Bug: Prism 5 can get confused by blank cells at the bottom of a grouped table

This is a bug in Prism 5 windows and mac, fixed in Prism 6. 

Two-way ANOVA requires that every combination of row/column has at least one value, and Prism attempts to enforce this. But it gets confused when the empty cell is in the bottom of the table in the first data set. In the figure below, the first data table confuses Prism. Since there are only two rows of data in column A, it only looks for two rows of data in column B. Prism 5 reports the two-way ANOVA results, and ignores the data highlighted in yellow. In contrast, Prism 5 correctly deals with the second data table below, and refuses to report any two-way ANOVA results "because you have entered no data for some row-column combinations".

Notes:

• This problem is for ordinary two-way ANOVA. Repeated measures ANOVA has different criteria for handling missing cells, and these work fine.
• It is fixed in Prism 6. 
• Unequal n is fine for ordinary two-way ANOVA. What is not allowed are row-column combinations with no data at all.

Prism file