS-LATEX Project

it have the laborsaving abilities of LATEX for preparing indexes, bibliographies,

8.2 Fonts and Symbols in Formulae

u 6< v or a 6∈ A $u \not< v$ or $a \not\in \mathbf{A}$

â \hat{a} á \acute{a} ā \bar{a} ȧ \dot{a} ă \breve{a}

α \alpha β \beta γ \gamma δ \delta  \epsilon

± \pm ∩ \cap  \diamond ⊕ \oplus

≤ \leq,\le ≥ \geq,\ge ≡ \equiv |= \models ≺ \prec

← \leftarrow ←− \longleftarrow ↑ \uparrow

Table 8.6: Miscellaneous symbols (available in LATEX)

\arccos \cos \csc \exp \ker \limsup \min \sinh

↑ \uparrow ⇑ \Uparrow ↓ \downarrow ⇓ \Downarrow

z \digamma κ \varkappa i \beth k \daleth ‫ג‬ \gimel

p \ulcorner q \urcorner x \llcorner y \lrcorner

V \Rrightarrow \rightsquigarrow ⇔ \leftleftarrows

8 \nleftarrow 9 \nrightarrow : \nLeftarrow

5 \leqq 6 \leqslant 0 \eqslantless

u \dotplus r \smallsetminus e \Cap,\doublecap

~ \hbar } \hslash M \vartriangle

8.2.2 Names of Math Font Commands

To make an entire math formula bold (or as much of it as possible, depending

8.3 Compound Symbols, Delimiters, Operators

8.3.1 Multiple Integral Signs

8.3.2 Over and Under Arrows

A series H1 , H2 , . . . , a regional sum H1 +H2 + A series $H_1,H_2,\dotsc$,

8.3.4 Accents in Math

8.3.5 Superscripted Accents

8.3.6 Dot Accents

8.3.8 Boxed Formulae

8.3.9 Extensible Arrows

8.3.10 \overset, \underset, and \sideset

8.3.11 The \smash Command

8.3.12 The \text Command

..._{\mbox{\scriptsize word or phrase}}

8.3.13 Operator Names

kf k∞ = ess sup|f (x)|

The starred form \DeclareMathOperator* is like \DeclareMathOperator;

8.3.14 \mod and Its Relatives

8.3.15 Fractions and Related Constructions

Other examples are the following re-implementation of TEX’s fraction primitives.

Of course, if you want to use a particular notation implemented with

8.3.16 Continued Fractions

Left or right positioning of any of the numerators is achieved by using the

8.3.17 Big-g-g-g Delimiters

8.4 Matrix-Like Environments and

Notice the use of \text and the embedded math.

8.4.2 The Matrix Environments

The maximum number of columns is determined by the counter

A row of dots in a matrix, spanning a given number of columns, can be

8.4.3 The \substack command

Each line can be left-adjusted instead of centered by using the subarray

8.4.4 Commutative Diagrams

8.5 Alignment Structures for Equations

align align* alignment at a single place

Some of these multiline display environments allow you to align parts of

8.5.1 Equation Groups without alignment

α \alpha β \beta γ \gamma δ \delta \epsilon

± \pm ∩ \cap \diamond ⊕ \oplus

\begin{multline}\tag{[a]} ... \end{multline*}

\begin{align} ... \end{align}

\begin{alignat}{3} ... \end{alignat}