Hi Marcelo,

Hi Tarquin, thanks for your kind answer.

> These surveys  only
> use tape and clinometer, surveying triangles, starting from two
> coordinates known  stations.

That's ... both cool and insane at the same time.

I believe that a typical speleological topographer thing: being a little bit ingenious and a lot insane...

Pretty sure a Total Station can cope with this magnetic rock situation.
Perhaps someone else on this list can offer advice...

For sure, but total stations are expensive, cumbersome and time consuming. Our idea is to provide a cheap fast way to deal with magnetic anomalous caves..

> Is there a way to specify specific standard deviations for each leg in
> Therion?

As far as I know, SD affects all legs specified after it, until another
sd command changes it. Similar to the "calibrate" command or "units"
command.
sd compass 0.3 degrees
1 2 12.5 315 17
sd compass 0.5 degrees
2 3 7.2 220 33

It is a Survex command (most things inside a Therion centreline block
are a direct copy of their Survex equivalents, and can be passed through
to Survex for processing; group/endgroup is an exception, which is
begin/end in Survex syntax).

I tested it with a loop where every leg is set to sd 0.1, but one single
leg set to 200. Therion gave that single leg most of the error.

Thanks for the suggestion, I will try it for sure. Nevertheless I will prefer to use cartesian data and dX, dY and dZ notations. It is because transversal and longitudinal expected errors are very different in trilateration methodology and I have been able to estimate it in cartesian coordinates.

I am a little worried if it will also work in "leap-frogging" legs, but I will check it, for sure. 

If that doesn't work for you, you can use:
group
sd compass 0.3 degrees
1 2 12.5 315 17
endgroup
group
sd compass 0.5 degrees
2 3 7.2 220 33
endgroup

It's quite verbose, but that is explicitly stated as being possible in
the Therion book.

I will try this option too. Right now, that calculations have been done in an Excell spreadsheet,it is verbose, for sure. But the idea is to include in our TOPGRU software package (https://see.ufop.br/topgru) in the future. So, after coding, it will only lead to larger lines centerline blocks, that can be copied directly into Therion. 

TOPGRU right now is only in portuguese and have a more educational characteristics, but one of our 2021 goals is to release an english version and maybe moving it to GNU compliant.

Hope this helps.

It helped a lot, Tarquin. Many thanks.