capscale() + anova.cca() for unequal sample sizes

Dear all,

I perform canonical analysis of principal coordinates (vegan::capscale)
followed by permutation test (vegan::anova.cca) to determine the variance
explained by sevaral environmental factors in my experimental design. My
problem is, the data I have contain unequal sample sizes in some of the
factors and I am not confident about if it is okay to use CAP and

rhizo.cap <- capscale(rhizo.dist ~ AMF + Condition(Field + Dev_Stage +
Fertilizer), data = rhizo.env)
anova.cca(rhizo.cap, permutations = 5000)

I do the very same thing for the other three factors too (~ Field +
Condition(AMF + Dev_Stage + Fertilizer) and so on). Factor AMF has two
levels containing the same amount of samples in each but the other three
factors (field, dev_stage and fertilizer) have unequal sizes. Is my
approach correct? As far as I know unequal sample size is not suitable for
PERMANOVA but I don't know about permutation test. If my approach is wrong,
what can I use instead?

Thanks in advance for any help.

Best regards,

Akyol