The entire argument of
eqalign must be enclosed in braces.
Successive rows of the equation must be separated by the control
symbol
cr to tell TEX to make a carriage return.
Every row contains two parts (or templates) separated
by the special tabulation symbol &.
This symbol tells TEX what character you want to line up in each
line of the set of equations.
If you want half of some of the lines to be empty that is all right,
but the symbol & must always appear.
Left and right braces within a template must balance separately.
PLAIN TEX sometimes puts aligned equations too close together.
If you want to force
eqalign to spread the lines apart
somewhat there is a special control symbol for doing this.
All you have to do is type
cropen{space}
instead of
cr at the end of each line.
For example,
cropen{12pt} will force
the lines 12pt further apart.
The control sequence
crr is an abbreviation for
cropen{10pt}.
Although it is not obvious from the previous example,
eqalign
won't always put the equation number where you would like it.
For example, if one of the equation lines is very long and
pushes into the right hand margin, then
eqalign will
put the equation number on a separate line below the last equation
in the set.
If, as often happens, there is only one or two long lines and
there is room on one of the other lines for an equation number
you can force TEX to put it there if you wish.
The way to do this is to use the PLAIN TEX control sequence
eqinsert immediately following the
cr or
cropen ending the line you wish to have the equation on.
If you type
$$ … =& …
cr
eqinsert{
equation.number} $$
then you get the equation number you specify on the line
indicated.
The corresponding PHYZZX macros which allow you to do this
while at the same time automatically generating a name and number
for the aligned set of equations are
mideqn
name
If you don't want to name the equation at the same time that
you automatically generate a number for it you can simply use
eqinsert
eq