What's wrong with this?

Error message: "! Missing \right. inserted." (But I don't see it. Everythink works!)

U=\sum\limits_{n=0}^{\infty}\sum\limits_{k=0}^{\infty}\left[(a_{nk}\cos n\varphi + b_{nk}\sin n\varphi)\cos\frac{a\mu_{nk}t}{R} \\
+(c_{nk}\cos n\varphi + d_{nk}\sin n\varphi)\sin\frac{a\mu_{nk}t}{R}\right] J_n\left(\mu_{nk}\frac{\varrho}{R}\right)
  • You have to close the \left[ with \right. before the line break and open with \left. on the next line.
    – TeXnician
    Nov 19, 2018 at 17:54
  • 2
    Using \left and \right in different lines might result in mismatching sizes. It is better to size the brackets with \big or similar.
    – Skillmon
    Nov 19, 2018 at 17:57
  • @Hello I see a parade of questions coming from you over the last few hours. Before asking a new question, if u find the answer of your previous question satisfactory, please consider accepting it :) Nov 19, 2018 at 19:41
  • Since the multline material is automatically in display-style math mode, the \limits qualifiers are redundant.
    – Mico
    Nov 19, 2018 at 19:47

2 Answers 2


Don't overuse \left and \right to auto-size large parentheses. Not only will this get you into trouble if these directives span a line break -- as is the case in your example code -- it will quite often also fail to generate the typographically appropriate sizes for the parentheses. For the case at hand, I believe \Bigl( and \Bigr) is far more appropriate.

I would also like to suggest to you that you use a split environment instead of a multline environment, as there's a fairly natural way to align the two rows to make their common structures more readily apparent. Finally, consider using inline-fraction notation for the three expressions that contain R in the denominator. Oh, and do get rid of the two \limits directives: they do nothing here except add code clutter.

enter image description here

\usepackage{amsmath} % for 'split' env.
U=\sum_{n=0}^{\infty} \sum_{k=0}^{\infty}
\Bigl[&(a^{}_{nk}\cos n\varphi + b^{}_{nk}\sin n\varphi)
\cos(a\mu^{}_{nk}t/R) \\
{}+{}&(c^{}_{nk}\cos n\varphi + d^{}_{nk}\sin n\varphi)
  • 1
    \big size is more than sufficient.
    – egreg
    Nov 19, 2018 at 21:06
  • @egreg - I agree that \big suffices in a way, especially if one goes with a multline environment. If one employs a split environment, though, using \Big helps to reveal, aesthetically speaking, the overall structure of the equation. In my opinion, the equation shows above looks a bit weak if one uses \big instead of \Big. Of course, as the ancient Romans were amply aware, de gustibus non est disputandum...
    – Mico
    Nov 19, 2018 at 22:04

Just for fun:


\newcommand{\bracketed}[1]{\left[\vphantom{\let\\=\relax #1}\right. #1 \left.\vphantom{\let\\=\relax #1}\right]}

U=\sum\limits_{n=0}^{\infty}\sum\limits_{k=0}^{\infty}\bracketed{(a_{nk}\cos n\varphi + b_{nk}\sin n\varphi)\cos\frac{a\mu_{nk}t}{R} \\
+(c_{nk}\cos n\varphi + d_{nk}\sin n\varphi)\sin\frac{a\mu_{nk}t}{R}} J_n\left(\mu_{nk}\frac{\varrho}{R}\right)


  • 1
    Might be for fun, but works.
    – Skillmon
    Nov 19, 2018 at 18:52

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .