One more

diff --git a/examples/step-50/doc/intro.dox b/examples/step-50/doc/intro.dox
index 21d34cdb1cfdb36bc7b271c753e6fdb2c383c7c9..10c6c0dad8fd967c362ce37d9e0c34c00f03f709 100644 (file)
--- a/examples/step-50/doc/intro.dox
+++ b/examples/step-50/doc/intro.dox
@@ -29,10 +29,10 @@ in step-16 (but for parallel computations) and a matrix-free version
 discussed in step-37. The goal is to find out which approach leads to
 the best solver for large parallel computations.
 
-This tutorial is based on one of the numerical examples in @cite
-clevenger_par_gmg. Please see that publication for a detailed background on
-the multigrid implementation in deal.II. We will summarize some of the results
-in the following text.
+This tutorial is based on one of the numerical examples in
+@cite clevenger_par_gmg. Please see that publication for a detailed background
+on the multigrid implementation in deal.II. We will summarize some of the
+results in the following text.
 
 Algebraic multigrid methods are obviously the easiest to implement
 with deal.II since classes such as TrilinosWrappers::PreconditionAMG