Hello Anthi. With the information you provide I can only guess here, and I would think that you are probably not reading properly your data. Can you make sure that your matrices are read well on your scripts? I don't believe your problem is related to a bug...

Best regards.

the factors are read as factors and the numeirc variables are also read well.

See also the example in adonis

data(dune)

data(dune.env)

str(dune.env)

adonis(dune ~ Management*A1, data=dune.env, permutations=99)

although A1 is a numeric variable the result provides 1 df.

Thanks in advance

I see. The permanova test tries to explain some data given some grouping, and in the example you show (dune dataset), the management variable is grouped while the A1 is simply a numeric variable, thus it is unclear to me at least how that variable is treated by this test (the Df for the other grouped variables is correctly shown as the number of groups minus one, for example). Maybe you should think about another tests like Generalized Linear Models (GLM) for your data?

Please take a look at this link as they discuss something related to your concern:

https://stats.stackexchange.com/questions/312302/adonis-in-vegan-order-of-variables-non-nested-with-one-degree-of-freedom-for?rq=1

Best regards.