Mark,

When you run two-way ANOVA with fixed effect, the sample size for each factor level combination has to be more than one. For example, you need at least one repeat for observation of factor A at level 1 and factor B at level 1, and same for other factor level combinations. The reason for this is because, without repeat, there is no way to obtain the variance due to sampling errors and thus there is no way to test the significance of factor effect.

From the 3dANOVA2 script you provided, I can see why you cann't run variance analysis in this case because each factor level combination only bears one sample dataset.

If you want further help, could you please provide more information about your experiment design? For example, what are the factors and their levels? What kind of statistical tests would you want to run?

Gang