Citation
A study of the interstellar medium in NGC 185 and other early-type galaxies

Material Information

Title:
A study of the interstellar medium in NGC 185 and other early-type galaxies
Creator:
Johnson, Douglas William
Publication Date:
Language:
English
Physical Description:
viii, 121 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Antennas ( jstor )
Early type galaxies ( jstor )
Galaxies ( jstor )
Galaxy formation ( jstor )
Hydrogen ( jstor )
In kind support and maintenance ( jstor )
Milky Way Galaxy ( jstor )
Star formation ( jstor )
Telescopes ( jstor )
Velocity ( jstor )
Astronomy thesis Ph. D ( lcsh )
Dissertations, Academic -- Astronomy -- UF ( lcsh )
Galaxies ( lcsh )
Interstellar matter ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1980.
Bibliography:
Includes bibliographical references (leaves 114-119).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Douglas William Johnson.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Douglas William Johnson. 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:
07093208 ( OCLC )
0023422734 ( ALEPH )

Downloads

This item has the following downloads:


Full Text











A STUDY OF THE INTERSTELLAR MEDIUM IN NGC 185
AND OTHER EARLY-TYPE GALAXIES




By

DOUGLAS WILLIAM JOHNSON
















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

1980































to Maryfran













ACKNOWLEDGEMENTS



There are many people who have contributed invaluable resources

and encouragement to me in the completion of the research for this dis-

sertation. Although it is not possible to thank each one individually,

there are some I would like to acknowledge in particular.

The Northeast Regional Data Center is acknowledged for providing

a vigorous and stimulating computing environment. The operators of the

11 m NRAO telescope were very helpful in assisting me during my observing

sessions and the assistance of Telescope Engineer Rick Howard was

especially appreciated. I would also like to thank Dan McGuire for his

assistance during my first observing session.

I thank my supervisory committee members, Dr. Stephen T. Gottesman,

Dr. Thomas D. Carr, Dr. Kwan-Yu Chen, Dr. Gary Ihas, and Dr. William

Weltner, for their attention and interest in my work.

I thank the Physics Department, and especially Dr. Richard Garrett,

for the assistantships that have allowed me to pursue this course of

study.

The Division of Sponsored Research is thanked for its Seed Money

Grant Competition, providing financial support for Dr. Gottesman and me

during the course of our investigations. The Graduate School's Supple-

mentary Fellowship for 1979-80 was also greatly appreciated.

Finally, I wish to thank Steve Gottesman and his family for their

efforts to make Maryfran's and my stay in Gainesville a wonderful time



iii







in our lives. Steve's assistance and discussions with me (to say

nothing of his witticisms) were above and beyond the call of duty.

I thank my parents, Bill and Mary Ann, for their support and love

and good humor throughout the years. But I reserve perhaps my sincerest

appreciation for Maryfran and the "kids" (Smokey, Phantom, Harpo, and

Marble) for making it all worthwhile.










































iv











TABLE OF CONTENTS


Page
ACKNOWLEDGEMENTS . . . . . . . . . . iii

ABSTRACT . . . . . . . . . . . . vii

CHAPTER

I INTRODUCTION . . . . . . . .... . . 1
The Nature of Early-Type Galaxies . . . . . 1
Galaxy Classifying Schemes ............ . 2
The Morphology of Elliptical Galaxies . . . 3
The Morphology of Lenticular (SO) Galaxies. . . 6
The Formation and Evolution of Early-Type Galaxies . 7
Formation . . . . . . . .. . 7
Mass Accretion . . . . . .... . . . 11
Mass Loss Due to Stellar Evolution . . . ... 13
Stripping Mechanisms . . . . . . .. 15
Internally Driven Winds . . . . . .. . 15
Ram-Jet Stripping by an Intracluster Medium . . 18
Fate of Retained Gas . . . . . . . . . 23
Supermassive Objects in the Nucleus . . . . 23
Cyclic Bursts of Star Formation .. ....... 25
Continual Star Formation with a Skewed Initial
Mass Function . . . . . . . . 27
Statement of Dissertation Problem . . . . .. 29

II THE OBSERVATIONS . . . . . . . . . 31

Carbon Monoxide Observations at Kitt Peak, AZ. .. .... 31
Telescope Description . . . . . . . 31
Data Reduction Techniques . . . . ... . 34
Data Presentation . . . . . .. . 35
Neutral Hydrogen Observations at Green Bank, WV. .. .. 36
Telescope Description . . . . . . . 36
Data Reduction Techniques . . . . . . 38

III RESULTS . . . . . . . . . . 39

Carbon Monoxide. . . . . . . . . . 39
Positive Result in NGC 185 . . . . .. 39
Negative Results . . . . . . . .. 48
Neutral Hydrogen . . . . . . . . . 51

IV DISCUSSION . . . . . . . . . . . 68

NGC 185 .. . . . . . . . . .68
NGC 205. . . . . . . . . . . . 84


v







Page

Negative CO Results . . . . . . . . 87
Neutral Hydrogen and NGC 185 .. . . . . . . 93

V SUMMARY. . . . . . . . . ... ...... .96

APPENDIX

I CALIBRATION THEORY FOR CO OBSERVATIONS . . . . .. 100

II THE PHYSICS OF CO SPECTRAL LINE CALCULATIONS ...... 105

III THE GEOMETRY IN AN INCLINED DISK . . . . . .. 111

REFERENCES. . . . . . . . . ... ....... 114

BIOGRAPHICAL SKETCH . . . . . . . . ... .. ... 120






































vi











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 STUDY OF THE INTERSTELLAR MEDIUM IN NGC 185
AND OTHER EARLY-TYPE GALAXIES

By

Douglas William Johnson

August 1980

Chairman: Stephen T. Gottesman
Major Department: Astronomy

The question of an interstellar medium in early-type galaxies is

considered in light of the small amounts of gas detected as neutral

hydrogen (HI). It is apparent that there is some method of removal or

reprocessing that keeps the interstellar medium of early-type systems

gas and dust free in spite of mass loss from normal stellar evolution.

A detection of 12CO is presented for the dwarf elliptical system

NGC 185. The mechanisms of line formation of the J = 1-*0 transition

strongly imply that the emitting region is in a state of gravitational

collapse. These observations are consistent with the observed dust

content of the galaxy and its blue, presumably young stellar population.

It is nearly certain that the galaxy is reprocessing its interstellar

medium via star formation.

The radial velocity of the CO cloud can be combined with the

velocity data for two planetary nebulae within the system to allow

rough mass calculations. Coupling this information with luminosity

data indicates that the mass-to-luminosity ratio of the nuclear regions

(within 1 arcminute) is between 5 and 18, with a value near 8 being the

most probable.

vii







Observations of 10 other early-type systems are also presented

and discussed. The negative results imply that the gas is either

clumped and thus optically thick or has been removed from the system

through a galactic wind, ram-jet stripping, or has been consumed by

star formation.

The nature of the star formation must be somewhat different than

that in our own galaxy. The high-mass end of the initial mass function

for star formation would result in bluer colors than observed, the

star formation must be skewed towards the low-mass stars to be effec-

tive yet unobserved. Theoretical arguments that this is possible are

advanced, but more sensitive and highly resolved CO observations are

necessary to observe directly this scale of star formation.

Neutral hydrogen observations of NGC 185 obtained with the NRAO

43 m telescope are presented and discussed. There is apparently low-

level (20-40 mK) high-velocity hydrogen in the region of NGC 185. The

most likely source of the material is the Magellanic Stream which ter-

minates in this area. Superposed on this high-velocity material, at

the same velocity, is an excess of about 5 mK of HI at the location of

NGC 185. This was detected by using "off" spectra 1-2 beamwidths from

the galaxy in the four cardinal directions. The observations cannot

distinguish between an enhanced high-velocity feature projected onto

the galaxy and genuine emission from the galaxy itself, but the results

are very suggestive and should be followed up with observations of

greater resolution.








viii












CHAPTER I

INTRODUCTION


The Nature of Early-Type Galaxies

Since the discovery that the "spiral nebulae" were indeed island

universes somewhat like our own, the effort to study, classify, and

dissect them has been increasing with remarkable speed. In large

measure this is due as much to the expansion of the accessible electro-

magnetic spectrum as to the increasing sensitivity of instruments

within each spectral window.

The basic observational technique used for this dissertation, radio

astronomy, dates from the 1932 observations of the Milky Way by Karl

Jansky. The achievements with this relatively new technique have

accumulated steadily over the past half century and it is now recog-

nized as an invaluable tool in studying the'universe and its contents.

One of the most attractive characteristics of radio astronomy is

that it complements the endeavors of the oldest technique, optical

astronomy. The analysis of visible light almost always entails objects

with a temperature of at least 2000 K and the vast majority of objects

this hot are stars. Radio astronomy, utilizing much less energetic

quanta, is sensitive to objects of several hundred degrees or less.

Typically this is primarily the interstellar gas that fills the space

between the stars. There are important exceptions to this rule

(synchrotron radiation, thermal bremstrahlung, etc.) but it illustrates

well the complementary nature of optical and radio astronomy.



-1-




-2-


Galaxy Classifying Schemes


Early studies of galaxies made it clear that the general morphology

of these immense stellar systems allowed them to be grouped into a

relatively small number of types. The most successful early venture

was Hubble's "tuning fork" diagram published in 1936, and reproduced

here in Figure 1. The spherical systems (EO) are at one end with pro-

gressively flatter (E1-E7) systems leading up to a split in the diagram.

SO galaxies (also called lenticular galaxies) occupy the vertex of the

fork because Hubble believed that they were transitional systems. They

contain prominent elliptical bulges as well as a conspicuous disk

component. It is interesting to note that the nature of SO galaxies

is still being vigorously debated.










ELLIPTICAL NESULAE AO- SC

Eo E3 E7 i, s SB
SBc


SP1RALSC





Figure 1. Hubble's "tuning fork" diagram of the classification of
galaxies.




-3-


Forming the tines of the fork are two parallel sequences of spiral

galaxies; one with a bar, the other without. The trend along the tines

is from tightly wound spiral arms (Sa or SBa) to looser, more open

arms at the end (Sb-Sc or SBb-SBc).

In addition to these major players in the drama, some 3% of all

galaxies are irregular, possessing no dominant symmetrical structural

features.

Many additions and modifications have been made to this initial clas-

sifying scheme (Hubble 1936, Morgan 1958 and 1979, de Vaucouleurs 1959,

van den Bergh 1960a and b, Sandage 1961) but it has remained remarkably

unchanged over the years. In large measure the modifications are to take

into account the more extensive information available due to more sensi-

tive equipment and increasing access to other spectral regions.

The following sections describe the elliptical and lenticular

galaxies in more detail and lay the groundwork for the statement of the

thesis problem in the final section of Chapter I. For historical reasons

both elliptical and lenticular systems are commonly known as "early-type"

galaxies.


The Morphology of Elliptical Galaxies


As the name implies, the elliptical galaxies are characterized by

elliptical isophotes. The stellar population usually appears to be

well-evolved with little or no interstellar gas or dust.

The degree of ellipticity E is defined to be (a b)/a (where a

and b are the semi-major and semi-minor axes, respectively). The range

observed is 0.0-0.7 (EO-E7) with EO-E1 the most common and decreasing

in frequency at the flatter end of the range.




-4-


Taking into account the statistics of random projection on the

sky (for we are viewing a two-dimensional projection of a three-

dimensional object) it appears that the ellipticals are distributed

normally about a mean of E3.6 (e = 0.36) with a dispersion of 0.11

(de Vaucouleurs 1977). It appears that true EO and E5.5 (de Vaucouleurs

contends that E5.5 is the flattest bona fide elliptical) are rela-

tively rare.

The lack of flat systems, usually considered to be a dynamical

effect caused by instabilities in thin disks, suggests that perhaps

spiral density waves and the attendant star formation are suppressed

in disks of sufficient thickness.

The origin of the flatness of elliptical systems has long been

thought a natural consequence of rotation. The greater the rotational

velocity, the greater the degree of flattening. This of course implies

that the three-dimensional figure is an oblate ellipsoid (polar diameter

smaller than the equatorial diameter).

In recent years several rotation curves of elliptical galaxies

have been published (Bertola and Cappaccioli 1975, Illingworth 1977,

Peterson 1978, Sargent et al. 1978, and Young et al. 1978a)which cast

strong doubt on the validity of this simple approach. The small ob-

served ratio of the maximum rotational velocity to the central dis-

persion velocity mitigates strongly against models which use isotropic

velocity distributions and either oblate or prolate ellipsoids

(Schechter and Gunn 1979). It appears necessary to use both anisotropic

velocity distributions as well as rotation to account for the observed

rotation curves (Binney 1978, Schechter and Gunn 1979).




-5-


The origin of the anisotropy is still not clear but the most

likely source is remnant anisotropy from the collapse phase of the

galaxies' formation.

A further difficulty in our understanding of elliptical galaxies

is the existence of extreme population I ingredients in a significant

number of systems. Specifically:

OB clusters are observable in NGC 185 and 205 (Hodge 1963
and 1973)

ionized gas is seen in the nuclei of at least 15% of all ellipti-
cals (Osterbrock 1960 and 1962)

neutral hydrogen has been detected in at least 8 elliptical
systems:

NGC 1052 (Knapp et al. 1978b)
NGC 2974 (Bottinelli and Gouguenheim 1979b)
NGC 3904 (Bottinelli and Gouguenheim 1977b)
NGC 3962 (Bottinelli and Gouguenheim 1979a)
NGC 4105 (Bottinelli and Gouguenheim 1979b)
NGC 4278 (Gallagher et al. 1977)
NGC 4636 (Knapp et al. 1978a)
NGC 5846 (Bottinelli and Gouguenheim 1979b)

The presence of population I material is unusual for systems thought

to have ended all star formation long ago. Some possible explanations

are that the material was accreted relatively recently and thus it is

not representative of an elliptical galaxy's normal evolution, or that

through normal processes of stellar evolution the material was shed by

the stars and is observable in various forms today.

The thrust of the foregoing observations is that, as a class,

elliptical galaxies are not as well-understood as was earlier believed.

The apparently relaxed stellar distribution is most likely not relaxed

at all, but still contains velocity anisotropies which strongly in-

fluence the shape of the galaxy. It is disturbingly common for the

smooth isophotes to be blemished with dust obscuration or some other




-6-


form of detectable population I material. .The CO detection described

in Chapter III is another example of the incongruities present in

elliptical systems.


The Morphology of Lenticular (SO) Galaxies


The lenticular systems, occupying the vertex of Hubble's tuning

fork diagram, are as perplexing today as ever. The essential features

of these galaxies are an elliptical bulge and. a disk, both apparently

composed of population II stars. The disks are devoid of spiral arms

and the systems lack the obvious star formation proceeding in spiral

galaxies.

Suggestions as to the origin of the lenticulars range from van den

Bergh's proposal (1976b)that they are part of a complete sequence of

gas-poor, "anemic" spirals parallel to the S and SB types to the more

recent suggestion (Gunn and Gott 1972, Gisler 1979, and references

therein) that they are normal spirals which have been stripped of their

gas and dust by an interaction with an intracluster medium.

Much work has been done in recent years on these systems and

some interesting observations have been reported. Neutral hydrogen

observations by Balick et al. (1976) and a number of others place the

lenticulars roughly midway between the Sa-SBa and E systems. This

evidence supports the contention that they form another sequence of

spirals much like the normal sequence, only gas-poor.

Further, Burstein (1979a,b,c) finds significant differences in

the bulge-to-disk ratios of lenticulars and spirals, as well as

"thick" disks in lenticulars and not spirals. These are structural




-7-


features that are difficult to affect by stripping mechanisms, thus

supporting the parallel sequence hypothesis.

On the other hand, several people have investigated the spatial

distribution of lenticular systems and find that they are concentrated

within clusters, strongly suggesting that their current environment

is crucial in their.formation.

Regardless of the mechanism for the origin of the lenticular

systems they have been included in this study because they share

several significant properties with elliptical galaxies. The stellar

populations appear quite similar and both have a lack of dust and gas

within their interstellar media.



The Formation and Evolution of Early-Type Galaxies


Formation


The general outline of galaxy formation, and elliptical galaxy

formation in particular, is understood only in its coarsest features.

This section presents the scenario most widely agreed upon with emphasis

on early-type systems. Significant gaps in the scenario are also noted

with various suggestions that may, in the future, fill them.

In the study of galaxy formation one is inevitably forced to con-

sider earlier epochs. In astronomy this can be done easily by observing

more distant objects. The scale of the universe is such that the time

radiation has taken to arrive here on earth represents a significant

fraction of the object's existence. Continuing the effort to fainter

(more distant) objects gradually crosses over to cosmology and the

study of the origin of the universe.




-8-


Perhaps the most influential discovery in cosmology (and also

bearing heavily on galaxy formation) has been the 2.7 K microwave back-

ground. The existence of nearly isotropic (see Cheng et al. 1979 for

a discussion of a dipole anisotropy attributed to the earth's motion

with respect to the background) homogeneous radiation with an apparent

blackbody spectrum strongly constrains galaxy formation theories.

The cosmic background radiation is almost universally believed to

be the remnant from the era of decoupling; the time when the universe

had cooled enough to allow electrons and protons to recombine (about

3000 K). The effect of the formation of neutral hydrogen is to reduce

drastically the opacity of the matter with respect to the radiation.

Before recombination Thompson scattering of radiation off electrons

coupled very strongly the matter with the radiation, they were kept in

thermal equilibrium and cooled together. After decoupling the two com-

ponents evolved essentially independently of one another and the 2.7 K

background seen today is the remnant of the radiation component.

This forms the basis of the expectation that the early universe

was homogeneous and isotropic. Actually one can only say that it is

homogeneous and isotropic to at least the smallest scale observed,

currently that means the background is smooth to within 1 mK on a scale

of 70 (Smoot and Lubin 1979, see Sunyaev 1977 for a discussion of

fluctuations).

The condensations that have eventually evolved into clusters of

galaxies, galaxies and stars must have occurred during a later epoch.

The nature of these perturbations is currently a topic of great

interest. Several theories have been advanced to account for the

processes by which the perturbations form and grow (see Gott 1977 and




-9-


Field 1975 for reviews) but it will likely be many years before any

scenario is convincing and widely accepted.

Lynden-Bell (1967) proposed a theory in which the essential feature

of a galaxy's structure is determined by the timescale of collapse from

the background compared to the timescale of star formation. It can

be well illustrated by considering the various structural components

of our own galaxy.

Our galaxy consists of a disk of gas, dust, and relatively young

stars. The gas and dust are continually undergoing star formation

in which hot bright stars appear to be preferentially formed along

spiral arms. Spiral shock phenomena may be important in regulating

the star formation but regardless of the details there is a continuing

processing of gas and dust into stars.

Also present is a spheroidal halo which contains little if any

gas and dust. It is composed primarily of old stars and consequently

evolves only as fast as the stars that compose it.

Lynden-Bell's theory suggests that these structural features are

formed by varying rates of star formation occurring in the collapsing

proto-galactic cloud. Various components form as it separates out from

the cosmic expansion and begins to contract under its gravitational

force.

The novel feature of the theory is that it can explain how the

halo can relax in the time allowed. Essentially, the presence of a

changing gravitational potential will permit relaxation of the stellar

system much faster than would two-body encounters.

The stellar population of the halo is entirely old stars with no

gas or dust. During the collapse the stars form and are then




-10-


dynamically independent of the remaining material. That is, the

stars reflect the velocity dispersion of the cloud at the time of

their formation.

As the material collapses further, star formation proceeds and

eventually the gas becomes dense enough that damping forces become

important. At this stage the gas gives up a great deal of kinetic

energy to dissipative heating and, because of its angular momentum,

settles down into a disk. Within this disk various processes, especi-

ally star formation, continue to the present giving spiral galaxies

their distinctive optical appearance.

It takes only a minor modification of this scenario to produce an

elliptical galaxy. If star formation has proceeded to completion before

the dissipative effects can form a disk then a system will result which

has no disk component and is essentially a halo population of stars.

Models which can produce the observed luminosity profiles of ellipti-

cals have been constructed (Larson and Tinsley 1974) and add credence

to this basic concept.

Within this framework of galaxy formation many details are still

to be worked out. As previously mentioned, the nature of the initial

perturbations is not at all understood. Also the processes of frag-

mentation and collapse are poorly understood (Field 1975). There is

even a question of whether gravitational instabilities or turbulence

is responsible for the necessary condensations. Jones (1976) provides

an excellent review of these and other problems in the study of galaxy

formation.

Since the early star formation was apparently so efficient as to

consume all the primordial gas, one might not expect any continuing




-11-


star formation in elliptical galaxies. It will be shown, however,

that other processes operate which alter this simple picture and lead

to radically different expectations. The most important effects are

mass loss from the normal evolutionary processes of the stellar popula-

tion and perhaps accretion of extragalatic material; both are discussed

in the following sections.



Mass Accretion


The contention that early-type galaxies, ellipticals in particular,

accrete material is relatively new. The motivation is to explain the

cD galaxies (giant ellipticals usually located in the center of rich

clusters and described by Bautz and Morgan 1970) that are often strong

radio sources. The accretion described here is full scale cannibalism

of other galaxies during close encounters (Ostriker and Tremaine 1975,

White 1976, Ostriker and Hausman 1977, Hausman and Ostriker 1978).

The idea can be linked to cluster types as described by Bautz and

Morgan (1970) and Oemler (1974). This cluster classification scheme

ranks the clusters based on their richness. Observationally it is

found that the densest clusters (Type I in the notation of Bautz and

Morgan 1970) often contain giant elliptical galaxies near their center.

Further, the cD galaxies are often radio sources that are widely

suspected to be caused by material falling into a massive object.

There are several other pieces of circumstantial evidence that indi-

cate that this process may indeed be significant in the evolution of

elliptical galaxies (see Ostriker 1977 and Hausman and Ostriker 1978

for details).




-12-


A more moderate form of mass accretion is suggested by several

authors (Bottinelli and Gouguenheim 1977a, Gallagher et al. 1977, and

Knapp, Kerr, and Williams 1978c)to explain the inclined disk of

NGC 4278. The differing directions of the angular momentum of the

stellar component of the galaxy and the neutral hydrogen make an

internal origin of the matter difficult to believe.

A similar situation occurs in NGC 1052 (Knapp, Faber, and

Gallagher 1978a, Fosbury et al. 1978, Reif, Mebold, and Goss 1978) and

the accretion of an intergalactic HI cloud is suggested. The major

objection to this hypothesis is the lack of sufficiently massive clouds

available for accretion.

The results of Mathewson et al. (1975) purporting to find HI

clouds in the Sculptor group have been disputed by Haynes and Roberts

(1979). The latter group contend that the material is a portion of the

Magellanic Stream. Further, Lo and Sargent (1979) have searched nearby

groups for detached HI clouds and find none more massive than

4 x 107 Me.

A number of other authors have examined clusters of galaxies for

HI emission (Haynes et al. 1978, Baan et al. 1978) while others have

examined the line of sight to quasars for HI absorption (Roberts and

Steigerwald 1977, Shostak 1978). No isolated HI clouds are seen in

emission in the clusters and the HI absorption measurements show that

large clouds of neutral hydrogen are almost never seen outside

galaxies.

Thus it appears difficult to reconcile the several times 108 M

of HI found in NGC 4278 and 1052 with the dearth of sufficiently

massive candidates for accretion. Silk and Norman (1979) propose an




-13-


alternative hypothesis, the accretion of gas-rich dwarf galaxies.

They find that the gas component of the dwarfs will lose energy through

dissipation and fall to the central regions of the accreting galaxy.

The infalling gas, depending on the individual cloud mass, may either

form stars or continue to fall into the nucleus where it may fuel a

radio source. The stars will also be incorporated into the accreting

galaxy but with less visible effects.

Silk and Norman (1979) also consider the interaction of a Mathews

and Baker (1971) type wind and the infalling material. If the amount

of this material is sufficiently large the resulting supernovae (from

the high-mass stars formed) will help in driving the galactic wind.

However, an enhanced wind has the effect of inhibiting mass infall and

the process slows itself. The net effect may be for star formation to

proceed in cyclical bursts--a notion also suggested by van den Bergh

(1975) in a somewhat different context.

Both of the mass accretion processes described so far deal with

normal or giant cD elliptical galaxies. In order to be effective in

capturing and assimilating other systems the accreting galaxy must be

large. The evolutionary mechanism discussed in the following section,

mass loss from stellar evolution, operates in all systems. This in-

cludes the dwarf ellipticals NGC 185 and 205 considered in greater

detail in Chapter III.


Mass Loss Due to Stellar Evolution

It has recently been appreciated that the normal evolution of

stars in an early-type galaxy will be a source of interstellar material.




-14-


Small mass stars have stellar winds, Mira variables are known to eject

mass during certain stages, Type I supernovae occur in population II

stars, and planetary nebulae have been observed in early-type galaxies

of the local group.

The calculation of the contribution by stellar evolution to the

interstellar medium (ISM) of early-type galaxies depends more on theo-

retical estimates than observational evidence. To date, the most impor-

tant observational evidence is the detection of planetary nebulae in

nearby dwarf ellipticals (including NGC 185) by Ford and Jenner (1975).

Considering the uncertainties in the observations, the observed planetary

birthrate of 0.012 yr-1 (109 L )-1 agrees well with Larson and Tinsley's

(1974) estimate of 0.05 yr-1 (109 Le)-1
Following the reasoning of Faber and Gallagher (1976) and adopting

a mass per planetary of 0.2 Me results in a mass loss rate of 0.010

yr-1 (10 L )-1 from planetaries. Consideration of Mira-type variables

leads to a final assumed mass ejection rate of 0.015 yr-1 (109 Le)-.

The conservative nature of this calculation is apparent when one

considers that the present mass loss rate is certainly lower than that

of earlier epochs. This is primarily because any high-mass stars

would have evolved quickly and cycled their mass back to the ISM early

in the galaxy's evolution.

Further, the contribution of mass from Type I supernovae (ap-

parently confined to population II stars, Tammann 1974) and Type II

supernovae (massive progenitors) earlier in the galaxy's evolution

have been ignored. Even this conservative approach leads to contra-

dictions in the ISM of early-type galaxies after 10 -1010 years (Faber

and Gallagher 1976).




-15-


Stripping Mechanisms


The various efforts to determine how interstellar material shed

by stars can be removed from early-type systems can be divided into

two classes. The first is an internally driven "galactic wind" and

second is ram-jet stripping by an intracluster medium. Both of these

processes will be discussed in the following sections.



Internally Driven Winds


The possibility of a galactic wind driven by an internally driven

energy source was suggested by Johnson and Axford (1971) and considered

more quantitatively by Mathews and Baker (1971). In essence the mech-

anism operates by coupling the energy from Type I supernovae to the

general interstellar medium. The addition of this high-energy

(8 x 10 K) low-mass component significantly heats the interstellar

material to a high enough temperature to escape from the system by

evaporation. A later study by Coleman and Worden (1977) shows that

the energy released by flare stars is by itself enough to drive a

galactic wind of this type.

The parameters that are most important in the establishment and

maintenance of a galactic wind are the Type I supernovae rate, the

energy output from each supernova, the efficiency of the coupling of

the supernova's energy to the ISM, and the amount of "pre-heating"

of the ISM by the velocity dispersion within the galaxy. Each of

these quantities are known to probably a factor of 2 at best and in

certain instances various authors disagree by factors of 10 or more.




-16-


For example, Mathews and Baker (1971). assume a coupling effi-

ciency between the expanding supernova shell and the ISM of 1; that

is, all the kinetic energy of the supernova is converted into thermal

energy of the ISM. Gisler (1976) takes exception to this number and

notes that Larson (1974) uses an efficiency of 0.1. Given the various

uncertainties it appears that while a galactic wind will most likely

prevail in some instances, perhaps even a majority of elliptical

galaxies, there are cases in which it simply does not operate.

Indeed, Mathews and Baker (1971) find solutions in which a wind is

not supported and the material collapses to the center of the system.

They further propose that the hot, ionized gas will only be able to

form massive objects, thus linking the lack of a galactic wind to the

formation of radio sources in early-type galaxies.

Again, Gisler (1976) points out an inconsistency in this line of

reasoning. From observations one finds that strong radio sources were

more common in earlier epochs. Gisler notes that the earlier stellar

content of ellipticals is more likely to produce supernova. This

follows from the observation that only Type I supernovae occur in popu-

lation II (old) stars and the precursors are probably low mass stars

(Tammann 1974). In the earlier stages of an elliptical's life the

supernova rate can only be augmented by Type II supernovae (whose

progenitors are young, massive stars). In addition it is at the

earlier epochs that the galaxy will not have had time to collect a

significant amount of gas from the evolution of its stellar component.

For these two reasons it would appear that the ellipticals are better

able to support a galactic wind at earlier epochs--just the period

when the greatest fraction must also be radio sources.




-17-


Faber and Gallagher (1976) have considered the problem of a

galactic wind from a slightly different approach. Since the initial

heating of the ISM is through collision of the gas clouds shed by

stars in the process of normal evolution, the velocity dispersion of

the stars can be used as an indicator of the stellar equivalent temper-

ature, Ts. From Mathews and Baker (1971) Ts = v2 mH/6k where vs is

the stellar dispersion velocity and mH is the mass of the proton, and k

is Boltzmann's constant. Along with the condition that the supernova-

induced rise in temperature must roughly double the kinetic energy of

the gas to remove it from the system to infinity, one finds


"sn sn ssTs (1)

where asn and as are the specific rates of mass injection by super-

novae and stars, respectively, Tsn and Ts are the equivalent tempera-

tures of the supernovae and the stars.

Using the values quoted in Faber and Gallagher (1976) as =

5 x 1020 sec-1 and sn Tsn = 1.6 x 1012 K sec- yields T = 3.7 x 10

K or vs f 1260 km/sec. The line-of-sight velocity dispersion is

< 1260//T = 730 km/sec (assuming 3 dimensional isotropy of the velocity

dispersion), which exceeds by a factor of two the largest velocity

dispersions measured (Faber and Jackson 1975).

This calculation would make it appear that virtually all elliptical

systems should have a galactic wind scouring the ISM of any material.

However, Gisler (1976) recomputes this same quantity, taking into

account the 0.1 efficiency of energy coupling found by Larson (1974),

and finds that the maximum velocity dispersion that will still allow

the supernovae to double the energy is about 200 km/sec, a number

comparable to the observed velocity dispersions. In other words,





-18-


if a galaxy has a velocity dispersion greater than 200 km/sec then the

supernovae contribution will not be able to double the kinetic energy

of the ISM and a galactic wind cannot be established.

Even more damaging to the galactic wind hypothesis is the de-

tection of HI in any elliptical. In order for the wind to operate

the ISM must be hot (, 107 K). All of the gas in a galaxy would thus

be ionized and according to Mathews and Baker (1971) quite unobservable

by present techniques.

A final argument against the universal existence of galactic winds

is that if all other conditions were the same, one would expect the

more spherical systems to be better able to support a galactic wind.

The reasoning is that the spherical systems have less surface area per

unit volume through which to radiate excess energy, keeping the ISM

as hot as possible.

Contradicting this expectation, the neutral hydrogen observations

of 8 elliptical systems show detections significantly skewed towards

the more spherical galaxies. The systems detected in neutral hydrogen

have the following classifications: 2-EO, 2-E1, 1-E2, 1-E3, 2-E4.

Conspicuously absent are the flatter systems that one would expect to

be better able to radiate energy away and thus retain their ISM.


Ram-Jet Stripping by an Intracluster Medium


The proposal that ram-jet stripping of an ISM could be significant

in the evolution of a system was treated first by Gunn and Gott (1972).

Later, more sophisticated treatments by Tarter (1975), Gisler (1976),

and Lea and De Young (1976) all support the notion that stripping can

be an effective process.




-19-


The thrust of much work in this area has been to determine if

SOs can be formed by stripping spiral galaxies of their gas and dust

(Gisler 1979). This would quench star formation and significantly change

the optical appearance of the galaxy.

Another development that has spurred interest in the interaction

of an ISM with an intergalactic medium (IGM) is the discovery of

head-tail radio galaxies. The most straightforward explanation of

this phenomenon being just such an interaction.

In spite of the varied motivations for these studies many of the

numerical simulations are directly applicable to the analysis of an

elliptical system passing through an IGM.

The primary results of these studies are to confirm that under

appropriate conditions there is an effective sweeping out of material

from the galaxy. The procedures and model parameters used to arrive

at this conclusion vary substantially for the different experiments,

all .agree however, that some material tightly bound near the nucleus

may be retained. Gisler (1979) explores the situation further and

finds that the rate of gas replenishment is important, possibly stopping

the stripping effect entirely if it is high enough.

In spite of this it appears that stripping can be at least par-

tially effective over a broad range of galaxy velocities and IGM

densities. This finding agrees well with the observation that a large

fraction of galaxies in rich clusters are SOs and ellipticals (Oemler

1974, 1977).

It would seem also that the evidence of a positive correlation

between X-ray luminosity of clusters (presumably from a hot intracluster

component) and SO/spiral ratios (Tytler and Vidal 1978) argues strongly




-20-


that the cluster environment does indeed have a significant influence

on the structure and evolution of its component galaxies.

An evolutionary effect may also have been observed by Butcher

and Oemler (1978). Their study found that a rich cluster observed at

a redshift of 0.4 contains many more blue galaxies than a similar rich

cluster observed in the current epoch. Their conclusion is that as

late as 4 x 10 years ago the galaxies in this cluster had not yet

been stripped of their ISM. They were consequently undergoing at

least moderate star formation, thus producing the blue colors observed.

One further piece of evidence that fits in quite well with the

general hypothesis of ram-jet stripping is the common coincidence of a

strong radio galaxy at the center of a dense cluster (McHardy 1974,

Guthrie 1974, and Riley 1975). The argument in this case is that the

central galaxies have a small velocity with respect to the intra-

cluster medium and will be more likely to retain gas shed by its

stellar population. The material collapses to the center of the

galaxy, apparently forming a massive object and producing the ob-

served radio source.

In view of the variety of indications that imply a substantial

interaction between an ISM and an intracluster medium it seems clear

that environmental factors can be important in the evolution and

structure of galaxies in clusters. But in the particular case of

the detected ellipticals NGC 4278 and 4636 there is reason to believe

the IGM is unimportant.

These are the only two galaxies which have an HI distribution

that is extended enough to map. In both instances the HI distribution

appears to be considerably wider than the photometric diameter of the




-21-


system (Knapp et al. 1978a,Bottinelli and Gouguenheim 1979a). More

importantly, the distributions appear to be reasonably symmetric, a

condition difficult to reconcile with stripping or partial stripping

by an intracluster medium. This discrepancy is pursued further in

Chapter IV.

Other evidence that ram-jet stripping may not be as effective as

the numerical analyses indicate is found in the Hercules cluster, a

rather loose cluster composed almost entirely of spiral galaxies.

The inconsistency is that it is also an X-ray source. Our current

understanding of cluster X-ray sources necessitates a hot (a 107 K)

intracluster medium as the origin of the radiation. How the Hercules

galaxies have remained spirals and not been stripped is not understood

within the framework of current research. Even more perplexing is the

origin of the intracluster gas, it is usually thought to have been the

gas removed from the galaxies.

A recent statistical study of cluster morphology by Gisler (1980)

shows that the anticipated presence of Sc galaxies in clusters is not

found. The Sc galaxies are expected to be highly resistant to having

their ISM swept because they have a high rate of gas replenishment

(Gisler 1979). The apparent underabundance in rich clusters indicates

that ram-jet stripping cannot be the dominant mode of SO production.

Dressler (1980) comes to similar conclusion based on a study of

the morphology of the galaxies in 55 rich clusters. He finds a

significant number of SO systems in clusters which have too low a

density to accomplish any stripping. Further, the study finds a

difference between the bulge-to-disk ratios of spirals and SOs as well

as a tendency towards thicker disks in SO systems. From these





* 4




-22-


observations Dressier (1980) concludes that spirals do not evolve

into SOs. He concludes that the galaxy types are affected more by the

initial conditions at the time of their formation than by environmental

factors such as ram-jet stripping.




-23-


Fate of Retained Gas


Supermassive Objects in the Nucleus


Mathews and Baker (1971) suggested that if their galactic wind

should fail, the gas in the system would fall to the nucleus in an

ionized state. They further argue that the Jeans radius


R r TrkT 2
J= L161MG(p + p) (2)


where T is the temperature, v is the molecular mass, G is the gravita-

tional constant, p and p, are the densities of the gas and stars,

respectively, and k is Boltzmann's constant is determined primarily

by the stellar density. That is, as the gas collapses it responds to

the gravitational field of the stars. This will continue until the gas

becomes more dense than the stellar component. The gas is dense and

collapsing quickly at this stage and Mathews and Baker suggest that

there may not be enough time for fragmentation to take place. The

collapsing material then forms a massive object rather than fragmenting

and forming stars with a normal distribution of masses.

Another argument for supermassive objects is the observation that

cD galaxies with radio emission are often located in the center of

dense, rich clusters. The evidence is largely circumstantial but if

ram-jet stripping is important in the evolution of ellipticals then

it follows that the central members of a cluster will be least affected

by this mechanism. Of course the step from ineffective ram-jet

stripping to a supermassive object in the galaxy's nucleus is by no

means secure. It rests on the assumption that the retained material




-24-


either forms the supermassive object or at least provides fuel for the

radio emission.

These arguments actually rest on much firmer ground due to recent

work on the velocity dispersions and light distributions within the

nuclei of supergiant cD galaxies found in the centers of rich clusters.

Young et al. (1978b)obtained luminosity profiles of the supergiant

elliptical NGC 4486 (M87) which, when examined with the velocity dis-

persions determined by Sargent et al. (1978), show that the nucleus

contains a massive dark object. The nature of the dark object cannot

yet uniquely be determined, but it must contain 5 x 109 M of material

and have a radius less than or equal to 100 parsecs (pc). Young et al.

also determined that the mass-to-luminosity ratio must be greater than

60. Several possibilities are advanced but Young et al. find the most

plausible to be a massive black hole of 5 x 109 Me. The attraction

of this hypothesis is that the 1042 erg sec-1 energy output of NGC 4486

can be explained by supposing a mass infall of about 10-2 M yr-1 with

only a 0.002 conversion efficiency into radiation.

De Vaucouleurs and Nieto (1979) confirmed the results of the

Young et al. (1978b)study and found essentially similar results for

the dark mass at the nucleus. The earlier results were obtained with

a CCD (charge-coupled device) camera, while the work by de Vaucouleurs

and Nieto was with more conventional photographic and photoelectric

photometry.

Young et al. (1979) also examined the luminosity profiles of NGC

4874, 4889, and 6251 and found that only NGC 6251 requires a supermassive

object at its center to fit the data. Further, only NGC 6251 is a radio

source amongst the three.




-25-


It appears from these observations that there may be a correla-

tion between radio galaxies in the center of clusters and anomalous

nuclei. Since current understanding of radio sources usually involves

massive objects (which can also explain the anomalous nuclei) the

circumstantial connection between a lack of ram-jet stripping and a

supermassive object in a galaxy's nucleus is established.

However, within this scenario it is quite unclear whether the

retained gas actually formed the massive object or just fuels it. The

possibility that the galaxy formed with a massive nucleus cannot be

overlooked; indeed, one of the central questions yet to be answered is

how important are the initial conditions under which the galaxy formed.



Cyclic Bursts of Star Formation


The contention that the star formation rate in an elliptical or

SO galaxy is strongly dependent on time gains credibility only recently.

Van den Bergh (1975) cites several examples of elliptical galaxies

experiencing anomalously vigorous star formation.

NGC 5128 (also known as the radio source Centaurus A) has recently

been shown by van den Bergh (1976a) to be undergoing very active star

formation along and interior to its prominent equatorial dust band. He

also suggests that the source of the unusual dust and gas in the system

is stellar debris shed by the stars. The galaxy is apparently a rare

field elliptical. It does not belong to a rich cluster and presumably

lacks any ram-jet stripping which may exist in such an environment.

A strong argument against this hypothesis is the finding by

Graham (1979) that the old stellar population rotates much slower than






4i




-26-


the equatorial band of dust and gas. This. is difficult to reconcile

with the interstellar material originating with the stars.

Most other researchers in the field tend toward the explanation that

the system may actually be a collision between a gas cloud or small

galaxy and an elliptical (Graham 1979). Thus the relevance of NGC

5128 to elliptical galaxy evolution cannot be assessed until these

questions are answered.

Van den Bergh (1975) has also suggested that NGC 1275 is an ex-

ample of an elliptical galaxy caught in a burst of star formation. The

underlying stellar distribution in the system is elliptical and the

galaxy is located near the center of the Perseus cluster. The small

velocity difference between the cluster's mean velocity and that of the

galaxy again suggests inefficient ram-jet stripping.

However, more recent work on the system (Kent and Sargent 1979,

Rubin et al. 1977) indicate that this is a collision between a fore-

ground spiral system and NGC 1275 in progress. Thus both NGC 5128 and

1275 may be atypical as far as early-type galaxies are concerned.

However, these two examples support the view that a massive (" 109

Me) influx of material over a relatively short time period results in

vigorous star formation. As both these systems are also radio sources,

one could argue that another effect of the mass accretion is to

either form, or fuel a previously existing, massive object in the

nucleus, producing the radio source.

Further support for this hypothesis, although not in an early-type

system, may be found in the M81-82 system. The interaction between

these galaxies seems to be resulting in a substantial mass infall to

M82 (Gottesman and Weliachew 1977, Killian 1978). It is likely that




-27-


the infall is connected with the peculiar structure of M82 and its

vigorous star formation.

On a gentler scale there are several examples of star formation

in dwarf ellipticals in the local group. NGC 185 and 205 both contain

dust patches and a sprinkling of blue, presumably young stars. The CO

observations described in Chapter III show that the star formation

is probably continuing in NGC 185. NGC 205 is a close companion of M31

and it may be argued that an interaction is taking place, although HI

maps of the region do not support this idea. Regardless, NGC 185 is

considerably further from M31 and an interaction does not appear likely

(see the HI observations presented in Chapter III concerning this

possibility).

Consequently, it appears that these two dwarf ellipticals are

reprocessing their ISM back into stars. The most likely source of

the material is the normal products of evolution of stars. Large scale

interactions are apparently not supported by the HI observations of

the galaxies.

The more sedate pace of star formation, and various arguments

suggesting an initial mass function for the stars shifted to lower

masses, are examined in the following section.


Continual Star Formation with a Skewed Initial Mass Function


A third fate that may befall gas retained in early-type systems

is star formation. The most stringent requirements on the nature of

this star formation are set by the observations of the colors or early-

type galaxies (see Larson and Tinsley 1978 for a discussion and earlier

references).




-28-


Larson and Tinsley (1974) have calculated models for elliptical

galaxies with star formation rates, continuing to the present, capable

of consuming the gas shed by other stars. While the integrated colors

of the models are consistent with observed galaxies, the expected

gradient of increasingly blue colors in the nucleus is not widely

observed. The most obvious drawback in their modeling is the use of

an initial mass function (IMF) which is fairly rich in hot stars.

As Faber and Gallagher (1976) comment "Since we have no a priori

knowledge of the IMF in ellipticals, star formation might conceivably

be confined to stars of small mass and low luminosity" (p. 370). They

go on to point out that the star formation must proceed efficiently

since very little, if any, interstellar material is seen in most

elliptical galaxies.

There is, however, observational evidence for star formation with

an anomalous IMF. Van den Bergh (1976c) suggests that an IMF deficient

in high mass stars is the most likely explanation for the lack of HII

regions in the Sa galaxy NGC 4594 (M104). Knots of young, blue stars

are observed near the prominent dust lanes. Normal, massive 0 and B

stars would form prominent HII regions under such conditions. Further,

this galaxy was not detected by Gallagher et al. (1975) in HI even

though a normal dust-to-gas ratio indicates it should have been easily

seen. Van den Bergh suggests that the lack of HI is due to its being

converted into molecular hydrogen, thus escaping detection.

On theoretical ground also, an IMF skewed away from massive stars

may be expected. Jura (1977) finds that one effect of reduced heating

of interstellar clouds (expected in elliptical galaxies) is to allow

clouds with much smaller masses to become gravitationally unstable




-29-


and collapse. Indeed, he finds that the critical mass for collapse under

such conditions is less than 10 Me. Because of fragmentation the

result of the collapse of a 10 M cloud will almost certainly be a

number of small mass stars rather than one 10 M star.

Thus while no compelling reason can be advanced to accept star

formation with a skewed IMF, the possibility cannot be rejected either.

The main intention of this dissertation as outlined in the following

section is to examine this possibility to determine if it is a sig-

nificant process in the evolution of early-type galaxies.



Statement of the Dissertation Problem

From the previous sections it is clear that our understanding of

early-type galaxies is incomplete. There are a large number of inter-

esting ideas concerning their formation and evolution, but as usual, in-

sufficient data are available to assess adequately their importance.

The primary goal of the observations presented and discussed in this

dissertation is to investigate the significance of star formation in the

evolution of early-type galaxies. The results of this study are most

directly applicable to the recycling of material by star formation in

an environment much different than our own galaxy. The results also

bear on other settings in which the star formation is poorly under-

stood.

It is not overstating the current situation to say that we only

have a dim view of how star formation occurs in our own galaxy, and an

even dimmer view of the process in drasically different environments.

Faber and Gallagher (1976) argue that possible star formation in




-30-


early-type systems must be very efficient; if so, it may be generically

related to the star formation that occurred as the galaxy collapsed

originally. Judging from the lack of primordial material observed,

that process was also very efficient.

This work is also of importance in the formation and continued

activity of radio sources in early-type galaxies. From the available

information it seems that radio sources and star formation are com-

peting for the same interstellar debris in a galaxy. If one is very

successful, it may be at the expense of the other.

All of this is not to say that galactic winds and ram-jet stripping

never occur, simply that all possibilities need to be analyzed and

evaluated to determine their relative importance in the overall

scheme of galaxy evolution.













CHAPTER II

THE OBSERVATIONS



This chapter presents the data acquisition and handling procedures

used for the work described in this dissertation. The instruments used are

described giving particular attention to the equipment and techniques

which aided this work immensely. The method of data presentation is

also explained.



Carbon Monoxide Observations at Kitt Peak, AZ


Telescope Description


The observations searching for the 2.6 mm transition of 12C160

were made at the National Radio Astronomy Observatory (NRAO) Millimeter

Wave Telescopel at Kitt Peak, AZ. The telescope is an 11 m paraboloid

which can be driven in altitude and azimuth. Tracking of celestial

objects, data acquisition,monitoring of system status, as well as

initial data reduction is handled by an on-line PDP 11/40 computer.

The observations were taken during three separate observing

sessions

December 23-26, 1977
July 7-9, 1978
December 10-16, 1979

The National Radio Astronomy Observatory is operated by Associated
Universities, Inc., under contract with the National Science Foundation.



-31-




-32-


The cooled mixer cassegrain receiver operating at the 115.2712 GHz

assumed rest frequency of the J = 1+0 CO transition was used for all

observations. The halfpower beam width at this frequency is about

65".

Pointing for the telescope is initially determined by the observa-

tory personnel. Pointing correction data are taken in all parts of the

sky on bright mm point sources. Analytic functions are then fitted to

the data, interpolated over regions with few data points. The observer

then checks these corrections by making five-point maps of bright sources,

usually the planets. Rarely are the corrections more than 5 arcseconds

at this stage.

However, the 1979 observations were hindered by large errors in

regions far from the celestial equator due to a lack of adequate data

for the fitting equations. This problem is discussed further in

Chapter III.

Nominally, the sky signal is separated into two linearly polarized

orthogonal components. In practice the system operated with only one

polarization, except during the 1979 observations when both channels

were available. These two signals are then fed to two Shottky-barrier

mixer diodes. Also fed to the mixer is the local oscillator (LO)

signal via a tunable injection cavity. This process of mixing two

frequencies to get a third (usually the difference of the two) is

known as heterodyning. The first intermediate frequency (IF) signal

emerges at 4750 MHz and is amplified by a pair of low-noise parametric

amplifiers. All of the previous equipment is enclosed in a dewar

cooled to 15 K.




-33-


Once outside the dewar the 4750 MHz first IF is amplified again

by room temperature transistor amplifiers. The signal is heterodyned

further at this point to 1328 MHz.

A third and final mixer produces a 150 MHz third IF which is

detected by two banks of 256 square-law detectors. One has a choice of

filter widths for the filter banks; our observations always used either

the 0.5 or 1.0 MHz (per channel) banks, corresponding to 1.3 and 2.6

km/sec velocity resolution, respectively. The December 1977 and July

1978 observations were made with only a single polarization operating, so

the separate channels were actually the same signal and could not be

averaged to reduce noise. Consequently the 0.5 MHz (1.3 km/sec) and 1.0

MHz (2.6 km/sec) filter banks were both used. At the time of the Decem-

ber 1979 observations the receiver operated in a proper two-channel mode

with two independent IFs. Both were fed to 1 MHz filterbanks since they

could now be added with a resultant decrease in noise.

Calibration was accomplished by alternately observing blank sky

and an ambient temperature microwave absorber. Details of the cali-

bration procedure are explained in Appendix I.

As a result of the calibration procedure one obtains the corrected

antenna temperature of the source

T* = f(l e)[J( TE) J(vs,Tg)] (1)

where the explanations for the symbols are found in Appendix I. Note

that there are three unknowns (nf, r, TE) related to the particular

source being observed; these cannot be determined without additional

information. For extragalactic work one can make estimates for

T and TE based on galactic studies and then calculate nf. The




-34-


errors are large in this approach and they give only a general idea

of the nature of the molecular cloud being observed.


Data Reduction Techniques


Each spectrum consists of two channels of 256 points each. For

December 1977 and July 1978 only one polarization was available and it

was detected in both the 256 MHz wide filter bank as well as the 128

MHz wide filter bank. The December 1979 observations took advantage

of the two orthogonal polarizations then available and detected them

separately in two 256 MHz filter banks. At the rest frequency of the

J = 1+0 transition of 12CO the 1 MHz resolution corresponds to 2.6

km/sec.

The on-site PDP 11/40 computer writes the data for each spectrum

(typically representing 5-10 minutes of observing) on disk. Real time

analysis can be done at the telescope and allows the observing time to

be optimally used. At the completion of the observing session the

disk is copied onto a binary coded tape.

The next step in the processing is to rewrite the tape with the

IBM 360 computer at the NRAO in Charlottesville, VA, to put the infor-

mation into an IBM-compatible format. This is entirely a translation

and no analyses are performed.

The final reduction is done at the University of Florida using

the resources of the Northeast Regional Data Center. Each spectrum is

examined for an unusually high root-mean-square deviation; all of the

offending spectra are rejected. The baselines are then examined and

if a polynomial of more than first degree is required to fit the base-

line, the spectrum is discarded.




-35-


At this stage of the reduction process each point of the spectrum

is examined to see if it is greater than five times the standard

deviation for that spectrum. If it is, and the adjacent points are not,

the point is replaced with the mean of the two adjacent points. This

procedure is used to remove interference that involves only a single

channel; multichannel filter banks are particularly susceptible to

this type of interference.

Each spectrum is then weighted by the inverse square of its

standard deviation and combined with all other spectra on the same

source.

A preliminary first order baseline is fit and subtracted and the

resulting spectrum is examined for possible features. If any are

apparent the channels involved are eliminated from the baseline fitting

procedure and a new first order baseline is calculated and subtracted.

The resulting spectrum represents the best data on a given source.

Various smoothing functions can be applied; the most common for this

work has been smoothing with a rectangular function of about the same

width as the suspected spectral feature. The effect of this is to

maximize the signal-to-noise ratio at the expense of velocity resolu-

tion.

The basic format of each spectrum presented is explained in

Figure 2.


Data Presentation


The data are displayed with antenna temperature as the ordinate

and velocity along the abscissa. Strictly speaking the abscissa




-36-


represents frequency, but since the frequency of the molecular

transition is already known the axis is calibrated in km/sec using

the following relation


vs = {r obslc Vf} / (1 vf/c) (2)


where vs is the center velocity of the observed band, vf is the velocity

of the source (corrected for the earth's rotation and revolution, i.e.

heliocentric), vr is the rest frequency of the spectral features, and

Vobs is the observed frequency.
The ordinate, corrected antenna temperature (T*), is found by

calibration as described in Appendix I. The unit is Kelvins (K), and

it is related to the flux density (S) by the following equation


S= 2 k nf TA/Ae (3)

where n is the antenna efficiency due to the loss of elements

(spillover, blockage, and ohmic losses in the antenna), A is the

effective area of the antenna, and k is Boltzmann's constant.


Neutral Hydrogen Observations at Green Bank, WV

Telescope Description

The neutral hydrogen observations were made with the NRAO 43 m

radio telescope at Green Bank, West Virginia. The telescope is an

equatorially mounted instrument completely under computer control. The

observations were taken during one session from July 13 to July 23, 1978.




-37-


The 21 cm cooled cassegrain receiver was used for all data acquisi-

tion. The system has two channels provided by linearly polarized,

orthogonal feeds. After initial amplification by a cooled upconverter

amplifier the signal is heterodyned and amplified through various stages

in much the same fashion as the process described for the 11 m tele-

scope at Kitt Peak. Typical system temperatures for the 43 m system

were 50-60 K.

The standard NRAO "back end" uses a 150 MHz IF which is fed into

a Model II autocorrelator spectrometer. The IF signal is autocorrelated

and the resulting autocorrelation function is Fourier transformed to

produce the power spectrum. The formation of the spectrum using auto-

correlation techniques is described in more detail by Blackman and

Tukey (1958) and Cooper (1976).

A 10 MHz bandwidth was chosen for all observations to provide an

adequate baseline. Also in the interest of baseline stability a

position-switched mode of observing was adopted. Ten minutes of data

are taken at the "off" position followed by 10 minutes at the "on"

position. The final spectrum is found by differencing the two spectra

thus acquired.

A number of "off" positions were used in an effort to deduce the

distribution of HI in the region around NGC 185. The majority of the

data were taken either with an "off" 16m to the west or as a five-point

map. The arrangement for the five-point map is with the "on" at the

center and "offs" taken successively to west, east, south, and north

at a distance of either 48' or 24'. The half-power beam-width at 21 cm

is about 20.5'.




-38-


Data Reduction Techniques


The output of the autocorrelator is two 192 channel spectra, each

being linearly polarized but orthogonal to the other. The spectra are

recorded by an on-line disk drive which can also be accessed by the

on-line reduction computer. In operation, the near instantaneous access

to the data just taken enables the observer to monitor the quality of

the system operation and to update the observing procedure based on the

preliminary results.

A characteristic of the autocorrelation method of spectral analysis

is that the strong galactic hydrogen within the bandpass produces a

sinusoidal ripple in the spectrum. This is known as "ringing" and its

removal is accomplished by convolving the spectrum with a hanning

function. This function is a weighting scheme in which 1 the value of

the channel on either side is added to I the value of the central

channel to produce the new value for that channel. Application of

this smoothing worked very well and all data presented here from the

43 m telescope have been smoothed with the hanning function.

A baseline is removed from the data by fitting a low-order poly-

nomial to the spectrum in regions removed from either galactic emission

or suspected NGC 185 emission. In practice the order of the fitting

polynomial was 2 to 4.

Calibration for this system is done under computer control by

periodically firing a noise tube within the receiver and comparing

the system output with and without the additional noise. The data are

then scaled to this system temperature. The stability of the system

was monitored by observing several sources throughout the session. No

unexplained drifts in system performance were seen.













CHAPTER III

RESULTS


In this chapter the results of the observations are presented.

The carbon monoxide data are treated first, followed by the neutral

hydrogen study of NGC 185 and the surrounding region. Only the results

are considered in this chapter; the analysis and interpretation of the

observations are found in Chapter IV.



Carbon Monoxide

Positive Result in NGC 185


The dwarf elliptical galaxy NGC 185 was detected at the 115.2712

GHz frequency of 12CO. Figure 2 shows the best spectrum obtained on

the source, averaging data from all three observing sessions. The

256 MHz total bandwidth corresponds to about 660 km/sec. The feature

is not discernible in the 128 MHz bandwidth spectrum, probably because

of the lower signal-to-noise (SN) ratio. Table 1 summarizes the

parameters of the spectral feature as measured in Figure 2.

Since the line is only slightly greater than 3 standard deviations,

it is very important to be certain of the reality of the feature. One

method is to smooth the spectrum with a function (in this case rect-

angular) of about the same width as the suspected spectral feature.

This procedure has the effect of increasing the SN ratio at the expense




-39-





















Figure 2. The CO spectrum for the nucleus of NGC 185. Data from all three
observing sessions have been averaged for this spectrum. The
total bandwidth for the observation appears in the upper left
along with the total integration time on the source. In the lower
right the position (1950.0 coordinates) observed is given and the
standard deviation for the spectrum is also shown. The baseline
at 0.0 Kelvin is shown along with horizontal lines at 3 standard
deviations.










(X101 )
-58. 00 -1,8.00 -38 .00 -28. 00 -18.00 -8.00 2.00 1230 ,

256 MHZ


LI L-)









cc
o l I Ln I t



LLJ. ,
- I II





.- R 0g H 36: 1 EL:,3


R MS = 25.29 MKELV I I
-5 .0 -118. f111 -38.- -28. fFI -18.00 -R.01 2. 0 1 2. 0h
VELiC ITiT (KM/SEC) (X101 )




-42-


of velocity resolution. Figure 3 is the result of this convolution,

and indeed the reality of the feature is strongly supported.



Table 1

Line Parameters for 12CO in NGC 185


Peak Antenna Temperature (K) 0.081 0.025

Velocity of Line (heliocentric, km/sec) -175 2

Full-width at Half-maximum (FWHM) (km/sec) 12 2

Position Observed Oh 36m 10s3

480 3' 53.'3

Total Integration Time on Source (min) 484.5



Other support for the reality of the line is found by careful

examination of the spectra obtained during each session of observing

(Figures 4 to 6). The spectral characteristics vary somewhat for each

session. But this is to be expected since each individual spectrum

has a relatively low SN ratio. In particular the line strengths for

each session are

December 1977 0.20 0.07 K
July 1978 0.20 0.08 K
December 1979 0.075 0.028 K

where the quoted errors are 1 standard deviation. In view of the

agreement here, and the pointing uncertainty during the December 1979

observations, the discrepancy among the observations is not large.

During the December 1979 observations the cooled receiver operated

with two independent channels, linearly polarized and orthogonal to

one another. The feature does not appear in the spectrum from the










(X101
S-58.00 -O8.00 -38.00 -28.00 -18. 0 -8. n 2.00 12 ~0
I I I I I -- ,_
256 MHZ
S8L .50 M IN NG 1 U 5















RA OH | "
D-C tl 0 3 [53P31
RMS 1 12 M.,E L.,. ME V I
r,









Sn -28.00 8.00 e.n n
LIU
PA ." 0 H 36 10s 3
[JEC : -3 53'1'3
NBO)X= 5 .
C-' I t I- I t I "
-58.00 -v:8.01] -38.0i -2;.011 -[R.0n- -F.0in 2 n 12. iO
VELOC IT (IT 'KM/SEC) (X101 )

Figure 3. The spectrum of Figure 2 smoothed with a 5 channel (13 km/sec) rectangular
function.










(X101 )
1-59.00 -.9.00 -39-00 -29.01 -19.00 -9.00 1. 00 n 1
1 1 I 1 I 1 -,_
256 MHZ
73.50 MIN N G C 1 8







'" 1 S 0
ci
WI r

t I


I RA 0 H 3F 103 IJ
DEC 148 3 '31'
C)
RMS = 72.01 M ELVIN --
I I I I I - --
-59.00 -9. 00 -39.00 -29.nn -199.0] 1. -i00i- 1i .
VELOCIT'Y ('KM/SEC) (X101

Figure 4. The 12CO data from December 1977.











(Xi01 )
-518.00 -18.0 -38. o -2.0 -1800 -8.00 2.00 t12

256 MHZ
72.00 MIN N G 185









ci D ,VD


CcL
LLJ
c cr



PRA 0 H 3" 10?3
DEC 418B 3 5 31'3
D FiRMS = 80.39 MKELVIlN 'LI
n o I I I I
-5.00 -118.00 -38.00 -2 8. -1 .O0 -e.O0 2.011 12. h
VELOCITY I(KM/SEC) (XI10'

Figure 5. The 12CO data from July 1978.










(X101
E-58.00 -4. 00 -38.00 -28.00 -18.00 -8.00 2.00 12L


339.00 MIN 1 b 1
SI f I I i ,

I I I I I-
ru I II It j | ' p
CD




C)



IJ=i
0-

p\RA 0 H 36 10 2
DEC 8. 3 53 '7
NB,= 'i RMS = 28. 4R MKELVIN1
1 I 1 IF I
'-53. 0 --8. n -3 .00 -28.00 -18.o -8.00 2. 1211
VEL OCT IT (KM/SEC) (,X 101 1

Figure 6. The 12CO data from December 1979, channel A.










CU fj-(X1 D )
2-58.00 -o8.00 -38. 0 -28 00 -18.00 -8.00 2.00 12 D10

256 MHZ
R3, nn MTN N [ 1 -i 1 1__ -1_


3 3





E 480 5 1 ii
cc I I I I I
L -i I
CI--

0B.-0 2.00 h
LLJ
CD
-- -RA 0 H 36, ]0 2
DEC i48 3 53:'7
N RMS = 26.93 MKELVI_ ;-
D I I -t '
'-58.0 --8.00 -38. [I -28.00J -18.0L -.00I 2.0U 121.hI
VEL OCIT (KM/SEC) (X 01

Figure 7. The 12CO data from December 1979, channel B.




-48-


second channel of the receiver (Figure 7). We have no physical reason

to expect a polarized feature and indeed, this transition has never been

observed to be polarized in our own galaxy. Also, the tracking procedure

rotates the observed polarization planes with respect to the sky; even if

the signal were linearly polarized we should be able to detect it in each

channel. The lack of a feature in this channel is almost certainly due

to the low SN ratio.

A further complication during the December 1979 session was poor

pointing corrections for the telescope, especially in regions far from

the celestial equator (the declination of NGC 185 is 480). Consequently

we feel that the lack of a feature in the second channel is within the

statistical and instrumental uncertainties of the experiment.


Negative Results

A total of 11 early-type systems were observed during the three

observing sessions. Our criteria for selecting candidates for CO

emission were based on the notion that a system able to retain HI or

dust would be more likely to have significant star formation.

Consequently we chose several of the early-type systems that have

been detected in HI. Added to the list were galaxies that are close

(< 15 Mpc) and also show some form of an ISM. Often this was a nota-

tion in the Reference Catalog of Bright Galaxies (RCBG, de Vaucouleurs

and de Vaucouleurs 1964) that the system contained obscuring matter.

Table 2 lists the 11 galaxies selected for the CO study. Included

are the positions observed, the assumed distance and radial velocities

as well as the Holmberg magnitude and the total luminosity. The dates

in which useful data were collected are also noted. The final quantity










Table 2

Parameters for Early-Type Galaxies Searched for 12C0

NGC # a(1950.0) 6(1950.0) Type D V mHol L Dates Observed 1 Standard
S om pg Deviation
(Mpc) (km/sec) (109L0) Dec 1977 July 1978 Dec 1979 (K)

185 00h36m10s3 48003'34" E3p 0.69 -245 11.0 0.0267 X X X Detected--
(DP-I) see text

185 00 36 12.2 48 03 27 X 0.025
(DP-II)

205-11 00 37 40.7 41 25 33 E5p 0.69 -239 10.1 0.0611 X 0.041

205-12 00 37 41.4 41 24 05 X 0.026

404 01 06 39.1 35 27 08 SO 1.5 -35 11.3 0.0955 X 0.067

1052 02 38 37.0 -8 28 05 E4 11.0 1439 11.1 6.18 X X 0.069

2685 08 51 41.7 58 56 01 SBOp 12.0 868 12.0 3.0 X X 0.042

3226 10 20 42.9 20 09 11 E2p 16.5 1356 12.3 4.6 X 0.062

4150 12 08 01.2 30 40 53 SO 3.8 970 12.2 0.267 X X 0.043

4278 12 17 35.7 29 33 33 *El 9.6 659 11.2 4.29 X X 0.057

4636 12 40 17.3 02 57 42 EO 12.5 979 10.4 15.2 X 0.089

5846 15 03 56.5 01 47 50 EO 16.5 2353 11.1 13.9 X 0.140

5866 15 05 07.4 55 57 18 SO 13.8 692 11.0 10.7 X X 0.036











Table 2

Continued



Notes: The positions for NGC 185 and 205 were measured from the Palomar Sky Survey. The
positions for the other galaxies are from Gallouet and Heidmann (1971), Gallouet
et al. (1973), or Gallouet et al. (1975). The galaxy types and heliocentric
velocities are from de Vaucouleurs et al. (1976). The Holmberg magnitude is cal-
culated as described by Gallagher et al. (1975).





0I




-51-


listed in the table is the standard deviation for all observations of a

particular source.

Figures 8-18 show the spectra obtained with the 256 MHz filterbank.

Also indicated in the figures are the radial velocity assumed for each

source. In many of the spectra there are single channels above or below

the 3 standard deviation limit. These are not significant because they

fail to persist in further observations. This type of interference is

relatively common in multi-channel filterbank receivers, and arises from

stray voltages within the equipment. The implications of the negative

results are discussed in Chapter IV.



Neutral Hydrogen


The region surrounding NGC 185 was examined to sensitive limits

to detect any neutral hydrogen (HI) associated with the galaxy. The

43 m telescope at Green Bank was used in a total power mode as de-

scribed in Chapter II. The arrangement of "offs" was varied throughout

the course of the observing in an effort to delineate the HI detected.

One series of observations used an "off" at about 16m west to allow a

reasonable estimate of the total HI content of the region.

Another sequence of observing used closely spaced "offs" (24' or

48') to the west, east, south, and north to get a better idea of the

behavior of hydrogen around a specific region. For those observations

the "on" was always at the center of the "offs."

Neutral hydrogen was found at significant levels in several re-

gions, both at the position of NGC 185 and several points to the north.

Figure 19, which used a distant "off," shows a strong signal of about

14.2 milliKelvin (mK) at a radial velocity of -200 km/sec. The










5 58.00 -.800 -38.o00 -28a.00 18.00 -8.00 2.o 00

256 MHZ '
235. 00 MIN NG C205















RA cl 37 40?7
SDEC 15 N

'-58.00 -^ o00 -D3.I0 -28.00 -18.00 -8. 00 2.0-]b
C3





CVELCT (K/SEC
LFigure 8. The 12 data taken in July 1978 on dust patch no. 11 of Hodge (1973) in NGC 185.


1 n
Fi gJ
D E 4 --) 2 -_ -3-. -



VELOCITY (tH/SEU) (X10 t )
Figure 8. The 12CO data taken in July 1978 on dust patch no. 11 of Hodge (1973) in NGC 185.










(XIOI
8-58.0 0-48.0O -38.00 -283.00 -18.00 -8.00 2.00 1250

256 '1, H I







=D 1 I I I


LU



-DEC n 'C
^- RA 0 H 371 [:--!1

--------------------------[-- -------- BE- 1 2'--J .I---
RMS = 25.R1 MKELVIN
CD I ;_ I -r
'-S8.00 8.00 -38.00 -28.00 -18. 0J -8.00 2.00 L 2.
VELOCITY (KM/SE C) (X101 )

Figure 9. The 12C0 spectrum obtained in December 1979 on dust patch no. 12 of Hodge (1973)
in NGC 205.











(X 01 )
E--37.00 -27.00 -17.00 7.00 3:..0 1[3.00 231-0 33 0
SI- -----------
256 MHZ
75. O00 MIN N G C- 4


CCd
(-j o






LJi
CD



0 S'V Ii I I'" "1 I f lil ,=
LA I1 H F .








lit' 5' 7
-37 -27 -7 .
I---
CE

S--VELOI

PA 1 H 6 39s.
.. KE.C 35E U!7 7.5
_, H P___S = 56.56 MK E ELV ",.]

'-37.00 -27.00 -17.00 -7.0 3.F l.o 20 n1nn n
VELOCITY (KO.M'S"EC [i I

Figure 10. The 12CO spectrum for NGC 404. The negative feature at 200 km/sec is probably
not real, but rather an artifact of the multichannel filterbank.












(X 1n1
S10. l 1 0.00 130.00 140. 00 150.00 160.00 1.70.00 1 ; iR
CA_ _- _- -
256 MHZ
110 50" MIN 10 2






u I i f

t-----


CC.
WI __
LL
i *n





CA
RA 2 H _3Q 3S69
DEC -AD 28' H5.l
II -
RMS = 8.6 I MKELVIN I
S-- --- ------ --------------------
110.00 120.00 1 .00 1 1110.00 15!0.00 r O. 170.1 0 1 wl ..
VE L OCIT TY (K/ SE C) I X101

Figure 11. The 12CO spectrum for NGC 1052,










(X1D1
S .0 0I 63.O 73.00 i8 0 9 0 103.0 13. I
.--..K I-I1- ]-I- ____ _I___ l-__-_____.__.I l II Li::' : H-ll
2_ H_-----I I"-I- -- -...-- -.-. ..----------- .
2 56 M H 7
164.00 M INr, NG I2 l '











LL .I C
TLU __ __ _- '- -
1-- [l F1
RA B H 51 4 : "6
DEC 58 n 3
SRMS = 1. 69 HKELV NI tl
5 3.00: 73.0n 83.10 9 3.0 103.00 1113 10 II
VELOCITY I(M/SEC) iX1,01

Figure 12. The 12CO spectrum for NGC 2685.











(X1 1
002. -I0 112.0__00 31Ln 2 12.100 52-.00 162.00 17
2 5--1 F M H..---,- -
256 MHZ I_



E- ---- --- --_____ ___ I__




00 I N



Um a
o *
LU
I- P H~O 1~
FRA 10" 20 42s9
DEC 2-0 9 I1''

______MS = 62. 39 KELV IN
-E'.'-- --1 F- ------....
102.00 12.00 122.00 132. I 2. oo 1 00 1 2. 0 17", n
VEL0CJT ( M /SEiCI (X t1 01-

Figure 13. The 12CO spectrum for NGC 3226.










(XMi1 )
--q 0 1.00 11. o .0 21. 00 31 I ] Il .n 0 01 0 I .n
I ------_------_-------------------- ------- o
256 MHZ
1 0.OO MIN 3 1 ;






i-Li-l I jlii

u_. f
-_ i-, I I


LL
- ----------- ^i. --- .
DEC 30- 4i0 52' 3
RJ F"RMS = 2.73 MI'ELVIN ':
CD ---1-- --- ------.-----------I -------_____
'- F. l 1.00 11.00 21.00 31.00] q i1. o0 51.00 61. I h
VEL CI T (K M/SEC (X1 01

Figure 14. The 1CO spectrum for NGC 4150. De Vaucouleurs and de Voucouleurs (1964) give
a radial velocity of 244 km/sec, but de Vaucouleurs et al. (1976) list it as
970 km/sec.










(X101 1
." 'IL2. 6( !52. [ 6i An'. (-9 I, _
_.00 .2. 00 52.00 62.00 72.00 82.00 92.00 Ct 00
d-| ------------ --t-------------------------~--------------
,,7 o ii I I I G
256 MHIHZ

1 | 0t V,





LLJ I
LjJ I fl

lU

LI
f0i I F I

t- j2H .7IA


f"RMS = 56._7 MELVIN

32.0n0 L2.00 5;2. 0 2.0 72.00 82.0n 92.00 l ...00
VELOC IT' (tKM/'SE[C) (I X11 )

Figure 15. The 12CO spectrum for NGC 4278.











(X 101 'I
54:, o0 71. 00 ,0 191.. 00 10.00 114 .00 1 40001 13 A00

256 MHi
52. n MIN N G C I F F ;

p~j ---------------- t----------------------------.-------- C3




^ cD I 1 I '^ II
C II




CT
1



LU
F--
A. 12 41 1 7 3
DEC 2 N 57' 41 'LI
CD a
RMS = 88.6 MKELVIN '
I I I- --------
6 it. Fn 74. 0I 8 1:.00 911. 0I 1OW. 00 1 ? 4. 0 21 I 3iLI 1-13 li
VEL CiITT (,KM/SEC) (X O'10

Figure 16. The 12CO spectrum for NGC 4636.











(X101 )
38.oo 18.o0 158.00 1 1 8.00 178.110 188.0' 19A.n8 20O 0
,t ---- ----- 1------------I---- \--- ~--
256 MHZ
.-.--....I---------_-::,.--.




4 P 0 | '1 (



C1-

f I


LU
LLi
I--


RA 15" 3 M S5B
DE-C 1 47' 49.'6

I-F- 1'RMS = 139. 4 1MK ELVIN J
13 8..00 18 00 158.0 0 168.00 178.00 18.00 198.00 20'. 100
VEL OCITT ( :M/ SEC [X 101 )

Figure 17. The 1CO spectrum for NGC 5846. De Vaucouleurs and de Vaucouleurs (1964) list
a radial velocity of 1771 km/sec, but de Vaucouleurs et al. (1976) list it as
2353 km/sec.










(X1D )
-1- 1 55 00 I2 0 0S

256 MHZ

3 9 O ON h i 1 ,\ i , C--



LIWJ il' I N l I





CE

a l
I-RA 15 H H I 4u
DE: 5 n 5 A
)3 D 3S-P
'---MS = 3 15 MELV I N

35.. ..5.00 55.00 65.00 S.0B 7r 105.00
VE LCIT (K/E C I t 10 1


Figure 18. The 12C0 spectrum for NGC 5866.
or --------------I --A----^-----------------.(--------- o~










(XI10
-100. oo -n. 00 -60. l0 -01.00 i -20.00 0.00 2.00 IJ ol
n I I I I ,
10 MAHZ r- C1 0 6 1




I,
J118.00 MT n I Nq9
J I UI T r1













- RA-" 0 H 1- q- )
DEU F q 4 4




...M.) 1.99 M KE-LVI
4 ---P---- -








0.0 -80.00 -60.00 -.10.0 -20-00 0.00 21 ..0
LU vI H io w ^ \

Lf ?'F cr>


VEL DIT-N /SE0 C 10' I I



Figure 19. The 21 cm HI spectrum on NGC 185. The off position is indicated in the lower
riht. The features between -100 and 0 km/sec are local hydroen.
'-00. o0 -80.00 -60.00 -1[0.00 -20.00 o.oo 20.1. Il-.u
VELI.CITT (KM/SEC (X10 "

Figure 19. The 21 cm HI spectrum on NGC 185. The off position is indicated in the lower
right. The features between -100 and 0 km/sec are local hydrogen.




-64-


position of the "on" is at NGC 185. A full-width at half-maximum

(FWHM) for the feature is about 60 km/sec.

It is apparent though, from Figure 20, that all of this HI may

not be associated with NGC 185. The "offs" for Figure 20 are taken at

either 24' or 48' (they have been mixed to increase the SN ratio for

this spectrum) in the four cardinal directions. The fact that the signal

is only 5.2 mK indicates that some, but not all, of the line has been

subtracted out in the "offs."

Further, spectra were generated separately for each of the four

directions. Thoseto the west, east, and south show about a 3

standard deviation positive signal. However, the spectrum with an
"off" to the north shows no significant feature in the NGC 185 velocity

range.

Continued examination of points to the north of the galaxy re-

veals a strong 40 mK feature about 70' directly north. The radial

velocity of this feature is -180 km/sec with a RWHM of 40 km/sec.

Five point maps of the region to the north also indicate HI signals

as far as 20 north and 48' east with a radial velocity of about -180

km/sec. There is only weak HI emission at 49' north of the galaxy.

This is shown in Figure 21; the feature is around 15 mK. It appears

that there is a ridge or plume of HI that is strongest (43 mK) between
1-2 directly north of NGC 185. It extends further north and east at

a level of around 20 mK.

To the south towards the galaxy the plume decreases in strength to

about 10-15 mK. From the sensitive 5-point maps at the position of

NGC 185, an excess of about 5 mK is found for the galaxy's location.

These observations are summarized in Table 3. The interpretation of

these observations is found in Chapter IV.










(X101 )
100. 00 -l.00 -60. 0 -o1.0o0 -2o.nn n0.0 20n.n uo,'o
SI I I I I
10 MHZ
1 366. 0 M IN NGC 1 5
14575 TO 16009


14" I -- 2



v C


P_ I I -

-,.l l = 1 ,ltELV [




II I --- o
101.10 -80. 00 0. iO -10. 0i -20.00 0.00 20.1 ir 11. [1in
VELOCITY (KM/SEC- (X iO J

Figure 20. The HI spectrum of NGC 185 with offs 1-2 beam widths to the west, east, south,
and north.
[-



o~ ---- I ----------------- I--------- o_---Cr




and north.










(XlO )
,--100. 00 -80.00 -60.00 -.0 -20. 00 o.0 20.00 Lr0 n0

10 MHZ
360,00 MIN IF 185
L 5 f U08 TO 161Y96

cc









EC-CFF 40 3 4 4
SI -- --





Figure 21. The HI spectrum at 49' north of NGC 185.
1-_

\- A ?A-OFF np 0 f H ;19 2
"E-"FF L 3 I L' --

S ".S = 2.216 1 -KELVI ,

-100.00 -BO.00 -6i0-o -4o.00 -2O.00 0.01-1 1l.0i '40 .1
VELOCITY (K_/SEC (X1D1 ). ,-

Figure 21. The HI spectrum at 49' north of NGC 185.











Table 3

HI Observations in the Vicinity of NGC 185

ON OFF Full-width Time
Position of observation observation Standard Velocity Half-maximum on
ON wrt Temperature Deviation (km/sec) (km/sec) Source
NGC 185 a(1950.0) 6(1950.0) a(1950.0) 6(1950.0) (mK) (mK) 5 km/sec 3 K/sec (min)

NGC 185a 00h36ml3 48003'45" 00h19m43S9 48003'44" 14.4 2.0 -190 63 378

49' Northb 00 36 11.4 48 52 38 00 19 27.8 48 03 44 15.7 2.3 -175 35 360

NGC 185c 00 36 11.4 48 03 44 00 31 19.6 48 03 44 6.1 2.0 -200 40 344
00 33 45.0 48 03 45
NGC 185 00 41 3.2 48 03 44 6.3 1.9 -200 45 344
00 38 37.7 48 03 44
NGC 185 00 36 11.4 47 14 50 6.3 1.9 -200 45 344
00 36 11.4 47 39 17
NGC 185 00 36 11.5 48 52 38 <5 2.6 -- 344.
00 36 11.4 48 28 12
120' North 00 36 11.3 50 03 44 00 41 16.1 50 03 44 -17.3 5.1 -200 38 60
00 36 11.4 49 14 50 -34.6 4.7 -180 33 60
71' North 00 36 11.4 49 14 50 00 19 20 48 03 44 39.8 8.6 -180 33 13


Notes: These measurements are from spectra reduced using a third order baseline and hanning smoothed. The
spectra on NGC 185 with two "off" positions are from data taken at both "off" positions. Approxi-
mately 27% of the integration time was with the first "off," the balance with the second "off."
aReproduced as Figure 19.

Reproduced as Figure 21.

cThe sum of this and the three spectra following is reproduced as Figure 20.












CHAPTER IV

DISCUSSION


In this chapter the implications of the observations presented in

Chapter III are considered. The simplest calculations from these data

involve assumptions about the excitation temperature and optical depth.

The portion of the dissertation based on such assumptions is found in

this chapter.



NGC 185


This dwarf elliptical galaxy is usually considered to be a com-

panion of M31. It is located about 6 to the north of M31 and the

difference in radial velocities is about 55 km/sec. For a discussion

of the virial mass of the M31 system and comments on membership in

that system see Rood (1979). A distance of 0.69 Mpc is assumed for

NGC 185 throughout the following discussion. At this distance 1"

1.04 x 1019 cm = 3.3 pc.

The classification given by de Vaucouleurs et al. (1976) is E3p.

A detailed photographic study of the system by Hodge (1963) shows

clearly that the ellipticity (defined by e = 1 b varies from 0.18
a
near the nucleus to 0.26 for the outer contours at 350" from the

nucleus. An increase in ellipticity with radius is commonly observed

for elliptical galaxies. The E3 classification (which is based on the

outer contours only) is confirmed by Hodge's photometry.


-68-




-69-


The system is peculiar because of the presence of two dark dust

patches within about 20 arcseconds of the nucleus. Figure 22 is an

optical photograph of NGC 185 and a schematic of the dust regions is

reproduced from Hodge (1963) in Figure 23. Careful comparison of the

two figures reveals the extent of the dust in the optical photograph.

Table 4 tabulates the positions of the patches used for observation

and the areas in square arcseconds and pc2 as measured from Figure 23.

Note that dust patch I (DP-I) has 2 entries, la and I. DP-Ia is the

dark core of the northwest patch and DP-I is the larger, less dense

patch (but also including DP-Ia).


Table 4

Positions, Areas, and Telescope Filling Factors for
Dust Patches in NGC 185

Object a(1950.0) 6(1950.0) Arem 2 Filling
("1) pc Factor

DP-Ia Oh36m103 48003'54" 35 380 0.0072

DP-I Same as DP-Ia 115 1250 0.0235

DP-II Oh36m12 2 48003'28" 350 3800 0.0716

DP-I+II Oh36m114 48003'44" 465 5050 0.0951
(Center of
NGC 185)



The CO observations were taken at the position of DP-I in December

1977 and July 1978. The December 1979 observations were made at both

DP-I and -II as well as the center of the galaxy. However, as the

FWHM for the 11 m telescope indicates in Figure 22, observations of one

patch do not entirely exclude the other. Further, the pointing for





-70-





-IU-











HPBW







'* *








*
*









30"

0





Figure 22. NGC 185 showing the dust patches. North is up and west is
to the right. The half-power beam-width is indicated.
This photograph is reproduced through the courtesy of Lick
Observatory.




-71-






\ czr n AN







10

o









MAJOR
AXIS










Figure 23. A schematic of the dust regions near the nucleus of
NGC 185 (Hodge 1973). DP-I is northwest of the center,
DP-II is southeast and DP-Ia is darker region within
DP-I.




-72-


December 1979 is unreliable to perhaps 20 arcseconds and at times

30-40 arcseconds. Consequently we have no confidence in the apparent

identification of the CO source as DP-I based on the absolute pointing.

Because of the uncertainty of the source for the CO emission, the

following analysis considers three cases for the calculations of column

and particle densities. The individual cases and the assumed CO sources

are

Case A DP-Ia
Case B DP-I
Case C DP-I+II

To estimate the kinetic temperature of the molecule we can assume

the clouds are optically thick (T >> 1). Observations of CO sources

within our galaxy indicate that this is often the case with the J = 1 0

transition (Zuckerman and Palmer 1974). We find (from Appendix II)

that

T = nfTex (1)

where nf is the forward beam coupling efficiency (also known as the

filling factor or the beam dilution) from Ulich and Haas (1976). Since we

have no detailed information on the brightness distribution we will assume

that ff = 0.7 (source area/beam area). The beam area is found from the

65" half-power beamwidth and the factor of 0.7 arises because 30% of the

power enters the antenna through the very broad error pattern. Given

these assumptions Table 5 gives the expected antenna temperature for the

three cases and several excitation temperatures.

Clearly we cannot distinguish between the several possibilities

in Table 5 which are consistent with the detected temperature of 0.081 K.

But if the CO is optically thick it is unlikely to be above 25 K or to

originate in both the clouds.




-73-


Table 5

Expected Antenna Temperatures (K) for T >> 1


Source T =10 K 25 K 100 K 103 K


Case A 0.071 K 0.179 0.714 7.14

Case B 0.235 0.588 2.35 23.5

Case C 0.951 2.38 9.51 95.1



Another approach is to compare the dust clouds observed in NGC 185

to clouds more easily observed in our own galaxy. The disadvantage

is that it is not certain that the nature of the clouds will be the

same. It does, however, provide a reference point for their study.

The extinction in the dust clouds was measured by Hodge (1963).

He finds DP-I to have a mean visual absorption of 0.3 mag but as high

as 1 mag in certain parts. The mean for DP-II is measured as 0.15 mag

of absorption in the visual.

At this level of absorption the clouds are similar to the diffuse

clouds studied by Knapp and Jura (1976). Their study involved clouds

situated in front of stellar sources that show color excesses of

E(B V) 0.3 mag. Assuming a normal Av/E(B V) = 3 indicates

Av 1 mag, similar to the absorption in the NGC 185 clouds.

Knapp and Jura suspect optically thin 12CO emission based on their

inability to observe 13CO at an intensity of 2-6 times less than the
12CO (usually possible if 12CO is optically thick), and they found

antenna temperatures of 1-2 K even though the sources probably filled

the beam and had kinetic temperatures of 20 K or more. Using the thin




-74-


cloud models of Lucas (1974) and assuming a kinetic temperature of

20 K (Morton 1975) they find that the column density is given by

N(CO) u 6.0 x 1014(TA/nf)Av (2)

where AV is the full-width half-maximum of the line in km/sec. Using

this expression and taking into account beam dilution (nf) we find

column densities for NGC 185 as listed in Table 6.


Table 6

Column density of 12CO for Optically Thin, Diffuse Clouds

SourceN(C) N(CO) x 2 x 4 = N(H2)
Source 2m

Case A 8.1 x 1016 1.6 x 1021

Case B 2.5 x 1016 5.0 x 1020

Case C 6.1 x 1015 1.2 x 1020



The mass of the molecular cloud can be found by multiplying the

column density of H2 first by the area of the cloud and then by the

mass per particle. The result is 9680 Me and is independent of which

case is actually correct. It is independent because we are assuming

that the source is optically thin and we are detecting all the CO

within a given region. Further, it represents a firm lower limit if

the colliding particles are H2; if they are HI the minimum mass is

4840 M .

The last column of Table 6, total column density of H2, is found

by assuming N(CO)/N(H2) = 5 x 105(Martin and Barrett 1978).




-75-


A further finding by Knapp and Jura (.1976) is that a necessary,

but not sufficient, condition for CO in emission is a particle density
3 -3
S10 cm3. This agrees with the theoretical result noted in Appendix

II. We thus feel confident that the CO emission in NGC 185 is

originating in a region at least as dense as 103 cm-3.

It would be useful for calculations of the space density of various

particles if the physical size of the absorbing regions were known.

This would help constrain the various possible sources of the CO emission

and help in determinations of the mass of gas involved. An estimate

of the cloud volume can be made by assuming they have a dimension along

the line-of-sight equal to the average width. This approach will tend

to underestimate the volumes unless the clouds are flat and oriented

broadside to the line-of-sight. Table 7 tabulates the result of this

calculation. The gas mass listed in the last column is found by

assuming the entire cloud is filled with molecular hydrogen at a
3 -3
density of 10 cm.



Table 7

Estimates of Cloud Volumes for NGC 185 Dust Clouds

Dust Depth of Cloud Volume Mass of Gas
Cloud (pc) (cm3) (pc3) (M )

DP-Ia 6.6 7.37 x 1058 2.5 x 102 6.2 x 104

DP-I 16.5 6.06 x 1059 2.1 x 104 5.1 x 105

DP-II 42.9 4.79 x 1060 1.6 x 105 4.0 x 106




Hodge (1963) has estimated the mass of dust in DP-I and DP-II.

Using the dust masses and the total gas content for the optically thin




*




-76-


case (Table 6) and the minimum calculated gas mass from Table 7 allows

a calculation of the gas-to-dust (G/D) ratio. The results appear in

Table 8.


Table 8

Gas-to-Dust Ratios

Cloud Dust Mass Gas Mass G/D Gas Mass G/D
(M0) T < 1 (M0) T >> 1 (M)

DP-I 25 9680 390 5.1 x 105 2 x 104

DP-I+II 215 9680 45 4.5 x 106 2 x 104



Considering that the gas-to-dust ratio in our own galaxy is usually

quoted as 200, the most attractive source for the CO emission appears

to be DP-I.

Using the minimum mass consistent for DP-Ia (Table 6) allows a

calculation of the Jeans length for gravitational instability

AT T3
irkT, TKR3c 1
= [ ] 1.6[K p (3)
J 4Gpp M

where Me is the mass of the cloud in solar masses. Taking the width

of the cloud as a characteristic dimension, R = 6.6 pc, TK = Tex =

20 K, and M = 4.84 x 103 M0


j = 1.1 pc

Since the cloud is substantially larger than the Jeans length,it is

most likely in a state of collapse.




-77-


The Jeans mass can be calculated under these conditions from


S3 23 (TK/P)3/2
M= Po 10 1/2 gm (4)
J 6 0 1 1/2
p

with T = 20 K, y = 2, p = 3.34 x 10-21 g/cm3 (corresponding to 103 cm-3

of H2) we find M = 27 Me. If the cloud remains isothermal as it

collapses, and this appears likely because of the effective cooling,

M decreases and fragmentation is likely to occur.

Another calculation that confirms the collapsing nature of the

cloud in NGC 185 is provided by Rowan-Robinson (1979). He estimates

the cloud density, in an essentially virial calculation, by the

equation

n(H2) = 10390 (AV)2 r-2 K1 (5)

where n is 0.8 for a cloud in equilibrium, r is the radius of the cloud

in units of 1018 cm, and K2 is a dimensionless quantity related to a

hot core for the cloud (it is always greater than 1.5).

The density calculated for the NGC 185 cloud, assuming 6.6 pc

for r (the characteristic width for DP-Ia) is 3.3 x 103 cm-3. If the

cloud is in free-fall n = 0.2 and the derived density is 8.3 x 102 cm-3

Thus considering that the NGC 185 cloud masses are uncomfortably large

(Table 7 ) for higher particle densities, we consider the lower density

more likely.

From these arguments it appears that either DP-Ia or DP-I is most

likely in a state of collapse. Further, the lack of emission from

ionized gas (Humason et al. 1956) implies one of two possibilities.

The first is that star-formation ceased in the system long ago, such




-78-


that the HII regions associated with massi.ve stars have since dis-

sipated.

Considering that typical lifetimes for HII regions are 106 years

and Hodge (1973) estimates the age of the OB complex in NGC 205 (very

similar to that of NGC 185) to be 5 x 106 years, it is certainly

plausible that the early HII regions have evaporated. Since the time

for the observed clouds to collapse to stars is about 105 years, it

seems that the system may be experiencing bursts of star formation

similar to that envisioned by van den Bergh (1975).

However, there is a serious flaw in this scenario. Hodge (1963)

estimates that the population I stellar component of NGC 185 is about

2 x 105 M Using Faber and Gallagher's (1976) mass loss rate of

0.015 Me yr-1 (109L )-1 and the luminosity of NGC 185 (2.67 x 107 L ),

we find that it took 5 x 108 yrs for the material to collect. During

the 5 x 106 years since the last star-forming episode only 2 x 103 M

of gas should have been able to appear. Our CO observations show at

least 5 x 103 M0 (for optically thin CO, assuming atomic hydrogen for

the colliding particles) and it is likely that there is actually more.

An even stronger case is made if the total mass of dust (215 M )

is multiplied by an assumed gas-to-dust ratio of 200. The system is

seen to contain 4.3 x 104 Me of gas, more than 10 times the amount

from stellar mass loss.

The second possibility that may be occurring in NGC 185 is con-

tinual star formation but with no 0 or B stars, and consequently no HII

regions. It is very difficult to prove that this is the process taking

place. If indeed low-mass stars are being formed they will be too

faint to see and deductive reasoning is necessary to support this




-79-


possibility. Specifically, since the regions observed are dense

enough to form stars yet no HII regions are observed, either no stars

have formed yet (leading to the burst hypothesis) or only low-mass

stars are forming.

The idea of a skewed mass distribution for stars is explored in

some details by Mezger and Smith (1977). They find that clusters of

massive stars (which produce giant HII regions) are found only along

spiral arms. It is probable that the spiral shock front of a density

wave triggers their formation. It appears though, that the low-mass

stars can form out of small but dense clouds (Herbig 1970) that are

far more widespread than large clouds.

Jura (1977) also proposes that the expected radiation environment

of elliptical galaxies (low UV flux) will lead to low-mass star forma-

tion. The reasoning is that the low flux results in less heating of the

clouds. A cooler cloud will have a lower critical mass which can

separate out and collapse to form a star.

The dynamics of the nuclear region will be considered in the fol-

lowing paragraphs. The difficulties in measuring radial velocities for

early-type galaxies are immense. There are usually no HII regions in

these galaxies so their sharp emission lines are not available for

measurement. The absorption lines from the stars are intrinsically wide

and further broadened by the stellar velocity dispersion, resulting in

large errors.

The radial velocity data for NGC 185 are actually better than that

for most early-type galaxies. Ford et al. (1973) identified four

planetary nebulae in the system and later Ford et al. (1977) succeeded

in measuring the radial velocities of two of these objects, NGC 185-1 and




-80-


-2 in their notation. These measurements, coupled with the radial

velocity of DP-I, can be used to calculate a rough mass interior to the

object and, when combined with the isophotal contours of Hodge (1963),

the mass-to-luminosity ratio (MLR) can be computed.



Table 9

Positions and Radial Velocities of Points Within NGC 185


Object Radial Distance from Distance from
Velocity Minor Axis Major Axis
(km/sec) (pc) (pc)

185-1 -212.5 144.2 75.6

185-2 -200 107.3 103.6

DP-I -175 0 43.9



Table 9 lists the three radial velocities determined for the system,

as well as the projected distances from the major and minor axes. We

make the simple assumption that the ellipticity is caused by rotation

about the minor axis, thus the figure of the system is an oblate spheroid.

We further assume that the radial velocities are the projected results

of purely circular velocities about the nucleus. We can then make a

straightforward calculation of the mass interior to the measured point.

Implicit in this approach is the assumption that the points lie in the

plane of the galaxy.

Recalling the work of de Vaucouleurs (1977), described in Chapter I,

on the relative frequencies of various ellipticities, we can make an

informed guess of the true ellipticity of NGC 185. The galaxy is most

likely to be an E3.6, but we can be more flexible and also consider




-81-


the possibility that it is an E5.5 (the flattest elliptical according

to de Vaucouleurs 1977). Of course the system may be an E3 seen

face-on, but the velocity gradient for the three measured points strongly

suggests rotation.

The details of the determination of the inclination angle (i,

defined to be the angle between the plane of the galaxy and the plane

of the sky) and its effects on other parameters are found in Appendix

III. Under the assumption that the system as is an E3.6 or an E5.5 the

inclination angle is 53.10 or 68.00, respectively.

The azimuthal angle, e', can be deprojected to find the true azi-

muthal angle, e. The results are contained in Table 10 for the two

planetary nebulae. Since DP-I lies on the minor axis we assume it has

no radial velocity due to rotation; consequently we take its radial

velocity, -175 km/sec, as the systemic velocity for the galaxy.



Table 10

Values for the Observed and True Azimuthal Angles


Object e' e (degrees)
(degrees)
i = 53.10 i = 68.00

185-1 26.8 40.1 53.4

185-2 44.0 58.1 68.8



The true circular velocity for the nebulae can thus be calculated from


vc = vr/cos e sin i (6)




-82-


The true radius can also be computed and is listed with the velocities

in Table 11. Finally, the mass interior to the point, assuming Kepler-

ian motion, is given by


M = 2.36 x 102 v2 r (7)


where M is in solar masses, v is the circular velocity in km/sec, and

r is the true radius in parsecs.


Table 11

Circular Velocities and Radial Distances in NGC 185


Object vr vc (km/sec) r (pc)
km/sec i = 53.10 i = 68.00 i = 53.10 i = 68.0'

185-1 37.5 61.3 67.8 191.4 248.0

185-2 25 59.2 74.6 203.2 296.6



The isophotal contours of Hodge (1963) were measured to find the

luminosity interior to a given point. The contours measured were the

innermost (1) and the next brightest (2), corresponding to 20.47 and

21.22 mag/arcsec2, respectively. Using a distance modulus of 23.9 and

assuming absorption within our own galaxy according to the cosecant law

(Av = 0.26 csc b = 1.0 mag for NGC 185) allows the calculation of the

luminosity interior to the first and second contours, listed in

Table 12.




-83-


Table 12

Luminosities in the Nucleus of NGC 185


Contour Brightness2 Area 2 Luminosity
(mag/arcsec ) (arcsec ) (L

1 20.47 1690 1.03 x 107

2 21.22 7715a 2.35 x 107a

aExcludes contribution interior to contour 1.



S The mass calculated from equation 7 is then divided by the appro-

priate luminosity from Table 12 to derive the MLR interior to each

object. The Planetary nebulae, 185-1, lies on the second contour,

making the luminosity interior to it the sum of the luminosities inside

contours 1 and 2. The other nebula, 185-2, lies midway between con-

tours 1 and 2 so we use the luminosity inside 1 plus half the luminosity

inside 2. Table 13 lists the results for the MLR derived with these

values.


Table 13

Mass-to-Luminosity Ratios for the Nucleus of NGC 185


Object Mass Interior (M ) Luminosity MLR (solar units)
(L )
i = 53.1 i = 68.00 i = 53.10 i = 68.00

185-1 1.7 x 108 2.7 x 108 3.38 x 107 5.0 8.0

185-2 1.7 x 108 3.9 x 108 2.21 x 107 7.7 17.6




-84-


While these values for the MLR are very reasonable for early-type

galaxies in general (Faber and Gallagher 1979), there is a significant

disagreement with the MLR of 1.8 used by Ford et al. (1977) for NGC 185

and NGC 147. This MLR is ostensibly derived from observations of M32.

We feel that the most likely cause for the discrepancy is that Ford

et al. (1977) do not consider possible rotation in their analysis.

The radial velocity of DP-I is most likely nearest the systemic velocity

of NGC 185 since it is fortuitously on the minor axis.

We find that the MLR of the nuclear regions of NGC 185 is between

5 and 18. Considering the non-linear effects of the inclincation and

azimuthal angle, the most likely value for the MLR is around 8.

Assuming that it is independent of radius, the total mass of NGC 185

is found to be 1.3 x 109 M .



NGC 205

Another dwarf elliptical companion of M31, NGC 105, was also observed

to sensitive limits for CO emission, with negative results. Figures 8 and

9 show the averaged spectra from the July 1978 and December 1979 sessions,

respectively. The spectra are taken at two different positions within

the galaxy (dust patches 11 and 12 in Hodge's (1973) notation). Hodge

(1973) describes and diagrams the dust content of NGC 205. The distri-

bution is reminiscent of NGC 185 but with a larger number of discrete

clouds over a wider region in the nucleus. The July 1978 observations




-85-


were taken on dust patch number 11 while the December 1979 observations

were of dust patch number 12 on the other side of the nucleus. Table 14

lists the positions observed for each patch as well as Hodge's (1973)

estimates of the dust content. The dimensions listed by Hodge are

used to find a rough area in square arcseconds and pc2 (the assumed

distance is 0.69 Mpc). Hodge (1973) also presents microphotometer

tracings of these dust regions indicating a visual absorption of about

0.2-0.3 magnitudes. It is clear that these regions are not as dense

as DP-Ia in NGC 185. Beyond this difference the two galaxies are

very similar in dust and young star content.



Table 14

Dust Regions in NGC 205

Region a(1950.0) 5(1950.0) Area of Cloud Dust Mass
no. (")2 pc2 M0

11 0h37m40s7 41025'33" 360 3.9 x 103 160

12 0 37 41.3 41 24 05 880 9.6 x 103 460



A similar calculation as was made for the NGC 185 dust clouds,

under the assumption of optically thin emission (see equation (2) of

this chapter), can be performed. The antenna temperature used is 3

times the standard deviation for each cloud. The velocity width is taken to

be four times the channel width (Av =10.4 km/sec). The results appear

in Table 15. The column densities are modest when compared to those

of NGC 185 (Table 6). The computed total masses are even less than

one may expect from the dust masses and a gas-to-dust ratio of 200,






4*




-86-


3.2 x 104 and 9.2 x 104 MQ for clouds 11 and 12, respective-

ly.


Table 15

Maximum Column Densities of CO and H2 for NGC 205, < < 1

Cloud no. N(CO) N(H2) M (molecular)
(cm-2) (cm-2) M0

11 7.0 x 1015 1.4 x 1020 4.4 x 103

12 2.8 x 1015 5.6 x 1019 5.2 x 103



The situation does not change if one assumes the clouds are optically

thick. The NGC 205 clouds are relatively large compared to those in

NGC 185, and any reasonable kinetic temperature will produce a large

antenna temperature, as shown in Table 5, very easily.

The question is then, why is NGC 205 not detected while NGC 185

with similar morphology (and actually less dust) is detected at the

frequency of 12C0?

We feel that the key difference in the two systems is that the

NGC 205 clouds are less dense than those in NGC 185. Hodge's (1963 and

1973) microphotometer tracings show this, and even optical photographs

of the two galaxies show the NGC 185 dust patches to be more prominent.

As Knapp and Jura (1976) found, even a density of 103 cm-3 is not

sufficient to ensure detection of 12CO in emission. Further, their

observations were made on nearby clouds where one could reasonably

expect a filled beam. The NGC 205 observations involve substantial

beam dilution which considerably worsens prospects for detection.




-87-


While many of the arguments about continuing star formation in

NGC 205 can be advanced with much the same reasoning as those for NGC

185 in the previous section, much of this support is lost because we

cannot be sure the NGC 205 clouds are in a state of collapse. However,

it is very likely that they will eventually form stars. This is clear

because NGC 205 has an even larger young stellar component than NGC

185; Hodge (1973) calculates some 2 x 106 M0 for the young stars.

We seem to be seeing an earlier stage of the process in an ellip-

tical galaxy which converts the ISM into stars. In a few times 105

years the dust clouds in NGC 205 will be much more condensed and CO

emission should be detectable. In the future stars will be forming and,

if these are high-mass stars, the galaxy will retain its unusual popu-

lation I component.



Negative CO Results


The galaxies that were not detected at the 12CO J = 1-0 transition

are listed in Table 2. Several attributes are also listed, along with

the limits of our observations. These limits can be interpreted on the

basis of two assumptions about the CO.

First is that the gas is optically thin and the emission was not

detected because of beam dilution or an exceedingly small optical depth

(T << 1). Second, if the line is optically thick then beam dilution is

the only mechanism to lower the antenna temperature below our detection

limits.

Another variable in the interpretation of these negative results

is the width of the undetected CO line. If it is several hundred km/sec

broad, representing the contributions of many small clouds within the




-88-


beam, it is more difficult to detect than if it were a narrow feature

arising from only a single, larger cloud.

If T >> 1 no information can be gained about the amount of molecu-

lar material stored. Essentially any quantity of material can be

stored in clumped, optically thick clouds. The material remains un-

detected because of the very small filling factor (nf). Consequently

Table 16 lists various parameters under the assumption that the CO

clouds are optically thin.

Knapp and Jura's (1976) finding that the 12CO line is probably

thin if the visual absorption is less than 1 magnitude indicates that

most CO in elliptical galaxies will be thin since deep absorption

features are rare.

Equation (2) from this chapter is used along with the assumption

of nf = 0.05. The antenna temperature is taken to be 3 times the
standard deviation listed in Table 2. The width of the undetected

feature is taken to be 10 km/sec for one set of calculations and 200

km/sec for a second set. The first width is appropriate for a single

cloud while the second is more typical of an expected global profile

for many clouds (Rickard et al. 1977).

Also listed is the column density of H2 assuming N(CO)/N(H2)

5 x 10-5 (Martin and Barrett 1978). The molecular mass is calculated

by (Schneps et al. 1978)

M = ( )()(T/nf)AvD2 (8)

where M is the molecular mass in solar masses, n(Tex) incorporates the

effect of temperature on the population levels of the CO, T* is taken

to be 3 times the standard deviation, Av is the assumed velocity width










Table 16
Maximum Molecular Masses Derived from CO Observations

S< 1, Av = 10 km/sec T < 1, Av = 200 km/sec
NGC # 3xrms N(CO) N(H2) Mass N(CO) N(H2) Mass MHI MHI
116 -2 20 2 T=20 K 16 20 2 T=20 K Detected Expected
(K) (10 cm ) (10 cm 2) (M) (10 c ) (10 cm2) (M ) (M ) (Me)

205-11 0.123 1.48 3.0 1.2 x 104 29.6 59.2 2.4 x 105 -- 2.7x106
205-12 0.078 0.936 1.9 7.4 x 103 18.7 37.4 1.5 x 105 3 x 105a 2.7x106

404 0.201 2.41 4.8 9.0 x 104 48.2 96.2 1.8 x 106 -- 4.2x106
1052 0.207 2.48 5.0 5.0 x 106 49.6 99.2 1.0 x 108 1.1x109b 2.72x108
2685 0.126 1.51 3.0 3.6 x 106 30.2 60.4 7.2 x 107 9.5x108c 1.32x108
3226 0.186 2.23 4.5 1.0 x 107 44.6 89.2 2.0 x 108 2.02x08

4150 0.129 1.55 3.1 3.7 x 105 31.0 62.0 7.4 x 106 1.2x107

4278 0.171 2.05 4.1 3.2 x 106 41.0 82.0 6.4 x 107 2.5x108d 1.88x108

4636 0.267 3.20 6.4 8.3 x 106 64.0 128 1.7 x 108 3.2x108e 6.69x108
5846 0.420 5.04 10.1 2.3 x 107 101 202 4.6 x 108 5.0x08f 6.12x108

5866 0.108 1.30 2.6 4.1 x 106 26.0 52.0 8.2 x 107 -- 4.71x108

aunwin (1980) CGallagher et al. (1978) eBottinelli and Gouguenheim
Reif et al. (1978) dKnapp et al. (1978c) (1977b)
Huchtmeieret al. (1977)




-90-

in km/sec, and D is the distance to the source in Kpc. For

Tex < 3000 K the vibrational levels of CO are not populated and
ex3
(Tex) 10-3 T

Also listed in Table 16 is the mass of HI that has been detected

in the galaxy. The last column gives the expected HI content from

Knapp et al. (1978c)


MHI= 4.4 x 107 (Lpg/109 ) (9)

where MHI is in solar masses and Lg is the photographic luminosity of

the galaxy. This relation uses the mass loss rate given by Faber and

Gallagher (1976) of 0.015 yr-1 (109 Le)-1 and assumes a time for

accumulation of 4 x 109 yrs, the time over which we are sure that

normal elliptical galaxies have existed (Gunn and Oke 1975). A de-

crease of 25% is made to allow for an undetectable helium contribution.

Caution must be exercised in comparing the expected HI content

with the maximum molecular content as derived here. First they are

correct only if the CO is optically thin. While there is reason to

believe this is the case, it is not certain. For an optically thick

cloud only the surface is seen and no information is available about

the total mass.

Second, the CO observations are made with a 65" HPBW while the

galaxies are typically several times this size. The CO-implied esti-

mates of the total gas are thus only true for the central regions,

whereas the expected HI is for the entire galaxy. Indeed, in the cases

of the detected HI the beam is usually about the size of the galaxy.

This discrepancy in the size of the regions sampled may not be

an important consideration. This is due to the difficulty of finding




-91-


a mechanism that would allow the gas to maintain the same spatial dis-

tribution as the stars that shed it. As long as the gas can cool itself

effectively, usually via line radiation, it must collapse into the po-

tential well of the galaxy. If it cannot cool itself, a galactic wind

sets up and the system is swept clean. The point has been made pre-

viously that galactic winds are not universally effective, consequently

one expects the interstellar material to preferentially collect in the

nucleus.

Comparing the maximum reasonable amount of molecular material that

could escape detection under these assumptions (T < 1, Av = 200 km/sec)

with the expected HI content reveals that molecular storage is not a

dominating feature of the galaxy's ISM. The largest discrepancy, for

NGC 205, may not be significant because the region sampled for CO is

small with respect to the dust patches.

For the other galaxies the discrepancy may be due to one, or a

combination, of three processes: the gas may be clumped and thus

optically thick, the material could have been removed either by a

galactic wind or ram-jet stripping, or finally, star formation could be

consuming the gas with an IMF deficient in high-mass, bright stars.

A fourth possibility, that the clouds are similar to Knapp and

Jura's (1976) thin clouds with n 103 cm-3 but with no CO emission, is

unlikely. The CO observations included a large portion of the galaxy's

nucleus and it seems highly improbable that all the clouds will be in

this nether region of collapse with no CO emission. Certainly some

fraction may be without CO emission but others, further collapsed,

should produce CO emission.




-92-


It is interesting to note that the three galaxies with a detected

HI content larger than expected (NGC 1052, 2685, and 4278) are all

suspected of accretion. All three systems show peculiar dynamics for

the HI that make an extra-galactic source the most plausible explanation

(Reif et al. 1978, Shane 1980, Knapp et al. 1978c).




Full Text
Table 2
12
Parameters for Early-Type Galaxies Searched for CO
NGC #
a( 1950.0)
6(1950.0)
Type
D
V
0
mHo1m
Lpg
Dates Observed
1 Standard
Deviation
(K)
(Mpc)
(km/sec)
(1o9lq)
Dec
1977 July 1978 Dec 1979
185
(DP-I)
00h36m10^3
4803
'34"
E3p
0.69
-245
11.0
0.0267
X
X X
Detected--
see text
185
(DP-11)
00 36 12.2
i
48 03
27
X
0.025
205-11
00 37 40.7
41 25
33
E5p
0.69
-239
10.1
0.0611
X
0.041
205-12
00 37 41.4
41 24
05
X
0.026
404
01 06 39.1
35 27
08
SO
1.5
-35
11.3
0.0955
X
0.067
1052
02 38 37.0
-8 28
05
E4
11.0
1439
11.1
6.18
X
X
0.069
2685
08 51 41.7
58 56
01
SBOp
12.0
868
12.0
3.0
X
X
0.042
3226
10 20 42.9
20 09
11
E2p
16.5
1356
12.3
4.6
X
0.062
4150
12 08 01.2
30 40
53
SO
3.8
970
12.2
0.267
X
X
0.043
4278
12 17 35.7
29 33
33
El
9.6
659
11.2
4.29
X
X
0.057
4636
12 40 17.3
02 57
42
EO
12.5
979
10.4
15.2
X
0.089
5846
15 03 56.5
01 47
50
EO
16.5
2353
11.1
13.9
X
0.140
5866
15 05 07.4
55 57
18
SO
13.8
692
11.0
10.7
X
X
0.036
-49-


-15-
Strippinq Mechanisms
The various efforts to determine how interstellar material shed
by stars can be removed from early-type systems can be divided into
two classes. The first is an internally driven "galactic wind" and
second is ram-jet stripping by an intracluster medium. Both of these
processes will be discussed in the following sections.
Internally Driven Winds
The possibility of a galactic wind driven by an internally driven
energy source was suggested by Johnson and Axford (1971) and considered
more quantitatively by Mathews and Baker (1971). In essence the mech
anism operates by coupling the energy from Type I supernovae to the
general interstellar medium. The addition of this high-energy
/ 9 x
(8 x 10 K) low-mass component significantly heats the interstellar
material to a high enough temperature to escape from the system by
evaporation. A later study by Coleman and Worden (1977) shows that
the energy released by flare stars is by itself enough to drive a
galactic wind of this type.
The parameters that are most important in the establishment and
maintenance of a galactic wind are the Type I supernovae rate, the
energy output from each supernova, the efficiency of the coupling of
the supernova's energy to the ISM, and the amount of "pre-heating"
of the ISM by the velocity dispersion within the galaxy. Each of
these quantities are known to probably a factor of 2 at best and in
certain instances various authors disagree by factors of 10 or more.


CHAPTER I
INTRODUCTION
The Nature of Early-Type Galaxies
Since the discovery that the "spiral nebulae" were indeed island
universes somewhat like our own, the effort to study, classify, and
dissect them has been increasing with remarkable speed. In large
measure this is due as much to the expansion of the accessible electro
magnetic spectrum as to the increasing sensitivity of instruments
within each spectral window.
The basic observational technique used for this dissertation, radio
astronomy, dates from the 1932 observations of the Milky Way by Karl
Jansky. The achievements with this relatively new technique have
accumulated steadily over the past half century and it is now recog
nized as an invaluable tool in studying the universe and its contents.
One of the most attractive characteristics of radio astronomy is
that it complements the endeavors of the oldest technique, optical
astronomy. The analysis of visible light almost always entails objects
with a temperature of at least 2000 K and the vast majority of objects
this hot are stars. Radio astronomy, utilizing much less energetic
quanta, is sensitive to objects of several hundred degrees or less.
Typically this is primarily the interstellar gas that fills the space
between the stars. There are important exceptions to this rule
(synchrotron radiation, thermal bremstrahlung, etc.) but it illustrates
well the complementary nature of optical and radio astronomy.
-1-


-21-
system (Knapp et al. 1978a,Bottinelli and Gouguenheim 1979a). More
importantly, the distributions appear to be reasonably symmetric, a
condition difficult to reconcile with stripping or partial stripping
by an intracluster medium. This discrepancy is pursued further in
Chapter IV.
Other evidence that ram-jet stripping may not be as effective as
the numerical analyses indicate is found in the Hercules cluster, a
rather loose cluster composed almost entirely of spiral galaxies.
The inconsistency is that it is also an X-ray source. Our current
understanding of cluster X-ray sources necessitates a hot ('v 10^ K)
intracluster medium as the origin of the radiation. How the Hercules
galaxies have remained spirals and not been stripped is not understood
within the framework of current research. Even more perplexing is the
origin of the intracluster gas, it is usually thought to have been the
gas removed from the galaxies.
A recent statistical study of cluster morphology by Gisler (1980)
shows that the anticipated presence of Sc galaxies in clusters is not
found. The Sc galaxies are expected to be highly resistant to having
their ISM swept because they have a high rate of gas replenishment
(Gisler 1979). The apparent underabundance in rich clusters indicates
that ram-jet stripping cannot be the dominant mode of SO production.
Dressier (1980) comes to similar conclusion based on a study of
the morphology of the galaxies in 55 rich clusters. He finds a
significant number of SO systems in clusters which have too low a
density to accomplish any stripping. Further, the study finds a
difference between the bulge-to-disk ratios of spirals and SOs as well
as a tendency towards thicker disks in SO systems. From these


APPENDIX II
THE PHYSICS OF CO SPECTRAL LINE CALCULATIONS
In this appendix the formation of CO line emission is treated.
This is followed by consideration of column density calculations and
the assumptions involved therein.
The one-dimensional time-independent equation of transfer is
dl
~j~ = -K I + £
ds v v v
(1)
where k is the volume absorption coefficient, e is the volume emission
v v
coefficient, and I is the specific intensity at frequency v.
This can be integrated to give
-x (Sr>)
I (sj = I (0)e v 0 + e
V O V
,_n/ -T (s ) O T (s)
v o v o r v
fee ds
(2)
s
where xy(s) = / K^ds is the optical depth integrated along the line of
sight from an initial point (s = 0) to the observer's position (s ).
This can be further simplified by assuming the source is a uniform
homogeneous cloud and that k is due to an atomic or molecular transi
tion. We now have
-x (s ) e -T (s )
Us0) = Io(0)e v 0 + (^)(1 e 0 0
(3)
The first term is the attenuation of the background radiation as
it passes through the cloud and the second term is the emission
-105-


-9-
Field 1975 for reviews) but it will likely be many years before any
scenario is convincing and widely accepted.
Lynden-Bell (1967) proposed a theory in which the essential feature
of a galaxy's structure is determined by the timescale of collapse from
the background compared to the timescale of star formation. It can
be well illustrated by considering the various structural components
of our own galaxy.
Our galaxy consists of a disk of gas, dust, and relatively young
stars. The gas and dust are continually undergoing star formation
in which hot bright stars appear to be preferentially formed along
spiral arms. Spiral shock phenomena may be important in regulating
the star formation but regardless of the details there is a continuing
processing of gas and dust into stars.
Also present is a spheroidal halo which contains little if any
gas and dust. It is composed primarily of old stars and consequently
evolves only as fast as the stars that compose it.
Lynden-Bell's theory suggests that these structural features are
formed by varying rates of star formation occurring in the collapsing
proto-galactic cloud. Various components form as it separates out from
the cosmic expansion and begins to contract under its gravitational
force.
The novel feature of the theory is that it can explain how the
halo can relax in the time allowed. Essentially, the presence of a
changing gravitational potential will permit relaxation of the stellar
system much faster than would two-body encounters.
The stellar population of the halo is entirely old stars with no
gas or dust. During the collapse the stars form and are then


-72-
December 1979 is unreliable to perhaps 20 arcseconds and at times
30-40 arcseconds. Consequently we have no confidence in the apparent
identification of the CO source as DP-I based on the absolute pointing.
Because of the uncertainty of the source for the CO emission, the
following analysis considers three cases for the calculations of column
and particle densities. The individual cases and the assumed CO sources
are
Case A DP-Ia
Case B DP-I
Case C DP-I+II
To estimate the kinetic temperature of the molecule we can assume
the clouds are optically thick (t 1). Observations of CO sources
within our galaxy indicate that this is often the case with the J = 1 0
transition (Zuckerman and Palmer 1974). We find (from Appendix II)
that
ex
(1)
where is the forward beam coupling efficiency (also known as the
filling factor or the beam dilution) from Ulich and Haas (1976). Since we
have no detailed information on the brightness distribution we will assume
that 0.7 (source area/beam area). The beam area is found from the
65" half-power beamwidth and the factor of 0.7 arises because 30% of the
power enters the antenna through the very broad error pattern. Given
these assumptions Table 5 gives the expected antenna temperature for the
three cases and several excitation temperatures.
Clearly we cannot distinguish between the several possibilities
in Table 5 which are consistent with the detected temperature of 0.081 K.
But if the CO is optically thick it is unlikely to be above 25 K or to
originate in both the clouds.


-90-
in km/sec, and D is the distance to the source in Kpc. For
T < 3000 K the vibrational levels of CO are not populated and
ca
a 103 Tex'
Also listed in Table 16 is the mass of HI that has been detected
in the galaxy. The last column gives the expected HI content from
Knapp et al. (1978c)
mhi 4-4 x lo7 V1()9 L0>
(9)
where is in solar masses and Lp^ is the photographic luminosity of
the galaxy. This relation uses the mass loss rate given by Faber and
Gallagher (1976) of 0.015 yr-* (10^ L )* and assumes a time for
0
9
accumulation of 4 x 10 yrs, the time over which we are sure that
normal elliptical galaxies have existed (Gunn and Oke 1975). A de
crease of 25% is made to allow for an undetectable helium contribution.
Caution must be exercised in comparing the expected HI content
with the maximum molecular content as derived here. First they are
correct only if the CO is optically thin. While there is reason to
believe this is the case, it is not certain. For an optically thick
cloud only the surface is seen and no information is available about
the total mass.
Second, the CO observations are made with a 65" HPBW while the
galaxies are typically several times this size. The C0-implied esti
mates of the total gas are thus only true for the central regions,
whereas the expected HI is for the entire galaxy. Indeed, in the cases
of the detected HI the beam is usually about the size of the galaxy.
This discrepancy in the size of the regions sampled may not be
an important consideration. This is due to the difficulty of finding


-7-
features that are difficult to affect by stripping mechanisms, thus
supporting the parallel sequence hypothesis.
On the other hand, several people have investigated the spatial
distribution of lenticular systems and find that they are concentrated
within clusters, strongly suggesting that their current environment
is crucial in their formation.
Regardless of the mechanism for the origin of the lenticular
systems they have been included in this study because they share
several significant properties with elliptical galaxies. The stellar
populations appear quite similar and both have a lack of dust and gas
within their interstellar media.
The Formation and Evolution of Early-Type Galaxies
Formation
The general outline of galaxy formation, and elliptical galaxy
formation in particular, is understood only in its coarsest features.
This section presents the scenario most widely agreed upon with emphasis
on early-type systems. Significant gaps in the scenario are also noted
with various suggestions that may, in the future, fill them.
In the study of galaxy formation one is inevitably forced to con
sider earlier epochs. In astronomy this can be done easily by observing
more distant objects. The scale of the universe is such that the time
radiation has taken to arrive here on earth represents a significant
fraction of the object's existence. Continuing the effort to fainter
(more distant) objects gradually crosses over to cosmology and the
study of the origin of the universe.


Page
Negative CO Results 87
Neutral Hydrogen and NGC 185 93
V SUMMARY 96
APPENDIX
ICALIBRATION THEORY FOR CO OBSERVATIONS 100
IITHE PHYSICS OF CO SPECTRAL LINE CALCULATIONS 105
IIITHE GEOMETRY IN AN INCLINED DISK Ill
REFERENCES. 114
BIOGRAPHICAL SKETCH 120
vi



PAGE 1

A STUDY OF THE INTERSTELLAR MEDIUM IN NGC 185 AND OTHER EARLY-TYPE GALAXIES By DOUGLAS WILLIAM JOHNSON 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 1980

PAGE 2

to Maryfran fe^a^^ ff i rt^ wT'
PAGE 3

ACKNOWLEDGEMENTS There are many people who have contributed invaluable resources and encouragement to me in the completion of the research for this dissertation. Although it is not possible to thank each one individually, there are some I would like to acknowledge in particular. The Northeast Regional Data Center is acknowledged for providing a vigorous and stimulating computing environment. The operators of the 11 m NRAO telescope were very helpful in assisting me during my observing sessions and the assistance of Telescope Engineer Rick Howard was especially appreciated. I would also like to thank Dan McGuire for his assistance during my first observing session. I thank my supervisory committee members. Dr. Stephen T. Gottesman, Dr. Thomas D. Carr, Dr. Kwan-Yu Chen, Dr. Gary Ihas, and Dr. William Weltner, for their attention and interest in my work. I thank the Physics Department, and especially Dr. Richard Garrett, for the assistantships that have allowed me to pursue this course of study. The Division of Sponsored Research is thanked for its Seed Money Grant Competition, providing financial support for Dr. Gottesman and me during the course of our investigations. The Graduate School's Supplementary Fellowship for 1979-80 was also greatly appreciated. Finally, I wish to thank Steve Gottesman and his family for their efforts to make Maryfran's and my stay in Gainesville a wonderful time 111 ^S***?^^' -<* f?***'* "-"w "^ ; yrsifti/V^

PAGE 4

in our lives. Steve's assistance and discussions with me (to say nothing of his witticisms) were above and beyond the call of duty. I thank my parents. Bill and Mary Ann, for their support and love and good humor throughout the years. But I reserve perhaps my sincerest appreciation for Maryfran and the "kids" (Smokey, Phantom, Harpo, and Marble) for making it all worthwhile. TV >•# •**.^r^ii.#i^§.A'ixVgi^-jaicv f>,T jp.-^*^^*^^ W4 n ,imi m'^' m miGAtim^

PAGE 5

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS iii ABSTRACT vii CHAPTER I INTRODUCTION 1 The Nature of Early-Type Galaxies 1 Galaxy Classifying Schemes 2 The Morphology of Elliptical Galaxies 3 The Morphology of Lenticular (SO) Galaxies 6 The Formation and Evolution of Early-Type Galaxies ... 7 Formation 7 Mass Accretion 11 Mass Loss Due to Stellar Evolution 13 Stripping Mechanisms 15 Internally Driven Winds 15 Ram-Jet Stripping by an Intracluster Medium .... 18 Fate of Retained Gas 23 Supermassive Objects in the Nucleus 23 Cyclic Bursts of Star Formation 25 Continual Star Formation with a Skewed Initial Mass Function 27 Statement of Dissertation Problem 29 II THE OBSERVATIONS 31 Carbon Monoxide Observations at Kitt Peak, AZ 31 Telescope Description 31 Data Reduction Techniques 34 Data Presentation 35 Neutral Hydrogen Observations at Green Bank, WV 36 Telescope Description 36 Data Reduction Techniques ..... 38 III RESULTS 39 Carbon Monoxide 39 Positive Result in NGC 185 39 Negative Results 48 Neutral Hydrogen 51 IV DISCUSSION ..... 68 NGC 185 68 NGC 205 84 -an aw iw***u

PAGE 6

Page Negative CO Results 87 Neutral Hydrogen and NGC 185 93 V SUMMARY 96 APPENDIX I CALIBRATION THEORY FOR CO OBSERVATIONS 100 II THE PHYSICS OF CO SPECTRAL LINE CALCULATIONS 105 III THE GEOMETRY IN AN INCLINED DISK Ill REFERENCES. 114 BIOGRAPHICAL SKETCH 120 VI 'i^i ^^-<'^^*u'•'^m^it.^^,^^\,\^^i.mm\ml -tf^M^l•lfi-^•*t>-.'^*^^*tt^.•\J*
PAGE 7

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 STUDY OF THE INTERSTELLAR MEDIUM IN NGC 185 AND OTHER EARLY-TYPE GALAXIES By Douglas William Johnson August 1980 Chairman: Stephen T. Gottesman Major Department: Astronomy The question of an interstellar medium in early-type galaxies is considered in light of the small amounts of gas detected as neutral hydrogen (HI). It is apparent that there is some method of removal or reprocessing that keeps the interstellar medium of early-type systems gas and dust free in spite of mass loss from normal stellar evolution. 12 A detection of CO is presented for the dwarf elliptical system NGC 185. The mechanisms of line formation of the J = 1^ transition strongly imply that the emitting region is in a state of gravitational collapse. These observations are consistent with the observed dust content of the galaxy and its blue, presumably young stellar population. It is nearly certain that the galaxy is reprocessing its interstellar medium via star formation. The radial velocity of the CO cloud can be combined with the velocity data for two planetary nebulae within the system to allow rough mass calculations. Coupling this information with luminosity data indicates that the mass-to-luminosity ratio of the nuclear regions (within 1 arcminute) is between 5 and 18, with a value near 8 being the most probable. vii

PAGE 8

Observations of 10 other early-type systems are also presented and discussed. The negative results imply that the gas is either clumped and thus optically thick or has been removed from the system through a galactic wind, ram-jet stripping, or has been consumed by star formation. The nature of the star formation must be somewhat different than that in our own galaxy. The high-mass end of the initial mass function for star formation would result in bluer colors than observed, the star formation must be skewed towards the low-mass stars to be effective yet unobserved. Theoretical arguments that this is possible are advanced, but more sensitive and highly resolved CO observations are necessary to observe directly this scale of star formation. Neutral hydrogen observations of NGC 185 obtained with the NRAO 43 m telescope are presented and discussed. There is apparently lowlevel (20-40 mK) high-velocity hydrogen in the region of NGC 185. The most likely source of the material is the Magellanic Stream which terminates in this area. Superposed on this high-velocity material, at the same velocity, is an excess of about 5 mK of HI at the location of NGC 185. This was detected by using "off" spectra 1-2 beamwidths from the galaxy in the four cardinal directions. The observations cannot distinguish between an enhanced high-velocity feature projected onto the galaxy and genuine emission from the galaxy itself, but the results are \jery suggestive and should be followed up with observations of greater resolution. vm 4#-fii-l-, ,—1 '" m. -^t r— l"^*•••fc.4t>— •^•'lH|^**J)iat*l n iiwil' It-

PAGE 9

CHAPTER I INTRODUCTION The Nature of Early-Type Galaxies Since the discovery that the "spiral nebulae" were indeed island universes somewhat like our own, the effort to study, classify, and dissect them has been increasing with remarkable speed. In large measure this is due as much to the expansion of the accessible electromagnetic spectrum as to the increasing sensitivity of instruments within each spectral window. The basic observational technique used for this dissertation, radio astronomy, dates from the 1932 observations of the Milky Way by Karl Jansky. The achievements with this relatively new technique have accumulated steadily over the past half century and it is now recognized as an invaluable tool in studying the universe and its contents. One of the most attractive characteristics of radio astronomy is that it complements the endeavors of the oldest technique, optical astronomy. The analysis of visible light almost always entails objects with a temperature of at least 2000 K and the vast majority of objects this hot are stars. Radio astronomy, utilizing much less energetic quanta, is sensitive to objects of several hundred degrees or less. Typically this is primarily the interstellar gas that fills the space between the stars. There are important exceptions to this rule (synchrotron radiation, thermal bremstrahlung, etc.) but it illustrates well the complementary nature of optical and radio astronomy. -1-C^>|'*''r>tJrt Si*i.w ii*. > ,ww--.-~iiM > >--uV *^<>l*B ? ..^*W fc-*£^

PAGE 10

Galaxy Classifying Schemes Early studies of galaxies made it clear that the general morphology of these immense stellar systems allowed them to be grouped into a relatively small number of types. The most successful early venture was Bubble's "tuning fork" diagram published in 1936, and reproduced here in Figure 1, The spherical systems (EO) are at one end with progressively flatter (E1-E7) systems leading up to a split in the diagram. SO galaxies (also called lenticular galaxies) occupy the vertex of the fork because Hubble believed that they were transitional systems. They contain prominent elliptical bulges as well as a conspicuous disk component. It is interesting to note that the nature of SO galaxies is still being vigorously debated. VVR^^'^ ELUPTICAL NE3ULAE EO E3 i^ALs Figure 1. Hubble's "tuning fork" diagram of the classification of galaxies.

PAGE 11

-3Forming the tines of the fork are two parallel sequences of spiral galaxies; one with a bar, the other without. The trend along the tines is from tightly wound spiral arms (Sa or SBa) to looser, more open arms at the end (Sb-Sc or SBb-SBc). In addition to these major players in the drama, some 3% of all galaxies are irregular, possessing no dominant symmetrical structural features. Many additions and modifications have been made to this initial classifying scheme (Hubble 1936, Morgan 1958 and 1979, de Vaucouleurs 1959, van den Bergh 1960a and b, Sandage 1961) but it has remained remarkably unchanged over the years. In large measure the modifications are to take into account the more extensive information available due to more sensitive equipment and increasing access to other spectral regions. The following sections describe the elliptical and lenticular galaxies in more detail and lay the groundwork for the statement of the thesis problem in the final section of Chapter I. For historical reasons both elliptical and lenticular systems are commonly known as "early-type" galaxies. The Morphology of Elliptical Galaxies As the name implies, the elliptical galaxies are characterized by elliptical isophotes. The stellar population usually appears to be well-evolved with little or no interstellar gas or dust. The degree of ellipticity E is defined to be (a b)/a (where a and b are the semi -major and semi -mi nor axes, respectively). The range observed is 0.0-0.7 (E0-E7) with EO-El the most common and decreasing in frequency at the flatter end of the range. —•l^CaWlttf,***—*-*!*. ••(w*^ ,^rMm^m^m •l^'w^-.a/m^-^*^>W.

PAGE 12

-4Taking into account the statistics of random projection on the sky (for we are viewing a two-dimensional projection of a threedimensional object) it appears that the ellipticals are distributed normally about a mean of E3.6 (e = 0.36) with a dispersion of 0.11 (de Vaucouleurs 1977). It appears that true EO and E5.5 (de Vaucouleurs contends that E5.5 is the flattest bona fide elliptical) are relatively rare. The lack of flat systems, usually considered to be a dynamical effect caused by instabilities in thin disks, suggests that perhaps spiral density waves and the attendant star formation are suppressed in disks of sufficient thickness. The origin of the flatness of elliptical systems has long been thought a natural consequence of rotation. The greater the rotational velocity, the greater the degree of flattening. This of course implies that the three-dimensional figure is an oblate ellipsoid (polar diameter smaller than the equatorial diameter). In recent years several rotation curves of elliptical galaxies have been published (Bertola and Cappaccioli 1975, Illingworth 1977, Peterson 1978, Sargent et al 1978, and Young et al 1978a) which cast strong doubt on the validity of this simple approach. The small observed ratio of the maximum rotational velocity to the central dispersion velocity mitigates strongly against models which use isotropic velocity distributions and either oblate or prolate ellipsoids (Schechter and Gunn 1979). It appears necessary to use both anisotropic velocity distributions as well as rotation to account for the observed rotation curves (Binney 1978, Schechter and Gunn 1979).

PAGE 13

-5The origin of the anisotropy is still not clear but the most likely source is remnant anisotropy from the collapse phase of the galaxies' formation, A further difficulty in our understanding of elliptical galaxies is the existence of extreme population I ingredients in a significant number of systems. Specifically: • OB clusters are observable in NGC 185 and 205 (Hodge 1963 and 1973) • ionized gas is seen in the nuclei of at least 15% of all ellipticals (Osterbrock 1960 and 1962) • neutral hydrogen has been detected in at least 8 elliptical systems : NGC 1052 (Knapp et al 1978b) NGC 2974 (Bottinelli and Gouguenheim 1979b) NGC 3904 (Bottinelli and Gouguenheim 1977b) NGC 3952 (Bottinelli and Gouguenheim 1979a) NGC 4105 (Bottinelli and Gouguenheim 1979b) NGC 4278 (Gallagher et al 1977) NGC 4636 (Knapp et al 1978a) NGC 5846 (Bottinelli and Gouguenheim 1979b) The presence of population I material is unusual for systems thought to have ended all star formation long ago. Some possible explanations are that the material was accreted relatively recently and thus it is not representative of an elliptical galaxy's normal evolution, or that through normal processes of stellar evolution the material was shed by the stars and is observable in various forms today. The thrust of the foregoing observations is that, as a class, elliptical galaxies are not as well -understood as was earlier believed. The apparently relaxed stellar distribution is most likely not relaxed at all, but still contains velocity anisotropies which strongly influence the shape of the galaxy. It is disturbingly common for the smooth isophotes to be blemished with dust obscuration or some other v1T>v
PAGE 14

form of detectable population I material. The CO detection described in Chapter III is another example of the incongruities present in elliptical systems. The Morphology of Lenticular (SO) Galaxies The lenticular systems, occupying the vertex of Hubble's tuning fork diagram, are as perplexing today as ever. The essential features of these galaxies are an elliptical bulge and. a disk, both apparently composed of population II stars. The disks are devoid of spiral arms and the systems lack the obvious star formation proceeding in spiral galaxies. Suggestions as to the origin of the lenticulars range from van den Bergh's proposal (1976b) that they are part of a complete sequence of gas-poor, "anemic" spirals parallel to the S and SB types to the more recent suggestion (Gunn and Gott 1972, Gisler 1979, and references therein) that they are normal spirals which have been stripped of their gas and dust by an interaction with an intracluster medium. Much work has been done in recent years on these systems and some interesting observations have been reported. Neutral hydrogen observations by Balick et al (1976) and a number of others place the lenticulars roughly midway between the Sa-SBa and E systems. This evidence supports the contention that they form another sequence of spirals much like the normal sequence, only gas-poor. Further, Burstein (1979a, b,c) finds significant differences in the bulge-to-disk ratios of lenticulars and spirals, as well as "thick" disks in lenticulars and not spirals. These are structural

PAGE 15

features that are difficult to affect by stripping mechanisms, thus supporting the parallel sequence hypothesis. On the other hand, several people have investigated the spatial distribution of lenticular systems and find that they are concentrated within clusters, strongly suggesting that their current environment is crucial in their formation. Regardless of the mechanism for the origin of the lenticular systems they have been included in this study because they share several significant properties with elliptical galaxies. The stellar populations appear quite similar and both have a lack of dust and gas within their interstellar media. The Formation and Evolution of Early-Type Galaxies Formation The general outline of galaxy formation, and elliptical galaxy formation in particular, is understood only in its coarsest features. This section presents the scenario most widely agreed upon with emphasis on early-type systems. Significant gaps in the scenario are also noted with various suggestions that may, in the future, fill them. In the study of galaxy formation one is inevitably forced to consider earlier epochs. In astronomy this can be done easily by observing more distant objects. The scale of the universe is such that the time radiation has taken to arrive here on earth represents a significant fraction of the object's existence. Continuing the effort to fainter (more distant) objects gradually crosses over to cosmology and the study of the origin of the universe. •J ^t90 ^'VUnwf ^11*

PAGE 16

-8Perhaps the most influential discovery in cosmology (and also bearing heavily on galaxy formation) has been the 2.7 K microwave background. The existence of nearly isotropic (see Cheng et al 1979 for a discussion of a dipole anisotropy attributed to the earth's motion with respect to the background) homogeneous radiation with an apparent blackbody spectrum strongly constrains galaxy formation theories. The cosmic background radiation is almost universally believed to be the remnant from the era of decoupling; the time when the universe had cooled enough to allow electrons and protons to recombine (about 3000 K). The effect of the formation of neutral hydrogen is to reduce drastically the opacity of the matter with respect to the radiation. Before recombination Thompson scattering of radiation off electrons coupled very strongly the matter with the radiation, they were kept in thermal equilibrium and cooled together. After decoupling the two components evolved essentially independently of one another and the 2.7 K background seen today is the remnant of the radiation component. This forms the basis of the expectation that the early universe was homogeneous and isotropic. Actually one can only say that it is homogeneous and isotropic to at least the smallest scale observed, currently that means the background is smooth to within 1 mK on a scale of 7 (Smoot and Lubin 1979, see Sunyaev 1977 for a discussion of fluctuations). The condensations that have eventually evolved into clusters of galaxies, galaxies and stars must have occurred during a later epoch. The nature of these perturbations is currently a topic of great interest. Several theories have been advanced to account for the processes by which the perturbations form and grow (see Gott 1977 and w .M l I iiBgaijr i%sa^ftJg *Mj ( aiti'*> i gin^ft4fc<'t^tf*i w i*pi f i i.> v iuv a**fti*^ OMt c^ ;t-i' -.i-4k.:..-*;€*nSAAW. ^<,'i>:-i->c*>^i.^.^ki>.ciu— ^>* av*;^o>
PAGE 17

Field 1975 for reviews) but it will likely be many years before any scenario is convincing and widely accepted. Lynden-Bell (1957) proposed a theory in which the essential feature of a galaxy's structure is determined by the timescale of collapse from the background compared to the timescale of star formation. It can be well illustrated by considering the various structural components of our own galaxy. Our galaxy consists of a disk of gas, dust, and relatively young stars. The gas and dust are continually undergoing star formation in which hot bright stars appear to be preferentially formed along spiral arms. Spiral shock phenomena may be important in regulating the star formation but regardless of the details there is a continuing processing of gas and dust into stars. Also present is a spheroidal halo which contains little if any gas and dust. It is composed primarily of old stars and consequently evolves only as fast as the stars that compose it. Lynden-Bell's theory suggests that these structural features are formed by varying rates of star formation occurring in the collapsing proto-galactic cloud. Various components form as it separates out from the cosmic expansion and begins to contract under its gravitational force. The novel feature of the theory is that it can explain how the halo can relax in the time allowed. Essentially, the presence of a changing gravitational potential will permit relaxation of the stellar system much faster than would two-body encounters. The stellar population of the halo is entirely old stars with no gas or dust. During the collapse the stars form and are then I n ^iM n il I iiii : I ^ ^ f^ irmmi^ }te^

PAGE 18

-10dynamically independent of the remaining material. That is, the stars reflect the velocity dispersion of the cloud at the time of their formation. As the material collapses further, star formation proceeds and eventually the gas becomes dense enough that damping forces become important. At this stage the gas gives up a great deal of kinetic energy to dissipative heating and, because of its angular momentum, settles down into a disk. Within this disk various processes, especially star formation, continue to the present giving spiral galaxies their distinctive optical appearance. It takes only a minor modification of this scenario to produce an elliptical galaxy. If star formation has proceeded to completion before the dissipative effects can form a disk then a system will result which has no disk component and is essentially a halo population of stars. Models which can produce the observed luminosity profiles of ellipticals have been constructed (Larson and Tinsley 1974) and add credence to this basic concept. Within this framework of galaxy formation many details are still to be worked out. As previously mentioned, the nature of the initial perturbations is not at all understood. Also the processes of fragmentation and collapse are poorly understood (Field 1975). There is even a question of whether gravitational instabilities or turbulence is responsible for the necessary condensations. Jones (1976) provides an excellent review of these and other problems in the study of galaxy formation. Since the early star formation was apparently so efficient as to consume all the primordial gas, one might not expect any continuing V*r''"*^"."V— s-' "^J ''.-— Trrr'**V''l**T^'''*-*^^-II?r.--a ,9 *~•,(.;,_ w— j-fa.-,i;^;i*i|.jj-

PAGE 19

•11star formation in elliptical galaxies. It will be shown, however, that other processes operate which alter this simple picture and lead to radically different expectations. The most important effects are mass loss from the normal evolutionary processes of the stellar population and perhaps accretion of extragalatic material; both are discussed in the following sections. Mass Accretion The contention that early-type galaxies, ellipticals in particular, accrete material is relatively new. The motivation is to explain the cD galaxies (giant ellipticals usually located in the center of rich clusters and described by Bautz and Morgan 1970) that are often strong radio sources. The accretion described here is full scale cannibalism of other galaxies during close encounters (Ostriker and Tremaine 1975, White 1976, Ostriker and Hausman 1977, Hausman and Ostriker 1978). The idea can be linked to cluster types as described by Bautz and Morgan (1970) and Oemler (1974). This cluster classification scheme ranks the clusters based on their richness. Observational ly it is found that the densest clusters (Type I in the notation of Bautz and Morgan 1970) often contain giant elliptical galaxies near their center. Further, the cD galaxies are often radio sources that are widely suspected to be caused by material falling into a massive object. There are several other pieces of circumstantial evidence that indicate that this process may indeed be significant in the evolution of elliptical galaxies (see Ostriker 1977 and Hausman and Ostriker 1978 for details)

PAGE 20

12A more moderate form of mass accretion is suggested by several authors (Bottinelli and Gouguenheim 1977a, Gallagher et al. 1977, and Knapp, Kerr, and Williams 1978c) to explain the inclined disk of NGC 4278. The differing directions of the angular momentum of the stellar component of the galaxy and the neutral hydrogen make an internal origin of the matter difficult to believe. A similar situation occurs in NGC 1052 (Knapp, Faber, and Gallagher 1978a, Fosbury et al 1978, Reif, Mebold, and Goss 1978) and the accretion of an intergalactic HI cloud is suggested. The major objection to this hypothesis is the lack of sufficiently massive clouds available for accretion. The results of Mathewson et al (1975) purporting to find HI clouds in the Sculptor group have been disputed by Haynes and Roberts (1979). The latter group contend that the material is a portion of the Magellanic Stream. Further, Lo and Sargent (1979) have searched nearby groups for detached HI clouds and find none more massive than '^4 X 10^ M„. A number of other authors have examined clusters of galaxies for HI emission (Haynes et al 1978, Baan et al. 1978) while others have examined the line of sight to quasars for HI absorption (Roberts and Steigerwald 1977, Shostak 1978). No isolated HI clouds are seen in emission in the clusters and the HI absorption measurements show that large clouds of neutral hydrogen are almost never seen outside galaxies. Thus it appears difficult to reconcile the several times 10^ M of HI found in NGC 4278 and 1052 with the dearth of sufficiently massive candidates for accretion. Silk and Norman (1979) propose an I *agiwwg?g- eM UMi ifffnKi

PAGE 21

•13alternative hypothesis, the accretion of gas-rich dwarf galaxies. They find that the gas component of the dwarfs will lose energy through dissipation and fall to the central regions of the accreting galaxy. The inf ailing gas, depending on the individual cloud mass, may either form stars or continue to fall into the nucleus where it may fuel a radio source. The stars will also be incorporated into the accreting galaxy but with less visible effects. Silk and Norman (1979) also consider the interaction of a Mathews and Baker (1971) type wind and the infalling material. If the amount of this material is sufficiently large the resulting supernovae (from the high-mass stars formed) will help in driving the galactic wind. However, an enhanced wind has the effect of inhibiting mass infall and the process slows itself. The net effect may be for star formation to proceed in cyclical bursts--a notion also suggested by van den Bergh (1975) in a somewhat different context. Both of the mass accretion processes described so far deal with normal or giant cD elliptical galaxies. In order to be effective in capturing and assimilating other systems the accreting galaxy must be large. The evolutionary mechanism discussed in the following section, mass loss from stellar evolution, operates in all systems. This includes the dwarf ellipticals NGC 185 and 205 considered in greater detail in Chapter III. Mass Loss Due to Stellar Evoluti on It has recently been appreciated that the normal evolution of stars in an early-type galaxy will be a source of interstellar material i i air~ i TaBa Bffgnifim**^a'na-* ii w>L Mi a > aafiiMii r >iii.g3. r7 ttr m-^ n^^t.w ik<.>.-

PAGE 22

14Small mass stars have stellar winds, Mira variables are known to eject mass during certain stages. Type I supernovae occur in population II stars, and planetary nebulae have been observed in early-type galaxies of the local group. The calculation of the contribution by stellar evolution to the interstellar medium (ISM) of early-type galaxies depends more on theoretical estimates than observational evidence. To date, the most important observational evidence is the detection of planetary nebulae in nearby dwarf ellipticals (including NGC 185) by Ford and Jenner (1975). Considering the uncertainties in the observations, the observed planetary 19 1 birthrate of > 0.012 yr (10 L^)" agrees well with Larson and Tinsley's (1974) estimate of 0.05 yr'"^ (10^ ^q^~^ • Following the reasoning of Faber and Gallagher (1976) and adopting a mass per planetary of 0.2 M^ results in a mass loss rate of 0.010 yr' (10 Lg)" from planetaries. Consideration of Miratype variables leads to a final assumed mass ejection rate of 0.015 yr" (10 Lj' The conservative nature of this calculation is apparent when one considers that the present mass loss rate is certainly lower than that of earlier epochs. This is primarily because any high-mass stars would have evolved quickly and cycled their mass back to the ISM early in the galaxy's evolution. Further, the contribution of mass from Type I supernovae (apparently confined to population II stars, Tammann 1974) and Type II supernovae (massive progenitors) earlier in the galaxy's evolution have been ignored. Even this conservative approach leads to contradictions in the ISM of early-type galaxies after 10-10 years (Faber and Gallagher 1976). —i mnT i — rT; 'irr'i~i i irr a i i n i ^n vm ru ri n ? ^ irf' M aT i f i i m • — ii -i i^ -t i i tii -rrrm -^T i^r i ~ '^ ^w'Sivr*<.'ja.ia m i i .i;

PAGE 23

15Stn'pping Mechanisms The various efforts to determine how interstellar material shed by stars can be removed from earlytype systems can be divided into two classes. The first is an internally driven "galactic wind" and second is ramjet stripping by an intracluster medium. Both of these processes will be discussed in the following sections. Internally Driven Winds The possibility of a galactic wind driven by an internally driven energy source was suggested by Johnson and Axford (1971) and considered more quantitatively by Mathews and Baker (1971). In essence the mechanism operates by coupling the energy from Type I supernovae to the general interstellar medium. The addition of this high-energy / 9 \ (8 X 10 K) low-mass component significantly heats the interstellar material to a high enough temperature to escape from the system by evaporation. A later study by Coleman and Worden (1977) shows that the energy released by flare stars is by itself enough to drive a galactic wind of this type. The parameters that are most important in the establishment and maintenance of a galactic wind are the Type I supernovae rate, the energy output from each supernova, the efficiency of the coupling of the supernova's energy to the ISM, and the amount of "pre-heating" of the ISM by the velocity dispersion within the galaxy. Each of these quantities are known to probably a factor of 2 at best and in certain instances various authors disagree by factors of 10 or more. • tii., tr i Kiiw. ^MM>Cfr*ivUJu*^,>>iC^',>t^'ti ^'m *>*n vii.'-i'-*i-wv

PAGE 24

-16For example, Mathews and Baker (1971). assume a coupling efficiency between the expanding supernova shell and the ISM of 1; that is, all the kinetic energy of the supernova is converted into thermal energy of the ISM. Gisler (1976) takes exception to this number and notes that Larson (1974) uses an efficiency of 0.1. Given the various uncertainties it appears that while a galactic wind will most likely prevail in some instances, perhaps even a majority of elliptical galaxies, there are cases in which it simply does not operate. Indeed, Mathews and Baker (1971) find solutions in which a wind is not supported and the material collapses to the center of the system. They further propose that the hot, ionized gas will only be able to form massive objects, thus linking the lack of a galactic wind to the formation of radio sources in early-type galaxies. Again, Gisler (1976) points out an inconsistency in this line of reasoning. From observations one finds that strong radio sources were more common in earlier epochs. Gisler notes that the earlier stellar content of ellipticals is more likely to produce supernova. This follows from the observation that only Type I supernovae occur in population II (old) stars and the precursors are probably low mass stars (Tammann 1974). In the earlier stages of an elliptical's life the supernova rate can only be augmented by Type II supernovae (whose progenitors are young, massive stars). In addition it is at the earlier epochs that the galaxy will not have had time to collect a significant amount of gas from the evolution of its stellar component. For these two reasons it would appear that the ellipticals are better able to support a galactic wind at earlier epochs--just the period when the greatest fraction must also be radio sources.

PAGE 25

17Faber and Gallagher (1976) have considered the problem of a galactic wind from a slightly different approach. Since the initial heating of the ISM is through collision of the gas clouds shed by stars in the process of normal evolution, the velocity dispersion of the stars can be used as an indicator of the stellar equivalent temper2 ature, T^. From Mathews and Baker (1971 ) T^ = v^ mu/6k where v^ is s s s n s the stellar dispersion velocity and mj, is the mass of the proton, and k is Boltzmann's constant. Along with the condition that the supernovainduced rise in temperature must roughly double the kinetic energy of the gas to remove it from the system to infinity, one finds 'sn ^sn ^s"^s (D where a and a are the specific rates of mass injection by supernovae and stars, respectively, T and T are the equivalent temperatures of the supernovae and the stars. Using the values quoted in Faber and Gallagher (1976) a = 5 X 10"^ sec"-^ and a T_ = 1.6 x 10"^^ K sec"-^ yields T = 3.7 x 10^ sn sn -^ s K or V < 1260 km/sec. The line-of-sight velocity dispersion is < 1260/v^ = 730 km/sec (assuming 3 dimensional isotropy of the velocity dispersion), which exceeds by a factor of two the largest velocity dispersions measured (Faber and Jackson 1975). This calculation would make it appear that virtually all elliptical systems should have a galactic wind scouring the ISM of any material. However, Gisler (1976) recomputes this same quantity, taking into account the 0.1 efficiency of energy coupling found by Larson (1974), and finds that the maximum velocity dispersion that will still allow the supernovae to double the energy is about 200 km/sec, a number comparable to the observed velocity dispersions. In other words, -•urii i f C JT !— ^.:^ii>fc-4'^ •*

PAGE 26

-18if a galaxy has a velocity dispersion greater than 200 km/sec then the supernovae contribution will not be able to double the kinetic energy of the ISM and a galactic wind cannot be established. Even more damaging to the galactic wind hypothesis is the detection of HI in any elliptical. In order for the wind to operate the ISM must be hot (^ 10 K). All of the gas in a galaxy would thus be ionized and according to Mathews and Baker (1971) quite unobservable by present techniques. A final argument against the universal existence of galactic winds is that if all other conditions were the same, one would expect the more spherical systems to be better able to support a galactic wind. The reasoning is that the spherical systems have less surface area per unit volume through which to radiate excess energy, keeping the ISM as hot as possible. Contradicting this expectation, the neutral hydrogen observations of 8 elliptical systems show detections significantly skewed towards the more spherical galaxies. The systems detected in neutral hydrogen have the following classifications:' ,; 2-EO, 2-El, 1-E2, 1-E3, 2-E4. Conspicuously absent are the flatter systems that one would expect to be better able to radiate energy away and thus retain their ISM. Ram-Jet Stripping by an Intracluster Medium The proposal that ram-jet stripping of an ISM could be significant in the evolution of a system was treated first by Gunn and Gott (1972). Later, more sophisticated treatments by Tarter (1975), Gisler (1976), and Lea and De Young (1976) all support the notion that stripping can be an effective process. :^a*...^w>^j^^ i i' — '^''-j-MlLJi'if -~i:yT* t". t liif 'T-jr"''"''^ AM ti'l'^' a ^^'\^

PAGE 27

19The thrust of much work in this area has been to determine if SOs can be formed by stripping spiral galaxies of their gas and dust (Gisler 1979). This would quench star formation and significantly change the optical appearance of the galaxy. Another development that has spurred interest in the interaction of an ISM with an intergal actio medium (IGM) is the discovery of head-tail radio galaxies. The most straightforward explanation of this phenomenon being just such an interaction. In spite of the varied motivations for these studies many of the numerical simulations are directly applicable to the analysis of an elliptical system passing through an IGM. :;, The primary results of these studies are to confirm that under appropriate conditions there is an effective sweeping out of material from the galaxy. The procedures and model parameters used to arrive at this conclusion vary substantially for the different experiments, all .agree however, that some material tightly bound near the nucleus may be retained. Gisler (1979) explores the situation further and finds that the rate of gas replenishment is important, possibly stopping the stripping effect entirely if it is high enough. In spite of this it appears that stripping can be at least partially effective over a broad range of galaxy velocities and IGM densities. This finding agrees well with the observation that a large fraction of galaxies in rich clusters are SOs and ellipticals (Oemler 1974, 1977). It would seem also that the evidence of a positive correlation between X-ray luminosity of clusters (presumably from a hot intracluster component) and SO/spiral ratios (Tytler and Vidal 1978) argues strongly

PAGE 28

-20that the cluster environment does indeed have a significant influence on the structure and evolution of its component galaxies. An evolutionary effect may also have been observed by Butcher and Oemler (1978). Their study found that a rich cluster observed at a redshift of 0.4 contains many more blue galaxies than a similar rich cluster observed in the current epoch. Their conclusion is that as Q late as 4 X 10 years ago the galaxies in this cluster had not yet I been stripped of their ISM. They were consequently undergoing at least moderate star formation, thus producing the blue colors observed. One further piece of evidence that fits in quite well with the general hypothesis of ram-jet stripping is the common coincidence of a strong radio galaxy at the center of a dense cluster (McHardy 1974, Guthrie 1974, and Riley 1975). The argument in this case is that the central galaxies have a small velocity with respect to the intracluster medium and will be more likely to retain gas shed by its stellar population. The material collapses to the center of the galaxy, apparently forming a massive object and producing the observed radio source. In view of the variety of indications that imply a substantial interaction between an ISM and an intracluster medium it seems clear that environmental factors can be important in the evolution and structure of galaxies in clusters. But in the particular case of the detected ellipticals NGC 4278 and 4636 there is reason to believe the IGM is unimportant. These are the only two galaxies which have an HI distribution that is extended enough to map. In both instances the HI distribution appears to be considerably wider than the photometric diameter of the

PAGE 29

21system (Knapp et a1 1978a, Bottinelli and Gouguenheim 1979a). More importantly, the distributions appear to be reasonably symmetric, a condition difficult to reconcile with stripping or partial stripping by an intracluster medium. This discrepancy is pursued further in Chapter IV. Other evidence that ram-jet stripping may not be as effective as the numerical analyses indicate is found in the Hercules cluster, a rather loose cluster composed almost entirely of spiral galaxies. The inconsistency is that it is also an X-ray source. Our current understanding of cluster X-ray sources necessitates a hot ('^^ 10 K) intracluster medium as the origin of the radiation. How the Hercules galaxies have remained spirals and not been stripped is not understood within the framework of current research. Even more perplexing is the origin of the intracluster gas, it is usually thought to have been the gas removed from the galaxies. A recent statistical study of cluster morphology by Gisler (1980) shows that the anticipated presence of Sc galaxies in clusters is not found. The Sc galaxies are expected to be highly resistant to having their ISM swept because they have a high rate of gas replenishment (Gisler 1979). The apparent underabundance in rich clusters indicates that ram-jet stripping cannot be the dominant mode of SO production. Dressier (1980) comes to similar conclusion based on a study of the morphology of the galaxies in 55 rich clusters. He finds a significant number of SO systems in clusters which have too low a density to accomplish any stripping. Further, the study finds a difference between the bulge-to-disk ratios of spirals and SOs as well as a tendency towards thicker disks in SO systems. From these

PAGE 30

-22observations Dressier (1980) concludes that spirals do not evolve into SOs. He concludes that the galaxy types are affected more by the initial conditions at the time of their formation than by environmental factors such as ramjet stripping. TfcpBry7i>->.H"'>'i<^ w i ^ -^f;^'rf>i"1i1g>i "w B| n -rWMttttaiaBMii w ;<*i<

PAGE 31

23Fate of Retained Gas Supermassive Objects in the Nucleus Mathews and Baker (1971) suggested that if their galactic wind should fail, the gas in the system would fall to the nucleus in an ionized state. They further argue that the Jeans radius 1 D r TrkT -1^ ^J 46yMG(p + p^)-J (2) where T is the temperature, y is the molecular mass, G is the gravitational constant, p and p^ are the densities of the gas and stars, respectively, and k is Boltzmann's constant is determined primarily by the stellar density. That is, as the gas collapses it responds to the gravitational field of the stars. This will continue until the gas becomes more dense than the stellar component. The gas is dense and collapsing quickly at this stage and Mathews and Baker suggest that there may not be enough time for fragmentation to take place. The collapsing material then forms a massive object rather than fragmenting and forming stars with a normal distribution of masses. Another argument for supermassive objects is the observation that cD galaxies with radio emission are often located in the center of dense, rich clusters. The evidence is largely circumstantial but if ram-jet stripping is important in the evolution of ellipticals then it follows that the central members of a cluster will be least affected by this mechanism. Of course the step from ineffective ram-jet stripping to a supermassive object in the galaxy's nucleus is by no means secure. It rests on the assumption that the retained material

PAGE 32

-24either forms the supermassive object or at least provides fuel for the radio emission. These arguments actually rest on much firmer ground due to recent work on the velocity dispersions and light distributions within the nuclei of supergiant cD galaxies found in the centers of rich clusters. Young et al (1978b) obtained luminosity profiles of the supergiant elliptical NGC 4486 (M87) which, when examined with the velocity dispersions determined by Sargent et al (1978), show that the nucleus contains a massive dark object. The nature of the dark object cannot yet uniquely be determined, but it must contain 5 x 10^ M of material and have a radius less than or equal to 100 parsecs (pc). Young et al. also determined that the mass-to-luminosity ratio must be greater than 60. Several possibilities are advanced but Young et al find the most plausible to be a massive black hole of 5 x 10 M The attraction 4? 1 of this hypothesis is that the 10^ erg sec" energy output of NGC 4486 can be explained by supposing a mass infall of about 10"^ M yr~^ with only a 0.002 conversion efficiency into radiation. De Vaucouleurs and Nieto (1979) confirmed the results of the Young et al (1978b) study and found essentially similar results for the dark mass at the nucleus. The earlier results were obtained with a COD (charge-coupled device) camera, while the work by de Vaucouleurs and Nieto was with more conventional photographic and photoelectric photometry. Young et al (1979) also examined the luminosity profiles of NGC 4874, 4889, and 6251 and found that only NGC 6251 requires a supermassive object at its center to fit the data. Further, only NGC 6251 is a radio source amongst the three. m^-.rT-".~'-^^^(n'^>ni'*mfitn^-^-' "-" j ^ p P. t f*vfa<-;r>y>— n^'.-t

PAGE 33

-25It appears from these observations that there may be a correlation between radio galaxies in the center of clusters and anomalous nuclei. Since current understanding of radio sources usually involves massive objects (which can also explain the anomalous nuclei) the circumstantial connection between a lack of ram-jet stripping and a supermassive object in a galaxy's nucleus is established. However, within this scenario it is quite unclear whether the retained gas actually formed the massive object or just fuels it. The possibility that the galaxy formed with a massive nucleus cannot be overlooked; indeed, one of the central questions yet to be answered is how important are the initial conditions under which the galaxy formed. Cyclic Bursts of Star Formati on The contention that the star formation rate in an elliptical or SO galaxy is strongly dependent on time gains credibility only recently. Van den Bergh (1975) cites several examples of elliptical galaxies experiencing anomalously vigorous star formation. NGC 5128 (also known as the radio source Centaurus A) has recently been shown by van den Bergh (1976a) to be undergoing very active star formation along and interior to its prominent equatorial dust band. He also suggests that the source of the unusual dust and gas in the system is stellar debris shed by the stars. The galaxy is apparently a rare field elliptical. It does not belong to a rich cluster and presumably lacks any ram-jet stripping which may exist in such an environment. A strong argument against this hypothesis is the finding by Graham (1979) that the old stellar population rotates much slower than

PAGE 34

26the equatorial band of dust and gas. This, is difficult to reconcile with the interstellar material originating with the stars. Most other researchers in the field tend toward the explanation that the system may actually be a collision between a gas cloud or small galaxy and an elliptical (Graham 1979). Thus the relevance of NGC 5128 to elliptical galaxy evolution cannot be assessed until these questions are answered. Van den Bergh (1975) has also suggested that NGC 1275 is an example of an elliptical galaxy caught in a burst of star formation. The underlying stellar distribution in the system is elliptical and the galaxy is located near the center of the Perseus cluster. The small velocity difference between the cluster's mean velocity and that of the galaxy again suggests inefficient ram-jet stripping. However, more recent work on the system (Kent and Sargent 1979, Rubin et al 1977) indicate that this is a collision between a foreground spiral system and NGC 1275 in progress. Thus both NGC 5128 and 1275 may be atypical as far as early-type galaxies are concerned. However, these two examples support the view that a massive (a, 10^ Mg) influx of material over a relatively short time period results in vigorous star formation. As both these systems are also radio sources, one could argue that another effect of the mass accretion is to either form, or fuel a previously existing, massive object in the nucleus, producing the radio source. Further support for this hypothesis, although not in an early-type system, may be found in the M81-82 system. The interaction between these galaxies seems to be resulting in a substantial mass infall to M82 (Gottesman and Weliachew 1977, Killian 1978). It is likely that

PAGE 35

-27the infall is connected with the peculiar structure of M82 and its vigorous star formation. On a gentler scale there are several examples of star formation in dwarf ellipticals in the local group. NGC 185 and 205 both contain dust patches and a sprinkling of blue, presumably young stars. The CO observations described in Chapter III show that the star formation is probably continuing in NGC 185. NGC 205 is a close companion of M31 and it may be argued that an interaction is taking place, although HI maps of the region do not support this idea. Regardless, NGC 185 is considerably further from M31 and an interaction does not appear likely (see the HI observations presented in Chapter III concerning this possibility). Consequently, it appears that these two dwarf ellipticals are reprocessing their ISM back into stars. The most likely source of the material is the normal products of evolution of stars. Large scale interactions are apparently not supported by the HI observations of the galaxies. The more sedate pace of star formation, and various arguments suggesting an initial mass function for the stars shifted to lower masses, are examined in the following section. Continual Star Formation with a Skewed Initial Mass Functi on A third fate that may befall gas retained in early-type systems is star formation. The most stringent requirements on the nature of this star formation are set by the observations of the colors or earlytype galaxies (see Larson and Tinsley 1978 for a discussion and earlier references). ^--••(7****i>i.,titfe,r*t>,^|^.,n*.f>.J[4*i**(V-i

PAGE 36

-28Larson and Tinsley (1974) have calculated models for elliptical galaxies with star formation rates, continuing to the present, capable of consuming the gas shed by other stars. While the integrated colors of the models are consistent with observed galaxies, the expected gradient of increasingly blue colors in the nucleus is not widely observed. The most obvious drawback in their modeling is the use of an initial mass function (IMF) which is fairly rich in hot stars. As Faber and Gallagher (1976) comment "Since we have no a priori knowledge of the IMF in ellipticals, star formation might conceivably be confined to stars of small mass and low luminosity" (p. 370). They go on to point out that the star formation must proceed efficiently since very little, if any, interstellar material is seen in most elliptical galaxies. There is, however, observational evidence for star formation with an anomalous IMF. Van den Bergh (1976c) suggests that an IMF deficient in high mass stars is the most likely explanation for the lack of HII regions in the Sa galaxy NGC 4594 (M104). Knots of young, blue stars are observed near the prominent dust lanes. Normal, massive and B stars would form prominent HII regions under such conditions. Further, this galaxy was not detected by Gallagher et al (1975) in HI even though a normal dust-to-gas ratio indicates it should have been easily seen. Van den Bergh suggests that the lack of HI is due to its being converted into molecular hydrogen, thus escaping detection. On theoretical ground also, an IMF skewed away from massive stars may be expected. Jura (1977) finds that one effect of reduced heating of interstellar clouds (expected in elliptical galaxies) is to allow clouds with much smaller masses to become gravitationally unstable ^...^£^.JV^.^:li-t>v^fol^^^i,.>j^^ ^^ll ^ f^ v ^,li A l^^ ^• '*.'*^l^l^^C^

PAGE 37

-29and collapse. Indeed, he finds that the critical mass for collapse under such conditions is less than 10 M^. Because of fragmentation the result of the collapse of a 10 M^ cloud will almost certainly be a number of small mass stars rather than one 10 M star. Thus while no compelling reason can be advanced to accept star formation with a skewed IMF, the possibility cannot be rejected either. The main intention of this dissertation as outlined in the following section is to examine this possibility to determine if it is a significant process in the evolution of early-type galaxies. Statement of the Dissertation Problem From the previous sections it is clear that our understanding of early-type galaxies is incomplete. There are a large number of interesting ideas concerning their formation and evolution, but as usual, insufficient data are available to assess adequately their importance. The primary goal of the observations presented and discussed in this dissertation is to investigate the significance of star formation in the evolution of early-type galaxies. The results of this study are most directly applicable to the recycling of material by star formation in an environment much different than our own galaxy. The results also bear on other settings in which the star formation is poorly understood. It is not overstating the current situation to say that we only have a dim view of how star formation occurs in our own galaxy, and an even dimmer view of the process in drasically different environments. Faber and Gallagher (1976) argue that possible star formation in U< >£C*gaJf.i^Ji)tyDtji-m.V I n .! —I t>an.~^i j"i i m i-iQi 1^

PAGE 38

-30early-type systems must be very efficient; if so, it may be generically related to the star formation that occurred as the galaxy collapsed originally. Judging from the lack of primordial material observed, that process was also very efficient. This work is also of importance in the formation and continued activity of radio sources in early-type galaxies. From the available information it seems that radio sources and star formation are competing for the same interstellar debris in a galaxy. If one is very successful, it may be at the expense of the other. All of this is not to say that galactic winds and ramjet stripping never occur, simply that all possibilities need to be analyzed and evaluated to determine their relative importance in the overall scheme of galaxy evolution.

PAGE 39

CHAPTER II THE OBSERVATIONS This chapter presents the data acquisition and handling procedures used for the work described in this dissertation. The instruments used are described giving particular attention to the equipment and techniques which aided this work immensely. The method of data presentation is also explained. Carbon Monoxide Observations at Kitt Peak, AZ Telescope Description The observations searching for the 2.6 mm transition of CO were made at the National Radio Astronomy Observatory (NRAO) Millimeter Wave Telescope at Kitt Peak, AZ. The telescope is an 11 m paraboloid which can be driven in altitude and azimuth. Tracking of celestial objects, data acquisition, monitoring of system status, as well as initial data reduction is handled by an on-line POP 11/40 computer. The observations were taken during three separate observing sessions December 23-26, 1977 July 7-9, 1978 December 10-16, 1979 The National Radio Astronomy Observatory is operated by Associated Universities, Inc., under contract with the National Science Foundation. -31-

PAGE 40

-32The cooled mixer cassegrain receiver operating at the 115.2712 GHz assumed rest frequency of the J = l->0 CO transition was used for all observations. The halfpower beam width at this frequency is about 65". Pointing for the telescope is initially determined by the observatory personnel. Pointing correction data are taken in all parts of the sky on bright mm point sources. Analytic functions are then fitted to the data, interpolated over regions with few data points. The observer then checks these corrections by making five-point maps of bright sources, usually the planets. Rarely are the corrections more than 5 arcseconds at this stage. However, the 1979 observations were hindered by large errors in regions far from the celestial equator due to a lack of adequate data for the fitting equations. This problem is discussed further in Chapter III. Nominally, the sky signal is separated into two linearly polarized orthogonal components. In practice the system operated with only one polarization, except during the 1979 observations when both channels were available. These two signals are then fed to two Shottky-barrier mixer diodes. Also fed to the mixer is the local oscillator (LO) signal via a tunable injection cavity. This process of mixing two frequencies to get a third (usually the difference of the two) is known as heterodyning. The first intermediate frequency (IF) signal emerges at 4750 MHz and is amplified by a pair of low-noise parametric amplifiers. All of the previous equipment is enclosed in a dewar cooled to 15 K. *W'''t.1*lJfl"*>1k*-r—*f-*^ -^>.^'^>^^^=^1^"s^~-y^*^*-*y^^#-,^>^a)^> ^' ^.J n' ji>^^<^,ii t -:f in i a Mj)i l^yyT^f*=^t^-^^^ :fa:S 1^ lyiaA^-'

PAGE 41

-33Once outside the dewar the 4750 MHz first IF is amplified again by room temperature transistor amplifiers. The signal is heterodyned further at this point to 1328 MHz. A third and final mixer produces a 150 MHz third IF which is detected by two banks of 256 square-law detectors. One has a choice of filter widths for the filter banks; our observations always used either the 0.5 or 1.0 MHz (per channel) banks, corresponding to 1.3 and 2.6 km/sec velocity resolution, respectively. The December 1977 and July 1978 observations were made with only a single polarization operating, so the separate channels were actually the same signal and could not be averaged to reduce noise. Consequently the 0.5 MHz (1.3 km/sec) and 1.0 MHz (2.6 km/sec) filter banks were both used. At the time of the December 1979 observations the receiver operated in a proper two-channel mode with two independent IFs, Both were fed to 1 MHz filterbanks since they could now be added with a resultant decrease in noise. Calibration was accomplished by alternately observing blank sky and an ambient temperature microwave absorber. Details of the calibration procedure are explained in Appendix I. As a result of the calibration procedure one obtains the corrected antenna temperature of the source TJ = rifd e"^)[J(v3,T^) J(v3,T^g)] (1) where the explanations for the symbols are found in Appendix I. Note that there are three unknowns (n^, x, T^) related to the particular source being observed; these cannot be determined without additional information. For extragalactic work one can make estimates for T and T^ based on galactic studies and then calculate n^ The a i g ig ar i g it g iCiB i ..^aitf *ri*

PAGE 42

-34errors are large in this approach and they give only a general idea of the nature of the molecular cloud being observed. Data Reduction Techniques Each spectrum consists of two channels of 256 points each. For December 1977 and July 1978 only one polarization was available and it was detected in both the 256 MHz wide filter bank as well as the 128 MHz wide filter bank. The December 1979 observations took advantage of the two orthogonal polarizations then available and detected them separately in two 256 MHz filter banks. At the rest frequency of the 12 J = KO transition of CO the 1 MHz resolution corresponds to 2.6 km/sec. The on-site PDF 11/40 computer writes the data for each spectrum (typically representing 5-10 minutes of observing) on disk. Real time analysis can be done at the telescope and allows the observing time to be optimally used. At the completion of the observing session the disk is copied onto a binary coded tape. The next step in the processing is to rewrite the tape with the IBM 360 computer at the NRAO in Charlottesville, VA, to put the information into an IBM-compatible format. This is entirely a translation and no analyses are performed. The final reduction is done at the University of Florida using the resources of the Northeast Regional Data Center. Each spectrum is examined for an unusually high root-mean-square deviation; all of the offending spectra are rejected. The baselines are then examined and if a polynomial of more than first degree is required to fit the baseline, the spectrum is discarded. i-annni' 11

PAGE 43

-35At this stage of the reduction process each point of the spectrum is examined to see if it is greater than five times the standard deviation for that spectrum. If it is, and the adjacent points are not, the point is replaced with the mean of the two adjacent points. This procedure is used to remove interference that involves only a single channel; multichannel filter banks are particularly susceptible to this type of interference. Each spectrum is then weighted by the inverse square of its standard deviation and combined with all other spectra on the same source. A preliminary first order baseline is fit and subtracted and the resulting spectrum is examined for possible features. If any are apparent the channels involved are eliminated from the baseline fitting procedure and a new first order baseline is calculated and subtracted. The resulting spectrum represents the best data on a given source. Various smoothing functions can be applied; the most common for this work has been smoothing with a rectangular function of about the same width as the suspected spectral feature. The effect of this is to maximize the signal-to-noise ratio at the expense of velocity resolution. The basic format of each spectrum presented is explained in Figure 2. Data Presentation The data are displayed with antenna temperature as the ordinate and velocity along the abscissa. Strictly speaking the abscissa

PAGE 44

-36represents frequency, but since the frequency of the molecular transition is already known the axis is calibrated in km/sec using the following relation \ = {t^^V-^^^ ^f^ / (1 V^) (2) where v is the center velocity of the observed band, v^ is the velocity of the source (corrected for the earth's rotation and revolution, i.e. heliocentric), v^ is the rest frequency of the spectral features, and V is the observed frequency. The ordinate, corrected antenna temperature (T^), is found by calibration as described in Appendix I. The unit is Kelvins (K), and it is related to the flux density (S) by the following equation S = 2 k n, T*/A^ (3) where n^^ is the antenna efficiency due to the loss of elements (spillover, blockage, and ohmic losses in the antenna), A is the effective area of the antenna, and k is Boltzmann's constant. Neutral Hydrogen Observations at Green Bank, WV Telescope Description The neutral hydrogen observations were made with the NRAO 43 m radio telescope at Green Bank, West Virginia. The telescope is an equatorially mounted instrument completely under computer control. The observations were taken during one session from July 13 to July 23, 1978. ^..-.:-.. ^^. ^^--^.J Uj.,!...... x-i:...)^. ..--— .-.. -1-^.. .;-r.:^^--^^-Tw.wj.-Y^^ ^m.£. >tV— ....^.i-.;.^^^-.

PAGE 45

37The 21 cm cooled cassegrain receiver was used for all data acquisition. The system has two channels provided by linearly polarized, orthogonal feeds. After initial amplification by a cooled upconverter amplifier the signal is heterodyned and amplified through various stages in much the same fashion as the process described for the 11 m telescope at Kitt Peak. Typical system temperatures for the 43 m system were 50-60 K. The standard NRAO "back end" uses a 150 MHz IF which is fed into a Model II autocorrelator spectrometer. The IF signal is autocorrelated and the resulting autocorrelation function is Fourier transformed to produce the power spectrum. The formation of the spectrum using autocorrelation techniques is described in more detail by Blackman and Tukey (1958) and Cooper (1976). A 10 MHz bandwidth was chosen for all observations to provide an adequate baseline. Also in the interest of baseline stability a position-switched mode of observing was adopted. Ten minutes of data are taken at the "off" position followed by 10 minutes at the "on" position. The final spectrum is found by differencing the two spectra thus acquired. A number of "off" positions were used in an effort to deduce the distribution of HI in the region around NGC 185. The majority of the data were taken either with an "off" 16"^ to the west or as a five-point map. The arrangement for the five-point map is with the "on" at the center and "offs" taken successively to west, east, south, and north at a distance of either 48' or 24'. The half-power beam-width at 21 cm is about 20.5',

PAGE 46

-38Data Reduction Techniques The output of the autocorrelator is two 192 channel spectra, each being linearly polarized but orthogonal to the other. The spectra are recorded by an on-line disk drive which can also be accessed by the on-line reduction computer. In operation, the near instantaneous access to the data just taken enables the observer to monitor the quality of the system operation and to update the observing procedure based on the preliminary results. A characteristic of the autocorrelation method of spectral analysis is that the strong galactic hydrogen within the bandpass produces a sinusoidal ripple in the spectrum. This is known as "ringing" and its removal is accomplished by convolving the spectrum with a hanning function. This function is a weighting scheme in which i the value of the channel on either side is added to i the value of the central channel to produce the new value for that channel. Application of this smoothing worked very well and all data presented here from the 43 m telescope have been smoothed with the hanning function. A baseline is removed from the data by fitting a low-order polynomial to the spectrum in regions removed from either galactic emission or suspected NGC 185 emission. In practice the order of the fitting polynomial was 2 to 4. Calibration for this system is done under computer control by periodically firing a noise tube within the receiver and comparing the system output with and without the additional noise. The data are then scaled to this system temperature. The stability of the system was monitored by observing several sources throughout the session. No unexplained drifts in system performance were seen. ~ ^Si.^mSz.^^^^:':^ii!}BUff^[i^^Aix''fiy^..^^£i^^^i^^^ '^-^ ^>.'i;t-^-,*'^.-^.^.yv"^Pp* == ''~'< =? ** ^-

PAGE 47

CHAPTER III RESULTS In this chapter the results of the observations are presented. The carbon monoxide data are treated first, followed by the neutral hydrogen study of NGC 185 and the surrounding region. Only the results are considered in this chapter; the analysis and interpretation of the observations are found in Chapter IV. Carbon Monoxide Positive Result in NGC 185 The dwarf elliptical galaxy NGC 185 was detected at the 115.2712 12 GHz frequency of CO. Figure 2 shows the best spectrum obtained on the source, averaging data from all three observing sessions. The 256 MHz total bandwidth corresponds to about 660 km/sec. The feature is not discernible in the 128 MHz bandwidth spectrum, probably because of the lower signal-to-noise (SN) ratio. Table 1 summarizes the parameters of the spectral feature as measured in Figure 2. Since the line is only slightly greater than 3 standard deviations, it is very important to be certain of the reality of the feature. One method is to smooth the spectrum with a function (in this case rectangular) of about the same width as the suspected spectral feature. This procedure has the effect of increasing the SN ratio at the expense -39HIT -![i"tTT ~ iT" ir n"" '-— '• -, — •^-•— '•^'T— :r >.-,i ^p. --|^.-'..--... -.^ .. ^,-, -., -j.-.^-j ^m— -an-iB r-iii-I --i— -*-tf^^"'^^'l.£i'^fr

PAGE 48

o (-> c 0) O) SOJ ^ CL 3 S_ +-> O (1) 0) CL-c: 00 +J CO to CXD C_) C5 4O to CD u 3 QJ o -o CD s_ > 3-0 3 O HI o c to > x: •rSl/l CU 0) to .C I/) o S4-> JD 00 to O t— -a SfO XJ E to +-> to CO +1 ro to CL 0-— • CL to n3 OJ O) C Tro O 4-> E fO rc +J E O N O c -Q O o 3 S+-> O CU CL t/) o o CM E -O O £rO -t-> o fO O SCDO CD • +-> o c tn rCT^ +-> o E o CU o lj_ 4-> o +J rJE +J 4-> to 4-> CU -rro to "O -E to •!— CU -1+-> O > to 3 CL CU o ^ -o CD E +-> CU E (O TJ= -O IX! S +J S> ro S1 — cn+J "o (U to E J= E to +-) O C71 (O ja O t— •.+-> O -P to i. to cn E O to E o to to E > r— to CU E i=^ o •r— o +-> (O o f— > 4-> CU
PAGE 49

-41SO'D i\\] 3yniuy3dH3i oroI (>f = J A-'*.^ltf*r***JV1

PAGE 50

-42of velocity resolution. Figure 3 is the result of this convolution, and indeed the reality of the feature is strongly supported. Table 1 Line Parameters for CO in NGC 185 Peak Antenna Temperature (K) 0.081 0.025 Velocity of Line (heliocentric, km/sec) -175 2 Full-width at Half-maximum (FWHM) (km/sec) 12 2 Position Observed 0*^ 36^" 10^3 48 3' 53'.' 3 Total Integration Time on Source (min) 484.5 Other support for the reality of the line is found by careful examination of the spectra obtained during each session of observing (Figures 4 to 6). The spectral characteristics vary somewhat for each session. But this is to be expected since each individual spectrum has a relatively low SN ratio. In particular the line strengths for each session are December 1977 0.20 + 0.07 K July 1978 0.20 0.08 K December 1979 0.075 0.028 K where the quoted errors are 1 standard deviation. In view of the agreement here, and the pointing uncertainty during the December 1979 observations, the discrepancy among the observations is not large. During the December 1979 observations the cooled receiver operated with two independent channels, linearly polarized and orthogonal to one another. The feature does not appear in the spectrum from the '*>*Jlqc:y^mj^=gnJ'-wy:=afi *jta::vw*-^;^^nhi*j-; .-in j^-^j^^-.-y--^^4r*^— jw^^rtv jfcjfc ^^^ >" la t ^' infc'^Mitg^aw iic' a *-^K^tf-=*sp^% .^= ^ >"4-^afpr^'^J>.t' .•^^Ji^?*^|Ul>^;v-*^•*— i,p^ y*t?%>?V**M t3 i*^f

PAGE 51

-4301 "0Lf"! ro O.J O CO rn 1 — 1 II LU Lr"! II (T3 O 5O -i<; n cu u cu o o iCD 0) o Q.T1/5 +-> O I— 14DI '0 SO'O oo'o so"noro-

PAGE 52

-44a-/ ro -U" (, ) I + r^^ ^ Lj~ ( n „ LTi 0"; OS'O 02 TiU 3 u 00 >Sl o fO r—l LJ-i T— 1 0"i 0-1 o x: o n,i 33 r-O =H II < LJ ''•jLL 1 — 1 oz 09 "0 02 '0 02-Qr I LJJ CO r-j I -" 09-0i^j 3yniby3dW3i 00 -Iscu E O) <_) cu a E o s-p (0 •o o o CM O) iwnTiiW'W t*^^'-* rti. ^w ; ift-T fc w ia o ^iffi gawff^n ^rn i a 'lj m !*<

PAGE 53

-45c'j I CO en — I i I 'M Ut'O QZ"Ci03-0 ^ :-^=~ DS'Dn n : Ll.' D 01 o_ T— 1 u-i IE ijii CTi 0"i CTj rr.i 4 X o 1-1 CO CO D II CO 1 i V -, — ; rj Q -:3 = [d E o CO X'i1 fO p o o o o 1 — 1 CM r— t O) LO .— 1 OJ o 3 03 LO 1 •

PAGE 54

45£1 o B o O o o IX) orod' O) l_l i:3 CO en 1 Ll. 3uniuy3cJW3i igg't^ii^>iC=i^M W i^ia o o^ I mim r M ^'.n ^ i M >^i* T^U--:fag; >.C^ ..*;nU^<^

PAGE 55

-47B'D CD 173 CO I U" r>-4 CD CT LT) rr 2ro fiO'O fiOTj?ro 02 '0^1^^-=^ LU T— 1 ID n--;i D o CO ij2i II Q ij"i E ^Q-ri TiO'Osro[>n 3yniHy3dN3i OB-QCQ — (V c C_) c LU JC CO o '^--^ n n cn 1^ ", !— 1 II S-I -^
PAGE 56

-48second channel of the receiver (Figure 7), We have no physical reason to expect a polarized feature and indeed, this transition has never been observed to be polarized in our own galaxy. Also, the tracking procedure rotates the observed polarization planes with respect to the sky; even if the signal were linearly polarized we should be able to detect it in each channel. The lack of a feature in this channel is almost certainly due to the low SN ratio. A further complication during the December 1979 session was poor pointing corrections for the telescope, especially in regions far from the celestial equator (the declination of NGC 185 is 48). Consequently we feel that the lack of a feature in the second channel is within the statistical and instrumental uncertainties of the experiment. Negative Results A total of 11 early-type systems were observed during the three observing sessions. Our criteria for selecting candidates for CO emission were based on the notion that a system able to retain HI or dust would be more likely to have significant star formation. Consequently we chose several of the early-type systems that have been detected in HI. Added to the list were galaxies that are close (< 15 Mpc) and also show some form of an ISM. Often this was a notation in the Reference Catalog of Bright Galaxies (RCBG, de Vaucouleurs and de Vaucouleurs 1964) that the system contained obscuring matter. Table 2 lists the 11 galaxies selected for the CO study. Included are the positions observed, the assumed distance and radial velocities as well as the Holmberg magnitude and the total luminosity. The dates in which useful data were collected are also noted. The final quantity tf7>gi^''^->'y VX^^^'J i ^ ra w g -t^=t;=>;i.i*M.i>.catrj>-wr-'^ -..*^^,.i— ,^ m M -y i.;^,.^*— •i-i**---??!!,;;! -^i^Tis. fc5*i

PAGE 57

-49so -a 4->--^ C n3 :^ (O -r---— +J > 00 QJ CTl CTl CM J3 O C_J CM io qO) J= o SCD Ul CO OJ •rX ta I— CD 0) Q. io cu CO O LO cu +j fO Q o cu Q CO en CTl o cu Q I a +J CD X +-> CU o +-> OJ +j cu CU cu X Lf5 CM O t— I CO (--. "=dCM CD O O O CTl CM CM OO 1 — en O CO CD •vtCD =f Ln 03 =dOO O o O o o O T—i o o cna\ a. o o OJ to E o CD. CD CM o cu Qo LO en o en =te C_) CJ3 LO CM cn CO oo LU OO OO o CO oo CO • o CD OO o o LO CO Q. CM OO o CO CM CM CD OO o o LO OCO Q en U3 LiJ OO 00 LO CM o CM LO CD LO o CM o CO o CM LO OO UJ LO o CO CM CO I CD. O Q. CQ CM C>0 UJ CO OO o 00 CD en LO CM r-l 00 I— 1 LO CO LO CO LO en o o CM o o OO OO OO OO en en CM CM LO CM o LO CD o UJ o LO o OO o CD LO o CM OO o o I LO o CM CTl OO CD o o 00 CO 00 CM o CM LO o I— I "=dCM •-t o LO CM CO CD CO o LO LD 00 CM CO CM CM OO ^ o LO LO ro I— I CM CO CM OO o CD OO CO CD LO 00 o CO •^ CO LO t— 1 LO l-l LO r-(£> cn CO CO cn CM cn r-^ O o I— 1 o CD CM CM • • • • • • LO OO ro O o CO 00 =Jo "=dI-H 1— 1 1— H i-f 00 I— 1 o oo CM CM >=*• I— 1 o o I— 1 I-H CM CNJ CM t— 1 o .— 1 1— ( an en CO 00 CO UO 00 CO LO o en CT< LO CM CM CO =dCD 00 rLO t^ OO cn 1 1 r— 1 CO 1—4 en CO
PAGE 58

•500) x: +J 1— (— QJ (O 3 O o • 1 — CJ LO >,,_ .,_ .^ O) (C s> CD 4-> OJ sc -a 3 •> (U 3 00-— u +-> r-H o •^>)r--Ic .i<: cyi 1 — en OO <— 1 QJ (T3 ---^ E s_ rtS C T3 en E E C SO fO ra 0) >— E JD ,3r ^ 4-> -p -a OJ E >,jc: E (0 X t— O n3 i+-> c— tOJ (0 • 3 C7). — ~ -O O UD-^ QJ .— OJ r-~ Ln sI— x: cr> r-^ 3 fO (— >— 1 CTl T3 c/1 CD 1— 1 3 (0 E (T3 QJ O r— (C 1— CJ r— fd +-> 3 (B Ln Dl QJ (O CD CO 3 > .-1 SO >, QJ r— QJ .Q O -C 1— "O CD -P rd T3 S O CD E O) O JD SQJ S_ S-rO ^ O HS<+-P O "0 1/1 to S_-~. SOJ E O rO fO T3 o Mr-~ •1— of^ to to •P to .-H QJ (0 rE-— -rto O -P ID O T. TQJ 0-+J ,— CJ -p •r(a o fd QJ t/1 .— 1— -E O -P QJ 3 1— QQJ > O tn QJ +J O

PAGE 59

-51listed in the table is the standard deviation for all observations of a particular source. Figures 8-18 show the spectra obtained with the 255 MHz filterbank. Also indicated in the figures are the radial velocity assumed for each source. In many of the spectra there are single channels above or below the 3 standard deviation limit. These are not significant because they fail to persist in further observations. This type of interference is relatively common in multi -channel filterbank receivers, and arises from stray voltages within the equipment. The implications of the negative results are discussed in Chapter IV. Neutral Hydrogen The region surrounding NGC 185 was examined to sensitive limits to detect any neutral hydrogen (HI) associated with the galaxy. The 43 m telescope at Green Bank was used in a total power mode as described in Chapter II. The arrangement of "offs" was varied throughout the course of the observing in an effort to delineate the HI detected. One series of observations used an "off" at about 16*^ west to allow a reasonable estimate of the total HI content of the region. Another sequence of observing used closely spaced "offs" (24' or 48') to the west, east, south, and north to get a better idea of the behavior of hydrogen around a specific region. For those observations the "on" was always at the center of the "offs." Neutral hydrogen was found at significant levels in several regions, both at the position of NGC 185 and several points to the north. Figure 19, which used a distant "off," shows a strong signal of about 14.2 milliKelvin (mK) at a radial velocity of -200 km/sec. The

PAGE 60

5201 '0 1—1 V:; CO • — CTi CO t—i OJ >, 1 — =s 1 — 1 "D (-:> c ET) r— o 1 — ) c 1 1 1 OJ r>*i ^ 1 — +-> "3 -P to -a t' 1 o 1 — 1 C_3 CM r-H =(J 02 -Q oro oo'o oro03-0,ffi O) CD i.W-J— til— J .''.^i^^*^!!**!* '•-"-

PAGE 61

-53o en a o o o E x: u 1—1 4J 1 — 1 n3 Q. a2< +J (/) 3 •a — 1 E O r^. ^ 1 CTl '— ,.-x:. r-^ -^ — Ol i-H 1 CJi i. LU 0) .Q ^ ^^_ OJ -T" o O) ,— 1 -^ a E 00 \-..J -a 1 0) 1 c '1— ""^ fO F~l 4-5 =^ o ro>. S1 4-> O O) Q. CO LO O O CM o CM O .-( CD D 3yniyy3dN3i 01 "0ri'*i*---i'"-i-ciW^**-^^

PAGE 62

-5402 "0 02'D]S'C[. J ui rrI '^ o LTiLO rUJ rf i"i iri OJ UT =c CZ' I'O Ul ij2! 1 — 1 rn II < r '-I ij-: l_C t^ cc 09-0 OS'O 09 "000' I>> (O O SCL t/1 QJ on -^ F c j^ fO J2 o io r~ •)-> >, — (O 4O! 1 ici; 3 E 4-> C r— •1— 3 +-> F "3 en QJ OJ ^ E -P OJ HJ= O 1— 4-> O ITS ;J4o •1— <:1+J iO fd CD 2: c rrt io SMOJ x: E +J 3 fO Ss+J o +J cu Z3 Q-J3 t/l n O — o fO CM CIJ r-l t. QJ -u x: o 0) en ^VW*")!"" I^|i='S

PAGE 63

-55OE'OXU") ( J |7 ''L. "->4:^ Cj2i Lj-1 cn r-j ij^i Ci < LJ 2: o CD o 3 S)-> o O) CL to o CM (1) 0) C71 Oc'IJ 02-0Qq-nCMJ 3idrilUU3dW3I 00 "I-

PAGE 64

56J I u J 1 = 1 i '— LI ,' ds-0 1 n 01 Q: L_i n~ 3!driiuy3dWji ^1 LU r-CO to CM o CD o Z3 S_ 4-> O cu Ol o o CM CM 0) s'0-

PAGE 65

57m D ( i ru Od I I ( I n cvj r-Li.yo c o T 1 D izrj ri~. 1 — i o O r-.j lD s: 09 '0j 3yniuy3dW3i _-<£> — ,-CM l_l -^ CM t ""^ 00 rr> O ., 1 ..: cu I— 2: 1 — So i'^~'l <+'— 1 E Ri=O OJ a. (A o CJ CM ZZ' 1—1 — CD : i -C 00' I
PAGE 66

58> •r(/I CD ta •.+J >* • r<£ . rC OJ 4J -o QJ -o lo c in3 3 0) t/l r— S3 13 O 0) o r^ =s rs (0 o> o 3 CLl fC T3 > +J OJ 3 Q J3 „ O O cu LO c/1 — 1 •^ =1E Jii o o ^ 2: OJ So 4itO F >^ rs -P sr4-) O O O QJ ^ Q. dj (/) > o 0) O — CO O "--> CM b r-l T3 J,i (0 0) so -C r-> 1— res o^ Ol Uc Oil II 3 Id 01 "0'"llHld3dh 02 "031

PAGE 67

•59S Q I -., ZIP ( ;) r-Ki if) roj rL' J. ijy u ==3= -^^, =^=SL r-ro r-1 ro r-IjZ< OJ iT.i IjZi LiO < LJ :^ 02'! i\\) did I 09 '0by3dN31 u } t: — CO .__ r~^ r-, ^i^ OJ o — •!:lO uli — C!3 1 — 1 so ( q,_, Zj 3 i^lJ Su O) Q. lO o o CM n n O) 0) s=5 en

PAGE 68

-60I i f I n Lie: U I I J ( C] 1^ ,1 x: r^ ij-j iS'O - u cu Q. CO O O CM O)
PAGE 69

-61Ill iXi i; i?0 iJli m IT! :^J s -f r-CO 0": =f x: o rr, U-i LTJC 0J = otj 02 "0 Qc'Lln9"0c>i;i 3yr!iuyddW3. 00 +J (/) 01 fO •1— '~ -t-> •r— •^ +J vo t/) en r— I— 1 r— ,_^ (A MD Sr-^ =5 Cn CD I—) 1 — *— -* =3 o • o r— 3 ra rO > 4J 0) -a on o 13 c OJ fO r— 2J 10 o i. CJ =s rj 0) m f— > 3 o cu o-o =5 fO +-> > =3 OJ Q • cu kO lo ^^ CO b Lf) J^ o 1—1 CO r^ I— 1 so MMO E >> 3 4-' i'r(-> (J o o OJ r— • Q. cu o to > O) 00 o p— ^•-^ o fO R CM •rJii t— 1 a re CO 0) sin -C oo 1ro OJ
PAGE 70

-62U4 'U OS 0£"0. ^ r-I L M ,1ri' IX! '--' i^' s I SI r1 — 1 TTrb-"i UT u-j i_i X o ijIi LTj Li-1 CO 1— i UT 11 CO < ID s: orci02 "0>i) 3yniuy3dW3i OE'DCO o CJ3 o +-> u o o CM cu 00 S-

PAGE 71

-53I-U L CI •[i^; ZIP U~ : — I. 1 X—L r>-4 r--~ CO J en ijzi a ZT < lij ro r* X"^ CO ^L_5 \tlL OT OT to OJ 1 L LU fO LxJ 02 "0 01 T! 00 '0 ni "0( T-oixi L>n 3yniu!d3dN3. 02 "Di. CIJ o — cu JSZ +> E • r• E -o Ol O) cn +-> o fO SO T3 •— >, •o JC c •r— r—^ (0 10 u •r— o 1 — C o aj •rS+-> ra •r01 o O cu Q, (/) ^F H^ o o 0) j= -a 1— c (0 • o tn o 00 <— 1 I— 1 I CJ c CD o cu JD E Z3 (/I S_ CI) 4-> sO 3 CU -M Qfd lO 0) <+1 — 1 :e CU x: E 1u r-H a CNJ 4-> ^ OJ CD x: r-^ hia> cu rs en

PAGE 72

-64position of the "on" is at NGC 185. A full-width at half-maximum (FWHM) for the feature is about 60 km/sec. It is apparent though, from Figure 20, that all of this HI may not be associated with NGC 185. The "offs" for Figure 20 are taken at either 24' or 48' (they have been mixed to increase the SN ratio for this spectrum) in the four cardinal directions. The fact that the signal is only 5.2 mK indicates that some, but not all, of the line has been subtracted out in the "offs." Further, spectra were generated separately for each of the four directions. Those to the west, east, and south show about a 3 standard deviation positive signal. However, the spectrum with an "off" to the north shows no significant feature in the NGC 185 velocity range. Continued examination of points to the north of the galaxy reveals a strong 40 mK feature about 70' directly north. The radial velocity of this feature is -180 km/sec with a RWHM of 40 km/sec. Five point maps of the region to the north also indicate HI signals as far as 2 north and 48' east with a radial velocity of about -180 km/sec. There is only weak HI emission at 49' north of the galaxy. This is shown in Figure 21; the feature is around 15 mK. It appears that there is a ridge or plume of HI that is strongest (43 mK) between l-2 directly north of NGC 185. It extends further north and east at a level of around 20 mK. To the south towards the galaxy the plume decreases in strength to about 10-15 mK. From the sensitive 5-point maps at the position of NGC 185, an excess of about 5 mK is found for the galaxy's location. These observations are summarized in Table 3. The interpretation of these observations is found in Chapter IV. -r*^T'i'-n 7Tft ^ i

PAGE 73

-65cs. I I UJ CO F (' n Li-:' r-Ci CO T— -I '. — I I i-OIXi UIJ'O =i^ '-. -I =H ;=: -• L i g Oc'O 01 "0 00 "0 01 Tl( i-oix) m 3!druby3 o to +-> CO 4-> (/I 0) o T3 (U o 4= 00 o CD 4O S+-> o O) • to +-> sI— I o O) T3 o OJ scn — .--^wr > w fg i ;

PAGE 74

-66ra o CD if) iEZi CO ^ CO ( ") o r-'-j i; ^1 O J (J. -I 0" 02' CI I-UIX. CID'CI c. ( I-OIX 00 '0 oroi.Tl IT) CO PJ S OT O =H lie < LJ < LJ ^ oro>i;i 3yniuy3dN3i LU 'J-J 02'0CO o CD o So en ^ E s+J o Q. O) cu SCD

PAGE 75

-67OJ-— ^ tj c E S-rO 3 E O — T3 E 3 03 O 0) O) CO E 3 ro -^ 00 -13Z-' +1 S E I 1 I — 4— 00 CO CJ3 4O o o >>^~cu +-> O (/) •1— O) ""^ O CO E o -^j^ <— E >-— +1 a o S-1rO +J--^ T3 ro i^ s= •-E rcs >•— • +J cu C/0 Q cu s_ Z5 4-> rC-— Si^ CU E E cu cu I— cu (d 4-> +-> n3 > cu to O o o •r— to > icu CO X! O O CJl O o Ln CTi E O •r— +J fO z: > o scu CO o o un o en CX3 ro CO o CTl o CM co o o * CTl C/1 • fO "vlE ^ CJl o o uo ^ ro O 00 CO CO • C£) OO o o o 00 00 fO ro o o n UD m r1 in CO un o LO ^1=:Jo o o o o o CM CM CXJ 00 00 OO OO OO c^ o o o O CO CO CM t-H t— I I I I 00 CM o CM CM i-H r^ CO tn "=30C3 LO CO 00 CO OO CO 00 to CO LO r-. "* en V .— I 00 OO I I OO o 00 co CM O O 00 OO CM un CO CO OO o o LO 00 00 o o 00 CO co o en LO I— 1 ro 00 00 o o o o OO OO o o CO CO CM • • r-00 OO .-H CO d00 o o o o o r-^ Ln t— 1 "* en .-H 00 00 CM 00 I— f CM CO Ln CM CO CO ^ O "* "=t Ln '=;3OO ^ OO O <— I o o cri CO Ln cd"vj"^ Ln >::ir— I ^ CO I— 1 o i-H I— I CM CO lO OO en o o o o CO CO 00 OO o o o o .—I CO CTl o o o o o o 00 o CO 5dLD OO o o =doo o o Ln CO 00 o o o Ln en co OO CD o ctO J3 JE SI E 4-> Ln (13 +J s_ +J O SCO LO SLO Ln Ln Ln s_ t; ^^ CO CO 00 00 CO +-> 1 — 1 2r t — 1 .—1 — 1 t-H •r2: C^ CO CJ3 C_) 2: CD cn CD CIS CIS CJ3 CM 1 — I Q^ <=i^ zi 2: — < r^ cu a'. -E 1 1— -.X • s_ -a Q-ctO) CLCJ^ =c +J = • T3 E CO E 00 E CD-rCU E 4-J CO •1 — TE (/) CU E x: fO C2.-l-> JT = -E -a 4+-> E 4-rfO 3 cu ~ — CU JD r— • 00 fO fO +-> JD CM J2 n3 CU OJ sE x: Scu O) -P 3 T3 -i^ XJ> S_ re = •r4-5 u. cjO fO 4in s+-> •rfO = -E T3 0 +J -t-> 0) E 00 ro s3 S•0 c^mct_ E s> •rCU CU Ol OO ix: a> 3 n3 )-' £ T3 CO x: to tU E 4J •!=3 •.5 cn -a +-> c cu TCO SCO ro 3 2 ro Q. r— SCU ^E "o CJ-rt cu S+-> ro a. qiU) x: CD cu 4-> cu OJ S-r+J OJ ro 2 E £. •r— J= CO in p -i-j CO cu • • E .-1 J= cn t-H CU cu -M T-H CM x: E +J cu CIS 4cu cu s2: iia 3 3 r3 E CO E ^3 C7) cn ro (0 r-•r— • 1 — cu CM u. Ll_ 00 E ro •ri>, 00 r/i x: cu -f-> 1— (O ra +-> CO tJ cu cu cu +-> -o -a Mx: Q. ro cu cu h(O E CJ rs 3 E tD a 3 • CO C/l sScu Q. Q. cu +J cu cu x: cu q; h'Z. ro X! CJ tovM^i^r^it'-

PAGE 76

CHAPTER IV DISCUSSION In this chapter the implications of the observations presented in Chapter III are considered. The simplest calculations from these data involve assumptions about the excitation temperature and optical depth. The portion of the dissertation based on such assumptions is found in this chapter. NGC 185 This dwarf elliptical galaxy is usually considered to be a companion of M31. It is located about 6 to the north of M31 and the difference in radial velocities is about 55 km/sec. For a discussion of the virial mass of the M31 system and comments on membership in that system see Rood (1979). A distance of 0.69 Mpc is assumed for j NGC 185 throughout the following discussion. At this distance 1" = 1.04 X 10-^^ cm = 3.3 pc. The classification given by de Vaucouleurs et al (1976) is E3p. A detailed photographic study of the system by Hodge (1963) shows clearly that the ellipticity (defined by e = 1 -) varies from 0.18 a near the nucleus to 0.26 for the outer contours at 350" from the nucleus. An increase in ellipticity with radius is commonly observed for elliptical galaxies. The E3 classification (which is based on the outer contours only) is confirmed by Hodge's photometry. -68I 1 A/, ., inl9

PAGE 77

-69The system is peculiar because of the presence of two dark dust patches within about 20 arcseconds of the nucleus. Figure 22 is an optical photograph of NGC 185 and a schematic of the dust regions is reproduced from Hodge (1963) in Figure 23. Careful comparison of the two figures reveals the extent of the dust in the optical photograph. Table 4 tabulates the positions of the patches used for observation and the areas in square arcseconds and pc^ as measured from Figure 23. Note that dust patch I (DP-I) has 2 entries, la and I. DP-Ia is the dark core of the northwest patch and DP-I is the larger, less dense patch (but also including DP-Ia). Table 4 \ Positions, Areas, and Telescope Filling Factors for Dust Patches in NGC 185 Object a(1950.0) 6(1950.0) Area PC^ Filling 4803'54" (n)2 35 Factor DP-Ia 0^36'^10!3 380 0.0072 DP-I Same as DP-Ia 115 1250 0.0235 DP-II o'^36'^12!2 4803'28" 350 3800 0.0716 DP-I+II (Center of NGC 185) o'^36'"ll?4 48=03 '44" 465 5050 0.0951 The CO observations were taken at the position of DP-I in December 1977 and July 1978. The December 1979 observations were made at both DP-I and -II as well as the center of the galaxy. However, as the FWHM for the 11 m telescope indicates in Figure 22, observations of one patch do not entirely exclude the other. Further, the pointing for

PAGE 78

-70•1.• • HPBW *' 30" Figure 22. NGC 185 showing the dust patches. North is up and west is to the right. The half-power beam-width is indicated. This photograph is reproduced through the courtesy of Lick Observatory.

PAGE 79

-71Figure 23. A schematic of the dust regions near the nucleus of NGC 185 (Hodge 1973). DP-I is northwest of the center, DP-II is southeast and DP-Ia is darker region within DP-I.

PAGE 80

-72December 1979 is unreliable to perhaps 20 arcseconds and at times 30-40 arcseconds. Consequently we have no confidence in the apparent identification of the CO source as DP-I based on the absolute pointing. Because of the uncertainty of the source for the CO emission, the following analysis considers three cases for the calculations of column and particle densities. The individual cases and the assumed CO sources are Case A DP-Ia Case B DP-I Case C DP-I+II To estimate the kinetic temperature of the molecule we can assume the clouds are optically thick (t 1). Observations of CO sources within our galaxy indicate that this is often the case with the J = 1 transition (Zuckerman and Palmer 1974). We find (from Appendix II) that ^A = Vex (1) where n^ is the forward beam coupling efficiency (also known as the filling factor or the beam dilution) from Ulich and Haas (1976). Since we have no detailed information on the brightness distribution we will assume that n^ = 0.7 (source area/beam area). The beam area is found from the 65" half-power beamwidth and the factor of 0.7 arises because 30% of the power enters the antenna through the very broad error pattern. Given these assumptions Table 5 gives the expected antenna temperature for the three cases and several excitation temperatures. Clearly we cannot distinguish between the several possibilities in Table 5 which are consistent with the detected temperature of 0.081 K. But if the CO is optically thick it is unlikely to be above 25 K or to originate in both the clouds.

PAGE 81

73Table 5 Expected Antenna Temperatures (K) for t 1 Source ^ex 10 K 25 K 100 K 10^ K Case A 0.071 K 0.179 0.714 7.14 Case B 0.235 0.588 2.35 23.5 Case C 0.951 2.38 9.51 95.1 Another approach is to compare the dust clouds observed in NGC 185 to clouds more easily observed in our own galaxy. The disadvantage is that it is not certain that the nature of the clouds will be the same. It does, however, provide a reference point for their study. The extinction in the dust clouds was measured by Hodge (1963). He finds DP-I to have a mean visual absorption of 0.3 mag but as high as 1 mag in certain parts. The mean for DPII is measured as 0.15 mag of absorption in the visual. At this level of absorption the clouds are similar to the diffuse clouds studied by Knapp and Jura (1976). Their study involved clouds situated in front of stellar sources that show color excesses of E(B V) ^ 0.3 mag. Assuming a normal A /E(B V) = 3 indicates Ay 'V 1 mag, similar to the absorption in the NGC 185 clouds. 1? Knapp and Jura suspect optically thin CO emission based on their 13 inability to observe CO at an intensity of 2-6 times less than the 12 1? CO (usually possible if CO is optically thick), and they found antenna temperatures of 1-2 K even though the sources probably filled the beam and had kinetic temperatures of 20 K or more. Using the thin

PAGE 82

-74cloud models of Lucas (1974) and assuming a kinetic temperature of 20 K (Morton 1975) they find that the column density is given by N(CO) ^ 5.0 X 10^^(T*/nf)AV (2) where Av is the full -width half-maximum of the line in km/sec. Using this expression and taking into account beam dilution (n^) we find column densities for NGC 185 as listed in Table 6. Table 6 Column density of -^^CO for Optically Thin, Diffuse Clouds Source ^^^9^ N(CO) x 2 x 10^ = n{ti^) cm 'cm -2 Case A 8.1 x 10^^ 1.6 x 10^^ Case B 2.5 x 10^^ 5.0 x 10^ Case C 6.1 x 10"^^ 1.2 x 10^ The mass of the molecular cloud can be found by multiplying the column density of H^ first by the area of the cloud and then by the mass per particle. The result is 9680 M^ and is independent of which case is actually correct. It is independent because we are assuming that the source is optically thin and we are detecting all the CO within a given region. Further, it represents a firm lower limit if the colliding particles are H2; if they are HI the minimum mass is 4840 M The last column of Table 6, total column density of H2, is found by assuming N(C0)/N(H2) ^ ^ 10"^(Martin and Barrett 1978).

PAGE 83

75A further finding by Knapp and Jura (1976) is that a necessary, but not sufficient, condition for CO in emission is a particle density 3 -3 ^10 cm This agrees with the theoretical result noted in Appendix II. We thus feel confident that the CO emission in NGC 185 is originating in a region at least as dense as 10"^ cm""^. It would be useful for calculations of the space density of various particles if the physical size of the absorbing regions were known. This would help constrain the various possible sources of the CO emission and help in determinations of the mass of gas involved. An estimate of the cloud volume can be made by assuming they have a dimension along the line-of-sight equal to the average width. This approach will tend to underestimate the volumes unless the clouds are flat and oriented broadside to the line-of-sight. Table 7 tabulates the result of this calculation. The gas mass listed in the last column is found by assuming the entire cloud is filled with molecular hydrogen at a density of 10^ cm"^. Table 7 Estimates of Cloud Volumes for NGC 185 Dust Clouds Dust Cloud Depth of (pc) Cloud Vol (cm3) ume (pc3) Mass of Gas (M^) DP-Ia 6.6 7.37 X 10^^ 2.5 X 10^ 6.2 X 10^ DP-I 16.5 6.06 X 10^^ 2.1 X 10^ 5.1 X 10^ DP1 1 42.9 4.79 X 10^ 1.6 X 10^ 4.0 X 10^ Hodge (1963) has estimated the mass of dust in DP-I and DP-II. Using the dust masses and the total gas content for the optically thin

PAGE 84

76case (Table 6) and the minimum calculated gas mass from Table 7 allows a calculation of the gas-to-dust (G/D) ratio. The results appear in Table 8. Table 8 Gas-to-Dust Rat ios Cloud Dust Mass Gas Mass T < 1 (M^) G/D Gas Mass T 1 (Mg) G/D DP1 25 9680 390 5.1 X 10^ 2 X 10^ DP-I+II 215 9680 45 4.5 X 10^ 2 X 10^ Considering that the gas-to-dust ratio in our own galaxy is usually quoted as 200, the most attractive source for the CO emission appears to be DPI. Using the minimum mass consistent for DP-Ia (Table 6) allows a calculation of the Jeans.: length for gravitational instability 3 ^J = [45^]' = 1.6[-^]^ (3) where M„ is the mass of the cloud in solar masses. Taking the width of the cloud as a characteristic dimension, R = 6.6 pc, T,, = T •^ K ex 20 K, and M = 4.84 x 10^ M^ X. = 1.1 pc Since the cloud is substantially larger than the Jeans length, it is most likely in a state of collapse.

PAGE 85

-77The Jeans mass can be calculated under these conditions from (Vy) 3/2 Mj = f Po ^J ^ 10^3 -L__ gn, (4) P with T = 20 K, y = 2, p = 3.34 x 10"^-^ g/cm^ (corresponding to 10"^ cm~^ of H2) we find Mj = 27 M^. If the doud remains isothermal as it collapses, and this appears likely because of the effective cooling, Mj decreases and fragmentation is likely to occur. Another calculation that confirms the collapsing nature of the cloud in NGC 185 is provided by Rowan-Robinson (1979). He estimates the cloud density, in an essentially virial calculation, by the equation n(H2) = 10^-^ n (av)^ r'^ ^^ (5) where n is 0.8 for a cloud in equilibrium, r is the radius of the cloud 18 in units of 10 cm, and K^^ is a dimensionless quantity related to a hot core for the cloud (it is always greater than 1.5). The density calculated for the NGC 185 cloud, assuming 5.6 pc for r (the characteristic width for DP-Ia) is 3.3 x 10^ cm""^. If the cloud is in free-fall n = 0.2 and the derived density is 8.3 x 10 cm""^. Thus considering that the NGC 185 cloud masses are uncomfortably large (Table 7 ) for higher particle densities, we consider the lower density more likely. From these arguments it appears that either DP-Ia or DPI is most likely in a state of collapse. Further, the lack of emission from ionized gas (Humason et al 1956) implies one of two possibilities. The first is that star-formation ceased in the system long ago, such

PAGE 86

-78that the HII regions associated with massi.ve stars have since dissipated. Considering that typical lifetimes for HII regions are 10 years and Hodge (1973) estimates the age of the OB complex in NGC 205 {very similar to that of NGC 185) to be 5 x 10 years, it is certainly plausible that the early HII regions have evaporated. Since the time for the observed clouds to collapse to stars is about 10 years, it seems that the system may be experiencing bursts of star formation similar to that envisioned by van den Bergh (1975). However, there is a serious flaw in this scenario. Hodge (1963) estimates that the population I stellar component of NGC 185 is about 5 2 x 10 Mg. Using Faber and Gallagher's (1976) mass loss rate of 0.015 H^ yr"-^ (lO^L^)"^ and the luminosity of NGC 185 (2.67 x 10^ L ), we find that it took 5 x 10 yrs for the material to collect. During the 5 X 10 years since the last star-forming episode only 2 x 10 M_ of gas should have been able to appear. Our CO observations show at 3 least 5 X 10 M^ (for optically thin CO, assuming atomic hydrogen for the colliding particles) and it is likely that there is actually more. An even stronger case is made if the total mass of dust (215 M ) is multiplied by an assumed gas-to-dust ratio of 200. The system is seen to contain 4.3 x 10 M^ of gas, more than 10 times the amount from stellar mass loss. The second possibility that may be occurring in NGC 185 is continual star formation but with no or B stars, and consequently no HII regions. It is very difficult to prove that this is the process taking place. If indeed low-mass stars are being formed they will be too faint to see and deductive reasoning is necessary to support this

PAGE 87

-79possibility. Specifically, since the regions observed are dense enough to form stars yet no HII regions are observed, either no stars have formed yet (leading to the burst hypothesis) or only low-mass stars are forming. The idea of a skewed mass distribution for stars is explored in some details by Mezger and Smith (1977). They find that clusters of massive stars (which produce giant HII regions) are found only along spiral arms. It is probable that the spiral shock front of a density wave triggers their formation. It appears though, that the low-mass stars can form out of small but dense clouds (Herbig 1970) that are far more widespread than large clouds. Jura (1977) also proposes that the expected radiation environment of elliptical galaxies (low UV flux) will lead to low-mass star formation. The reasoning is that the low flux results in less heating of the clouds. A cooler cloud will have a lower critical mass which can separate out and collapse to form a star. The dynamics of the nuclear region will be considered in the following paragraphs. The difficulties in measuring radial velocities for earlytype galaxies are immense. There are usually no HII regions in these galaxies so their sharp emission lines are not available for measurement. The absorption lines from the stars are intrinsically wide and further broadened by the stellar velocity dispersion, resulting in large errors. The radial velocity data for NGC 185 are actually better than that for most early-type galaxies. Ford et al (1973) identified four planetary nebulae in the system and later Ford et al. (1977) succeeded in measuring the radial velocities of two of these objects, NGC 185-1 and

PAGE 88

-80-2 in their notation. These measurements, coupled with the radial velocity of DP-I, can be used to calculate a rough mass interior to the object and, when combined with the isophotal contours of Hodge (1963), the mass-to-luminosity ratio (MLR) can be computed. Table 9 Positions and Radial Velocities of Points Within NGC 185 Object Radial Velocity (km/sec) Distance Minor Ax (pc) from is Distance from Major Axis (pc) 185-1 -212.5 144.2 75.5 185-2 -200 107.3 103.6 DP-I -175 43.9 Table 9 lists the three radial velocities determined for the system, as well as the projected distances from the major and minor axes. We make the simple assumption that the ellipticity is caused by rotation about the minor axis, thus the figure of the system is an oblate spheroid. We further assume that the radial velocities are the projected results of purely circular velocities about the nucleus. We can then make a straightforward calculation of the mass interior to the measured point. Implicit in this approach is the assumption that the points lie in the plane of the galaxy. Recalling the work of de Vaucouleurs (1977), described in Chapter I, on the relative frequencies of various ellipticities, we can make an informed guess of the true ellipticity of NGC 185. The galaxy is most likely to be an E3.6, but we can be more flexible and also consider

PAGE 89

-81the possibility that it is an E5.5 (the flattest elliptical according to de Vaucouleurs 1977). Of course the system may be an E3 seen face-on, but the velocity gradient for the three measured points strongly suggests rotation. The details of the determination of the inclination angle (i, defined to be the angle between the plane of the galaxy and the plane of the sky) and its effects on other parameters are found in Appendix III. Under the assumption that the system as is an E3.6 or an E5.5 the inclination angle is 53.1 or 68.0, respectively. The azimuthal angle, e', can be deprojected to find the true azimuthal angle, e. The results are contained in Table 10 for the two planetary nebulae. Since DPI lies on the minor axis we assume it has no radial velocity due to rotation; consequently we take its radial velocity, -175 km/sec, as the systemic velocity for the galaxy. Table 10 Values for the Observed and True Azimuthal Angles Object e' 9 (degrees) (degrees) i = 53.1 i = 68.0' 185-1 26.8 40.1 53.4 185-2 44.0 58.1 68.8 The true circular velocity for the nebulae can thus be calculated from ^c v^/cos e sin i (6)

PAGE 90

-82The true radius can also be computed and is listed with the velocities in Table 11. Finally, the mass interior to the point, assuming Keplerian motion, is given by M = 2.36 X 10^ v^ r (7) where M is in solar masses, v is the circular velocity in km/sec, and r is the true radius in parsecs. Table 11 Circular Velocities and Radial Distances in NGC 185 Object km/sec V (km/sec) r (pc) i = 53.1 i = 68.0 i = 53.1 i = 68.0 185-1 185-2 37.5 25 61.3 67.8 59.2 74.6 191.4 203.2 248.0 296.6 The isophotal contours of Hodge (1963) were measured to find the luminosity interior to a given point. The contours measured were the innermost (1) and the next brightest (2), corresponding to 20.47 and 2 21.22 mag/arcsec respectively. Using a distance modulus of 23.9 and assuming absorption within our own galaxy according to the cosecant law (Ay = 0.26 CSC b = 1.0 mag for NGC 185) allows the calculation of the luminosity interior to the first and second contours, listed in Table 12.

PAGE 91

-salable 12 Luminosities in the Nucleus of NGC 185 Contour Brightness^ (mag/arcsec ) Area „ (arcsec ) Luminosity 1 2 20.47 21.22 1690 7715' 1.03 X 10 7 2.35 X 10 7a Excludes contribution interior to contour 1. i The mass calculated from equation 7 is then divided by the appropriate luminosity from Table 12 to derive the MLR interior to each object. The Planetary nebulae, 185-1, lies on the second contour, making the luminosity interior to it the sum of the luminosities inside contours 1 and 2. The other nebula, 185-2, lies midway between contours 1 and 2 so we use the luminosity inside 1 plus half the luminosity inside 2. Table 13 lists the results for the MLR derived with these values. Object 185-1 185-2 Table 13 Mass-to-Luminosity Ratios for the Nucleus of NGC 185 Mass Interior (M ) i = 53.1 1.7 X 10 1.7 X 10' 8 8 i = 68.0^ 2.7 X 10 3.9 X 10 8 8 Luminosity MLR (solar units) (L,) 3.38 X 10' 2.21 X lO' i 53. r 5.0 7.7 i = 68.0 8.0 17.6

PAGE 92

-84While these values for the MLR are very reasonable for early-type galaxies in general (Faber and Gallagher 1979), there is a significant disagreement with the MLR of 1.8 used by Ford at al. (1977) for NGC 185 and NGC 147. This MLR is ostensibly derived from observations of M32. We feel that the most likely cause for the discrepancy is that Ford et al. (1977) do not consider possible rotation in their analysis. The radial velocity of DP-I is most likely nearest the systemic velocity of NGC 185 since it is fortuitously on the minor axis. We find that the MLR of the nuclear regions of NGC 185 is between 5 and 18. Considering the non-linear effects of the inclincation and azimuthal angle, the most likely value for the MLR is around 8. Assuming that it is independent of radius, the total mass of NGC 185 is found to be 1.3 X 10^ M NGC 205 Another dwarf elliptical companion of M31, NGC 105, was also observed to sensitive limits for CO emission, with negative results. Figures 8 and 9 show the averaged spectra from the July 1978 and December 1979 sessions, respectively. The spectra are taken at two different positions within the galaxy (dust patches 11 and 12 in Hodge's (1973) notation). Hodge (1973) describes and diagrams the dust content of NGC 205. The distribution is reminiscent of NGC 185 but with a larger number of discrete clouds over a wider region in the nucleus. The July 1978 observations

PAGE 93

-85were taken on dust patch number 11 while the December 1979 observations were of dust patch number 12 on the other side of the nucleus. Table 14 lists the positions observed for each patch as well as Hodge's (1973) estimates of the dust content. The dimensions listed by Hodge are used to find a rough area in square arcseconds and pc (the assumed distance is 0.69 Mpc). Hodge (1973) also presents microphotometer tracings of these dust regions indicating a visual absorption of about 0.2-0.3 magnitudes. It is clear that these regions are not as dense as DP-Ia in NGC 185. Beyond this difference the two galaxies are very similar in dust and young star content. Table 14 Dust Regions in NGC 205 Region no. a(1950.0) 6(1950.0) Area of Cloud Dust Mass (")2 pc2 M^ 11 Q^37"^40!7 4125'33" 350 3.9 X 10-^ 160 12 37 41.3 41 24 05 880 9.6 X 10^ 460 A similar calculation as was made for the NGC 185 dust clouds, under the assumption of optically thin emission (see equation (2) of this chapter), can be performed. The antenna temperature used is 3 times the standard deviation for each cloud. The velocity width is taken to be four times the channel width (av =10.4 km/sec). The results appear in Table 15. The column densities are modest when compared to those of NGC 185 (Table 6). The computed total masses are even less than one may expect from the dust masses and a gas-to-dust ratio of 200,

PAGE 94

-863.2 X 10^ and 9.2 x 10^ M^^ for clouds 11 and 12, respectively. Table 15 Maximum Column Densities of CO and Hp for NGC 205, x < 1 Cloud no. N(CO) N(H2) M (molecular) (cm-2) [rm-2\ M (cm-^) (cm-^) M_ 11 7.0 X 10^^ 1.4 X 10^ 4.4 X 10^ 12 2.8 X 10^^ 5.6 X 10^^ 5.2 x 10^ The situation does not change if one assumes the clouds are optically thick. The NGC 205 clouds are relatively large compared to those in NGC 185, and any reasonable kinetic temperature will produce a large antenna temperature, as shown in Table 5, very easily. The question is then, why is NGC 205 not detected while NGC 185 with similar morphology (and actually less dust) is detected at the 1 9 frequency of CO? We feel that the key difference in the two systems is that the NGC 205 clouds are less dense than those in NGC 185. Hodge's (1963 and 1973) microphotometer tracings show this, and even optical photographs of the two galaxies show the NGC 185 dust patches to be more prominent. As Knapp and Jura (1976) found, even a density of 10'^ cm""^ is not sufficient to ensure detection of ^^CO in emission. Further, their observations were made on nearby clouds where one could reasonably expect a filled beam. The NGC 205 observations involve substantial beam dilution which considerably worsens prospects for detection.

PAGE 95

-87While many of the arguments about continuing star formation in NGC 205 can be advanced with much the same reasoning as those for NGC 185 in the previous section, much of this support is lost because we cannot be sure the NGC 205 clouds are in a state of collapse. However, it is very likely that they will eventually form stars. This is clear because NGC 205 has an even larger young stellar component than NGC 185; Hodge (1973) calculates some 2 x 10 M_ for the young stars. We seem to be seeing an earlier stage of the process in an elliptical galaxy which converts the ISM into stars. In a few times 10 years the dust clouds in NGC 205 will be much more condensed and CO emission should be detectable. In the future stars will be forming and, if these are high-mass stars, the galaxy will retain its unusual population I component. Negative CO Results IP The galaxies that were not detected at the CO J = 1^0 transition are listed in Table 2. Several attributes are also listed, along with the limits of our observations. These limits can be interpreted on the basis of two assumptions about the CO. First is that the gas is optically thin and the emission was not detected because of beam dilution or an exceedingly small optical depth (t 1). Second, if the line is optically thick then beam dilution is the only mechanism to lower the antenna temperature below our detection limits. Another variable in the interpretation of these negative results is the width of the undetected CO line. If it is several hundred km/sec broad, representing the contributions of many small clouds within the

PAGE 96

-88beam, it is more difficult to detect than if it were a narrow feature arising from only a single, larger cloud. If T >> 1 no information can be gained about the amount of molecular material stored. Essentially any quantity of material can be stored in clumped, optically thick clouds. The material remains undetected because of the very small filling factor (n^). Consequently Table 16 lists various parameters under the assumption that the CO clouds are optically thin. Knapp and Jura's (1976) finding that the ^^CO line is probably thin if the visual absorption is less than 1 magnitude indicates that most CO in elliptical galaxies will be thin since deep absorption features are rare. Equation (2) from this chapter is used along with the assumption of n^ = 0.05. The antenna temperature is taken to be 3 times the standard deviation listed in Table 2. The width of the undetected feature is taken to be 10 km/sec for one set of calculations and 200 km/sec for a second set. The first width is appropriate for a single cloud while the second is more typical of an expected global profile for many clouds (Rickard et al 1977). Also listed is the column density of H2 assuming N(C0)/N(H2) = 5 X 10 (Martin and Barrett 1978). The molecular mass is calculated by (Schneps et al 1978) M = ^(Tgx)(T^/nf)AvD2 (8) where M is the molecular mass in solar masses, f2(T ) incorporates the effect of temperature on the population levels of the CO, T* is taken to be 3 times the standard deviation, av is the assumed velocity width

PAGE 97

-89-a OJ 00 CO CO CO CO CO CO +j CO CO CO o o o r--. o o o o E O'— o O o 1-H 1— 1 -H o I-H 1—4 1—4 1—4 l-HQJ (S r-H r-l I— I X X X 1—4 X X X X CD nro-s: X X X CM CM CM X CO . CM r-v CO o CM CO CO 1—4 r~-. c: r--v UJ • • • • • • • • • OJ r--v CM CM =JCM 1—4 CM 1—4 1—4 CO CO =33 cn cn 1—4 o Z3 QJ fC -Q O •a 03 Ho 4-> cn cn CO CO CO CO Q • • O) CO I— 1 cn CM CO cn •rsLD cn CD CO r-- 03 CO r-~CO CO r-^ r™ OJ i^ O o O o o O o o o o o 0) •-— CU CO — .— 1 1 1—1 1—1 1—1 1—1 1—4 1—4 1—4 1—4 1—4 c -Q E i/l O O •1— r~^ +-> fO CM 2: X X X X X X X X X X X +J r^ .c :s N — M cn CJ o h=dun 00 o CM o ^ >=dr--. CD CM o 1—4 3 a> • • • • • • • • • • cn — ZH to CO CM r-l 1—1 1-H r-^ CM r-v. CO 1—4 00 0) <+o "e — ^ p. .^ CM 4-) 1 fd o E > o (J i. CM ,— ^ CM =dCM CM =( CM O o o OJ CM O • • • • • CO CM • CO II in CM cn r-. CO cn C~i cn OJ CM CM O CM -Q -— o en CO cn cn CO CO CD CO 1—4 CM cn o > <) z: .—1 o o r— 1 E ^— ^,-.,.^ o V CM CO S1 r^ MH E u 1—4 e> -o .'—^ CO r-~. CM CD CM CO O o o o CO OJ O CO • • • • • • 1—4 • r-^ to > O r-H en CO CO cn o vf1—4 1 — 4 ^ O CD cn t— 1 -I— --O CM I-H ':3CO =1CO CO 1—4 CM 1 „ r— 4 S^ r-f '.a cu dj •. 1— Q .f_> J3 m 10 ^ CO "=dCO CO r^ cn CO CO r-CO CD Its f0) ^ O o o O o o o o o o O J_ t/1 C/l '— ^ i-H r-l 1— ( v—t 1— 1 1— 1 1—1 1—4 1 — i 1—4 1—4 CD 4-) CO c/l O s: CD to ro CM s: X X X X X X X X X X X cn s: s: II •--' CO Q* 1— CM <* o o CO o r-^ OJ CO CO 1—4 D. fO O .— ( • cn • cn CO 1—4 CO CO CO CM ^ (O to p~* CO CD :^ 3 """^^ ^~^ tJ -a O E OJ cu j:^ 1 r— E O O u s: r-H ^-*. 1—4 CM O o CD CO o o cn 1—4 1—4 >=}CD E II a: CM • • • • • • • o 3 --o CO 1—1 =* cn CO iz .-1 "r< s.^^^ X (0 # s: r-H V H CM E o CO -"-^ CO CO 1—1 CO 1—) CO cn cn o "^ o O • • • • -— O r-H o CM CXJ 1—1 CM 1—4 CM CO cn 1—4 ^ ^ s: T-i CO at t/i CO 00 1—1 r^ CD CO cn 1—4 r^ CD CO "^ — ^ E — CM r-^ o O OJ 00 CM r--. CO CM o O SNi t-H o CM CM 1—1 1—4 1—1 1—4 CM ^ 1—4 CO X •-— • • • • • • • cn to oo o CD CM o o o o o o o o o 1—4 +-> QJ =te r— 1 r-H M1 1 CM cn CD o 00 CO CO CD s c_> cn cn "53in CO CM cn r-^ CO ^ CO c CD CD o o o o CO CM 1—4 C^J CD 00 CO rD Di ^ CM CM >=11—1 CM CO ^ <^ dLO cn (0 XI

PAGE 98

-90in km/sec, and D is the distance to the source in Kpc. For """ex *^ ^"^"^^ ^ ^^^ vibrational levels of CO are not populated and ex' ex Also listed in Table 16 is the mass of HI that has been detected in the galaxy. The last column gives the expected HI content from Knapp et al (1978c) M^j = 4.4 x 10^ (Lpg/lO^ L^) (9) where M^j is in solar masses and L is the photographic luminosity of the galaxy. This relation uses the mass loss rate given by Faber and Gallagher (1976) of 0.015 yr~^ (10^ L^)"-^ and assumes a time for g accumulation of 4 x 10 yrs, the time over which we are sure that normal elliptical galaxies have existed (Gunn and Oke 1975). A decrease of 25% is made to allow for an undetectable helium contribution. Caution must be exercised in comparing the expected HI content with the maximum molecular content as derived here. First they are correct only if the CO is optically thin. While there is reason to believe this is the case, it is not certain. For an optically thick cloud only the surface is seen and no information is available about the total mass. Second, the CO observations are made with a 65" HPBW while the galaxies are typically several times this size. The CO-implied estimates of the total gas are thus only true for the central regions, whereas the expected HI is for the entire galaxy. Indeed, in the cases of the detected HI the beam is usually about the size of the galaxy. This discrepancy in the size of the regions sampled may not be an important consideration. This is due to the difficulty of finding

PAGE 99

91a mechanism that would allow the gas to maintain the same spatial distribution as the stars that shed it. As long as the gas can cool itself effectively, usually via line radiation, it must collapse into the potential well of the galaxy. If it cannot cool itself, a galactic wind sets up and the system is swept clean. The point has been made previously that galactic winds are not universally effective, consequently one expects the interstellar material to preferentially collect in the nucleus. Comparing the maximum reasonable amount of molecular material that could escape detection under these assumptions (t < 1, Av = 200 km/sec) with the expected HI content reveals that molecular storage is not a dominating feature of the galaxy's ISM. The largest discrepancy, for NGC 205, may not be significant because the region sampled for CO is small with respect to the dust patches. For the other galaxies the discrepancy may be due to one^ or a combination, of three processes: the gas may be clumped and thus optically thick, the material could have been removed either by a galactic wind or ram-jet stripping, or finally, star formation could be consuming the gas with an IMF deficient in high-mass, bright stars. A fourth possibility, that the clouds are similar to Knapp and Jura's (1976) thin clouds with n -x. 10 cm" but with no CO emission, is unlikely. The CO observations included a large portion of the galaxy's nucleus and it seems highly improbable that all the clouds will be in this nether region of collapse with no CO emission. Certainly some fraction may be without CO emission but others, further collapsed, should produce CO emission.

PAGE 100

-92It is interesting to note that the three galaxies with a detected HI content larger than expected (NGC 1052, 2685, and 4278) are all suspected of accretion. All three systems show peculiar dynamics for the HI that make an extra-galactic source the most plausible explanation (Reif et al 1978, Shane 1980, Knapp et al 1978c).

PAGE 101

93Neutral Hydrogen and NGC 185 As described in Chapter III, we have detected neutral hydrogen at the position of NGC 185. There is also detectable HI at least as far as 2 north of the galaxy. It is apparent that any HI physically associated with NGC 185 is being confused with high-velocity hydrogen belonging to our own galaxy. The mass of HI in a particular feature is given by (Wright 1974) M^j = 2.356 X 10^ D^ / S dV (10) where the mass is in solar masses, D is in Mpc, S is the flux density in Janskys (1 Oansky = 10"^^ Wm"^ Hz"^), and dV is in km/sec. This equation is based on the assumption of low optical depth in the 21 cm line. While this assumption may not be valid in the plane of the galaxy, it is quite likely true for HI out of the plane (NGC 185 is at a galactic latitude of -15). The confusion with high-velocity local HI is primarily due to the Magellanic Stream, an arc of HI clouds stretching from the Magellanic clouds through the south galactic pole and terminating in the vicinity of NGC 185 (Mathewson et al 1974). That there are indeed discrete clouds in the velocity range around -200 km/sec and within 4-6 of NGC 185 is confirmed by Giovanelli (1979) and Hulsbosch (1980). Hulsbosch's observations of the area show nothing at either the position of the galaxy or within 2 to the north. This lack of confirmation of our findings is not surprising because Hulsbosch's sensitivity is only about 400 mJy while ours is about 25 to 40 mJy (the 43 m telescope has a sensitivity of about 4 Jy/K)".

PAGE 102

-94Based on our observations it is clear that there is HI around NGC 185 with a distribution greater than 24'. Figure 19 shows a feature with about 15 mK of HI when the "off" is located many beamwidths to the west. When a spectrum is taken with the "offs" only 24' or 48' to the west, east, south, and north, the strength of the feature drops to around 5 mK. This clearly indicates that we are subtracting out about 10 mK of broadly distributed HI. Analysis of the individual spectra with an "off" in one of the four cardinal directions shows that there is an excess of HI with respect to all directions except perhaps the north. The intriguing possibility is that while there is clearly a broadly distributed HI components around NGC 185, the spectra show that there is excess line radiation from the position of the galaxy when compared with areas 1 to 2 beamwidths away. Of course it is possible that an enhancement in the high-velocity cloud coincidental ly projects onto NGC 185. Our observations are not sensitive enough nor spatially resolved enough to answer the question definitively. It Is interesting to note that if the 5 mK excess does originate in NGC 185, the mass associated with it is about 9 x 10 M This is quite reasonable since the minimum molecular mass needed to generate the observed CO signal is 10 M^^. Indeed, a calculation of the mass of HI expected in the galaxy from normal stellar evolution (see equation 7, in the previous section) is 1.2 x 10 M A final possibility suggested by the detection of a plume between r and 2 further north of the galaxy is that we are seeing a tail or streamer of HI associated with NGC 185. Since the galaxy is a

PAGE 103

-95companion of M31 an interaction of this sort may be occurring. The difficulty is that M31 is to the south of NGC 185 and is over 6 away, a rather larger distance for an interaction considering the small mass of NGC 185. Thus while the intergalactic plume is an intriguing possibility, it is not well -supported by our observations. Higher spatial resolution measurements are required in the area centered on the galaxy and at least 20-30' around. It is necessary to examine the local features well enough to allow their subtraction from possible HI belonging to NGC 185. Any HI associated with the galaxy would probably have a distribution less than 5' across; the galaxy itself has a major diameter of around 10'. The VLA can easily achieve this resolution, and an investigation of the NGC 185 region should be undertaken to determine the kinematics of the HI at the position of the galaxy.

PAGE 104

CHAPTER V SUMMARY The current understanding of elliptical galaxy formation and evolution has several important shortcomings. It is not understood how elliptical, or for that matter any, galaxies separate out from the background and collapse. Even the general process of whether the clusters of galaxies form before the individual galaxies is not clear. The deceptively relaxed-looking stellar distributions are apparently not relaxed at all. It is quite possible that the 3-dimensional figure of elliptical galaxies is actually a triaxial ellipsoid. Further inconsistencies are apparent in the disposition of gas shed by the stars within the galaxy. If it recycled into stars with a normal IMF the nuclei of all early-type systems should be bluer than observed. Recent theoretical work has shown that two gas removal mechanisms may be present in these systems. One is a galactic wind in which supernovae provide an energy source to heat the ISM of the galaxy to a temperature high enough to evaporate from the system. The crucial points of this method are the assumed Type I supernova, rate and the efficiency of energy coupling between the expanding supernova shells and the general ISM. The other mechanism is ram-jet stripping. The process is basically a hydrodynamic interaction between an assumed intracluster medium and the ISM of the galaxy. There is circumstantial evidence for this -96-

PAGE 105

-97process; early-type galaxies are preferentially found in rich clusters that would be most effective in retaining an intracluster medium. However, it is not clear that the presence of earlytype systems in rich clusters is not a result of initial conditions at the time of the cluster formation. Indeed recent work has tended to support this view. Observations of nearby galaxies do not fit well with either of the two mechanisms. Some have been detected in HI when the galaxy should be able to expel its ISM. Worse, in at least two cases the HI has a broad distribution, entirely contrary to expectations for a partially operating galactic wind or ram-jet stripping. The suggestion has been made that the HI has recently been accreted, but this is difficult to support based on the space density of HI clouds large enough to account for the material. Other early-type galaxies show patches of obscuring matter firmly indicating that mass removal processes are not completely efficient. 12 Observations of CO were undertaken for early-type galaxies to determine the role of star formation in the removal of the ISM. The CO molecule was chosen because it is widely distributed, resistant to dissociation, and has a transition frequency accessible to highly sensitive radio telescopes. 12 The J = 1^ transition of CO was detected in the dwarf elliptical galaxy NGC 185. The physics of the line formation process strongly imply that the emitting region is the north-west dust patch about 15 arcseconds from the nucleus. The fact that the line is seen at all indicates that the region is at least as dense as 10 cm The minimum mass consistent with the observations is about 10 M

PAGE 106

-98A region as large as this dust patch is gravitationally unstable and is collapsing, probably forming stars at a later epoch. The nucleus of NGC 185 contains several bright blue, presumably young stars which have formed in the relatively recent past. Observations of another dwarf elliptical galaxy, NGC 205, failed to find CO emission even though this galaxy contains a larger population of young stars and more total dust content distributed in 12 I patches. The densest of the regions are not as dense as the northwest dust patch of NGC 185. It is suggested that this is the primary reason for not detecting NGC 205 while obtaining a positive result on NGC 185. Negative results for 9 other early-type systems are presented and the implications are discussed. The major finding is that the material shed by the stars cannot be contained in optically thin clouds and still elude detection in these observations. Alternatives are that the gas is clumpy and thus optically thick, it has been swept out by either a galactic wind or ram-jet stripping, or it is being consumed by star formation with an IMF skewed towards low-mass stars. Theoretical support for a skewed IMF is presented, but the question cannot be answered by observations with present-day equipment. More sensitive CO data are needed and the interaction between the intracluster medium and the ISM needs to be explored further. Using radial velocities from two planetary nebulae in the system we find convincing evidence that the galaxy is rotating about its minor axis. The data can be further used to derive the mass interior to the measured points. Of course some basic assumptions about the true ellipticity of the galaxy and the effect of projection onto the plane

PAGE 107

99of the sky preclude firm conclusions about the dynamics of the system. We find, however, that the mass-to-luminosity ratio in the nucleus (the planetary nebulae are both within 1 arcminute of the nucleus) is between 5 and 18, with the most likely value around 8. This number is quite reasonable for an early-type system, but the difficulty in determining the ratio, and the lack of agreement from different methods, prohibit any conclusions beyond those already mentioned. Neutral hydrogen observations of the vicinity of NGC 185 are presented. They reveal the probable existence of highvelocity hydrogen at -180 km/sec at and near NGC 185. The observations also show an enhancement of about 5 mK antenna temperature at the position of the galaxy with respect to 1-2 beamwidths away. The data are inadequate to determine if the material is in NGC 185 or is just an enhancement in the local material coincidental ly superposed on NGC 185.

PAGE 108

APPENDIX I CALIBRATION THEORY FOR CO OBSERVATIONS The intent of the calibration procedure is to correct the data for several factors that distinguish a real telescope at the earth's surface from an ideal telescope at the top of the atmosphere. The major correction factors are ohmic losses in the telescope, spillover, blockage, and the radiation/attenuation of the atmosphere. The following paragraphs describe the calibration procedure used at the 11m NRAO telescope as described by Ulich and Haas (1976). The method of choice to calibrate mm-wavelength observations is to use a rotating chopper which alternately covers the feed horn with an ambient temperature microwave absorber. The method is attractive because it automatically compensates for changes in atmospheric absorption (Penzias and Burrus 1973). The ambient temperature absorber is placed over the feed horn aperture and the antenna temperature measured is Tload = V(^s'Tamb)-^S-^(^-'Tarnb) (D where G is the gain in the receiver, T is the ambient temperature of the absorber, the subscripts s and i indicate the signal and image sidebands, respectively, and ^'-^' exptKvkT) 1 (2) -100ii | ,4;w^t. '' : "'

PAGE 109

-101is the effective radiation temperature of a blackbody of temperature T at the frequency v. Note that the receiver amplifies not only the signal sideband but also the image sideband. The image sideband could be filtered out before the signal is amplified but this invariably involves the introduction of more noise at a crucial and sensitive point in the system. Since this work is concerned primarily with detection and requires the most sensitive arrangement possible, the image rejection filter was not used. Consequently the following equations include terms for the image sideband as well as the signal sideband. When the telescope is pointed at blank sky the antenna temperature is Tsky Gstn,J(v3 J3) + (1 n,) ^i^,J,^,U + similar terms for the image sideband (3) where n. is the telescope efficiency considering spillover, blockage, and ohmic losses. The T is the apparent brightness temperature from the spillover, blockage, and ohmic losses. The T and T. (in the image sideband terms) are the brightness temperatures of the sky at the signal and image frequencies, respectively. They are given by 0(v^,T3) J(v3,Tj^)[l exp(-T^A)] + J(v5,T^g)exp(-T^A) (4) and a similar image sideband equation. The T is the mean atmospheric temperature, T, is the brightness temperature of the cosmic background radiation, and t^ and t. are the atmospheric zenith optical depths at the signal and image frequencies, A is the air mass in the observed direction. -< ftf -f^ io T M i tflri i nr ai w i ia iiSiSc5itgaLaftr^-g

PAGE 110

-102The calibration signal is found by taking the difference between the load temperature and the antenna temperature looking at blank sky ^^cal "^load "^sky = '^sl^^(^''^amb) V^^'^s^ (1 ^o)^('^s'"'"sbr^-^ "*" ^'i'"''^^'" image sideband terms (5) When observing a spectral line source the antenna temperature is ^source = ^s^\^' ^f)^i-s'h^ ^ r,,r,,m-,,\)ll exp(-T)]exp(-T^A) + J(vg,T|^ )exp(-T)exp(-TgA) + J(^^s.T^)[l exp(-T^A)]} + (1 n^)J(v2,Tsbr)^ + image terms for the sky brightness temperature (6) and Tsbr where T^ is the excitation temperature of the molecule, x is the optical depth of the source, and n^ is the forward coupling efficiency given by 'f = 1/2^ P^(fi)df2 source ,, n. = (7) where P^ is the normalized antenna power pattern, B is the normalized apparent source brightness distribution, fi is the direction of peak antenna gain, ^ is the direction of maximum source brightness, and df2 and d^ are infinitesimal solid angles. Essentially, n^ is the filling factor for the source within the telescope power pattern. If one observed an extended, uniformly bright source n^ would be 1. Otherwise it is less than 1 and dilutes the antenna temperature as is apparent from equation (8).

PAGE 111

• 103The difference in antenna temperature, between the source and the sky is source source ~ sky = Ggn^nfexp(-T^A)(1 e"'')[J(v2,T^) J(v5,T^g)] (8) The corrected antenna temperature of the source is defined as cai where T is obviously c •' \ = ^T^al/t'^s\^^P(-^s^)^ (10) Consequently T* = n^d e~^)[J(v^,T^) JCv^.T^^g)] (11) Note that several parameters do not appear, namely the receiver gain (G ), atmospheric extinction exp(-T A), and antenna losses (nJ. These have been corrected for by an appropriate choice of T From equation ( g ) it can be seen that we need only the ratio of the source temperature to the calibration temperature in addition to T to determine the source antenna temperature. If we assume that the IF is small with respect to the LO frequency (a very safe assumption for this work) then J(v.j,T) ^ J(v ,T) ; that is, the signal and image temperatures are close enough to be considered equal then

PAGE 112

-104T,.(l + G./G^)[J(v^,y -J(v3,T,g)] + (1 + G./G^)[exp(x^A)][J(v^,T^^^) J(v^,T^)] + (G./G3){exp[(T^ T.)A] l}[J(v^,T^) J(v3,Tbg)] + (1 ^ S./63)[exp(x3A)/n,][J(v3,T^^^) H^^J^^)! (12) In general evaluation of this equation is difficult and an easier method is to use the definition of T [equation (10)]. Thus, only AT ,, G c ca I s n^, and t^ need to be measured as a function of elevation to determine T^ at a given frequency. Since the atmosphere changes little at high altitudes in the mm region, T^ is a constant function that varies slowly if at all with time. By comparing equation (10) with equation (12), T (the mean atmospheric temperature) can be calculated. This is perhaps the most difficult value of equation (12) to determine. The others can be measured or estimated with reasonable confidence. Calibration for this work was done using equation (12) with the following values adopted; ^ = G. = 1 T^ = 280 K \,-'-'^ Tsbr = 280 K Tamb = 290 K A = sec (zenith angle) n^ = 0.87 T. = 0.085 T = 0.35 s

PAGE 113

APPENDIX II THE PHYSICS OF CO SPECTRAL LINE CALCULATIONS In this appendix the formation of CO line emission is treated. This is followed by consideration of column density calculations and the assumptions involved therein. The one-dimensional time-independent equation of transfer is dl -T^= -K I + e (1) ds V V V where k is the volume absorption coefficient, e is the volume emission coefficient, and I is the specific intensity at frequency v. This can be integrated to give M^o) = MO)^ V +e ^ / B^e ^ ds (2) s where t (s) = / k ds is the optical depth integrated along the line of sight from an initial point (s = 0) to the observer's position (s ). This can be further simplified by assuming the source is a uniform homogeneous cloud and that k is due to an atomic or molecular transition. We now have -T (s„) e„ -T„(s„) The first term is the attenuation of the background radiation as it passes through the cloud and the second term is the emission -105-

PAGE 114

-105by atoms or molecules in the cloud, corrected for self -absorption. The quantity measured with a radio telescope is most closely related to the intensity at the transition frequency, [I (s )] minus the intensity near the transition frequency [I (s )]. Since I,(s„) I„(0) By convention radio astronomical measurements are usually made in temperature units and one can substitute the radiation temperature 2 2 [J(Tg) = c I /2kv ] in the preceding equation. The brightness temperature is found from o o hv/kTn K = ^(Tg) = 2hv-^/c^(e ^ 1)] (5) Obviously B^ is the Planck function and Tg is the temperature of the blackbody that subtends the same solid angle and emits the same specific intensity at v. If the RayleighJeans limit (hv/kTg 1) applies then J(Tg) -^ Tg. This is not always true for CO measurements. The equation for the excess line temperature is then (Zuckerman and Palmer 1974) AJ(Tl£) = J(Tl) J(T^) = [J(Tg^) J(T^)](1 e'"^) (6) where T|^ is the line brightness temperature, Tp is the brightness temperature of the continuum, and T is the excitation temperature of the molecule. Assuming local thermodynamic equilibrium (LTE) gives

PAGE 115

107... I ( a Boltzmann distribution of the rotationallevels which can be written n g -hv/kT Ji = _u e s^ (7) "1 9l ^ ^ q where n and n-, are the populations (cm ) in the upper and lower states, respectively, and g and g, are the statistical weights of the levels. Note that t^^(s) = / K^ds and (Lang 1974) _2 n. ^ K 2 ^^P (^)tl exp(iT^)]A (8) ^ 8ttv^ Av "^'ex "^ ex where Av is the full-width at ha If -maximum for the line, T is again ex ^ the excitation temperature and the Einstein coefficient (A) for the spontaneous electric dipole transition is A = 64lV, |2 ^gj 3hc-^ "^ where I |2 2 (J + 1) ,.., I^jl ^ (2J + 3) (1) for the J+1 -> J transition. Since y = 0.112 x 10" in cgs units, |yjl^ = 5.02 X 10"^^ for the J = 1-K) transition, A = 8.9 x 10"^ sec"^ s Setting N-, = / n,ds we can write V^)=!\li^>
PAGE 116

108Sirvk hc'-A 27 J ^v^^ex)%^^ "TZri V^ hc^A (12) Since T^ = ^if^DJ the column density in the upper state (N ) can now be found. To calculate the column density in all states one assumes LTE, then the number of molecules in the J state is given by (Herzberg 1950) N., = ^^^ N,„,exp(^M40iJi) U "tot kT (13) ID 1 ? where B is the rotational constant (5.764 x 10 Hz for CO) and U is the partition function U = I (2J + 1) exp -'^^'^(j ^ ^^ a=o kT (14) Table AII-1 gives values of U and '^tot/'^j-i "^^ several temperatures, Table AII-1 Values of U and N tot/^J= =1 fo*" 'ho T^x = 10 K 25 50 100 io3 U 3.97 9.37 18.40 54.89 361.7 "tot/No=l 2.30 3.90 6.85 19.34 121.1 Equation (12) and Table AII-1 can be used to find the total column density of CO under the various assumptions noted in the derivation. This estimate is a lower limit due to the assumption that t 1; clearly, if the clouds are optically thick no information on column density is available.

PAGE 117

109Another line of reasoning can give space densities for a molecule if a few conditions are satisfied (see Rank et al 1971 for a discussion). If a molecule is observed in emission, well-removed from discrete sources of radiation, it is likely that collisions raise T C A well above T^ (= 2.7 K, the cosmic background radiation). The relaxation rate between two molecular states for isotropic radiation is (Rank et al. 1971) ^ 3hc''(l e ^) The collision rate can be written 7-=n"^v)^ (16) c m where the summation is over all colliding species, n is the density of each species, a is the effective cross-section, and v is the average relative velocity for the collisions. In most clouds the only particles abundant enough to contribute to equation (16) are atomic and molecular hydrogen. Under these conditions the effective temperature of the tradition is 'eff T + T ^^^> r c where T is the temperature of the colliding particles. Since the molecule is seen in emission it follows that t is not much larger than T The implication is that in order to see the molecule in

PAGE 118

-noemission it must be elevated above Tp by some process (collisions in this case). Thus, a minimum colliding particle density can be found by equating t and t This gives n = 647r |y| ^ A ,,o^ -hv/kTp -hv/kTn ^^^' 3hX"^(l e ^) (l e ^) The usual approach to determine is to use the geometric crosssection ('\. 10"-^^ cm^) and v = Skljm^ '\^ 10^ cm/sec for T, = 100 K and m = m(H2). For CO this gives a minimum colliding particle density of 8.6 X 10^ cm"-^.

PAGE 119

APPENDIX III THE GEOMETRY IN AN INCLINED DISK Projection effects serve to transform the desired quantities of circular velocity (v ), azimuthal angle (e) and radius (r) into their foreshortened counterparts, v^', e' and r' In this appendix the relationships between these quantities are derived, also the calculation of the inclination is considered. The inclination angle is defined to be the angle between the plane of the disk and the plane of the sky. Figure 24 illustrates the various parameters that will be used in the following material. The effect of inclination is to foreshorten, or project, the minor axis; the major axis is unaffected by this. The velocity component v (always perpendicular to the line of sight) is also unaffected by inclination, but it is not measurable. A component of V is observed as the radial velocity (v ) and v^ = Vy sin i (1) The inclination angle is available, provided we know the true axial ratio (q) as well as the projected axial ratio (q'). Since q' = b/a, where b is the semi -mi nor axis and a is the semi -major axis as measured in the sky, and e = (a b)/a then q' = 1 £ (2) The quantity q is calculated similarly, using the true axial ratio for -111f

PAGE 120

-112Figure 24. A disk system viewed face-on (i = 0). The x axis coincides with the major axis and the y axis is also the minor axis. The circular velocity, v and its x and y components are indicated. Also shown is the azimuthal angle, e.

PAGE 121

-113the system. The inclination is then given by 2 2 cos^= ^' ~ ^ (3) 1 q^ The true azimuthal angle, e, is found from the projected angle, e', by tan 9 = tan e'/cos i (4) and V is related to v by r c -^ ^c ^r^^^^ ^''" ^' ^^^ Finally, the true radius of the point in question is given by r = [x^ + (y/cos i)^]* (6) where x and y are the distances along the major and minor axes, respectively, and y' = y/cos i. Since the primed quantities can all be measured, the true values can be calculated using the relations given. The two assumptions made in the derivation are that the true axial ratio is known and that the figure of the system is an oblate ellipsoid supported by rotation about the minor axis.

PAGE 122

REFERENCES Baan, W.A., Haschick, A.D., and Burke, B.F. 1978, Ap. J., 225, 339. Bailey, M.E. 1980, M.N.R.A.S., m_, 195. Balick, B., Faber, S.M., and Gallagher, J.S. 1976, Ap. J., 209, 710. Bautz, L.P. and Morgan, W.W. 1970, Ap. J. (Letters), 162, L149. Bertola, F. and Capaccioli, M. 1975, Ap. J., 200, 439. Binney, J. 1978, M.N.R.A.S., 183, 501. Blackman, R.B. and Tukey, J.W. 1958, "The Measurement of Power Spectra' (Dover: New York). Bottinelli, L. and Gouguenheim, L. 1977a, Astron. & Astrophys., 54 641. Bottinelli, L. and Gouguenheim, L. 1977b, Astron. & Astrophys. ^ 60, 223. Bottinelli, L. and Gouguenheim, L. 1979a, Astron. & Astrophys., 76, 176. Bottinelli, L. and Gouguenheim, L. 1979b, Astron. & Astrophys., 74, 172. Burstein, D. 1979a, Ap. J. Supplement, j41, 435. Burstein, D. 1979b, Ap. J., 234, 435. Burstein, D. 1979c, Ap. J., 234, 829. Butcher, H. and Oemler, A. 1978, Ap. J., 219, 18. Cheng, E.S., Saulson, P.R., Wilkinson, D.T., and Corey, B.E. 1979, Ap. J. (Letters), 232, L139. Coleman, G.D. and Worden, S.P. 1977, Ap. J., 218, 792. Cooper, B.F.C, 1976, in "Methods of Experimental Physics," 12^ part B, ed. M.L. Meeks (Academic Press: New York), 280. de Vaucouleurs, G. 1959, Handbuch der Physik, 53, 275. 114-

PAGE 123

115de Vaucouleurs, G. 1977, in "The Evolution of Galaxies and Stellar Populations" (Yale University Observatory: New Haven), 74. de Vaucouleurs, G. and de Vaucouleurs, A. 1964, "Reference Catalog of Bright Galaxies" (University of Texas Press: Austin). de Vaucouleurs, G. de Vaucouleurs, A., and Corwin, H. 1976, "Second Reference Catalog of Bright Galaxies" (University of Texas Press: Austin). de Vaucouleurs, G. and Neito, J.-L. 1979, Ap. J., 230, 697. Dressier, A. 1980, Ap. J., 236 351. Faber, S.,M. and Gallagher, J.S. 1976, Ap. J., 204, 365. Faber, S.M. and Gallagher, J.S. 1979, Ann. Rev. Astron. &Astrophys., 17, 135. Faber, S.M. and Jackson, R. 1976, Ap. J., 204, 668. Field, G.B. 1975, in "Galaxies and the Universe," ed. A. Sandage, M. Sandage, and J. Kristian (Univ. of Chicago Press: Chicago). Ford, H.C., Jacoby, G. and Jenner, D.C. 1977, Ap. J., 212, 18. Ford, H.C. and Jenner, D.C. 1975, Ap. J., 202^, 365. Ford, H.C, Jenner, D.C, and Epps, H.W. 1973, Ap. J. (Letters), 183 L73. Fosbury, R.A.E., Mebold, U., Goss, W.M., and Dopita, M.A. 1978, M.N.R.A.S., 183, 549. Gallagher, J.S., Faber, S.M., and Balick, B. 1975, Ap. J., 202, 7. Gallagher, J.S., Knapp, G.R., Faber, S.M., and Balick, B. 1977, Ap. J., 215, 463. Gallouet, L. and Heidmann, N. 1971, Astron. & Astrophys. Suppl 3, 325. Gallouet, L., Heidmann, N., and Dampierre, F. 1973, Astron. & Astrophys. Suppl., 12, 89. Gallouet, L., Heidmann, N., and Dampierre, F. 1975, Astron. & Astrophy. Suppl 19, 1. Giovanelli, R. 1979, Private communication. Gisler, G.R. 1976, Astron. & Astrophys., 5^, 137. Gisler, G.R. 1979, Ap. J., 228, 385.

PAGE 124

116Gisler, G.R. 1980, A. J., 85, 623. Gott, J.R. 1977, Ann. Rev. Astron. & Astrophys., 15, 235. Gottesman, S.T. and Weliachew, L. 1977, Ap. J., 211 47. Graham, J. A. 1979, Ap. J., 232, 60. Gunn, J.E. and Gott, J.R. 1972, Ap. J., j76, 1. Gunn, J.E. and Oke, J.B. 1975, Ap. J., 195 255. Guthrie, B.N.G. 1974, M.N.R.A.S., 168, 15. Hausman, M.A. and Ostriker, J. P. 1978, Ap. J., 2Z4, 320. Haynes, M.P., Brown, R.L., and Roberts, M.S. 1978, Ap. J., 221, 414. Haynes, M.P. and Roberts, M.S. 1979, Ap. J., 337, 767. Herbig, G.H. 1970, "Proc. of the XVIth Li^ge Symposium," 59, 13. Herzberg, G. 1950, "Spectra of Diatomic Molecules" (Van Nostrand: New York), 125. Hodge, P.W. 1963, A. J., 68, 691. Hodge, P.W. 1973, Ap. J., 182, 671. Hubble, E.P. 1926, Ap. J., 64, 321. Hubble, E.P. 1936, "The Realm of the Nebulae" (Yale University Press New Haven), 36. Huchtmeier, U.K., Tammann, G.A., and Wendker, H.J. 1977. Astron. & Astrophys. 57, 313. Hulsbosch, A. 1980, Private communication. Humason, M.L., Mayall, N.U., and Sandage, A.R. 1956, A. J., 61^, 97. Illingworth, G. 1977, Ap. J. (Letters), 218, L43. Johnson, H.E. and Axford, W.I. 1971, Ap. J., 165^, 381. Jones, B.T. 1976, Rev. Mod. Phys., 48, 107. Jura, M. 1977, Ap. J., 212^, 634. Kent, S.M. and Sargent, W.L.W. 1979, Ap. J., 230, 667. Killian, D.J. 1978, Ph.D. Dissertation, University of Florida.

PAGE 125

ii/Knapp, G.R., Faber, S.M., and Gallagher, J.S. 1978a, A. J., 83, 11. Knapp, G.R., Gallagher, J.S., and Faber, S.M. 1978b, A. J., 83, 139. Knapp, G.R. and Jura, M. 1976, Ap. J., 209, 782. Knapp, G.R., Kerr, F.J., and Williams, B.A. 1978c, Ap. J., 222, 800. Lang, K.R. 1974, "Astrophysical Formulae" (Springer-Verlag: New York), 161. Larson, R.B. 1974, M.N.R.A.S., m, 229. Larson, R.B. and Tinsley, B.M. 1974, Ap. J., 192, 293. Larson, R.B. and Tinsley, B.M. 1978, Ap. J., 219, 46. Lea, S.M, and De Young, D.S. 1976, Ap. J., 210, 647. Lo, K.Y. and Sargent, W.L.W. 1979, Ap. J., 227, 756. Lucas, R. 1974, Astron. & Astrophys., 36, 465. Lynden-Bell, D. 1967, M.N.R.A.S., 136, 101. Martin, R.N. and Barrett, A.H. 1978, Ap. J. Suppl 36_, 1. Mathews, W.G. and Baker, J.C. 1971, Ap. J., 170, 241. Mathewson, D.S., Cleary, M.N., and Murray, J.D. 1974, Ap. J., 190, 291, Mathewson, D.S., Cleary, M.N., and Murray, J.D. 1975, Ap. J. (Letters), 195 L97. McHardy, I.M. 1974, M.N.R.A.S., 169, 522. Mezger, P.G. and Smith, L.F. 1977, in "Star Formation, I.A.U. Symp. No. 75," ed. T. de Jong and A. Maeder (D. Reidel : Dordrecht), 133. Morgan, W.W. 1958, Pub. A.S.P., 70, 364. Morgan, W.W.. 1959, Pub. A.S.P., 71, 394. Morton, D.C. 1975, Ap. J., 197, 85. Oemler, A. 1974, Ap. J., 194, 1. Oemler, A. 1977, Highlights of Astronomy, 4, 253. Osterbrock, D. 1960, Ap. J., j^, 325.

PAGE 126

-118Osterbrock, D. 1962, in "Interstellar Matter in Galaxies" (Benjamin: New York), 111. Ostriker, J. P. 1977, in "The Evolution of Galaxies and Stellar Populations," ed. B.M. Tinsley and R.B. Larson (Yale University Observatory: New Haven), 369. Ostriker, J. P. and Hausman, M.A. 1977, Ap. J. (Letters), 2j7, L125. Ostriker, J. P. and Tremaine, S.D. 1975, Ap. J. (Letters), 202, L113. Penzias, A. A. and Burrus, C.A. 1973, Ann. Rev. Astron. & Astrophys., 11, 51. Peterson, C.J. 1978, Ap. J., 222, 81. Rank, D.M., Townes, C.H., and Welch, W.J. 1971, Science, YT, 1083. Reif, K., Mebold, U., and Goss, W.M. 1978, Astron. & Astrophys., 67_, LI. Rickard, L.J., Palmer, P., Morris, M. Turner, B.E., and Zuckerman, B. 1977, Ap. J., 213, 673. Riley, J.M. 1975, M.N.R.A.S., jTO, 53. Roberts, M.S. and Steigerwald, D.G. 1977, Ap. J., 217, 883. Rood, H. 1979, Ap. J., 232, 699. Rowan-Robinson, M. 1979, Ap. J., 234, 111. Rubin, V.C, Ford, W.K., Peterson, C.J., and Oort, J.H. 1977^ Ap. J., 211 693. Sandage, A.R. 1961, "The Hubble Atlas of Galaxies" (Carnegie Inst, of Washington: Washington). Sargent, W.L.W., Young, P.J., Boksenberg, A., Shortridge, K. Lynds, C.R., and Hartwick, F.D.A. 1978, Ap. J., 221, 731. Schechter, P.L. and Gunn, J.E. 1979, Ap. J., 229, 472. Schneps, M.H., Ho, P.T.P., Barrett, A.H., Buxton, R.B., and Myers, P.C. 1978, Ap. J., 225, 808. Shane, W.W. 1980, Astron. & Astrophys., 82, 314. Shostak, G.S. 1978, Astron. & Astrophys., 54, 919. Silk, J. and Norman, C. 1979, Ap. J., 234, 86. Smoot, G.F. and Lubin, P.M. 1979, Ap. J. ILetters), 234, L83.

PAGE 127

-119Sunyaev, R.A. 1977, in "Radio Astronomy and Cosmology, I.A.U. Symp. No. 74," ed. D.L. Jauncey (D. Reidel : Dordrecht). Tammann, G.A. 1974, in "Supernovae and Supernovae Remnants," ed. C.B. Cosmovici (D. Reidel: Dordrecht). Tartar, J. 1975, Ph.D. Dissertation, University of California, Berkeley. Tytler, D. and Vidal, N. 1978, M.N.R.A.S., 182, 33p, Ulich, B.L. and Haas, R.W. 1976, Ap. J. Suppl 30, 247. Unwin, S.C. 1980, M.N.R.A.S., 190, 551. van den Bergh, S. 1960a, Ap. J., 13^, 215. van den Bergh, S. 1960b, Ap. J., 131, 558. van den Bergh, S. 1975, Ann. Rev. Astron. & Astrophys., 21, 217. van den Bergh, S. 1976a, Ap. J., 208, 673. van den Bergh, S. 1976b, Ap. J., 206, 883. van den Bergh, S. 1976c, A. J., Sl^, 797. White, S. 1976, M.N.R.A.S., 174, 19. Wright, M.C.H. 1974, in "Galactic and Extra-Galactic Radio Astronomy," ed. G.L. Verschuur and K.I. Kellerman (Springer-Verlag: New York), 291. Young, P.J., Boksenberg, A., Lynds, C.R., and Hartwick, F.D.A. 1978a, Ap. J., 222, 450. Young, P.J., Sargent, W.L.W., Kristian, J., and Westphal J. A. 1979, Ap. J., 234, 76. Young, P.J., Westphal, J. A., Kristian, J., Wilson, C.P., and Landauer, F.P., 1978b, Ap. J., 221, 721. Zuckerman, B. and Palmer, P. 1974, Ann. Rev. Astron. & Astrophys., 12, 279.

PAGE 128

BIOGRAPHICAL SKETCH On reaching the age of 13, Douglas William Johnson had managed to take apart every watch, toaster, alarm clock, and gadget in his parents' wood frame home in both Gas City, Indiana, and Connellesville, Pennsylvania. His sisters, Rhonda and Peggy, grew up with the enduring fear their hair dryer would fall prey to his eager little hands. Doug's mother comforted herself with the conviction that her second eldest son would go on to become a mechanic. Anyone who so loved to work with his hands, she reasoned, must be born to things mechanical Mary Ann Johnson's instincts about her son were good, but she had not reckoned on the draw of the night sky. Those brilliant, changing lights--completely out of reach--caught Doug's interest effortlessly. When his father's job as a Postal Inspector brought the family east to New Jersey, in 1968, the would-be astronomer constructed his own telescope to ply his trade in the Johnson's lush and overgrown backyard. Doug attended Boonton High School and pursued his favorite subjects of physics and math. With a well -nurtured respect for the universe and the deep curiosity of a scientist, Doug went on to college at Rensselaer Polytechnic Institute in Troy, New York. In 1975, equipped with a bachelor's degree in physics, and a wife gracious enough to work full-time, he arrived in Gainesville,. It was late summer, a time designed to kill Yankees with 120-

PAGE 129

-121its stifling, muggy heat. But the Johnsons survived and prospered, spreading their prosperity among the animal kingdom through the addition of four highly personable cats and a sweet-faced mongrel from the county pound to their family. When lured away from his computers, mounds of graph paper, or blackboards, Doug holds true to his earliest love of tinkering. He is an accomplished carpenter, a competent mechanic, and makes a mean pitcher of frozen daquiris. Though not given over to an abandoned social life, he misses few science fiction movies. His friends are a peculiar mixture of astronomy graduate students and newspaper types--a situation forced on him by his wife, Maryfran, the tolerant breadwinner of the family for the last five years.

PAGE 130

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. /^4v.^ /j,^ !,'>>mJur. 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. (ii4 i4^^i.^^ /,. r-^ c. iL^.-, I)r. Thomas D. Carr 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. KAAI7.VV '^hv Uf'vXvv. Dr. Kwan-Yu Chen Professor of Astronomy

PAGE 131

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. %-'-vM Dr. Gary G. Ihas Associate Professor of Physics 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. jx.LQi:... H^^Q-fr. Dr. William Weltner, Jr. Professor of Chemistry 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. August 1980 Dean, Graduate School


EM PERA TURE (K)
-0.60 -0.20 0.20 0.6,0
(XI o1 )
Figure 5. The ^CO data from July 1978.
-45-


-12-
A more moderate form of mass accretion is suggested by several
authors (Bottinelli and Gouguenheim 1977a, Gallagher et al. 1977, and
Knapp, Kerr, and Williams 1978c)to explain the inclined disk of
NGC 4278. The differing directions of the angular momentum of the
stellar component of the galaxy and the neutral hydrogen make an
internal origin of the matter difficult to believe.
A similar situation occurs in NGC 1052 (Knapp, Faber, and
Gallagher 1978a, Fosbury et al. 1978, Reif, Mebold, and Goss 1978) and
the accretion of an intergalactic HI cloud is suggested. The major
objection to this hypothesis is the lack of sufficiently massive clouds
available for accretion.
The results of Mathewson et al. (1975) purporting to find HI
clouds in the Sculptor group have been disputed by Haynes and Roberts
(1979). The latter group contend that the material is a portion of the
Magellanic Stream. Further, Lo and Sargent (1979) have searched nearby
groups for detached HI clouds and find none more massive than
% 4 x 107 M .
A number of other authors have examined clusters of galaxies for
HI emission (Haynes et al. 1978, Baan et al. 1978) while others have
examined the line of sight to quasars for HI absorption (Roberts and
Steigerwald 1977, Shostak 1978). No isolated HI clouds are seen in
emission in the clusters and the HI absorption measurements show that
large clouds of neutral hydrogen are almost never seen outside
galaxies.
O
Thus it appears difficult to reconcile the several times 10 Mq
of HI found in NGC 4278 and 1052 with the dearth of sufficiently
massive candidates for accretion. Silk and Norman (1979) propose an


CHAPTER III
RESULTS
In this chapter the results of the observations are presented.
The carbon monoxide data are treated first, followed by the neutral
hydrogen study of NGC 185 and the surrounding region. Only the results
are considered in this chapter; the analysis and interpretation of the
observations are found in Chapter IV.
Carbon Monoxide
Positive Result in NGC 185
The dwarf elliptical galaxy NGC 185 was detected at the 115.2712
12
GHz frequency of CO. Figure 2 shows the best spectrum obtained on
the source, averaging data from all three observing sessions. The
256 MHz total bandwidth corresponds to about 660 km/sec. The feature
is not discernible in the 128 MHz bandwidth spectrum, probably because
of the lower signal-to-noise (SN) ratio. Table 1 summarizes the
parameters of the spectral feature as measured in Figure 2.
Since the line is only slightly greater than 3 standard deviations,
it is very important to be certain of the reality of the feature. One
method is to smooth the spectrum with a function (in this case rect
angular) of about the same width as the suspected spectral feature.
This procedure has the effect of increasing the SN ratio at the expense
-39-


-14-
Small mass stars have stellar winds, Mira variables are known to eject
mass during certain stages, Type I supernovae occur in population II
stars, and planetary nebulae have been observed in early-type galaxies
of the local group.
The calculation of the contribution by stellar evolution to the
interstellar medium (ISM) of early-type galaxies depends more on theo
retical estimates than observational evidence. To date, the most impor
tant observational evidence is the detection of planetary nebulae in
nearby dwarf ellipticals (including NGC 185) by Ford and Jenner (1975).
Considering the uncertainties in the observations, the observed planetary
-1 9 -1
birthrate of > 0.012 yr (10 L ) agrees well with Larson and Tinsley's
(1974) estimate of 0.05 yr-^ (10^ L )*.
Following the reasoning of Faber and Gallagher (1976) and adopting
a mass per planetary of 0.2 M^ results in a mass loss rate of 0.010

-1/9 -1
yr (10 L ) from planetaries. Consideration of Mira-type variables
leads to a final assumed mass ejection rate of 0.015 yr (10 L ) .
The conservative nature of this calculation is apparent when one
considers that the present mass loss rate is certainly lower than that
of earlier epochs. This is primarily because any high-mass stars
would have evolved quickly and cycled their mass back to the ISM early
in the galaxy's evolution.
Further, the contribution of mass from Type I supernovae (ap
parently confined to population II stars, Tammann 1974) and Type II
supernovae (massive progenitors) earlier in the galaxy's evolution
have been ignored. Even this conservative approach leads to contra-
9 10
dictions in the ISM of early-type galaxies after 10 -10 years (Faber
and Gallagher 1976).


-36-
represents frequency, but since the frequency of the molecular
transition is already known the axis is calibrated in km/sec using
the following relation
(2)
where vg is the center velocity of the observed band, v^ is the velocity
of the source (corrected for the earth's rotation and revolution, i.e.
heliocentric), v is the rest frequency of the spectral features, and
vobs bservec* frequency.
The ordinate, corrected antenna temperature (T^), is found by
calibration as described in Appendix I. The unit is Kelvins (K), and
it is related to the flux density (S) by the following equation
(3)
s = 2 k TA/Ae
where n£ is the antenna efficiency due to the loss of elements
(spillover, blockage, and ohmic losses in the antenna), A is the
G
effective area of the antenna, and k is Boltzmann's constant.
Neutral Hydrogen Observations at Green Bank, WV
Telescope Description
The neutral hydrogen observations were made with the NRAO 43 m
radio telescope at Green Bank, West Virginia. The telescope is an
equatorially mounted instrument completely under computer control. The
observations were taken during one session from July 13 to July 23, 1978.


-25-
It appears from these observations that there may be a correla
tion between radio galaxies in the center of clusters and anomalous
nuclei. Since current understanding of radio sources usually involves
massive objects (which can also explain the anomalous nuclei) the
circumstantial connection between a lack of ram-jet stripping and a
supermassive object in a galaxy's nucleus is established.
However, within this scenario it is quite unclear whether the
retained gas actually formed the massive object or just fuels it. The
possibility that the galaxy formed with a massive nucleus cannot be
overlooked; indeed, one of the central questions yet to be answered is
how important are the initial conditions under which the galaxy formed.
Cyclic Bursts of Star Formation
The contention that the star formation rate in an elliptical or
SO galaxy is strongly dependent on time gains credibility only recently.
Van den Bergh (1975) cites several examples of elliptical galaxies
experiencing anomalously vigorous star formation.
NGC 5128 (also known as the radio source Centaurus A) has recently
been shown by van den Bergh (1976a) to be undergoing very active star
formation along and interior to its prominent equatorial dust band. He
also suggests that the source of the unusual dust and gas in the system
is stellar debris shed by the stars. The galaxy is apparently a rare
field elliptical. It does not belong to a rich cluster and presumably
lacks any ram-jet stripping which may exist in such an environment.
A strong argument against this hypothesis is the finding by
Graham (1979) that the old stellar population rotates much slower than


-116-
Gisler, G.R. 1980, A. J., 85, 623.
Gott, J.R. 1977, Ann. Rev. Astron. & Astrophys., 15_, 235.
Gottesman, S.T. and Weliachew, L. 1977, Ap. J., 211, 47.
Graham, J.A. 1979, Ap. J., 232, 60.
Gunn, J.E. and Gott, J.R. 1972, Ap. J., 176, 1.
Gunn, J.E. and Oke, J.B. 1975, Ap. J., 195, 255.
Guthrie, B.N.G. 1974, M.N.R.A.S., 168, 15.
Hausman, M.A. and Ostriker, J.P. 1978, Ap. J., 224, 320.
Haynes, M.P., Brown, R.L., and Roberts, M.S. 1978, Ap. J., 221, 414.
Haynes, M.P. and Roberts, M.S. 1979, Ap. J., 337, 767.
Herbig, G.H. 1970, "Proc. of the XVIth Lige Symposium," 59., 13.
Herzberg, G. 1950, "Spectra of Diatomic Molecules" (Van Nostrand:
New York), 125.
Hodge, P.W. 1963, A. J., 68, 691.
Hodge, P.W. 1973, Ap. J., 182, 671.
Hubble, E.P. 1926, Ap. J., 64, 321.
Hubble, E.P. 1936, "The Realm of the Nebulae" (Yale University Press
New Haven), 36.
Huchtmeier, W.K., Tammann, G.A., and Wendker, H.J. 1977. Astron. &
Astrophys., 57, 313.
Hulsbosch, A. 1980, Private communication.
Humason, M.L., Mayall, N.U., and Sandage, A.R. 1956, A. J., 6_1, 97.
Illingworth, G. 1977, Ap. J. (Letters), 218, L43.
Johnson, H.E. and Axford, W.I. 1971, Ap. J., 165, 381.
Jones, B.T. 1976, Rev. Mod. Phys., 48, 107.
Jura, M. 1977, Ap. J., 212, 634.
Kent, S.M. and Sargent, W.L.W. 1979, Ap. J., 230, 667.
Killian, D.J. 1978, Ph.D. Dissertation, University of Florida.


TEMPERATURE (K) CX10"1 )
-0.10 0.00 0.10 0.2,0
(X101 )
Figure 21. The HI spectrum at 49' north of NGC 185.


-4-
Taking into account the statistics of random projection on the
sky (for we are viewing a two-dimensional projection of a three-
dimensional object) it appears that the ellipticals are distributed
normally about a mean of E3.6 (e = 0.36) with a dispersion of 0.11
(de Vaucouleurs 1977). It appears that true E0 and E5.5 (de Vaucouleurs
contends that E5.5 is the flattest bona fide elliptical) are rela
tively rare.
The lack of flat systems, usually considered to be a dynamical
effect caused by instabilities in thin disks, suggests that perhaps
spiral density waves and the attendant star formation are suppressed
in disks of sufficient thickness.
The origin of the flatness of elliptical systems has long been
thought a natural consequence of rotation. The greater the rotational
velocity, the greater the degree of flattening. This of course implies
that the three-dimensional figure is an oblate ellipsoid (polar diameter
smaller than the equatorial diameter).
In recent years several rotation curves of elliptical galaxies
have been published (Bertola and Cappaccioli 1975, Illingworth 1977,
Peterson 1978, Sargent et al. 1978, and Young et al. 1978a)which cast
strong doubt on the validity of this simple approach. The small ob
served ratio of the maximum rotational velocity to the central dis
persion velocity mitigates strongly against models which use isotropic
velocity distributions and either oblate or prolate ellipsoids
(Schechter and Gunn 1979). It appears necessary to use both anisotropic
velocity distributions as well as rotation to account for the observed
rotation curves (Binney 1978, Schechter and Gunn 1979).


-28-
Larson and Tinsley (1974) have calculated models for elliptical
galaxies with star formation rates, continuing to the present, capable
of consuming the gas shed by other stars. While the integrated colors
of the models are consistent with observed galaxies, the expected
gradient of increasingly blue colors in the nucleus is not widely
observed. The most obvious drawback in their modeling is the use of
an initial mass function (IMF) which is fairly rich in hot stars.
As Faber and Gallagher (1976) comment "Since we have no a priori
knowledge of the IMF in ellipticals, star formation might conceivably
be confined to stars of small mass and low luminosity" (p. 370). They
go on to point out that the star formation must proceed efficiently
since very little, if any, interstellar material is seen in most
elliptical galaxies.
There is, however, observational evidence for star formation with
an anomalous IMF. Van den Bergh (1976c) suggests that an IMF deficient
in high mass stars is the most likely explanation for the lack of HII
regions in the Sa galaxy NGC 4594 (M104). Knots of young, blue stars
are observed near the prominent dust lanes. Normal, massive 0 and B
stars would form prominent HII regions under such conditions. Further,
this galaxy was not detected by Gallagher et al. (1975) in HI even
though a normal dust-to-gas ratio indicates it should have been easily
seen. Van den Bergh suggests that the lack of HI is due to its being
converted into molecular hydrogen, thus escaping detection.
On theoretical ground also, an IMF skewed away from massive stars
may be expected. Jura (1977) finds that one effect of reduced heating
of interstellar clouds (expected in elliptical galaxies) is to allow
clouds with much smaller masses to become gravitationally unstable


-13-
alternative hypothesis, the accretion of gas-rich dwarf galaxies.
They find that the gas component of the dwarfs will lose energy through
dissipation and fall to the central regions of the accreting galaxy.
The infalling gas, depending on the individual cloud mass, may either
form stars or continue to fall into the nucleus where it may fuel a
radio source. The stars will also be incorporated into the accreting
galaxy but with less visible effects.
Silk and Norman (1979) also consider the interaction of a Mathews
and Baker (1971) type wind and the infalling material. If the amount
of this material is sufficiently large the resulting supernovae (from
the high-mass stars formed) will help in driving the galactic wind.
However, an enhanced wind has the effect of inhibiting mass infall and
the process slows itself. The net effect may be for star formation to
proceed in cyclical bursts--a notion also suggested by van den Bergh
(1975) in a somewhat different context.
Both of the mass accretion processes described so far deal with
normal or giant cD elliptical galaxies. In order to be effective in
capturing and assimilating other systems the accreting galaxy must be
large. The evolutionary mechanism discussed in the following section,
mass loss from stellar evolution, operates in all systems. This in
cludes the dwarf ellipticals NGC 185 and 205 considered in greater
detail in Chapter III.
Mass Loss Due to Stellar Evolution
It has recently been appreciated that the normal evolution of
stars in an early-type galaxy will be a source of interstellar material.


-99-
of the sky preclude firm conclusions about the dynamics of the
system.
We find, however, that the mass-to-luminosity ratio in the nucleus
(the planetary nebulae are both within 1 arcminute of the nucleus) is
between 5 and 18, with the most likely value around 8. This number is
quite reasonable for an early-type system, but the difficulty in deter
mining the ratio, and the lack of agreement from different methods,
prohibit any conclusions beyond those already mentioned.
Neutral hydrogen observations of the vicinity of NGC 185 are
presented. They reveal the probable existence of high-velocity hydrogen
at -180 km/sec at and near NGC 185. The observations also show an
enhancement of about 5 mK antenna temperature at the position of the
galaxy with respect to 1-2 beamwidths away. The data are inadequate
to determine if the material is in NGC 185 or is just an enhancement
in the local material coincidentally superposed on NGC 185.


REFERENCES
Baan, W.A., Haschick, A.D., and Burke, B.F. 1978, Ap. J., 225, 339.
Bailey, M.E. 1980, M.N.R.A.S., 191, 195.
Balick, B., Faber, S.M., and Gallagher, J.S. 1976, Ap. J., 209, 710.
Bautz, L.P. and Morgan, W.W. 1970, Ap. J. (Letters), 162, L149.
Bertola, F. and Capaccioli, M. 1975, Ap. J., 200, 439.
Binney, J. 1978, M.N.R.A.S., 183, 501.
Blackman, R.B. and Tukey, J.W. 1958, "The Measurement of Power Spectra"
(Dover: New York).
Bottinelli, L. and Gouguenheim, L. 1977a, Astron. & Astrophys., 54,
641.
Bottinelli, L. and Gouguenheim, L. 1977b, Astron. & Astrophys., 60,
223.
Bottinelli, L. and Gouguenheim, L. 1979a, Astron. & Astrophys., 7J5,
176.
Bottinelli, L. and Gouguenheim, L. 1979b, Astron. & Astrophys., 74,
172.
Burstein, D. 1979a, Ap. J. Supplement, 41, 435.
Burstein, D. 1979b, Ap. J., 234, 435.
Burstein, D. 1979c, Ap. J., 234, 829.
Butcher, H. and Oemler, A. 1978, Ap. J., 219, 18.
Cheng, E.S., Saulson, P.R., Wilkinson, D.T., and Corey, B.E. 1979,
Ap. J. (Letters), 232, L139.
Coleman, G.D. and Worden, S.P. 1977, Ap. J., 218, 792.
Cooper, B.F.C. 1976, in "Methods of Experimental Physics," 12_ part B,
ed. M.L. Meeks (Academic Press: New York), 280.
de Vaucouleurs, G. 1959, Handbuch der Physik, 53, 275.
-114-


-108-
u
J Jv(Tex
hcA v ex
he
(12)
Since = n^Tg, the column density in the upper state (N ') can now be
found. To calculate the column density in all states one assumes LTE,
then the number of molecules in the J state is given by (Herzberg 1950)
N, ^ Ntotexp(^iH)
kT
(13)
ID 19
where B is the rotational constant (5.764 x 10 Hz for CO) and U is
the partition function
U I (2J + 1) exp -hBJ^ + 1}
0=0 K1
(14)
Table AII-1 gives values of U and fr several temperatures.
Table AII-1
Values of and N, ,/N, for 12C0
tot J=1
Tex 10 K
25
50
100
103
u
3.97
9.37
18.40
54.89
361.7
Ntot/NJ=X
2.30
3.90
6.85
19.34
121.1
Equation (12) and Table AII-1 can be used to find the total column
density of CO under the various assumptions noted in the derivation.
This estimate is a lower limit due to the assumption that 1;
clearly, if the clouds are optically thick no information on column
density is available.


-97-
process; early-type galaxies are preferentially found in rich clusters
that would be most effective in retaining an intracluster medium.
However, it is not clear that the presence of early-type systems in
rich clusters is not a result of initial conditions at the time of the
cluster formation. Indeed recent work has tended to support this view.
Observations of nearby galaxies do not fit well with either
of the two mechanisms. Some have been detected in HI when the galaxy
should be able to expel its ISM. Worse, in at least two cases the HI
has a broad distribution, entirely contrary to expectations for a par
tially operating galactic wind or ram-jet stripping. The suggestion
has been made that the HI has recently been accreted, but this is dif
ficult to support based on the space density of HI clouds large enough
to account for the material. Other early-type galaxies show patches
of obscuring matter firmly indicating that mass removal processes are
not completely efficient.
12
Observations of CO were undertaken for early-type galaxies to
determine the role of star formation in the removal of the ISM. The
CO molecule was chosen because it is widely distributed, resistant to
dissociation, and has a transition frequency accessible to highly
sensitive radio telescopes.
12
The J = 1-K) transition of CO was detected in the dwarf ellipti
cal galaxy NGC 185. The physics of the line formation process strongly
imply that the emitting region is the north-west dust patch about 15
arcseconds from the nucleus. The fact that the line is seen at all
3 -3
indicates that the region is at least as dense as 10 cm The minimum
mass consistent with the observations is about 10^ M .


TEMPERATURE K)
(XI O1 )
j j IJ IJ ilij II l_l [ m m IIII
VELOCITY (KM/SECHXio1 I
Figure 3. The spectrum of Figure 2 smoothed with a 5 channel (13 km/sec) rectangular
function.


CHAPTER IV
DISCUSSION
In this chapter the implications of the observations presented in
Chapter III are considered. The simplest calculations from these data
involve assumptions about the excitation temperature and optical depth.
The portion of the dissertation based on such assumptions is found in
this chapter.
NGC 185
This dwarf elliptical galaxy is usually considered to be a com
panion of M31. It is located about 6 to the north of M31 and the
difference in radial velocities is about 55 km/sec. For a discussion
of the virial mass of the M31 system and comments on membership in
that system see Rood (1979). A distance of 0.69 Mpc is assumed for
NGC 185 throughout the following discussion. At this distance 1" =
1.04 x 10 y cm = 3.3 pc.
The classification given by de Vaucouleurs et al. (1976) is E3p.
A detailed photographic study of the system by Hodge (1963) shows
clearly that the ellipticity (defined by e = 1 ) varies from 0.18
near the nucleus to 0.26 for the outer contours at 350" from the
nucleus. An increase in ellipticity with radius is commonly observed
for elliptical galaxies. The E3 classification (which is based on the
outer contours only) is confirmed by Hodge's photometry.
-68-


TEMPERA TURE (K)
(XI O1 )
. The ^CO spectrum for NGC 3226,
Figure 13


-3-
Forming the tines of the fork are two parallel sequences of spiral
galaxies; one with a bar, the other without. The trend along the tines
is from tightly wound spiral arms (Sa or SBa) to looser, more open
arms at the end (Sb-Sc or SBb-SBc).
In addition to these major players in the drama, some 3l of all
galaxies are irregular, possessing no dominant symmetrical structural
features.
Many additions and modifications have been made to this initial clas
sifying scheme (Hubble 1936, Morgan 1958 and 1979, de Vaucouleurs 1959,
van den Bergh 1960a and b, Sandage 1961) but it has remained remarkably
unchanged over the years. In large measure the modifications are to take
into account the more extensive information available due to more sensi
tive equipment and increasing access to other spectral regions.
The following sections describe the elliptical and lenticular
galaxies in more detail and lay the groundwork for the statement of the
thesis problem in the final section of Chapter I. For historical reasons
both elliptical and lenticular systems are commonly known as "early-type"
galaxies.
The Morphology of Elliptical Galaxies
As the name implies, the elliptical galaxies are characterized by
elliptical isophotes. The stellar population usually appears to be
well-evolved with little or no interstellar gas or dust.
The degree of ellipticity E is defined to be (a b)/a (where a
and b are the semi-major and semi-minor axes, respectively). The range
observed is 0.0-0.7 (E0-E7) with E0-E1 the most common and decreasing
in frequency at the flatter end of the range.


-69-
The system is peculiar because of the presence of two dark dust
patches within about 20 arcseconds of the nucleus. Figure 22 is an
optical photograph of NGC 185 and a schematic of the dust regions is
reproduced from Hodge (1963) in Figure 23. Careful comparison of the
two figures reveals the extent of the dust in the optical photograph.
Table 4 tabulates the positions of the patches used for observation
2
and the areas in square arcseconds and pc as measured from Figure 23.
Note that dust patch I (DP-I) has 2 entries, la and I. DP-Ia is the
dark core of the northwest patch and DP-I is the larger, less dense
patch (but also including DP-Ia).
Table 4
Positions, Areas, and Telescope Filling Factors for
Dust Patches in NGC 185
Object
a(1950.0)
6(1950.0)
Area
(-)2
PC2
Fi 11ing
Factor
DP-Ia
0h36m10^3
4803'54"
35
380
0.0072
DP-I
Same as
DP-Ia
115
1250
0.0235
DP-I I
0h36m12?2
4803128"
350
3800
0.0716
DP-I+II
(Center of
NGC 185)
0h36ml1?4
4803'44"
465
5050
0.0951
The CO
observations were
taken at the
position
of DP-I
in December
1977 and July 1978. The December 1979 observations were made at both
DP-I and -II as well as the center of the galaxy. However, as the
FWHM for the 11 m telescope indicates in Figure 22, observations of one
patch do not entirely exclude the other. Further, the pointing for


-8-
Perhaps the most influential discovery in cosmology (and also
bearing heavily on galaxy formation) has been the 2.7 K microwave back
ground. The existence of nearly isotropic (see Cheng et al. 1979 for
a discussion of a dipole anisotropy attributed to the earth's motion
with respect to the background) homogeneous radiation with an apparent
blackbody spectrum strongly constrains galaxy formation theories.
The cosmic background radiation is almost universally believed to
be the remnant from the era of decoupling; the time when the universe
had cooled enough to allow electrons and protons to recombine (about
3000 K). The effect of the formation of neutral hydrogen is to reduce
drastically the opacity of the matter with respect to the radiation.
Before recombination Thompson scattering of radiation off electrons
coupled very strongly the matter with the radiation, they were kept in
thermal equilibrium and cooled together. After decoupling the two com
ponents evolved essentially independently of one another and the 2.7 K
background seen today is the remnant of the radiation component.
This forms the basis of the expectation that the early universe
was homogeneous and isotropic. Actually one can only say that it is
homogeneous and isotropic to at least the smallest scale observed,
currently that means the background is smooth to within 1 mK on a scale
of 7 (Smoot and Lubin 1979, see Sunyaev 1977 for a discussion of
fluctuations).
The condensations that have eventually evolved into clusters of
galaxies, galaxies and stars must have occurred during a later epoch.
The nature of these perturbations is currently a topic of great
interest. Several theories have been advanced to account for the
processes by which the perturbations form and grow (see Gott 1977 and


TEMPERATURE (K)
(X101 )
12
Figure 10. The CO spectrum for NGC 404. The negative feature at 200 km/sec is probably
not real, but rather an artifact of the multichannel filterbank.


-48-
second channel of the receiver (Figure 7). We have no physical reason
to expect a polarized feature and indeed, this transition has never been
observed to be polarized in our own galaxy. Also, the tracking procedure
rotates the observed polarization planes with respect to the sky; even if
the signal were linearly polarized we should be able to detect it in each
channel. The lack of a feature in this channel is almost certainly due
to the low SN ratio.
A further complication during the December 1979 session was poor
pointing corrections for the telescope, especially in regions far from
the celestial equator (the declination of NGC 185 is 48). Consequently
we feel that the lack of a feature in the second channel is within the
statistical and instrumental uncertainties of the experiment.
Negative Results
A total of 11 early-type systems were observed during the three
observing sessions. Our criteria for selecting candidates for CO
emission were based on the notion that a system able to retain HI or
dust would be more likely to have significant star formation.
Consequently we chose several of the early-type systems that have
been detected in HI. Added to the list were galaxies that are close
(< 15 Mpc) and also show some form of an ISM. Often this was a nota
tion in the Reference Catalog of Bright Galaxies (RCBG, de Vaucouleurs
and de Vaucouleurs 1964) that the system contained obscuring matter.
Table 2 lists the 11 galaxies selected for the CO study. Included
are the positions observed, the assumed distance and radial velocities
as well as the Holmberg magnitude and the total luminosity. The dates
in which useful data were collected are also noted. The final quantity


-6-
form of detectable population I material. The CO detection described
in Chapter III is another example of the incongruities present in
elliptical systems.
The Morphology of Lenticular (SO) Galaxies
The lenticular systems, occupying the vertex of Hubble's tuning
fork diagram, are as perplexing today as ever. The essential features
of these galaxies are an elliptical bulge and. a disk, both apparently
composed of population II stars. The disks are devoid of spiral arms
and the systems lack the obvious star formation proceeding in spiral
galaxies.
Suggestions as to the origin of the lenticulars range from van den
Bergh's proposal (1976b)that they are part of a complete sequence of
gas-poor, "anemic" spirals parallel to the S and SB types to the more
recent suggestion (Gunn and Gott 1972, Gisler 1979, and references
therein) that they are normal spirals which have been stripped of their
gas and dust by an interaction with an intracluster medium.
Much work has been done in recent years on these systems and
some interesting observations have been reported. Neutral hydrogen
observations by Balick et al. (1976) and a number of others place the
lenticulars roughly midway between the Sa-SBa and E systems. This
evidence supports the contention that they form another sequence of
spirals much like the normal sequence, only gas-poor.
Further, Burstein (1979a,b,c) finds significant differences in
the bulge-to-disk ratios of lenticulars and spirals, as well as
"thick" disks in lenticulars and not spirals. These are structural


APPENDIX III
THE GEOMETRY IN AN INCLINED DISK
Projection effects serve to transform the desired quantities of
circular velocity (v ), azimuthal angle (e), and radius (r) into their
foreshortened counterparts, vc, 0', and r. In this appendix the
relationships between these quantities are derived, also the calculation
of the inclination is considered. The inclination angle is defined to
be the angle between the plane of the disk and the plane of the sky.
Figure 24 illustrates the various parameters that will be used in
the following material. The effect of inclination is to foreshorten,
or project, the minor axis; the major axis is unaffected by this. The
velocity component v (always perpendicular to the line of sight) is
X
also unaffected by inclination, but it is not measurable. A component
of Vy is observed as the radial velocity (vr) and
vr = vy sin i (1)
The inclination angle is available, provided we know the true
axial ratio (q) as well as the projected axial ratio (q1). Since
q1 = b/a, where b is the semi-minor axis and a is the semi-major axis
as measured in the sky, and e = (a b)/a then
q' = 1 e (2)
The quantity q is calculated similarly, using the true axial ratio for
-111-


-98-
A region as large as this dust patch .is gravitationally unstable
and is collapsing, probably forming stars at a later epoch. The
nucleus of NGC 185 contains several bright blue, presumably young
stars which have formed in the relatively recent past.
Observations of another dwarf elliptical galaxy, NGC 205, failed
to find CO emission even though this galaxy contains a larger popula
tion of young stars and more total dust content distributed in 12
patches. The densest of the regions are not as dense as the northwest
dust patch of NGC 185. It is suggested that this is the primary reason
for not detecting NGC 205 while obtaining a positive result on NGC 185.
Negative results for 9 other early-type systems are presented and
the implications are discussed. The major finding is that the material
shed by the stars cannot be contained in optically thin clouds and
still elude detection in these observations. Alternatives are that
the gas is clumpy and thus optically thick, it has been swept out by
either a galactic wind or ram-jet stripping, or it is being consumed
by star formation with an IMF skewed towards low-mass stars.
Theoretical support for a skewed IMF is presented, but the question
cannot be answered by observations with present-day equipment. More
sensitive CO data are needed and the interaction between the intra
cluster medium and the ISM needs to be explored further.
Using radial velocities from two planetary nebulae in the system
we find convincing evidence that the galaxy is rotating about its minor
axis. The data can be further used to derive the mass interior to the
measured points. Of course some basic assumptions about the true
ellipticity of the galaxy and the effect of projection onto the plane


-94-
Based on our observations it is clear that there is HI around
NGC 185 with a distribution greater than 24'. Figure 19 shows a fea
ture with about 15 mK of HI when the "off" is located many beamwidths
to the west. When a spectrum is taken with the "offs" only 24' or 48'
to the west, east, south, and north, the strength of the feature drops
to around 5 mK. This clearly indicates that we are subtracting out
about 10 mK of broadly distributed HI.
Analysis of the individual spectra with an "off" in one of the four
cardinal directions shows that there is an excess of HI with respect to
all directions except perhaps the north. The intriguing possibility is
that while there is clearly a broadly distributed HI components around
NGC 185, the spectra show that there is excess line radiation from the
position of the galaxy when compared with areas 1 to 2 beamwidths
away.
Of course it is possible that an enhancement in the high-velocity
cloud coincidentally projects onto NGC 185. Our observations are not
sensitive enough nor spatially resolved enough to answer the question
definitively.
It is interesting to note that if the 5 mK excess does originate
in NGC 185, the mass associated with it is about 9 x 104 M This is
@
quite reasonable since the minimum molecular mass needed to generate
4
the observed CO signal is 10 Mq. Indeed, a calculation of the mass
of HI expected in the galaxy from normal stellar evolution (see equation
f.
7, in the previous section) is 1.2 x 10 Mq.
A final possibility suggested by the detection of a plume between
Io and 2 further north of the galaxy is that we are seeing a tail
or streamer of HI associated with NGC 185. Since the galaxy is a


-113-
the system. The inclination is then given by
2.
COS 1
(3)
The true azimuthal angle, e, is found from the projected angle, e1,
by
tan 9 = tan e'/cos i (4)
and v is related to v by
r c
vc = vr/cos 0 sin i (5)
Finally, the true radius of the point in question is given by
r = [x2 + (y/cos i)2]2 (6)
where x and y are the distances along the major and minor axes, respec
tively, and y' = y/cos i.
Since the primed quantities can all be measured, the true values
can be calculated using the relations given. The two assumptions made
in the derivation are that the true axial ratio is known and that the
figure of the system is an oblate ellipsoid supported by rotation about
the minor axis.


-71-
\
MAJOR
AXIS
gure 23. A schematic of the dust regions near the nucleus of
NGC 185 (Hodge 1973). DP-I is northwest of the center,
DP-II is southeast and DP-Ia is darker reqion within
DP-I.


-34-
errors are large in this approach and they give only a general idea
of the nature of the molecular cloud being observed.
Data Reduction Techniques
Each spectrum consists of two channels of 256 points each. For
December 1977 and July 1978 only one polarization was available and it
was detected in both the 256 MHz wide filter bank as well as the 128
MHz wide filter bank. The December 1979 observations took advantage
of the two orthogonal polarizations then available and detected them
separately in two 256 MHz filter banks. At the rest frequency of the
12
J = 1-M3 transition of CO the 1 MHz resolution corresponds to 2.6
km/sec.
The on-site PDP 11/40 computer writes the data for each spectrum
(typically representing 5-10 minutes of observing) on disk. Real time
analysis can be done at the telescope and allows the observing time to
be optimally used. At the completion of the observing session the
disk is copied onto a binary coded tape.
The next step in the processing is to rewrite the tape with the
IBM 360 computer at the NRAO in Charlottesville, VA, to put the infor
mation into an IBM-compatible format. This is entirely a translation
and no analyses are performed.
The final reduction is done at the University of Florida using
the resources of the Northeast Regional Data Center. Each spectrum is
examined for an unusually high root-mean-square deviation; all of the
offending spectra are rejected. The baselines are then examined and
if a polynomial of more than first degree is required to fit the base
line, the spectrum is discarded.


Table 2
Continued
Notes: The positions for NGC 185 and 205 were measured from the Palomar Sky Survey. The
positions for the other galaxies are from Gallouet and Heidmann (1971), Gallouet
et al. (1973), or Gallouet et al. (1975). The galaxy types and heliocentric
velocities are from de Vaucouleurs et al. (1976). The Holmberg magnitude is cal
culated as described by Gallagher et al. (1975).
I
cn
O


-106-
by atoms or molecules in the cloud, corrected for self-absorp
tion.
The quantity measured with a radio telescope is most closely
related to the intensity at the transition frequency, [I (s )] minus
the intensity near the transition frequency [Iv(s )]. Since
I (S ) V; I (0)
V 0 0
en Tn(sn)
Al = I (s ) I (0) = [ I (0)](1 e 0 0 )
n o o o k o '
o
(4)
By convention radio astronomical measurements are usually made in
temperature units and one can substitute the radiation temperature
2 2
[J(Tg) = c Iy/2kv ] in the preceding equation. The brightness tempera
ture is found from
I
v
= Bv(Tb) = 2hv3/c2(e
hv/kTg
- 1)1
(5)
Obviously B is the Planck function and Tg is the temperature of the
blackbody that subtends the same solid angle and emits the same specific
intensity at v. If the Rayleigh-Jeans limit (hv/kTg 1) applies then
J(Tg) Tg. This is not always true for CO measurements.
The equation for the excess line temperature is then (Zuckerman
and Palmer 1974)
AJ(Tle) = J(Tl) J(TC) = [J(Tgx) J(TC)](1 e Tv) (6)
where T^ is the line brightness temperature, T^ is the brightness tem
perature of the continuum, and T is the excitation temperature of
CA
the molecule. Assuming local thermodynamic equilibrium (LTE) gives


-81-
the possibility that it is an E5.5 (the flattest elliptical according
to de Vaucouleurs 1977). Of course the system may be an E3 seen
face-on, but the velocity gradient for the three measured points strongly
suggests rotation.
The details of the determination of the inclination angle (i,
defined to be the angle between the plane of the galaxy and the plane
of the sky) and its effects on other parameters are found in Appendix
III. Under the assumption that the system as is an E3.6 or an E5.5 the
inclination angle is 53.1 or 68.0, respectively.
The azimuthal angle, 6', can be deprojected to find the true azi
muthal angle, e. The results are contained in Table 10 for the two
planetary nebulae. Since DP-I lies on the minor axis we assume it has
no radial velocity due to rotation; consequently we take its radial
velocity, -175 km/sec, as the systemic velocity for the galaxy.
Table 10
Values for the Observed and True Azimuthal Angles
Object
e1
(degrees)
0 (degrees)
i = 53.1
i = 68.0
185-1
26.8
40.1
53.4
185-2
44.0
58.1
68.8
The true circular velocity for the nebulae can thus be calculated from
vc = vr/,cos 9 S1'n 1
(6)


-33-
Once outside the dewar the 4750 MHz first IF is amplified again
by room temperature transistor amplifiers. The signal is heterodyned
further at this point to 1328 MHz.
A third and final mixer produces a 150 MHz third IF which is
detected by two banks of 256 square-law detectors. One has a choice of
filter widths for the filter banks; our observations always used either
the 0.5 or 1.0 MHz (per channel) banks, corresponding to 1.3 and 2.6
km/sec velocity resolution, respectively. The December 1977 and July
1978 observations were made with only a single polarization operating, so
the separate channels were actually the same signal and could not be
averaged to reduce noise. Consequently the 0.5 MHz (1.3 km/sec) and 1.0
MHz (2.6 km/sec) filter banks were both used. At the time of the Decem
ber 1979 observations the receiver operated in a proper two-channel mode
with two independent IFs. Both were fed to 1 MHz filterbanks since they
could now be added with a resultant decrease in noise.
Calibration was accomplished by alternately observing blank sky
and an ambient temperature microwave absorber. Details of the cali
bration procedure are explained in Appendix I.
As a result of the calibration procedure one obtains the corrected
antenna temperature of the source
T* = nf(l e"T)[J(vs,TE) J(vs,Tbg)] (1)
where the explanations for the symbols are found in Appendix I. Note
that there are three unknowns (nf, x, T^.) related to the particular
source being observed; these cannot be determined without additional
information. For extragalactic work one can make estimates for
x and T^ based on galactic studies and then calculate ty. The


-87-
While many of the arguments about continuing star formation in
NGC 205 can be advanced with much the same reasoning as those for NGC
185 in the previous section, much of this support is lost because we
cannot be sure the NGC 205 clouds are in a state of collapse. However,
it is very likely that they will eventually form stars. This is clear
because NGC 205 has an even larger young stellar component than NGC
185; Hodge (1973) calculates some 2 x 10 for the young stars.
We seem to be seeing an earlier stage of the process in an ellip-
5
tical galaxy which converts the ISM into stars. In a few times 10
years the dust clouds in NGC 205 will be much more condensed and CO
emission should be detectable. In the future stars will be forming and,
if these are high-mass stars, the galaxy will retain its unusual popu
lation I component.
Negative CO Results
12
The galaxies that were not detected at the CO J = 1+0 transition
are listed in Table 2. Several attributes are also listed, along with
the limits of our observations. These limits can be interpreted on the
basis of two assumptions about the CO.
First is that the gas is optically thin and the emission was not
detected because of beam dilution or an exceedingly small optical depth
(t 1). Second, if the line is optically thick then beam dilution is
the only mechanism to lower the antenna temperature below our detection
1imits.
Another variable in the interpretation of these negative results
is the width of the undetected CO line. If it is several hundred km/sec
broad, representing the contributions of many small clouds within the


-10-
dynamically independent of the remaining material. That is, the
stars reflect the velocity dispersion of the cloud at the time of
their formation.
As the material collapses further, star formation proceeds and
eventually the gas becomes dense enough that damping forces become
important. At this stage the gas gives up a great deal of kinetic
energy to dissipative heating and, because of its angular momentum,
settles down into a disk. Within this disk various processes, especi
ally star formation, continue to the present giving spiral galaxies
their distinctive optical appearance.
It takes only a minor modification of this scenario to produce an
elliptical galaxy. If star formation has proceeded to completion before
the dissipative effects can form a disk then a system will result which
has no disk component and is essentially a halo population of stars.
Models which can produce the observed luminosity profiles of ellipti
cals have been constructed (Larson and Tinsley 1974) and add credence
to this basic concept.
Within this framework of galaxy formation many details are still
to be worked out. As previously mentioned, the nature of the initial
perturbations is not at all understood. Also the processes of frag
mentation and collapse are poorly understood (Field 1975). There is
even a question of whether gravitational instabilities or turbulence
is responsible for the necessary condensations. Jones (1976) provides
an excellent review of these and other problems in the study of galaxy
formation.
Since the early star formation was apparently so efficient as to
consume all the primordial gas, one might not expect any continuing


TEMPERATURE K'l
UTS. i
¡3.0
93.00
CD
CO
CD
I
i i
CO
CD
I
' r 1 1- ^ - l - i - 1 ~
256 MHZ ( _
1 6 4 0 U M I N I n 1 j [ (j (j ,j
i
i ¡iUlJ1 lj\ A JHMi,i )H
tj]Ak jlLj .(i) ,:
U,/
L
If
If
W| v* njf/'f \h'
f r
V <
tflvh '
s' 1 II
i i.. i
RA B H 51
DEC 58 56 j
RMS = 41.69
i- i ~
41
:'3
H K E L.'
/1 M
63.
73.00 03.00 9 3. 0 0 108.0 0
VELOCITY (KM/SEC) CXI O1 J
113.
I
CD
03
I
Figure 12. The 12C0 spectrum for NGC 2685.


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.
/jjcfi}vw ft] rj:c
m\ 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.
(L¡4 U HAn. ~ T- e
(/
t)r. Thomas D. Carr
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.
vx vv
Dr. Kwan-Yu Chen
Professor of Astronomy


-78-
that the HII regions associated with massi.ve stars have since dis
sipated.
c
Considering that typical lifetimes for HII regions are 10 years
and Hodge (1973) estimates the age of the OB complex in NGC 205 (very
similar to that of NGC 185) to be 5 x 10 years, it is certainly
plausible that the early HII regions have evaporated. Since the time
5
for the observed clouds to collapse to stars is about 10 years, it
seems that the system may be experiencing bursts of star formation
similar to that envisioned by van den Bergh (1975).
However, there is a serious flaw in this scenario. Hodge (1963)
estimates that the population I stellar component of NGC 185 is about
5
2 x 10 M|g. Using Faber and Gallagher's (1976) mass loss rate of
0.015 yr-1 (io\0)_1 and the luminosity of NGC 185 (2.67 x 107 L ),
O
we find that it took 5 x 10 yrs for the material to collect. During
6 3
the 5 x 10 years since the last star-forming episode only 2 x 10 M0
of gas should have been able to appear. Our CO observations show at
least 5 x 10 M0 (for optically thin CO, assuming atomic hydrogen for
the colliding particles) and it is likely that there is actually more.
An even stronger case is made if the total mass of dust (215 Mq)
is multiplied by an assumed gas-to-dust ratio of 200. The system is
4
seen to contain 4.3 x 10 M0 of gas, more than 10 times the amount
from stellar mass loss.
The second possibility that may be occurring in NGC 185 is con
tinual star formation but with no 0 or B stars, and consequently no HII
regions. It is very difficult to prove that this is the process taking
place. If indeed low-mass stars are being formed they will be too
faint to see and deductive reasoning is necessary to support this


-73-
Table 5
Expected Antenna Temperatures (K) for x 1
Source
Tex = 10 K
25 K
100 K
CO
o
11
Case A
0.071 K
0.179
0.714
7.14
Case B
0.235
0.588
2.35
23.5
Case C
0.951
2.38
9.51
95.1
Another approach is to compare the dust clouds observed in NGC 185
to clouds more easily observed in our own galaxy. The disadvantage
is that it is not certain that the nature of the clouds will be the
same. It does, however, provide a reference point for their study.
The extinction in the dust clouds was measured by Hodge (1963).
He finds OP-I to have a mean visual absorption of 0.3 mag but as high
as 1 mag in certain parts. The mean for DP-II is measured as 0.15 mag
of absorption in the visual.
At this level of absorption the clouds are similar to the diffuse
clouds studied by Knapp and Jura (1976). Their study involved clouds
situated in front of stellar sources that show color excesses of
E(B V) £ 0.3 mag. Assuming a normal Av/E(B V) = 3 indicates
A 'v 1 mag, similar to the absorption in the NGC 185 clouds.
12
Knapp and Jura suspect optically thin CO emission based on their
13
inability to observe CO at an intensity of 2-6 times less than the
12 12
CO (usually possible if CO is optically thick), and they found
antenna temperatures of 1-2 K even though the sources probably filled
the beam and had kinetic temperatures of 20 K or more. Using the thin


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.
Associate Professor of Physics
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.
This dissertation was submitted to the Graduate Faculty of the Depart
ment 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.
August 1980
Dean, Graduate School


TEMPERA TURE CK)
.10 -0.05 0.00
(XI o1
)
2-58.00 -48.00 -38.00 -28.00 -18.00 -8.00 2.00 12
O
Ln
r~i
O
O
.bo
12
The CO spectrum obtained in December 1979 on dust patch no. 12 of Hodqe (1973)
in NGC 205.
Figure 9.
-0.05 0.00


-101-
is the effective radiation temperature of a blackbody of temperature T
at the frequency v.
Note that the receiver amplifies not only the signal sideband but
also the image sideband. The image sideband could be filtered out
before the signal is amplified but this invariably involves the intro
duction of more noise at a crucial and sensitive point in the system.
Since this work is concerned primarily with detection and re
quires the most sensitive arrangement possible, the image rejection
filter was not used. Consequently the following equations include
terms for the image sideband as well as the signal sideband.
When the telescope is pointed at blank sky the antenna tempera
ture is
Tsky GsKJ + (1 V J(VTsbr)]
+ similar terms for the image sideband (3)
where is the telescope efficiency considering spillover, blockage,
and ohmic losses. The T ^ is the apparent brightness temperature
from the spillover, blockage, and ohmic losses. The Tg and T^ (in the
image sideband terms) are the brightness temperatures of the sky at the
signal and image frequencies, respectively. They are given by
J(vsV = J^vs,Tm^1 expKA)] + vsTbg)exp(TsA)
and a similar image sideband equation. The Tm is the mean atmospheric
temperature, T^ is the brightness temperature of the cosmic background
radiation, and and t. are the atmospheric zenith optical depths at
the signal and image frequencies, A is the air mass in the observed
direction.


-86-
3.2 x 10^ and 9.2 x 10^ Mq for clouds 11 and 12, respective
ly.
Table 15
Maximum Column Densities of CO and H^ for NGC 205, t < 1
Cloud no.
N(C0]
(cm_2)
n(h2)
(cm-2)
M (molecular)
11
15
7.0 x 10iD
1.4 x 1020
4.4 x 103
12
2.8 x 1015
1 9
5.6 x 10iy
5.2 x 103
The situation does not change if one assumes the clouds are optically
thick. The NGC 205 clouds are relatively large compared to those in
NGC 185, and any reasonable kinetic temperature will produce a large
antenna temperature, as shown in Table 5, very easily.
The question is then, why is NGC 205 not detected while NGC 185
with similar morphology (and actually less dust) is detected at the
frequency of 12C0?
We feel that the key difference in the two systems is that the
NGC 205 clouds are less dense than those in NGC 185. Hodge's (1963 and
1973) microphotometer tracings show this, and even optical photographs
of the two galaxies show the NGC 185 dust patches to be more prominent.
O O
As Knapp and Jura (1976) found, even a density of 10 cm is not
12
sufficient to ensure detection of CO in emission. Further, their
observations were made on nearby clouds where one could reasonably
expect a filled beam. The NGC 205 observations involve substantial
beam dilution which considerably worsens prospects for detection.


-110-
emission it must be elevated above TR by some process (collisions in
this case).
Thus, a minimum colliding particle density can be found by
equating xr and t This gives
n
64-rr4 1 U [ 2
o -hv/kTp
3hXJ(l e K)
(l
A
-hv/kTp
- e R)
(18)
The usual approach to determine is to use the geometric cross-
section (% 10 ^ cm^) and v = 8kTR/irm2 v 10^ cm/sec for T^ = 100 K and
m = m^). For CO this gives a minimum colliding particle density of
8.6 x 10^ crrf^.


-22-
observations Dressier (1980) concludes that spirals do not evolve
into SOs. He concludes that the galaxy types are affected more by the
initial conditions at the time of their formation than by environmental
factors such as ram-jet stripping.


-42-
of velocity resolution. Figure 3 is the result of this convolution,
and indeed the reality of the feature is strongly supported.
Table 1
Line Parameters
for 12C0 in
NGC 185
Peak Antenna Temperature (K)
0.081
0.025
Velocity of Line (heliocentric, km/sec)
-175 2
>
Full-width at Half-maximum (FWHM)
(km/sec)
12 2
Position Observed
0h 36m
10?3
48 3'
5 33
Total Integration Time on Source
(min)
484.5
Other support for the reality of the line is found by careful
examination of the spectra obtained during each session of observing
(Figures 4 to 6). The spectral characteristics vary somewhat for each
session. But this is to be expected since each individual spectrum
has a relatively low SN ratio. In particular the line strengths for
each session are
December 1977 0.20 0.07 K
July 1978 0.20 0.08 K
December 1979 0.075 0.028 K
where the quoted errors are 1 standard deviation. In view of the
agreement here, and the pointing uncertainty during the December 1979
observations, the discrepancy among the observations is not large.
During the December 1979 observations the cooled receiver operated
with two independent channels, linearly polarized and orthogonal to
one another. The feature does not appear in the spectrum from the


-17-
Faber and Gallagher (1976) have considered the problem of a
galactic wind from a slightly different approach. Since the initial
heating of the ISM is through collision of the gas clouds shed by
stars in the process of normal evolution, the velocity dispersion of
the stars can be used as an indicator of the stellar equivalent temper-
ature, T From Mathews and Baker (1971) T = v mu/6k where v is
S S S n S
the stellar dispersion velocity and m^ is the mass of the proton, and k
is Boltzmann's constant. Along with the condition that the supernova-
induced rise in temperature must roughly double the kinetic energy of
the gas to remove it from the system to infinity, one finds
"sn ^sn
C£
- asTs
(1)
where agn and ag are the specific rates of mass injection by super
novae and stars, respectively, Tsn and Ts are the equivalent tempera
tures of the supernovae and the stars.
Using the values quoted in Faber and Gallagher (1976) a =
5 x IQ"20 sec"1 and a T = 1.6 x 10"12 K sec'1 yields T = 3.7 x 107
sn sn J s
K or v$ < 1260 km/sec. The line-of-sight velocity dispersion is
< 1260/v^S = 730 km/sec (assuming 3 dimensional isotropy of the velocity
dispersion), which exceeds by a factor of two the largest velocity
dispersions measured (Faber and Jackson 1975).
This calculation would make it appear that virtually all elliptical
systems should have a galactic wind scouring the ISM of any material.
However, Gisler (1976) recomputes this same quantity, taking into
account the 0.1 efficiency of energy coupling found by Larson (1974),
and finds that the maximum velocity dispersion that will still allow
the supernovae to double the energy is about 200 km/sec, a number
comparable to the observed velocity dispersions. In other words,


Knapp, G.R., Faber, S.M., and Gallagher, J.S. 1978a, A. J., 83, 11.
Knapp, G.R., Gallagher, J.S.., and Faber, S.M. 1978b, A. J., 83, 139.
Knapp, G.R. and Jura, M. 1976, Ap. J., 209, 782.
Knapp, G.R., Kerr, F.J., and Williams, B.A. 1978c, Ap. J., 222, 800.
Lang, K.R. 1974, "Astrophysical Formulae" (Spr.inger-Verlag: New
York), 161.
Larson, R.B. 1974, M.N.R.A.S., 169, 229.
Larson, R.B. and Tinsley, B.M. 1974, Ap. J., 192, 293.
Larson, R.B. and Tinsley, B.M. 1978, Ap. J., 219, 46.
Lea, S.M. and De Young, D.S. 1976, Ap. J., 210, 647.
Lo, K.Y. and Sargent, W.L.W. 1979, Ap. J., 22J_, 756.
Lucas, R. 1974, Astron. & Astrophys., 36., 465.
Lynden-Bell, D. 1967, M.N.R.A.S., 136, 101.
Martin, R.N. and Barrett, A.H. 1978, Ap. J. Suppl., 3(5, 1.
Mathews, W.G. and Baker, J.C. 1971, Ap. J., 170, 241.
Mathewson, D.S., Cleary, M.N., and Murray, J.D. 1974, Ap. J., 190, 291.
Mathewson, D.S., Cleary, M.N., and Murray, J.D. 1975, Ap. J. (Letters),
195, L97.
McHardy, I.M. 1974, M.N.R.A.S., 169, 522.
Mezger, P.G. and Smith, L.F. 1977, in "Star Formation, I.A.U. Symp.
No. 75," ed. T. de Jong and A. Maeder (D. Reidel: Dordrecht),
133.
Morgan, W.W. 1958, Pub. A.S.P., 70, 364.
Morgan, W.W, 1959, Pub. A.S.P., 71, 394.
Morton, D.C. 1975, Ap. J., 197_, 85.
Oemler, A. 1974, Ap. J., 194, 1.
Oemler, A. 1977, Highlights of Astronomy, 4, 253.
Osterbrock, D. 1960, Ap. J., 132, 325.


-30-
early-type systems must be very efficient; if so, it may be genericaliy
related to the star formation that occurred as the galaxy collapsed
originally. Judging from the lack of primordial material observed,
that process was also very efficient.
This work is also of importance in the formation and continued
activity of radio sources in early-type galaxies. From the available
information it seems that radio sources and star formation are com
peting for the same interstellar debris in a galaxy. If one is very
successful, it may be at the expense of the other.
All of this is not to say that galactic winds and ram-jet stripping
never occur, simply that all possibilities need to be analyzed and
evaluated to determine their relative importance in the overall
scheme of galaxy evolution.


-76-
case (Table 6) and the minimum calculated gas mass from Table 7 allows
a calculation of the gas-to-dust (G/D) ratio. The results appear in
Table 8.
Table 8
Gas-to-Dust Ratios
Cloud
Dust Mass
Gas Mass
* < 1 (M0)
G/D
Gas
T
Mass
1
G/D
DP-I
25
9680
390
5.1 x
105
2 x 104
DP-I+II
215
9680
45
4.5 x
106
2 x 104
Considering that the gas-to-dust ratio in our own galaxy is usually
quoted as 200, the most attractive source for the CO emission appears
to be DP-1.
Using the minimum mass consistent for DP-Ia (Table 6) allows a
calculation of the Jeans, length for gravitational instability
irkT
-4G
PP
= i.6[-
TKR£c]i
M J

(3)
where Mq is the mass of the cloud in solar masses. Taking the width
of the cloud as a characteristic dimension, R = 6.6 pc, T^, = Tgx =
20 K, and M = 4.84 x 103 H
0
Aj = 1.1 pc
Since the cloud is substantially larger than the Jeans length,it is
most likely in a state of collapse.


TEMPERATURE (K 3
(X101 )
Figure 18. The ^CO spectrum for NGC 5866.


-84-
While these values for the MLR are very reasonable for early-type
galaxies in general (Faber and Gallagher 1979), there is a significant
disagreement with the MLR of 1.8 used by Ford et al. (1977) for NGC 185
and NGC 147. This MLR is ostensibly derived from observations of M32.
We feel that the most likely cause for the discrepancy is that Ford
et al. (1977) do not consider possible rotation in their analysis.
The radial velocity of DP-I is most likely nearest the systemic velocity
of NGC 185 since it is fortuitously on the minor axis.
We find that the MLR of the nuclear regions of NGC 185 is between
5 and 18. Considering the non-linear effects of the inclincation and
azimuthal angle, the most likely value for the MLR is around 8.
Assuming that it is independent of radius, the total mass of NGC 185
is found to be 1.3 x 10^ M .
9
NGC 205
Another dwarf elliptical companion of M31, NGC 105, was also observed
to sensitive limits for CO emission, with negative results. Figures 8 and
9 show the averaged spectra from the July 1978 and December 1979 sessions,
respectively. The spectra are taken at two different positions within
the galaxy (dust patches 11 and 12 in Hodge's (1973) notation). Hodge
(1973) describes and diagrams the dust content of NGC 205. The distri
bution is reminiscent of NGC 185 but with a larger number of discrete
clouds over a wider region in the nucleus. The July 1978 observations


-29-
and collapse. Indeed, he finds that the critical mass for collapse under
such conditions is less than 10 M Because of fragmentation the
result of the collapse of a 10 Mq cloud will almost certainly be a
number of small mass stars rather than one 10 M star.

Thus while no compelling reason can be advanced to accept star
formation with a skewed IMF, the possibility cannot be rejected either.
The main intention of this dissertation as outlined in the following
section is to examine this possibility to determine if it is a sig
nificant process in the evolution of early-type galaxies.
Statement of the Dissertation Problem
From the previous sections it is clear that our understanding of
early-type galaxies is incomplete. There are a large number of inter
esting ideas concerning their formation and evolution, but as usual, in
sufficient data are available to assess adequately their importance.
The primary goal of the observations presented and discussed in this
dissertation is to investigate the significance of star formation in the
evolution of early-type galaxies. The results of this study are most
directly applicable to the recycling of material by star formation in
an environment much different than our own galaxy. The results also
bear on other settings in which the star formation is poorly under
stood.
It is not overstating the current situation to say that we only
have a dim view of how star formation occurs in our own galaxy, and an
even dimmer view of the process in drasically different environments.
Faber and Gallagher (1976) argue that possible star formation in


-32-
The cooled mixer Cassegrain receiver operating at the 115.2712 GHz
assumed rest frequency of the J = l->0 CO transition was used for all
observations. The halfpower beam width at this frequency is about
65".
Pointing for the telescope is initially determined by the observa
tory personnel. Pointing correction data are taken in all parts of the
sky on bright mm point sources. Analytic functions are then fitted to
the data, interpolated over regions with few data points. The observer
then checks these corrections by making five-point maps of bright sources,
usually the planets. Rarely are the corrections more than 5 arcseconds
at this stage.
However, the 1979 observations were hindered by large errors in
regions far from the celestial equator due to a lack of adequate data
for the fitting equations. This problem is discussed further in
Chapter III.
Nominally, the sky signal is separated into two linearly polarized
orthogonal components. In practice the system operated with only one
polarization, except during the 1979 observations when both channels
were available. These two signals are then fed to two Shottky-barrier
mixer diodes. Also fed to the mixer is the local oscillator (LO)
signal via a tunable injection cavity. This process of mixing two
frequencies to get a third (usually the difference of the two) is
known as heterodyning. The first intermediate frequency (IF) signal
emerges at 4750 MHz and is amplified by a pair of low-noise parametric
amplifiers. All of the previous equipment is enclosed in a dewar
cooled to 15 K.


-27-
the infall is connected with the peculiar structure of M82 and its
vigorous star formation.
On a gentler scale there are several examples of star formation
in dwarf ellipticals in the local group. NGC 185 and 205 both contain
dust patches and a sprinkling of blue, presumably young stars. The CO
observations described in Chapter III show that the star formation
is probably continuing in NGC 185. NGC 205 is a close companion of M31
and it may be argued that an interaction is taking place, although HI
maps of the region do not support this idea. Regardless, NGC 185 is
considerably further from M31 and an interaction does not appear likely
(see the HI observations presented in Chapter III concerning this
possibility).
Consequently, it appears that these two dwarf ellipticals are
reprocessing their ISM back into stars. The most likely source of
the material is the normal products of evolution of stars. Large scale
interactions are apparently not supported by the HI observations of
the galaxies.
The more sedate pace of star formation, and various arguments
suggesting an initial mass function for the stars shifted to lower
masses, are examined in the following section.
Continual Star Formation with a Skewed Initial Mass Function
A third fate that may befall gas retained in early-type systems
is star formation. The most stringent requirements on the nature of
this star formation are set by the observations of the colors or early-
type galaxies (see Larson and Tinsley 1978 for a discussion and earlier
references).


(XI O1 )


Table 3
HI Observations in the Vicinity of N6C 185
ON OFF Full-width Time
Posi
tion of
observat'
ion
observation
Standard
Velocity
Half-maximum
on
0!\
1 wrt
Temperature Deviation
(km/sec)
(km/sec)
Source
NGC
: 185
a(1950.0)
6(1
L950.0)
a
(1950.0)
6(1950.0)
(mK)
(mK)
5 km/sec
3 K/sec
(min)
NGC
185a
OO1
h36
*US.3
48
303
'45"
oo1
n19m43?9
48
5031
44"
14.4
2.0
-190
63
378
49'
Northb
00
36
11.4
48
52
38
00
19
27.8
48
03"
44
15.7
2.3
-175
35
360
NGC
185c
00
36
11.4
48
03
44
00
31
19.6
48
03
44
6.1
2.0
-200
40
344
00
33
45.0
48
03
45
NGC
185
00
41
3.2
48
03
44
6.3
1.9
-200
45
344
00
38
37.7
48
03
44
NGC
185
00
36
11.4
47
14
50
6.3
1.9
-200
45
344
00
36
11.4
47
39
17
NGC
185
00
36
11.5
48
52
38
<5
2.6
--
--
344
00
36
11.4
48
28
12
120'
North
00
36
11.3
50
03
44
00
41
16.1
50
03
44
-17.3
5.1
-200
38
60
00
36
11.4
49
14
50
-34.6
4.7
-180
33
60
71'
North
00
36
11.4
49
14
50
00
19
20
48
03
44
39.8
8.6
-180
33
13
Notes: These measurements are from spectra reduced using a third order baseline and hanning smoothed. The
spectra on NGC 185 with two "off" positions are from data taken at both "off" positions. Approxi
mately 27% of the integration time was with the first "off," the balance with the second "off."
Reproduced as Figure 19.
^Reproduced as Figure 21.
Q
The sum of this and the three spectra following is reproduced as Figure 20.


-119-
Sunyaev, R.A. 1977, in "Radio Astronomy and Cosmology, I.A.U. Symp.
No. 74," ed. D.L. Jauncey (D. Reidel: Dordrecht).
Tammann, G.A. 1974, in "Supernovae and Supernovae Remnants," ed.
C.B. Cosmovici (D. Reidel: Dordrecht).
Tartar, J. 1975, Ph.D. Dissertation, University of California, Berkeley.
Tytler, D. and Vidal, N. 1978, M.N.R.A.S., 182, 33p.
Ulich, B.L. and Haas, R.W. 1976, Ap. J. Suppl., 30, 247.
Unwin, S.C. 1980, M.N.R.A.S., 190, 551.
van den Bergh, S. 1960a, Ap. J., 131, 215.
van den Bergh, S. 1960b, Ap. J., 131, 558.
van den Bergh, S. 1975, Ann. Rev. Astron. & Astrophys., 13, 217.
van den Bergh, S. 1976a, Ap. J., 208, 673.
van den Bergh, S. 1976b, Ap. J., 206, 883.
van den Bergh, S. 1976c, A. J., 81, 797.
White, S. 1976, M.N.R.A.S., 174, 19.
Wright, M.C.H. 1974, in "Galactic and Extra-Galactic Radio Astronomy,"
ed. G.IL. Verschuur and K.I. Kellerman (Springer-Verlag: New York),
291.
Young, P.J., Boksenberg, A., Lynds, C.R., and Hartwick, F.D.A. 1978a,
Ap. J., 222, 450.
Young, P.J., Sargent, W.L.W., Kristian, J., and Westphal, J.A. 1979,
Ap. J., 234, 76.
Young, P.J., Westphal, J.A., Kristian, J., Wilson, C.P., and Landauer,
F.P. 1978b, Ap. J., 221, 721.
Zuckerman, B. and Palmer, P. 1974, Ann. Rev. Astron. & Astrophys.,
12, 279.


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 STUDY OF THE INTERSTELLAR MEDIUM IN NGC 185
AND OTHER EARLY-TYPE GALAXIES
By
Douglas William Johnson
August 1980
Chairman: Stephen T. Gottesman
Major Department: Astronomy
The question of an interstellar medium in early-type galaxies is
considered in light of the small amounts of gas detected as neutral
hydrogen (HI). It is apparent that there is some method of removal or
reprocessing that keeps the interstellar medium of early-type systems
gas and dust free in spite of mass loss from normal stellar evolution.
12
A detection of CO is presented for the dwarf elliptical system
NGC 185. The mechanisms of line formation of the J = 1-K) transition
strongly imply that the emitting region is in a state of gravitational
collapse. These observations are consistent with the observed dust
content of the galaxy and its blue, presumably young stellar population.
It is nearly certain that the galaxy is reprocessing its interstellar
medium via star formation.
The radial velocity of the CO cloud can be combined with the
velocity data for two planetary nebulae within the system to allow
rough mass calculations. Coupling this information with luminosity
data indicates that the mass-to-luminosity ratio of the nuclear regions
(within 1 arcminute) is between 5 and 18, with a value near 8 being the
most probable.
vii


12
Figure 2. The CO spectrum for the nucleus of NGC 185. Data from all three
observing sessions have been averaged for this spectrum. The
total bandwidth for the observation appears in the upper left
along with the total integration time on the source. In the lower
right the position (1950.0 coordinates) observed is given and the
standard deviation for the spectrum is also shown. The baseline
at 0.0 Kelvin is shown along with horizontal lines at 3 standard
deviations.


-109-
Another line of reasoning can give space densities for a molecule
if a few conditions are satisfied (see Rank et al. 1971 for a dis
cussion). If a molecule is observed in emission, well-removed from
discrete sources of radiation, it is likely that collisions raise T
t. A
well above TR (= 2.7 K, the cosmic background radiation).
The relaxation rate between two molecular states for isotropic
radiation is (Rank et al. 1971)
4 3
64tt v
3hcJ(l e
-hv/kT
R>
(15)
The collision rate can be written
= l (nov) (16)
t L 'm v '
c m
where the summation is over all colliding species, n is the density of
each species, a is the effective cross-section, and v is the average
relative velocity for the collisions. In most clouds the only par
ticles abundant enough to contribute to equation (16) are atomic and
molecular hydrogen.
Under these conditions the effective temperature of the tradi
tion is
r = yTm + tcTR
eff t + t
r c
(17)
where Tm is the temperature of the colliding particles. Since the
molecule is seen in emission it follows that xc is not much larger
than t The implication is that in order to see the molecule in


-112-
y
Figure 24. A disk system viewed face-on (i = 0). The x axis coincides
with the major axis and the y axis is also the minor axis.
The circular velocity, v and its x and y components are
indicated. Also shown is the azimuthal angle, e.


APPENDIX I
CALIBRATION THEORY FOR CO OBSERVATIONS
The intent of the calibration procedure is to correct the data
for several factors that distinguish a real telescope at the earth's
surface from an ideal telescope at the top of the atmosphere. The
major correction factors are ohmic losses in the telescope, spillover,
blockage, and the radiation/attenuation of the atmosphere.
The following paragraphs describe the calibration procedure used
at the 11m NRAO telescope as described by Ulich and Haas (1976).
The method of choice to calibrate mm-wavelength observations is
to use a rotating chopper which alternately covers the feed horn with
an ambient temperature microwave absorber. The method is attractive
because it automatically compensates for changes in atmospheric ab
sorption (Penzias and Burrus 1973).
The ambient temperature absorber is placed over the feed horn
aperture and the antenna temperature measured is
load
^(vW + M^.Tw,)
i v i amb'
(1)
where G is the gain in the receiver, T ^ is the ambient temperature
of the absorber, the subscripts s and i indicate the signal and image
sidebands, respectively, and
hv/k
exp(hv/kT) 1
J(v,T) =
(2)


(XI O1 )
12
Figure 17. The CO spectrum for NGC 5846. De Vaucouleurs and de Vaucouleurs (1964) list
a radial velocity of 1771 km/sec, but de Vaucouleurs et al. (1976) list it as
2353 km/sec.


-35-
At this stage of the reduction process each point of the spectrum
is examined to see if it is greater than five times the standard
deviation for that spectrum. If it is, and the adjacent points are not,
the point is replaced with the mean of the two adjacent points. This
procedure is used to remove interference that involves only a single
channel; multichannel filter banks are particularly susceptible to
this type of interference.
Each spectrum is then weighted by the inverse square of its
standard deviation and combined with all other spectra on the same
source.
A preliminary first order baseline is fit and subtracted and the
resulting spectrum is examined for possible features. If any are
apparent the channels involved are eliminated from the baseline fitting
procedure and a new first order baseline is calculated and subtracted.
The resulting spectrum represents the best data on a given source.
Various smoothing functions can be applied; the most common for this
work has been smoothing with a rectangular function of about the same
width as the suspected spectral feature. The effect of this is to
maximize the signal-to-noise ratio at the expense of velocity resolu
tion.
The basic format of each spectrum presented is explained in
Figure 2.
Data Presentation
The data are displayed with antenna temperature as the ordinate
and velocity along the abscissa. Strictly speaking the abscissa


-26-
the equatorial band of dust and gas. This, is difficult to reconcile
with the interstellar material originating with the stars.
Most other researchers in the field tend toward the explanation that
the system may actually be a collision between a gas cloud or small
galaxy and an elliptical (Graham 1979). Thus the relevance of NGC
5128 to elliptical galaxy evolution cannot be assessed until these
questions are answered.
Van den Bergh (1975) has also suggested that NGC 1275 is an ex
ample of an elliptical galaxy caught in a burst of star formation. The
underlying stellar distribution in the system is elliptical and the
galaxy is located near the center of the Perseus cluster. The small
velocity difference between the cluster's mean velocity and that of the
galaxy again suggests inefficient ram-jet stripping.
However, more recent work on the system (Kent and Sargent 1979,
Rubin et al. 1977) indicate that this is a collision between a fore
ground spiral system and NGC 1275 in progress. Thus both NGC 5128 and
1275 may be atypical as far as early-type galaxies are concerned.
However, these two examples support the view that a massive (a- 10^
M ) influx of material over a relatively short time period results in
vigorous star formation. As both these systems are also radio sources,
one could argue that another effect of the mass accretion is to
either form, or fuel a previously existing, massive object in the
nucleus, producing the radio source.
Further support for this hypothesis, although not in an early-type
system, may be found in the M81-82 system. The interaction between
these galaxies seems to be resulting in a substantial mass infall to
M82 (Gottesman and Weliachew 1977, Killian 1978). It is likely that


-2-
Galaxy Classifying Schemes
Early studies of galaxies made it clear that the general morphology
of these immense stellar systems allowed them to be grouped into a
relatively small number of types. The most successful early venture
was Hubble's "tuning fork" diagram published in 1936, and reproduced
here in Figure 1. The spherical systems (EO) are at one end with pro
gressively flatter (E1-E7) systems leading up to a split in the diagram.
SO galaxies (also called lenticular galaxies) occupy the vertex of the
fork because Hubble believed that they were transitional systems. They
contain prominent elliptical bulges as well as a conspicuous disk
component. It is interesting to note that the nature of SO galaxies
is still being vigorously debated.
Figure 1. Hubble's "tuning fork" diagram of the classification of
galaxies.


TEMPERATURE (K)
-0.20 -0.12 -O.O14 0.04
(X ID1 )
33-58.00 -ts.no -38.00 -28.00 -18.00 -8.00 2.00 12330
j ij m IJ IJ C. O U U J. 1 1 M |J
VELOCITY (KM/SEC) (X101 )
Figure 7. The ^C0 data from December 1979s channel B.


-118-
Osterbrock, D. 1962, in "Interstellar Matter in Galaxies" (Benjamin:
New York), 111.
Ostriker, J.P. 1977, in "The Evolution of Galaxies and Stellar Popula
tions," ed. B.M. Tinsley and R.B. Larson (Yale University Observa
tory: New Haven), 369.
Ostriker, J.P. and Hausman, M.A. 1977, Ap. J. (Letters), 217, L125.
Ostriker, J.P. and Tremaine, S.D. 1975, Ap. J. (Letters), 202, LI13.
Penzias, A.A. and Burrus, C.A. 1973, Ann. Rev. Astron. & Astrophys.,
11, 51.
Peterson, C.J. 1978, Ap. J., 222, 81.
Rank, D.M., Townes, C.H., and Welch, W.J. 1971, Science, 174, 1083.
Reif, K., Mebold, U., and Goss, W.M. 1978, Astron. & Astrophys., 6)7, LI.
Rickard, L.J., Palmer, P., Morris, M., Turner, B.E., and Zuckerman, B.
1977, Ap. J., 213, 673.
Riley, J.M. 1975, M.N.R.A.S., 170, 53.
Roberts, M.S. and Steigerwald, D.G. 1977, Ap. J., 217, 883.
Rood, H. 1979, Ap. J., 232, 699.
Rowan-Robinson, M. 1979, Ap. J., 234, 111.
Rubin, V.C., Ford, W.K., Peterson, C.J., and Oort, J.H. 1977, Ap. J.,
211, 693.
Sandage, A.R. 1961, "The Hubble Atlas of Galaxies" (Carnegie Inst, of
Washington: Washington).
Sargent, W.L.W., Young, P.J., Boksenberg, A., Shortridge, K., Lynds,
C.R., and Hartwick, F.D.A. 1978, Ap. J., 221, 731.
Schechter, P.L. and Gunn, J.E. 1979, Ap. J., 229, 472.
Schneps, M.H., Ho, P.T.P., Barrett, A.H., Buxton, R.B., and Myers,
P.C. 1978, Ap. J., 225, 808.
Shane, W.W. 1980, Astron. & Astrophys., 82, 314.
Shostak, G.S. 1978, Astron. & Astrophys., 54^, 919.
Silk, J. and Norman, C. 1979, Ap. J., 234, 86.
Smoot, GF. and Lubin, P.M. 1979, Ap. J. ILetters), 234, L83.


TEMPERATURE CK) (
(X101 )
-100.00 -80.00 -60.00 -40.00 -20.00 0.00 20.00 40 0)0
1 1 1 1
10 MHZ
1 3 6 6 0 0 M I N h
1
A.
1 1 i
4A
n
i 1
r i q a
L- 1 ij vj
. 7 S 1(1 IhiiOH
1 s
AAA A A.. a, a \ ,
/ v \ / v v v V \ / 1
l' ¡1
1
30 O TO
m
AO
re
ts
1 0 11 36M 11? 3
M 48 3 1 4426
= 3.18 HKELV1N
I 1 .
00.00 -80.00 -60.00 -40.00 -20.00 0.00 20.00 40. llO
VELOCITY (KM/SEC) (X101 )
Figure 20. The HI spectrum of NGC 185 with offs 1-2 beam widths to the west, east, south
and north.


-92-
It is interesting to note that the three galaxies with a detected
HI content larger than expected (NGC 1052, 2685, and 4278) are all
suspected of accretion. All three systems show peculiar dynamics for
the HI that make an extra-galactic source the most plausible explanation
(Reif et al. 1978, Shane 1980, Knapp et al. 1978c).


-li
star formation in elliptical galaxies. It will be shown, however,
that other processes operate which alter this simple picture and lead
to radically different expectations. The most important effects are
mass loss from the normal evolutionary processes of the stellar popula
tion and perhaps accretion of extragalatic material; both are discussed
in the following sections.
Mass Accretion
The contention that early-type galaxies, ellipticals in particular,
accrete material is relatively new. The motivation is to explain the
cD galaxies (giant ellipticals usually located in the center of rich
clusters and described by Bautz and Morgan 1970) that are often strong
radio sources. The accretion described here is full scale cannibalism
of other galaxies during close encounters (Ostriker and Tremaine 1975,
White 1976, Ostriker and Hausman 1977, Hausman and Ostriker 1978).
The idea can be linked to cluster types as described by Bautz and
Morgan (1970) and Oemler (1974). This cluster classification scheme
ranks the clusters based on their richness. Observationally it is
found that the densest clusters (Type I in the notation of Bautz and
Morgan 1970) often contain giant elliptical galaxies near their center.
Further, the cD galaxies are often radio sources that are widely
suspected to be caused by material falling into a massive object.
There are several other pieces of circumstantial evidence that indi
cate that this process may indeed be significant in the evolution of
elliptical galaxies (see Ostriker 1977 and Hausman and Ostriker 1978
for details).


Table 16
Maximum Molecular Masses Derived from CO Observations
NGC #
3xrms
(K)
T < 1,
Av = 10 km/sec
T < ]
l, Av = 200
km/sec
mhi
Detected
Mr j
Expected
N(C0)
(1016 cm-2)
n(h2)
(1020 cm"2)
Mass
T=20 K
(M )
N (CO)
(1016 cm'2)
N(H2)
(10 cm
Mass
0 T=20 K
2) (M,)
205-11
0.123
1.48
3.0
1.2
X
104
29.6
59.2
2.4
X
105

2.7xl06
205-12
0.078
0.936
1.9
7.4
X
103
18.7
37.4
1.5
X
105
3 x 105a
2.7xl06
404
0.201
2.41
4.8
9.0
X
104
48.2
96.2
1.8
X
106
--
4.2xl06
1052
0.207
2.48
5.0
5.0
X
106
49.6
99.2
1.0
X
108
1.Ixl09b
2.72X108
2685
0.126
1.51
3.0
3.6
X
106
30.2
60.4
7.2
X
107
9.5x108c
1.32x10s
3226
0.186
2.23
4.5
1.0
X
107
44.6
89.2
2.0
X
108

2.02x10s
4150
0.129
1.55
3.1
3.7
X
105
31.0
62.0
7.4
X
106

1.2xl07
4278
0.171
2.05
4.1
3.2
X
106
41.0
82.0
6.4
X
107
2.5xl08d
1.88xl08
4636
0.267
3.20
6.4
8.3
X
106
64.0
128
1.7
X
108
3.2xl08e
6.69xl08
5846
0.420
5.04
10.1
2.3
X
107
101
202
4.6
X
108
5.0xl08f
6.12xlOS
5866
0.108
1.30
2.6
4.1
X
106
26.0
52.0
8.2
X
107
--
4.71xlOS
aUnwin (1980)
bReif et al. (1978)
c6allagher et al. (1978)
dKnapp et al. (1978c)
eBottinelli and Gouguenheim
(1977b)
SHuchtmeier et al. (1977)
-89-


Figure 19. The 21 cm HI spectrum on NGC 185. The off position is indicated in the lower
right. The features between -100 and 0 km/sec are local hydrogen.


A STUDY OF THE INTERSTELLAR MEDIUM IN NGC 185
AND OTHER EARLY-TYPE GALAXIES
By
DOUGLAS WILLIAM JOHNSON
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
1980

to Maryfran

ACKNOWLEDGEMENTS
There are many people who have contributed invaluable resources
and encouragement to me in the completion of the research for this dis
sertation. Although it is not possible to thank each one individually,
there are some I would like to acknowledge in particular.
The Northeast Regional Data Center is acknowledged for providing
a vigorous and stimulating computing environment. The operators of the
11 m NRAO telescope were very helpful in assisting me during my observing
sessions and the assistance of Telescope Engineer Rick Howard was
especially appreciated. I would also like to thank Dan McGuire for his
assistance during my first observing session.
I thank my supervisory committee members, Dr. Stephen T. Gottesman,
Dr. Thomas D. Carr, Dr. Kwan-Yu Chen, Dr. Gary Ihas, and Dr. William
Weltner, for their attention and interest in my work.
I thank the Physics Department, and especially Dr. Richard Garrett,
for the assistantships that have allowed me to pursue this course of
study.
The Division of Sponsored Research is thanked for its Seed Money
Grant Competition, providing financial support for Dr. Gottesman and me
during the course of our investigations. The Graduate School's Supple
mentary Fellowship for 1979-80 was also greatly appreciated.
Finally, I wish to thank Steve Gottesman and his family for their
efforts to make Maryfran's and my stay in Gainesville a wonderful time
iii

in our lives. Steve's assistance and discussions with me (to say
nothing of his witticisms) were above and beyond the call of duty.
I thank my parents, Bill and Mary Ann, for their support and love
and good humor throughout the years. But I reserve perhaps my sincerest
appreciation for Maryfran and the "kids" (Smokey, Phantom, Harpo, and
Marble) for making it all worthwhile.
iv

TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS iii
ABSTRACT vii
CHAPTER
IINTRODUCTION 1
The Nature of Early-Type Galaxies 1
Galaxy Classifying Schemes 2
The Morphology of Elliptical Galaxies 3
The Morphology of Lenticular (SO) Galaxies 6
The Formation and Evolution of Early-Type Galaxies ... 7
Formation 7
Mass Accretion 11
Mass Loss Due to Stellar Evolution 13
Stripping Mechanisms 15
Internally Driven Winds 15
Ram-Jet Stripping by an Intracluster Medium .... 18
Fate of Retained Gas 23
Supermassive Objects in the Nucleus 23
Cyclic Bursts of Star Formation 25
Continual Star Formation with a Skewed Initial
Mass Function 27
Statement of Dissertation Problem 29
IITHE OBSERVATIONS 31
Carbon Monoxide Observations at Kitt Peak, AZ 31
Telescope Description 31
Data Reduction Techniques 34
Data Presentation 35
Neutral Hydrogen Observations at Green Bank, WV 36
Telescope Description 36
Data Reduction Techniques 38
IIIRESULTS 39
Carbon Monoxide 39
Positive Result in NGC 185 39
Negative Results 48
Neutral Hydrogen 51
IVDISCUSSION ..... 68
NGC 185 68
NGC 205 84
v

Page
Negative CO Results 87
Neutral Hydrogen and NGC 185 93
V SUMMARY 96
APPENDIX
ICALIBRATION THEORY FOR CO OBSERVATIONS 100
IITHE PHYSICS OF CO SPECTRAL LINE CALCULATIONS 105
IIITHE GEOMETRY IN AN INCLINED DISK Ill
REFERENCES. 114
BIOGRAPHICAL SKETCH 120
vi

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 STUDY OF THE INTERSTELLAR MEDIUM IN NGC 185
AND OTHER EARLY-TYPE GALAXIES
By
Douglas William Johnson
August 1980
Chairman: Stephen T. Gottesman
Major Department: Astronomy
The question of an interstellar medium in early-type galaxies is
considered in light of the small amounts of gas detected as neutral
hydrogen (HI). It is apparent that there is some method of removal or
reprocessing that keeps the interstellar medium of early-type systems
gas and dust free in spite of mass loss from normal stellar evolution.
12
A detection of CO is presented for the dwarf elliptical system
NGC 185. The mechanisms of line formation of the J = 1-K) transition
strongly imply that the emitting region is in a state of gravitational
collapse. These observations are consistent with the observed dust
content of the galaxy and its blue, presumably young stellar population.
It is nearly certain that the galaxy is reprocessing its interstellar
medium via star formation.
The radial velocity of the CO cloud can be combined with the
velocity data for two planetary nebulae within the system to allow
rough mass calculations. Coupling this information with luminosity
data indicates that the mass-to-luminosity ratio of the nuclear regions
(within 1 arcminute) is between 5 and 18, with a value near 8 being the
most probable.
vii

Observations of 10 other early-type systems are also presented
and discussed. The negative results imply that the gas is either
clumped and thus optically thick or has been removed from the system
through a galactic wind, ram-jet stripping, or has been consumed by
star formation.
The nature of the star formation must be somewhat different than
that in our own galaxy. The high-mass end of the initial mass function
for star formation would result in bluer colors than observed, the
star formation must be skewed towards the low-mass stars to be effec
tive yet unobserved. Theoretical arguments that this is possible are
advanced, but more sensitive and highly resolved CO observations are
necessary to observe directly this scale of star formation.
Neutral hydrogen observations of N6C 185 obtained with the NRAO
43 m telescope are presented and discussed. There is apparently low-
level (20-40 mK) high-velocity hydrogen in the region of NGC 185. The
most likely source of the material is the Magellanic Stream which ter
minates in this area. Superposed on this high-velocity material, at
the same velocity, is an excess of about 5 mK of HI at the location of
NGC 185. This was detected by using "off" spectra 1-2 beamwidths from
the galaxy in the four cardinal directions. The observations cannot
distinguish between an enhanced high-velocity feature projected onto
the galaxy and genuine emission from the galaxy itself, but the results
are very suggestive and should be followed up with observations of
greater resolution.
vm

CHAPTER I
INTRODUCTION
The Nature of Early-Type Galaxies
Since the discovery that the "spiral nebulae" were indeed island
universes somewhat like our own, the effort to study, classify, and
dissect them has been increasing with remarkable speed. In large
measure this is due as much to the expansion of the accessible electro
magnetic spectrum as to the increasing sensitivity of instruments
within each spectral window.
The basic observational technique used for this dissertation, radio
astronomy, dates from the 1932 observations of the Milky Way by Karl
Jansky. The achievements with this relatively new technique have
accumulated steadily over the past half century and it is now recog
nized as an invaluable tool in studying the universe and its contents.
One of the most attractive characteristics of radio astronomy is
that it complements the endeavors of the oldest technique, optical
astronomy. The analysis of visible light almost always entails objects
with a temperature of at least 2000 K and the vast majority of objects
this hot are stars. Radio astronomy, utilizing much less energetic
quanta, is sensitive to objects of several hundred degrees or less.
Typically this is primarily the interstellar gas that fills the space
between the stars. There are important exceptions to this rule
(synchrotron radiation, thermal bremstrahlung, etc.) but it illustrates
well the complementary nature of optical and radio astronomy.
-1-

-2-
Galaxy Classifying Schemes
Early studies of galaxies made it clear that the general morphology
of these immense stellar systems allowed them to be grouped into a
relatively small number of types. The most successful early venture
was Hubble's "tuning fork" diagram published in 1936, and reproduced
here in Figure 1. The spherical systems (EO) are at one end with pro
gressively flatter (E1-E7) systems leading up to a split in the diagram.
SO galaxies (also called lenticular galaxies) occupy the vertex of the
fork because Hubble believed that they were transitional systems. They
contain prominent elliptical bulges as well as a conspicuous disk
component. It is interesting to note that the nature of SO galaxies
is still being vigorously debated.
Figure 1. Hubble's "tuning fork" diagram of the classification of
galaxies.

-3-
Forming the tines of the fork are two parallel sequences of spiral
galaxies; one with a bar, the other without. The trend along the tines
is from tightly wound spiral arms (Sa or SBa) to looser, more open
arms at the end (Sb-Sc or SBb-SBc).
In addition to these major players in the drama, some 3l of all
galaxies are irregular, possessing no dominant symmetrical structural
features.
Many additions and modifications have been made to this initial clas
sifying scheme (Hubble 1936, Morgan 1958 and 1979, de Vaucouleurs 1959,
van den Bergh 1960a and b, Sandage 1961) but it has remained remarkably
unchanged over the years. In large measure the modifications are to take
into account the more extensive information available due to more sensi
tive equipment and increasing access to other spectral regions.
The following sections describe the elliptical and lenticular
galaxies in more detail and lay the groundwork for the statement of the
thesis problem in the final section of Chapter I. For historical reasons
both elliptical and lenticular systems are commonly known as "early-type"
galaxies.
The Morphology of Elliptical Galaxies
As the name implies, the elliptical galaxies are characterized by
elliptical isophotes. The stellar population usually appears to be
well-evolved with little or no interstellar gas or dust.
The degree of ellipticity E is defined to be (a b)/a (where a
and b are the semi-major and semi-minor axes, respectively). The range
observed is 0.0-0.7 (E0-E7) with E0-E1 the most common and decreasing
in frequency at the flatter end of the range.

-4-
Taking into account the statistics of random projection on the
sky (for we are viewing a two-dimensional projection of a three-
dimensional object) it appears that the ellipticals are distributed
normally about a mean of E3.6 (e = 0.36) with a dispersion of 0.11
(de Vaucouleurs 1977). It appears that true E0 and E5.5 (de Vaucouleurs
contends that E5.5 is the flattest bona fide elliptical) are rela
tively rare.
The lack of flat systems, usually considered to be a dynamical
effect caused by instabilities in thin disks, suggests that perhaps
spiral density waves and the attendant star formation are suppressed
in disks of sufficient thickness.
The origin of the flatness of elliptical systems has long been
thought a natural consequence of rotation. The greater the rotational
velocity, the greater the degree of flattening. This of course implies
that the three-dimensional figure is an oblate ellipsoid (polar diameter
smaller than the equatorial diameter).
In recent years several rotation curves of elliptical galaxies
have been published (Bertola and Cappaccioli 1975, Illingworth 1977,
Peterson 1978, Sargent et al. 1978, and Young et al. 1978a)which cast
strong doubt on the validity of this simple approach. The small ob
served ratio of the maximum rotational velocity to the central dis
persion velocity mitigates strongly against models which use isotropic
velocity distributions and either oblate or prolate ellipsoids
(Schechter and Gunn 1979). It appears necessary to use both anisotropic
velocity distributions as well as rotation to account for the observed
rotation curves (Binney 1978, Schechter and Gunn 1979).

-5-
The origin of the anisotropy is still not clear but the most
likely source is remnant anisotropy from the collapse phase of the
galaxies' formation.
A further difficulty in our understanding of elliptical galaxies
is the existence of extreme population I ingredients in a significant
number of systems. Specifically:
OB clusters are observable in NGC 185 and 205 (Hodge 1963
and 1973)
ionized gas is seen in the nuclei of at least 15% of all ellipti
cals (Osterbrock 1960 and 1962)
neutral hydrogen has been detected in at least 8 elliptical
systems:
NGC 1052 (Knapp et al. 1978b)
NGC 2974 (Bottinelli and Gouguenheim 1979b)
NGC 3904 (Bottinelli and Gouguenheim 1977b)
NGC 3962 (Bottinelli and Gouguenheim 1979a)
NGC 4105 (Bottinelli and Gouguenheim 1979b)
NGC 4278 (Gallagher et al. 1977)
NGC 4636 (Knapp et al. 1978a)
NGC 5846 (Bottinelli and Gouguenheim 1979b)
The presence of population I material is unusual for systems thought
to have ended all star formation long ago. Some possible explanations
are that the material was accreted relatively recently and thus it is
not representative of an elliptical galaxy's normal evolution, or that
through normal processes of stellar evolution the material was shed by
the stars and is observable in various forms today.
The thrust of the foregoing observations is that, as a class,
elliptical galaxies are not as well-understood as was earlier believed.
The apparently relaxed stellar distribution is most likely not relaxed
at all, but still contains velocity anisotropies which strongly in
fluence the shape of the galaxy. It is disturbingly common for the
smooth isophotes to be blemished with dust obscuration or some other

-6-
form of detectable population I material. The CO detection described
in Chapter III is another example of the incongruities present in
elliptical systems.
The Morphology of Lenticular (SO) Galaxies
The lenticular systems, occupying the vertex of Hubble's tuning
fork diagram, are as perplexing today as ever. The essential features
of these galaxies are an elliptical bulge and. a disk, both apparently
composed of population II stars. The disks are devoid of spiral arms
and the systems lack the obvious star formation proceeding in spiral
galaxies.
Suggestions as to the origin of the lenticulars range from van den
Bergh's proposal (1976b)that they are part of a complete sequence of
gas-poor, "anemic" spirals parallel to the S and SB types to the more
recent suggestion (Gunn and Gott 1972, Gisler 1979, and references
therein) that they are normal spirals which have been stripped of their
gas and dust by an interaction with an intracluster medium.
Much work has been done in recent years on these systems and
some interesting observations have been reported. Neutral hydrogen
observations by Balick et al. (1976) and a number of others place the
lenticulars roughly midway between the Sa-SBa and E systems. This
evidence supports the contention that they form another sequence of
spirals much like the normal sequence, only gas-poor.
Further, Burstein (1979a,b,c) finds significant differences in
the bulge-to-disk ratios of lenticulars and spirals, as well as
"thick" disks in lenticulars and not spirals. These are structural

-7-
features that are difficult to affect by stripping mechanisms, thus
supporting the parallel sequence hypothesis.
On the other hand, several people have investigated the spatial
distribution of lenticular systems and find that they are concentrated
within clusters, strongly suggesting that their current environment
is crucial in their formation.
Regardless of the mechanism for the origin of the lenticular
systems they have been included in this study because they share
several significant properties with elliptical galaxies. The stellar
populations appear quite similar and both have a lack of dust and gas
within their interstellar media.
The Formation and Evolution of Early-Type Galaxies
Formation
The general outline of galaxy formation, and elliptical galaxy
formation in particular, is understood only in its coarsest features.
This section presents the scenario most widely agreed upon with emphasis
on early-type systems. Significant gaps in the scenario are also noted
with various suggestions that may, in the future, fill them.
In the study of galaxy formation one is inevitably forced to con
sider earlier epochs. In astronomy this can be done easily by observing
more distant objects. The scale of the universe is such that the time
radiation has taken to arrive here on earth represents a significant
fraction of the object's existence. Continuing the effort to fainter
(more distant) objects gradually crosses over to cosmology and the
study of the origin of the universe.

-8-
Perhaps the most influential discovery in cosmology (and also
bearing heavily on galaxy formation) has been the 2.7 K microwave back
ground. The existence of nearly isotropic (see Cheng et al. 1979 for
a discussion of a dipole anisotropy attributed to the earth's motion
with respect to the background) homogeneous radiation with an apparent
blackbody spectrum strongly constrains galaxy formation theories.
The cosmic background radiation is almost universally believed to
be the remnant from the era of decoupling; the time when the universe
had cooled enough to allow electrons and protons to recombine (about
3000 K). The effect of the formation of neutral hydrogen is to reduce
drastically the opacity of the matter with respect to the radiation.
Before recombination Thompson scattering of radiation off electrons
coupled very strongly the matter with the radiation, they were kept in
thermal equilibrium and cooled together. After decoupling the two com
ponents evolved essentially independently of one another and the 2.7 K
background seen today is the remnant of the radiation component.
This forms the basis of the expectation that the early universe
was homogeneous and isotropic. Actually one can only say that it is
homogeneous and isotropic to at least the smallest scale observed,
currently that means the background is smooth to within 1 mK on a scale
of 7 (Smoot and Lubin 1979, see Sunyaev 1977 for a discussion of
fluctuations).
The condensations that have eventually evolved into clusters of
galaxies, galaxies and stars must have occurred during a later epoch.
The nature of these perturbations is currently a topic of great
interest. Several theories have been advanced to account for the
processes by which the perturbations form and grow (see Gott 1977 and

-9-
Field 1975 for reviews) but it will likely be many years before any
scenario is convincing and widely accepted.
Lynden-Bell (1967) proposed a theory in which the essential feature
of a galaxy's structure is determined by the timescale of collapse from
the background compared to the timescale of star formation. It can
be well illustrated by considering the various structural components
of our own galaxy.
Our galaxy consists of a disk of gas, dust, and relatively young
stars. The gas and dust are continually undergoing star formation
in which hot bright stars appear to be preferentially formed along
spiral arms. Spiral shock phenomena may be important in regulating
the star formation but regardless of the details there is a continuing
processing of gas and dust into stars.
Also present is a spheroidal halo which contains little if any
gas and dust. It is composed primarily of old stars and consequently
evolves only as fast as the stars that compose it.
Lynden-Bell's theory suggests that these structural features are
formed by varying rates of star formation occurring in the collapsing
proto-galactic cloud. Various components form as it separates out from
the cosmic expansion and begins to contract under its gravitational
force.
The novel feature of the theory is that it can explain how the
halo can relax in the time allowed. Essentially, the presence of a
changing gravitational potential will permit relaxation of the stellar
system much faster than would two-body encounters.
The stellar population of the halo is entirely old stars with no
gas or dust. During the collapse the stars form and are then

-10-
dynamically independent of the remaining material. That is, the
stars reflect the velocity dispersion of the cloud at the time of
their formation.
As the material collapses further, star formation proceeds and
eventually the gas becomes dense enough that damping forces become
important. At this stage the gas gives up a great deal of kinetic
energy to dissipative heating and, because of its angular momentum,
settles down into a disk. Within this disk various processes, especi
ally star formation, continue to the present giving spiral galaxies
their distinctive optical appearance.
It takes only a minor modification of this scenario to produce an
elliptical galaxy. If star formation has proceeded to completion before
the dissipative effects can form a disk then a system will result which
has no disk component and is essentially a halo population of stars.
Models which can produce the observed luminosity profiles of ellipti
cals have been constructed (Larson and Tinsley 1974) and add credence
to this basic concept.
Within this framework of galaxy formation many details are still
to be worked out. As previously mentioned, the nature of the initial
perturbations is not at all understood. Also the processes of frag
mentation and collapse are poorly understood (Field 1975). There is
even a question of whether gravitational instabilities or turbulence
is responsible for the necessary condensations. Jones (1976) provides
an excellent review of these and other problems in the study of galaxy
formation.
Since the early star formation was apparently so efficient as to
consume all the primordial gas, one might not expect any continuing

-li
star formation in elliptical galaxies. It will be shown, however,
that other processes operate which alter this simple picture and lead
to radically different expectations. The most important effects are
mass loss from the normal evolutionary processes of the stellar popula
tion and perhaps accretion of extragalatic material; both are discussed
in the following sections.
Mass Accretion
The contention that early-type galaxies, ellipticals in particular,
accrete material is relatively new. The motivation is to explain the
cD galaxies (giant ellipticals usually located in the center of rich
clusters and described by Bautz and Morgan 1970) that are often strong
radio sources. The accretion described here is full scale cannibalism
of other galaxies during close encounters (Ostriker and Tremaine 1975,
White 1976, Ostriker and Hausman 1977, Hausman and Ostriker 1978).
The idea can be linked to cluster types as described by Bautz and
Morgan (1970) and Oemler (1974). This cluster classification scheme
ranks the clusters based on their richness. Observationally it is
found that the densest clusters (Type I in the notation of Bautz and
Morgan 1970) often contain giant elliptical galaxies near their center.
Further, the cD galaxies are often radio sources that are widely
suspected to be caused by material falling into a massive object.
There are several other pieces of circumstantial evidence that indi
cate that this process may indeed be significant in the evolution of
elliptical galaxies (see Ostriker 1977 and Hausman and Ostriker 1978
for details).

-12-
A more moderate form of mass accretion is suggested by several
authors (Bottinelli and Gouguenheim 1977a, Gallagher et al. 1977, and
Knapp, Kerr, and Williams 1978c)to explain the inclined disk of
NGC 4278. The differing directions of the angular momentum of the
stellar component of the galaxy and the neutral hydrogen make an
internal origin of the matter difficult to believe.
A similar situation occurs in NGC 1052 (Knapp, Faber, and
Gallagher 1978a, Fosbury et al. 1978, Reif, Mebold, and Goss 1978) and
the accretion of an intergalactic HI cloud is suggested. The major
objection to this hypothesis is the lack of sufficiently massive clouds
available for accretion.
The results of Mathewson et al. (1975) purporting to find HI
clouds in the Sculptor group have been disputed by Haynes and Roberts
(1979). The latter group contend that the material is a portion of the
Magellanic Stream. Further, Lo and Sargent (1979) have searched nearby
groups for detached HI clouds and find none more massive than
% 4 x 107 M .
A number of other authors have examined clusters of galaxies for
HI emission (Haynes et al. 1978, Baan et al. 1978) while others have
examined the line of sight to quasars for HI absorption (Roberts and
Steigerwald 1977, Shostak 1978). No isolated HI clouds are seen in
emission in the clusters and the HI absorption measurements show that
large clouds of neutral hydrogen are almost never seen outside
galaxies.
O
Thus it appears difficult to reconcile the several times 10 Mq
of HI found in NGC 4278 and 1052 with the dearth of sufficiently
massive candidates for accretion. Silk and Norman (1979) propose an

-13-
alternative hypothesis, the accretion of gas-rich dwarf galaxies.
They find that the gas component of the dwarfs will lose energy through
dissipation and fall to the central regions of the accreting galaxy.
The infalling gas, depending on the individual cloud mass, may either
form stars or continue to fall into the nucleus where it may fuel a
radio source. The stars will also be incorporated into the accreting
galaxy but with less visible effects.
Silk and Norman (1979) also consider the interaction of a Mathews
and Baker (1971) type wind and the infalling material. If the amount
of this material is sufficiently large the resulting supernovae (from
the high-mass stars formed) will help in driving the galactic wind.
However, an enhanced wind has the effect of inhibiting mass infall and
the process slows itself. The net effect may be for star formation to
proceed in cyclical bursts--a notion also suggested by van den Bergh
(1975) in a somewhat different context.
Both of the mass accretion processes described so far deal with
normal or giant cD elliptical galaxies. In order to be effective in
capturing and assimilating other systems the accreting galaxy must be
large. The evolutionary mechanism discussed in the following section,
mass loss from stellar evolution, operates in all systems. This in
cludes the dwarf ellipticals NGC 185 and 205 considered in greater
detail in Chapter III.
Mass Loss Due to Stellar Evolution
It has recently been appreciated that the normal evolution of
stars in an early-type galaxy will be a source of interstellar material.

-14-
Small mass stars have stellar winds, Mira variables are known to eject
mass during certain stages, Type I supernovae occur in population II
stars, and planetary nebulae have been observed in early-type galaxies
of the local group.
The calculation of the contribution by stellar evolution to the
interstellar medium (ISM) of early-type galaxies depends more on theo
retical estimates than observational evidence. To date, the most impor
tant observational evidence is the detection of planetary nebulae in
nearby dwarf ellipticals (including NGC 185) by Ford and Jenner (1975).
Considering the uncertainties in the observations, the observed planetary
-1 9 -1
birthrate of > 0.012 yr (10 L ) agrees well with Larson and Tinsley's
(1974) estimate of 0.05 yr-^ (10^ L )*.
Following the reasoning of Faber and Gallagher (1976) and adopting
a mass per planetary of 0.2 M^ results in a mass loss rate of 0.010

-1/9 -1
yr (10 L ) from planetaries. Consideration of Mira-type variables
leads to a final assumed mass ejection rate of 0.015 yr (10 L ) .
The conservative nature of this calculation is apparent when one
considers that the present mass loss rate is certainly lower than that
of earlier epochs. This is primarily because any high-mass stars
would have evolved quickly and cycled their mass back to the ISM early
in the galaxy's evolution.
Further, the contribution of mass from Type I supernovae (ap
parently confined to population II stars, Tammann 1974) and Type II
supernovae (massive progenitors) earlier in the galaxy's evolution
have been ignored. Even this conservative approach leads to contra-
9 10
dictions in the ISM of early-type galaxies after 10 -10 years (Faber
and Gallagher 1976).

-15-
Strippinq Mechanisms
The various efforts to determine how interstellar material shed
by stars can be removed from early-type systems can be divided into
two classes. The first is an internally driven "galactic wind" and
second is ram-jet stripping by an intracluster medium. Both of these
processes will be discussed in the following sections.
Internally Driven Winds
The possibility of a galactic wind driven by an internally driven
energy source was suggested by Johnson and Axford (1971) and considered
more quantitatively by Mathews and Baker (1971). In essence the mech
anism operates by coupling the energy from Type I supernovae to the
general interstellar medium. The addition of this high-energy
/ 9 x
(8 x 10 K) low-mass component significantly heats the interstellar
material to a high enough temperature to escape from the system by
evaporation. A later study by Coleman and Worden (1977) shows that
the energy released by flare stars is by itself enough to drive a
galactic wind of this type.
The parameters that are most important in the establishment and
maintenance of a galactic wind are the Type I supernovae rate, the
energy output from each supernova, the efficiency of the coupling of
the supernova's energy to the ISM, and the amount of "pre-heating"
of the ISM by the velocity dispersion within the galaxy. Each of
these quantities are known to probably a factor of 2 at best and in
certain instances various authors disagree by factors of 10 or more.

-16-
For example, Mathews and Baker (1971) assume a coupling effi
ciency between the expanding supernova shell and the ISM of 1; that
is, all the kinetic energy of the supernova is converted into thermal
energy of the ISM. Gisler (1976) takes exception to this number and
notes that Larson (1974) uses an efficiency of 0.1. Given the various
uncertainties it appears that while a galactic wind will most likely
prevail in some instances, perhaps even a majority of elliptical
galaxies, there are cases in which it simply does not operate.
Indeed, Mathews and Baker (1971) find solutions in which a wind is
not supported and the material collapses to the center of the system.
They further propose that the hot, ionized gas will only be able to
form massive objects, thus linking the lack of a galactic wind to the
formation of radio sources in early-type galaxies.
Again, Gisler (1976) points out an inconsistency in this line of
reasoning. From observations one finds that strong radio sources were
more common in earlier epochs. Gisler notes that the earlier stellar
content of ellipticals is more likely to produce supernova. This
follows from the observation that only Type I supernovae occur in popu
lation II (old) stars and the precursors are probably low mass stars
(Tammann 1974). In the earlier stages of an elliptical's life the
supernova rate can only be augmented by Type II supernovae (whose
progenitors are young, massive stars). In addition it is at the
earlier epochs that the galaxy will not have had time to collect a
significant amount of gas from the evolution of its stellar component.
For these two reasons it would appear that the ellipticals are better
able to support a galactic wind at earlier epochs--just the period
when the greatest fraction must also be radio sources.

-17-
Faber and Gallagher (1976) have considered the problem of a
galactic wind from a slightly different approach. Since the initial
heating of the ISM is through collision of the gas clouds shed by
stars in the process of normal evolution, the velocity dispersion of
the stars can be used as an indicator of the stellar equivalent temper-
ature, T From Mathews and Baker (1971) T = v mu/6k where v is
S S S n S
the stellar dispersion velocity and m^ is the mass of the proton, and k
is Boltzmann's constant. Along with the condition that the supernova-
induced rise in temperature must roughly double the kinetic energy of
the gas to remove it from the system to infinity, one finds
"sn ^sn
C£
- asTs
(1)
where agn and ag are the specific rates of mass injection by super
novae and stars, respectively, Tsn and Ts are the equivalent tempera
tures of the supernovae and the stars.
Using the values quoted in Faber and Gallagher (1976) a =
5 x IQ"20 sec"1 and a T = 1.6 x 10"12 K sec'1 yields T = 3.7 x 107
sn sn J s
K or v$ < 1260 km/sec. The line-of-sight velocity dispersion is
< 1260/v^S = 730 km/sec (assuming 3 dimensional isotropy of the velocity
dispersion), which exceeds by a factor of two the largest velocity
dispersions measured (Faber and Jackson 1975).
This calculation would make it appear that virtually all elliptical
systems should have a galactic wind scouring the ISM of any material.
However, Gisler (1976) recomputes this same quantity, taking into
account the 0.1 efficiency of energy coupling found by Larson (1974),
and finds that the maximum velocity dispersion that will still allow
the supernovae to double the energy is about 200 km/sec, a number
comparable to the observed velocity dispersions. In other words,

-18-
if a galaxy has a velocity dispersion greater than 200 km/sec then the
supernovae contribution will not be able to double the kinetic energy
of the ISM and a galactic wind cannot be established.
Even more damaging to the galactic wind hypothesis is the de
tection of HI in any elliptical. In order for the wind to operate
the ISM must be hot lO'7 K). All of the gas in a galaxy would thus
be ionized and according to Mathews and Baker (1971) quite unobservable
by present techniques.
A final argument against the universal existence of galactic winds
is that if all other conditions were the same, one would expect the
more spherical systems to be better able to support a galactic wind.
The reasoning is that the spherical systems have less surface area per
unit volume through which to radiate excess energy, keeping the ISM
as hot as possible.
Contradicting this expectation, the neutral hydrogen observations
of 8 elliptical systems show detections significantly skewed towards
the more spherical galaxies. The systems detected in neutral hydrogen
have the following classifications: 2-E0, 2-El, 1-E2, 1-E3, 2-E4.
Conspicuously absent are the flatter systems that one would expect to
be better able to radiate energy away and thus retain their ISM.
Ram-Jet Stripping by an Intracluster Medium
The proposal that ram-jet stripping of an ISM could be significant
in the evolution of a system was treated first by Gunn and Gott (1972).
Later, more sophisticated treatments by Tarter (1975), Gisler (1976),
and Lea and De Young (1976) all support the notion that stripping can
be an effective process.

-19-
The thrust of much work in this area has been to determine if
SOs can be formed by stripping spiral galaxies of their gas and dust
(Gisler 1979). This would quench star formation and significantly change
the optical appearance of the galaxy.
Another development that has spurred interest in the interaction
of an ISM with an intergalactic medium (IGM) is the discovery of
head-tail radio galaxies. The most straightforward explanation of
this phenomenon being just such an interaction.
In spite of the varied motivations for these studies many of the
numerical simulations are directly applicable to the analysis of an
elliptical system passing through an IGM.
The primary results of these studies are to confirm that under
appropriate conditions there is an effective sweeping out of material
from the galaxy. The procedures and model parameters used to arrive
at this conclusion vary substantially for the different experiments,
all agree however, that some material tightly bound near the nucleus
may be retained. Gisler (1979) explores the situation further and
finds that the rate of gas replenishment is important, possibly stopping
the stripping effect entirely if it is high enough.
In spite of this it appears that stripping can be at least par
tially effective over a broad range of galaxy velocities and IGM
densities.. This finding agrees well with the observation that a large
fraction of galaxies in rich clusters are SOs and ellipticals (Oemler
1974, 1977).
It would seem also that the evidence of a positive correlation
between X-ray luminosity of clusters (presumably from a hot intracluster
component) and SO/spiral ratios (Tytler and Vidal 1978) argues strongly

-20-
that the cluster environment does indeed have a significant influence
on the structure and evolution of its component galaxies.
An evolutionary effect may also have been observed by Butcher
and Oemler (1978). Their study found that a rich cluster observed at
a redshift of 0.4 contains many more blue galaxies than a similar rich
cluster observed in the current epoch. Their conclusion is that as
Q
late as 4 x 10 years ago the galaxies in this cluster had not yet
been stripped of their ISM. They were consequently undergoing at
least moderate star formation, thus producing the blue colors observed.
One further piece of evidence that fits in quite well with the
general hypothesis of ram-jet stripping is the common coincidence of a
strong radio galaxy at the center of a dense cluster (McHardy 1974,
Guthrie 1974, and Riley 1975). The argument in this case is that the
central galaxies have a small velocity with respect to the intra
cluster medium and will be more likely to retain gas shed by its
stellar population. The material collapses to the center of the
galaxy, apparently forming a massive object and producing the ob
served radio source.
In view of the variety of indications that imply a substantial
interaction between an ISM and an intracluster medium it seems clear
that environmental factors can be important in the evolution and
structure of galaxies in clusters. But in the particular case of
the detected ellipticals NGC 4278 and 4636 there is reason to believe
the IGM is unimportant.
These are the only two galaxies which have an HI distribution
that is extended enough to map. In both instances the HI distribution
appears to be considerably wider than the photometric diameter of the

-21-
system (Knapp et al. 1978a,Bottinelli and Gouguenheim 1979a). More
importantly, the distributions appear to be reasonably symmetric, a
condition difficult to reconcile with stripping or partial stripping
by an intracluster medium. This discrepancy is pursued further in
Chapter IV.
Other evidence that ram-jet stripping may not be as effective as
the numerical analyses indicate is found in the Hercules cluster, a
rather loose cluster composed almost entirely of spiral galaxies.
The inconsistency is that it is also an X-ray source. Our current
understanding of cluster X-ray sources necessitates a hot ('v 10^ K)
intracluster medium as the origin of the radiation. How the Hercules
galaxies have remained spirals and not been stripped is not understood
within the framework of current research. Even more perplexing is the
origin of the intracluster gas, it is usually thought to have been the
gas removed from the galaxies.
A recent statistical study of cluster morphology by Gisler (1980)
shows that the anticipated presence of Sc galaxies in clusters is not
found. The Sc galaxies are expected to be highly resistant to having
their ISM swept because they have a high rate of gas replenishment
(Gisler 1979). The apparent underabundance in rich clusters indicates
that ram-jet stripping cannot be the dominant mode of SO production.
Dressier (1980) comes to similar conclusion based on a study of
the morphology of the galaxies in 55 rich clusters. He finds a
significant number of SO systems in clusters which have too low a
density to accomplish any stripping. Further, the study finds a
difference between the bulge-to-disk ratios of spirals and SOs as well
as a tendency towards thicker disks in SO systems. From these

-22-
observations Dressier (1980) concludes that spirals do not evolve
into SOs. He concludes that the galaxy types are affected more by the
initial conditions at the time of their formation than by environmental
factors such as ram-jet stripping.

-23-
Fate of Retained Gas
Supermassive Objects in the Nucleus
Mathews and Baker (1971) suggested that if their galactic wind
should fail, the gas in the system would fall to the nucleus in an
ionized state. They further argue that the Jeans radius
R.i = [
rkT
X
2
16yMG((
P*J
(2)
where T is the temperature, y is the molecular mass, G is the gravita
tional constant, p and p* are the densities of the gas and stars,
respectively, and k is Boltzmann's constant is determined primarily
by the stellar density. That is, as the gas collapses it responds to
the gravitational field of the stars. This will continue until the gas
becomes more dense than the stellar component. The gas is dense and
collapsing quickly at this stage and Mathews and Baker suggest that
there may not be enough time for fragmentation to take place. The
collapsing material then forms a massive object rather than fragmenting
and forming stars with a normal distribution of masses.
Another argument for supermassive objects is the observation that
cD galaxies with radio emission are often located in the center of
dense, rich clusters. The evidence is largely circumstantial but if
ram-jet stripping is important in the evolution of ellipticals then
it follows that the central members of a cluster will be least affected
by this mechanism. Of course the step from ineffective ram-jet
stripping to a supermassive object in the galaxy's nucleus is by no
means secure. It rests on the assumption that the retained material

-24-
either forms the supermassive object or at least provides fuel for the
radio emission.
These arguments actually rest on much firmer ground due to recent
work on the velocity dispersions and light distributions within the
nuclei of supergiant cD galaxies found in the centers of rich clusters.
Young et al. (1978b)obtained luminosity profiles of the supergiant
elliptical NGC 4486 (M87) which, when examined with the velocity dis
persions determined by Sargent et al. (1978), show that the nucleus
contains a massive dark object. The nature of the dark object cannot
Q
yet uniquely be determined, but it must contain 5 x 10 M of material

and have a radius less than or equal to 100 parsecs (pc). Young et al.
also determined that the mass-to-luminosity ratio must be greater than
60. Several possibilities are advanced but Young et al. find the most
g
plausible to be a massive black hole of 5 x 10 M The attraction

42 -1
of this hypothesis is that the 10 erg sec energy output of NGC 4486
-2 -1
can be explained by supposing a mass infall of about 10 M0 yr with
only a 0.002 conversion efficiency into radiation.
De Vaucouleurs and Nieto (1979) confirmed the results of the
Young et al. (1978b) study and found essentially similar results for
the dark mass at the nucleus. The earlier results were obtained with
a CCD (charge-coupled device) camera, while the work by de Vaucouleurs
and Nieto was with more conventional photographic and photoelectric
photometry.
Young et al. (1979) also examined the luminosity profiles of NGC
4874, 4889, and 6251 and found that only NGC 6251 requires a supermassive
object at its center to fit the data. Further, only NGC 6251 is a radio
source amongst the three.

-25-
It appears from these observations that there may be a correla
tion between radio galaxies in the center of clusters and anomalous
nuclei. Since current understanding of radio sources usually involves
massive objects (which can also explain the anomalous nuclei) the
circumstantial connection between a lack of ram-jet stripping and a
supermassive object in a galaxy's nucleus is established.
However, within this scenario it is quite unclear whether the
retained gas actually formed the massive object or just fuels it. The
possibility that the galaxy formed with a massive nucleus cannot be
overlooked; indeed, one of the central questions yet to be answered is
how important are the initial conditions under which the galaxy formed.
Cyclic Bursts of Star Formation
The contention that the star formation rate in an elliptical or
SO galaxy is strongly dependent on time gains credibility only recently.
Van den Bergh (1975) cites several examples of elliptical galaxies
experiencing anomalously vigorous star formation.
NGC 5128 (also known as the radio source Centaurus A) has recently
been shown by van den Bergh (1976a) to be undergoing very active star
formation along and interior to its prominent equatorial dust band. He
also suggests that the source of the unusual dust and gas in the system
is stellar debris shed by the stars. The galaxy is apparently a rare
field elliptical. It does not belong to a rich cluster and presumably
lacks any ram-jet stripping which may exist in such an environment.
A strong argument against this hypothesis is the finding by
Graham (1979) that the old stellar population rotates much slower than

-26-
the equatorial band of dust and gas. This, is difficult to reconcile
with the interstellar material originating with the stars.
Most other researchers in the field tend toward the explanation that
the system may actually be a collision between a gas cloud or small
galaxy and an elliptical (Graham 1979). Thus the relevance of NGC
5128 to elliptical galaxy evolution cannot be assessed until these
questions are answered.
Van den Bergh (1975) has also suggested that NGC 1275 is an ex
ample of an elliptical galaxy caught in a burst of star formation. The
underlying stellar distribution in the system is elliptical and the
galaxy is located near the center of the Perseus cluster. The small
velocity difference between the cluster's mean velocity and that of the
galaxy again suggests inefficient ram-jet stripping.
However, more recent work on the system (Kent and Sargent 1979,
Rubin et al. 1977) indicate that this is a collision between a fore
ground spiral system and NGC 1275 in progress. Thus both NGC 5128 and
1275 may be atypical as far as early-type galaxies are concerned.
However, these two examples support the view that a massive (a- 10^
M ) influx of material over a relatively short time period results in
vigorous star formation. As both these systems are also radio sources,
one could argue that another effect of the mass accretion is to
either form, or fuel a previously existing, massive object in the
nucleus, producing the radio source.
Further support for this hypothesis, although not in an early-type
system, may be found in the M81-82 system. The interaction between
these galaxies seems to be resulting in a substantial mass infall to
M82 (Gottesman and Weliachew 1977, Killian 1978). It is likely that

-27-
the infall is connected with the peculiar structure of M82 and its
vigorous star formation.
On a gentler scale there are several examples of star formation
in dwarf ellipticals in the local group. NGC 185 and 205 both contain
dust patches and a sprinkling of blue, presumably young stars. The CO
observations described in Chapter III show that the star formation
is probably continuing in NGC 185. NGC 205 is a close companion of M31
and it may be argued that an interaction is taking place, although HI
maps of the region do not support this idea. Regardless, NGC 185 is
considerably further from M31 and an interaction does not appear likely
(see the HI observations presented in Chapter III concerning this
possibility).
Consequently, it appears that these two dwarf ellipticals are
reprocessing their ISM back into stars. The most likely source of
the material is the normal products of evolution of stars. Large scale
interactions are apparently not supported by the HI observations of
the galaxies.
The more sedate pace of star formation, and various arguments
suggesting an initial mass function for the stars shifted to lower
masses, are examined in the following section.
Continual Star Formation with a Skewed Initial Mass Function
A third fate that may befall gas retained in early-type systems
is star formation. The most stringent requirements on the nature of
this star formation are set by the observations of the colors or early-
type galaxies (see Larson and Tinsley 1978 for a discussion and earlier
references).

-28-
Larson and Tinsley (1974) have calculated models for elliptical
galaxies with star formation rates, continuing to the present, capable
of consuming the gas shed by other stars. While the integrated colors
of the models are consistent with observed galaxies, the expected
gradient of increasingly blue colors in the nucleus is not widely
observed. The most obvious drawback in their modeling is the use of
an initial mass function (IMF) which is fairly rich in hot stars.
As Faber and Gallagher (1976) comment "Since we have no a priori
knowledge of the IMF in ellipticals, star formation might conceivably
be confined to stars of small mass and low luminosity" (p. 370). They
go on to point out that the star formation must proceed efficiently
since very little, if any, interstellar material is seen in most
elliptical galaxies.
There is, however, observational evidence for star formation with
an anomalous IMF. Van den Bergh (1976c) suggests that an IMF deficient
in high mass stars is the most likely explanation for the lack of HII
regions in the Sa galaxy NGC 4594 (M104). Knots of young, blue stars
are observed near the prominent dust lanes. Normal, massive 0 and B
stars would form prominent HII regions under such conditions. Further,
this galaxy was not detected by Gallagher et al. (1975) in HI even
though a normal dust-to-gas ratio indicates it should have been easily
seen. Van den Bergh suggests that the lack of HI is due to its being
converted into molecular hydrogen, thus escaping detection.
On theoretical ground also, an IMF skewed away from massive stars
may be expected. Jura (1977) finds that one effect of reduced heating
of interstellar clouds (expected in elliptical galaxies) is to allow
clouds with much smaller masses to become gravitationally unstable

-29-
and collapse. Indeed, he finds that the critical mass for collapse under
such conditions is less than 10 M Because of fragmentation the
result of the collapse of a 10 Mq cloud will almost certainly be a
number of small mass stars rather than one 10 M star.

Thus while no compelling reason can be advanced to accept star
formation with a skewed IMF, the possibility cannot be rejected either.
The main intention of this dissertation as outlined in the following
section is to examine this possibility to determine if it is a sig
nificant process in the evolution of early-type galaxies.
Statement of the Dissertation Problem
From the previous sections it is clear that our understanding of
early-type galaxies is incomplete. There are a large number of inter
esting ideas concerning their formation and evolution, but as usual, in
sufficient data are available to assess adequately their importance.
The primary goal of the observations presented and discussed in this
dissertation is to investigate the significance of star formation in the
evolution of early-type galaxies. The results of this study are most
directly applicable to the recycling of material by star formation in
an environment much different than our own galaxy. The results also
bear on other settings in which the star formation is poorly under
stood.
It is not overstating the current situation to say that we only
have a dim view of how star formation occurs in our own galaxy, and an
even dimmer view of the process in drasically different environments.
Faber and Gallagher (1976) argue that possible star formation in

-30-
early-type systems must be very efficient; if so, it may be genericaliy
related to the star formation that occurred as the galaxy collapsed
originally. Judging from the lack of primordial material observed,
that process was also very efficient.
This work is also of importance in the formation and continued
activity of radio sources in early-type galaxies. From the available
information it seems that radio sources and star formation are com
peting for the same interstellar debris in a galaxy. If one is very
successful, it may be at the expense of the other.
All of this is not to say that galactic winds and ram-jet stripping
never occur, simply that all possibilities need to be analyzed and
evaluated to determine their relative importance in the overall
scheme of galaxy evolution.

CHAPTER II
THE OBSERVATIONS
This chapter presents the data acquisition and handling procedures
used for the work described in this dissertation. The instruments used are
described giving particular attention to the equipment and techniques
which aided this work immensely. The method of data presentation is
also explained.
Carbon Monoxide Observations at Kitt Peak, AZ
Telescope Description
-jo -i r
The observations searching for the 2.6 mm transition of iCiD0
were made at the National Radio Astronomy Observatory (NRAO) Millimeter
Wave Telescope1 at Kitt Peak, AZ. The telescope is an 11 m paraboloid
which can be driven in altitude and azimuth. Tracking of celestial
objects, data acquisition,monitoring of system status, as well as
initial data reduction is handled by an on-line PDP 11/40 computer.
The observations were taken during three separate observing
sessions
December 23-26, 1977
July 7-9, 1978
December 10-16, 1979
^The National Radio Astronomy Observatory is operated by Associated
Universities, Inc., under contract with the National Science Foundation.
-31-

-32-
The cooled mixer Cassegrain receiver operating at the 115.2712 GHz
assumed rest frequency of the J = l->0 CO transition was used for all
observations. The halfpower beam width at this frequency is about
65".
Pointing for the telescope is initially determined by the observa
tory personnel. Pointing correction data are taken in all parts of the
sky on bright mm point sources. Analytic functions are then fitted to
the data, interpolated over regions with few data points. The observer
then checks these corrections by making five-point maps of bright sources,
usually the planets. Rarely are the corrections more than 5 arcseconds
at this stage.
However, the 1979 observations were hindered by large errors in
regions far from the celestial equator due to a lack of adequate data
for the fitting equations. This problem is discussed further in
Chapter III.
Nominally, the sky signal is separated into two linearly polarized
orthogonal components. In practice the system operated with only one
polarization, except during the 1979 observations when both channels
were available. These two signals are then fed to two Shottky-barrier
mixer diodes. Also fed to the mixer is the local oscillator (LO)
signal via a tunable injection cavity. This process of mixing two
frequencies to get a third (usually the difference of the two) is
known as heterodyning. The first intermediate frequency (IF) signal
emerges at 4750 MHz and is amplified by a pair of low-noise parametric
amplifiers. All of the previous equipment is enclosed in a dewar
cooled to 15 K.

-33-
Once outside the dewar the 4750 MHz first IF is amplified again
by room temperature transistor amplifiers. The signal is heterodyned
further at this point to 1328 MHz.
A third and final mixer produces a 150 MHz third IF which is
detected by two banks of 256 square-law detectors. One has a choice of
filter widths for the filter banks; our observations always used either
the 0.5 or 1.0 MHz (per channel) banks, corresponding to 1.3 and 2.6
km/sec velocity resolution, respectively. The December 1977 and July
1978 observations were made with only a single polarization operating, so
the separate channels were actually the same signal and could not be
averaged to reduce noise. Consequently the 0.5 MHz (1.3 km/sec) and 1.0
MHz (2.6 km/sec) filter banks were both used. At the time of the Decem
ber 1979 observations the receiver operated in a proper two-channel mode
with two independent IFs. Both were fed to 1 MHz filterbanks since they
could now be added with a resultant decrease in noise.
Calibration was accomplished by alternately observing blank sky
and an ambient temperature microwave absorber. Details of the cali
bration procedure are explained in Appendix I.
As a result of the calibration procedure one obtains the corrected
antenna temperature of the source
T* = nf(l e"T)[J(vs,TE) J(vs,Tbg)] (1)
where the explanations for the symbols are found in Appendix I. Note
that there are three unknowns (nf, x, T^.) related to the particular
source being observed; these cannot be determined without additional
information. For extragalactic work one can make estimates for
x and T^ based on galactic studies and then calculate ty. The

-34-
errors are large in this approach and they give only a general idea
of the nature of the molecular cloud being observed.
Data Reduction Techniques
Each spectrum consists of two channels of 256 points each. For
December 1977 and July 1978 only one polarization was available and it
was detected in both the 256 MHz wide filter bank as well as the 128
MHz wide filter bank. The December 1979 observations took advantage
of the two orthogonal polarizations then available and detected them
separately in two 256 MHz filter banks. At the rest frequency of the
12
J = 1-M3 transition of CO the 1 MHz resolution corresponds to 2.6
km/sec.
The on-site PDP 11/40 computer writes the data for each spectrum
(typically representing 5-10 minutes of observing) on disk. Real time
analysis can be done at the telescope and allows the observing time to
be optimally used. At the completion of the observing session the
disk is copied onto a binary coded tape.
The next step in the processing is to rewrite the tape with the
IBM 360 computer at the NRAO in Charlottesville, VA, to put the infor
mation into an IBM-compatible format. This is entirely a translation
and no analyses are performed.
The final reduction is done at the University of Florida using
the resources of the Northeast Regional Data Center. Each spectrum is
examined for an unusually high root-mean-square deviation; all of the
offending spectra are rejected. The baselines are then examined and
if a polynomial of more than first degree is required to fit the base
line, the spectrum is discarded.

-35-
At this stage of the reduction process each point of the spectrum
is examined to see if it is greater than five times the standard
deviation for that spectrum. If it is, and the adjacent points are not,
the point is replaced with the mean of the two adjacent points. This
procedure is used to remove interference that involves only a single
channel; multichannel filter banks are particularly susceptible to
this type of interference.
Each spectrum is then weighted by the inverse square of its
standard deviation and combined with all other spectra on the same
source.
A preliminary first order baseline is fit and subtracted and the
resulting spectrum is examined for possible features. If any are
apparent the channels involved are eliminated from the baseline fitting
procedure and a new first order baseline is calculated and subtracted.
The resulting spectrum represents the best data on a given source.
Various smoothing functions can be applied; the most common for this
work has been smoothing with a rectangular function of about the same
width as the suspected spectral feature. The effect of this is to
maximize the signal-to-noise ratio at the expense of velocity resolu
tion.
The basic format of each spectrum presented is explained in
Figure 2.
Data Presentation
The data are displayed with antenna temperature as the ordinate
and velocity along the abscissa. Strictly speaking the abscissa

-36-
represents frequency, but since the frequency of the molecular
transition is already known the axis is calibrated in km/sec using
the following relation
(2)
where vg is the center velocity of the observed band, v^ is the velocity
of the source (corrected for the earth's rotation and revolution, i.e.
heliocentric), v is the rest frequency of the spectral features, and
vobs bservec* frequency.
The ordinate, corrected antenna temperature (T^), is found by
calibration as described in Appendix I. The unit is Kelvins (K), and
it is related to the flux density (S) by the following equation
(3)
s = 2 k TA/Ae
where n£ is the antenna efficiency due to the loss of elements
(spillover, blockage, and ohmic losses in the antenna), A is the
G
effective area of the antenna, and k is Boltzmann's constant.
Neutral Hydrogen Observations at Green Bank, WV
Telescope Description
The neutral hydrogen observations were made with the NRAO 43 m
radio telescope at Green Bank, West Virginia. The telescope is an
equatorially mounted instrument completely under computer control. The
observations were taken during one session from July 13 to July 23, 1978.

-37-
The 21 cm cooled cassegrain receiver was used for all data acquisi
tion. The system has two channels provided by linearly polarized,
orthogonal feeds. After initial amplification by a cooled upconverter
amplifier the signal is heterodyned and amplified through various stages
in much the same fashion as the process described for the 11 m tele
scope at Kitt Peak. Typical system temperatures for the 43 m system
were 50-60 K.
The standard NRAO "back end" uses a 150 MHz IF which is fed into
a Model II autocorrelator spectrometer. The IF signal is autocorrelated
and the resulting autocorrelation function is Fourier transformed to
produce the power spectrum. The formation of the spectrum using auto
correlation techniques is described in more detail by Blackman and
Tukey (1958) and Cooper (1976).
A 10 MHz bandwidth was chosen for all observations to provide an
adequate baseline. Also in the interest of baseline stability a
position-switched mode of observing was adopted. Ten minutes of data
are taken at the "off" position followed by 10 minutes at the "on"
position. The final spectrum is found by differencing the two spectra
thus acquired.
A number of "off" positions were used in an effort to deduce the
distribution of HI in the region around NGC 185. The majority of the
data were taken either with an "off" 16m to the west or as a five-point
map. The arrangement for the five-point map is with the "on" at the
center and "offs" taken successively to west, east, south, and north
at a distance of either 48' or 24'. The half-power beam-width at 21 cm
is about 20.51.

-38-
Data Reduction Techniques
The output of the autocorrelator is two 192 channel spectra, each
being linearly polarized but orthogonal to the other. The spectra are
recorded by an on-line disk drive which can also be accessed by the
on-line reduction computer. In operation, the near instantaneous access
to the data just taken enables the observer to monitor the quality of
the system operation and to update the observing procedure based on the
preliminary results.
A characteristic of the autocorrelation method of spectral analysis
is that the strong galactic hydrogen within the bandpass produces a
sinusoidal ripple in the spectrum. This is known as "ringing" and its
removal is accomplished by convolving the spectrum with a hanning
function. This function is a weighting scheme in which the value of
the channel on either side is added to i the value of the central
channel to produce the new value for that channel. Application of
this smoothing worked very well and all data presented here from the
43 m telescope have been smoothed with the hanning function.
A baseline is removed from the data by fitting a low-order poly
nomial to the spectrum in regions removed from either galactic emission
or suspected NGC 185 emission. In practice the order of the fitting
polynomial was 2 to 4.
Calibration for this system is done under computer control by
periodically firing a noise tube within the receiver and comparing
the system output with and without the additional noise. The data are
then scaled to this system temperature. The stability of the system
was monitored by observing several sources throughout the session. No
unexplained drifts in system performance were seen.

CHAPTER III
RESULTS
In this chapter the results of the observations are presented.
The carbon monoxide data are treated first, followed by the neutral
hydrogen study of NGC 185 and the surrounding region. Only the results
are considered in this chapter; the analysis and interpretation of the
observations are found in Chapter IV.
Carbon Monoxide
Positive Result in NGC 185
The dwarf elliptical galaxy NGC 185 was detected at the 115.2712
12
GHz frequency of CO. Figure 2 shows the best spectrum obtained on
the source, averaging data from all three observing sessions. The
256 MHz total bandwidth corresponds to about 660 km/sec. The feature
is not discernible in the 128 MHz bandwidth spectrum, probably because
of the lower signal-to-noise (SN) ratio. Table 1 summarizes the
parameters of the spectral feature as measured in Figure 2.
Since the line is only slightly greater than 3 standard deviations,
it is very important to be certain of the reality of the feature. One
method is to smooth the spectrum with a function (in this case rect
angular) of about the same width as the suspected spectral feature.
This procedure has the effect of increasing the SN ratio at the expense
-39-

12
Figure 2. The CO spectrum for the nucleus of NGC 185. Data from all three
observing sessions have been averaged for this spectrum. The
total bandwidth for the observation appears in the upper left
along with the total integration time on the source. In the lower
right the position (1950.0 coordinates) observed is given and the
standard deviation for the spectrum is also shown. The baseline
at 0.0 Kelvin is shown along with horizontal lines at 3 standard
deviations.

TEMPERATURE K)
it?

-42-
of velocity resolution. Figure 3 is the result of this convolution,
and indeed the reality of the feature is strongly supported.
Table 1
Line Parameters
for 12C0 in
NGC 185
Peak Antenna Temperature (K)
0.081
0.025
Velocity of Line (heliocentric, km/sec)
-175 2
>
Full-width at Half-maximum (FWHM)
(km/sec)
12 2
Position Observed
0h 36m
10?3
48 3'
5 33
Total Integration Time on Source
(min)
484.5
Other support for the reality of the line is found by careful
examination of the spectra obtained during each session of observing
(Figures 4 to 6). The spectral characteristics vary somewhat for each
session. But this is to be expected since each individual spectrum
has a relatively low SN ratio. In particular the line strengths for
each session are
December 1977 0.20 0.07 K
July 1978 0.20 0.08 K
December 1979 0.075 0.028 K
where the quoted errors are 1 standard deviation. In view of the
agreement here, and the pointing uncertainty during the December 1979
observations, the discrepancy among the observations is not large.
During the December 1979 observations the cooled receiver operated
with two independent channels, linearly polarized and orthogonal to
one another. The feature does not appear in the spectrum from the

TEMPERATURE K)
(XI O1 )
j j IJ IJ ilij II l_l [ m m IIII
VELOCITY (KM/SECHXio1 I
Figure 3. The spectrum of Figure 2 smoothed with a 5 channel (13 km/sec) rectangular
function.

TEMPERATURE (K)
(XIO1 )
-f=>
I
Figure 4
The 12C0 data from December 1977.

EM PERA TURE (K)
-0.60 -0.20 0.20 0.6,0
(XI o1 )
Figure 5. The ^CO data from July 1978.
-45-

TEMPERATURE ¡K)
tX 101 )
JO IJIJ 4;U U U i O U U £_j U U In. II M
VEL0C ITT (KM/SEC) (X I O1 )
Figure 6. The 12C0 data from December 1979, channel A.

TEMPERATURE (K)
-0.20 -0.12 -O.O14 0.04
(X ID1 )
33-58.00 -ts.no -38.00 -28.00 -18.00 -8.00 2.00 12330
j ij m IJ IJ C. O U U J. 1 1 M |J
VELOCITY (KM/SEC) (X101 )
Figure 7. The ^C0 data from December 1979s channel B.

-48-
second channel of the receiver (Figure 7). We have no physical reason
to expect a polarized feature and indeed, this transition has never been
observed to be polarized in our own galaxy. Also, the tracking procedure
rotates the observed polarization planes with respect to the sky; even if
the signal were linearly polarized we should be able to detect it in each
channel. The lack of a feature in this channel is almost certainly due
to the low SN ratio.
A further complication during the December 1979 session was poor
pointing corrections for the telescope, especially in regions far from
the celestial equator (the declination of NGC 185 is 48). Consequently
we feel that the lack of a feature in the second channel is within the
statistical and instrumental uncertainties of the experiment.
Negative Results
A total of 11 early-type systems were observed during the three
observing sessions. Our criteria for selecting candidates for CO
emission were based on the notion that a system able to retain HI or
dust would be more likely to have significant star formation.
Consequently we chose several of the early-type systems that have
been detected in HI. Added to the list were galaxies that are close
(< 15 Mpc) and also show some form of an ISM. Often this was a nota
tion in the Reference Catalog of Bright Galaxies (RCBG, de Vaucouleurs
and de Vaucouleurs 1964) that the system contained obscuring matter.
Table 2 lists the 11 galaxies selected for the CO study. Included
are the positions observed, the assumed distance and radial velocities
as well as the Holmberg magnitude and the total luminosity. The dates
in which useful data were collected are also noted. The final quantity

Table 2
12
Parameters for Early-Type Galaxies Searched for CO
NGC #
a( 1950.0)
6(1950.0)
Type
D
V
0
mHo1m
Lpg
Dates Observed
1 Standard
Deviation
(K)
(Mpc)
(km/sec)
(1o9lq)
Dec
1977 July 1978 Dec 1979
185
(DP-I)
00h36m10^3
4803
'34"
E3p
0.69
-245
11.0
0.0267
X
X X
Detected--
see text
185
(DP-11)
00 36 12.2
i
48 03
27
X
0.025
205-11
00 37 40.7
41 25
33
E5p
0.69
-239
10.1
0.0611
X
0.041
205-12
00 37 41.4
41 24
05
X
0.026
404
01 06 39.1
35 27
08
SO
1.5
-35
11.3
0.0955
X
0.067
1052
02 38 37.0
-8 28
05
E4
11.0
1439
11.1
6.18
X
X
0.069
2685
08 51 41.7
58 56
01
SBOp
12.0
868
12.0
3.0
X
X
0.042
3226
10 20 42.9
20 09
11
E2p
16.5
1356
12.3
4.6
X
0.062
4150
12 08 01.2
30 40
53
SO
3.8
970
12.2
0.267
X
X
0.043
4278
12 17 35.7
29 33
33
El
9.6
659
11.2
4.29
X
X
0.057
4636
12 40 17.3
02 57
42
EO
12.5
979
10.4
15.2
X
0.089
5846
15 03 56.5
01 47
50
EO
16.5
2353
11.1
13.9
X
0.140
5866
15 05 07.4
55 57
18
SO
13.8
692
11.0
10.7
X
X
0.036
-49-

Table 2
Continued
Notes: The positions for NGC 185 and 205 were measured from the Palomar Sky Survey. The
positions for the other galaxies are from Gallouet and Heidmann (1971), Gallouet
et al. (1973), or Gallouet et al. (1975). The galaxy types and heliocentric
velocities are from de Vaucouleurs et al. (1976). The Holmberg magnitude is cal
culated as described by Gallagher et al. (1975).
I
cn
O

-51-
1 isted in the table is the standard deviation for all observations of a
particular source.
Figures 8-18 show the spectra obtained with the 256 MHz filterbank.
Also indicated in the figures are the radial velocity assumed for each
source. In many of the spectra there are single channels above or below
the 3 standard deviation limit. These are not significant because they
fail to persist in further observations. This type of interference is
relatively common in multi-channel filterbank receivers, and arises from
stray voltages within the equipment. The implications of the negative
results are discussed in Chapter IV.
Neutral Hydrogen
The region surrounding NGC 185 was examined to sensitive limits
to detect any neutral hydrogen (HI) associated with the galaxy. The
43 m telescope at Green Bank was used in a total power mode as de
scribed in Chapter II. The arrangement of "offs" was varied throughout
the course of the observing in an effort to delineate the HI detected.
One series of observations used an "off" at about 16m west to allow a
reasonable estimate of the total HI content of the region.
Another sequence of observing used closely spaced "offs" (24' or
48') to the west, east, south, and north to get a better idea of the
behavior of hydrogen around a specific region. For those observations
the "on" was always at the center of the "offs."
Neutral hydrogen was found at significant levels in several re
gions, both at the position of NGC 185 and several points to the north.
Figure 19, which used a distant "off," shows a strong signal of about
14.2 milliKelvin (mK) at a radial velocity of -200 km/sec. The

TEMPERATURE (K)
-0.10 0.00 0.10 0.2,0
(X101 )
The ^CO data taken in July 1978 on dust patch no. 11 of Hodge (1973) in NGC 185.
Figure 8

TEMPERA TURE CK)
.10 -0.05 0.00
(XI o1
)
2-58.00 -48.00 -38.00 -28.00 -18.00 -8.00 2.00 12
O
Ln
r~i
O
O
.bo
12
The CO spectrum obtained in December 1979 on dust patch no. 12 of Hodqe (1973)
in NGC 205.
Figure 9.
-0.05 0.00

TEMPERATURE (K)
(X101 )
12
Figure 10. The CO spectrum for NGC 404. The negative feature at 200 km/sec is probably
not real, but rather an artifact of the multichannel filterbank.

(XI O1 )

TEMPERATURE K'l
UTS. i
¡3.0
93.00
CD
CO
CD
I
i i
CO
CD
I
' r 1 1- ^ - l - i - 1 ~
256 MHZ ( _
1 6 4 0 U M I N I n 1 j [ (j (j ,j
i
i ¡iUlJ1 lj\ A JHMi,i )H
tj]Ak jlLj .(i) ,:
U,/
L
If
If
W| v* njf/'f \h'
f r
V <
tflvh '
s' 1 II
i i.. i
RA B H 51
DEC 58 56 j
RMS = 41.69
i- i ~
41
:'3
H K E L.'
/1 M
63.
73.00 03.00 9 3. 0 0 108.0 0
VELOCITY (KM/SEC) CXI O1 J
113.
I
CD
03
I
Figure 12. The 12C0 spectrum for NGC 2685.

TEMPERA TURE (K)
(XI O1 )
. The ^CO spectrum for NGC 3226,
Figure 13

TEMPERATURE CK)
1.00
11.00
11 .00
-9.no
f X101 )
21.00 31.00
VELOCITY (KM
Figure 14. The 12C0 spectrum for NGC 4150. De Vaucouleurs and de Voucouleurs (1964) give
a radial velocity of 244 km/sec, but de Vaucouleurs et al. (1976) list it as
970 km/sec.
I
on
CO
i

TEMPERAT U R E (K)
42. 00
52.00
C X101 )
62.00 72.00
92.00
lofc'ur
VELOCITY (KM/SEC) (X1D1 )
Figure 15. The 12C0 spectrum for NGC 4278.
I
c_n

TEMPERATURE CK)
-0.60 -0.20 0.20
(X101 )
Figure 16. The ^C0 spectrum for NGC 4636.
-60-

(XI O1 )
12
Figure 17. The CO spectrum for NGC 5846. De Vaucouleurs and de Vaucouleurs (1964) list
a radial velocity of 1771 km/sec, but de Vaucouleurs et al. (1976) list it as
2353 km/sec.

TEMPERATURE (K 3
(X101 )
Figure 18. The ^CO spectrum for NGC 5866.

Figure 19. The 21 cm HI spectrum on NGC 185. The off position is indicated in the lower
right. The features between -100 and 0 km/sec are local hydrogen.

-64-
position of the "on" is at NGC 185. A full-width at half-maximum
(FWHM) for the feature is about 60 km/sec.
It is apparent though, from Figure 20, that all of this HI may
not be associated with NGC 185. The "offs" for Figure 20 are taken at
either 24' or 48' (they have been mixed to increase the SN ratio for
this spectrum) in the four cardinal directions. The fact that the signal
is only 5.2 mK indicates that some, but not all, of the line has been
subtracted out in the "offs."
Further, spectra were generated separately for each of the four
directions. Those to the west, east, and south show about a 3
standard deviation positive signal. However, the spectrum with an
"off" to the north shows no significant feature in the NGC 185 velocity
range.
Continued examination of points to the north of the galaxy re
veals a strong 40 mK feature about 70' directly north. The radial
velocity of this feature is -180 km/sec with a RWHM of 40 km/sec.
Five point maps of the region to the north also indicate HI signals
as far as 2 north and 48' east with a radial velocity of about -180
km/sec. There is only weak HI emission at 49' north of the galaxy.
This is shown in Figure 21; the feature is around 15 mK. It appears
that there is a ridge or plume of HI that is strongest (43 mK) between
l-2 directly north of NGC 185. It extends further north and east at
a level of around 20 mK.
To the south towards the galaxy the plume decreases in strength to
about 10-15 mK. From the sensitive 5-point maps at the position of
NGC 185, an excess of about 5 mK is found for the galaxy's location.
These observations are summarized in Table 3. The interpretation of
these observations is found in Chapter IV.

TEMPERATURE CK) (
(X101 )
-100.00 -80.00 -60.00 -40.00 -20.00 0.00 20.00 40 0)0
1 1 1 1
10 MHZ
1 3 6 6 0 0 M I N h
1
A.
1 1 i
4A
n
i 1
r i q a
L- 1 ij vj
. 7 S 1(1 IhiiOH
1 s
AAA A A.. a, a \ ,
/ v \ / v v v V \ / 1
l' ¡1
1
30 O TO
m
AO
re
ts
1 0 11 36M 11? 3
M 48 3 1 4426
= 3.18 HKELV1N
I 1 .
00.00 -80.00 -60.00 -40.00 -20.00 0.00 20.00 40. llO
VELOCITY (KM/SEC) (X101 )
Figure 20. The HI spectrum of NGC 185 with offs 1-2 beam widths to the west, east, south
and north.

TEMPERATURE (K) CX10"1 )
-0.10 0.00 0.10 0.2,0
(X101 )
Figure 21. The HI spectrum at 49' north of NGC 185.

Table 3
HI Observations in the Vicinity of N6C 185
ON OFF Full-width Time
Posi
tion of
observat'
ion
observation
Standard
Velocity
Half-maximum
on
0!\
1 wrt
Temperature Deviation
(km/sec)
(km/sec)
Source
NGC
: 185
a(1950.0)
6(1
L950.0)
a
(1950.0)
6(1950.0)
(mK)
(mK)
5 km/sec
3 K/sec
(min)
NGC
185a
OO1
h36
*US.3
48
303
'45"
oo1
n19m43?9
48
5031
44"
14.4
2.0
-190
63
378
49'
Northb
00
36
11.4
48
52
38
00
19
27.8
48
03"
44
15.7
2.3
-175
35
360
NGC
185c
00
36
11.4
48
03
44
00
31
19.6
48
03
44
6.1
2.0
-200
40
344
00
33
45.0
48
03
45
NGC
185
00
41
3.2
48
03
44
6.3
1.9
-200
45
344
00
38
37.7
48
03
44
NGC
185
00
36
11.4
47
14
50
6.3
1.9
-200
45
344
00
36
11.4
47
39
17
NGC
185
00
36
11.5
48
52
38
<5
2.6
--
--
344
00
36
11.4
48
28
12
120'
North
00
36
11.3
50
03
44
00
41
16.1
50
03
44
-17.3
5.1
-200
38
60
00
36
11.4
49
14
50
-34.6
4.7
-180
33
60
71'
North
00
36
11.4
49
14
50
00
19
20
48
03
44
39.8
8.6
-180
33
13
Notes: These measurements are from spectra reduced using a third order baseline and hanning smoothed. The
spectra on NGC 185 with two "off" positions are from data taken at both "off" positions. Approxi
mately 27% of the integration time was with the first "off," the balance with the second "off."
Reproduced as Figure 19.
^Reproduced as Figure 21.
Q
The sum of this and the three spectra following is reproduced as Figure 20.

CHAPTER IV
DISCUSSION
In this chapter the implications of the observations presented in
Chapter III are considered. The simplest calculations from these data
involve assumptions about the excitation temperature and optical depth.
The portion of the dissertation based on such assumptions is found in
this chapter.
NGC 185
This dwarf elliptical galaxy is usually considered to be a com
panion of M31. It is located about 6 to the north of M31 and the
difference in radial velocities is about 55 km/sec. For a discussion
of the virial mass of the M31 system and comments on membership in
that system see Rood (1979). A distance of 0.69 Mpc is assumed for
NGC 185 throughout the following discussion. At this distance 1" =
1.04 x 10 y cm = 3.3 pc.
The classification given by de Vaucouleurs et al. (1976) is E3p.
A detailed photographic study of the system by Hodge (1963) shows
clearly that the ellipticity (defined by e = 1 ) varies from 0.18
near the nucleus to 0.26 for the outer contours at 350" from the
nucleus. An increase in ellipticity with radius is commonly observed
for elliptical galaxies. The E3 classification (which is based on the
outer contours only) is confirmed by Hodge's photometry.
-68-

-69-
The system is peculiar because of the presence of two dark dust
patches within about 20 arcseconds of the nucleus. Figure 22 is an
optical photograph of NGC 185 and a schematic of the dust regions is
reproduced from Hodge (1963) in Figure 23. Careful comparison of the
two figures reveals the extent of the dust in the optical photograph.
Table 4 tabulates the positions of the patches used for observation
2
and the areas in square arcseconds and pc as measured from Figure 23.
Note that dust patch I (DP-I) has 2 entries, la and I. DP-Ia is the
dark core of the northwest patch and DP-I is the larger, less dense
patch (but also including DP-Ia).
Table 4
Positions, Areas, and Telescope Filling Factors for
Dust Patches in NGC 185
Object
a(1950.0)
6(1950.0)
Area
(-)2
PC2
Fi 11ing
Factor
DP-Ia
0h36m10^3
4803'54"
35
380
0.0072
DP-I
Same as
DP-Ia
115
1250
0.0235
DP-I I
0h36m12?2
4803128"
350
3800
0.0716
DP-I+II
(Center of
NGC 185)
0h36ml1?4
4803'44"
465
5050
0.0951
The CO
observations were
taken at the
position
of DP-I
in December
1977 and July 1978. The December 1979 observations were made at both
DP-I and -II as well as the center of the galaxy. However, as the
FWHM for the 11 m telescope indicates in Figure 22, observations of one
patch do not entirely exclude the other. Further, the pointing for

-70-
t
I I
30"
%
Figure 22. NGC 185 showing the dust patches. North is up and west is
to the right. The half-power beam-width is indicated.
This photograph is reproduced through the courtesy of Lick
Observatory.

-71-
\
MAJOR
AXIS
gure 23. A schematic of the dust regions near the nucleus of
NGC 185 (Hodge 1973). DP-I is northwest of the center,
DP-II is southeast and DP-Ia is darker reqion within
DP-I.

-72-
December 1979 is unreliable to perhaps 20 arcseconds and at times
30-40 arcseconds. Consequently we have no confidence in the apparent
identification of the CO source as DP-I based on the absolute pointing.
Because of the uncertainty of the source for the CO emission, the
following analysis considers three cases for the calculations of column
and particle densities. The individual cases and the assumed CO sources
are
Case A DP-Ia
Case B DP-I
Case C DP-I+II
To estimate the kinetic temperature of the molecule we can assume
the clouds are optically thick (t 1). Observations of CO sources
within our galaxy indicate that this is often the case with the J = 1 0
transition (Zuckerman and Palmer 1974). We find (from Appendix II)
that
ex
(1)
where is the forward beam coupling efficiency (also known as the
filling factor or the beam dilution) from Ulich and Haas (1976). Since we
have no detailed information on the brightness distribution we will assume
that 0.7 (source area/beam area). The beam area is found from the
65" half-power beamwidth and the factor of 0.7 arises because 30% of the
power enters the antenna through the very broad error pattern. Given
these assumptions Table 5 gives the expected antenna temperature for the
three cases and several excitation temperatures.
Clearly we cannot distinguish between the several possibilities
in Table 5 which are consistent with the detected temperature of 0.081 K.
But if the CO is optically thick it is unlikely to be above 25 K or to
originate in both the clouds.

-73-
Table 5
Expected Antenna Temperatures (K) for x 1
Source
Tex = 10 K
25 K
100 K
CO
o
11
Case A
0.071 K
0.179
0.714
7.14
Case B
0.235
0.588
2.35
23.5
Case C
0.951
2.38
9.51
95.1
Another approach is to compare the dust clouds observed in NGC 185
to clouds more easily observed in our own galaxy. The disadvantage
is that it is not certain that the nature of the clouds will be the
same. It does, however, provide a reference point for their study.
The extinction in the dust clouds was measured by Hodge (1963).
He finds OP-I to have a mean visual absorption of 0.3 mag but as high
as 1 mag in certain parts. The mean for DP-II is measured as 0.15 mag
of absorption in the visual.
At this level of absorption the clouds are similar to the diffuse
clouds studied by Knapp and Jura (1976). Their study involved clouds
situated in front of stellar sources that show color excesses of
E(B V) £ 0.3 mag. Assuming a normal Av/E(B V) = 3 indicates
A 'v 1 mag, similar to the absorption in the NGC 185 clouds.
12
Knapp and Jura suspect optically thin CO emission based on their
13
inability to observe CO at an intensity of 2-6 times less than the
12 12
CO (usually possible if CO is optically thick), and they found
antenna temperatures of 1-2 K even though the sources probably filled
the beam and had kinetic temperatures of 20 K or more. Using the thin

-74-
cloud models of Lucas (1974) and assuming a kinetic temperature of
20 K (Morton 1975) they find that the column density is given by
N(CO) ^ 6.0 x 1014(T*/nf)Av (2)
where AV is the full-width half-maximum of the line in km/sec. Using
this expression and taking into account beam dilution (nf) we find
column densities for NGC 185 as listed in Table 6.
Table 6
Column density of 12C0 for Optically Thin, Diffuse Clouds
Source
N(CQ)
cm-2
N(C0) x 2 x 104 =
cm-2
n(h2)
Case A
8.1 x 1016
21
1.6 x 10
Case B
2.5 x 1016
5.0 x 1020
Case C
6.1 x 1015
1.2 x 1020
The mass of the molecular cloud can be found by multiplying the
column density of first by the area of the cloud and then by the
mass per particle. The result is 9680 M0 and is independent of which
case is actually correct. It is independent because we are assuming
that the source is optically thin and we are detecting all the CO
within a given region. Further, it represents a firm lower limit if
the colliding particles are if they are HI the minimum mass is
4840 M .
0
The last column of Table 6, total column density of H2, is found
by assuming N(C0)/N(H2) = 5 x 10^(Martin and Barrett 1978).

-75-
A further finding by Knapp and Jura (1976) is that a necessary,
but not sufficient, condition for CO in emission is a particle density
3 -3
£10 cm This agrees with the theoretical result noted in Appendix
II. We thus feel confident that the CO emission in NGC 185 is
3 -3
originating in a region at least as dense as 10 cm .
It would be useful for calculations of the space density of various
particles if the physical size of the absorbing regions were known.
This would help constrain the various possible sources of the CO emission
and help in determinations of the mass of gas involved. An estimate
of the cloud volume can be made by assuming they have a dimension along
the 1ine-of-sight equal to the average width. This approach will tend
to underestimate the volumes unless the clouds are flat and oriented
broadside to the 1ine-of-sight. Table 7 tabulates the result of this
calculation. The gas mass listed in the last column is found by
assuming the entire cloud is filled with molecular hydrogen at a
density of 108 cm-8.
Table 7
Estimates of Cloud Volumes for NGC 185 Dust Clouds
Dust
Cloud
Depth of Cloud
(pc)
Volume
(cm3)
(pc3)
Mass of Gas
(M )
DP-Ia
6.6
7.37 x 1058
2.5 x 102
6.2 x 104
DP-I
16.5
6.06 x 1059
2.1 x 104
5.1 x 105
DP-11
42.9
4.79 x 1060
1.6 x 105
4.0 x 106
Hodge (1963) has estimated the mass of dust in DP-I and DP-II.
Using the dust masses and the total gas content for the optically thin

-76-
case (Table 6) and the minimum calculated gas mass from Table 7 allows
a calculation of the gas-to-dust (G/D) ratio. The results appear in
Table 8.
Table 8
Gas-to-Dust Ratios
Cloud
Dust Mass
Gas Mass
* < 1 (M0)
G/D
Gas
T
Mass
1
G/D
DP-I
25
9680
390
5.1 x
105
2 x 104
DP-I+II
215
9680
45
4.5 x
106
2 x 104
Considering that the gas-to-dust ratio in our own galaxy is usually
quoted as 200, the most attractive source for the CO emission appears
to be DP-1.
Using the minimum mass consistent for DP-Ia (Table 6) allows a
calculation of the Jeans, length for gravitational instability
irkT
-4G
PP
= i.6[-
TKR£c]i
M J

(3)
where Mq is the mass of the cloud in solar masses. Taking the width
of the cloud as a characteristic dimension, R = 6.6 pc, T^, = Tgx =
20 K, and M = 4.84 x 103 H
0
Aj = 1.1 pc
Since the cloud is substantially larger than the Jeans length,it is
most likely in a state of collapse.

-77-
The Jeans mass can be calculated under these conditions from
M TT
MJ 6 po
n in23
Aj ^ 10
(TK/y)
3/2
1/2
gm
(4)
with T = 20 K, y = 2, p = 3.34 x 10~^ g/cm3 (corresponding to 103 cm
of H2) we find Mj = 27 M If the cloud remains isothermal as it
collapses, and this appears likely because of the effective cooling,
Mj decreases and fragmentation is likely to occur.
Another calculation that confirms the collapsing nature of the
cloud in NGC 185 is provided by Rowan-Robinson (1979). He estimates
the cloud density, in an essentially virial calculation, by the
equation
n(H2) = 103*90 n (Av)2 r"2 K^1 (5)
where n is 0.8 for a cloud in equilibrium, r is the radius of the cloud
18
in units of 10 cm, and is a dimensionless quantity related to a
hot core for the cloud (it is always greater than 1.5).
The density calculated for the NGC 185 cloud, assuming 6.6 pc
for r (the characteristic width for DP-Ia) is 3.3 x 103 cm-3. If the
2 -3
cloud is in free-fall n = 0.2 and the derived density is 8.3 x 10 cm .
Thus considering that the NGC 185 cloud masses are uncomfortably large
(Table 7 ) for higher particle densities, we consider the lower density
more likely.
From these arguments it appears that either DP-Ia or DP-I is most
likely in a state of collapse. Further, the lack of emission from
ionized gas (Humason et al. 1956) implies one of two possibilities.
The first is that star-formation ceased in the system long ago, such

-78-
that the HII regions associated with massi.ve stars have since dis
sipated.
c
Considering that typical lifetimes for HII regions are 10 years
and Hodge (1973) estimates the age of the OB complex in NGC 205 (very
similar to that of NGC 185) to be 5 x 10 years, it is certainly
plausible that the early HII regions have evaporated. Since the time
5
for the observed clouds to collapse to stars is about 10 years, it
seems that the system may be experiencing bursts of star formation
similar to that envisioned by van den Bergh (1975).
However, there is a serious flaw in this scenario. Hodge (1963)
estimates that the population I stellar component of NGC 185 is about
5
2 x 10 M|g. Using Faber and Gallagher's (1976) mass loss rate of
0.015 yr-1 (io\0)_1 and the luminosity of NGC 185 (2.67 x 107 L ),
O
we find that it took 5 x 10 yrs for the material to collect. During
6 3
the 5 x 10 years since the last star-forming episode only 2 x 10 M0
of gas should have been able to appear. Our CO observations show at
least 5 x 10 M0 (for optically thin CO, assuming atomic hydrogen for
the colliding particles) and it is likely that there is actually more.
An even stronger case is made if the total mass of dust (215 Mq)
is multiplied by an assumed gas-to-dust ratio of 200. The system is
4
seen to contain 4.3 x 10 M0 of gas, more than 10 times the amount
from stellar mass loss.
The second possibility that may be occurring in NGC 185 is con
tinual star formation but with no 0 or B stars, and consequently no HII
regions. It is very difficult to prove that this is the process taking
place. If indeed low-mass stars are being formed they will be too
faint to see and deductive reasoning is necessary to support this

-79-
possibility. Specifically, since the regions observed are dense
enough to form stars yet no HII regions are observed, either no stars
have formed yet (leading to the burst hypothesis) or only low-mass
stars are forming.
The idea of a skewed mass distribution for stars is explored in
some details by Mezger and Smith (1977). They find that clusters of
massive stars (which produce giant HII regions) are found only along
spiral arms. It is probable that the spiral shock front of a density
wave triggers their formation. It appears though, that the low-mass
stars can form out of small but dense clouds (Herbig 1970) that are
far more widespread than large clouds.
Jura (1977) also proposes that the expected radiation environment
of elliptical galaxies (low UV flux) will lead to low-mass star forma
tion. The reasoning is that the low flux results in less heating of the
clouds. A cooler cloud will have a lower critical mass which can
separate out and collapse to form a star.
The dynamics of the nuclear region will be considered in the fol
lowing paragraphs. The difficulties in measuring radial velocities for
early-type galaxies are immense. There are usually no HII regions in
these galaxies so their sharp emission lines are not available for
measurement. The absorption lines from the stars are intrinsically wide
and further broadened by the stellar velocity dispersion, resulting in
large errors.
The radial velocity data for NGC 185 are actually better than that
for most early-type galaxies. Ford et al. (1973) identified four
planetary nebulae in the system and later Ford et al. (1977) succeeded
in measuring the radial velocities of two of these objects, NGC 185-1 and

-80-
-2 in their notation. These measurements, coupled with the radial
velocity of DP-I, can be used to calculate a rough mass interior to the
object and, when combined with the isophotal contours of Hodge (1963),
the mass-to-luminosity ratio (MLR) can be computed.
Table 9
Positions and Radial Velocities of Points Within NGC 185
Object
Radial
Velocity
(km/sec)
Distance from
Minor Axis
(pc)
Distance from
Major Axis
(pc)
185-1
-212.5
144.2
75.6
185-2
-200
107.3
103.6
DP-1
-175
0
43.9
Table 9 lists the three radial velocities determined for the system,
as well as the projected distances from the major and minor axes. We
make the simple assumption that the ellipticity is caused by rotation
about the minor axis, thus the figure of the system is an oblate spheroid.
We further assume that the radial velocities are the projected results
of purely circular velocities about the nucleus. We can then make a
straightforward calculation of the mass interior to the measured point.
Implicit in this approach is the assumption that the points lie in the
plane of the galaxy.
Recalling the work of de Vaucouleurs (1977), described in Chapter I,
on the relative frequencies of various ellipticities, we can make an
informed guess of the true ellipticity of NGC 185. The galaxy is most
likely to be an E3.6, but we can be more flexible and also consider

-81-
the possibility that it is an E5.5 (the flattest elliptical according
to de Vaucouleurs 1977). Of course the system may be an E3 seen
face-on, but the velocity gradient for the three measured points strongly
suggests rotation.
The details of the determination of the inclination angle (i,
defined to be the angle between the plane of the galaxy and the plane
of the sky) and its effects on other parameters are found in Appendix
III. Under the assumption that the system as is an E3.6 or an E5.5 the
inclination angle is 53.1 or 68.0, respectively.
The azimuthal angle, 6', can be deprojected to find the true azi
muthal angle, e. The results are contained in Table 10 for the two
planetary nebulae. Since DP-I lies on the minor axis we assume it has
no radial velocity due to rotation; consequently we take its radial
velocity, -175 km/sec, as the systemic velocity for the galaxy.
Table 10
Values for the Observed and True Azimuthal Angles
Object
e1
(degrees)
0 (degrees)
i = 53.1
i = 68.0
185-1
26.8
40.1
53.4
185-2
44.0
58.1
68.8
The true circular velocity for the nebulae can thus be calculated from
vc = vr/,cos 9 S1'n 1
(6)

-82-
The true radius can also be computed and is listed with the velocities
in Table 11. Finally, the mass interior to the point, assuming Kepler-
ian motion, is given by
M = 2.36 x 102 v2 r (7)
where M is in solar masses, v is the circular velocity in km/sec, and
r is the true radius in parsecs.
Table 11
Circular Velocities and Radial Distances in NGC 185
Object
vr
vc (km/sec)
r (pc)
km/sec
i = 53.1
i = 68.0
i = 53.1
i.
II
O'!
00
o
o
185-1
37.5
61.3
67.8
191.4
248.0
185-2
25
59.2
74.6
203.2
296.6
The isophotal contours of Hodge (1963) were measured to find the
luminosity interior to a given point. The contours measured were the
innermost (1) and the next brightest (2), corresponding to 20.47 and
2
21.22 mag/arcsec respectively. Using a distance modulus of 23.9 and
assuming absorption within our own galaxy according to the cosecant law
(Av = 0.26 esc b = 1.0 mag for NGC 185) allows the calculation of the
luminosity interior to the first and second contours, listed in
Table 12.

-83-
Table 12
Luminosities in the Nucleus of NGC 185
Contour
Brightness2
Area ^
Luminosity
(mag/arcsec )
(arcsec )

1
20.47
1690
1.03 x 107
2
21.22
7715a
2.35 x 107a
aExcludes contribution interior to contour 1.
The mass calculated from equation 7 is then divided by the appro
priate luminosity from Table 12 to derive the MLR interior to each
object. The Planetary nebulae, 185-1, lies on the second contour,
making the luminosity interior to it the sum of the luminosities inside
contours 1 and 2. The other nebula, 185-2, lies midway between con
tours 1 and 2 so we use the luminosity inside 1 plus half the luminosity
inside 2. Table 13 lists the results for the MLR derived with these
values.
Table 13
Mass-to-Luminosity Ratios for the Nucleus of NGC 185
Object Mass Interior (Ml Luminosity MLR (solar units)
5 (L0)
i = 53.1 i = 68.0 u i = 53.1 i = 68.0
185-1 1.7 x 108 2.7 x 108 3.38 x 107 5.0 8.0
1.7 x 108 3.9 x 108 2.21 x 107
185-2
7.7
17.6

-84-
While these values for the MLR are very reasonable for early-type
galaxies in general (Faber and Gallagher 1979), there is a significant
disagreement with the MLR of 1.8 used by Ford et al. (1977) for NGC 185
and NGC 147. This MLR is ostensibly derived from observations of M32.
We feel that the most likely cause for the discrepancy is that Ford
et al. (1977) do not consider possible rotation in their analysis.
The radial velocity of DP-I is most likely nearest the systemic velocity
of NGC 185 since it is fortuitously on the minor axis.
We find that the MLR of the nuclear regions of NGC 185 is between
5 and 18. Considering the non-linear effects of the inclincation and
azimuthal angle, the most likely value for the MLR is around 8.
Assuming that it is independent of radius, the total mass of NGC 185
is found to be 1.3 x 10^ M .
9
NGC 205
Another dwarf elliptical companion of M31, NGC 105, was also observed
to sensitive limits for CO emission, with negative results. Figures 8 and
9 show the averaged spectra from the July 1978 and December 1979 sessions,
respectively. The spectra are taken at two different positions within
the galaxy (dust patches 11 and 12 in Hodge's (1973) notation). Hodge
(1973) describes and diagrams the dust content of NGC 205. The distri
bution is reminiscent of NGC 185 but with a larger number of discrete
clouds over a wider region in the nucleus. The July 1978 observations

-85-
were taken on dust patch number 11 while the December 1979 observations
were of dust patch number 12 on the other side of the nucleus. Table 14
lists the positions observed for each patch as well as Hodge's (1973)
estimates of the dust content. The dimensions listed by Hodge are
used to find a rough area in square arcseconds and pc (the assumed
distance is 0.69 Mpc). Hodge (1973) also presents microphotometer
tracings of these dust regions indicating a visual absorption of about
0.2-0.3 magnitudes. It is clear that these regions are not as dense
as DP-la in NGC 185. Beyond this difference the two galaxies are
very similar in dust and young star content.
Table 14
Dust Regions in NGC 205
Region
no.
a(1950.0)
6(1950.0)
Area
()2
of Cloud
pc2
Dust Mass
Me
11
0h37m40!7
4125133"
360
3.9 x 103
160
12
0 37 41.3
41 24 05
880
9.6 x 103
460
A similar calculation as was made for the NGC 185 dust clouds,
under the assumption of optically thin emission (see equation (2) of
this chapter), can be performed. The antenna temperature used is 3
times the standard deviation for each cloud. The velocity width is taken to
be four times the channel width (av =10.4 km/sec). The results appear
in Table 15. The column densities are modest when compared to those
of NGC 185 (Table 6). The computed total masses are even less than
one may expect from the dust masses and a gas-to-dust ratio of 200,

-86-
3.2 x 10^ and 9.2 x 10^ Mq for clouds 11 and 12, respective
ly.
Table 15
Maximum Column Densities of CO and H^ for NGC 205, t < 1
Cloud no.
N(C0]
(cm_2)
n(h2)
(cm-2)
M (molecular)
11
15
7.0 x 10iD
1.4 x 1020
4.4 x 103
12
2.8 x 1015
1 9
5.6 x 10iy
5.2 x 103
The situation does not change if one assumes the clouds are optically
thick. The NGC 205 clouds are relatively large compared to those in
NGC 185, and any reasonable kinetic temperature will produce a large
antenna temperature, as shown in Table 5, very easily.
The question is then, why is NGC 205 not detected while NGC 185
with similar morphology (and actually less dust) is detected at the
frequency of 12C0?
We feel that the key difference in the two systems is that the
NGC 205 clouds are less dense than those in NGC 185. Hodge's (1963 and
1973) microphotometer tracings show this, and even optical photographs
of the two galaxies show the NGC 185 dust patches to be more prominent.
O O
As Knapp and Jura (1976) found, even a density of 10 cm is not
12
sufficient to ensure detection of CO in emission. Further, their
observations were made on nearby clouds where one could reasonably
expect a filled beam. The NGC 205 observations involve substantial
beam dilution which considerably worsens prospects for detection.

-87-
While many of the arguments about continuing star formation in
NGC 205 can be advanced with much the same reasoning as those for NGC
185 in the previous section, much of this support is lost because we
cannot be sure the NGC 205 clouds are in a state of collapse. However,
it is very likely that they will eventually form stars. This is clear
because NGC 205 has an even larger young stellar component than NGC
185; Hodge (1973) calculates some 2 x 10 for the young stars.
We seem to be seeing an earlier stage of the process in an ellip-
5
tical galaxy which converts the ISM into stars. In a few times 10
years the dust clouds in NGC 205 will be much more condensed and CO
emission should be detectable. In the future stars will be forming and,
if these are high-mass stars, the galaxy will retain its unusual popu
lation I component.
Negative CO Results
12
The galaxies that were not detected at the CO J = 1+0 transition
are listed in Table 2. Several attributes are also listed, along with
the limits of our observations. These limits can be interpreted on the
basis of two assumptions about the CO.
First is that the gas is optically thin and the emission was not
detected because of beam dilution or an exceedingly small optical depth
(t 1). Second, if the line is optically thick then beam dilution is
the only mechanism to lower the antenna temperature below our detection
1imits.
Another variable in the interpretation of these negative results
is the width of the undetected CO line. If it is several hundred km/sec
broad, representing the contributions of many small clouds within the

-88-
beam, it is more difficult to detect than if it were a narrow feature
arising from only a single, larger cloud.
If t >> 1 no information can be gained about the amount of molecu
lar material stored. Essentially any quantity of material can be
stored in clumped, optically thick clouds. The material remains un
detected because of the very small filling factor (n^). Consequently
Table 16 lists various parameters under the assumption that the CO
clouds are optically thin.
12
Knapp and Jura's (1976) finding that the CO line is probably
thin if the visual absorption is less than 1 magnitude indicates that
most CO in elliptical galaxies will be thin since deep absorption
features are rare.
Equation (2) from this chapter is used along with the assumption
of n.p = 0.05. The antenna temperature is taken to be 3 times the
standard deviation listed in Table 2. The width of the undetected
feature is taken to be 10 km/sec for one set of calculations and 200
km/sec for a second set. The first width is appropriate for a single
cloud while the second is more typical of an expected global profile
for many clouds (Rickard et al. 1977).
Also listed is the column density of assuming N(C0)/N(H2) =
5 x 10 (Martin and Barrett 1978). The molecular mass is calculated
by (Schneps et al. 1978)
M = ^(Tex)(T*/nf)AvD2 (8)
where M is the molecular mass in solar masses, n(T ) incorporates the
CA
effect of temperature on the population levels of the CO, T* is taken
to be 3 times the standard deviation, av is the assumed velocity width

Table 16
Maximum Molecular Masses Derived from CO Observations
NGC #
3xrms
(K)
T < 1,
Av = 10 km/sec
T < ]
l, Av = 200
km/sec
mhi
Detected
Mr j
Expected
N(C0)
(1016 cm-2)
n(h2)
(1020 cm"2)
Mass
T=20 K
(M )
N (CO)
(1016 cm'2)
N(H2)
(10 cm
Mass
0 T=20 K
2) (M,)
205-11
0.123
1.48
3.0
1.2
X
104
29.6
59.2
2.4
X
105

2.7xl06
205-12
0.078
0.936
1.9
7.4
X
103
18.7
37.4
1.5
X
105
3 x 105a
2.7xl06
404
0.201
2.41
4.8
9.0
X
104
48.2
96.2
1.8
X
106
--
4.2xl06
1052
0.207
2.48
5.0
5.0
X
106
49.6
99.2
1.0
X
108
1.Ixl09b
2.72X108
2685
0.126
1.51
3.0
3.6
X
106
30.2
60.4
7.2
X
107
9.5x108c
1.32x10s
3226
0.186
2.23
4.5
1.0
X
107
44.6
89.2
2.0
X
108

2.02x10s
4150
0.129
1.55
3.1
3.7
X
105
31.0
62.0
7.4
X
106

1.2xl07
4278
0.171
2.05
4.1
3.2
X
106
41.0
82.0
6.4
X
107
2.5xl08d
1.88xl08
4636
0.267
3.20
6.4
8.3
X
106
64.0
128
1.7
X
108
3.2xl08e
6.69xl08
5846
0.420
5.04
10.1
2.3
X
107
101
202
4.6
X
108
5.0xl08f
6.12xlOS
5866
0.108
1.30
2.6
4.1
X
106
26.0
52.0
8.2
X
107
--
4.71xlOS
aUnwin (1980)
bReif et al. (1978)
c6allagher et al. (1978)
dKnapp et al. (1978c)
eBottinelli and Gouguenheim
(1977b)
SHuchtmeier et al. (1977)
-89-

-90-
in km/sec, and D is the distance to the source in Kpc. For
T < 3000 K the vibrational levels of CO are not populated and
ca
a 103 Tex'
Also listed in Table 16 is the mass of HI that has been detected
in the galaxy. The last column gives the expected HI content from
Knapp et al. (1978c)
mhi 4-4 x lo7 V1()9 L0>
(9)
where is in solar masses and Lp^ is the photographic luminosity of
the galaxy. This relation uses the mass loss rate given by Faber and
Gallagher (1976) of 0.015 yr-* (10^ L )* and assumes a time for
0
9
accumulation of 4 x 10 yrs, the time over which we are sure that
normal elliptical galaxies have existed (Gunn and Oke 1975). A de
crease of 25% is made to allow for an undetectable helium contribution.
Caution must be exercised in comparing the expected HI content
with the maximum molecular content as derived here. First they are
correct only if the CO is optically thin. While there is reason to
believe this is the case, it is not certain. For an optically thick
cloud only the surface is seen and no information is available about
the total mass.
Second, the CO observations are made with a 65" HPBW while the
galaxies are typically several times this size. The C0-implied esti
mates of the total gas are thus only true for the central regions,
whereas the expected HI is for the entire galaxy. Indeed, in the cases
of the detected HI the beam is usually about the size of the galaxy.
This discrepancy in the size of the regions sampled may not be
an important consideration. This is due to the difficulty of finding

-91-
a mechanism that would allow the gas to maintain the same spatial dis
tribution as the stars that shed it. As long as the gas can cool itself
effectively, usually via line radiation, it must collapse into the po
tential well of the galaxy. If it cannot cool itself, a galactic wind
sets up and the system is swept clean. The point has been made pre
viously that galactic winds are not universally effective, consequently
one expects the interstellar material to preferentially collect in the
nucleus.
Comparing the maximum reasonable amount of molecular material that
could escape detection under these assumptions (t < 1, av = 200 km/sec)
with the expected HI content reveals that molecular storage is not a
dominating feature of the galaxy's ISM. The largest discrepancy, for
NGC 205, may not be significant because the region sampled for CO is
small with respect to the dust patches.
For the other galaxies the discrepancy may be due to one, or a
combination, of three processes: the gas may be clumped and thus
optically thick, the material could have been removed either by a
galactic wind or ram-jet stripping, or finally, star formation could be
consuming the gas with an IMF deficient in high-mass, bright stars.
A fourth possibility, that the clouds are similar to Knapp and
3 -3
Jura's (1976) thin clouds with n ^ 10 cm but with no CO emission, is
unlikely. The CO observations included a large portion of the galaxy's
nucleus and it seems highly improbable that all the clouds will be in
this nether region of collapse with no CO emission. Certainly some
fraction may be without CO emission but others, further collapsed,
should produce CO emission.

-92-
It is interesting to note that the three galaxies with a detected
HI content larger than expected (NGC 1052, 2685, and 4278) are all
suspected of accretion. All three systems show peculiar dynamics for
the HI that make an extra-galactic source the most plausible explanation
(Reif et al. 1978, Shane 1980, Knapp et al. 1978c).

-93-
Neutral Hydrogen and NGC 185
As described in Chapter III, we have detected neutral hydrogen at
the position of NGC 185. There is also detectable HI at least as far
as 2 north of the galaxy. It is apparent that any HI physically
associated with NGC 185 is being confused with high-velocity hydrogen
belonging to our own galaxy.
The mass of HI in a particular feature is given by (Wright 1974)
Mhi = 2.356 x 105 D2 / S dV (10)
where the mass is in solar masses, D is in Mpc, S is the flux density
in Janskys (1 Jansky = 10 Wm Hz ), and dV is in km/sec. This
equation is based on the assumption of low optical depth in the 21 cm
line. While this assumption may not be valid in the plane of the
galaxy, it is quite likely true for HI out of the plane (NGC 185 is at
a galactic latitude of -15).
The confusion with high-velocity local HI is primarily due to the
Magellanic Stream, an arc of HI clouds stretching from the Magellanic
clouds through the south galactic pole and terminating in the vicinity
of NGC 185 (Mathewson et al. 1974).
That there are indeed discrete clouds in the velocity range around
-200 km/sec and within 4-6 of NGC 185 is confirmed by Giovanelli
(1979) and Hulsbosch (1980). Hulsbosch's observations of the area
show nothing at either the position of the galaxy or within 2 to the
north. This lack of confirmation of our findings is not surprising
because Hulsbosch's sensitivity is only about 400 mJy while ours is
about 25 to 40 mJy (the 43 m telescope has a sensitivity of about
4 Jy/K).

-94-
Based on our observations it is clear that there is HI around
NGC 185 with a distribution greater than 24'. Figure 19 shows a fea
ture with about 15 mK of HI when the "off" is located many beamwidths
to the west. When a spectrum is taken with the "offs" only 24' or 48'
to the west, east, south, and north, the strength of the feature drops
to around 5 mK. This clearly indicates that we are subtracting out
about 10 mK of broadly distributed HI.
Analysis of the individual spectra with an "off" in one of the four
cardinal directions shows that there is an excess of HI with respect to
all directions except perhaps the north. The intriguing possibility is
that while there is clearly a broadly distributed HI components around
NGC 185, the spectra show that there is excess line radiation from the
position of the galaxy when compared with areas 1 to 2 beamwidths
away.
Of course it is possible that an enhancement in the high-velocity
cloud coincidentally projects onto NGC 185. Our observations are not
sensitive enough nor spatially resolved enough to answer the question
definitively.
It is interesting to note that if the 5 mK excess does originate
in NGC 185, the mass associated with it is about 9 x 104 M This is
@
quite reasonable since the minimum molecular mass needed to generate
4
the observed CO signal is 10 Mq. Indeed, a calculation of the mass
of HI expected in the galaxy from normal stellar evolution (see equation
f.
7, in the previous section) is 1.2 x 10 Mq.
A final possibility suggested by the detection of a plume between
Io and 2 further north of the galaxy is that we are seeing a tail
or streamer of HI associated with NGC 185. Since the galaxy is a

-95-
companion of M31 an interaction of this sort may be occurring. The
difficulty is that M31 is to the south of NGC 185 and is over 6 away,
a rather larger distance for an interaction considering the small mass
of NGC 185. Thus while the intergalactic plume is an intriguing pos
sibility, it is not well-supported by our observations.
Higher spatial resolution measurements are required in the area
centered on the galaxy and at least 20-30' around. It is necessary
to examine the local features well enough to allow their subtraction
from possible HI belonging to NGC 185.
Any HI associated with the galaxy would probably have a distribu
tion less than 5' across; the galaxy itself has a major diameter of
around 10'. The VLA can easily achieve this resolution, and an in
vestigation of the NGC 185 region should be undertaken to determine
the kinematics of the HI at the position of the galaxy.

CHAPTER V
SUMMARY
The current understanding of elliptical galaxy formation and
evolution has several important shortcomings. It is not understood
how elliptical, or for that matter any, galaxies separate out from the
background and collapse. Even the general process of whether the
clusters of galaxies form before the individual galaxies is not clear.
The deceptively relaxed-looking stellar distributions are apparently
not relaxed at all. It is quite possible that the 3-dimensional figure
of elliptical galaxies is actually a triaxial ellipsoid. Further in
consistencies are apparent in the disposition of gas shed by the stars
within the galaxy. If it recycled into stars with a normal IMF the
nuclei of all early-type systems should be bluer than observed.
Recent theoretical work has shown that two gas removal mechanisms
may be present in these systems. One is a galactic wind in which
supernovae provide an energy source to heat the ISM of the galaxy to
a temperature high enough to evaporate from the system. The crucial
points of this method are the assumed Type I supernova, rate and the
efficiency of energy coupling between the expanding supernova shells
and the general ISM.
The other mechanism is ram-jet stripping. The process is basically
a hydrodynamic interaction between an assumed intracluster medium and
the ISM of the galaxy. There is circumstantial evidence for this
-96-

-97-
process; early-type galaxies are preferentially found in rich clusters
that would be most effective in retaining an intracluster medium.
However, it is not clear that the presence of early-type systems in
rich clusters is not a result of initial conditions at the time of the
cluster formation. Indeed recent work has tended to support this view.
Observations of nearby galaxies do not fit well with either
of the two mechanisms. Some have been detected in HI when the galaxy
should be able to expel its ISM. Worse, in at least two cases the HI
has a broad distribution, entirely contrary to expectations for a par
tially operating galactic wind or ram-jet stripping. The suggestion
has been made that the HI has recently been accreted, but this is dif
ficult to support based on the space density of HI clouds large enough
to account for the material. Other early-type galaxies show patches
of obscuring matter firmly indicating that mass removal processes are
not completely efficient.
12
Observations of CO were undertaken for early-type galaxies to
determine the role of star formation in the removal of the ISM. The
CO molecule was chosen because it is widely distributed, resistant to
dissociation, and has a transition frequency accessible to highly
sensitive radio telescopes.
12
The J = 1-K) transition of CO was detected in the dwarf ellipti
cal galaxy NGC 185. The physics of the line formation process strongly
imply that the emitting region is the north-west dust patch about 15
arcseconds from the nucleus. The fact that the line is seen at all
3 -3
indicates that the region is at least as dense as 10 cm The minimum
mass consistent with the observations is about 10^ M .

-98-
A region as large as this dust patch .is gravitationally unstable
and is collapsing, probably forming stars at a later epoch. The
nucleus of NGC 185 contains several bright blue, presumably young
stars which have formed in the relatively recent past.
Observations of another dwarf elliptical galaxy, NGC 205, failed
to find CO emission even though this galaxy contains a larger popula
tion of young stars and more total dust content distributed in 12
patches. The densest of the regions are not as dense as the northwest
dust patch of NGC 185. It is suggested that this is the primary reason
for not detecting NGC 205 while obtaining a positive result on NGC 185.
Negative results for 9 other early-type systems are presented and
the implications are discussed. The major finding is that the material
shed by the stars cannot be contained in optically thin clouds and
still elude detection in these observations. Alternatives are that
the gas is clumpy and thus optically thick, it has been swept out by
either a galactic wind or ram-jet stripping, or it is being consumed
by star formation with an IMF skewed towards low-mass stars.
Theoretical support for a skewed IMF is presented, but the question
cannot be answered by observations with present-day equipment. More
sensitive CO data are needed and the interaction between the intra
cluster medium and the ISM needs to be explored further.
Using radial velocities from two planetary nebulae in the system
we find convincing evidence that the galaxy is rotating about its minor
axis. The data can be further used to derive the mass interior to the
measured points. Of course some basic assumptions about the true
ellipticity of the galaxy and the effect of projection onto the plane

-99-
of the sky preclude firm conclusions about the dynamics of the
system.
We find, however, that the mass-to-luminosity ratio in the nucleus
(the planetary nebulae are both within 1 arcminute of the nucleus) is
between 5 and 18, with the most likely value around 8. This number is
quite reasonable for an early-type system, but the difficulty in deter
mining the ratio, and the lack of agreement from different methods,
prohibit any conclusions beyond those already mentioned.
Neutral hydrogen observations of the vicinity of NGC 185 are
presented. They reveal the probable existence of high-velocity hydrogen
at -180 km/sec at and near NGC 185. The observations also show an
enhancement of about 5 mK antenna temperature at the position of the
galaxy with respect to 1-2 beamwidths away. The data are inadequate
to determine if the material is in NGC 185 or is just an enhancement
in the local material coincidentally superposed on NGC 185.

APPENDIX I
CALIBRATION THEORY FOR CO OBSERVATIONS
The intent of the calibration procedure is to correct the data
for several factors that distinguish a real telescope at the earth's
surface from an ideal telescope at the top of the atmosphere. The
major correction factors are ohmic losses in the telescope, spillover,
blockage, and the radiation/attenuation of the atmosphere.
The following paragraphs describe the calibration procedure used
at the 11m NRAO telescope as described by Ulich and Haas (1976).
The method of choice to calibrate mm-wavelength observations is
to use a rotating chopper which alternately covers the feed horn with
an ambient temperature microwave absorber. The method is attractive
because it automatically compensates for changes in atmospheric ab
sorption (Penzias and Burrus 1973).
The ambient temperature absorber is placed over the feed horn
aperture and the antenna temperature measured is
load
^(vW + M^.Tw,)
i v i amb'
(1)
where G is the gain in the receiver, T ^ is the ambient temperature
of the absorber, the subscripts s and i indicate the signal and image
sidebands, respectively, and
hv/k
exp(hv/kT) 1
J(v,T) =
(2)

-101-
is the effective radiation temperature of a blackbody of temperature T
at the frequency v.
Note that the receiver amplifies not only the signal sideband but
also the image sideband. The image sideband could be filtered out
before the signal is amplified but this invariably involves the intro
duction of more noise at a crucial and sensitive point in the system.
Since this work is concerned primarily with detection and re
quires the most sensitive arrangement possible, the image rejection
filter was not used. Consequently the following equations include
terms for the image sideband as well as the signal sideband.
When the telescope is pointed at blank sky the antenna tempera
ture is
Tsky GsKJ + (1 V J(VTsbr)]
+ similar terms for the image sideband (3)
where is the telescope efficiency considering spillover, blockage,
and ohmic losses. The T ^ is the apparent brightness temperature
from the spillover, blockage, and ohmic losses. The Tg and T^ (in the
image sideband terms) are the brightness temperatures of the sky at the
signal and image frequencies, respectively. They are given by
J(vsV = J^vs,Tm^1 expKA)] + vsTbg)exp(TsA)
and a similar image sideband equation. The Tm is the mean atmospheric
temperature, T^ is the brightness temperature of the cosmic background
radiation, and and t. are the atmospheric zenith optical depths at
the signal and image frequencies, A is the air mass in the observed
direction.

-102-
The calibration signal is found by taking the difference between
the load temperature and the antenna temperature looking at blank sky
ATcal Tload Tsky GstJ^vsTamb) VKV
- (1 \)J(vs,Tsbr)] + similar image sideband terms (5)
When observing a spectral line source the antenna temperature is
We = W1 nf)J(vs,Ts) + VftJ(vs>TE)tl
- exp(-T)]exp(-T$A) + J(vs,Tbg)exp(-T)exp(-TsA)
+ exp(-TsA)3> (1 VJ(VTsbl.)]
+ image terms for the sky brightness temperature (6)
^sbr
where is the excitation temperature of the molecule, t is the optical
depth of the source, and n^. is the forward coupling efficiency given by
I! PnU n)B U)dij>
nf = (7)
//27r Pn(Q)dQ
where Pn is the normalized antenna power pattern, Bn is the normalized
apparent source brightness distribution, Q is the direction of peak
antenna gain, ^ is the direction of maximum source brightness, and
d£J and d^ are infinitesimal solid angles.
Essentially, n^ is the filling factor for the source within the
telescope power pattern. If one observed an extended, uniformly bright
source n^ would be 1. Otherwise it is less than 1 and dilutes the
antenna temperature as is apparent from equation (8).

-103-
The difference in antenna temperature, between the source and
the sky is
. T- *r t
A source = source ~ sky
= G n0nfexp(-x A)(1 e"T)[J(vc,Tr) J(vc,T. )] (8)
S A"f
s EJ
s bg'
The corrected antenna temperature of the source is defined as
TA 5 iTsource/Gs\exP(-Tsfl> =
(9)
cal
where Tc is obviously
Tc s ATcal/[GsVxp(-Tsfl)]
(10)
Consequently
T* = nf(1 e"T)[J(vs,TE) J(vs,Tbg)] (11)
Note that several parameters do not appear, namely the receiver
gain (G ), atmospheric extinction exp(-x A), and antenna losses (n0)-
These have been corrected for by an appropriate choice of T From
equation ( g ) it can be seen that we need only the ratio of the source
temperature to the calibration temperature in addition to Tc to deter
mine the source antenna temperature.
If we assume that the IF is small with respect to the L0 frequency
(a very safe assumption for this work) then J(v^,T) ^ J(v ,T); that is,
the signal and image temperatures are close enough to be considered
equal then

-104-
Tc (1 + VGs>[J JJ
+ d + Si/Gs)[exp(xsA)]CJ(vs,Tsbr) J(vs,Tm)]
+ (G1/Gs){exp[(Ts Ti)A] l}[J(vs,Tm) J(VTbg)]
+ d + Gi/Gs)Cexp(.sA)/nJ,][J(vs,Tamb) J(vs,Tbg)] (12)
In general evaluation of this equation is difficult and an easier method
is to use the definition of Tc [equation (10)]. Thus, only ATcal, Gs,
n5 and t$ need to be measured as a function of elevation to determine
Tc at a given frequency. Since the atmosphere changes little at high
altitudes in the mm region, Tc is a constant function that varies
slowly if at all with time.
By comparing equation (10) with equation (12), (the mean
atmospheric temperature) can be calculated. This is perhaps the most
difficult value of equation (12) to determine. The others can be
measured or estimated with reasonable confidence. Calibration for this
work was done using equation (12) with the following values adopted:
Gs Gi 1
T = 280 K
m
Tbg=2-7K
Tsbr 280 K
Tan,b 290 K
A = sec (zenith angle)
\ = 87
xi = 0.085
t = 0.35
s

APPENDIX II
THE PHYSICS OF CO SPECTRAL LINE CALCULATIONS
In this appendix the formation of CO line emission is treated.
This is followed by consideration of column density calculations and
the assumptions involved therein.
The one-dimensional time-independent equation of transfer is
dl
~j~ = -K I + £
ds v v v
(1)
where k is the volume absorption coefficient, e is the volume emission
v v
coefficient, and I is the specific intensity at frequency v.
This can be integrated to give
-x (Sr>)
I (sj = I (0)e v 0 + e
V O V
,_n/ -T (s ) O T (s)
v o v o r v
fee ds
(2)
s
where xy(s) = / K^ds is the optical depth integrated along the line of
sight from an initial point (s = 0) to the observer's position (s ).
This can be further simplified by assuming the source is a uniform
homogeneous cloud and that k is due to an atomic or molecular transi
tion. We now have
-x (s ) e -T (s )
Us0) = Io(0)e v 0 + (^)(1 e 0 0
(3)
The first term is the attenuation of the background radiation as
it passes through the cloud and the second term is the emission
-105-

-106-
by atoms or molecules in the cloud, corrected for self-absorp
tion.
The quantity measured with a radio telescope is most closely
related to the intensity at the transition frequency, [I (s )] minus
the intensity near the transition frequency [Iv(s )]. Since
I (S ) V; I (0)
V 0 0
en Tn(sn)
Al = I (s ) I (0) = [ I (0)](1 e 0 0 )
n o o o k o '
o
(4)
By convention radio astronomical measurements are usually made in
temperature units and one can substitute the radiation temperature
2 2
[J(Tg) = c Iy/2kv ] in the preceding equation. The brightness tempera
ture is found from
I
v
= Bv(Tb) = 2hv3/c2(e
hv/kTg
- 1)1
(5)
Obviously B is the Planck function and Tg is the temperature of the
blackbody that subtends the same solid angle and emits the same specific
intensity at v. If the Rayleigh-Jeans limit (hv/kTg 1) applies then
J(Tg) Tg. This is not always true for CO measurements.
The equation for the excess line temperature is then (Zuckerman
and Palmer 1974)
AJ(Tle) = J(Tl) J(TC) = [J(Tgx) J(TC)](1 e Tv) (6)
where T^ is the line brightness temperature, T^ is the brightness tem
perature of the continuum, and T is the excitation temperature of
CA
the molecule. Assuming local thermodynamic equilibrium (LTE) gives

-107-
a Boltzmann distribution of the rotational- levels which can be
written
nu 9u -hv/kTex
"7 = 5?
(7)
o
where nu and n-j are the populations (cm ) in the upper and lower states,
respectively, and g and g-, are the statistical weights of the levels.
s
Note that tv(s) = / lyds and (Lang 1974)
k = exp (fe )Cl exp(-¡-y^-)]A (8)
v 8ttv Av K ex K ex
where Av is the full-width at half-maximum for the line, T is again
GX
the excitation temperature and the Einstein coefficient (A) for the
spontaneous electric dipole transition is
4 3
o4tt v
3hc,^
(9)
where
= y
W
+ l
+ 3
(10)
-18
for the J+l -> J transition. Since y = 0.112 x 10~ in cgs units,
|yj|^ = 5.02 x 10"^ for the J = l->0 transition, A = 8.9 x 10"^ sec"''.
s
Setting N-, = J n,ds we can write
1 0 1
\> ~2 ^ P P (if-)] (11)
8-rrv ex ex
Substituting (7) into (11), assuming hv kT and rearranging
CA
gives

-108-
u
J Jv(Tex
hcA v ex
he
(12)
Since = n^Tg, the column density in the upper state (N ') can now be
found. To calculate the column density in all states one assumes LTE,
then the number of molecules in the J state is given by (Herzberg 1950)
N, ^ Ntotexp(^iH)
kT
(13)
ID 19
where B is the rotational constant (5.764 x 10 Hz for CO) and U is
the partition function
U I (2J + 1) exp -hBJ^ + 1}
0=0 K1
(14)
Table AII-1 gives values of U and fr several temperatures.
Table AII-1
Values of and N, ,/N, for 12C0
tot J=1
Tex 10 K
25
50
100
103
u
3.97
9.37
18.40
54.89
361.7
Ntot/NJ=X
2.30
3.90
6.85
19.34
121.1
Equation (12) and Table AII-1 can be used to find the total column
density of CO under the various assumptions noted in the derivation.
This estimate is a lower limit due to the assumption that 1;
clearly, if the clouds are optically thick no information on column
density is available.

-109-
Another line of reasoning can give space densities for a molecule
if a few conditions are satisfied (see Rank et al. 1971 for a dis
cussion). If a molecule is observed in emission, well-removed from
discrete sources of radiation, it is likely that collisions raise T
t. A
well above TR (= 2.7 K, the cosmic background radiation).
The relaxation rate between two molecular states for isotropic
radiation is (Rank et al. 1971)
4 3
64tt v
3hcJ(l e
-hv/kT
R>
(15)
The collision rate can be written
= l (nov) (16)
t L 'm v '
c m
where the summation is over all colliding species, n is the density of
each species, a is the effective cross-section, and v is the average
relative velocity for the collisions. In most clouds the only par
ticles abundant enough to contribute to equation (16) are atomic and
molecular hydrogen.
Under these conditions the effective temperature of the tradi
tion is
r = yTm + tcTR
eff t + t
r c
(17)
where Tm is the temperature of the colliding particles. Since the
molecule is seen in emission it follows that xc is not much larger
than t The implication is that in order to see the molecule in

-110-
emission it must be elevated above TR by some process (collisions in
this case).
Thus, a minimum colliding particle density can be found by
equating xr and t This gives
n
64-rr4 1 U [ 2
o -hv/kTp
3hXJ(l e K)
(l
A
-hv/kTp
- e R)
(18)
The usual approach to determine is to use the geometric cross-
section (% 10 ^ cm^) and v = 8kTR/irm2 v 10^ cm/sec for T^ = 100 K and
m = m^). For CO this gives a minimum colliding particle density of
8.6 x 10^ crrf^.

APPENDIX III
THE GEOMETRY IN AN INCLINED DISK
Projection effects serve to transform the desired quantities of
circular velocity (v ), azimuthal angle (e), and radius (r) into their
foreshortened counterparts, vc, 0', and r. In this appendix the
relationships between these quantities are derived, also the calculation
of the inclination is considered. The inclination angle is defined to
be the angle between the plane of the disk and the plane of the sky.
Figure 24 illustrates the various parameters that will be used in
the following material. The effect of inclination is to foreshorten,
or project, the minor axis; the major axis is unaffected by this. The
velocity component v (always perpendicular to the line of sight) is
X
also unaffected by inclination, but it is not measurable. A component
of Vy is observed as the radial velocity (vr) and
vr = vy sin i (1)
The inclination angle is available, provided we know the true
axial ratio (q) as well as the projected axial ratio (q1). Since
q1 = b/a, where b is the semi-minor axis and a is the semi-major axis
as measured in the sky, and e = (a b)/a then
q' = 1 e (2)
The quantity q is calculated similarly, using the true axial ratio for
-111-

-112-
y
Figure 24. A disk system viewed face-on (i = 0). The x axis coincides
with the major axis and the y axis is also the minor axis.
The circular velocity, v and its x and y components are
indicated. Also shown is the azimuthal angle, e.

-113-
the system. The inclination is then given by
2.
COS 1
(3)
The true azimuthal angle, e, is found from the projected angle, e1,
by
tan 9 = tan e'/cos i (4)
and v is related to v by
r c
vc = vr/cos 0 sin i (5)
Finally, the true radius of the point in question is given by
r = [x2 + (y/cos i)2]2 (6)
where x and y are the distances along the major and minor axes, respec
tively, and y' = y/cos i.
Since the primed quantities can all be measured, the true values
can be calculated using the relations given. The two assumptions made
in the derivation are that the true axial ratio is known and that the
figure of the system is an oblate ellipsoid supported by rotation about
the minor axis.

REFERENCES
Baan, W.A., Haschick, A.D., and Burke, B.F. 1978, Ap. J., 225, 339.
Bailey, M.E. 1980, M.N.R.A.S., 191, 195.
Balick, B., Faber, S.M., and Gallagher, J.S. 1976, Ap. J., 209, 710.
Bautz, L.P. and Morgan, W.W. 1970, Ap. J. (Letters), 162, L149.
Bertola, F. and Capaccioli, M. 1975, Ap. J., 200, 439.
Binney, J. 1978, M.N.R.A.S., 183, 501.
Blackman, R.B. and Tukey, J.W. 1958, "The Measurement of Power Spectra"
(Dover: New York).
Bottinelli, L. and Gouguenheim, L. 1977a, Astron. & Astrophys., 54,
641.
Bottinelli, L. and Gouguenheim, L. 1977b, Astron. & Astrophys., 60,
223.
Bottinelli, L. and Gouguenheim, L. 1979a, Astron. & Astrophys., 7J5,
176.
Bottinelli, L. and Gouguenheim, L. 1979b, Astron. & Astrophys., 74,
172.
Burstein, D. 1979a, Ap. J. Supplement, 41, 435.
Burstein, D. 1979b, Ap. J., 234, 435.
Burstein, D. 1979c, Ap. J., 234, 829.
Butcher, H. and Oemler, A. 1978, Ap. J., 219, 18.
Cheng, E.S., Saulson, P.R., Wilkinson, D.T., and Corey, B.E. 1979,
Ap. J. (Letters), 232, L139.
Coleman, G.D. and Worden, S.P. 1977, Ap. J., 218, 792.
Cooper, B.F.C. 1976, in "Methods of Experimental Physics," 12_ part B,
ed. M.L. Meeks (Academic Press: New York), 280.
de Vaucouleurs, G. 1959, Handbuch der Physik, 53, 275.
-114-

-115-
de Vaucouleurs, G. 1977, in "The Evolution'of Galaxies and Stellar
Populations" (Yale University Observatory: New Haven), 74.
de Vaucouleurs, G. and de Vaucouleurs, A. 1964, "Reference Catalog of
Bright Galaxies" (University of Texas Press: Austin).
de Vaucouleurs, G., de Vaucouleurs, A., and Corwin, H. 1976, "Second
Reference Catalog of Bright Galaxies" (University of Texas Press:
Austin).
de Vaucouleurs, G. and Neito, J.-L. 1979, Ap. J., 230, 697.
Dressier, A. 1980, Ap. J., 236, 351.
Faber, S.M. and Gallagher, J.S. 1976, Ap. J., 204, 365.
Faber, S.M. and Gallagher, J.S. 1979, Ann. Rev. Astron. & Astrophys.,
17, 135.
Faber, S.M. and Jackson, R. 1976, Ap. J., 204, 668.
Field, G.B. 1975, in "Galaxies and the Universe," ed. A. Sandage, M.
Sandage, and J. Kristian (Univ. of Chicago Press: Chicago).
Ford, H.C., Jacoby, G., and Jenner, D.C. 1977, Ap. J., 213, 18.
Ford, H.C. and Jenner, D.C. 1975, Ap. J., 202, 365.
Ford, H.C., Jenner, D.C., and Epps, H.W. 1973, Ap. J. (Letters),
183, L73.
Fosbury, R.A.E., Mebold, U., Goss, W.M., and Dopita, M.A. 1978,
M.N.R.A.S., 183, 549.
Gallagher, J.S., Faber, S.M., and Balick, B. 1975, Ap. J., 202, 7.
Gallagher, J.S., Knapp, G.R., Faber, S.M., and Balick, B. 1977, Ap.
J., 215, 463.
Gallouet, L. and Heidmann, N. 1971, Astron. & Astrophys. Suppl., _3,
325.
Gallouet, L., Heidmann, N., and Dampierre, F. 1973, Astron. & Astrophys.
Suppl., 12_, 89.
Gallouet, L., Heidmann, N., and Dampierre, F. 1975, Astron. & Astrophy.
Suppl., 19_, 1.
Giovanelli, R. 1979, Private communication.
Gisler, G.R. 1976, Astron. & Astrophys., 51^, 137.
Gisler, G.R. 1979, Ap. J., 228, 385.

-116-
Gisler, G.R. 1980, A. J., 85, 623.
Gott, J.R. 1977, Ann. Rev. Astron. & Astrophys., 15_, 235.
Gottesman, S.T. and Weliachew, L. 1977, Ap. J., 211, 47.
Graham, J.A. 1979, Ap. J., 232, 60.
Gunn, J.E. and Gott, J.R. 1972, Ap. J., 176, 1.
Gunn, J.E. and Oke, J.B. 1975, Ap. J., 195, 255.
Guthrie, B.N.G. 1974, M.N.R.A.S., 168, 15.
Hausman, M.A. and Ostriker, J.P. 1978, Ap. J., 224, 320.
Haynes, M.P., Brown, R.L., and Roberts, M.S. 1978, Ap. J., 221, 414.
Haynes, M.P. and Roberts, M.S. 1979, Ap. J., 337, 767.
Herbig, G.H. 1970, "Proc. of the XVIth Lige Symposium," 59., 13.
Herzberg, G. 1950, "Spectra of Diatomic Molecules" (Van Nostrand:
New York), 125.
Hodge, P.W. 1963, A. J., 68, 691.
Hodge, P.W. 1973, Ap. J., 182, 671.
Hubble, E.P. 1926, Ap. J., 64, 321.
Hubble, E.P. 1936, "The Realm of the Nebulae" (Yale University Press
New Haven), 36.
Huchtmeier, W.K., Tammann, G.A., and Wendker, H.J. 1977. Astron. &
Astrophys., 57, 313.
Hulsbosch, A. 1980, Private communication.
Humason, M.L., Mayall, N.U., and Sandage, A.R. 1956, A. J., 6_1, 97.
Illingworth, G. 1977, Ap. J. (Letters), 218, L43.
Johnson, H.E. and Axford, W.I. 1971, Ap. J., 165, 381.
Jones, B.T. 1976, Rev. Mod. Phys., 48, 107.
Jura, M. 1977, Ap. J., 212, 634.
Kent, S.M. and Sargent, W.L.W. 1979, Ap. J., 230, 667.
Killian, D.J. 1978, Ph.D. Dissertation, University of Florida.

Knapp, G.R., Faber, S.M., and Gallagher, J.S. 1978a, A. J., 83, 11.
Knapp, G.R., Gallagher, J.S.., and Faber, S.M. 1978b, A. J., 83, 139.
Knapp, G.R. and Jura, M. 1976, Ap. J., 209, 782.
Knapp, G.R., Kerr, F.J., and Williams, B.A. 1978c, Ap. J., 222, 800.
Lang, K.R. 1974, "Astrophysical Formulae" (Spr.inger-Verlag: New
York), 161.
Larson, R.B. 1974, M.N.R.A.S., 169, 229.
Larson, R.B. and Tinsley, B.M. 1974, Ap. J., 192, 293.
Larson, R.B. and Tinsley, B.M. 1978, Ap. J., 219, 46.
Lea, S.M. and De Young, D.S. 1976, Ap. J., 210, 647.
Lo, K.Y. and Sargent, W.L.W. 1979, Ap. J., 22J_, 756.
Lucas, R. 1974, Astron. & Astrophys., 36., 465.
Lynden-Bell, D. 1967, M.N.R.A.S., 136, 101.
Martin, R.N. and Barrett, A.H. 1978, Ap. J. Suppl., 3(5, 1.
Mathews, W.G. and Baker, J.C. 1971, Ap. J., 170, 241.
Mathewson, D.S., Cleary, M.N., and Murray, J.D. 1974, Ap. J., 190, 291.
Mathewson, D.S., Cleary, M.N., and Murray, J.D. 1975, Ap. J. (Letters),
195, L97.
McHardy, I.M. 1974, M.N.R.A.S., 169, 522.
Mezger, P.G. and Smith, L.F. 1977, in "Star Formation, I.A.U. Symp.
No. 75," ed. T. de Jong and A. Maeder (D. Reidel: Dordrecht),
133.
Morgan, W.W. 1958, Pub. A.S.P., 70, 364.
Morgan, W.W, 1959, Pub. A.S.P., 71, 394.
Morton, D.C. 1975, Ap. J., 197_, 85.
Oemler, A. 1974, Ap. J., 194, 1.
Oemler, A. 1977, Highlights of Astronomy, 4, 253.
Osterbrock, D. 1960, Ap. J., 132, 325.

-118-
Osterbrock, D. 1962, in "Interstellar Matter in Galaxies" (Benjamin:
New York), 111.
Ostriker, J.P. 1977, in "The Evolution of Galaxies and Stellar Popula
tions," ed. B.M. Tinsley and R.B. Larson (Yale University Observa
tory: New Haven), 369.
Ostriker, J.P. and Hausman, M.A. 1977, Ap. J. (Letters), 217, L125.
Ostriker, J.P. and Tremaine, S.D. 1975, Ap. J. (Letters), 202, LI13.
Penzias, A.A. and Burrus, C.A. 1973, Ann. Rev. Astron. & Astrophys.,
11, 51.
Peterson, C.J. 1978, Ap. J., 222, 81.
Rank, D.M., Townes, C.H., and Welch, W.J. 1971, Science, 174, 1083.
Reif, K., Mebold, U., and Goss, W.M. 1978, Astron. & Astrophys., 6)7, LI.
Rickard, L.J., Palmer, P., Morris, M., Turner, B.E., and Zuckerman, B.
1977, Ap. J., 213, 673.
Riley, J.M. 1975, M.N.R.A.S., 170, 53.
Roberts, M.S. and Steigerwald, D.G. 1977, Ap. J., 217, 883.
Rood, H. 1979, Ap. J., 232, 699.
Rowan-Robinson, M. 1979, Ap. J., 234, 111.
Rubin, V.C., Ford, W.K., Peterson, C.J., and Oort, J.H. 1977, Ap. J.,
211, 693.
Sandage, A.R. 1961, "The Hubble Atlas of Galaxies" (Carnegie Inst, of
Washington: Washington).
Sargent, W.L.W., Young, P.J., Boksenberg, A., Shortridge, K., Lynds,
C.R., and Hartwick, F.D.A. 1978, Ap. J., 221, 731.
Schechter, P.L. and Gunn, J.E. 1979, Ap. J., 229, 472.
Schneps, M.H., Ho, P.T.P., Barrett, A.H., Buxton, R.B., and Myers,
P.C. 1978, Ap. J., 225, 808.
Shane, W.W. 1980, Astron. & Astrophys., 82, 314.
Shostak, G.S. 1978, Astron. & Astrophys., 54^, 919.
Silk, J. and Norman, C. 1979, Ap. J., 234, 86.
Smoot, GF. and Lubin, P.M. 1979, Ap. J. ILetters), 234, L83.

-119-
Sunyaev, R.A. 1977, in "Radio Astronomy and Cosmology, I.A.U. Symp.
No. 74," ed. D.L. Jauncey (D. Reidel: Dordrecht).
Tammann, G.A. 1974, in "Supernovae and Supernovae Remnants," ed.
C.B. Cosmovici (D. Reidel: Dordrecht).
Tartar, J. 1975, Ph.D. Dissertation, University of California, Berkeley.
Tytler, D. and Vidal, N. 1978, M.N.R.A.S., 182, 33p.
Ulich, B.L. and Haas, R.W. 1976, Ap. J. Suppl., 30, 247.
Unwin, S.C. 1980, M.N.R.A.S., 190, 551.
van den Bergh, S. 1960a, Ap. J., 131, 215.
van den Bergh, S. 1960b, Ap. J., 131, 558.
van den Bergh, S. 1975, Ann. Rev. Astron. & Astrophys., 13, 217.
van den Bergh, S. 1976a, Ap. J., 208, 673.
van den Bergh, S. 1976b, Ap. J., 206, 883.
van den Bergh, S. 1976c, A. J., 81, 797.
White, S. 1976, M.N.R.A.S., 174, 19.
Wright, M.C.H. 1974, in "Galactic and Extra-Galactic Radio Astronomy,"
ed. G.IL. Verschuur and K.I. Kellerman (Springer-Verlag: New York),
291.
Young, P.J., Boksenberg, A., Lynds, C.R., and Hartwick, F.D.A. 1978a,
Ap. J., 222, 450.
Young, P.J., Sargent, W.L.W., Kristian, J., and Westphal, J.A. 1979,
Ap. J., 234, 76.
Young, P.J., Westphal, J.A., Kristian, J., Wilson, C.P., and Landauer,
F.P. 1978b, Ap. J., 221, 721.
Zuckerman, B. and Palmer, P. 1974, Ann. Rev. Astron. & Astrophys.,
12, 279.

BIOGRAPHICAL SKETCH
On reaching the age of 13, Douglas William Johnson had managed to
take apart every watch, toaster, alarm clock, and gadget in his parents'
wood frame home in both Gas City, Indiana, and Connell esvil le, Pennsyl
vania. His sisters, Rhonda and Peggy, grew up with the enduring fear
their hair dryer would fall prey to his eager little hands.
Doug's mother comforted herself with the conviction that her
second eldest son would go on to become a mechanic. Anyone who so
loved to work with his hands, she reasoned, must be born to things
mechanical.
Mary Ann Johnson's instincts about her son were good, but she had
not reckoned on the draw of the night sky. Those brilliant, changing
lights--completely out of reach--caught Doug's interest effortlessly.
When his father's job as a Postal Inspector brought the family east to
New Jersey, in 1968, the would-be astronomer constructed his own tele
scope to ply his trade in the Johnson's lush and overgrown backyard.
Doug attended Boonton High School and pursued his favorite subjects
of physics and math.
With a well-nurtured respect for the universe and the deep curi
osity of a scientist, Doug went on to college at Rensselaer Polytechnic
Institute in Troy, New York. In 1975, equipped with a bachelor's degree
in physics, and a wife gracious enough to work full-time, he arrived in
Gainesville. It was late summer, a time designed to kill Yankees with
-120-

-121-
its stifling, muggy heat. But the Johnsons survived and prospered,
spreading their prosperity among the animal kingdom through the addition
of four highly personable cats and a sweet-faced mongrel from the county
pound to their family.
When lured away from his computers, mounds of graph paper, or
blackboards, Doug holds true to his earliest love of tinkering. He
is an accomplished carpenter, a competent mechanic, and makes a mean
pitcher of frozen daquiris.
Though not given over to an abandoned social life, he misses few
science fiction movies. His friends are a peculiar mixture of astronomy
graduate students and newspaper types--a situation forced on him by his
wife, Maryfran, the tolerant breadwinner of the family for the last
five years.

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.
/jjcfi}vw ft] rj:c
m\ 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.
(L¡4 U HAn. ~ T- e
(/
t)r. Thomas D. Carr
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.
vx vv
Dr. Kwan-Yu Chen
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.
Associate Professor of Physics
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.
This dissertation was submitted to the Graduate Faculty of the Depart
ment 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.
August 1980
Dean, Graduate School



-115-
de Vaucouleurs, G. 1977, in "The Evolution'of Galaxies and Stellar
Populations" (Yale University Observatory: New Haven), 74.
de Vaucouleurs, G. and de Vaucouleurs, A. 1964, "Reference Catalog of
Bright Galaxies" (University of Texas Press: Austin).
de Vaucouleurs, G., de Vaucouleurs, A., and Corwin, H. 1976, "Second
Reference Catalog of Bright Galaxies" (University of Texas Press:
Austin).
de Vaucouleurs, G. and Neito, J.-L. 1979, Ap. J., 230, 697.
Dressier, A. 1980, Ap. J., 236, 351.
Faber, S.M. and Gallagher, J.S. 1976, Ap. J., 204, 365.
Faber, S.M. and Gallagher, J.S. 1979, Ann. Rev. Astron. & Astrophys.,
17, 135.
Faber, S.M. and Jackson, R. 1976, Ap. J., 204, 668.
Field, G.B. 1975, in "Galaxies and the Universe," ed. A. Sandage, M.
Sandage, and J. Kristian (Univ. of Chicago Press: Chicago).
Ford, H.C., Jacoby, G., and Jenner, D.C. 1977, Ap. J., 213, 18.
Ford, H.C. and Jenner, D.C. 1975, Ap. J., 202, 365.
Ford, H.C., Jenner, D.C., and Epps, H.W. 1973, Ap. J. (Letters),
183, L73.
Fosbury, R.A.E., Mebold, U., Goss, W.M., and Dopita, M.A. 1978,
M.N.R.A.S., 183, 549.
Gallagher, J.S., Faber, S.M., and Balick, B. 1975, Ap. J., 202, 7.
Gallagher, J.S., Knapp, G.R., Faber, S.M., and Balick, B. 1977, Ap.
J., 215, 463.
Gallouet, L. and Heidmann, N. 1971, Astron. & Astrophys. Suppl., _3,
325.
Gallouet, L., Heidmann, N., and Dampierre, F. 1973, Astron. & Astrophys.
Suppl., 12_, 89.
Gallouet, L., Heidmann, N., and Dampierre, F. 1975, Astron. & Astrophy.
Suppl., 19_, 1.
Giovanelli, R. 1979, Private communication.
Gisler, G.R. 1976, Astron. & Astrophys., 51^, 137.
Gisler, G.R. 1979, Ap. J., 228, 385.


-82-
The true radius can also be computed and is listed with the velocities
in Table 11. Finally, the mass interior to the point, assuming Kepler-
ian motion, is given by
M = 2.36 x 102 v2 r (7)
where M is in solar masses, v is the circular velocity in km/sec, and
r is the true radius in parsecs.
Table 11
Circular Velocities and Radial Distances in NGC 185
Object
vr
vc (km/sec)
r (pc)
km/sec
i = 53.1
i = 68.0
i = 53.1
i.
II
O'!
00
o
o
185-1
37.5
61.3
67.8
191.4
248.0
185-2
25
59.2
74.6
203.2
296.6
The isophotal contours of Hodge (1963) were measured to find the
luminosity interior to a given point. The contours measured were the
innermost (1) and the next brightest (2), corresponding to 20.47 and
2
21.22 mag/arcsec respectively. Using a distance modulus of 23.9 and
assuming absorption within our own galaxy according to the cosecant law
(Av = 0.26 esc b = 1.0 mag for NGC 185) allows the calculation of the
luminosity interior to the first and second contours, listed in
Table 12.


-70-
t
I I
30"
%
Figure 22. NGC 185 showing the dust patches. North is up and west is
to the right. The half-power beam-width is indicated.
This photograph is reproduced through the courtesy of Lick
Observatory.


CHAPTER V
SUMMARY
The current understanding of elliptical galaxy formation and
evolution has several important shortcomings. It is not understood
how elliptical, or for that matter any, galaxies separate out from the
background and collapse. Even the general process of whether the
clusters of galaxies form before the individual galaxies is not clear.
The deceptively relaxed-looking stellar distributions are apparently
not relaxed at all. It is quite possible that the 3-dimensional figure
of elliptical galaxies is actually a triaxial ellipsoid. Further in
consistencies are apparent in the disposition of gas shed by the stars
within the galaxy. If it recycled into stars with a normal IMF the
nuclei of all early-type systems should be bluer than observed.
Recent theoretical work has shown that two gas removal mechanisms
may be present in these systems. One is a galactic wind in which
supernovae provide an energy source to heat the ISM of the galaxy to
a temperature high enough to evaporate from the system. The crucial
points of this method are the assumed Type I supernova, rate and the
efficiency of energy coupling between the expanding supernova shells
and the general ISM.
The other mechanism is ram-jet stripping. The process is basically
a hydrodynamic interaction between an assumed intracluster medium and
the ISM of the galaxy. There is circumstantial evidence for this
-96-


A STUDY OF THE INTERSTELLAR MEDIUM IN NGC 185
AND OTHER EARLY-TYPE GALAXIES
By
DOUGLAS WILLIAM JOHNSON
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
1980


-79-
possibility. Specifically, since the regions observed are dense
enough to form stars yet no HII regions are observed, either no stars
have formed yet (leading to the burst hypothesis) or only low-mass
stars are forming.
The idea of a skewed mass distribution for stars is explored in
some details by Mezger and Smith (1977). They find that clusters of
massive stars (which produce giant HII regions) are found only along
spiral arms. It is probable that the spiral shock front of a density
wave triggers their formation. It appears though, that the low-mass
stars can form out of small but dense clouds (Herbig 1970) that are
far more widespread than large clouds.
Jura (1977) also proposes that the expected radiation environment
of elliptical galaxies (low UV flux) will lead to low-mass star forma
tion. The reasoning is that the low flux results in less heating of the
clouds. A cooler cloud will have a lower critical mass which can
separate out and collapse to form a star.
The dynamics of the nuclear region will be considered in the fol
lowing paragraphs. The difficulties in measuring radial velocities for
early-type galaxies are immense. There are usually no HII regions in
these galaxies so their sharp emission lines are not available for
measurement. The absorption lines from the stars are intrinsically wide
and further broadened by the stellar velocity dispersion, resulting in
large errors.
The radial velocity data for NGC 185 are actually better than that
for most early-type galaxies. Ford et al. (1973) identified four
planetary nebulae in the system and later Ford et al. (1977) succeeded
in measuring the radial velocities of two of these objects, NGC 185-1 and


TEMPERATURE ¡K)
tX 101 )
JO IJIJ 4;U U U i O U U £_j U U In. II M
VEL0C ITT (KM/SEC) (X I O1 )
Figure 6. The 12C0 data from December 1979, channel A.


TEMPERATURE (K)
(XIO1 )
-f=>
I
Figure 4
The 12C0 data from December 1977.


-104-
Tc (1 + VGs>[J JJ
+ d + Si/Gs)[exp(xsA)]CJ(vs,Tsbr) J(vs,Tm)]
+ (G1/Gs){exp[(Ts Ti)A] l}[J(vs,Tm) J(VTbg)]
+ d + Gi/Gs)Cexp(.sA)/nJ,][J(vs,Tamb) J(vs,Tbg)] (12)
In general evaluation of this equation is difficult and an easier method
is to use the definition of Tc [equation (10)]. Thus, only ATcal, Gs,
n5 and t$ need to be measured as a function of elevation to determine
Tc at a given frequency. Since the atmosphere changes little at high
altitudes in the mm region, Tc is a constant function that varies
slowly if at all with time.
By comparing equation (10) with equation (12), (the mean
atmospheric temperature) can be calculated. This is perhaps the most
difficult value of equation (12) to determine. The others can be
measured or estimated with reasonable confidence. Calibration for this
work was done using equation (12) with the following values adopted:
Gs Gi 1
T = 280 K
m
Tbg=2-7K
Tsbr 280 K
Tan,b 290 K
A = sec (zenith angle)
\ = 87
xi = 0.085
t = 0.35
s


TEMPERATURE CK)
1.00
11.00
11 .00
-9.no
f X101 )
21.00 31.00
VELOCITY (KM
Figure 14. The 12C0 spectrum for NGC 4150. De Vaucouleurs and de Voucouleurs (1964) give
a radial velocity of 244 km/sec, but de Vaucouleurs et al. (1976) list it as
970 km/sec.
I
on
CO
i


ACKNOWLEDGEMENTS
There are many people who have contributed invaluable resources
and encouragement to me in the completion of the research for this dis
sertation. Although it is not possible to thank each one individually,
there are some I would like to acknowledge in particular.
The Northeast Regional Data Center is acknowledged for providing
a vigorous and stimulating computing environment. The operators of the
11 m NRAO telescope were very helpful in assisting me during my observing
sessions and the assistance of Telescope Engineer Rick Howard was
especially appreciated. I would also like to thank Dan McGuire for his
assistance during my first observing session.
I thank my supervisory committee members, Dr. Stephen T. Gottesman,
Dr. Thomas D. Carr, Dr. Kwan-Yu Chen, Dr. Gary Ihas, and Dr. William
Weltner, for their attention and interest in my work.
I thank the Physics Department, and especially Dr. Richard Garrett,
for the assistantships that have allowed me to pursue this course of
study.
The Division of Sponsored Research is thanked for its Seed Money
Grant Competition, providing financial support for Dr. Gottesman and me
during the course of our investigations. The Graduate School's Supple
mentary Fellowship for 1979-80 was also greatly appreciated.
Finally, I wish to thank Steve Gottesman and his family for their
efforts to make Maryfran's and my stay in Gainesville a wonderful time
iii


CHAPTER II
THE OBSERVATIONS
This chapter presents the data acquisition and handling procedures
used for the work described in this dissertation. The instruments used are
described giving particular attention to the equipment and techniques
which aided this work immensely. The method of data presentation is
also explained.
Carbon Monoxide Observations at Kitt Peak, AZ
Telescope Description
-jo -i r
The observations searching for the 2.6 mm transition of iCiD0
were made at the National Radio Astronomy Observatory (NRAO) Millimeter
Wave Telescope1 at Kitt Peak, AZ. The telescope is an 11 m paraboloid
which can be driven in altitude and azimuth. Tracking of celestial
objects, data acquisition,monitoring of system status, as well as
initial data reduction is handled by an on-line PDP 11/40 computer.
The observations were taken during three separate observing
sessions
December 23-26, 1977
July 7-9, 1978
December 10-16, 1979
^The National Radio Astronomy Observatory is operated by Associated
Universities, Inc., under contract with the National Science Foundation.
-31-


BIOGRAPHICAL SKETCH
On reaching the age of 13, Douglas William Johnson had managed to
take apart every watch, toaster, alarm clock, and gadget in his parents'
wood frame home in both Gas City, Indiana, and Connell esvil le, Pennsyl
vania. His sisters, Rhonda and Peggy, grew up with the enduring fear
their hair dryer would fall prey to his eager little hands.
Doug's mother comforted herself with the conviction that her
second eldest son would go on to become a mechanic. Anyone who so
loved to work with his hands, she reasoned, must be born to things
mechanical.
Mary Ann Johnson's instincts about her son were good, but she had
not reckoned on the draw of the night sky. Those brilliant, changing
lights--completely out of reach--caught Doug's interest effortlessly.
When his father's job as a Postal Inspector brought the family east to
New Jersey, in 1968, the would-be astronomer constructed his own tele
scope to ply his trade in the Johnson's lush and overgrown backyard.
Doug attended Boonton High School and pursued his favorite subjects
of physics and math.
With a well-nurtured respect for the universe and the deep curi
osity of a scientist, Doug went on to college at Rensselaer Polytechnic
Institute in Troy, New York. In 1975, equipped with a bachelor's degree
in physics, and a wife gracious enough to work full-time, he arrived in
Gainesville. It was late summer, a time designed to kill Yankees with
-120-


-19-
The thrust of much work in this area has been to determine if
SOs can be formed by stripping spiral galaxies of their gas and dust
(Gisler 1979). This would quench star formation and significantly change
the optical appearance of the galaxy.
Another development that has spurred interest in the interaction
of an ISM with an intergalactic medium (IGM) is the discovery of
head-tail radio galaxies. The most straightforward explanation of
this phenomenon being just such an interaction.
In spite of the varied motivations for these studies many of the
numerical simulations are directly applicable to the analysis of an
elliptical system passing through an IGM.
The primary results of these studies are to confirm that under
appropriate conditions there is an effective sweeping out of material
from the galaxy. The procedures and model parameters used to arrive
at this conclusion vary substantially for the different experiments,
all agree however, that some material tightly bound near the nucleus
may be retained. Gisler (1979) explores the situation further and
finds that the rate of gas replenishment is important, possibly stopping
the stripping effect entirely if it is high enough.
In spite of this it appears that stripping can be at least par
tially effective over a broad range of galaxy velocities and IGM
densities.. This finding agrees well with the observation that a large
fraction of galaxies in rich clusters are SOs and ellipticals (Oemler
1974, 1977).
It would seem also that the evidence of a positive correlation
between X-ray luminosity of clusters (presumably from a hot intracluster
component) and SO/spiral ratios (Tytler and Vidal 1978) argues strongly


-77-
The Jeans mass can be calculated under these conditions from
M TT
MJ 6 po
n in23
Aj ^ 10
(TK/y)
3/2
1/2
gm
(4)
with T = 20 K, y = 2, p = 3.34 x 10~^ g/cm3 (corresponding to 103 cm
of H2) we find Mj = 27 M If the cloud remains isothermal as it
collapses, and this appears likely because of the effective cooling,
Mj decreases and fragmentation is likely to occur.
Another calculation that confirms the collapsing nature of the
cloud in NGC 185 is provided by Rowan-Robinson (1979). He estimates
the cloud density, in an essentially virial calculation, by the
equation
n(H2) = 103*90 n (Av)2 r"2 K^1 (5)
where n is 0.8 for a cloud in equilibrium, r is the radius of the cloud
18
in units of 10 cm, and is a dimensionless quantity related to a
hot core for the cloud (it is always greater than 1.5).
The density calculated for the NGC 185 cloud, assuming 6.6 pc
for r (the characteristic width for DP-Ia) is 3.3 x 103 cm-3. If the
2 -3
cloud is in free-fall n = 0.2 and the derived density is 8.3 x 10 cm .
Thus considering that the NGC 185 cloud masses are uncomfortably large
(Table 7 ) for higher particle densities, we consider the lower density
more likely.
From these arguments it appears that either DP-Ia or DP-I is most
likely in a state of collapse. Further, the lack of emission from
ionized gas (Humason et al. 1956) implies one of two possibilities.
The first is that star-formation ceased in the system long ago, such


-80-
-2 in their notation. These measurements, coupled with the radial
velocity of DP-I, can be used to calculate a rough mass interior to the
object and, when combined with the isophotal contours of Hodge (1963),
the mass-to-luminosity ratio (MLR) can be computed.
Table 9
Positions and Radial Velocities of Points Within NGC 185
Object
Radial
Velocity
(km/sec)
Distance from
Minor Axis
(pc)
Distance from
Major Axis
(pc)
185-1
-212.5
144.2
75.6
185-2
-200
107.3
103.6
DP-1
-175
0
43.9
Table 9 lists the three radial velocities determined for the system,
as well as the projected distances from the major and minor axes. We
make the simple assumption that the ellipticity is caused by rotation
about the minor axis, thus the figure of the system is an oblate spheroid.
We further assume that the radial velocities are the projected results
of purely circular velocities about the nucleus. We can then make a
straightforward calculation of the mass interior to the measured point.
Implicit in this approach is the assumption that the points lie in the
plane of the galaxy.
Recalling the work of de Vaucouleurs (1977), described in Chapter I,
on the relative frequencies of various ellipticities, we can make an
informed guess of the true ellipticity of NGC 185. The galaxy is most
likely to be an E3.6, but we can be more flexible and also consider


-83-
Table 12
Luminosities in the Nucleus of NGC 185
Contour
Brightness2
Area ^
Luminosity
(mag/arcsec )
(arcsec )

1
20.47
1690
1.03 x 107
2
21.22
7715a
2.35 x 107a
aExcludes contribution interior to contour 1.
The mass calculated from equation 7 is then divided by the appro
priate luminosity from Table 12 to derive the MLR interior to each
object. The Planetary nebulae, 185-1, lies on the second contour,
making the luminosity interior to it the sum of the luminosities inside
contours 1 and 2. The other nebula, 185-2, lies midway between con
tours 1 and 2 so we use the luminosity inside 1 plus half the luminosity
inside 2. Table 13 lists the results for the MLR derived with these
values.
Table 13
Mass-to-Luminosity Ratios for the Nucleus of NGC 185
Object Mass Interior (Ml Luminosity MLR (solar units)
5 (L0)
i = 53.1 i = 68.0 u i = 53.1 i = 68.0
185-1 1.7 x 108 2.7 x 108 3.38 x 107 5.0 8.0
1.7 x 108 3.9 x 108 2.21 x 107
185-2
7.7
17.6


-95-
companion of M31 an interaction of this sort may be occurring. The
difficulty is that M31 is to the south of NGC 185 and is over 6 away,
a rather larger distance for an interaction considering the small mass
of NGC 185. Thus while the intergalactic plume is an intriguing pos
sibility, it is not well-supported by our observations.
Higher spatial resolution measurements are required in the area
centered on the galaxy and at least 20-30' around. It is necessary
to examine the local features well enough to allow their subtraction
from possible HI belonging to NGC 185.
Any HI associated with the galaxy would probably have a distribu
tion less than 5' across; the galaxy itself has a major diameter of
around 10'. The VLA can easily achieve this resolution, and an in
vestigation of the NGC 185 region should be undertaken to determine
the kinematics of the HI at the position of the galaxy.


-102-
The calibration signal is found by taking the difference between
the load temperature and the antenna temperature looking at blank sky
ATcal Tload Tsky GstJ^vsTamb) VKV
- (1 \)J(vs,Tsbr)] + similar image sideband terms (5)
When observing a spectral line source the antenna temperature is
We = W1 nf)J(vs,Ts) + VftJ(vs>TE)tl
- exp(-T)]exp(-T$A) + J(vs,Tbg)exp(-T)exp(-TsA)
+ exp(-TsA)3> (1 VJ(VTsbl.)]
+ image terms for the sky brightness temperature (6)
^sbr
where is the excitation temperature of the molecule, t is the optical
depth of the source, and n^. is the forward coupling efficiency given by
I! PnU n)B U)dij>
nf = (7)
//27r Pn(Q)dQ
where Pn is the normalized antenna power pattern, Bn is the normalized
apparent source brightness distribution, Q is the direction of peak
antenna gain, ^ is the direction of maximum source brightness, and
d£J and d^ are infinitesimal solid angles.
Essentially, n^ is the filling factor for the source within the
telescope power pattern. If one observed an extended, uniformly bright
source n^ would be 1. Otherwise it is less than 1 and dilutes the
antenna temperature as is apparent from equation (8).


-121-
its stifling, muggy heat. But the Johnsons survived and prospered,
spreading their prosperity among the animal kingdom through the addition
of four highly personable cats and a sweet-faced mongrel from the county
pound to their family.
When lured away from his computers, mounds of graph paper, or
blackboards, Doug holds true to his earliest love of tinkering. He
is an accomplished carpenter, a competent mechanic, and makes a mean
pitcher of frozen daquiris.
Though not given over to an abandoned social life, he misses few
science fiction movies. His friends are a peculiar mixture of astronomy
graduate students and newspaper types--a situation forced on him by his
wife, Maryfran, the tolerant breadwinner of the family for the last
five years.


TEMPERATURE CK)
-0.60 -0.20 0.20
(X101 )
Figure 16. The ^C0 spectrum for NGC 4636.
-60-


TEMPERAT U R E (K)
42. 00
52.00
C X101 )
62.00 72.00
92.00
lofc'ur
VELOCITY (KM/SEC) (X1D1 )
Figure 15. The 12C0 spectrum for NGC 4278.
I
c_n


-20-
that the cluster environment does indeed have a significant influence
on the structure and evolution of its component galaxies.
An evolutionary effect may also have been observed by Butcher
and Oemler (1978). Their study found that a rich cluster observed at
a redshift of 0.4 contains many more blue galaxies than a similar rich
cluster observed in the current epoch. Their conclusion is that as
Q
late as 4 x 10 years ago the galaxies in this cluster had not yet
been stripped of their ISM. They were consequently undergoing at
least moderate star formation, thus producing the blue colors observed.
One further piece of evidence that fits in quite well with the
general hypothesis of ram-jet stripping is the common coincidence of a
strong radio galaxy at the center of a dense cluster (McHardy 1974,
Guthrie 1974, and Riley 1975). The argument in this case is that the
central galaxies have a small velocity with respect to the intra
cluster medium and will be more likely to retain gas shed by its
stellar population. The material collapses to the center of the
galaxy, apparently forming a massive object and producing the ob
served radio source.
In view of the variety of indications that imply a substantial
interaction between an ISM and an intracluster medium it seems clear
that environmental factors can be important in the evolution and
structure of galaxies in clusters. But in the particular case of
the detected ellipticals NGC 4278 and 4636 there is reason to believe
the IGM is unimportant.
These are the only two galaxies which have an HI distribution
that is extended enough to map. In both instances the HI distribution
appears to be considerably wider than the photometric diameter of the


-5-
The origin of the anisotropy is still not clear but the most
likely source is remnant anisotropy from the collapse phase of the
galaxies' formation.
A further difficulty in our understanding of elliptical galaxies
is the existence of extreme population I ingredients in a significant
number of systems. Specifically:
OB clusters are observable in NGC 185 and 205 (Hodge 1963
and 1973)
ionized gas is seen in the nuclei of at least 15% of all ellipti
cals (Osterbrock 1960 and 1962)
neutral hydrogen has been detected in at least 8 elliptical
systems:
NGC 1052 (Knapp et al. 1978b)
NGC 2974 (Bottinelli and Gouguenheim 1979b)
NGC 3904 (Bottinelli and Gouguenheim 1977b)
NGC 3962 (Bottinelli and Gouguenheim 1979a)
NGC 4105 (Bottinelli and Gouguenheim 1979b)
NGC 4278 (Gallagher et al. 1977)
NGC 4636 (Knapp et al. 1978a)
NGC 5846 (Bottinelli and Gouguenheim 1979b)
The presence of population I material is unusual for systems thought
to have ended all star formation long ago. Some possible explanations
are that the material was accreted relatively recently and thus it is
not representative of an elliptical galaxy's normal evolution, or that
through normal processes of stellar evolution the material was shed by
the stars and is observable in various forms today.
The thrust of the foregoing observations is that, as a class,
elliptical galaxies are not as well-understood as was earlier believed.
The apparently relaxed stellar distribution is most likely not relaxed
at all, but still contains velocity anisotropies which strongly in
fluence the shape of the galaxy. It is disturbingly common for the
smooth isophotes to be blemished with dust obscuration or some other


-88-
beam, it is more difficult to detect than if it were a narrow feature
arising from only a single, larger cloud.
If t >> 1 no information can be gained about the amount of molecu
lar material stored. Essentially any quantity of material can be
stored in clumped, optically thick clouds. The material remains un
detected because of the very small filling factor (n^). Consequently
Table 16 lists various parameters under the assumption that the CO
clouds are optically thin.
12
Knapp and Jura's (1976) finding that the CO line is probably
thin if the visual absorption is less than 1 magnitude indicates that
most CO in elliptical galaxies will be thin since deep absorption
features are rare.
Equation (2) from this chapter is used along with the assumption
of n.p = 0.05. The antenna temperature is taken to be 3 times the
standard deviation listed in Table 2. The width of the undetected
feature is taken to be 10 km/sec for one set of calculations and 200
km/sec for a second set. The first width is appropriate for a single
cloud while the second is more typical of an expected global profile
for many clouds (Rickard et al. 1977).
Also listed is the column density of assuming N(C0)/N(H2) =
5 x 10 (Martin and Barrett 1978). The molecular mass is calculated
by (Schneps et al. 1978)
M = ^(Tex)(T*/nf)AvD2 (8)
where M is the molecular mass in solar masses, n(T ) incorporates the
CA
effect of temperature on the population levels of the CO, T* is taken
to be 3 times the standard deviation, av is the assumed velocity width


in our lives. Steve's assistance and discussions with me (to say
nothing of his witticisms) were above and beyond the call of duty.
I thank my parents, Bill and Mary Ann, for their support and love
and good humor throughout the years. But I reserve perhaps my sincerest
appreciation for Maryfran and the "kids" (Smokey, Phantom, Harpo, and
Marble) for making it all worthwhile.
iv


TEMPERATURE K)
it?


-23-
Fate of Retained Gas
Supermassive Objects in the Nucleus
Mathews and Baker (1971) suggested that if their galactic wind
should fail, the gas in the system would fall to the nucleus in an
ionized state. They further argue that the Jeans radius
R.i = [
rkT
X
2
16yMG((
P*J
(2)
where T is the temperature, y is the molecular mass, G is the gravita
tional constant, p and p* are the densities of the gas and stars,
respectively, and k is Boltzmann's constant is determined primarily
by the stellar density. That is, as the gas collapses it responds to
the gravitational field of the stars. This will continue until the gas
becomes more dense than the stellar component. The gas is dense and
collapsing quickly at this stage and Mathews and Baker suggest that
there may not be enough time for fragmentation to take place. The
collapsing material then forms a massive object rather than fragmenting
and forming stars with a normal distribution of masses.
Another argument for supermassive objects is the observation that
cD galaxies with radio emission are often located in the center of
dense, rich clusters. The evidence is largely circumstantial but if
ram-jet stripping is important in the evolution of ellipticals then
it follows that the central members of a cluster will be least affected
by this mechanism. Of course the step from ineffective ram-jet
stripping to a supermassive object in the galaxy's nucleus is by no
means secure. It rests on the assumption that the retained material


-16-
For example, Mathews and Baker (1971) assume a coupling effi
ciency between the expanding supernova shell and the ISM of 1; that
is, all the kinetic energy of the supernova is converted into thermal
energy of the ISM. Gisler (1976) takes exception to this number and
notes that Larson (1974) uses an efficiency of 0.1. Given the various
uncertainties it appears that while a galactic wind will most likely
prevail in some instances, perhaps even a majority of elliptical
galaxies, there are cases in which it simply does not operate.
Indeed, Mathews and Baker (1971) find solutions in which a wind is
not supported and the material collapses to the center of the system.
They further propose that the hot, ionized gas will only be able to
form massive objects, thus linking the lack of a galactic wind to the
formation of radio sources in early-type galaxies.
Again, Gisler (1976) points out an inconsistency in this line of
reasoning. From observations one finds that strong radio sources were
more common in earlier epochs. Gisler notes that the earlier stellar
content of ellipticals is more likely to produce supernova. This
follows from the observation that only Type I supernovae occur in popu
lation II (old) stars and the precursors are probably low mass stars
(Tammann 1974). In the earlier stages of an elliptical's life the
supernova rate can only be augmented by Type II supernovae (whose
progenitors are young, massive stars). In addition it is at the
earlier epochs that the galaxy will not have had time to collect a
significant amount of gas from the evolution of its stellar component.
For these two reasons it would appear that the ellipticals are better
able to support a galactic wind at earlier epochs--just the period
when the greatest fraction must also be radio sources.


-93-
Neutral Hydrogen and NGC 185
As described in Chapter III, we have detected neutral hydrogen at
the position of NGC 185. There is also detectable HI at least as far
as 2 north of the galaxy. It is apparent that any HI physically
associated with NGC 185 is being confused with high-velocity hydrogen
belonging to our own galaxy.
The mass of HI in a particular feature is given by (Wright 1974)
Mhi = 2.356 x 105 D2 / S dV (10)
where the mass is in solar masses, D is in Mpc, S is the flux density
in Janskys (1 Jansky = 10 Wm Hz ), and dV is in km/sec. This
equation is based on the assumption of low optical depth in the 21 cm
line. While this assumption may not be valid in the plane of the
galaxy, it is quite likely true for HI out of the plane (NGC 185 is at
a galactic latitude of -15).
The confusion with high-velocity local HI is primarily due to the
Magellanic Stream, an arc of HI clouds stretching from the Magellanic
clouds through the south galactic pole and terminating in the vicinity
of NGC 185 (Mathewson et al. 1974).
That there are indeed discrete clouds in the velocity range around
-200 km/sec and within 4-6 of NGC 185 is confirmed by Giovanelli
(1979) and Hulsbosch (1980). Hulsbosch's observations of the area
show nothing at either the position of the galaxy or within 2 to the
north. This lack of confirmation of our findings is not surprising
because Hulsbosch's sensitivity is only about 400 mJy while ours is
about 25 to 40 mJy (the 43 m telescope has a sensitivity of about
4 Jy/K).


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 E5JEGB87T_QXVKK9 INGEST_TIME 2014-12-05T23:48:15Z PACKAGE AA00026480_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


-74-
cloud models of Lucas (1974) and assuming a kinetic temperature of
20 K (Morton 1975) they find that the column density is given by
N(CO) ^ 6.0 x 1014(T*/nf)Av (2)
where AV is the full-width half-maximum of the line in km/sec. Using
this expression and taking into account beam dilution (nf) we find
column densities for NGC 185 as listed in Table 6.
Table 6
Column density of 12C0 for Optically Thin, Diffuse Clouds
Source
N(CQ)
cm-2
N(C0) x 2 x 104 =
cm-2
n(h2)
Case A
8.1 x 1016
21
1.6 x 10
Case B
2.5 x 1016
5.0 x 1020
Case C
6.1 x 1015
1.2 x 1020
The mass of the molecular cloud can be found by multiplying the
column density of first by the area of the cloud and then by the
mass per particle. The result is 9680 M0 and is independent of which
case is actually correct. It is independent because we are assuming
that the source is optically thin and we are detecting all the CO
within a given region. Further, it represents a firm lower limit if
the colliding particles are if they are HI the minimum mass is
4840 M .
0
The last column of Table 6, total column density of H2, is found
by assuming N(C0)/N(H2) = 5 x 10^(Martin and Barrett 1978).


-38-
Data Reduction Techniques
The output of the autocorrelator is two 192 channel spectra, each
being linearly polarized but orthogonal to the other. The spectra are
recorded by an on-line disk drive which can also be accessed by the
on-line reduction computer. In operation, the near instantaneous access
to the data just taken enables the observer to monitor the quality of
the system operation and to update the observing procedure based on the
preliminary results.
A characteristic of the autocorrelation method of spectral analysis
is that the strong galactic hydrogen within the bandpass produces a
sinusoidal ripple in the spectrum. This is known as "ringing" and its
removal is accomplished by convolving the spectrum with a hanning
function. This function is a weighting scheme in which the value of
the channel on either side is added to i the value of the central
channel to produce the new value for that channel. Application of
this smoothing worked very well and all data presented here from the
43 m telescope have been smoothed with the hanning function.
A baseline is removed from the data by fitting a low-order poly
nomial to the spectrum in regions removed from either galactic emission
or suspected NGC 185 emission. In practice the order of the fitting
polynomial was 2 to 4.
Calibration for this system is done under computer control by
periodically firing a noise tube within the receiver and comparing
the system output with and without the additional noise. The data are
then scaled to this system temperature. The stability of the system
was monitored by observing several sources throughout the session. No
unexplained drifts in system performance were seen.


-51-
1 isted in the table is the standard deviation for all observations of a
particular source.
Figures 8-18 show the spectra obtained with the 256 MHz filterbank.
Also indicated in the figures are the radial velocity assumed for each
source. In many of the spectra there are single channels above or below
the 3 standard deviation limit. These are not significant because they
fail to persist in further observations. This type of interference is
relatively common in multi-channel filterbank receivers, and arises from
stray voltages within the equipment. The implications of the negative
results are discussed in Chapter IV.
Neutral Hydrogen
The region surrounding NGC 185 was examined to sensitive limits
to detect any neutral hydrogen (HI) associated with the galaxy. The
43 m telescope at Green Bank was used in a total power mode as de
scribed in Chapter II. The arrangement of "offs" was varied throughout
the course of the observing in an effort to delineate the HI detected.
One series of observations used an "off" at about 16m west to allow a
reasonable estimate of the total HI content of the region.
Another sequence of observing used closely spaced "offs" (24' or
48') to the west, east, south, and north to get a better idea of the
behavior of hydrogen around a specific region. For those observations
the "on" was always at the center of the "offs."
Neutral hydrogen was found at significant levels in several re
gions, both at the position of NGC 185 and several points to the north.
Figure 19, which used a distant "off," shows a strong signal of about
14.2 milliKelvin (mK) at a radial velocity of -200 km/sec. The


-85-
were taken on dust patch number 11 while the December 1979 observations
were of dust patch number 12 on the other side of the nucleus. Table 14
lists the positions observed for each patch as well as Hodge's (1973)
estimates of the dust content. The dimensions listed by Hodge are
used to find a rough area in square arcseconds and pc (the assumed
distance is 0.69 Mpc). Hodge (1973) also presents microphotometer
tracings of these dust regions indicating a visual absorption of about
0.2-0.3 magnitudes. It is clear that these regions are not as dense
as DP-la in NGC 185. Beyond this difference the two galaxies are
very similar in dust and young star content.
Table 14
Dust Regions in NGC 205
Region
no.
a(1950.0)
6(1950.0)
Area
()2
of Cloud
pc2
Dust Mass
Me
11
0h37m40!7
4125133"
360
3.9 x 103
160
12
0 37 41.3
41 24 05
880
9.6 x 103
460
A similar calculation as was made for the NGC 185 dust clouds,
under the assumption of optically thin emission (see equation (2) of
this chapter), can be performed. The antenna temperature used is 3
times the standard deviation for each cloud. The velocity width is taken to
be four times the channel width (av =10.4 km/sec). The results appear
in Table 15. The column densities are modest when compared to those
of NGC 185 (Table 6). The computed total masses are even less than
one may expect from the dust masses and a gas-to-dust ratio of 200,


-107-
a Boltzmann distribution of the rotational- levels which can be
written
nu 9u -hv/kTex
"7 = 5?
(7)
o
where nu and n-j are the populations (cm ) in the upper and lower states,
respectively, and g and g-, are the statistical weights of the levels.
s
Note that tv(s) = / lyds and (Lang 1974)
k = exp (fe )Cl exp(-¡-y^-)]A (8)
v 8ttv Av K ex K ex
where Av is the full-width at half-maximum for the line, T is again
GX
the excitation temperature and the Einstein coefficient (A) for the
spontaneous electric dipole transition is
4 3
o4tt v
3hc,^
(9)
where
= y
W
+ l
+ 3
(10)
-18
for the J+l -> J transition. Since y = 0.112 x 10~ in cgs units,
|yj|^ = 5.02 x 10"^ for the J = l->0 transition, A = 8.9 x 10"^ sec"''.
s
Setting N-, = J n,ds we can write
1 0 1
\> ~2 ^ P P (if-)] (11)
8-rrv ex ex
Substituting (7) into (11), assuming hv kT and rearranging
CA
gives


Observations of 10 other early-type systems are also presented
and discussed. The negative results imply that the gas is either
clumped and thus optically thick or has been removed from the system
through a galactic wind, ram-jet stripping, or has been consumed by
star formation.
The nature of the star formation must be somewhat different than
that in our own galaxy. The high-mass end of the initial mass function
for star formation would result in bluer colors than observed, the
star formation must be skewed towards the low-mass stars to be effec
tive yet unobserved. Theoretical arguments that this is possible are
advanced, but more sensitive and highly resolved CO observations are
necessary to observe directly this scale of star formation.
Neutral hydrogen observations of N6C 185 obtained with the NRAO
43 m telescope are presented and discussed. There is apparently low-
level (20-40 mK) high-velocity hydrogen in the region of NGC 185. The
most likely source of the material is the Magellanic Stream which ter
minates in this area. Superposed on this high-velocity material, at
the same velocity, is an excess of about 5 mK of HI at the location of
NGC 185. This was detected by using "off" spectra 1-2 beamwidths from
the galaxy in the four cardinal directions. The observations cannot
distinguish between an enhanced high-velocity feature projected onto
the galaxy and genuine emission from the galaxy itself, but the results
are very suggestive and should be followed up with observations of
greater resolution.
vm


-24-
either forms the supermassive object or at least provides fuel for the
radio emission.
These arguments actually rest on much firmer ground due to recent
work on the velocity dispersions and light distributions within the
nuclei of supergiant cD galaxies found in the centers of rich clusters.
Young et al. (1978b)obtained luminosity profiles of the supergiant
elliptical NGC 4486 (M87) which, when examined with the velocity dis
persions determined by Sargent et al. (1978), show that the nucleus
contains a massive dark object. The nature of the dark object cannot
Q
yet uniquely be determined, but it must contain 5 x 10 M of material

and have a radius less than or equal to 100 parsecs (pc). Young et al.
also determined that the mass-to-luminosity ratio must be greater than
60. Several possibilities are advanced but Young et al. find the most
g
plausible to be a massive black hole of 5 x 10 M The attraction

42 -1
of this hypothesis is that the 10 erg sec energy output of NGC 4486
-2 -1
can be explained by supposing a mass infall of about 10 M0 yr with
only a 0.002 conversion efficiency into radiation.
De Vaucouleurs and Nieto (1979) confirmed the results of the
Young et al. (1978b) study and found essentially similar results for
the dark mass at the nucleus. The earlier results were obtained with
a CCD (charge-coupled device) camera, while the work by de Vaucouleurs
and Nieto was with more conventional photographic and photoelectric
photometry.
Young et al. (1979) also examined the luminosity profiles of NGC
4874, 4889, and 6251 and found that only NGC 6251 requires a supermassive
object at its center to fit the data. Further, only NGC 6251 is a radio
source amongst the three.


-18-
if a galaxy has a velocity dispersion greater than 200 km/sec then the
supernovae contribution will not be able to double the kinetic energy
of the ISM and a galactic wind cannot be established.
Even more damaging to the galactic wind hypothesis is the de
tection of HI in any elliptical. In order for the wind to operate
the ISM must be hot lO'7 K). All of the gas in a galaxy would thus
be ionized and according to Mathews and Baker (1971) quite unobservable
by present techniques.
A final argument against the universal existence of galactic winds
is that if all other conditions were the same, one would expect the
more spherical systems to be better able to support a galactic wind.
The reasoning is that the spherical systems have less surface area per
unit volume through which to radiate excess energy, keeping the ISM
as hot as possible.
Contradicting this expectation, the neutral hydrogen observations
of 8 elliptical systems show detections significantly skewed towards
the more spherical galaxies. The systems detected in neutral hydrogen
have the following classifications: 2-E0, 2-El, 1-E2, 1-E3, 2-E4.
Conspicuously absent are the flatter systems that one would expect to
be better able to radiate energy away and thus retain their ISM.
Ram-Jet Stripping by an Intracluster Medium
The proposal that ram-jet stripping of an ISM could be significant
in the evolution of a system was treated first by Gunn and Gott (1972).
Later, more sophisticated treatments by Tarter (1975), Gisler (1976),
and Lea and De Young (1976) all support the notion that stripping can
be an effective process.


TEMPERATURE (K)
-0.10 0.00 0.10 0.2,0
(X101 )
The ^CO data taken in July 1978 on dust patch no. 11 of Hodge (1973) in NGC 185.
Figure 8


TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS iii
ABSTRACT vii
CHAPTER
IINTRODUCTION 1
The Nature of Early-Type Galaxies 1
Galaxy Classifying Schemes 2
The Morphology of Elliptical Galaxies 3
The Morphology of Lenticular (SO) Galaxies 6
The Formation and Evolution of Early-Type Galaxies ... 7
Formation 7
Mass Accretion 11
Mass Loss Due to Stellar Evolution 13
Stripping Mechanisms 15
Internally Driven Winds 15
Ram-Jet Stripping by an Intracluster Medium .... 18
Fate of Retained Gas 23
Supermassive Objects in the Nucleus 23
Cyclic Bursts of Star Formation 25
Continual Star Formation with a Skewed Initial
Mass Function 27
Statement of Dissertation Problem 29
IITHE OBSERVATIONS 31
Carbon Monoxide Observations at Kitt Peak, AZ 31
Telescope Description 31
Data Reduction Techniques 34
Data Presentation 35
Neutral Hydrogen Observations at Green Bank, WV 36
Telescope Description 36
Data Reduction Techniques 38
IIIRESULTS 39
Carbon Monoxide 39
Positive Result in NGC 185 39
Negative Results 48
Neutral Hydrogen 51
IVDISCUSSION ..... 68
NGC 185 68
NGC 205 84
v


-91-
a mechanism that would allow the gas to maintain the same spatial dis
tribution as the stars that shed it. As long as the gas can cool itself
effectively, usually via line radiation, it must collapse into the po
tential well of the galaxy. If it cannot cool itself, a galactic wind
sets up and the system is swept clean. The point has been made pre
viously that galactic winds are not universally effective, consequently
one expects the interstellar material to preferentially collect in the
nucleus.
Comparing the maximum reasonable amount of molecular material that
could escape detection under these assumptions (t < 1, av = 200 km/sec)
with the expected HI content reveals that molecular storage is not a
dominating feature of the galaxy's ISM. The largest discrepancy, for
NGC 205, may not be significant because the region sampled for CO is
small with respect to the dust patches.
For the other galaxies the discrepancy may be due to one, or a
combination, of three processes: the gas may be clumped and thus
optically thick, the material could have been removed either by a
galactic wind or ram-jet stripping, or finally, star formation could be
consuming the gas with an IMF deficient in high-mass, bright stars.
A fourth possibility, that the clouds are similar to Knapp and
3 -3
Jura's (1976) thin clouds with n ^ 10 cm but with no CO emission, is
unlikely. The CO observations included a large portion of the galaxy's
nucleus and it seems highly improbable that all the clouds will be in
this nether region of collapse with no CO emission. Certainly some
fraction may be without CO emission but others, further collapsed,
should produce CO emission.


-64-
position of the "on" is at NGC 185. A full-width at half-maximum
(FWHM) for the feature is about 60 km/sec.
It is apparent though, from Figure 20, that all of this HI may
not be associated with NGC 185. The "offs" for Figure 20 are taken at
either 24' or 48' (they have been mixed to increase the SN ratio for
this spectrum) in the four cardinal directions. The fact that the signal
is only 5.2 mK indicates that some, but not all, of the line has been
subtracted out in the "offs."
Further, spectra were generated separately for each of the four
directions. Those to the west, east, and south show about a 3
standard deviation positive signal. However, the spectrum with an
"off" to the north shows no significant feature in the NGC 185 velocity
range.
Continued examination of points to the north of the galaxy re
veals a strong 40 mK feature about 70' directly north. The radial
velocity of this feature is -180 km/sec with a RWHM of 40 km/sec.
Five point maps of the region to the north also indicate HI signals
as far as 2 north and 48' east with a radial velocity of about -180
km/sec. There is only weak HI emission at 49' north of the galaxy.
This is shown in Figure 21; the feature is around 15 mK. It appears
that there is a ridge or plume of HI that is strongest (43 mK) between
l-2 directly north of NGC 185. It extends further north and east at
a level of around 20 mK.
To the south towards the galaxy the plume decreases in strength to
about 10-15 mK. From the sensitive 5-point maps at the position of
NGC 185, an excess of about 5 mK is found for the galaxy's location.
These observations are summarized in Table 3. The interpretation of
these observations is found in Chapter IV.


-103-
The difference in antenna temperature, between the source and
the sky is
. T- *r t
A source = source ~ sky
= G n0nfexp(-x A)(1 e"T)[J(vc,Tr) J(vc,T. )] (8)
S A"f
s EJ
s bg'
The corrected antenna temperature of the source is defined as
TA 5 iTsource/Gs\exP(-Tsfl> =
(9)
cal
where Tc is obviously
Tc s ATcal/[GsVxp(-Tsfl)]
(10)
Consequently
T* = nf(1 e"T)[J(vs,TE) J(vs,Tbg)] (11)
Note that several parameters do not appear, namely the receiver
gain (G ), atmospheric extinction exp(-x A), and antenna losses (n0)-
These have been corrected for by an appropriate choice of T From
equation ( g ) it can be seen that we need only the ratio of the source
temperature to the calibration temperature in addition to Tc to deter
mine the source antenna temperature.
If we assume that the IF is small with respect to the L0 frequency
(a very safe assumption for this work) then J(v^,T) ^ J(v ,T); that is,
the signal and image temperatures are close enough to be considered
equal then


-75-
A further finding by Knapp and Jura (1976) is that a necessary,
but not sufficient, condition for CO in emission is a particle density
3 -3
£10 cm This agrees with the theoretical result noted in Appendix
II. We thus feel confident that the CO emission in NGC 185 is
3 -3
originating in a region at least as dense as 10 cm .
It would be useful for calculations of the space density of various
particles if the physical size of the absorbing regions were known.
This would help constrain the various possible sources of the CO emission
and help in determinations of the mass of gas involved. An estimate
of the cloud volume can be made by assuming they have a dimension along
the 1ine-of-sight equal to the average width. This approach will tend
to underestimate the volumes unless the clouds are flat and oriented
broadside to the 1ine-of-sight. Table 7 tabulates the result of this
calculation. The gas mass listed in the last column is found by
assuming the entire cloud is filled with molecular hydrogen at a
density of 108 cm-8.
Table 7
Estimates of Cloud Volumes for NGC 185 Dust Clouds
Dust
Cloud
Depth of Cloud
(pc)
Volume
(cm3)
(pc3)
Mass of Gas
(M )
DP-Ia
6.6
7.37 x 1058
2.5 x 102
6.2 x 104
DP-I
16.5
6.06 x 1059
2.1 x 104
5.1 x 105
DP-11
42.9
4.79 x 1060
1.6 x 105
4.0 x 106
Hodge (1963) has estimated the mass of dust in DP-I and DP-II.
Using the dust masses and the total gas content for the optically thin


-37-
The 21 cm cooled cassegrain receiver was used for all data acquisi
tion. The system has two channels provided by linearly polarized,
orthogonal feeds. After initial amplification by a cooled upconverter
amplifier the signal is heterodyned and amplified through various stages
in much the same fashion as the process described for the 11 m tele
scope at Kitt Peak. Typical system temperatures for the 43 m system
were 50-60 K.
The standard NRAO "back end" uses a 150 MHz IF which is fed into
a Model II autocorrelator spectrometer. The IF signal is autocorrelated
and the resulting autocorrelation function is Fourier transformed to
produce the power spectrum. The formation of the spectrum using auto
correlation techniques is described in more detail by Blackman and
Tukey (1958) and Cooper (1976).
A 10 MHz bandwidth was chosen for all observations to provide an
adequate baseline. Also in the interest of baseline stability a
position-switched mode of observing was adopted. Ten minutes of data
are taken at the "off" position followed by 10 minutes at the "on"
position. The final spectrum is found by differencing the two spectra
thus acquired.
A number of "off" positions were used in an effort to deduce the
distribution of HI in the region around NGC 185. The majority of the
data were taken either with an "off" 16m to the west or as a five-point
map. The arrangement for the five-point map is with the "on" at the
center and "offs" taken successively to west, east, south, and north
at a distance of either 48' or 24'. The half-power beam-width at 21 cm
is about 20.51.


to Maryfran