[vsnet-chat 3586] re R Normae

["." <crawl@zoom.co.uk>]

Hi Sebastian

Sebastian Vasarao wrote:

>This is to ask for some comments on R Normae. the star was mentioned as
one of
>the changing-period Miras. But it wasn't specified if the star is showing
shorter or longer
>periods. I thimk it's the opposite of R Centauri, so its period is
lenghtening. Is this true?

Yes it is, and no, not necessarily.

Coincidentally I've got a part analysis of R Nor festering away in one
corner of my hard disk, and I've just had a quick look at it.

At present the "trend" is for ever longer durations between maximum (this
does not necessarily mean that any one maximum to maximum cycle is
necessarily longer than the previous one).

However, when such things are tested, whether via O-C or some method of
fourier decomposition or other, there is a reference datum against which
the period of time is tested.

It is relative to this that the behaviour of the lightcurve, or parts
thereof, is tested.

So, strictly speaking, the visual measures constituting the lightcurve of R
Normae when tested against a representative period [whether said
representative period is derived via fourier transform or linear regression
of O-C residuals] suggests that an increasing trend in main period
currently exists.

This does not necessarily mean that any evolutionary change is occuring in
the star.  Indeed, even if the change _is_ a physical one it is not
necessarily the case that it is anything more than a stochastic variation
which in the current short term appears as a trend when viewed at "local
temporal resolution", for want of a phraseology.

Analysis of LPVs is problematic.  In my view much work that is done on
these objects nowadays, with arguments over pulsation modes etc [some of
these arguments are nearly older than I am], misses the point.

ACTUAL investigations of what behaviour Miras actually exhibit in terms of
their lightcurve are rarely undertaken. On these rare occasions, exemplary
on this front are only Isles and Saw's work of about 20 years ago, and
continuing work of Szatmary and Kiss et al over the past decade.  Some
effort has also been put into this by Foster et al at AAVSO via wavelet
analysis.  In the UK John Howarth (in tandem with some bloke called
Greaves) has recently made some interesting observartions using fourier
decomposition methods analogous to those pioneered in SPVs by the likes of
Simon and Teays. Koen and Lombard and others of South Africa have made some
interesting points re the use of O-C across the board in variable star
work.  O-C is a means of determining elements for a variable, but a lot of
people turn it upside down and try to use it to show evolutionary change in
period for many variables.  Koen and Lombard have rigorously tested the
statistics of this in many ways, both for LPVs and SPVs.  For people forget
that the "test period" and base epoch that they are using via O-C to look
for evolutionary changes were themselves derived from O-C.

On the other hand, and in my opinion, the professional community is still
primarily stuck in assessing  all work on LPVs only as part of the age old
barney twixt Wood and Feast about LPVpulsation mode, and thus all other
aspects of LPV lightcurve behaviour are just often ignored.

So, given such caveats of a digressionary nature, note that the following
statement is probably true, but definitely "unfashionable" and deemed
suspect in professional circles, if only because it is irrelevant to their
agenda.

R Normae has a 12,000 day or so cyclicity in period and will probably be
declining in period again in ten years time, or more properly be in "a
trend of declining periodicity".  I can't be more exact with the long term
figure, as with only about 20,000 days of data I haven't got two full cycles.

Note also that the full range of periodicities about the mean is quite
small, though I haven't quantified it as yet.

In ALL cases, the double peaked nature of this Mira makes analysis
problematic : in O-C because of the possibility of cycle count ambiguities
where data is sparse; in fourier transformation based methods like AMPSCAN
and Wavelet Analysis because the lightcurve is somewhat divorced from that
of a pure sinusoid.

Anyway, keep watching, cos it'll take about another twenty years of visual
data to see if I'm right or not!!!!  ;^) 

Cheers

John

VSNET Home Page

Return to Daisaku Nogami

---

vsnet-adm@kusastro.kyoto-u.ac.jp