Group Title: morphology and energetics of discrete optical events in compact extragalactic objects /
Title: The morphology and energetics of discrete optical events in compact extragalactic objects /
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00099363/00001
 Material Information
Title: The morphology and energetics of discrete optical events in compact extragalactic objects /
Physical Description: ix, 177 leaves : ill. ; 28 cm.
Language: English
Creator: Pollock, Joseph Thomas, 1950-
Publication Date: 1982
Copyright Date: 1982
 Subjects
Subject: Quasars   ( lcsh )
N stars   ( lcsh )
Optical measurements   ( lcsh )
Variable stars   ( lcsh )
Pulsating stars   ( lcsh )
Astronomy thesis Ph. D   ( lcsh )
Dissertations, Academic -- Astronomy -- UF   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Thesis: Thesis (Ph. D.)--University of Florida, 1982.
Bibliography: Bibliography: leaves 95-97.
Statement of Responsibility: by Joseph Thomas Pollock.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00099363
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000318978
oclc - 09248080
notis - ABU5827

Downloads

This item has the following downloads:

morphologyenerge00poll ( PDF )


Full Text


















THE MORPHOLOGY AND ENERGETIC OF DISCRETE OPTICAL
EVENTS IN COMPACT EXTRAGALACTIC OBJECTS











BY

JOSEPH THOMAS POLLOCK


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


1982















ACKNOWLEDGMENTS

With this dissertation a quarter of a century of my education ends.

Throughout this time I have received the constant encouragement and

support of my parents, Joseph and Cecelia Pollock. They instilled the

common sense, inner strength and sense of humor necessary to complete

this endeavor, and it is to these two very special people that this work

is lovingly dedicated.

Dr. Peter D. Usher and Dr. Alex G. Smith, my master's and doctoral

advisors, together have converted a fledgling graduate student into a

professional astronomer. I can only hope that my research and teaching

abilities, and general professionalism can someday approach the examples

of these two men. To each of them goes my deep gratitude. I would also

like to thank the other members of my committee, Dr. T. D. Carr, Dr. J. P.

Oliver, Dr. H. C. Smith and Dr. P. Kumar.

It is a pleasure to recognize three other fine professionals who

have opened doors of opportunity. Mr. Stanley Walker allowed me to

obtain advanced mathematical training in high school. Dr. Satoshi

Matsushima gave a shakey undergraduate the chance to prove himself in

graduate school. Dr. William Liller provided the opportunity for me to

do extensive research and to publish the same.

All of my fellow graduate students have provided invaluable

support, both scientific and spiritual. My special thanks go to Betty

Whitmire, Greg Fitzgibbons, my shadow Andy Pica, Dan Caton and Jim

Webb.












Special thanks go as well to Irma Smith, my typist, whose ability

and knowledge made creation of the final copy so painless.

My thanks go as well to all the other people who, from day 1,

have helped to smooth out the rough spots.

For as long as I can remember my Grandmother asked me, "Are you

finished school yet?" This past Christmas day she passed away at the

age of 96. I hope she knows that I finally made it.


iii
















TABLE OF CONTENTS

ACKNOWLEDGMENTS . . ..............

LIST OF TABLES . . . . . . .

LIST OF FIGURES . . . . ... .

ABSTRACT . . . . . . . .

Chapter

I. INTRODUCTION . . . . . .


Historical Background ..
Group Subdivisions .. ...
Overall Problems Relating to
Objectives of This Study .


Compact Sources


TI. OPTICAL VARIATIONS OF COMPACT OBJECTS .. ....

Discovery of Optical Variability .. .....
General Variability Characteristics .. ...
Data Acquisition and Reduction Techniques .. ..
Sources of Compact Object Variability Data .. ...


III. TIE VARIABILITY PROBLEM . ....


Apparent Magnitude .. .....
Absolute Magnitude or Emitted Flux .
Rest Frame-Standard Wavelength Flux.


IV. DATA HOMOGENIZING AND TRANSFORMATION .. .....

Photometric Systems .. . .......
Standard Rest Frame System .. . .......


V. EVENT DETERMINATION AND ANALYSIS .

Event Definition . ..
Identification of Events . .
Event Objects . ....
Event Parameter Calculations .
Extreme Luminosity Changes .
Normalized Event Luminosity Curves
Event Morphology . ..
Individual Event Morphologies. .


21


. . 30

. . 30
. . 31
. . 36
. . 39
. . 45
. . 45
. . 47
. . 49


I I I : j I I (












Individual Event Comments. . . .
Variability and Emitting Region Sizes . .
Bolometric Luminosities . . .

VI. COMPARISONS WITH THEORETICAL MODELS . . .

General Energy and Emission Constraints . .
Black Hole Models . . . .
Supermassive Rotating Magnetoplasmic Body .
Multiple Supernovae . . . .
Final Comments and Suggestions for Future Work .


LIST OF REFERENCES . . . . .

APPENDICES . . . . .

A. LIST OF VARIABLES AND SYMBOLS . .

B. REFERENCES TO OPTICAL VARIABILITY DATA .

C. NORMALIZED EVENT LUMINOSITY CURVES . .

BIOGRAPHICAL SKETCH . . . . .


53
56
58

68

69
75
91
92
92

95

98

99

102

110

177
















LIST OF TABLES

1. EVENT MAGNITUDE CHANGES. ... .. ........... 21

2. PG MAGNITUDE CORRECTIONS . . . .. 26

3. EVENT OBJECT PARAMETERS. .... . . . 37

4. EVENT PARAMETERS I .... .... . . . 40

5. EVENT PARAMETERS II. . ... . .... 42

6. FRACTIONAL FLUX DURATIONS. .... . . 50

7. FRACTIONAL LUMINOSITY FACTOR . . . 57

8. EMITTING REGION SIZES AND VARIABILITY TIMESCALES . .. 59

9. EVENT LUMINOSITIES AND ENERGIES. ... ... ..... 64

10. EVENT ENERGY DENSITIES AND MAGNETIC FIELDS . ... 70

11. EVENT PARAMETERS III . . . . 73

12. MASS ACCRETION RATES .. . . . . 77

13. EVENT ENERGIES AND CONVERTED MASSES. . . 78

14. UPPER AND LOWER MASS LIMITS. . . . . 84

15. DISK PULSATION PERIODS ..... . . . 86

16. MAGNETIC FLARE MODEL PARAMETERS. . . . 89















LIST OF FIGURES

1. Sample light curves containing long term trends. ... 32

2. Sample light curves containing discrete events . .. 33

3. Sample light curve containing intermediate events. .. ... .34

4. Normalized event luminosity curve for a type I supernova 48

5. Log of the variability timescale in seconds versus the log
of the bolometric luminosity in ergs/sec . ... 81















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


THE MORPHOLOGY AND ENERGETIC OF DISCRETE OPTICAL
EVENTS IN COMPACT EXTRAGALACTIC OBJECTS

By

Joseph Thomas Pollock

May 1982

Chairman: Alex G. Smith
Major Department: Astronomy

All available optical variability data for compact extragalactic

objects were assembled. For objects showing evidence for one or more

discrete events, this magnitude and data information was digitized,

composite light curves generated and discrete events identified. All

event data were converted to rest frame 2500 R monochromatic lumin-

osities. The base luminosity was determined for and subtracted from

each set of event data, and normalized event luminosity curves were

generated. The morphology of these events was examined, both for

single events and for groups of events.

Determinations were made of event parameters such as the rest

frame durations, 2500 A peak and base luminosities, event to base

energy ratios, total event energies and shortest significant variability

timescales. From these and object parameters, values were obtained

for lower limits to the peak bolometric luminosities and event energies,

as well as the maximum size of the emitting region and resulting limits

on the energy densities.


viii












Under the assumptions of various models for these types of

objects (mostly involving mass accretion onto a central black hole),

further parameters such as mass accretion rates, total accreted masses,

upper and lower limits on the mass of the central object and magnetic

field strength limits were determined.

In only a few cases were these objects found to be emitting at

a rate greater than their Eddington luminosities. A possible correlation

between higher luminosities and shorter variability timescales was

found, indicative of beamed radiation. Not all events could be explained

in terms of non-radial pulsations of accretion disks. A model employ-

ing quiescent luminosity from mass accretion and event luminosity from

magnetic flares on the accretion disk appears promising in terms of

both energetic and emitting region sizes. The masses converted during

the events were found to lie within the range of typical stellar masses.















CHAPTER I
INTRODUCTION

Historical Background

N-galaxies

The objects known as N-galaxies (for nuclear) are so classified

because their optical appearance is that of a relatively faint galaxy

with an extremely bright, compact nucleus. The classification is a

morphological one and is thus dependent on the distance to the given

object.

A group of objects, later to be designated a subset of the class

of N-galaxies, was examined by C. Seyfert in his doctoral dissertation

in 1943. He investigated a small set of spiral galaxies which exhibited

two properties not seen in most other "normal" spirals. This set of

objects, now designated as Seyfert galaxies, showed extremely bright,

compact nuclei (thus making them N-galaxies) and their spectra displayed

strong, broad emission lines. Later studies have shown some of these

objects to be strong radio emitters and X-ray sources, and to be vari-

able in their radio, infrared and optical emission. These additional

properties are more recent discoveries that resulted, to some extent

at least, from the discovery and study of a second class of compact

extragalactic objects. These are the quasi-stellar objects or "quasars"

as they are now commonly referred to. It should be noted that not all

N-galaxies exhibit the Seyfert characteristic of strong emission lines,

a point which is often confused.












Quasars

The discovery of quasars was a direct result of the increase in

the accuracy of positional measurements for radio sources in the late

1950's. T. Matthews and A. Sandage in 1963 found that a 16th magnitude

"star" coincided with the position of the Cambridge radio source 3C 48.

Optical spectra of this "radio star" contained unidentifiable, broad

emission lines unlike those found in the spectra of normal emission

line stars. A stellar object with characteristics similar to 3C 48

was then found at the position of SC 273. Finally Schmidt (1963)

recognized the spectral lines of 3C 273 to be the Balmer series of

hydrogen, but highly redshifted from their normal rest wavelength.

This was tremendously significant since redshifts of this nature were

seen only in extragalactic objects whose distance was then determinable

using Hubble's law. Using the distance derived in this way, the bolo-

metric luminosity of 3C 273 could be calculated from a knowledge of

the flux received from the object at all observable wavelengths (in

principle). The result was about 3 [47] ergs per second (this notation

for powers of ten will be maintained from this point on), which is

about 6000 times that emitted by our own galaxy and about 300 times

that of NGC 1275, one of the brightest known Seyfert galaxies. However,

3C 273 was of a size such that it could not be resolved by optical

telescopes.

There is a caveat which must be placed here, however, concerning

the determination of the distance to 3C 273 and all other quasars.

It has been tacitly assumed that the origin of the redshifts of these

objects is cosmological--that is, the redshift is a result of the

object's taking part in the general expansion of the universe. If these












redshifts originate from another mechanism, such as gravitational red-

shift or local acceleration, then distances obtained through the use

of Hubble's law are meaningless and the calculated luminosities will

be wrong. This, of course, grossly affects the nature of such problems

as the method of energy generation. Debate over quasar redshift origins

has raged for almost twenty years. The presently prevailing opinion is

that the redshifts are indeed cosmological; however smaller groups of

vocal dissenters continue to feel that their evidence points to a non-

cosmological origin. It will be assumed from here on in this work that

the redshifts of these objects are valid indicators of the distance to

them.

Group Subdivisions

As is always the case when a class of objects is scrutinized

closely and is expanded in total membership, subdivisions based on

differing observed properties are created. Major subdivisions of

Seyfert galaxies and quasars are briefly described below.

Seyfert Galaxies

Seyfert galaxies are divided into types I and II depending on

the width of the forbidden emission lines in their spectra. Type I

exhibit narrow forbidden lines whereas those seen in type II are broad,

often having widths of several thousand kilometers per second.

Quasars

There is a variety of criteria which are employed to divide up

the general population of quasars as indicated below.

(a) Optical spectral features
(b) Presence or absence of strong radio emission
(c) Centimeter "excess" radio sources












(d) Nature of optical variability
(e) Method of discovery
(f) X-ray emission

The above list is by no means exhaustive. The three earliest subdivi-

sions which developed were "classical" quasars, "radio-quiet" quasars and

BL Lacertae objects. These classes can be compared and contrasted in

the following ways.

3C 273 is a prototype of the classical quasars. It is starlike in

its optical appearance (excepting a faint jet of material apparently

emanating from the central region), has a highly redshifted emission

line spectrum,-has a partially polarized non-thermal optical continuum

showing both ultraviolet and infrared excesses and is a strong radio

emitter.

Not all quasars, however, emit a detectable radio flux. Those

which do not are the radio-quiet quasars which are usually discovered

by searching for objects with ultraviolet excesses and then obtaining

their spectra in order to confirm their nature. Many of these blue

stellar objects (BSO's) were catalogued, well before the identification

of 3C 273, by means of multi-color surveys carried out at observatories

such as Tonantzintla and Mount Palomar in programs to search for very

blue stars. However, spectra were never obtained for them and their

actual nature remained unknown at that time. A second search technique

now extensively employed involves the examination of objective prism or

grating plates to locate and classify these objects in a single step.

Such surveys are beginning to indicate that a major fraction of all

quasars may be of the radio-quiet variety.













The third class is comprised of the BL Lacertae objects. The

prototype of this category, BL Lacertae, was originally classified as a

variable star but was later found to coincide with a strong radio

source. A major difference between these objects and the classical

quasars is that the BL Lacs have very weak emission lines or totally

continuous optical spectra. In addition they are usually neutral in

color, not having strong ultraviolet excesses. These properties limit

greatly our ability to identify this type of compact source and we must

rely, in general, on a positional coincidence with a radio source to

locate them. Some of the spectra, as mentioned earlier, are devoid of

any emission or absorption features. This makes it impossible to deter-

mine their distance from Hubble's law and thus one cannot calculate

directly their luminosities. Because BL Lacertae objects are usually

extremely violent optical variables, their light curves are very

conducive to the study of the phenomenon of variability. Unfortunately,

for some of these objects the physical parameters that one would like

to derive from such a study cannot be determined since the distance

to the object remains unknown.

Overall Problems Relating to Compact Sources

The ultimate goal of the study of compact extragalactic objects

is to devise a model that satisfactorily predicts the observed charact-

eristics. One would hope that at some point in the future these models

would be as successful as present day models for stellar atmospheres

and interiors. It is possible that these objects are so physically

complex and varied, however, that no such unified theory could ever

be devised. The potential exists as well, as it did with the source of











stellar energy at the turn of this century, that the physical principles

necessary to the solution of the problem have not yet been discovered.

A few of the major questions to be answered are given below. Some

of these seem to have fairly satisfactory explanations while others are

little understood.

(a) What is the "engine" which can provide the vast quantities of
energy which are emitted?

(b) What causes the luminosities to change by a factor of two in
a day or one hundred in a month in some of the most violent
objects?

(c) Are normal galaxies, Seyfert galaxies and quasars linked by
morphology alone, or do they represent different phases
within a single object's evolution?

(d) What are the physical conditions which give rise to the
emission and absorption line systems?

(e) What is the mechanism which is responsible for the polarized,
non-thermal continuum?

(f) What causes the apparent super-relativistic velocities seen
in some radio interferometric observations?

There are many more detailed problems as well.

Objectives of This Study

Since 1963, a tremendous amount of data on the optical variability

of these objects has been gathered. In few instances, however, have the

data been thoroughly analyzed in terms of the physical parameters they

imply for the emitting objects. Even less has been done to intercompare

the results so obtained among a large sample of these objects.

In order to make a comprehensive study of this sort it is necessary

to proceed in a fashion similar to the one which follows. First, locate

and gather together all of the available optical variability data refer-

ences. This is a somewhat large and unexciting chore, which may explain












why it has not been done before, at least as thoroughly as is necessary.

Then the data must be organized so that all observations of an individual

object can be gathered together and examined. Next, all of the data

to be used in the analysis need to be put into machine readable form,

which consists of typing thousands of pieces of data by hand. These

data must then be converted to a standard system and finally, using

other known object parameters, these standardized data are converted

to physical quantities in the rest frame of the object. Once this

stage has been reached it is possible to analyze the information and

intercompare results within the group of objects under study.

The focus of this study is what can be described as discrete

optical events (commonly referred to as flares or outbursts). It is

during these periods of intense activity that the most extreme physical

conditions occur. These extreme physical conditions place the severest

constraints on the "physics of what is going on," including the energy

generation mechanism, emission and transfer of radiation processes and

actual macroscopic structural changes. In addition, these events allow

a determination of the actual size of the region where the emission is

taking place. It is the intention of this study to obtain realistic

and self-consistent values for those extreme physical conditions in as

large a sample of objects as is possible. Then these parameters will

be compared with the predictions of present theories in hopes of narrow-

ing down the possible processes which are occurring in these enigmatic

objects.















CHAPTER II
OPTICAL VARIATIONS OF COMPACT OBJECTS

Discovery of Optical Variability

The earliest known information on the optical brightness of compact

objects was recorded in the late 1880's as photographic astronomy developed

and sky mapping and photographic patrol programs, such as those of the

Harvard College Observatory, were initiated. These observations were,

of course, serendipitous. Several variable "stars," which in recent

times were found to be quasars, were observed and cataloged in the early

part of this century. Among these were AP Librae, W and X Comae and

BL Lacertae.

The first variability observations of an identified compact source

were made by Matthews and Sandage (1965) for the source 3C 48. Although

the total range in magnitude was relatively small (B=16.44 to 16.68), it

pointed the way to more intensive studies of the optical behavior of

these objects. Matthews and Sandage also reported significant variations

on a timescale of 15 minutes. Such "short-timescale" variability will be

examined in more detail later. Smith and Hoffleit (1963) soon published

the first historical light curve for a QSO (3C 273) derived from the

Harvard photographic archives. In about 80 years of coverage 3C 273

varied irregularly between B=11.8 and B=13.2, about one and one-half

magnitudes. Also, it was seen to change a magnitude or more in just a

few months. At this point various groups initiated optical monitoring

of the few objects of this type known at that time. In addition, more

historical light curves were being derived and the data accumulated very

rapidly.












General Variability Characteristics

What these early studies indicated, and what studies since then

have confirmed, is that there is no known compact extragalactic object

whose variations have any strict periodicity. Several cases of short

term quasi-periodicity have been reported (most notably in 3C 345 by

Kinman et al., 1968). This case led Morrison (1969) to propose a model

of these sources as giant pulsars. Sadly, future cycles of this period-

icity failed to appear, leaving the interpretation of the earlier observa-

:tions, in doubt.

If one examines historical light curves of these objects extending

for 30 or more years, one sees little to suggest any regular behavior.

Some 60 light curves of this type have been published (see the References

to Optical Variability Data Appendix) and the end result is that the

only regular feature of the light curves is their irregularity. One

important point has emerged, however. The number of non-variable objects

(having this type of coverage) is small.

Shorter term contemporary monitoring studies (5-15 years) such as

those in progress at Herstmonceux, the University of Florida and the

Trudy Observatory of Leningrad do indicate some differentiation in

variability behavior over this sort of period. An initial classification

of the objects, based on their variability, vas made by Penston and

Cannon (1970). Objects which displayed changes of a magnitude or more on

a timescale of days or weeks were designated as OVV's (Optically Violent

Variables), whereas objects not yet found to vary in this way were called

non-OVV's. If one has a sufficient number of datum points and a long

enough time baseline, other general light curve subclasses can be defined.











Subclass I includes objects whose behavior is dominated by rapid flicker-

ing without underlying long term trends. Objects of Subclass II, con-

versely, show long term changes in their mean level which are much

larger than any short term flickering. In Subclass III the short and

long term effects are of comparable amplitudes. Finally, Subclass IV

is episodic, displaying long intervals of quiescence interrupted by

short periods of activity. As Pollock et al. (1979) point out, however,

some (or perhaps all) of these objects may alter their variability

characteristics from epoch to epoch (see, e.g., OJ 287, Pollock, 1975).

Thus it is preferable to consider these classifications as representative

of recent behavior, rather than as necessarily indicating some intrinsic

differences between the sources. It seems quite likely that the overall

variability classes are related to the macroscopic structure of the

objects in some way. The OVV, non-OVV classification is subject to

temporal variations as well, as shown by the historical light curves

of 1156+295 and 3C 454.3.

Large variations, on the order of one magnitude, are well docu-

mented for timescales as short as a day. Significant variations (that

is variations which are well above the experimental errors involved and

which are confirmed by two or more observers) are nearly as well docu-

mented for periods of a few hours. Changes of this type, not surpris-

ingly, are usually detected during violent outbursts of objects and

often are a segment in a larger, longer variation in flux.

Attempts to detect significant variations on timescales of a half-

hour or less have met with mixed success and great controversy. In at

least two cases, detection by one observer of rapid variations well












above the formal error of the observations was not confirmed by a second,

simultaneous but independent set of observations. Detection of small

changes (less than 0.10 magnitude) in short time periods is not a simple

procedure, especially when the variations are irregular in nature.

Capabilities and limitations of various observational methods are detailed

in the next section. It seems at present that caution will reign con-

cerning very short timescale variability measurements, and that it will

take two or more simultaneous and independent observations of the same

variation to instill confidence concerning their reality.

Data Acquisition and Reduction Techniques

The two major observational techniques employed to obtain magni-

tudes for compact extragalactic objects are direct photography and

photoelectric photometry. Each is discussed in more detail below con-

cerning its applications, characteristics and accuracy.

Photographic Photometry

A majority of the compact source magnitudes used in this study were

derived by means of photography, utilizing either existing plates taken

for some other purpose or plates taken specifically of an object or

objects of interest.

In the case of already existing exposures, such as those in the

Harvard Archives, one must deal with plate materials which vary greatly

in terms of types of emulsions and filters used, plate limiting magnitudes,

types of recording instruments and off-centering of the field of interest,

along with general plate quality (which deteriorates with time). Fortun-

ately, most of the exposures taken employed a plate-filter-instrument












combination which yielded a photometric system reasonably close to the

standard Johnson B-magnitudes. Indeed, a majority of the blue spectral

region plates taken since 1960 used an -0 spectroscopic emulsion and a

Schott GG13 filter, which is the standard photographic Johnson B com-

bination. The two main limiting factors in the use of archival data

to study the variability of compact sources are the plate limits and

the temporal coverage. Few plates will have limiting B-magnitudes of

fainter than 16 or so, and a majority of the quasars are fainter than

this on the average. For 3C 273 (B%13.0) Smith and Hoffleit (1963)

found over one thousand Harvard archival images, whereas for 1538+14

(B%17.0) Pollock (1975) found only 18 images. The number of plates

containing a given field will vary as well, especially for the non-

patrol plate series.

How well an object is covered in time is another factor. For

2254+07 (Pollock, 1975) 86 images were found, but 65 of these were from

the period 1938-1940. This gave excellent coverage for that period but

a poor determination of the long term light curve characteristics.

The second source of brightness determinations is from monitoring

programs such as that conducted at the University of Florida. Here

controls are exerted so as to produce standard system plates having

adequate plate limits for quasar study. In addition, the temporal

coverage can be defined as well.

Once an exposure has been obtained, how is an actual magnitude

for the object extracted from it? The two standard approaches are iris

photometry and visual estimation (the Argelander method) by comparison

with a nearby set of standard stars.












Iris photometry involves the determination of the relationship

between object apparent magnitude and image size and density for each

individual plate. This relation, the so called calibration curve, can

be determined only if there exists a series of stars of known apparent

magnitude in the vicinity of the object on the plate. By measuring the

object in the same manner as the stars used to determine the calibration

curve, the object's magnitude can then be determined by means of that

calibration curve. In general, for an object of typical quasar magnitude

(14 to 20), the estimated error in the magnitude determination ranges

from about 0.05 to 0.15 magnitude as indicated by the rms scatter of the

comparison stars about the adopted calibration curve. In the Florida

program the time necessary to measure the object and from 10 to 20

comparison stars, plus that required to computer-fit a least squares

calibration curve, is about 15 minutes. To reduce 1000 plates of 3C 273

by this method would thus have required about 10 man-days.

The quicker, but somewhat less accurate, method of visual inter-

polation may instead be employed, if the object's comparison sequence

meets certain criteria. Both the object and the sequence should be

simultaneously viewable with a magnifier. The magnitude limits of the

sequence should encompass the variability range of the object and the

magnitude interval between successive comparison stars should be no more

than 0.5 magnitude. Then the trained eye can estimate the object's

magnitude to within about +0.10 of the value determined off the same

plate by iris photometry (Pollock, 1975). This implies a total estimated

error of about +0.20 magnitude from the "true" value. An estimate of

this type takes about thirty seconds, a factor of 30 less than that of

iris photometry.












Uniformity of the data, as pointed out earlier, is of great

importance to this study. The zero-point errors inherent in either of

these techniques depend upon the accuracy of the comparison sequence.

Ideally one wishes to have photoelectrically determined magnitudes for the

comparison stars to be used. If carefully done, there should be no

significant zero-point error and less than 0.05 magnitude of random

error in each star's magnitude. Many objects, however, do not yet have

such sequences available.

Some sort of photographic transfer is the next choice. This usually

involves taking four equal length exposures on the same plate, two of

which are the object field and two are of a nearby field which contains

a sequence of photoelectrically calibrated stars. Four exposures are

necessary to eliminate the errors due to changes in plate sensitivity,

the extinction coefficient, seeing and focus (Stock and Williams, 1962).

The calibration curves derived for the standard fields are then used to

determine magnitudes of the selected comparison stars in the object

field. This technique requires relatively constant sky transparency and

guiding, equal exposure times and constant and uniform emulsion sensi-

tivity. If all these conditions are met, zero point errors should be

small (less than 0.10 magnitude), while random errors in the initial

magnitudes will be typical of iris photometry.

A third, rough calibration technique has been, rather unfortunately,

extensively used by some workers in recent archival studies. The "diameter

of the stars on the POSS" (Palomar Observatory Sky Survey) technique

involves measuring the diameter of the comparison star images on the POSS

print and then applying a general relationship between magnitude and


~












image diameter which is assumed to hold for all of the Sky Survey fields.

The relation most often quoted is that given by Liller and Liller (1975).

Differences in transparency, seeing and plate sensitivity cause individual

plates to deviate from this relation. This plate-to-plate variation was

estimated by Liller and Liller to be +0.4 magnitudes. If one examines

the results of the extensive study of the POSS by Dorschner and Gurtler

(1963) the situation does not appear so good. They list, for 33 blue

prints, the deviation of the individual calibration curves (Am). For

ten of the prints the magnitude of Am exceeds 0.40, with four of these

greater than 0.85. The total range of deviations from the mean is from

0.69 to -1.52 magnitudes. This gives a zero-point difference of 2.2

magnitudes between the two "extreme" prints. A great potential exists

for unacceptable zero-point errors when using this technique. Data

reduced employing sequences of this type have been used in this study

only if it was possible to determine that no significant zero-point error

existed.

Nuclear brightnesses of compact objects which exhibit a significant

non-stellar component are difficult to measure accurately photographically.

For iris photometry to yield valid, consistent results for these objects

exposure times must be adjusted (empirically) so as to record on the

plate only the stellar component. This becomes more and more difficult

to accomplish as the nebular component becomes comparable in surface

brightness to the nuclear component. Archival plates rarely meet this

criterion and magnitude estimates made from them of non-stellar objects

must be treated with caution. In the worst cases the nebular contribution












can cause the measured brightness to be too bright by a magnitude or

more. Even the use of a fixed iris diameter (in terms of the angular

size on the sky) does not eliminate the problem because of the non-

linear response of photographic emulsions. Each case of a non-stellar

object must be considered separately.

In summary, using photographic techniques it is possible to obtain

compact object brightnesses to +0.10 magnitude with time resolution of

a minute or so. The time resolution obtainable is a function of the

aperture and focal ratio of the instrument and the time necessary to

insert a new plate. Photography's main advantage is that it allows

relatively small instruments to record faint objects, the only limiting

factors being the night sky brightness and patience. In addition, one

can observe through skies of poorer photometric quality than one requires

for photoelectric measurements, allowing more continuous monitoring.

Photoelectric Photometry

All photoelectric (p.e.) work on compact objects has been carried

out in the last 20 years. Its advantages are high accuracy (0.01 to

0.05 magnitude), high time resolution (as short as I second for compact

sources) and automatic digitization and reduction of the resulting data.

Its major disadvantage is that for objects fainter than about B=15,

that is 95% of all known quasars, a telescope of 30 inches or more

aperture is needed. There are few telescopes this size which can have

large portions of their time devoted to a single observing project.

Consequently, most p.e. photometry programs for compact objects are of

short duration and/or the interval between observing sessions is very

long. In addition, highly photometric sky conditions are required.












It is with photoelectric photometry that the best hope for the

detection of very short timescale variability lies. More specifically,

it lies with the two-star photometer. When using a single-channel

photometer it is necessary to move back and forth between the object

under study and the comparison star. In addition, frequent sky measure-

ments must be made. The two-star design utilizes two separate channels,

one which constantly monitors the object while the other observes the

comparison star. This gives a 100% duty cycle and the results are not

affected by small changes in transparency. To date only a few attempts

have been made to examine an active compact object with this type of

instrument. If such an instrument could be devoted exclusively to the

study of short timescale variability, many of the questions about it

would be quickly answered.

Studies of the radial light distribution of non-stellar objects

are more easily accomplished photoelectrically. If the proper observ-

ations are made, precise corrections for the non-stellar contributions

can be determined, as we shall see in the case of NGC 4151.

Sources of Compact Object Variability Data

Location of Data

Aside from the data obtained from the Florida quasar monitoring

programs, all data used in this study were obtained from the general

astronomical literature. This includes all astronomical journals as

well as available circulars and observatory reports and publications.

One is fortunate in that these objects have been studied only since

about 1963. For the period from 1969 through the present the Astronomy











and Astrophysics Abstracts provide references to and descriptions of

articles published in the general astronomical literature. From this

it was possible to assemble all references or potential references to

optical measurements of compact extragalactic objects. A significant

number of these references were not listed in either the "Revised Catalog

of Quasi-stellar Objects" of Hewitt and Burbidge (1980) or the "Optical

Catalog of Radio Galaxies" of Burbidge (1979). For the period from

1963 through 1968, it was necessary to examine each journal individually.

Many observatory reports and publications from this period were not

available to the author, so it is possible that some data from these

sources were missed.

Once all of the potentially useful references were found and

sorted by journal it was then necessary to examine each article individu-

ally. If it contained no useful information it was discarded. Otherwise

the reference was assigned an identification number and notes were made

of which objects were reported on and the nature of the observations.

Objects were recorded initially by their co-ordinate designation to avoid

the confusion associated with multiple names for the same object arising

from various radio source catalogs. An identification number was

assigned to each object as well. All references were then photocopied.

This resulted in about 180 references containing potentially useful

data on 260 different objects. This information was then arranged so

that on a 4x6 card were the cross reference numbers to all papers present-

ing data on a given object. At this point it became possible to determine

rapidly if a given object had sufficient information to warrant further

study.












The data in these studies ranged in scope from a single (but

occasionally critical) point to hundreds of points spread over as much

as 80 years. The formats of data presentation varied greatly. In

terms of this study, the ideal format would have been a combination of

both tabular and graphical data. The graphs would be of sufficient

scale that significant, continuous (multi-point) events were recognizable.

The tables would present the magnitude, rms error and Julian Date (to

thousandths of a day) for each observation. In addition, if photoelectric

photometry was employed, an indication of the diaphragm size which was

used would be given.

Deviations from the above format were handled in various ways.

Times of observations were often given as decimal U.T. (Universal Time)

Date or U.T. Date and time. These were converted to Julian Date using

auxiliary computer programs after the data had been initially digitized.

Occasionally the date (J.D. or otherwise) was quoted only to the nearest

day. In this case a "probable time of exposure" was calculated on the

basis of the location of the observatory. The given date was then

corrected by this amount and, when digitized, these points were annotated

as to their nature.

When no individual errors were quoted for the measurements,the

"typical error" given by the author of the paper was used for each of

the datum points. If no typical error was given, errors were assigned

on the basis of observational and reduction techniques. Errors of 0.05,

0.15 and 0.20 were adopted as standards for photoelectric and iris photo-

metry and visual interpolation respectively.

Data reproduced only in graphical form presented other difficulties.

In a few cases it was possible to obtain the tabular data used to create











the original plots. When this was not possible one was faced with the

task of reading the data directly off the graph. Test measurements were

made using the University of Florida binocular microscope XY-measuring

engine on published light curves from the Florida program. An accuracy

of +0.01 magnitude and +2.0 days was attainable for these plots whose

scales were about three magnitudes per inch and 800 days per inch. Con-

sidering the time necessary to determine the co-ordinates of a given

point and the relative inaccuracy of the date determination, only a few

key points and events were digitized in this way.

In order to perform various transformations and analyses on the

light curves it was necessary to have the data in machine readable form.

All data from the Florida monitoring program were available in this

form; however, all other data had to be hand entered. To facilitate

this process it was decided to encode only data for objects which, upon

initial examination of all pertinent references, showed definite or

possible discrete events. Nonetheless, over 10,000 individual points

(representing 40,000 pieces of data) were entered.















CHAPTER III
THE VARIABILITY PROBLEM

In present usage a celestial object is termed optically variable

if one is able to detect a statistically significant change in its

apparent magnitude. This standard definition and conception of vari-

ability can severely mask and distort certain physical characteristics

of an object or group of objects, and it disguises some observational

selection effects.

The typical 2500A monochromatic absolute magnitudes M for compact

objects run from about -21 to -30. Consider an event, an added source

of energy, to have (unto itself) a peak absolute magnitude of -24.

Suppose now that identical events occur within objects whose quiescent

or background absolute magnitudes range from -21 to -30. The resulting

maximum magnitude change AM for each object is given in Table 1. The

TABLE 1

EVENT MAGNITUDE CHANGES



M AM


-21 3.07
-22 2.16
-23 1.36
-24 0.75
-25 0.36
-26 0.16
-27 0.07
-28 0.03
-29 0.01
-30 0.00


21












observed results of the event in the various objects are quite differ-

ent. In the -21 object a major outburst would be reported, a violent

event, and in the -25 object a minor brightening would occur. Photo-

graphic and photoelectric techniques could not reliably detect the

event for objects brighter than -26 and -28, respectively.

The question then becomes, are the objects "variable?" In terms

of an absolute change in luminosity, yes, but in terms of an observable

change in magnitude, employing present techniques, no. There are physically

meaningful interpretations of each of these results.

By examination of Table 1 it can be seen that, if event energies

were not proportional to the luminosity of the object involved, there

would be fewer events detectable in the brighter objects and the morph-

ology of those events would be different as well. In addition, because

intrinsically fainter objects are more difficult to detect and study,

one might conclude, perhaps erroneously, that high-redshift objects

were less "variable" than their low-redshift counterparts. Analyses of

the variability phenomenon which deal strictly with changes in apparent

magnitude necessarily have these problems inherent in them.

For extragalactic objects variability data can be examined in

various forms. What assumptions and additional object parameters are

required and what can be learned physically from each of these forms

are examined below.

Apparent Magnitude

An apparent magnitude light curve derived directly from observ-

ations involves no assumptions other than reliable technique and equip-

ment. It is sometimes done for multiple received wavelengths. The












term "received wavelength" is important here because of the large range

of Doppler shifts occurring in these types of objects.

For an individual object one can determine whether or not the

object varies in magnitude and the morphology of those variations

(shapes, periodicities, etc.). Colors (spectral indices), variation

in color, and the timescale of variations can also be determined. The

only absolute physical parameter which can be obtained is an estimate

of the upper limit to the size of the emitting region during an out-

burst. This is derived using light travel time arguments. For a

group of objects qualitative similarities in light curve morphology and

color can be obtained. The problem with intercomparing objects in this

situation is that the information obtained in a given bandpass (say

the B at 4400A) represents different emitted frequencies in the rest

frames of the objects. For objects having z=0.05 and z=2.50, the 4400A

received radiation corresponds to 4190C and 1257A emitted radiation,

respectively. In addition, the rest time frames are different. A 10-

day observed event actually spanned 10/(l+z) days at the objects, that

is 9.5 and 2.9 days, respectively.

Absolute Magnitude or Emitted Flux

Knowledge of the absolute magnitude or the emitted flux implies a

knowledge of the distance and, for these objects, that means a redshift

determination and the use of Hubble's law.

Often a value of the absolute B-magnitude is quoted, calculated

from the apparent B-magnitude, the distance and the inverse square law.

This does not represent, however, the flux emitted by the object at












4400R, unless the optical spectral distribution is flat or z=0. What

properly should be derived is the emitted flux at whatever wavelength

4400A represents in the rest frame of the object. When one examines a

group of objects with differing redshifts, the rest frame flux curves

derived will be for different emitted wavelengths.

Rest Frame-Standard Wavelength Flux

If the redshift and the spectral energy distribution of an object

are available, one can in principle determine the flux at any wavelength

within the known range from the flux at one wavelength. It is also

possible to extrapolate the spectral distribution to regions not directly

measured, but such a process requires great care. Thus for two objects

with observed B light curves, differing redshifts and known spectral

indices, it is possible to transform their apparent magnitudes at 4400X

into emitted fluxes at, say 2500A. This then allows a direct comparison

of the physical behavior of the objects at that wavelength. One can

create a "standard rest wavelength" to which all light curves will be

transformed. In the process of this standardization the time axes of

these light curves will be compressed by a factor of 1+z. If one wishes

to determine the total amount of energy being emitted in all directions

by the object it is necessary to assume some angular dependence of the

emitted radiation. The usual approach is to assume isotropic emission.

If, in fact, the radiation is generated by a beaming process of some

sort, the above assumption could lead to an overestimation of the lumin-

osity of the object. Ramifications of various assumptions will be

discussed in later sections.















CHAPTER IV
DATA HOMOGENIZING AND TRANSFORMATION

Photometric Systems

Nearly all of the raw data which was employed in this study was

in one of the following forms.

(a) Johnson U,B and/or V-magnitudes
(b) International Photographic (P) magnitudes
(c) Millijansky flux at (or about) 4400D

Over 75% of the data was in the form of B-magnitudes, with only a small

quantity as U-magnitudes and millijansky fluxes. The remaining data

were about equally divided between V and P-magnitudes.

For a given object, in order to be able to combine data from all

available sources it was necessary to put the data on the same photo-

metric system. Because a majority of the data was already in the form

of B-magnitudes) the Johnson B system was chosen as the standard. If

U or V-magnitudes and U-B and B-V color indices were given for each

point, the determination of the B-magnitudes was straightforward. If

individual color indices were not given, then color indices obtained

for that object from other observations were used. Use of this single

value for the color index assumes that no temporal color changes occur.

Color changes of a few tenths of a magnitude have been seen in some

objects. Fortuitously, few conversions utilizing "typical" color

indices were necessary. For all objects where U and V-magnitudes were

quoted, color indices were available in one of the above forms.

Millijansky 4400A received fluxes were converted to B-magnitudes

using the standard flux-magnitude conversion formulae of Johnson (1966).

25












Conversion to magnitude form was necessary in order to perform other

corrections discussed in the next section.

The P-magnitude system has a bandpass which encompasses the

entire Johnson B and U ranges. Thus the value B-P, needed to convert

to the B system, is a function of the color of the object. This was

demonstrated by Arp (1961) for main sequence stars and by Lu (1972)

for QSO's. The correction for QSO's averages about 0.35 magnitude but

ranges between 0.20 to 0.50 for objects whose U-B values range from

-1.25 to 0.00. If one or more nearly simultaneous observations in the B

and P were available, the average value of B-P obtained from them was

used to convert the P-magnitudes to the B system. Otherwise empirical

corrections as listed in Table 2 were made.

TABLE 2

PG MAGNITUDE CORRECTIONS



U-B B-P


0.00 0.20
0.00 to -0.50 0.30
-0.50 to -1.00 0.40
-1.00 0.50
Unknown 0.35


Once all the observations of a single object were on the B system

a composite data set and light curve could be generated and the search

for discrete events undertaken. However, before data examination and

manipulation were actually begun a final intercomparison of the data

from different sources was made. After digitization the data were

sorted by Julian Date. If two or more points on the same date from












different sources were found, these were carefully examined for indications

of zero point differences, even if the data were taken in the "same"

photometric system. Once this was done the data were pronounced fit for

analysis and transformation.

Standard Rest Frame System

As mentioned earlier, in order to intercompare these objects it is

necessary to convert the available data back to the rest frame of the

objects. In addition, one wishes to choose a single standard rest wave-

length. The approach taken is essentially that of Schmidt (1968) and

Richstone and Schmidt (1980), and is discussed below.

The standard rest wavelength was chosen as 2500 because of the

relative lack of strong emission lines in this region of the typical

QSO spectrum, and because this wavelength range is redshifted into one

of the Johnson bandpasses for a significant percentage of the known

range of z's for these objects. To obtain the monochromatic luminosity

at 2500A in the rest frame of an object, the following procedure is

employed. The case examined below is for conversion from B-magnitudes,

but the method is basically the same for U and V-magnitudes.

First, it is necessary to correct for galactic extinction in one

of two ways. Multicolor photometry of standard stars in the same

direction as the object of interest can be obtained and the absorption

coefficient in that direction can be specifically determined. Instead,

if one is working away from the galactic plane, one can employ a standard

extinction law where the absorption coefficient is a function of the

object's galactic latitude. This is satisfactory for most extragalactic

objects and the extinction law given by Tapia et al. (1976) has been used

in this study.












The spectral index a (for power law spectra f(v)=v-t) must be

calculated in most cases from the U, B and V-magnitudes for the object.

Since a defines the slope of the non-thermal continuum in the optical

region, it is not dependent on the nature of the emission lines. However,

the broadband UBV magnitudes contain not only continuum radiation but

any line radiation which happens to lie in their respective bandpasses

as well. Thus, before these magnitudes can be used to determine a they

must first be corrected for the presence of emission line radiation.

Which lines contribute to the various bandpasses depends upon the redshift

of the object. The amount of the contribution depends on the equivalent

width of the line. Thus, in the calculations of a done here, if the

equivalent width of a line for a given object is known (and is signifi-

cant) it is corrected for.

Whenever possible the values of U-B, B-V and z from Hewitt and

Burbidge are used. For some objects not given in this catalog,parameter

values are obtained elsewhere. A summary of the adopted parameters is

given in Table 3 in the next chapter.

At this point the B bandpass received flux is determined from

the formula of Johnson (1966) as

f44 = 4.44 x 10 (15.0 Bc)/2.5

where Bc is the B-magnitude corrected for galactic extinction and the

presence of emission lines. Now, using the value of a previously cal-

culated, the 2500R flux received can be expressed as

f25 = [2500(l+z)/4409a f44









29


where f25 has units of ergs/sec cm2 Hz. Assuming isotropic emission of

radiation and an empty (qo=0) Friedman universe, the 2500 monochromatic

luminosity is given by

L25 = 4TTHc(z+z2)]2 f25

where L25 has units of ergs/sec Hz.

For this study the values chosen for Hubble's constant Ho and

the cosmological constant qo were 75 km/sec Mpc (lying in the middle of

the currently accepted range of values) and 0 respectively.















CHAPTER V
EVENT DETERMINATION AND ANALYSIS

Event Definition

It is the objective of this study to examine the discrete events

in the light curves of compact extragalactic objects. Thus it is

necessary to define what such an event is and how they are picked out

of the actual light curves.

In the broadest sense an event occurs whenever there is an observa-

tionally significant difference between two temporally adjacent flux

measurements. Identification of such an "event" involves no subjectivity

(other than the choice of the significance level), but also does not

yield more than the rate of change of flux with time, an important

quantity but far from a complete description of the behavior of the

object.

The problem can be approached in a manner similar to the recog-

nition and analysis of emission features in the spectra of astronomical

objects. In most astronomical spectra any emission lines are super-

imposed upon a continuum. To be recognized as discrete emission features

they must be of sufficient intensity to cause a discontinuity in the

continuum. In order to determine the actual amount of flux present in

the emission line, the continuum level must be subtracted. In addition,

these spectral lines may be rendered more difficult to detect if they

are broadened by any of a number of mechanisms. Also, the continuum may

not be smooth and monotonic (consider, for example, the Balmer discontin-

uity).












Looking for and analyzing events in compact object light curves

follows essentially the same path. Two useful features are often found

in spectral analysis and rarely in light curve analysis. First, for light

curves, the data sampling rate is generally non-uniform, often showing

large gaps. Secondly, one does not have an a priori knowledge of where

the features are expected to occur.

The flux variations that occur in these objects are of two major

types, which at some point blend into one another. Most objects show

long term trends (duration >2 years), some examples of which are shown

in Figure 1. These, and the other light curves discussed in this section,

are reproduced from Pollock et al. (1979). Then there are discrete

events such as those seen in the light curves of 0235+164, 2345-16 and

3C 446 shown in Figure 2. In these cases the object goes through a

relatively rapid, monotonic increase in brightness followed by a return

to the base level that existed before the event.

Figure 3 shows the recent behavior of OJ 287, 3C 345 and 0420-01.

For each of these there appear events which have the characteristics of

both types of variations discussed above. Are these events or continuum

level changes, and at this point are the two in any way distinguishable?

In this study they have been analyzed in the same manner as the shorter-

term outbursts. Similarities and differences in the calculated event

parameters can therefore be examined.

Identification of Events

Using our present definition, the smallest number of points required

to define an event (to be able to specify its duration, peak and total

energy) is three. Obviously this would be a most unsatisfactory situation.





























13

14 3C 120
B r





17 1968 1969 970 1971 1972 9 1
1I --- |I--J_ 1971 1972 -1973 I9714 1975 1976 1977 1978 19)



12

13 3C 371
P I
14

15 -

16 I I I
968 1969 1970 1971 1972 1973 1974 1975 1976 977 1978 197

Figure 1. Sample light curves containing long term trends.






















AO 0235+164

I )
















I
-

i
1977 1978 1979





PKS 2345-16



I .

S-. 1,*.. 7 "



1968 1969 1970 1971 1972 1973 !97 175 1976 1977 1978 1979


15

16 3C 4q6 .
I : r .
17 -

18 ":

I I II
1968 1969 1970 1971 1972 1973 197 94 1975 1976 1977 1978 1979

Figure 2. Sample light curves containing discrete events.


19

IS

S16

17

18

19



14

15

P 16

17

18

19




















II t k t I

13 OJ 287

I .













15 3C 3115
17I









16
1StI I -[












I .
18 .
I I.-- I I I I I I



16 --- 1 -- I --- --- --- --I--- I --- ] -- i --- ] I --


17 PKS 0420-0 1 ,' "






20 1968 1969 1970 1971 1972 1973 197 1975 1976 1977 1978 1979

Figure 3. Sample light curve containing intermediate events.











It would be of no use unless the magnitude change was quite large, so

as to cast no doubt on the single peak point's validity, and of very

short duration so that a reasonably high average sampling rate still

existed. Here the average sampling rate is simply the total number of

event measurements divided by the duration of the event in days. No

cases of three-point events were examined in this study. The single

four-point and two five-point events in this study had durations of 1.6,

2.3 and 11 days and magnitude changes of 2.0, 1.8 and 1.6, respectively.

The actual procedure used for picking out events was as follows.

Magnitude data from all available sources were digitized and sorted by

Julian Date as described earlier. A data file and table of this informa-

tion was then produced. A line printer plot of magnitude versus Julian

Date was created with a resolution of <0.04 magnitudes/line and <7 days/

line. Visual examination of the plot and table were then used to identify

events and to establish the base level of the object at the time of each

event. A subset of the original object data set was created containing

only those points within the defined event. The lower of the two event

endpoints was defined as the base level for that event. The event

magnitudes and the corresponding object parameters (redshift, spectral

index, right ascension, declination and emission line strengths) were

supplied to a program which produced a data set having for each point

(a) Julian Date
(b) Object Event Date
(c) Log of received flux at 2500A
(d) Log of the luminosity at 2500D
(e) Logs of the errors in the 2500C luminosity












The Object Event Date gives, for each point, the number of rest

frame days which have elapsed since the start of the event. The errors

in the 2500S luminosity are the uncertainties resulting from the observa-

tional errors for a given point. The upward and downward errors are not

equal because of the nature of the magnitude system and will be used in

later analyses to determine the significance of luminosity changes.

Event Objects

Twenty-four compact object composite light curves contained useful

events. Table 3 lists these objects in right ascension order and contains

various important object parameters. Column 1 gives the co-ordinate

designation, 2 the most common name (if different from 1). Objects from

this point on will be referred to by their common names. Column 3 gives

the adopted U-B and B-V colors, 4 the calculated spectral index, 5 the

correction for galactic absorption, 6 the redshift and 7 the classifica-

tion of the object. Emission line corrections were necessary for only

two objects, 3C 345 and 3C 446, and were 0.14 and 0.04 magnitude respect-

ively. That there were so few corrections of this type is not surprising.

A majority of these objects are OW's, which, as a class, are usually

weak-line or BL Lacertae objects.

The general characteristics of all but six of these objects have

been discussed by either Pollock et al. (1979) or Pica et al. (1980). A

short characterization of the remaining six objects is given below

before beginning the general discussion of the events themselves.

0039+40.--Zwicky et al. (1969) note that this Seyfert galaxy shows

a variable nucleus superimposed on a faint halo. They comment that the

absolute energy distribution bears a "striking resemblance" to 3C 120,










TABLE 3

EVENT OBJECT PARAMETERS


Co-ordinate Common Adopted Spec. Gal.
Designation Name U-B B-V Index Abs. z T

0039+40 -- -- -- 0.84a 0.42 0.103 S
0219+42 3C 66A -0.58 0.33 1.30 0.64 0.444 B
0235+164 -- 0.14 0.96 3.89 0.12 0.852 B
0420-01 -- -- 0.70b 0.20 0.915 Q
0440-00 NRAO 190 -1.05 0.37 1.00 0.28 0.850 Q
0521-36 -- -0.30 0.67 2.27 0.21 0.055 N
0735+17 -- -0.58 0.47 0.97 0.58 0.424 B
0851+20 OJ 287 -0.64 0.39 0.98 0.17 0.306 B
0906+01 -- -- 0.70b 0.24 1.018 0
1156+295 -- -0.56 0.44 1.38 0.00 0.728 B
1209+39 NGC 4151 -- -- 0.05c 0.00 0.003 S
1253-05 3C 279 -0.56 0.26 1.55 0.00 0.538 Q
1308+326 -- -0.59 0.40 1.23 0.00 0.996 B
1510-08 -- -0.74 0.17 0.90 0.11 0.361 0
1545+21 3C 323.1 -0.85 0.11 0.26 0.01 0.264 N
1633+38 4C 38.41 -- -- 0.70b 0.09 1.814 0
1638+39 NRAO 512 -- -- 0.70b 0.10 1.6d Q
1641+39 3C 345 -0.50 0.29 1.11 0.10 0.595 Q
1730-13 NRAO 530 -- -- 0.70b 1.10 0.9e Q
1807+69 3C 371 -0.25 0.83 2.68 0.27 0.050 N
2200+42 BL LAC -0.10 0.97 1.90 1.15 0.069 B
2223-05 3C 446 -0.90 0.44 1.77 0.01 1.404 0
2251+15 3C 454.3 -0.66 0.47 1.49 0.14 0.859 0
2345-16 -- -- -- 0.70b 0.00 0.600 Q

Notes to Table 3:
a. From published spectrum of Zwicky et al. (1969)
b. Adopted "typical" value
c. Deconvolved nuclear spectrum of Penston et al. (1971)
d. Private communication, B. Lynds
e. Tentative, private communication, A. Marscher
f. Final column abbreviations:
N, N-galaxy
B, BL Lacertae Object
S, Seyfert galaxy
0, QSO












with the absolute flux at a rest wavelength of 5000A differing by only

about 20%. Unlike 3C 120, 0039+40 is radio quiet, with S(408 MHz)

<0.012 flux units. Optical monitoring at Assiago has shown 0039+40

to vary erratically between B=16-18.

0521-36.--Eggen (1970) summarizes the data available for 0521-36.

It is an N-galaxy and a strong radio source (19.0 flux units at 1410

MHz). The object's color becomes redder as it decreases in brightness,

similarly to 3C 371 and 3C 390.3. Shen et al. (1972) found, from the

Harvard Archives, short-term variability superimposed on a long-term

component, the later variation spanning on the order of a decade.

1156+295.--This object was not widely known until the detection

of a major outburst in 1981. It is a BL Lac object whose historical light

curve (Pollock, 1982) showed the object to be very quiescent until

recently.

NGC 4151.--This prototype Seyfert galaxy is extremely well known.

Cannon et al. (1971) summarize references to its optical variability.

Photometery of the object is complicated by a bright, extended galaxy

component, which, through a 10" aperture, is brighter than the central

point source from the V to K bands (Penston et al., 1971). They have

used a series of photoelectric observations through various diameter

apertures to deconvolve the galaxy and nuclear component spectra. The

spectral index used in this work for NGC 4151 was determined from their

nuclear component spectrum.

In addition Penston et al. (1971) determined the galaxy flux

contribution through a given aperture to be












f(D") = f(10") (D/10)a

where they found a = 0.6 + 0.05. For the B-bandpass we thus obtain

f(10") = 1.45 x 10-2 Janskys

which in combination with their earlier expression gives

fB(D") = 1.45 x 10-2 D0.6(.1)0.6

fg(D") = 3.64 D0.6 millijanskys

Since, in this work, we are interested in only the compact central

source, the following special reduction procedures were employed for

NGC 4151. Photoelectric observations, which reported the aperture size

which was used, were first converted to B-magnitudes, if necessary.

These were then converted to received fluxes, corrected for the galaxy

contribution and then converted back to B-magnitudes to be treated in

the standard fashion.

1545+21.--This source, also known as 3C 323.1, was identified by

Wyndham (1966) as a slightly diffuse object with a very blue nucleus.

Schmidt (1968) determined its redshift and Lu (1972) carried out the

variability observations used in this work. The Harvard historical

light curve of Angione (1973) showed variations between B=15 and 17.

4C 38.41.--Other than its identification as an optical counterpart

of a radio source and its redshift determination little has been published

concerning this quasar.

Event Parameter Calculations

Various parameters describing each event have been determined and

are summarized in Tables 4 and 5. Given below are descriptions of the

contents and format of Tables 4 and 5 and, when necessary, an explanation

of the procedure used to obtain a given parameter.










TABLE 4

EVENT PARAMETERS I


Julian Date B-Magnitude=========


Event

0039+40 A

3C 66A A

0235+164 A
B
C
D

0420-01 A
B
C

NRAO 190 A
B

0521-36 A
B

0735+17 A
B
C

OJ 287 A
B
C
D

0906+01 A
B
C

1156+295 A

NGC 4151 A

3C 279 A
B
C

1308+326 A
B
C
D


Julian Date
Start End

41513.465-41927.597

42667.847-42834.440

42688.7 -42780.7
43751.853-44136.843
44136.843-44305.547
44487.779-44629.580

42089.544-42700.932
43183.612-43892.603
43892.603-44487.801

43054.894-43137.726
43462.817-43544.557

39817.300-40560.300
43097.0 -43221.0

42047.323-42148.550
42717.873-42928.584
42928.584-43489.702

36526.598-36667.301
41653.639-41822.708
42341.911-42488.250
40625.463-42149.566

40862.886-40899.885
40943.757-41009.590
41011.667-41041.736

44640.708-44729.708

39505.0 -40019.488

28279.0 -28330.0
28601.0 -28722.0
27814.0 -29028.0

43169.877-43310.750
43543.750-43728.578
44395.648-44455.578
44673.806-44754.681


Duration
days

375.460

115.369

49.7
175.497
91.093
76.566

319.263
370.230
310.808

44.774
44.184

704.265
117.536

71.090
147.971
394.044

107.736
129.456
112.051
1167.001

18.334
32.623
14.900

51.505

512.8

33.2
78.7
789.3

70.578
92.599
30.025
40.519


Se
pts/d

0.04

0.10

0.26
0.19
0.18
0.22

0.05
0.06
0.05

0.20
0.16

0.02
0.09

0.25
0.51
0.22

0.19
1.17
0.87
0.50

0.44
0.89
0.81

0.60

0.10

0.45
0.17
0.09

0.26
0.63
0.83
0.30


B-Magnitude
Peak Base

16.90-17.80

15.10-16.00

15.25-18.04
15.44-19.39
17.16-19.39
17.04-19.54

17.24-19.19
16.86-19.03
16.38-18.94

17.13-18.93
17.48-19.42

14.60-15.95
15.22-16.22

15.07-16.56
14.37-16.08
13.90-16.11

13.12-15.68
13.25-14.65
14.66-15.95
12.36-16.04

16.24-17.38
16.44-17.65
16.77-17.58

13.15-16.78

11.73-14.32

12.03-13.98
11.27-14.90
11.27-17.32

15.21-16.95
15.20-17.54
14.95-17.68
15.79-18.82










TABLE 4 (continued)


Julian Date Duration Se B-Magnitude
Event Start End days pts/d Peak Base


1510-08 A

1545+21 A

4C 38.41 A

NRAO 512 A
B
C
D
E
F
G
H

3C 345 A
B
C
D
E
F
G
H


32671.0 -32800.0

40596.920-40840.580

41087.4 -41119.5

40703.4 -40709.5
40709.5 -40739.4
40739.4 -40743.621
40743.621-40779.769
41134.742-41246.518
41397.890-41458.875
41458.875-41506.709
41134.742-41866.726

39017.7 -39063.6
39358.7 -39387.7
39534.69 -39548.73
39569.630-39616.4
39675.510-39709.490
40298.8 -40451.8
41120.4 -41175.25
39290.7 -39943.52


94.8 0.16 11.80-15.00

192.769 0.06 16.06-16.58

11.0 0.45 15.85-17.40


2.3
11.5
1.6
13.903
42.991
23.456
18.398
281.532

28.8
18.8
8.803
29.3
21.304
95.9
34.4
409.3


2.13
0.78
2.46
1.58
0.56
0.17
0.38
0.21

0.97
1.01
1.14
0.48
0.76
0.96
1.13
0.62


17.00-18.80
17.60-18.80
16.67-18.70
17.76-18.81
17.30-19.47
17.39-18.50
17.52-18.50
17.30-19.61

16.25-17.21
16.27-17.18
15.41-16.26
15.77-16.40
15.49-16.25
15.53-17.50
14.99-16.20
15.41-17.45


NRAO 530 A 43280.838-43420.565
B 44764.794-44792.709

3C 371 A 40827.380-40879.260
B 41471.360-41518.470


73.541 0.20 15.76-18.52
23.593 0.68 16.94-18.70

44.824 1.63 14.37-15.53
49.363 0.59 14.54-15.62


BL LAC A
B
C

3C 446 A
B
C
D
E


40384.570-40747.348
41891.280-41899.510
32043.0 -32073.7

38944.9 -40127.38
40500.34 -40918.30
42301.656-42687.740
43018.733-44550.545
44572.535-44792.852


3C 454.3 A 39856.3 -40150.8
B 44057.796-44246.540

2345-16 A 40504.691-41123.864


83.983
7.699
28.718

491.9
173.86
160.606
637.205
91.646:


0.83
7.27
0.45

0.30
0.10
0.17
0.11
0.09


158.4 0.25
75.214: 0.13

386.983 0.12


14.29-16.10
15.00-16.14
14.51-16.32

15.67-19.00
16.15-18.94
15.44-18.80
15.35-18.57
16.33-18.35

16.08-17.21
16.22-17.62

15.87-17.96


----------------------------------------- ---------











TABLE 5

EVENT PARAMETERS II


Event


0039+40 A

3C 66A A

0235+164 A
B
C
D

0420-01 A
B
C

NRAO 190 A
B

0521-36 A
B

0735+17 A
B
C

OJ 287 A
B
C
D

0906+01 A
B
C

1156+295 A

NGC 4151 A

3C 279 A
B
C

1308+326 A
B
C
D


Lp
- --------

1.55 [293

2.18 [313

7.48 [313
6.30 [313
1.26 [313
1.44 [313

1.28 [313
1.81 [313
2.82 [313

1.28 [313
9.29 [303

1.40 [293
7.91 [283

2.04 [313
3.88 [313
5.98 E313

3.95 [313
3.51 [313
8.49 [303
9.78 [313

4.32 [313
3.60 [313
2.65 [313

2.58 E323

1.50 [283

3.20 [323
6.44 [323
6.44 [323

1.45 [323
9.25 E313
1.15 [323
5.31 [313


e


LB 1p
ergs/sec Hz -- -

6.78 [283 8.75 [283

9.53 [303 1.23 [313

1.65 [303 7.32 E313
1.65 [303 6.13 E313
1.65 [303 1.10 [313
1.45 [303 1.30 [313

2.12 [303 1.06 [313
2.46 [303 1.57 E313
2.67 E303 2.55 [313

2.44 [303 1.04 [313
1.56 [303 7.73 E303

4.11 [283 9.88 E283
3.15 [283 4.76 E283

5.16 [30) 1.52 E313
8.04 [303 3.08 [313
7.82 [303 5.20 [313

3.74 E303 3.58 [313
9.66 E303 2.54 [313
2.91 [303 5.57 [303
2.68 [303 7.69 E313

1.51 [313 2.81 E313
1.18 [313 2.42 [313
1.20 [313 1.45 E313

9.12 E303 2.49 [323

1.38 [273 1.36 [283

5.31 [313 2.67 E323
2.28 1313 6.21 [323
2.45 [303 6.42 [323

1.04 [313 1.34 [323
1.06 [313 8.19 [313
9.31 [303 1.06 E323
3.26 [303 4.98 [313


==================================~=====


Ee
ergs/Hz

9.71 [353

3.31 [373

1.51 [383
1.04 [383
1.61 [373
2.92 E373

1.03 [383
2.09 [383
8.47 [373

1.76 [373
1.00 [373

1.99 [363
2.67 [353

4.59 [373
1.52 [383
4.50 E383

1.37 E383
1.12 [383
2.06 E373
2.42 [393

2.81 [373
3.00 [373
7.44 [363

3.15 [383

2.53 [353

2.77 [383
8.84 [383
3.62 [393

3.70 [383
2.45 [383
6.94 E373
6.23 C373


Re


0.44

0.35

21.26
2.92
1.24
3.03

1.76
2.67
1.18

1.86
1.69

0.80
0.83

1.45
1.48
1.69

3.94
1.04
0.73
8.96

1.17
0.90
0.48

7.77

4.13

1.71
5.72
21.68

5.73
2.89
2.87
5.47










TABLE 5 (continued)


Event Lp LB Ip
--- -- ergs/sec Hz -- -


E
ergs/Hz


R
e


1510-08 A 1.89 C323 9.91 [303 1.79 [323 4.84 1383 5.96

1545+21 A 2.08 E303 1.26 E303 8.13 C293 6.00 [363 0.29

4C 38.41 A 2.47 1322 5.93 [313 1.88 C323 8.32 1373 1.47

NRAO 512 A 6.14 [313 1.17 1313 4.97 E313 4.18 [363 1.78
B 3.53 [313 1.17 1313 2.36 [313 1.21 [373 1.05
C 8.32 E313 1.28 E313 7.04 [313 4.44 [363 2.73
D 3.05 [313 1.16 [313 1.89 [313 8.78 C363 0.63
E 4.66 E313 6.31 C303 4.02 E313 5.38 [373 2.30
F 4.28 E313 1.54 E313 2.74 [313 2.74 [373 0.88
G 3.80 [313 1.54 E313 2.26 [313 1.14 [373 0.47
H 4.66 [313 5.55 [303 4.10 E313 3.65 [383 2.53

3C 345 A 8.67 [303 3.58 [303 5.09 E303 4.17 [363 0.47
B 8.51 [303 3.68 E303 4.83 E303 2.33 [363 0.37
C 1.88 [313 8.59 E303 1.02 [313 3.77 [363 0.58
D 1.35 [313 7.55 [303 5.94 E303 7.48 [363 0.39
E 1.75 E31) 8.67 [303 8.79 [303 8.50 [363 0.53
F 1.68 [313 2.74 E303 1.41 [313 2.62 E373 1.15
G 2.77 [313 9.08 E303 1.86 [313 1.74 [373 0.64
H 1.88 [313 2.87 C303 1.59 E313 1.76 E383 1.91

NRAO 530 A 1.10 [323 8.63 E303 1.01 [323 8.26 E373 1.51
B 3.70 [313 7.31 E303 2.97 E313 2.15 [373 1.44

3C 371 A 1.26 E293 4.32 [283 8.26 [283 1.64 E353 1.06
B 1.08 [293 3.98 [283 6.78 E283 9.43 [343 0.51

BL LAC A 8.65 [293 1.42 [293 7.23 E293 2.10 E363 2.04
B 4.50 [293 1.44 [293 3.06 [293 1.01 [353 1.05
C 7.06 [293 1.33 [293 5.73 [293 9.23 C353 2.79

3C 446 A 1.83 [323 8.51 [303 1.74 323 1.72 E393 4.75
B 1.18 [323 9.00 [303 1.08 [323 6.23 [383 4.61
C 2.26 [323 1.02 [313 2.16 [323 4.03 [383 2.84
D 2.46 E323 1.26 [313 2.33 E323 4.38 [393 6.29
E 9.95 C313 1.55 [313 8.40 [313 2.57 [383 2.09

3C 454.3 A 3.11 E313 1.06 [313 2.05 C313 7.97 E373 0.55
B 2.73 [313 7.53 C303 1.98 E313 3.80 [371 0.78

2345-16 A 1.35 [313 1.97 C303 1.15 [313 6.82 E373 1.04


-=======~r======












Table 4--Description

Column 1 gives the event designation by using the object's common

name followed by a capital letter suffix.

Column 2 gives the starting and ending Julian Dates.

Column 3 gives the rest frame event duration, Te, where.

T = [JD(final)-JD(initial)/(l+z)
e
Column 4 lists the data sampling rate, Se, where

S = (number of event points)/T-

It should be noted here that the actual temporal distribution of measure-

ments is often highly non-uniform.

Column 5 gives the peak and base B-magnitudes.

Table 5--Description

Column 1 gives the event designation by using the object's common

name followed by a capital letter suffix.

Column 2 gives the peak total 2500D luminosity, Lp.

Column 3 gives the base event 25006 luminosity, LB.

Column 4 gives the peak event 2500S luminosity, Ip, where

Lp = LB + p

Column 5 lists the total event 2500X energy, E where Ee is the

total energy emitted during the event (corrected for the base level) in

a one Hertz bandwidth at 2500R. This was determined from a simple

point-by-point integration of the event luminosity at point a, la, where

for an event having m points

Ea,a+1 = (la + la+l)ATa,a+1/2

e a=l a,a+l


Column 6 gives the ratio of the event energy to the base energy,


R where
e











Re = E /LB Te


Extreme Luminosity Changes

The change in luminosity, AL, was calculated for all possible

point pairs. This value was taken to be statistically significant if

ALab/eab was >2, where eab is the sum of the individual root mean square

errors associated with La and Lb, ea and eb respectively. As discussed

earlier, for an individual point the positive and the negative errors

are not the same. The luminosity at point a is


L ea+ where lel > I eaj
-ea-

Thus, for an upward luminosity change from points a to b, the signifi-

cance condition becomes

(Lb La)/(ea+ + eb-) >2

whereas for a downward change from points c to d we have

(Lc Ld)/(ec_ + ed+) >2

For all statistically significant luminosity changes the following

quantities were examined, AT, the rest time interval for the change to

occur; AL/AT, the rate of change of the luminosity; and, most importantly,

AT L/AL, whose application is discussed in the section on emitting

region sizes.

Normalized Event Luminosity Curves

In order to examine the general morphology of individual events

and groups of events, plots of normalized event luminosity as a function

of object event date were created and are shown in Appendix C. The normal-

ized luminosity at a point a is given by

1' = /I
a a p












The x-axis is the Object Event Date, with divisions being 0.1 of the

indicated axis extent. The first and last event points are then at

t=O and t=Te. Temporally adjacent points are connected by a dotted line.

This aids in following the point-by-point evolution of the event and

also better defines the general event shape. The figures in Appendix C

show the normalized luminosity curves for the events studied in this

work, ordered by right ascension.

Plotting of the rms error bars associated with each point would

have resulted in cluttered and confused graphs. The typical size of

the error involved in the original luminosity measurement is about 10%

(although this can be as small as 2% for photoelectric values or as large

as 20% for eye estimates). This percentage, however, is of the total

luminosity rather than the base level corrected luminosity.

The "normalized" error for a given point a (at the 10% error

level) is-

ea = 0.10 La/1p

= 0.10 (1a + L )/lp

= 0.10 E[1 + (L /lp)]

For normalized luminosity values of 0.4 and 0.8, respectively, this

becomes

e' = 0.04 + 0.10 (Lp/lp) 1' = 0.4

e' = 0.08 + 0.10 (Lg/11) 1' = 0.8

It is evident that, as the ratio of base event luminosity to peak event

luminosity becomes small, the value of e' approaches 10% of 1'. Values

of this ratio range, in the events studied, were between 0.01 and 1.












Event Morphology

The shape of an event, its morphology, describes the way that

the energy release took place. This can, hopefully, give some insight

into the energy generation mechanism and/or the surrounding macroscopic

structure. Before dealing with the individual events, a brief discussion

is given below of the so-called supernova-type light curve to illustrate

why light curve shapes must be carefully analyzed.

The Supernova-Type Light Curve

Some models for the energy generation mechanisms for compact

objects employ the occurrence of one or more supernovae. One of the argu-

ments against this proposition is that few outburst light curves resemble

those of classical supernovae. The long "tail" of these events is not

seen.

Figure 4 shows a schematic representation of a type I supernova

light curve (Zwicky, 1964) given in normalized luminosity. The points

span the period of time during which the luminosity is between 16% and

100%. Even without any background contribution to mask the event, it

appears quite symmetric. If one were to add observational scatter,

less regular data spacing and a background of only 15% of the peak

luminosity, the "tail" would be undetectable and the remaining, less

pronounced asymmetry would be masked as well. This fabricated event

would remain basically smooth in its rise and decline, however, which

is not usually observed in compact object events. It should be remembered

though that galactic supernovae are relatively isolated, while the physical

conditions in the nuclei of Seyfert galaxies and QSO's are not likely to

be so quiescent and uniform.








48


.------1--


C)



4-,
C3

>










C)
rC
0








C)
c)
3







C)






-j



c)
yi


a,












N
C)
C)
^13










z
n

C)
*3
*tO
c-3


a cLo = cuj D
tC3C
1 1 II I












Individual Event Morphologies

One qualitative morphological measure of the events shown in

Appendix C was made. For luminosities of 20%, 40%, 60%, and 80%,

the fraction of the event duration spent at or above each level was

determined. In some cases the luminosity dropped below a given level

only to rise above it again. In that case the partial durations above

that level were summed to obtain the final result.

Table 6 lists, by event, the decimal fraction of the event spent

at or above each level. Several events are not included in this table

because they are relatively incomplete (i.e., the ingress or decline

is not well documented). Values followed by a colon indicate less

severe, but still significant incompleteness.

A symmetrical, linear rise and decline would yield values of .80,

.60, .40 and .20 for levels of 20%, 40%, 60% and 80%, respectively. It

should be mentioned that it is possible to devise more complicated

event shapes which would yield these same results, but a glance at their

normalized flux curve would quickly reveal their true nature. There

are only three well-sampled events in Table 6 which match these values

reasonably well (0735+17 A, 3C 345 E and NRAO 512 C). The rates of

energy release are thus generally non-linear. The supernova light curve

in Figure 4 yields values of .85, .70, .56 and .40 for an event duration

of 45 days. Only one event (3C 371 B, .89, .78, .55, .15, 49.363 days)

matches these values relatively well, although its 80% duration is rather

short. In addition, its light curve shows a pair of peaks (100%, 95%)

separated by a decline to below 60%.










TABLE 6

FRACTIONAL FLUX DURATIONS


Event


0039+40 A

3C 66A A

0235+164 A
B
C
D

0420-01 A
B
C

NRAO 190 A
B

0521-36 A
B

0735+17 A
B
C

OJ 287 A
B
C
D

0906+01 A
B
C

1 T'+205 A

NGC 4151 A

3C 279 A
B
C

1308+326 A
B
C
D


Duration 20% 40%. 60. 0%.
days

375.460 .72 .36 .17 .03

115.369 .69 .09 .05 .01

49.676 .87 .67 .22 .11
175.497 .10 .04 .03 .02
91.093 .15 .11 .07 .04
76.566 .73 .31 .16 .06

319.263 .71 .31 .17 .07
370.230 .78 .52 .22 .05
310.808 .13 .03 .02 .01

44.774 .69 .52 .35 .17
44.184 .54 .36 .24 .12

704.265 .50 .41 .19 .11
117.536 .96 .68 .44 .22

71.090 .82 .57 .40 .16
147.971 .89 .42 .13 .02
394.044 .75 .19 .08 .01

107.736 .82 .60 .23 .01
129.456 .75 .47 .26 .08
112.051 .76 .47 .20 .08
1167.001 .70 .28 .17 .01

18.334 .94 .79 .63 .35
32.623 .81 .67 .31 .05
14.900 .65: .42 .27 .18

51.505 .45 .27 .18 .09

512.849 .94: .53 .18 .04

33.200 .64 .41 .20 .07
78.700 .29 .20 .14 .04
789.337 .07 .02 .01 .01

70.578 .95: .71 .14 .07
92.599 .80: .52 .16 .04
30.025 .46 .27 .15 .03
40.519 .65 .33 .21 .12


======_=================================










TABLE 6 (continued)


Event


1510-08 A

1545+21 A

4C 38.41 A

NRAO 512 A
B
C
D
E
F
G
H

3C 345 A
B
C
D
E
F
G
H

NRAO 530 A
B

3C 371 A
B

BL LAC A
B
C

3C 446 A
B
D

3C 454.3 A
B

2345-16 A


Duration 20% 40% 60% 80%
days

94.783 .66 .37 .19 .05

192.769 .79 .55 .29 .14

11.016 .74 .55 .36 .18

2.346 .69 .45 .30 .15
11.500 .82: .64: .43: .07
1.623 .69 .52 .35 .18
13.903 .87 .39 .10 .03
42.991 .68: .41 .28 .01
23.456 .84 .61 .37 .14
18.398 .63 .23 .15 .11
281.532 .80 .35 .15 .01

28.777 .87 .38 .07 .03
18.909 .52 .34 .21 .10
8.803 .82 .58 .38 .17
29.323 .79 .62 .45 .14
21.304 .89 .55 .43 .28
95.925 .45 .21 .01 .01
34.389 .77 .21 .08 .05
409.292 .63 .25 .06 .03

73.541 .25 .02 .01. .01
23.593 .67 .38 .19 .10

44.824 .64 .24 .09 .03
49.363 .89 .78 .55 .15

83.983 .67 .59 .32 .02
7.699 .68 .58 .42 .26
28.718 .98: .82: .66: .31

491.880 .46 .22 .09 .01
173.860 .88 .40 .04 .02
637.205 .66 .33 .13 .12

158.436 .62 .25 .07 .03
75.214 .95 .71 .51 .23

386.983 .30 .17 .05 .01


========================================












A careful examination of the light curves presented in Appendix C

reveals little to suggest that there is any typical shape to these

events. This is borne out as well by the diversity in the duration

fractions presented in Table 6. Even for events occurring in the same

object, there is no apparent pattern. This is reminiscent of the situ-

ation fon the long- erm light curves of compact objects and suggests

that there is little order or regularity in the emission and/or radiation

transfer process.

For some events (e.g. 1308+326 C, 3C 345 B and BL Lac B) the

decline from maximum luminosity is much steeper than the rise phase.

It is difficult to come up with an "explosive" type of energy generation

mechanism which will display this characteristic. If the variations

are due to the covering and uncovering of a constant source by obscuring

clouds of material, this type of behavior could be produced. The edges

of the "holes" would have to be rather sharply defined to produce the

type of events such as 0235+164 B, 0420-01 C and NRAO 530 A. Large

variations in optical depth are required as well to account for such

large luminosity changes. In addition, events which are correlated

over a large frequency range require "material" which is opaque to that

entire range. Delayed correlations would present even larger problems

to any obscuration model.

What is the relative position of the point of peak luminosity in

these events; that is does the event peak early, late, or is it symmetric?

Examination of the light curves by eye gives 8 which peak early, 27

symmetric, 14 late and 18 inconclusive. This again points out the

diversity of these events.












Individual Event Comments

Given below are a few descriptive comments for each of the events.

The term "consecutive" implies that the end point of one event marks

the beginning of the next event.

0039+40 A Multiple peaks.

3C 66A A Very fast rise, decline to 40%.

0235+164 A -

B Very sharp peak, actual event shorter?

C Same as above.

D Multiple peaks.

0420-01 A Dual peak.

B B and C are consecutive.

C Sharp peak, actual event shorter?

NRAO 190 A A and B quite similar.

B -

0521-36 A -

B Very symmetric.

0735+17 A Shape similar to type I supernova.

B Two events?

C At least two sharp peaks.

OJ 287 A Multiple peaks, incomplete decline.

B Fast rise, slow decline.

C Slow rise, quicker decline, multiple peaks.

D Longest duration event, contains B

0906+01 A -

B Dual peaks.

C -










1156+295 A Secondary peak.

NGC 4151 A Ingress incomplete.

3C 279 A Triple peak.

B Actual event shorter?

C Contains A and B, questionable event.

1308+326 A Incomplete ingress, secondary peak.

B Complex, multiple peaks

C Very rapid decline.

D -

1510-08 A Two events?

1545+21 A -

NRAO 512 A A,B,C, and D are consecutive.

B. Incomplete ingress, two events?

C Shortest event, 1.6 days, 2 magnitudes.

D Multiple peaks.

E Incomplete decline, two events?

F F and G are consecutive.

G -

H Extremely complex, contains E,F and G.

3C 345 A Sharp decline.

B Sharp decline.

C-

D -

E -

F Very sharp peak.

G Complex multiple peaks.

HI Contains B,C,D and E.












NRAO 530 A Extremely sharp peak, actual event shorter?

B Double peak.

3C 371 A Sharp rise, slow decline.

B. Decline incomplete.

BL LAC A Flat topped event.

B Flat topped event.

C Two events?

3C 446 A Very active, 6-8 sharp peaks.

B Multiple peaks, ingress incomplete.

C Decline only, sharp drop, dual peaks.

D Late peak, precipitous decline.

E Ingress only.

3C 454.3 A Multiple sharp peaks, ingress poorly defined.

B Slow ingress, decline incomplete.

2345-16 A Dual sharp peaks.

Matese and Whitmire (1978) have discussed the effects of strong

gravity on the optical appearance of rapidly fluctuating astronomical

sources. In their model of a delta function profile flash occurring in

the vicinity of a strong gravitational source, the flash is modified such

that the observer (depending on his orientation with respect to the

flash position) sees a broadened event. For some orientations a second

event is observed as a result of photons initially traveling away from

the observer having their paths bent around the mass and back toward the

observer. Secondary brightenings are fairly common, but are of a much

larger amplitude than the 1% expected from the above model. Satellite

observatories, such as the Space Telescope could potentially detect such

secondary events.












Variability and Emitting Region Sizes

Knowledge of the absolute physical dimensions of the region where

emission occurs in these objects is crucial to the determination of volume

dependent quantities such as particle and energy densities. Linear

dimension estimations (equated to such quantities as the Schwarzschild

radius of a black hole) are extremely important as well. For sources

which are not expanding relativistically, upper limits to the emitting

region sizes are usually estimated using luminosity variations and

light travel time arguments. Often the limiting diameter of the region

responsible for an observed variation is given as

D< cAT/(l+z)

where c is the speed of light, AT is the duration of the observed vari-

ation and z is the redshift. The use of this approximation has become

widespread, but in some cases it has been employed in situations where

it is no longer valid. Terrell (1967) derived a more precise expression

for this limiting diameter as

4cAT L
D<( (+z) AL

which reduces to

D< 1.2732 AT (L/AL)

where AT is in rest frame days, L is the average luminosity during the

variation, AL is the luminosity change and D is the diameter of the emit-

ting region in light days. This formulation takes into account not only

the timescale of the variation, but also the fractional variation as

well. For a relatively small luminosity change (a small change in the

object's magnitude) the quantity in parentheses can become quite signifi-

cant. We rewrite the above equation as












D< k(L, AL)AT
o
Table 7 shows the variation of k(L, AL), tne fractional luminosity

factor, for various changes in object magnitude.

TABLE 7

FRACTIONAL LUMINOSITY FACTOR



Am L2/L1 k(L,AL) V


0.01 1.0093 137.5 2.6 [ 06]
0.02 1.0186 69.1 3.3 [ 05]
0.05 1.0471 27.7 2.1 [ 04]
0.10 1.0965 13.8 2.6 [ 03]
0.20 1.2023 6.9 3.3 [ 02]
0.50 1.5849 2.8 2.2 [ 01]
1.00 2.5119 1.5 3.2 [ 00]
1.63 4.5040 1.0 1.0 [ 00]
2.00 6.3096 0.9 6.7 [-01]
4.00 39.8107 0.7 3.0 [-01]


For very small magnitude changes the factor k can be larger than

100. An object which is observed to vary 0.02 magnitudes in 5 minutes

(0.0035 days) has a maximum emitting region size of 0.24 light days, not

0.0035 light days as the original simple formulation would imply. As

showninthe last column of Table 7, this represents a volume factor, V,

of 3.3 [05] in the volume of the emitting region. Values of volume

dependent quantities based on the simpler diameter estimation will thus

be a factor of 3.3 [05] too large. Only for a variation of 1.63 mag-

nitudes will the simpler form be correct, and for very large variations

it leads to an overestimation of the emitting region size.

In terms of the object parameters previously discussed, for two

measurements a and b, we have

Lab (La + b)/2

Lab = La Lb












Substituting these values into the expression for D gives

D <0.6366 ATo (La + L )/La Lbh

For a given event, the minimum value for the emitting region

diameter will occur wnen AT o/AL is a minimum. Table 8 presents this

limiting diameter for the emitting region (in light days and in cm) as

well as the log of the variability timescale t in seconds for each of
v
the objects in this study. The values for certain objects are larger

than some previously quoted values. This results, in part, from the

effect of the fractional luminosity factor in the expression for D as

well as from the fairly strict criterion for a "statistically significant

variation."

Bolometric Luminosities

As with physical dimensions, a knowledge of the total energy

emitted by these objects is most important in both directly and indirectly

distinguishing between potential models for their structure and energy

generation and transfer mechanisms.

There are many difficulties involved in the accurate determination

of this quantity, which we will refer to as BL from now on. Two important

questions are; what does the complete spectrum of the object look like

and how does the shape and level of this spectrum change when an event,

such as those under study here, occurs?

There are certain regions of the spectrum which are inaccessible to

ground based observations. One is between 1 [12] and 1 [14] Hz and the

other is >1 [153 Hz. Satellite observations have begun to fill in the

latter, but the former remains relatively unexplored. As indicated by










TABLE 8

EMITTING REGION SIZES AND VARIABILITY TIMESCALES


Event Emitting Region Size log (t )
1. days cm sec.

0039+40 A 12.69 3.29 [163 6.04

3C 66A A 4.21 1.09 E163 5.56

0235+164 A 6.75 1.75 1163 5.77
B 0.72 1.87 153 4.79
C 0.79 2.05 E153 4.83
D 1.86 4.81 [153 5.21

0420-01 A 12.22 3.17 163 6.02
B 16.05 4.16 E163 6.14
C 3.18 8.25 1153 5.44

NRAO 190 A 7.70 2.00 [163 5.82
B 3.62 9.39 [153 5.50

0521-36 A 42.03 1.09 173 6.56
B 66.14 1.71 [173 6.76

0735+17 A 4.09 1.06 [163 5.55
B 0.71 1.85 E153 4.79
C 2.16 5.59 [153 5.27

OJ 287 A 2.43 6.30 1153 5.32
B 0.54 1.39 [153 4.67
C 0.38 9.83 [143 4.54
D 0.54 1.39 [153 4.67

0906+01 A 5.32 1.38 [163 5.66
B 1.41 3.65 [153 5.09
C 5.30 1.37 [163 5.66

1156+295 A 0.20 5.05 E143 4.22

NGC 4151 A 10.86 2.82 [163 5.97

3C 279 A 2.24 5.82 E153 5.29
B 2.66 6.90 153 5.36
C 1.40 3.64 E153 5.08

1308+326 A 5.82 1.51 C163 5.70
B 0.23 5.97 E143 4.30
C 0.80 2.06 E153 4.84
D 4.50 1.16 E163 5.59










TABLE 8 (continued)


Event Emitting Region Size
1. days cm


1510-08 A

1545+21 A

4C 38.41 A

NRAO 512 A


3C 345 A
B
C
D
E
F
G
H


2.47

175.54

3.35


0.62
0.57
0.44
1.77
1.74
16.95
6.67
1.74

4.28
2.82
1.10
24.75
5.63
0.34
1.90
1.10


6.40 [153


4.55

8.69

1.62
1.49
1.13
4.58
4.50
4.39
1.73
4.50

1.11
7.30
2.84
6.41
1.46
8.82
4.93
2.84


[171

[153

[153
[153
[153
[153
[153
[163

[153

[163
[153
[153
[163
[163
1143
[15)
[153
E153


NRAO 530 A 0.64
B 1.07

3C 371 A 2.28
B 2.71

BL LAC A 0.49
B 1-82
C 2.74

3C 446 A 0.44
B 2.82
C 0.79
D 1.22
E 45.03

3C 454.3 A 1.27
B 21.55


----------------------------------------


1.67 [153 4.74
2.76 [153 4.97


5.90
7.02

1.27
4.71
7.11


[153
[153

E153
[153
[153


5.29
5.37

4.63
5.20
5.37


1.13 [153 4.58
7.32 [153 5.39
2.05 [153 4.83
3.18 [153 5.02
1.17 [173 6.59

3.29 E153 5.04
5.58 (163 6.27

4.40 [153 5.17


log (tv)
sec.

5.33

7.18

5.46

4.73
4.69
4.58
5.18
5.18
6.17
5.76
5.18

5.57
5.39
4.98
6.33
5.69
4.47
5.22
4.98


2345-16 A


1.70











Oke et al. (1970), estimation of BL depends strongly on this region and

extrapolation to it using the optical and radio spectral indices can

lead to quite different results.

Many compact extragalactic objects are known to vary from the

radio to the X-ray regions, although not necessarily simultaneously.

Thus, to obtain an accurate representation of the instantaneous spectrum,

observations at all frequencies must be made at nearly the same time.

Because satellite time is limited and usually assigned months in advance,

such simultaneity is difficult to achieve. It is even more difficult if

one wishes to obtain such a spectrum during a large outburst.

Correlated optical-radio variability has been examined for a large

sample of these objects by Pomphrey et al. (1976) and Balonek (1981).

Certain OW's (0235+164, 0420-01, OJ 287 e.g.) have shown a high degree

of correlation, but many others have not. During the recent optical

outburst of 1156+295 Bregman (1982) observed a similar brightening

in the near and far UV regions accessible to the IUE (International

Ultraviolet Explorer satellite). No X-ray observations of this event

were possible since the Einstein satellite malfunctioned several months

earlier. Little is known at present about correlated X-ray-optical

variability.

Ground based and airborne investigations in the infrared, such as

those of Rieke and Kinman (1974), Rieke et al. (1976), O'Dell et al.

(1978) and Moore et al. (1980), suggest that from the optical out to

3-10 microns variations are well correlated and of similar amplitude.

For OJ 287 identical rates of decrease of flux at 0.44 and 10.5 microns

were found over a two-year period, during which the object's brightness












decreased by a factor of 15. It should be kept in mind that these are

observed, rather than emitted, wavelengths.

By Lower Limits

From present evidence it is reasonable to assume that when an

increase of flux is observed in the B-bandpass (0.44 microns), that a

similar increase will be seen, in the rest frame of the object, at least

between 0.2 and 2.0 microns. For the object in our sample with the

largest redshift, 4C 38.41 at z=1.814, the radiation emitted between

those limits will be observed between 0.56 and 5.6 microns. This is

within the limits of typical correlated variability.

Integration of the flux distribution between 0.2 and 2.0 microns

will yield a lower limit to the bolometric luminosity. This limit is

actually fairly conservative, excluding radio, far-TR and X-ray regions,

but until complete spectra can be obtained during an outburst, the

relative contributions of these regions are difficult to assess.

Calculation of BL

For power-law spectra, calculation of BL is straightforward. If

significant curvature is present, the procedure given below must be

modified somewhat. Curvature implies that the spectral index is a

function of frequency. For 0235+164, Rieke et al. (1976) pointed out

that between 10.6 and 0.36 microns (5.7 and 0.19 microns in the rest

frame) the spectral index ranged from 1.49 to 4.14. The spectra of

three objects were divided, in this work, into two sections, 0.2-0.6

and 0.6-2.0 microns. The infrared spectral indices were obtained from

O'Dell et al. (1978) and were 1.00, 0.64 and 0.86 for 0235+164, 0735+17












and BL Lac, respectively. The spectra of these objects were then

treated as two separate power laws between the above defined limits.

For a power-law spectrum we have

L(v) = kv

The value of the constant k is given by

k = (1.199 [151)a L2. Hz' ergs/sec

where L25 is the known luminosity at a rest wavelength of 2500R, which

corresponds to a frequency of 1.199 [153 Hz. The total luminosity over

the desired frequency range is then given by


BL 2 L(v) dv



= k -2 -dv

1+1
= [k/(1-_)]( 1 +1) for a e 1

= k(logv2 logy1) for a = 1

It is necessary then to supply a and L2S and perform the integration.

The limits of integration will be taken as 1 [14] to 1 [15] Hz (between

2 and .2 microns).

Table 9 lists the results of these calculations for each of the

events. Column 2 gives the peak bolometric luminosity (as we have

defined it), in ergs/sec, and column 3 gives the total energy, in ergs,

emitted during each event. Listed in columns 4 and 5 are the logs of

the variability timescales t and t which will be discussed later.
9 v
The peak luminosity recorded for any object in this sample was

1.23 [49] ergs/sec during the 1975 outburst of 0235+164, and this was










TABLE 9

EVENT LUMINOSITIES AND ENERGIES


Event


0039+40 A

3C 66A A

0235+164 A
B
C
D

0420-01 A
B
C

NRAO 190 A
B

0521-36 A
B

0735+17 A
B
C

OJ 287 A
B
C
D

0906+01 A
B
C

1156+295 A

NGC 4151 A

3C 279 A
B
C

1308+326 A
B
C
D


L log (tv)=====~=========


L
ergs/sec

3.71 [443

8.11 [463

1.23 E493
1.05 E493
2.09 1483
2.40 E[483

2.75 E463
3.89 [463
6.01 [463

3.57 [463
2.57 [463

1.75 [453
9.77 [443

5.13 E463
9.77 E463
1.53 [473

1.07 [471
9.55 [463
2.29 [463
2.65 E473

9.21 [463
7.59 [463
5.50 [463

1.05 E483

2.08 E433

1.58 C483
3.16 E483
3.16 [483

5.01 [473
3.24 [473
3.98 [473
1.86 [473


Ee
ergs

2.32 [513

1.23 [533

2.48 [553
1.71 [553
2.65 [543
4.80 E543

2.20 [533
4.45 [53
1.81 [533

4.91 [523
2.79 C523

2.49 E523
3.34 E513

1.17 [533
3.89 E533
1.15 [543

3.71 [533
3.03 [533
5.58 [523
6.56 [533

5.99 [523
6.40 [523
1.59 [523

1.28 [543

3.51 E503

1.36 [543
4.34 [543
1.78 [553

1.283 543
8.48 [533
2.40 [533
2.16 [533


log (t )
sec.

2.79

5.13

7.31
7.24
6.54
6.60

4.66
4.81
5.00

4.77
4.63

3.46
3.21

4.93
5.21
5.40

5.25
5.20
4.58
5.64

5.18
5.10
4.96

6.24

1.54

6.42
6.72
6.72

5.92
5.73
5.82
5.49


log (t )
sec.

6.04

5.56

5.77
4.79
4.83
5.21

6.02
6.14
5.44

5.82
5.50

6.56
6.76

5.55
4.79
5.27

5.32
4.67
4.54
4.67

5.66
5.09
5.66

4.22

5.97

5.29
5.36
5.08

5.70
4.30
4.84
5.59










TABLE 9 (continued)


Event L Ee log (tt) log (tv)
ergs/sec ergs sec. sec.

1510-08 A 4.77 [473 1.22 [543 5.90 5.33

1545+21 A 3.25 [453 9.38 [503 3.73 7.18

4C 38.41 A 5.27 [473 1.78 [533 5.94 5.46

NRAO 512 A 1.32 [473 8.89 [513 5.34 4.73
B 7.59 [463 2.57 [523 5.10 4.69
C 1.77 E473 9.45 [513 5.47 4.58
D 6.46 E463 1.87 [523 5.03 5.18
E 1.00 [473 1.14 [533 5.22 5.18
F 9.12 [463 5.83 [523 5.18 6.17
8 8.13 E463 2.43 [513 5.13 5.76
H 1.00 [473 7.77 E533 5.22 5.18

3C 345 A 2.69 [463 1.28 [523 4.65 5.57
B 2.63 [463 7.14 [513 4.64 5.39
C 5.75 [463 1.16 [523 4.98 4.98
D 4.17 [463 2.29 [523 4.84 6.33
E 5.37 [463 2.61 [523 4.95 5.69
F 5.25 [463 8.03 [523 4.94 4.47
G 8.49 [463 5.33 [523 5.15 5.22
H 5.75 [463 5.39 [533 4.98 4.98

NRAO 530 A 2.35 [473 1.76 [533 5.59 4.74
B 7.94 [463 4.59 [523 5.12 4.97

3C 371 A 2.89 E453 3.76 [513 3.68 5.29
B 2.45 [453 2.16 [513 3.61 5.37

BL LAC A 6.04 [453 1.47 [523 4.00 4.63
B 3.09 [453 7.05 [503 3.71 5.20
C 4.90 E453 6.44 [513 3.91 5.37

3C 446 A 1.17 [483 1.10 [553 6.29 4.58
B 7.59 [473 4.00 [543 6.10 5.39
C 1.45 [483 2.59 [533 6.38 4.83
D 1.58 E483 2.81 [553 6.42 5.02
E 6.46 [473 1.65 [543 6.03 6.59

3C 454.3 A 1.43 [473 3.66 [533 5.38 5.04
B 1.29 [473 1.75 [533 5.33 6.27

2345-16 A 2.88 1463 1.45 [533 4.68 5.17












nearly equaled (1.05 [493) during the subsequent event. Only three

other objects have exceeded 1 [48], 1156+295, 3C 279 and 3C 446.

Representative of the lower end of the peak luminosities recorded are

the compact galaxies NGC 4151 (2.09 [43]), 0039+40 (3.72 [441) and

0521-36 (1.74 [45]). In terms of peak luminosities the fainter end of

the quasar-BL Lac group merges with the brighter end of the compact

galaxies. The total range of peak luminosities spans almost six orders

of magnitude.

The bolometric event energies give the total energy (in ergs)

radiated during the event, corrected for the object's base level. These

values range from 3.51 E503 ergs for NGC 4151 A to 2.81 [55] for 3C 446 D.

What these luminosities and energies represent in terms of mass conversion

will be examined in the next chapter.

The third column of Table 9 gives the log of the minimum variability

timescale parameter (t ) devised by Fabian and Rees (1979) for a spher-

ically symmetric, homogeneous cloud of radiating matter. The shortest

timescale for variability occurs when the optical depth of the cloud is

unity and is given by

tz = 5 [-431 BL/f seconds

where BL is the bolometric luminosity as defined earlier, and f is the

mass to energy conversion efficiency factor. Here f was taken as 0.3, a

value referred to again in the section on black hole accretion. Reduc-

ing f will, of course, increase tz for a given BL. Values of the log of

tn for each event are given in column 3 of Table 9. Column 4 gives the

log of the variability timescale, t derived from observation.
v








67


Thirteen of the sources (27 of the events) show values of log(t )
V
which are more than 0.2 less than their corresponding log(t.), with the

largest difference being 2.4, that is, 250 times shorter than expected.

Significantly, perhaps, none of these discrepant objects are galaxies.

The value of t. could be lowered a bit by increasing f, but f must

remain less than 1 so this does not provide much help. If only one or

two objects were involved, one might suspect the values of BL, but

this is not the case. It seems that this simple model of Fabian and

Rees (1979) does not apply to a significant number of these sources.

At this point we will begin to examine various proposed mechanisms and

models for these objects.















CHAPTER VI
COMPARISONS WITH THEORETICAL MODELS

Much of the focus of early theoretical studies into the quasar

phenomenon was directed towards explaining the enormous energy gener-

ation rates, the shapes of the optical and radio continue and the creation

of the emission and absorption features. Only more recently has there

been an increase in the number of models which attempt to account for

variability as well. Given below is a summary of the major past and

present models.

(a) Massive Black Holes

i. Spherical accretion
ii. Disk accretion-general
iii. Disk oscillation and accretion
iv. Disk hot spots and flares
v. Beamed radiation models

(b) Supermassive Rotating Magnetoplasmic Bodies

i. Giant pulsars
ii. Non-degenerate oblique rotators

(c) Compact Star Clusters (Multiple Supernovae)

(d) Esoteric Models

i. Matter-antimatter anihilation
ii. Quark fusion
iii. White holes
iv. Plasma laser stars

The presently most accepted category of models is the first one,

and this category also provides a majority of the tests applicable to

the variability data presented in this work. Categories (b) and (c) will

be briefly examined, while the esoteric models given in category (d) are












generally not sufficiently developed to predict variability character-

istics.

General Energy and Emission Constraints

Almost all of the present models invoke the mechanism of incoherent

synchrotron emission from electrons as the source of the non-thermal

continuum. Early on, Hoyle et al. (1966) pointed out that, for the

energy densities apparently existing in these objects, inverse Compton

scattering could compete with the synchrotron process. The energy

densities were determined assuming isotropic emission of radiation,

emission region sizes calculated on the basis of variability timescales

and cosmological origin of the redshifts. Unless the magnetic energy

density exceeds the radiation energy density, inverse Compton losses in

these objects would be prohibitive. Field strengths of thousands of

gauss were necessary to avoid these losses. The constraints could be

eased by "moving" the objects closer than their cosmological z would

imply, by invoking non-isotropic emission and bulk relativistic motions.

As mentioned earlier, the first of these suggestions is still a source

of intense debate.

Columns 2 and 3 of Table 10 list, for each event, the log of the

radiation density (in ergs/cm) for the peak event luminosity and for

the average event luminosity. Emitting volumes were derived from observed

variability timescales. In some cases the values in column 3 exceed

1 [061. Columns 4 and 5 list the logs of the magnetic field strength in

gauss such that the inverse Compton losses just balance the synchrotron

emission, that is:










TABLE 10

EVENT ENERGY DENSITIES AND MAGNETIC FIELDS


Energy Densities Magnetic Fields
Event Log Peak Log Avg. Log Peak Log Avg.
ergs/sec - gauss -

0039+40 A 0.44 0.24 0.92 0.82

3C 66A A 3.74 3.51 2.57 2.45

0235+164 A 5.50 5.19 3.45 3.30
B 7.39 6.55 4.40 3.98
C 6.61 6.07 4.01 3.74
D 5.91 5.51 3.66 3.46

0420-01 A 2.35 2.01 1.88 1.70
B 2.26 1.95 1.83 1.68
C 3.85 3.17 2.63 2.28

NRAO 190 A 2.86 2.60 2.13 2.00
B 3.36 3.02 2.38 2.21

0521-36 A 0.07 -0.21 0.74 0.60
B -0.58 -0.71 0.41 0.34

0735+17 A 3.56 3.36 2.48 2.38
B 5.36 5.08 3.38 3.24
C 4.59 4.14 3.00 2.77

OJ 287 A 4.34 4.01 2.87 2.70
B 5.59 5.34 3.50 3.37
C 5.23 5.01 3.32 3.20
D 6.03 4.47 3.72 2.93

0906+01 A 3.59 3.48 2.50 2.44
B 4.65 4.45 3.03 2.93
C 3.37 3.21 2.39 2.31

1156+295 A 7.53 7.02 4.47 4.21

NGC 4151 A -0.67 -1.00 0.37 0.20

3C 279 A 5.57 5.25 3.49 3.32
B 5.73 5.10 3.57 3.25
C 6.29 5.23 3.85 3.31

1308+326 A 4.25 3.94 2.83 2.67
B 6.86 6.50 4.13 3.95
C 5.87 5.37 3.64 3.38
D 4.04 3.63 2.72 2.52










TABLE 10 (continued)


Energy Densities Magnetic Fields
Event Log Peak Log Avg. Log Peak Log Avg.
ergs/sec - gauss -

1510-08 A 4.97 4.53 3.18 2.97

1545+21 A -0.90 -2.01 0.25 -0.31

4C 38.41 A 4.75 4.53 3.08 2.96

NRAO 512 A 5.61 5.33 3.51 3.37
B 5.45 5.27 3.43 3.34
C 6.04 5.76 3.72 3.58
D 4.40 4.20 2.90 2.80
E 4.59 4.23 3.00 2.82
F 2.57 2.40 1.99 1.90
G 3.34 2.11 2.37 1.75
H 4.59 4.24 3.00 2.82

3C 345 A 3.24 3.02 2.32 2.21
B 3.59 3.38 2.50 2.39
C 4.75 4.61 3.08 3.00
D 1.91 1.80 1.66 1.60
E 3.30 3.18 2.35 2.29
F 5.73 5.27 3.57 3.33
6 4.44 4.17 2.92 2.79
H 4.75 4.36 3.08 2.88

NRAO 530 A 5.84 5.13 3.62 3.27
B 4.91 4.59 3.16 3.00

3C 371 A 2.83 2.65 2.12 2.02
B 2.60 2.39 2.00 1.89

BL LAC A 4.47 4.17 2.94 2.78
B 3.04 2.87 2.22 2.13
C 2.90 2.76 2.15 2.08

3C 446 A 6.86 6.29 4.13 3.84
B 5.05 4.68 3.23 3.04
C 6.45 5.80 3.93 3.60
D 6.11 4.31 3.76 2.86
E 2.58 --- 1.99

3C 454.3 A 5.03 4.75 3.22 3.07
B 2.52 ---- 1.96

2345-16 A 4.07 3.54 2.74 2.47












p= pM= B2/8
PR M
Average magnetic field strengths often reached 1 [033 gauss, on the same

order of magnitude as that found in sunspots, but over a vastly larger

region.

A second, recently determined, constraint concerns an analog

for electrons of the optical depth for photons. Caviliere and Morrison

(1980) point out that when

L45/R15 > 1

(where L45 is the bolometric luminosity in units of 1 [45] ergs and R1I

is the emitting region size in units of 1 [15lcm) electrons will undergo

repeated scatterings as they traverse the emission region. Strong

replenishment of the electron energies is then crucial for highly variable

sources, independently of the nature of their non-thermal radiation

emission mechanism.

Table 11 lists the values of L45/R15 for each event. The values

range from 1 [-03] to 5 [03], indicating that re-acceleration mechanisms

are required in some but not all of these objects. Caviliere and Morrison

point out that constant injection of new, high energy electrons results

in a density of "spent" electrons which would wash out rapid variability,

high polarization and even power-law spectral shapes.

Reduction of the implied luminosity by employing radiating needle

beam models (e.g. Blandford and Rees, 1979) would lower the values of

L45/R15. For the largest determined values of L45/R15 in Table 11 this

directivity would have to be very large. One consequence of this type

of model, an implied correlation between shorter variability timescales

and higher luminosity, is examined later in this chapter.










TABLE 11

EVENT PARAMETERS III


Event

0039+40 A

3C 66A A

0235+164 A
B
C
D

0420-01 A
B
C

NRAO 190 A
B

0521-36 A
B

0735+17 A
B
C

0J 287 A
B
C
D

0906+01 A
B
C

1156+295 A

NGC 4151 A

3C 279 A
B
C

1308+326 A
B
C
D


----------


_________


============================r===========


LS/1 R15

1.12 E-023

7.41 E 00]

7.08 E 023
5.62 I 033
1.02 C 033
5.01 E 023

8.71 [-013
9.33 E-013
7.24 E 00)

1.78 C 003
2.75 I 003

1.58 E-023
5.75 (-033

4.79 1 00]
5.25 [ 013
2.69 E 013

1.70 E 013
6.92 [ 013
2.34 013
1.91 C 023

6.61 C 00]
2.09 013
3.98 E 00]

2.09 C 033

7.41 C-043

2.75 C 023
4.57 C 023
8.71 C 023

3.31 1 013
5.37 E 023
1.95 1 023
1.62 E 023


Log (Te/t
---- -




-6.96
-4.94
-5.29
-5.93

-6.53
-6.64
-5.67

-7.08
-6.60

-6.99
-8.07

-6.47
-5.01
-5.31

-5.95
-4.89
-4.76
-3.94

-7.23
-6.12
-7.32






-6.41
-6.14
-4.72

-6.70
-4.48
-5.78
-6.78









TABLE 11 (continued)


Event Lk5/ R15 Log (Te /

1510-08 A 7.41 01 --

1545+21 A 7.08 C-033

4C 38.41 A 6.03 E 01 --

NRAO 512 A 8.13 E 013 -6.73
B 5.13 E 013 -5.97
C 1.58 C 023 -6.83
D 1.41 E 013 -6.63
E 2.24 C 013 -6.14
F 2.09 E 00] -7.88
6 4.68 C 003 -7.38
H 2.24 C 013 -5.32

3C 345 A 2.40 C 00] -6.90
B 3.63 C 003 -6.81
C 2.04 C 013 -6.53
D 6.46 E-013 -8.03
E 3.72 C 00] -7.21
F 5.89 E 013 -4.72
6 1.74 C 013 -6.29
H 2.04 C 013 -4.86

NRAO 530 A 1.41 E 023 -5.24
s 2.88 E 013 -6.08

3C 371 A 4.90 [-013 -6.28
B 9.12 [-013 -6.36

BL LAC A 4.79 C 003 -5.02
B 6.61 E-013 -6.91
C 6.92 (-013 -6.60

3C 446 A 1.05 E 033 -4.18
B 1.05 C 023 -5.84
C 7.08 [ 023 -5.04
D 5.01 E 023 -4.73
E 5.49 E 003 --

3C 454.3 A 4.37 C 013 -5.36
B 2.29 C 00) --

2345-16 A 6.61 00 --












Black Hole Models

Accretion of matter onto a supermassive black hole has been and

remains the most popular model for compact sources. Various models of

this type have been examined by Salpeter (1964), Lynden-Bell (1969),

Lynden-Bell and Rees (1971), Hills (1975), Shields and Wheeler (1976),

Young et al. (1977), Vila (1979) and Pineault (1980). The major variable

among these models is the nature and structure of the accreting material.

We will begin by examining the general black hole accretion process and

then look in detail at more specialized cases.

Mass Accretion

For matter accreting onto a black hole, the resulting rate of

energy release is given by.

L = f M c2

where f is the efficiency factor for the energy conversion process, M

is the mass accretion rate and c is the speed of light. Values of f for

three types of black holes are given by Thorne (1974) as 0.06 for a

Schwarzschild configuration, 0.30 when the hole is torqued by accretion

and damped by radiation and 0.42 for a maximal Kerr condition. Altering

the above equation somewhat gives

L = 1.70 E463] M ergs/sec

where M. is in solar masses per year and f is equal to 0.30. A second

relation of interest is

E = f c2

= 5.37 [531 M ergs

This expression gives the total energy released during the accretion of

M solar masses of material where f is, again, equal to 0.3.












Table 12 presents, for all 24 event objects, the mass accretion

rates in solar masses per year which correspond to each object's lowest

base and highest peak bolometric luminosities. In Table 13 the total

mass accreted during each event is shown.

The peak accretion rate was for 0235+164, corresponding to 7.2

[02] M,/year. It should be kept in mind that this peak rate was main-

tained for only a small fraction of the duration of the event. The

largest base rate seen, again in 0235+164, was about 14 M/year. Only

seven objects had base accretion rates greater than 1 ME/year.

Table 13 details the total converted mass for each of the events,

which ran from about 0.001 to 50 solar masses. These values closely

approximate the lower and upper bounds for the masses of normal stars.

This result will be discussed in detail in the section on accretion of

stars by black holes.

Black Hole Masses and Eddington Luminosities

The luminosity of an accreting black hole must be less than the

Eddington luminosity (LEDD), determined by the condition that the radi-

ation pressure on the infalling matter cannot exceed the gravitational

force (Eddington, 1921). We find that

LEDD = 3.2 [04] Le 'to
where L, is the luminosity of the sun and MHo is the mass of the black

hole in solar masses. The bolometric luminosity of an object thus sets

a lower limit on the mass of the black hole.

The minimum variability timescale, based on the black hole model,

is related to the gravitational radius of the hole, given by









TABLE 12

MASS ACCRETION RATES


Object L M L M
ergs/sec Me /yr ergs/sec Me/yr

0039+40 3.71 E 443 2.18 E-023 1.62 [ 443 9.53 E-033
3C 66A 8.11 E 463 4.77 C-023 3.55 [ 453 2.09 E 003
0235+164 1.23 E 493 7.24 E 023 2.38 [ 473 1.40 E 013
0420-01 6.01 E 463 3.54 C 003 4.52 E 453 2.66 E-013
NRAO 190 3.57 E 463 2.10 E 003 4.35 C 453 2.56 (-013
0521-36 1.75 E 453 1.03 E-013 3.94 [ 443 2.32 E-023
0735+17 1.53 C 473 9.00 C 003 1.32 C 463 7.76 E-013
OJ 287 2.65 C 473 1.56 E 013 7.26 C 453 4.27 E-013
0906+01 9.21 E 463 5.42 E 003 2.51 C 463 1.48 E 003
1156+295 1.05 C 483 6.18 E 013 3.71 E 463 2.18 E 003
NBC 4151 2.08 E 433 1.22 E-033 1.92 C 423 1.13 [-043
3C 279 3.16 C 483 1.86 E 023 1.20 C 463 7.06 C-013
1308+326 3.98 ( 473 2.34 C 013 1.13 C 463 6.65 E-013
1510-08 4.75 C 473 2.81 E 013 2.50 [ 463 1.47 E 003
1545+21 3.25 C 453 1.91 1-013 1.97 E 453 1.16 E-013
4C 38.41 5.27 C 473 3.10 C 013 1.26 E 473 7.41 E 003
NRAO 512 1.77 C 473 1.04 C 013 1.18 E 463 6.94 E-013
3C 345 8.49 E 463 4.99 E 003 1.10 E 463 6.47 E-013
NRAO 530 2.35 C 473 1.38 ( 013 1.56 C 463 9.18 E-013
3C 371 2.89 C 453 1.70 C-013 9.11 E 443 5.36 (-023
BL Lac 6.04 C 453 3.55 (-013 9.28 C 443 5.46 E-023
3C 446 1.58 [ 483 9.29 C 013 5.46 E 463 3.21 C 003
3C 454.3 1.43 E 473 8.41 C 003 3.46 E 463 2.04 C 003
2345-16 2.88 E 463 1.69 E 003 4.20 C 453 2.47 E-013










TABLE 13

EVENT ENERGIES AND CONVERTED MASSES


Event


0039+40 A

3C 66A A

0235+164 A
B
C
D

0420-01 A
B
C

NRAO 190 A
B

0521-36 A
B

0735+17 A
B
C

OJ 287 A
B
C
D

0906+01 A
B
C

1156+295 A

NGC 4151 A

3C 279 A
B
C

1308+326 A
B
C
D


Ee
ergs

2.32 C 513

1.23 C 533

2.48 C 553
1.71 C 553
2.65 E 543
4.80 C 543

2.20 U 533
4.45 E 533
1.81 C 533

4.91 [ 523
2.79 E 523

2.49 C 523
3.34 [ 513

1.17 E 533
3.89 [ 533
1.15 C 543

3.71 [ 533
3.03 E 533
5.58 [ 523
6.56 [ 533

5.99 [ 523
6.40 E 523
1.59 C 523

1.28 C 543

3.51 C 503

1.36 E 543
4.34 E 543
1.78 C 553

1.28 C 543
8.48 E 533
2.40 E 533
2.16 [ 533


Conv. Mass
M
.-----_
4.32 [-033

2.29 [-013

4.62 E 013
3.18 E 013
4.93 E 00)
8.94 C 003

4.10 E-013
8.29 [-013
3.37 [-013

9.14 [-023
5.20 [-023

4.64 [-023
6.22 [-033

2.18 E-013
7.24 [-013
2.14 C 003

6.91 [-013
5.64 [-013
1.04 [-013
1.22 [ 003

1.12 [-013
1.19 [-013
2.96 C 00)

2.38 C 00]

6.54 [-043

2.53 E 00]
8.08 E 003
3.31 C 013

2.38 C 003
1.58 C 003
4.47 [-013
4.02 E-013










TABLE 13 (continued)


Event


1510-08 A

1545+21 A

4C 38.41 A

NRAO 512 A
B
C
D
E
F
3
G
H

3C 345 A
B
C
D
E
F
G
H

NRAO 530 A
B

3C 371 A
B

BL LAC A
B
C

3C 446 A
B
C
D
E

3C 454.3 A
B

2345-16 A


=====================Ee==


Ee
ergs

1.22 C

9.38 E

1.78 E

8.89 I
2.57
9.45 C
1.87 E
1.14 E
5.83 1
2.43 C
7.77 C

1.28 :
7.14 C
1.16 E
2.29 :
2.61 E
8.03 E
5.33 C
5.39 E

1.76 C
4.59 E

3.76 E
2.16 C

1.47 C
7.05 C
6.44 E

1.10 C
4.00 C
2.59 E
2.81 1
1.65 E

3.66 E
1.75 C

1.45 C


Conv. Mass
M,

2.27 C 00]

1.75 C-033

3.31 C-013

1.66 [-023
4.79 C-023
1.76 (-023
3.48 C-023
2.12 [-013
1.09 [-01o
4.53 E-033
1.45 C 00]

2.38 [-023
1.33 [-023
2.16 C-023
4.26 [-023
4.86 C-023
1.50 [-013
9.93 [-023
1.00 E 00]

3.28 [-013
8.55 C-023

7.00 C-033
4.02 [-033

2.74 [-023
1.31 C-033
1.20 C-023

2.05 [ 013
7.45 C 00]
4.82 [-013
5.23 C 013
3.07 E 00]

6.82 [-013
3.26 E-013

2.70 [-013












Rg = G H/c2

= 1.5 [05] MHe cm

The variability timescale will then be

tv = n Rg/c (n>l)

= 1.5 [05] n MH%/c

We note that the minimum variability timescale sets an upper limit on

the mass of the black hole. Assuming that the object, at its peak

luminosity Lp, is radiating at the Eddington limit, and relating tv

and Lp gives.

S 1.5 105] Lp n
v 3.2 1041 LP c

= 4.0 [-44] n Lp seconds

log t = log n + log Lp -43.4

Figure 5 is a plot of log t versus log L for each of the events

studied in this work, and parallels the plot of Young et al. (1977).

The lines represent the relation between minimum variability timescales

and luminosities for n = 1, 2, 6 and 20 R Values of n = 1 and 6
g
correspond to the last stable orbits for maximal Kerr and Schwarzschild

black holes respectively, while n = 2 is the value used by Elliot and

Shapiro (1974) in a similar investigation. The last case, n = 20, is

based on Lynden-Bell's (1969) value of 22R for the radius of greatest

luminosity of the accretion disk. The filled circles, empty circles

and triangles represent events occurring in BL Lac objects, QSO's

and compact galaxies, respectively.

All but two of the events lie near or above the n = 1 line, making

their variability timescales at least marginally consistent with the
































c.

o

0


























0
0
4

C)
0
0






t
0
oo







-0
vi

0>.v





CC)
00~
04-,









00
.0l
004-

0u

00
0











general black hole model. Two events (0235+164 B and 1156+295 A) lie

well below the n = 1 line. There are many possible reasons for this.

If there is bulk relativistic motion, then the physical size of

the emitting region can be larger than that indicated by variability

arguments (Rees, 1966). There does not seem to be any problem with

the original photometry; in fact there are other pairs of measurements

which yield only slightly larger values of t for these events. Any of

the earlier mentioned factors which work to reduce the actual luminosity

could come into play. The emission directivity factor in a beaming

model for 0235+164 would have to be 0.15 to sufficiently lower its

luminosity.

In this regard, it was mentioned earlier that for beaming models

there should be a correlation between shorter variability timescales and

higher luminosities. Cursory examination of Figure 5 reveals some

indication of such a correlation between tv and Lp. The faint end of

the luminosity range is undersampled, however, which should be remedied

in the future by more vigorous observations of intrinsically faint

sources. Scatter will be introduced into the correlation mentioned

above if there is a wide range in the directivity factors for objects of

similar intrinsic luminosities.

A final comment concerning the position of the points on Figure

5 should be made. The actual bolometric luminosities of these objects

may be greater than that used here, if the contributions from the parts

of the spectrum not included in its calculation contribute significantly

to the total luminosity. If this were indeed the case, then the points

on Figure 5 will be shifted to the right, perhaps resulting in more

objects lying below the R = 1 line.
g











Upper and lower mass limits of the black holes involved can be

determined as indicated earlier from the shortest observed variability

timescale and the assumption that the peak luminosity equals the Eddington

luminosity. This differs slightly from the method used to create

Figure 5, in that these two conditions may not appear during the same

event.

Table 14 presents these lower and upper mass limits. For three

objects, 0235+164, 1156+295 and 3C 446, the lower limit exceeds the

upper limit by factors of 7, 2 and 1.5 respectively. One also notes

that the mass of the central object associated with 3C 279 seems rather

well determined.

Accretion of Stars

In the models of Hills (1975) and Young et al. (1977) matter

accreted onto a black hole was assumed to be from stars which were

tidally disrupted. Both investigators found that once the central mass

exceeded about 3 [081 solar masses it would swallow stars whole, with

a resulting decline in the luminosity from its peak value of about

8 [451 ergs/sec. This result, of course, assumes no other source of

accreting material. Examination of the base luminosities for the event

objects given in Table 13 shows thirteen objects whose base luminosities

exceed 8 [45] ergs/sec. Of these, however, only three (0235+164,

4C 38.41 and 3C 446) exceed this value by a factor of 5 or more. Conse-

quently, accretion of tidally disrupted stars seems able to account for

for the base luminosities in most of these objects.

If the base energy emission is due to tidally disrupted stars, what

is the source of the discrete event energy? Assuming the event energy is










TABLE 14

UPPER AND LOWER MASS LIMITS


Object Lower Limit Upper Limit
MHo MHo


0039+40
3C 66A
0235+164
0420-01
NRAO 190
0521-36
0735+17
OJ 287
0906+01
1156+295
NGC 4151
3C 279
1308+326
1510-08
1545+21
4C 38.41
NRAO 512
3C 345
NRAO 530
3C 371
BL Lac
3C 446
3C 454.3
2345-16


2.85
6.24
9.46
4.62
2.75
1.35
1.18
2.04
7.08
8.08
1.60
2.43
3.06
3.67
2.50
4.05
1.36
6.53
1.81
2.22
4.65
1.22
1.10
2.22


C063
[08J
[10]
[083
[083
[073
E093
C093
[083
[093
E[053
[103
[093
[093
C073
[093
[093
E083
[093
[073
[073
C103
[093
E083


2.19
7.27
1.25
5.50
6.26
7.27
1.23
6.55
2.43
3.37
1.88
2.43
3.98
4.27
3.03
5.79
7.53
5.88
1.11
3.93
8.47
7.53
2.19
2.93


C113
C103
C103
[103
[103
C113
C113
C093
E103
E093
[113
[103
C093
[103
[123
[103
C093
C093
E103
C103
[093
[093
C103
[103
[I0]









85

released as the result of mass accreted onto the black hole, then this

mass has two potential origins. It may be material which has never been

in stellar form and thus does not require disruption. Alternatively,

it may be material from stars which have been disrupted in some manner

other than tidally by the black hole. This seems the more likely of

the two possibilities. Such a process would provide discrete "bursts" of

accretable material, rather than an amorphous source of "free" mass. In

addition, as pointed out earlier in this chapter, the equivalent converted

masses for the energies emitted during these events lie almost exactly

within the range of known stellar masses.

Two processes for stellar disruption (other than tidal effects)

may be occurring in these objects. If the concentration of stars is

high, collisions between them become more likely. This collision

process can lead to the mechanical disruption of one or both stars or

may even initiate a supernova, the former producing accretable mass and

the latter both accretable mass and emitted radiation. Judging by the

typical frequency of these events one would require only a few such

collisions per year. Supernovae resulting from stellar evolution,

rather than collisions, could also contribute radiation and accretable

material.

Accretion Disk Timescales

The formation of accretion disks around the black holes in these

objects has been modeled by a number of people. Paczynski (1978)

presented a model where turbulence is driven by local disk instabil-

ities and the accompanying viscosity makes possible the accretion onto

the central object. He also found, however, that the timescale for such












accretion was on the order of 10 million years, but that local gravi-

tational instabilities could decrease this to hundreds of years. This

is still much too long to explain the known variations in these objects.

Pulsation of the accretion disk has been examined by Ledoux (1951),

Goldreich and Lynden-Bell (1965), Stewart (1975) and others. Vila (1980)

examined adiabatic oscillations in these disks and found that

P M1/2 r-3/2 = 0.035 days

where P is the oscillation period, M is the mass of the central object

(in grams) and r is the radius of the disk (in cm) at which the oscil-

lation occurred.

Table 15 shows the oscillation periods in days expected for various

typical combinations of r and M, where M is in solar masses.

TABLE 15

DISK PULSATION PERIODS



Disk Radius (cm)

MH1 1 [143 1 [ 15] 1 [161 1 E17]


1 [07] 2.5 [-01] 7.8 [ 00] 2.5 [02] 7.8 [03]
1 [09] 2.5 [-02] 7.8 [-01] 2.5 [013 7.8 [02]
1 [11] 2.5 [-03] 7.8 [-02] 2.5 E003 7.8 [01]


These values are in the same range as the durations of the events

studied here. Could these events represent a cycle of one of these

oscillations? If so, then for multiple events in a given object (so

that M is the same in all cases) we would find












P r-/2 = constant

which also means that

T /t3/2 = constant

3/2
Table 11 lists the log of T /t2 for all objects with more than one

event. For some objects (e.g. 0420-01) the values for each of the events

are quite similar, but for many objects the ratio varies by several

orders of magnitude from event to event. This result does not preclude

the possibility that some of these events are due to this type of

oscillation, but certainly not all of the events are so created.

Accretion Disk Flares

Few, if any, models prior to that of Pineault (1980) have suggested

that the origin of the background luminosity is significantly different

than that of the flare luminosity. Pineault proposes that the background

(orbase luminosity as we have referred to it) is a result of the central

object accreting at or below its Eddington luminosity, while the outburst

energy results from scaled up versions of solar flares occurring on the

accretion disk and resulting from the release of stored magnetic energy.

Models of this type are quite attractive because, for all of the objects

in this study, their base level luminosities are well below their Edding-

ton luminosities. This obviates the need to invoke beamed radiation, at

least for the base luminosity.

Pineault has applied his model to the 1975 outburst of 0235+164.

Following his procedures we will calculate similar paramters for the

events in this study. We find from his scaling laws that

RF = RS T/TS

BF BS (EF/ES)1/2 (T )3/2











where R is the flare region size, T is the flare duration and EF is
F e F
the energy in the flare. The laboratory scaling values are RS = 1 o01]

cm, TS = 1 [-06] seconds, BS 1 [043 gauss and ES = 1 [8O ergs. Sub-

stituting these values in the above two equations gives

RF = 1 [07] T cm

BF = 1 [-09] EF 1/2/T 3/2 gauss

where

EF = ERF g(e)/j

where ERF is the radiative flare energy, j is the fraction of the flare

energy given off as radiation, and g(g) is the correction factor allow-

ing for non-isotropic emission of the flare energy. Values of RF and

BFJ /g(9)1/2 are given in Table 16. All of the values of RF are

smaller than their emission region size counterparts calculated from

variability timescales.

We can compare the size of the flare region to the thickness of

the accretion disk using the model of a thin disk by Novikov and Thorne

(1973) in which the thickness H is given by

H = 1.34 [14] M cm

where M. is in solar masses per year. The base level values of M, were

given in Table 12. The third column of Table 16 lists (next to the

first object event) the value of H obtained using MO.

In a majority of cases H >RF, indicating that, even in the

extreme case of the viewing the disk edge on, the flare region solid

angle will be less than that of the disk. In the cases where RF > H,

the disk could be being viewed at a non-zero inclination. Also the










TABLE 16

MAGNETIC FLARE MODEL PARAMETERS


Event


0039+40 A

3C 66A A

0235+164 A
B
C
D

0420-01 A
B
C

NRAO 190 A
B

0521-36 A
B

0735+17 A
B
C

OJ 287 A
B
C
D

0906+01 A
B
C

1156+295 A

NGC 4151 A

3C 279 A
B
C

1308+326 A
B
C
D


Flare
Extent
RF (cm)

3.24 [143

9.97 E133

4.29 [133
1.52 [143
7.87 [133
6.62 [133

2.76 [143
3.20 [143
2.69 E143

3.87 [133
3.82 [133

6.08 1143
1.02 [143

6.14 E133
1.28 [143
3.40 E143

9.31 E133
1.12 E143
9.68 E133
1.01 [153

1.58 [133
2.82 [13]
1.29 E133

4.45 [133

4.43 [143

2.87 [131
6.80 [133
6.82 [143

6.10 [133
8.00 [133
2.59 [133
3.50 E133


Disk
Thickness
H (cm)

1.28 [12]

2.80 [143

1.88 [153




3.56 [133



3.43 E133


3.11 [123


1.04 [143



5.72 [133





1.98 [143


2.92

1.51

9.46


8.91 133


BF (j/g(B))
gauss

2.61 [053

1.11 [073

5.60 [08]
7.00 [073
7.37 [073
1.29 [086

3.24 [063
3.69 [063
3.06 [063

2.91 [073
2.24 [073

3.32 [053
1.79 [063

2.25 [073
1.36 [073
5.40 [063

2.14 [073
1.47 [073
7.84 E063
8.00 [05]

1.23 [083
5.35 [073
8.63 [073

1.21 [083

6.35 [043

2.40 [083
1.17 [083
7.49 [063

7.51 [073
4.07 [07]
1.17 [083
7.10 E073


E143

[103

[133


====================~===================










TABLE 16 (continued)


Flare Disk
Event Extent Thickness BF (j/(8))
R (cm) H (cm) gauss
------------------- --------------- -
1510-08 A 8.19 E133 1.97 E143 4.71 E073

1545+21 A 1.67 E143 1.55 [133 4.51 [053

4C 38.41 A 9.50 [123 9.93 E143 4.55 E083

NRAO 512 A 1.99 [123 9.30 E133 1.06 [093
B 9.94 E123 1.62 E083
C 1.38 123 1.89 E093
D 1.20 133 1.04 E083
E 3.71 E133 4.72 E073
F 2.03 1133 8.37 E073
8 1.59 E133 2.46 E073
H 2.43 E143 7.35 E063

3C 345 A 2.49 E133 8.67 E133 2.88 E073
B 1.62 E133 4.08 [073
C 7.61 E123 1.62 E083
D 2.53 E133 3.76 E073
E 1.84 E133 6.47 E073
F 8.29 E133 1.19 E073
6 2.97 C133 4.51 E073
H 3.54 E143 3.49 E063

NRAO 530 A 6.35 [133 1.23 E143 2.62 C073
B 2.04 [133 7.36 E073

3C 371 A 3.87 E133 7.18 E123 8.05 C063
B 4.26 E133 5.28 063

BL LAC A 7.26 E131 7.32 C123 6.20 E063
B 6.65 E123 4.89 E073
C 2.48 133 2.05 E073

3C 446 A 4.25 E143 4.30 C143 1.20 E073
B 1.50 C143 3.44 E073
C 1.39 E143 9.85 E063
D 5.51 E143 1.30 073


3C 454.3 A 1.37 E143 2.73 E143 1.19 [073
B


2345-16--A- 3.34-- 14)-- 3.31--13)-- 1.97-


1.97 E063


2345-16 A


3.34 E143 3.31 C131












value of M. (and thus H) goes inversely as the mass to energy conver-

sion efficiency factor, which for these calculations was 0.30, but which

certainly could be much lower.

This magnetic flare model seems to be able to account for many of

the properties seen in these objects and does not suffer so badly from

energetic problems as do others. It would be extremely useful if the

model could be developed in such a way as to predict the time dependence

of the flare emission process so that it could be compared to the events

studied here.

Supermassive Rotating Magnetoplasmic Body

The model of compact objects as supermassive rotating magnetoplas-

mic bodies is summarized by Ginzburg and Ozneroy (1977). The radiation

is a combination of thermal and synchrotron contributions, generated

by the rotator in the course of collapse and rotational braking. One

of the main points of this type of model is its ability to explain certain

quasi-periodic behavior in the observed light curves. These periodi-

cities, determined by Fourier analysis of light curves, generally do not

persist for many cycles, require phase drift and seem rather questionable

at best as to their reality. Many false periods can arise from observ-

ational selection effects. We note that in the table of periodicities

of Ginsburg and Ozneroy (1977) three of the periods are within 20 days

or so of one year, while the period of 29.5 days listed for NGC 1275

is exactly the synodic period of the moon. Because the observing

seasons for objects vary as a function of the object's declination

combined with the observer's latitude, checks for this type of false

period must be done as well.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs