# Strange behavior by \qedhere in align environment

I noticed some strange behavior in the align* (or just align) environment. The code

\begin{proof}
\begin{align*}
C=1 \qedhere
\end{align*}
\end{proof}


puts overlaps the tombstone with the equation, whereas

\begin{proof}
\begin{align*}
C &=1 \qedhere
\end{align*}
\end{proof}


properly puts the tombstone to the right.

What is causing this strange behavior? What exactly is the ampersand doing here?

-
The align environment without an & is useless; there's gather for centering equations. –  egreg Oct 19 '11 at 16:07
@egreg I've always used align* without an & and only one equation as if it were equivalent to the displaymath environment. (This is how I noticed the strange behavior I'm asking about.) Is it somehow or incorrect or bad to use align* as displaymath? –  Quinn Culver Oct 19 '11 at 16:10
They aren't equivalent. For one equation there's equation*, for more non aligned equations there's gather*; align* is for aligned equations (say at a relation symbol). Yes, it's bad to use align* as displaymath. –  egreg Oct 19 '11 at 16:13
@egreg I still don't see why they're not equivalent (and hence why it's bad to use align* as displaymath). Is there a technical difference? –  Quinn Culver Oct 19 '11 at 16:26
show 1 more comment

The comments explain why you should use the align environment only when you needs things to be aligned. Otherwise you should use the gather environment.

Not that I am recommending this but, if you insist on using the align environment without really aligning anything, you could just add the & before the \qedhere:

\begin{proof}
\begin{align*}
C=1 &\qedhere
\end{align*}
\end{proof}


or after the end of the align environment. Of course this results in the \qed symbol on the following line

\begin{proof}
\begin{align*}
C=1
\end{align*}\qedhere
\end{proof}


If you do not like the \qed symbol on the subsequent line, you can adjust \belowdisplayskip as follows:

\begin{proof}\belowdisplayskip=-12pt
\begin{align*}
C=1
\end{align*}\qedhere
\end{proof}

-
Thanks. But in the case that I only have one line of math, shouldn't I use the displaymath environment (r, equivalently (I think), $...$) instead of gather? –  Quinn Culver Nov 2 '11 at 18:45
Sure you can, but I thought the question was about using the \qed symbol with align. –  Peter Grill Nov 2 '11 at 18:52
Yes, it was. I'm just wondering if the gather recommendation was appropriate in this case. Is gather* with one line of math equivalent to displaymath? –  Quinn Culver Nov 2 '11 at 19:02
You should post a separate question, so that if others have the same question in the future they will be able to find it. –  Peter Grill Nov 2 '11 at 19:05
Good idea. Done. Thanks. –  Quinn Culver Nov 2 '11 at 19:12
show 1 more comment