Transcript kdtalk_jun2005 - The HST Treasury Program on Eta Carinae

kd 2005 June 8 -- talk at STScI
“... OK, suppose we do an experiment and it
confirms theory. What have we learned?
Nothing! Progress comes from experiments
where theory fails.”
-- mid-20th century physicists’ truism
kd 2005 June 8 -- talk at STScI
An Astrophysical Counter-Paradigm:*
The Hubble Treasury Project for Eta Carinae
* ( Which is the counter-paradigm?
-- Well, mainly I mean
Eta Carinae; but our project is also somewhat unconventional.)
kd 2005 June 8 -- talk at STScI
F
TRUTH-IN-ADVERTISING DISCLAIMER
This talk shouldn’t be considered
“a report on the h Car Treasury Program”,
exactly -- because ...
E
kd 2005 June 8 -- talk at STScI
The topic is honestly too large!
We only have time for a rather breathless
outline of just a fraction of the project.
( More about this later. )
kd 2005 June 8 -- talk at STScI
Carina: A constellation in the far south of the sky;
formerly Argo, Argo Navis, or even Robur Carolinum.
(But that last one is another story – history, not astronomy.)
...The most spectacular spiral
arm in our quadrant of the
Galaxy lies in that direction.
For example, at galactic
longitude 287°, latitude -1°,
we can see ...
kd 2005 June 8 -- talk at STScI
NGC 3372 , a big star-formation region 7500 lightyears away.
kd 2005 June 8 -- talk at STScI
A closer view of NGC 3372 (now North is at the top)
kd 2005 June 8 -- talk at STScI
Eta Carinae and its famous ejecta
Strictly speaking, h Car is the star at
the center. The bipolar structure is the
“Homunculus Nebula”, ejected in a
titanic eruption observed from 1836 to
1860. The ejected mass was 5 Msun
or more, but the central star survived.
After 160 years of expansion at 650 km/s, the Homunculus’
polar diameter is currently about 0.2 pc or 0.65 lightyear.
Here, though, we’re mainly interested in h Car itself.
kd 2005 June 8 -- talk at STScI
This object is uniquely significant
for astrophysics in a variety of
ways. In fact it really amounts
to a surprisingly broad topic, and
the hard part of describing it is
to choose where to start.
For example, we might begin with its impressive
list of superlatives ...
kd 2005 June 8 -- talk at STScI
h Car’s pages in the Guinness book
1. Most luminous star that we’re sure about:
Certainly L > 3 million L sun , and probably
L  5 million L sun .
-- Which implies it’s the most massive, too:
Initial mass  140 to 200 M sun ,
Present-day mass  110 to 150 M sun .
2. Most extreme stellar wind. Mass-loss rate is
around 0.001 M sun per year. That’s 300
to 1000  more than a normal massive star.
kd 2005 June 8 -- talk at STScI
Guinness list, continued
3. Brightest extra-solar object in the sky at mid-IR
wavelengths.
4. Probably the strongest, hottest thermal X-ray
source among known stars. (kT ~ 5 keV)
5. Biggest non-terminal stellar explosion that we
know much about. (5—10 M sun ejected, total
energy > 10 49 ergs)
6. The only naked-eye star that we really don’t
understand.
kd 2005 June 8 -- talk at STScI
ST. JAMES’S GATE, DUBLIN
...These superlatives are potentially
valuable, because often one learns
something by exploring the extremes
of parameter space.
In particular, that can be a good
way to learn which parts of theory
don’t work.
The next page shows a real example.
kd 2005 June 8 -- talk at STScI
The Eddington Limit is fundamental and ubiquitous in astrophysics
-- re. stars, black holes, accretion disks, AGN’s, etc.
A few years ago Nir Shaviv discovered that it doesn’t work the way
that everyone assumed. His best, most concrete example was the
Great Eruption of Eta Carinae.
kd 2005 June 8 -- talk at STScI
A second way to begin: The most massive stars
have been fashionable among some cosmologists.
(And what could be more fashionable than cosmology?)
According to astrophysical folklore dating back about 40 years
– and supported by theory, sort of – the first generation of stars
included a larger fraction of very massive objects, M > 60 M sun ,
than we have today. They helped to re-ionize the universe,
with various consequences.
Two or three years ago, when questioned closely, enthusiasts
typically replied, “Very massive stars are straightforward.
Simple Thomson-scattering opacity, no convection in the outer
layers, nothing to worry about. What could possibly go wrong?”
Alas, most of them didn’t know the following history.
kd 2005 June 8 -- talk at STScI
THE H-R DIAGRAM HAS A SLOPING EMPIRICAL UPPER LIMIT
– NO YELLOW, ORANGE, RED SUPER-SUPERGIANTS. WHY?
kd 2005 June 8 -- talk at STScI
LIKELY EXPLANATION: ERUPTIVE MASS LOSS -- A COMPLICATED
SURFACE INSTABILITY WHICH ARISES NEAR THE EDDINGTON LIMIT.
kd 2005 June 8 -- talk at STScI
In extreme cases, we think, the giant
eruption is a SUPERNOVA IMPOSTOR.
This is the modern term for Zwicky’s
notorious “Type V Supernova”:
-- Peak brightness lower than a real SN,
-- Much longer duration than a SN,
-- Total energy radiated as photons is
roughly the same as in a SN,
-- The star survives! (Most of it, anyway)
kd 2005 June 8 -- talk at STScI
(Supernova Impostors, continued:)
The two classic examples
kd 2005 June 8 -- talk at STScI
Other Supernova Impostors are also known.
A research group interested in SN 1954J is
reportedly planning a TV series on Impostors:
kd 2005 June 8 -- talk at STScI
At any rate, in terms of physics, Supernova Impostors
are more mysterious than real Supernovae.
Unlike SNae, the instability mechanism has not been
modeled and is not even known for certain. We
certainly can’t read about it in textbooks!
kd 2005 June 8 -- talk at STScI
ETA CARINAE APPEARS TO BE THE MOST EXTREME OBSERVABLE
EXAMPLE. INDEED IT INSPIRED THE BASIC IDEA.
kd 2005 June 8 -- talk at STScI
HERE’S THE POINT (or points):
-- No one predicted this phenomenon, which
represented a serious gap in theory.
-- It still does, in fact; even 25 years after the
basic idea arose, the eruption mechanism
is only vaguely understood.
(Turbulence + radiation + rotation:
a Combination From Hell.)
-- Eta Car has done a lot of other things that
also were not predicted by theorists !
Indeed, theory has a downright poor
record for this star.
kd 2005 June 8 -- talk at STScI
That’s one reason why we call Eta Car
a counter-paradigm.
By the philosopher’s criterion (*quote*),
this object is a good experiment,
with implications for several branches
of astrophysics – not just stars.
It’s also a sobering & healthy rebuke to
theoretical complacency.
kd 2005 June 8 -- talk at STScI
A third way to begin (even if it’s almost a digression):
The GRB Connection – Cosmic Gamma-Ray Bursters.
kd 2005 June 8 -- talk at STScI
The connection seems pretty obvious,
given that Eta’s mass is very large.
And this is one star whose rotation axis we know ! *
* ( Probably. )
kd 2005 June 8 -- talk at STScI
... Which brings us to a related point about
rotation and stellar winds – an important recent
observational + theoretical development,
in which HST/STIS observations of h Car
played a unique, unconventional, and
absolutely essential role.
kd 2005 June 8 -- talk at STScI
Logarithmic H-alpha profiles
(see N. Smith’s PhD thesis)
We can observe the spectrum from the
polar direction, via reflection by dust.
It’s different in a very significant way!
kd 2005 June 8 -- talk at STScI
--- Latitude about 40°
Probable interpretation:
The wind is densest at the
poles, even though this
seems counter-intuitive.
(Stan Owocki has a fairly
logical explanation.)
--- Approx. polar view
kd 2005 June 8 -- talk at STScI
The photosphere (which is located in the wind)
is most likely prolate, with a range of temperatures.
kd 2005 June 8 -- talk at STScI
5th motivation: A DIVERSITY OF SUBTOPICS,
embracing multiple branches of astrophysics.
Stellar structure: instabilities, evolution,
nature of the Eddington limit
Stellar wind physics: extreme parameters.
Practically the only stellar wind we can
view from various latitudes (reflection)
Exotic nebular excitation processes (unique)
Dust formation in Nitrogen-rich circumstances (dust grains
in the Homunculus are known to be highly unusual)
Gas-dynamics in the ejecta (odd structures)
Etcetera -- this ain’t a complete list.
kd 2005 June 8 -- talk at STScI
6th motivation (not entirely scientific) :
THIS OBJECT IS EXTRAORDINARILY
WELL - MATCHED TO THE HST.
For instance, we need the highest spatial resolution,
but that resolution is attainable -- because Eta is bright.
We can “push” the instruments more than most other
users try to do.
...Indeed, if we didn’t know better, one would almost
imagine that the STIS was designed specifically to observe
Eta Carinae. It wasn’t, but it turned out almost that way.
kd 2005 June 8 -- talk at STScI
Ground-based spatial resolution -- even 0.5 -is simply not adequate for this object.
kd 2005 June 8 -- talk at STScI
WITH HST, A DIVERSITY OF TYPES OF SPECTRA,
EACH UNIQUE AMONG KNOWN OBJECTS:
kd 2005 June 8 -- talk at STScI
kd 2005 June 8 -- talk at STScI
Quite honestly, this has been one of the most successful of
all HST targets, using most of the available instruments.
Since 1991 it has consistently held the record for highest
angular resolution of any complex spectroscopy.
· FOS 1991 (pre-COSTAR): First spectrum ever obtained of
the star itself; enormous mass-loss rate. Discovered the
basic nature of the Weigelt blobs.
· WFC-PC 1991-1993: Revealed that the Homunculus Nebula
is bipolar, with a mysterious equatorial skirt.
· WFPC2(PC) 1994-1996: One of the most familiar of all HST
color images, quickly became a popular-astronomy icon.
· FOS 1996-1997: One of the very few objects ever observed
with the tiny 0.1 FOS aperture.
( continued on next page )
kd 2005 June 8 -- talk at STScI
· FOC 1991-94, then GHRS 1996: Showed that the Weigelt blobs
were ejected long after the Great Eruption, and are moving at
slow speeds that defy intuition.
· FOS 1996-97: Discovered several different and unfamiliar
classes of exotic emission spectra throughout the Homunculus,
each of them unique among known objects.
· GHRS 1996: Pseudo-laser lines near 2507 Å – again, unique
in astrophysics.
· STIS 1998-2000: Discovered the UV “iron curtain” during a
spectroscopic event, disproved existing orbital velocity models,
revealed many other details of the 1998 event.
· STIS 1998-99 and later 2002-04: The central star has
recently brightened at an amazing rate.
( continued on next page )
kd 2005 June 8 -- talk at STScI
· WFPC2 1994-2000: Proper motion studies of the Homunculus.
Dominant ejection date was 1843, during the Great Eruption.
· STIS 2000: Spectra seen from various latitudes via reflection
by dust. Result: The stellar wind is essentially polar -- a fairly
revolutionary result for wind theory in general.
· STIS 1998-2001: [ Sr II ] and other weird emission lines.
· STIS 2000: Three-dimensional shape and orientation of the
Homunculus. Found that the equatorial debris was ejected
in two or more separate events.
· STIS 1998-2000: Discovery of a “Little Homunculus” hidden
inside the two large lobes. Apparently ejected 50 years
after the Giant Eruption.
· STIS and ACS 2002 et seq.: Numerous Treasury Program
results, some of them outlined in this talk.
kd 2005 June 8 -- talk at STScI
Digression: Pop culture, HST, and explosions in space
1977: In the original version of Star Wars, the Death
Star blew up spherically.
Late 1990’s: Mysteriously, the same explosion had
now become bipolar – including an equatorial skirt.
Now why do you suppose that happened?
kd 2005 June 8 -- talk at STScI
20 YEARS AGO: ANOTHER MYSTERY,
OR MAYBE A CLUE TO ETA’S STRUCTURE?
“SPECTROSCOPIC EVENTS”
High-excitation emission lines temporarily
disappeared at various times –
· 1948 (Gaviola data)
· 1964 (Rodgers & Searle data)
· 1981 (classic Zanella et al. paper)
· 1986 (noticed by various observers)
kd 2005 June 8 -- talk at STScI
THE CLEAREST EARLY DESCRIPTION OF A
SPECTROSCOPIC EVENT IN ETA ...
kd 2005 June 8 -- talk at STScI
THE CLEAREST EARLY DESCRIPTION OF A
SPECTROSCOPIC EVENT IN ETA ...
kd 2005 June 8 -- talk at STScI
...And later Viotti
noticed that the
strange 2507 Å
emission had
also disappeared
for a while in
1981 (IUE data)
kd 2005 June 8 -- talk at STScI
THE POINT:
IF WE CAN FIGURE OUT WHAT CAUSES A
SPECTROSCOPIC EVENT, IT MAY TELL US
SOMETHING ABOUT ETA CAR’S STRUCTURE
AND INSTABILITIES.
(In fact, Zanella et al. expressed a promising idea
-- more about that later.)
BUT IT’S HARD TO DIAGNOSE SPORADIC,
POORLY OBSERVED OCCASIONS.
kd 2005 June 8 -- talk at STScI
Anyway: High-excitation, high ionization emission
lines temporarily disappeared in
1948, 1964, 1981, 1986.
1992: Another event -- observed (so far as I know)
only by Augusto Damineli in Brazil.
1996: Based on the 1992 event, Damineli realized
that a recurrence period of 5.5 years would fit
all the known instances.*
1997-98: The next event occurred at the predicted
time.
* ( Today we know the period is about 5.54 yr. )
kd 2005 June 8 -- talk at STScI
DOUBLE SIGNIFICANCE OF THE
5.5-YEAR SPECTROSCOPIC CYCLE
1. The periodicity itself may be a clue.
2. --And, given the 5.5-year period, for the first
time we knew when such an event would occur!
kd 2005 June 8 -- talk at STScI
A binary companion is the most obvious
“explanation” for the periodicity.
Its orbit must be quite eccentric, though.
kd 2005 June 8 -- talk at STScI
Results of the 1997-98 event
1. 2-10 keV X-rays increased tremulously, then
crashed.
2. A few orbits with STIS showed various changes in
the spectrum of the star itself – e.g., the UV
“iron curtain”. Strong low-excitation absorption
-- Fe II, Ca II, etc. – and temporary P Cyg
absorption in Ha ; etc. Practically no temporal
coverage, though, and obs were late in the event.
kd 2005 June 8 -- talk at STScI
Thus, by 2001, the topic had reached a sort of crisis.
1. We knew about the 5.5-year period.
2. We had obtained some minimal spectroscopy
in 1998 using the HST’s STIS instrument.
3. The next predicted “event” would occur in
mid-2003.
4. HST data were absolutely necessary.
5. We also recognized that HST spectroscopy
will not be possible for the next event after
that, in 2008-2009. Nor later !
kd 2005 June 8 -- talk at STScI
Idea (2001): A Treasury Project!
Criteria for such a program:
-- Data of broad significance.
-- Extensive enough to justify and construct
a long-term data archive.
-- Similar observations will not be possible
after the HST ceases to operate.
This topic matched the criteria perfectly.
kd 2005 June 8 -- talk at STScI
(Parenthetical: an unprecedented situation)
No instrument with HST/STIS’s capabilities
will be available again for a long time –
conceivably a long, long time.
Therefore: Many observations, possible
from 1997 to early 2004, will now be
impossible for the forseeable future.
But the case of Eta Car is even worse! By
the time that some capable new instrument
becomes available, perhaps 10 to 30 years
from now, this star will have changed.
kd 2005 June 8 -- talk at STScI
Hence our project. Main goals & components:
1. A huge set of observations during 2002—2004 with
STIS, the Space Telescope Imaging Spectrograph.
This is, quite honestly, close to the most intensive
spectroscopy that HST will ever do.
2. We are developing better software for reducing
and analyzing STIS data. (Badly needed for
other users as well.)
3. Supplemental images of Eta Car using two other
HST instruments, the WFPC2 and the ACS.
kd 2005 June 8 -- talk at STScI
(Project goals & components, continued):
4. Parallel observations using the European Southern
Observatory’s “VLT/UVES” instrument in Chile.
5. The whole caboodle will go into a special, highly
non-routine Data Archive that astronomers
will still be able to use 30 or 40 years from now.
kd 2005 June 8 -- talk at STScI
Project participants (CoI’s) – listed in more or less
inverted order
CoI’s at other institutions who
in using the data:
Manuel Bautista (Venezuela)
Augusto Damineli (Brazil)
John Hillier (U. Pittsburgh)
are mainly interested
Mike Corcoran (NASA/GSFC)
Fred Hamann (U. Florida)
Nolan Walborn (STScI)
CoI’s in Germany who have obtained the ESO/VLT data:
Otmar Stahl
Kerstin Weis
Preparation of the big observing plan, and help with
material from STIS IDT: Ted Gull (NASA/GSFC)
kd 2005 June 8 -- talk at STScI
Project participants, continued
Initial data reduction and modified software:
Kazunori Ishibashi (MIT)
System maintenance & linux sanity:
J. T. Olds (U MN)
Data analyses & some dandy discoveries:
Michael Koppelman (Clockwork Mpls. & U MN)
Data archive & database planning, etc.:
Roberta M. Humphreys (U MN)
kd 2005 June 8 -- talk at STScI
Project participants, continued
Linux building, system maintenance, software,
data archive, etc. etc.:
Matt Gray
(U MN, later Clockwork)
Data analysis, techniques development, some
discoveries and most of the hard work:
John C. Martin
(U MN).
kd 2005 June 8 -- talk at STScI
Our web site:
http://etacar.umn.edu
The initial data archive is located there and
is beginning to be publicly available.
(STScI will have a modified version
of the archive.)
kd 2005 June 8 -- talk at STScI
SPECTROGRAPH
SLIT
kd 2005 June 8 -- talk at STScI
SPECTROGRAPH
SLIT
OFFSET FROM STAR
kd 2005 June 8 -- talk at STScI
Concerning the amount of data ...
Each complete spectrum covered the wavelength range from
170 nm in the UV to about 1 mm in the near-IR. To do this,
we needed 30 grating tilts. Eta Car is almost unique in getting
complete wavelength coverage with STIS.
Since this object is so bright, we typically obtained more than
a dozen independent integrations per orbit. This is highly
unusual, and is the chief reason why our data volume is
among the largest obtained in any HST program.
Altogether we have several hundreds of separate wavelength
observations obtained with an 800 x 800 CCD detector.
In addition, far-UV data were obtained with a different
type of detector.
( explain about special processing )
kd 2005 June 8 -- talk at STScI
Ca II and hydrogen lines near 390 nm
kd 2005 June 8 -- talk at STScI
COMPLEX STRUCTURE IN UV DATA NEAR 290 nm
kd 2005 June 8 -- talk at STScI
kd 2005 June 8 -- talk at STScI
Vicinity of H b, c. 490 nm
kd 2005 June 8 -- talk at STScI
Behavior of the H a emission line profile
during the 1998—2003 cycle. (our STIS data)
kd 2005 June 8 -- talk at STScI
Only a small fraction of the project can be included
in this talk. Some of the topics we must skip:
· Improvements in data processing for STIS. Some of
our techniques may be helpful for HST users in general.
· Data archive details. The STIS/CCD data are publicly
available now in convenient form, other data will
follow. (Standard STScI archiving is not very suitable
for these unusually intensive observations.)
· ACS imaging data obtained for the Treasury Program.
In most respects these are the highest-quality images
of the Homunculus so far.
· STIS / MAMA / echelle observations in the UV.
-- continued on next page --
kd 2005 June 8 -- talk at STScI
-- more topics not covered here --
· Ground-based VLT / UVES spectroscopy which
supplements the HST data. (ESO; O. Stahl & K. Weis)
· Observations of the diffuse ejecta. Today we’ll restrict
ourselves to the star itself (or rather, its wind).
· Routine “as expected” results on the star, essential
for testing models but not particularly exciting.
· Possible single-star models for the event. Contrary
to what some authors have said, the companion star
has not, repeat not been confirmed or proven. Every
existing observation might in principle be explained with
just one star. Since a model of that type seems less
likely, however, here we’ll save time by assuming that
h Car is a binary system.
kd 2005 June 8 -- talk at STScI
The 2003.5 event did
occur on schedule.
Here are some of the
observables used for
timing it.
kd 2005 June 8 -- talk at STScI
First point: Eclipse or
mass ejection?
This question strongly
affects the basic
importance of the
5.5–year cycle.
( explain )
kd 2005 June 8 -- talk at STScI
First point: Eclipse or
mass ejection?
This question strongly
affects the basic
importance of the
5.5–year cycle.
( explain )
kd 2005 June 8 -- talk at STScI
Frankly, the eclipse idea never was very appealing. It was
motivated by X-ray analogies (0.01% of the luminosity),
its predictions even for the X-rays generally failed, and
it was unclear for other parts of the spectrum.
kd 2005 June 8 -- talk at STScI
(Parenthetical complaint): This mini-debate would make a
good case study of scientific procedure. Normal rules of
evidence, predictions, etc., have been perceptibly “bent”.
Just about any outcome is customarily hailed as a great
success of the eclipse theory. Sorry if this sounds grumpy,
but it’s been true!
kd 2005 June 8 -- talk at STScI
... And the alternative is older and far more interesting.
(Please don’t take the word “shell” too literally, though.)
kd 2005 June 8 -- talk at STScI
(Imagine what might happen near periastron.)
kd 2005 June 8 -- talk at STScI
(Imagine what
might happen
near periastron.)
kd 2005 June 8 -- talk at STScI
... Some emission-line velocities (hydrogen, helium)
appear to make trouble for the eclipse model.
But new and unexpected evidence has appeared in
the He II 4687 emission. This feature was first
reported by Damineli last year, but its true brightness,
other details, and analysis became evident only with
our STIS data.
kd 2005 June 8 -- talk at STScI
He II 4687 emission
The peak amount was
terrific compared to
expectations, and the
post-maximum decline
was catastrophically
fast.
kd 2005 June 8 -- talk at STScI
Desperate ansatz to explain the He II emission:
Soft X-rays from the shocked gas (region 3)
create He++ in region 4 ...
This sketch is VERY idealized (explain).
kd 2005 June 8 -- talk at STScI
... Which presents three technical difficulties:
1. Energy budget requires either an extra mass
ejection or else a special trick (see next
slide);
kd 2005 June 8 -- talk at STScI
Energy-level diagram for He II
kd 2005 June 8 -- talk at STScI
... Which presents three technical difficulties:
1. Energy budget requires either an extra mass
ejection or else a special trick;
2. But the special trick requires high gas densities
– so we need a mass ejection anyway ...
kd 2005 June 8 -- talk at STScI
... Which presents three technical difficulties:
1. Energy budget requires either an extra mass
ejection or else a special trick;
2. But the special trick requires high gas densities
– so we need a mass ejection anyway ...
3. And the timing of the peak! (next slide)
kd 2005 June 8 -- talk at STScI
The He II emission
rose rapidly, after
the hard X-rays had
already declined
substantially. This
doesn’t seem very
eclipse-like.
kd 2005 June 8 -- talk at STScI
Moreover, several
other bits of
evidence strongly
suggest that the
orbit is oriented
roughly as shown
here – so periastron
is near quadrature.
That’s where the
event happens!
kd 2005 June 8 -- talk at STScI
Some other unexpected
developments may be
related to each other.
For instance, the star’s
apparent brightness
increased at an almost
alarming rate in 2003.
This is a figure from
Martin & Koppelman’s
paper.
kd 2005 June 8 -- talk at STScI
We suspect that something basic is changing
right now, i.e. in this decade! (explain)
kd 2005 June 8 -- talk at STScI
kd 2005 June 8 -- talk at STScI
kd 2005 June 8 -- talk at STScI
kd 2005 June 8 -- talk at STScI
kd 2005 June 8 -- talk at STScI
The point: Thermal and rotational recovery
timescales following an eruption may possibly
produce cycles of various lengths.
(Terrestrial precedent: geysers.)
Moreover, we shouldn’t be very surprised if
the star’s surface conditions suddenly change.
(Explain: sensitive to parameters.)
kd 2005 June 8 -- talk at STScI
If developments continue at this rapid pace –
Within 15 years, the star will appear brighter than
its Homunculus Nebula and will be 4th magnitude.
To the unaided eye it will again seem as it did to
Halley three centuries ago. And the wind may
become appreciably weaker.
kd 2005 June 8 -- talk at STScI
If developments continue at this rapid pace –
Within 15 years, the star will appear brighter than
its Homunculus Nebula and will be 4th magnitude.
To the unaided eye it will again seem as it did to
Halley three centuries ago. And the wind may
become appreciably weaker.
In saying this, however, I’m ignoring the
time-honored, repeatedly vindicated,
STANDARD ETA CARINAE CARTOON:
kd 2005 June 8 -- talk at STScI
kd 2005 June 8 -- talk at STScI
... What origin can we ascribe to these sudden
flashes and relapses? What conclusions are we to
draw as to the comfort or habitability of a system
depending for its supply of light and heat on so
uncertain a source?
-- J.F.W. Herschel, 1847
kd 2005 June 8 -- talk at STScI
----
kd 2005 June 8 -- talk at STScI
-----