diff --git a/Tools/DirectIOART/README.md b/Tools/DirectIOART/README.md
index f61db2abb6b96d0cb762db7195a14e87122756e8..de314f6af076e482fbbf94ea4182171a10e27015 100644
--- a/Tools/DirectIOART/README.md
+++ b/Tools/DirectIOART/README.md
@@ -12,12 +12,12 @@ Detailed lists of the tests that are performed within the particular software re
 * **TFile::Open** (using an AOD input file)
   * master/Athena
   * 21.0/Athena
+  * 21.0/AthSimulation
   * 21.2/AthAnalysis
   * 21.2/AthDerivation
   * 21.2/AnalysisBase
   * 21.2/AnalysisTop
   * 21.3/Athena
-  * 21.3/AthSimulation
   * 21.9/Athena
 * **`Reco_tf.py`** with RAW input file:
   * 21.0/Athena
@@ -122,4 +122,4 @@ art.py grid -v ../source/athena/Tools/DirectIOART/ DirectIOART_20180724_1
 After calling `"art.py"`, the keyword `"grid"` tells ART to prepare for grid submission. Again, the `"../source.DirectIOART/"` path has to be given, but now with a sequence tag which is composed of, firstly, the package name, secondly, the submission date, and third, a counting index. Every `"art.py grid ..."` command will create a separate job containing all executed tests as single tasks. Actually, what the sequence tag looks like doesn't matter as long as it is unique, since it will be part of the grid job's output name as displayed on your [bigpanda](https://bigpanda.cern.ch/) page, e.g.:
 ```
 user.flfische.atlas.21.2.AnalysisBase.x86_64-slc6-gcc62-opt.2018-07-04T0335.DirectIOART_20180706_2.DirectIOART/
-```
\ No newline at end of file
+```