These are all restrictions of one rank-3 temperament, that is a bit off-topic in this subforum (and some may say that the phi temperaments themselves, having half octave periods, are off topic, but I think they are closer in spirit to this sub forum than to the "Notations for other tunings" one and if one of the mods wants he can probably move it there, as well as my discussion about the rank-3 pental temperament that also has an octave-fraction period, or just change this subforum's name to "Regulat Temperament notations"?) but I think a good starting point (and an objective in and of itself) would be to a phi chain in this tuning, with major thirds in either side of it. This tuning has 3 generators to a perfect fifth, and with a half octave and 3 chains this means we'll need floor(3 * 2 * 3 / 2) = floor(18 / 2) = 9 symbols (without the natural). This rank-3 tuning is also Keenanismic, which already suggests using , and which are all useful at this point. comes in a bundle with , which, when subtracting gives , which is useful as well, and we've already have everything we need for the 11-limit intervals. Another thing I found useful is to use (which usually has the 7-limit meaning of 5120/5103) as , the 13/11k or 352/351, which can also be combined with to get for the 13/7 S-diesis, or 1701/1664 (which is in Olympian).

However, here I got a bit stuck. First of all, in this tuning the Pythagorean comma is both negative and splits in half, which means we need a sagittal for the negative half pythagorean comma, or [9.5, -6⟩, which, if the golden ratio is tuned pure, is 2

^{12.5}/ϕ

^{18}, but there is no secondary comma that maps to it. However, there are two intervals with n2d3p9 of less than 307 which map to it: 208/207 which is an alternative meaning for and 1024/1023 which is an alternative meaning of , so both and can technically be used for this interval. The next one to consider is 5ϕ/8, which has 2 sagittals mapped to it: (144/143) which is technically fine and (65/64) which is rejected because is already taken. There is also that can take its place, but we already have and and they sum to something else so that is rejected too, at least for now. is also not ideal because actually this comma is greater than the alteration notated by , and is actually the sum of it and the negative half pythagorean comma. However, these can't be notated by the same flag, since the only Prometheans that use <> are itself, which is already used, which is rejected because the value is not -+ and which is rejected because it comes in a bundle with which doesn't get the value of 2(-). The third hole is for 5√2/ϕ

^{4}, which has many secondaries that are rejected for similar reasons, except for . However, these three cannot be used together, since the last one gives a value to the flag <> which is equal to neither the one of the minus half pythagorean comma nor the one gotten from it and the 5ϕ/8 if we decide they get and respectuvely. Also, if the 5s are tuned pure, this has the value of almost 54 cents, which is more than half an apotome, so notating it with a single shaft sagittal and its complement with a double shaft seems wrong to me.

I have also run the same search for 5ϕ/8 and got 640/637, 78/77, 120/119, 135/133, 190/189, 85/84, 144/143, 324/319, 2025/1984, 145/144, 65/64, 189/187, 1701/1672, 133/132, 357/352, 105/104, 945/928 and 217/216, and for 5√2/ϕ

^{4}and got 12160/11907, 1360/1323, 1024/1001, 580/567, 65/63, 192/187, 216/209, 304/297, 34/33, 40/39, 30/29, 248/243, 91/88, 35/34, 315/304, 665/648, 595/576, 147/143, 3185/3072, and I guess all of which can be used for these intervals if deemed necessary.

This is a WIP and I just tried to summarize all I have found in this post, but it's a bit late here now so I'll hopefully go back to it tomorrow. In the meantime I'll be very glad to receive help if anyone else is also interested, by searching directions I rejected (or didn't even considered), continuing from here or notating some of the rank-2 tunings with the 6 sagittals I've already established (+tentative placeholders for the 3 others?), and maybe some that are mapped to a chain of fifths a third above and below what I considered, like .