PHOTOGRAPHIC STUDIES OF QUASI-STELLAR OBJECTS
AND OTHER ACTIVE RADIO SOURCES
ROGER LEONARD SCOTT
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
The author is extremely grateful to his advisor and
Graduate Committee chairman, Dr. Alex G. Smith, for guidance
and help throughout the course of his graduate studies, and
for the intriguing photographic research that Dr. Smith and
he have performed together. The contributions made by Drs.
G. R. Lebo, C. N. Olsson, F. B. Wood, and T. L. Bailey as
members of his Graduate Committee are acknowledged and appre-
ciated. Drs. N. F. Six and S. S. Ballard are thanked for
help received at various times during the author's graduate
The photographic studies of quasars in this dissertation
would not have been possible without the help of A. G. Smith,
R. J. Leacock, K. R. Hackney, R. L. Hackney, G. 11. Folsom,
B. Q. McGimsey, and P. L. Edwards in the acquisition and
reduction of the data. Special thanks are extended to G. H.
Folsom and II. R. Miller for their help in the BL Lac and PKS
0735+17 campaigns, and for kindly allowing data taken at the
Bradley Observatory of Agnes Scott College to be included in
E. E. Graves, II. W. Schrader, and W. W. Richardson con-
structed and maintained much of the equipment at Rosemary
Hill, and donated much time and effort to helping the author
with various problems during his graduate work. Special
recognition is given to E. E. Graves for the careful con-
struction of the Cassegrain camera and the cold emulsion
cassette used in the author's studies. The recently installed
Rosemary Hill Ritchey-Chr6tien 18-inch telescope was acquired
and refurbished largely through the efforts of J. P. Oliver,
A. G. Smith, and E. W. Ludington. The final adjustments
necessary to allow the telescope to become photographically
operational were performed by R. L. Hackney and K. R. Hackney,
assisted by a special grant from the National Science Founda-
tion. P. L. Coleman and R. B. Warren are thanked for helpful
advice pertinent to the cold emulsion investigations.
Special gratitude is extended to M. A. Lynch, R. L.
Hackney, S. S. Prasad, and P. L. Edwards for the many hours
of help given in the photographic study of Jupiter. Thanks
are given to A. E. S. Green and to R. D. McPeters for the use
of their microdensitometer to scan the Jovian images. The
proper operation of the densitometer was assured by the in-
stallation of parts by J. P. Oliver, borrowed from a similar
instrument loaned to the department by the National Aeronau-
tics and Space Administration.
The author was supported during his studies by graduate
assistantships from the Department of Physics and Astronomy,
the Graduate School, the National Science Foundation, and the
College of Arts and Sciences. The quasar research program
at Rosemary Hill is partially supported by the National Sci-
ence Foundation through its grant NSG 23456. The careful
preparation of the final manuscript by Mrs. Elizabeth Godey
is most appreciated. Gratitude is expressed to Mrs. C. J.
Kerrick, the department secretary, for her friendship, help,
The author is especially grateful to Karen R. Hackney
and Richard L. Hackney for their help and friendship during
the course of this study. Their companionship and encourage-
ment have been much appreciated. The author's most profound
gratitude is to his father, Mr. Sankey Scott. During the
author's long graduate career his father's support and
encouragement have never faltered, and have been a major
factor in its successful completion. It is to his father
that the author dedicates this work.
TABLE OF CONTENTS
ACKNOWLEDGML.I S . .
LIST OF TABLES . . .
LIST OF FIGURES . . .
ABSTRACT . . . . .
T TMTDOnnr C' T
Quasars: An Astronomical Enigma .
Variability of Quasars . . .
The Redshift-Luminosity Paradox
Quasars and Related Objects . .
Models . . . . . . .
Supernova Theory . . . .
Sninar Theory . . . . .
Irtron Model . . . . .
Gravitational Lens Hypothesis .
The Florida Program . . . . .
Jupiter, Another Active Radio Source
II RESEARCH TECHNIQUES AND INSTRUMENTATION
Photographic Studies of Quasars . .
Program Techniques . . . .
Magnitude Systems . . . .
A Photographic Attempt to Detect the
Jovian "Hotspot" . . . . .
The Cassegrain Camera . . .
Plates and Filters . . . .
Choosing Observation Times . .
Observing Techniques . . . .
. . . . ii
. . . . ix
. . . . xi
. . . . xvi
. . 23
. . 29
. . 36
. . 48
. . 51
. . 56
TABLE OF CONTENTS Continued
III NITROGEN BAKING AND EMULSION COOLING AS
METHODS OF HYPERSENSITIZATION . . .
Introduction . . . . . . .
Ilypersensitization of Kodak 103a-0
Plates by Nitrogen Baking . . . .
Experimental Procedure . . . .
Results . . . . . . .
Photometric Tests . . . . .
Summary . . . . . . .
Hypersensitization by Emulsion Cooling
Experimental Procedure . . . .
Results . . . . . . .
Summary . . . . . . .
IV A STUDY OF TIE OPTICAL VARIABILITY OF A
SAMPLE OF QUASARS AND RELATED OBJECTS .
Introduction . .
PKS 0048-09 . .
PKS 0420-01 . .
PKS 0725+14 . .
PKS 0735+17 . .
PKS 0736+01. .
PKS 0906+01 . .
PKS 1004+13 . .
PKS 1055+20 . .
PKS 1116+12 . .
4C 09.42 . . .
PKS 1510-08 . .
PKS 1645+17 . .
PKS 2145+06 . .
PKS 2349-01 . .
Summary . . . .
V COLOR STUDIES OF QUASARS
. . 59
. . 59
. . 62
. . 63
. . 64
. . 72
. . 73
. . 74
. . 75
. . 81
. . 89
. . 91
. . . . . 91
d Discussion . ... 94
. . . . . 94
. . . . . 95
. . . . . 98
. . . . . . 98
. . . . . . 101
. . . . . . 104
. . . . . . 10'9
. . . . . . 112
. . . . . . 112
. . . . . . 119
. . . . . . 119
. . . . . . 126
. . . . . . 126
. . . . . . 131
. . . . . . 131
AND RELATED OBJECTS
Introduction . . . . . . .
The Recent Photometric History of the
Florida Objects . . . . . . .
3C 120 . . . . . . . .
3C 390.3 . . . . . . . .
TABLE OF CONTENTS Continued
BL Lac . . . . . . . .
0, 287 . . . . . . . .
3C 273 . . . . . . . .
3C 345 . . . . . . . .
3C 454.3 . . . . . . .
PKS 2134+004 . . . . . .
4C 05.34 . . . . . . .
Color Variations of the Florida Objects
The Short-Term Color Trends of OJ 287
and 3C 345 . . . . . . .
Possible Relationships Between Classes
of Objects . . . . . . .
The Color-Redshift Relation for Quasi-
Stellar Objects . . . . . .
VI A STUDY OF TIE SHORT-TERM OPTICAL ACTIVITY
OF FIVE LACERTIDS . . . . .
Introduction . . . . . . .
The Optical Activity of PKS 1514-24,
ON 231 and ON 325 . . . . . .
PKS 1514-24 . . . . . .
ON 231 . . . . . . . .
ON 325 . . . . . . . .
A Search for Intraday Optical Activity
in BL Lac and PKS 0735+17 . . . .
BL Lac . . . . . . . .
PKS 0735+17 . . . . . .
Summary . . . . . . . .
VII A PHOTOGRAPHIC SEARCH FOR AN IO-RELATED
JOVIAN HOTSPOT . . . . . . .
Introduction . . . . . . .
The Estimated Brightness of the Spot
The Jovacentric Coordinates of the Spot
Digitizing the Jovian Images . . .
The Computer Programs . . . . .
Examination of the Computer Output .
. . 293
TABLE OF CONTENTS Continued
VIII CONCLUSIONS . . . . . .
I LIBRARY OF JUPITER IMAGES . . .
II THE SUPERIMPOSED ARRAYS FROM JSUPER .
Index . . . . . . . .
BIBLIOGRAPHY . . . .
BIOGRAPHICAL SKETCH . . .
. . . 295
. . . 309
. . . 319
. . 320
. . . 349
. . . 356
LIST OF TABLES
1 Observational Similarities Between Quasars,
and Seyfert and N-Type Galaxies . . .
2 Speed Gains Relative to Unbaked 1L1 for Five
Emulsion Batches Baked at 650C . . . .
3 The Effect of Cooling on Emulsion Speed .
4 Effect of Cooling on Baked IIIa-J . . .
5 A Sample of Objects from the Florida Program
6 The Photographic Comparison Star Magnitudes
for PKS 0048-09 . . . . . . .
7 The Objects Used for the Color Studies . .
8 The Rosemary Hill U, B, and V Observations
of 3C 390.3, 3C 273, 3C 454.3, PKS 2134+004,
and 4C 05.34 . . . . . . . .
9 The Comparison Stars for 3C 120 . . .
10 The Comparison Stars for 3C 390.3 . . .
11 The Comparison Stars For BL Lac . . .
12 The Comparison Stars for OJ 287 . . .
13 The Comparison Stars for 3C 273 . . .
14 The Comparison Stars for 3C 454.3 . . .
15 The Comparison Stars for PKS 2134+004 . .
16 The Comparison Stars for 4C 05.34 . . .
17 The Observed Short-Term Ranges in Magnitude
and Color of the Florida Objects . . .
18 The Comparison Stars for ON 231 . . .
19 The Comparison Stars for ON 325 . . .
LIST OF TAI.liS Continued
Intraday Observations of BL Lac . . ... 238
Intraday Observations of PKS 0735+17 . . .. .259
A Comparison of the Expected Brightness of
the Auroral Hotspot and the Jovian Limb . . 269
LIST OF FIGURES
1 Observed correlation on the redshifts of
quasars and radio galaxies, with angular
size . . . . . . . . . . 10
2 The continuum spectra of three Seyfert
galaxies, NGC 1275, NGC 1068, and 3C 120,
and the quasar 3C 273 . . . . ... 14
3 Color-color diagram for quasars, Seyfert
galaxies, and N-type galaxies . . . .. 17
4 A polar diagram of Jupiter's decametric
radio sources .. . . . . . . . 25
5 Current sheets in the Io flux tube . . .. 27
6 The Rosemary Hill Observatory 30-inch Newton-
ian/Cassegrain Tinsley reflector, with both
Newtonian and Cassegrain cameras mounted . . 31
7 Comet Kohoutek in the brilliant zodical
light, as seen from Rosemary Hill Observa-
tory on the evening of January 10, 1974,
at 2030 EDT . . . . . . . . . 33
8 The Rosemary Hill Observatory 18-inch
Ritchey-Chritien telescope showing offset
guiding telescope and camera . . . . .. 35
9 The infrared system response function used
in the QSO studies, obtained by combining
the Kodak type I-N emulsion with a Schott
RG-8 infrared filter . . . . . . . 40
10 The Cassegrain camera mounted on the guide-
box of the 30-inch reflector showing offset
guider, counterweights, base tube, rotation
clamps, eyepiece projection tube, filter
holder, shutter, with cable release, and
camera bed, with dark slide . . . . .. 43
11 Tricolor separation print of Jupiter from
negatives taken on July 29, 1972 . . .. 47
LIST OF FIGURES Continued
12 Spectral reflectivity of Jupiter . . ... .50
13 The ultraviolet response function used in the
Jupiter studies, obtained by combining the
Kodak type 103a-0 emulsion with a Kodak #18 A
filter . . . . . . . . .. .. . 53
14 The infrared system response functions used
in the Jupiter studies, obtained by combining
the Kodak type I-Z emulsion with a Schott RG-8
filter for broad band coverage, and a Corning
7-56 filter for narrow band coverage . . .. 55
15 Speed and fog curves for emulsion batches 1L1
and 1K2, baked in nitrogen at 65C . . .. 66
16 Speed and fog curves for batch 1E2 baked at
650C in air and in dry nitrogen . . . .. 69
17 Speed and fog curves for batches 1L3 and 1K2
baked in dry nitrogen at 72.50C . . .. 71
18 Schematic diagram of the cold cassette ... 78
19 Photograph of the cold cassette . . ... .80
20 Speed and fog curves obtained by cooling Kodak
Tri-X Pan and Kodak Contrast Process Pan . . 85
21 Speed and fog curves obtained by cooling Kodak
103a-0 and Kodak IIIa-J . . . . ... 88
22 The Florida light curves of PKS 0048-09, PKS
0420-01, and PKS 0725+14 . . . . . . 97
23 The sequence finder for PKS 0725+14 . . .. 100
24 The Florida light curves of PKS 0735+17 and
PKS 0736+01 . . . . .. . . . . 103
25 The sequence finder for PKS 0736+01 . . .. .106
26 The Florida light curve of PKS 0906+01 . . 108
27 The Florida light curves of PKS 1004+13, PKS
1055+20 and PKS 1116+12 . . . . ... 111
28 The sequence finder for PKS 1004+13 . . . 114
LIST OF FIGURES Continued
29 The sequence finder for PKS 1055+20 . .
30 The sequence finder for PKS 1116+12 . .
31 The Florida light curves of 4C 09.42, PKS
1510-08 and PKS 1645+17 . . . . .
32 The sequence finder for 4C 09.42 . . .
33 The sequence finder for PKS 1510-08 . .
34 The sequence finder for PKS 1645+17 ..
35 The Florida light curves of PKS 2145+06 and
PKS 2349-01 . . . . . . . .
36 The sequence finder for PKS 2145+06 . .
37 The sequence finder for PKS 2349-01 . .
38 The Florida four-color light curve of 3C 120
39 Apparent variations in the color indices of
3C 120 . . . . . . . . . .
40 The sequence finder for the N galaxy
3C 390.3 . . . . . . . . .
41 The Florida four-color light curve of BL Lac
42 The color indices of BL Lac corresponding
to the light curves shown in Figure 41 . .
43 The Florida four-color light curve of OJ 287
44 The color indices of OJ 287 corresponding
to the light curves of Figure 43 . . .
45 The Rosemary Hill Observatory light curve
of OJ 287 from early 1970 to the present,
showing an increase in the long-term compon-
ent before the long decline shown in the
color observations . . . . . . .
46 The sequence finder for the lacertid OJ 287
47 The sequence finder for the quasar 3C 273
LIST OF FIGURES Continued
48 The Florida three-color light curve of 3C 345. 175
49 The color indices of 3C 345 corresponding
to the light curves of Figure 48 . . ... .177
50 The sequence finder for the quasar 3C 454.3 . 181
51 The sequence finder for the quasar PKS
2134+004 . . . . . . . .... . .184
52 The sequence finder for the quasar 4C 05.34 188
53 The normalized color regression plots of
OJ 287; the long-term decline in intensity
has been removed . . . . . . . . 195
54 The normalized color regression plots of
3C 345; the long-term declines in intensity
has been removed . . . . . . . . 197
55 A model to explain the color trends suggested
by OJ 287 and 3C 345 . . . . . ... 202
56 The positions of the objects used in the
color study on a color-color diagram . . .. .208
57 The observed correlations between the color
indices of quasars and their redshifts . .. .214
58 The Florida light curve of PKS 1514-24 . .. .219
59 The sequence finder for PKS 1514-24 . . . 221
60 The Florida B and V light curves of ON 231 . 224
61 The sequence finder for ON 231 . . ... . 227
62 The Florida B and V light curves of ON 325 . 230
63 The sequence finder for ON 325 . . . . 233
64 Light curves of BL Lac intraday campaigns
held on August 9 and October 2 at Rosemary
Hill Observatory .. . . . . . ... 240
65 Light curves of the joint Florida-Georgia
BL Lac intraday campaign held on October 4,
1973 . . . . . . . . ... . . 242
LIST OF FIGURES Continued
66 Light curves of BL Lac intraday campaigns
held at Rosemary Hill on June 23 and July 16
of 1974 using the 30-inch and 18-inch
reflectors . . . . . . . . . 244
67 Light curve of PKS 0735+17 intraday campaign
held at Rosemary Hill on January 24, 1974 . .251
68 Light curves of PKS 0735+17 intraday campaigns
held at Bradley Observatory on January 31,
February 1, and February 12 of 1974 . . .. .253
69 Light curves of PKS 0735+17 intraday campaigns
held at Bradley Observatory on February 18,
February 27, and March 15 of 1974 . . .. 255
70 Light curves of joint Florida-Georgia PKS
0735+17 intraday campaigns held on February
25 and March 18 of 1974 . . . . . . 257
71 The projected Jovacentric coordinates X
and Y of the hotspot . . . . . . . 272
72 A polar view of the flux tube projected
positional geometry . .. . . . . 274
73 The model MK III C Joyce-Loebl scanning
microdensitometer . . . . . . . 279
74 The analog-to-digital converter designed and
built by Dr. M. A. Lynch . . . . . . 281
75 The 10-level pixel array of infrared image
no. 72 generated by JSCAN, compared with a
standard print of image no. 72 . . . .. 290
76 The 35-level centered pixel array of image
no. 72 generated by JSUPER compared with a
standard print of image no. 72 . . . . 292
Abstract of Dissertation Pri'-:ited to the Graduate Council
of the University of Florid:a in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
PHOTOGRAPHIC STUDIES OF QUASI-STELLAR OBJECTS
AND OTHER ACTIVE RADIO SOURCES
Roger Leonard Scott
Chairman: Dr. Alex G. Smith
Major Department: Astronony
Nitrogen baking of the Kodak type 103a-0 spectroscopic
emulsion has proved to be an effective means of hypersensiti-
zation, yielding speed gains of from 2 to 3 over an untreated
plate. The preliminary results of this process were reported
by R. L. Scott, A. G. Smith, R. J. Leacock, B. Q. McGimsey,
P. L. Edwards, K. R. Hackney, and R. L. Hackney in "Response
of 103a-0 Spectroscopic Emulsion to Controlled Baking and Its
Advantages for Photographic Monitoring of Quasi-Stellar
Objects," presented at the 40th Meeting of the Southeastern
Section of the American Physical Society, 1973 November 8-10,
in Winston-Salem, North Carolina. After further experimenta-
tion, the final results were reported by R. L. Scott and
A. G. Smith in "llypersensitization of Kodak type 103a-0
Plates by Nitrogen Baking," Astronomical Journal, 79, 656
Studies of the optical variability of a sample of 14
quasars and similar objects at Rosemary Hill Observatory
detected statistically significant fluctuations in the light
curves of PKS 0048-09, PKS 0420-01, PKS 0735+17, PKS 0736+01,
PKS 0906+01, PKS 1510-08, and PKS 2349-01. The preliminary
results of this study were partially communicated by R. L.
Scott, A. G. Smith, R. J. Leacock, B. Q. McGimsey, P. L.
Edwards, K. R. Hackney, R. L. Hackney, and G. H. Folsom in
"Optical Variations of a Sample of Quasars and Related
Objects," presented at the 41st Meeting of the American Astro-
nomical Society, 1973 December 2-6, in Tucson, Arizona.
Apparent color variations observed in a sample of QSO's
at Rosemary Hill appear to be generally compatible with the
expanding source model. However, OJ 287 and 3C 345 appear to
be exceptions, as variations in their color indices suggest
that changes in the slope of the optical spectrum may not be
simply correlated with brightness. A spectral model composed
of thermal and non-thermal components is offered as a possible
explanation for the color variations observed in the objects
studied. The optical variations of OJ 287 are discussed by
A. G. Smith, R. L. Scott, R. J. Leacock, B. Q. McGimsey, P. L.
Edwards, R. L. Hackney, and K. R. Hackney in "Four-Color
Photometry of OJ 287 During Its Recent Three-Magnitude
Decline," Publications of the Astronomical Society of the
Pacific, in press (1975).
Studies of the lacertids PKS 1514-24, ON 231, and ON 325
suggest that the optical behavior of PKS 1514-24 and ON 231
is similar to that of the prototype lacertid BL Lac, while
ON 325 shows more moderate short-term optical behavior super-
imposed on long-term fluctuations of several months' duration.
A search for intraday activity in BL Lac detected statis-
tically significant variations on two nights; the largest
such variation was a decline of 01.6 in 33 minutes. Observa-
tions of PKS 0735+17 suggested that intraday activity may
have been marginally detected. The preliminary results of
this study were partially communicated by R. L. Scott, A. G.
Smith, R. J. Leacock, B. Q. McGimsey, P. L. Edwards, and
K. R. Hackney in "Recent Observations of the Lacertids BL Lac
and OJ 287, and a Possible Lacertid, PKS 0735+17," presented
at the 38th Annual Meeting of the Florida Academy of Sciences,
1974 March 21-23, in Orlando, Florida.
A search for an lo-related Jovian auroral hotspot has
ended with essentially negative results. During the 1973
Jovian apparition, a total of 160 Jovian images were taken at
Rosemary Hill in low-albedo ultraviolet and infrared regions.
Visual examination and computer reductions revealed no trace
of the hotspot. The results were presented by R. L. Scott,
A. G. Smith, M. A. Lynch, S. S. Prasad, P. L. Edwards, R. L.
Hackney, and K. R. Hackney in "A Search for an lo-Related
Jovian Hotspot," 144th Meeting of the American Astronomical
Society, 1974 December 10-13, in Gainesville, Florida.
Quasars: An Astronomical Enigma
More than a decade has passed since Thomas Matthews and
Allan Sandage discovered the first quasi-stellar radio source,
but perhaps no event has had as great an impact upon modern
astronomy. In 1960, using plates taken with the 200-inch
telescope, they identified the radio source 3C 48 with a 16th-
magnitude stellar object, completely starlike except for
traces of surrounding nebulosity. The pecularity of 3C 48
was immediately evident, for no star other than the sun had
ever been found to be a radio source. In addition, its spec-
trum was very different from that of a normal star; it con-
tained a large ultraviolet excess and several broad unknown
emission lines. Several more such objects were soon dis-
covered, but their true significance was not realized until
1963, when Maartin Schmidt recognized that the lines in the
optical spectrum of 3C 273 were the ordinary Balmer lines of
hydrogen, greatly redshifted. Schmidt's discovery enabled
lines in the spectra of the four other such objects then
known to be likewise identified as familiar lines greatly
displaced toward longer wavelengths.
At present, more than ftr hundred such objects have been
identified with known radio sources; they are often called
"quasars," short for quasi-stellar radio sources, or simply
QSO's (quasi-stellar objects). Many radio-quiet QSO's have
also been discovered, on the basis of their colors and red-
shifts. Quasars have confronted astronomers with a most
perplexing problem, for placing them at the cosmological dis-
tances implied by their redshifts necessitates intrinsic
luminosities considerably in excess of that of an entire
galaxy. Their observed variability implies source sizes of
a few light days or years in diameter, and creates severe
luminosity requirements for source emission mechanisms.
There are advocates of non-cosmological redshifts, such as
large Doppler shifts due to ejection from relatively nearby
objects, and redshifts of a gravitational nature. But the
consensus of opinion favors a cosmological interpretation,
and current source models must satisfy cosmological luminosity
Variability of Quasars
The optical variability of the quasar prototype 3C 48
was firmly established by Matthews and Sandage (1963), whose
observations showed a 0 4 change within thirteen months.
Subsequent observations of the other known quasars have shown
that the majority of them are variable at optical and radio
wavelengths. Due to the difficulty in establishing their
identification, little work has been done with radio-quiet
objects. However, those that have been monitored regularly
show the same type of variability as radio-emitting quasars.
The nature and time scale of observed fluctuations vary
widely from object to object. Many quasars exhibit moderate
variations of a few tenths of a magnitude over a time of
weeks or months, then suddenly undergo changes of a magnitude
or more in a few days. Such objects are known as optically
violent variables (OVV's). A classic example is NRAO 512.
In June of 1970, Folsom observed this object to flare 1.2
with a rise time of only two days, and then decline even more
rapidly (Folsom et al. 1970).
Other objects appear to undergo quiescent periods, where
activity ceases for an extended period of time. The optical
counterpart of PKS 1004+13 dropped 0 '3 between 1967 and 1969
according to Hunter and lii (1969). Li (1972) reports a
decline of a magnitude between March of 1969 and May of 1970.
However, no statistically significant variations were observed
from March of 1972 to May of 1973 (Scott et al. 1973).
An apparent periodicity may have been observed in the
activity of the OVV 3C 345. Kinman et al. (1968) reported
flares in the light curve of this object which appeared to
occur in a phase-locked pattern of four, with an average
interval of 80 days between individual outbursts, and a
period of about 320 days for recurrence of the group. HIow-
ever, observations of Lii (1972) and Hackney (1973) suggest
that these outbursts were not permanent features of the
object's light curve.
The optical counterpart of 4C 05.34, which until recently
had the largest known redshift, does not exhibit any activity
of a violent nature, but appears to undergo long-term ampli-
tude variations of approximately 0.1 per year (Hackney et al.
The peculiar object BL Lac is characterized by variabil-
ity of an extremely violent nature. BL Lac is the protytype
of a class of objects now called lacertids, which resemble
quasars but have featureless optical spectra. Day to day
fluctuations on the order of a magnitude were observed by
DuPuy et al. (1969), and other observers have reported simi-
lar activity. Intraday variations greater than 0'n5 in a few
hours were reported by Bertaud et al. (1969), and Weistrop
(1973) reports a decline of 1 2 in 23 minutes. While obser-
vations by Hackney (1973) and Scott et al. (1973) show intra-
day variations of a less extreme nature, the object apparently
spends a large portion of its time in violent activity. BL
Lac is also variable at radio wavelengths, and Hackney et al.
(1972b) report that radio outbursts seem to lag optical
events by several months. Recently, Oke and Gunn (1974) have
succeeded in obtaining a redshift of 0.07 for the faint enve-
lope of this object whose spectrum resembles that of a com-
pact galaxy. The envelope is very faint and is visible only
in long exposures with large instruments.
Studies of the optical variations of quasars provide
many clues as to the nature of these objects. Assuming qua-
sar redshifts are distance indicators, the implied luminosities
combined with the observed r,;pid fluctuations create an
intriguing paradox for astronomers.
The Redshift-Luminosity Paradox
The known range of quasar redshifts runs from 0.06 for
B 234 to the enormous value of 3.53 for the recently measured
OQ 172 (Wampler et al. 1973). Most galaxies have redshifts
greater than 0.001 but less than 0.01, with 3C 295 having the
largest known galaxian redshift of 0.46 (Morrison 1973).
While the quasar and galaxian redshift ranges do overlap,
many quasars have redshifts greater than 1, and the implied
cosmological distances place them far beyond the most distant
galaxy. The resulting intrinsic luminosities, as high as 1047
ergs/sec (Colgate 1969), are more than 100 times those of the
giant elliptical galaxies at optical wavelengths. According
to Terrell (1964,1967), the time required for significant
variations to occur sets an upper limit to the size of the
active region, assuming non-relativistic motion in the source.
Since many quasars have been observed to undergo significant
fluctuations on a time scale of hours or days, a cosmological
interpretation of their redshifts confronts astronomers with
an object of solar system size which has a luminosity much
greater than that of an entire galaxy.
Terrell (1964) attempted to overcome these difficulties
by suggesting that quasars were local objects, being ejected
from the central part of our galaxy at relativistic speeds.
This idea found little acceptance, due to the problems in
imagining such an object bei,;, accelerated to relativistic
velocities while retaining a compact form. Another argument
against the ejection theory is the absence of blue shifts.
Objects moving toward the observer should appear brighter
than those moving away, and in a survey down to a given
apparent magnitude, the blue-shifted objects should predomi-
While no quasars have as yet been found to have blue
shifts, there are certain objects, according to Chiu and
Morrison (1973), which are likely candidates. These objects
are the half-dozen or so lacertids; their flat line free
spectra could well represent a region ordinarily seen in the
infrared. However, this theory remains an interesting specu-
lation at the present, and Oke's measurement of a redshift
for the envelope of BL Lac seems to discourage this hypothe-
sis (Oke 1974).
Attempts to explain the redshifts as due to the effects
of an intense gravitational field on escaping photons, as
predicted by general relativity theory, have also faced diffi-
culties. Greenstein and Schmidt (1964) have shown that in
conventional models the potential gradient should broaden
spectral lines more than observations show. These models
assume the lines arise from a thin shell surrounding a highly
compact object. Hoyle and Fowler (1967) overcame this diffi-
culty by constructing a model where the emitting gas is con-
centrated well in the center of a massive object, which could
consist of a cluster of neutron stars or white dwarfs in
order to be transparent to the line radiation. However,
there are objections to this model on theoretical grounds,
and at present no satisfactory model of this type has been
Observational evidence against cosmological distances
for quasars and related objects has been put forward by Halton
Arp and the Burbidges. Arp (1971) reports the discovery of a
luminous filament connecting the nucleus of the spiral galaxy
NGC 4319 with the quasar Iakarian 205, two objects with quite
different redshifts. Burbidge et al. (1971) report the appar-
ent proximities of four quasars to small-redshift galaxies,
and the existence of a physical connection between the com-
paratively nearly galaxy IC 1746 and the radio-quiet QSO PHL
1226. Burbidge et al. (1972) claim that the blue stellar
object Weedman 2, a possible lacertid, appears to be associ-
ated with the galaxy NGC 2992. In all these cases the QSO and
associated galaxy have different redshifts except for Weedman
2, whose redshift has never been measured due to its feature-
less optical spectrum. This would rule out cosmological red-
shifts for the involved QSOs if the physical connections are
While these apparent associations between galaxies and
quasars are extremely interesting, they are seen in only a
small sample of the entire quasar population. In addition,
Sanitt (1972) warns against i literal interpretation of fila-
ments connecting adjacent images on a photographic plate, as
peculiar photographic effects sometimes occur which give the
appearance of a physical connecting bridge. If these appar-
ent physical relationships could be proven, it would show
that at least some quasars have non-cosmological redshifts,
but widespread acceptance awaits more convincing evidence.
Perhaps the most convincing evidence in favor of a cos-
mological interpretation is the "redshift-angular size" rela-
tionship (Miley 1971). Figure 1 shows the angular separation
at a wavelength of 11.1 centimeters of the double radio com-
ponents of quasars (open circles) and radio galaxies (closed
circles) plotted versus redshift. Because the data represent
the component separation of three-dimensional radio sources
projected on the sky, the solid line is an upper envelope.
As Figure 1 shows, a definite decrease in the angular size of
radio components occurs with increasing redshift. This,
together with the apparent continuity between the angular size
properties of radio galaxies and quasars, suggests that the
origin of the redshifts of the two classes of objects is not
appreciably different, and that the redshifts are actual dis-
Additional support for a cosmological interpretation is
supplied by the investigation of the "magnitude-redshift rela-
tion" for quasars which are most luminous optically in the
redshift range from 0.2 to beyond 2.0 (Bahcall and Hills 1973).
Bahcall and Hills have found that the distribution of
Figure 1. Observed correlation of the redshifts of
quasars and radio galaxies, with angular
size. On this log scale plot, the solid
line is a best fit to the envelope of all
the points. The quasars are represented
by open circles; the radio galaxies by
closed circles. After Miley (1971).
0 RADIO GALAXY
0 0 O0 O
* 0 00
0 0 0
00 00 C
0 0 O
"""" ....L~._ ~......~.. _-...u
I I I I I I
redshifts relative to magnitudes is consistent with quasars
being at the cosmological distances implied by their red-
The nature of quasar redshifts is still unresolved, but
the preponderance of opinion favors a cosmological interpre-
tation. Even if a few quasars do prove to be associated with
galaxies, and have redshifts of a non-cosmological origin,
the majority of these objects appear to be very distant, and
participating in the general expansion of the universe.
Quasars and Related Objects
Seyfert galaxies, named for Carl Seyfert, who first
studied them in 1943, are compact spirals with a bright star-
like nucleus which is far more luminous than the nucleus of
an average galaxy. They represent about 1 percent of the
total galaxy population. Both quasars and the nuclei of Sey-
fert galaxies have a large ultraviolet excess, and appear
very blue in photographs. Seyfert galaxies are the most lumi-
nous of the spiral galaxies, and many of them have radio
power outputs in the range of average radio galaxies, while
others emit no detectable radio emissions. Like quasars,
they emit the largest portion of their energy in the infrared.
The most luminous galaxies at optical wavelengths are
N-type galaxies, which appear similar to Seyferts, except
that their envelopes are smaller and more amorphous with no
well-defined spiral structure. The ultraviolet excess and
infrared emissions of their nuclei are even greater than
those of Seyfert galaxies, and they are among the most power-
ful radio emitters.
Except for their surrounding galactic envelopes, Seyfert
and N-type nuclei appear strikingly similar to quasars. All
three classes of objects produce starlike images on a photo-
graphic plate, and have spectra characterized by broad emis-
sion lines up to 100 A in width, superimposed on a blue con-
tinuum. While the spectra of Seyfert and N-type galaxies are
identical, there are two minor differences between their
spectra and those of quasars. The 0 II emission line at 3727
X is faint or missing in quasar spectra (Burbidge 1968), and
the spectra of some quasars contain narrow absorption lines,
especially around z = 1.95.
Predominantly, the typical quasar is a non-thermal source
whose emission-line intensity corresponds to 10 percent of the
total optical intensity. The optical energy distribution of
quasars may be described by a typical synchrotron electron
continuum, strongly polarized, with a power law intensity
shape described by
I c -n (1.1)
where I is the observed spectral intensity, v is the fre-
quency, and n is the spectral index (Morrison 1973). This
form of spectral energy distribution also describes the opti-
cal continuum spectra of Seyfert and N-type galaxies. Figure
2 shows the spectra of the Seyfert galaxies NGC 1275, NGC
1068, and 3C 120, compared with the quasar 3C 273 (Colgate
The continuum spectra of three Seyfert
galaxies, NGC 1275, NGC 1068, and 3C 120,
and the quasar 3C 273. Large variations
are known to exist in the frequency range
109 to 1011 Hz, so little weight should
be placed union this portion of the curves.
After Colgate (1969). Reproduced by per-
mission of Physics Today.
I I I
I8 108 012
10 10E0 1012
1 I 1 .
1969). At optical wavelen.' all resemble a typical single
peaked synchrotron spectrum.
The optical variability observed for the nuclei of Sey-
fert and N-type galaxies is similar in character to that of
quasars though in general of lesser amplitude. An exception
is the Seyfert galaxy 3C 120, which was observed by Folsom
et al. (1971) to undergo fluctuations of a magnitude or more
on a time scale of days or weeks. The Florida observations
of this object also show considerable activity at infrared
wavelengths as discussed in Chapter V.
On a color-color diagram, all three classes of objects
lie well above the main sequence, with the region occupied by
quasars and Seyfert galaxies overlapping to some extent.
This is shown in Figure 3. Color studies by Hackney (1973)
and others show that in general these objects all become pro-
gressively bluer as they brighten optically, and redder as
they decline. Such observations, together with the observed
similarities in spectral energy distribution and variability,
suggest that quasars and the nuclei of Seyfert and N-type
galaxies have similar mechanisms for the release of energy,
the process being less violent in the latter objects.
Table 1 is a list of observational similarities between
quasars, Seyfert galaxies, and N-type galaxies, compiled by
Colgate (1969). These observations, and those previously
mentioned, suggest a relationship between classes of objects.
If the redshifts of quasars are indeed cosmological, then the
majority of them are seen at a very early epoch in the age of
Figure 3. Color-color diagram for quasars (Q), Seyfert
galaxies (S), and N-type galaxies (N). The
curved line represents the main sequence
(MS). After Burbidge and Iloyle (1966) and
I I -- I I I I
I I I I
0 UU ? >,
*- 0 0 ct r r c
Q0 C) a 2 0 o
) C.1 0 H1
vn v, O \ L
,H 0 0 0 0 ~ U
0 0-4- 4-)
+j,- ', 41 Q) 7: 3; >. n>.
S-3 OkO H H OQ u
m C CD 0 "I C-0 C I
,-H Ct m + o c (t
t) 4o U u- 0 0 -H bl) ) 0C O 4- 4-
Ct t ed V) U 0 CD 0 Ct A U) O-
G) -} 4 5 ( -D
nCt oU) -n U) 00 0
4-J 3 ) U) UO C.) 0 D
0 O : 0- 0 0
m ct ri 0 4J j ., u
( >,0 Lfl \ .H 'H ;0 .
Q 4- Ct -It t- 4-) r Ct
!-4 'HI .' CD CD 0 bU 0 j 0n
0 j ct M
0 H 4-0 CD D
r-l ?-l ::, U) 't 4 U 0 4- t C LC 0
Ct 0t C16 Ln 0f ct
C "- c CD n 0 0 c d
SHd n o
nCt 0o -S 0 0
E rl 4 l- n L ) O
0 U) +U) U)
U)) 4-4 4-J 4)
4- M i0 bO 0 W, of 3
c'' 0 + CD r-C
-H 0 0
0U 4> +-i Ct l] i T-A 0
0 r- r-t U) 0 c t to 4 0- (
0 0 0 0 D E .C4 N U) L (
H OE ^ f i 0 0 o-*He r 0
4 4 ct *H 0 U r t N .) 'H 0 u
Ct -H 4 ;- + f 4-J S U C4-) C 4-) 0-
(f EU -O*o QU o U ct ag
0 U)Q 0 'H 0)
0 E 4- 4-0 0 4-) o c0 t 00 0I
U H 0 0 0 U .H c 0 Ct *H *H H
r-4 0 US US 4-- 0 'H U Uo 0 Ct u
'- 0 0 U -l 0 4-- -1 0 0 0
71 Q3 P 0 o llzU r-HO rHl 0 'H 'H 4-H
r0 U-q U)C U) C ) 0 -1 ,-1 44- l !-< r-1 -PL 4-1 4-J 0
C' 0 0 Ct C t 4-' 0) Ct U) $2-0- ;i
ct u 0 0 Ct u u o 04-i Uo 0 0 .4
-I H *H 0 H 0 .- H H H r-O c H H' 0 0
S4-) I : Ct Ct -4- 4-J 4-) '-1U) ') C0 ) 4-J 'r -, uV J)
0 0 CO 4 C14 4-) 0 O0 C t: t 0 Ct Ct 4
the universe. This implication, together with the apparent
similarities between classes of objects, suggests an evolu-
tionary relationship, starting with quasars and ending with
galaxies such as our own. Seyferts and N-types are inter-
preted as intermediate stages. Such a relationship would
imply a gradation of average properties relative to distance
and retarded time. Kleinmann and Low (1970) suggest that the
spectral energy distribution for all galaxies is very similar,
excluding variability, lending further weight to the evolu-
Observations of quasars, Seyfert galaxies, and N-type
galaxies have suggested various source models. Since evidence
favors cosmological distances for most quasars, models must
not only explain the similarities between classes of objects,
but also must conform to cosmological luminosity requirements.
While no current quasar model completely satisfies all obser-
vational requirements, a few of the more interesting are now
Colgate (1969) has proposed that the observed character-
istics of quasars may be explained by a model in which a
dense cluster of stars coalesces to form stars of 50 M@,
which rapidly proceed to the supernova state. A supernova
rate of 5 per year yields a peak luminosity of 1047 ergs/sec,
in good agreement with that calculated for quasars. Colli-
sional excitation of expanding supernovae envelopes gives
rise to the rapid fluctuations of the continuum radiation.
The optical lines, arising from the excitation of the whole
gas cloud, would not be subject to these fluctuations, in
agreement with observations. Colgate argues that the large(
infrared flux observed for quasars could be due to the exci-
tation of a surrounding dust cloud by intense ultraviolet
radiation. Since this does not explain the rapid variations
observed at infrared and millimeter wavelengths, he proposes
an alternate to the dust cloud theory, in which the high-
energy fraction of the supernovae ejecta excites a weak two-
stream instability in the ionized gas cloud, the infrared and
millimeter emissions being due to plasma oscillations. A
scaled-down version of this model is used by Colgate to ex-
plain the energy released in the nuclei of Seyfert galaxies.
This theory, proposed by Cavaliere et al. (1970) and
Morrison (1973), imagines a pulsar-like mechanism operating
in quasars, consisting of a rapidly rotating central "spinar"
having a mass of 10 M with an attached coherent magnetic
field. The rotating central body, well under one-tenth of a
light year in size, has an extermely high peripheral speed
of perhaps 0.01 to 0.1 of the velocity of light. Using the
pulsar mechanism, such an object can efficiently eject parti-
cles at relativistic velocities. The high infrared flux is
explained as synchrotron emission, occurring when the parti-
cles interact with the innermost portion of the magnetic
field. The optical and x-ray continue arise from Compton-
recoil emission, due to curvature of the magnetic field lines,
as the particles progress outward. Moving outward, clouds of
expanding particles give rise to the radio frequency emission.
Hot and cool gas filaments give rise to the optical lines.
Low (1970) envisions clusters of identical infrared
sources called irtrons to be characteristic of the central
regions of quasars and the nuclei of all galaxies. Continu-
ous creation of matter and antimatter within the irtrons, and
the subsequent annihilation, convert a large portion of the
created mass directly into relativistic particles, which radi-
ate by the synchrotron process when interacting with a mag-
netic field of about 100 gauss. Such a process creates a
peak in the flux distribution for 100 MeV electrons at approx-
imately 75 microns, which agrees well with observations.
Luminosity fluctuations are explained as interactions between
irtrons, or between irtrons and stars, which affect the in-
jection of particles into the magnetic field. As an object
ages, the number of irtrons diminishes and activity decreases.
Low suggests an evolutionary chain, starting with quasars and
ending with galaxies similar to our own.
Gravitational Lens Iypothesi
A somewhat different approach is presented by Barnothy
and Barnothy (1968). They theorize that quasars and N-type
galaxies may not be real objects, but rather the images of
Seyfert galaxy nuclei, amplified by gravitational lenses in
the form of intervening galaxies. In the case of a quasar,
the deflector galaxy is very distant and invisible, but in
N-type galaxies it is close enough so that part of it can be
seen surrounding the brilliant nucleus, which is the image
of the Seyfert galaxy behind it. The observed emission lines
could originate either in the object or the deflector, or
both. Variability is explained as a combination of intrinsic
brightness fluctuations of the object, combined with the
changing aspect between the object and lens, due to their
differential velocity across the line of sight. This some-
what novel theory has received little support, the majority
of opinion being that quasars are indeed a distinct class of
The Florida Program
In 1968 a program for photographic monitoring of the
optical variability of quasars was initiated by A. G. Smith
and G. HI. Folsom, using the University of Florida's newly
installed 30-inch Tinsley reflector. K. R. Hackney and R. L.
Hackney have contributed extensively to the expansion and con-
tinuation of this program. At present, observations and data
reduction are carried out bi A. G. Smith, R. J. Leacock, B.
Q. McGimsey, P. L. Edwards, and the author. Through the
efforts of these observers, and through cooperation with
other observatories, the Florida program has amassed consid-
erable knowledge about the nature of quasar variability.
Over 160 objects are now being monitored; the majority are
quasars, but Seyfert and N-type galaxies, and lacertids are
The author has selectively monitored several of the more
active sources in an attempt to learn more about these par-
ticular objects. In addition, to study apparent similarities
between classes of objects and possible evolutionary relation-
ships, the average properties of representative classes of
objects have been compared, and the observed properties of
several quasars have been compared with their redshifts.
Various photographic techniques which are directly or in-
directly related to the author's studies have been investi-
Jupiter, Another Active Radio Source
While the majority of the author's research is concerned
with extragalactic radio sources, as a further exercise in
photographic techniques a study was made of another active
radio source, the planet Jupiter.
Since its discovery in 1955 by Kenneth Franklin and
Bernard Burke, the decametric radio emission of Jupiter has
intrigued astronomers with lihe problem of understanding the
origin of these energetic radio bursts. Histograms plotting
the probability of receiving emissions as a function of cen-
tral meridian longitude show three well-defined peaks or
"sources," which are illustrated in Figure 4 (Smith 1969).
In 1964 the Australian statistician E. K. Bigg showed
that the position of the satellite Io in its orbit had a
definite effect on the probability of reception (Bigg 1964).
Bigg found that when Io was at 900 and 240 from superior
geocentric conjunction, the probability of reception was
definitely enhanced. When Io is at 90, source B is activated,
and when lo reaches 2400, sources A and C are activated. In
addition, source A emits some radiation which is apparently
unrelated to lo.
Thus far, no theory has been able to account for all the
observed phenomena. One of the most interesting theories
proposed to explain the role of Io is that of Goldreich and
Lynden-Bell (1969). There have been several variations of
this theory nut forth in recent years, but the general idea
remains approximately the same. Goldreich and Lynden-Bell
theorize that To may be considered a unipolar generator which
develops an emf of 7 x 105 volts across its radial diameter,
as seen from a frame of reference rotating with Jupiter. It
is assumed that lo's orbit is circular, that the Jovian mag-
netic field has permeated Io, and that o1 always keeps the
same side toward Jupiter. Assuming further that Io is a
fairly good conductor, at least as good as the upper mantle
Figure 4. A polar diagram of Jupiter's decametric radio
sources. The radial coordinate is the proba-
bility of receiving Jovian emissions. After
Smith (1969) .
of the earth, the electric Iield will have to vanish inside
the tube of flux enclosing the satellite. Any plasma en-
closed by the flux tube will be frozen to lo, and as Jupiter
rotates the "feet" or intersections of the flux tube with the
Jovian ionosphere will slip. This induces a voltage of
approximately 700,000 volts, which drives a current of a
million amperes across each foot of the flux tube. The cur-
rent is assumed to flow up one side of the tube, cross the
magnetic field in lo, and then flow back down the other side
of the tube. This current circuit is illustrated in Figure 5.
Instabilities in these current sheets are theorized as being
responsible for the lo-induced decametric bursts. The geom-
etry of the beaming of the bursts suggests coherent cyclotron
radiation as the emission mechanism, the coherence being pro-
duced by bunching of the electrons in the current sheets.
When energetic charged particles from the sun stream
along the earth's magnetic lines of force into the earth's
ionosphere, they often cause intense auroral emissions.
Similar emissions may radiate from the feet of the proposed
Jovian flux tube, and if observed could provide substantial
support for the Goldreich and Lynden-Bell theory. Photo-
graphic detection of this "auroral hotspot," as it is some-
times called, was the purpose of this study.
Since it was felt that this project could be carried
out most advantageously at the Cassegrain focus of the 30-
inch reflector, a Cassegrain camera was designed by the
author and A. G. Smith. This camera was designed with a
JUPITER FEET Io
Current sheets in the Io flux tube (not to
scale). After Goldreich and Lynden-Bell (1969).
variety of uses in mind; bce;ides planetary photography, it
has been used to monitor one of the brighter active quasars
on moonlit nights, and it is currently being used by B. Q.
McGimsey in polarization studies of quasars. The Cassegrain
camera is described in Chapter II, together with the observa-
tional techniques used in the remainder of the program.
RESEARCH TECHNIQUES AND INSTRUMENTATION
Photographic Studies of Quasars
The University of Florida's 30-inch aperture Tinsley
reflector is well suited for a program of optical monitoring
of quasars. This instrument, located at Rosemary Hill Obser-
vatory near Bronson, Florida, may be used in either an f/4
Newtonian or f/16 Cassegrain focal ratio by exchanging secon-
dary mirrors. Figure 6 shows the 30-inch reflector and the
Newtonian and Cassegrain cameras. Figure 7, taken on the
evening of January 10, 1974, shows Comet Kohoutek beside the
At present, most program objects are photographed with
the 30-inch reflector. However, the University's recently
acquired 18-inch f/10.3 Ritchey-Chr6tien telescope has proven
useful for monitoring several of the program's brighter
sources. This telescope, which was acquired and refurbished
under the guidance of Professor John Oliver, was used during
the summer of 1974 by visiting professors Karen and Richard
Hackney of Western Kentucky University to study short-term
variations of lacertids. These data have also contributed
significantly to the author's own studies. The 18-inch reflec-
tor is shown in Figure 8; its dome is located about 100 yards
north of the 30-inch dome.
Figure 6. The Rosemary lill Observatory 30-inch Newton-
ian/Cassegrain Tinsley reflector, with both
Newtonian (N) and Cassegrain (C) cameras
Comet Kohoutek in the brilliant zodical light,
as seen from Rosemary Hill Observatory on the
evening of January 10, 1974, at 2030 EDT.
This print was made from a contrast enhanced
internegative. The original negative was
taken with a 50 mm f/1.4 Nikkor lens using
an exposure of 30 seconds on Tri-X Pan,
which was force developed for added speed
Figure 8. The Rosemary Hill Observatory 18-inch Ritchey-
Chr6tien telescope showing offset guiding
telescope (T) and camera (C). The camera bed
is similar to the camera beds used at the
Newtonian and Cassegrain positions of the
30-inch telescope, and will accept the same
plate cassettes as the 30-inch cameras.
The majority of the pro gram objects are between 15th and
19th magnitude in the spectral regions studied, which is
beyond the practical limit for photoelectric observations
with the 30-inch telescope, this being about 14th magnitude
due to the limitations of aperture and sky brightness. The
technique used is that of photographic photometry; a sky-
limited photographic magnitude of 20 to 20.5 can be reached
in approximately 15 to 20 minutes at the Newtonian focus, on
unfiltered Kodak 103a-0 plates. When hypersensitized by
nitrogen baking, as described in Chapter III, such plates can
reach the Newtonian limiting magnitude in 10 minutes or less,
and they have enabled satisfactory exposures to be made at
the slower Cassegrain focus, making possible polarization
studies and occasional monitoring of bright sources during
periods of moonlight. However, the majority of program objects
are photographed at the fast Newtonian focus.
For regular monitoring of many sources, photographic
techniques have a further advantage over photoelectric photom-
etry. Studies by Hackney (1973) indicate that on the average
66 percent of the 12 nights per month allotted to quasar
monitoring are useful photographically, while only 35 percent
of the nights are useful photoelectrically.
Detailed descriptions of the telescope, Newtonian camera,
observing procedures, and data reduction are given by Folsom
(1970), R. L. Hackney (1972a), and K. R. Hackney (1973). In
general, objects are photogri;hed at the Newtonian focus of
the 30-inch telescope; plates are exposed in sealed cassettes
charged with dry nitrogen to prevent speed losses due to the
excessive humidity of the Florida air. The observer care-
fully guides the telescope during the exposure, to ensure
images of usable quality. Plates are developed immediately,
at the observatory, in a darkroom set up for that purpose.
The plates are reduced on a Cuffey iris astrophotometer; an
on-line Hewlett-Packard 9810 A calculator fits a least-squares
calibration curve to the iris readings and magnitudes of a
series of comparison stars, and calculates the magnitude of
the object. Comparison star magnitudes are smoothed for
local field effects with an IBM 370/165; the computer is also
used to perform chi-squared tests on the data to determine
the confidence level for variability (Hackney 1973).
Most program objects are photographed on unfiltered 103a-0
plates, yielding a photographic magnitude m For studying
color changes in a select group of objects, K. R. Hackney
used a photographic UBV system which closely approximates the
U, B, and V magnitudes of the Johnson-Morgan system (Stock
and Williams 1966). A Kodak IIa-0 emulsion is combined with
Schott UG-2 and GG-13 filters, to yield photographic U and B
magnitudes, respectively. A IIa-D emulsion and a Schott GG-14
filter define the photographic V magnitude. The system
response functions of the Rosemary Hill UBV system are
described by K. R. Hackney (1973). The IIa-O and IIa-D
plates were chosen initially because it was felt their fine-
grained emulsions would contribute to photometric precision.
Both plates are baked in dry nitrogen, approximately doubling
their speed. Investigations by the author and A. G. Smith,
described in Chapter III, have shown that the faster, coarser-
grained 103a-0 may be substituted for the IIa-0, with a small
loss in photometric precision unimportant in routine monitor-
The author has used the above UBV system to compare the
color indices of various classes of objects, and to investi-
gate the relation between the color indices of quasars and
their redshifts. In addition, an infrared system, also
described by Stock and Williams (1966), has been used in
these studies. Kodak I-N plates combined with a Schott RG-8
filter define the infrared magnitude system. Figure 9 shows
the normalized infrared system response function. The I-N
emulsion is inherently slow; it must be hypersensitized at
the observatory and used within a few hours to obtain suffi-
cient speed for practical use. Using the ammonia hypering
procedures described by Barker (1968) and Babcock et al.
(1974), sufficient sensitivity has been obtained to reach a
limiting infrared magnitude of about 16 with an exposure of
25 minutes at the Newtonian focus.
The infrared system response function used in
the QSO studies, obtained by combining the
Kodak type I-N emulsion with a Schott RG-8
infrared filter. The curve shows the normal-
ized relative response RN versus wavelength,
and was obtained by a point-by-point multi-
plication of the plate sensitivity and percent
transmission of the filter. The spectral
response of the I-N emulsion was obtained
from the booklet "Kodak Plates and Films for
Scientific Photography," Eastman Kodak Com-
pany, 1973. The transmission characteristics
of the RG-8 Filter were obtained from the
Schott "Color Filter Catalog No. 365e."
A Photographic Attempt to Detect
the Jovian "Hotspot"
The Cassegrain Camera
The Rosemary Hill Observatory Cassegrain camera was
designed by the author and A. G. Smith in the summer of 1971.
The camera was built by Eli Graves, of the Department of
Physics and Astronomy research shop. Though initially con-
ceived with a variety of uses in mind, the major goal at the
time of construction was to obtain a convenient image size
for the author's photographic study of Jupiter.
Assuming an apparent Jovian diameter of 35 arcsec, the
image diameter at the f/4 Newtonian focus will be approximately
1/50 of an inch. It was felt this image size should be in-
creased relative to plate grain in order to obtain adequate
resolution. Also, a larger image would aid in the reduction
of the plates, yielding a larger area for more convenient
The image size may be boosted by a factor of 4 by working
at the f/16 Cassegrain focus. In addition, larger increases
can be obtained by eyepiece projection, a standard procedure
for increasing image size in planetary photography.
Figure 10 shows the various parts of the Cassegrain
camera. Essentially, the camera consists of a flat baseplate
connected to a camera bed by a spacing tube designed to align
the focal plane of the camera with the f/16 Cassegrain focal
plane of the 30-inch reflector. The camera bed is similar to
that of the Newtonian camera, and will accept the sealed
The Cassegrain camera mounted on the guidebox
of the 30-inch reflector showing offset
guider (G), counterweights (W), base tube
(T), rotation clamps (R), eyepiece projection
tube (E) filter holder (11) shutter (S),
with cable release (C), and camera bed (B),
with dark slide (D).
cassettes used in the quasar program if desired. The bed may
be rotated to any desired position angle relative to the
guidebox by loosening the rotation clamps.
A shutter is provided for exposures of short duration.
It can be set at 1/2 sec, 1/5 sec, 1/10 sec, 1/25 sec, 1/50
sec, "time" and "bulb." For the Jupiter project, where most
of the exposures were seconds or sometimes minutes, the
shutter was set on "bulb" and the exposure duration was con-
trolled by the observer, using a cable release. For very
long exposures, the shutter is set on "time" and opened, and
the exposure is controlled with the dark slide.
After removal of the photoelectric photometer, the camera
bolts directly to the Cassegrain guidebox of the 30-inch re-
flector as shown in Figure 10. A series of six counterweights
are used to compensate for the 80-lb weight differential
between the camera and the photometer. The existing Casse-
grain offset guiding mechanism, with slight modification by
A. G. Smith, was found to be satisfactory for the guiding of
An eyepiece projection tube accepting standard 1-1/4-inch
outside-diameter eyepieces can be used if image enlargement
is desired over that provided at f/16. With the camera bed
unscrewed, the projection tube screws directly onto the base
tube; the camera bed is then screwed onto the end of the pro-
jection tube. The eyepiece projection tube is shown in Fig-
ure 10. The camera is constructed mostly of aluminum; excep-
tions are the stainless steel dark slide, the steel shutter
(from a commercial source), and the eyepiece holder and filter
slide for the projection tube, both of which are made of brass.
A filter slide was constructed for the eyepiece projec-
tion tube, to allow exposures to be taken through standard
tricolor filters in rapid succession. This has enabled the
author to construct several tricolor separation prints of
Jupiter; Figure 11 is an example. The color print in Figure
11 is a composite of the three black and white exposures taken
on fine-grained Kodak Contrast Process Pan film through a red
#25, blue #47 B, and green #58 filter, respectively.
To produce such a color print, each of the three black
and white negatives is projected one at a time upon ordinary
color print paper. Each negative is projected through the
same filter used to produce it at the telescope. Accurate
superposition is accomplished by use of a cardboard mask,
upon which details of the Jovian image are traced with extreme
care. With the mask covering the print paper, a negative is
carefully aligned with the tracing on the mask and the proper
filter is placed in the enlarger. The mask is then removed
and the exposure is made. Then the mask is carefully replaced
and another negative is aligned with it. When all three nega-
tives have been exposed through their filters, the paper is
developed in its usual manner.
Tricolor separation is a standard method used to produce
color pictures of planets and, occasionally, celestial objects
too faint to photograph successfully on color films without
using cooled emulsion techniques. Such a procedure produces
Tricolor separation print of Jupiter from
negatives taken on July 29, 1972. A print
from the red negative is at upper right, the
blue at lower left, and the green at lower
right. The red and green exposures were
each 25 seconds, the blue 70 seconds. A 25
mm eyepiece was used to project the primary
Cassegrain image onto the film plane yielding
an equivalent focal length of f/64 or 160
feet. All exposures were made on Kodak
Contrast Process Pan 4" x 5" sheet film. The
three negatives were superimposed through the
corresponding filters to produce the Unicolor
print. In the color print the Great Red
Spot predominates, along with a double equa-
torial band crossed with white wisps, and a
white spot to the right of the red spot.
South is up.
a sharper print than a single exposure, due to suppression
of emulsion grain and seeing effects. By manipulating the
relative exposures of the tricolor negatives, the desired
results may usually be obtained after a little experimentation.
Plates and Filters
According to Goldreich and Lynden-Bell, the intersection
of the Io-linked flux tube with Jupiter's ionosphere has an
elliptical cross section, with a major axis of 230 km and a
minor axis of 120 km. If a Jovian angular diameter of 35
arcsec is assumed, the angular diameter of the foot of the
flux tube is approximately 0.05 arcsec. The theoretical
resolving power of the 30-inch reflector is 0.15 arcsec, and
seeing usually limits the resolution to at least 2 arcsec.
However, the real difficulty in detection of the spot is the
interference from reflected sunlight. Unresolved stellar
images are easily seen or photographed because of their great
brightness relative to the sky background. Since Jupiter
exhibits considerable limb darkening, confining the search
for the hotspot to those times when it should be near the
Jovian limb would lessen interference from reflected sunlight.
Pilcher and McCord (1971) show the albedo of Jupiter to
be at a minimum between 0.3 and 0.4 microns, and between 0.95
and 1.05 microns. Their curve is shown in Figure 12. Since
interference from reflected sunlight would be minimized,
these spectral regions are well suited for an attempt at
detecting the hotspot, especially the infrared region.
Spectral reflectivity of Jupiter. After
Pilcher and McCord (1971). Pilcher and
McCord's original albedo curve was normal-
ized to unity at 0.56p1; the normalized
albedos have been replaced with estimates
of the true albedo obtained from Danielson
0.5 0.7 0.9
Kodak 103a-0 plates, combined with a Kodak i18 A filter,
approximate the above ultraviolet region. Kodak I-Z plates,
combined with a Schott RG-8 or Corning 7-56 filter, will give
broad and narrow band coverage of the above infrared region
respectively. The normalized system response Functions of
these plate and filter combinations are shown in Figures 13
To obtain maximum speeds and reduce exposure times, the
103a-0 plates were nitrogen baked. The I-Z plates, like most
infrared emulsions, must be hypersensitized before use. They
were treated with ammonia, following a procedure described by
Pope and Kirby (1967).
Choosing Observation Times
A computer-generated ephemeris, programmed by G. R. Lebo
and W. W. Richardson, is used by the University of Florida
Radio Observatory in predicting lo-related Jovian radio
storms. This ephemeris also gives the geocentric longitude
of lo, Europa, and Ganymede for each hour of Universal Time,
for each day of the year. The longitude is computed relative
to superior geocentric conjunction, the point at which the
satellite is farthest from earth in its orbit. This ephemeris
was used to select optimum times for photographing the hotspot,
by determining what times Jo, and thus the hotspot, would be
near the Jovian limb.
Goldreich and Lynden-Bell suggest that the spot may lead
Io in longitude by as much as 12', depending upon the conduc-
tivity of the Jovian ionosphere. Such a lead should be at a
The ultraviolet response function used in the
Jupiter studies, obtained by combining the
Kodak type 103a-0 emulsion with a Kodak #18 A
filter. The curve shows the normalized rela-
tive response RN versus wavelength, and was
obtained by a point-by-point multiplication
of the plate sensitivity and percent trans-
mission of the filter. The spectral response
of the 103a-0 emulsion was obtained from the
booklet "Kodak Plates and Films for Scien-
tific Photography," Eastman Kodak Company,
1973. The transmittance of the #18 A filter
was obtained from the booklet "Kodak Filters
for Scientific and Technical Uses," Eastman
Kodak Company, 1970.
0CiD ,D 4- 4-J C- r : LL CT Lr)
7U4-) 1 ::4-) rd *H A ~-C C3
.4-JO U-0o 4-0 C 0 0 -
4-J g--0 H 0 -1
m 0f c o m-~
4J Cd 0 V t") 4-J r( 0
.1 4 ( ), 4-) (
CD f4 '-04 4CD-1 -) 0 0. 4-
H C,6 Ct 1V1)4 U mD\DU
5-- C, -cn 4 ctc
4-) 0c 4-~) ct C4Lf
D4-)U 0 F4
0 D D H,- 02 U4 0o D
d flr OOCCDO U
( H) 0 i CD 0 :)
7 l Q) V)kTn U Mv
a CD04-)l 4 0 Q) 011)
0 Q)4-) ---1 0 -o0 $-j
4-3 0 U) )c -I Q) o-0 4-
U 4- r -C > 0 CD C 4-k 0 4- -1
;1 '- H 4- 4-J V) 4-) -H
7L4 (DZ -Jr, 4tfm.H L,
ct () uo 4 t iqr: m F
cd) o 0ctoH ;: n0
CD tr-9 cD 'J C) OrH0
0 0 D 00, -44- ` DI- 4 0-
o-I -r-' 0 0' U :44) ()r4 ~-A
1) 4J0 U r-H- 4-HOO.HCDU )
S-4bLU -4 r4 c, b, E-c
C7 F04-j 0 0 -1
-0 Q4 ) 4 0 0 4-J 1 C 3
4-) r-4 ., -I v) 4 :
U) 0 73 0 4-)MU, 04-1
P n P4 7 U 4-J P .H 4J 0 0 C
0 0 0 r- 4-'1 V
P 4-4 LH -H rH C C CD
---CD H- 0 4-) 0 H U 0L4C-
r-5 4 PrH1 05 U F:0 4-
r-4 C) C) t 0 Q D I 0' --P
cl, Ap j t ro t;
Ej-, -0 -H 5.pm -i C'U6
cF- D 4-4(4-
'HEe c ~l cd a O a
10 0 In
minimum as the spot emerges From the terminator on the east
limb of Jupiter, and at a maximum as the spot approaches the
western limb. Assuming a 120 lead, the spot will be on the
east limb when Jo has a geocentric longitude of 780, and on
the west limb when lo is at longitude 2580. Because of the
uncertainty in the lead in longitude of the hotspot relative
to lo, "hotspot events" were considered to start when Io was
near 780 or 2580, and to end when Io reached 90 or 270,
respectively. Since lo moves about 8.4 per hour in its
orbit, such an event lasts about 1-1/2 hours. If possible,
Jupiter was photographed continuously during these times.
Due to the short exposures required for this study it
was not felt necessary to use the sealed quartz-window cas-
settes used in the quasar program. Instead, nonsealed cas-
settes containing no optical windows were used; they can be
unloaded and loaded quickly, making it possible to take a
series of exposures in rapid succession.
At the start of observing sessions, the camera was given
a careful knife-edge focus. Focus plates have shown that
there is no significant change in focus introduced by the
ultraviolet and infrared filters. After focusing, the wide
field Erfle eyepiece used in the quasar program was inserted,
and Jupiter was placed near the "top" of the field. Then the
eyepiece was removed and the desired filter was placed in the
camera. A series of exposures was now taken; they were
spaced by sliding the cassette relative to a scale on the
side of the camera bed. Usually five or six exposures were
taken in such a series. Then the eyepiece was replaced,
Jupiter placed in the center or "bottom" of the field, and
another series of exposures was taken. In this manner,
several rows of Jovian images could be placed on a single
plate. The starting time of each exposure was determined to
the nearest second with WWV and this, together with the dura-
tion of each exposure, was recorded in a logbook.
Since the expected brightness of the hypothetical spot
is known only roughly at best, exposures were bracketed in
such a manner as to range from images that showed detail in
the Jovian cloud features to burned-in disks which showed no
surface features at all. Because of the slowness of the I-Z
emulsion, infrared exposures were taken at f/16, and required
exposure times of nearly a minute to several minutes in length.
Using the RG-8 filter, infrared exposures of about 40 seconds
showed the best detail in Jovian cloud features. Longer
exposures of several minutes showed little detail and were
often surrounded by a halo. When using the 7-56 filter,
exposures of approximately 2 minutes showed the best Jovian
cloud detail, while exposures of up to 5 minutes could be
made without burning in the planetary disk. The longer expo-
sures were guided with the Cassegrain offset guiding mechanism.
The fast 103a-0 plates yielded satisfactory ultraviolet
images with exposures of only a few seconds, and occasionally
eyepiece projection was used to increase the size of the
image. At f/16, ultraviolet exposures of 1 to 1/2 seconds
showed the best cloud detail, while exposures of 5 seconds
and greater produced totally burned-in images with no visible
detail. The reduction of the plates is discussed in Chapter
NITROGEN BAKING AND EMULSION COOLING
AS METHODS OF HYPERSENSITIZATION
Two hypersensitization techniques are discussed in this
chapter: the effect of nitrogen baking on the Kodak type
103a-0 spectroscopic emulsion, and the speed gains obtained
by cooling various emulsions. The nitrogen baking of the
103a-0 was found to be a practical aid in the routine monitor-
ing of QSO's at Rosemary 1ill; a discussion of this technique
was published in the Astronomical Journal (Scott and Smith
The speed gains obtained by nitrogen baking or emulsion
cooling apparently arise due to a lessening of "low intensity
reciprocity failure." According to the reciprocity law for
photochemical reactions, the product of a photochemical reac-
tion is dependent upon the total energy involved (Eastman
Kodak 1973). Applying the reciprocity law to photographic
emulsions would imply that the density of the image formed
is dependent upon the energy of the exposure; the energy may
be expressed as the product of the intensity of the exposure
and the exposure duration. For most photographic materials
the density produced by a long exposure of low intensity will
be less than that produced Ib a shorter exposure of higher
intensity, even if the total energy in the two cases is the
same. This is an example of low intensity reciprocity fail-
ure, and it shows that image density is dependent upon both
intensity and time of exposure.
The currently accepted theory of latent image formation
is that of Gurney and Mott (1938). The typical photographic
emulsion is composed of silver halides, usually a mixture of
silver bromide and silver iodide, suspended in gelatin. The
latent image is assumed to be formed of metallic silver and
arises from the combination of photoelectrons and silver ions,
which forms a speck of metallic silver on or in the silver
halide crystal. The latent image photolysis reactions for a
silver bromide emulsion are given below (Stock and Williams
Br + hv + Br + e (3.1)
e + Ag+ Ag (3.2)
where hv is a quantum of radiation and e an electron. At
low intensity levels many of the liberated photoelectrons are
believed to recombine with positive holes before a stable
latent image can be formed; this is the suggested explanation
for low intensity reciprocity failure.
Kodak spectroscopic plates of "a" designation, such as
IIIa-J or 103a-0, are chemically treated by Eastman Kodak
during manufacture to partially compensate for low intensity
reciprocity failure. Nitrogen baking can usually be used to
hypersensitizc "a" plates; the baking presumably extends the
chemical sensitization initiated by Kodak, reducing the
probability that photoelectrons will recombine with positive
holes. Oxygen and water are also removed by baking; their
presence in the emulsion is also believed to hinder latent
image formation. Though impressive speed gains may be ob-
tained by baking, they are accompanied by an increase in
background fog level. The shelf life of baked materials is
also greatly reduced unless stored at a low temperature in
nitrogen. It should be pointed out that it has long been a
standard practice at many observatories to bake plates in
air; the current more efficient practice of baking in nitro-
gen was pioneered by Dr. A. G. Smith at the University of
Florida (Smith et al. 1971).
The cooling of photographic emulsions is believed to
retard "thermal regression" of the latent image (Hoag 1964).
At low intensity levels thermal activity tends to break up
the latent image specks before they are stable enough for
development. Cooling of the emulsion reduces this effect,
presumably increasing the number of photoelectrons trapped
by silver ions, while reducing the mobility of positive holes
(Eastman Kodak 1973). Unlike baking, which is carried out
prior to exposure, the cooling process must be carried out
during the exposure interval. Cooling is most effective for
materials designed for short exposures, such as Tri-X Pan or
Ektachrome-X. Such materials have a large inherent low in-
tensity reciprocity failure. Kodak spectroscopic materials
which have been chemically Ireated during manufacture to
reduce low intensity reciprocity failure do not respond well
to cooling. Though intricate equipment is usually required
for emulsion cooling, an advantage over baking is that con-
siderable speed gains can be achieved without the gain in
background fog that baking produces.
IIypersensiti zation of Kodak 103a-0 Plates
by Nitrogen Baking
Earlier work at Rosemary Hill by Smith et al. (1971)
showed that baking Kodak type IIIa-J spectroscopic plates in
an atmosphere of dry nitrogen provided the greatest gain in
speed of any method evaluated. Because of its simplicity and
reproducibility this process has also been adopted for the
Kodak type IIa-0 and IIa-D emulsions. These three baked emul-
sions, together with untreated type 103a-0 plates, have been
used extensively in the Rosemary Hill quasar monitoring pro-
gram described in Chapters I and II.
Since the 103a-0 emulsion approaches its limiting magni-
tude with an exposure of 15 to 20 minutes at the f/4 Newtonian
focus of the 30-inch reflector, there has been little incen-
tive to increase the speed of this plate. However, photo-
graphic monitoring of bright quasars recently was inaugurated
at the f/16 Cassegrain focus to extend light curves through
periods of moonlight, and to compensate for the considerably
increased exposure times hypersensitization of the 103a-0
emulsion has now been investigated.
All plates were baked in specially designed aluminum
boxes whose lids are sealed with high temperature 0-rings.
Dry nitrogen from a commercial cylinder flowed continuously
through the boxes at a rate of 0.2 1/min. To insure dryness,
the gas passed through a silica gel column prior to entering
the oven, and inside the oven it was preheated by flowing
through a copper coil before entering the baking box.
To simulate a reasonable period of storage before use at
the telescope each plate was held for 16 hours at 5C in a
static atmosphere of the gas in which it had been baked (the
exception to this was the 1L3 batch, which is discussed later).
The plate was then transferred to a sealed cassette charged
with nitrogen (again to simulate practice at the telescope)
and exposed for 10 minutes on a Smithsonian-type tube sensi-
tometer (Latham 1969). The plates were developed 9 minutes
in MWP-2 developer (Diffley 1968) and the characteristic
curves were obtained by means of a commercial densitometer
that measures diffuse densities. Plate speed was defined as
the reciprocal of the exposure producing a density of 0.6
above the background fog. Unless baking times are excessive,
the effect is merely that of shifting the characteristic
curve of the conventional D vs log E plot to the left without
changing the slope of the linear Dortion of the curve. Thus
the contrast or gamma is unaltered except in the toe of the
curve, which is of course elevated as the fog level rises.
Since it is well known that emulsions vary significantly
from batch to batch, five different emulsion batches (desig-
nated IL1, 1E2, 1F2, 1K2, and 1L3) were tested. The initial
investigation was conducted with an oven temperature of 650C,
the standard adopted for other emulsions at Rosemary Hill.
The speed gains achieved in these runs are summarized in
Speed Gains Relative to Unbaked 1L1 for
Five Emulsion Batches Baked at 650C
in hours 1L1 1E2 1F2 1K2 1L3
0 1.00 1.10 1.01 1.07 1.00
8 1.47 2.16 2.44 1.74
16 1.55 2.33 2.59 1.94
30 2.27 3.77 2.20 3.22 2.44
50 2.31 3.66 4.00
Figure 15 shows the speed and fog curves for batches
1L1 and 1K2, which are representative. In this plot, the
speed of an unbaked 1L1 plate has been taken as unity. The
anticipated difference between batches is immediately evident,
as well as the fact that a speed gain of three can be achieved
before fog becomes excessive. In testing the 1E2 batch, the
effect of the nitrogen atmosphere was compared with that of
ordinary air by simultaneously baking a second plate in a box
Figure 15. Speed and fog curves for emulsion batches 1L1
(dashed curves) and 1K2 (solid curves), baked
in nitrogen at 650C. Speed is measured rela-
tive to the untreated 1L1 emulsion.
wa3 y -.
0- 2 C
S I I I I -0. 60
0 -- -- ----0- - - - - - - . . @ (.
O 10 20 30 40 50
BAKING TIME IN HOURS AT 65 C
whose valves were open to the oven air. The results of this
experiment are shown in Figure 16. The fog levels in air
are somewhat higher, but this does not become disastrous
until excessive baking times are reached. Similarly, the
speed curves are comparable up to baking times of about 30
hours. The plateau that occurs in both the speed and fog
curves at baking times near 15 hours appears to be a real
feature, since it is present in all three sets of data in
Figures 15 and 16; it suggests the completion of one ripening
process and the onset of another.
Inspection of the several runs indicates that a baking
time of about 30 hours at 650C represents a reasonable com-
promise between speed and fog. In each case, a speed gain
from about two to four over an untreated plate was realized.
It is interesting that the plates which were initially fast-
est tended to reach the highest final speed. This is some-
what in contrast to the behavior of the IIIa-J emulsion,
where the slower plates seem to be speeded up more so that
all tend to reach a similar ultimate speed.
Because 30 hours is an inconveniently long baking time,
a higher temperature was tried as a means of shortening the
process. In view of Miller's (1970) warning against possible
emulsion damage at temperatures in excess of 750C, 72.5C was
selected for additional tests on three emulsion batches, 1F2,
1K2, and 1L3. The results shown in Figure 17 suggest that
eight to ten hours at 72.5C is roughly the equivalent of 30
Figure 16. Speed and fog curves for batch 1E2 baked at
650C in air (dashed curves) and in dry nitro-
gen (solid curves). Speed is measured rela-
tive to the untreated 1L1 emulsion.
0 10 20 30 40 50
BAKING TIME IN HOURS AT 65 OC
Figure 17. Speed and fog curves for batches 1L3 (dashed
lines) and 1K2 (solid lines) baked in dry
nitrogen at 72.50C. Speed is measured rela-
tive to the untreated 1L1 (and 1L3) emulsions.
The "X" shows the small loss in speed of a
20-hour 1L3 plate stored for 16 hours.
BAKING TIME IN HOURS AT 72.5 C
I I II
~pe \L3 ..o
- -- ----.--3 x
I~t I- I
_ __ _
hours at 650C, and Roscmary Hill has now adopted 72.5C as
its standard temperature for baking the 103a-0 emulsion.
In testing the 1L3 batch at both temperatures, plates
carried to the maximum baking times were sensitometered
immediately after cooling, while comparison plates were held
16 hours in nitrogen at 50C before testing. As the "X" in
Figure 17 shows, there was a small drop in speed of the
stored plate. Additional storage tests have shown that there
is no further appreciable change in speed or fog for baked
plates stored for two weeks at 40C in dry nitrogen. Plates
stored for a month at -260C show a moderate increase in fog
but little loss in speed.
Sensitometric tests conducted with the 1L3 emulsion have
shown that, as anticipated, the speed gains derived from
baking are even greater for the long exposures that might be
encountered, for example, in Cassegrain photography. For a
ten-minute exposure, a plate baked eight hours at 72.50C in
nitrogen was 2.76 times faster than an unbaked plate (Figure
17). At an exposure of one hour, the gain rose to 3.07, and
at two hours it was 4.80; at such exposures the reciprocity
failure of the unbaked emulsion is of course becoming
An obvious question is whether hypersensitization impairs
the precision of photometric measurements. Fortunately, the
author was able to compare a sample of 96 unbaked plates of
the variable source BL Laccrtae with a more recent sample of
92 baked plates for the same field. All plates were reduced
with an on-line calculator that generates a least-squares
calibration curve from the magnitudes and iris readings of
the comparison stars. The rms scatter of the comparison
stars relative to this curve is taken as a measure of the
quality of each plate.
In the case of the two samples of BL Lac plates, the
unbaked plates showed an average rms of 0.053, while the
corresponding figure for the baked plates was 0.074. Appli-
cation of standard statistical tests indicates that the dif-
ference between the baked and unbaked plates is highly sig-
nificant for samples of this size; that is, the hypersensiti-
zation process does slightly degrade the photometric quality
of the plates. However, the observed small difference is not
regarded as important in routine monitoring.
The response of Kodak type 103a-0 plates to nitrogen
baking varies considerably from batch to batch, both in speed
gain and in increase in background fog. In general, the
speed can be increased by a factor of from two to three with-
out excessive fog by baking for about 30 hours at 650C or for
eight to ten hours at 72.5C. Similar results may be pro-
duced by baking in air if nitrogen is unavailable, but the
fog tends to increase more rapidly. Plates stored in nitrogen
at refrigerator or freezer temperatures have adequate shelf
life for most practical purposes. Because of batch variabil-
ity, it is probably wise to determine the optimum baking time
for each batch.
Extensive experience at the telescope confirms the sensi-
tometer tests, showing the exposures can safely be reduced by
a factor of two to three. This experience also shows that
there is only a small loss in photometric precision with the
baked plates. Baking of the 103a-0 emulsion has been adopted
as standard procedure at Rosemary Hill for exposures at both
the Newtonian and Cassegrain foci of the 30-inch reflector.
Baked plates are of course also used with the recently
installed 18-inch reflector.
Hypersensitization by Emulsion Cooling
The fact that photographic emulsions can be made more
sensitive to exposures of low intensity by cooling is not a
newly discovered effect; rather this work was pioneered in
the early 1900's by E. S. King of Harvard Observatory (Hoag
1964). King found that plates cooled to exterior winter tem-
peratures yielded half a magnitude gains in exposure, when
compared with plates warmed over a heat register in his labo-
ratory. King also discovered that cooling reduces the sensi-
tivity for intense exposures, probably due to temperature-
imposed limits on the rate of migration of silver ions to
emulsion sensitivity centers (Hoag 1964), which limits the
response of the emulsion. The discovery that moisture lowers
emulsion speed is also attributed to King, arising from his
experiments in cooling plates with snowballs.
Much of the modern work in cold emulsion photography was
initiated by IHoag (1961,1964). In his early experiments I-oag
used dry ice (-78C) and a frozen liquid mercury mixture as
refrigerants, and he experimented with evacuated plate cham-
bers and a plate chamber flushed with cold dry oxygen, in
order to prevent moisture from forming on the emulsion.
Hoag's latest camera uses a thermoelectric device for cool-
ing; the plates are enclosed in a vacuum to prevent frost
formation. In all Hoag's cameras plates are cooled by physi-
cal contact with a refrigerated metal platen. Using tempera-
tures as low as -65C, Iloag has studied the effect of cooling
on a number of emulsions. Hoag's work has shown that many
amateur films such as Panatomic-X respond well to cooling,
yielding not only increased speed but better exposure lati-
tude, enabling objects with a large brightness differential,
such as many nebulae, to be photographed with more detail in
both bright and faint areas. Many of IHoag's experiments were
with color reversal emulsions, showing that cooling greatly
increases their speed and color balance, enabling more accu-
rate color renditions to be made of faint astronomical objects.
The range of temperatures achievable with dry ice refrig-
eration is very restricted, and thermoelectric devices are
relatively expensive. In order to achieve a reasonable range
in temperature without resort ing to thermoelectric devices,
liquid nitrogen was used as the refrigerant in the author's
studies, following procedures outlined by Stong (1969). The
exposures were made in a "cold cassette" designed by the
author and A. G. Smith for a Smithsonian-type tube sensitom-
eter (Latham 1969). A schematic diagram of the cassette is
shown in Figure 18; Figure 19 is a photo of the cassette with
its lid removed.
The principle of the cassette was suggested by R. L.
Hackney (1972b); the cooling is accomplished by a flow of
cold nitrogen directly across the emulsion surface; the
nitrogen then flows across the back of the plate and exits.
A double window separates the emulsion from the air; warm dry
nitrogen flows continuously through the window to prevent
frost formation. The warm nitrogen was supplied from a com-
mercial cylinder at room temperature (20'C). The cold nitro-
gen flow was supplied by a dewar of liquid nitrogen; a 300 ohm
power resistor was used as a heating element to obtain suffi-
cient cold gas flow for adequate cooling. To provide ade-
quate insulation for the cold chamber the cassette was con-
structed of Bakelite, a material of extremely low thermal
conductivity. The cold chamber temperature was monitored by
thermocouples mounted on brass springs which pressed directly
against the emulsion when the plate was positioned for expo-
sure. Valves were mounted at the cold-gas inlet and outlet
to provide additional control over the flow of cold gas.
n d 0 4-) C Ut On H 'H7
- ~0 4- -i o -o i, 0
S -1 4- E o o 7 .4 H -0 0 o o
U c -~U o o I o
oS--, cC ,-4 o -H
o 0 4od0 4 U. co
0 0 0 re C) o 0 e 4- OI>
00r 0 O 0 0- 0 ,cy 0 i
Ou rV) C*H 4) d i n ,-I _'0 -i i
ed 0 .H C4 1 4 -"1 O -H 0 0
C' r-p O 0 1 .C) 1 a) C 0U
0o 0- o 0 40 0 H 0 0 4 0J mo
A ,- 0 ,-4 o 0- -4 4 t" M 4-
0 oU-- ,-> 0 4- 4 o a b- -
- t.n r- C U- O OtiA r- Ct 4
O= U0 U00.H Hi d
- l i r C) I Oz ( 'd 1-il l-^ C -
a. 0 -0 0 ~ O
t4- UO InO *H5
) o -) o 4-o o _!0 u -(
C)U 0 ri () 0C 0 4-- H -'l o t
0 4-) bf 4 0 l-H 0 0 1 -1 1 0 c
o i t 0 4- t 0 q-.i O I 0 m
L o o -I) : o u ) (nl U r I
cO C o C)i o L-C o r *H m r4-
-o 0 cj 0v o ui (u o rA
0 0Inu t 4-)O C004-4 04
4-4 1--0 n 7:5 E e Q *4-)O
0 O, r0 4 C 4 c0 i0--0 4- 4 M
Ud Oi- *t n- o i aj r OO 4oo
KH UP p *I U 0J C U 0 0 M0 0 .
4- -A OP L O Od r- - 2
U 4- -4 O(D -1 0 4-) C
ft 4-l I04-) C-4J 04 + n0UV 10
U' U- d U
t.0 b 70 0 U L4-ed O 4 O C .C
*d r .- 4- i r, 1 QU +j
A0 Q >, 0 0 0 E0 3 (D
0 k O C) ) 0 0 4- )n
Sf) 0 0 U- O7 t > : > et 4 + d 't
SAb do OE c,
* C- U4-) U 0 0 0 -J
U U -- F-' 0 n 4-J o 0 tU 0
C glC3 cli
0 X z
0 w u-
.1 i i ': V - :I
I: I: : *A
I I'' I KS I >.:
-A F- T FT
.; w L CO
0W 7 a
2 0 70 y I
~~LLJ~ ~ ~ L. NI L L
-4 ---0--- 0--
-j -- H ca
Figure 19. Photograph of the cold cassette. The removable
lid is shown at the top; the cassette with the
cold chamber exposed is at the bottom. The
cold-gas inlet and exit valves are mounted on
the right; the vertical pipes are the warm-gas
inlet and exit. The wires are the thermocouple
leads; the thermocouples, which can be placed
at any position in the cold chamber, are mounted
on the triangular brass springs. The window
mask has holes that exactly align with those on
the tube stack of the sensitometer. The 0-ring
seal can be seen around the perimeter of the
Before it was cooled the cold chamber was flushed with
dry nitrogen at room temperature to remove moisture that
might condense on the emulsion or inner window. The cold
chamber temperature was regulated by the cold gas flow; about
30 minutes was required for the temperature to stabilize
before starting each exposure. A rather vigorous gas flow
was required to maintain the coldest temperatures; a temper-
ature of about -68'C was the lowest that could be achieved.
A thermal gradient of about 15C was present in the cold
chamber during the exposures; the coldest end of the plate
was near the cold-gas inlet valve. As soon as the cassette
temperature stabilized heavy frost would form on the cassette
and cold-gas feed tube; at the lowest temperature traces of
frost formed on the outer window surface in spite of the
warm gas flow between the windows.
The response of four emulsions to cooling was tested:
Kodak Tri-X Pan, Kodak Contrast Process Pan, Kodak 103a-0
(1E2), and Kodak IIIa-J (IGl). The results are summarized in
Table 3. Figure 20 shows speed gains obtained by cooling
Tri-X Pan and Contrast Process Pan, both in a 4" x 5" format.
The characteristic curves and speeds for the cold emulsion
work were determined in the same manner as those for the
nitrogen baking experiments. In Figure 20 the speeds have
been normalized to exposures made at room temperature. The
Tri-X showed the largest increase in speed of any emulsion
0 0 3
4) C- )l
E I ien .o
r 0 Ul -
0 o U
0 I 'n-'
*H 6 U
4 0 4J
vr~ o oo
cj r-4 .
4J 4-J -)
0 U -
,H oH *
+3 O O05
S* o U -
*Hi ,Cn4 0 G)
*H )-1 3
ci 0 0 -
0t 'H f C) C)
F- 0 0 0
H U < 4o