Citation
A neutral hydrogen survey of blue compact galaxies

Material Information

Title:
A neutral hydrogen survey of blue compact galaxies
Creator:
Gordon, David, 1950-
Publication Date:
Copyright Date:
1979
Language:
English
Physical Description:
vii, 148 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Compact galaxies ( jstor )
Galaxies ( jstor )
Hydrogen ( jstor )
Late type galaxies ( jstor )
Luminosity ( jstor )
Markarian galaxies ( jstor )
Nebulae ( jstor )
Spiral galaxies ( jstor )
Telescopes ( jstor )
Velocity ( jstor )
Astronomy thesis Ph. D ( lcsh )
Dissertations, Academic -- Astronomy -- UF ( lcsh )
Galaxies ( lcsh )
Interstellar hydrogen ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 144-147.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by David Gordon.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
023341555 ( AlephBibNum )
06566571 ( OCLC )
AAL2888 ( NOTIS )

Downloads

This item has the following downloads:


Full Text













A NEUTRAL HYDROGEN SURVEY
OF BLUE COMPACT GALAXIES











BY

DAVID GORDON


















A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE
UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA





























To My Parents,

Dora Denby Gordon

and

Maurice Gordon















ACKNOWLEDGEMENTS


I wish to express my appreciation to Dr. Stephen T. Gottesman, who

introduced me to the field of extragalactic neutral hydrogen studies.

It was his suggestions and ideas which led to this study. His ex-

perience and his suggestions, comments and encouragement have been

invaluable in the HI observations and in the analysis of the data.

Many others deserve mention here. At the National Radio Astronomy

Observatory, Dr. Richard Fisher and Dr. Patrick Crane were most helpful

in making observations with the 91 meter telescope. I am grateful to

the many telescope operators and engineers who helped make the observa-

tions a success and to Al Wu, in particular, who did an excellent job

of maintaining the receiver system.

I am also grateful to the many telescope operators and engineers

at Arecibo Observatory for performing an excellent job. I particularly

wish to thank Dr. Edward K. Conklin and Dr. Michael M. Davis for their

expert assistance in making the spectral line observations on the 305

meter telescope. Special thanks are due to Dr. Nathan A. Krumm for

his invaluable help in reducing the observations at Arecibo Observatory

in 1977. Special thanks must also be extended to Dr. Martha P. Haynes,

who provided computer programs which allowed me to read and reduce the

1978 Arecibo data tapes at the University of Florida.

Special thanks are also extended to Dr. Gerard de Vaucouleurs and

Dr. Harold G. Corwin for calculating blue magnitudes from Zwicky









photographic magnitudes for several dozen galaxies. Without their help,

the luminosities of many systems could not have been determined as

accurately as they were.

Further thanks are extended to Dr. Michael D. Desch and Mr. David R.

Florkowski for help in miscellaneous areas, inspiration, and many

interesting discussions.
















TABLE OF CONTENTS


CHAPTER PAGE

ACKNOWLEDGEMENTS . . . . . . . . . iii

ABSTRACT . . . . . . . . ... . . vi

I INTRODUCTION . . . . . . . . ... . 1

II BLUE COMPACT GALAXIES. . . . . . . . . 8

III THE OBSERVATIONS . . . . . . . . ... .16

The 21 cm Line . . . . . . . . ... .16
21 cm Spectral Line Observations . . . ... 19
Observations on the 91 Meter Telescope ...... 22
Observations on the 305 Meter Telescope. . . ... 28

IV THE DATA . . . . . . . . ... .. . .36

Optical Parameters . . . . . . .... .36
21 cm Global Parameters. . . . . . . ... 50

V ANALYSIS OF THE DATA . . . . . . .... .73

Selection Effects and Completeness of the Sample . 74
Luminosities . . . . . . .... . .. 79
Neutral Hydrogen Content . . . . . .... 88
HI Dimensions. . . . . . . . . ... 90
The HI Profiles. . . . . . . . . ... 100
Total Indicative Masses. . . . . . . ... 115
Comparisons Between Color, HI Mass, Luminosity and
Indicative Mass ............... 118
Comparison With Low Surface Brightness Systems . 132
Correlations . . . . . . . . . . 134

VI CONCLUSIONS. . . . . . . . . . ... 138

Suggestions for Future Study ...... .... 142

BIBLIOGRAPHY . . . . . . . . . . 144

BIOGRAPHICAL SKETCH. .. . . . . . . . 148















Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


A NEUTRAL HYDROGEN SURVEY
OF BLUE COMPACT GALAXIES

By

David Gordon

December, 1979

Chairman: Stephen T. Gottesman
Major Department: Astronomy

Compact galaxies are galaxies showing high surface brightnesses

and small angular dimensions. They are, morphologically, a very in-

homogeneous group, ranging from featureless objects barely distinguish-

able from stars on plates of the Palomar Sky Survey, to objects showing

various irregularities, such as multiple cores, jets, filaments, bridges

to nearby objects or other irregularities. Most of the compact galaxies

catalogued by Zwicky are red or very red systems. These appear to be

early type galaxies deficient in their outer envelopes. A minority of

the Zwicky compact galaxies, perhaps 10-20%, are blue or very blue.

These appear similar to the compact systems found on the Haro and

Markarian lists of galaxies with strong UV excesses. A fairly common

characteristic of blue compact galaxies is the presence of emission

lines in their spectra, either with or without absorption lines. The

excitation level of the emission in these objects indicates that the

lines originate in HII regions excited by hot young stars. Several








of the objects in this class, mostly dwarf systems, have been found to

be underabundant in elements heavier than helium. Possible explanations

are that these are young galaxies just commencing star formation, or

that they are old galaxies experiencing a brief burst of star formation,

to be followed by a long quiescent period.

This study is a report of the 21 cm neutral hydrogen line observa-

tions of 99 compact galaxies, most of them blue or very blue. Of these,

72 were detected in HI. These systems are found to be richer in HI

than late type galaxies of the Hubble sequence and also to have higher

surface luminosities. Approximately half of these blue compact systems

are bluer than the bluest late type galaxies. Their HI velocity profiles

are often narrow and single peaked, most often for the lower luminosity

systems. However, the evidence suggests that the compact galaxies are

rotating disk systems, perhaps with a strong central concentration of

HI. The observations also indicate that at least some of these systems

are quite extensive in their overall HI dimensions.

It is possible that these HI rich, blue compact galaxies are old

systems which have not yet completed their collapse out of the inter-

galactic medium. Primordial gas may be condensing into their central

regions and could be the source of gas for intermittent, brief bursts

of star formation.














CHAPTER I
INTRODUCTION


In addition to the many normal types of galaxies which fit nicely

into the Hubble sequence of galaxies, astronomers now recognize the

existence of various types of galaxies which are described as being in

excited states, or "active." These types of active galaxies are known

mainly under the names of Haro galaxies, Markarian galaxies, compact

galaxies, Seyfert galaxies, N galaxies and quasars (van den Bergh,

1975). There is currently wide discussion as to where, if at all, these

systems belong in the Hubble sequence of galaxies. The following study

is concerned with the compact and high surface brightness galaxies on

the lists by Zwicky, Haro, and Markarian.

Zwicky (1964) defined a compact galaxy as an object of abnormally

high surface brightness just distinguishable from stars on plates taken

with the Mount Palomar 48 inch Schmidt telescope. During the years

1960-1968, while examining the Palomar Sky Survey plates for compilation

of the Catalogue of Galaxies and of Clusters of Galaxies (Zwicky et al.,

1961, 1965; Zwicky and Herzog, 1963, 1966, 1968; Zwicky and Kowal, 1968),

Zwicky produced seven lists of what he called compact galaxies, galaxies

with compact parts, and post-eruptive galaxies. Some 30 years before

this, Zwicky claims to have known of the existence of galaxies with

super-dense stellar populations (described by Zwicky, 1971). He

believed that these compact galaxies were the super-dense stellar

systems he had predicted (Fairall, 1978).








Zwicky distributed his seven lists privately to various astronomers

and later published them in book form. The Catalogue of Selected

Compact Galaxies and of Post-Eruptive Galaxies (Zwicky, 1971) contains

some 2300 systems. Morphologically, these systems make up a very

inhomogeneous group. As seen on the Palomar Sky Survey, many of them

have a small saturated image surrounded by a small amount of

nebulosity, while others show various irregularities, such as jets,

rings, bright knots, distorted spiral arms, multiple nuclei and bridges

to nearby objects. A common feature among these systems is a region or

regions of high surface brightness (Sargent, 1970b). The majority of

these galaxies have normal reddish colors, as seen in E and SO galaxies

and have normal absorption line spectra. A minority, perhaps 10-20%,

are blue or very blue and often have emission lines in their spectra,

indicative that they have regions of ionized hydrogen. The designation

of color by Sargent (1970b) is used here, where a "blue" image is of

approximately equal brightness on both plates of the Palomar Sky

Survey; "very blue" is brighter on the blue plate; "red" is brighter

on the red plate; and "very red" is much brighter on the red plate.

The Markarian galaxies are blue galaxies discovered on objective

prism plates taken with the 1 meter Schmidt telescope of the

Byurakan Observatory by Markarian and his co-workers. They have

published 11 lists so far, containing 1095 galaxies (Markarian, 1967,

1969a, 1969b; Markarian and Lipovetskii, 1971, 1972, 1973, 1974,

1976a, 1976b; Markarian et al., 1977a, 1977b). A 1.50 prism was used,

yielding a dispersion of 2500 Ao/mm at H The aim of their study is

to catalog galaxies with intense ultraviolet continue between apparent

magnitudes 13 and 17.









On these lists, the objects are classified into one of two groups

based on the appearance of their objective prism spectra (Markarian,

1967). Objects having sharp spectra (like stellar spectra) are called

type s. A bright nucleus is the source of their UV continue. Some of

these s types are spiral galaxies with bright nuclei and some are

Seyfert galaxies with broad emission lines arising in their nuclei. In

the second group, type d, the boundaries of the spectra are diffuse,

indicating that the UV continuum originates in an extended region and

not just the nucleus. Intermediate types are designated as types sd or

ds, depending on which type they resemble most. Additionally, a common

characteristic of Markarian galaxies is the presence of narrow emission

lines in their spectra.

Haro (1956) reported on the discovery of 44 blue and very blue

galaxies. These were detected in the course of an objective prism

search for very blue stars, using the Schmidt telescope of the

Tonontzintla Observatory. Like the Markarian galaxies, they show

strong UV excesses. These systems generally fit into the diffuse

(class d or ds) spectroscopic subgroup of the Markarian objects. Like

the Markarian galaxies, emission lines are a fairly common feature of

their spectra.

Many of the Markarian and Haro galaxies are similar to the blue and

very blue high surface brightness compact and post-eruptive systems

listed by Zwicky (1971), in that they are of similar colors, are of

small angular dimensions and have a region or regions of high surface

brightness. Some of these have circular or elliptical shapes like

some of the featureless Zwicky compacts, while others display various

irregularities, such as jets, filaments, multiple cores and other









irregularities. Thus, morphologically, the three groups show many

common properties. The overlap of their properties is further illus-

trated by the fact that of the 99 galaxies reported on in this study,

20 are on both the Markarian and Zwicky lists; 10 are on both the Haro

and Markarian lists; and 2 are on all three lists. Thus, there is

ample reason to group objects from these three lists together in a

study of the properties of blue compact galaxies.

The original definition of compactness (Zwicky, 1964) as being

just distinguishable from stars on plates taken with the 48" Palomar

Schmidt telescope and with diameters of 2-5 arc seconds has proven too

restrictive to be of practical use. This definition has not been

rigorously followed by workers in the field, including Zwicky himself.

Zwicky's (1971) definition of compactness also requires that the

surface brightness be brighter than the 20th magnitude per square arc

second. This requires surface photometry and has not been attempted for

very many compacts. The more or less accepted procedure is to classify

as compact those systems which show, on the Palomar Sky Survey, a small

saturated core, either regular or irregular, surrounded by none or

small amounts of nebulosity, showing no normal structure (such as

spiral arms) but possibly showing irregular structure, such as jets,

filaments, bridges or double cores. This is the definition that is

used in this study to select compact galaxies from the lists of Zwicky,

Haro and Markarian. For the benefit of the reader, we reproduce in

Figure I-1 several examples of compact galaxies. These are negative

prints, taken from the Palomar Sky Survey.

In the following study, we report on neutral hydrogen observations

of 99 compact galaxies. Almost all of these galaxies were blue or very

















Examples of blue compact galaxies taken from the
Palomar Sky Survey.

M7 (= VIIZwl53) has an irregular shaped bright
core surrounded by a small amount of nebulosity.

M297 (= Arp209) has a "hat" shaped core composed
of two bright knots side by side, and a small
amount of nebulosity.

IZw207 is a "boomerang" shaped arc of bright
knots.

IIZw40 is a dwarf blue compact showing a bright,
nearly stellar core and a fainter nebulosity
containing two filaments.

Haro29 (= Izw36 = M209) is a dwarf blue compact
showing two bright condensations, a jet, and a
"fan" shaped nebulosity.

Haro25 (= M727) shows a bright stellar core with
a very small amount of nebulosity and is barely
distinguishable from a star.


Figure I-1.













.







*
S '. .'















- 'iZw_ 20---









,.














**
....,:.:.-. ,. . ": :
















A 297












S. . *








r .
S T. O 2. : . "













. .
i, .% "
" ,. .





"g.-7 n '


-e-






7


blue, and thus our discussion will be concentrated on this type. We

shall be concerned with what can be learned from their neutral hydrogen

global properties when considered in conjunction with their known

optical properties. For our discussion and analysis, a Hubble constant

of 75 km/sec/Mpc will be used throughout. A more detailed discussion

of blue compact galaxies follows in Chapter II. In Chapter III, we

describe how the 21 cm spectral line observations were made and reduced.

Optical and neutral hydrogen global data are presented in Chapter IV.

Discussion and analysis of the data are given in Chapter V. Conclusions

and directions for future work are given in Chapter VI.















CHAPTER II
BLUE COMPACT GALAXIES


The lists of Zwicky, Markarian and Haro contain some 3500 galaxies.

The majority of these objects are of small angular dimensions and of

high enough surface brightnesses to be classified as compact galaxies.

The Haro and Markarian compact galaxies are all blue or very blue as a

result of their selection criteria. Zwicky compacts, on the other hand,

cover a wide range of colors--from as red as the reddest Hubble sequence

galaxies to bluer than the bluest type I irregular galaxies. No more

than 20% of these are in the blue or very blue category (Sargent, 1970b).

The objects cataloged by Zwicky (1971) were selected at random,

but they were not chosen in any uniform manner over the entire sky and

are not complete down to any limiting apparent magnitude. The Markarian

lists, on the other hand, appear to be complete down to a limiting

photographic magnitude of mp = 15.5 (Sargent, 1972).

In 1969, Zwicky began a complete survey for compact galaxies on

12 plates taken with the 48" Schmidt telescope (Zwicky, 1971). He

appears not to have completed or published more than a few preliminary

results of this study before his death in 1974. However, Rodgers et al.

(1978) studied the compacts marked on two of these plates. Out of 348

objects marked on one plate, using Sargent's (1970b) color designations,

they found that 232 were very red, 79 were red, 25 were blue and 12 were

very blue. Thus perhaps as few as 10% of the Zwicky compacts are blue

or very blue.








Studies of these red and very red compacts are reported by

Fairall (1971, 1978), Kormendy (1977), Sargent (1970b) and Rodgers

et al. (1978). These systems typically show normal absorption line

spectra typical of early type galaxies. Surface photometry of these

systems indicates that their central surface brightnesses are not so

much higher than is seen in normal early type galaxies (Kormendy, 1977;

Rodgers et al., 1978). Their primary difference from normal E and SO

galaxies seems to be a deficiency in their outer envelopes (Kormendy,

1977; Fairall, 1978; Rodgers et al., 1978). Thus it appears that these

systems do not have the abnormally high stellar densities expected by

Zwicky. Fairall (1978) has suggested that these systems have lost

their outer envelopes as a result of tidal stripping.

We shall now concentrate our discussion on the properties of blue

compact galaxies, with which this study is primarily concerned. DuPuy

(1968, 1970) has studied Haro galaxies. Sargent (1970a, 1972) has

surveyed Markarian galaxies. Sargent (1970b) made a study of Zwicky

compacts which was heavily weighted towards blue objects.

These blue systems are characterized by a strong UV continuum,

colors which are often bluer than the bluest late type galaxies, and

moderate or strong emission lines--typically the Balmer lines and

forbidden lines of oxygen, neon and nitrogen. The emission lines and

the UV continuum are generally concentrated in, but not restricted to,

the nuclear regions. About half of these systems also show absorption

lines typical of the stellar populations in late type galaxies. All 31

Haro galaxies studied by DuPuy (1968, 1970) showed emission lines.

Those brightest in the UV showed the strongest emission lines. Among

the Markarian objects, some 80% seem to show emission lines (Sargent,








1972). Sargent (1970b) found that all Zwicky compact with (U-B)

between +0.20 and -0.15 had both emission and absorption lines and

that all those bluer than (U-B) = -0.15 had emission lines only.

The levels of excitation seen in blue compacts are comparable to

those of galactic HII regions. Thus it is generally agreed that the

emission lines originate in HII regions excited by hot stars. Their

strong UV continue appear to be stellar in nature, the result of a

high number of hot stars (DuPuy, 1970; Sargent, 1970a, 1970b, 1972;

Searle and Sargent, 1972; Forrester, 1973).

The space density of blue, UV intense galaxies has been estimated

by Sargent (1972). For absolute photographic magnitude M between -20

and -22, about 2.5% of all galaxies are UV intense. Their abundance

increases towards lower luminosities. At M = -17, about 7% of all

galaxies are UV intense and at Mp = -14, it may be as high as 10%.

The luminosities of blue compact galaxies cover a broad range. At

the high end, they are as luminous as normal galaxies, and at the low

end, they are as intrinsically faint as the smallest dwarf galaxies.

In the surveys by Sargent (1970b, 1972), the absolute photographic

magnitudes of the Zwicky and Markarian narrow emission line galaxies

were found to extend from -21.6 to -14.9, and from -21.5 to -13.9,

respectively. DuPuy (1970) finds a similar range in the absolute

visual magnitudes of Haro galaxies, from -21.6 to -13.7.

With their broad range of luminosities, blue compact galaxies are

not a homogeneous group. However, the low luminosity dwarf systems may

form a fairly homogeneous subgroup. Sargent (1970b) noted several low

luminosity blue systems fainter than M = -19 among the Zwicky compacts.








Eleven similar dwarf systems fainter than Mp = -16 were identified by

Sargent (1972) from the Markarian lists. Several of the Haro galaxies

also fit into this subgroup. These systems are similar in that they are

all nearly uniformly high in excitation levels, have strong emission

lines relative to their continuum, have physically small emission line

regions,typically a few hundred parsecs across, and are often bluer

than the bluest galaxies of the Hubble sequence (Sargent, 1972). An

exact upper luminosity limit for considering an object a dwarf is not

clearly established. An absolute photographic magnitude of approxi-

mately -17 or -18 appears to be a reasonable upper limit.

Two such blue dwarf compact systems were labelled as "isolated

extragalactic HII regions" by Sargent and Searle (1970). These types

of systems are similar to the largest and highest excitation HII regions

found in the spiral arms of giant late type galaxies, and yet they often

appear as isolated systems. IlZw40 and IZwl8 were studied, using high

resolution spectroscopy. Searle and Sargent (1972) reported on the

abundances found in these two systems. They were found to have normal

and presumably primordial abundances of helium, but significant under-

abundances of heavier elements. IIZw40 and IZwl8 were found to be

underabundant in oxygen and neon by factors of approximately 3 and 7,

respectively, compared to normal HII regions in the solar neighborhood.

(Bergeron (1977), using the same data, gets greater underabundances,

of 6 10 for IIZw40 and 30 50 for IZwl8.) These were the first

metal poor, Population I systems found, raising the question as to

whether they are young galaxies. Obviously they could not have been

producing 0 and B stars at their current observed rates for 1010 years

and have remained metal poor. IIZw40 would have converted approximately








1/4 of its total mass into elements heavier than helium if this were

the case (Searle and Sargent, 1972).

Other examples of metal poor blue dwarf systems have been identi-

fied since. Neugebauer et al. (1976) found that Haro3 (= M35) and

M59 are underabundant in oxygen by a factor of -2 and Haro4 (= M36) is

underabundant by a factor of ~4, with respect to the solar neighbor-

hood. Five additional metal poor systems are found in a study by

Ulrich (1971). Alloin et al. (1978) analyzed spectroscopic data for

these five systems and six others from Sargent and Searle (1970) and

Neugebauer et al. (1976). All 11 of these systems are underabundant

in oxygen, neon and nitrogen. They find oxygen to be underabundant by

factor of ~2 for M19, ~3 for Haro3, -10 for Haro4, ~6 for IIZw40,

-7 for M162, -11 for M193, -6 for M156, -45 for M171, -40 for IZwl8,

-5 for M108 and -5 for M59.

Searle et al. (1973) consider the nature of the bluest galaxies

known--the isolated, metal poor, dwarf blue compacts. These systems

also appear to have a large fraction of their total mass in the form of

HI gas (Searle and Sargent, 1972; Gottesman and Weliachew, 1972). Their

sizes and their stellar and gaseous content are similar to giant HII

regions seen in the spiral arms of giant Sc galaxies. These systems,

which all seem to be bluer than (B-V) = +0.3, are explainable by two

possible hypotheses, according to Searle et al. (1973). Either they are

younger than 10 years, or they are of normal ages (-1010 years) and

have a rate of star formation at this current epoch which greatly

exceeds their past average rate.

Model calculations by Searle et al. (1973) indicate that a

galaxy -2 x 108 years old with a Salpeter initial luminosity function








and a uniform rate of star formation would have the colors of the

bluest dwarf compact galaxies. If these systems are young, and if

their rate of formation has not increased with time, then there should

exist at least 5 times as many 10 years old and at least 50 times

as many 110 years old. The models by Searle et al. (1973) indicate

that such a galaxy would brighten by -0.5 magnitudes in aging from

2 x 108 years to 109 years and by -1.2 magnitudes in aging to 1010

years. Thus old dwarf galaxies should be observationally more abundant

than young, very blue dwarf galaxies by a factor of at least 50 to 100.

As stated earlier, at M = -14, 1 galaxy in 10 appears to be a very

blue dwarf compact (Sargent, 1972). This high an abundance argues

against the young galaxy hypothesis.

In the second hypothesis, called the flashing galaxy or star burst

model, star formation occurs in brief intense bursts. The blue dwarf

compacts are said to be undergoing a burst of star formation equal to

s times their star formation rate averaged over their total lifetime

(Searle et al., 1973). About 1/4 of the light from a galaxy 110 years

old comes from stars no more than 108 years old. Thus a flash of star

formation of strength s = 4 or more would have profound effects on the

colors of a galaxy. The colors of the bluest galaxies would require

flashes of strength s = 10 20 to produce. Such a flash would increase

the luminosity by 1 2 magnitudes and would require -3 x 108 years for

the color and luminosity to return to normal. Searle et al. (1973)

propose that the blue dwarfs have ages of -1010 years and have under-

gone perhaps 5 10 brief bursts of rapid star formation, each lasting

-108 years. Nonflashing galaxies of this type should have a space

density -2 5 times those of the very blue dwarfs. Flashing galaxies









will be brightest during and just after a flash, so at a given absolute

magnitude, the ratio of interflashing to flashing galaxies will be

fairly low, perhaps as low as 1.

Searle et al. (1973) conclude that extremely blue galaxies of

high luminosity must be very rare. They consider a galaxy as composed

of statistically independent cells in which star formation is either not

occurring or is occurring in a burst. They use a cell size approxi-

mately equal to the smallest blue dwarfs, or M = -14. Thus a galaxy

of M = -19 has 100 such cells. They calculate the probability of at

least half of the cells in a galaxy undergoing flashes at any given

time. These probabilities are 0.16 at Mp = -15, .02 at Mp = -16 and

10-4 at M = 17. Thus extremely blue galaxies brighter than M = -17

should be extremely rare, according to this model.

Nondwarf blue compact systems have not received as much attention

as the blue dwarf systems. These galaxies are similar in size and

luminosity to late type Hubble sequence galaxies, but they do not show

any obvious spiral structure. Some of these systems do show hints of

spiral structure on direct plates more sensitive and with a greater

plate scale than the Palomar Sky Survey (Fairall, 1978).

O'Connell and Kraft (1972) obtained a rotation curve of IZw129, a

blue luminous compact galaxy with a bright central core and a faint fila-

mentary structure. Its absolute visual magnitude is -21.1. They find

clear evidence for rotation and calculate the total mass interior to

the last point of their rotation curve. The mass determined for this

system is quite low compared to its luminosity.

O'Connell (1979) studied nine nondwarf blue compacts, all brighter

than M = -19. Slit spectra were taken of these systems. Some








internal structure was seen in their HII distribution, such as evidence

of ring-like or of one-armed spiral structure. Good rotation curves

were obtained for six of these systems. The remaining three showed

no evidence of rotation. Two of these were face-on while the third,

IVZw153, was a double system and needs more study. The rotation curves

for the six systems were found to be consistent with circular motion.

These rotation curves resemble those of Sc and Sd galaxies. Unfortu-

nately, only one can be traced to a velocity turnover point. The total

mass-to-luminosity ratios for these six systems averages to -1.2

(scaled to a Hubble constant of 75) within the regions observed. This

is about 1/3 of the values typically found in the interiors of late

type spirals.

Neutral hydrogen studies of blue compact galaxies have generally

been quite successful. Bottinelli et al. (1973) detected at least nine

Haro galaxies in HI with the Nancay radio telescope. The HI masses

found were large compared to their total luminosities and estimated

total masses. Markarian galaxies also were found to be rich in HI by

Bottinelli et al. (1975). Blue Zwicky compacts showing emission lines

were shown to HI rich as well (Chamaraux, 1977).

Gottesman and Weliachew (1972) made a low resolution, aperture

synthesis study of the HI in IIZw40. It was found to be very extensive

in HI. This object has an HI core approximately equal in size to its

optical halo and an HI halo approximately six times larger than this.

Approximately half of its HI mass is in this HI halo. They found non-

conclusive evidence of rotation and were able to estimate a lower

limit to the total mass. The total mass is found to be at least twice

the HI mass (adjusted to a Hubble constant of 75).















CHAPTER III
THE OBSERVATIONS


Observations of the 21 cm line of neutral atomic hydrogen were made

using the 91 meter (300 foot) radio telescope of the National Radio

Astronomy Observatory (NRAO) in Green Bank, West Virginia, and the

305 meter (1000 foot) radio telescope of Arecibo Observatory, near

Arecibo, Puerto Rico. The NRAO is operated by Associated Universities,

Inc., under contract with the National Science Foundation. The Arecibo

Observatory is part of the National Astronomy and Ionosphere Center,

which is operated by Cornell University under contract with the National

Science Foundation. A discussion of the 21 cm emission line and how

the observations were made follows in this chapter.


The 21 cm Line


A discussion of the nature of the 21 cm emission line is given by

Kerr (1968). The ground state of atomic hydrogen is split into two

hyperfine sublevels, the upper state being metastable. In the lower

energy state, the magnetic dipole moments, or spins, of the proton

and the electron are in opposite directions, or antiparallel. In the

higher energy state, the two spins are parallel. The difference in

energy between these two states corresponds to the energy of a photon

of wavelength 21.11 cm and frequency 1420.4058 MHz. The emission or

absorption of a 21 cm photon is produced by a transition between these

two energy levels, often called a spin flip transition.









For an HI region in thermodynamic equilibrium, the relative popula-

tion of the lower and upper levels, no and n1, respectively, is given

by the Boltzmann equation,


n, 91 -hv
= exp(- ), III-1
n0 g0 kTs


where k is Boltzmann's constant, h is Planck's constant, Ts is known as

the spin temperature and v is the frequency, 1420.4058 MHz. The

statistical weights of the lower and the upper energy levels, g0 and gl,

are given by gi = 2Fi + 1, where Fi is the sum of the spin of the

electron (1/2) and the proton (+1/2). Thus F0 = 0, F1 = 1, g0 = 1

and g1 = 3. The term (hv/k) = 0.07 K and will always be much less than

the spin temperature. Thus, exp(-hv/kTs) will always be close to unity.

For example, at Ts = 10 K, nl/n0 = 2.9806; at Ts = 100 K, nl/n0 = 2.9981;

and at Ts = 1000 K, n /n0 = 2.9998. Thus about 3/4 of the HI atoms in

a galaxy's interstellar medium will be in the upper ground state at all

times. This means that it is not necessary to know the spin temperature

of an HI region in order to determine the number of neutral hydrogen

atoms present.

The Einstein coefficient, A21, for the 21 cm spin flip transition

is 2.85 x 10-15 sec-1. Thus the radiative lifetime of the upper level

is very long, approximately 1.1 x 107 years. In the interstellar

medium of a galaxy, the time between collisions will be much shorter

than this, so that most transitions will be due to collisions. A

typical HI atom will move up and down between the two levels about

every 400 years due to collisions, but only about every 11 million

years will it undergo a spontaneous downward transition with the









emission of a 21 cm photon. If collisions are dominant in populating

the two levels, then the spin temperature will effectively be equal to

the kinetic temperature.

The equation of transfer can be solved for the brightness

temperature, TB, of an HI region. One gets


TB = Ts (1 e-) III-2


where T is the optical depth of the HI region. It is generally

believed that the optical depth in HI regions is small. Thus

equation III-2 reduces to


TB TTs 11I-3


Assuming small optical depth, the total number, NH, of hydrogen

atoms per cm2 along the line of sight is



NH = 3.88 x 1017 f_ TB d III-r



where v is in KHz. However, HI spectra are normally plotted as a

function of velocity, v, in km/sec. The number of hydrogen atoms per

cm2 along the line of sight is then



NH = 1.823 x 1018 1 T *dv III-5



Flux, rather than brightness temperature, is the observed quantity

in radio astronomy observations. Equation III-5 can be reduced to the

total neutral hydrogen mass by








MHI = 2.356 x 105 D S -dv, III-6



where MHI is the HI mass in units of solar masses, S is the flux in

Janskys, and D is the distance to the HI region in Mpc. The integral

need only be evaluated over the velocity range of the HI region and will

simply be the area under the HI profile, if it is plotted in the usual

units of Janskys versus km/sec.


21 cm Spectral Line Observations


For the type of radio telescope considered here, the desired signal

from an astronomical source is focused by a reflector onto the front end

of a radio frequency (RF) receiver, known as the antenna, or feed,

mounted at the focus of the reflector. For 21 cm observations, the re-

ceiver is designed to receive both orthogonal polarizations, and to have

its maximum sensitivity at or near the expected Doppler frequency of

the spectral line. The RF signal from the feed is combined with a

local oscillator (LO) frequency to produce a lower, intermediate fre-

quency (IF) signal. In spectral line observations, the LO must be

computer controlled to correct for the constantly changing motion of

the earth. The IF power is passed into a baseband mixer and

then into an autocorrelator which produces the spectrum.

An autocorrelator works in arbitrary units. In order to determine

the total amplitude of the signal, the system temperature, Tsys, must

be determined independently of the autocorrelator. This is done by

monitoring the total IF power and comparing it to a source of known

noise temperature. The power from the noise source, of noise tempera-

ture TN, is periodically injected into the receiver. Continuum








receivers measure the power when the noise source is on, P on' and

when it is off, Poff* The system temperature is given by

(Pon+ Poff )
sys 2(Pon Poff) -7

The total bandwidth B is fed into the autocorrelator, which uses

N channels per polarization. Each channel has a time delay of n -At,

where n is the channel number (from 1 to N) and At = 1/2 B. The auto-

correlator computes the autocorrelation function, ACF(n At), of the

spectral bandpass. The spectral bandpass can be considered to be a

function of time, y(t). Then the autocorrelation function is

T
ACF(n At) = limit 1 T y(t) -y(t +n -At). II-8
2

Each channel of the autocorrelator gives one point of the autocorrela-

tion function. The power spectrum is obtained by taking the Fourier

transform of this autocorrelation function, usually by an on-line

computer. The result is an N channel spectra of total bandwidth B.

The channels will be spaced B/N apart and each channel will have a half

power width, or resolution, of Af = 1.21 B/N (Shalloway et al., 1968).

This spectrum is not in a readily usable form. The channel values

are scaled arbitrarily; the bandpass is not smooth or flat; and the

bandpass is usually dominated by the system noise temperature. For these

reasons, 21 cm spectral line observations of external galaxies are

usually made in what is called the "total power" mode. In this scheme,

an "ON" spectrum is taken at the position of the source, and an "OFF"

spectrum is taken at a blank sky position. The OFF spectrum will be the









receiver passband function, while the ON spectrum will be the receiver

passband function plus the source signal. Let ON(i) and OFF(i) be the

relative intensities in the ith channel of the ON and the OFF spectra,

respectively. Then the antenna temperature, T(i), of the source in

channel i is


T(i) = ON(i) OFF(i)
T(i) OFF(i) sys 111-9



In practice, it is necessary to get a good match between the ON and

the OFF spectra. This is usually accomplished by using the same

integration time and center frequency and by following the same range

of altitude and azimuth in the sky so that any instrumental effects

will be repeated in both the ON and the OFF spectra. The two should

also be taken closely together in time to avoid slow changes in the

gain of the receiver.

The system sensitivity will be the RMS noise temperature, ATrms

given by

y'T
AT sys T III-10
rms Af-t


where t is the total integration time and y is a constant (21) which

depends on the system and how it is operated.

The RMS flux is given by


2k' ATr
AS rms III-11
rms Ae
e








where Ae is the effective area of the reflecting dish, typically

50-60% of its geometric area. In practice, the minimum detectable flux

is about three times the RMS flux.


Observations on the 91 Meter Telescope


The 91 meter (300 foot) diameter telescope of the NRAO is located

in the mountains of West Virginia, near the town of Green Bank. Its

parabolic reflecting surface is an aluminum mesh which is effectively

a solid reflector at radio wavelengths down to about 6 cm. Being

solely a transit telescope, it points always at the meridian and is

movable in declination between the North Celestial Pole and -19

declination. The front end of the 21 cm receiver is located at the

focus on a travelling mount which allows the feed to be moved up to

1/2 degree to the east and to the west of the meridian. This allows

a transiting source to be tracked for 4 sec 6 minutes ( =

declination).

The orthogonal linear polarizations are fed into cooled parametric

amplifier receivers, giving system temperatures of typically 50 K in

each polarization. A bandwidth of 10 MHz was used for all the observa-

tions. The spectra were produced with the NRAO Mark III autocorrelation

receiver (described by Shalloway et al., 1968). This is a one bit

digital machine containing 384 channels. Thus 192 channels were used

for each polarization. The channel spacing was approximately 11.1 km/sec

and the channel resolution was approximately 13.4 km/sec. Employing

this system, the constant y in equation III-10 is approximately 1.6.

For each source, an OFF scan was usually taken first, selected to

end 70 seconds before the ON scan was to begin. This allowed time to








move the feed back to its starting position before beginning the ON

scan. Both the ON and the OFF scan were of the same duration and

covered the same range in altitude and azimuth. The autocorrelator

was set to integrate for 20 seconds. Each of these 20 second spectra

is called a "record." All the records taken during a single transit

are called a "scan." These individual records were initially recorded

on a 7-track magnetic tape. All the records in each scan were daily

averaged together by the computer staff in Charlottesville, Virginia,

and written onto a 9-track tape.

We received 8 days of observing time in March 1977 and 12 days in

September 1977. The sources were observed each transit, excluding

periods when the telescope was shut down for maintenance and malfunc-

tions. Most of the sources were observed for at least 5 transits and

several were observed for as many as 20 transits.

For proper calibration of the system, drift scans of 21 continuum

sources were taken at various times during gaps in the observing

schedule. Only unconfused sources from Bridle et al. (1972) were

used. A drift scan consisted of setting the telescope at the meridian

and at the declination of the calibration source several minutes before

transit and allowing the source to drift through the beam of the tele-

scope. The noise tubes were fired just before and after the source

drifted through the beam. The signal due to the noise tubes and the

source were recorded as deflections on a strip chart. Since the

temperature of the noise tubes and the flux strength of the calibration

sources were known, the sensitivity of the system in Jy/K could be

found. The results of all drifts in both polarizations were averaged

together. The final results for the 21 calibration sources are








presented in Table III-1. Column (1) gives the name of the calibration

source. Column (2) gives the 1950 declination. Column (3) gives the

sensitivity of the system in Jy/K. Column (4) gives the number of

drift scans used. Column (5) is the flux of the calibration source in

Janskys at 1400 MHz, from Bridle et al. (1972). The sensitivity was

found to be a fairly smooth function of the declination. A parabola

has been fit, by the method of least squares, relating the sensitivity

of the system S, in Jy/K, to the declination 6, in degrees. This

parabola is given by


S = 0.99775 0.004408 + 0.0000749 62 III-12


The RMS uncertainty is 0.02 Jy/K, or about 2%. The data points and

the parabolic fit are shown in Figure III-1.

The reason for this declination dependency is presumably due to

deformation of the parabolic reflector. The calibration curve is not

centered on the zenith as expected. Greatest sensitivity occurs at

approximately +290 declination, whereas the zenith is at approximately

+380. This calibration curve agrees quite well with one obtained for

the 91 meter telescope by Fisher and Tully (1975). However, their

calibration curve reaches its maximum sensitivity very near the zenith.

A small correction was also required for the variation of sensi-

tivity with hour angle. When east or west of the meridian, the para-

bolic reflector is used off axis, resulting in decreased efficiency.

Several continuum sources were tracked and the average flux was found

to be approximately 4.5% less than the peak flux at the meridian. Thus,

in the final analysis, all spectral intensities were multiplied by

1.045.

















Table III-1.



Source

(1)
P2128-12
3C422
3C C8
3C 132
3C234
3C48
3C21 7
3C197. 1
3C1 47
3C325
3C371
aC-02. 79
DWL 716 *00
3C435
3CA49
3C0
PK 0 11 +19
3C2;86
3C131
4C33.48
3C351


Sources Used for Calibrating the
91 Meter Telescope.


Dec Jy/K


(2)
-12 2C
-2 47
10 18
22 45
29 2
32 54
38 1
47 12
49 50
62 50
69 4C
-2 32
0 40
7 20
13 38
15 24
19 25
30 46
31 24
33 24
60 49


(3)
1.0605
1.0268
0.9302
0.9246
0.9296
0.9198
0.9460
0.9528
0.9076
1.0 131
1. 0284
0.9958(
1.0243
0.9567
0. 9696
0.9267
0.9591
0.0612
C.9491
0.8986
1.0468


# Flux


(5)
1.46
2.24
9.56
3.25
5.35
15.29
2.12
1.87
21.20
3.62
2.59
2.01
2.18
2.01
2.69
1.98
6.73
14.78
2.90
3.76
3.52
















Figure III-1. Sensitivity curve used for calibrating the 91 meter spectra. The data
points are taken from columns 2 and 3 of Table III-1. A parabola has
been fit to these points using the method of least squares. The sensi-
tivity S, in Jy/K, is related to the declination 6, in degrees, by

S = 0.99775 0.004408- 6 + 0.0000749- 62 + .02.

In order to calibrate each source, the spectral values of each channel
are multiplied by the appropriate value of S.













Si--- -- --i-- - --H-


N'

K
K


4 /

/










fli) 21' Sa* I li I F. gifl Sp'r









At the end of each of the two observing sessions, several days were

spent reducing the data at the NRAO headquarters in Charlottesville,

Virginia. The work was done on a CRT terminal using the NRAO T Power

and S Power programs. For each source, the ON and its corresponding

OFF were difference according to equation 111-9, and then all transits

were averaged. Any bad data, of course, were excluded. For sources

in which a spectral line was present, a polynomial curve was fit to

the baseline on either side of the spectral line. The lowest order

polynomial as was suitable was used, seldom going beyond third order.

This curve was then subtracted from the spectrum to leave a flat base-

line. The spectra were then recorded on a 9-track magnetic tape and

taken back to the University of Florida for further reduction.

At the University of Florida, corrections for the variation of

system sensitivity with declination and with hour angle were applied,

converting the spectral values from units of temperature to units of

flux. The two polarizations were averaged together and the spectra

were plotted on a Gould electrostatic plotter. Final RMS uncertainties

were on the order of 5-10 mJy per channel. These spectra are shown

in the next chapter. The computation of various parameters from these

spectra is also developed in the next chapter.


Observations on the 305 Meter Telescope


The 305 meter (1000 foot) telescope of Arecibo Observatory is

located in a mountainous region of Puerto Rico, approximately 15 km

south of the city of Arecibo. Its reflecting surface is a fixed

spherical aluminum dish, 305 meters in diameter and covering 18 acres.

A huge platform is suspended above the dish, supported by cables from









three large towers. The feeds are mounted beneath this platform and

can be steered to point anywhere within about 200 of the zenith.

Spherical aberration is corrected for by using line feeds. The 21 cm

feed has a range in declination between about -1 and +38.

The Arecibo observations described here were made in two observ-

ing sessions. A 17 day observing run was made in April and May of

1977. As these observations were quite successful, additional time

was requested. The second observing session, covering eight days, was

made in June and July of 1978.

The feed used for the 21 cm observations was a 40 foot line feed

which was illuminated by an annulus of approximately 210 meters outer

diameter and 90 meters inner diameter. The receivers were uncooled

parametric amplifiers which accepted the two circular polarizations

from the feed. For zenith angles of 100 or less, the feed was il-

luminated entirely by the reflecting surface of the telescope. Within

this range, the half power beamwidth was -3.3 arc minutes and system

temperatures were typically 70-80 K. At zenith angles greater than

100, the feed began picking up ground radiation, increasing the system

temperature. Other side effects were an increase in and a distortion

of the beamwidth, and a reduction in system sensitivity. This effect

was not too serious inside a 140 zenith angle. However, near the

zenith angle limit, the sensitivity dropped to -60% of its maximum

and the system temperatures rose to -150 K. For these reasons, most

of our observations were made within about 14 of the zenith, but a

few sources had to be observed at less favorable parts of the sky.

One of the biggest problems with this telescope is the uncertainty

in the position of the beam. The RMS pointing accuracy is on the order









of 30" of arc. Thus pointing errors of 1' of arc are not uncommon.

This uncertainty has the effect of limiting the determination of HI

masses to an accuracy of -20%.

The spectra were produced using 504 channels of a 1008 channel

autocorrelator. Each polarization was split into a spectrum of 252

channels. Total bandwidths of either 5 MHz or 10 MHz were used. The

channel spacing was approximately 4.2 km/sec with a resolution of ap-

proximately 5.1 km/sec at 5 MHz and twice these values at 10 MHz. With

a 5 MHz bandwidth, three level sampling of the data could be employed.

In this mode, the constant y in equation III-10 was -1.2. With a

10 MHz bandwidth, the autocorrelator was usable only as a one bit

sampler. In this case, y was -1.6. For a 5 MHz and a 10 MHz spectrum

of identical integration times and smoothed to the same velocity

resolution, the RMS uncertainty of the 5 MHz spectrum would be ap-

proximately 3/4 as much as the 10 MHz spectrum. Thus it was advan-

tageous to use a 5 MHz bandwidth.

All of the Arecibo observations were made in the total power mode.

An ON scan of five minutes duration was taken at the source position,

followed shortly afterwards with a five minute OFF scan covering the

same range in zenith angle and azimuth. The observations were monitored

on-line using a Harris Datacraft computer and a CRT terminal. The

Arecibo HI system is undoubtedly the most sensitive in the world. In

almost every case where a spectral line was present in the bandpass, it

was evident, though perhaps noisy, after the first set of ON and OFF

scans. In several cases, this allowed changing the observed velocity

to place the spectral line in the center of the bandpass. Most of the

observations were made using a 5 MHz bandwidth. A 10 MHz bandwidth was









employed at times to search for several Haro galaxies with unknown

velocities and several galaxies which did not show a spectral line at

the published optically determined velocities. If a signal was found,

the system was usually switched back to a 5 MHz bandwidth. Typically,

between three and ten sets of ON and OFF scans were taken for each

source to get a good signal to noise ratio and a low RMS uncertainty.

The 21 cm feed is tunable in frequency by moving it up or down in

its housing. For instance, to get maximum sensitivity at -1390 MHz, the

feed is moved 14 inches closer to the reflector than the normal focus

position for a frequency of 1415 MHz. Unfortunately, the feed cannot

be adjusted individually for each source as it requires several hours

for a single adjustment. For the two sets of observations reported

here, the feed was set for optimum sensitivity at a frequency of

-1415 MHz for about 60% of the observations andat -1390 MHz for about

40% of the observations.

The variation of telescope sensitivity with frequency for these two

settings of the feed has been determined by the Arecibo staff. Plots

of system sensitivity in K/Jy versus frequency were obtained from

Dr. Mike Davis, allowing the data to be calibrated and corrected for

this effect. Figure 111-2 shows the telescope sensitivity versus

frequency for the 1415 MHz setting. The peak sensitivity is -8.5 K/Jy.

Figure III-3 shows the same information for the 1390 MHz setting. The

peak sensitivity is -8.1 K/Jy for this setting. The sensitivity of

the system over the velocities observed ranged from -8.5 K/Jy to

-6 K/Jy.

Zenith angle corrections have also been determined by the Arecibo

staff. No correction is required for zenith angles of 100 or less.
























S6



N)
5








1395 MHz 1400 MHiz 1405 MHz 1410 MHz 1415 MHz

FREQUENCY

Figure III-2. Sensitivity versus frequency curve for the 305 meter telescope with
the feed set for optimum efficiency at a frequency of 1415 MHz, as
determined by the Arecibo Observatory staff.



















7 7
E-








5





1370 MHz 1375 MHz 1380 MHz 1385 MHz 1390 MHz 1395 MHz 1400 MHz 1405 MHz

FREQUENCY


Figure 111-3. Sensitivity versus frequency curve for the 305 meter telescope with the feed
set for optimum efficiency at a frequency of 1390 MHz, as determined by the
Arecibo Observatory staff.









For zenith angles greater than this, the relative gain is given by


Gain = exp [(-.00521) (ZA 10)2], III-13


where ZA is the zenith angle in degrees.

The noise tubes used for the 21 cm observations had to be cali-

brated, as their noise temperatures were not accurately known. This was

accomplished by making drift scans of continuum sources, as was done on

the 91 meter telescope. Sources were taken from Bridle et al. (1972).

Owing to pointing errors, a single drift scan was not sufficient for an

accurate calibration. Usually drifts were taken at the source declina-

tion and at declinations 1' of arc to the north and to the south. The

noise tubes were fired just before and after the source drifted through

the beam. The drift scans were recorded on a strip chart. With three

such drifts, the peak flux and the error in pointing could be obtained.

Using 1400 MHz fluxes from Bridle et al. (1972), the values of the

noise tubes were found in flux units. These values were then converted

to temperature units by multiplying by 8.5 K/Jy at the 1410 MHz

setting, and by 8.1 K/Jy at the 1390 MHz setting.

The observations taken in 1977 were partially reduced at Arecibo

Observatory using several programs written by Nathan Krumm. The

individual scans were averaged together and the spectra were punched

out on computer cards for further reduction at the University of

Florida. For the observations taken in 1978, the data were brought

back on magnetic tape and all the reductions were made at the University

of Florida. One program written by Martha Haynes and Steven Peterson

was used to convert the Arecibo data tape into an IBM readable tape.

Several other reduction programs, incorporating elements written by






35


Nathan Krumm and Martha Haynes, were developed by the author. The

spectra were calibrated for the variation of sensitivity due to zenith

angle, according to equation 111-13, and for the variation of sensi-

tivity with frequency, according to Figures 111-2 and III-3. The

spectra were plotted on a Gould electrostatic plotter. The final RMS

uncertainties were typically 1 -5 mJy per channel. The calculation

of various parameters from these spectra is given in the next chapter.















CHAPTER IV
THE DATA


Optical Parameters


Most of the optical parameters of the 99 compact galaxies observed

for this study are given in Table IV-1. The columns are numbered and

are explained below by column number. Following Table IV-1 is a list

of notes on the individual objects, giving alternate names and a brief

description.


Column 1: Source name.


Columns 2 and 3: Right Ascension (RA) and Declination (Dec) of the

source in 1950 coordinates. Positions have been measured by the

author on the Palomar Sky Survey (hereafter abbreviated PSS) plates

using an overlay program and a least squares fitting program. Posi-

tions are believed accurate to 2-3 arc seconds.


Column 4: V heliocentric velocity. The Doppler shifted velocities

with respect to the sun are taken from de Vaucouleurs et al. (1976)

unless otherwise noted (S implies Sargent, 1970b). De Vaucouleurs

et al. (1976) use weighted means of published velocities, most of which

are optically determined.


Column 5: The major axis (a") and the minor axis (b") in seconds of

arc. These are given in the system of de Vaucouleurs et al. (1976).








They are the approximate dimensions to a limiting surface brightness

level of the 25th magnitude per square arc second. The author has

measured the dimensions of 150 compact and noncompact active galaxies

on the blue plates-of the PSS using a low power microscope. For 138

of these objects whose dimensions are given in de Vaucouleurs et al.

(1976), the following regression formulae are found.

50 ......
For eSS 50 : e25 = 1.4 eSS 10 15 ,

IV-1

and for SS 2 50 : 25 = 1.01 pSS + 9 16 ,


where ePSS is the major or minor axis in arc seconds as measured on the

PSS, and 625 is the major or minor axis in arc seconds in the system of

de Vaucouleurs et al. (1976). Dimensions in column 5 identified with

a G are reduced from their PSS dimensions using equation IV-1. The

others are taken from de Vaucouleurs et al. (1976).


Column 6: Axial ratio, b/a. These are not strictly the ratio of the

optical dimensions in column 5 because these dimensions often include

irregularities such as double cores, filaments or jets. Since the

axial ratio is used to calculate the inclination of the system, such

irregularities are ignored in estimating the axial ratio. For multiple

or very peculiar systems, axial ratios are not given. These estimated

axial ratios cannot be considered extremely accurate due to the gen-

erally small dimensions and often irregular morphology of these sys- s.


Column 7: Inclination i, in degrees. The inclination is used to

estimate the total masses of these systems, assuming they are rot,- :ng

disk systems. It is not certain that blue compact galaxies are








rotating disk systems. O'Connell and Kraft (1972) and O'Connell (1979)

have shown that some of the blue nondwarf systems appear to rotate with

rotation curves similar to late type disk galaxies. Gottesman and

Weliachew (1972) found inconclusive evidence of rotation in one dwarf

compact galaxy. For compact galaxies, the observed axial ratios are

higher, on the average, than is seen in late type spiral galaxies.

This is an indication that, if these are disk systems, their intrinsic

axial ratios are greater than those of spiral galaxies. This problem

will be discussed more fully in Chapter V. For now, we will calculate

an inclination assuming that these are disk systems. For a disk of

intrinsic axial ratio q, and apparent axial ratio b/a, the inclination i

is given by

2 (b/a)2 q2
cos2 i = (b/a)2 q2 IV-2
1 q2

For normal spiral galaxies, q is believed to be between 0.2 and 0.25.

Very few spiral galaxies are found with b/a > 0.2. Only one galaxy in

our sample has b/a > 0.3. For compact galaxies then, we use q = 0.3

in equation IV-2 to compute the inclinations. This is similar to the

value of 0.33 used by Chamaraux (1977) for blue compacts. As in

column 6, inclinations are not given for multiple or very peculiar

systems.


Column 8: BT, blue magnitude, in the system of de Vaucouleurs et al.

(1976). Slightly less than half of our sources have BT magnitudes

listed in de Vaucouleurs et al. (1976). These are listed in column 8

with no further designation. Those marked with an H are taken from

Huchra (1977) and those marked with a P are from Huchra (1979). The








remainder, marked with a Z, are reduced from the photographic magni-

tudes given in Zwicky's Catalogue of Galaxies and of Clusters of

Galaxies (Zwicky et al., 1961, 1965; Zwicky and Herzog, 1963, 1966,

1968; Zwicky and Kowal, 1968) to B magnitudes. De Vaucouleurs and

Corwin (1979) have determined regression formulae to perform these

conversions and have personally calculated these magnitudes and com-

municated them to the author in advance of publication of the regression

techniques.


Column 9: BT, the total blue magnitude, in the system of de Vaucouleurs

et al. (1976). These are calculated from the BT magnitudes of column 8,

correcting them to the face-on, zero extinction, zero redshift values,

according to the precepts given in de Vaucouleurs et al. (1976).


Column 10: MBn, the absolute blue magnitude. These are calculated
T
using the BT magnitudes of column 9 and the distances given in either

Table IV-2 or Table IV-3.


Columns 11 and 12: (B-V)T and (U-B)T, color indexes, in the system of

de Vaucouleurs et al. (1976). These color indexes are corrected to

the face-on zero extinction, zero redshift values, according to the

precepts given in de Vaucouleurs et al. (1976). The symbol in column

12 gives the source of the color information. Colors from de Vaucouleurs

et al. (1976) are given with no symbol; H implies Huchra (1977); K

implies Khachikian and Weedman (1974); S implies Sargent (1970b); D

implies Du Puy (1970); and I implies Hiltner and Iriarte (1958).










Table IV-1. Optical Parameters


RA Dec Ve a" b" b/a i BT BT Mgo (B-V)T (U-B)T
RA Te T BTBT)


(2) (3)


M335
HAROU 1 4
I i I Z',:1 2
HAR( ) 16


1 I1 ZW33

VZL;;31
111I Z'42
1 I I ZW-t3.
HAR)20
VZN372
IIZ~ I1
11ZY 23
i lZw2S
I 1Z',33
11 Z452'
II ZW3b


11.6
VI IZW153
V11Zib01
H AR U
4385
1022

ZW4055
M 1 05
I Z w 11
.1 4 J 2
1 z I ? 1
H A -22
HA Fr:J23
11 /,4 4
HIAIA)2
M1l48
HARC3I
HAR1025
HA104


h m
0 3 4<
0 43 Ii
0 45 1;
0 45
0 57
1 13 1'
1 41 1.
1 41 4
1 54 5:
2 8 5(
2 11
3 25 5
4 10 4
4 35 5.
4 47
4 5)
5 a 1I
b 1 4 2
5 53
6( 0 2
5 45 4
7 22 1
7 23 3
7 31 3
3 0 2
8 4 2
8 32 2
a 55 4
9 1 4+
9 30 3
9 32 2
9 43 1
9 47
10 3 2
10 12 2
i0 29 2
10 32 3
10 42 1
13 4b
11 2 1


(4) (5) (6) (7) (8)


s o 1
5.1 19 53
6.5 -15 52
7.1 22 1
4.0 -12 59
3.6 31 33
9.5 32 49
J.9 16 43
7.9 16 51
3.. 27 37
0.5 13 4 1
3.8 .3 52
7.2 -17 35
).2 29 1
3.0 11 8
7.2 3 14
4.2 3 30
38. -2 44
4.6 0 52
4.Y 3 23
. 7 7 49
3.4 74 29
8.7 72 40
6.0 72; 13
9.1 3 21S1
7.3 25 14
1.0 39 ')
3.2 30 42
8.3 6 31
3.1 /1 J
0.0 5b: 27
2.0 33 37
3.1 45 159
7.H ;2:3 14
6.5 29 11
9.4 21 21
3.0 54 39
7.2 44 34
7.0 5b 13
0.1 26 19
5.4 29 24


7445
915554
59 541

6415
4477
4793
7964

8041
7743
3438a
1834
5342
4393
54 03
5i649
2B13 67
26607
72 20

5391!
5600

3470

804 1
7075
7147
3501
3562

7289
4965

1402
6166
b 1 6 (
1 4 4 6
7147
865

636


1 IG

1 7G
52G
52G
41
30
205
25G
5G66
48
4 .
90
22C
71
1 3.
47
16G
7 7G
24
76
59)
76
01l
4uG
4 0
37
47
2't+G
4 4 G
44G
o5
ol
49
35
900
6 7
95
19
19


'3 -33
0. 7 5
0 .90

0.751


0.97
0 0

0.50

0.31
.).77



1 .00
0. 33
1, 00
1 .0o
0.74
0 :,
0 50
(). 9.0
0.70
0 .* 1o
0. 73

0 6.3

0. i,
1 .00
0. 84
0.7
0 7 30
0. 76
0.60
0 /0
1.00
0. ;J1


(9) (10) (11) (12)


1'I 10
13.70
a15. 31z
13.90
1 7 /
15. 12
14 .89/

I 1. 3 1 /
I4.318Z
1 4. 31Z
14. 90


14 55Z
15.5
14.20Z

15.64P
17.60
14. 507
14.09Z
13. t69Z
12. 2

14.667
15.02Z
14 .73L
I s 027-
14 7 4 Z
16. ')
16.34H

14 .8 Z
15.2 0
14.00
15.94Z
1.3 0 0

13.20
15. 70
15.80H1


13.7'5
1', .* >

1 3 .415
14.2 4
14 .09
1 4 6

15.00
14 .51
14.0 5
14.44


13.33
14.94

13. t~h
14 22

1-3 *. .3
13.40
13.1 1
1 2. 3
14.57
14.33
14/.50
14.31
15 1
15.00

14.59
14 4
14.33
15.75
13.15

12.81
15 .52
15.51


3.23

0. 1 3

0.4
0.61
0.28


-3. 76
-3 .23
-3 .i '

-0 .66
-0.02
-3 3.)H


0. 3 -3. 0 S
0.15 -0. 3



0.26 -,.53
0.203 -0.63)


-21.35
-17.15
- i 9.05 5
- 2.1 .20
-1 9.72
-19.4 4
-20.o5

-20.20
-20 .(63
-19.31
-1 .45


-21 .32
-20.32
-19.30

-15.'55
-10.92
-20 5
1 9 7 7
-20.31
-21. 109
-20.02
--20 u-2
-20 .53
-20.44
-19.01
-1 /. 57
-14. 36

-19 54
-16.40
-16. 95
-110. O?
-1 3
0.0
-1. 7. 3 '
-1I9.51
-14.01


-0. 34.
-0 .0
-0.07<
- C 0 C)<
-0. '4
-0 43-1
-0 .* 9


0. 3 -0.2?
0.13 -0.7?-


4.20
03. 2 6
0. ?d




0. .11
0.31
0.31


-.3.35
-0.11
-0.223
-0.50
- 0). *7 0

-3.3)1
-0 .7 H


Source


(1)


0. 17
0.23

0.17
0.3'
0 :'









Table IV-1. Continued.


RA Dec


V a" b" b/a


i B BT B (B-V) (U-
T T MB0 (B-V)T (U-B)T
T


(2) (3) (4) (5) (6) (7) (8)


Source


(1)

I W26 f
m 1 6 59
HAR;]27
M 11 )
I I ZA"57
HARiGd

H Rn)29
4121 3
1.2R3J2
MR23J 1
.-,,r03j




II!Zw68
11 Z.67
r57
M23'5
4M241
L Z ,53
I I 1.H56
HAR C38
1270
Me 75
H A R'J 39
"iA 1 '142

HAi R J4
1 1.470
I wI 2 7 1



SZW 11 5
I [Z 1 1

1ZW123
VIl Z-631
4297


h 22m 4b.S
23 52.9
37 40.0
62 43.2
2 7 0.0
13 17.4
1 10 36.95
S20 50.0
1 23 50.5
30 10.7
241 32.0
2 42 11.7
12 ? J 37.
1 44 40.0
S44 42.1
2 't 16.9
2 55 37.1
2 56 10.0
2 3b 12.0
2 37 39.0
3 3 58.0
3 11 38.9
j Id 17.1
3 33 17.0
3 30 41.9
3 2:3 25.'
: 56 5.4
4 2 56. 5
4 33 5b./
4 41 14.1
14 4-3 55'.2
4 4)' 13.1
4 -32 47.3
4 53 23.4
5 1 50.0
5 31 19 .
5 34 .3.
5 3 4 B.0
S'33 50.5
. 3 1.2


6147
1318
2970
7202
6620
131 8
1502
7034
237

4 0 0S
95t,
7002
322

4239
5700
7550
7042
7445
7743
5043
0004
840
2697
S041

4 470


1147
122(
2530
54P5
48605
669
5520
663

4700


)97
46
51

93
72
JO1
1,4G
5.?
'3
87
47
:3 -
37
19G
lyo
1 6
79
71
3G G
30

70
104
7f,
Z l
79
5'G.
44
2201
261
,0
71

72
2 3
L03
U3
59
18
135
57


1.00
0. 5
0.64
O. y O


0. 75

0.90

0 60
0.60
0.9 )
o. t2
0.80
0.9 5
0.55
0.00
0. f
0 . 0

0.03
0. 50



0.30
0 73
0.32

0.3'
0.20
0 .* c


0.90


0 70
0.30

0.7;)


14.33Z
14.531
14.90


1 4 )J 3 2


14 .130





15.40
14. -30
14 0 0
1 .37
15.37
1 u. 1 1 l


15.20
134.0 Z

15.23Z
15.711
15. 79Z

14.75
14. 30
14.91Z
1 4 2 Z

15 1- 4Z
15.14Z

15. 73Z
13.35
13.45


13 .95
14. 13
14 5:
13.oO
13.21
14.49

14.24
14.147
13. /0
14.30
1 4. 86
1 .67

13.1.5
14.34
13,609
15.03
15.07
1 3 .17

14 10
1 4 7
13.1

14. 49
1i .74
1 ~ 0


13.34
14 4 7
1 .91
1 19
1'. 53
14.13
1 5 .tv
12.9'
12 98


0.54 -0.33H
0.40 -0.271
0.66 0.03


0.39 -0.'OD


0.11
0.4-,
3.27
0.21
0.4*
0.51


-0.43
-0.19
-0.330
-3.41
-3.25
-*.2 10


-17.20
-17.77
-20.37
-21.12
-1t.90
-16.90

-14.17
-20.01
-20.43
-10.10
-19.9
-14.32
-1 9 93

-14 .03

-20. 0o
-21 .33
-19.9 t
-19.93
-10.93


-20.75
-15.52
-19.29
-20.5)
-18.23
-10.19
-17.00

-17.24
-17. 44
-1 3 1
-20.4 7
-1 9 7
-1 5. 1 7
-23.23
-14.6 3
-20.D98
-21 0-


0.4' -0.29

0.77

0.33 -'.2?H
0.51 -0.12H


0.50 -0.35
0.42 -0.22K


0.33 -0.47D


0.11 '-0.85
0.3- -0. 1')



0.33 -0.14P

0.49 -0.53H

0.33 -0.46


(9) (10) (11) (12)













RA Dec


Table IV-1. Continued.


a" b" b/a i


BT BT MBo (B-V)T (U-B)T
T T BT T7


(2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)


16 22 2.6
16 34 7. t
16 47 3.0
17 39 6.3
17 49 15.3
16 33 10.4
20 20 42.4
21 0 20.0
22 7 15.2
22 12 20.5
22 13 46.9
22 14 1.4
22 23 44.7
22 3' 0.0
23 3 30.5
23 17 35.0
23 23 11.7
21 27 40.3
23 35 9.


54 16 0
52 1I 56
480 47 34
47 45 20
56 41 10
b5 14 1 7
0 30 10
36 29 47
17 24 57
13 35 30
22 41 6
16 13 13
30 4) 17
23 o 46
16 20 0
25 56 25
23 10. 48
25 15 22
29 51 10


5432
26-2
7679
5793
5300
55/3
3975

70i80

3685
74+45
6710
7130



1308
*33510
i jaa


16 19G
15 18
30 30
12 135
27 37
17 43G
48 55J
45 5?G
14 175
61 61
43 2 .3
25 44
40 46
39 43
47 69
1a 2 1
97 102
1 255
25 35


1.00
0 a 40
0.93
1.00

0.837


1.00
0.63

0.90
0 630
0.753




O. ,3
0. 7'j
0.90

0.90


16.06Z 15.70 -17.20
15.31Z 15.03 -20.10


0.33 -0.30H


15.loZ 14 .82 1 .52

14.20 13.46 -20.2d 0.64


15.712 15.25 -19.05
13.70 13.23 -21.92


15.20




13.00
15.217
b1. 75Z


14.52
13.52
13.94
13.45
14.89
12.6o-
14 3
15.37


-20.44
-21.39
-21.02
-1a.93
-2).20
-20.75
-19.68
-16.24


0.90 0.45

0.4B -0.16H


0.39 -0.38H

0.25 -0.22

0.57 -0.12H


Source


IZW147
1Zo159
I Zw 166

I ZW 191
1 '4207
I I Zw 2
IVZ n 7
1 1L 153
IIZ a 72
1VZ 93.
:1333
Z:2220
11Z.'1850
11 2 1 3 3
IV! 1 4

1 VZ 149
1 vZ o 1 5 3
2I L A 3 -
Z W2.33





43


Notes to Table IV-1


M335 = A0003+19. Seyfert galaxy. Almost stellar; slight trace of
nebulosity; b/a rather uncertain.

Harol4 = MGC-3-3-3. Has large bright core not centered in nebulosity;
bright knot in ENE part of nebulosity (NGC244).

IIIZwl2 = M347 = IC1586. Nearly stellar core; hint of nebulosity.

Harol5 = M960 = MGC-2-3-19 = A0046-12. Large bright core with
nebulosity.

M352 = A0057+31. Seyfert. Small, nearly circular core; slight amount
of nebulosity; weak jet to W.

M1 = NGC449 = MGC5-4-9. Seyfert. Small, bright elliptical core with
a slight amount of nebulosity. Correctly identified in de Vaucouleurs,
et al. (1976), but incorrectly in Nilson (1973).

IIIZw33 = M360 = MGC3-5-13 = A0141+16. Bright elongated core; two
bright knots N of core imbedded in nebulosity; somewhat irregular,
nonuniform nebulosity. May be a distorted Scd (Huchra, 1977).

IIIZw35. Double compact system; in contact. Southern compact has
circular core; little nebulosity. Northern compact has bright
elliptical core; little nebulosity.

VZwl55 = M364 = A0154+27. Featureless; small circular core; very little
nebulosity.

IIIZw42 = M366 = A0208+13. Bright, nearly stellar, slightly elongated
core; clumpy irregular nebulosity not centered on core.

IIIZw43 = M589 = MGC1-6-56 = UGC1716 = A0211+03. Bright, nearly
stellar core; faint nebulosity.

Haro20 = MGC-3-9-45 = A0325-17. Small, uniform, elliptical core;
faint nebulosity.

VZw372 = MGC5-10-12 = A0410+29 = UGC2989(?). Bright, stellar nucleus;
faint, grainy and extensive nebulosity.

IIZwl8 = A0435+11. Round, bright, stellar core; some nebulosity; jet
to SE. May have faint outer spiral arms (Sargent, 1970b).

IIZw23 = M1087(?) = A0447+03 = UGC3179. Elliptical core with faint
jets; faint nebulosity.

IIZw28 = A0459+03. Round; slightly irregular; clumpy; little
nebulosity. Has a ring (Sargent, 1970b).









IIZw33 = M1094 = MGCO-14-10 = A0508-02. Elliptical, patchy, compact
core; perhaps composed of three clumps; irregular, patchy nebulosity.

IIZw35. Two stellar knots; very faint nebulosity.

IIZw40 = A0553+03. Approximately stellar core; very faint, grainy
nebulosity; two faint jets to SE; b/a very uncertain.

IIZw42 = MGC1-16-0 = A0600+07= UGC3393. Bright stellar core; little
nebulosity.

M6 = IC450 = MGC12-7-18 = UGC3547. Seyfert. Small, bright, elliptical
core; faint, fuzzy, irregular nebulosity.

VIIZwl53 = M7 = MGC12-7-38 = A0722+72 = UGC3838. Irregular, "tank"
shaped, compact core; small amount of nebulosity; b/a very uncertain.
U shaped core of knots (Huchra, 1977).

VIIZwl56 = M8 = IC2184 = MGC12-7-41 = UGC3852. Triangular shaped
clumpy core; triangular shaped nebulosity; b/a uncertain. Ring of
HII regions (Huchra, 1977).

Harol = NGC2415 = MGC6-17-21 = UGC3930. Slightly ellipsoidal,
irregular core; irregular nebulosity.

M385 = A0800+25. Bright, approximately stellar core; small amount of
nebulosity.

M622 = A0804+39 = UGC4229. Small, bright, almost circular core;
fairly dense, featureless nebulosity.

M390 = MGC5-20-28 = A0832+30. Bright, ellipsoidal, compact core; faint
nebulosity.

Zw0855 = MGC1-23-13 = A0855+06 = UGC4703. Unnamed object in Zwicky
(1971). Two tiny clumps, -87" apart; connected by a faint, thin
bridge; very strange object.

M105 = A0915+17. Small bright core; trace of nebulosity.

IZwl8 = M116 = A0930 A and B. Pair of interconnected compacts;
figure 8 shaped core with some nebulosity.

M402 = A0932+30. Elliptical core; very little nebulosity.

IZw21 = MGC8-18-30 = A0943+46 = UGC5225. Circular, nearly stellar core;
very faint nebulosity.

Haro22 = MGC5-23-40 = A0947+28. Bright, slightly elongated core;
small amount of nebulosity.

Haro23 = MGC5-24-11 = A1003+29. Small, ellipsoidal core, slight amount
of nebulosity.





45


IIZw44 = A1012+21. Small, semistellar core; main object is only -20"
in diameter; several possible nebulous objects within -45" are used
for the total dimensions.

Haro2 = M33= Arp233 = MGC9-17-70 = UGC5720 = A1029+54. Featureless,
ellipsoidal core and nebulosity.

M148 = A1032+44 = UGC5747. Bright irregular core; jets to SE and NW;
slight amount of nebulosity.

Haro3 = M35 = NGC3353 = MGC9-18-22 = UGC5860. Small bright core;
faint irregular nebulosity.

Haro25 = M727 = MGC4-26-9 = A1046+26. Stellar core; very slight trace
of nebulosity. DuPuy (1970) misidentified Haro25 with a red object.
Hiltner and Iriarte (1958) looked at the correct object.

Haro4 = M36 = MGC5-26-46 = A1102+29. Small, slightly elongated core;
small amount of nebulosity.

IZw26 = M40 = Arpl51 = VV144 = MGC9-19-73 = A1122+54. Seyfert. Thin
filament with two knots in it; very irregular object.

M169 = IC691 = MGC10-16-139 = UGC6447. Slightly elongated core; some
nebulosity; faint jet to S.

Haro27 = MGC5-28-10 = A1137+28 = UGC6637. Bright elongated core with
some nebulosity.

M198 = MGC8-22-73 = A1206+47. Bright, nearly stellar core; elliptical,
slightly irregular nebulosity.

IIZw57 = MGC3-31-49 = A1207+17. Bright elongated core; clumpy
nebulosity.

Haro28 = NGC4218 = MGC8-22-88 = UGC7283. Large, bright, slightly
irregular, elliptical core; small amount of nebulosity.

Haro8 = M49 = MGC1-31-50 = A1216+04 = UGC7354. Stellar core; slight
amount of nebulosity, mostly to SE.

M50 = A1220+02. Seyfert. Bright, nearly stellar core; slight trace
of nebulosity.

Haro29 = IZw36 = M209 = MGC8-23-35 = A1223+48. Strange appearance;
fan shaped nebulosity; bright compact core at western tip and compact
core in center; small bright object or jet to N; b/a very uncertain.

M215 = MGC8-23-52 = A1230+46. Small, bright, slightly elongated core;
small amount of nebulosity.

Haro32 = IZw41 = M220 and M221 = MGC9-21-33 and 34 = A1241+55 A and B
= UGC7905. Close compact pair; very irregular. Southern component has





46


bright core; slight amount of nebulosity. Northern component is very
irregular; bright elongated core; extensive curving nebulous region to
N and E.

Haro33 = MGC5-30-70 = A1242+28. Bright central knot; smaller knots to
ESE and WNW, interconnected with nebulosity.

Haro34 = IC3730. Bright, approximately stellar core; irregular
nebulosity; curving filament to N, like a single spiral arm.

Haro36 = MGC9-21-47 = UGC7950 = A1244+51. Bright central core; weak,
somewhat irregular nebulosity.

Haro35. Bright uniform core, shaped like a "combat" hat; virtually no
nebulosity.

Haro37 = M444 = MGC6-28-32 = A1246+34. Approximately stellar appear-
ance; faint haze; hint of faint jets.

IIIZw68 = MGC5-31-38 = A1255+27B = UGC8080. Bright, nearly circular
core; uniform elliptical nebulosity.

IIZw67 = NGC4853 = MGC5-31-48 = UGC8092. Slightly elongated core;
some nebulosity.

M57 = A1256+27B. Small, elongated, slightly irregular core; very
little nebulosity; faint jet to S.

M235 = MGC6-29-10 = A1257+33. Small elongated core, pointed at one
end; little nebulosity.

M241 = A1303+33. Bright elongated core; very little nebulosity.

IZw53 = A1311+35. Nearly stellar core; slightly elongated; virtually
no nebulosity.

IZw56 = IC883 = Arp193 = MGC6-29-0 = UGC8387. Bright elongated core;
some nebulosity; two jets to SE and to SW.

Haro38 = MGC5-32-41 = A1333+29 = UGC8578. Bright elongated core with
some nebulosity.

M270 = NGC5283 = MGC11-17-7 = UGC8672(?). Seyfert. Bright, nearly
stellar core; faint nebulosity.

M275 = MGC5-33-2 = A1346+31. Elongated, somewhat irregular core;
nebulosity with wispy features, perhaps hint of spiral structure.

Haro39. Bright, very elongated, cigar shaped core; very little
nebulosity.

Haro42 = M685 = MGC5-34-61 = A1428+27. Elongated core; faint
nebulosity; filaments to east.








Haro43 = MGC5-34-80. Very elongated bright core; very little nebulosity.

Haro44 = MGC5-35-8. Irregular, tear shaped core; small amount of
nebulosity; possibly a jet to N.

IIZw70 = M829 = MGC6-33-2 = VV324B = A1448+35 = UGC9560. Pair with
IIZw71, -250" apart; slightly elongated, nearly stellar core; nebulosity
mostly along major axis, like two streamers.

IIZw71 = MGC6-33-4 = VV324A = A1449+35 = UGC9562. Elongated with
central bulge; faint halo; perhaps an edge on galaxy.

IZw97 = A1452+42. Elongated core; very little nebulosity.

IZw98 = NGC5787 = MGC7-31-8 = UGC9599. Bright elongated core; exten-
sive elliptical nebulosity; axes of the core and of the nebulosity
differ by -45.

IZwlOl = IC1090 = MGC7-31-25. Small elongated core; slight amount of
nebulosity.

IZw115 = MGC8-28-38 = A1531+46 = UGC9893. Bright, crescent shaped
core embedded in elongated clumpy nebulosity.

IZwll7 = MGC7-32-30 and 31 = A1534+38 = UGC9922. Elongated core;
compact knot at northern tip; slight nebulosity; b/a rather uncertain.

IZw123 = M487 = A1535+55. Round, almost stellar core; slight amount
of nebulosity.

VIIZw631 = Seyfert's sextet = NGC6027 and companions = VV115 =
MGC4-38-5,6,7,8,9 and 10 = UGC10116. Compact group of six compact
galaxies; in a tight triangular cluster.

M297 = NGC6052 = Arp209 = VV86 = MGC4-38-22 = UGC10182. Irregular
object; two bright knots side by side; "hat" shaped appearance;
slight amount of nebulosity; b/a very uncertain.

IZw147 = A1622+54. Approximately stellar with fuzzy edges.

IZwl59 = A1634+52. Small, nearly stellar core; little nebulosity.

IZw166 = M499 = A1647+48A. Nearly stellar core; small amount of
nebulosity.

IZwl91 = A1739+47. Approximately stellar, fuzzy edges.

IZw199 = MGC9-29-39 and 40 = A1749+56 A and B. Small, interconnected
pair of compacts; -20" apart; small amount of nebulosity.

IZw207 = A1830+55. Seems very blue; arc shaped clumpy object;
"boomerang" shaped; very irregular.









IIZw82 = IC1317 = MGCO-52-4 = UGC11546. Elongated core; faint
nebulosity.

IVZw67 = MGC6-46-0(?) = UGC11668. Elongated core, perhaps double or
with a star at one tip; little nebulosity; possibly several faint jets.

IIZw168 = A2207+17. Nearly stellar, just slightly oblate core; slight
trace of nebulosity.

IIZw172 = NGC7236 and 7237 = Arpl69 = MGC2-56-23 and 24 = UGC11958.
Three small round compacts in a row; surrounded by nebulosity.

IVZw93 = A2213+22. Triangular shaped nucleus; perhaps made up of
three stellar or semistellar knots; surrounded by faint nebulosity.

M303 = NGC7244 = MGC3-56-21. Elongated core; clumpy nebulosity.

Zw2220 = A2220+30 A and B = UGC12011. Unnamed object in Zwicky (1971).
Interconnected pair of compacts; 16" apart. Western component has a
small elongated core; small amount of nebulosity. Eastern component
has a smaller elongated core; jet and small amount of nebulosity.

IIZw185 = IC5243 = MGC4-53-11 = UGC12153. Irregular; mostly dense
core; little nebulosity; curving tail, or perhaps single spiral arm;
b/a rather uncertain.

M314 = NGC7468 = MGC3-58-0 = UGC12329. Oblate, tear shaped core; some
nebulosity; faint thin jets extending from each side of major axis.

IVZwl42 = M322 = A2317+25. Bright elongated core; fuzzy edges.

IVZwl49 = NGC7673 = M325 = MGC4-55-14 = UGC12607. Large bright core,
pointed at southern end; small amount of nebulosity; short jet to N.

IVZwl53 = A2327+25. Double core; small amount of nebulosity.

Zw2335 = M328 = A2335+29. Bright, tear shaped core; small amount of
nebulosity.



Abbreviations used:

MGC: designation in the Morphological General Catalogue (Vorontsou-
Velyaminov et al., 1962).

UGC: designation in the Uppsala General Catalogue of Galaxies
(Nilson, 1973).

A: designation for sources without NGC or IC numbers in de Vaucouleurs
et al. (1976).

M: Markarian number.





49


Arp: designation in the Atlas of Peculiar Galaxies (Arp, 1966).

VV: designation in the Atlas and Catalogue of Interacting Galaxies
(Vorontsov-Velyaminov, 1958).

N: north

S: south

E: east

W: west









21 cm Global Parameters


Table IV-2 lists the global properties of the 72 systems which

were detected in HI. The columns are explained by column number and

line number.


Column 1, line 1: Source name.


Column 1, line 2: Designation for the telescope used. Thus N implies

the 91 meter NRAO telescope; A implies the 305 meter Arecibo Ob-

servatory telescope; AN implies both telescopes.


Column 2, line 1: V21, the heliocentric velocity, in km/sec, meas-

ured from the 21 cm spectrum. The velocity centroid is taken as the

velocity of the system. This is defined by

max
E V(n) -S(n)
n = min
21 max IV
E S(n)
n = min

where V(n) is the velocity of the center of the nth channel and S(n)

is the flux of the nth channel. The summations are carried out from

the minimum channel to the maximum channels containing the spectral

line.


Column 2, line 2: 6V21, the uncertainty in V21, in km/sec. This is

found by differentiating equation IV-3, which gives

max
(F V(n)) N V21
6V = n-mn .AS IV-4
21 max rms
E S(n)
n = min








where N is the number of channels containing the spectral line and

ASrms is the RMS uncertainty of the baseline, calculated from the

channels on each side of the spectral line. Values of 6V21 are rounded

up to the nearest integer. Equation IV-4 is strictly correct only if

the uncertainties are due entirely to random noise. But in practice,

errors in the pointing of the telescope and in fitting a baseline will

further limit the accuracy of V21.


Column 3, line 1: D, distance to the source in Mpc. The heliocentric

velocities, V21, are corrected to the velocity seen from the center of

the galaxy, VGC, by


VGC = V21 + 250 *sin L* cos B, IV-5


where L and B are the galactic longitude and latitude of the source.

The distance D is then found by dividing VGC by a Hubble constant of

75 km/sec/Mpc.


Column 3, line 2: A, the linear size of the major axis converted to

units of kpc. These are calculated using the distances from column 3,

line 1 and the major axis, a", given in Table IV-1.


Column 4: AV, the width of the 21 cm profile in km/sec. This velocity

width is measured at the points equal to 20% of the average flux after

boxcar smoothing over three channels, or at the 2a level after boxcar

smoothing over three channels if the average flux is too weak. The

RMS uncertainty, a, is calculated after smoothing over three channels.

The velocity widths listed are corrected for the boxcar smoothing.








Column 5, line 1: BCF, beam correction factor, applied to the HI

masses to correct for the finite width of the beam.

A radio telescope will underestimate the total flux from a radio

source, unless it is a point source centered in the beamwidth. For

extended sources, some way of correcting for this underestimation is

necessary. The 91 meter telescope has a 21 cm beamwidth of approxi-

mately 10:8 of arc at half power points. The shape of the beam is

approximately Gaussian, except near the first null. All of the sources

observed with the 91 meter telescope were much smaller than the beam-

width and so this deviation from Gaussian shape need not be taken into

account. The flux detected will be the convolution of the source flux

distribution and the beam pattern. If the HI surface density of the

source is distributed as an elliptical Gaussian with major and minor

axes half power dimensions of A and B arc minutes, then the ratio of

the total flux to the detected flux (from Fisher and Tully, 1975) will

be


BCF = (1 + (A/10.8)2)/2 (1 + (B/10.8)2) 1/2 IV-6


Unfortunately, the values of A and B are not known. For sim-

plicity, we assume that they are some multiple of the optical dimen-

sions, a" and b". Converting to minutes of arc, we have A = F -a"/60

and B = F b"/60, where F is this unknown multiple. For spiral galaxies,

the HI distribution is typically found to be somewhat greater than

the optical distribution. Thus we expect F to be greater than unity.

The 305 meter telescope has a half power beamwidth of approxi-

mately 3.3 of arc at 21 cm. This is not small compared to the dimen-

sions of many of our sources. Additionally, the departure of the beam









pattern from a Gaussian is significant enough that an equation of the

form of equation IV-6 is not sufficiently accurate. The sidelobes of

the beam may also contribute to the detected flux and need to be con-

sidered as well. A numerical model of the Arecibo beam, including its

first two sidelobes, was constructed. The source size and distribution

is considered as in the previous case. The beam pattern and the source

shape are then convolved numerically to determine a beam correction

factor. Due to the pointing errors of the 305 meter telescope, this

convolution also includes the effects of a 30" of arc pointing error.

Again the factor F is unknown.

However, several sources were observed successfully with both

telescopes. In the analysis of the HI masses of these sources, a value

of F was found that led to the same corrected HI masses on both tele-

scopes. The values of F found for 13 sources ranged from 1.0 up to 5.2

with an average of 2.4393. Thus for the other sources, a value for

F of 2.4 is assumed, with a few exceptions, all of those being multiple

galaxies.


Column 5, line 2: F, the ratio of optical to HI dimension assumed

for the calculation of the beam correction factor of column 5, line 1.


Column 6, line 1: MHI, the HI mass expressed in solar mass units and

corrected for the beamwidth effect. The uncorrected HI mass is found

by converting equation III-6 to a summation of the form


1 5 2 max
MHI= 2.356 x 10 D E S(v) Av, IV-7
n = min


where the summation is carried out from the minimum to the maximum








channel numbers containing the spectral line, and Av is the channel

spacing in km/sec. The corrected HI mass is then found by multiplying

by BCF of column 5.


Column 6, line 2: AMHI, the uncertainty in the HI mass. This is found

by differentiating the equation for corrected HI mass. One gets


AMHI = 2.356 x 105 D2 VN Av ASrms BCF. IV-8



Column 7: LBo, total blue luminosity in solar units. These are
T
calculated from the absolute magnitudes, BT given in Table IV-7 and

assuming the absolute B magnitude of the sun to be +5.41 (Allen, 1964).


Column 8: Mi, indicative total mass, in solar units. We use the

equation used by Shostak (1978) for late type galaxies:


M. = 2.45 x 104 A (AV/sin i)2 IV-9


Total masses will be discussed in more detail in Chapter V.


Column 9: MHI/L, column 6 divided by column 7.


Column 10: M./L, column 8 divided by column 7.


Column 11: MHI/Mi, column 6 divided by column 8.


Table IV-3 lists a limited amount of data for 27 systems which

were not detected in HI. The columns are explained by number.


Column 1: Source name.










Table IV-2. Global Parameters of Detected Sources.


Source V2 D AV BCF MHI L M MHI/L M /L MH/M
Telescope 21 HI BT HIi
6V21 A F AMHI
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

lHAR- 14 940 13.2 04 1.0H03 1.805C 0!9 1.057LZ 09 4.184!L 09 0.171 3. '5 0.043
N 3 5.1 2.400 1.6 7 07
I11ZW12 5830 U0.0 200 1.0992 3.Cr00 09 1.070, 10 9 .54F 09 V .280 0. .892 0.314
A 11 6.6 2. .00 2.03E 0"I
HARalS 6407 86.0 220 1.0239 5..l ;I 0 4.3)5F 10 J.t'- 4E 10 0.126 0.331 0.152
1 10 21.7 2.400 1.79E 09
'1352 4442 62.0 242 1.3211 3.142 '- 09 1 .12'9 10 3. 1 87 10 0.276 2.023 0.)93
A 12 15.6 2. 400 2. 1L 0"
M1 4039 67.0 259 1 .2b4 3.339C 09 B.743' 09 4.551E 10 0.'38 5-.202 3.073
A 12 13.3 2.4/)0 2.347 0)4
1:1 ZW3 83057 139.0 245 1.1 ,40 C.2)'F' 0' 2.606E 10 6.473F 10 0.315 2.434 0.132
A 3 15. 2.400 3.31F 0 oL
!lIZW35 0222 111.0 109 1.0941 3.309F 09'
A 6 10.8 2.40J 2.j4E 06
1!1Z'42 7 49' 107.0 89 1.4170 1.IGOti' 099 2.614E 10 1.424' 10 0.051 0.51'5 0.1 J
A 6 29.0 2.4'00 2. :i r 0
111 -43 3444 47.0 186 1.30,7 1.014 l 09 7.727 09 3.030E 10 0.131 3.9:3 0.333
AN 16 10.9 1.000 3.00 0'I
HAR020 1553 24.0 141 1.01?4 4..202- 03 1.402E 09 3.7209 09 0.300 2.650 0.113
N 1b 5.6 2. 400 7. 1 5t 07
VZW372 5454 73.0 101 1.3749 1.140E 0 9. 655 09- 0.1 1
4 4 31.8 2.400 2.35; O'H
1IZ,.18 44*i5 59.0 280 1.1107 1.511F 0,) 2.701_ 10 0.0O
A 5 6.3 2.400 '). 2ii 07
IIZA23 8330 110.0 270 1.2303 1.71 t:. 1 4.934E' 10 1.7 1E 11 0.349 3.,10 0.096
A 10 37* ,.400 1.32": r,,
!L ?33 2820 36.0 153 1.0235 1.7'0C F : 7.f6926 09 8.391E 0) 3.222 1.091 0. '03
'1 2 3.2 2.400 91- 0
IIZw40 795 9.0 173 1.0722 J.45'[. 09 2.41y 0 g 6.011 09 1.430 24. 35) ?.)0
AN 2 3.3 2.4,1 2.04e 07
1IZ.42 5263 6).*0 1.1440 5.A b:? 0' 4.' '20 09 119
A 6 4f.0 2.4 '0 9. 92 0
VII 1 153 3350 43.0 232 1.0233 2.!10!1- 0) 1.177E 10 2.115E 10 0.231 1.797 0.132
N 2 12.3 2.403 j .GC 01
V11ZW156 3595 50.0 243 1.J676 5 .60; 0' 1.974E 10 5.734E 10 0.233 3.11 0.06) j
N 3 16.4 2.400 3.12 0.-
HARG1 3797 53.0 250 1 .0367 96.039- 9 4 .0j0- 10 1.O R4F 11 0. 150 2..:) 0.0'i,
AN 1 14.0 ,.03 2..1' 3










Table IV-2. Continued.


Mi MHI/L M /L MHI/Mi


Source
Telescope

(1)


A
,490
A
7 9055
A


M4 02
A
HAHU022
-\N
HAG,23
A
1 7 A44
AN


'I
HAR3
N

A
HIA R204

11 6)
N
HARG027
A
I 1 Z 1i'j 7
A


HAO 32
N
H ANOS
A
H -3]29)

HARO32
N


(11)
0 036

0. 97


V21

V21
(2)

0300
100
7640
22
3733
915
75)
2
7416

14'43
2
1372
2
6201
10
1467
4
943
2
7630
5
640
2
1233
2
1936


12
6(14
3
721
2
1515
3
281
2
493-3
3


D

A

(3)

109.0
24 .3
101 .0
1 .1
10 .





13.1
2.3

98.0

1.3
2n .0
6.5

10; .0
9 .3
5 .0
0. 7
17.0

3 .13l
3.8
24 .0
5.9)
JO )
3 ).7

3.7
19.0n
3 .b
4 .'3
1.?
57 .0
29 .*


BCF

F

(5)

1.29-49
2.4 00
1 2053
2.4 00
2.400
1 0 1
2.. 43
1.1 3 4-
2.400
1. 3133
1. 36
1 .23'.9
2.400
1.325 3

1.00 '1
2 .4C0
1 0 '39
. 400
1.113
?. 400
1.10.54
1 1 0 P,
2. 4 J0

2.4 0
1 .30O1

2. '400
2 .3 7 .'&
2? ? ,0
1 04'
2.400
2 .' 1 0
1.0337
2. 410 ,,)
1.0 707
-' 4 0 0


MHI

AMHI
(6)

1.027' 09
4 .49 2 03
6.571r O0
1.')IE 011
7. 815 Os
'r..t.r 07
E.313i 07
.3. S F
2 01 ',1 0 ')

2.5 E ,} 0
2. '31: 07
o .9554F o 0
'. 30' 0,O
1 .21 :, 10
8. 3.':1 0-
4.71.37 In
3.(67c 07
5. JO'U 01
1.2;" 07
1.7";3: ).
S 7 3,1 )E )
1 *, ,.37 ( 7




9 1. E- 0
2 ., 01-. 0 .
3. 4 0 ')
2.O ".OS 'OI


o.' 7E O
5.00 1.- 0 7
.. 32:' 0
R5.?21 0I


LB.o
BT

(7)
2.583B 10

2. 192 10

1.8453 3 09

0. 125E 07



5.2,351 Ob

U. 79 6 0'

4.924- 09O

3.21 lEc 09

1 .') 7:" 0-)

9..'-13 O'1 )



1. Y191' 01)

1. 366- 09

4.10 7C1 10

b.373,- 0,O

H. 43J'E 0O

G.7 74' 07

2. 1657 10


(8)
2.892F 10

&.710F 10




2.'568E 10

5.193E 00

1.1 7-E 0o

0.0

1.483E 10

5.0 t 1 0 -)



4.471 00

4.462E 0 )

6. 0250 0O

1 31i 3 9 -.' I 1

3.750C 09

1.0 ',7' 09

1.002E 0q


(9) (10)
0.040 1.120

0.300 3.079

0. 134

1 053

0.0 0.3

0.4 3 9.5335

0.075 1.340

2.463 0. )

0. 14I? 4.620

0.270 2.534

0. 620

0.'30 7.03

0.440 4.027

0.124 1.229

0. 95 3.260

0.253 4. 49 -

0. 1 2.347

0.73) 14. 712

.0 407


3.114

0.050

0 .056

3.0

O.032

0. 10i



0. ) 43










0.0'30
O) .01 1






1.050










Table IV-2. Continued.


Mi MHI/L M/L MHI/M


Source
Telescope

(1)

HAR033


A
H AR036
N

A



A
M53

1241

1 Z 53
A
HAR038
A
S27 5
A
HAI PC39
AN
H AR342
A
HA 0,Z43
A
H A R<44
A
I1 Zw70
A
11 ZW7 1
A
I Z 10 1
N
I Z115
IZiI17
IZN117
N


V21.

6V21
(2)

93,9
2
70 '1
2
491
2
7452
4
4280
2


7959
3
5143
100
8,4

7937
4
25' 7
3
4465
-3
1012
2
3725

1207

12501

4933

652
2
,596
4


D

A

(3)

12.5
2.
93 .
10.4
7.9
3.3
99.0
9.1
58.0
4.5
102.0
13.8
105.0
14.2
69.0
2.3
12.0
4.4
1 06.0
40 .6
35.0
9.5
G1 .0
13.0
27.0

51 .0
6.4



6.2
6 0
7.
11.0
4 '.
77.0
22 .,


BCF

F

(5)

1. 0162?
12.134
1. 137
2. 400
1.03 H44
2 i) )

2. 40')
1 '001
2. 400
1.1036
2. 'tOO
1 .16 2 ?

2. 0I '0
2. ) 02
2.400
1 .4I ;'

2.400
2.3'96?





2. 400
1.1217
2. 400

2. '400
1 3 117
2. 4 )





1.7531
2. 400
1.4031

1 .006
2. '0 0
1.0531
2. 400
1.02 70
2 400


MHI
AMHI
(6)

1. 9 ,Ji 0 1
1.03= 07
7.039L 09
1.93:C 0

1.O1E 07
2. 'JJF 09
1. ')01L 0-
i ri r0 '
7. '4E 07
2.21I. 0'i

L3.t, 3: 0o'

2.Q-)E 0O

2.4 0 0'O
3. 31LC 07

1 5"Ic o0
4.5.EC 0j
1 500" 00
9.OL 07
3. 312' 09
!. 03 0(.>'
'.923. f O,-;
1.70'" 07
1.213'C 0
'. o*,'E 0
3.2ho0 O3
7.51F 05
9.21 3E Oi
I .3:E 0'7

4.02E O-0
1.43 7 01
6. 7' 00
'1 1 O'Li-
1 3 1 0*J


LB0
T

(7)

4.003E OB

1.435;E 10

1.237: ON0



4.151- 09-

1.4 J 3 10

7. 185 0')



2.358E 00

2. 54l'i 10

2.'0U50o 00

. '20E3 09

'9.370E 06O



1 .1571 09

2. 10'J 0')

5.071C 0'3

2.715 U00

1. 062 10


(9) (10)

0.490 4.995

0.491 6.9-13

1.0 39 I').225



0.24'5 2.331

0.157 3.195

0.272 1.869



0.381 11.079

0.33'3 1. 86

0. 561 3. 16

0.479 2.4

0.600 1.651


(8)

2.000E 09

1.02E 1i

2.037E 09

4.0 3F 0

9.67-jE 0i

4.515E 10

1. 3'3E 1 0



2. 01E 09

4.301E 10

1.116E 10

1.7220F 10

1.6300 0'-

5.71T7e 09

3.103' 09

I.211: 09

1. 720'i 10


2.710

3. 204

3. 04


0 .542

5.U62E 10 0.434


(11)

0.093

0.070


0 .529

0.103

0. 34'.





0.032

0. 17

0.143

0.192
0 1')2
0 .3. 3

0.212

3.106

0.12 -

0.227


3.109 3.140


0.28')

0.421

0. 92










Table IV-2. Continued.


Source
Telescope

(1)

IZW123
N
VI IZ r631
A
M:'.? 7
AN
IZW166
0
T 7./#.07
N
IIZI 168
A
IVz ?93
AN
A

ZW2220

11ZW185
AN

AN
IVZW14A 2
A
IV 149

IVZ! 153
AN.
ZwA335
AN


V21

6V21

(2)

677
2
4542
.3
4 7331
10
773J
53)
5631
3
8046
11
3855
6
7543
3
671 '
53

207~3
2
2063

4
3402
2
573+
5
1363
3


Mi MHI/L Mi/L MHI/Mi


D

A

(3)


(8) (9) (10)

1.639E 09 0.43b 12.735
0.075

1.835E 11 0.29'9 4.777
1 .50'3


(11)
0.034


3.063


AV BCF

F

(4) (5)

118 1.00O40
2. 40)
233 1 .? t39
1. 500
490 1.3040
2.567
1. 1 372 -
4. 000
109 1.0 14i
2.400
145 1.0972
2. 400
186 1. 10

287 1.23:25
2. 400
189 1 .0255
2. 400
180 1.074 1
4 J 1 ,
223 1.0213
.01
1.1077
2.400
220 2.7023
2 40')
294 1.3012
1.000
134 1.05-0
1* 't


MHI

AMHI
(6)

s 60" "4
9 .56
2.r699y
1 .431
1.179
5. 1,j5
2.404F

2.194F
3. 6 53E
3. 7

I 7 -
5. 1Ob 4
3.:34T

9.50-
1.192"
6.17C
2.602C
1. 04
2 .55 "'
(. 40 i

1 .63-:
6. '146!
4. 712
I .57 )9
2.87F


11.5
1.0
62.0
40.3


1 06 5 0
13.3
106.0
15. .3


113.0
.1 .
54.0
16.2
103.0
22.0
9 .0
21 .4
91.0
2 0 4
0 .4
30.0
10.3
109.0
11.1




3.5


0.076



0 .130

0. 141



0.072



0021


1.399F. 10 1.531F 10 0.182 1.094 0..166


I 2 7E/
1.2072

3.003-

3. ')4,5I

1 .5 7C


2. 162 10

5.243F 10

3. 721E 10

5.' %5 0o)

1.911E 10

2. 91 JC 10

1. 0')91 10

4. 561 E 011


u.6(33 10 0.231 3.046
0. 109

3.5b0O 10 0.320 2.300

1 .844E 10 0.476 3. 3 7
0. 134

1.348F 1 1 0.332 4.0L23

0 .627

7.471E 09 0.346 16.380








Notes to Table IV-2

VZw372: The spectra contains a strong source at V : 5800 km/sec in
the sidelobe of the 305 meter telescope.

Zw0855: The spectra shows two emission features. It is uncertain
whether both features are from this double object.

IIZw44: There is a difference of -4 times in the uncorrected HI masses
determined with the 91 meter and the 305 meter telescope. This could
be due to a giant HI halo. This object needs more careful study.

Haro22 and Haro39: No optical velocities were known. Their HI
emission lines were found by searching over a large range of veloci-
ties with the 91 meter telescope.

Haro35, Haro43 and Haro44: No optical velocities were known. Their
HI emission lines were found by searching over a large range of
velocities with the 305 meter telescope.

IZwl66: Confused in HI with two or three nearby galaxies. The HI
mass is for all these sources. It is not known which feature is due
to IZw166. This mass is not used in the analysis in the next chapter.

VIIZw631: Unfortunately, a 10 MHz bandwidth could not be used at the
time of observation, due to autocorrelator problems. There could be
emission outside of our 5 MHz bandwidth from this group of six
galaxies, which would have been missed. This HI mass is not used in
the analysis in the next chapter.








Column 2: D, distance to the source in Mpc. These are calculated in

the same manner as in Table IV-2, except that the heliocentric veloci-

ties are taken from the optical velocities of Table IV-1.


Column 3: A, major axis in kpc, calculated the same way as in

Table IV-2.


Column 4: MHImax upper limit to the HI mass. This is found by
max
taking three times the uncertainty defined by equation IV-8, and

assuming the velocity width is no greater than 400 km/sec. An N or

an A indicates which telescope was used, as in Table IV-2.


Column 5: L o, total blue luminosity, in solar units, calculated as
T
in Table IV-2.


Spectra of the detected systems are presented in Figure IV-1. The

source name is given in the upper left corner. The telescope used,

either the 305 meter or the 91 meter telescope, is indicated in the

upper right corner. The X axis gives the velocity in km/sec. For

spectra taken with the 91 meter telescope, the velocities are helio-

centric. For those taken with the 305 meter telescope, the velocities

are with respect to the Local Standard of Rest. The Y axis gives the

intensity of the channel values in Janskys. The two vertical bars,

one to each side of the spectral line, indicates the velocity range

which was integrated to determine the HI mass. All of the spectra

have been boxcar smoothed over three channels.













Table IV-3. Nondetected Sources.


Source D A MHI LBo
Hmax T
(1) (2) (3) (4) (5)

M335 103.0 5.3 1.2E 0-, 9AN 5.047E 10
VZWIl5 111.0 13 3 2.7E 09A 1.7 59 10
IIZ.v2y 113.0 '.9 2.2E 09A 1.61 E 10
11L~35 96.0 7.4 1.2E 0OA
N, 76.0 23.0 2.-: 3 09N 2.402 10
:4522 94. 0 13.2 1 .7E OJN 2.33"- 11)
"4105 49.3 .7 5.1E ,O)N 1.5b52 09
7IZ,,21 61.0 19.q 2.5E iN ').b94 09
114. 96.0 44.2 3.8E 99N
.Z 3 >' .3 o. 2.4E 09N
'1l8 97.0 20.7 2.3[ 09N 2.050E 10
150 92.0 o.2 1.3E 09A
.121 5 79.3 0.9 ?.2E 09 1 1.479: 10
IIlZWb8 76.0 20.1 7.5E 03A 1.5~7E 10
I IZWL 7 101.0 34. 1. O 034, A 4.'398A- 10
%12.3 10).0 14. '5 1].E6= A I 1C,'5: 1 ,
I w56 43.3 46.9 3.OE 094 2..900r 1 0
73 0 .38.0 14.9 3.5E 0() 7.577F' 0')
17 7 36.0 9.1 2.7E 08N 3.07 3 00f)
1ZA'-)? 75.0 26.2 1 091H 2.242E 10
lZw147 75.0 .9 9.7E 0 N
17 15 38.0 3.3 3. 5 03 N 1.1031r 09
IZW 191 10.1) '5,) I .oI 0C.N
ILZ. 9) 74.0 13.3 1.I1 n. N 9. '42. 0')
I Z." ,32 56.3 14.9 6.SE 0 N 1., ,3 10
IVZ .VG7 39.0 9 3 .C s A9
11 ZW172 107.0 31.6 1 .7t 0')A 2'-:F 10





















The HI spectra. The source name is given in the upper left corner of each spectrum. The
telescope used is given in the upper right corner. Velocities run along the horizontal
axis, in km/sec. For spectra taken with the 91 meter telescope, the velocities are with
respect to the sun. For those taken with the 305 meter telescope, the velocities are with
respect to the local standard of rest. The vertical axis is the flux strength in Janskys.
The vertical lines to each side of the spectral feature indicates the velocity range over
which the HI mass is calculated.


Figure IV-1.

















L-F: I ilLO F i4A F.8 0 1
-LL I JL I Li







']~ 7'
~FI ~--.1 J-F F Fil

H A F 1FIETi


200 1 ll 10' -1 ".111
5:111 9:nO riiloi11


T~4. KLWII


1' ni i in 7011 1 L II





FU I


*;II I L J iL .iii


'I- IETEF



E~~1IJ


II_. 1 III ._


q -0I fI Ift` III UI Iti It .II I 11 1ii


[VFT1 r7 ViFI



I 1- 1 A RLI iw41 I Fi



'nIll I-
1i -

T F i~uI iJ! ii


?fhflU 7ulr Pil-Irj I'l 2 1 0 '.4w







Ff_ - - I- TF---- F [ 1 I



S0 I2 Tiii,
",, ._ f V, Z i 1", i ,
I l _L L _L L J_ I J i_ J! L__.. I_ _i _. ._ .Jl I I
I I ;l I i I I -hl) l i I l"l I l I : l "1 L1i Jl 1 [ l I OI lI i i, 1 ll
- . -- I I I- - l-- ....F]..[. .I- -I .--- 111Th ii---i--i---
F- II, -I- F /
LTFIE. T M E
f1 IZ.1 I I




I.i :I ; r j -lA I I I -I L i 0 II .ii L I I 0 nI ,I 1 11 I .
I _ ..-T i ._- 11 I --1 -.
I M U F; LI HEL
,-i- i- r "-i r r- ... T -,r ,** ;-*| T ^---,r -, ,- 1


i, llIi 'II.

-1 . -A-. -- t iI
SL _. I _. .L__ -. _...._L ... l _l l_ 1 I I ..-. _. [ L_ I
,0 ,-, i -I IJ j 8 1 I.I ll iK .1 I iO 'l ,-l'i!1i 4 I I 4 i 'tI L. f l- I, l l i L- i !, -1,)
Table IV-1. (Continued)









-I ---.- ----T
[ l r i ME ir n








) :lijpil i... 4~'00


I ZLI'I iV'r


L Li
2L00 isIui 32i0ii i:ii0


,-U--F, E
S-I-IT--r~
jI MFicEr








i -
J_ 1__ L_'iI _
i.eOr-n 3aO0 i iln


21lOI 2 11iLI 3200 00i 1 IiiF1l
Table IV-1. Continued.


i L L __ I _I _
7,",n ,11I11 77i,


_L_._L_
1;,,lU F, l II U-,Z !I"


....__L.L L_ LLL_I_I.L_J.._ I_. I


7? iil 7800 Oj'ilp
.... -...... n-- -




_I LI_ I L_




'4- 01 i 1 II
00 j n ~n ^nun


I--I 1 I




ILL
S i
c-j_..~


T


i14


TT -Fi -
30. HiF rrr;


I_ .i. ..j L....

C41 HFTRE
76


r;i- on 1 n li I IIni I l in l Rr iin I i l 00 1 i ''. li Fl FI

1 1 I 1T f
:n; F i I 1- FTr,








P Li I L L jI
ln 1 i20n ?i11io S*:,i i Ui l '_ iII i i U 10 I|'iiij


- ,


lid Vfl,


I







VVVFT-
L-I I I Ii- ''1


--F-T--"~-T-7--
zFT -v IL-


Vi I 1F I LI Tf U-V v

I
LiEEI9 ,i _


I ifI 5l H 'I F1 H11A



Table Ip (Co t Ln ue1 I oji l I ' I (I I1II
--- ~ ~ ~ ~ ~ ~ ~ ~ Ir 1---r- l-- l-1-1---


r~ 1-II iIU ,1T 1_- I Yr F LL

hiL~liJ; 'ii

Ii Li-..l LI L..1...I __L h _! 1~. 1
TableL JV-l._1 (Cotnued


I










- - I-r-- I T -i -I F
I I Z idi5 10 ; 1EThLh









t o uCi|i-ii f6 0n 0 rlO.io 7cuii



1 I -Al -- i I








L I I L -_1 il
1:I ll Il il ; 11[111


Table IV-l .
Table IV-l. (Ci


1 ntinue
continued


I'L~iJI


- IUI


1 H[tTEi
9-1 HtLT[ i


1II._I_.L A L _
LI11 i, I


I1 1il ,
-r -,_ l ", /--r -l- ]r-',-J- T *---





SJ_ I _J.__l__1 L..... L _. I_
I J Ei l [I ll'l I' -,111]. ,'iiillj .


I i r .i..----|- -- r r TFH iLf










-L L I,
lI hI l l l i ii 6ill Ci i 1MULI. I liULl Iijii L, .li I d .i
)


' -






S i'1-1 ",i I I '.IU A 1 l i' l i U

,0 1-P










t_* __
I I 1 i l


O __ur_ 1I l_0 _.0J


[1W~


-! L


__ I_ l


SnA h .iL-E


_I ._L_. LI_


F--


r-
c


I Oi" "1 11 60-fU 511l':I 5110110 U '.I







T -F- T- T--- 1'-- -7 -i,.,-F 1 -1 1 l-* 1^7 |P, 3"i
: I-fHE h 1 r- 1,- - F1



-;i^ ^ ^ [ It j
I L-I
I -- i II 1 3
__L_. j _L h_ IU -. ' I l t _l .. ._ _F







]* c '- '_ 0 L l l I- 1 l u ii u 7 iU.1 I I j7 ^ 0 0 7 I '0 7 'd-l 1 ; 0 4 IO u 01 I 'I F 0 I O o80 1 J .ul- i )l r U1 U
.I.--I, _b 9 1. I- t I F Hl- -' ,FERl F -, -;- -. -..


j-oil 7 0 10 400 0 F0
k E r;
i ILIf1 I


h 0l.1 b'll l, i 7101.1 I
Table IV-1. (Continued).







T-1-~-I-- -I- rr--T-I--T







SFJLili o Tii i O '-F I I I -U L I r siu

II B F MEHTE H

1 (II
\J .

I I E

F L_J ..1 L__LL I__Ij_L_L__. .
-_ Hill l 'L- I ,Ii l PTE _
HO ^


.I C'LI_ l
Table IV-1.


II I i LI
2 I11] I I l I0
(Continued)


I- n -
" /I I'i Ii--


T



T iJ

Ill0 iI I l J u 1 :il l i IU



S*I1
i__ J_ L__I I__- _zjiL .i _LJ




I i - --1 l -i ... I .. -i
1 r' i Lit,.l _,ru ~ ,Il- r!i


I -7
1/Wilt 1


- I ML I F


_.J._-_ !i .L..-_J__ ..--
i\.i :ji' ,l 11 I i n ,LI ,i



fAI j .1 A, I


1 1 F TI I Wl ', 1 111
I - I 111r_ ,- - I li T L -


7Lij 91 m 1 l 1100 1:i0 iH









I: _. V I Z -







i I I 0
9 ~I. j I 9


!._'t i kI


I. I i l l r,.-. I'i


i I


LLLJ


J 1 t .I .L I,
Ta_.L_ LV-. (C_1

Table IV-1. (C


-I----- ... M
:05 METER


I 11710 4 l'll'l 5 1
T T-[....T---
1 II _1LILF










-F 1 .
;)I I 11110 ZL 11OJ



S1 I





:I h 500 H i:,:20 I
continued)


S.. I -1
I' c3 7A J 91

L l




I __F F_ .-- --




l l I tL l {I A I l _L



s ,t -


TT -
METER







_J__. I__


IL [II'g





I_! II i


I1 r I IJ'I qi I I I 11


M1 O 3


T ---1- TF
0 8-


n-r~
i T-. METL i
C M1 ETE VI -


1711U 1I 711i Fl L ? --l iF .uI. U
SI[ II-;li II ^*I^ 1








S__i _i .._I- .. -LL--_l-_ L ..- J
i, I I1 7t~lll: :::;I.HII I b l~ i E l I, _-I-


-I 1- Li -
i I'l,I i _111


F ul fltl t








b i l l i l ti l:l


,'J


72u0 7/, 11 I I 10 LI I &L0Q1







F-r----r T Fr-T 1F- -nI--r-- 1 r- T
STi JHETEh L MEr Ii





L Li LI11h -IJ 1 i-L-A -\.4 f1--- '-- I.
li lIil Jl L L_ L


LII Ci' l I Eclll iiiO 71 j''I ,*lSIII
I I F 1i
I i j1't II MLtFii

li tI~I


L__ 11_.. ..LLj L _
IT0 b;h I :i 'U Itii1l w "
Table IV-1. (Continued)


1|'1.1i


I IF q I T I I I







11 Sl --_ -- .l.. .l T ---J .....1-_ --r- f-l. 1 I_ I.-U- I- L- LF_ --. I
0 1






.- - /Itl; -li i *i
I vvm '-'/JI


1' 111, 1. 1
L UI L I H I i H iIt ,i ,' ,





- LI..-I i-..l. -. l L _J__ .._ l
', 1 l l. .. I -1, i I ~i 111 1 L I I IM
Shlull Fu 0 .(Hi 70HU1 Ibul oilll itiuUl liil O 2011I C1ison


,I ;-I I- I1Ij111 11 1[11 i "









F--F I -


C. I I,





ii I

'F~ II'l I;h l I 1



I t _(' L I' .' i _j 'I *
:1?1 /\111 ';1


_.I .1 S -
I,~i"i~._L__ L.


;1i iU i ,
Table IV- .


Lt IIII I 11-1
- - ]
_ _.J .-_^ ^-,


I _.l_ I.._. I


- I i II
_J 61H 1c'H


'i.U l l i l
(Continued)


I LI


-I I 15~C! Ii '








~I I I I
I l I i ,11


r
F















CHAPTER V
ANALYSIS OF THE DATA


The properties of blue compact galaxies seem to be more similar to

those of late type galaxies of the Hubble sequence than to any other

type. O'Connell (1979) found that the rotation curves of six luminous

blue compact systems were quite similar to those of late type galaxies.

Total masses, however, were found to be somewhat low for late type

galaxies. Studies by Fairall (1978) and O'Connell (1979) indicate that

many of the more luminous compact galaxies show some faint underlying

structure, indicative of late type systems. Neutral hydrogen studies

by Chamaraux (1977) indicate that the total masses and the HI masses of

blue compact galaxies extend into the same range as those of late type

galaxies. Finally, Huchra (1977) concludes that Markarian galaxies are

mostly late type systems.

We would like to compare the properties of the blue compact

galaxies in our sample with those of late type systems to see how

closely related the two types are. The best sample currently available

of the optical and HI properties of late type galaxies has been pub-

lished by Shostak (1978). Shostak's study utilizes the same system of

dimensions and magnitudes as our study of blue compacts. Thus we are

able to compare the properties of the two samples without any conversion

between the two. Of Shostak's (1978) survey of 169 late type systems,

we omit any which are peculiar, multiple, or confused in their HI.

This leaves 139 late type systems for comparison against the 72 blue









compact systems detected in HI. After a discussion of the selection

effects involved in our sample of blue compact galaxies, we will

compare their basic global properties to those of these late type

galaxies.


Selection Effects and Completeness of the Sample


Our list of 99 compact systems was drawn from the lists of Zwicky,

Markarian and Haro. Only the Markarian catalog is a reasonably

complete list.

Our selection of compact objects from the Zwicky (1971) list was

made from among the systems with published Doppler velocities. Many

of these were determined by Sargent (1970b), whose study was heavily

weighted toward the bluest Zwicky compacts, with the expectation that

they would be the most interesting to study. Since all the Zwicky

compacts with known velocities could not be observed in this study,

blue systems were selected preferentially over red objects, for it was

believed they would contain more neutral hydrogen. Thus, the sample

of Zwicky compact galaxies on our list is heavily weighted towards the

blue and very blue systems. Any statistics determined for Zwicky com-

pacts in general are not applicable to this study. However, the blue

and very blue Zwicky compacts can be considered together with Markarian

compact galaxies, owing to their similar morphological and spectro-

scopic properties.

The Haro galaxies were found at random and by accident. But their

number is so small that they obviously are a very incomplete sample.

A minority of these systems appear to be spiral or peculiar spiral

galaxies seen nearly edge-on and are not included in this study. The









majority can be classified as compact galaxies, because they were dis-

covered in a manner similar to that for the Markarian galaxies, and

because of their morphological and spectroscopic similarities. There-

fore, they can be considered a subset of the Markarian catalog and

considered together with the Markarian galaxies rather than as a

separate group.

Only a limited number of Markarian galaxies were placed on our

list for this study. Mainly, these were systems whose Doppler veloci-

ties were published after 1975. This was done because several neutral

hydrogen studies of Markarian galaxies by other observers were apparently

in progress, and we did not want to duplicate these studies. However,

no other criteria were used in choosing the Markarian galaxies.

Employing these criteria, our initial observing list contained

some 150 Zwicky, Haro, Markarian and Seyfert galaxies. Approximately

1/3 of these systems were not compact galaxies and they are not reported

in this study.

Our sample of compact galaxies can be considered, for the most

part, a subset of the Markarian catalog. A few of our sources were

red objects without Markarian-like properties, but these were mostly

undetected in HI. As this study represents only a selected subset of

the Markarian catalog, we must be very careful in applying the statis-

tics found for Markarian galaxies in general. We need to estimate

the fraction of Markarian galaxies that are compact galaxies. The

study of Huchra (1977) showed that approximately half of all Markarian

galaxies may belong in the normal Hubble sequence, mostly as late type

systems. Most of the remaining Markarian galaxies can be classified

as compact galaxies. Thus perhaps half of all Markarian galaxies are









compact galaxies. On the original list of 150 sources, 24 were selected

only because they were Markarian galaxies (36 other Markarian galaxies

were selected because they were also Zwicky, Haro, or Seyfert galaxies).

Of these, we have classified 16 as compact and 8 as noncompact. Thus,

the estimate that half of all Markarian galaxies are compact is probably

not an overestimate. Indeed, the low luminosity, dwarf systems in the

Markarian catalog may all be compact systems.

The Markarian catalog appears to be substantially complete down to

an apparent photographic magnitude of -15.5 (Sargent, 1972). The pro-

portion of all galaxies that are of the UV intense, Flarkarian type has

been estimated in three studies. Sargent (1972) finds this proportion

to be -2.5% for -22 < M < -20; -7% for Mp = -17; and perhaps as high

as 10% for M = -14. Huchra and Sargent (1973) find this proportion

to be 5-10% for -22 < M < -14. Huchra (1977) finds it to be -6% for

-21 < M < -14. Huchra (1977) included Seyfert galaxies in his es-

timate while the first two studies did not. There is a lot of uncer-

tainty at the low luminosity end. Huchra and Sargent (1973) point out

that the statistics for low luminosity systems is rather poorly known

for both field galaxies and Markarian galaxies.

The majority of the Markarian galaxies are not extremely blue sys-

tems. In the color-color diagram, approximately 75% of the Markarian

galaxies overlap the colors of field galaxies (Huchra, 1977). Thus,

only -1/4 of the Markarian galaxies are exceptionally blue systems.

Huchra (1977) made a rough estimate of the space density of galaxies

as a function of their colors. For (U-B) 5 -0.4 and (B-V) < +0.35, it

seems that over half of all galaxies are Markarian galaxies. Thus, the

Markarian lists seem to be fairly complete for the bluest systems, but









probably less complete towards the redder systems. This bias must be

present among the objects in our study and is probably accentuated owing

to the method by which the objects were chosen. Thus, we expect the

ratio of very blue to blue compact galaxies in this study to be greater

than the actual ratio of these systems.

We can make rough estimates of the proportion of all galaxies

which are blue or very blue compact systems. Overall, it is perhaps

1-5%. For -22 < M < -20, it is -1-3%; for M = -17, it is -2-4%;

and for Mp = -14, it is -3-10%.

Our sample of blue compact galaxies is a magnitude limited sample.

The Markarian catalog is restricted to 13 < M < 17. The Zwicky and

Haro lists have similar, but less well defined limits. In our sample,

only one source is brighter than BT = 13.0 and only one is fainter than

BT = 16.4. This has the effect of producing a strong distance dependency

in the absolute magnitudes. This can be seen in Figure V-l, where we

plot absolute magnitude, MBo, versus log D. The absolute magnitudes
T
must be confined to a band -3-4 magnitudes wide with a slope of -5.

For example, systems of exactly B' = 15.0 would lie along the straight

line defined by MB = -5 log D 10.0. A least squares regression fit

to the data of Figure V-l gives MBg = -4.82 (.27) log D 10.96

(.45).

Thus, nearby, very luminous compacts and distant, intrinsically

faint compacts are excluded from this study. Obviously, the distant,

dwarf systems are too faint to detect. However, might there be any

nearby, very luminous compacts, excluded from our survey, that could

be studied to learn more about the distant systems? Huchra and Sargent

(1973) have estimated the space density of Markarian galaxies. Assuming







































.2

C
C



C
C
C
C
CC
C
C0


CU






'II













-I
II






C1



CI


Figure V-1. Absolute magnitude, MBo, versus the log of the

distance, in Mpc, for blue compact galaxies
distance, in Mpc, for blue compact galaxies.


2.31


C

C


C C


C O
C


L5 G
iLi









that half of the Markarian galaxies are compact, then at M = -22, we

should expect only one blue compact within -50 Mpc. We find none this

luminous in our study. At M = -21, we should expect only one blue

compact within -20 Mpc; at M = -20, we should expect only one within

-12 Mpc; and at M = -19, we should expect only one within -10 Mpc.

Thus, there is a strong selection effect against bright, blue compact

galaxies close to the Milky Way and our list may be deficient by a

very small number of luminous, nearby systems.


Luminosities


Compact galaxies are defined to be systems with higher than average

surface brightnesses. We would like to see how the luminosities and

surface brightnesses of the compact galaxies in our sample compare to

those of more normal galaxies. We first compare their luminosities to

those of the late type galaxies taken from the survey by Shostak (1978).

In Figure V-2, we plot log LBo versus log A. The blue compacts do not
T
extend to as high luminosities or dimensions as the late type. This

difference does not seem to be great. However, the blue compacts do

extend to much lower luminosities than late type galaxies. Late types

appear to have a cutoff in luminosity at LBo 10 L and only -8% in

this sample are fainter than LB = 2 x 10 L. For the blue compacts
T
of our sample, -32% are fainter than LBo = 2 x 109 L and -22% are
9 T
fainter than LBo = 10 L Thus, on the order of 1/3 of the compacts
T
in our sample seem to be dwarf systems.

In the region of overlap in Figure V-2, blue compact galaxies are

somewhat more luminous, on the average, than late type galaxies of

equal dimensions. Least squares regression lines have been fit to the














O


--







a
a

--







a
N










O







Oh

a
C

~L3,

i


23




0
C!


M


LOG P


1.25


Figure V-2. Log LB; as a function of log A. The squares are blue

compact galaxies. The crosses are late type galaxies.
compact galaxies. The crosses are late type galaxies.


++f
-4--





-1i+-














two groups. For the blue compact galaxies, the luminosities are

given by


log L = 1.88 log A + 7.86 V-1
T (.11) (.11)

For late type galaxies, it is given by


log LBo = 1.63 log A + 7.93 V-2
T (.09) (.12)


The difference in the slopes of the two lines appears to be sig-

nificant, considering the uncertainties. On the average then, blue

compact galaxies are more luminous than late type galaxies by a factor

of -1.5 at A = 10 kpc and by a factor of -1.9 at A = 25 kpc.

Next, we consider the surface brightnesses of blue compact galaxies.

Sargent (1970b) found that the surface brightnesses of Zwicky compact

galaxies were -100 times greater than is found in normal Hubble sequence

galaxies. Sargent, however, made the mistake of comparing Holmberg

dimensions for normal galaxies with the dimensions of only the bright

cores of the compact galaxies. These two systems are not compatible.

However, O'Connell (1979), in a study of blue compacts, found that their

average surface brightness was about one magnitude greater than that in

late type galaxies.

A thorough study of surface brightnesses of blue compacts would

determine their central surface brightnesses. Unfortunately, in this

study, we can only find the surface brightnesses averaged over the

total dimensions. This will dilute the effect of the high surface

brightness cores of these objects.








For a measure of the average surface brightness, we divide the

luminosity by the square of the major axis. The log of this is plotted

against color, (B-V)T, in Figure V-3, for blue compacts and for late

type galaxies. Blue compact galaxies are found to be typically bluer

and of higher surface brightnesses than late type galaxies. For 63

blue compacts, the average of their surface brightnesses is 1.998 z 2

times as high as that of the late type galaxies. Among the blue com-

pacts, the highest surface brightnesses are -10 times the average for

late type galaxies.

As approximately 1/3 of our blue compacts are dwarf systems, we

would like to know if the surface luminosities are different for dwarfs

and nondwarfs. The HI masses are an indication of sizes, as will be

seen in the next section. Thus, we plot surface luminosity versus HI

mass for compact systems in Figure V-4. The data for blue compacts are

quite scattered and no correlation is seen. Thus there does not seem

to be any significant difference between dwarf and nondwarf blue

compacts, in terms of their surface luminosities.

Searle et al. (1973) used a model of star bursts to argue that very

luminous galaxies could not be of the extremely blue type, such as the

low metal abundance dwarf systems. The model they used predicted that

at M = -17, the fraction of all galaxies that are extremely blue would

be -10-4, and much smaller at greater luminosities. Thus very blue

systems brighter than M -17 (or Lo = 109 L) should be extremely

rare. However, two of the 11 metal poor systems studied by Alloin et al.

(1978) appear to be of greater luminosities than this limit. The iso-

lated system M162 has a similar metal abundance as IIZw40 but has

Mp -19. A double system, M171, is similar to IZwl8 in metal abundance

but has M = -21.































_. '1
"







c!



C




AT




CD

'-F---.
N

WLC |








Do
5- I



io

C3
C3 1
ml a


0
0


G n
+.3 a +


C- +_+ t' -

,- I t -- -



++

4-


[3-y To


o.eo


O.So


Figure V-3. Log (LBo/A2) as a function of color index, (B-V)'
B- T.
T
The squares are blue compact galaxies. The crosses
are late type galaxies.


nD
a -
CJ-


---


+


;.2






















3
E
a i, =a

"0 C





al M
0
am 3
0
0
0 0
0 -
0 a


9. -0


10.20


11. 0


Log (LBo/A2) as a function of log MHI for blue


compact galaxies.


0
0

On
5
0~
0


0
22


C i
rrJ




















cci








Zt0


L 5. u uC H

LOG M1,


Figure V-4.


, r


! ; i


j








To look more deeply at this hypothesis, we plot, in Figure V-5,

log LB0 versus (B-V)T and, in Figure V-6, log LBo versus (U-B)T for blue

compact galaxies and for late type galaxies. For late type galaxies,

it is seen that -2% are bluer than (B-V)T = +0.30, and -15% are bluer

than (B-V)- = +0.40. For blue compacts, we find that -34% are bluer

than (B-V)T = +0.3, and -66% are bluer than (B-V)- = +0.4. Approxi-

mately 1/3 of the blue compacts are bluer than the bluest late type

galaxies in (U-B)T. From Figures V-5 and V-6, it can be seen that the

bluest compacts are dwarf systems, but that extreme blueness is not

limited to the dwarf galaxies. It appears that, of the blue compacts

in our sample brighter than LB = 109 L at least 20% are very blue
Tvr
objects.

From Sargent (1972) and Huchra (1977), it appears that, of all

galaxies brighter than M = -17, on the order of 3-7% are the UV intense,

Markarian types. If half of these are blue compacts and if 20% of

these blue compacts are extremely blue, then on the order of .3-.7%

of all galaxies brighter than Mp = -17 are extremely blue. We must

keep in mind that the Markarian lists may be more complete at the

extremely blue end. Thus, this proportion may actually be as low as

.1 or .2%. This is still considerably greater than the proportion

predicted by Searle et al. (1973). However, this finding does not

invalidate the star burst model. Rather, it indicates that the model

of Searle et al. (1973), of flashes of star formation occurring in

small, statistically independent cells in a galaxy, is not a good

description for the mechanism of the star bursts in these systems.

What is needed is a model which takes into account all of the properties

of these systems and allows star bursts at all luminosities.









0
i3

7:T











o
















,i
O-l



*-n
.-- _







u !u
H:

5921


--
1 -
_+ -
.__
+ + '

a 4- a
a
-I -~
+ -


a
4 5
S -45


0.20
( 3 -/
T


0.10


0. 0


3.30


Figure V-5. Log (LBo) as a function of color index, (B-V)T
T
The square are blue compact galaxies. The crosses
are late type galaxies.


I_
























+


++
-r + -i
+ +
M

f;+


7-m5 1 -C.511
(LI P.I


0. 5


Figure V-6. Log LBo as a function of color index, (U-B)T.
T
The squares are blue compact galaxies. The crosses
are late type galaxies.


-ii



KT




.w]


-0.:5









Neutral Hydrogen Content


We would also like to see how the neutral hydrogen masses of blue

compact galaxies compare to those of late type galaxies. In Figure V-7,

we plot log MHI versus log A. The two groups follow a similar relation-

ship. For the blue compact galaxies, the HI masses are given by


log MI = 1.51 log A + 7.68 V-3
HI (.16) (.12)


For the late type galaxies, they are given by


log MHI = 1.61 log A + 7.43 V-4
(+.07) (.10)


The differences between the two regression lines is slightly less

than significant. The indication is, though not certain, that blue

compact galaxies are more HI rich than late type galaxies of the same

dimensions. A similar result was found by Chamaraux (1977). At A =

6 kpc, the blue compacts contain, on the average, -1.5 times as much

HI as late type galaxies. However, at A = 40 kpc, this is reduced to

a factor of -1.2. We stress that these values are rather uncertain,

but at the least, blue compact galaxies have marginally more neutral

hydrogen for their sizes, than late type galaxies.

As was seen with the luminosities, about 1/3 of the blue compacts

in our sample are dwarf systems. Late type galaxies have a fairly

sharp cutoff at A = 6 kpc. This may, however, be a selection effect

in the classification of low luminosity galaxies. Very few of these

were included in Shostak's (1978) sample, and several of these we

excluded from this discussion because they were classified as peculiar

systems.







89












CD
cc
o



















23


















-r
-L C S






































Figure V-7. Log 11HI as a function of log A. The squares are blue
compact galaxies. The crosses are late type galaxies.
! 2
(-S" -^ -e



2 2


CD

-I -

















Figure V-7. Log MHI as a function of log A. The squares are blue

compact galaxies. The crosses are late type galaxies.








We next consider the HI surface densities of these systems, which

can be represented by dividing the HI mass by the square of the major

axis. In Figure V-8, this is compared with color index, (B-V)T. The

blue compacts can be seen to have significantly greater HI surface

densities than late type galaxies. For 70 of the blue compacts detected

in HI, the average HI surface density is 1.98 z 2 times as high as in

the late type galaxies. The HI surface densities of some of the blue

compacts are much higher than this, up to 12 times the average for

late type systems. This HI surface density does not seem to differ

between dwarf and nondwarf blue compacts. This is seen in Figure V-9,

where HI surface density is compared to luminosity. The HI surface

densities considered here are derived from the optical dimensions of

these systems. Ideally, they should be derived from the HI dimensions,

but these are not easily obtained. In the next section, we shall dis-

cuss the HI dimensions of blue compact galaxies and investigate whether

their true HI surface densities are actually greater than those of late

type galaxies.

As was done for the luminosities, we plot log MHI versus (B-V)T in

Figure V-10 and log MHI versus (U-B)T in Figure V-ll for blue compact

galaxies and late type galaxies. The results are similar to those found

for the luminosities. The bluest systems are seen to have low HI

masses but extreme blueness is not limited to the low HI mass, dwarf

systems.


HI Dimensions


As outlined in the last chapter, the HI masses were corrected for

beamwidth effects by assuming that the HI in these systems was











a

















C-2 _






a a a-


44-
T






























Figure V-8. Log (MI/A2) as a function of color index, (B-V)4
The squares are blue compact galaxies. The crosses
are late type galaxies.
C' I
e V. L (A a f o cojor ie B





























are late type galaxies.













7










C



Sn9

r-
'"oT














+ I
c In

T9
+ -4-
+ 4_


+
n c
'2719 O1 D L IDL. OF =









F igr V-9 Lo (M1/ as a fucto of lo (L^ orbu









BT
CL3 -T + ^ -












FiueV9 og([IA2 a a func ion oflg( O frbu

T













compact galaxies (squares) and for late type
galaxies (crosses).
i +_+-^- T *' +



3~ ~ -r ++ :




o i
'"'.o ~ a .0 i 0n oois o








FigreV- 0 Lo (,,/ )a a f cin. ofo g (L-o) o bu





compact galaxies (squares) and for late type
galaxies (crosses).






















+




+ +





-4- r
o Ci --+--
+ '


+ n -
+ 3 _
7-


-.1







I-rL


0.i 4


0.60


Figure V-10. Log MHI versus color index, (B-V)T, for blue

compact galaxies (squares) and for late type
galaxies (crosses).


0.20
(B


i I


0.30




Full Text

PAGE 1

A NEUTRAL HYDROGEN SURVEY OF BLUE COMPACT GALAXIES BY DAVID GORDON A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IIVERSITY OF FLORIDA 1979

PAGE 2

To My Parents, Dora Denby Gordon and Maurice Gordon

PAGE 3

ACKNOWLEDGEMENTS I wish to express my appreciation to Dr. Stephen T. Gottesman, who introduced me to the field of extragalactic neutral hydrogen studies. It was his suggestions and ideas which led to this study. His experience and his suggestions, comments and encouragement have been invaluable in the HI observations and in the analysis of the data. Many others deserve mention here. At the National Radio Astronomy Observatory, Dr. Richard Fisher and Dr. Patrick Crane were most helpful in making observations with the 91 meter telescope. I am grateful to the many telescope operators and engineers who helped make the observations a success and to Al Wu, in particular, who did an excellent job of maintaining the receiver system. I am also grateful to the many telescope operators and engineers at Arecibo Observatory for performing an excellent job. I particularly wish to thank Dr. Edward K. Conklin and Dr. Michael M. Davis for their expert assistance in making the spectral line observations on the 305 meter telescope. Special thanks are due to Dr. Nathan A. Krumm for his invaluable help in reducing the observations at Arecibo Observatory in 1977. Special thanks must also be extended to Dr. Martha P. Haynes, who provided computer programs which allowed me to read and reduce the 1978 Arecibo data tapes at the University of Florida. Special thanks are also extended to Dr. Gerard de Vaucouleurs and Dr. Harold G. Corwin for calculating blue magnitudes from Zwicky

PAGE 4

photographic magnitudes for several dozen galaxies. Without their help, the luminosities of many systems could not have been determined as accurately as they were. Further thanks are extended to Dr. Michael D. Desch and Mr. David R. Florkowski for help in miscellaneous areas, inspiration, and many interesting discussions. IV

PAGE 5

TABLE OF CONTENTS CHAPTER PAGE ACKNOWLEDGEMENTS iii ABSTRACT vi I INTRODUCTION 1 II BLUE COMPACT GALAXIES 8 III THE OBSERVATIONS 16 The 21 cm Line 16 21 cm Spectral Line Observations 19 Observations on the 91 Meter Telescope 22 Observations on the 305 Meter Telescope 28 IV THE DATA 36 Optical Parameters 36 21 cm Global Parameters 50 V ANALYSIS OF THE DATA 73 Selection Effects and Completeness of the Sample . . 74 Luminosities 79 Neutral Hydrogen Content 88 HI Dimensions 90 The HI Profiles 100 Total Indicative Masses 115 Comparisons Between Color, HI Mass, Luminosity and Indicative Mass 118 Comparison With Low Surface Brightness Systems . . . 132 Correlations 134 VI CONCLUSIONS 138 Suggestions for Future Study ... 142 BIBLIOGRAPHY 144 BIOGRAPHICAL SKETCH 148

PAGE 6

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy A NEUTRAL HYDROGEN SURVEY OF BLUE COMPACT GALAXIES By David Gordon December, 1979 Chairman: Stephen T. Gottesman Major Department: Astronomy Compact galaxies are galaxies showing high surface brightnesses and small angular dimensions. They are, morphologically, a \jery inhomogeneous group, ranging from featureless objects barely distinguishable from stars on plates of the Palomar Sky Survey, to objects showing various irregularities, such as multiple cores, jets, filaments, bridges to nearby objects or other irregularities. Most of the compact galaxies catalogued by Zwicky are red or very red systems. These appear to be early type galaxies deficient in their outer envelopes. A minority of the Zwicky compact galaxies, perhaps 10-20%, are blue or very blue. These appear similar to the compact systems found on the Haro and Markarian lists of galaxies with strong UV excesses. A fairly common characteristic of blue compact galaxies is the presence of emission lines in their spectra, either with or without absorption lines. The excitation level of the emission in these objects indicates that the lines originate in HII regions excited by hot young stars. Several VI

PAGE 7

of the objects in this class, mostly dwarf systems, have been found to be underabundant in elements heavier than helium. Possible explanations are that these are young galaxies just commencing star formation, or that they are old galaxies experiencing a brief burst of star formation, to be followed by a long quiescent period. This study is a report of the 21 cm neutral hydrogen line observations of 99 compact galaxies, most of them blue or very blue. Of these, 72 were detected in HI. These systems are found to be richer in HI than late type galaxies of the Hubble sequence and also to have higher surface luminosities. Approximately half of these blue compact systems are bluer than the bluest late type galaxies. Their HI velocity profiles are often narrow and single peaked, most often for the lower luminosity systems. However, the evidence suggests that the compact galaxies are rotating disk systems, perhaps with a strong central concentration of HI. The observations also indicate that at least some of these systems are quite extensive in their overall HI dimensions. It is possible that these HI rich, blue compact galaxies are old systems which have not yet completed their collapse out of the intergalactic medium. Primordial gas may be condensing into their central regions and could be the source of gas for intermittent, brief bursts of star formation. vn

PAGE 8

CHAPTER I INTRODUCTION In addition to the many normal types of galaxies which fit nicely into the Hubble sequence of galaxies, astronomers now recognize the existence of various types of galaxies which are described as being in excited states, or "active." These types of active galaxies are known mainly under the names of Haro galaxies, iMarkarian galaxies, compact galaxies, Seyfert galaxies, N galaxies and quasars (van den Bergh, 1975). There is currently wide discussion as to where, if at all, these systems belong in the Hubble sequence of galaxies. The following study is concerned with the compact and high surface brightness galaxies on the lists by Zwicky, Haro, and Markarian. Zwicky (1964) defined a compact galaxy as an object of abnormally high surface brightness just distinguishable from stars on plates taken with the Mount Palomar 48 inch Schmidt telescope. During the years 1960-1968, while examining the Palomar Sky Survey plates for compilation of the Catalogue of Galaxies and of Clusters of Galaxies (Zwicky et al., 1961, 1965; Zwicky and Herzog, 1963, 1966, 1968; Zwicky and Kowal , 1968), Zwicky produced seven lists of what he called compact galaxies, galaxies with compact parts, and post-eruptive galaxies. Some 30 years before this, Zwicky claims to have known of the existence of galaxies with super-dense stellar populations (described by Zwicky, 1971). He believed that these compact galaxies were the super-dense stellar systems he had predicted (Fairall, 1978). 1

PAGE 9

Zwicky distributed his seven lists privately to various astronomers and later published them in book form. The Catalogue of Selected Compact Galaxies and of Post-Eruptive Galaxies (Zwicky, 1971) contains some 2300 systems. Morphologically, these systems make up a very inhomogeneous group. As seen on the Palomar Sky Survey, many of them have a small saturated image surrounded by a small amount of nebulosity, while others show various irregularities, such as jets, rings, bright knots, distorted spiral arms, multiple nuclei and bridges to nearby objects. A common feature among these systems is a region or regions of high surface brightness (Sargent, 1970b), The majority of these galaxies have normal reddish colors, as seen in E and SO galaxies and have normal absorption line spectra. A minority, perhaps 10-20%, are blue or "jery blue and often have emission lines in their spectra, indicative that they have regions of ionized hydrogen. The designation of color by Sargent (1970b) is used here, where a "blue" image is of approximately equal brightness on both plates of the Palomar Sky Survey; "^^ery blue" is brighter on the blue plate; "red" is brighter on the red plate; and "^ery red" is much brighter on the red plate. The Markarian galaxies are blue galaxies discovered on objective prism plates taken with the 1 meter Schmidt telescope of the Byurakan Observatory by Markarian and his co-workers. They have published 11 lists so far, containing 1095 galaxies (Markarian, 1967, 1969a, 1969b; Markarian and Lipovetskii, 1971, 1972, 1973, 1974, 1976a, 1976b; Markarian et al., 1977a, 1977b). A 1.5° prism was used, yielding a dispersion of 2500 A°/mm at H^. The aim of their study is p to catalog galaxies with intense ultraviolet continua between apparent magnitudes 13 and 17.

PAGE 10

On these lists, the objects are classified into one of two groups based on the appearance of their objective prism spectra (Markarian, 1967). Objects having sharp spectra (like stellar spectra) are called type s. A bright nucleus is the source of their UV continua. Some of these s types are spiral galaxies with bright nuclei and some are Seyfert galaxies with broad emission lines arising in their nuclei. In the second group, type d, the boundaries of the spectra are diffuse, indicating that the UV continuum originates in an extended region and not just the nucleus. Intermediate types are designated as types sd or ds, depending on which type they resemble most. Additionally, a common characteristic of Markarian galaxies is the presence of narrow emission lines in their spectra. Haro (1956) reported on the discovery of 44 blue and wery blue galaxies. These were detected in the course of an objective prism search for \jery blue stars, using the Schmidt telescope of the Tonontzintla Observatory. Like the Markarian galaxies, they show strong UV excesses. These systems generally fit into the diffuse (class d or ds) spectroscopic subgroup of the Markarian objects. Like the Markarian galaxies, emission lines are a fairly common feature of their spectra. Many of the Markarian and Haro galaxies are similar to the blue and very blue high surface brightness compact and post-eruptive systems listed by Zwicky (1971), in that they are of similar colors, are of small angular dimensions and have a region or regions of high surface brightness. Some of these have circular or elliptical shapes like some of the featureless Zwicky compacts, while others display various irregularities, such as jets, filaments, multiple cores and other

PAGE 11

irregularities. Thus, morphologically, the three groups show many common properties. The overlap of their properties is further illustrated by the fact that of the 99 galaxies reported on in this study, 20 are on both the Markarian and Zwicky lists; 10 are on both the Haro and Markarian lists; and 2 are on all three lists. Thus, there is ample reason to group objects from these three lists together in a study of the properties of blue compact galaxies. The original definition of compactness (Zwicky, 1964) as being just distinguishable from stars on plates taken with the 48" Palomar Schmidt telescope and with diameters of 2-5 arc seconds has proven too restrictive to be of practical use. This definition has not been rigorously followed by workers in the field, including Zwicky himself. Zwicky' s (1971) definition of compactness also requires that the surface brightness be brighter than the 20th magnitude per square arc second. This requires surface photometry and has not been attempted for very many compacts. The more or less accepted procedure is to classify as compact those systems which show, on the Palomar Sky Survey, a small saturated core, either regular or irregular, surrounded by none or small amounts of nebulosity, showing no normal structure (such as spiral arms) but possibly showing irregular structure, such as jets, filaments, bridges or double cores. This is the definition that is used in this study to select compact galaxies from the lists of Zwicky, Haro and Markarian. For the benefit of the reader, we reproduce in Figure I-l several examples of compact galaxies. These are negative prints, taken from the Palomar Sky Survey. In the following study, we report on neutral hydrogen observations of 99 compact galaxies. Almost all of these galaxies were blue or very

PAGE 12

Figure 1-1. Examples of blue compact galaxies taken from the Palomar Sky Survey. M7 (= VIIZwl53) has an irregular shaped bright core surrounded by a small amount of nebulosity. M297 (= Arp209) has a "hat" shaped core composed of two bright knots side by side, and a small amount of nebulosity. IZw207 is a "boomerang" shaped arc of bright knots. IIZw40 is a dwarf blue compact showing a bright, nearly stellar core and a fainter nebulosity containing two filaments. Haro29 (= Izw36 = M209) is a dwarf blue compact showing two bright condensations, a jet, and a "fan" shaped nebulosity. Haro25 (= M727) shows a bright stellar core with a '^ery small amount of nebulosity and is barely distinguishable from a star.

PAGE 13

/ M 297; J ' .*. t"7. -1' /'f ^"^ II Z^ 40 "7.;M •?!; •J'* /"^ V >

PAGE 14

blue, and thus our discussion will be concentrated on this type. We shall be concerned with what can be learned from their neutral hydrogen global properties when considered in conjunction with their known optical properties. For our discussion and analysis, a Hubble constant of 75 km/sec/Mpc will be used throughout. A more detailed discussion of blue compact galaxies follows in Chapter II. In Chapter III, we describe how the 21 cm spectral line observations were made and reduced. Optical and neutral hydrogen global data are presented in Chapter IV. Discussion and analysis of the data are given in Chapter V. Conclusions and directions for future work are given in Chapter VI.

PAGE 15

CHAPTER II BLUE COMPACT GALAXIES The lists of Zwicky, Markarian and Haro contain some 3500 galaxies. The majority of these objects are of small angular dimensions and of high enough surface brightnesses to be classifed as compact galaxies. The Haro and Markarian compact galaxies are all blue or very blue as a result of their selection criteria. Zwicky compacts, on the other hand, cover a wide range of colors — from as red as the reddest Hubble sequence galaxies to bluer than the bluest type I irregular galaxies. No more than 20% of these are in the blue or ^^ery blue category (Sargent, 1970b). The objects cataloged by Zwicky (1971) were selected at random, but they were not chosen in any uniform manner over the entire sky and are not complete down to any limiting apparent magnitude. The Markarian lists, on the other hand, appear to be complete down to a limiting photographic magnitude of m = 15.5 (Sargent, 1972). In 1969, Zwicky began a complete survey for compact galaxies on 12 plates taken with the 48" Schmidt telescope (Zwicky, 1971). He appears not to have completed or published more than a few preliminary results of this study before his death in 1974. However, Rodgers et al . (1978) studied the compacts marked on two of these plates. Out of 348 objects marked on one plate, using Sargent's (1970b) color designations, they found that 232 were \/ery red, 79 were red, 25 were blue and 12 were ^ery blue. Thus perhaps as few as 10% of the Zwicky compacts are blue or ^ery blue.

PAGE 16

studies of these red and very red compacts are reported by Fairall (1971, 1978), Kormendy (1977), Sargent (1970b) and Rodgers et al. (1978). These systems typically show normal absorption line spectra typical of early type galaxies. Surface photometry of these systems indicates that their central surface brightnesses are not so much higher than is seen in normal early type galaxies (Kormendy, 1977; Rodgers et al., 1978). Their primary difference from normal E and SO galaxies seems to be a deficiency in their outer envelopes (Kormendy, 1977; Fairall, 1978; Rodgers et al., 1978). Thus it appears that these systems do not have the abnormally high stellar densities expected by Zwicky. Fairall (1978) has suggested that these systems have lost their outer envelopes as a result of tidal stripping. We shall now concentrate our discussion on the properties of blue compact galaxies, with which this study is primarily concerned. DuPuy (1968, 1970) has studied Haro galaxies. Sargent (1970a, 1972) has surveyed Markarian galaxies. Sargent (1970b) made a study of Zwicky compacts which was heavily weighted towards blue objects. These blue systems are characterized by a strong UV continuum, colors which are often bluer than the bluest late type galaxies, and moderate or strong emission 1 ines--typical ly the Balmer lines and forbidden lines of oxygen, neon and nitrogen. The emission lines and the UV continuum are generally concentrated in, but not restricted to, the nuclear regions. About half of these systems also show absorption lines typical of the stellar populations in late type galaxies. All 31 Haro galaxies studied by DuPuy (1968, 1970) showed emission lines. Those brightest in the UV showed the strongest emission lines. Among the Markarian objects, some 80% seem to show emission lines (Sargent,

PAGE 17

10 1972). Sargent (1970b) found that all Zwicky compact with (U-B) between +0.20 and -0.15 had both emission and absorption lines and that all those bluer than (U-B) = -0.15 had emission lines only. The levels of excitation seen in blue compacts are comparable to those of galactic HII regions. Thus it is generally agreed that the emission lines originate in HII regions excited by hot stars. Their strong UV continua appear to be stellar in nature, the result of a high number of hot stars (DuPuy, 1970; Sargent, 1970a, 1970b, 1972; Searle and Sargent, 1972; Forrester, 1973). The space density of blue, UV intense galaxies has been estimated by Sargent (1972). For absolute photographic magnitude M between -20 and -22, about 2.5% of all galaxies are UV intense. Their abundance increases towards lower luminosities. At M = -17, about 7% of all galaxies are UV intense and at M = -14, it may be as high as 10%. The luminosities of blue compact galaxies cover a broad range. At the high end, they are as luminous as normal galaxies, and at the low end, they are as intrinsically faint as the smallest dwarf galaxies. In the surveys by Sargent (1970b, 1972), the absolute photographic magnitudes of the Zwicky and Markarian narrow emission line galaxies were found to extend from -21.6 to -14.9, and from -21.5 to -13.9, respectively. DuPuy (1970) finds a similar range in the absolute visual magnitudes of Haro galaxies, from -21.6 to -13.7. With their broad range of luminosities, blue compact galaxies are not a homogeneous group. However, the low luminosity dwarf systems may form a fairly homogeneous subgroup. Sargent (1970b) noted several low luminosity blue systems fainter than M = -19 among the Zwicky compacts.

PAGE 18

11 Eleven similar dwarf systems fainter than M^ -16 were identified by Sargent (1972) from the Markarian lists. Several of the Haro galaxies also fit into this subgroup. These systems are similar in that they are all nearly uniformly high in excitation levels, have strong emission lines relative to their continuum, have physically small emission line regions, typically a few hundred parsecs across, and are often bluer than the bluest galaxies of the Hubble sequence (Sargent, 1972). An exact upper luminosity limit for considering an object a dwarf is not clearly established. An absolute photographic magnitude of approximately -17 or -18 appears to be a reasonable upper limit. Two such blue dwarf compact systems were labelled as "isolated extragalactic HII regions" by Sargent and Searle (1970). These types of systems are similar to the largest and highest excitation HII regions found in the spiral arms of giant late type galaxies, and yet they often appear as isolated systems. IIZw40 and IZwlS were studied, using high resolution spectroscopy. Searle and Sargent (1972) reported on the abundances found in these two systems. They were found to have normal and presumably primordial abundances of helium, but significant underabundances of heavier elements. IIZw40 and IZwlS were found to be underabundant in oxygen and neon by factors of approximately 3 and 7, respectively, compared to normal HII regions in the solar neighborhood. (Bergeron (1977), using the same data, gets greater underabundances, of 6 10 for IIZw40 and 30 50 for IZwlS.) These were the first metal poor. Population I systems found, raising the question as to whether they are young galaxies. Obviously they could not have been producing and B stars at their current observed rates for 10 years and have remained metal poor. IIZw40 would have converted approximately

PAGE 19

12 1/4 of its total mass into elements heavier than helium if this were the case (Searle and Sargent, 1972). Other examples of metal poor blue dwarf systems have been identified since. Neugebauer et al . (1976) found that Haro3 (= M35) and M59 are underabundant in oxygen by a factor of ~2 and Haro4 (= M35) is underabundant by a factor of ~4, with respect to the solar neighborhood. Five additional metal poor systems are found in a study by Ulrich (1971). Alloin et al. (1978) analyzed spectroscopic data for these five systems and six others from Sargent and Searle (1970) and Neugebauer et al . (1976). All 11 of these systems are underabundant in oxygen, neon and nitrogen. They find oxygen to be underabundant by factor of ~2 for M19, -3 for Haro3, -10 for Haro4, -6 for IIZw40, ~7 for Ml 62, -11 for Ml 93, ~6 for Ml 56, -45 for Ml 71, -40 for IZwl8, ~5 for Ml 08 and -5 for M59. Searle et al . (1973) consider the nature of the bluest galaxies known--the isolated, metal poor, dwarf blue compacts. These systems also appear to have a large fraction of their total mass in the form of HI gas (Searle and Sargent, 1972; Gottesman and Weliachew, 1972). Their sizes and their stellar and gaseous content are similar to giant HII regions seen in the spiral arms of giant Sc galaxies. These systems, which all seem to be bluer than (B-V) = +0.3, are explainable by two possible hypotheses, according to Searle et al . (1973). Either they are 9 10 younger than 10 years, or they are of normal ages (-10 years) and have a rate of star formation at this current epoch which greatly exceeds their past average rate. Model calculations by Searle et al . (1973) indicate that a Q galaxy -2 x 10 years old with a Salpeter initial luminosity function

PAGE 20

13 and a uniform rate of star formation would have the colors of the bluest dwarf compact galaxies. If these systems are young, and if their rate of formation has not increased with time, then there should g exist at least 5 times as many 10 years old and at least 50 times as many 10 years old. The models by Searle et al . (1973) indicate that such a galaxy would brighten by -0.5 magnitudes in aging from 2 X 10 years to 10 years and by -1.2 magnitudes in aging to 10 years. Thus old dwarf galaxies should be observationally more abundant than young, Mery blue dwarf galaxies by a factor of at least 50 to 100. As stated earlier, at M = -14, 1 galaxy in 10 appears to be a '>jery blue dwarf compact (Sargent, 1972). This high an abundance argues against the young galaxy hypothesis. In the second hypothesis, called the flashing galaxy or star burst model, star formation occurs in brief intense bursts. The blue dwarf compacts are said to be undergoing a burst of star formation equal to s times their star formation rate averaged over their total lifetime (Searle et al . , 1973). About 1/4 of the light from a galaxy 10^° years o old comes from stars no more than 10 years old. Thus a flash of star formation of strength s = 4 or more would have profound effects on the colors of a galaxy. The colors of the bluest galaxies would require flashes of strength s = 10 20 to produce. Such a flash would increase Q the luminosity by 1 2 magnitudes and would require ~3 x 10 years for the color and luminosity to return to normal. Searle et al . (1973) propose that the blue dwarfs have ages of -10 years and have undergone perhaps 5-10 brief bursts of rapid star formation, each lasting Q -10 years. Nonflashing galaxies of this type should have a space density -2-5 times those of the Mery blue dwarfs. Flashing galaxies

PAGE 21

14 will be brightest during and just after a flash, so at a given absolute magnitude, the ratio of interflashing to flashing galaxies will be fairly low, perhaps as low as 1. Searle et al . (1973) conclude that extremely blue galaxies of high luminosity must be very rare. They consider a galaxy as composed of statistically independent cells in which star formation is either not occurring or is occurring in a burst. They use a cell size approximately equal to the smallest blue dwarfs, or M = -14. Thus a galaxy of M = -19 has 100 such cells. They calculate the probability of at least half of the cells in a galaxy undergoing flashes at any given time. These probabilities are 0.16 at M = -15, .02 at M = -16 and 10 at M = 17. Thus extremely blue galaxies brighter than '1 = -17 should be extremely rare, according to this model. Nondwarf blue compact systems have not received as much attention as the blue dwarf systems. These galaxies are similar in size and luminosity to late type Hubble sequence galaxies, but they do not show any obvious spiral structure. Some of these systems do show hints of spiral structure on direct plates more sensitive and with a greater plate scale than the Palomar Sky Survey (Fairall, 1978). O'Connell and Kraft (1972) obtained a rotation curve of IZwl29, a blue luminous compact galaxy with a bright central core and a faint filamentary structure. Its absolute visual magnitude is -21.1. They find clear evidence for rotation and calculate the total mass interior to the last point of their rotation curve. The mass determined for this system is quite low compared to its luminosity. O'Connell (1979) studied nine nondwarf blue compacts, all brighter than M = -19. Slit spectra were taken of these systems. Some

PAGE 22

15 internal structure was seen in their HII distribution, such as evidence of ring-like or of one-armed spiral structure. Good rotation curves were obtained for six of these systems. The remaining three showed no evidence of rotation. Two of these were face-on while the third, IVZwl53, was a double system and needs more study. The rotation curves for the six systems were found to be consistent with circular motion. These rotation curves resemble those of Sc and Sd galaxies. Unfortunately, only one can be traced to a velocity turnover point. The total mass-to-luminosity ratios for these six systems averages to -1.2 (scaled to a Hubble constant of 75) within the regions observed. This is about 1/3 of the values typically found in the interiors of late type spirals. Neutral hydrogen studies of blue compact galaxies have generally been quite successful. Bottinelli et al . (1973) detected at least nine Haro galaxies in HI with the Nancay radio telescope. The HI masses found were large compared to their total luminosities and estimated total masses. Markarian galaxies also were found to be rich in HI by Bottinelli et al . (1975). Blue Zwicky compacts showing emission lines were shown to HI rich as well (Chamaraux, 1977). Gottesman and Weliachew (1972) made a low resolution, aperture synthesis study of the HI in IIZw40. It was found to be very extensive in HI. This object has an HI core approximately equal in size to its optical halo and an HI halo approximately six times larger than this. Approximately half of its HI mass is in this HI halo. They found nonconclusive evidence of rotation and were able to estimate a lower limit to the total mass. The total mass is found to be at least twice the HI mass (adjusted to a Hubble constant of 75).

PAGE 23

CHAPTER III THE OBSERVATIONS Observations of the 21 cm line of neutral atomic hydrogen were made using the 91 meter (300 foot) radio telescope of the National Radio Astronomy Observatory (NRAO) in Green Bank, West Virginia, and the 305 meter (1000 foot) radio telescope of Arecibo Observatory, near Arecibo, Puerto Rico. The NRAO is operated by Associated Universities, Inc., under contract with the National Science Foundation. The Arecibo Observatory is part of the National Astronomy and Ionosphere Center, which is operated by Cornell University under contract with the National Science Foundation. A discussion of the 21 cm emission line and how the observations were made follows in this chapter. The 21 cm Line A discussion of the nature of the 21 cm emission line is given by Kerr (1968). The ground state of atomic hydrogen is split into two hyperfine sublevels, the upper state being metastable. In the lower energy state, the magnetic dipole moments, or spins, of the proton and the electron d,re in opposite directions, or antiparal lei . In the higher energy state, the two spins are parallel. The difference in energy between these two states corresponds to the energy of a photon of wavelength 21.11 cm and frequency 1420.4058 MHz. The emission or absorption of a 21 cm photon is produced by a transition between these two energy levels, often called a spin flip transition. 16

PAGE 24

17 For an HI region in thermodynamic equilibrium, the relative population of the lower and upper levels, ng and n-i , respectively, is given by the Boltzmann equation. where k is Boltzmann's constant, h is Planck's constant, T is known as the spin temperature and v is the frequency, 1420.4058 MHz. The statistical weights of the lower and the upper energy levels, g^, and g., , are given by g. = 2F. + 1, where Fis the sum of the spin of the electron (±1/2) and the proton (+1/2). Thus Fq = 0, F, = 1 , g^, = 1 and g-j = 3. The term (hv/k) = 0.07 K and will always be much less than the spin temperature. Thus, exp(-hv/kT2) will always be close to unity. For example, at T^ 10 K, n^/n^ 2.9806; at T^ = 100 K, n-|/nQ = 2.9981; and at T^ 1000 K, n,/n = 2.9998. Thus about 3/4 of the HI atoms in a galaxy's interstellar medium will be in the upper ground state at all times. This means that it is not necessary to know the spin temperature of an HI region in order to determine the number of neutral hydrogen atoms present. The Einstein coefficient, Ao-, , for the 21 cm spin flip transition -15 -1 is 2.85 X 10 sec . Thus the radiative lifetime of the upper level is \ery long, approximately 1.1 x 10 years. In the interstellar medium of a galaxy, the time between collisions will be much shorter than this, so that most transitions will be due to collisions. A typical HI atom will move up and down between the two levels about every 400 years due to collisions, but only about ewery 11 million years will it undergo a spontaneous downward transition with the

PAGE 25

1! emission of a 21 cm photon. If collisions are dominant in populating the two levels, then the spin temperature will effectively be equal to the kinetic temperature. The equation of transfer can be solved for the brightness temperature, Tn, of an HI region. One gets Tg = T^ (1 e"") , III-2 where x is the optical depth of the HI region. It is generally believed that the optical depth in HI regions is small. Thus equation III-2 reduces to Tg^xT^. III-3 Assuming small optical depth, the total number, Nm, of hydrogen atoms per cm along the line of sight is N^ = 3.88 X 10^^ /_„ """b ' ^"^ ' ^^^'"^ where v is in KHz. However, HI spectra are normally plotted as a function of velocity, v, in km/sec. The number of hydrogen atoms per 2 cm along the line of sight is then N,^ 1.823 X 10^^ /_ Tg • dv . III-5 Flux, rather than brightness temperature, is the observed quantity in radio astronomy observations. Equation III-5 can be reduced to the total neutral hydrogen mass by

PAGE 26

19 M„, = 2.356 X 10^ • D^ / S • dv, III-6 Hi -^ where M^y is the HI mass in units of solar masses, S is the flux in Janskys, and D is the distance to the HI region in Mpc. The integral need only be evaluated over the velocity range of the HI region and will simply be the area under the HI profile, if it is plotted in the usual units of Janskys versus km/sec. 21 cm Spectral Line Observations For the type of radio telescope considered here, the desired signal from an astronomical source is focused by a reflector onto the front end of a radio frequency (RF) receiver, known as the antenna, or feed, mounted at the focus of the reflector. For 21 cm observations, the receiver is designed to receive both orthogonal polarizations, and to have its maximum sensitivity at or near the expected Doppler frequency of the spectral line. The RF signal from the feed is combined with a local oscillator (LO) frequency to produce a lower, intermediate frequency (IF) signal. In spectral line observations, the LO must be computer controlled to correct for the constantly changing motion of the earth. The IF power is passed into a baseband mixer and then into an autocorrelator which produces the spectrum. An autocorrelator works in arbitrary units. In order to determine the total amplitude of the signal, the system temperature, T must be determined independently of the autocorrelator. This is done by monitoring the total IF power and comparing it to a source of known noise temperature. The power from the noise source, of noise temperature T[u, is periodically injected into the receiver. Continuum

PAGE 27

20 receivers measure the power when the noise source is on, P , and on when it is off, P^^The system temperature is given by T = on off . T TTT 7 '^' ^(Pon Poff^ ^' The total bandwidth B is fed into the autocorrelator, which uses N channels per polarization. Each channel has a time delay of n • At, where n is the channel number (from 1 to N) and At ^ 1/2 B. The autocorrelator computes the autocorrelation function, ACF(n • At), of the spectral bandpass. The spectral bandpass can be considered to be a function of time, y(t). Then the autocorrelation function is T f 2 ACF(n.At) = yi'l i J_^ y(t) .y(t + n.At) . III-8 Each channel of the autocorrelator gives one point of the autocorrelation function. The power spectrum is obtained by taking the Fourier transform of this autocorrelation function, usually by an on-line computer. The result is an N channel spectra of total bandwidth B. The channels will be spaced B/N apart and each channel will have a half power width, or resolution, of Af = 1.21 B/N (Shalloway et al., 1968). This spectrum is not in a readily usable form. The channel values are scaled arbitrarily; the bandpass is not smooth or flat; and the bandpass is usually dominated by the system noise temperature. For these reasons, 21 cm spectral line observations of external galaxies are usually made in what is called the "total power" mode. In this scheme, an "ON" spectrum is taken at the position of the source, and an "OFF" spectrum is taken at a blank sky position. The OFF spectrum will be the

PAGE 28

21 receiver passband function, while the ON spectrum will be the receiver passband function plus the source signal. Let ON(i) and OFF(i) be the relative intensities in the i^h channel of the ON and the OFF spectra, respectively. Then the antenna temperature, T(i), of the source in channel i is ,.. _ ON(i) OFF(i) . ^ '^'^ " OFFTT) sys^^^-9 In pactice, it is necessary to get a good match between the ON and the OFF spectra. This is usually accomplished by using the same integration time and center frequency and by following the same range of altitude and azimuth in the sky so that any instrumental effects will be repeated in both the ON and the OFF spectra. The two should also be taken closely together in time to avoid slow changes in the gain of the receiver. The system sensitivity will be the RMS noise temperature, aT , given by AT = _^ys_ Ill-lO where t is the total integration time and y is a constant (-1) which depends on the system and how it is operated. The RMS flux is given by 2k • AT , AS r^^ , III-ll rms A e

PAGE 29

22 where A^ is the effective area of the reflecting dish, typically 50-60% of its geometric area. In practice, the minimum detectable flux is about three times the RMS flux. Observations on the 91 Meter Telescope The 91 meter (300 foot) diameter telescope of the NRAO is located in the mountains of West Virginia, near the town of Green Bank. Its parabolic reflecting surface is an aluminum mesh which is effectively a solid reflector at radio wavelengths down to about 6 cm. Being solely a transit telescope, it points always at the meridian and is movable in declination between the North Celestial Pole and -19° declination. The front end of the 21 cm receiver is located at the focus on a travelling mount which allows the feed to be moved up to 1/2 degree to the east and to the west of the meridian. This allows a transiting source to be tracked for 4 • sec 5 minutes (6 = declination). The orthogonal linear polarizations are fed into cooled parametric amplifier receivers, giving system temperatures of typically 50 K in each polarization. A bandwidth of 10 MHz was used for all the observations. The spectra were produced with the NRAO Mark III autocorrelation receiver (described by Shalloway et al . , 1968). This is a one bit digital machine containing 384 channels. Thus 192 channels were used for each polarization. The channel spacing was approximately 11.1 km/sec and the channel resolution was approximately 13.4 km/sec. Employing this system, the constant y in equation III-IO is approximately 1.6. For each source, an OFF scan was usually taken first, selected to end 70 seconds before the ON scan was to begin. This allowed time to

PAGE 30

23 move the feed back to its starting position before beginning the ON scan. Both the ON and the OFF scan were of the same duration and covered the same range in altitude and azimuth. The autocorrelator was set to integrate for 20 seconds. Each of these 20 second spectra is called a "record." All the records taken during a single transit are called a "scan." These individual records were initially recorded on a 7-track magnetic tape. All the records in each scan were daily averaged together by the computer staff in Charlottesville, Virginia, and written onto a 9-track tape. We received 3 days of observing time in March 1977 and 12 days in September 1977. The sources were observed each transit, excluding periods when the telescope was shut down for maintenance and malfunctions. Most of the sources were observed for at least 5 transits and several were observed for as many as 20 transits. For proper calibration of the system, drift scans of 21 continuum sources were taken at various times during gaps in the observing schedule. Only unconfused sources from Bridle et al. (1972) were used. A drift scan consisted of setting the telescope at the meridian and at the declination of the calibration source several minutes before transit and allowing the source to drift through the beam of the telescope. The noise tubes were fired just before and after the source drifted through the beam. The signal due to the noise tubes and the source were recorded as deflections on a strip chart. Since the temperature of the noise tubes and the flux strength of the calibration sources were known, the sensitivity of the system in Jy/K could be found. The results of all drifts in both polarizations were averaged together. The final results for the 21 calibration sources are

PAGE 31

24 presented in Table III-l. Column (1) gives the name of the calibration source. Column (2) gives the 1950 declination. Column (3) gives the sensitivity of the system in Jy/K. Column (4) gives the number of drift scans used. Column (5) is the flux of the calibration source in Janskys at 1400 MHz, from Bridle et al . (1972). The sensitivity was found to be a fairly smooth function of the declination. A parabola has been fit, by the method of least squares, relating the sensitivity of the system S, in Jy/K, to the declination 6, in degrees. This parabola is given by S = 0.99775 0.004408 • 6 + 0.0000749 • 6^ . III-12 The RMS uncertainty is ±0.02 Jy/K, or about 2%. The data points and the parabolic fit are shown in Figure III-l. The reason for this declination dependency is presumably due to deformation of the parabolic reflector. The calibration curve is not centered on the zenith as expected. Greatest sensitivity occurs at approximately +29° declination, whereas the zenith is at approximately +38°. This calibration curve agrees quite well with one obtained for the 91 meter telescope by Fisher and Tully (1975). However, their calibration curve reaches its maximum sensitivity very near the zenith. A small correction was also required for the variation of sensitivity with hour angle. When east or west of the meridian, the parabolic reflector is used off axis, resulting in decreased efficiency. Several continuum sources were tracked and the average flux was found to be approximately 4.5% less than the peak flux at the meridian. Thus, in the final analysis, all spectral intensities were multiplied by 1.045.

PAGE 32

25 Table III-l. Sources Used for Calibrating the 91 Meter Telescope. Source (1) Dec (2) Jy/K (3) Flux (5) P212312

PAGE 33

"3

PAGE 34

27

PAGE 35

28 At the end of each of the two observing sessions, several days were spent reducing the data at the NRAO headquarters in Charlottesville, Virginia. The work was done on a CRT terminal using the NRAO T Power and S Power programs. For each source, the ON and its corresponding OFF were differenced according to equation III-9, and then all transits were averaged. Any bad data, of course, were excluded. For sources in which a spectral line was present, a polynomial curve was fit to the baseline on either side of the spectral line. The lowest order polynomial as was suitable was used, seldom going beyond third order. This curve was then subtracted from the spectrum to leave a flat baseline. The spectra were then recorded on a 9-track magnetic tape and taken back to the University of Florida for further reduction. At the University of Florida, corrections for the variation of system sensitivity with declination and with hour angle were applied, converting the spectral values from units of temperature to units of flux. The two polarizations were averaged together and the spectra were plotted on a Gould electrostatic plotter. Final RMS uncertainties were on the order of 5-10 mJy per channel. These spectra are shown in the next chapter. The computation of various parameters from these spectra is also developed in the next chapter. Observations on the 305 Meter Telescope The 305 meter (1000 foot) telescope of Arecibo Observatory is located in a mountainous region of Puerto Rico, approximately 15 km south of the city of Arecibo. Its reflecting surface is a fixed spherical aluminum dish, 305 meters in diameter and covering 18 acres. A huge platform is suspended above the dish, supported by cables from

PAGE 36

29 three large towers. The feeds are mounted beneath this platform and can be steered to point anywhere within about 20° of the zenith. Spherical aberration is corrected for by using line feeds. The 21 cm feed has a range in declination between about -1° and +38°. The Arecibo observations described here were made in two observing sessions. A 17 day observing run was made in April and May of 1977. As these observations were quite successful, additional time was requested. The second observing session, covering eight days, was made in June and July of 1978. The feed used for the 21 cm observations was a 40 foot line feed which was illuminated by an annulus of approximately 210 meters outer diameter and 90 meters inner diameter. The receivers were uncooled parametric amplifiers which accepted the two circular polarizations from the feed. For zenith angles of 10° or less, the feed was illuminated entirely by the reflecting surface of the telescope. Within this range, the half power beamwidth was -3.3 arc minutes and system temperatures were typically 70-80 K. At zenith angles greater than 10°, the feed began picking up ground radiation, increasing the system temperature. Other side effects were an increase in and a distortion of the beamwidth, and a reduction in system sensitivity. This effect was not too serious inside a 14° zenith angle. However, near the zenith angle limit, the sensitivity dropped to -60?^ of its maximum and the system temperatures rose to -150 K. For these reasons, most of our observations were made within about 14° of the zenith, but a few sources had to be observed at less favorable parts of the sky. One of the biggest problems with this telescope is the uncertainty in the position of the beam. The RMS pointing accuracy is on the order

PAGE 37

30 of 30" of arc. Thus pointing errors of 1 ' of arc are not uncommon. This uncertainty has the effect of limiting the determination of HI masses to an accuracy of -20%. The spectra were produced using 504 channels of a 1008 channel autocorrelator. Each polarization was split into a spectrum of 252 channels. Total bandwidths of either 5 MHz or 10 MHz were used. The channel spacing was approximately 4.2 km/sec with a resolution of approximately 5.1 km/sec at 5 MHz and twice these values at 10 MHz. With a 5 MHz bandwidth, three level sampling of the data could be employed. In this mode, the constant y in equation III-IO was -1.2. With a 10 MHz bandwidth, the autocorrelator was usable only as a one bit sampler. In this case, y was -1.6. For a 5 MHz and a 10 MHz spectrum of identical integration times and smoothed to the same velocity resolution, the RMS uncertainty of the 5 MHz spectrum would be approximately 3/4 as much as the 10 MHz spectrum. Thus it was advantageous to use a 5 MHz bandwidth. All of the Arecibo observations were made in the total power mode. An ON scan of five minutes duration was taken at the source position, followed shortly afterwards with a five minute OFF scan covering the same range in zenith angle and azimuth. The observations were monitored on-line using a Harris Datacraft computer and a CRT terminal. The Arecibo HI system is undoubtedly the most sensitive in the world. In almost every case where a spectral line was present in the bandpass, it was evident, though perhaps noisy, after the first set of ON and OFF scans. In several cases, this allowed changing the observed velocity to place the spectral line in the center of the bandpass. Most of the observations were made using a 5 MHz bandwidth. A 10 MHz bandwidth was

PAGE 38

31 employed at times to search for several Haro galaxies with unknown velocities and several galaxies which did not show a spectral line at the published optically determined velocities. If a signal was found, the system was usually switched back to a 5 MHz bandwidth. Typically, between three and ten sets of ON and OFF scans were taken for each source to get a good signal to noise ratio and a low RMS uncertainty. The 21 cm feed is tunable in frequency by moving it up or down in its housing. For instance, to get maximum sensitivity at -1390 MHz, the feed is moved 14 inches closer to the reflector than the normal focus position for a frequency of 1415 MHz. Unfortunately, the feed cannot be adjusted individually for each source as it requires several hours for a single adjustment. For the two sets of observations reported here, the feed was set for optimum sensitivity at a frequency of -1415 MHz for about 60% of the observations and at -1390 MHz for about 40% of the observations. The variation of telescope sensitivity with frequency for these two settings of the feed has been determined by the Arecibo staff. Plots of system sensitivity in K/Jy versus frequency were obtained from Dr. Mike Davis, allowing the data to be calibrated and corrected for this effect. Figure III-2 shows the telescope sensitivity versus frequency for the 1415 MHz setting. The peak sensitivity is -8.5 K/Jy. Figure III-3 shows the same information for the 1390 MHz setting. The peak sensitivity is -8.1 K/Jy for this setting. The sensitivity of the system over the velocities observed ranged from -8.5 K/Jy to -6 K/Jy. Zenith angle corrections have also been determined by the Arecibo staff. No correction is required for zenith angles of 10° or less.

PAGE 39

32 ,.. I . .. I ... I OJ

PAGE 40

33 ' ' i_ I cu

PAGE 41

34 For zenith angles greater than this, the relative gain is given by Gain = exp [(-.00521) • (ZA 10)^], III-13 where ZA is the zenith angle in degrees. The noise tubes used for the 21 cm observations had to be calibrated, as their noise temperatures were not accurately known. This was accomplished by making drift scans of continuum sources, as was done on the 91 meter telescope. Sources were taken from Bridle et al. (1972). Owing to pointing errors, a single drift scan was not sufficient for an accurate calibration. Usually drifts were taken at the source declination and at declinations 1' of arc to the north and to the south. The noise tubes were fired just before and after the source drifted through the beam. The drift scans were recorded on a strip chart. With three such drifts, the peak flux and the error in pointing could be obtained. Using 1400 MHz fluxes from Bridle et al. (1972), the values of the noise tubes were found in flux units. These values were then converted to temperature units by multiplying by 8.5 K/Jy at the 1410 MHz setting, and by 8.1 K/Jy at the 1390 MHz setting. The observations taken in 1977 were partially reduced at Arecibo Observatory using several programs written by Nathan Krumm. The individual scans were averaged together and the spectra were punched out on computer cards for further reduction at the University of Florida. For the observations taken in 1978, the data were brought back on magnetic tape and all the reductions were made at the University of Florida. One program written by Martha Haynes and Steven Peterson was used to convert the Arecibo data tape into an IBM readable tape. Several other reduction programs, incorporating elements written by

PAGE 42

35 Nathan Krumm and Martha Haynes, were developed by the author. The spectra were calibrated for the variation of sensitivity due to zenith angle, according to equation III-13, and for the variation of sensitivity with frequency, according to Figures III-2 and III-3. The spectra were plotted on a Gould electrostatic plotter. The final RMS uncertainties were typically 1 -5 mJy per channel. The calculation of various parameters from these spectra is given in the next chapter.

PAGE 43

CHAPTER IV THE DATA Optical Parameters Most of the optical parameters of the 99 compact galaxies observed for this study are given in Table IV-1 . The columns are numbered and are explained below by column number. Following Table IV-1 is a list of notes on the individual objects, giving alternate names and a brief description. Column 1 : Source name. Columns 2 and 3: Right Ascension (RA) and Declination (Dec) of the source in 1950 coordinates. Positions have been measured by the author on the Palomar Sky Survey (hereafter abbreviated PSS) plates using an overlay program and a least squares fitting program. Positions are believed accurate to 2-3 arc seconds. Column 4: V_ , heliocentric velocity. The Doppler shifted velocities with respect to the sun are taken from de Vaucouleurs et al . (1976) unless otherwise noted (S implies Sargent, 1970b). De Vaucouleurs et al. (1976) use weighted means of published velocities, most of which are optically determined. Column 5: The major axis (a") and the minor axis (b") in seconds of arc. These are given in the system of de Vaucouleurs et al. (1976). 36

PAGE 44

37 They are the approximate dimensions to a limiting surface brightness level of the 25 magnitude per square arc second. The author has measured the dimensions of 150 compact and noncompact active galaxies on the blue plates 'of the PSS using a low power microscope. For 138 of these objects whose dimensions are given in de Vaucouleurs et al . (1976), the following regression formulae are found. For 9p33 < 50 : 625 1.4 9p33 10 ± 15 , and for Once ^ 50 : Q^^ = 1.01 Qr^^r + 9 ± 16 , IV-1 where 9p<^c; "is the major or minor axis in arc seconds as measured on the ji PSS, and e^c is the major or minor axis in arc seconds in the system of de Vaucouleurs et al. (1976). Dimensions in column 5 identified with a G are reduced from their PSS dimensions using equation IV-1. The others are taken from de Vaucouleurs et al. (1976). Column 6: Axial ratio, b/a. These are not strictly the ratio of the optical dimensions in column 5 because these dimensions often include irregularities such as double cores, filaments or jets. Since the axial ratio is used to calculate the inclination of the system, such irregularities are ignored in estimating the axial ratio. For multipl-^ or '
PAGE 45

38 rotating disk systems. O'Connell and Kraft (1972) and O'Connell (1979) have shown that some of the blue nondwarf systems appear to rotate with rotation curves similar to late type disk galaxies. Gottesman and Weliachew (1972) found inconclusive evidence of rotation in one dwarf compact galaxy. For compact galaxies, the observed axial ratios are higher, on the average, than is seen in late type spiral galaxies. This is an indication that, if these are disk systems, their intrinsic axial ratios are greater than those of spiral galaxies. This problem will be discussed more fully in Chapter V. For now, we will calculate an inclination assuming that these are disk systems. For a disk of intrinsic axial ratio q, and apparent axial ratio b/a, the inclination i is given by cos2 i = ibZaji^ _ ,,_, 1 q^ For normal spiral galaxies, q is believed to be between 0.2 and 0.25. \lery few spiral galaxies are found with b/a > 0.2. Only one galaxy in our sample has b/a > 0.3. For compact galaxies then, we use q = 0.3 in equation IV-2 to compute the inclinations. This is similar to the value of 0.33 used by Chamaraux (1977) for blue compacts. As in column 6, inclinations are not given for multiple or very peculiar systems. Column 8: B^, blue magnitude, in the system of de Vaucouleurs et al. (1976). Slightly less than half of our sources have By magnitudes listed in de Vaucouleurs et al. (1976). These are listed in column 8 with no further designation. Those marked with an H are taken from Huchra (1977) and those marked with a P are from Huchra (1979). The

PAGE 46

39 remainder, marked with a Z, are reduced from the photographic magnitudes given in Zwicky's Catalogue of Galaxies and of Clusters of Galaxies (Zwicky et al., 1961, 1965; Zwicky and Herzog, 1963, 1966, 1968; Zwicky and Kowal , 1968) to B magnitudes. De Vaucouleurs and Corwin (1979) have determined regression formulae to perform these conversions and have personally calculated these magnitudes and communicated them to the author in advance of publication of the regression techniques. Column 9: B^, the total blue magnitude, in the system of de Vaucouleurs et al. (1976). These are calculated from the By magnitudes of column 8, correcting them to the face-on, zero extinction, zero redshift values, according to the precepts given in de Vaucouleurs et al. (1976). Column 10: M^o, the absolute blue magnitude. These are calculated o ^ using the B^ magnitudes of column 9 and the distances given in either Table IV-2 or Table IV-3. Columns 11 and 12: (B-V), and (U-B)j, color indexes, in the system of de Vaucouleurs et al . (1976). These color indexes are corrected to the face-on zero extinction, zero redshift values, according to the precepts given in de Vaucouleurs et al . (1976). The symbol in column 12 gives the source of the color information. Colors from de Vaucouleurs et al . (1976) are given with no symbol; H implies Huchra (1977); K implies Khachikian and Weedman (1974); S implies Sargent (1970b); D implies Du Puy (1970); and I implies Hiltner and Iriarte (1958).

PAGE 47

40

PAGE 48

41 X —

PAGE 49

42 oh— (Tl lo m c» c? oj n ':^ r tn <* c* o
PAGE 50

43 Notes to Table IV-1 M335 = A0003+19. Seyfert galaxy. Almost stellar; slight trace of nebulosity; b/a rather uncertain. Harol4 = MGC-3-3-3. Has large bright core not centered in nebulosity; bright knot in ENE part of nebulosity (NGC244). IIIZwlZ = M347 = IC1586. Nearly stellar core; hint of nebulosity. HarolB = iM960 = MGC-2-3-19 = A0046-12. Large bright core with nebulosity. M352 = A0057+31 . Seyfert. Small, nearly circular core; slight amount of nebulosity; weak jet to W. Ml = NGC449 = MGC5-4-9. Seyfert. Small, bright elliptical core with a slight amount of nebulosity. Correctly identified in de Vaucouleurs, et al. (1975), but incorrectly in Nilson (1973). IIIZw33 M360 = MGC3-5-13 = A0141+15. Bright elongated core; two bright knots N of core imbedded in nebulosity; somewhat irregular, nonuniform nebulosity. May be a distorted Scd (Huchra, 1977). IIIZw35. Double compact system; in contact. Southern compact has circular core; little nebulosity. Northern compact has bright elliptical core; little nebulosity. VZwl55 = M364 = AOl 54+27. Featureless; small circular core; very little nebulosity. IIIZw42 = M366 A0208+13. Bright, nearly stellar, slightly elongated core; clumpy irregular nebulosity not centered on core. IIIZw43 M589 MGCl-6-55 = UGC1716 A0211+03. Bright, nearly stellar core; faint nebulosity. Haro20 = MGC-3-9-45 = A0325-17. Small, uniform, elliptical core; faint nebulosity. VZw372 = MGC5-10-12 = A0410+29 = UGC2989(?). Bright, stellar nucleus; faint, grainy and extensive nebulosity. IIZwlS = A0435+11. Round, bright, stellar core; some nebulosity; jet to SE. May have faint outer spiral arms (Sargent, 1970b). IIZw23 = M1087(?) = A0447+C3 = UGC3179. Elliptical core with faint jets; faint nebulosity. IIZw28 = A0459+03. Round; slightly irregular; clumpy; little nebulosity. Has a ring (Sargent, 1970b).

PAGE 51

44 IIZw33 = M1094 = MGCO-14-10 = A0508-02. Elliptical, patchy, compact core; perhaps composed of three clumps; irregular, patchy nebulosity. IIZw35. Two stellar knots; very faint nebulosity. IIZw40 == A0553+03. Approximately stellar core; very faint, grainy nebulosity; two faint jets to SE; b/a very uncertain. IIZw42 = MGCl-16-0 = A0600+07 = UGC3393. Bright stellar core; little nebulosity. M6 = IC450 = MGC12-7-18 = UGC3547. Seyfert. Small, bright, elliptical core; faint, fuzzy, irregular nebulosity. VIIZwl53 = M7 MGC12-7-38 = A0722+72 = UGC3838. Irregular, "tank" shaped, compact core; small amount of nebulosity; b/a very uncertain. U shaped core of knots (Huchra, 1977). VIIZwl56 M8 IC2184 = MGC12-7-41 = UGC3852. Triangular shaped clumpy core; triangular shaped nebulosity; b/a uncertain. Ring of HII regions (Huchra, 1977). Harol NGC2415 = MGC6-17-21 = UGC3930. Slightly ellipsoidal, irregular core; irregular nebulosity. M385 = A0800+25. Bright, approximately stellar core; small amount of nebulosity. M622 A0804+39 = UGC4229. Small, bright, almost circular core; fairly dense, featureless nebulosity. M390 = M6C5-20-28 = AG832+30. Bright, ellipsoidal, compact core; faint nebulosity. Zw0855 = MGCl-23-13 A0855+06 = UGC4703. Unnamed object in Zwicky (1971). Two tiny clumps, -87" apart; connected by a faint, thin bridge; very strange object. M105 = A0915+17. Small bright core; trace of nebulosity. IZwl8 = M116 = A0930 A and B. Pair of interconnected compacts; figure 8 shaped core with some nebulosity. M402 = A0932+30. Elliptical core; very little nebulosity. IZw21 = MGC8-18-30 = A0943+46 = UGC5225. Circular, nearly stellar core; very faint nebulosity. Haro22 = MGC5-23-40 = A0947+28. Bright, slightly elongated core; small amount of nebulosity. Haro23 = MGC5-24-11 = A1003+29. Small, ellipsoidal core, slight amount of nebulosity.

PAGE 52

45 IIZw44 = A1012+21. Small, semistel lar core; main object is only -20" in diameter; several possible nebulous objects within ~45" are used for the total dimensions. Haro2 = M33= Arp233 MGC9-17-70 = UGC5720 = Al 029+54. Featureless, ellipsoidal core and nebulosity. M148 = A1032+44 = UGC5747. Bright irregular core; jets to SE and NW; slight amount of nebulosity. Haro3 = M35 = NGC3353 MGC9-18-22 = UGC5860. Small bright core; faint irregular nebulosity. Haro25 = M727 = MGC4-26-9 Al 046+26. Stellar core; very slight trace of nebulosity. DuPuy (1970) misidentif ied Haro25 with a red object. Hiltner and Iriarte (1958) looked at the correct object. Haro4 = M36 = MGC5-26-46 = All 02+29. Small, slightly elongated core; small amount of nebulosity. IZw26 = M40 = Arpl51 = VV144 = MGC9-19-73 = All 22+54. Seyfert. Thin filament with two knots in it; wery irregular object. Ml 69 IC691 = MGClO-16-139 = UGC6447. Slightly elongated core; some nebulosity; faint jet to S. Haro27 = MGC5-28-10 = All 37+28 = UGC6637. Bright elongated core with some nebulosity. M198 = MGC8-22-73 A1206+47. Bright, nearly stellar core; elliptical, slightly irregular nebulosity. IIZw57 MGC3-31-49 A1207+17. Bright elongated core; clumpy nebulosity. Haro28 = NGC4218 = MGC8-22-88 = UGC7283. Large, bright, slightly irregular, elliptical core; small amount of nebulosity. Haro8 M49 MGCl-31-50 A1216+04 = UGC7354. Stellar core; slight amount of nebulosity, mostly to SE. M50 = A1220+02. Seyfert. Bright, nearly stellar core; slight trace of nebulosity. Haro29 = IZw36 = M209 = MGC8-23-35 = A1223+48. Strange appearance; fan shaped nebulosity; bright compact core at western tip and compact core in center; small bright object or jet to N; b/a very uncertain. M215 = MGC8-23-52 = A1230+46. Small, bright, slightly elongated core; small amount of nebulosity. Haro32 = IZw41 = M220 and M221 = MGC9-21-33 and 34 = A1241+55 A and B = UGC7905. Close compact pair; very irregular. Southern component has

PAGE 53

46 bright core; slight amount of nebulosity. Northern component is very irregular; bright elongated core; extensive curving nebulous region to N and E. Haro33 = MGC5-30-70 = Al 242+28. Bright central knot; smaller knots to ESE and WNW, interconnected with nebulosity. Haro34 = IC3730. Bright, approximately stellar core; irregular nebulosity; curving filament to N, like a single spiral arm. Haro36 = MGC9-21-47 = UGC7950 = A1244+51. Bright central core; weak, somewhat irregular nebulosity. Haro35. Bright uniform core, shaped like a "combat" hat; virtually no nebulosity. Haro37 = M444 = iMGC6-28-32 = Al 246+34. Approximately stellar appearance; faint haze; hint of faint jets. IIIZw68 = M6C5-31-38 = A1255+27B = UGC8080. Bright, nearly circular core; uniform ell iptical nebulosity. IIZw67 = NGC4853 = MGC5-31-48 = UGC8092. Slightly elongated core; some nebulosity. M57 = A1256+27B. Small, elongated, slightly irregular core; '^ery little nebulosity; faint jet to S. M235 = MGC6-29-10 = A1257+33. Small elongated core, pointed at one end; little nebulosity. M241 = Al 303+33. Bright elongated core; very little nebulosity. IZw53 = A1311+35. Nearly stellar core; slightly elongated; virtually no nebulosity. IZw56 = IC883 = Arpl93 = MGC6-29-0 = UGC8387. Bright elongated core; some nebulosity; two jets to SE and to SW. Haro38 = MGC5-32-41 = Al 333+29 = UGC8578. Bright elongated core with some nebulosity. M270 = NGC5283 = MGCll-17-7 = UGC8672(?). Seyfert. Bright, nearly stellar core; faint nebulosity. M275 ^ MGC5-33-2 A1346+31. Elongated, somewhat irregular core; nebulosity with wispy features, perhaps hint of spiral structure. Haro39. Bright, very elongated, cigar shaped core; ^ery little nebulosity. Haro42 = M685 = MGC5-34-61 = Al 428+27. Elongated core; faint nebulosity; filaments to east.

PAGE 54

47 Haro43 = MGC5-34-80. Very elongated bright core; very little nebulosity. Haro44 = MGC5-35-8. Irregular, tear shaped core; small amount of nebulosity; possibly a jet to N. IIZw70 = M829 MGC6-33-2 = VV324B = Al 448+35 = UGC9560. Pair with IIZw71, -250" apart; slightly elongated, nearly stellar core; nebulosity mostly along major axis, like two streamers. IIZw71 MGC5-33-4 = VV324A = A1449+35 = UGC9562. Elongated with central bulge; faint halo; perhaps an edge on galaxy. IZw97 = A1452+42. Elongated core; very little nebulosity. IZw98 = NGC5787 = MGC7-31-8 UGC9599. Bright elongated core; extensive elliptical nebulosity; axes of the core and of the nebulosity differ by -45°. IZwlOl = IC1090 = MGC7-31-25. Small elongated core; slight amount of nebulosity. IZwll5 = MGC8-28-38 = A1531+46 = U6C9893. Bright, crescent shaped core embedded in elongated clumpy nebulosity. IZwllZ = MGC7-32-30 and 31 = Al 534+38 UGC9922. Elongated core; compact knot at northern tip; slight nebulosity; b/a rather uncertain. IZwl23 = M487 = A1535+55. Round, almost stellar core; slight amount of nebulosity. VIIZw631 = Seyfert's sextet = NGC6027 and companions = VV115 = MGC4-38-5,6,7,8,9 and 10 = UGC10116. Compact group of six compact galaxies; in a tight triangular cluster. M297 = NGC6052 Arp209 = VV86 MGC4-38-22 = UGC10182. Irregular object; two bright knots side by side; "hat" shaped appearance; slight amount of nebulosity; b/a vevy uncertain. IZwl47 = A1622+54. Approximately stellar with fuzzy edges. IZwl59 = A1634+52. Small, nearly stellar core; little nebulosity. IZwl66 = M499 = A1647+48A. Nearly stellar core; small amount of nebulosity. IZwl91 = A1739+47. Approximately stellar, fuzzy edges. IZwl99 MGC9-29-39 and 40 = A1749+56 A and B. Small, interconnected pair of compacts; -20" apart; small amount of nebulosity. IZw207 = A1830+55. Seems very blue; arc shaped clumpy object; "boomerang" shaped; 'very irregular.

PAGE 55

48 IIZw82 = IC1317 = MGCO-52-4 UGC11546. Elongated core; faint nebulosity. IVZw67 = MGC6-46-0(?) = UGC11668. Elongated core, perhaps double or with a star at one tip; little nebulosity; possibly several faint jets. IIZwl68 = A2207+17. Nearly stellar, just slightly oblate core; slight trace of nebulosity. IIZwl72 = NGC7236 and 7237 = Arpl69 = MGC2-56-23 and 24 = UGC11958. Three small round compacts in a row; surrounded by nebulosity. IVZw93 = A2213+22. Triangular shaped nucleus; perhaps made up of three stellar or semistellar knots; surrounded by faint nebulosity. M303 = NGC7244 = MGC3-56-21 . Elongated core; clumpy nebulosity. Zw2220 = A2220+30 A and B = UGC120n. Unnamed object in Zwicky (1971). Interconnected pair of compacts; 16" apart. Western component has a small elongated core; small amount of nebulosity. Eastern component has a smaller elongated core; jet and small amount of nebulosity. IIZW185 = IC5243 = MGC4-53-n = UGC12153. Irregular; mostly dense core; little nebulosity; curving tail, or perhaps single spiral arm; b/a rather uncertain. M314 = NGC7468 = MGC3-58-0 = UGC12329. Oblate, tear shaped core; some nebulosity; faint thin jets extending from each side of major axis. IVZwl42 = M322 = A2317+25. Bright elongated core; fuzzy edges. IVZwl49 = NGC7573 = M325 = M6C4-55-14 = UGC125G7. Large bright core, pointed at southern end; small amount of nebulosity; short jet to N. IVZwl53 = A2327+25. Double core; small amount of nebulosity. Zw2335 = M328 = A2335+29. Bright, tear shaped core; small amount of nebulosity. Abbreviations used: MGC: designation in the Morphological General Catalogue (VorontsouVelyaminov et al., 1962). UGC: designation in the Uppsala General Catalogue of Galaxies (Nil son, 1973). A: designation for sources without N6C or IC numbers in de Vaucouleurs et al. (1976). M: Markarian number.

PAGE 56

49 Arp: designation in the Atlas of Peculiar Galaxies (Arp, 1966). VV: designation in the Atlas and Catalogue of Interacting Galaxies (Vorontsov-Velyaminov, 1958). north south east west

PAGE 57

50 21 cm Global Parameters Table IV-2 lists the global properties of the 72 systems which were detected in HI. The columns are explained by column number and line number. Column 1, line 1: Source name. Column 1, line 2: Designation for the telescope used. Thus N implies the 91 meter NRAO telescope; A implies the 305 meter Arecibo Observatory telescope; AN implies both telescopes. Column 2, line 1: V^i , the heliocentric velocity, in km/sec, measured from the 21 cm spectrum. The velocity centroid is taken as the velocity of the system. This is defined by max E V(n) • S(n) ,, n = min Vol = , IV-3 21 max ' '^ -^ E S(n) n min where V(n) is the velocity of the center of the n**^ channel and S(n) is the flux of the n channel. The summations are carried out from the minimum channel to the maximum channels containing the spectral line. Column 2, line 2: "SV^i , the uncertainty in V^-, , in km/sec. This is found by differentiating equation IV-3, which gives max (Z V(n)) N. V„, 6V„, "^-^^ . AS , IV-4 21 max rms ' ^ S(n) n = min

PAGE 58

51 where N is the number of channels containing the spectral line and AS is the RMS uncertainty of the baseline, calculated from the channels on each side of the spectral line. Values of 5V„, are rounded up to the nearest integer. Equation IV-4 is strictly correct only if the uncertainties are due entirely to random noise. But in practice, errors in the pointing of the telescope and in fitting a baseline will further limit the accuracy of Vp-,. Column 3, line 1: D, distance to the source in Mpc. The heliocentric velocities, M^-,, are corrected to the velocity seen from the center of the galaxy, V„„, by ^GC " ^21 * ^^° * ^''" L • cos B , IV-5 where L and B are the galactic longitude and latitude of the source. The distance D is then found by dividing Vpp by a Hubble constant of 75 km/sec/Mpc. Column 3, line 2: A, the linear size of the major axis converted to units of kpc. These are calculated using the distances from column 3, line 1 and the major axis, a", given in Table IV-1. Column 4: aV, the width of the 21 cm profile in km/sec. This velocity width is measured at the points equal to 20% of the average flux after boxcar smoothing over three channels, or at the Za level after boxcar smoothing over three channels if the average flux is too weak. The RMS uncertainty, a, is calculated after smoothing over three channels. The velocity widths listed are corrected for the boxcar smoothing.

PAGE 59

52 Column 5, line 1: BCF, beam correction factor, applied to the HI masses to correct for the finite width of the beam. A radio telescope will underestimate the total flux from a radio source, unless it is a point source centered in the beamwidth. For extended sources, some way of correcting for this underestimation is necessary. The 91 meter telescope has a 21 cm beamwidth of approximately lo!8 of arc at half power points. The shape of the beam is approximately Gaussian, except near the first null. All of the sources observed with the 91 meter telescope were much smaller than the beamwidth and so this deviation from Gaussian shape need not be taken into account. The flux detected will be the convolution of the source flux distribution and the beam pattern. If the HI surface density of the source is distributed as an elliptical Gaussian with major and minor axes half power dimensions of A and B arc minutes, then the ratio of the total flux to the detected flux (from Fisher and Tully, 1975) will be o V2 1/2 BCF = (1 + (A/10. 8)^) • (1 + (B/10.8)^) . IV-6 Unfortunately, the values of A and B are not known. For simplicity, we assume that they are some multiple of the optical dimensions, a" and b". Converting to minutes of arc, we have A = F • a"/60 and B = F* b"/60, where F is this unknown multiple. For spiral galaxies, the HI distribution is typically found to be somewhat greater than the optical distribution. Thus we expect F to be greater than unity. The 305 meter telescope has a half power beamwidth of approximately 3.3 of arc at 21 cm. This is not small compared to the dimensions of many of our sources. Additionally, the departure of the beam

PAGE 60

53 pattern from a Gaussian is significant enough that an equation of the form of equation IV-6 is not sufficiently accurate. The sidelobes of the beam may also contribute to the detected flux and need to be considered as well. A numerical model of the Arecibo beam, including its first two sidelobes, was constructed. The source size and distribution is considered as in the previous case. The beam pattern and the source shape are then convolved numerically to determine a beam correction factor. Due to the pointing errors of the 305 meter telescope, this convolution also includes the effects of a 30" of arc pointing error. Again the factor F is unknown. However, several sources were observed successfully with both telescopes. In the analysis of the HI masses of these sources, a value of F was found that led to the same corrected HI masses on both telescopes. The values of F found for 13 sources ranged from 1.0 up to 5.2 with an average of 2.4393. Thus for the other sources, a value for F of 2,4 is assumed, with a few exceptions, all of those being multiple galaxies. Column 5, line 2: F, the ratio of optical to HI dimension assumed for the calculation of the beam correction factor of column 5, line 1. Column 6, line 1: M , the HI mass expressed in solar mass units and HI corrected for the beamwidth effect. The uncorrected HI mass is found by converting equation III-6 to a summation of the form r o max M^j = 2.356 X 10 • D . z S(v) • Av , IV-7 n min where the summation is carried out from the minimum to the maximum

PAGE 61

54 channel numbers containing the spectral line, and Av is the channel spacing in km/sec. The corrected HI mass is then found by multiplying by BCF of column 5. Column 6, line 2: t\H^r, the uncertainty in the HI mass. This is found by differentiating the equation for corrected HI mass. One gets AM^j = 2.356 X 10^ • D^ • /N . AV . aS^^^ • BCF. IV-8 Column 7: Lpo, total blue luminosity in solar units. These are ^T calculated from the absolute magnitudes. By given in Table IV-7 and assuming the absolute B magnitude of the sun to be +5.41 (Allen, 1964). Column 8: M-, indicative total mass, in solar units. We use the equation used by Shostak (1978) for late type galaxies: M. = 2.45 X 10^ • A(AV/sin i)^ . IV-9 Total masses will be discussed in more detail in Chapter V. Column 9: Muj/L, column 6 divided by column 7. Column 10: M./L, column 8 divided by column 7. Column 11: Mut/M,-, column 6 divided by column 8. Hi I Table IV-3 lists a limited amount of data for 27 systems which were not detected in HI. The columns are explained by number. Column 1 : Source name.

PAGE 62

55 n

PAGE 63

56 .— o -H lo m n ^ r "^ ^ -I r: -< r> ^ o o o a-' ,-N l^ -< ^ O ^ ^ O nI^ O '^ o ") 'T — l.";;l. lUi :LI:.l!Jci' fv-^ -< -H jT ,< f«-i --i' if C •jT »'XJ 'Tl •C o c o -r T h^ N ^T f^ I " T "^ T C ~' ' c c o o c c c o c c c e o _' u 11 i.. J. a, . '.'.:uu. l; l; "ji-M'. " -l; r' ^ ^^ o •^'' r T" N:^ r. r^ T ": o c C '-'"' '-'^ T O ^^O "^ O ^'. O -< O ^ O V o " -^ Jo rw r o ^ o rn o K r; " o -o a ' i • T • o • o • — • -< cv; .^ »r: . <} »n .r. .-. .r: • r^ .r: • I "^ • n c7 o;o^-'?'^C) <; -' o o rj (T -• c -J t: ro 01 m ; 0CM0-<--C---^ «5'^' OJ CMCVJ oooA.o.n-^JovccoJAJoJ-o^--+or.|':>J^o
PAGE 64

57 TON cr N vC o c c ^ o O 'O v^ f^ •-* N C ^ IT: (.O ri CD "^ :* O ^ J^ r^ {\, -^ 1—1 (Tl C 'l" N

PAGE 65

58 o ^ —

PAGE 66

59 Notes to Table IV-2 VZw372: The spectra contains a strong source at V = 5800 km/sec in the sidelobe of the 305 meter telescope. Zw0855: The spectra shows two emission features. It is uncertain whether both features are from this double object. IIZw44: There is a difference of ~4 times in the uncorrected HI masses determined with the 91 meter and the 305 meter telescope. This could be due to a giant HI halo. This object needs more careful study. Haro22 and Haro39: No optical velocities were known. Their HI emission lines were found by searching over a large range of velocities with the 91 meter telescope. Haro35, Haro43 and Haro44: No optical velocities were known. Their HI emission lines were found by searching over a large range of velocities with the 305 meter telescope. IZwl66: Confused in HI with two or three nearby galaxies. The HI mass is for all these sources. It is not known which feature is due to IZwl65. This mass is not used in the analysis in the next chapter. VIIZw631: Unfortunately, a 10 MHz bandwidth could not be used at the time of observation, due to autocorrelator problems. There could be emission outside of our 5 MHz bandwidth from this group of six galaxies, which would have been missed. This HI mass is not used in the analysis in the next chapter.

PAGE 67

60 Column 2: D, distance to the source in Mpc. These are calculated in the same manner as in Table IV-2, except that the heliocentric velocities are taken from the optical velocities of Table IV-1. Column 3: A, major axis in kpc, calculated the same way as in Table IV-2. Column 4: Ml,. , upper limit to the HI mass. This is found by "^max taking three times the uncertainty defined by equation IV-S, and assuming the velocity width is no greater than 400 km/sec. An N or an A indicates which telescope was used, as in Table IV-2. Column 5: 1^° > total blue luminosity, in solar units, calculated as ^T in Table IV-2. Spectra of the detected systems are presented in Figure IV-1. The source name is given in the upper left corner. The telescope used, either the 305 meter or the 91 meter telescope, is indicated in the upper right corner. The X axis gives the velocity in km/sec. For spectra taken with the 91 meter telescope, the velocities are heliocentric. For those taken with the 305 meter telescope, the velocities are with respect to the Local Standard of Rest. The Y axis gives the intensity of the channel values in Janskys. The two vertical bars, one to each side of the spectral line, indicates the velocity range which was integrated to determine the HI mass. All of the spectra have been boxcar smoothed over three channels.

PAGE 68

61 coo C O (T V' ^^_ ^^OO -TN c. -< oj n oj < i2. >^ c c; i> C> o ^c^ '• '-') -^ o c o o c. o c^ c o c •" c c ^ ^^^^.-.OO--' o o-^ •-< c c> '-^ '-"> o sr^rj !^ -> " . •> r.-t .T! o o f^ f^ ^ a '".! "c ''I o '^ cj f ^. r^ c\j ^ ,, -t ^ : I N n f ^ 7 -z. 7 y ~' "Z < T < < < '^ ^ 7. 'Z y. 7.'^7 ^< < CO -> :o T! o > ri ~.^ i" UJ UJ li! U UJ -J U.I UJ Ul U U U; UJ U: U UJ U U.I liJ "J u 'Jl ';J '-! M U ;J M -^ a' M Ti K -• 1.'^ cc v-t rrro ro -.0 o .n o 10 N 0^ N in ^ 'a . ^ ^ "-• cvi CJ •-' o.' .-I L'"w r'^ fvj oj •" f\. N i n n (\j ^ C> f^. '^ — '^ "^ lO ro .-r. <} o ! N T OJ O r ^ <}• ^ (M -< -< O oooooooooooooooooooooocoooo n o r* .* O ?^ O r^ »-• ri O £)-"->'"" CO O ^/"-'^ CO o ^ '0 C> l^ o ^ ^ :» IV. T .-t vC c> " i? ?* ^~ f^ o o c" r. "T N N<" cr N 'r. r^ o IT c; ri >-« ^ ip ^ >: "< f\J Li f\J or.' " i CD n ?= ^J "^ "J "^ 5 O a N ro .rf Ul j> a^ CCo -< n M ? I'; w^ o (^ r> -I '^ 'H '• "•^»• -' rj. o '^ — SI rT S ^ ^ "< •?: ^ 7C ?: "Ni M '-I — ,r) M -> -' -D O — ^; -> M -• "^ CI -• — rj M r K rsj M n; Ni -vJ — > —

PAGE 69

-C

PAGE 70

63 i [T 1 — 2: I '0 C0"0

PAGE 71

64 [U— > I i r> I i i-^-\ •^ ^^^, == I ,-. ^-^

PAGE 72

65 L i^ I I r..j h: r H H rS tiQ'J'C! 000'

PAGE 73

66 -i r.T

PAGE 74

67 4 ' Slr^

PAGE 75

68 i tr b1 r-<• 4J%; -> i ! -1 s I = 1 ::a^_ I '! -I. r,r^E^^ U 1

PAGE 76

69 _

PAGE 77

70

PAGE 78

71 (rr r: H -L^l j^ ~^^\ -H 5j

PAGE 79

72 rs J en I U> U j:j L I'J U

PAGE 80

CHAPTER V ANALYSIS OF THE DATA The properties of blue compact galaxies seem to be more similar to those of late type galaxies of the Hubble sequence than to any other type. O'Connell (1979) found that the rotation curves of six luminous blue compact systems were quite similar to those of late type galaxies. Total masses, however, were found to be somewhat low for late type galaxies. Studies by Fairall (1978) and O'Connell (1979) indicate that many of the more luminous compact galaxies show some faint underlying structure, indicative of late type systems. Neutral hydrogen studies by Chamaraux (1977) indicate that the total masses and the HI masses of blue compact galaxies extend into the same range as those of late type galaxies. Finally, Huchra (1977) concludes that Markarian galaxies are mostly late type systems. We would like to compare the properties of the blue compact galaxies in our sample with those of late type systems to see how closely related the two types are. The best sample currently available of the optical and HI properties of late type galaxies has been published by Shostak (1978). Shostak's study utilizes the same system of dimensions and magnitudes as our study of blue compacts. Thus we are able to compare the properties of the two samples without any conversion between the two. Of Shostak's (1978) survey of 169 late type systems, we omit any which are peculiar, multiple, or confused in their HI. This leaves 139 late type systems for comparison against the 72 blue 73

PAGE 81

74 compact systems detected in HI. After a discussion of the selection effects involved in our sample of blue compact galaxies, we will compare their basic global properties to those of these late type galaxies. Selection Effects and Completeness of the Sample Our list of 99 compact systems was drawn from the lists of Zwicky, Markarian and Haro. Only the Markarian catalog is a reasonably complete list. Our selection of compact objects from the Zwicky (1971) list was made from among the systems with published Doppler velocities. Many of these were determined by Sargent (1970b), whose study was heavily weighted toward the bluest Zwicky compacts, with the expectation that they would be the most interesting to study. Since all the Zwicky compacts with known velocities could not be observed in this study, blue systems were selected preferentially over red objects, for it was believed they would contain more neutral hydrogen. Thus, the sample of Zwicky compact galaxies on our list is heavily weighted towards the blue and very blue systems. Any statistics determined for Zwicky compacts in general are not applicable to this study. However, the blue and very blue Zwicky compacts can be considered together with Markarian compact galaxies, owing to their similar morphological and spectroscopic properties. The Haro galaxies were found at random and by accident. But their number is so small that they obviously are a very incomplete sample. A minority of these systems appear to be spiral or peculiar spiral galaxies seen nearly edge-on and are not included in this study. The

PAGE 82

75 majority can be classified as compact galaxies, because they were discovered in a manner similar to that for the Markarian galaxies, and because of their morphological and spectroscopic similarities. Therefore, they can be considered a subset of the Markarian catalog and considered together with the Markarian galaxies rather than as a separate group. Only a limited number of Markarian galaxies were placed on our list for this study. Mainly, these were systems whose Doppler velocities were published after 1975. This was done because several neutral hydrogen studies of Markarian galaxies by other observers were apparently in progress, and we did not vyant to duplicate these studies. However, no other criteria were used in choosing the Markarian galaxies. Employing these criteria, our initial observing list contained some 150 Zwicky, Haro, Markarian and Seyfert galaxies. Approximately 1/3 of these systems were not compact galaxies and they are not reported in this study. Our sample of compact galaxies can be considered, for the most part, a subset of the Markarian catalog. A few of our sources were red objects without Markarian-1 ike properties, but these were mostly undetected in HI. As this study represents only a selected subset of the Markarian catalog, we must be yery careful in applying the statistics found for Markarian galaxies in general. We need to estimate the fraction of Markarian galaxies that are compact galaxies. The study of Huchra (1977) showed that approximately half of all Markarian galaxies may belong in the normal Hubble sequence, mostly as late type systems. Most of the remaining Markarian galaxies can be classified as compact galaxies. Thus perhaps half of all Markarian galaxies are

PAGE 83

76 compact galaxies. On the original list of 150 sources, 'l^ were selected only because they were flarkarian galaxies (36 other Markarian galaxies were selected because they were also Zwicky, Haro, or Seyfert galaxies). Of these, we have classified 16 as compact and 8 as noncompact. Thus, the estimate that half of all Markarian galaxies are compact is probably not an overestimate. Indeed, the low luminosity, dwarf systems in the Markarian catalog may all be compact systems. The Markarian catalog appears to be substantially complete down to an apparent photographic magnitude of -15.5 (Sargent, 1972). The proportion of all galaxies that are of the UV intense, Markarian type has been estimated in three studies. Sargent (1972) finds this proportion to be -2.5% for -22 < M < -20; -1% for M = -17; and perhaps as high as lO'o for M = -14. Huchra and Sargent (1973) find this proportion to be 5-10°; for -22 < M < -14. Huchra (1977) finds it to be -6% for -21 < M < -14. Huchra (1977) included Seyfert galaxies in his estimate while the first two studies did not. There is a lot of uncertainty at the low luminosity end. Huchra and Sargent (1973) point out that the statistics for low luminosity systems is rather poorly known for both field galaxies and Markarian galaxies. The majority of the Markarian galaxies are not extremely blue systems. In the color-color diagram, approximately 75% of the Markarian galaxies overlap the colors of field galaxies (Huchra, 1977). Thus, only -1/4 of the Markarian galaxies are exceptionally blue systems. Huchra (1977) made a rough estimate of the space density of galaxies as a function of their colors. For (U-B) $ -0.4 and (B-V) < +0.35, it seems that over half of all galaxies are Markarian galaxies. Thus, the Markarian lists seem to be fairly complete for the bluest systems, but

PAGE 84

77 probably less complete towards the redder systems. This bias must be present among the objects in our study and is probably accentuated owing to the method by which the objects were chosen. Thus, we expect the ratio of very blue to blue compact galaxies in this study to be greater than the actual ratio of these systems. We can make rough estimates of the proportion of all galaxies which are blue or very blue compact systems. Overall, it is perhaps ]-5%. For -22 < M < -20, it is -1-3%; for M^ -17, it is -2-4%; P P and for M = -14, it is -3-10%. Our sample of blue compact galaxies is a magnitude limited sample. The Markarian catalog is restricted to 13 < M^ < 17. The Zwicky and P Haro lists have similar, but less well defined limits. In our sample, only one source is brighter than By = 13.0 and only one is fainter than Ej = 16.4. This has the effect of producing a strong distance dependency in the absolute magnitudes. This can be seen in Figure V-1 , where we plot absolute magnitude, Mpo, versus log D. The absolute magnitudes T must be confined to a band -3-4 magnitudes wide with a slope of -5. For example, systems of exactly By = 15.0 would lie along the straight line defined by M„o = -5 log D 10.0. A least squares regression fit ^T to the data of Figure V-1 gives Mgo = -4.82 (±.27) • log D 10.96 (±.45). Thus, nearby, very luminous compacts and distant, intrinsically faint compacts are excluded from this study. Obviously, the distant, dwarf systems are too faint to detect. However, might there be any nearby, very luminous compacts, excluded from our survey, that could be studied to learn more about the distant systems? Huchra and Sargent (1973) have estimated the space density of Markarian galaxies. Assuming

PAGE 85

!T 78 n ^ ri3 rju u.-ij LOG Figure V-1 . Absolute magnitude, Mgo, versus the log of the distance, in Mpc, for blue compact galaxies.

PAGE 86

79 that half of the Markarian galaxies are compact, then at M = -22, we should expect only one blue compact within -50 Mpc. We find none this luminous in our study. At M = -21, we should expect only one blue compact within -20 Mpc; at M = -20, we should expect only one within -12 Mpc; and at M = -19, we should expect only one within ~10 Mpc. Thus, there is a strong selection effect against bright, blue compact galaxies close to the Milky Way and our list may be deficient by a wery small number of luminous, nearby systems. Luminosities Compact galaxies are defined to be systems with higher than average surface brightnesses. We would like to see how the luminosities and surface brightnesses of the compact galaxies in our sample compare to those of more normal galaxies. We first compare their luminosities to those of the late type galaxies taken from the survey by Shostak (1973). In Figure V-2, we plot log Lpo versus log A. The blue compacts do not "^T extend to as high luminosities or dimensions as the late type. This difference does not seem to be great. However, the blue compacts do extend to much lower luminosities than late type galaxies. Late types 9 appear to have a cutoff in luminosity at Lno = 10 L , and only -8% in D-r © Q ' this sample are fainter than Lno =2x10 L . For the blue compacts 9 of our sample, -32% are fainter than Lpo =2x10 L and -22% are D-p @ 9 fainter than Lno = 10 L„. Thus, on the order of 1/3 of the compacts Bj in our sample seem to be dwarf systems. In the region of overlap in Figure V-2, blue compact galaxies are somewhat more luminous, on the average, than late type galaxies of equal dimensions. Least squares regression lines have been fit to the

PAGE 87

80 ++ CD

PAGE 88

81 two groups. For the blue compact galaxies, the luminosities are given by log L o = 1.88 . log A + 7.86 . V-1 ^T (±.11) (±.11) For late type galaxies, it is given by log Lro = 1.63 • log A + 7.93 . V-2 ^T (±.09) (±.12) The difference in the slopes of the two lines appears to be significant, considering the uncertainties. On the average then, blue compact galaxies are more luminous than late type galaxies by a factor of -1.5 at A = 10 kpc and by a factor of ~1.9 at A = 25 kpc. Next, we consider the surface brightnesses of blue compact galaxies. Sargent (1970b) found that the surface brightnesses of Zwicky compact galaxies were -100 times greater than is found in normal Hubble sequence galaxies. Sargent, however, made the mistake of comparing Holmberg dimensions for normal galaxies with the dimensions of only the bright cores of the compact galaxies. These two systems are not compatible. However, O'Connell (1979), in a study of blue compacts, found that their average surface brightness was about one magnitude greater than that in late type galaxies. A thorough study of surface brightnesses of blue compacts would determine their central surface brightnesses. Unfortunately, in this study, we can only find the surface brightnesses averaged over the total dimensions. This will dilute the effect of the high surface brightness cores of these objects.

PAGE 89

82 For a measure of the average surface brightness, we divide the luminosity by the square of the major axis. The log of this is plotted against color, (B-V)°, in Figure V-3, for blue compacts and for late type galaxies. Blue compact galaxies are found to be typically bluer and of higher surface brightnesses than late type galaxies. For 63 blue compacts, the average of their surface brightnesses is 1.998 = 2 times as high as that of the late type galaxies. Among the blue compacts, the highest surface brightnesses are ~10 times the average for late type galaxies. As approximately 1/3 of our blue compacts are dwarf systems, we would like to know if the surface luminosities are different for dwarfs and nondwarfs. The HI masses are an indication of sizes, as will be seen in the next section. Thus, we plot surface luminosity versus HI mass for compact systems in Figure V-4. The data for blue compacts are quite scattered and no correlation is seen. Thus there does not seem to be any significant difference between dwarf and nondwarf blue compacts, in terms of their surface luminosities. Searle et al . (1973) used a model of star bursts to argue that very luminous galaxies could not be of the extremely blue type, such as the low metal abundance dwarf systems. The model they used predicted that at M = -17, the fraction of all galaxies that are extremely blue would be -10 , and much smaller at greater luminosities. Thus very blue g systems brighter than M = -17 (or Lgo = 10 L^) should be extremely rare. However, two of the 11 metal poor systems studied by Alloin et al . (1978) appear to be of greater luminosities than this limit. The isolated system M162 has a similar metal abundance as IIZw40 but has M z -19. A double system, M171, is similar to IZwlS in metal abundance but has M = -21.

PAGE 90

XI 1 >. Oo '^!uo 83 ^ +3 a -V""ZlEtr [R-vIt 2 Figure V-3. Log (Ldo/A ) as a function of color index, (B-V)° Dj T The squares are blue compact galaxies. The crosses are late type galaxies.

PAGE 91

pa (JD 84 " n a . t^' LOG M HI y.4[l Figure V-4. Log iLr.o//\ ) as a function of log M,,, for blue oj ^ HI compact galaxies.

PAGE 92

85 To look more deeply at this hypothesis, we plot, in Figure V-5, log L o versus (B-V)° and, in Figure V-6, log Lro versus (U-B)° for blue Dj \ b^ 1 compact galaxies and for late type galaxies. For late type galaxies, it is seen that ~Z% are bluer than (B-V)° = +0.30, and -15% are bluer than (B-\/)° = +0.40. For blue compacts, we find that -34% are bluer than (B-V)° = +0.3, and ~66% are bluer than (B-V)° = +0.4. Approximately 1/3 of the blue compacts are bluer than the bluest late type galaxies in (L)-B)j. From Figures V-5 and V-6, it can be seen that the bluest compacts are dwarf systems, but that extreme blueness is not limited to the dwarf galaxies. It appears that, of the blue compacts 9 in our sample brighter than Lgo = 10 L , at least 20% are very blue objects. From Sargent (1972) and Huchra (1977), it appears that, of all galaxies brighter than M = -17, on the order of 3-7% are the UV intense, Harkarian types. If half of these are blue compacts and if 20% of these blue compacts are extremely blue, then on the order of .3-. 7% of all galaxies brighter than M = -17 are extremely blue. We must keep in mind that the Markarian lists may be more complete at the extremely blue end. Thus, this proportion may actually be as low as .1 or .2%. This is still considerably greater than the proportion predicted by Searle et al . (1973). However, this finding does not invalidate the star burst model. Rather, it indicates that the model of Searle et al . (1973), of flashes of star formation occurring in small, statistically independent cells in a galaxy, is not a good description for the mechanism of the star bursts in these systems. What is needed is a model which takes into account all of the properties of these systems and allows star bursts at all luminosities.

PAGE 93

86 03 "^ + in '-f'3' . 'j. 1-n a n 0.^0 Figure V-5. Log (L„o) as a function of color index, (B-V)° b-j'T The square are blue compact galaxies. The crosses are late type galaxies.

PAGE 94

87 * ^+ a -ia ^ jx" -i.GO C.JG Figure V-6. Log L^o as a function of color index, (U-B)^. The squares are blue compact galaxies. The crosses are late type galaxies.

PAGE 95

Neutral Hydrogen Content We would also like to see how the neutral hydrogen masses of blue compact galaxies compare to those of late type galaxies. In Figure V-7: we plot log I'] , versus log A. The two groups follow a similar relationship. For the blue compact galaxies, the HI masses are given by log M^, = 1.51 • log A + 7.68 . V-3 "^ (±.16) (±.12) For the late type galaxies, they are given by log MuT = 1.61 • log A + 7.43 . V-4 "^ (±.07) (±.10) The differences between the two regression lines is slightly less than significant. The indication is, though not certain, that blue compact galaxies are more HI rich than late type galaxies of the same dimensions. A similar result was found by Chamaraux (1977). At A = 6 kpc, the blue compacts contain, on the average, -1.5 times as much HI as late type galaxies. However, at A = 40 kpc, this is reduced to a factor of -1.2. We stress that these values are rather uncertain, but at the least, blue compact galaxies have marginally more neutral hydrogen for their sizes, than late type galaxies. As was seen with the luminosities, about 1/3 of the blue compacts in our sample are dwarf systems. Late type galaxies have a fairly sharp cutoff at A = 6 kpc. This may, however, be a selection effect in the classification of low luminosity galaxies, ^ery few of these were included in Shostak's (1978) sample, and several of these we excluded from this discussion because they were classified as peculiar systems.

PAGE 96

89 #^ c ^ ^ "til ^~ = 3 3 -n "'^ LOG H 1.75 Figure V-7. Log M^^ as a function of log A. The squares are blue compact galaxies. The crosses are late type galaxies.

PAGE 97

90 We next consider the HI surface densities of these systems, which can be represented by dividing the HI mass by the square of the major axis. In Figure V-8, this is compared with color index, (B-\/)° The blue compacts can be seen to have significantly greater HI surface densities than late type galaxies. For 70 of the blue compacts detected in HI, the average HI surface density is 1.98 = 2 times as high as in the late type galaxies. The HI surface densities of some of the blue compacts are much higher than this, up to 12 times the average for late type systems. This HI surface density does not seem to differ between dwarf and nondwarf blue compacts. This is seen in Figure V-9, where HI surface density is compared to luminosity. The HI surface densities considered here are derived from the optical dimensions of these systems. Ideally, they should be derived from the HI dimensions, but these are not easily obtained. In the next section, we shall discuss the HI dimensions of blue compact galaxies and investigate whether their true HI surface densities are actually greater than those of late type galaxies. As was done for the luminosities, we plot log M^,, versus (B-V)° in Hi T Figure V-10 and log Mj,, versus (U-B)° in Figure V-11 for blue compact galaxies and late type galaxies. The results are similar to those found for the luminosities. The bluest systems are seen to have low HI masses but extreme blueness is not limited to the low HI mass, dwarf systems. HI Dimensions As outlined in the last chapter, the HI masses were corrected for beamwidth effects by assuming that the HI in these systems was

PAGE 98

nti 91 ^. CO r R \i ] ° cm: L.UFigure V-8. Log (M,^j/A ) as a function of color index, (B-V)° The squares are blue compact galaxies. The crosses are late type galaxies.

PAGE 99

92 Oo H^, "^ —3 _T--;_'^ /.uu . uU 9.00 .OG L_o li .00 i^.oi: Figure V-9. Log (M /A ) as a function of log (L„J for blue HI T compact galaxies (squares) and for late type galaxies (crosses).

PAGE 100

93 + a ^ -f+ a Q-CO G.20 0.40 0.60 u^on ['on Figure V-10. Log M^^^ versus color index, (B-V)° for blue compact galaxies (squares) and for late type galaxies (crosses).

PAGE 101

94 4—a -Q."'^ -0.50 r 1 ! _ & 1 ° T Figure V-11. Log M versus color index, (U-B)^, for blue compact galaxies (squares) and for late type galaxies (crosses) .

PAGE 102

95 distributed as an elliptical Gaussian with major and minor axes of Fa" and F. b". For 13 sources observed with both the 91 meter telescope and the 305 meter telescope, the average value of F was found to be -2.4. For sources observed with only one telescope, the HI masses were corrected assuming F = 2.4. For observations with the 91 meter telescope, the HI mass corrections are all small--only one is greater than 8.5%. These are not highly sensitive to the value of F and can be considered quite accurate. For observations with the 305 meter telescope, the HI mass corrections are considerably larger. The smallest correction is -7% and the largest is -170%. However, only 3 out of 47 of the HI mass corrections exceed 100%. To show the validity of these HI mass corrections, we can compare the corrected HI masses to the linear dimensions of these systems. For uncorrected HI masses, those found with the 305 meter telescope are systematically lower than those found with the 91 meter telescope. This is expected, of course, because of the larger beamwidth of the 91 meter telescope. Since there were not systematic differences in the two samples, they should follow the same relationship between corrected HI mass and linear diameter. In Figure V-12, we plot log M , versus log A with separate symbols for the 91 meter observations (crosses) and the 305 meter observations (triangles). Regression curves have been calculated and plotted for the two samples. The two samples seem to follow almost the same relation, but the two regression lines differ slightly. For the 91 meter data, we get log M = 1.80 • log A + 7.43 , V-5 "^ (±.13) (±.12)

PAGE 103

96 Figure V-12. Log Muj versus log A for blue compact galaxies. Crosses are for observations with the 91 meter telescope. Triangles are for observations with the 305 meter telescope. The boldface line is fit to the 91 meter data, the lighter line to the 305 meter data.

PAGE 104

97 2 with a correlation coefficient of r = 0.87. For the 305 meter data, we get log M^. = 1.44 • log A + 7.78 , V-6 "^ (±.15) (±.16) with a correlation coefficient of 0.65. The fit between the two regression lines cannot be improved by using larger or smaller values of F, The scatter in the data is such that the two samples cannot be used to find the best value of F. We can only show here that the two sets of data are closely corrected to the same system. The differences in the two regression curves show that the two sets of data agree in HI mass to better than a factor of 2, even at the extreme limits of log A. The assumption that the HI in these blue compact galaxies is Gaussian distributed is probably an oversimplification. Gottesman and Weliachew (1972) found that the HI in IIZw40 was distributed in the form of a double Gaussian. An HI Gaussian core with F = 1 and a low density Gaussian halo with F = 5-6 were found, each component containing about half of the HI mass. To fit a single Gaussian to their observations would require an F parameter somewhere between these two values, close to the value of F = 2.46 obtained in this study. We observed the blue compact galaxy M314 with the 91 meter and 305 meter telescopes. A value for F of 1.6015 was obtained. Additionally, this object was observed with the three element hydrogen line interferometer of NRAO in 1977 by Gordon and Gottesman (unpublished). Baselines of 100 meters, 200 meters and 300 meters were obtained. The flux detected at these baselines, averaged over all hour angles, can

PAGE 105

be used to estimate the HI extent. We can consider the 91 meter observation as the zero meter spacing, giving an estimate of the total flux. The data obtained are given below in Table V-1. Table V-1. Flux of M314 Baseline Integrated Flux (meters) (wavelengths) (Jykm/sec) 11.64 ± .50 100 470 10.26 ± 1.34 200 940 9.39 + .89 300 1410 5.29 ± .94 A plot of these data, in Figure V-13, shows that the HI is probably not distributed as a simple Gaussian. A best fit Gaussian through the error bars of these data points has a half power width of -1200 wavelengths. This indicates a half power width for the source of ~76" arc. This compares with the major and minor optical dimensions of (47" x 69") of arc. The parameter F is therefore between 1.1 and 1.6. This agrees fairly well with the value of F = 1.6015 found for M314 in this study. Shostak (1978) corrected the HI masses of the late type galaxies in a somewhat different manner. To get the same beam correction factors used by Shostak requires F : 1.3 in our beam correction program. This is considerably less than the average value of F = 2.4 found for the blue compacts. While a direct comparison between the two is not conclusive, the implication is that blue compact galaxies are more extensive in HI, compared to their optical dimensions, than late type

PAGE 106

99

PAGE 107

100 galaxies. This also implies that the HI surface densities of blue compact galaxies are actually not greater, on the average, than those of late type galaxies. Some evidence is available to support this contention. Bottinelli (1971) found that the ratio of HI diameter to optical diameter was not independent of the type of galaxy. The ratio increased systematically towards later types, increasing by a factor of -2.1 from types Sb to Irr. Kellman and Black (1973), using 21 cm dimensions of 51 galaxies from Bottinelli (1971), found that the mean HI surface densities were no more than weakly correlated with the type of galaxy. These densities were found to be either constant with galaxy type, or possibly decreasing towards later types. Thus the ratio of the HI dimensions to optical dimensions for the blue compact systems seem to fit the progression found for normal galaxies. Compact galaxies could represent an extension of the Hubble sequence to later types. The HI Profiles We next consider the 21 cm velocity profiles and their velocity widths. This subject has been investigated thoroughly by Roberts (1978) for spiral galaxies. The shape of the 21 cm profile of a galaxy is determined by the combined effects of the motion and the distribution of HI in the galaxy. For spiral galaxies, the profiles generally show two horns, with steep, nearly vertical boundaries, and a central depression. Of the 21 cm profiles shown in Figure IV-1 , IIZwlS is a good example of a double horn profile. Such a profile is caused by the gradient of the radial velocity field of the galaxy. The peaks are produced by HI near the major axis and beyond the region of solid body

PAGE 108

101 rotation, where the spatial area is large and the velocity gradient is small. The peak rotational velocity occurs at the edges of the profile and not at the peaks of the horns. For a uniform HI distribution in a rotating disk system, two horns and a moderate central depression will be seen in the HI profile. If the HI has a ring distribution, as is often the case for spiral galaxies, the horns will be more pronounced and the central depression will be deeper. If the HI is centrally concentrated, a single peak profile will be seen. Thus when a double horn profile is found for a galaxy, it is taken as clear evidence for rotation in a disk system. When a profile with steep sides but with a flat or rounded peak is found, it can be taken as evidence of rotation and indicative of a greater central concentration of HI. A good many of the HI profiles of the blue compact galaxies in Figure IV-1 can thus be interpreted as due to rotation. However, many of the HI profiles of Figure IV-1 show single peaked, almost Gaussian shaped profiles. Single peaks can occur if the HI is strongly concentrated in the central regions, or if the Doppler component of the random motion exceeds the ordered motion, as in a face-on spiral galaxy or a nonrotating galaxy. Irregular galaxies often show single peaked profiles, probably due to a central concentration of HI and large random motions. Since 35 of the 72 detected galaxies show single peaked HI profiles, we cannot immediately assume that they are rotating disk systems. The majority of these show large values of axial ratio. Thus, they could be nearly face-on, rotating disk systems, or they could have a strong central concentration of HI. Alternatively they could have shapes like the elliptical galaxies and not be rotating disk systems at all.

PAGE 109

102 It is also seen that the least luminous galaxies are most likely to show single peaked profiles. In fact, all of the blue compacts in this study fainter than M^o = -17.7 show single peaked, almost Gaussian profiles. For systems brighter than this, the majority have profiles indicative of rotation. Thus, it seems to be a general property of dwarf blue compact galaxies to show single peaked 21 cm profiles. If these systems are not rotating disk systems, then the observed line widths, aV, should not vary with axial ratio, b/a. On the other hand, if they are rotating disk systems seen at some inclination, then the observed line widths should vary with b/a, in the sense that AV increases for decreasing values of b/a. In Figure V-13, we plot log AV versus b/a for the blue compacts fainter than Mo = -18, the systems "t that can be considered dwarf. A fairly good correlation is found to exist, in the sense expected if these are inclined rotating disk systems. When corrected for inclination, the dependency of line widths on axial ratio should disappear. The corrected line width, aV , is found by dividing AV by sin i, where i is the assumed inclination. In Figure V-14, we plot aV^ versus b/a for the systems fainter than Mno = -18. Systems with inclinations of less than 30° are excluded '^T because of the large uncertainties in their inclinations. An almost horizontal line can be fit to the points of Figure V-14, indicative that the correction for inclination is not a bad one. We make a similar comparison for the blue compacts of medium luminosity, those between an absolute magnitude, NLo, of -18 and -20. ^T Figure V-15 shows log AV versus b/a for these systems. As expected, there seems to be a trend toward greater values of AV with decreasing axial ratios. Figure V-15 shows more scatter than Figure V-13 and may

PAGE 110

103 r_2"rj Co.;2»J ).0C n -.on J . L. U 0.40 'R Q.SQ 3.30 1.00 Figure V-13. Log AV versus axial ratio, b/a, for the low luminosity systems (Mp,o I -18). The observed line width, AV, ^T " is seen to correlate with axial ratio, as expected if these are inclined rotating disk systems.

PAGE 111

< .± 104 Q.UU 0.4D B/fl G.SC 0.30 1.00 Figure V-14. Log aV versus axial ratio, b/a, for the systems with M„o > -18 and inclination, i > 30°. When corrected for inclination, the line width, AV., does not appear to correlate with axial ratio.

PAGE 112

105 be less convincing. Figure V-16 shows the corrected line widths for these systems with inclinations greater than 30°. The resulting corrected line widths are approximately independent of axial ratios. We lastly do the same for the most luminous systems. Figure V-17 shows log aV versus b/a for the systems more luminous than Moo= -20. ^T Again, AV appears to increase with decreasing b/a. As with Figure V-IS, the data are somewhat scattered and not as convincing as Figure V-13. Figure V-18 shows the same systems, with inclinations greater than 30°, corrected for assumed inclination. The corrected line widths appear to be less dependent on the axial ratio. From a study of Figures V-13 through V-18, it appears that the dwarf systems show the strongest evidence of being inclined rotating disk systems. This is exactly the opposite of what is expected. Further evidence for rotation in dwarf systems is given by Gottesman and Weliachew (1972) for IIZw40. Thus, we make the assumption that dwarf blue compacts are rotating disk systems. The medium and highest luminosity systems often show HI profiles which indicate rotation. O'Connell (1979) found clear evidence for rotation in six nondwarf blue compacts. Thus, we also make the assumption that the nondwarfs are rotating disk systems. The corrected 21 cm line widths of spiral galaxies have been found to be correlated with their absolute magnitudes and their major axes. We would like to investigate whether such a correlation exists for our sample of blue compact galaxies. Since there existed a distance dependent selection effect in the absolute magnitudes of this sample, a similar effect should be seen in the corrected line widths, if they are indeed correlated with the absolute magnitudes.

PAGE 113

nT < CD 106 1].rjo , :-'D 0.40 B/R 0. 0.30 GO Figure V-15. Log aV versus b/a for systems with -18 > Mpo ^ -20. With a few exceptions, the observed line widths appear to correlate with axial ratio.

PAGE 114

0.00 107 20 0.40 B/R o.so . 8 Figure V-16. Log AV versus b/a for the systems with -18 > Mpo > -20 and inclination, i > 30°. The ^T corrected line widths appear to be almost independent of axial ratio.

PAGE 115


PAGE 116

< CD CDOJ 109 0.00 Q. 3/R Q.SO Figure V-18. Log aV versus b/a for the systems with M„o < -20 B^ and inclination, i 30°. The corrected line widths appear almost independent of axial ratio.

PAGE 117

no In Figure V-19, we plot log AV versus log D, for late type and for blue compact galaxies. For the late type galaxies, a fairly strong correlation is found. The largest values of AV^ are found at the greatest distances; the smallest at the nearest distances. A least squares regression line for the late type galaxies yields log AV = 0.39 • log D + 2.05 . V-7 (±.04) (±.05) The correlation coefficient is 0.45. For the blue compact systems, the data are more scattered and the correlation is weaker. The regression 1 ine is log AV = 0.20 • log D + 2.05 . V-8 ° (±.06) (±.10) The correlation coefficient here is only 0.21. Thus, there appears to be some distance dependent selection effect in the corrected line widths but it is not nearly as strong as was seen for the absolute magnitudes. In Figure V-20, we plot log AV^ versus absolute magnitude, Mpo , to see if the corrected line widths are correlated with the absolute magnitudes. For the late type galaxies, a regression line fit to the data gives log AVq = -0.092 • Mgo + 0.731 . V-9 (±.008) ^ (±.151) The correlation is fairly good, with a correlation coefficient of 0.57. For blue compact galaxies, the data are more scattered, but a correlation does exist. A regression line fit to the data yields

PAGE 118

m Figure V-19. Log aV versus log D for blue compact galaxies (squares) and for late type galaxies (crosses).

PAGE 119

! 112 -^ Bip Figure V-20. Log aV versus absolute magnitude, M„o, for blue B^ compact galaxies (squares) and for late type galaxies (crosses).

PAGE 120

113 log AV^ = -0.056 • Mgo + 1.345 . V-10 (±.010) ^ (±.181) The correlation coefficient is 0.45. At the higher luminosities, the corrected line widths are marginally less for the blue compact galaxies than for the late type galaxies of equal MpoAt M„o = -21, they differ B^ ^T by -40% on the average. The relationship between AV^ and M„o is often called the TullyDj Fisher relation. It has been used to estimate the distances to galaxies and to estimate the Hubble constant (Tully and Fisher, 1977; Shostak, 1978). A sample of nearby galaxies with well known distances, inclinations, morphological types and 21 cm line widths are required as calibrators. No such calibrators exist of the blue compact type. Thus, we cannot use our data to determine the Hubble constant. In Figure V-21 , we plot log AV^ versus log A for blue compact galaxies and for late type galaxies. Again, the correlation is best for the late type galaxies. The smallest systems tend to have the smallest values of aV • the largest tend to have the greatest values of AV . -^ The least squares regression line for the late type galaxies is log AV = 0.39 • log A + 2.04 . V-11 (±.04) (±.06) The correlation coefficient is 0.42. For the blue compact systems, the regression curve is log AV^ = 0.21 • log A + 2.17 . V-12 ° (±.06) (±.06) The correlation coefficient is somewhat small, only 0.22. The larger blue compacts appear to have smaller corrected line widths than the

PAGE 121

114 Figure \/-21 . Log aV^ versus log A for blue compact galaxies (squares) and for late type galaxies (crosses),

PAGE 122

115 late type galaxies of the same size. The consequences of smaller values of aV^ will be discussed in the next section. Total Indicative Masses The corrected 21 cm profile widths can be used to estimate the total masses of these systems, assuming they are rotating disk systems. In Chapter IV, a model described and used by Shostak (1978) for late type galaxies was used for estimating the total masses. The assumptions made are that the maximum rotational velocity is AV /2 and that this velocity occurs at a radius from the center of the galaxy of 0.14 * A kpc. The quantity, M^. , defined by equation IV-9, is called the indicative mass. It should not be taken literally as being the total mass of the galaxy but rather as a model estimate of the total mass. We wish to compare these indicative masses of the blue compact galaxies to those of the late type galaxies in Shostak's (1978) sample. In Figure V-22, we plot log M^ versus log A for the two groups. This plot is similar to Figures V-2 and V-7, which showed log L^o and log M,,t ^ Bj ^ HI versus log A. Figure V-22 shows the opposite effect though; indicative masses of the blue compacts are less than those of the late type galaxies. A regression line fit to the blue compacts yields log M. = 1.37 • log A + 8.85 , V-13 (±.11) (±.11) with a correlation coefficient of 0.75. A regression line fit to late type galaxies yields log M 1.62 • log A + 8.66 , V-14 ^ (±.10) (±.13)

PAGE 123

116 o+ -4Ti -i^"^=^ m.3^3 r -U n 3 -t3", 0.20 0.60 LOG fl 1.40 20 3.00 Figure V-22. Log of the indicative mass, M^ , versus log A, for blue compact galaxies (squares) and late type galaxies (crosses).

PAGE 124

117 with a correlation coefficient of 0.67. At the dwarf end, indicative masses seem greater for the blue compacts; but this quickly reverses. At A = 10 kpc, the average indicative masses of the late type galaxies are -23% greater than for blue compacts. This difference increases to -45% at A = 20 kpc and to ~10% at A = 40 kpc. We cannot, however, interpret these results as meaning blue compacts are less massive than late type galaxies of the Hubble sequence. This is because of the uncertainty in how well Shostak's (1978) model applies to late type galaxies and to blue compact galaxies. The reason the indicative masses are lower for blue compacts is because their corrected line widths are less than for late type galaxies of the same linear dimension. Roberts (1978) found that values of aV decreased toward later type galaxies of equal absolute magnitude. The blue compact galaxies appear to follow this progression to narrower corrected line widths. This could be an indication that they represent later types than the normal late type galaxies of the Hubble sequence. Another property of disk systems is that the early types appear to have greater central mass densities than later types. Roberts (1978) points out that the peak of the rotation curve (or the velocity turnover point) is more centrally located for early types than for late types. The masses inside this turnover radius have been found to be about equal for all types, at constant absolute magnitude. The central mass densities are thus found to be as much as 25 times greater for the early types as for the latest types. O'Connell (1979) found total central masses for six luminous, blue compact systems. His rotation curves covered, primarily, only the optical cores of these objects, and not their entire optical extent.

PAGE 125

118 His rotation curves do not extend to the velocity turnover points for five of these systems. For the sixth, it is not certain that a turnover point is actually reached. This is not so unusual, however. Rotation curves of late type galaxies are often flat, or nearly flat at large distances from the center. O'Connell (1979) points out that his rotation curves resemble those of normal late type spirals. He calculates the total masses out to the last point in the velocity curves. The masses found are somewhat small, approximately 1/3 of that found in the interiors of late type galaxies. But since these are masses inside the velocity turnover point, we do not have the entire interior masses. Three of O'Connell 's sources were observed in this study (IIZw23, IIIZwl2, and IIIZw43). O'Connell's masses are only -1-2 times the HI masses we find for these systems. More work needs to be done in this area before any real conclusions can be reached. In the next section, we will consider relations between colors, diameters, M^., M. and Lnonil Dj Comparisons Between Color, HI Mass, Luminosity and Indicative Mass The position of a galaxy in a color-color diagram can tell a lot about its nature. In Figure V-23, we present a color-color plot for the blue compact galaxies and the late type galaxies. Although there is a considerable amount of overlap, the blue compacts extend to considerably bluer values of (B-V)° and (U-B)°. For 44 of the blue compact galaxies, a least squares regression line fit to the data yields (U-B)° = 0.67 • (B-V)° 0.57 , V-15 ' (±.21) ' (±.08)

PAGE 126

0_H CDl7 119 + , + 45

PAGE 127

120 with a correlation coefficient of only 0.20. For 47 of the late type galaxies, we get (U-B); 0.96 • (B-V)° 0.60 , V-16 ' (±.11) ' (±.05) with a correlation coefficient of 0.64. In spite of the differences in the two regression lines, the two groups appear to form a fairly continuous band in the color-color plot. The greater scatter in the data for blue compacts may be partially due to the fact that their colors were taken from a variety of sources and may consequently have systematic differences and greater uncertainties. For the late type galaxies, colors were all from de Vaucouleurs et al . (1976). It can be seen, in Figure V-23, that something like 1/3 to 1/2 of the blue compacts are bluer in (U-B)° and/or (B-V)j than the bluest late type systems. Huchra (1977) found that -1/4 of the Markarian galaxies were bluer than the bluest Hubble sequence galaxies. This indicates that our sample of blue compact galaxies is overabundant in the very blue systems by no more than a factor of two. We next compare luminosities with HI masses. In Figure V-24, we plot log Lno versus log Mht for blue compact galaxies and for late type galaxies. The two parameters are found to be closely correlated. A large HI mass is accompanied by a high luminosity. A low HI mass is accompanied by a low luminosity. For 61 blue compacts, a least squares regression fit gives log Lpo = 1.02 • log M^,, + 0.39 , V-17 ^T (±.06) "^ (±.54)

PAGE 128

121 ~P^ -f '3" -a -I' 7. CO LCh M HI Figure V-24. Log Lgo versus log M^^ for blue compact galaxies (squares) and for late type galaxies (crosses).

PAGE 129

122 with a correlation coefficient of 0.84. For 139 late type galaxies, we get log L„o = 0.86 • log M^t +1.84 , V-18 ^T (±.05) "^ (±.51) with a correlation coefficient of 0.56. The two sets of data do not appear to systematically differ. Over the values of M|,r present, the two curves give similar values of log Lpo. The similarity between the two groups is further demonstrated by comparing the HI mass-to-luminosity ratios. For both groups, the average of f'liT/Lno is 0.39. For the blue compacts, this does not vary significantly between dwarf and nondwarf systems. In Figure V-25, we plot M j/Lno versus log A for blue compact galaxies and for late type galaxies. No correlation is found between these two parameters, although some higher values of Ml,t/L(,o are found at small values of A for Ml D-p blue compact galaxies. In Figure V-26, we plot Muj/Lno versus (B-V)°, to see if there is any correlation with color. No correlation is found for the blue compact systems. An extremely weak correlation is found for the late type systems, in the sense that bluer systems tend to have slightly greater average values of M|,j/Lno. The quantity, M ,/L o, is a ^^ery important parameter because it is independent of distance. For the late type galaxies used here, the values of M^j/L^o were averaged over all morphological types. When these are broken down into morphological classes, it is found that the M j/Lno ratios tend to increase towards later types (Roberts, 1969; Shostak, 1978). The average value of Muj/Lno found for the blue compacts is approximately consistent with the values found for Scd and Sd galaxies by Shostak (1978).

PAGE 130

a I 123 I m + on -3 a -3 ^^-_-fEit^l*-." -0. a. 75 1 Figure V-25. HI mass-to-luminosity ratio, Mmt/L o, versus log A for blue compact galaxies (squares) and for late type galaxies (crosses).

PAGE 131

OH 124 + '3' 2! C3 ; a 3-74.rjTi-_i_ ^j-sL — ;::: nr a ^ +"" 0.40 .V/ 1° Figure V-26. HI mass-to-luminosity ratio, ^^uT/Lpoversus (B-V)° for blue compact galaxies (squares) and for late type galaxies (crosses).

PAGE 132

125 We next consider the relation between total indicative masses and luminosities. In Figure \/-27, we plot log Mversus log L„o for blue compact galaxies and for late type galaxies. Total indicative masses are seen to correlate well with luminosities. For 51 blue compacts, a least squares regression fit gives log M. = 0.74 • log Lro + 3.60 , V-19 ( + .05) "^T (±.46) with a correlation coefficient of 0.83. For 138 late type galaxies, we get log M. = 0.86 • log Lpo + 2.08 , V-20 ^ (±.05) ^T (±.48) with a correlation coefficient of 0.71. Late type systems appear to have greater values of M^, on the average, than the blue compact systems at equal values of Lpo. At the low luminosity end, the opposite trend may exist, although there is too little data to be sure. When we consider the total indicative mass-to-luminosity ratios, the blue compact galaxies are found to have considerably smaller values than the late type galaxies. For the blue compact systems, the average of M^VLgo is 4.73; for late type systems, it is 5.41. In Figure V-28, we plot M^/Lno versus log A for blue compact galaxies and for late type galaxies. For blue compact systems, the highest values of M./Lpo are found at the dwarf end. However, there is no real overall correlation found. In Figure V-29, we plot M./L„o versus color, (B-V)°, for blue compact galaxies and for late type galaxies. Again, no correlations are found.

PAGE 133

126 -d^ '-rfCT -p ' — 3 a a. -5 .a 3/ . J ij c: . ij u y . U IJ LuG Figure \/-27. Log M^. versus log L o for blue compact galaxies (squares) and for late type galaxies (crosses).

PAGE 134

eP 127 ^ . --;r ^ a a ^^ -iU.^t. u.'d: LOG M Figure V-28. Total indicative mass-to-luminosity ratio, M./Lco, versus log A for blue compact galaxies (squares) and for late type galaxies (crosses).

PAGE 135

LI ^P.J j I uo 128 + -I

PAGE 136

129 The total mass-to-luminosity ratios of spiral galaxies have been found to decrease with increasing type (Faber and Gallagher, 1979). This conclusion may be somewhat weak because only a small number of late type galaxies were used. The fact that the blue compacts have considerably smaller values of M./Lpo than late types could be an indication that they follow this trend towards types later than the normal Hubble sequence galaxies. This is somewhat uncertain, owing to the questionable accuracy of the total indicative mass estimates for the blue compact systems. We lastly consider the relation between total indicative masses and HI masses. We plot, in Figure V-30, log Mversus log M , for blue compact galaxies and for late type galaxies. Total indicative masses appear to correlate quite well with HI masses. For 56 blue compact systems, a least squares regression fit yields log M. = 0.82 • log M^,. + 2.71 , V-21 ^ (±.05) "^ (±.48) with a correlation coefficient of 0.82. For 138 late type systems, we get log M. 0.86 • log 1%. + 2.58 , V-22 ^ (±.06) "^ (±.56) with a correlation coefficient of 0.61. Although the two curves are not significantly different, it can be seen that the blue compacts are slightly more HI rich than the late types at equal values of M^. This can be seen more easily in the ratios of HI mass to total indicative mass. In Figure V-31 , we plot M|,,/M^ versus log A for the blue compact galaxies and for the late type galaxies. The average

PAGE 137

^.JO 130 T"" rr-i ^_ + _ ^ _ ^3 __ ^ t^z^^~^ 3 n on H LOG Y HI Figure V-30. Log Mversus log Muj for blue compact galaxies (squares) and for late type galaxies (crosses).

PAGE 138

131 n a, a Ha ^ -0.c5 1.25 1.^5 LOG fl Figure V-31 . Ratio of Muj/M^. versus log A for blue compact galaxies (squares) and for late type galaxies (crosses).

PAGE 139

132 value of Mmt/Mis -0.11 for the blue compact systems, and -0.07 for the late type systems. The indication is that blue compacts have a larger proportion of neutral hydrogen than late type systems. However, due to the uncertainty in the indicative mass estimates, this is not definite. It can be seen, in Figure V-31 , that f^ur/^l-j is not correlated with the size of the galaxy. In Figure V-32, we plot Muj/M^versus (B-V)°. A weak trend appears to exist, in the sense that bluer systems have slightly greater proportions of HI. However, the correlation coefficients are rather small--only 0.18 for late types and only 0.04 for blue compacts. The fact that the blue compacts appear to have a greater proportion of HI than the late types is consistent with their being later in type than the latest types of the normal Hubble sequences. Roberts (1969), Shostak (1978) and many others have found such a trend towards larger values of HI mass-to-total mass ratios towards later type galaxies. Comparison With Low Surface Brightness Systems Searle et al. (1973) suggested that the very blue compact systems were undergoing brief, intense bursts of star formation. If so, they reasoned, there should exist similar objects in a quiescent phase. Fisher and Tully (1975) and Thuan and Seitzer (1979) have made HI observations of a class of low surface brightness galaxies. There is a possibility that these low surface brightness systems represent the quiescent phase of the very blue, high surface brightness systems. This type of galaxy is characterized by a very low surface brightness and little or no central concentration of light. Both dwarf and

PAGE 140

133 T 3 _a _j ^ ^ ' — I ~ ' ' -1.' _ u . U l1 U i B V' J T Figure V-32. Ratio of M^j/M^ versus (B-V)° for blue compact galaxies (squares) and for late type galaxies (crosses).

PAGE 141

134 nondwarf low surface brightness systems are found with these properties. The nearby, low surface brightness systems studied by Fisher and Tully (1975) and by Thuan and Seitzer (1979) appear to have HI masses in the same range as the blue compact galaxies in this study. Additionally, their HI profile widths and shapes are similar to those of the blue compact galaxies. The analysis by Thuan and Seitzer (1979) indicates that the low surface brightness systems have greater ratios of HI mass to photographic luminosity and of HI mass to total mass than late type galaxies (from Shostak, 1978) of comparable luminosities. The low surface brightness systems are seen to be more diffuse than late type systems and to have converted a smaller percentage of their HI into stars. Their ratio of total mass to photographic luminosity is comparable to that found in late type galaxies. More work needs to be done on these systems to see if they are related to blue compact systems. In the next chapter, we summarize the analysis of blue compact galaxies and attempt to explain some of their properties. Correlations In many of the figures, we found correlations or possible correlations between two variables. We would like to sum up the analysis with a list of such correlations. This is given in Table V-Z. The first two columns give the figure number and the parameters plotted. The square roots, r, of the correlation coefficients for the blue compact galaxies (rg p ) and the late type galaxies (r, j ) are given in columns 3 and 6. Positive values of r imply a positive slope of the regression line, while negative values imply a negative slope. The number of data points

PAGE 142

135 1 — CQ

PAGE 143

136 en 1—1

PAGE 144

137 used in the regression analysis are given in columns 4 and 7, for the blue compacts and the late type galaxies, respectively. In columns 5 and 8, we list the 5% critical limits, rro/, for blue compact galaxies and late type galaxies, respectively. These are from Crow et al . (1960, page 241) and depend on the number of data points. If \r\ is greater than rro,, then there is less than a 5% probability of such a high value of |r| occurring randomly.

PAGE 145

CHAPTER VI CONCLUSIONS The blue compact galaxies studied here have been found to be similar in their global properties compared to the late type galaxies studied by Shostak (1978). Their average values of luminosity, HI mass, and surface brightness are slightly higher than is seen in late type galaxies of the Hubble sequence, while their indicative masses are somewhat lower. The nondwarf, blue compact galaxies are seen to be rotating disk systems, probably similar to spiral galaxies, but without any obvious spiral structure. Approximately 1/3 of the blue compacts in our sample are seen to be dwarf systems. The properties of the blue compacts seem to run continuously from the dwarf systems to the most luminous systems, and there does not seem to be any clear distinction betv;een the two. More work in this area is needed to see if the dwarfs and nondwarfs are distinct classes or not. Blue dwarf systems, almost without exception, show single peaked HI profiles. At the same time, evidence is found to indicate that these dwarfs are inclined rotating disk systems. It is quite likely then that they have a strong concentration of HI gas in their central regions. Approximately half of these dwarf systems are found to be wery blue. These may be expected to be underabundant in elements heavier than helium, as is seen in IIZw40 and IZwl8. These dwarf systems are unlikely to be young galaxies. As pointed out by 138

PAGE 146

139 Searle et al . (1973), there are simply too many blue dwarf systems in comparison to nonblue dwarf systems for them to be young galaxies. Many of the properties of blue compact galaxies are such as to make them appear "later" in type than the latest normal galaxies of the Hubble sequence. Blue compacts appear to be more extensive in HI than late type galaxies. Among spiral galaxies, the ratio of HI to optical dimensions appears to increase from early to late types. The blue compacts appear to follow this progression. Blue compact galaxies have smaller corrected HI velocity widths than late type galaxies at comparable luminosities. This follows the progression seen in spiral galaxies of decreasing velocity widths toward later type systems. Blue compact galaxies have slightly greater HI masses than late type galaxies of comparable dimensions. They also appear to have a greater proportion of HI mass than late types. This follows the progression of increasing HI content towards later types. Blue compact galaxies are seen to be as blue as, or bluer than, the latest systems of the Hubble sequence. This follows a trend towards increasing blueness with later types. The ratio of indicative mass-to-luminosity appears to be less for blue compact galaxies than for late types. This follows an apparent trend towards decreasing mass-to-luminosity ratio with increasing type. The HI mass-to-luminosity ratios are approximately the same for blue compacts and for late types. This is consistent with blue compacts being at least as late as the latest types. The rotation curves of several luminous blue compacts have been found to be similar to those of late type spiral galaxies. Their central mass densities may be lower than is found in late type galaxies of comparable luminosity. If so, they would be following the progression towards decreasing central mass

PAGE 147

140 density with later types. Thus, it is quite probable that blue compact galaxies merely extend the Hubble sequence to later types, rather than being a separate type of galaxy. The low surface brightness galaxies show some similarities to blue compacts. Their HI content and their 21 cm profile shapes and widths are similar. These low surface brightness galaxies appear to have greater values of HI mass-to-luminosity ratio and total mass-toluminosity ratio than blue compact galaxies. If we could remove the bright optical cores from blue compact galaxies, they would much more closely resemble the low surface brightness systems, with corresponding increases in HI mass-to-luminosity and total mass-to-luminosity ratios. Thus, it is possible that low surface brightness systems represent the quiescent phase of blue compact galaxies. Much more work is needed in this area before any conclusions can be drawn. Extreme blueness is not restricted to dwarf compact systems, as was predicted by Searle et al. (1973). In addition, two of the metal poor systems studied by Alloin et al. (1978) are not dwarf systems. Thus, what is needed is a model which can explain extreme blueness and low metal abundances, if necessary, in galaxies of normal age and covering a broad range of luminosities and total masses. One possible explanation is that these systems have recently encountered a cloud of primordial gas. Accretion of such a cloud could trigger a burst of star formation and the resulting HII regions would be seen to be metal poor. A second possibility is that uncondensed gas, left over from the initial formation of the galaxy, is still in the process of condensing into the galaxy.

PAGE 148

141 This second possibility seems to fit into the context of a theory of galaxy formation proposed by Larson (1977). Larson proposes that the rate at which the material originally collapsed during the initial stage of galaxy formation determines the type of galaxy formed. Elliptical galaxies formed in regions where the protogalactic gas density was high or the velocity dispersion of the protogalaxies was high. Spiral galaxies formed in regions where the protogalactic gas density was low or where the gas was more quiescent. Elliptical galaxies condensed the most rapidly. They are often found concentrated in clusters, where the densities and the velocities of motion are higher than average. Spiral galaxies collapsed more slowly, allowing time for the gas to condense into a disk. Spiral galaxies tend to avoid dense clusters and are found mostly between clusters or in small groups, where the densities and velocities are less than average, as expected from this model. The outer regions of spiral galaxies will take the longest time to collapse into the galaxy. Thus, in this theory, an infall of low density gas may continue long after the galaxy has formed. In some spiral galaxies, this infall may be continuing into the present epoch, explaining why they still have large amounts of HI gas. Early type spirals would have condensed more rapidly than late type spirals. This could explain the apparent decrease in central mass density from early to late types. In the context of this model, blue compact galaxies would be the systems which collapsed most slowly, and though they are of normal ages, this process of collapse is far from completed. They would then be expected to have significant amounts of gas outside of their optically visible regions. This is just what is indicated by this study and by

PAGE 149

142 the study of IIZw40 by Gottesman and Weliachew (1972). This gas would be condensing into the optical regions, and when the density reaches a high enough level, a burst of star formation could begin. This gas would be primordial and could thus account for the low metal abundances found in some of these systems. The range in observed metal abundances could be explained by various amounts of mixing of this primordial gas with gas recycled from star formation. Thus, in this model, blue compact galaxies are seen as old galaxies which have not yet completed their collapse out of the intergalactic medium. Suggestions for Future Study There are many avenues for the future study of blue compact galaxies. More UBV photometry is needed to provide a larger sample of colors and luminosities. Low resolution spectroscopic surveys of Zwicky compacts could identify many interesting objects. High resolution spectroscopic studies are needed to investigate more fully the metal licity of both the dwarf and the nondwarf types of blue compacts. Optical rotation curves, as were done by O'Connell (1979), would help shed further light onto the structure, kinematics and masses of blue compacts. Large scale, high sensitivity, direct photography could go a long way towards showing any structural similarities between blue compact galaxies and late type galaxies. Surface photometry would also help in understanding these systems. Much more 21 cm spectral line work could be done. There are hundreds of blue compact galaxies remaining which could be observed with large, single dish radio telescopes. Also, high resolution 21 cm line observations are needed to determine the neutral hydrogen dimensions

PAGE 150

143 and distribution in these systems. Such information is crucial to an understanding of the nature of these systems. It should be possible to make such observations with the VLA in the near future. Resolutions on the order of 5 10 arc seconds should be obtainable with this instrument. High resolution HI maps from the VLA will also allow a determination of the velocity fields of the systems, allowing their total masses to be calculated. A better understanding of blue compact galaxies could lead to a better understanding of the formation of galaxies.

PAGE 151

BIBLIOGRAPHY Allen, C. W. 1964, "Astrophysical Quantities," London: University of London Press. Alloin, D., Bergeron, J,, and Pelat, D. 1978, Astron. Astrophys., 70, 141. Arp, H. C. 1966, Ap. J. Suppl . , U, 1. Bergeron, J. 1977, Ap. J., 2V\_, 62. van den Bergh, S. 1975, J.R.A.S.C, 6^, 105. Bottinelli, L. 1971, Astron. Astrophys., 1_0, 437. Bottinelli, L., Chamaraux, P., Gouguenheim, L., and Heidnian, J. 1973. Astron. Astrophys., 2^, 217. Bottinelli, L., Deflot, R. , Gouguenheim, L., and Heidman, J. 1975, Astron. Astrophys., 4j_, 61. Bridle, A. H., Davis, M. M., Fomalont, E. B., Lequeux, J. 1972, A. J., 77, 405. Chamaraux, P. 1977, Astron. Astrophys., 60^, 67. Crow, E. L., Frances, A. P., Maxfield, M. W. 1960, "Statistics Manual," page 241. New York: Dover Publications, Inc. DuPuy, D. L. 1968, P.A.S.P., 80, 29. DuPuy, D. L. 1970, A. J., 75, 1143. Faber, S. M. , and Gallagher, J. S. 1979, "Annual Review of Astronomy and Astrophysics," Vol. 17, page 135. Burbidge, G. , Layzer, D. , and Phillips, J. G., editors, Palo Alto: Annual Reviews, Inq, Fairall, A. P. 1971, M.N.R.A.S., 153, 383. Fairall, A. P. 1978, The Observatory, 98, 1. Fisher, J. R. , and Tully, R. B. 1975, Astron. Astrophys., 44, 151. Forrester, W. T. 1973, Ap. J., 181, 633. 144

PAGE 152

145 Gottesman, S. T., and Weliachew, L. 1972, Ap. Lett., VZ, 63. Haro, G. 1956, Bol . Obs. Tonantzintla y Tacubaya, 1^, 8. Hiltner, W. A., and Iriarte, B. 1958, Ap. J., J_28, 443. Huchra, J. P. 1977, Ap. J. Suppl . , 35, 171. Huchra, J. P. 1979, Private communication. Huchra, J. P., and Sargent, W. L. W. 1973, Ap. J., 186, 433. Kellman, S. A., and Black, D. C. 1973, Ap. J., 1_84, 753. Kerr, F. J. 1968, "Nebulae and Interstellar Matter," Vol. VII of "Stars and Stellar Systems." Middlehurst, B. M. and Aller, L. H., editors, Chicago: University of Chicago Press. Khachikian, E. Y., and Weedman, D. W. 1974, Ap. J., 192, 581. Kormendy, J. 1977, Ap. J., 2H, 359. Larson, R. B. 1977, American Scientist, 65^, 189. Markarian, B. E. 1967, Astrofizika, 3, 55. Markarian, B. E. 1969a, Astrofizika, 5, 443. Markarian, B. E. 1969b, Astrofizika, 5, 581. Markarian, B. E., and Lipovetskii, V. A. 1971, Astrofizika, 7_, 511. Markarian, B. E. , and Lipovetskii, V. A. 1972, Astrofizika, 8, 155. Markarian, B. E. , and Lipovetskii, V. A. 1973, Astrofizika, 9, 487. Markarian, B. E., and Lipovetskii, V. A. 1974, Astrofizika, 1_0, 307. Markarian, B. E., and Lipovetskii, V. A. 1976a, Astrofizika, J2^, 389. Markarian, B. E., and Lipovetskii, V. A. 1976b, Astrofizika, U, 657. Markarian, B. E. , Lipovetskii, V. A., Stepanyan, D. A. 1977a, Astrofizika, 13^, 225. Markarian, B. E., Lipovetskii, V. A., Stepanyan, D. A. 1977b, Astrofizika, j_3, 397. Neugebauer, G., Becklin, E. E. , Oke, J. B., and Searle, L. 1976, Ap. J. 205, 29. Nilson, P. 1973, "Uppsala General Catalogue of Galaxies," Uppsala Astron. Obs. Ann., Band 6.

PAGE 153

146 O'Connell, R. W. 1979, "Proceedings of the Texas Galaxy Conference," in press, Austin: University of Texas Press. O'Connell, R. W. , and Kraft, R. P. 1972, Ap. J., U5, 335. Roberts, M. S. 1969, A. J., 74, 859. Roberts, M. S. 1978, A. J., 83, 1026. Rodgers, A. W., Peterson, B. A., and Harding, P. 1978, Ap. J., 225 , 768. Sargent, W. L. W. 1970a, Ap. J., J^, 765. Sargent, W. L. W. 1970b, Ap. J., 160, 405. Sargent, W. L. W. 1972, Ap. J., 1_73, 7. Sargent, W. L. W. , and Searle, L. 1970, Ap. J. Lett., 162, L155. Searle, L., and Sargent, W. L. W. 1972, Ap. J., 173, 25. Searle, L., Sargent, W. L. W. , and Bagnuola, W. G. 1973, Ap. J., 179 , 427. Shalloway, A. M. , Mauzy, R., Greenhalgh, J., and Weinreb, S. 1968, "Autocorrelation Receiver Model II: Operational Description," NRAO Electronics Division Internal Report No. 75. Shostak, G. S. 1978, Astron. Astrophys., 68, 321. Thuan, T. X., and Seitzer, P. 0. 1979, Ap. J., 231, 327. Tully, R. B., and Fisher, J. R. 1977, Astron. Astrophys., 54, 661. Ulrich, M. J. 1971, Ap. J., 163, 441. de Vaucouleurs, G. , and Corwin, H. G. , Jr. 1979, Private communication, de Vaucouleurs, G. , de Vaucouleurs, A., and Corwin, H. G. , Jr. 1976, "Second Reference Catalogue of Bright Galaxies," Austin: University of Texas Press. Vorontsov-Velyaminov, B. A. 1958, "Atlas and Catalogue of Interacting Galaxies," Moscow: Sternberg Institute, Moscow State University. Vorontsov-Velyaminov, B. A., Krasnogorskaja, A., and Arkipova, V. P. 1962, "Morphological General Catalogue," Moscow: Sternberg Institute, Moscow State University. Zwicky, F. 1964, Ap. J., 140, 1467.

PAGE 154

147 Zwicky, F. 1971, "Catalogue of Selected Compact Galaxies and of PostEruptive Galaxies," Guemligen: F. Zwicky. Zwicky, F., and Herzog, E. 1963, "Catalogue of Galaxies and of Clusters of Galaxies," Vol. II, California Institute of Technology, Switzerland: Offsetdruck L. Speich Zuerich. Zwicky, F., and Herzog, E. 1966, Ibid., Vol. III. Zwicky, F., and Herzog, E. 1968, Ibid., Vol. IV. Zwicky, F., Herzog, E., and Wild, P. 1961, Ibid., Vol. I. Zwicky, F., Karpowicz, M. , and Kowal , C. T. 1965, Ibid., Vol. V. Zwicky, F., and Kowal, C. T. 1968, Ibid., Vol. VI.

PAGE 155

BIOGRAPHICAL SKETCH David Gordon was born on September 14, 1950, in Dallas, Texas. He lived in Irving, Texas, from 1951 until 1963. His family moved to Dallas, Texas, in 1963. He graduated from Thomas Jefferson High School in May, 1969. He attended the University of Arizona from September, 1969, until May, 1973, when he graduated with a Bachelor of Science degree in astronomy. In September, 1973, Mr. Gordon began his graduate studies toward the doctorate at the University of Florida. He has received financial support from the Department of Physics and Astronomy and the Department of Astronomy in the form of a teaching assistantship. He has worked under the supervision of Dr. Stephen T. Gottesman since 1975. The Doctor of Philosophy is expected to be conferred in December, 1979. Mr. Gordon has studied Cuong Nhu Karate since June, 1974, and was promoted to black belt in June, 1979. 148

PAGE 156

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. 1 /. . ; Stephen T. Gottesman, Chairman Associate Professor of Astronomy I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. D TFiomas D. Ca rr Professor of Astronomy 12 'i^<~yLT[ I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Oliver lAssbciate Professor of Astronomy I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. _./ Haywood C. Smith Associate Professor of Astronomy

PAGE 157

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Gary G. Ihaas Assistant Professor of Physics This dissertation was submitted to the Graduate Faculty of the Department of Astronomy in the College of Liberal Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December, 1979 Dean, Graduate School

PAGE 158

UNIVERSITY OF FLORIDA 3 1262 08666 309 2


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EX6GQ97IC_8Q051Q INGEST_TIME 2017-07-14T22:10:30Z PACKAGE UF00099520_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES