Dear Florent, Ole and friends,

Florent, a~n~na is an adjective, so it follows the rule of
adjectives, i.e. it is declined in case, gender and number of the
succeeding word.

In this case, the succeeding word is

vaapiya.m - locative, feminine, singular: in/on/at the tank.

Therefore, a~n~nissa.m should also be locative, feminine, singular:
in/on/at the other.

Hence, a~n~nissa.m vaapiya.m: in/on/at the other tank.

Then, the next question is why a~n~nissa.m? I believe it is a typo,
it should be a~n~nassa.m, unless (as you suggested) a~n~na can also
be declined like ima. Otherwise, it should be corrected as

a~n~nassa.m vaapiya.m: in/on/at the other tank


metta,
Yong Peng.

--- In Pali@yahoogroups.com, flrobert2000 wrote:

> The reading is ok. vaapii (cf. Sanskrit vaapii) is feminine, and so
is the loc. aññissa.m of añña: other, different.

according to Buddhatdatta añña is declined as ya (p49) and the
closest form I found in his declension table (and also in Pali
Lookup) was aññesa.m and not aññisa.m. Hence my choice of a gen.
plural which I rendered by "in the tank belonging to others". By
looking again at the table I found the form aññassa.m/aññaaya.m for
this pronoum in the feminine locative singular, but not aññissa.m.
Always according to Buddhadatta, aññissa.m would be correct if it was
similarly declined as the demonstrative pronoum ima.