BURSTEIN : Many of your E galaxy luminosity profiles show deviations
from an r1/4 law at faint surface brightnesses. Can
you prove
that such deviations cannot be due to sky subtraction? Sky
subtraction is a problem we have to deal with at the 1-2% level in
the data on E galaxies being taken for our own survey. It is a
straightforward test to make.
SCHOMBERT : Of course I cannot prove that the deviations at
faint surface photometry levels are indeed real. The behavior of elliptical
profiles at the 27 to 28 mag arcsec-2 level is in the
realm of 0.5% of sky brightness and are very sensitive to proper
sky subtraction as you noted. However, there is a great deal of overlap
in surface photometry in the literature; i.e. same galaxy from different
telescopes and observers. A comparison shows that there is very good
agreement down to 28 mag arcsec-2. In other words,
faint surface photometry is repeatable. I believe everyone in this
field has be very careful and very honest about their techniques.
The burden of proof here is for someone to show that there is a legitimate
reason to believe that everyone has been systematically wrong in their
sky subtraction.
SURMA : At Heidelberg Observatory we have performed a study of deep
surface photometry of E's, where we also looked at the galaxies in
Kormendy's `tidal class' galaxy sample. What we have found is that
the occurrence of so-called `tidal halos' is strongly coupled
with the problem of accurate sky subtraction. M. Cappaccioli has been
emphasizing the importance of sky subtraction in order to reach the
faint outer regions of E's, SO's and also cD's. It is evident that
you need a sky subtraction accurate to << 1% if you want to reach
brightness levels of 24 mag arcsec-2, and this is the
region you need to reach in order to see cD envelopes! I wonder how
much of the discussion about cD's is influenced by these problems.
SCH0MBERT : First, I strongly disagree with your statement that better
than 1% sky is need to obtain the 24 arcsec-2 photometry
level. A sky of 1% (in the Johnson V band) corresponds to 26 not
24 mag arcsec-2. The exact shape of the outer halos
of ellipticals is certainly open to debate and new work; however,
CCD's have achieved 0.1% flattening, on average, for many years now.
And many observers, with different telescopes and detectors, using
different techniques and software, have time and time again proven
the repeatability of faint surface photometry. I cannot accept the
claim that everyone has performed their sky subtraction incorrectly
for 20 years. cD envelopes have been known for 15 years and cannot
be the result of sky errors.
KENT : D and gE galaxies have identically shaped profiles
(i.e. r1/4).
Hence, they can only differ in the value of µ at
re. Is
your "diffuse" characteristic of D galaxies another
way of saying that they have lower µe?
SCHOMBERT : No, D galaxies are noted by their shallower
r1/4
fits. This places them larger and brighter in the
µe-re
diagram (see Figure 6). I prefer
to interpret this as enlarged re
as predicted by N-body simulations.
BUTA : What are the three-dimensional shapes of cD's, particularly
the envelopes?
SCHOMBERT : Alan Porter's thesis (Caltech 1988) addresses this problem
more fully. His answer is that cD isophotes become more flattened
with radius, presumably shifting from the galaxy potential to the
cluster potential. However, it is an excellent future project, with
the advent of reimaging cameras, to study the behavior of cD envelopes
with respect to twisting and central shifts.
D JORGOVSKI : Two more bits of information come from Porter's thesis.
First, BCM's show systematic ellipticity gradients, in the sense of
becoming flatter outwards, whereas the normal ellipticals show no
such trend. Second, the envelopes tend to be well aligned with the
parent clusters (the Binggeli effect). To me, this also suggests that
the origin of envelopes is at the epoch of cluster formation. The
BCM galaxy and the cD envelope may be independent entities, sharing
the same special location (i.e. the bottom of the cluster potential
well).
KING : I didn't understand your argument that the extended envelopes
must be primordial. Can you distinguish against the case of a merger
of galaxies that already consisted only of old stars? In answering,
please note that the merging process favors massive galaxies, which
have colors like those of cD envelopes. Also, w'th regard to
Djorgovski's
arguments about the alignment of cD envelopes with the major axis
of their cluster, this could be simply due to the greater velocities
of galaxies in this direction.
SCHOMBERT : If tidal stripping is the origin of cD envelopes, then
a majority of their luminosity must be taken from low luminosity members
of the cluster. The mass-metallicity relation predicts that this material
will be bluer than the stellar population of the cD parent body. Although
only a handful of cD envelopes have been studied by myself (in Johnson
B and V plus a Stromgren narrow band system) all show very flat color
profiles. The merging of bright ellipticals can maintain a red cD
interior, but the outer envelope is going to be influenced by lower
luminosity E's and, of course, any disk galaxies in the cluster. I
agree that the elongation of the cD envelope could be due to the orbits
of galaxies being stripped if that is the mechanism of formation.
However, a more likely explanation is that the cD envelope is reflecting
the shape of the cluster potential in which it was born which, if
galaxies are on strongly radial orbits, is the same statement.
MELNICK : I disagree with your statement that there is no correlation
between cluster parameters and cD envelopes. I see in your plots a
rather strong correlation which, of course, eventually disappears
if one eliminates critical data (such as the brightest envelopes).
My question is, is there any physical or statistical arguments to
eliminate the points that give a strong correlation?
SCHOMBERT : I agree that the extreme points in these diagrams are
the "best" cD galaxies, but a simple correlation
test gives only marginal significance. My statement is not that the
correlations do not exist, but that they are surprising weak, and,
therefore, have been partially erased during cluster collapse.
MERRITT : Is there any evidence that the presence of a cD galaxy correlates
with the dynamical properties of a cluster, such as its velocity
dispersion?
SCHOMBERT : A comparison of cD envelope properties with cluster dynamical
properties (velocity dispersion, Bautz-Morgan type, etc... ) is an
exploration into the world of bad statistics. The correlations exist,
but they are only marginally significant, strongly weighted by the
very few brightest cD's. I argue that this merely reflects the early
cluster state into the present epoch, and that cD envelopes are linked
to the local conditions (i.e. its original subcluster) at birth as
you predicted in 1955.
MERRITT : Does this fact argue for a primordial origin of the cD's,
as well as for the diffuse envelopes?
SCHOMBERT : There is only a weak correlation between the underlying
galaxy luminosity and the cD envelope luminosity. Therefore, I don't
believe we can conclude that the merging that produced the parent
galaxy was only primordial. In fact, bound populations exist today
with dynamical friction timescales of only 5 × 109 yrs,
so the process of dynamical evolution must still be ongoing. My
interpretation
of the parent to envelope correlations is that parallel dynamical
formation processes are being echoed from early epochs (i.e. mergers
versus tidal stripping)
NIETO : Can you comment on the flat central profiles of BCM's? Are
they intrinsically flat or is your material resolution limited?
SCHOMBERT : Due to the large variety of surface photometry for the
sample; some from CCD's, others from large scale photographic plates,
I cut the inner limit at 2 kpc and, therefore, can say very little
on the core properties of BCM's.