OPTICAL MORPHOLOGY OF RADIO

GALAXIES

A thesis submitted to the

UNIVERSITY OF PUNE
for the degree of
DOCTOR OF PHILOSOPHY
in Physics

by
ASHISH MAHABAL
Inter-University Centre for
Astronomy and Astrophysics
Post Bag 4, Ganeshkhind
Pune 411 007, INDIA

March 1998
 Declaration
This is to certify that the work incorporated in the thesis \Optical mor-
phology of radio galaxies" submitted by Ashish Mahabal was carried
out by the candidate under my supervision. Such material as has been ob-
tained from other sources has been duly acknowledged in the thesis.

Place: Pune Ajit Kembhavi

Date: 23 March 1998 (Thesis supervisor)
Contents
Contents i
Acknowledgments ix
Abstract 1
1 Overview 6
1.1 Introduction : : : : : : : : : : : : : : : : : : : : : : : : : : : : 6
1.2 Present Work : : : : : : : : : : : : : : : : : : : : : : : : : : : 9
1.3 Future Possibilities : : : : : : : : : : : : : : : : : : : : : : : : 10
2 Observations: the raw material 13
2.1 The sample : : : : : : : : : : : : : : : : : : : : : : : : : : : : 13
2.2 Observations : : : : : : : : : : : : : : : : : : : : : : : : : : : : 20
2.3 Calibration : : : : : : : : : : : : : : : : : : : : : : : : : : : : 24
2.3.1 Preprocessing : : : : : : : : : : : : : : : : : : : : : : : 25
2.3.2 Standardization : : : : : : : : : : : : : : : : : : : : : : 34
2.3.3 Correction for Galactic Extinction : : : : : : : : : : : : 39
2.3.4 K-correction : : : : : : : : : : : : : : : : : : : : : : : : 39
2.4 Control Sample : : : : : : : : : : : : : : : : : : : : : : : : : : 40
3 Brightness distribution 43
3.1 Introduction : : : : : : : : : : : : : : : : : : : : : : : : : : : : 43

i
 3.2 The radial prole : : : : : : : : : : : : : : : : : : : : : : : : : 48
3.2.1 Introduction : : : : : : : : : : : : : : : : : : : : : : : : 48
3.2.2 Obtaining the t : : : : : : : : : : : : : : : : : : : : : 49
3.2.3 Procedure Adopted for the Sample : : : : : : : : : : : 52
3.3 The dierent components : : : : : : : : : : : : : : : : : : : : : 53
3.3.1 Ellipticals and bulges : : : : : : : : : : : : : : : : : : : 53
3.3.2 The disk : : : : : : : : : : : : : : : : : : : : : : : : : : 58
3.3.3 The point source : : : : : : : : : : : : : : : : : : : : : 59
3.4 Decomposing the prole : : : : : : : : : : : : : : : : : : : : : 60
3.5 Consistency checks : : : : : : : : : : : : : : : : : : : : : : : : 65
3.6 2D distribution : : : : : : : : : : : : : : : : : : : : : : : : : : 69
3.7 Summary : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 73
4 Scale lengths at dierent wavelengths 75
4.1 Results of prole ts : : : : : : : : : : : : : : : : : : : : : : : 76
4.1.1 Goodness of t : : : : : : : : : : : : : : : : : : : : : : 76
4.1.2 Disk parameters : : : : : : : : : : : : : : : : : : : : : : 77
4.1.3 Bulge parameters : : : : : : : : : : : : : : : : : : : : : 80
4.2 Color gradients : : : : : : : : : : : : : : : : : : : : : : : : : : 89
4.2.1 Results for the samples : : : : : : : : : : : : : : : : : : 91
4.2.2 Color gradients and scale length ratios : : : : : : : : : 94
4.2.3 Preference for scale lengths : : : : : : : : : : : : : : : : 95
4.2.4 Discussion : : : : : : : : : : : : : : : : : : : : : : : : : 96
4.3 Extrapolated proles : : : : : : : : : : : : : : : : : : : : : : : 99
4.4 Conclusions : : : : : : : : : : : : : : : : : : : : : : : : : : : : 108
5 Radio galaxies with disks 109
5.1 Disky galaxies : : : : : : : : : : : : : : : : : : : : : : : : : : : 110
5.1.1 Radio galaxies : : : : : : : : : : : : : : : : : : : : : : : 113
5.1.2 Control Sample : : : : : : : : : : : : : : : : : : : : : : 115
5.2 Morphological Signatures of disks : : : : : : : : : : : : : : : : 117

ii
 5.3 Discussion : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 120
6 Dusty galaxies 124
6.1 Introduction : : : : : : : : : : : : : : : : : : : : : : : : : : : : 124
6.2 Detection techniques : : : : : : : : : : : : : : : : : : : : : : : 126
6.2.1 Color maps : : : : : : : : : : : : : : : : : : : : : : : : 126
6.2.2 Small scale features : : : : : : : : : : : : : : : : : : : : 127
6.2.3 Extrapolated proles : : : : : : : : : : : : : : : : : : : 129
6.3 Dust lane galaxies : : : : : : : : : : : : : : : : : : : : : : : : : 129
6.4 Discussion and conclusions : : : : : : : : : : : : : : : : : : : : 134
7 Radio structure: using a dierent eye 138
7.1 Introduction : : : : : : : : : : : : : : : : : : : : : : : : : : : : 138
7.2 Optical-radio correlations : : : : : : : : : : : : : : : : : : : : : 145
7.2.1 Disky galaxies : : : : : : : : : : : : : : : : : : : : : : : 146
7.2.2 Blue centers : : : : : : : : : : : : : : : : : : : : : : : : 146
7.2.3 Optical luminosity : : : : : : : : : : : : : : : : : : : : 148
7.2.4 Morphological peculiarities and dust : : : : : : : : : : 148
7.2.5 Harmonic components : : : : : : : : : : : : : : : : : : 152
7.3 Induced star formation : : : : : : : : : : : : : : : : : : : : : : 153
7.4 Conclusions : : : : : : : : : : : : : : : : : : : : : : : : : : : : 155
8 Ages: how recent is the blue 156
8.1 Introduction : : : : : : : : : : : : : : : : : : : : : : : : : : : : 156
8.2 The existing picture : : : : : : : : : : : : : : : : : : : : : : : : 157
8.3 Synthetic spectra : : : : : : : : : : : : : : : : : : : : : : : : : 159
8.3.1 Galaxy spectra : : : : : : : : : : : : : : : : : : : : : : 160
8.3.2 Modus operandi : : : : : : : : : : : : : : : : : : : : : : 161
8.4 Age of the most recent starburst : : : : : : : : : : : : : : : : : 162
9 Conclusions 164
9.1 The emerging picture : : : : : : : : : : : : : : : : : : : : : : : 164
iii
 9.2 The work goes on : : : : : : : : : : : : : : : : : : : : : : : : : 167
A Morphological processing 169
A.1 Dilation and erosion : : : : : : : : : : : : : : : : : : : : : : : 169
A.2 Opening and closing : : : : : : : : : : : : : : : : : : : : : : : 171
A.3 Gradient : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 171
B Individual objects 173
References 178

iv
List of Tables
2.1 MST parameters : : : : : : : : : : : : : : : : : : : : : : : : : 13
2.2 The low redshift MRC 1 Jy sample : : : : : : : : : : : : : : : 18
2.3 The sample : : : : : : : : : : : : : : : : : : : : : : : : : : : : 21
2.4 The observing runs : : : : : : : : : : : : : : : : : : : : : : : : 22
2.5 Filters used : : : : : : : : : : : : : : : : : : : : : : : : : : : : 23
2.6 Detectors used : : : : : : : : : : : : : : : : : : : : : : : : : : : 23
2.7 Broad band standard stars observed : : : : : : : : : : : : : : : 36
2.8 NIR standard stars observed : : : : : : : : : : : : : : : : : : : 38
3.1 Tests on NGC 661 : : : : : : : : : : : : : : : : : : : : : : : : 67
3.2 2D tests : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 70
4.1 Distribution of 2 for the two samples : : : : : : : : : : : : : 77
4.2 D=B values for the two samples : : : : : : : : : : : : : : : : : 80
4.3 Comparative re(B ) and re(R) values for the two samples : : : 81
4.4 Scale lengths for the radio sample : : : : : : : : : : : : : : : : 86
4.5 Scale lengths for the control sample : : : : : : : : : : : : : : : 87
4.6 Color gradient details using two methods : : : : : : : : : : : : 96
4.7 Details of excess near center in the radio and control samples : 100
4.8 Absorption and excess emission for the two samples : : : : : : 105
5.1 Radio galaxies with D=B > 0:3 : : : : : : : : : : : : : : : : : 116
5.2 Control galaxies with D=B > 0:3 : : : : : : : : : : : : : : : : 116

v
6.1 Statistics of objects with prominent dust features as indicated
by B ? R color maps : : : : : : : : : : : : : : : : : : : : : : : 128
6.2 Properties of the radio galaxies with dust lanes : : : : : : : : 134
6.3 Correlation of dust mass with radio and optical power : : : : : 136
7.1 Optical properties of FR I and FR II galaxies : : : : : : : : : 148
7.2 Morphological peculiarities in FR I and FR II galaxies : : : : 151
7.3 B4 and A4 distributions for the radio and control galaxies : : : 153

vi
List of Figures
2.1 Histograms of S408 for the thesis sample and the complete sample 16
2.2 KS test on S408 of the thesis sample and the complete sample 17
2.3 Cut across a atelded eld : : : : : : : : : : : : : : : : : : : 30
3.1 Hubble's tuning-fork diagram : : : : : : : : : : : : : : : : : : 44
3.2 A4 ; B4 coecients : : : : : : : : : : : : : : : : : : : : : : : : : 51
3.3 NGC 661 as a good example of de Vaucouleurs' law prole : : 57
3.4 Eect of starting 1D ts at dierent points : : : : : : : : : : : 68
4.1 2 as a function of redshift : : : : : : : : : : : : : : : : : : : : 78
4.2 Comparison of 2 values in B and R bands : : : : : : : : : : : 79
4.3 A hypothetical de Vaucouleurs' law galaxy : : : : : : : : : : : 83
4.4 The distribution of re(B )=re(R) for the radio and control samples 84
4.5 The eective radius re for the two samples : : : : : : : : : : : 85
4.6 Color gradients for the radio and the control samples : : : : : 88
4.7 Dependence of color gradients on absolute magnitude : : : : : 90
4.8 Color gradients for radio galaxies as a function of linear scale
length : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 92
4.9 Color gradients for the control sample galaxies as a function
of scale length : : : : : : : : : : : : : : : : : : : : : : : : : : : 93
4.10 Scale length ratio against color gradients : : : : : : : : : : : : 97
4.11 Fitted prole of 0446-206 : : : : : : : : : : : : : : : : : : : : : 102
4.12 Fitted prole of 0520-289 : : : : : : : : : : : : : : : : : : : : : 103

vii
4.13 Fitted prole of 1103-244 : : : : : : : : : : : : : : : : : : : : : 106
4.14 Comparison of excesses in B and R bands : : : : : : : : : : : 107
5.1 D=B as a function of Hubble type : : : : : : : : : : : : : : : : 110
5.2 D=B distribution for radio and control samples : : : : : : : : 111
5.3 D=B in R for the radio and control samples : : : : : : : : : : 112
5.4 Filled contours of disky radio galaxies : : : : : : : : : : : : : : 118
5.5 Filled contours of two more disky radio galaxies : : : : : : : : 119
5.6 Morphological gradient images of radio galaxies : : : : : : : : 121
5.7 A disky galaxy and the eect of the gradient lter : : : : : : : 123
6.1 Color maps of dust lane radio galaxies : : : : : : : : : : : : : 130
6.2 Color maps of three more dust lane radio galaxies : : : : : : : 131
6.3 Correlation of dust mass with radio power and absolute B
magnitude : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 135
7.1 3C 449: Example of an FR I galaxy : : : : : : : : : : : : : : : 140
7.2 MRC 1346-252: an example of an FR I galaxy : : : : : : : : : 141
7.3 3C 47: example of an FR II quasar : : : : : : : : : : : : : : : 142
7.4 MRC 1222-252: an example of an FR II galaxy : : : : : : : : 143
7.5 Disk-to-bulge ratio as a function of radio power : : : : : : : : 147
7.6 Scale length ratio in B and R as a function of radio power : : 149
7.7 R magnitude as a function of radio power : : : : : : : : : : : 150
7.8 Radio (1.4 GHz) and B images of 1006-214 : : : : : : : : : : 154
8.1 Synthetic B ? R and R ? K colors for E+A galaxies : : : : : 163

viii
 Acknowledgments
Following my project work during the graduate school, I had decided to
work on Sersic-Pastoriza galaxies. During the next two years I made trips to
several telescopes but failed to get good data. I had almost started worrying
then. Thankfully, Patrick McCarthy lent a helping hand at that critical
juncture. It was through our collaboration with Pat that the excellent data
on which this thesis is based could be obtained. I am greatly indebted to
him. The only other person who equals and indeed surpasses a say in this
thesis is Ajit Kembhavi who has stood by me during dataless periods as well.
I am greatly thankful to him.
If you tell someone on an average Saturday that IUCAA is a lively place,
your sanity is sure to be doubted. But IUCAA is a very lively place in
an entirely dierent way. Every now and then it goes abuzz with activity
hosting tens of people during various workshops. It is through several such
workshops and schools that I chanced to meet innumerable number of people
who have in uenced me one way or another. Too many names to mention.
But I have to mention George Djorgovski and Kenneth Freeman. Some of
their insights and ideas have made a great impact on me.
I have been helped in many ways by discussions with Paddy, Somak, Sri-
anand, Sridhar, S. K. Pandey, Anupama and Devendra Sahu. But a singular
impact in my early observing days is due to Tushar Prabhu. He was with me
during my rst observing run at the VBT. I still remember how he answered
in the greatest possible detail every small query that I had. The observing
knowledge that I gained then has helped me all through.
My sincere thanx also go to the NAs, TAs and MAs at the various ob-
servatories who kept me awake during the observing runs (Oscar, thank you
for the Spanish lessons!). And to the lively people at LCO and OCIW.
Going back a few years, MSc nal was a bad year as I was down and
stuck to the bed. No entrance exams, no interviews. A telegram from IU-
CAA conveyed the news that based on my VSP work I had been chosen for
ix
the graduate school. It was a ray of hope. My stay at IUCAA has been
memorable. I thank Jayant Narlikar and all the members of IUCAA for pro-
viding excellent and ecient working atmosphere. Names worth mentioning
include Lata, who helped during the anxious ying moments, and Chella for
patiently regiving lost forms. The most notable facilities at IUCAA are the
library and the computer network. I am thankful to the computers for pro-
viding company during large parts of the days, and indeed nights. And to
friends when the computers were down. But seriously, it was great to have
a friend like Ramaprakash around.
I will risk sounding like a Radio Ceylon DJ and say that I am thankful
to Ramaprakash, Srikumar, Yogesh, Srini, Sunu, Sonia, Sukanya, Ranjeev,
Rama, Sharadini, Archana, Ajit, Asha, Ramana, Ishwar and many others
for wonderful discussions/treks/dinners. I have enjoyed sharing puzzles and
discussing sports with Bala, Hariki, Srianand, Arun, Santosh and Pravin. I
have also enjoyed debating cricketing issues with Sriram and playing cricket
with Santosh, Anoop and Rajesh. You all have helped keep me sane during
the last 5+ years.
I am most thankful to Somak, Indrajit, Yogesh, Niranjan and Anu for
useful suggestions and help regarding the thesis. I also thank Naresh for
help with the technical supervision of the thesis.
I must also mention those who have in uenced my life and have been re-
sponsible, at least partly, for my being here in the rst place. I am thankful
to my father for introducing me to the night sky, to Vivek Wagh for intensi-
fying the interest, and to Anjutai and Yashodhan Paranjapye for being there
when most needed.
Finally, and most importantly, Anu and Jay have patiently labored during
the last several months (years? ages?) hoping that I will spend more time at
home after the thesis. They keep reminding me that there exists a universe
beyond astronomy. I owe a very great deal to them.
Dedicating this thesis to aai and mai ...

x
 Abstract
This thesis is about the optical and near-infrared morphology of low-
redshift (z < 0:3) radio galaxies. It is based on observations of a radio
selected sample of galaxies from the Molonglo Reference Catalogue. We
have observed these galaxies with the 1.0m and 2.5m telescopes at the Las
Campanas Observatory in Chile, using broad band B and R and near-infrared
K 0 lters. The morphology and structure of the galaxies have been studied
through a series of steps involving the analysis of the 2-dimensional brightness
distribution of the galaxies. We outline below the motivation behind this
work, summarize the techniques and mention important results.
Traditionally, radio galaxies have been believed to be ellipticals consisting
of a coeval population of old stars, and almost no dust or gas. However,
photometric studies carried out over the past few years have demonstrated
that elliptical galaxies not only have dust and/or gas, but also possess ne
structure indicating some amount of activity in the past  102 million years.
Radio galaxies host an active galactic nucleus (AGN) and are associated with
highly energetic phenomena like the radio jets which transport a very large
amount of energy over hundreds of kiloparsecs. Such phenomena are likely
to be associated with morphological features not found in normal elliptical
galaxies. A motivation of this thesis has been to identify such features in the
radio galaxies.
We have (1) studied the 2-dimensional brightness distribution, in the B ,
R and K 0 bands, of the galaxies in the plane of the sky by tting elliptical
shapes to isophotes; (2) tted standard bulge and disk laws to 1-dimensional
major-axis proles extracted from the elliptical model isophotes; (3) studied
distributions of tted parameters and correlations between them; (4) studied
color distribution in the galaxies using dierent techniques; (5) correlated
radio, optical and near-IR properties; (6) used morphological image process-
ing techniques to identify faint features and (7) carried out simulations and
modeling to conrm the validity of the extracted features.
1
 In order to dierentiate between the properties of radio galaxies as a class
from those of normal galaxies, we have compared radio galaxy properties with
those of a control sample consisting of non-radio selected early-type galaxies.
The comparison sample has been drawn from the CCD frames of the radio
galaxies. This ensures that the radio and control samples are subject to
identical data processing and photometric calibration. Comparison has also
been made with properties of galaxies from the literature wherever adequate
data has been available.
We begin the thesis by presenting an overview of the properties of radio
galaxies and elliptical galaxies. We explain the importance of the study of
the morphology of galaxies to the understanding of the physical phenomena
that give rise to these morphologies.
We then go on to describe in detail (1) the sample that has been used
for the present study; (2) the instruments used and the data acquisition
techniques; (3) preprocessing to remove the signature of the instruments
from the data; (4) calibration using the observations of standard stars and (5)
various corrections applied to the data to account for the Galactic extinction
and the redshift of the object. Finally we describe the control sample used
in this thesis.
We next discuss the isophotal shapes of the program and control galaxies.
We expect the projected shapes of isophotes of elliptical galaxies to be el-
lipses. We describe a program, which, for an isophote at a given semi-major
axis length, provides the best tting ellipse characterized by an ellipticity
and position angle. A series of such ellipses with dierent semi-major axis
lengths then provide an intensity prole for the galaxy image. The devia-
tion of the isophotes from the best tting ellipses yield important parameters
which are indicative of disky or boxy nature of the isophotes. We describe the
various correlations that we nd amongst the parameters which are obtained
from the ellipse tting exercise. These parameters provide useful information
about the 3-D shapes of the galaxies. A section is devoted to the present

2
understanding of the 3-D intensity distribution in elliptical galaxies.
The radial intensity prole of a galaxy is well described by two compo-
nents viz. the bulge, which is well approximated by de Vaucouleurs' r1=4
law, and the disk which is exponential in nature. Each of these components
is characterized by a scale length (re for the bulge, rs for the disk) and the
intensity at a characteristic radius. In case of an elliptical galaxy, the bulge
dominates over the disk. In fact, until a few years ago, it was believed that
elliptical galaxies do not contain a disk component at all. We have obtained
the scale lengths for each galaxy by tting the intensity prole, in the dif-
ferent lters available, with a bulge-disk combination. Such a decomposition
into bulge and disk parameters plays an important role in determining vari-
ous properties related to the morphology. We present details of the technique
along with a working algorithm and its implementation.
We then describe the distribution of various tted parameters and the
relations between them. Since the bulge is dominant in ellipticals, we partic-
ularly emphasize bulge related properties. While absorption due to dust, age
and metallicity tend to redden a galaxy in its central region, star formation
occurring near the center causes it to have a greater concentration of blue
light there. All these factors aect the scale lengths, and the scale lengths
in dierent lters (e. g. B , R, K 0) then indicate the more dominant factor
in each lter. For normal galaxies, one expects that re is larger at shorter
wavelengths since normal age and metallicity eects dominate over star for-
mation. From a study of bulge scale lengths we show that radio galaxies
tend to have excess blue colors in their central regions compared to normal
ellipticals. The origin of the extra blue light, seen over several kpc from the
centers of these galaxies, is likely to be the formation of stars over the last
several 100 Myr. The modeling that we have carried out in this context is
described later on in the thesis.
Galaxies that have a disk-to-bulge ratio (D=B ) > 0.3 are classied as
lenticulars or spirals. We nd that  20% of the galaxies in our radio sample

3
have D=B > 0:3. We have investigated these disky ellipticals in detail. In
two cases we nd that the D=B value, as well as parameters describing the
disk-like structure, are similar to those obtained for Sb and Sc galaxies. In
the remaining cases the D=B value is high because of contribution from a
disk-like structure having a small scale length. The bulge is dominant and
the disk scale length to bulge scale length ratio (rs=re) for these galaxies is
much smaller than in lenticulars and spirals.
To better isolate these disks and other faint features in the 2-D images,
we have developed a number of morphological lters. As the name suggests,
these lters deal with shapes within images. The morphological gradient lter
has been found to be especially useful in locating changes in the luminosity
levels since it depends less on edge directionality than the Sobel operator. We
describe in detail the gradient and other morphological lters.
We then discuss the detailed morphology of individual galaxies in the
sample. We say a galaxy is disturbed if it possess at least one of the follow-
ing. (1) A secondary nucleus (a non-stellar brightness peak within 10 kpc of
the galaxy brightness peak); (2) tidal features; (3) other features that result
in distorted isophotes. We look for these signs of disturbance by making
use of color maps, distribution of intensity prole parameters and morpho-
logical image processing techniques. We show that radio galaxies, though
they possess a morphology very similar to that of elliptical galaxies, more
often exhibit signs of disturbance than normal elliptical galaxies: there is a
higher incidence of secondary nuclei in radio galaxies, their isophotes devi-
ate from being elliptical more often than for normal galaxies and there is a
greater incidence of blue structures that are indicative of recent star forming
activity.
We explore in some detail the connection between radio properties and
features and parameters in the optical and near-IR bands. We nd that FR II
radio galaxies are more likely to possess a disk than FR I radio galaxies. We
nd that blue central regions preferentially occur in the more powerful radio

4
galaxies. In some of the radio galaxies, star formation is seen to preferen-
tially lie along the radio axis and is likely to have been induced by the radio
emission. Such a phenomenon has been noted before but for higher redshift
radio galaxies.
We model the age of the recent star formation using synthetic spectra in
the optical and near-IR bands. We assume that the galaxy is formed at some
epoch in the past, with the Scalo stellar initial mass function (IMF). To such
a galaxy we add a starburst, again with the Scalo IMF. The spectrum of the
galaxy, at the present time, is the sum of the individual spectra of appro-
priately evolved stars originally formed, and in the starburst. We compare
broad band colors of the spectrum with observed colors, and determine the
burst mass and epoch for the best correspondence.
Finally, we provide a summary of the thesis and discuss the new lines of
investigation, for studying the morphology of radio galaxies, that have been
suggested by the present work.

5
Chapter 1
Overview
1.1 Introduction
Powerful radio sources are known to be hosted almost universally by early
type galaxies. Recent observations with high dynamic range detectors have
shown that elliptical galaxies hosting radio sources show the presence of faint
structure in various forms. In this thesis we study the optical and near in-
frared morphology of low redshift (z < 0:3) radio galaxies with particular
emphasis on morphologically peculiar features and the distribution of dust.
In the following sections of this chapter, we brie y summarize known prop-
erties of normal elliptical and radio galaxies, mention the most important
results we have obtained, and indicate what we have left for the future.

Elliptical galaxies
In the 1930's Hubble introduced the concept of elliptical galaxies as a sep-
arate class of objects that did not contain any substructure. This \tradi-
tional" view persisted till the 1970s. As a result, most research on elliptical
galaxies concentrated on studying the properties of their stellar component.
Initially, it was thought that the attening of elliptical galaxies is due to a

6
rotation about the polar axis (Osterbrock, 1960). More detailed observations
indicated that elliptical galaxies rotated much more slowly than required to
support the attening (Bertola and Cappacioli, 1975; Illingworth, 1977). It
was also shown that the elliptical galaxies were supported by anisotropic ran-
dom velocities of stars (Binney, 1976) and that these galaxies are likely to be
triaxial bodies rather than axisymmetric oblate spheroids. These results led
to dynamical modeling of elliptical galaxies. Modeling related to accretions
and mergers produced features like arcs, shells and ripples which had been
seen in elliptical galaxies (e. g. Malin and Carter, 1983; Quinn, 1984). We
have extensively used existing techniques and developed a few of our own
to carefully look for and quantify the occurrence of small scale structure in
radio galaxies and in a control sample. However, we have not carried out any
modeling as part of this thesis work.
As per the denition of elliptical galaxies, they are diskless. Galaxies
whose morphological characteristics are intermediate between disk-dominated
spirals and spheroidal ellipticals are classied as S0 galaxies, also called lentic-
ulars. Recently, it has been suggested that elliptical galaxies with disky
isophotes are the bridge between the normal, diskless ellipticals and S0 galax-
ies (Bender et al. , 1993). The disky ellipticals are likely to be rapidly rotat-
ing, low-luminosity, coreless, nearly isotropic and oblate spheroidal galaxies
(Kormendy and Bender, 1996). We nd that  17% of the radio galaxies
in our sample contain an appreciable disk component (D=B > 0:3). These
galaxies are described in detail in Chapter 5.
The radio galaxy Cen A (NGC 5128) is a peculiar elliptical galaxy with
a prominent dust lane. Bertola and Galletta (1978) showed that it is a pro-
totype of a new kind of objects, the dust lane galaxies. Subsequent searches
showed that several galaxies contained dust lanes (e. g. Hawarden et al. ,
1981) and astronomers began to ponder on the origin of dust and the possi-
ble correlations of the presence of dust with other properties. Studies have
shown that ellipticals with boxy isophotes are associated with radio-loud

7
galaxies (Bender et al. , 1989) and with essentially nonrotating giant ellipti-
cals (Kormendy and Bender, 1996). A dust lane along the major-axis leads
to boxy isophotes and it may be possible that radio galaxies, in general, are
dust lanes galaxies. One of the triggers of this thesis has been to examine
the incidence and distribution of dust in radio galaxies. These issues have
been covered in Chapter 6.

Radio galaxies
If the radio luminosity of a galaxy is comparable to or exceeds its optical
luminosity (LR =Lopt >1), then the galaxy can be called a radio galaxy. The
dividing line, however, is not sharp since the criteria for separating radio
loud and radio quiet objects tends to be subjective. Cygnus A was the rst
radio source to be associated with a galaxy. Jennison and Das Gupta (1953)
showed that the radio source had a double lobed structure. The list of radio
galaxies grew rapidly thereafter. About a tenth of giant elliptical galaxies
are radio loud (Lradio >1041 erg sec?1 ).
Early observations indicated that the hosts of all powerful radio sources
are elliptical galaxies. Seyfert nuclei, on the other hand, are almost always
associated with spiral hosts. Seyferts are towards the low luminosity end
of the radio galaxy luminosity function and may form the bridge between
powerful radio galaxies and spirals. Recent work using detectors with higher
dynamic range has indicated that though the radio galaxies appear to be of
the early type, there are morphological peculiarities associated with them.
Heckman et al. (1985, 1986), showed that the enhanced local galaxy density
and galaxy mergers can act as producers of powerful radio sources in early-
type galaxies. It is now believed that strong radio sources are associated with
elliptical galaxies or recent merger remnants (see e. g. Urry and Padovani,
1995).
Fanaro and Riley (1974) showed that, from their radio structures, one
can divide radio galaxies into two distinct classes viz. FR I and FR II. This

8
dichotomy is seen to extend to several properties associated with the radio
galaxies. More recently, Owen and White (1991) showed that the FR I/II
division depends not only on radio power but also on the optical luminosity.
In Chapter 7 we discuss the ndings for our radio sample in this context.

1.2 Present Work

In this thesis we have studied 30 radio galaxies from the Molonglo Reference
Catalogue (MRC) which have a redshift < 0:3. A brief description of the
work has been presented in the abstract. Here we take a look at the contents
from the point of view of new achievements.
The main objective of the thesis has been to study the optical and near
infrared morphology of radio galaxies. In the course of the work we have
shown that radio galaxies do not have highly distorted morphology. Their
surface brightness proles can, in  85% of the cases, be t rather well by
a de Vaucouleurs' law model, with the addition, in some cases, of a signi-
cant disk component. The good ts are obtained right upto the maximum
redshift of 0.3 of our sample. Over half the radio galaxies show morphologi-
cal peculiarities superposed on the rather regular large scale distribution of
light. The peculiarities occur more frequently in case of the radio galaxies
than they do for the control sample galaxies.
From our model prole ts to the observed surface brightness proles we
have found that the distribution of the ratio of de Vaucouleurs scale lengths
in the B and R bands re(B )=re(R) is dierent for the two samples. A value
< 1 for this ratio indicates that the color becomes bluer towards the center,
while a value > 1 indicates redder color towards the center. The ratio has a
simple relation with the color gradient obtained directly from color proles.
But the ratio can be used as an indicator of the radial dependence of color,
even when the measured gradient is too noisy to serve the purpose. Simply
on the basis of the conventionally measured color gradients we could not

9
have been able to make the distinction between radio and control samples.
Extrapolation of model proles to inner regions, which are not used in the
t, provide a measure of star formation activity and dust absorption there.
One of the objectives of the thesis was to look for signatures of dust in
radio galaxies. With the work on scale length ratios we have been able to
present a new technique to look for dust as well as star formation. As far
as the distribution of dust is concerned, we nd that dust is present at the
centers of radio galaxies more frequently than in the control sample galaxies.
The dust in radio galaxies is seen to be more coherent than in the control
sample. We also found that in case of the radio sample, the mass of dust
in lanes is correlated to the radio luminosity of the galaxy. The dust mass,
however, correlates only weakly with the absolute B magnitude.
According to the Hubble classication scheme, an elliptical galaxy is to-
tally devoid of dust and gas and possesses no substructure. Whenever a
galaxy was seen to possess a disk, it was classied as either a lenticular or
a spiral. The bulge and disk parameters obtained by tting the intensity
prole of each galaxy with a bulge-disk combination indicate that  17%
radio galaxies in our sample have D=B > 0:3 and on that basis alone could
be classied as lenticulars or spirals. To better isolate these disks and other
faint features in the 2-D images, we have adapted the morphological gradi-
ent lter. One of our disky galaxy clearly shows the presence of spiral-like
features. Using the morphological gradient lter we have been able to bring
out spiral-like features in two more galaxies.

1.3 Future Possibilities

This section brie y outlines the future possibilities. These include projects
that could not be done due to lack of appropriate data and projects that
could not be undertaken in the time available. Of the many projects that
can be put in these categories, we mention a few here.

10
 The fundamental plane (Djorgovski and Davies, 1987; Dressler et al. ,
1987) for ellipticals is a two-dimensional manifold in the three-dimensional
space formed by the global structural parameters (1) scale length, re, (2)
velocity dispersion, , and (3) mean eective surface brightness hSB ie. The
fundamental plane is likely to be independent of environment (Jrgensen,
1996). The face-on view of the fundamental plane provides information about
formation processes and evolution. Lack of availability of velocity dispersion
for the radio and control galaxies has prevented us from locating these galax-
ies on the fundamental plane. However, we hope to do this as part of a future
study.
With the available data we could have positioned the galaxies on the
photometric fundamental plane, as well as ranked the galaxies within the
surrounding region. However, even the group membership of certain galaxies
close to the radio galaxy was uncertain. We will pursue these lines in the
near future.
We have shown that on the average radio galaxies tend to possess excess
blue color in the central region as compared to control galaxies. It is likely
that the excess blue has its origin in a star formation episode in the recent
past. High resolution spectra will allow us to determine the age and strength
of such a burst of star formation.
It is now known that nuclear density proles of elliptical galaxies continue
to rise upto the smallest observable radius. It has been debated whether
there exist two types of galaxies viz. those with a core and those with a
cusp (e. g. Ferrarese et al. , 1994) or it is a continuous family (Merritt, 1997)
or, an artifact, to some extent, due to selection eects like distance to the
objects (Faber et al. , 1997). Dynamical modeling based on light distribution
of the radio galaxies in our sample will provide valuable information on their
intrinsic shapes. In case of dust lane galaxies, we have information about
the radio source orientation and the position angle of the dust lane thus
providing us with an additional vector to study the 3D shape than is usually

11
available. We would like to pursue dynamical modeling in the future.

12
Chapter 2
Observations: the raw material
2.1 The sample
The sample chosen for this study has been drawn from the Molonglo Refer-
ence Catalogue (MRC, Large et al. , 1981). The MRC was the result of a 408
MHz survey made from 1968 to 1978 using the Molonglo Synthesis Telescope
(MST). The parameters of this meridian transit instrument are summarized
in Table 2.1.
The limiting ux density (5) of the MST for the detection of a point
source near the zenith was less than 0:3 Jy. The MRC itself had a limiting
ux density of 0:7 Jy and hence thermal noise and confusion could not have

Table 2.1: MST parameters

Operating Bandwidth Beamwidth Beamwidth Eective
frequency   latitude a
408 MHz 2.5 MHz 20:62 20:86sec( ? a) -35 deg:5
a denotes the eective latitude

13
seriously aected it. The catalogue is believed to be 99.9% complete for ux
density  0:95 Jy.
The MRC covers 7:85 steradian between ?85:0 deg and +18:5 deg dec-
lination, excluding a 3 deg strip near the Galactic plane. The total num-
ber of sources is 12141, which corresponds to an average source density of
1500 ster?1. Of these, 7347 sources have a ux density > 1 Jy. C. R. Subrah-
manya (see McCarthy et al. , 1990, 1991) dened a subset of this sample by
limiting the right ascension and declination covered: ?30 deg  (1950) 
?20 deg, 09h20m  (1950)  14h04m or 20h20m  (1950)  06h04m. This
contains 558 sources.
An additional constraint was used to create the sample used in the present
study. Since one of the primary aims of this project is to study the optical
morphology of radio galaxies, it was necessary that the objects being imaged
subtend areas that are several arcsec across so that various techniques could
be reliably used. With this in mind a redshift cuto of 0.3 was introduced.
Further, sources identied with quasars and BL Lac objects were excluded.
Thus the present study includes radio galaxies from the MRC that satisfy
the following criteria:
 S408  0:95 Jy
 ?30 deg  (1950)  ?20 deg
 09h20m  (1950)  14h04m or 20h20m  (1950)  06h04m
 z  0:3
This sample will henceforth be referred to as the low redshift MRC 1 Jy
sample. It consists of the 83 objects given in Table 2.2.
During the two rounds of observations when time was available for the
project, 30 galaxies from the sample could be observed. These are the ones
we will be using in our analysis. We will often refer to these galaxies as the
sample or the radio sample.
14
 Basic information about the sample has been included in Table 2.3. A
Kolmogrov-Smirnov test (Figure 2.2) shows, at a signicance level > 99:9%,
that the ux selected subsample and the original complete sample of 83 radio
galaxies belong to the same parent population.
A few comparable radio selected samples exist in the literature. These
have been reviewed by McCarthy (1993). Specically, the 3CRR sample
(Laing et al. , 1983) is a subset from the 3CR sample. The 3CR itself was
badly aected by confusion. The 3CRR sample contains sources with S178 
10 Jy and is believed to be 96% complete. The other constraints are  
10 deg and jbj  10 deg. Allington-Smith (1982) studied a complete sample
from the Bologna catalogue (Colla et al. , 1970). The selection criteria for
the complete sample were 1  S408  2 Jy, b > 30 deg, 34 deg    40 deg
and  < 13h02m. This sample consists of 59 sources and is in many ways
similar to the present sample.
Besides these northern hemisphere studies, the Parkes Selected Regions
(Downes et al. , 1986) form an important sample from the southern hemi-
sphere. This catalogue consists of 188 sources with S2700  0:1 Jy, is a subset
of the Parkes catalogue, and overlaps with the MRC. However, a fth of the
overlapping Parkes sources are not included in the MRC since the latter
has a more sharply dened ux density cut-o. Also, the 2700 MHz Parkes
catalogue contains many at-spectrum sources which fall below the MRC
cut-o.
A complete optical and NIR study has not been done on any of these sam-
ples. When presenting results from this work, comparison will be frequently
made with whatever results are available on the above samples. Work on
the higher redshift MRC sources (e. g. McCarthy et al. , 1996, Athreya et al. ,
1997) will also be used for comparison.

15
Figure 2.1: The histograms of radio ux at 408 MHz for the complete low-
redshift 1 Jy sample and the sample observed for the thesis. The two his-
tograms are similar in nature. A quantitative comparison is shown in Fig-
ure 2.2 using a Kolmogrov-Smirnov test.

16
Figure 2.2: The Kolmogrov-Smirnov test on the 408 MHz ux of the complete
low-redshift sample and the thesis sample conrms at a signicance level of
> 99:9% that the two have been drawn from the same parent population.

17
 Table 2.2: The low redshift MRC 1 Jy sample
IAU z S408 S5000  RS LAS RA(1950) Dec(1950)
mJy mJy 408/5000 arcsec
0001-233 0.100 1230 341 y 0.99 y - 112.0 00 01 28.31 -23 23 41
0017-205 0.197 1960 184 . 0.95 . - 27.6 00 17 01.35 -20 32 53
0028-223 0.206 1180 165 . 0.79 . D 23.2 00 28 32.40 -22 23 43
0038-209 0.091 1270 614 y 0.56 y S < 6:0 00 38 18.90 -21 00 08
0045-255 0.001 6120 .. .. - . 00 45 00.00 -25 30 00
0048-233 0.110 1180 399 y 0.84 y C 45.0 00 48 25.34 -23 23 32
0055-256 0.199 1180 106 . 0.97 . D 22.8 00 55 31.10 -25 38 49
0100-221 0.058 2470 133 y 2.26 y - 35.0 01 00 15.32 -22 09 03
0115-261 0.268 2580 990 y 0.74 y R 10.0 01 15 52.77 -26 07 35
0137-263 0.155 1460 186 . 0.83 . D 77.1 01 37 32.60 -26 22 19
0149-260 0.144 1050 98 . 0.95 . T 96.5 01 49 32.40 -26 02 03
0155-212 0.159 2390 619 . 1.04 y T 86.0 01 55 32.30 -21 16 52
0208-240 0.230 1870 296 . 0.74 . D 67.1 02 08 59.50 -24 05 27
0226-284 0.210 1480 .. .. - . 02 26 14.25 -28 26 50
0229-208 0.087 1870 .. .. - . 02 29 18.66 -20 53 36
0247-207 0.086 1510 .. .. - . 02 47 19.10 -20 42 48
0253-220 0.112 1120 149 . 0.81 . S . 02 53 23.90 -22 05 58
0302-273 0.115 1110 144 . 0.82 . R 6.3 03 02 12.60 -27 22 45
0305-226 0.268 4950 564 . 0.88 . T 88.0 03 05 17.80 -22 36 51
0313-271 0.216 1820 418 y 1.14 y D 227.0 03 13 55.12 -27 09 00
0326-288 0.108 4030 510 . 0.83 . T 17.0 03 26 31.60 -28 52 08
0344-291 0.137 2140 840 y 0.72 y S < 5:0 03 44 58.67 -29 09 40
0349-278 0.066 8750 .. .. - . 03 49 31.66 -27 53 31
0357-247 0.205 2160 252 . 0.87 . T 30.3 03 57 49.19 -24 42 22
0420-263 0.130 2820 479 . 0.71 . D 219.0 04 20 30.85 -26 23 30
0434-225 0.067 2260 .. .. - . 04 34 28.50 -22 32 24
0442-282 0.150 18850 1704 . 0.97 . T 85.6 04 42 37.49 -28 15 18
0446-206 0.073 2060 .. .. - . 04 46 19.50 -20 37 32
0452-260 0.140 980 124 . 0.83 . R 12.4 04 52 31.69 -26 02 42
0453-206 0.035 11250 715 . .. - . 04 53 14.07 -20 36 59
0457-247 0.185 1250 222 . 0.70 . C > 60:0 04 57 50.00 -24 44 04
0503-284 0.037 1030 834 y .. - . 05 03 42.00 -28 29 00
0508-220 0.160 5100 680 . 0.81 . T 38.5 05 08 53.08 -22 05 32
0512-200 0.133 980 178 . 0.69 . T 64.5 05 12 43.09 -20 00 47
0520-289 0.068 1360 202 . 0.77 . D 7.8 05 20 03.40 -28 58 58
0522-263 0.290 1360 418 y 0.91 y S < 5:0 05 22 07.88 -26 22 50
0952-224 0.230 1710 367 . 0.62 . D 6.2 09 52 20.48 -22 27 03
1006-214 0.250 1350 492 y 0.78 y T 191.0 10 06 13.00 -21 24 30
1033-251 0.110 1530 521 . 0.83 y D 84.3 10 33 49.50 -25 09 30
1048-238 0.207 1310 484 y 0.76 y - 82.0 10 48 09.83 -23 50 02
1053-282 0.061 3510 591 y 1.38 y R 30.0 10 53 10.09 -28 15 34
1056-272 0.250 970 88 . 0.97 . T 7.1 10 56 32.07 -27 14 24
1103-244 0.050 3900 350 . 0.97 . - 17.5 11 03 45.50 -24 28 32
1126-246 0.155 1100 174 . 0.74 . T 51.7 11 26 34.41 -24 41 09

18
 Table 2.2: The low redshift MRC 1 Jy sample (continued)
IAU z S408 S5000  RS LAS RA(1950) Dec(1950)
mJy mJy 408/5000 arcsec
1153-231 0.080 1940 190 . 0.94 . R 45.0 11 53 36.17 -23 10 35
1215-215 0.075 1500 192 . 0.83 . S . 12 15 09.04 -21 30 48
1222-252 0.077 1820 155 . 0.99 . D 133.0 12 22 13.38 -25 12 13
1226-211 0.190 3280 448 . 0.80 . D 29.3 12 26 08.76 -21 09 41
1230-244 0.257 1930 191 . 0.93 . S . 12 30 05.22 -24 29 18
1251-289 0.058 4220 839 y 1.20 y C 25.0 12 51 58.39 -28 57 30
1254-268 0.135 1140 182 . 0.74 . D 24.5 12 54 47.28 -26 47 33
1257-253 0.045 1500 622 y 0.68 y T 43.0 12 57 07.29 -25 23 35
1258-229 0.127 1400 408 . 0.50 . R? . 12 58 17.41 -22 56 07
1303-215 0.120 1460 222 . 0.76 . D 25.5 13 04 00.87 -21 31 50
1309-211 0.300 1670 214 . 0.83 . D 53.7 13 09 10.13 -21 06 47
1323-271 0.044 2170 .. .. - . 13 23 25.59 -27 10 49
1329-257 0.190 3730 411 . 0.89 . T 48.8 13 29 45.68 -25 44 21
1334-296 0.001 1650 .. .. - . 13 34 10.00 -29 36 37
1343-253 0.129 1050 401 y 0.74 y T 55.6 13 43 55.55 -25 22 32
1344-241 0.019 1020 .. .. - . 13 44 26.50 -24 07 58
1346-252 0.125 1270 248 y 1.26 y T 55.0 13 46 05.10 -25 12 26
1354-251 0.037 1080 .. .. - . 13 54 27.29 -25 08 53
2020-211 0.054 1590 46 . 1.43 . D 14.9 20 20 04.40 -21 06 39
2040-219 0.204 1200 82 . 1.08 . D 7.1 20 40 20.42 -21 55 24
2040-267 0.040 3450 .. .. - . 20 40 46.47 -26 43 52
2053-201 0.155 6370 867 . 0.80 . D 29.7 20 53 11.98 -20 07 50
2058-282 0.038 6870 .. .. - . 20 58 39.96 -28 14 04
2101-214 0.198 960 158 . 0.73 . T 31.9 21 01 24.98 -21 24 59
2104-256 0.037 13250 .. .. - . 21 04 24.88 -25 38 19
2117-269 0.103 2600 386 . 0.77 . T 31.3 21 17 49.08 -26 57 28
2134-281 0.070 2120 344 . 0.73 . T 30.2 21 34 15.50 -28 08 55
2206-237 0.087 3780 2009 y 0.51 . S < 2:0 22 06 32.58 -23 46 39
2206-251 0.158 2040 203 . 0.93 . T 104.0 22 06 34.37 -25 08 23
2308-214 0.151 990 126 . 0.83 . T 58.5 23 08 09.20 -21 24 49
2317-277 0.173 5440 706 . 0.82 . C 79.0 23 17 15.46 -27 43 50
2321-228 0.114 1030 51 . 1.21 . D 36.6 23 21 24.50 -22 51 25
2324-259 0.286 1440 240 . 0.72 . D 21.7 23 24 59.78 -25 58 19
2327-215 0.280 2050 231 . 0.88 . S . 23 27 02.89 -21 30 18
2345-284 0.028 960 9. 1.87 . S . 23 45 09.04 -28 25 07
2356-210 0.097 990 225 . 0.60 . T 17.5 23 56 45.68 -21 04 45
Notes:
y: Value at 1400 MHz used instead of 5000 MHz
: Radio spectral index
LAS: Linear Angular Size
RS: Radio structure. C: Complex, D: Double, R: Resolved, S: Single, T: Triple

19
2.2 Observations
Optical and NIR observations were carried out in two rounds at the Las
Campanas Observatory (LCO), Chile. The telescopes used were the 1.0m
f/7 Swope telescope and the 2.5m f/7.5 Du Pont telescope of the Carnegie
observatory. Table 2.4 gives a summary of the two observing runs. The lters
and detectors used are listed in Tables 2.5 and 2.6 respectively.
A large number of twilight ats were taken, both in the evening and in the
morning. During the day dome ats were also taken, but these were not used
eventually since the quality of the twilight ats was acceptable. An overscan
bias region was present on the CCD. Still, a few bias frames were obtained
at the beginning of the night and also every few hours during each observing
session, just to ensure that the bias level remained constant for the length
of the exposures used. Dark current is very low in modern CCD detectors
and was neglected. No dark current contribution was later discernible in the
data.
Object scheduling was done so that objects were closest to zenith when
observed. As far as possible, objects were sandwiched between observations
of standard stars. Standard stars from the Standard Areas (SA) 98 and 104
were observed at dierent air masses for eective calculation of the apparent
magnitudes. Though all objects in the sample were meant to be observed
during the runs, when only one of two objects close to each other could be
observed, the brighter one was preferred. Because of this, a bias in redshift
selection of the sample might have crept in with preference having been given
to the nearer (brighter) object. It has already been demonstrated (Figure 2.2)
that no radio ux density selection bias has been introduced.
Multiple frames of each object were taken with exposure time adjusted
to avoid saturation. Each object was also moved around in the frame during
successive exposures, so that bad pixels do not aect any particular object
in all the frames. The typical total exposures in the B and R lters were

20
 Table 2.3: The sample
IAU z S408 RS RA(1950) DEC(1950) Galactic
Jy l b
0344-291 0.137 2.140.07 03 44 58 .90.2 -29 09 40 3 226.3 -51.4
0349-278 0.066 8.750.38 E 03 49 36 .10.2 -27 52 47 3 224.6 -50.2
0354-263 0.000 1.200.07 03 54 46 .30.3 -26 21 31 5 222.6 -48.8
0420-263 0.130 2.820.13 E 04 20 31 .30.3 -26 23 49 4 224.5 -43.2
0434-225 0.067 2.260.08 E 04 34 28 .50.2 -22 32 24 3 220.8 -39.1
0446-206 0.073 2.060.04 E 04 46 19 .50.2 -20 37 32 2 219.6 -35.9
0453-206 0.035 11.250.16 04 53 14 .10.2 -20 38 58 2 220.3 -34.4
0503-284 0.037 1.030.05 C 05 03 42 .30.3* -28 28 26 4* 230.2 -34.4
0520-289 0.068 1.360.06 05 20 02 .60.3* -28 58 52 4 231.9 -31.1
0938-205 0.000 1.350.03 09 38 31 .50.2 -20 33 52 3 254.1 23.5
0950-239 0.000 1.380.08 09 50 53 .90.3 -23 58 12 5 258.9 23.0
0952-224 0.230 1.710.08 09 52 20 .30.2 -22 26 51 4 258.0 24.4
1006-214 0.250 1.350.06 E 10 06 11 .90.3 -21 24 40 4 259.9 27.3
1033-251 0.110 1.530.05 10 33 49 .70.2 -25 09 43 3 268.2 28.2
1048-238 0.207 1.310.08 10 48 10 .40.4 -23 50 00 5 270.5 31.1
1053-282 0.061 3.510.09 10 53 10 .10.2 -28 15 34 3 274.2 27.8
1056-272 0.250 0.970.05 10 56 32 .60.3 -27 14 30 4 274.4 29.0
1103-244 0.050 3.900.18 11 03 45 .30.2 -24 28 32 4 274.5 32.3
1126-246 0.155 1.100.04 11 26 34 .20.2 -24 41 11 3 280.2 34.3
1153-231 0.075 1.940.05 11 53 36 .60.2 -23 10 38 3 286.9 37.7
1215-215 0.077 1.500.06 12 15 09 .40.2 -21 30 45 4 292.6 40.4
1251-289 0.058 4.220.17 12 51 58 .40.2 -28 57 30 4 303.8 33.6
1254-268 0.135 1.140.04 12 54 47 .80.2 -26 51 39 3 304.6 35.7
1257-253 0.045 1.500.04 12 57 07 .30.2 -25 23 35 3 305.3 37.2
1303-215 0.120 1.460.06 13 03 59 .90.2* -21 31 52 4 307.6 40.9
1323-271 0.044 2.170.08 E 13 23 25 .50.2 -27 10 49 3 312.3 34.8
1329-257 0.190 3.730.10 13 29 45 .70.2 -25 44 26 3 314.4 36.0
1343-253 0.129 1.050.04 13 43 54 .30.2 -25 23 08 3 318.3 35.6
1344-241 0.019 1.020.09 E 13 44 26 .50.3 -24 07 58 7 318.8 36.8
1346-252 0.125 1.270.04 13 46 06 .90.2 -25 12 13 3 318.9 35.6
1354-251 0.037 1.080.04 13 54 27 .30.2 -25 08 53 3 321.2 35.1
Notes: E indicates extended and type C complex radio structure. Redshifts are
from McCarthy et al. (1996), the other data is from MRC.

21
 Table 2.4: The observing runs
40" TEK#1y 40" TEK#5y 100 arcsec IRCAMy
Date Object Date Object Date Object
25 Jan 1995 0434-225 17 Feb 1996 0420-263 27 Feb 1997 0503-284
0520-289 0952-224 1048-238
1053-282 1344-241 1103-244
1222-252 1346-252 1215-215
26 Jan 1995 0446-206 18 Feb 1996 0354-263 1251-289
0503-284 1048-238 1344-241
1103-244 1254-268 1323-271
1251-289 20 Feb 1996 0354-263 1346-252
27 Jan 1995 0349-278 0952-224 29 Feb 1996 0344-291
0950-239 1048-238 0349-278
1215-215 1323-271 0354-263
28 Jan 1995 0453-206 21 Feb 1996 0420-263 0446-206
1033-251 1006-214 0503-284
1257-253 1346-252 0938-205
1354-251 22 Feb 1996 0354-263 1048-238
29 Jan 1995 0344-291 0938-205 1103-244
1056-272 1303-215 1126-246
1126-246 23 Feb 1996 0938-205 1215-215
1343-253 1329-257 1323-271
1354-251 1344-241
1346-252
y: TEK#1, TEK#5, IRCAM are the CCDs and the IR camera used at LCO.
Details are in Table 2.6.

22
 Table 2.5: Filters used
Filter Bandwidth Central wavelength
micron micron
Visible Johnson's B 0.10 0.44
Cousin's R 0.10 0.65
NIR K 0 0.33 2.16
K 0 is a varient of Johnson's K .

Table 2.6: Detectors used

Detector Size Read noise Pixel size Scale
type pixel  pixel electrons microns arcsec/pixel
Visible TEK#1 CCDa 10241024 8 24 0.7
TEK#5 CCD 20482048
b 8 24 0.7
NIR HgCdTe array 256256 40c 40 0.348
a UV enhanced AR back coating
b Thinned
c For one double correlated read, with no endpoint averaging

23
60 and 20 minutes respectively. The standard stars were observed in focus
with exposures of few to several seconds, and also out of focus with longer
exposures so that the bright peak gets distributed into a torus. This is useful
for standard stars since one is interested in total magnitudes and not in radial
proles of the stars. The longer exposures also help avoid the eect of false
structures being introduced in the image due to the opening and closing of
the shutter. The stars were also moved about in the CCD frame.
The procedure adopted was slightly dierent in the case of IR observa-
tions. The sky is brighter in IR so that exposures have to be much shorter to
avoid saturation. Exposures of 35 seconds were therefore used. Seven such
exposures were taken with the object at the same point within the array.
The telescope was then moved a little so that the object shifted within the
array and seven more exposures were taken. Several such sets were taken
per object. In case the object covered a large part of the IR array, separate
sky frame sets were taken between every two object sets. The dark current
for the IR detector is not negligible and every day a large number of dark
exposures of durations 2, 3, 5 and 35 seconds were taken.
The K 0 standards used for standardization were HD 38921, Gl 347a and
HD 161743 from Elias (1982). These were observed several times at dierent
airmasses throughout the night.

2.3 Calibration
We wish to obtain surface brightness of each galaxy, in standard magnitudes
per arcsec2 , as a function of position. Obtaining this from the raw data,
which is in the form of CCD counts per pixel in arbitrary units, involves
several steps.
1. Removing the signatures of the telescope and instruments from the
data.

24
 2. Conversion of CCD counts to instrumental magnitudes.
3. Removing eects of the atmosphere.
4. Conversion of instrumental magnitudes to standard magnitudes.
5. Accounting for absorption in our Galaxy.
6. Transforming the observed magnitudes in xed lter passbands to galaxy
rest frame magnitudes in the same passbands.
The preprocessing, standardization and corrections necessary for this will be
discussed in this section.

2.3.1 Preprocessing
Photons incident on a photodetector give rise to electrons. These electrons
can then be transferred to an output stage and counted. This is the principle
used in a Charge Coupled Device (CCD). A CCD is a 2-dimensional array of
photodetectors (picture elements, or pixels). The basic functions of the CCD
include injection of charges, their storage, their transfer and their readout.
The CCD is exposed to incident illumination for a period of time and then
the accumulated charge, proportional to the incident light, is transferred to
the output.
For a given pixel, (i; j ), the total number of electrons, Nij , is a sum of
three components.
Nij = bij + dij + qij  sij (2.1)
where, sij is the incident signal, qij is the quantum eciency of the (i; j )th
pixel of the CCD, bij is the bias applied to the CCD, and dij is the dark noise
(thermal electrons).
Several preprocessing steps have to be carried out in order to obtain a
single clean image per lter per object. These are described below. Tasks

25
from IRAF1 and STSDAS2 were used for all basic optical reductions. For IR
reductions, command language scripts of IRAF and STSDAS developed by
S. E. Persson (Observatories of the Carnegie Institute of Washington), were
used.

Dark Subtraction The dark current is the dij component in Equation 2.1.
Electrons that constitute the count are excited into the conduction band by
photons. However, thermal energy can also excite electrons into the conduc-
tion band. The resulting thermal electrons constitute the undesirable dark
current, which makes a contribution even when there is no incident light. In
modern CCD detectors, dark current is negligible when the CCD is oper-
ated at a temperature near 150 K and can be ignored. However, the NIR
detectors are made of semi-conductor material that has a smaller energy gap
between the valence and the conduction band and hence these devices have to
be operated at much lower temperatures (e. g. 70 K for a HgCdTe detector).
Consequently, the dark current is sensitive to thermal background radiation
from the ambient environment.
If one does not have dark frames for the exact duration for which an object
has been observed in the NIR (e. g. 30 sec), dark frames of dierent durations
have to be interpolated to obtain the exact dark current corresponding to
the object's exposure. Keeping that in mind, several tens of dark frames of
duration 2, 3, 5 and 35 seconds were obtained everyday during the NIR run.
The at eld and object exposures were chosen from the above set so that
no interpolation was required.
1 IRAF is distributed by the National Optical Astronomy Observatories, which is oper-
ated by the Association of Universities, Inc. (AURA) under cooperative agreement with
the National Science Foundation.
2 The Space Telescope Science Data Analysis System STSDAS is distributed by the

Space Telescope Science Institute.

26
Bias Subtraction The bias is the bij component in Equation 2.1. The
bias is the pedestal level (DC oset) of the CCD introduced at the time of
readout. Bias subtraction is done either by calculating the bias from periodi-
cally taken special bias frames (e. g. Jedrzejewski, 1987) or from the overscan
region (e. g. Peletier et al. , 1990). The overscan is a series of readouts of the
CCD amplier with no exposure. It is a convenient provision for reading o
the oset level. The main advantages, in using an overscan region for bias
subtraction, are (1) no separate bias frames are needed and (2) the bias level
obtained is closer in time to the actual observation than a separate frame can
provide. At LCO the overscan region was used for bias subtraction during
both the runs. A 11-pixel wide region served this purpose. Though the bias
was found to be constant throughout the night, separate bias frames were
taken at the beginning and end of the night as also once or twice during the
night. If at any point some problem was suspected with the chip, additional
bias frames were taken and inspected.
The bias is usually constant over the chip. If it does vary a little from
pixel to pixel the variation is generally a slight function of the row number
i. e., it varies along the columns. The overscan regions spans a few columns.
A median column is obtained by extracting the median value of each row.
This column is then t as a function of the row number. This t, which is
often a constant, is then subtracted from each column of the image. The
image is then trimmed to get rid of the overscan region. All this was done
using the IRAF task colbias 3.
For the rst few runs colbias was applied interactively, the t was ex-
amined and it was found to be satisfactory. Thereafter the procedure was
applied automatically to each frame.
3 IRAF and STSDAS tasks will be denoted by italics and when the context is clear, it
will not be mentioned separately that these are IRAF or STSDAS tasks. Italics are of
course used in other contexts as well.

27
Flat elding Flat elding involves accounting for the qij component in
Equation 2.1. Unlike dark current and bias, this is a multiplicative eect
and comes in due to the dierent response that each pixel has to incident
radiation. Thus, the Quantum Eciency (QE) of a pixel is dened as the
probability that an electron in the pixel will be excited at the incidence of a
photon, i. e.,
average number of detected photons
QE =
average number of incident photons
: (2.2)
The QE for a pixel, qij , thus is in the interval [0,1]. This is dierent for each
pixel in a CCD, since each pixel is an individual semiconductor unit and has
its own characteristics, dependent on factors like thickness which cannot be
fully controlled. Other external factors like vignetting contribute to the CCD
response being non-uniform. Thus if the CCD is illuminated by a uniform
signal sij , which is the same over all the pixels, the output is qij  sij , which
varies from pixel to pixel. To correct for this non-uniform response one uses
the technique of at elding, which is described below.
Let the CCD be exposed to a uniform source of illumination f . The count
in the ij th pixel is
fij = qij  f (2.3)
Now let the CCD be exposed to a source which provides sij photons at the
ij th pixel. The count in the pixel is then
scij = qij  sij : (2.4)
From Equations 2.3 and 2.4 we get
sc
sij = fijc  f: (2.5)
ij
We have from this operation recovered sij , from the observed quantities
scij and fij , each of which is similarly aected by the pixel-to-pixel non-
uniformity. The constant f is not known, but is accounted for during the
28
preprocessing. The technique depends on having a uniform eld of illumi-
nation, with a suciently high count fijc , so that shot noise in the at eld
does not contribute signicantly to the overall noise in the signal.
Flat elds are of three types viz. , dome ats, twilight ats and dark sky
ats. Dome ats are obtained by uniformly exposing a at white surface on
the inside of the dome. Twilight ats are images of empty sky taken shortly
after sunset or before sunrise. Twilight ats are considered superior because
the wavelength dependence of the twilight is closer to that of the night sky.
For the same reason dark sky ats are sometimes obtained. Here the CCD
is exposed to an empty region of the night sky to obtain the pixel to pixel
variation. But the procedure requires spending a large amount of telescope
time which could otherwise be used for observing objects.
Another way of obtaining a at eld, without devoting extra telescope
time, is to get the median at. For this one needs a large number of images
taken over the night(s). One considers all the frames together and at each
pixel (i; j ) obtains the median of the (i; j )th pixel of all the frames. This
method proves eective when a given pixel is more likely to be not illuminated
by any source than it is likely to be illuminated by one.
The optimal count to obtain in each at eld is  10000. At this level
the CCD is in the linear range of the response curve, there is no saturation
of the pixels and the error in the count is kept down to  1%. By suitably
arranging several such frames to produce a master at, the noise can be
further reduced.
During our two runs over 50 twilight ats were obtained in each lter.
The exposures were chosen so that the counts reached 40-70% of the CCD
saturation. The minimum exposure used for the at elds was three seconds
so that the shutter's opening and closing does not give rise to structure
in the at frames due to non-uniform exposure. No gradient was seen in
the at elds. The at elds were combined using atcombine taking into
consideration the dierent exposure times and a normalized master at was

29
Figure 2.3: A cut across atelded eld of 0446-206 R. The high quality
of atelding is demonstrated by the at line. The lower panel shows a
magnied version of the same cut. The peak near line 200 is due to a star.

30
created in each lter. To test its accuracy, a at eld not used in making the
master at was divided by the master at. Variations in the mean count,
over a small area, across the attened image were noted, and were found
to be less than 1%. A cut across one such at elded image is reproduced
in Figure 2.3. Each object frame was divided by the master at for the
appropriate lter.

Cosmic Ray Removal Cosmic rays can interact with the CCD semicon-
ductor substrate to produce a large amount of electrical charge. Often such
an event is characterized by being localized at a single pixel on the CCD
frame. However, at times, the incidence can be of a grazing nature and sev-
eral adjacent pixels can be aected. Similar events giving rise to spurious
counts can also be generated within the instrument itself by either the elec-
tronics, or some radioactive material like the CCD window glass (Buil, 1991).
Since isolated pixels are aected, it is easy to detect the isolated peaks in
intensity.
Cosmicrays is used to remove the cosmic rays. To begin with, a window
size, e. g. 5  5 pixels, is chosen and the window is slid across the image.
Intensity peaks in the window are detected and their intensity is compared
with the mean intensity of the remaining window. If the ratio of the peak
intensity to mean intensity exceeds a threshold, say ve, the pixel is marked
as a cosmic ray. The window size and threshold depend on the point spread
function (PSF) and the exposure. The task is run interactively the rst few
times to determine 'reasonable' values for the threshold and window size. A
second pass is made to detect cosmic ray events in neighboring pixels and
for grazing incidence.

Sky Subtraction When a CCD is exposed to a region of night sky seem-

ingly devoid of discrete objects, it still registers non-zero counts coming from
the sky and monotonically increasing with the exposure duration. On a

31
starry, moonless night, at a good site, the sky can be as bright as 22B . The
sky brightness is the result of the following factors (Dube et al. , 1972): (1)
photochemical processes in the upper atmosphere. This component has an
irregular spectrum, (2) Particles in the solar system scattering sunlight and
giving rise to zodiacal light, (3) Faint and unresolved stars in our Galaxy
and (4) Diuse extragalactic light from unresolved galaxies. The standard
deviation of the background becomes an important factor in determining the
validity of the detection of a feature. Thus only when the count in an object
exceeds (3 ? 5)  sky , say, can one consider the feature to be real. The sky
background aects extended images the most. At every point of the image,
one overestimates the count equivalent to the background, resulting in a gross
error if the background is not accounted for. As a result, sky subtraction is
a very vital factor in all photometry and various methods have been devised
for measuring the sky level in an image.
When a galaxy image is small compared to the CCD frame, the \boxes
method" is preferred (e. g. Peletier et al. , 1990). In this method the average
count in boxes in dierent corners of the CCD, well removed from stars and
galaxies, is obtained. The MRC objects are several arcsec across. The CCD
eld was at least 110 for all observations. Hence there was plenty of empty
region for sky estimation within the eld, well removed from all sources. For
each frame over 20 boxes of size 5  5 pixels were used for sky estimation.
The mean of the medians in these boxes was accepted as the sky value. The
estimated accuracy is  1 percent.
If the galaxy image is large, regions free of the image may not be available
for sky subtraction. One then needs to apply a correction to the initial
boxes estimate. This is done iteratively by (1) subtracting the estimated
sky from the frame and (2) determining the luminosity prole at outer and
intermediate values to see if the two fall similarly. If the outer prole falls
too steeply, the sky has been over subtracted. If it falls slower than the
intermediate prole, the sky has not been fully subtracted. In either case,

32
the sky estimate is updated and the process repeated. The accuracy expected
using this method is  2 percent. A polynomial too can be tted to the sky
for an accurate estimation (e. g. Fasano et al. , 1996).
The sky in the NIR is much brighter (> 100 times) compared to the
optical sky. Absolute sky uctuations are also greater. Air glow and tem-
perature uctuations can lead to a sky level variation of factor 2 during the
night. Accurate photometry at fainter levels is thus more dicult in the NIR.
Additionally, large format NIR arrays have not been developed yet. Hence,
except in the case of very small galaxies (few arcsec) one cannot reliably use
the boxes technique.
The \sandwiching technique" is used to estimate the sky background
when the galaxy covers over a quarter of the array. In this technique an
exposure of the galaxy is sandwiched between two exposures of an empty
sky. One starts by pointing the telescope to a region close to the galaxy
which is aected neither by the galaxy nor by any other bright source. The
length of the sky exposure is kept the same as that of the galaxy frame. The
galaxy is observed next. It is again followed by a sky exposure of an equal
duration. The sky is then estimated from the two sky images.
For all our objects we carefully determined the sky from the galaxy frames
or from the sandwiched separate sky frames.

Aligning multiple frames The longer the total exposure is, the better
the signal-to-noise will be. However, because of the contribution of the sky
and the fast saturating bright stars, very long exposures are not advisable.
The solution is to split the exposure into two or more shorter exposures. It
is very rare that the images will turn out to be perfectly aligned in successive
exposures. In fact, to avoid bad pixels an object is often moved around in the
CCD eld during dierent exposures. As a result, one needs to realign the
images. In general, alignment requires translation and rotation of frames.
Alignment is accomplished to an accuracy of 0.1 pixels or better using

33
geomap and geotran. The procedure is as follows: (1) In each lter we have
several frames per object. One of these frames is taken to be the reference
frame for all the lters. (2) 3-4 bright, unsaturated stars well distributed
around the program object are chosen as reference stars and their frame
coordinates are obtained using imexam to an accuracy of 0.01 pixels. (3)
Coordinates of the same stars are located in one of the other frames. (4)
Geotran is used to obtain the coordinate dierences between the two frames.
The task uses the coordinates of all the reference stars chosen. (5) Geomap
is then used to map the chosen frame onto the reference frame. (6) The
procedure is repeated for all the frames, in all the lters, for the given object.
At the end of the procedure we have a set of frames in which corresponding
objects have the same frame coordinates within  0:1 pixel. The frames are
ready to be combined in any manner required for improving the signal-to-
noise ratio, obtaining color maps etc.

Combining multiple frames Imcombine is used to combine aligned, mul-

tiple frames of objects. The task allows weighing the frames by their expo-
sure. Thus a frame with a longer exposure gets a higher weight. If there are
a large number of frames, a rejection criterion too can be adopted whereby
the outliers are ignored during the nal averaging. We were able to use the
rejection criterion in cases where the galaxy was close to a bright star and
hence several very short exposures had to be taken in each lter. In the end
there exists a single image per lter per object.

2.3.2 Standardization
After the preprocessing steps the data have to be converted to a standard
magnitude system, like the Johnson-Morgan B, V, R system, so that they
can be compared with the observations carried out by others. To obtain
the required transformation coecients, from instrumental to standard mag-
nitudes, one observes several standard stars a few times during each night
34
along with the program objects. These stars are non-variable and have been
standardized by repeated observations with various telescopes over a period
of time and their magnitudes in dierent standard passbands are known to
 0:001 magnitude. One then measures the eect of conditions of the at-
mosphere on these standards and uses the information to correct for the
magnitudes of the program objects.
The procedure is to obtain instrumental magnitudes for a number of
stars, and then to derive the transformations from instrumental to standard
magnitudes. To rst order, the transformation involves the color of the stars
(see Henden and Kaitchuck, 1982).
A large format CCD has the advantage that several stars can be simulta-
neously observed. Thus, if a region containing many standard stars is chosen,
a larger number of data points are obtained for the standardization. During
the two runs, two elds from the Standard Areas (Landolt, 1992) viz. , SA 98
and SA 104 were used. The standard stars observed are listed in Table 2.7.
The CCD counts for the stars were converted to instrumental magnitudes
by using
b = ?2:5 log Ib ;
r = ?2:5 log Ir ; (2.6)
where, Ib and Ir refer to the total number of counts per second from the
individual stars in the B and R lters respectively. Since the absorption and
diusion in the earth's atmosphere aect the light, the resulting atmospheric
extinction has to be removed by transforming the instrumental magnitudes
to their extra-atmospheric values. The amount of extinction depends on the
altitude of the observatory, atmospheric conditions, wavelength used and the
zenith distance of the object. The zenith distance, z, is given by
sec z = 1=(sin sin  + cos cos  cos h); (2.7)
where, is the latitude of the observatory,  is the declination of the object,
and h is the object's hour angle.
35
 Table 2.7: Broad band standard stars observed
(2000) (2000) B B ? R
SA 98 666 06h52m10s -00o2303400 12.896 0.255
671 06h52m12s -00o1802200 13.453 1.543
670 06h52m12s -00o1901700 12.286 2.079
676 06h52m14s -00o1902100 14.214 1.829
675 06h52m14s -00o1904100 15.307 2.991
685 06h52m19s -00o2001900 12.417 0.753
SA 104 456 12h42m54s -00o3200600 12.984 0.979
457 12h42m54s -00o2804900 16.801 1.237
460 12h43m03s -00o2802100 14.173 2.100
461 12h43m07s -00o3202100 10.181 0.765
350 12h43m15s -00o3302100 14.307 1.056
470 12h43m22s -00o2905600 15.042 1.027

36
 The term sec z, often denoted by , is called the air mass. It denotes the
thickness of the atmosphere crossed by the light rays. It is dened to be unity
at the zenith and increases to innity at the horizon. The extra-atmospheric
magnitudes, b0 and r0 , are obtained from the instrumental magnitudes by
making use of the zenith distance of the object and the extinction factor in
magnitudes at the zenith. Thus,
b0 = b ? kb ;
r0 = r ? kr ; (2.8)
where,  is the airmass, b and r are the instrumental magnitudes, and kb and
kr are the extinction coecients in b and r lters respectively. The standard
stars are observed for a range of air-masses. The values of kb and kr are
obtained by tting a straight line to the observed magnitudes versus the
airmass. The slope gives the extinction factors whereas the intercept gives
the extra-atmospheric magnitude (magnitude at  = 0).
Though it is customary to refer to the k term above as the extinction
coecient for the lter , it is actually a composite of two factors: (1) a zero-
point, k0 , which is an extinction term that depends only on the transparency
of the atmosphere. It changes from night to night and (2) a color coecient,
k00, which is a color dependent term arising from the dierent response that
broadband lters have to objects with dierent colors. For a given site, this
term is generally constant for a season for a given combination of lters.
Thus,
K = K0 + K00 (B ? R): (2.9)
The transformations to the standard B and R are given by
R = r0 + (B ? R) + r ;
B ? R = (b ? r)0 + br ; (2.10)
where,  and  are the color coecients and r and br are the zero-points.

37
 Table 2.8: NIR standard stars observed
star (1950) (1950) J K L
h m s 0
HD 38921 05 45 41.0 -38 deg14 51" 7.570 7.535 7.525
Gl 347a 09h26m25.0s -07 deg08030" 8.410 7.630 7.425
HD 161743 17h45m31.8s -38 deg-6 011" 7.620 7.615 7.605

We followed the procedure outlined above to obtain the extinction cor-

rection and transformation coecients.
For NIR calibration, standards from Elias (1982) were observed in J
and K 0 bands and are tabulated in Table 2.8. The stars are bright (7th
magnitude) and hence they were always observed out-of-focus so as to form
a torus. This avoids saturating the array and allows one to extract the total
magnitude. Each standard was observed several times each night. During
each round, the star was rst placed in the rst quadrant of the IR array and
seven exposures obtained. Then it was similarly observed in the second, third
and forth quadrants. The sky too was observed by moving the standard out
of the frame and making sure no other bright star was in the frame. Since
the sky in NIR can change fast, it is very important to observe the sky close
to the standard.
The NIR standardization is slightly dierent from the broad band proce-
dures. First a nominal sky frame is created for each wavelength observed. A
note is made of all the contaminating stars that may be present in the im-
ages. These are then removed and the sky is estimated to a better accuracy
and subtracted out. The magnitudes are then extracted and using the array
sensitivity values and the standard magnitudes, the instrumental constant
for each night is calculated.

38
2.3.3 Correction for Galactic Extinction
The absorption due to our galaxy depends on the direction of the line of
sight through the Galaxy. It is maximum in the plane of the galaxy and
least towards the poles. We adopted an absorption free polar-cap model
(e. g. Sandage, 1973) to calculate the galactic absorption. It is given in mag-
nitudes by
Av = 0:1(cosec jbj ? 1); jbj  500;
Av = 0; jbj > 500: (2.11)
All the sample objects have jbj  520 and therefore we always used the
expression for jbj  500. Extinction in the B and R bands is obtained from
Ab = 1:33Av ;
Ar = 0:75Av : (2.12)

2.3.4 K-correction
When galaxies at dierent redshifts are seen through a xed passband, the
collected light comes from dierent wavelength ranges in the restframe of the
respective galaxies. K -correction is used to transform the restframe magni-
tudes of each object to the passband of the lter. The K -correction depends
on the spectrum of the object, and is given by (Oke and Sandage, 1968)
2 R1 3
S (
K (z)  2:5 log(1 + z) + 2:5 log 4 R 10   ) R ( ) d 5; (2.13)
0 S 1+z R ( ) d
where, S () is the energy ux density in the reference frame of the galaxy and
R() is the response function of the lter used. The K -corrected magnitude,
mc, is then obtained from the uncorrected magnitude m using
mc = m + K: (2.14)

39
 We have used an approximate form of the K -correction, applicable for
small z, from Whitford (1971) and Persson (1979). For the three lters we
have,
KK 0 = ?3z;
KB = 5z;
KR = z: (2.15)
The magnitudes corrected in accordance with this approximate form have
been used for subsequent correlations and calculations.

2.4 Control Sample

Powerful radio sources have been believed to be hosted by early-type galaxies.
One of the main aims of this thesis is to test this claim in the face of numerous
examples of morphological distortions found in radio galaxy hosts following
better observations. The aim is to see if there exist any systematic dierences
between radio galaxy hosts and normal galaxies.
We have compared our results on the MRC radio galaxies with other
radio samples and with normal ellipticals mentioned in the literature. In
addition, we have carried out exhaustive comparison with a sample of early-
type galaxies extracted from the radio galaxy elds and processed in a fashion
identical to the radio galaxies. We detail below the procedure used to extract
this sample and the advantages of having such a comparison sample.

Selecting the Comparison Sample

The radio galaxies have angular sizes which range from several arcsec to a
few 10s of arcsec. However, the CCD used in 1995 was 110  110 and that
used in 1996 was 230  230. As a result several other galaxies, besides the
object of primary interest, are present in each frame. It is from these images

40
that the control sample has been chosen. The advantages of such a selection
are:
1. The control objects are observed using the same telescope, lters and
CCD as the radio sample. As a result, during at elding etc. the
instrument signature removal is automatically identical to that for the
radio galaxies.
2. The objects are observed on the same nights as the program objects,
thereby allowing one to use exactly the same extinction corrections.
3. The accuracy of the various reduction techniques used is same both for
the comparison sample and the actual sample.
4. Since the raw data is available just as in the case of the actual sam-
ple, everything that is done for the actual sample can be done for the
comparison sample.
We selected from these galaxies, for the comparison sample, those galaxies
which have a semi-major axis length of at least 15 arcsec in both the B and R
bands. To decide this objectively we went through the following steps: (1)
determine the background level and its , (2) subtract the background, (3)
set points having a count below 3 to zero and (4) display the frame.
The resulting set contained 82 galaxies including several program objects.
Classication by eye was done to group the objects into ve classes viz.
E/S0, spiral, multiple (including dumbbell), distorted, and marginal. The
marginal included galaxies that seemed to be bulge dominant but had at least
one feature that looked either like a spiral arm or a tidal interaction. The
bulge disk decomposition program was run on all these galaxies to obtain
a quantitative measure of the various parameters. The resulting numbers
were used to test the classication by eye. It was found that in eight cases
the two diered. These included two ellipticals, four edge-on spirals and
two marginal cases. The program is known not to be able to handle edge-on
41
cases where the disk dominates and the bulge and disk ellipticities tend to be
very dierent. In the two marginal cases the program detected a dominant
disk (D=B > 3). One of the galaxies put into the class of ellipticals was
reclassied as a marginal case as it did have some distorting features in the
form of faint spiral arms. The other elliptical galaxy where the eye and the
program diered in judgment had a kink at the point where ellipse tting was
begun. Changing the starting point made the program give proper elliptical
parameters. Finally, only galaxies that were classied as E/S0 galaxies by
the program as well as the eye were chosen as member of the control sample.
These numbered 30, coincidentally the same as the number in the radio
sample.
The comparison sample thus consists of bulge-dominant galaxies having
semi-major axis length > 10 arcsec.

42
Chapter 3
Brightness distribution
3.1 Introduction
Galaxies have a wide range of appearance. Even the appearance of the
same galaxy at dierent wavelengths can be quiet dierent. Nevertheless,
galaxies of regular shape can be broadly divided into two types viz. , ellipticals
and spirals. Galaxies that possess no obvious symmetry are called irregular
galaxies. We describe here certain morphological properties of these dierent
type of galaxies and also of the intermediate classes. The following discussion
will be best understood with the help of Figure 3.1.
Elliptical galaxies appear smooth and structureless and have elliptical
isophotes. These galaxies are denoted by En, with n = 10[1 ? (b=a)], where
a and b are the major- and minor-axis respectively of the projected image.
Thus E0 galaxies appear round whereas E6 galaxies are among the most
elongated galaxies observed. Ellipticals on the whole are not very attened.
The intrinsic shape of elliptical galaxies could either be spheroidal or triaxial.
It was believed at one point that the moderate attening is a result of
rotation of the galaxy about a preferred axis. However, following the rotation
measurements of elliptical galaxies we now know that the rotation is not suf-
cient to support the galaxy and that the support comes from the anisotropy

43
 normal spirals
Sc

Sb

E0 E3 E6 S0 Sa

SB0 SBa

ellipticals
SBb

SBc

barred spirals

Figure 3.1: Hubble's tuning-fork diagram.

44
in the random velocity of stars. The nding has also lead to the suggestion
that most elliptical galaxies are triaxial in nature i. e., the ratios between the
three axes change as a function of distance from the center. Isophotal twists
observed in many ellipticals further supports this idea. Isophotes are equal
surface brightness contours and, if a galaxy is axisymmetric, the isophotes
will be seen to have the same major-axis even if viewed from any direction.
Spiral galaxies, as the name suggests, have obvious spiral features, or
arms. Spiral galaxies comprise of a central brightness condensation resem-
bling an elliptical, called the bulge, and a thin disk. It is this disk that
contains the spiral arms. In addition a bar may be present interior to the
spiral arms. Depending on the presence of a bar, a spiral is classied as a
barred (SB) or an unbarred or normal spiral (S). Spiral galaxies are further
subclassied depending on how tightly wound the spiral arms are. Sa galax-
ies have their arms tightly wound and have conspicuous bulges whereas Sc
galaxies have open arms.
Intermediate between the ellipticals and spirals are the lenticular galaxies,
denoted by S0. These galaxies have a smooth central brightness condensation
surrounded by a large intrinsically at structure. This latter component is
like the disk component of spirals, but lacks spiral arms. Like a spiral, a
lenticular may (SB0) or may not (S0) have a bar.
Irregular galaxies can be broadly subdivided into IrrI and IrrII galaxies.
IrrII have an irregular mass as well as light distribution. IrrI galaxies have
their mass much more symmetrically distributed than IrrII, but owing to the
irregular light distribution, still look irregular. Recent close encounter with
one or more nearby galaxies is likely to be the cause for the irregularity in
these galaxies.
In any of the above type of galaxies an active galactic nucleus (AGN)
may also be present. An AGN is a point-like source at the center of a galaxy,
which in extreme cases can outshine the rest of the galaxy. We will discuss
AGN in detail below.

45
 Hubble's tuning fork diagram (Figure 3.1) summarizes the belief once
held that the two type of galaxies form an evolutionary sequence with round
ellipticals evolving into spirals. We still believe that ellipticals are more
closely associated with lenticulars that with spirals, and that barred spirals
are more closely associated with each other than with unbarred spirals. In
the above sense Hubble's diagram continues to remain valid. However the
interpretation that elliptical galaxies evolve into spirals is now known not to
be true. Ellipticals, for instance, are known to have the highest and lowest
masses among galaxies, with spirals galaxies having a small intermediate
range of masses. Also, ellipticals possess small quantities of gas compared
to spiral galaxies. In fact, if anything, we now believe that it is the spirals
that merge and give rise to ellipticals. This is borne out by studies and
simulations of interacting galaxies (Schweizer, 1982; Barnes and Hernquist,
1992) as well as the observation that the concentration of ellipticals is more
in regions where the galaxy density is greater.
About 1% of galaxies have appreciable emission at radio wavelengths
(radio luminosity exceeding 1022 W Hz?1). These galaxies are termed radio
galaxies. The morphology of the radio galaxies as captured by photographic
plates in early observations was seen to be elliptical, and it is even now
believed that radio galaxies are elliptical in nature (e. g. Spinrad et al. , 1985).
However, recent high dynamic range images of radio galaxies have revealed
that several of these galaxies, though basically elliptical in morphology, are
distorted to dierent extent. On the other hand, it has also been found
that almost all galaxies once classied as ellipticals do contain variable, but
essentially non-zero, amounts of dust and gas, particularly the former, along
with signs of clear departures from a classical elliptical. Dust is often visible
in the form of a lane (e.g. Kotanyi and Ekers, 1979, Hawarden et al. , 1981,
Mahabal et al., 1996). The source of distortion can also be dynamical in
nature, e. g. , a stellar disk. (Gonzalez-Serrano, 1989).
It has also been known that many ellipticals possess other interesting

46
features like shells (Forbes and Thomson, 1992; Sahu et al. , 1996), infalling
HI gas (Van Gorkom et al. , 1989) and secondary nuclei (Forbes et al. , 1994).
The possible reasons for the distortions include interactions, mergers, radio
jets and cooling ows. In the earlier classication, a galaxy having even a
faint disk was classied as an S0, a class intermediate between ellipticals and
spirals. However, better instruments and larger telescopes have revealed that
some galaxies originally classied as ellipticals contain a disk as well.
The dierent morphological components are important pointers towards
the history of the galaxy, and hence the process of galaxy formation and
evolution in general. For a physical understanding it becomes important to
quantify the various components present in dierent galaxies and classify
the galaxies in accordance with the nature of the components. For that
purpose it is necessary to express the structure of a galaxy using a small set
of parameters. One can then describe the deviations in individual galaxies
as being over and above the range of parameters for the \standard" set. One
way of doing this is to obtain a radial brightness prole1 of the galaxy i. e.,
how the galaxy brightness is distributed as a function of distance from the
center. One can then attempt to describe the radial distribution in terms of a
simple model with a few parameters, which describe the overall morphology
of the galaxy.
In the rest of this chapter we elaborate on the process of obtaining an
intensity prole by tting ellipses to galaxy isophotes. We discuss the dier-
ent components of a galaxy at length along with the various tting laws that
are used to separate these components. The program that we use for de-
composing a prole into its components will be described along with several
tests on articial galaxies as well as control ellipticals. We will also describe
a two-dimensional (2D) decomposition technique and compare it with the
one-dimensional (1D) technique.
1 The radial brightness prole will often be simply referred to as the prole hereafter

47
3.2 The radial prole
3.2.1 Introduction
Historically, an intensity prole was obtained by simply taking a cut along
the major axis of the galaxy or along a few selected axes. With the advent of
computers various routines that can average light along isophotes have been
devised and used in studying the light distribution.
Throughout, it is important to remember that galaxies are three-dimensional
(3D) objects and what is observed is just their 2D projections on the sky.
Thus when a prole is obtained, the 2D distribution is being further averaged
to a 1D distribution leading to loss of information. Due to the regularities on
a large scale in galaxies it is often enough to consider the 1D prole to obtain
a fair idea of the light distribution. However, when dealing with particular
morphological distortions and localized perturbations, it is better to consider
the 2D distributions.
Isophotes of galaxies2 are well approximated by ellipses. One can use this
prior knowledge to obtain a radial brightness prole. One can thus t ellipses
to the isophotes and obtain mean intensity as a function of radius. Besides
the mean intensity, one also obtains the radial proles of ellipticity, position
angle etc., each with its errors. All these parameters can be used to obtain
some idea regarding the 3D shape of the underlying galaxy. A set of Fourier
components A3; B3; A4 ; B4 (described below) obtained are indicative of the
deviations of the shapes of the tted isophote from true ellipses. These have
been used over the years to argue over the presence of dust, disks etc. The
present section details the tting of ellipses, the precautions to be taken, the
information that the resulting parameters reveal and the dependability of
the errors involved.
Unless specied otherwise, terms like isophotes of a galaxy will mean isophotes of a
2

2D projected galaxy

48
3.2.2 Obtaining the t
The 2D CCD image obtained is the projection on to the plane of the sky of a
3D galaxy. As mentioned previously, the major components of a galaxy are
the spheroidal bulge and a thin disk. In projection, either component gives
rise to elliptical isophotes. The two components taken together will also yield
elliptical isophotes, with the isophotes containing information from both the
components. Measuring various parameters of these projected ellipses is an
important step in understanding the brightness distribution of a galaxy. In
this subsection we describe in detail the procedure of tting ellipses to galaxy
isophotes.
Any ellipse can be fully described by ve parameters viz. semi-major axis
length, position angle, ellipticity and the coordinates of the center, xc ; yc.
The aim here is to t to the isophotes a series of ellipses with dierent semi-
major axis lengths. An ellipse with a given semi-major axis will then have an
ellipticity (), a position angle (PA) and an x- and y-center. The ellipticity
is given  = 1 ? b=a where a and b are the major- and minor-axis lengths
respectively. PA is the angle (0o < PA < 180o) made by the major-axis of
the galaxy anticlockwise with the +ve x-axis. The peak central intensity of
the galaxy denes xc; yc.
For a smooth and featureless spheroidal galaxy the projected ellipses are
concentric with the same ellipticity, PA, xc and yc. A straightforward way of
tting an ellipse to an individual isophote is to use (x; y) pairs on the isophote
and to use some method, like least squares, to obtain the ellipse parameters.
However, only a nite number of points are used in this procedure.
We have used the IRAF task ellipse which follows the procedure outlined
by Jedrzejewski (1987; See also Young, 1976; Kent, 1984). To begin with,
one chooses a trial ellipse at a given semi-major axis length. The intensity
along the trial ellipse changes as one goes along it. The change in intensity

49
is periodic and can hence be expanded in a Fourier series.
X
1
I = I0 + (Ansin(nE ) + Bncos(nE )) ; (3.1)
n=1
where I0 is the mean intensity along the ellipse, E is the ellipse eccentric
anomaly and An; Bn(n = 1; 2; :::) are harmonic amplitudes.
If the isophotes are ideal ellipses, only the rst four coecients viz.
A1; A2 ; B1; B2 are non-zero. All the higher coecients are zero. For a real
galaxy, the isophotes are rarely perfectly elliptical. One however begins by
assuming that they are elliptical and a trial ellipse is obtained. A1 ; A2; B1; B2
are expected to be non-zero. The values are obtained by a least square t to
I = I0 + A1sin(E ) + B1 cos(E ) + A2sin(2E ) + B2cos(2E ): (3.2)
Each of the four coecients is related to one ellipse parameter with the mag-
nitude of each coecient dependent on the misalignment between the trial
ellipse and the actual isophote. Correction factors to the trial ellipse are now
derived by changing the initial ellipse parameters. The parameter that cor-
responds to the largest amplitude is varied, and the image is resampled. The
procedure continues iteratively till one of the following criteria is satised:
 The magnitude of the largest amplitude has gone below a threshold or
 The maximum number of iterations have been performed.
At the end of the exercise, the center,  and PA are obtained for the best
tting ellipse along with the mean intensity, I0.
After the best tting parameters have been determined, third and fourth
harmonics are obtained using a least-squares t to
I = I0 + A3 sin(3E ) + B3 cos(3E ) + A4 sin(4E ) + B4cos(4E ): (3.3)
Here A3 ; A4; B3; B4 determine the deviations of the isophote from the perfect
elliptical shape. For a perfectly elliptical isophote, the coecients of these
50
Figure 3.2: 2% deviated ellipses. Deviations due to A4 (top) and B4 (bottom)
are shown. The 4-fold symmetry is clearly visible. Even in images of galaxies,
if the deviation is  1%, it is discernible to the eye.

higher harmonics will all be zero. When non-zero, the A3 and B3 compo-
nents indicate that the isophotes deviate from ellipses and that the deviation
has a 3-fold symmetry. Similarly, the presence of A4 and B4 denote 4-fold
symmetric deviations. Typically the deviations are less than  1%. Even
a deviation to this extent distorts the ellipse considerably (see Figure 3.2).
The B4 component is especially important since it indicates the presence of
a disky structure (B4 > 0) or that of boxiness (B4 < 0).
The procedure is repeated for the next semi-major axis length. For succes-
sive ellipses the semi-major axis is increased by, say, 10% over the previous

51
one. One continues in this manner till the desired outer limit is reached.
This is determined either by the size of the galaxy (one should typically have
more than 60% of the ellipse on the CCD frame) or by the noise levels (with
modern CCDs one can reliably go upto a few percent of the sky level).
The ellipticity and position angles (as also other parameters) are not very
reliable near the center. This is because the inner light gets distributed to
larger radii due to the PSF. Since the PSF is typically azimuthally symmetric,
this results in making the isophotes near the center circular.
As far as the complete galaxy image is concerned, it often contains regions
which one needs to ignore during the process of tting. Bad pixels are agged
so that they get a zero weight during tting. Stars can be masked and
interpolated over. Additionally upto 40% of the top pixels on a given isophote
can be masked to get rid of faint stars and other spurious features. This
ensures that unwanted objects do not aect the tted ellipses.

3.2.3 Procedure Adopted for the Sample

The images used for the work in this thesis have been taken in the B , R and
K 0 bands. The CCD eld in case of B and R and the camera eld in case
of K 0 were large enough to contain the whole galaxy within it. We stopped
the ellipse t when I=I0 exceeded 0.1, which corresponds to an error in
magnitude of
m  0:1. Often the ellipses were t well beyond this limit to
see how the prole progressed. However, during all subsequent analyses these
additional outer points were ignored. To look for a possibly varying center,
the center was always left free during ellipse tting. However, the typical
xc; yc shifts obtained were less than half a pixel. (This is another indication
that there was no articial gradient in the images. If there is a gradient,
the ellipse centers are seen to shift.) Only when we wanted to look at some
particular aspect of an individual galaxy that needed xing the ellipse centers
did we actually x it.
The tting was begun at a semi-major axis length of ve pixels. Suc-
52
cessive ellipses had major-axis length longer by ten percent and the tting
was continued until the 1 error in the mean intensity (I0 ) was less than
0:1  I0 corresponding to an error of
m  0:1 magnitude. Ellipses were
tted to isophotes smaller than the starting one at ve pixels, by reducing
the semi-major axis-length by 10% at every step. It is necessary to begin
at an intermediate semi-major axis length, and then move inwards, as the
parameter values obtained near the center are not reliable, due to the small
number of pixels involved in their measurement, as well as the modication
produced by the PSF. Using such starting values can have an adverse eect
on parameters derived for the larger isophotes. Clipping of 10% was used
to make the program ignore the top 10% pixels during the t. In some of
the cases a kink was present at 5 pixels from the center. In such a case we
started at a nearby point without the abnormality and proceeded as above.

3.3 The dierent components

In order to describe the 1D distributions we need simple model proles for
the dierent components. The bulge and disk components of spiral galaxies
have dierent prole shapes. The bulges have proles similar to ellipticals
while the disk component is exponential in nature. In this section we will
discuss a few simple model proles that are in use to describe the galaxy
components.

3.3.1 Ellipticals and bulges

Hubble's Law
The so called Hubble's Law is really due to Reynolds (1913). It bears the
name of Hubble because it was he (Hubble, 1930) who made extensive use

53
of it. According to this law the surface brightness (r) is given by

(r) = (0) ; (3.4)

(1 + r=r0)2
where r is the major axis.  stands for the surface brightness and (0) is
the surface brightness at r = 0. The scale length r0 is the distance at which
the surface brightness falls to a quarter of its central value.
Hubble's law describes very well the radial brightness distribution of most
ellipticals within certain bounds of r. However, the light contained within a
given r diverges logarithmically with increasing r. This certainly cannot be
true for real galaxies.

King's Proles

King (1966) introduced two dynamical models of self-gravitating stellar sys-
tems. These were successfully applied to globular clusters and later extended
to elliptical galaxies (King, 1982). A King prole can be described in terms
of the following parameters:
1. Central brightness, 0 .
2. Core radius, rc, at which the brightness is 0 =2.
3. Tidal radius, rt . The brightness vanishes at this point.
Generally the curves are labeled in terms of a quantity c = log (rt =rc). For
suciently small r, all King curves have a zero gradient. The model with
c = 1 has a slope of ?1 at large r and corresponds to the isothermal sphere.
King's proles are based on specic dynamical models, and provide good
ts to globular cluster surface brightness. However, recent studies (Lauer et
al. , 1995) have shown that these are not really suitable for elliptical galaxies.

54
de Vaucouleurs' law
de Vaucouleurs' law is perhaps the most widely used empirical law to describe
the surface brightness prole of an elliptical galaxy. Also known as the r1=4
law (de Vaucouleurs, 1953), it is given by
(r) = e expf?7:67[(r=re)1=4 ? 1]g: (3.5)
The scale length, re, known as the eective radius, contains half of the total
light of the galaxy. The total light B of the bulge is given by
Z1 Z re
B= (r) 2  r dr = 2 (r) 2  r dr = 7:22  re2 e: (3.6)
0 0
e is the surface brightness at r = re. de Vaucouleurs' law gives (0) 
2000 e. A plot of the logarithm of surface brightness versus r1=4 gives a
straight line if de Vaucouleurs' law is valid (see Figure 3.3).
de Vaucouleurs' law is empirical in nature. It does not necessarily t all
elliptical galaxies over all ranges of r. Burkert (1993) has shown that de
Vaucouleurs' law describes very well the brightness distribution in ellipticals
for 0:1 re  r  1:5 re.
Studies have shown that the r1=4 law could be the end result of the re-
laxation of merged galaxies. For example, NGC 7252 is a peculiar galaxy
with two large luminous tails. Schweizer (1982) has shown that the central
region of the radial luminosity prole of this galaxy, which is made of two
disk galaxies, obeys the r1=4 law. However, two purely stellar disks cannot
give rise to an elliptical galaxy. This is because stars in an elliptical galaxy
are more tightly packed than in a spiral. Hence, to form an elliptical, the
participation of gas which can pack during a merger is necessary.
For de Vaucouleurs' law the deprojected mass density (r), corresponding
to an r1=4 intensity distribution, is not analytically tractable (Young, 1976).
As a result, de Vaucouleurs' law is not very useful for theoretical analyses.
Attempts have hence been made to obtain density proles resembling that of
the r1=4 law and having simple analytic forms (e.g. Binney, 1982; Bertin and
55
Stiavelli, 1984, 1989 and Hernquist, 1990). However, these analytic forms are
not as convenient to use as de Vaucouleurs' law. Moreover, though empirical
in nature, de Vaucouleurs' law is seen to provide adequately good ts to
many ellipticals. The approach that has been taken in this thesis is to t de
Vaucouleurs' law to the observed proles, and to analyze the distribution of
the tted parameters as well as the deviations from the best-t law. As we
shall see, the latter provide important physical information.
de Vaucouleurs' law can be generalized by replacing the 1/4 by 1/n, where
n is a variable. The r1=n law is then given by
(r) = e10f?bn[(r=re )1=n ?1]g; (3.7)
with bn chosen so that half of the luminosity is contained within r < re. By
tting a large number of galaxy surface brightness proles with a generalized
de Vaucouleurs' law, it has been shown that dwarf galaxies have n < 4 while
luminous galaxies have n > 4. (Caon et al. , 1988; Graham et al. , 1996)

Petrosian's  function
Petrosian (1976) introduced another useful way of understanding the bulge.
He introduced a dimensionless monotonic function (r), which is the ratio of
the average surface brightness up to a radius r to the surface brightness at r.
It is scale-less and does not assume any particular cosmological model as it
involves only the structural parameters of the galaxy. It is useful for studies
involving, say, the dependence of surface brightness as a function of redshift
without reference to any particular cosmological model.

The Jae Prole
Jae (1983) proposed that elliptical galaxies have a density distribution given
by
(r) / r?2 (r + rJ )?2: (3.8)

56
Figure 3.3: NGC 661 is a good example of de Vaucouleurs' law prole. The
prole is a very good approximation of a straight line in the r1=4 ?  plane.
The curvature at small r is mostly due to the eect of the PSF.

57
Quantities like surface brightness, gravitational potential, projected velocity
dispersion etc. can be calculated easily from this law. Hence this prole is
widely used in numerical studies of spherical galaxies.

Dehnen models
Dehnen (1993) proposed a family of density proles for which the total lu-
minosity within the projected radius r does not diverge as r ! 1:
j (r) = 3 4? r (r +
Lra
ra )4? : (3.9)
Here L is the total luminosity of the model and ra is the scale length. This
form of j (r) yields an analytically tractable projected density when is an
integer or a half-integer. The models of Jae (1983) and Hernquist (1990)
described above correspond to = 2 and = 1 respectively. The closest
approximation to the r1=4 law is given by = 23 . The Dehnen models will
be discussed further when we consider the 3-D distribution of luminosity in
elliptical galaxies.

3.3.2 The disk

It has been found (Freeman, 1970; Kormendy, 1977) that the surface bright-
ness d (r) of the disk component of a spiral galaxy has an exponential form:
d (r) = (0) exp(?r=rs); (3.10)
where rs is the scale length. The total light in the disk is given by
Z1
D= (r)2rdr = 2rs2(s); (3.11)
0
where (s) = (r = rs).
The disk-to-bulge ratio,
D = 0:28  rs 2 (s) ; (3.12)
B re (e)
58
is also the ratio of the luminosities of the disk and the bulge. The ratio is
sometimes expressed in terms of the bulge-to-total light ratio,
B= 1
T (D=B + 1) : (3.13)

The disk becomes less and less prominent as one goes from spirals to lentic-
ulars to ellipticals (Kent, 1985). A galaxy classied as Sc can easily have a
D=B ratio of 4, while a typical S0 galaxy will have D=B > 0:3. An elliptical
galaxy, by extension, would be expected to have a D=B < 0:3. The disk in a
disk galaxy is likely to be made of an old disk, a young disk and a dust layer
(e. g. Wainscoat, 1989). Burstein (1979) showed for a sample of S0 galaxies
that they possess a thick disk and a thin disk in addition to a r1=4 bulge.
The orientation of the disk and the presence of dust complicates the accurate
determination of the components. The subtleties of the disk-bulge decom-
position will be described in the next section. Not much is known about the
faint disks being detected in galaxies originally classied as ellipticals.

3.3.3 The point source

A small fraction of galaxies (< 10%) are known to possess an Active Galactic
Nucleus (AGN). An AGN is a compact object at the center of the galaxy
and is believed to be powered by accretion of gas onto a supermassive black
hole at the center of the galaxy. AGN luminosities can reach 1048 erg sec?1
and can outshine the host galaxy by over a factor of  103. This factor is
highest for quasars and least for normal galaxies.
When present in a galaxy, an AGN corresponds to a sharp peak in the
surface brightness prole at r = 0. The extent of the AGN is small compared
to the PSF. However, the light from the AGN gets distributed upto a few
arcsec during convolution by the PSF. As a result, if the AGN is weak, it
is very dicult to reliably detect it from ground-based observations. AGNs
are readily detectable in Seyferts, but we nd that in radio galaxies they are

59
generally too weak to be extracted from our data. If such a source is present
it will be weak and indistinguishable from other compact features like a small
scale disk. Therefore we have not considered point sources during the prole
decomposition for the sample. In cases where there seem to be strong reasons
for believing that an AGN is present, additional processing was done to look
for the point source.

3.4 Decomposing the prole

In the absence of a point source, the prole can be viewed as a sum of the
bulge and disk components.
(r) = e expf?7:67[(r=re)1=4 ? 1]g + s exp(?r=rs): (3.14)
In order to t models to the observed prole, use has been made of a program
written by Ajit Kembhavi in standard Fortran 77. The program makes use
of the Minuit 94.1 (James, 1994) minimization package from CERN and
routines from Numerical Recipes (Press et al. , 1986). A detailed discussion
of the program follows.

Free Parameters The program ts a de Vaucouleurs prole to the bulge

and an exponential to the disk. The parameters used are:
1. re: The bulge scale length.
2. rs: The disk scale length.
3. D=B : The disk-to-bulge ratio.
4. P=B : The ratio of the luminosity of the point source to the bulge
luminosity. As stated earlier, for the sample, P=B was always xed at
zero.
5. e(0): The central magnitude.
60
 6. eb: The bulge ellipticity.
7. ed: The disk ellipticity.
The bulge and disk ellipticities are invisible parameters and the D=B
ratio appears as the relative luminosity of the disk and the bulge.
One can specify the range within which a specic parameter is to vary
and the step size. Alternately, one can hold a parameter xed at a particular
value if that value is known. If, for instance, one knows that a galaxy does
not have a disk component, D=B for that galaxy can be xed at zero.

Input parameters The brightness prole for which a best t is sought

forms the main input for the decomposition program. The prole contains
the surface brightness along with 1 errors on it as a function of r. The trial
values viz. re; rs; D=B; P=B; e(0); eb and ed are put together in a le which
we call the minimization le. The FWHM of the PSF, as extracted from the
input galaxy image is also included in the minimization le. Finally, the le
also contains information as to which points from the input prole are to be
actually used during a given run. This facility helps us exclude pathological
parts of an image. It is also desirable to leave out certain inner and outer
points from the prole during the process of tting since it is possible that
de Vaucouleurs' law does not hold right upto the center (inner points) and
since the signal-to-noise ratio decreases away from the center (outer points).

PSF The minimization process involves the convolution of the model galaxy
with a PSF. The FWHM of the PSF, is obtained by tting a Gaussian to a
number of stars in the object frame using psfstar. The convolution step is
crucial since the light near the center of the galaxy is highly aected by the
seeing and it is not possible to carry out a satisfactory deconvolution of the
image.

61
Model generation The decomposition of the surface brightness prole is
done in an iterative fashion. At each stage a model prole is constructed
from the current guess parameters. At initialization, the tting program
reads in the intensity prole. The minimization le is also read in. Using
Equation 3.14 a model galaxy is built. It is then convolved with a Gaussian
PSF with FWHM as read in from the input. The generated galaxy has a xed
position angle viz. zero since the position angle is not being t. A prole is
now obtained along the major-axis of the model at the points corresponding
to the input prole. This prole is then compared with the observed galaxy
prole, the guess parameters adjusted and the model generation repeated
until the best t is obtained.

Goodness of t The best t model is obtained by minimizing the value

of the reduced chi-squared function 2 , given by
n2  o ? m 2
X
2 = 1 i

i
; (3.15)
i=n1 i
where, oi and mi are the observed and model surface brightness at the ith
point and i is the standard deviation in the surface brightness. Here  is
the number of degrees of freedom and is given by
 = n ? f; (3.16)
where, f is the number of free parameters of the t, and n is the number of
data points used.
The 2 minimization and error evaluation is handled by the powerful
package minuit (described below). Very rarely, for pathological proles, the
program may terminate with incorrect values if the maximum number of
iterations is exceeded before reaching convergence. In such a case a ag
in the output marks the reason for the error exit. The information on the
goodness of t is output along with the error matrix, warnings, if any, arising
during tting and an optional set of 2 contours for parameter pairs.
62
Output The program ends by creating several les. The various les con-
tain the best t model prole, a summary of parameters being t, their initial
values and the best t values along with the error bars. The output prole is
in the form of a tted bulge prole, tted disk prole and a summed prole.
The 2 value is output as well.

Procedure for minimization Minuit: Function Minimization and Error

Analysis is a package from CERN, Geneva, Switzerland, oering a host of
minimization routines. Minuit is a tool to nd the minimum value of a
multi-parameter function and analyze the shape of the function around the
minimum. Minuit acts on a user dened Fortran function containing the
dierent parameters.
Minuit is a smart program and hence rather than plodding along a regular
grid in a multidimensional space, it is capable of making sudden jumps to
dierent parts of the space to check if it can nd a better minimum. A
ag, called the strategy ag, can be set within minuit. This ag decides
whether a very careful search is to be carried out around the minimum that
has been located. If the strategy is thus set, one can avoid obtaining only a
local minimum rather than a global minimum. The cost is several additional
function calls.
The shape of the function close to the minima decides the errors on the
best t parameters. Normally these are 1 errors. However, one can set
this number to whatever one desires, say, 2. In the prole tting program,
use has been made of three minuit processors viz. migrad, hesse and minos.
Migrad and hesse produce an error matrix which is the inverse of the matrix of
second derivatives of the function being minimized. The minuit error matrix
takes into account all parameter correlations, but not the non-linearities. It
is assumed that the problem was of a linear nature and that the error bars
are symmetric. Thus, 2 errors are always twice the 1 errors. On the other
hand minos cam handle asymmetric errors in case of a non-linear problem.

63
Minos starts with the error matrix generated by a method like migrad. It then
actually traverses along a parameter to see when a 1 error bar is reached.
It does this on both sides and for each parameter. Minos is thus essential
to estimate the correct parameter errors when the problem is signicantly
non-linear.
To sum up, at the end of the day you have your extracted parameters
along with their errors.

Fitting Considerations The region over which the tting is actually car-
ried out can turn out to be crucial for the accuracy of the extracted parame-
ters. In the literature one nds that a number of dierent methods are used
in selecting the data for tting. Once the data has been t, several additional
parameters are obtained for specic purposes.
The prole can be t with de Vaucouleurs' law, either separately or in
combination with an exponential disk and/or a point source. In case of
nearby galaxies that subtend large angles on the CCD, one has the luxury of
being able to go to several tens of arcsec from the center. Sometimes a metric
radius of, say, 13 kpc is chosen and the t done up to that point (Gonzalez-
Serrano et al., 1992). However, such a radius may cover the whole galaxies
in the case of dwarf galaxies, but would cover only a part of cD galaxies.
However, in case of the radio galaxies of the sample the angle subtended
on the CCD is small. Therefore one has to stop at only several arcsec when
the background noise starts being dominant. The outermost points we used
were those with an error of 0.1 magnitude in the surface brightness.
The innermost point which can be included in the prole t has also been
widely debated. In the inner region the eect of seeing is very dominant and
the validity of de Vaucouleurs' law close to the center is also suspect. For
the above reason, the inner few points have to be disregarded during the t
even if the convolution of the model prole with the PSF is carried out. The
decision on the innermost point to be included was taken following a series

64
of tests described below.
During the program runs the ellipticity was tted for (a constant rather
than as a function of r), but the position angle was not. This is because in
case of a galaxy showing isophotal twist the prole that is output by ellipse
is along the twisting major axis. If the ellipticity and position angle are not
being t, the value of PA and  at re=2 can be used. Values at a ducial
magnitude (e.g. mB = 24:5; Owen and Laing, 1989) have been used too. In
our work, the prole was always obtained along the major axis. However,
since we are using a single value for the ellipticity, the corresponding proles
along the minor-axis, or at any other PA will provide similar results since
these proles are related to the major axis prole through the ellipticity.

3.5 Consistency checks

Before obtaining parameters for the radio galaxy sample, a series of test runs
were carried out on galaxies previously studied, and on simulated galaxies
in order to test the validity and robustness of the results provided by the
program. It is important that the test runs are carried out on a single galaxy
or a set of galaxies. The results can then be used to decide the approach for
the galaxies to be studied. The tests were also meant to decide on the range
of radius over which the tting could be reliably done. With the above in
mind, the following kinds of ts/tests were tried:
 Fit just the eective radius and central bulge intensity.
 Fit the eective radius, bulge ellipticity and central bulge intensity.
 Fit both the bulge and disk scales and the respective ellipticities. Also
t the central bulge intensity.
 Fit including all points of the prole upto the point where the error as
indicated by ellipse falls to 0.1 magnitude in the surface brightness.
65
  Fit leaving out a dierent number of innermost points i. e., start at
1 arcsec, 1:5 arcsec, 2 arcsec, 2:5 arcsec, 3 arcsec and 4 arcsec from the
center.
 Fit proles obtained with dierent lters for the same galaxy.
The results for NGC 661 are tabulated in Table 3.1 for the V , R and I
lters. Here n1 and n2 refer to the innermost and outermost points from the
prole that were used for the t. The rst column indicates the corresponding
numbers in arcsec along with the lter and t type, b indicating a bulge only
t and d a bulge+disk t. The graphical results are shown in Figure 3.4. The
extracted re is shown as a function of the innermost prole point that was
used. The last point was always chosen to be the one where I=I0 exceeded
0.1. For NGC 661, we also carried out ts with a point source. However,
we did not nd evidence for a point source. Any such source which may be
present is weak.
The results obtained reveal a number of interesting points.
1. If only the bulge parameters are free, the seeing aected central points,
which also have the highest signal to noise, tend to aect the results.
The value of re rises gradually from 2 arcsec to about 12 arcsec when
successively larger number of points are left out. The FWHM for the
PSF for the images in the three lters is 1:9 arcsec and the re is seen
to stabilize when the points in the inner 3 arcsec are omitted. The
stabilized re is  25% smaller than that obtained by other observers
(e. g. 16:2  0:3 arcsec in the rG band obtained by Djorgovski and Davis,
1987). The value of 16 arcsec will be called the actual value during the
course of this discussion.
2. When the disk as well as the bulge parameters are free, the re obtained
is close to the actual value (11:4 arcsec?16 arcsec). This is because
the inner points are absorbed in a seemingly false disk (seeing disk?)

66
 Table 3.1: Tests on NGC 661
code re rs D=B n1 n2 2
arcsec arcsec
Ib 2.3 | | 1 41 8328.89
Ib15 4.3 | | 11 41 6983.84
Ib20 8.3 | | 14 41 1084.20
Ib25 10.8 | | 16 41 6.77
Ib30 10.8 | | 18 41 1.78
Ib40 11.0 | | 21 41 1.85
Id 14.5 1.4 0.059 1 41 0.53
Id15 13.6 1.4 0.058 11 41 0.27
Id20 13.8 1.3 0.183 14 41 0.28
Id25 13.1 1.3 0.133 16 41 0.38
Id30 15.1 1.5 0.233 18 41 0.18
Id40 15.2 1.5 0.213 21 41 0.19
Rb 2.2 | | 1 42 10959.91
Rb15 4.6 | | 11 42 11404.35
Rb20 7.5 | | 14 42 2746.72
Rb25 11.9 | | 16 42 48.28
Rb30 11.4 | | 18 42 3.76
Rb40 11.6 | | 21 42 4.10
Rd 15.8 1.5 0.088 1 42 1.06
Rd15 15.0 1.4 0.209 11 42 0.43
Rd20 15.4 1.5 0.103 14 42 0.44
Rd25 15.9 1.5 0.110 16 42 0.40
Rd30 16.0 1.5 0.157 18 42 0.42
Rd40 16.0 1.5 0.129 21 42 0.47
Vb 1.8 | | 1 40 9486.18
Vb15 4.4 | | 11 40 9656.22
Vb20 9.7 | | 14 40 486.81
Vb25 11.3 | | 16 40 6.73
Vb30 11.3 | | 18 40 1.37
Vb40 11.4 | | 21 40 1.52
Vd 14.2 1.3 0.210 1 40 0.85
Vd15 11.2 0.1 0.000 11 40 2.42
Vd20 11.3 0.1 0.000 14 40 1.63
Vd25 11.4 0.1 0.000 16 40 1.52
Vd30 13.9 1.4 0.176 18 40 0.37
Vd40 13.4 1.2 0.174 21 40 0.35
Note: The \|" in columns rs and D=B indicates these parameters were not relevant in
the run because the disk parameters were not used during the ts. n1 is the innermost
point and n2 is the outermost point used during the t.

67
Figure 3.4: Eect of starting the ts at dierent points in case of B and B+D
ts. When only bulge parameters are free, the extracted bulge scale length
increases from  2 arcsec to  12 arcsec as succesively larger number of inner
points are excluded from the t. When both bulge and disk parameters are
free, the extracted bulge scale length does not depend on the number of inner
points left out. The result is consistent for the V; R and I lters. The error
bars are smaller than the symbol size used. Djorgovsky and Davis (1987)
had obtained a value of re = 16:2  0:3 arcsec in the rG band which matches
well with our value of 16  0:15 arcsec in the R band.

68
 which has a small increasing scale length (1:2 arcsec?1:5 arcsec) and a
decreasing disk-to-bulge ratio (0.23{0.06). For a typical disk, e. g. that
of a spiral or an S0 (Burstein, 1979) one expects rs to be in excess of
re and a D=B of well over 0.2. rs being smaller than the FWHM of
the PSF indicates that though the inner points are being absorbed in
a disk, the disk luminosity is small and is more likely to be a deviation
from de Vaucouleurs' law rather than a dynamical disk.
3. The typical 2 range for only-bulge (very large to about 1.3) and for
bulge-disk (exceeds 2 only once) indicates that a statistically better t
is obtained when the disk is left free. The reason why the 2 is very
large for the bulge-only case when the inner points are included, is that
the 1 errors on the innermost points are the smallest since the signal
there is highest. As a result, the parameters are constrained to t the
innermost points better. This decides the parameter values without
concern to the outer points. If there is even a small departure from,
say, the de Vaucouleurs model near the center, the parameters turn out
to be wrong and result in a large 2 .
Thus, the following conclusions can be drawn from the above results: (1)
Fitting for a disk as well as a bulge does not lead to the detection of a large
false disk if one is not present and (2) inner points within about 1:5 FWHM
of the PSF should be omitted for reliable results.
Hence, for tting the radio galaxies and the control sample galaxies, we
chose to t for bulge as well as disk parameters. We also chose to uniformly
mask out the center and not t for a central point source.

3.6 2D distribution
As noted earlier, the 1D decomposition of a prole involves averaging of the
prole along elliptical isophotes and obtaining a major axis prole. The

69
 Table 3.2: 2D tests
Filter Type of t re rs D=B eb ed 2
arcsec arcsec
I unmasked 14.28 9.00 0.000 0.30 0.67 2.881
mask of 300 11.44 35.03 0.072 0.31 0.00 2.239
mask of 500 11.25 35.98 0.079 0.31 0.00 2.216
64  64 14.93 1.45 0.077 0.30 0.78 1.771
R unmasked 14.67 9.47 0.000 0.30 0.64 1.716
mask of 300 15.19 35.90 0.000 0.29 0.82 1.495
mask of 500 11.69 19.84 0.103 0.31 0.00 1.278
64  64 16.16 1.67 0.060 0.30 0.57 1.264
V unmasked 14.11 1.00 0.000 0.30 0.92 1.613
mask of 300 14.45 0.59 0.000 0.30 0.94 1.470
mask of 500 14.86 0.09 0.000 0.30 0.82 1.423
64  64 13.76 1.08 0.000 0.31 0.98 1.370

70
averaging smears out faint features. A 2D tting scheme overcomes this
diculty. In the 2D t a model galaxy is generated using assumed values of
re, rs, eb, ed and 0. The model is convolved with an appropriate PSF and
a 2 minimization is then made using the minuit routines. The standard
deviation used at each pixel for the 2 minimization is obtained from the
image without subtracting the sky background. The iterative minimization
continues till either the maximum number of iterations is exceeded or till the
change in 2 goes below a specied tolerance. The procedure is described in
detail by Wadadekar et al. (1998).
An advantage of the 2D method is the detection of faint features. When
the best-t model is subtracted out from the input image, the features in
the residual stand out. In case of the 1D t, the 2D information has been
converted into a 1D prole by averaging along elliptical isophotes. In this
process the faint features are also averaged out. Even if the residual (input
prole { best t prole) indicates an excess at the semi-major axis where
the feature was, the spatial location of the feature is lost. An example is
that of shells in a galaxy. When a 2D t is made, interleaved shells in an
elliptical galaxy will stand out. On the other hand, if one wants to ignore
faint features and is only interested in the large scale light distribution of the
galaxy, the 1D ts prove more useful.
To be consistent with the 1D ts, we masked out the central regions of
the images before performing the t on the sample objects. The results of the
test ts that we carried out for 128  128pixel2 images of NGC 661 images
are given in Table 3.2.
The following points can be noted with respect to these results:
 2 values are acceptable for all the 64  64pixel2 images. Except for
the I images, where we get 2 > 2, the 2 values for 128  128pixel2
images in the other band are acceptable too.
 In 2 of the 3 cases where 2 is not acceptable we nd that a spurious disk

71
 has been detected. Though it has a large scale length (rs > 35 arcsec),
we have D=B < 0:2 indicating that the disk is weak.
 re values are similar to those obtained by the 1D program (See Ta-
ble 3.1).

Comparison with 1D
We mention here the functional dierences in the 1D and 2D tting programs.

2 : The rst noticeable thing is that the size of error bars in the 1D case
are much smaller than the 2D case. Though in both cases a 2 minimization
was done using the same minimization package, the denition of 2 was
dierent. In the 1D case the input parameter was mean intensity, I0 , along
dierent ellipses and the 1 error on this parameter was that given by the
IRAF task ellipse. I0 is obtained by tting an isophote with an elliptical
shape and averaging along the best t ellipse. A large number of points
having approximately the same intensity are used resulting in the 1 error
being small. In the 2D case the original photon shot noise is considered for
each pixel to evaluate the 2. Deviant points (e. g. those arising from faint
foreground stars) are clipped o during the 1D t. This can not be done in
the 2D program. However, unlike in the 1D case, the features do not get
averaged out either. If left unmasked, they contribute a large part to the
resulting 2 in the 2D case. The same holds true for the outer regions of
the galaxy frame. In 1D we used points in the prole for which the error
I=I0  0:1. In 2D when a 128  128 image is used the noisier regions also
get considered during the t. As seen from Table 3.2, for a galaxy like NGC
661, it does not make too much of a dierence to re but the 2 does reduce
in the smaller image.

72
Position Angle In both the programs the position angle is not varied with
radius during tting. In case of the 1D ts it poses less of a problem since
the prole output by ellipse is along the twisting semi-major axis direction
i. e., the position angle itself is output as a function of the semi-major axis
and the 'major-axis' prole is not strictly along a xed direction. Thus, when
there is a large position angle twist, the 1D program is still reliable whereas
the 2D program will not be. Additionally, the 2D program, since it involves
only a single position angle, is capable of nding only disks that are along
the major axis.

Ellipticity Neither the 1D program, nor the 2D program t for changing

ellipticity and the results need to be taken with caution when the ellipticity
varies a lot. When the ellipticity is approximately constant, the results should
be comparable.
We present the results of tting the program objects with the 1D proce-
dure in the next chapter.

3.7 Summary
In this chapter we have considered the surface brightness distribution of
dierent types of galaxies. We have noted that the surface brightness distri-
bution of almost all galaxies can be well described as being a combination of
a spheroidal bulge, a at disk and a point source.
From the point of view of understanding the galaxy morphology, it is very
important for us to be able to separate these components. Early observations
and rudimentary ts indicated that the bulge follows the de Vaucouleurs r1=4
law while the disk has an exponential distribution. Both these laws are
empirical in nature. We have also described some other laws that are in use.
To actually bring about the decomposition of the galaxy brightness dis-
tribution, a radial prole of the galaxy is rst obtained by tting ellipses to

73
the isophotes of the galaxy. Minimization routines are then used to obtain
the dierent parameters. Alternately one can use the 2D galaxy image di-
rectly to obtain the parameters. The 2D tting has been found to be very
promising in locating assymmetric features.
In the next chapter we provide the actual scale lengths obtained and
discuss the trends.

74
Chapter 4
Scale lengths at dierent
wavelengths
Traditionally, radio galaxies have been believed to be ellipticals consisting
of a coeval population of old stars, and almost no dust or gas. However,
photometric studies carried out over the past few years have demonstrated
that many elliptical galaxies not only have dust and gas, but also possess ne
structure indicating some amount of activity in the past  102 million years.
Radio galaxies host an active galactic nucleus (AGN) and are associated with
highly energetic phenomena like the radio jets which transport a very large
amount of energy over hundreds of kiloparsecs. Such phenomena are likely
to be associated with morphological features not found in normal elliptical
galaxies.
In this chapter we present the results of tting the 1D proles of galaxies
as described in the previous chapter. We show that the ratio re(B )=re(R) of
the bulge scale lengths in B and R lters, provides a measure of the color
gradient in the galaxy. The ratio is related to color gradients measured in
the conventional way, but is more robust: it can provide an estimate of the
color gradient even when the signal-to-noise ratio is not good enough for
the gradient to be measured unambiguously using conventional techniques.

75
Using the ratio we show that a large fraction of radio galaxies become bluer
towards the center. We compare in Chapter 6 these results with information
obtained from color maps.
Our model ts indicate that radio galaxies tend to deviate from de Vau-
couleurs' law near the center more often than control galaxies. We argue
that the deviations are due to excess blue emission or dust absorption in the
central region and obtain quantitative estimates.

4.1 Results of prole ts

We have t the surface brightness proles of radio galaxies and the control
sample galaxies with de Vaucouleurs' law and an exponential disk as de-
scribed in Chapter 3. We present in this section our ndings with regards
to the goodness of t, the values obtained for various parameters and their
dependence on the wavelength.

4.1.1 Goodness of t
We have obtained the scale lengths using a 2 minimization between the
observed radial brightness prole of a galaxy and a model prole. We say
that a t is very good for 2 < 1 and acceptable for 1 < 2 < 2. The limit
of 2 is chosen as a ducial boundary, and has no particular signicance.
The distribution of minimum 2 values are shown in Table 4.1. We note
the following points with respect to this distribution:
 We get very good or acceptable ts in  85% of the cases in the radio
as well as the control sample. Strong radio sources therefore do not
prefer highly distorted galaxies.
 The distribution of 2 values for the radio and control samples is similar
(See Table 4.1 and Figure 4.2). This indicates that the radio galaxies
are not very highly disturbed.
76
Table 4.1: Distribution of 2 for the radio and control samples. The number
of cases for which we obtain a good 2 is roughly the same for both the
samples.
Radio Control
B R B R
 <1
2 18 18 18 17
1< <2 8 7
2 7 8
2 > 2 4 5 5 5

 The 2 does not increase as a function of redshift, i. e., we get good ts
right upto the redshift of 0.3 that we have considered. (See Figure 4.1).
 2 (B ) is in general smaller than the corresponding 2 (R) (Figure 4.2).
This is partly due to the fact that both dust and star formation are
likely to aect the B light more thus increasing the (I ) which in turn
will lead to smaller 2 values provided all other factors remain the
same.

4.1.2 Disk parameters

Here we present the parameters obtained following the exercise of tting
the surface brightness proles with B+D models. So far as the disks are
concerned, we note the following:
 Half the radio galaxies show the presence of a disk with a disk{to{bulge
luminosity ratio D=B > 0:1 (Table 4.2). The fraction of such D=B
values is just a bit higher for the control sample. Values of D=B < 0:1
are not taken to be signicant in our data.

77
Figure 4.1: 2 in B and R lters is acceptable for most galaxies in the com-
plete range of redshift used. This suggests that higher redshift objects are
not necessarily more disturbed.

78
Figure 4.2: 2 values in B and R lters for the radio and the control samples.
 80% galaxies in both samples have acceptable ts.

79
Table 4.2: D=B ratio for the radio and control samples. rs is the disk scale
length.
Radio Control
B R B R
D=B > 10% 15 15 20 16
and
rs < PSF 8 10 10 7
rs > PSF 7 5 10 9

 About half the galaxies with D=B > 0:1 have exponential disk scale
length rs smaller than the FWHM of the PSF, for the radio as well as
the control sample.
As shown above, in a majority of cases, we nd that D=B is negligible.
In cases where D=B is appreciable (> 0:1) we nd that (1) rs(B )  rs(R)
and (2) (D=B )blue  (D=B )red. As a result we will, in general, not refer to
the disk component, while considering the dependence of tted parameters
on the observing band, even though we have t the prole with a bulge-disk
combination. There are cases where we do nd large D=B (> 0:3). We will
discuss these separately in Chapter 5.

4.1.3 Bulge parameters

We now turn to the bulge scale lengths in the B and R bands for the radio
galaxies and the control sample. We note the following points about bulge
scale lengths:
 The 1 errors on re are typically  10%.

80
Table 4.3: re(B ) and re(R) values for the samples. Scale lengths in the two
lters are taken to be equal if they dier by less than 1.
Radio Control
re(B )  re(R) 5 4
1 < re(B ) ? re(R) < 2 2 10
2 < re(B ) ? re(R) < 3 4 8
1 < re(R) ? re(B ) < 2 3 6
2 < re(R) ? re(B ) < 3 13 2
Indeterminate 3 0

 In the case of radio galaxies as well as the control sample, we have

galaxies with re(B ) < re(R); re(B )  re(R) and re(B ) > re(R). The
distribution amongst these classes is shown in Table 4.3.
 When re(B ) > re(R), the surface brightness in R increases more rapidly
towards the center than the surface brightness in B (See Figure 4.3). In
other words, half the red light from the galaxy is contained in a smaller
region than half the blue light. re(B ) > re(R) therefore implies that
the galaxy on the average becomes redder inwards.
 Similarly, when re(B ) < re(R), the galaxy becomes bluer as one moves
towards the center of the galaxy.
Figure 4.5 shows the scale lengths for the control and radio samples in the B
and R lters. Figure 4.5 and Table 4.3 indicate that for the control sample
galaxies typically the points are on or below the line of equality (re(B ) >
re(R)), whereas, for the radio galaxy sample a large number of points lie
above the line (re(B ) < re(R)). This suggests that the scale lengths for the
two samples could be worth comparing in detail from the point of view of
81
dierentiating between the host galaxies. With this in mind we obtained the
ratio re(B )=re(R) for the radio as well as the control sample.
We show the distribution of the re(B )=re(R) ratio in the form of a his-
togram in Figure 4.4. It is clear that there is a qualitative dierence in the
distribution for the two samples. The mean values of the ratio for the two
samples are

hre(B )=re(R)iradio = 0:87  0:15 (4.1)

and
hre(B )=re(R)icontrol = 1:25  0:10: (4.2)
Equation 4.2 indicates that the control galaxies become redder inward which
is consistent with previous studies (e. g. Sandage and Visvanathan, 1978).
However, the interesting new fact is that contrary to the behavior of control
galaxies, a majority of the radio galaxies become bluer as one approaches the
center, i. e., the color variation in radio galaxies is opposite of that in the
control galaxies. In the next section we obtain color gradients for the radio
and control samples using conventional techniques and compare the results
obtained with inferences drawn from the scale length ratios.
An important point to remember is that points within 1.5 times the
FWHM of the PSF from the center have been excluded from the t. The
corresponding physical distance is upto 4 kpc from the center of the galaxy
depending on its redshift. As a result, it cannot be the light from the AGN
that is responsible for the bluer colors of the radio galaxies. However, the
AGN may contribute indirectly, by triggering star formation which makes the
light blue. Later on in this chapter we quantify the blue excess and discuss
the possible mechanism giving rise to it.

82
Figure 4.3: A hypothetical de Vaucouleurs' law galaxy. On the x-axis we have
r1=4 and the surface brightness  in magnitude/ arcsec?2 on the y-axis. The
surface brightness proles for the galaxy in the B and R bands are shown as
B and R respectively. r1 and r2 are the two points on the radial prole
which we use to obtain color gradients in the conventional way (see text).

83
Figure 4.4: The distribution of re(B )=re(R) for the radio and control samples

84
Figure 4.5: The eective radius re in B and R lters for the radio sample
(top) and the control sample (bottom). The dotted line denotes re(B ) =
re(R). Points above it denote galaxies which become bluer inwards. There
is a larger number of such cases amongst the radio galaxies.

85
 Table 4.4: Scale lengths for the radio sample
B R
IAU z re rs D=B (0) 2 re rs D=B (0) 2
kpc kpc magnitude kpc kpc magnitude
0344{291 0.137 :::  ::: :::  ::: :::  ::: :::  ::: ::: :::  ::: :::  ::: :::  ::: :::  ::: :::
0349{278 0.066 5.25  0.74 8.53  0.87 1.493  0.205 16.694  0.190 0.33 20.70  5.44 0.44  0.08 0.055  0.009 16.831  0.495 0.25
0354{263 0.300 32.86  0.89 1.40  0.10 0.176  0.013 21.025  0.048 1.17 9.81  4.00 0.87  0.13 0.205  0.147 16.748  0.787 0.44
0420{263 0.130 11.78  1.05 0.55  36.89 0.001  7.402 17.904  0.159 2.27 21.78  1.33 1.51  0.39 0.082  0.047 16.627  0.105 0.33
0434{225 0.067 29.80  4.20 0.87  0.05 0.038  0.013 17.553  0.262 0.21 27.45  0.94 0.91  0.03 0.043  0.004 15.953  0.058 0.29
0446{206 0.073 13.13  0.28 1.04  0.03 0.140  0.007 17.283  0.040 0.42 14.32  1.29 1.02  0.03 0.144  0.008 16.047  0.163 0.61
0453{206 0.035 22.31  1.40 0.55  0.02 0.023  0.004 17.387  0.106 0.65 18.73  0.50 0.49  0.02 0.023  0.003 15.661  0.045 0.86
0520{289 0.068 12.59  0.45 0.86  0.06 0.127  0.016 17.279  0.069 0.35 16.29  1.17 0.88  0.03 0.157  0.011 16.491  0.126 1.01
0938{205 0.371 27.11  8.94 2.53  68.97 0.059  0.476 21.710  0.410 1.05 6.99  0.13 0.01  0.00 0.007  0.003 18.223  0.030 0.63
0950{239 0.300 :::  ::: :::  ::: :::  ::: :::  ::: ::: :::  ::: :::  ::: :::  ::: :::  ::: :::
0952{224 0.230 81.73  2.92 2.52  0.27 0.386  0.085 20.380  0.054 1.31 7.39  0.33 38.88  4.76 2.545  0.331 14.351  0.073 2.80
1006{214 0.250 6.29  2.34 5.74  0.31 8.841  1.100 18.107  0.786 1.19 31.86  8.35 4.86  0.16 5.353  1.440 19.060  0.600 1.14
1033{251 0.110 :::  ::: :::  ::: :::  ::: :::  ::: ::: :::  ::: :::  ::: :::  ::: :::  ::: :::
86

1048{238 0.207 5.29  0.18 0.22  2.35 0.001  0.079 15.558  0.062 13.50 20.77  8.09 1.13  0.17 0.544  0.081 16.746  0.720 0.33
1053{282 0.061 7.76  0.17 20.84  1.59 4.307  0.350 16.576  0.029 0.40 7.01  0.28 26.73  3.00 5.521  0.774 14.848  0.052 2.24
1056{272 0.250 61.97  7.28 3.89  10.70 0.006  0.438 20.101  0.192 0.32 89.52  3.13 2.29  0.35 0.040  0.008 18.439  0.056 0.22
1103{244 0.050 6.14  3.16 0.57  0.12 0.310  0.257 18.169  1.021 0.80 9.01  4.88 0.45  0.41 0.055  0.031 16.838  0.971 0.40
1126{246 0.155 36.02  1.00 57.01  132.97 0.001  0.119 17.644  0.059 2.33 79.34  12.99 3.27  0.53 0.024  0.009 17.362  0.262 1.41
1215{215 0.075 4.46  0.14 1.20  0.04 1.444  0.162 16.621  0.064 1.97 11.56  3.60 1.52  0.12 1.042  0.464 18.084  0.573 0.16
1222{252 0.077 51.95  7.52 5.23  0.36 1.494  0.441 21.201  0.436 1.12 22.96  0.30 3.50  0.18 2.506  0.074 19.072  0.027 1.03
1251{289 0.058 20.61  2.53 3.21  0.16 0.210  0.033 16.642  0.240 0.83 67.74  1.48 2.96  0.03 0.196  0.007 17.171  0.035 5.63
1254{268 0.135 15.40  1.95 153.53  165.93 7.739  8.814 17.853  0.210 0.56 21.28  1.40 0.69  0.27 0.033  0.011 15.874  0.115 1.43
1257{253 0.045 7.61  0.82 0.58  0.03 0.094  0.015 16.617  0.228 0.47 12.31  0.44 0.77  0.02 0.107  0.004 15.911  0.064 4.24
1303{215 0.120 5.53  0.16 1.30  0.49 0.144  0.064 16.939  0.060 1.73 11.82  3.11 1.66  0.22 0.502  0.196 16.635  0.558 0.49
1323{271 0.044 30.23  1.72 1.62  0.09 0.023  0.003 17.851  0.086 0.69 26.67  1.28 1.60  0.06 0.039  0.003 15.670  0.074 1.15
1329{257 0.190 19.84  2.79 0.36  236.51 0.001  6.517 18.146  0.222 3.30 37.03  2.74 1.12  0.23 0.116  0.033 17.231  0.116 0.84
1343{253 0.129 46.08  23.63 1.95  0.26 0.078  0.029 18.521  0.804 0.18 45.40  5.39 2.18  0.09 0.098  0.006 16.790  0.194 0.59
1344{241 0.019 5.42  0.17 0.16  0.02 0.020  0.004 15.664  0.057 1.13 21.94  0.29 0.69  0.01 0.090  0.002 16.039  0.022 44.34
1346{252 0.125 88.73  6.63 1.56  0.45 0.041  0.010 19.909  0.115 0.46 149.08  3.50 1.69  0.19 0.029  0.013 18.140  0.019 0.74
1354{251 0.037 15.62  1.72 1.27  0.06 0.112  0.011 17.209  0.184 0.97 30.51  1.54 1.23  0.02 0.135  0.005 16.871  0.078 1.15
 Table 4.5: Scale lengths for the control sample
B R
RA x y re rs D=B (0) 2 re rs D=B (0) 2
arcsec arcsec magnitudes arcsec arcsec magnitudes
0349 127.69 617.56 73.865  3.830 0.638  0.049 0.016  0.002 22.306  0.033 1.69 42.400  3.464 1.056  0.047 0.036  0.003 19.908  0.033 1.22
0349 434.07 456.51 11.508  0.094 2.301  0.086 0.119  0.002 18.736  0.000 5.85 8.844  0.299 1.931  0.056 0.486  0.014 17.300  0.088 3.62
0354 423.77 1261.58 31.200  3.825 4.994  2.559 0.027  0.077 19.501  0.185 0.33 18.369  0.910 1.718  0.337 0.078  0.031 21.658  0.070 0.25
0420 342.18 1325.86 16.700  0.016 1.822  0.032 0.119  0.000 19.916  0.007 14.79 9.263  3.018 2.240  0.256 0.263  0.150 17.071  0.118 0.39
0434 585.28 505.49 11.287  2.577 1.554  0.376 0.286  0.110 18.341  0.171 0.69 8.237  0.958 2.194  0.244 0.194  0.079 16.705  0.098 0.67
0446 384.59 541.63 12.459  0.084 1.058  0.023 0.431  0.039 18.496  0.000 0.38 13.207  0.497 1.198  0.053 0.083  0.004 17.124  0.060 0.71
0446 641.79 568.03 28.652  0.987 1.111  0.056 0.073  0.120 19.446  0.000 0.11 27.737  1.523 1.259  0.222 0.043  0.022 17.963  0.233 0.41
0446 644.08 48.45 38.851  3.225 5.291  0.766 0.291  0.081 20.419  0.090 0.92 14.053  3.568 4.267  3.075 1.520  1.108 18.577  0.324 0.26
0453 371.49 602.07 15.400  0.081 1.214  0.038 0.188  0.029 18.703  0.041 1.08 13.539  0.961 1.373  0.045 0.151  0.007 17.138  0.117 1.79
0503 603.37 928.14 22.100  0.282 2.400  0.050 0.089  0.023 20.306  0.000 1.55 19.429  0.180 1.235  0.128 0.050  0.025 18.106  0.000 0.80
0938 1028.82 1359.55 15.138  2.437 0.464  0.036 0.084  0.082 21.052  0.375 0.63 11.356  0.788 0.618  0.049 0.071  0.012 17.887  0.158 1.16
0938 621.36 1754.64 9.501  0.148 1.172  0.079 0.157  0.068 19.676  0.000 0.87 6.010  1.600 2.307  0.460 0.153  0.113 16.762  0.709 0.65
1053 421.36 332.66 8.402  0.696 9.370  9.086 0.723  0.633 19.440  0.124 0.41 9.600  0.343 0.716  0.117 0.108  0.136 17.571  0.144 0.57
87

1251 190.14 214.45 4.136  0.540 1.056  0.155 0.463  0.331 17.576  0.000 0.26 6.108  0.915 0.604  0.040 0.351  0.113 16.730  0.062 1.05
1251 733.71 750.91 19.285  1.508 10.000  6.585 0.001  0.013 18.340  0.125 2.46 28.300  0.687 0.933  0.047 0.068  0.025 17.385  0.027 1.20
1251 850.72 748.95 5.590  0.590 5.871  7.059 0.001  0.194 18.183  0.144 0.86 10.981  0.472 0.789  0.074 0.183  0.051 17.977  0.053 0.42
1257 403.28 962.45 9.226  0.151 1.016  0.067 0.220  0.215 18.436  0.000 0.42 10.541  0.852 0.956  0.048 0.159  0.013 17.504  0.144 1.46
1257 576.26 263.75 20.110  0.675 0.614  0.016 0.293  0.027 19.817  0.017 1.25 14.148  0.097 0.764  0.018 0.099  0.003 17.726  0.000 0.60
1257 674.69 812.88 11.754  1.929 0.795  0.033 0.423  0.040 18.579  0.286 0.98 11.617  3.610 0.850  0.053 0.081  0.013 17.124  0.260 0.39
1257 960.67 754.24 13.355  0.557 0.776  0.076 0.063  0.025 19.906  0.000 0.36 11.072  1.057 0.760  0.034 0.199  0.056 18.012  0.171 1.91
1303 636.63 769.88 22.272  0.766 2.295  0.261 0.230  0.032 20.179  0.084 1.07 0.000  0.000 0.000  0.000 0.000  0.000 28.996  0.000 0.00
1323 1080.31 384.50 24.090  0.253 1.231  0.060 0.087  0.023 19.002  0.043 1.46 29.480  9.627 1.027  0.053 0.046  0.010 17.216  0.508 0.94
1323 1477.09 1125.12 101.130  3.535 3.445  0.061 4.852  3.311 25.157  0.322 7.53 63.600  3.086 3.205  0.028 8.220  0.605 22.986  0.000 5.76
1323 1673.83 1341.04 9.437  0.917 3.363  0.229 0.519  0.426 18.096  0.000 1.42 11.713  0.221 4.063  0.199 0.458  0.011 16.522  0.017 3.05
1323 744.88 604.21 14.700  0.825 0.825  0.022 0.191  0.017 19.150  0.016 3.64 25.182  10.650 0.909  0.035 0.159  0.023 18.212  0.653 2.68
1329 231.71 1391.77 10.300  0.082 0.790  0.025 0.288  0.030 18.965  0.071 0.94 9.187  0.039 0.671  0.010 0.262  0.015 16.656  0.000 1.35
1329 234.65 1926.70 53.719  0.670 1.780  0.046 0.334  0.013 21.565  0.037 0.32 19.220  0.510 1.577  0.033 1.651  0.093 18.748  0.055 0.55
1346 1034.42 1342.65 9.024  3.619 2.531  0.229 0.445  0.339 20.360  0.187 0.91 13.600  2.871 0.712  0.472 0.030  0.058 17.768  0.044 0.79
1346 1239.18 1437.23 25.520  0.649 0.648  0.029 0.168  0.015 21.136  0.047 0.66 20.300  3.575 0.837  0.032 0.062  0.003 18.126  0.000 0.53
1346 971.00 1471.82 14.140  0.935 0.583  0.135 0.076  0.122 20.426  0.000 0.24 8.975  0.844 0.643  0.078 0.097  0.066 17.281  0.172 0.77
Figure 4.6: Color gradients for the radio sample and the control sample. A
negative gradient indicates that the galaxy becomes redder near the center.
It is seen that a higher proportion of control galaxies become redder towards
the center.

88
4.2 Color gradients
The color variation in a galaxy can be quantied using a color gradient
parameter, i. e.,

(B ? R)=
(log r); (4.3)
i. e., change in color per decade in radius. We describe below the proce-
dure for obtaining this parameter. We then present the color gradients that
we obtain for the radio and control samples and compare them with those
obtained by others for similar samples. We show that in principle the scale
lengths should provide the same information as the color gradients but that
the former are more robust indicators of color variation.
The change in color per decade in radius is almost linear in most cases and
the color gradient, e. g.
(B ? R)=
(log r), is obtained by tting a straight
line to the color prole between an inner radius r1 and an outer radius r2. The
choice of r1 and r2 has been described in Chapter 3. r1 is at a radial distance
of 1.5 times the FWHM of the PSF while r2 is the point along the major
axis at which the 1 error on the mean surface brightness of the tted ellipse
reaches 0.1 magnitude. The additional caveat is r2 = minfr2 (B ); r2(R)g.
When two small galaxies (angular diameter < 1500) and a quasar host
are excluded, we nd that the mean color gradients for the two samples in
magnitudes arcsec?2 per decade in radius are:
* +

(B ? R) = ?0:20  0:05; (4.4)

(log r) radio
and * +

(B ? R) = ?0:23  0:05: (4.5)

(log r) control
The distribution of gradients is shown in Figure 4.6.
We nd that the numbers that we obtain are larger than those obtained
by other authors. For a sample of normal ellipticals (with dusty galaxies
excluded) Peletier et al. (1990) had obtained a color gradient of {0.1. Zirbel
(1996) had obtained a value of {0.15 for a sample of radio galaxies.
89
Figure 4.7: Dependence of color gradients on absolute magnitude. Fainter
galaxies have larger scatter.

90
4.2.1 Results for the samples
Figure 4.5 shows that their are important dierence between the radio and
control samples with respect to the scale length ratio. One can especially see
that a larger proportion of radio galaxies become bluer towards the center.
However, such a distinction can not be made on the basis of color gradients.
In the case of radio galaxies as well as the control sample galaxies we nd
that a majority have a negative color gradient i. e., they become redder as one
approaches the center. This is in conrmation with color gradients in galaxies
in general (Sandage and Visvanathan, 1978). A comparative distribution of
color gradients for the two samples is given in Table 4.6.
A few properties related to color gradients hold true for both the samples:
 For optically brighter galaxies there is a hint that the color gradient
increases linearly with brightness (Figure 4.7). This is consistent with
elliptical galaxy formation models based purely on dissipation (Carl-
berg, 1984). As clouds collapse and give rise to galaxies, larger galax-
ies (larger scale length) are also seen to possess larger color gradients.
However, the correlation coecients indicate that this dependence is
not signicant, the best correlation being that between re( kpc) in B
for radio galaxies: the correlation coecient is {0.31 with a signicance
level of 98%. The plots for the radio and control samples are shown in
Figures 4.8 and 4.9.
 For fainter galaxies a larger scatter is seen in the color gradient (Fig-
ure 4.7). Such a behavior has been noted before (Vader et al. , 1988).
It has been suggested that dwarf ellipticals which are closer to disk
galaxies are likely to show larger scatter.

91
Figure 4.8: Color gradients for radio galaxies as a function of linear scale
length. Galaxies with larger scale lengths have larger color gradients. This
is in conrmation with galaxy formation by pure dissipation as suggested by
Carlberg (1984). However, the correlation is seen to be weak (see text).

92
Figure 4.9: Color gradients for the control sample galaxies as a function of
scale length.

93
4.2.2 Color gradients and scale length ratios
Color gradients and scale length ratios are both indicative of the change in
color with distance from the center. We present below a relation between
the two quantities so that one can be obtained from the other.
Let B and R denote the B and R proles of the galaxy in Figure 4.3.
We can then write,
0 1
B (r) = B (0) + @ m A 1=4
1=4 r ; (4.6)
re;B
and 0 1
R(r) = B (0) + C (0) + @ m A 1=4
1=4 r ; (4.7)
re;R
where, C (r) = B (r) ? R (r), m is a constant and re;B and re;R are the bulge
scale lengths in B and R.
The surface brightness prole of a de Vaucouleurs galaxy is a straight line
in the r1=4 ?  plane. Equations 4.6 and 4.7 provide the slopes of B and R
1=4 1=4
surface brightness proles (m=re;B and m=re;R ). One can also obtain these
values by considering two points, (r1 ; (r1)) and (r2; (r2)), on the respec-
tive straight lines representing the B and R surface brightness proles in the
r1=4 ?  plane. One can then obtain a relation between the scale length ratio
and the color variation by equating the ratio of the slopes (of the B and R
surface brightness proles) obtained by the two methods:
! !4
re;B = 1 ?
C
re;R B (r2 ) ? B (r1) ; (4.8)
where,
C = C (r2) ? C (r1).
The change in color can also be obtained from the color gradient as fol-
lows:
G 
(
Blog? R) =
C :
r log(r2 =r1)
Thus,

C = G log(r2 =r1) (4.9)
94
Using the value of
C from Equation 4.9 in Equation 4.8 and since
C is
small, we have,
re;B  1 ? 4G log(r2=r1) (4.10)
re;R B (r2) ? B (r1)
which further gives us:
!
m (r21=4 ? r11=4 ) r e;B
G  1=4 1?
re;R (4.11)
4re;B log(r2 =r1)
Thus, for galaxies that obey de Vaucouleurs' law the color gradient can be
obtained directly using the tted bulge scale lengths. We show in Table 4.6
the color gradients obtained from the color prole as well as that obtained
from the scale lengths. The conventional color gradient does not greatly
distinguish between the radio and control samples, although here too the
control sample does have a higher proportion of objects becoming redder
towards the center. When we turn to the color gradients as obtained from
the scale lengths, we see that the radio and the control sample are clearly
separated into two dierent categories. The scale lengths indicate that a
large fraction ( 66%) of the control sample galaxies become redder towards
the center while a large proportion ( 60%) of the radio galaxies are seen to
become bluer towards the center.
Since identical processing has been carried out for the radio and control
samples, it is unlikely that a bias in selection or procedure leads to the sig-
nicant dierence between the two samples as indicated by the scale lengths.

4.2.3 Preference for scale lengths

We have seen above that while the color gradients do not seem to distinguish
between radio galaxies and the control sample to a great extent, the results
obtained from the tted scale lengths clearly bring out the dierence very
vividly. We examine here the reasons behind this.
Obtaining the color gradient involves tting a straight line to the B ? R
prole. As a result, small variations, especially small 'kinks' close to the
95
Table 4.6: Number of galaxies with negative and positive color gradient for
the two samples as obtained from (1) their color proles and (2) their bulge
scale lengths. A negative color gradient is indicative of the galaxy having a
redder center compared to its outer regions.
Color gradient From color prole From scale lengths
Radio Control Radio Control
<0 20 24 9 20
>0 7 6 18 10
Indeterminate 3 0 3 0

center, cannot aect the t. The proles show that the radio galaxies are
more 'kinky', but the straight line ts miss that fact. In principle these small
changes can be crucial. Also, the magnitude of the color gradient is of the
order of the errors on the color prole, especially in the outer region.
The procedure for obtaining the scale lengths does not suer from the
above defects. We obtain scale lengths by tting an empirical formula sep-
arately to the B and R band proles. As discussed earlier, the 2 obtained
are well within acceptable limits in most cases and thus indicative of good
ts. The added advantage of using the eective radius for such a comparison
is that it provides a single descriptor for large scale trends in the distribution
of light.

4.2.4 Discussion
It is clear from the preceding discussion that in our sample the control galax-
ies seem to behave as expected, the radio galaxies show a larger scatter in
the scale length ratio. A few galaxies show evidence of large reddening as
one approaches the center while a majority show a tendency of bluer colors

96
Figure 4.10: Scale length ratio against color gradients for the radio sample
(top) and the control sample (bottom).

97
towards the center. We interpret these eects as being due to dust and star
formation respectively.
Following earlier studies made using reddening-free color indices, it was
assumed that the color gradient in a galaxy is a result of the metallicity
gradient (combined with age gradient) alone. Though metallicity is indeed
the main cause of color gradients, there are other factors like dust (Wise
and Silva, 1996) and star formation that can aect the color gradient and
hence the scale lengths. The eect of dust and star formation are in general
complimentary in that dust tends to redden whereas star formation leads to
bluer colors. If a galaxy is without dust or young stars or has dust that is
well mixed with stars, re(B )  re(R). Given below are three simple-minded
dust and star formation scenarios that aect re.
1. Young stars being formed in the central but extended region of the
galaxy: In such a case there will be excess blue light from the center
and the radius of the aperture enclosing half of the blue light will be
smaller than the corresponding radius for the red light, thus reducing
the re(B ).
2. Extended dust in the outer regions: Excess dust in the outer region
will cause larger amount of blue light to be absorbed there which is
equivalent of a red excess. As a result, as in the case above, the radius
of the aperture enclosing half of the blue light will be smaller than the
corresponding radius for the red light, thus leading to a smaller re(B ).
3. Dust in the inner region: Blue light will be more absorbed near the
center, as a result of which the blue surface brightness will appear to
rise slower than the red light towards the center, increasing re(B ).
The following calculation shows the quantitative eect of dust in changing
the scale length of a galaxy.

98
 If the opacity   1, then the change in magnitude due to absorption

m  1:06 . If the dust aects only the shorter wavelengths (e. g. B band)
then absorption is equivalent to the change in color,
m =
C . This change
occurs near the center of the galaxy, typically over a decade of radius. The
resulting color gradient, for   0:1, as given by Equation 4.9 is
G = log(
rC=r ) 
m  ?0:106:
2 1
The equivalent change in scale lengths is  10% i. e., re(B )=re(R) = 1:10.
Since the eect of dust is not easily discernible authors often prefer to
publish scale lengths for a single lter even when data are available in more
than one. Dust alone does not seem to be sucient in giving rise to the color
gradients, but substantial amounts of it can certainly lead to reddening. In
particular for the radio galaxies that have re(B ) > re(R) we do nd evidence
of dust in the color maps (Chapter 6). Thus the scale length ratio is also a
good indicator of dust.

4.3 Extrapolated proles

The 1D model prole tting has been done using points of the surface bright-
ness prole which are further from the center than 1.5 times the FWHM of
the PSF. This means that we are not using points which are closer to the
center than  1:5to2:1 arcsec in the tting. Having obtained the best t
proles as described above, we have extrapolated them to the center of the
galaxy ( 0:1 arcsec). The model proles for such extrapolation have been
obtained after convolving the best t model galaxy with the observed PSF.
The idea behind the extrapolation is to examine the departure of the
inner part of the galaxy from the simple standard model which provides
a good t beyond the omitted points. There are three basic possibilities{
(1) the inner points lie on the extrapolation within 1; (2) the inner points
have greater surface brightness than the extrapolation (we interpret this as
99
Table 4.7: Details of the excess emission found near the center in radio
galaxies and the control sample.
Radio Control Interpretation
% cases % cases
1 Babs > Rabs 30 27 Dust near the center
2 Bexc > Rexc 17 7 Weak excess emission
3 Bexc < Rexc 13 10 Excess emission + dust
4 Babs and Rexc 17 0 Excess emission + dust.
Dust more dominant.
5 Bexc and Rabs 3 0 Excess emission + dust.
Excess more dominant.
6 Babs < Rabs 3 6 Excess emission + dust.
Excess more dominant.
7 3+4+5+6 36 16 Excess emission + dust
8 2+3+4+5+6 53 23 Excess emission
9 1+3+4+5+6 66 43 Dust

100
excess emission) and (3) the inner points have less surface brightness than
the extrapolation, in which case we say that there is absorption.
We have assumed all along that we are dealing with a galaxy that obeys de
Vaucouleurs' law and perhaps has an additional exponential disk. However,
this assumption need not be true. In that case, some of the eects that we
attribute to excess blue emission or star bursts would be upper limits.
Figure 4.11 shows the ts and excess emission for 0446-206. The extrap-
olated prole shows hardly any deviation from the observed prole. A small
disk (rs = 1 arcsec; D=B = 0:14 in both lters) is seen to be present. 0520-
289, shown in Figure 4.12 is a good example of a galaxy that exhibits excess
emission. The excess in R exceeds the excess in B indicating the presence of
dust too. The B ? R colormap conrms the presence of dust. An example of
dominant dust absorption is 1103-244, shown in Figure 4.13. The absorption
in B exceeds the absorption in R typical of a dusty galaxy. We describe the
excess related results for the two samples below. A summary is presented in
Table 4.7.
1. Inner points of the same galaxy sometimes have dierent types (1, 2 or
3 above) of behavior in dierent lters. We are able to interpret this
using a combination of excess blue emission and dust absorption acting
simultaneously.
2. In  30% of radio galaxies, we nd that there is greater absorption in B
than in R at the center. This is consistent with the presence of dust in
the center, as dust absorption would be greater at shorter wavelengths.
Dust is seen in all but one of these galaxies in the color maps that
we have discussed above. Such behavior is seen in 27% of the control
galaxies. In 3/4th of these control galaxies we see the dust in color
maps as well.
3. In  17% of the radio galaxies we see greater emission in B than in
R. We cannot say much about the spatial extent of this excess blue
101
Figure 4.11: The tted prole of 0446-206 shows very little deviation from
the model prole. B and R prole ts are shown in the left and right columns
respectively. The top row depicts the observed prole (shown by points) and
dierent components of the tted prole: The dotted line is the bulge, the
dashed line is the disk and the summed prole is indicated by a solid line.
The disk has a sub-PSF scale length and D=B = 0:14 in both the lters. The
middle panel shows the residuals
m = mobserved ? mfit . The error bars on
the observed prole are also shown. The bottom panel shows the fractional
cumulative excess upto a semi-major axis length r as a function of r.

102
Figure 4.12: 0520-289 shows excess emission near the center. Parameters
same as in Figure 4.11.

103
 emission, because its angular size is comparable to that of the PSF.
The excess blue emission could be due to the presence of an AGN
which is too weak, in the optical, for us to have found it in model
prole ts. Only  7% of the control galaxies have such excess blue
emission.
4. In  13% of the radio galaxies, we have excess emission in B as well
as in R. However, the excess in B is less than the excess in R. This is
easily understood as follows: there is excess blue emission, as in point
3. But there is also dust, which absorbs more of the blue light than
the R light. The net result is excess emission in both cases, with less
excess in B as observed. In the control sample, this behavior is seen in
 10% of the galaxies.
5. In  17% of the radio galaxies we have absorption in B but emission
in R. This can be interpreted as in point 4, with the eect of dust in
the B band dominating the excess in that band. Some of the excess
emission in R manages to get through the dust. There is no example
of this behavior in the control sample.
6. In one radio galaxy we see excess emission in B and absorption in R.
We again interpret this as a blue excess plus presence of dust. The
optical depth is large enough for the absorption to become evident in
R. But the excess blue emission dominates the dust in B. There is no
such example in the control sample.
7. In one radio galaxy, we see absorption in both B and R, but the absorp-
tion in B is less than the absorption in R. This can be interpreted as
in point 6. The dierence in the two cases is only quantitative. There
are two galaxies in the control sample with such behavior.
8. From points 4 to 7, we see that excess emission aected by absorption
occurs in  33% of the radio galaxies. In the case of the control sample,
104
Table 4.8: Absorption and excess emission details for the radio and control
samples.
B R Radio Control
Excess Excess 7 13
Excess Decit 2 2
Decit Excess 10 3
Decit Decit 8 12

we get this behavior in  17% of galaxies.

9. Overall, we have excess blue emission in  53% of radio galaxies, and
in  23% of the control galaxies.
We have quantied the excess emission and the absorption due to dust
by accumulating the dierence between the model galaxy prole and the
observed surface brightness prole in the inner region. The procedure carried
out is as follows: (1) Obtain a model galaxy prole through a t to the surface
brightness prole as indicated earlier. (2) Along the major axis obtain the
dierence between the observed prole and the model prole. (3) Starting
from the center, at each point add the dierence to the sum of the dierences
at all inner points i. e., obtain a cumulative total of the dierence. The excess
is then given by the ratio of the cumulative sum to the luminosity of the de
Vaucouleurs bulge. Absorption due to dust is indicated by a negative excess.
We present in Figure 4.14 the fractional cumulative excess for the two
samples in the B and R lters upto 2:5 arcsec. These values denote upper
limits since the deviation from which we have calculated the excess could
partly be due to a departure from de Vaucouleurs' law as well. A numerical
comparison is presented in Table 4.8.
For a galaxy having mass 1011 M, a 1% excess translates to a mass of
105
Figure 4.13: 1103-244 has excess B absorption indicative of dust. Parameters
same as in Figure 4.11.

106
Figure 4.14: Fractional cumulative excess values in B and R lters for the
radio sample and the control sample.

107
 109M which is large for a starburst but cannot be ruled out over a period
of, say, 107 Myr. Our attempt at time-dating the star formation is presented
in Chapter 8. The dust extinction will be discussed further in Chapter 6
where we discuss various issues related to dust.

4.4 Conclusions
In this chapter we have described our results of tting 1D surface brightness
proles of galaxies to the radio and control samples. In particular we have
shown that:
1. The distribution of the re(B )=re(R) ratio for the control sample and ra-
dio galaxies show important dierences. Control galaxies seem to have
a quiet history whereas the radio galaxies are likely to have undergone
recent star formation in the central region.
2. Color gradients are re ected in the scale length ratio. However, the
latter are more sensitive at describing the average variation of color in
galaxies.
3. Extrapolated proles conrm enhanced star formation in the central
regions of radio galaxies.
In the next chapter we turn to discussing the disky galaxies.

108
Chapter 5
Radio galaxies with disks
Radio sources not selected to reside in rich cluster environments are known
to be hosted by elliptical galaxies (e. g. Heckman et al. , 1986). Host galaxies
in rich clusters are more homogeneous in nature and show a lower frequency
of merger activity (Ledlow and Owen, 1995). However, recently many exam-
ples of powerful radio sources have been found where the host galaxy shows
appreciable deviation from being an elliptical. One important distinguishing
feature between ellipticals and non-ellipticals is the presence of a disk. In
this chapter we study the morphology of the radio galaxies as well as the
control galaxies with prominent disks.
We present data on ve radio galaxies from our sample with prominent
disks. We present the images and obtain bulge and disk scale lengths (re and
rs) as well as the relative strengths of the two components (D=B ) using the
prole tting technique described in Chapter 3. We also introduce in this
chapter the morphological gradient lter which is especially useful in locating
disks in galaxies.

109
Figure 5.1: D=B as a function of Hubble type (data from Kent, 1985). El-
liptical galaxies do not have a disk component as per this classication. It is
the means for the respective galaxy types that have been plotted here. Later
type galaxies are seen to have a larger D=B ratio.

5.1 Disky galaxies

In Chapter 3 we have explained the procedure we follow to decompose the
surface brightness prole of a galaxy into de Vaucouleurs bulge (B ) and an
exponential disk (D) component. The ratio of the luminosities of the two
components, D=B , is the factor that indicates the presence of a disk. In
Figure 5.1 we show the typical D=B values for galaxies of dierent morpho-
logical types (Kent, 1985). As per Hubble's original classication scheme
(Hubble, 1926), elliptical galaxies do not have any structure. Whenever a
disk like component was suspected the galaxy was labeled S0. Recent studies
have shown, however, that elliptical galaxies can also be disky. It is likely
that such galaxies may be the lower end, in diskiness, of S0 galaxies.

110
Figure 5.2: D=B distribution for the radio and control samples in B and R
lters. There are ve radio galaxies with D=B > 0:3 (upper right quadrant)
and with rs > FWHM of the PSF in both B and R lters. 4 control sample
galaxies have D=B > 0:3 in both B and R lters.
111
Figure 5.3: Histogram of D=B values in the R lter for the radio and control
samples. The dotted line denotes D=B = 0:3. The disk parameters for the
disky radio galaxies are more signicant than for their counterparts in the
control sample (see text).

112
 In Figure 5.2 we show the distribution of the D=B ratio for the radio and
control samples. For the R lter we also depict the distribution as a histogram
for both the samples in Figure 5.3. It is seen that in general the D=B values
in both the lters are similar. Several galaxies in the radio as well as control
samples, have a D=B < 0:1. These are likely to be insignicant disks possibly
arising from deviation form the r1=4 law or a weak point source (when rs is
also small). We notice that there are a few galaxies with a D=B > 0:3 in
both the lters. In this section we consider these galaxies. The value of
D=B > 0:3 was chosen so that even if the scatter around the D=B values of
dierent galaxy types plotted in Figure 5.1 is taken into consideration, the
galaxies with these values will not be elliptical. In the radio sample, there
are ve such galaxies, while in the control sample there are four galaxies with
this lower limit for the ratio.
There is also a class of objects for which D=B > 0:3 but such that rs <
FWHM of the PSF. Such cases are likely to be the result of a genuine depar-
ture from de Vaucouleurs' law near the center or indicative of the presence
of a weak AGN. We do not consider these galaxies to be disky in the nor-
mal sense of the word and have not treated them as being signicant in the
following discussion.

5.1.1 Radio galaxies

We present here results on 5 radio galaxies with prominent disks. We examine
the host galaxy properties and the radio structure. 4 of the 5 galaxies have
FR II morphology and even the fth one is of type FR II?. The D=B ratio
is suggestive of spiral galaxies. However, we do not see spiral arms except in
one case. Relevant parameters for these galaxies are listed in Table 5.1.

1222-252 This galaxy is at a redshift of 0.075 and is by far the most in-
teresting one. It has S408 = 1:820 Jy and P408 = 1025:34 W Hz?1 . The radio
source is of type FR II. The morphology in the two lters shows a bar like
113
structure along with a linear structure at right angle to the 'bar', arising from
one end of the bar. There are also other occulent structures present. D=B
values are consistent with the galaxy being an Sb. The disk scale length is
in this case a few times the FWHM of the PSF, but only a fraction of the
bulge scale length. We are therefore again detecting a small disk relative to
the bulge.

1053-282 This galaxy is at a redshift of 0.061, has S408 = 3:510 Jy and

P408 = 1025:41 W Hz?1. The ellipticity of this galaxy changes from 0.1 near
the center to 0.4 in the outer region. There is a spiral galaxy 1500 away with
which the radio galaxy is likely to be interacting. The D=B values in both
lters are > 4. From this, and the disk and bulge scale lengths, the galaxy
can be classied as Sc. This is the most likely candidate from our radio
sample to be a normal spiral as dened by bulge and disk parameters. No
spiral structure is seen.

1006-214 This galaxy is at a redshift of 0.250. The radio source is of

type FR II with S408 = 1:350 Jy and P408 = 1026:23 W Hz?1 . The galaxy is
attened near the center (  0:55 near the center compared to   0:44 on
the outskirts). It has a bluer center and an armlike extension is seen along
the radio axis. The K 0 image does not show the armlike structure and hence
it is likely to be the outcome of recent star forming activity. This galaxy has
large D=B (> 5) in both lters and on that basis alone would be classied
as Sc. However the disk scale length is comparable to the FWHM of the
PSF. It is also much smaller than re in the R lter. In the B lter, re and
rs are comparable in size. However, the smaller bulge scale length in this
lter, relative to re in the R lter, is suspected to be a result of excess blue
emission in the central region. Taking all this into account, we conclude that
the prole t has detected a small scale disk-like structure. The linear scale
length of this structure in R is  3 kpc, and it is much smaller than the

114
bulge scale length of  20 kpc. The ratio of the disk to bulge scale lengths
is therefore much smaller than is observed for standard disk galaxies.

0952-224 This galaxy is at a redshift of 0.230 and is in a crowded region.

It has a 408 MHz radio ux S408 = 1:710 Jy and has a radio luminosity
of 1026:26 W Hz?1. The radio source is of type FR II. This galaxy has a
(D=B )red = 2:54  0:33, which is large enough for it to be classied as an
Sb spiral. However (D=B )blue = 0:39  0:01 is considerably smaller. This
dierence in parameter values is also seen in rs, as well as in bulge parameters.
Due to the dierence in values from one lter to another, it is not possible
to condently classify the galaxy. The dierences are caused by the presence
of dust as well as blue emission, which are seen in the color maps and prole
ts. We show the galaxy in Figure 5.1.2.

1215-215 This galaxy is at a redshift of 0.075. It has S408 = 1:500 Jy and

P408 = 1025:22 W Hz?1 . D=B values in B and R(> 1) are consistent with SA
classication. rs in B and R is comparable to the FWHM of the PSF.

5.1.2 Control Sample

We describe here objects from the control sample that have D=B > 0:3 in
both lters. The control galaxies have been chosen to be of early type and
we expect a few S0s but no spiral galaxies. Relevant parameters for the 4
disky galaxies are given in Table 5.2.

1251/190 In this case the D=B values are consistent with the galaxy being
a lenticular, but the error bars are large, and the D=B value, at least in the
B lter, is consistent with zero. The disk scale lengths are smaller than the
FWHM of the psf.

115
 Table 5.1: Radio galaxies with D=B > 0:3 in B or R or both
IAU z D=B Type rs ( arcsec) re ( arcsec) 2 Radio
B R B R B R B R
0952-224 0.230 0.39  0.01 2.54  0.33 Sb+ 0.8  0.08 12.3  1.50 25.7  0.90 2.3  0.1 1.31 2.80 FR II
1006-214 0.250 8.80  1.10 5.30  1.40 Sc+ 1.7  0.09 1.4  0.05 1.8  0.70 9.3  2.4 1.19 1.14 FR II
1053-282 0.061 4.30  0.35 5.50  0.79 Sc+ 24.8  1.90 31.7  3.60 9.2  0.20 8.3  0.3 0.40 2.24 FR II?
1215-215 0.075 1.44  0.16 1.04  0.46 Sa+ 1.2  0.04 1.5  0.12 4.3  0.13 11.2  3.5 1.97 0.16 FR II
1222-252 0.077 1.49  0.44 2.50  0.07 Sb+ 4.9  0.34 3.3  0.17 48.9  7.10 21.6  0.3 1.12 1.03 FR II
116

Table 5.2: Control galaxies with D=B > 0:3 in B and R

Object D=B Type re ( arcsec) rs ( arcsec) 2
B R B R B R B R
1251/190 0.46  0.33 0.35  0.11 S0 1.01  0.16 0.60  0.04 4.13  0.54 6.11  0.92 0.26 1.05
1323/1477 4.85  3.31 8.22  0.61 Sc+ 3.44  0.06 3.21  0.03 101.13  3.54 63.60  3.09 7.53 5.76
1323/1673 0.52  0.43 0.46  0.01 S0 3.36  0.23 4.06  0.20 9.44  0.92 11.71  0.22 1.42 3.05
1329/234 0.33  0.01 1.65  0.09 Sa+ 1.78  0.05 1.58  0.03 53.71  0.67 19.22  0.51 0.32 0.55
1323/1477 This galaxy has large D=B and appreciable bulge and disk
scale lengths in both lters. However, the t is poor in both lters.

1323/1673 In this case the t is poor in R, tolerable in B . In the latter,

the D=B value is consistent with zero.

1329/234 The disk in this case has a scale length comparable to the
FWHM of the psf, rs  re. The D=B in B is consistent with a lenticu-
lar and D=B in R consistent with a SA.

5.2 Morphological Signatures of disks

In Appendix A (Page 169) we describe the morphological image processing
techniques that we have applied to our galaxies to look for disks. For the
disks the morphological gradient technique is the most suitable. A disk,
when it is present, exists along with the underlying bulge. This additional
component provides a contrast over the 'background' of the bulge. It is this
contrast that the gradient lter is able to detect.
We have applied the gradient to all our galaxies. The lter is able to
bring out features like spiral arms and elongated structures especially well.
For normal elliptical galaxies, the lter does not introduce any articial struc-
tures. The images of the radio galaxies with disks, after processing with the
morphological gradient lter, are shown in Figure 5.2 along with that of a
normal galaxy. The lter involves manipulating the image and hence after
processing it is not straightforward to carry out photometry on the resultant
image.
We see that the spiral features of 1222-252 come out better after the
gradient processing. The disk in 1006-214 is also brought out well by the
lter with the blue feature possibly due to emission lines standing out at
one end. In 1053-282 we see an elongated disk like structure at the center

117
Figure 5.4: Filled contours of the disky radio galaxies in B (left) and R (right)
bands. North is to the top and east is to the left. The faintest contour level
in magnitudes is denoted in the top right corner with contour interval in the
bracket. (top): 0952-224; (middle): 1006-214; (bottom): 1053-282

118
Figure 5.5: Filled contours of the disky radio galaxies in B (left) and R (right)
bands. North is to the top and east is to the left. The faintest contour level
in magnitudes is denoted in the top right corner with contour interval in the
bracket. (top): 1215-215; (bottom): 1222-252

119
of the galaxy. However, it is not visible in the outer parts of the galaxy.
In 1215-215, hint of a spiral structure is seen to the north. In 0952-224 the
disky structure is seen north-south.

5.3 Discussion
Strong radio sources have been always identied with elliptical galaxies or
recent merger remnants. Even recent studies have pointed to this (Urry and
Padovani, 1995). When radio sources have been identied with spiral like
galaxies, the radio emission has been seen to be conned to the nucleus and
the disk rather than have extended structure. The emission is dominated by
thermal emission from HII regions and supernovae winds.
Fanaro and Riley (1974) classied powerful radio galaxies into FR I and
FR II types based on their radio morphology. In both types, the extended
radio structure is fueled by jets that transport the radio plasma from the
nucleus.
Seyfert galaxies are almost always associated with spiral hosts and in
some cases the radio emission is seen to be upto  100 kpc. Seyferts are
towards the low luminosity end of the radio galaxies and may form the bridge
between powerful radio galaxies and spirals. Meurs and Wilson (1984) have
studied the radio luminosity function of Seyferts, radio galaxies and spirals
and suggest that the luminosity functions do not smoothly merge.
We have shown here that in our radio galaxy sample there are ve sources
( 17%) which can be classied as disky galaxies. (There are additional
sources with D=B > 0:3 but with rs < FWHM of the PSF. We do not
consider these to be disky galaxies, since the \disk" cannot be distinguished
from a point source.) The optical morphology of these objects is complex.
Spiral arm like features are visible in only one of them. Examples have
been reported before of spiral galaxies being associated with FR II radio
sources. However, these have turned out to be chance superpositions (see

120
0952-224 1006-214

1053-282 1215-215

1222-252 0434-225

Figure 5.6: The morphological gradient images in B lter of ve radio galax-

ies and a normal galaxy. North is to the top and east is to the left. Top:
0952-224 (left), 1006-214 (right); middle: 1053-282 (left), 1215-215 (right);
bottom: 1222-252 (left), 0434-225 (right). 0434-225 is the normal galaxy.

121
e. g. Ronnback and Shaver, 1997). However, 1222-252 clearly seems to have
spiral like features.
Previous claims related to detection of disks in powerful radio galaxies
have been controversial e. g. 3C 293 and 3C 305 both have a disk of ro-
tating emission line gas but no evidence for a stellar disk (Heckman et al. ,
1985). 0313-192 in Abell 428 is an FR I source reported to be hosted by a
disk dominated galaxy showing spiral arms in the B band (Ledlow et al. ,
1998). Barring a possible exception of 1053-282, all the disk dominated radio
galaxies in our sample are FR II sources.
The D=B classication puts these galaxies in the Sa+ categories. How-
ever, when the re=rs ratio is considered, it is clear that the disks present
in these galaxies are of a dierent nature than in normal spiral galaxies, in
which rs  re.
Except in one case, we do not see spiral structure in case of the disky radio
galaxies. However, we have seen above that two more galaxies show hints of
spiral structure after they are operated on with the morphological gradient
lter. An interesting example of the ecacy of the lter is the anonymous
galaxy shown in Figure 5.7. We believed at one time that this galaxy was
the host of the radio source MRC 0503-284. We found subsequently that
this was a misidentication, and the anonymous galaxy was dropped from
the sample. Figure 5.7 shows the direct B image and the morphologically
processed image of this galaxy. The direct image clearly shows the presence
of a spiral structure to the south-east. There is a hint of an asymmetry
towards the north-west. The spiral structure in the south-east is seen much
more clearly after the image has passed through the morphological gradient.
In addition, the corresponding spiral structure in the north-west also stands
out. Thus, the morphological gradient operator is clearly of help in detecting
spiral structure.

122
Figure 5.7: Images of a disky galaxy previously misidentied and thought
to be the host of MRC 0503-284. The B image (top) shows the presence of
a spiral structure to the south-east. The morphological gradient brings out
the spiral structure more clearly and indicates the presence of a similar, but
weaker structure to the north-west. North is to the top and east to the left.

123
Chapter 6
Dusty galaxies
6.1 Introduction
As per the original classication of Hubble, elliptical galaxies were entirely
dust free. If observations indicated that a galaxy contained some amount of
dust, it was classied as an S0 or later type. High dynamic range imaging
and improved detection techniques over the last several years have indicated
that even classical elliptical galaxies can contain dust. Sadler and Gerhard
(1985) showed that this is true with  40% of ellipticals. More recently,
van Dokkum and Franx (1995) have inferred from a study based on HST
observations that 78%  16% early-type galaxies contain nuclear dust.
Though the dust can be in patchy structures, the more interesting and
frequently found conguration is that of a disk. Disks are seen in the form
of a dust lane when viewed nearly edge-on. Face on dust disks are likely to
go unnoticed unless the optical depth is suciently large.
A dust lane can be potentially useful in providing information about the
3D shape of the host galaxy. Elliptical galaxies are dynamically supported
by the anisotropy in velocity dispersion rather than rotation (Bertola and
Capaccioli, 1975; Illingworth, 1977). This results in the galaxy being triaxial
rather than axisymmetric (Binney, 1976). Schwarzschild (1979) has shown

124
that closed stable orbits in a triaxial system exist in the planes perpendicular
to the short and long axes of the system. Orbits in a plane perpendicular to
the intermediate axis are unstable and a gaseous disk in that orientation will
not last. Thus, when a dust lane is seen, the possibilities for the 3D shape
of the galaxy narrow down. If an elliptical is spheroidal, a minor-axis dust
lane will mean that the galaxy is prolate and a major-axis dust lane that it
is oblate (see, e. g. Kormendy and Djorgovski, 1989).
The situation is, however, complicated by the following facts: (1) elliptical
galaxies are, in general, triaxial, (2) the dust lane, if of external origin, could
be the result of gas trapped in a polar orbit and (3) the galaxy can have some
amount of rotation. Observations indicate that many dust lanes are warped
in the outer region indicating that the dust in the inner parts is settled in
a preferred plane but the dust in the outer parts has not settled into an
equilibrium plane yet. This further indicates that the dust is of an external
origin. The most plausible explanation is a merger. There is kinematic
evidence indicating that dust disks are often at random orientations with
respect to the major axis (e. g. Gallagher et al. , 1977) and can even be
counterrotating.
Radio galaxies with dust lanes are even more interesting since in addition
to the dust lane, we have the radio axis to provide us directional information.
Kotanyi and Ekers (1979) showed after studying a sample of 8 radio ellipti-
cals with dust lanes that there is a strong tendency of the dust lane to be
perpendicular to the radio source. Following studies of a sample of 26 ellipti-
cal galaxies with prominent dust lanes Mollenho et al. (1992) have further
conrmed that there exists a strong tendency for the radio axis to be aligned
perpendicular to the dust lane whereas the angle between the radio axis and
the major axis of the galaxy is consistent with being random. Though the
accretion disk associated with the AGN need not be aligned with the stellar
body or the large scale gaseous disk, it seems that it nevertheless is associated
with the latter. In fact, it was seen that 21 of these 26 dust lane ellipticals

125
were radio emitters (L1:5 GHz = 1020 ? 6  1022 W Hz?1 ) suggesting that the
existence of a dust lane itself may be very closely related to the host galaxy
being a radio galaxy. The selection criterion for this sample was simply that
each galaxy contain a prominent dust lane. For normal ellipticals, the rate
of radio detection is much smaller (64/216; Wrobel and Heeschen, 1991).
Dust lanes have been one of the more important triggers for the study that
we present in this thesis. In this chapter we present the dust conguration
and morphology of various galaxies in the radio sample with a particular
emphasis on the six galaxies where we detect dust lanes. In Section 6.2
we describe the various techniques that we have used for detecting dust,
in Section 6.3 we present details of the dust lane galaxies in the sample
and provide various related details including the estimated mass of dust
and extinction coecient. Finally we discuss the orientations and intrinsic
shapes.

6.2 Detection techniques

6.2.1 Color maps
Constructing a color map from images of a galaxy obtained in two dierent
lters is the most direct way of looking for dust using broadband imaging.
The B ? R color map is obtained using a procedure analogous to obtaining
the B ? R color at each pixel and can be described as follows: (1) degrade
the image which has the PSF with the smaller FWHM to the PSF with the
larger FWHM. It is the R images which most often have the smaller FWHM;
(2) divide the B image by the R image and convert the quotient image to
the magnitude scale.
Because of the greater absorption at shorter wavelengths, the B image of
a galaxy is more likely to be distorted by the presence of a dust feature than
the R image. The eect of obtaining a color map is to remove the smooth

126
background provided by the longer wavelength R image from the B image,
and highlighting the dust aected regions. A color map can thus reveal dust
features (as also blue features) anywhere in the galaxy.
We have obtained B ? R images for all the galaxies in the radio and
control samples. We nd that six ( 20%) radio galaxies have prominent
dust lanes (E (B ? R) > 0:1, dimension > 10 arcsec). Five ( 17%) more
radio galaxies have prominent dust patches (E (B ? R) > 0:1, size between
3 and 4 arcsec). For the control sample we nd that only  7% galaxies
have dust lanes, but  30% galaxies have dust patches. Thus, the number
of galaxies with dust features in both the samples is the same ( 37%), but
the dust in radio galaxies is more organized. The possible scenarios that can
explain this are: (1) dust in radio galaxies has recently settled into a plane
while the dust in normal ellipticals has passed that phase and has now settled
in the central region following the dissipation of angular momentum. This
implies that normal ellipticals have taken their time to settle down, that all
elliptical galaxies pass through an active phase and that the dust is accreted.
This is supported by the observation that at larger semi-major axis lengths,
all dust lanes are seen to be warped indicating that the outer regions are still
in the process of settling while the dust in the central region is better settled,
and (2) the dust in lanes and disks is accreted while the dust found at the
centers of galaxies is locally generated from mass loss. Detailed kinematic
observations of dust lanes could resolve the issue.
The dust properties inferred from color maps are summarized in Table 6.1
and the B and B ?R images of the radio galaxies with dust lanes are presented
in Figures 6.1 and 6.2.

6.2.2 Small scale features

Another way of looking for dust using the B and R images, recommended by
Sparks et al. (1985), is to smooth the R image and then divide the B image
with it before proceeding to convert it to the magnitude scale. This helps
127
Table 6.1: Statistics of objects with prominent dust features (E (B ?R) > 0:1)
in the radio and control samples as indicated by B ? R color maps and of
nuclear dust as indicated by extrapolated proles.

radio control
sample sample
Dust lanes 20% 7%
Dust patches 17% 30%
Dust in the center
(from extrapolated
proles) 66% 43%

to smooth out any distortions that may be present in the R image and the
distortions in the B image are better highlighted. The resulting smoothed
color map is qualitatively similar to the ordinary color map but the features
are sharper.
Yet another way to obtain the distortions introduced in a B image by the
presence of dust is to obtain a residual image by subtracting out a smooth
model of the image from the original image:
Iresidual = Ioriginal ? Ismooth: (6.1)
The smooth model obtained by interpolation of the surface brightness, ellip-
ticity and position angle proles provided by ellipse following the tting of
ellipses to the galaxy isophotes.
The disadvantage in the two methods outlined here is that the data no
longer remain calibrated and tasks related to photometry cannot be carried
out. In case of the radio and control samples, we have obtained residuals as
outlined above, but have used them only for the purpose of locating features.

128
When a feature was found, we always went back to the B ? R color map or
the original B and R images to obtain quantitative estimates.

6.2.3 Extrapolated proles

We have described in Section 4.3 how we t the surface brightness prole of
a galaxy by leaving out the central region, so that the high signal-to-noise
region where de Vaucouleurs' law may not hold true does not aect the t.
We also described how we then extrapolate the ts inward to quantitatively
measure the departure of the observed prole from the tted prole. The
departure is a combination of eects due to star formation, dust and genuine
departures from de Vaucouleurs' law. We have shown in the lower part of
Table 4.7 that as per the extrapolated proles  66% radio galaxies are likely
to have nuclear dust, whereas the incidence is  43% for the control sample.
The extrapolation uses a prole averaged along elliptical isophotes. It can
hence detect dust that is spread over a region with optical depth insucient
to be seen by a color map. From the numbers given above, radio galaxies are
seen to have dust features near the center far more often than the control
sample galaxies. We present these numbers in Table 6.1 along with those
expressing the incidence of larger scale dust features deduced from the B ? R
color maps.

6.3 Dust lane galaxies

We present in this section images of the dust lane galaxies in the radio sample
along with various parameters related to the dust lanes. The parameters we
provide include the angle of (mis)alignment between the dust lane and the
major axis of the galaxy and between the dust lane and the radio axis. The
latter is sometimes ambiguous owing to the poor quality of the radio maps.
We estimate the color excess in the dust lane by comparing the color

129
0354-263

0420-263

1254-268

Figure 6.1: Color maps of dust lane radio galaxies with north to the top
and east to the left. (left): B image (right):B ? R image. (top): 0354-263
(middle): 0420-263 (bottom): 1254-268

130
1257-253

1323-271

1344-241

Figure 6.2: Color maps of dust lane radio galaxies with north to the top
and east to the left. (left): B image (right):B ? R image. (top): 1257-253
(middle): 1323-271 (bottom): 1344-241

131
in it with that of the surroundings. For this purpose we used a square
aperture with side 2:1 arcsec (3 pixels) on the B and R images separately to
extract magnitudes along the lane and in a neighboring region free of dust.
The mean of the color excesses, E (B ? R), obtained in this manner is then
used for further calculations. We use the standard relation from Savage and
Mathis (1979) to calculate the optical depths in B and R:
AB = 2:303  E (B ? R) (6.2)
AR = 1:303  E (B ? R) (6.3)
 = 1:09  A: (6.4)
Here,  denotes wavelength and A and  respectively denote the extinction
and optical depth at . We have assumed here that the dust lane obscures
all stars.
Following Burstein and Heiles (1978), the column density of neutral hy-
drogen and the total neutral hydrogen content in the dust disk can be ob-
tained from
N (HI) = 1:03Z  1022 E (B ? R)atoms cm?2 ; (6.5)
HI = D 2 (dust)N (HI)d
;
Ntot (6.6)
where, D is the distance to the galaxy. We assume that the dust-to-gas ratio
in these galaxies is similar to the ratio in our Galaxy, Mgas=Mdust  1:3  102,
to obtain the dust mass. This is a lower limit, since not all stars are actually
obscured and the actual amount is at least 1=f larger where f < 1 is the
fraction of stars obscured.
We describe below dust related features found in the dust lane galaxies.

0354-263 This is one of the two galaxies in the radio sample for which we do
not have an accurate redshift. The redshift is estimated to be 0.3 (McCarthy,
private communication). The radio source is of the FR II type and the
host galaxy is well t by de Vaucouleurs' law (2 < 2 with a disk having
132
rs < 1:5 arcsec). The dust lane extent is 6  4 arcsec2 with E (B ? R) ' 0:2
and Mdust ' 5:5  107M. The dust lane makes an angle of  45 deg with the
major axis of the galaxy and  45 deg with the radio axis as well. There is
a small galaxy to the south-east (see Figure 6.1) with which 0354-263 could
be associated. A possible tidal feature is seen. Owing to the uncertainty in
redshift, this object has been excluded from subsequent calculations in this
chapter.

0420-263 This galaxy is well t by de Vaucouleurs' law in the R lter but

provides a marginal t in the B lter (2 = 2:27). The size of the dust lane
is 7  4 arcsec2 with E (B ? R) ' 0:15 and Mdust ' 9:0  106M. The dust
lane makes an angle of  10 deg with the major axis of the galaxy. We do
not have a radio map for this object.

1254-268 This galaxy too provides a good de Vaucouleurs' law t in the

R lter but a poor t in B lter (rs(B ) is large). The dust lane extent is
7  3 arcsec2 and is rather diuse. It has E (B ? R) ' 0:1 and Mdust '
4:9  106M. In this galaxy the radio axis is approximately along the major
axis of the galaxy and the dust lane makes an angle of  30 deg with both.

1257-253 This galaxy is t well by de Vaucouleurs' law in the B lter but

not in the R lter. The radio source has FR I structure. The dust lane
extent is 10  3 arcsec2 with a E (B ? R) ' 0:1 and Mdust ' 7:7  105M .
In the case of this galaxy the major axis is approximately perpendicular to
the radio axis. The dust lane makes an angle of  30 deg with the radio axis
and  60 deg with the major axis of the galaxy.

1323-271 This is an FR II radio source, well t by de Vaucouleurs' law in

B and R lters. There is a companion galaxy, which was masked during the
tting. Lack of distortion in the outer isophotes of both galaxies suggests that
it is only a chance superposition rather than an interacting pair. The dust
133
Table 6.2: Properties of the dust lane radio galaxies from our sample. The
dust mass is calculated as described in Section 6.3. The dust lanes are, in
general, not perpendicular to the radio axis.
Object z size E (B ? R) B Mdust B4 dust?radio dust?ma
arcsec2 magnitude M max deg deg
0354-263 0.30+ 64 0.20 0.42 3  107 -0.010 45 45
0420-263 0.130 74 0.15 0.32 2  106 -0.010 NA 10
1254-268 0.135 73 0.10 0.21 2  106 -0.003 30 30
1257-253 0.045 10  3 0.10 0.21 1  106 -0.005 30 60
1323-271 0.044 10  4 0.10 0.21 1  106 -0.003 30 00
1344-241 0.019 10  4 0.07 0.21 1  105 -0.005 00 00

lane extent is 10  4 arcsec2 with a E (B ? R) ' 0:1 and Mdust ' 9:9  105M .
In this case the dust lane in the galaxy is approximately along the galaxy
major axis while the radio axis makes an angle of  30 deg with both.

1344-241 This is most likely an S0 galaxy, well t by de Vaucouleurs' law in

B but not in R. It has a large dust lane (10  4 arcsec2 ) with E (B ? R) ' 0:07
and Mdust ' 1:3  105M. The dust lane is curved and along the major axis
of the galaxy. The radio structure is not very well determined.

6.4 Discussion and conclusions

From the analysis presented so far in this chapter, it is seen that dust is
present at the centers of radio galaxies more frequently than in the control
sample galaxies. Also, the dust is more likely to be organized in radio galax-
ies. Previous studies (e. g. Kotanyi and Ekers, 1979; Mollenho et al. , 1992;
van Dokkum and Franx, 1995) have indicated that dust lanes are likely to
be perpendicular to the radio axis.
However, from the small number of dust lane galaxies in our sample,

134
Figure 6.3: Correlation of dust mass in radio galaxies with dust lanes (lled
diamonds) or dust patches (open diamonds) at the center with absolute B
magnitude (top) and radio power (bottom). The dust mass can have an error
of upto 50% owing to the uncertainty of the dust orientation and the fraction
of stars that are obscured. The dust mass for dust lane galaxies is seen to
correlate strongly with the radio power but only weakly with the absolute
B magnitude. Neither correlation is very strong for the galaxies with dust
patches. The correlation coecients are listed in Table 6.3.
135
Table 6.3: Correlation coecients for dust mass in galaxies with dust lanes
and dust patches with radio power and absolute B magnitude. Five points
each were available for the dust lane and dust patch correlations. The com-
bined 10 points were used for the lanes+patches correlation. The condence
value for each correlation is also shown. The correlation of radio power with
dust mass in dust lane galaxies is the strongest.

dust lanes dust patches lanes + patches

correlation two-sided correlation two-sided correlation two-sided
coecient condence coecient condence coecient condence
radio-dustmass 1.00 0.99 0.20 0.38 0.64 0.99
optical-dustmass 0.60 0.86 0.60 0.86 0.51 0.96
radio-optical 0.60 0.86 0.20 0.38 0.24 0.67

we do not nd such a correlation (see Table 6.2). Though it is an attractive

scenario to be able to link the AGN with the large scale dust lanes, dynamical
considerations do not necessiate it: Rees (1978; 1984) has shown that the
direction of out ow can be insensitive to the direction of angular momentum
vector of the infalling matter owing to the Lense-Thirring precession of the
inertia frame around the black hole.
We have investigated the possibility of the dust content being related to
radio properties. We show in Figure 6.3 the dependence of the dust mass on
the radio power at 408 MHz as well as on the absolute B magnitude. We
have included here dust lane galaxies (shown by lled diamonds) as well as
galaxies with dust patches at the center (shown by open diamonds). We nd
that the dust mass is correlated with the radio power and that more powerful
radio galaxies contain a larger amount of dust. The correlation coecients
are listed in Table 6.3. Owing to the ambiguity in orientation parameters,
the estimated dust mass can have an error of upto 50%. We have veried that
the correlation is not an artifact introduced by the presence of the distance

136
term in radio power as well as dust mass. The correlation of dust mass with
the absolute B magnitude is weak. This latter result is in conrmation of
similar earlier results (e. g. Goudfrooij, 1994).
The picture that seems to emerge is that the incidence of dust as well as
its amount is correlated with radio emission in a galaxy. However, either the
dust lanes that we see have not dynamically stabilized, or their orientations
bear no correlation with the radio axis. Before accepting this conclusion,
however, it should be remembered that we have only six dust lane galaxies
in our radio sample and that the radio maps are not very good, which makes
the radio axis directions rather ambiguous.

137
Chapter 7
Radio structure: using a
dierent eye
7.1 Introduction
In the initial days of radio astronomy, several point sources of radio emission
were located. These objects came to be called radio stars. As the angular
resolution of the radio telescopes improved, it was seen that some of the
point sources were in fact extended in nature, and could be identied either
with gas clouds in our Galaxy or with other galaxies. Cygnus A was the rst
radio source to be associated with another galaxy. Since then we have learnt
a great deal about radio as well as optical morphology of radio galaxies.
In this chapter, we describe the radio properties of elliptical galaxies. We
describe the classication scheme of radio galaxies with examples from the
radio sample. We study the relationship of the radio properties of the ra-
dio galaxies with their optical properties. In particular, we show the strong
dependence of properties like dust content, diskyness, excess blue and mor-
phological peculiarities on the radio power.

138
Radio sources and ellipticals
Operationally, a galaxy is said to be a radio galaxy if its radio luminosity
exceeds the optical luminosity (LR=Lopt >1). However, the criteria for sep-
arating radio loud objects from radio quiet objects tends to be subjective
and gives rise to a boundary that is not very sharp. Powerful radio sources
(Lrad  1045 erg sec?1 ) have radio emission extending to Mpc scales. The
emission is non-thermal in nature and arises from relativistic electrons accel-
erated in a magnetic eld.
Objects that are not \radio loud" are not totally radio quiet. All ellip-
tical and spiral galaxies have some amount of radio emission at a level of
 1037 erg sec?1 . Radio emission from normal ellipticals almost always orig-
inates from a small, unresolved region at the center of the galaxy. In spiral
galaxies the radio emission primarily traces star formation in the disk. It
can either be non-thermal in nature (supernovae) or thermal in nature (H
II regions). Seyferts and star burst galaxies typically emit  1040 erg sec?1 .
For the powerful radio sources, the typical structure consists of two giant
radio lobes extending to several 100 kpc straddling a central galaxy. Fa-
naro and Riley (1974), based on a sample of 57 radio galaxies and quasars
from the 3CR catalogue, showed that the radio objects can be divided into
two distinct classes, based on the relative positions of regions of high and
low surface brightness in the radio lobes. They used the ratio RFR, of the
distance between the regions of highest surface brightness on opposite sides
of the central galaxy or quasar, to the total extent of the source upto the
lowest brightness contour in the map. Sources with RFR < 0:5 were called
FR I sources (also called edge-darkened sources) and sources with RFR > 0:5
were called FR II sources (also called edge-brightened sources). An example
of FR I type, from the 3C catalogue, 3C 449, is shown in Figure 7.1. An
example from our sample, 1346-252, is shown in Figure 7.2. An example of
FR II type, again from the 3C catalogue, 3C 47, is shown in Figure 7.3 and
an example from our sample, 1222-252, is shown in Figure 7.4.

139
Figure 7.1: 3C 449 is an FR I galaxy. This is a VLA map at 1.465 GHz
(Perley et al. , 1979)

140
Figure 7.2: MRC 1346-252: an example of an FR I galaxy.

141
Figure 7.3: 3C 47 is an FR II quasar. This is a VLA map at 5 GHz (Bridle
et al. , 1994)
142
Figure 7.4: MRC 1222-252: an example of an FR II galaxy.

143
 It turns out that this is a very basic distinction and is closely related to
several other properties, indicating a direct link between luminosity and the
energy transport and conversion mechanism. An important correlation is
that most FR II sources have radio luminosity L178 > 21025 W Hz?1 str?1 ,
while most FR I sources have L178 < 21025 W Hz?1 str?1 . This dividing
line is not very sharp and the luminosity overlap at higher frequencies can be
almost two orders of magnitude. FR I galaxies have a steeper radio spectral
index than FR II galaxies. The mean spectral index for FR II galaxies is
0.89 while that for FR I is 1.07. The incidence of jets in FR I type of sources
is common, while < 10% of FR II sources have jets. Owen and Laing (1989)
introduced the fat double class as being intermediate between the FR I and
FR II classes since several galaxies could not be unambiguously classied
in the original scheme. Fat double sources have characteristics of both FR
classes. Owen and White (1991) showed that the FR I/II division depended
not only on radio power but also on the optical luminosity. Theoretical
arguments have now been provided (Bicknell, 1995) for the slope of the line
in the radio/optical plane which separates the two classes.
FR I sources are found to be located in denser environments as compared
to FR II sources, though the immediate environment for both type of sources
is seen to be similar (Fanti, 1984; Ledlow and Owen, 1996).
Radio sources are found 3{5 times more frequently in distant clusters
(z > 0:2) than in nearby clusters (z < 0:1) (Owen et al. , 1996). Cluster
merging is seen at higher redshifts and there may be a relation between the
increased activity and the cluster merging and is likely to be related to blue
galaxies.
Heckman et al. (1986) have shown using a sample of 43 radio galaxies
that  30% of the powerful radio galaxies in the sample exhibit disturbed
optical morphologies. They found a lower incidence (< 10%) of such features
in less luminous radio galaxies (1023:5 < P408 < 1025:5 W Hz?1 ). Later work,
e. g. Gonzalez-Serrano et al. (1993), on low-luminosity radio galaxies however

144
showed that as many as  75% of the galaxies show peculiar optical features.
In the following sections we present our ndings related to the incidence
of peculiar features in the two FR classes as well as other correlations. We
nd in general that our results are intermediate between the two mentioned
above.

7.2 Optical-radio correlations

In this section we consider the correlations between the radio properties of
radio galaxies with their optical properties. The radio properties that have
been used are (1) the power at 408 MHz; (2) the spectral index between
408 MHz and 5000 MHz (when the ux is not available at 5000 MHz, the
ux at 1400 MHz has been used to calculate the spectral index) and (3) the
Fanaro-Riley (FR) type as determined from the radio maps at 408 MHz. It
has not been possible to extract detailed information from the available radio
maps because of their low resolution. In several cases even the Fanaro-Riley
classication cannot be readily made. In the plots we display FR I sources
by open circles, FR II sources by lled circles, and by diamonds radio sources
that cannot be classied as FR I or FR II due to lack of detail.
The optical properties that have been considered are (1) the extent to
which the galaxy becomes blue or red inwards, characterized by the re(B )=re(R)
ratio; (2) the D=B ratio and (3) the R magnitude. While the rst two quan-
tities are obtained by tting de Vaucouleurs' law to the radial prole of the
galaxy, the R magnitude is obtained by using aperture photometry.
We also consider the correlation of FR type with isophotal twists, ampli-
tudes of harmonic components obtained during ellipse ts, presence of dust
and morphological features like bridges and arcs.

145
7.2.1 Disky galaxies
As has been mentioned several times before, powerful radio galaxies have
always been identied with early-type galaxies. Though these galaxies often
have disturbed morphologies, there have been few claims of a spiral host
galaxy. Almost all such claims so far have turned out to be incorrect (a few
remain to be veried). In a few cases a nearby galaxy is superimposed over
the radio source (e. g. Shaver et al. , 1983; Ronnback and Shaver, 1997). In
other cases line emission give rise to spiral-like features (e. g. Holloway et al. ,
1996). S0? galaxies have also been reported to be hosts of powerful radio
sources (Zirbel, 1996). However, these do not possess spiral arms.
In Chapter 5 we have discussed that in the radio galaxy sample there are
5 galaxies having D=B > 0:3 in the B and R bands including one (1222-252)
which appears to have spiral structures. We consider in this subsection the
correlation of the D=B parameter with the radio power at 408 MHz. In
Figure 7.5 we have plotted D=B in the R band against the radio power at
408 MHz. The gure shows that there are seven galaxies with (D=B )R > 0:3.
Two of these galaxies have rs < FWHM, indicating that their 'disks' are very
small and could well be simply a departure from de Vaucouleurs' law or a
weak AGN. Four of the remaining ve possess an FR II radio morphology (as
determined by eye from the available radio maps). The radio morphology of
the fth source is not well determined and is likely to be FR II. It is therefore
seen that the hosts of FR II sources are more likely to show \diskiness" than
the hosts of FR I sources.

7.2.2 Blue centers

In Chapter 3 we have shown that galaxies with smaller re(B )=re(R) values
are bluer towards the center. Figure 7.6 shows the scale length ratio versus
radio power at 408 MHz. It is seen from the plot that the more powerful a
radio galaxy is, the bluer it is towards the center. The correlation coecient

146
Figure 7.5: Disk-to-bulge ratio in the R lter as a function of radio power.
The distribution is similar for the B band. Of the seven objects with sig-
nicant D=B , all four of type FR II have signicant rs as well. Only one of
the remaining three has a signicant rs. Disks do not seem to prefer galaxies
that host radio sources of type FR I.

147
 Table 7.1: Optical properties of FR I and FR II galaxies.

MB;24 MR;24 < B ? R >

FR I -19.27 -21.72  1:8
FR II -18.55 -20.99  1:5
FR I + FR II -18.65 -21.03  1:6

is {0.23 for the 24 objects, and the observed correlation is signicant at the
93% level.

7.2.3 Optical luminosity

Figure 7.7 shows absolute R magnitude against power at 408 MHz for the
radio galaxies. We conrm the observation of others (e. g. Lilly and Prestage,
1987) that FR I galaxies are optically brighter than FR II galaxies but are
less powerful in the radio.
For the complete sample we nd < MR;24 >= ?21:03; and < MB;24 >=
?18:65. For the FR I galaxies the numbers are < MR;24 >= ?21:72 and <
MB;24 >= ?19:27, while for the FR II galaxies they are < MR;24 >= ?20:99
and < MB;24 >= ?18:55. The B ? R color for the radio sample is  1:6
while for the FR I and FR II samples it is  1:8 and  1:5 respectively.
These numbers are summarized in Table 7.1.

7.2.4 Morphological peculiarities and dust

It has been shown time and again that though radio galaxies are basically el-
liptical, they possess unusual features to varying extent. Toomre and Toomre
(1972) have shown that the production of a long narrow feature in an inter-
action (or merger) requires that at least one of the objects involved possess

148
Figure 7.6: re(B )=re(R) as a function of radio power. The dashed line shows
the scale length ratio expected for a normal galaxy. There is a hint that
more powerful galaxies are bluer towards the center. However, its a weak
correlation. The disky galaxies are marked with an additional plus-symbol.
These galaxies are not amongst the galaxies showing the greatest blue excess.

149
Figure 7.7: R magnitude as a function of radio power at 408 MHz. FR I
sources tend to be optically bright while FR II are more powerful in radio.
The dashed pattern is the dividing line.

150
Table 7.2: Morphological peculiarities and dust in FR I and FR II galaxies.
FR I galaxies are seen to be more dusty and redder on an average. FR II
galaxies are more likely to be morphologically peculiar.

Morphologically Dust Common envelope Isophotal

peculiar multiple nuclei twist > 10 deg
FR I  35%  85%  50% 40
FR II  65%  45%  45% 50
FR I + FR II  55%  60%  45% 50

a dynamically cold component (disk) and further that for the production of
two such tails both the objects should possess a prominent disk before the
encounter.
We nd that  55% of radio galaxies from our sample possess some kind
of morphological peculiarity (a tail, a bridge or an arc). When FR I and FR
II sources are considered separately, we nd that  35% of our FR I radio
galaxies show a peculiar feature while as many as  65% FR II galaxies
show such a feature. Since the radio galaxies in our sample are not seen to
have a dominant companion, when an arc or tail like structure is present, the
structure is likely to be due to the gas in the radio galaxy. In the light of the
ndings of Toomre and Toomre (1972), this is consistent with the fact that
all our disky galaxies are FR II sources.
The situation with respect to dust is just the opposite of this. We nd
that  60% of the radio galaxies possess dust (detectable using color maps).
The division between the two types is that  85% FR I galaxies are seen to
contain dust while the corresponding number for FR II sources is  45%.
 50% radio galaxies ( 40% FR I and  50% FR II) are seen to exhibit

151
isophotal twists in excess of 10 deg. An almost equal number ( 55%) of
control sample galaxies also exhibit this feature.
The incidence of common envelopes and secondary nuclei is similar for
both types.  50% for FR I,  45% for FR II and  45% overall.
These numbers are summarized in Table 7.2.

7.2.5 Harmonic components

In Section 3.2.2 we have described how we obtain the deviations of isophotes
of a galaxy from perfect elliptical shapes. After the best tting ellipse has
been obtained, the higher harmonics indicative of deviations are obtained by
making a least squares t to
I = I0 + A3 sin(3E ) + B3cos(3E ) + A4sin(4E ) + B4cos(4E ):
Typically the deviations are < 1%.
Table 7.3 summarizes our ndings for the radio galaxies as well as the
control galaxies. The fourth harmonics are the more important ones. For
A4 and B4 harmonics, we have counted galaxies that have amplitude of the
harmonic exceeding 0.01 in a region away from the center.
We note that FR I sources have the greatest incidence of B4 < 0. A
negative B4 indicates boxy isophotes such that the surface brightness along
the major axis is reduced. While boxiness can come about due to dynamical
reasons, the presence of dust easily explains it as well (See e. g. Lauer, 1985;
Mahabal et al. , 1996). Dust along the major axis leads to a preferential
absorption of light along that axis leading to an eectively lower surface
brightness. This is consistent with our nding a greater incidence of dust
features in FR I galaxies.
Similarly, FR II galaxies possess the highest proportion of galaxies with
B4 > 0. A positive B4 is indicative of pointy isophotes i. e., isophotes that
are elongated along the major axis. This is a typical diagnostic of diskiness

152
Table 7.3: B4 and A4 distributions for the radio and control galaxies. FR I
galaxies are seen to be more boxy while FR II galaxies are disky.

B4 < ?0:1 B4 > 0:1 A4 < ?0:1 A4 > 0:1

FR I 10% 30% 10% 10%
FR II 20% 20% 0% 40%
FR I + FR II 10% 35% 15% 0%
Control 10% 25% 5% 15%

and is in conrmation with results from prole tting which demonstrate
that FR II galaxies often have a signicant disk component.

7.3 Induced star formation

In recent years several cases have been reported where star formation seems to
have been induced by the passage of the radio jets in regions close to the host
galaxy (See e. g. Best et al. , 1996). The studies are typically undertaken with
high resolution radio maps and equally good optical images. Unfortunately
we do not have very good radio maps to carry out any detailed analysis in this
respect. However, in the radio galaxy 1006-214 in our sample star formation
seems to be occurring in the region surrounding the galaxy (dark halo or
IGM). In addition, star formation is also occurring well within the galaxy.
1006-214 is one of the galaxies with a prominent disk. The twisted radio
source (Figure 7.8) is not very good at revealing ner details of the possible
induced star formation. It is quiet possible that all powerful radio sources
are associated with such starbursts. Better observations will be needed to
conrm this at such moderate-redshifts.

153
Figure 7.8: The radio (1.4 GHz) and B images of MRC 1006-214. North is
to the top and east to the left. The central radio blob covers a region slightly
greater than the optical image. It is along the radio axis (north-east lobe)
that star formation is seen to be taking place. (Radio observations obtained
by Patrick McCarthy).
154
7.4 Conclusions
The radio galaxies in our sample exhibit several properties that have been
noted before for other samples. To name a few:
 FR I galaxies are optically brighter but less luminous in radio as com-
pared to FR II galaxies.
 FR II galaxies show peculiarities more often than do the FR Is.
 Dust is more common in FR I galaxies. These galaxies are also boxy
in nature.
 FR II are more likely to contain a disk component. They have pointed
isophotes.
In the next chapter we show how ages for star bursts can be obtained.

155
Chapter 8
Ages: how recent is the blue
8.1 Introduction
We saw in Chapter 4 that the scale length ratios in the B and R bands clearly
indicate that a larger proportion of radio galaxies than the control sample
galaxies tend to have bluer colors near the center. The color gradients are
also indicative of this behavior. We further showed using extrapolated proles
near the centers of the radio and control galaxies that the radio galaxies tend
to have an excess emission near the center more frequently than the control
galaxies. We noted that the excess is over a few kpc, and is likely to be due
to recent star formation rather than an AGN.
In this chapter we attempt to understand the possible origin of this eect.
We begin by describing a number of similar cases observed in dierent types
of galaxies and models which are used to estimate the strength, i. e., the
gas mass undergoing the burst, and the age of the burst at the time of the
observation. We will see what factors are involved in the estimation and how
accurately one can expect to obtain these two important parameters. We
especially note how the age dating technique can be used to constrain the
age and strength with limited data.
Just as the interpretation of the color gradients is made dicult by the

156
age-metallicity degeneracy, so also the interpretation related to the excess
blue color is made dicult by the age-strength degeneracy: an old but strong
burst can produce features similar to what a young, weak burst can. We
describe techniques that are being developed by dierent groups to break
this degeneracy using line ratios in dierent parts of the spectrum.
Some recent observations have indicated that Wolf-Rayet stars can play
a very important role in the kind of study we are presenting here. The
signature of these stars is visible only for a few million years and hence they
are indicative of recent activity. We discuss the possibility of Wolf-Rayet
stars giving rise to the blue color we see in the radio galaxies.

8.2 The existing picture

Searle et al. (1973) proposed that abnormally powerful bursts of star forma-
tion are responsible for the unusually blue colors in some galaxies. The term
starburst is now used to denote an intense star formation event lasting over
a short period of time. The poststarburst phase lasts an order of magnitude
longer than the starburst phase. Poststarburst galaxies, also called the \E +
A" galaxies (Dressler and Gunn, 1983), are typically characterized by strong
Balmer absorption lines and weak or no emission lines. The E + A refers
to the spectroscopic similarity of the appearance of these galaxies to the su-
perimposition of an A-star component over an old elliptical galaxy spectrum
(Gunn and Dressler, 1988).
Butcher and Oemler (1984) showed that the proportion of such blue galax-
ies is greater in distant clusters. Dening fB to be the fraction of blue galaxies
in a cluster i. e., of galaxies whose rest-frame B ? V colors are at least 0.2
mag bluer than the early type galaxies at that magnitude, they nd that
fB  0:03 in nearby clusters (z < 0:1) but rises to  0:25 at z = 0:5.
Though the study refers not to individual blue galaxies but their propor-
tion in a cluster, the blue galaxies that are the cause of the Butcher-Oemler

157
eect are thought to be E + A galaxies. Thimm and Belloni (1994) have
observed galaxies at z > 0:8 which are well t by E + A spectra. E + A
galaxies are powerful indicators of galaxy evolution and several models have
been advanced to quantify the observable parameters (see e. g. Couch and
Sharples, 1987 and Belloni et al. , 1995).
Couch and Sharples (1987) classied E + A galaxies into blue, poststar-
burst galaxies (PSGs) and red, H strong galaxies (HDSs). Two-parameter
models (age and strength of burst) are reasonably successful in modeling
both types of galaxies. However, often the ambiguity in one of the param-
eters is large leading to degenerate solutions. For instance, Liu and Green
(1996) conclude after studying a heterogeneous sample of 8 E + A galaxies
that \the attempts to model such galaxies at high-redshifts using a simple
starburst-plus-galaxy model are inadequate and could lead to serious system-
atic errors if interpreted too boldly." They nd that 5/8 galaxies are well
described by E + A model, 2/8 by a burst-plus-spiral model and in one case
multiple bursts have to be invoked to describe the spectrum. The burst age
could vary by upto 30% in each case without loss of statistical signicance.
This leads to large degeneracy in mass as well.
Leonardi and Rose (1996) present a technique to overcome the age-strength
degeneracy. They use two spectral indices viz. H=4045 and one formed
from the ratio of the Ca II H + H line to the Ca II K. The former ratio
decreases in value as one proceeds from late type stars to earlier type stars
whereas the latter is constant in late type stars but decreases for stars earlier
than F2. Since the two indices have dierent behavior for dierent types of
stars, taken together they are able to resolve the degeneracy between age
and strength. While using two dierent indices to obtain better estimates
of the burst parameters is an attractive idea, the technique requires that in
addition to having good spectroscopic data, we also need to study the eect
of the history of star formation on these indices.
What kind of stars give rise to the excess blue color? If the starburst age is

158
under 10 million years, one interesting possibility is Wolf-Rayet (WR) stars.
Heckman et al. (1997) have observed Mrk 477, which hosts a powerful type
2 Seyfert nucleus. It is seen to have a circumnuclear starburst (dimension
< 1 kpc). Observation from UV to NIR indicate that the likely cause is Wolf-
Rayet stars numbering  30000. Such a starburst also leads to a pronounced
excess in the far-IR which is often seen in type 2 Seyfert nuclei. Thus, in
Seyfert galaxies, a WR starburst could be able to explain various properties.
Most of the models that have been tried so far have been for Solar metal-
licity, though invoking factors like WR stars may necessiate the use of higher
metallicities. One reason why other metallicities have not been extensively
tried is the age-strength degeneracy. If one more parameter is to be included,
the quality of the data and our understanding of the spectra will have to in-
crease manifold.
In the next section we describe how synthetic spectra are generated and
the procedure which is followed to compare colors of our radio galaxies with
those from the synthetic spectra.

8.3 Synthetic spectra

For the galaxies in the radio sample we have observations in B; R and K 0
bands for several objects and in B and R for the remaining. For the former
we have available two colors to compare with the synthetic spectra while
in the latter case we have only one color. With the degeneracy mentioned
in the previous section, it is not very meaningful to carry out a best t for
burst age and strength since such a t will provide a large range for both
the parameters. So, instead, we have plotted the curves showing the age-
strength relation. We have also plotted our objects on a similar diagram to
see what possible values the parameters are likely to take. In our work we
have used synthetic spectra kindly provided by Dr. Bruno Guiderdoni. A
detailed prescription of how these are generated can be found in Guiderdoni

159
and Rocca-Volmerange (1987). Details on models other than the burst model
can be found in Rocca-Volmerange and Guiderdoni (1988).

8.3.1 Galaxy spectra

The initial mass function (IMF, (m)) is one of the most important inputs
for the generation of synthetic spectra. The IMF gives the distribution of the
number of stars as a function of mass. The IMF and the law of star formation
together provide the number of stars forming in a given mass range at a given
time, t. These stars are then placed on the zero age main sequence (ZAMS) in
the Hertzprung-Russel (HR) diagram. Theoretical stellar tracks are used to
evolve the stars at each time step. The distribution of the stellar populations
in the HR diagram log Te ? log L=L , is estimated. Using a bolometric
correction and the luminosity class, log L=L is converted to Mv . A synthetic
stellar spectrum, F (; t) is then constructed. A nebular component Fneb (; t)
resulting from the Lyman continuum photons is added. A correction for
internal extinction Ai (t) is applied.
The synthetic spectra used here are based on the Scalo IMF. This has
the form (m) / m?x, with
x = 0:25 0:1 < m < 1:0(m)
x = 1:35 1:0 < m < 2:0(m)
x = 1:70 2:0 < m < 120(m): (8.1)
The stellar tracks used are from the Geneva group. The stellar spectra are
from Kurucz (1992) for hot stars (log Te > 3:5), and from Bessel et al. , Fluks
et al. (1994) for cold stars. The tracks include four evolutionary stages viz.
Main Sequence (MS), Giant Branch (GB), \red" Horizontal Branch (HB) and
Asymptotic Giant Branch (AGB). Along each evolutionary sage, the input
stellar tracks are interpolated in mass to get dierences of successive masses
lower than 1 Gyr. Mass loss of massive stars is taken into consideration.

160
 The basic spectra that we have made use of to obtain colors and mag-
nitudes cover a range of wavelength from 90:9  A ? 1:4106  A (1221 distinct
wavelengths). We used spectra corresponding to the following 219 distinct
ages:
0; 1; 2; ::9 Myr; 10; 20; ::90 Myr, 100; 200; ::900 Myr, 1; 1:1; ::1:9; 2; 2:1; ::20 Gyr.
The spectra were generated for dierent metallicities viz. 0.001, 0.004, 0.008,
0.020 and 0.040. However, we actually used only the solar metallicity spectra
with Y=0.28 to 0.30 and Z=0.02. The whole Hubble sequence is represented
with one metallicity since the mean metallicity of stars rapidly approaches
the solar value. Also, with the scanty data that we have, adding a further
variable Z would not have been vary meaningful.

8.3.2 Modus operandi

In this subsection we describe how colors were obtained for a E + A model
from the synthetic spectra.: (1) The age of the old galaxy was assumed to
be 20 Gyr. (2) A young burst was then superimposed onto the old galaxy.
The burst age was allowed to vary between 0 and 2 Gyr and it was allowed
to contain a mass fraction of upto 10%. (3) The resulting spectrum was
convolved with individual lter response curves, after appropriate corrections
for redshift and cosmological dimming were applied, to obtain magnitudes
for those lters.
Thus, for each burst age and each mass fraction, a number (magnitude)
is obtained for each lter. Colors and isochrones are then obtained from
the magnitudes. These are nally compared with the observed colors of the
galaxies to estimate the age and strength of the starburst. We in fact, shifted
the colors of the radio sample galaxies to a redshift of zero so that all the
galaxies can be simultaneously compared with the zero-redshift isochrones.

161
8.4 Age of the most recent starburst
We reproduce in Figure 8.1 the isochrones for a E + A galaxy with solar
metallicity. We have plotted the B ? R versus the R ? K colors resulting
from the prescription in the previous section. The rst panel shows just the
isochrones resulting from the superimposition of a 20 Gyr galaxy and a 0{2
Gyr burst of star formation involving 0{10% galaxy mass. The subsequent
panels show individual isochrones (mass 0{8%). Superimposed on these are
the colors of radio galaxies shifted to zero redshift. It is seen that the colors of
several galaxies do not lie on the 0% mass isochron (just the old population).
However, successive panels indicate that the colors are consistent with a
recent burst of star formation. Since we do not have several colors, it is not
meaningful to obtain a t to these numbers. For better prediction, a larger
number of colors, and perhaps high resolution spectra are needed. Dunlop
et al. (1989) have shown that colors can be used to decide between various
star formation models. This will form part of a future study.

162
Figure 8.1: B ? R and R ? K isochrones from synthetic spectra are shown in
the gure at top left. A 20 Gyr galaxy has been superimposed with a young
(0 { 2 Gyr) starburst containing 0{10% mass fraction of the galaxy. After
a few million years, bursts of dierent strengths are seen to produce similar
colors. The remaining ve panels show individual isochrones corresponding
to 0, 2, 4, 6 and 8% burstmass. Superimposed are the galaxy colors from the
radio sample shifted to zero redshift. Though the error bars are large, it is
clear that the colors are consistent with a recent burst of star formation.

163
Chapter 9
Conclusions
9.1 The emerging picture
In this thesis we have presented results related to the optical morphology
of a sample of radio galaxies from the MRC. The motivation of the thesis
has been to identify morphologically peculiar features in radio galaxies and
to try and relate them to the activity in these galaxies. The project had
its modest roots in looking for dust features in radio galaxies following our
nding a dust lane in NGC 4261 (Mahabal et al. , 1996).
It was soon realized that the prole ts are a great information resource.
We were able to devise dierent ways to make use of the information to study
the prole shapes as well as the galaxy components. de Vaucouleurs' law,
though empirical in nature, provides an excellent base t to normal ellipticals
as well as radio ellipticals. de Vaucouleurs' law has been extensively used
in literature. However, most authors tend to use the parameters in a single
photometric band. By studying parameters in more than one band, we have
shown that the distribution of the re(B )=re(R) ratio is a powerful indicator of
possible activity in a galaxy. Using the distribution of this ratio for the radio
galaxies, and a control sample of early type galaxies, we were able to show
that the radio galaxies are likely to have undergone recent star formation in

164
the central region whereas the control galaxies had a quiet history.
We have presented a simple relation between the scale length ratio and
the color gradient. Gradients obtained in this manner are well correlated
with conventionally measured color gradient, so the scale length ratio itself
can be used as a color indicator. We have found that the ratio is more robust
than the conventional gradient, in the sense that it can be used even when
the color prole is too noisy for the conventional color gradient to be useful.
Observations of a large number of galaxies have indicated that de Vau-
couleurs' law is valid in the interval 0:1re <r <1:5re (Burkert, 1993). For
r < 0:1re one expects departures from the law. The spreading due to the
PSF also makes it dicult to reliably obtain values close to the center. Hence
we have carried out all ts by leaving out points within 1.5 times the FWHM
of the PSF from the center. After obtaining the t we have looked for devia-
tions from de Vaucouleurs' law in the prole close to the center by comparing
the observed prole with extrapolation of the model prole right up to the
center. Our investigations indicate that for almost all control sample galaxies
de Vaucouleurs' law is valid close to the center in the sense that the observed
points lie within 1 of the extrapolated model. The radio galaxies, on the
other hand reveal excess emission in  53% of cases as compared to  23%
of the control sample galaxies. We attribute the excess emission to star-
burst. A decit relative to the extrapolation, indicating absorption due to
dust is found in 66% of the radio sample and 43% of the control sample (see
Table 4.7).
We nd that  55% of the radio galaxies exhibit some form of morpholog-
ical peculiarity with FR II galaxies exhibiting a peculiarity more frequently
( 65%) than the FR I galaxies ( 35%). A small note on individual galax-
ies is presented in Appendix B. A disk in a radio galaxy is supposed to be
particularly rare. In our sample, we have detected ve ( 17%) disky candi-
dates. One of them clearly possesses spiral-like structure where as two others
show some indication of such structures. To help us detect faint disks (or,

165
for that matter, any faint source hidden on top of a background for which an
empirical formula is available) we have developed a series of morphological
lters. The morphological gradient lter is able to detect changes in contrast
and using it we are able to see spiral-like structure in two galaxies. None of
the disky galaxies is an FR I source.
We nd that  20% radio galaxies have dust lanes and  17% have
prominent dust patches. In the control sample there are fewer galaxies with
dust lanes but more with patchy dust, indicating that the dust in radio
galaxies is more coherent. The dust content estimated from extrapolated
proles indicates that in nuclear regions radio galaxies are more likely to
have dust. We found that the dust mass in dust lane galaxies is proportional
to the radio power at 408 MHz.
One of the surprises of the study was our nding that the dust lanes
are, in general, not perpendicular to the radio axis. Earlier studies had
suggested that the two tend to be perpendicular (e. g. Kotanyi and Ekers,
1979; Mollenho et al. , 1992; van Dokkum and Franx, 1995). Our numbers
are small and it could be that it is just a statistical eect. However, it is a
result worth keeping in mind.
We nd that  85% FR I galaxies contain dust while only  45% FR II
galaxies are dusty. The incidence of isophotal twists and secondary nuclei is
comparable in FR I and FR II radio galaxies. FR II galaxies tend to possess
disky isophotes whereas FR I galaxies tend to possess boxy isophotes. It is
likely that in some of the FR II sources are inducing star formation along
the path of the radio source. Similarly, a higher incidence of blue centers is
seen in FR II galaxies.
To summarize, the optical morphology of radio galaxies indicates that
these galaxies are not smooth, classical ellipticals but frequently contain mor-
phologically peculiar features. They are characterized by bluer centers very
likely to be related to star formation in the recent past. Contrary to previous
beliefs, radio galaxies can contain disks too. It is FR II galaxies that are seen

166
to contain a higher frequency of morphological peculiarities and disks. FR I
galaxies are more likely to contain dust. Dust in radio galaxies is more likely
to be coherent.
These ndings point to a number of possible future studies. We discuss
some of these in the next section.

9.2 The work goes on

In the study presented here we have extracted a well dened control sample
from the radio galaxy elds. While this was a good sample of early-type
galaxies, we were hampered in studying certain distance related aspects of
these galaxies since we lacked the redshifts to these objects. During the
period of the thesis it was not possible to obtain these redshifts. Because
of their inherent uncertainties, photometric redshifts would not have served
the purpose. An alternative was to choose a dierent control sample made
of early-type galaxies but for which the redshifts are known. However, we
would then have lost the advantage of having identical processing for the
radio and the control sample.
Taking all the above points into consideration, obtaining photometric
redshifts for the control galaxies seems to be the lesser evil and we hope to
do that in the near future. That study will also help us examine cluster
characteristics and compare the cluster properties of FR I and FR II sources.
We will also be able to study the possible relationship between individual
galaxies with blue centers and the Butcher-Oemler eect.
In relation to the above point, high resolution spectra of the radio galaxies
will allow us to estimate various line strengths and hence metallicities and
enable us to model the recent star burst in the radio galaxies (e. g. Leonardi
and Rose, 1996). The spectra will also provide information on the nuclear
processes.
Recent studies of central regions of elliptical galaxies observed with the

167
HST have indicated that ellipticals can be divided into cuspy galaxies and
galaxies with a core (Crane et al. , 1993; Jae et al. , 1994; Forbes et al. , 1995;
Faber et al. , 1997). Our ground based data does not have the resolution to
sample the central parts of galaxies. However, the advantage of the ground
based data is that it extends to much larger angular sizes than the HST
data. We propose to obtain HST data for our galaxies and combine it with
our ground based data to see how the properties we have described here
relate to the two types of galaxies. It will also allow us to probe the central
regions better and conrm and further quantify the excess blue emission that
we have detected. We are thankful to Ken Freeman for the germ of this idea.
Time and again we felt the need of high resolution radio maps when trying
to examine the association between optical and radio features. Especially
interesting would be the possible association of recent star formation with the
radio jets (e. g. as shown by Best et al. , 1996, for z  1 galaxies), association
of spiral-like structures in various galaxies with radio enhancements and so
on. We plan to obtain high resolution radio maps of these southern galaxies
when VLA is in the right conguration (A + B).
We have used the morphological gradient lter for detecting disks in
galaxies. Initial tests with the morphological top hat lter have indicated
that it can be especially good at detecting faint absorption features since it
is particularly adapted to nding features \in shadows". The top hat and
other morphological lters will be studied in detail and calibrated as part of
the followup.
Studies have indicated that HI in galaxies is associated with disturbed
morphologies (e. g. Schiminovich et al. , 1994, 1995). We would like to ob-
tain HI observations for the sample. The Giant Meterwave Radio Telescope
(GMRT), near Pune in India, will be able to reach the southern declinations
of our sample.

168
Appendix A
Morphological processing
Morphological image processing techniques are useful for extracting image
components that are useful in representing and describing region shapes.
The lters can be described using set theoretic notation and implemented
using simple computer algorithms based on these. Almost all the material
in this chapter has been taken from Digital image processing (Gonzalez and
Woods, 1993).
In set theory notation a 2D image is said to be a member of R3 . For
simplicity we assume that the images are gray-scale images with their in-
tensity being an integer in the range [0,255]. In the following description
we consider only two gray levels viz. 0 and 255 which correspond to white
and black respectively (it is then called a binary image and belongs to Z 2 ).
However, all the discussion extends easily to discrete as well as continuous
gray-scale images.

A.1 Dilation and erosion

The two most basic morphological operators are dilation and erosion. Dila-
tion of the set A by set B , denoted by A  B , is obtained by rst re ecting
B about its origin and then translating the result by x. All x such that A
169
and re ected B translated by x that have at least one point in common form
the dilated set.
A  B = fx j (B^ )x \ A 6= g;
where, B^ denotes the re ection of B i. e., B^ = fx j x = ?b; for b 2 B g and
(B )x denotes the translation of B by x = (x1; x2 ) i. e., (B )x = fc j c =
b + x; for b 2 B g. Thus, dilation of A by B expands the boundary of A.
It is easier to implement for gray-scales than the above description sug-
gests.
(f  b)(s; t) = maxff (s ? x; t ? y)+ b(x; y) j (s ? x); (t ? y) 2 Df ; (x; y) 2 Db g:
(A.1)
Here, f and b denote images f (x; y) and b(x; y). f is being dilated and b
is called the structuring element and Df and Db are the domains of f and
b respectively. Thus, in dilation we choose the maximum value of f + b in
a neighborhood dened by b. If all elements of b are positive, the dilated
image is brighter than the original and the dark details are either reduced or
eliminated.
Erosion of A by B , denoted by A B , is the set of all x such that B
translated by x is completely contained in A, i. e.,
A B = fx j (B )x;  Ag:
For gray-scale images we have,
(f b)(s; t) = minff (s + x; t + y) ? b(x; y) j (s + x); (t + y) 2 Df ; (x; y) 2 Db g:
(A.2)
Erosion is thus based on choosing the minimum value of (f ? b) in a neighbor-
hood dened by the shape of b. If all elements of b are positive, the output
image is darker than the original and the eect of bright details in the input
image are reduced if they cover a region smaller than b.

170
A.2 Opening and closing
The subsequent morphological operators are dened in terms of erosion and
dilation. Their expressions for gray-scale and binary images are the same
and a distinction will not be henceforth made.
A is said to be opened by B if the erosion of A by B is followed by a
dilation of the result by B .
A  B = (A B )  B (A.3)
Opening results in removal of narrow peaks. The initial erosion removes the
small details and darkens the image. The following dilation increases the
brightness but does not reintroduce the details removed by erosion.
Similarly, A is said to be closed by B if A is rst dilated by B and the
result is then eroded by B . Thus,
A  B = (A  B ) B (A.4)
Closing is used to remove dark details from an image. The initial dilation
removes dark details and makes the image brighter. The erosion that follows
darkens the image but does not reintroduce the details removed by dilation.
Opening as well as closing are idempotent operators i. e., consecutive
openings do not change an image, nor do consecutive closings: (f b)b = f b
and (f  b)  b = f  b.

A.3 Gradient
The morphological gradient highlights sharp transitions in the input image.
It depends less on edge directionality than the Sobel operator and is useful
for locating faint but large scale structures. The morphological gradient is
dened by,
g = (f  b) ? (f b): (A.5)
171
It is thus the dierence between a dilated image and an eroded image. Dila-
tion removes small scale dark features and erosion removes small scale bright
features. Dilation brightens the image and erosion darkens it. When the dif-
ference is obtained, all small scale features are gone and the contrast between
the large scale features improved.

172
Appendix B
Individual objects
Morphology of MRC radio galaxies
We give here short notes on each of the MRC objects studied as part of this
thesis. Some of the more interesting objects have already been described
in earlier chapters. The main emphasis here is on the morphology evident
from the prole ts and the conclusions one can draw thereby. Several of
these objects need further investigation, of which a note has been made. The
radio parameters are listed in Table 2.3 and the scale lengths and related
parameters are presented in Table 4.3.

0344-291 0344-291 is a dumbbell galaxy belonging to Abell 3165. One

more galaxy is seen about 700 away. It is not clear if it is also at the same
redshift. The MRC radio map (Ekers, 1989) shows an unresolved component.
In the case of this galaxy, normal tting of the prole was not possible.
Fitting to individual components would also have involved being restricted
to a small region as the eect of the other component would have been very
strong. As a result, prole tting was not carried out.

173
0349-278 This galaxy is surrounded by 3 other galaxies at approximately
the same redshift. 0349-278 exhibits very interesting tidal features which
could very well be contaminated by line emission. Our morphological pro-
cessing shows that one of the companions might not be an elliptical. Danziger
et al. (1984) have reported their spectra to be consistent with expanding
rings. The radio source is of type FR II with the north-east lobe coinciding,
in projection, with one of the companions. The bulge and disk t indicates
the presence of a disk prominent in the B band. re(B ) < re(R).

0354-263 A 6  4 arcsec2 dust lane is seen in the color map. More details
of the dust lane are provided in Chapter 6. It has a possible spiral feature.
The bulge and disk t shows small matching disks and absorption consistent
with the dust seen. re(B ) > re(R).

0420-263 The color map of this galaxy exhibits a dust lane of dimensions
7  4 arcsec2. More details are provided in Chapter 6. The morphology
indicated by the B residual, obtained by subtracting a smooth model from
the original, is consistent with the presence of dust. re(B ) < re(R). It is one
of the few galaxies for which, the conventional color gradient indicates that
it gets bluer as one approaches the center.

0434-225 This galaxy is in Abell S0473. It has an interesting, elongated

companion. Radio map is in Ekers (1989). Large dust patch  4  4 arcsec2
is seen at the center. A spiral-like structure is also seen in the nuclear part
of this galaxy. re(B ) ' re(R).

0446-206 It is in Abell 0514, a dense cluster. It has 2 nearby companions

(in projection) and shows an armlike feature. re(B ) ' re(R).

0453-206 It is also identied as NGC 1692. It is an IRAS source and shows

a dust patch  3  3 arcsec2 in the B ? R color map. It has a previously
174
identied ring (Buta, 1995). re(B ) > re(R)

0520-289 The associated radio source is of type FR II with the radio axis
being almost perpendicular to the galaxy major axis. A dust patch of size
 5  4 arcsec2 is seen near the center.
0938-205 This galaxy is in a crowded eld. The radio source associated
with it is of type FR II. It has a redshift of 0.371 and was mistakenly included
in our sample. re(B ) > re(R).

0950-239 It subtends a very small angle on the CCD and the routine tech-
niques were not applied to it.

0952-224 This galaxy is in a crowded eld. It exhibits a dust patch near

the center with E (B ?R)  0:2 and of size  33 arcsec2 . The B t indicates
absorption consistent with the observed dust. The galaxy also hosts a disk,
details of which can be found in Chapter 5. re(B ) > re(R).

1006-214 It is one of the more interesting objects in the sample. It has a

blue, armlike extension towards the east. The extension curves, roughly, with
the FR II radio structure. The galaxy has a blue center and a prominent disk
seen better in the K 0 image. The radio source is of type FR II. re(B ) < re(R)

1033-251 This is another object in a crowded eld subtending a very small

angle on to the CCD. It was not subjected to the prole tting routines.

1048-238 1048-238 has been identied with a quasar. It was because of

the fuzz that is seen around this object that it got included in the sample.
The color map brings out the fuzz particularly well. The color gets bluer as
one approaches the center. re(B ) < re(R).

175
1053-282 It has a dust patch at the center that measures  5  4 arcsec2
and has an E (B ? R)  0:15. Radio map is in Ekers (1989). It has a disk
with a D=B > 5. More details are given in Chapter 5. A bridge like structure
extending towards a companion in projection is seen. re(B ) ' re(R).

1056-272 It is an IRAS source. The color map does not reveal the presence
of substantial amount of dust. It is likely to have a thin ring-like structure
around the nucleus. re(B ) < re(R)

1103-244 This galaxy belongs to Abell 1165. Color gradient indicates it

has a blue center. re(B ) < re(R).

1126-246 It has an armlike low-opacity dust feature almost perpendicular

to the FR II radio structure. re(B ) < re(R).

1215-215 This galaxy has a disk with D=B > 1. More details related to
the disk are given in Chapter 5. It has a blue center and a spiral arm-like
feature. The color gradient is consistent with a blue center. re(B ) < re(R).

1222-252 It exhibits a barlike structure with an elongation perpendicular

to it at the north-west end. Flocculent structure is also seen to the north-east.
Prole ts conrm the presence of a disk (see Chapter 5). re(B ) < re(R).

1251-289 It has a secondary nucleus which could be a cannibalized object.

It is in Abell 3528 and is an IRAS as well as X-Ray source. re(B ) < re(R).

1254-268 Linear red feature  600 is seen along the major axis of this
galaxy. E (B ? R)  0:1

1257-253 It has a diuse dust lane of size  7  3 arcsec2 with E (B ? R) 

0:1. It is an IRAS source. re(B ) < re(R).

176
1303-215 It has a blue center and possible spiral arms or ring. It is either
made of several closely interacting galaxies or a single rather distorted galaxy.
re(B ) < re(R).

1323-271 This galaxy is seen to have a companion with which it is inter-

acting. It belongs to Abell 1136. It exhibits isophotal twist. There are also
hints of shell-like or arm-like structure. re(B ) > re(R).

1329-257 This galaxy is another very interesting object. It has at least 5

components. The radio source is of type FR II. 1329-257 is an IRAS and
X-Ray source. K 0 band image of this galaxy shows only three of the ve
components prominently. The other two components are very blue in color
and it is likely that star formation has taken place at those locations in
the recent past. The radio source in this case does not seem to be directly
responsible. re(B ) < re(R)

1343-253 It belongs to Abell 1791 . It is an IRAS source and shows isopho-

tal twist.

1344-241 It is likely to be an S0 galaxy. It has a large dust lane, almost

along the major-axis, curving past the center. More details are given in
Chapter 6. re(B ) < re(R).

1346-252 It belongs to Abell 1791 and is an IRAS source. It is a Dumbbell

galaxy with a possible third component nearby. It is a head-tail FR I source.
re(B ) < re(R).

1354-251 It belongs to Abell 1822 and is an IRAS source. Soares et al.

(1995) have suggested that it is a galaxy pair. re(B ) < re(R).

177
References
[1] Allington-Smith J. R. 1982 MNRAS 199 611
[2] Athreya R. et al. 1997 MNRAS 289 525
[3] Barnes J. E. and Hernquit L. 1992 ARAA 30 705
[4] Belloni P., Bruzual A. G., Thimm G. J. and Roser H. -J. 1995 AA 297
61
[5] Bender R., Surma P., Dobereiner S., Mollenho C. and Madejsky R. 1989
AA 217 35
[6] Bender R. et al. 1993 in Structure, dynamics and chemical evolution
of elliptical galaxies eds. I. J. Danziger, W. W. Zeilinger and K. Kjar,
Garching: ESO conference and workshop proceedings No. 45, p. 3
[7] Bertin G. and Stiavelli. M 1984 AA 137 26
[8] Bertin G. and Stiavelli. M 1989 ApJ 338 723
[9] Bertola F. and Capaccioli M. 1975 ApJ 200 439
[10] Bertola F. and Galletta G. 1978 ApJL 226 L115
[11] Best P. N., Longair M. S. and Rottgering H. J. A. 1996 MNRAS 280
L9

178
[12] Bicknell G. V. 1995 ApJS 101 29
[13] Binney J. 1976 MNRAS 177 19
[14] Binney J. 1982 MNRAS 200 951
[15] Boroson T. A. 1983 AJ 88 1707
[16] Buil C. 1991 in CCD astronomy: construction and use of an astronomical
CCD camera (William Bell Inc., Virginia)
[17] Burkert A. 1993 AA 278 23
[18] Burstein D. ApJ 1979 234 435
[19] Burstein D. and Heiles C. 1978 ApJ 225 40
[20] Buta R. 1995 ApJS 96 39
[21] Butcher H. and Oemler A. Jr. 1984 ApJ 285 426
[22] Caon N., Capaccioli M. and D'Onofrio M. 1993 MNRAS 265 1013
[23] Carlberg R. G. 1984 ApJ 286 403
[24] Colla G. et al. 1970 A&AS 1 281
[25] Corture J. and Hardy E. 1988 AJ 96 867
[26] Couch A. J. and Sharples R. M. 1987 MNRAS 229 423
[27] Crane P. et al. 1993 AJ 106 1371
[28] Danziger I. J. et al. 1984 MNRAS 208 589
[29] Dehnen W. 1993 MNRAS 265 250
[30] de Vaucoulers G. 1953 MNRAS 113 134

179
[31] Djorgovski S. and Davis M. 1987 ApJ 313 59
[32] Downes A. J. B. et al. 1986 MNRAS 218 31
[33] Dressler A. and Gunn J. E. 1983 ApJ 270 7
[34] Dressler A. et al. 1987 ApJ 313 42
[35] Dube R. R., Wickes W. C. and Wilkinson D. T. 1972 ApJL 215 L51
[36] Dunlop J. S. et al. 1989 MNRAS 240 257
[37] Ekers R. D. et al. 1989 MNRAS 236 737
[38] Elias A. J. et al. 1982 AJ 87 1029
[39] Faber S. M. et al. 1997 AJ 114 1771
[40] Fanaro B. L. and Riley J. M. 1974 MNRAS 167 31p
[41] Fanti R. 1984 in Clusters and groups of galaxies eds. F. Mardirissian,
G. Giuricin and M. Mezzetti (Reidel, Dordrecht), p. 185
[42] Fasano G., Falomo R. and Scarpa R. 1996 MNRAS 282 40
[43] Ferrarese L. et al. 1994 AJ 108 1598
[44] Fluks M. A. et al. 1994 A&AS 105 311
[45] Forbes D. A. and Thomson R. C. 1992 MNRAS 254 723
[46] Forbes D. A. et al. 1994 AJ 107 1713
[47] Forbes D. A., Franx M. and Illingworth G. 1995 AJ 109 1988
[48] Franx M., Illingworth G. and Heckman T. 1989 AJ 98 537
[49] Freeman K. 1970 ApJ 160 811

180
[50] Gallagher J. S. et al. 1977 ApJ 215 463
[51] Gonzalez R. C. and Woods R. E. 1993 in Digital image processing (Ad-
dison Wesley), p. 518
[52] Gonzalez-Serrano J. I., Caballo R. and Perez-Fournon I. 1993 AJ 105
1710
[53] Gonzalez-Serrano, J. I. and Perez-Fournon, I. 1989 ApJL 338 L29
[54] Goudfrooij P. 1994 PhD thesis Amsterdam University
[55] Graham A. et al. 1996 ApJ 465 534
[56] Guiderdoni B. and Rocaa-Vomerange B. 1987 AA 186 1
[57] Gunn J. E. and Dressler A. 1988 in Towards understanding galaxies at
high redshift, eds. R. G. Kron and A. Renzini (Kluwer Academic, Dor-
drecht), p. 227
[58] Hawarden T. G., Elson R. A. W., Longmore A. J., Tritton S. B. and
Corwin Jr. H. G. 1981 MNRAS 196 747
[59] Heckman T. M. et al. 1985 ApJ 299 41
[60] Heckman T. M. et al. 1986 ApJ 311 526
[61] Heckman T. M. et al. 1997 ApJ 482 114
[62] Henden A. A. and Kaitchuck R. H. 1982 in Astronomical photometry
(Van Nostrand Reinhold Company)
[63] Hernquist L. 1990 ApJ 356 359
[64] Holloway A. J. et al. 1996 MNRAS 279 171
[65] Hubble E. 1926 ApJ 64 321

181
[66] Hubble E. 1930 ApJ 71 231
[67] Illingworth G. 1977 ApJL 218 L43
[68] Jae W. 1983 MNRAS 202 995
[69] Jae W. et al. 1994 AJ 108 1567
[70] James F. 1994 in Minuit Reference Manual (CERN, Geneva)
[71] Jedrzejewski R. I. 1987 MNRAS 226 747
[72] Jennison R. C. and Das Gupta, M. K. 1953 Nature 172 996
[73] Jrgensen I., Franx M. and Kjrgaard P. 1996 MNRAS 280 167
[74] Kent S. M. 1984 ApJS 56 105
[75] Kent S. M. 1985 ApJS 59 115
[76] King I. R. 1966 AJ 71 64
[77] King I. R. 1978 ApJ 222 1
[78] Kormendy J. 1977 ApJL 217 406
[79] Kormendy J. and Bender R. 1996 ApJ 464 L119
[80] Kormendy J. and Djorgovsky S. 1989 ARAA 27 235
[81] Kotanyi C. G. and Ekers R. D. 1979 AA 73 L1
[82] Laing R. A., Riley J. M. and Longair M. S. 1983 MNRAS 204 151
[83] Landolt A. U. 1992 AJ 104 340
[84] Large M. I. et al. 1981 MNRAS 194 693
[85] Lauer T. R. 1985 MNRAS 216 429
182
[86] Lauer T. R. et al. 1995 AJ 110 2622
[87] Ledlow M. J. and Owen F. N. 1995 AJ 110 1959
[88] Ledlow M. J. and Owen F. N. 1996 AJ 112 9
[89] Ledlow M. J., Owen F. N. and Keel W. C. 1998 ApJ 495 227
[90] Leonardi A. J. and Rose J. A. 1996 AJ 111 182
[91] Lilly S. J. and Prestage R. M. 1987 MNRAS 225 531
[92] Liu C. T. and Green R. F. 1996 ApJL 458 L63
[93] Mahabal A. A. et al. 1996 ApJ 457 598
[94] Malin D. F. and Carter D. 1983 ApJ 274 534
[95] McCarthy P. J. 1993 ARAA 31 639
[96] McCarthy P. J. et al. 1990 AJ 100 1014
[97] McCarthy P. J. et al. 1991 AJ 102 522
[98] McCarthy P. J. et al. 1996 ApJS 107 19
[99] McCarthy P. J. (private communication)
[100] Merritt D. 1997 in The nature of elliptical galaxies eds. M. Arnaboldi,
G. S. Da Costa and P. Saha ASPCS 116 p. 32
[101] Meurs E. J. A. and Wilson A. S. 1984 AA 136 206
[102] Mollenho C., Hummel E. and Bender R. 1992 AA 255 35
[103] Oke J. B. and Sandage A. 1968 ApJ 154 21
[104] Osterbrock D. E. 1960 ApJ 132 325

183
[105] Owen F. N. and Laing R. A. 1989 MNRAS 238 357
[106] Owen F. N. and White R. A. 1991 MNRAS 249 164
[107] Owen F. N. et al. 1996 in Energy transport in radio galaxies and
quasars Eds. P. E. Hardee, A. H. Bridle and J. A. Zensus ASPCS 100
p. 353
[108] Peletier R. F., Davies R. L., Illingworth G. D., Davis L. E. and Cawson
M. 1990 AJ 100 1091
[109] Persson S. E. et al. 1979 ApJS 39 61
[110] Petrosian V. 1976 ApJL 209 L1
[111] Press W. H. et al. 1986 in Numerical Recipes (CUP)
[112] Quinn P. J. 1984 ApJ 279 596
[113] Rees M. J. 1978 Nature 275 516
[114] Rees M. J. 1984 ARAA 22 471
[115] Reynolds J. H. 1913 MNRAS 74 132
[116] Rocca-Volmerange B. and Guiderdoni B. 1988 A&AS 75 93
[117] Ronnback J. and Shaver P. 1997 AA 322 38
[118] Sadler E. M. and Gerhard O. E. 1985 MNRAS 214 177
[119] Sahu D. K. et al. 1996 AA 314 721
[120] Sandage A. 1973 ApJ 183 711
[121] Sandage A. and Visvanathan N. 1978 ApJ 223 707
[122] Savage B. D. and Mathis J. S. 1979 ARAA 17 73

184
[123] Schawrzschild M. 1979 ApJ 232 236
[124] Schiminovich D. et al. 1994 ApJL 423 L101
[125] Schiminovich D. et al. 1995 ApJL 444 L77
[126] Schweizer F. 1982 ApJ 252 455
[127] Searle L., Sargent W. L. W. and Bagnuolo W. G. 1973 ApJ 179 427
[128] Shaver P. A. et al. 1983 MNRAS 205 819
[129] Soares D. S. L. et al. 1995 A&AS 205 819
[130] Sparks W. B. et al. 1985 MNRAS 217 87
[131] Spinrad H. et al. 1985 PASP 97 932
[132] Thimm G. J. and Belloni P. 1994 AA 289 L27
[133] Toomre A. and Toomre J. 1972 ApJ 178 623
[134] Urry C. M. and Padovani P. 1995 PASP 107 803
[135] Vader J. P. et al. 1988 AA 203 217
[136] van Dokkum P. G. and Franx M. 1995 AJ 110 2027
[137] van Gorkom J. H. et al. 1989 AJ 97 708
[138] Wadadekar Y., Robbason B. and Kembhavi A. 1998 (submitted)
[139] Wainscoat R. J., Freeman K. C. and Hyland A. R. 1989 ApJ 337 163
[140] White S. D. M. 1980 MNRAS 191 1P
[141] Whitford A. E. 1971 ApJ 169 215
[142] Wirth A. and Shaw R. 1983 AJ 88 171
185
[143] Wrobel J. M. and Heeschen D. S. 1991 AJ 101 148
[144] Young P. J. 1976 AJ 81 807
[145] Zirbel E. 1996 ApJ 473 713

186