Dealing with missing cells in ANOVA

I'm working on a project where I'm comparing sets of data from different measurement devices. I'm trying to do run several ANOVA tests in SPSS with the Welch test and Games-Howell post-hoc but am having a problem with my empty cells (they show up as " . "). The empty cells are missing pieces of data that were lost due to system errors or range limits. I'm fairly certain that SPSS is considering these cells 0.00 because the mean differences are so far off.

I tried using a small section containing a significantly large piece of missing data and tried running the test with "Exclude cases analysis by analysis" selected, then with "Exclude cases listwise" selected. The results do not change. I calculated the mean differences of this sample in Excel and it's nowhere close to what SPSS says.
Does anybody know how to deal with this problem? Is there something wrong with the way I'm inputting my data?