Inconsistency of instance number between initial data file and "visualize classifier errors" box

[Albert Kireev]

During lesson 1.6 (Visualizing your data) I repeated J48 classification procedure.  Instance: 92 from the box of visualize classifier error has class   Iris-virginica (sepallength 4.9, sepalwidth 2.5, petallength 4.5, petalwidth 1.7).  Instance: 92  from the initial file visualised by preprocess edit viewer has class Iris-versicolor (sepallength 6.1, sepalwidth 3.0, petallength 4.6, petalwidth 1.4).

The same happen with other wrongly classified instances.  Is it common inconsistency between instances numbers which a little bit confusing?