Not sure if this in in the correct forum, but I'll give it a try...

Old observations are sometimes hard to read. In the observations and comparison stars in attachment, I can find a kind of value where normally I would expect to find the brightness. Is there a way to calculate the visual magnitude from these values? Who can help solve this?

I found the observations and comp. stars here: http://cdsads.u-strasbg.fr/abs/1930AN....239..303L and http://cdsads.u-strasbg.fr/abs/1930AN....240..311L

I don't read German but it looks like these are step observations. It's too bad that we lost Kevin Paxon this year because he was the expert on archival observations. However, he did publish a paper in the JAAVSO about recovering and digitizing these old observations. Check JAAVSO Vol. 40, Page 1,

Digital Archiving: Where the Past Lives Again.Good luck.

...Tim (HTY)

thanks for the tip, nice reading!

I think I don't have steps, but "grades": "In very rare cases, magnitude estimates may exist as “decimal step” or “grade” data". This could be a clue: "When working with grade data, magnitudes must then be derived by regression of magnitude versus grade (tables of magnitudes and grades of the comparison stars are usually given within the paper).". Hmmm... now I'll have to find a formula that makes sense!

hmmm... if we take this observation: "t,u 16.2"

stars: t - variable - u

stages: 23.5 - 16.2 - 14.7

mag.: 8.9 - x - 9.2

How do I calculate "x"? And what to do with "t,u,v 13.0" or just "x 3.4"

We'll get there eventually... :-) Tips are welcome!!

[quote=BBI]hmmm... if we take this observation: "t,u 16.2"

stars: t - variable - u

stages: 23.5 - 16.2 - 14.7

mag.: 8.9 - x - 9.2

How do I calculate "x"? And what to do with "t,u,v 13.0" or just "x 3.4"

We'll get there eventually... :-) Tips are welcome!![/quote]

It resembles processing of visual estimates, isn't it?

Maybe it could be computed like this (basically using just linear regression using 2 points, one can try it easily on paper):

t

_{mag}+ (t_{mag}-u_{mag})*(t_{stage}-var_{stage})/(t_{stage}-u_{stage}) and therefore 8.9+0.3*((23.5-16.2)/(23.5-14.7)) = 9.15 magBest wishes,

Tõnis

Hello,

I think that I found a book where that method is described. I just remembered that my supervisor gave me 20 years ago a book "Peremennye zvezdy", translated (and a bit updated by GCVS-team) from:

C. Hoffmeister, G. Richter, W. Wenzel. 1990, "Veranderliche Sterne", Sternwarte Sonneberg

or in English: http://www.amazon.co.uk/Variable-Stars-Cuno-Hoffmeister/dp/0387134034/ref=asap_bc?ie=UTF8

I added a picture of a relevant page in that book. Don't be afraid that it's in Russian language, graphs, equations and example table are pretty self-describing. Though, there is more text, about 1 full page. That text explaines the concept etc. My Russian language is quite poor so unfortunately I can not translate it for you that well. Still, there may be a chance that English version of this book is available from AAVSO libary or in other public ones. Of course, I can take pictures of those other pages, too. Maybe you can find a person who can translate it to you.

Best wishes,

Tõnis

The main problem with these old (step) observations is they can be rather noisy. A step sequence is given, but when I plot a regression curve, by comparing the same sequence stars against modern APASS V mags, I find the following: t, w and x fall on a near perfect straight line. The others (u,v, y and z) fall well above or below the line. It becomes a matter of interpretation as to where the line of best fit lies. Matters are not helped by having three rather red stars in the mix - u,v and z. I've read off my own curve as to what are the best average values for each step value.

Step (V) : 0.0 (10.88); 1.0 (10.69); 2.0(10.50); 3.0(10.31); 4.0(10.19); 5.0(10.10);6.0(10.02); 7.0(9.94); 8.0(9.86 9.0(9.77); 10.0(9.69); 11.0(9.60); 12.0(9.52); 13.0(9.44); 14.0(9.36);15.0(9.28); 16.0(9.20);17.0 (9.11);18.0(9.03 19.0(8.95);20.0(8.86);21.0(8.78);22.0(8.70;23.0(8.62); 24.0(8.53).

It may be simpler to use a conversion table such as this.. A formula may be more precise, but the actual observations may be no less noisy than the sequence anyway.

Mati

.

Thank's for the information. It's complicated but it is fun!

Somebody pointet me out to an important part in the first paper (I should do something about my knowledge of German...):

So Lause corrects the value for comp. star w to mag. 9.75 and uses 0.095 steps. The other comparison stars can now be calculated:

t = 9.75-((23.7-8.6)*0.095) = 8.32

z = 9.75-((0.0-8.6)*0.095) = 10.57

and observations like this one:

242590.6 t,u 16.2 --> 9.75-((16.2-8.6)*0.095) = 9.03

so I could use: Wmag. - ((OBSstep - Wstep) * step) = OBSmag. to do the calculations.

(Just a new side question: It would be nice to see if Lause has the correct brightness values for the comparison stars. Is it possible to find these? I found the BD cataloque, but I'm not sure how to query the position for example comp. star t -4°1032 using http://vizier.u-strasbg.fr/viz-bin/VizieR-4 )

General form for BD catalogue should be:

BD -04 1032

BD +37 1234

etc.

You can query from Simbad as well:

http://simbad.u-strasbg.fr/simbad/sim-basic?Ident=bd+-04+1032&submit=SIMBAD+search

Tõnis

Thx Tõnis!

So the important comparison star "w" has following magnitudes:

in BD: 9.2 (given in the comp. star list in the first post)

value from Lause: 9.75

V 9.83 [0.03] (http://simbad.u-strasbg.fr/simbad/sim-basic?Ident=bd+-04+1037&submit=SIM...)

So Lauses value is close to the V value in Simbad.

Would it be better to use 9.83 for "w" in the calculations, or is this just mathematical (and historical) nonsense?

Hi Bruno,

I have no idea. I would use those magnitudes that are given in the paper. Maybe adding appropriate uncertainties. Those measurements were done on photographic plates (being digitized currently) and B or V you'll find from e.g. Simbad is not directly comparable with visual estimates from plates (typically blue sensitive before WWII?).

The story would be different when plates would be accessible in digital form - it is possible to create calibration curves for those plates, using e.g. APASS and other recent surveys. It is possible to get ~0.05 mag or sometimes just a bit smaller uncertainty in such case. Of course, one still has to assume constant brightness of comparison stars during the last century ;-)

Besides Harvard plates, see e.g. https://www.plate-archive.org/applause/

Best wishes,

Tonis