diff --git a/docs/source/statisticaltests.rst b/docs/source/statisticaltests.rst
index 45d0c33..fca3a0e 100644
--- a/docs/source/statisticaltests.rst
+++ b/docs/source/statisticaltests.rst
@@ -553,7 +553,7 @@ The introduction of composite hypotheses in Bayesian statistics transforms Bayes
 
 .. [#] To facilitate the distinction between symbolic constant expressions (a known background) and symbolic parameters (an unknown background) all constant symbols are marked with a tilde: i.e. :math:`\tilde{a}` is constant expression, whereas :math:`a` is a parameter.
 
-.. [#] An notable example of a distribution that has no well-defined mean or variance is the non-relativistic Breit-Wigner distribution.
+.. [#] An notable example of a distribution that has no well-defined variance is the non-relativistic Breit-Wigner distribution.
 
 
 Statistical inference with nuisance parameters
@@ -636,4 +636,4 @@ Response functions and subsidiary measurements
 ==============================================
 *"Sideband fits and systematic uncertainties"*
 
-.. [#] Where the condition is that the POI is fixed at the value :math:`\mu`, rather than allowed to float to the value :math:`\hat{\mu}` in the minimization, as is the case in the minimization of the unconditional estimate :math:`\hat{\theta}`
\ No newline at end of file
+.. [#] Where the condition is that the POI is fixed at the value :math:`\mu`, rather than allowed to float to the value :math:`\hat{\mu}` in the minimization, as is the case in the minimization of the unconditional estimate :math:`\hat{\theta}`